Heat Transfer

INDEX
S. No Topic Page No.

Week 1
1 Introduction to Heat Transfer 1
2 Introduction to Heat Transfer 11
3 Heat Diffusion Equation 22
4 Relevant Boundary Conditions in Conduction 31
5 One Dimensional Steady State Conduction 49
Week 2
6 Temperature Distribution in Radial Systems 60
7 Tutorial Problem on Critical Insulation Thickness 73
8 Heat Source Systems 83
9 Tutorial Problems of Heat Generating Systems 97
10 Transient Conduction 109
Week 3
11 Lumped Capacitance (Contd.) and Tutorial Problem 123
12 Transient heat Conduction 133
13 Transient Conduction - Heisler Chart 145
14 Heat Transfer from Extended Surface 156
15 Fins and General Conduction Analysis 169
Week 4
16 Fundamentals of Convection 182
17 Equations of Change for Non-isothermal Systems 197
18 Equations of Change for Non-isothermal Systems - continued 207
19 Tutorial on the Application of Energy Equation 224
20 Nusselt Number of a heated sphere in Stagnant Air 233
Week 5
21 Momentum and Thermal Boundary Layers 246
The Flat Plate in Parallel Flow - Hydrodynamics and Momentum
22 Transfer 252
23 The Flat Plate in Parallel Flow - Heat Transfer 265
24 The Effects of Turbulence 275
25 Turbulent External Flow 288
Week 6
26 Heat and Momentum Transfer Analogy 298
27 Mixed Boundary Layers 314
Tutorial Problem on External Flow and Behavior of Heat Transfer
28 Coefficient 322
29 Tutorial Problem in External Flow and Convection 333
Week 7
32 Internal Flow Heat Transfer 361
33 Internal Flow Heat Transfer (contd. 1) 375
34 Internal Flow Heat Transfer (Contd. 2) 383
35 Internal Flow and Heat Transfer (Contd. 3) 399
Week 8
36 Internal Flow and Heat Transfer (Tutorial) 408
37 Free Convection 419
38 Heat Exchangers 431
Week 9
41 Tutorial Problems on Heat Exchanger Calculations 467
42 Tutorial Problem on LMTD and Dirt Factor 482
43 Epsilon-NTU Method - 1 493
44 Epsilon-NTU Method - 1 (Continued) 504
45 Tutorial Problems on Epsilon - NTU Methods 520
Week 10
46 Tutorial Problems on Epsilon - NTU Methods 531
47 Boiling, Evaporation and Evaporators 543
48 Radiation - Fundamental Concepts 557
49 Spectral Blackbody Radiation Intesity and Emissive Power 568
Wein's Law, Stephen Boltzmann Law, Blackbody Radiation Function,
50 Tutorial Problem 579
Week 11
51 Kirchhoff's Law 593
Tutorial on Emissivity, Absroptivity and Blackbody Radiation
52 Functions 606
53 Solar Radiation and the Concept of View Factors 621
54 Determination of View Factors 633
55 Radiosity Blackbody Radiation Exchanges, Relevant Problem 643
Week 12
56 Network Method for Radiation Exchange in an Enclosure 654
57 Network Method - Two and Three Zone Enclosures 666
58 Tutorial Problem on Radiation Exhange using the Network Method 679
59 Radiation Shields 691
60 Gaseous Radiation (Participating Medium) 702
Heat Transfer
Prof. Sunando Dasgupta
Department of Chemical Engineering
Indian Institute of Technology, Kharagpur
Lecture - 01
Introduction to Heat Transfer
This would be the beginning of a new course - Heat Transfer, but first of all little
introduction about myself. My name is Sunando Dasgupta. I am a faculty of the
Chemical Engineering Department of IIT Kharagpur and together we are going to learn
Heat Transfer.
So, as you are well aware of the abundance of examples of heat transfer around us, it still
required, that we need to know what the governing principles are, the rules and the
correlations, which one should use in order to understand the concept of heat transfer.
And more importantly, apply heat transfer in a multitude of applications that we see
around us every day. So, from an automobile to a petroleum refinery, to any anything
that you can think or see around you, will involve examples of heat transfer.
Some of the concepts of fluid mechanics will also be used in heat transfer. But firstly, we
must understand; what is the distinguishing feature of heat transfer in comparison to
thermodynamics. Because thermodynamics also deals with temperature and we know
that heat transfer and thermodynamics are closely related, but they are not the same.
The first difference lies in that, heat transfer is energy in transit due to the presence of a
temperature gradient. So, thermodynamics tells us about the ends states of a process (i.e.
the initial and final states). However, heat transfer tells us more about the kinetics of
transfer of heat from one object to the other whenever there is a difference in temperature
between the two.
So, one can see heat transfer around us, but in most importantly, in many of the industrial
processes you will encounter that heat has to be transferred from one fluid to the other,
heat must be dissipated from a hot surface to the environment. And if you just look at
your laptop, the heat that the chip (processor) generates has to dissipated otherwise, you
would not be able to use your laptop. The safe operating temperature in an electronic
device is roughly about 80 oC. So, you would like to keep your laptop at a temperature
1
which is significantly lower than 80 oC. There are several ways by which people are
addressing this problem, which is also known as electronic packaging.
So, from your laptop to your cellphone and to your television set everywhere the control
of temperature within a certain level is important. If you look at the inside of an
automobile, if you just open the open the hood of a car, you would see that there are thin
fins (which are attached to the walls of the radiator, these fins are also known as
extended surfaces) which are attached to the radiator where the coolant is stored. The
coolant is pumped to keep the engine as a safe operating temperature; and when that
coolant gets heated it comes back to the radiator, where it must radiate heat to the
environment, in order to again reduce the temperature of the coolant such that it can be
pumped back to the engine. So, you would see the radiative heat transfer, the convective
heat transfer and so on.
If you ever visit any chemical plant you would see that there are several utilities, which
are present in a chemical plant. For example, you may have hot fluid, you may have a
reactor in which the reactant has to enter at an elevated temperature or you are getting a
product out which is coming out at a very high temperature. So, you do not want to waste
that heat. So, you need to design exchangers. Exchangers or heat exchangers are devices
in which a hot and a cold fluid exchange heat between one another without physically
coming in contact. You may have a tube (or a significantly large number of tubes)
through which a cold fluid is flowing, and all these tubes are encased in a shell-like
structure. So, through the shell the hot fluid will flow and through the tubes, which are
encased in the shell the cold fluid will flow. And as they flow they are going to exchange
heat among each other and by this way you can use steam to heat the reactant, which is
entering the distillation column, which is entering the reactor and so on or you can use
the product streams which are at a high temperature to heat the streams which are
coming into your process plant.
It is to be noted that heat transfer is not only important from a technological point of
view it is also extremely important, when you decide about the economics of the
processing plant. So, how do we minimize the waste of energy? How do we transfer
energy efficiently from one point to the other is essentially the subject of heat transfer.
2
There are several fundamental physics which are involved in describing heat transfer, but
at the same time the applicative examples are many as well. So, it is a course which is
designed to give you a flavor of the physics and then will move towards applications.
And, how the concept developed in the initial part of the course can be utilized to obtain,
generalized, governing equations, design equations, correlations, and how that can be
applied for the actual design of a heat exchange equipment.
The textbook that we should we are going to follow in this course are as follows:
(Refer Slide Time: 09:22)
1) Incropera and Dewitt and the name of this book is heat transfer heat and mass
transfer, by Incropera and Dewitt.
2) The second book that we would occasionally refer to is transport phenomena by

Bird, Stewart, and Lightfoot.
So, if you see the description of the course you would see both these books are listed as
prescribed textbook for the course.
As discussed previously, heat transfer is energy in transit due to a temperature difference.

Wherever and whenever there is a temperature difference you are going to have heat
transfer, but how heat is going to get transferred from one object to the other would
depend on a lot of factors. So, the modes of heat transfer that one needs to understand is
the mode of heat transfer prevalent in a specific situation.
3
So, we all understand that there can be three modes of heat transfer, the first is
conduction the second is convection and the third which is omnipresent is radiation. So,
conduction is a process when two adjacent molecules transfer energy (from the hotter to
the cooler molecule) due to their vibration and their interactions.
So, conduction requires the presence of a medium. And the heat gets transferred through
a solid or a liquid or a gas by means of these molecular vibrations. So, heat gets
transferred from the hot side to the cold side by means of conduction. However, the
molecules do not move physically, and are static at their place and they only fluctuate
and oscillate, thereby causing the transfer of energy from one point to the other without
leaving their average position.
On the other hand, there are situations in which the molecules around a hot object or a
cold object would start to move. So, you may have a hotplate which is kept in
atmosphere, where there is a slight breeze blowing on the plate. The molecules very
close to the solid surface will experience the heat and will attain the temperature of the
solid plate, but at the same time they will be swept away from the hot surface to be
replaced by other molecules which are coming from behind. So, there is an imposed or a
naturally occurring motion, which is the hall mark of the convection process. When I
said imposed and naturally occurring I am referring to two different types of convection
processes, which I will come in a minute. So, therefore, the convection is characterized
by the actual movement of the medium around the hot or the cold object, and in this way,
it is different from conduction.
Consider a hot plate which is just kept hanging in a room without any fans, therefore, it
is just a closed room where the air is stagnant. So, what is going to happen is the air
which is very close to the hot surface will get heated up, and its density will change. In
fact, its density will reduce, and it is starts to rise due to buoyancy along the plate and
then move away from the plate. So, the air current caused by the temperature difference
between the solid wall and the surrounding fluid creates a flow, which is not imposed by
an external agency; and these types of convections are generally known as natural
convection or free convection.
In many other situations mostly in engineering applications, you make the cold or the hot
fluid move, by the application of a pump or a blower. E.g. for the case of your laptop, a
4
fan is attached close to the hot where the temperature is a maximum. So, you are forcing
the air to move over the hot surface there by taking the heat out by convection; and this
is known as forced convection. Therefore, we have free or natural convection which is
buoyancy driven and forced convection, which requires an external agency for pumping
the fluid over the hot surface.
The total amount of heat transfer by convection depends on all the operational
parameters such as the velocity of the flowing stream, the thermo physical properties, the
geometry of the system etc. So, we must find out the appropriate relation for a specified
geometry, type of fluid and the velocity of the fluid. So, that is what we are going to do
in convective heat transfer the third one is unique because it does not require the
presence of a medium and we are referring to radiation.
So, any object which is at a temperature higher than the absolute zero will emit some
radiation and it is going to exchange heat with it is surrounding. So, it does not require
the presence of any medium. And the amount of radiant energy emitted by a body would
principally depend upon the temperature of the object. Theoretically, the maximum
emission, that you can obtain from an object at a given temperature is only possible for a
specific type of an object known as a black body.
So, black bodies can deliver the maximum amount of emissive energy compared to any
other object. The emissivity tells you how close the real surface is to a black body. So,
the emissivity has a value from 0 to 1 if the emissivity is 1 then; that means, the surface
is behaving like a black body. So, it is emitting the maximum amount of energy possible
at that given temperature. Secondly, any energy incident on an object is going to be
absorbed. Now, the fraction of the energy which gets absorbed in such a case is denoted
by another property, which is known as the absorptivity. Again, the black body has an
absorptivity equal to 1; that means any energy incident on the black body is would be
absorbed by it. There are some surfaces, which are highly polished (highly reflective)
and they do not absorb much of the incident energy. So, for those cases the absorptivity
will be close to 0.
5
We are now first going to start with conduction. So, when we start with conduction the
first law which I think many of you are aware of this is known as Fourier’s law. This
Fourier’s law connects the q '' , which is heat flux with the temperature gradient. Heat
flux is energy transferred per unit area per unit time. In other words, it is a rate of energy
which gets transported from 1 point to the other per unit area. So, if we have an object,
and one side of it is maintained at temperature T1 and the other side is maintained at T2 .
And let us assume that T1 is greater than T2 then of course, heat is going to move from
location 1 to location 2.
And the heat flux ( q '' ) is the amount of energy transferred from 1 to 2 per unit time and
per unit area, when I talk about per unit area it is the area which is perpendicular to the
flow direction.
Now, what Fourier observed is that the transfer of heat by conduction between 2 points it
is governed by a unique relationship and it is not a law which can be derived from first
principles. So, in that way it is similar to Ohm’s law. So, when you think of Ohm’s law it
states that the flow of the current is proportional to the potential gradient. So, the current
is proportional to the potential and the proportionality constant is known as the
resistance. So, more the resistance less would be the flow or the less the resistance more
would be the flow. So, Ohm’s law is obtained not from any first principles, but Ohm has
looked at a large number of experimental data of potential drop verses current and then
6
he understood that there is a linear relationship that exists between the potential drop and
the current.
Similarly, when you look at the amount of heat transfer, which is taking place due to a
temperature gradient, the heat flux is proportional to the temperature gradient. So, in that
sense it is this law is derived but this law is proposed by looking at a number of large
number of phenomena, which are related to conductive heat transfer.
So, in this way this Fourier’s law is a phenomenological law, which tries to express the
phenomenon that you observe over many systems. Interestingly, what is observed that
dT
this q '' , the heat flux is proportional to .
dx
dT
Therefore, according to Fourier’s law q ''   k . Where, k is the proportionality
dx
1
constant, and it is conceptually similar to resistance and it is essentially . So,
resistance
heat flows because of a temperature gradient and the proportionality constant is k. And,
you can see that with higher value of k more heat will get transferred by conduction for
the same temperature gradient. So, if you keep temperature gradient fixed and you
choose materials with higher and higher values of k then the heat transfer is going to be
more.
So, to give you an example let us think of this as a wooden block. And secondly, let us
think that this is a block of copper which has a higher thermal conductivity. So, the
materials can be characterized based on whether it is a good conductor of heat or a bad
conductor of heat. The same way it has been defined for the case of Ohm’s law whether
some material is a good conductor of electricity or a poor conductor of electricity also
known as insulators.
So, in heat transfer also based on the value of k you can classify all the materials which
you see every day. So, on one hand we have the very high thermal conductivity materials
for example, copper, aluminum, and so on. And on the other hand, you have materials
for which the thermal conductivity is very low, and they are known as the insulators.
7
So, the cause remains temperature gradient remember, it is the temperature gradient not
the temperature difference. That is, it is the temperature difference divided by the
dT
distance over which these temperatures are measured. So, it is not just T . So,
dx
conduction depends on the temperature gradient, it is proportional to temperature
gradient, and the proportionality constant is a property of the material, it is a thermo
physical property of the material, which is known as a thermal conductivity. And based
on thermal conductivity, you can differentiate between different materials. And you can
put them into different categories based on whether it is a very good conductor of heat or
a very poor conductor of heat and something in between.
dT
If you would look at carefully the equation or the law that I have written q ''   k , the
dx
minus sign in that law simply tells us that heat gets transferred in the direction of
decreasing temperature.
So, heat always flows from high temperature to low temperature. So, thermodynamics
tells us that energy has to move from high to low. Therefore, to denote the specific
directional nature of heat flow, the minus sign is provided in Fourier’s law of
conduction. Again, if you look at once again this q double prime is when you have
temperature difference in the x direction that is the temperature gradient in the x
direction that is the amount of heat which will travel in the x direction.
So, therefore, to denote the directional nature of heat flow sometimes the subscript x is
added to q double prime to denote that it is the heat which flows in x direction. If there
exists a temperature gradient in the y-direction then the same Fourier’s law can be
dT
written as q y ''  k .
dy
But, since now I understand that temperature is a function of x and y, so, I change the
ordinary differentials to partial differentials. Therefore, the equations for heat flux in the
three directions for a Cartesian coordinate system is given as:
T
qx ''  k
x
8
T
q y ''  k
y
T
qz ''  k
z
So, these 3 heat fluxes are nothing but the x, y and z components of the heat flux vector.
So, q '' the heat flux is the vector quantity. And qx '' , q y '' , and qz '' are the x, y, and z
components of that vector.
So, what we further understand is that heat flux can be expressed in terms of Fourier’s
law, it is proportional to temperature gradient the proportionality constant is a material
property denoted by k the thermal conductivity. The minus sign denotes that heat gets
transferred from high temperature to low temperature. And, we also understand that heat
flux has to be a vector with qx '' , q y '' , and qz '' being the x, y, and z components of the
heat flux vector.
In summary we have discussed - what is heat transfer, that is energy in transit, what is
the difference between thermodynamics and heat transfer, the different modes of heat
transfer- in terms of conduction convection and radiation and how radiation is distinctly
different from conduction and convection in as it does not require the presence of any
medium. Radiation also provides us some idea about the object which we are dealing
with, is it close to a black body such that it can emit maximum amount of energy
possible at a specific temperature.
So, what is its emissivity, if it is equal to 1 it will be a black body, but most of the most
of the real surfaces will have an emissivity less than 1, because black body is something
like an ideal surface. Similarly, when you have some energy incident on a surface it may
not absorb 100 percent of it which a black body does. So, absorptivity will also play a
role in defining the nature of the surface.
So, conduction, convection and then radiation will constitute the modes of heat transfer
which we generally encounter. And finally, we talked about the Fourier’s law which is
the fundamental governing principle of heat transfer from one point to the other where
the heat flow is proportional to the temperature gradient.
9
So, we will move in to the concepts of the control surface control volume energy balance
in the next class.
10
Heat Transfer
Lecture - 02
Introduction to Heat Transfer (Contd.)
So, by now we are familiar with the modes of Heat Transfer and the fundamental relation
of conduction, which is Fourier’s law, which gives you the heat flux vector as consisting
of components in 3 directions and the components in each direction can be expressed as
proportional to the temperature gradient in that direction.
In other words qx, which is the x component of the heat flux, vector q is going to be
proportional to δT/δx, which is the derivative and is the change in temperature in the x
direction and the proportionality constant being a property of the material which is
known as thermal conductivity. What we are going to do next is fundamental relation,
which is known to all of us which is conservation of energy. So, what does conservation
of energy tell us and how the conservation of energy can be related to heat transfer, in
order with an aim to obtain that temperature at a specific point at a given instant of time,
because in heat transfer that is what is required! You need to find out what is the
temperature of an object at a specific point and at a specific instant of time.
Since, we are studying conduction right now we understand that conduction is essentially
energy, which depends on temperature gradient. So, we need to precisely know what is
the temperature profile inside an object, in order to find what is its gradient, which when
multiplied with k would give us the flow of heat through a plane, which is perpendicular
to the point where the temperatures are uniform.
So, if we have 2 points at 2 different temperatures this point will call it as an isothermal
surface where everywhere the temperature is equal to T 1 and this is the T 2 another
surface and the amount of transfer between these two would simply be governed by
Fourier’s law, but the concept of control volume is going to be important.
Because not only you are going to have transport of heat between point 1 and point 2
there will be situations in which you are going to have heat generation between point 1
and point 2 as well. So, let us think of a current flowing through an electric wire. So, we
11
are going to have some sort of joule heating which is present in between two points that
you need to take into account in order to find what is the temperature of the object at any
given point?
So, when you are generating heat and when you are having some amount of heat flowing
in and some amount of heat flowing out. And as a result of all this the total energy
content of the volume that you see will keep on changing. In other words let us think of
just a block of copper, which you have placed on a hot surface and some amount of
energy enters the block of copper, because of its proximity with a hot surface. And as it
enters let us assume that no heat leaves from the copper surface to the surrounding. Let
us say to begin with the temperature of the copper surface is the same as that of the
surrounding, so, no heat transfer takes place. If it is different as the copper surface
becomes more and more hot, it is going to lose energy from all the other surfaces to the
surrounding.
So, you have some energy in by conduction, some energy which is going out of the
copper block to the atmosphere again by conduction. And let us say by some means you
are generating some amount of heat inside the inside the copper block and it is an
unsteady state process.
So, if I think of an instant of time the amount of energy which comes in to the copper
block, the amount of energy which goes out of the copper block, the amount of energy
which you are generating artificially inside the copper block, all these would result in the
total amount of energy which is stored in the copper block.
So, we are going to talk that is what conservation of energy is all about, the amount in
minus the amount out plus any heat generation, that is taking place must be equal to the
energy to the time rate of change of energy contained within the copper block. So, this
copper block is therefore, is termed as the control volume. So, control volume is what in
something it is an enclosed space, which has a specific mass and the form of
conservation law, which is applicable for such a control volume. I am going to write that
which is in minus out plus or minus, plus is for where energy is generated inside the
control volume.
12
And minus when there is some depletion of energy, which is getting depleted inside the
control volume and will see the examples of that in this. So, this must be equal to stored,
the time rate of change of energy stored in all these in and the out. So, this is
dE
E in  E out  E g  E st  st
dt
So, this is the conservation equation and the enclosed space with the fixed mass is known
as the control volume. So, this is a control volume and you have some amount of energy,
which is coming in and this is the time rate of energy, which is going in out and there
may be some amount of energy, which is generated in here as a result of which you are
going to have the rate of energy stored inside the control volume will also be different.
So, the conservation equation can be written for a control volume that has a fixed mass.
So, this is what control volume is and this is the form of the conservation equation that
you would expect you would expect for such a situation, then what is a control surface?
Control surface is again an imaginary surface, which together form the boundary of the
control volume. So, if you think of a copper block a rectangular copper block, then the 6
surfaces these and the 2 up and the top and the bottom, the 6 surfaces of the copper block
those are known as the control surfaces, which defines the control volume, which defines
what is the volume, what is the mass of copper, which is contained in the space. So, they
are known as control surfaces. By definition control surfaces do not have any mass of
13
their own whereas, control volumes do. So, the major distinction between control volume
and control surface is control surfaces do not have any mass. So, that is the difference
between control surface and control mass. So, the conservation equation, which we did
right for the control volume has to be modified for control surfaces as they do not have
any mass of their own. If they do not have any mass of their own, then they cannot store
any energy and if there is no mass of the control surface then they also cannot generate
any energy.
So, for a control surface in the conservation equation, you need to drop the E g term, the
time rate of change of energy generated in the system, as well as the E st that is the time
rate of change of energy stored in the control volume. So, for a control surface the
conservation equation would simply turn out to be E in = E out . So, this is what you are
going to get for a control surface. So, this is for a control surface and what you have here
is for a control volume. So, that is the difference between a control volume and the
control surface formulation as well as the conservation of energy is concerned.
So, this is primarily as a control surface (CS) as a CS does not have any mass of their
own and they simply define a control volume. And these are the forms of the
conservation equation. Now, if you look at this equation the conservation equation for a
control volume there is what is going to be.It is what is going to how what it would look
like if I apply it for a steady state process. What if there is no generation which is present
in such a system? It is a case of the block again where 2 surfaces are maintained at 2
different temperatures, ok.
However, there is no generation of heat in there and you allow the process, allow the
process sufficient time such that a steady state has reached. So, the moment a steady state
is reached the amount of energy stored inside the control volume does not change, cannot
change, ok. And, if there is no heat which is generated inside the control volume then
that term will also disappear. So, what you have for a system at steady state, so, for if
you have a system at steady state with no heat generation, this equation would simply be
equal to E in = E out .
So, when you when you when you look at these 2 equations in absence of any heat
generation and at steady state, this becomes equal and becomes identical with the control
14
surface equation. However, fundamentally conceptually you must appreciate what is the
difference between a control surface and a control volume. The next what I am going to
write is surface energy balance.
So, let us say that let us assume that I have a block this is my x direction and the over
here I have the temperature this is the temperature axis and this edge I am taking as my
control surface. So, the temperature is T1 here and the temperature is T2 at this point and
over here it is exposed to an ambient air ambiance with the temperature equal to T ∞ at
this point.
So, between this is a solid and here you have as air. So, in the solid the principle mode of
heat transfer in, since there cannot be any motion of the molecules. So, in a solid the heat
gets transferred from 1 point to the other by means of conduction only, as the molecules
cannot on an average move from their positions. So, in absence of any heat generation in
a system and we will see subsequently why in absence of heat generation and at steady
state the temperature distribution is going to be linear.
So, the temperature distribution in such a case will be linear, and therefore this is going
to be the difference and this is going to be the profile where this T is a function of x, but
it is a linear function of x. Now let us say that , this is my control surface. So, the red line
simply denotes the extend denotes that this is my this is this is my control surface ok.
15
So, you are going to have q” conduction which is moving towards the control surface
and over here from the control surface the heat is going towards air by q” convection. So,
you have heat gets transported in the solid by conduction, heat gets transported in the air
by convection and this is a reasonable approximation since we realize that thermal
conductivity of the solid is much more than thermal conductivity of air.
So, the principle mode of heat transfer on the air side of this solid interface is nothing,
but convection. So, you can also assume that the air is moving with certain velocity
along the exposed sides of the block. Not only convection you may also have, if the
temperature of the solid is large radiation which is going on.
So, this surface the control surface can lose heat as a result of convection as well as
conduction as convection as well as radiation. So, the heat that travels to the control
surface is; obviously, only by conduction and the heat which is lost from the surface is
convection and radiation. And, since we know that the control surface has no mass of its
own the conservation equation would simply be E in = E out .
And when you take this as the surface then q” conduction is simply going to be equal to
q double prime convection plus q double prime radiation. So, most of the cases the
radiation becomes important only when the temperature is quite large, in many realistic
situations the radiation does not play a significant role as compared to conduction and
convection.
So, when that happens we can also drop this radiation in here since it may not be that
significant. So, we have conduction equality of conduction and convection across a
control surface, 1 side of which faces the solid the other side is a flowing fluid of low
thermal conductivity because this is a prerequisite low thermal conductivity and
therefore, convection is the prevalent mode of heat transfer and your qconduction is equal to
qconvection.
But, how does the temperature profile go from here to T ∞. And therein lies something
which I will have to explain to you later. You will see more in greater detail that there
exists something, which is called as a boundary layer or a thermal boundary layer. What
is proposed is that all the temperature difference from T 2 to T∞, this drop in temperature
16
from T2 to T∞ takes place over a region, which is very close to the solid surface and then
it asymptotically approaches the value of the temperature T∞.
So, if I enlarge this section it would look something like this. So, this is my control
surface this is T2 and the temperature profile changes drastically now this is my T ∞. So,
temperature profile changes drastically over a region, which is very thin and this region
where the temperature changes from that of the base temperature to the temperature of
the free stream to the temperature of the free flowing air which is T ∞, this is known as the
thermal boundary layer.
So, in a thermal boundary layer as I have shown you the temperature drops suddenly the
temperature changes suddenly till it reaches the temperature of the free stream. We will
discuss more about why and what is the significance of thermal boundary layer, is there
any experimental proof that there is something called a thermal boundary layer.
So, we would discuss that in subsequently, but what you have to accept now for the time
being is that for any convection process, the temperature change is sharply near the solid
fluid interface. And it reduces or increases depending on whatever be the case sharply
enriches the value of the free stream temperature. And the thickness over which this
change over takes place is commonly known as the thermal boundary layer. I think in
your fluid mechanics you have you have already studied what is hydrodynamic boundary
layer thickness.
So, if you have a flat plate and a liquid is flowing over it then due to no slip condition the
velocity of the fluid on the solid liquid interface on the solid is going to be zero, that is
the no slip condition, but if you go above slightly above the plate the velocity here is
equal to the free stream velocity, which flows unperturbed over the solid surface if the
distance from the solid surface is beyond certain value.
So, the region over which this change in velocity from 0 to the free stream velocity takes
place is known as the hydrodynamic boundary layer. And you also probably know that
the viscous forces are important inside the thin hydro dynamic boundary layer, outside of
this boundary layer, the flow can be termed treated as inviscid or zero viscosity. So, you
have also probably have heard about 2 equations one is Navier Stokes equation and the
second is Euler’s equation.
17
So, Navier Stokes equation takes into account the viscosity the viscous forces, which are
present in a flowing fluid, whereas Euler’s equation assumes that the flow is inviscid and
the famous Bernoulli’s equation can be derived from Euler’s equation and in its true
form the Bernoulli’s equation can be used only for inviscid fluids. So, inside the the
momentum boundary layer, inside the hydrodynamic boundary layer, viscous forces are
present, outside of the boundary layer the viscous forces are unimportant, are not present.
So, this is that is the concept of hydrodynamic boundary layer. So, which I think you
know. So, based on same concepts the concept of thermal boundary layer is also
prescribed, where the temperature is going to be the same as that of the solid, when we
think about the fluid very close to the solid surface, but as we move along the
temperature changes sharply. And over a small thickness it varies from TS the
temperature of the solid to T∞, which is the temperature of the fluid, which is flowing
past the solid surface.
So, that is the thermal boundary layer, the important point here is what I need to show
you is what would the profile the temperature profile look like for such a case. So, this is
my solid and this is the temperature of the air and this side is maintained at a high
temperature TS1 and at a temperature T S2 over here as there is no heat generation an it is a
steady state case, it is going to be just a linear change from T S1 to TS2 inside the solid
where only conduction is present, outside of which the air or the air is flowing with some
velocity and a temperature constant temperature of T∞ which is lower than TS2.
So, the temperature change will be very sharp near the solid wall. So, this is what is
called as the thermal boundary layer and the complete profile would look something like
this. So, what you have here then is linear profile sharp change asymptotically
approaching the value of T∞. And, if you consider this point the flow due to conduction
must be equal to the flow due to convection, of course, assuming that there is there is no
radiation, radiation is not important in such a case. So, what is the methodology that one
has to follow and what is the conservation methodology analysis of application of
conservation laws?
18
So, how do we apply the conservation laws? The first step is you define an appropriate
control volume, ok. Appropriate control volume with control surfaces. Then the second
is identify the energy processes and write the conservation equation.
So, these are the 3 steps which one should follow in order to use in order to apply the
conservation laws. So, what you do then is first you identify the control volume and the
control volume is defined in terms of control surfaces, you identify the energy processes
is it a conduction, is it convection, is it both conduction and convection, whether
radiation is important or not you make those judgements and then you write the
conservation equations.
And we so far, we know only about what to write for conduction. We are not sure what
to write for convection or for radiation, since we have not discussed that, but I will show
you what what would be the form of the rate equations for that. So, identify whether it is
conduction, convection or radiation and express the heat flow in using certain laws for
example, Fourier’s law for the case of heat conduction.
And, then what you are going to get is a governing equation for the process and I will
show how this governing equation for any system can be derived, can be developed. One
more point before we conclude is that the conservation equations is valid for at
differential control volumes, as well as for an integral control volume.
19
So, we can assume that my control volume, its size is dx dy and dz. And then if we write
it appropriately what you are going to get is a differential equation. And, if you can solve
that differential equation based on an energy conservation, if you can solve that
differential equation you get temperature as a function of space coordinates, that is x y
and z and if it is an unsteady state process also as a function of time.
So, your temperature T as a function of x y z and time t is going to be the result of the
differential equation, which you have obtained for a differential control volume and the
approach by which the temperature at every point and at every instant, if it is, if it is
possible to obtain that the approach that gives you that is known as differential approach.
There is there is another approach, which does not care about how temperature varies at
every point in the in the control volume or at every instant. Though, in that approach we
will be satisfied with how does the system behave in an overall sense to obtain what
would be the overall behavior of this system as it gains energy, loses energy, and when it
comes towards a steady state.
So, the conservation equations can also be applied to an integral volume, which unlike
the differential approach has a finite length, height, and width. So, you have L 1 length,
width and height as well as well as it is a function of time. So, the approach which will
give you an overall approach, which is will give you an overall value of the temperature
at a plane not at every point is known as the integral approach. For times you are more
interested in how does the system behave generally rather than what is going to happen
to each point in the system.
So, on one hand you have a detailed approach, which is a differential approach that we
would see also you have an approach which can also be termed as an engineering
approach, which is a control volume approach, where you are not interested in obtaining
the specific value and average value would do, ok.
So, you do not want to know how temperature at every x varies, you just want to know
what is the temperature of this plane the average temperature of this plane. So, the
approach that gives you that is known as the control volume approach, the control
volume approach, integral approach and you also have the differential approach, and we
would see what those are so, but the bottom line is using your conservation equations
somehow you have to write the governing equation for the system.
20
Once you have the governing equation you should be able to solve it utilizing some of
the boundary conditions, because you know when you are integrating when you are
solving a differential equation you end up with integration constants. These integration
constants will have to be evaluated utilizing the physical description of the system. So,
what is going to be those physical descriptions of the systems that will act as boundary
conditions?
So, something has to be known for to convert the physical principles into boundary
conditions. One example could be, and I will give you all the detailed examples in the
next class, one example could be that the temperature at a specific point in your control
volume is known to you.
So, let us talk about that you have this pen whose you would like to find out the heat
transfer taking place in this pen, but let us say that this edge of the pen is firmly in
contact with a solid surface which is maintained at a constant temperature of a 100º. So,
what you can say then is at when x equals 0 the temperature is known let us say 100º,
temperature is known. So, at a spatial coordinate that the specific value of temperature
may be known to you and that can act as boundary can act as a boundary condition.
So, expressing the description of the problem, expressing the understanding of your
problem in physical terms would give you boundary conditions, which you are going to
need to solve the governing equation. So, what would we do in the next class is first
identify the common boundary conditions that one would expect and then what is going
to be the form of the boundary condition considering only conduction because in their
initial part we will restrict ourselves to conduction.
So, we will write what is the differential equation? What is the governing equation the
differential governing equation for conduction in a solid and what are the possible
boundary conditions? So, with the help of the equation and the boundary conditions we
should be able to obtain temperature profiles in different geometries under different
conditions and they that would give rise to some very interesting and useful results
which we will take up in the next class.
21
Heat Transfer
Lecture - 03
Heat Diffusion Equation
The first part of this course will be devoted towards understanding conduction, deriving
the relevant equations, and how those equations can be applied for specific practical
applications.
One of the fundamental relations, which is quite common in conductive heat transfer, is
Fourier’s law of conduction, which defines the material property - the thermal
conductivity. We have also seen that the heat flux, which is a vector, is proportional to
the temperature gradient and the proportionality constant defines the thermo physical
property “k” – the thermal conductivity. It is to be noted that the use of this expression
(Fourier’s law) for a control volume will give rise to an equation, where the temperature
can be expressed as a function of spatial coordinates; (e.g. the temperature at a specific x,
y, z in a fixed coordinate system as well as the time rate of variation of temperature at a
specific location).
The heat diffusion equation would allow us to find the temperature of the system as the
function of time and space. To derive the equation, the first thing one must do is define a
control volume. As discussed previously, the control volume has a fixed mass and it is
defined by control surfaces. The control surfaces do not have any mass of their own. So,
the conservation equation, which is applicable for a control surface would simply be
equal Quantity in – Quantity out = 0.
The conservation equation for a control volume would be rate of heat in minus rate of
heat out of the control volume, plus the rate of heat generation inside the control volume
is equal to the time rate of change of energy stored inside the control volume, i.e. the
algebraic sum of these three (heat in, heat out, and heat generated) must be equal to the
rate at which energy is stored inside the control volume. The generation of energy can be
by nuclear means (e.g.: a nuclear reaction is taking place in the control volume) or it can
also be due joule heating.
22
Once we have this equation we would see what the simplifications are, that can be made
to this equation and under those simplifications and with the help of appropriate
boundary conditions, we will identify the special form of the equation for temperature
distribution inside the control volume. Because the main aim of any conduction analysis
is to obtain the temperature profile of the system. Once we have the temperature profile;
for example, let us say T as a function of x, then we can find out by simply
dT dT
differentiating the profile what is at a specific location? If you know what is ,
dx dx
multiplying that with minus k; k being the thermal conductivity we would be able to
evaluate what is the heat flux at that location. So, it is of paramount importance to know;
what is the temperature distribution inside a solid at any given condition.
We will assume the control volume to be a cuboid defined by the length scale del x, del
y, and del z and we would see how the heat comes in to the system, goes out of the
system, what is the rate of heat generation? And the sum must be equal to the time rate of
change of energy storage inside the control volume. (The image below represents the
control volume and the subsequent energy-balance relations for the same)
This is my cuboid, which has a length of dx, dy, and dz. So therefore, this is the x
direction, this is the z direction, and this is the y direction. So, this is at this location is
xyz and the top obviously, is going to be x plus dx, y plus dy, and z plus dz.
23
We assume that qx is the amount of heat, which enters the control volume and it leaves
from this face as qxdx . qx denotes the amount of conductive heat flow to this control
volume through the surface at x.
So, the surface at x has an area equal to y  z . Similarly, the y face which is
perpendicular to the y direction has an area equal to x  z . And the z face which is of
area under at the bottom will be y  x . So, by x face, or y face, or z face we mean that
it is the area, which is perpendicular to the direction mentioned. So, the area x face is
perpendicular to the x direction and the amount of heat flow to the system would be
denoted by qx .
We are now going to write the conservation of energy for this control volume. In order to
do that, the first thing that we would do is we have to express qxdx , that is the amount
of energy which leaves the x face, as the amount of energy which comes in to the x face
plus the change of conductive heat flow in the x direction multiplied by dx.
So, this is a fundamental expression which is obtained by the Taylor series expansion of
qx and neglecting all higher order terms. Now, if we see what is the, if we understand
that dx is quite small and by writing the Taylor series expansion of qx around x and
qx
neglecting the higher order terms, we get: qx  dx  qx   dx
x
It is to be noted that this analysis is only valid if dx is small which is what we have
assumed at the very beginning while we defined our control volume.
Similarly, the equation fo the y and z-directions can be represented as:
q y
q y  dy  q y   dy
y
qz
qz  dz  qz   dz
z
24

Secondly, let us assume that q is the rate of energy generated inside the control volume.

Therefore, the total rate of energy generation per unit volume must be equal to q times
dx  dy  dz for the volume.
The last thing which remains is energy storage which would simply be equal to
 T
E   C p dx  dy  dz  (where, the dot denotes the time rate). This means the
t
change in energy inside the control volume because of conduction would simply be equal
to mC p T .
So, the final conservation equation can be represented as:
   
E in  E out  E gen  E stored
So, if we think of the energy conservation equation the rate of energy into the system,
which would be qx , q y , and qz ; and the rate of energy going out which would be
qxdx , q ydy and qz dz . The amount of energy generated inside the system, which is
 
E gen should be equal to q  dx  dy  dz  . The algebraic sum of all these must be equal to
the time rate of change in energy storage inside the system.
25
So, with the terms which are already written before, I am going to say that rate of energy
in plus rate of energy generated minus rate of energy out is equal to the time rate of
energy stored inside the control volume. So, the energy equation would be
 T
qx  q y  qz  q  dx  dy  dz   qx dx  q y  dy  q z  dz  C p dx  dy  dz  .
t
Substituting the Taylor series expansion for the relevant terms and upon proper
simplification, the final equation can be represented as:
qx q q  T
  dx  y  dy  z  dz  q dx  dy  dz   C p dx  dy  dz 
 
x y z t
T
We also understand this, from Fourier’s law qx  k  (dy  dz )  ; similarly,
x
T T
q y  k  (dz  dx)  and qz   k  (dx  dy )  . Substituting these equations in the
y z
above equation, and upon simplification, we get:
  T    T    T   T
k  k  k   q  C p
x  x  y  y  z  z  t
So, this is one of the more general forms of the heat diffusion equation. And, if you look
closely it gives you the variation of temperature with x, y, z, and time. And, this is
26
nothing, but the energy conservation equation where the only mode of energy transport is
by conduction.
If k is a constant; that means, it is not a function of x, y, and z; then the system is

isotropic. In an isotropic system, the properties do not change with direction, which is
true for many of the common systems that we deal with it, but for some special crystals
when the system size is reduced this k can be a function of the direction.
To simplify the above equation, we consider that our system is isotropic and hence, the k
can be taken out of the differentiation sign and what we get is the most common form of
heat diffusion equation, which would simply be:
  2T  2T  2T   T
k  2  2  2   q  C p
 x y z  t
This is probably the more common form of heat diffusion equation that you would
encounter.
A slight rearrangement gives another form of this equation:

 2T  2T  2T q  C p T
   
x 2 y 2 z 2 k k t
27

 2T  2T  2T q 1 T
 2  2  2  
x y z k  t
k
Where,   and if you put the units of k, rho, and C p in here you would see the
C p
units of alpha are simply meter square per second.
So, what this tells me is that alpha is known as the thermal diffusivity of the system.
Thermal diffusivity would tell us how fast a temperature front will propagate through the
solid to reach another point far from the original point.
So, conceptually the thermal diffusivity and the mass diffusivity are going to be the
same. And, you probably recollect that in fluid mechanics, you have studied you have
seen that mu the viscosity divided by rho, which is which is the density this mu by rho
also has units of meter square per second. And this mu by rho is known as the
momentum diffusivity.
So, what you would get 3 diffusivity is which you would see in one for the case of
momentum transfer, which is mu by rho the other is the case of thermal the energy
transport, that is k by rho C p and the third one is the more common more the one which
you commonly come across is the mass diffusivity which is denoted by DMA. The units
of all 3 quantities would be meter square per second, and the same units for these 3
quantities would lead to the unified treatment of heat mass and momentum transfer, that
you would come across in transport phenomena course in your later years, but the fact
that all 3 have the same unit would possibly have led people to think, that it is possible to
use the same type of analysis for heat mass and momentum transfer.
So, the fundamental equation which we now then have for conduction is this. So, this is
known as the heat diffusion equation. And this specific form is for constant thermal
conductivity system. Now many simplifications to this equation are possible for
example, if you assume that it is the heat conduction is taking place in a system in which

you do not have any heat generation. So, this q term would simply be equal to 0. If, you
assume that the system is at steady state and by steady state I mean that the temperature
is a could be a function of x y and z, but it is not a function of time. So, at steady state
28
this temperature change with time would simply be equal to 0 and you will be left with
only the 4 terms on the left-hand side.
The simplest possible situation describes the heat diffusion equation, being applied to a
case where you have only 1 dimensional conduction with no heat generation and it is at
steady state. So, that is the simplest possible thing which would give you that T is not a

function of y, T is not a function of z, there is no q present in the system and T is not a
 2T
function of time. So, if that is the case time then I will only have this term ( ) present
x 2
in it and I do not have to write the partial sign since T is a function only of position.
 2T
So, the governing equation for such a case would simply be  0 . If you solve this
x 2
equation, then T is a linear function of x. So, if in a solid let us say T is a function of x
only and it is at steady state, no heat is generated in the system, then for that special case
you are going to have a linear distribution of temperature inside the system. The same
type of equation can also be derived for the case of cylindrical systems and spherical
systems.
So, in cylindrical systems the temperature can be function of radial location r,  , or axial
location z. So, T is going to be a function of r,  , and z and in the case of spherical
systems the T can be a function of r,  and  . So, these two equations in spherical and
in cylindrical coordinates are available in your texts. Since, the geometry is slightly more
complicated for the case of spherical and cylindrical systems the heat diffusion equation
that you see for spherical and cylindrical systems do look more complicated, but
fundamentally there is no difference. They only use conservation of energy to explain the
physics of the system when only conductive heat transfer is present in the system.
So, what we do next is whenever you have a governing equation like the way we have
seen for the case of 1 dimensional steady state conduction with no heat generation my
aim is to obtain T as a function of x, y, z, and time. So, wherever you are solving the
governing equation, you are going to get integration constants. And the integration
constants can only be evaluated if you know what the boundary condition is. We would
29
see in the next class; what are some of the pertinent boundary conditions, which are
commonly encountered in conductive heat flow and how with the use of appropriate
boundary conditions we would be able to obtain the temperature profile inside a solid
undergoing conduction.
30
Heat Transfer
Lecture - 04
Relevant Boundary Conditions in Conduction
So, in the last class we have derived the heat diffusion equation, which is the
fundamental equation for conduction heat transfer and there, we have considered the
flow of heat, net flow of heat by conduction in the x direction, in the y and in the z
directions. And then we invoked the conservation of energy, which tells us that for a
control volume the rate net rate of energy in, the rate of energy in minus the rate of
energy out plus any heat, that can be generated that is generated inside the system the
algebraic sum of these three terms should result in a change in the energy stored of the
system.
Or in other words it is expressed as
dE
E in  E out  E g  E st  st
dt
that is rate of energy in minus the rate of energy out is equal to the rate of energy stored
in the system that is the time rate of energy stored in the system. And we know that the
energy of a system can be expressed as ρ, the density, Cp, that is the heat capacity times
the temperature minus the reference temperature. So, when we express that for a control
volume and from this equation, which is a difference equation we have obtained the
governing equation for conduction in Cartesian coordinate systems.
31
And this is the equation which we have derived in the last class, where for, Cartesian
coordinates systems the temperature change now, so, in all these cases T could be a
 2T
function of x, y, z and time. So, the and for y, for z and this q is the amount of heat
x 2
generated per unit volume. k is the thermal conductivity and this α as we have as we
k
have seen its units of meter square per second, it is defined as and its known as the
 Cp
thermal diffusivity.
So, this is the equation that one has to use one this is the starting point, for any
conduction analysis for a system, where the temperature could be a function of location,
where there could be some amount of heat generation and as a result of all these, the
temperature can also vary with time at fixed x y and z.
So, this is the fundamental equation, which is also known as the heat diffusion equation.
Similarly, the similar type of equation can be derived for cylindrical systems as well as
for spherical systems. I did not write the spherical systems which fundamentally,
conceptually there is no difference between these 2 except for the coordinate system. So,
here you see that T is essentially function of r, Φ, and axial location z and it could also be
a function of time. So, the same way the equation for the spherical coordinate systems
can also be written, it is there in your textbooks. So, am not writing it over here once
32
again and we have all discussed that how this equation, let us say this equation can be
simplified when we have for different conditions. For example, let us say we have a
steady state system. In a steady state system the temperature does not vary with time, ok.
So, the right hand side would be 0 and let us also assume that we have a situation in
which there is no heat generation in the system; so, this term would also be 0.
So, therefore, temperature is a function of x y and z only. Under certain under some
conditions it can also happen the temperature is a function only of x and not of y or of z.
So, if we think about steady state condition with no heat generation and temperature
being a function of only one spatial coordinates, then this equation would simply this
 2T
equation would then can simply be written as  0 . I do not need to use the partial
x 2
sign anymore, since t is function of time in which would T is a linear function of x.
So, this we have we have discussed and we also understand that this governing equation
can be solved with certain boundary conditions. So, in today’s class we will see what
those boundary conditions are, which could be used to solve to obtain the temperature
profile starting at the governing equation. So, now, we are in a position to know what is
the governing equation, what kind of simplifications I can make to that governing
equation depending on whether it is a steady state or an unsteady state, temperature is a
function of x or temperature is a function of both x and y and so on.
So, for all these cases if you look at the equation once again, you would see that you
would require 2 conditions on x, 2 conditions on y, 2 on z and one initial condition on
time to solve for the temperature profile. So, in the simplest possible situation of one
 2T
dimensional steady state conduction, where the governing equation is  0 . So, you
x 2
require 2 conditions on temperature. When you have an unsteady state situation;
however, the temperature is still a function only of one spatial coordinates, then again if
you look at the equation you would see only the first term on the left hand side and the
right hand side will remain in the equation.
So, therefore, your governing equation would be
33
 2T 1 T

x 2  t
 2T
So, this governing equation would then require 2 conditions on x since it is on the
x 2
left hand side, and it would require one condition on time that is a initial condition when
you look at the right hand side. So, for that specific case you need one initial condition
and 2 boundary conditions. So, it is important to understand, important to identify what
could those boundary conditions be. And the boundary conditions can also be coupled
with some other heat transfer process which is taking place in at the boundary.
For example you have you have a let us say cuboid, which is at a temperature high
temperature 100 degree centigrade, and which is exposed to a room where the
temperature is maintained at 25. So, inside the solid inside the solid cuboid the heat
transfer mode is conduction and if we neglect radiation, since the radiation is going to be
important only when the temperature is sufficiently high.
So, at the boundary of this cuboid what you get is convection. So, the cuboid is going to
lose heat to the ambient air by a process which is which all of us would know as
convection. So, let us say a small breeze is blowing over the cuboid making it cooler, and
in that condition at the interface between the solid the moving air what you have is
convection.
So, inside the solid its conduction once you reach the interface between the solid and the
air what you have is convection. So, in some of the some of the heat transfer cases
conditions, you will have the simultaneous presence of convection and conduction at the
interface. So, whenever we talk about convection or whenever we discuss about some
other modes of heat transfer, we need to identify what would be the governing equation
for that mode of heat transfer. Because we understand that in conduction the heat the
flow of heat is governed by Fourier’s law of conduction. So, what governs the
convective process? The convection is much more complicated than conduction because
in convection, you are also going to have flow of the fluid over the hot surface.
So, you need to know; what is the velocity profile, what kind of velocity you have on the
solid plate, what is the combined momentum and heat transfer process that is taking
place close to the interface. So, you not only have to have an idea of the fluid mechanics
34
the momentum transfer part of it, you would also have to have something which is
similar to Navier-stokes equation and applicable for energy transfer, which we would
cover later on when we have a more thorough study of convection. But it is suffice to say
that the process of convection is complicated because of the presence of fluid motion and
the associated momentum and heat transfer process with it.
So, in convection it is customary to define a convective heat transfer coefficient and it

has been found experimentally that the heat loss from a heated surface exposed to a
moving stream of fluid of different temperature, the amount of heat loss is proportional
to temperature difference. This temperature difference is temperature difference between
the solid and the fluid which is moving over it. So, the q of convection, the q the heat
loss or gain depending on whether the solid is at a higher temperature or the liquid is at a
higher temperature, this loss or gain of heat in convection is proportional to temperature
difference.
So, mark the difference with the conduction heat transfer; in conduction heat transfer its
the thermal gradient, which is deciding how much of heat is getting transferred between
the between 2 points. Whereas, in convection the heat flow the heat rate is proportional
to temperature difference not the gradient and this law which is essentially obtained by
looking at experimental results of convection over a wide variety of substrates and fluids
at different temperatures, where the fluid is flowing with different velocities on the
surface and so on this is known as the Newton’s law of cooling.
So, Newton’s law of cooling is a fundamental relation of convective heat transfer, which
again cannot be derived which is the result of a large number of experimental
observations in a generalized form. So, the Newton’s law of convection we are going to
use Newton’s law of cooling and this Newton’s law of cooling simply tells us that, the
heat loss or gain by convection is proportional to temperature difference T s minus T∞
where Ts is the temperature of the substrate that is the solid and T ∞ is the fluid which is in
contact with the solid.
So, if this is the solid and if the temperature on the solid side of the interface is T s and let
us say the temperature of the fluid is T∞ at a distance far from it, then this
q  hA(Ts  T ) ,
35
where A is the area which is in contact with the fluid. This h is known as the convection
heat transfer coefficient. So, this h is the convective heat transfer coefficient and most of
the studies of convection heat convection is to find what would be the expression of this
convective heat transfer. Now let us think what is this h is going to be a function of. It is
our common observation that if you want to cool a surface, cool a solid faster you have
to blow air at a higher velocity.
So, the velocity of the system velocity of the fluid which is flowing over the solid is
definitely going to dictate, what is going to be an important parameter in deciding; what
is the value of the heat transfer coefficient convective heat transfer coefficient. It is not
only the velocity it would also depend on the thermophysical properties of the fluid.
What is its density, what is its thermal conductivity and since it is a case of flow and heat
transfer, whenever there is a mention of flow we cannot neglect we cannot set aside the
important parameter of fluid flow which is viscosity.
And it is also going to depend on what kind of a thermal conductivity the fluid which is
flowing will have. So, it is going to depend on the density of the fluid, on the thermal
conductivity of the fluid, on the specific heat of the fluid and also on the viscosity of the
fluid. So, ρ, C p, μ and k, these are the thermo physical properties, which would be
relevant to decide what is the convective heat transfer coefficient.
There is an operating parameter which is the velocity with which these the fluid is made
to flow over the solid surface. So, that velocity let us call it as u is also going to be in the
functional form of h. So, if we think of the expression of h this functional form of h
should contain therefore, it should contain the following parameters. So, h is going to be
a function of what is the velocity, which is which is the fluid velocity its going to be a
function of ρ, C p, μ and k among others and it will be in some cases there are going to
be certain other parameters for example, its going to depend on what is geometry of
these surface. Is it a flow taking place over a flat surface, is it on a rounded surface, is it a
rough surface, or is it a smooth surface.
So, those extraneous factors that define the nature of the surface will also decide what is
the value of the heat transfer coefficient. So, if you look at the parameters which I have
written over there, that that is its v, ρ, μ and k, they are also going to be this geometric
parameters let us I am representing it by L, where L is the characteristic length and I do
36
not talk about anything else right now like for example, the nature of the surface, the
roughness and all.
So, if you think of these v, ρ, μ and L you can clearly see that they can be combined in
 vL
form of Reynolds number whereas, Reynolds number is simply . So, h is going to

be a function of Reynolds number of the system and when you think of Cp, μ and k, you
will know that Cp, μ and k , they are called Prandtl number.
Cp 
Where the Prandtl number (Pr) is defined as , this Pr is Prandtl number; one of the
k
famous scientists who has done extraordinary work not only on momentum transfer, the
concept of boundary layer but also on heat transfer. So, most likely the co correlation any
relation that you would see for forced convection, where you the fluid is made to flow by
external agencies is going to be a function of Reynolds and Prandtl number.
So, we will see this more when we discuss convection, but right now it would be
sufficient to know that the Newton’s law of cooling will decide how heat gets transferred
between the solid and the moving fluid, which would help us in deciding what are the
boundary conditions for this case.
37
So, let us see what are the boundary and initial conditions that one would one would get
when you come across this. So, let us talk about this is the solid and on this side I have a
fluid, and in the solid let us say this is my x = 0 and this is x = L. So, through the solid
you only have conduction and at this point from the interface you have convection. So,
the first condition possible condition is that, the surface temperature is constant, which
would tell you the T at x = 0 at any time t is maintained at some value of Ts.
So, the if the surface temperature at this point is constant, then the temperature over here
would also be constant. And let us also assume that some amount of heat is being added
to this surface, which we call as a q s”. So, this is the flux and the second condition could
be that constant surface heat flux. This constant surface heat flux let us call it as q s”, as
again double prime refers to the to the flux not the total amount. Now when this heat
comes to the solid, when this could be a heater this could be some sort of radiation which
is following on the surface and its getting absorbed. Once it gets absorbed delivered at
x=0, the heat then travels through the solid by a conduction process.
So, qs” which is the heat flux at x = 0, must be equal to,
 T 
qs "    k 
 x  x 0
So, this is the this is the second type of boundary conditions which you may see where
amount of delivered heat flux, which is to be carried into the solid by a conduction
process and by invoking Fourier’s law, I can write that the heat flow at x = 0 by
T
conduction is k and this is x equals 0 is a control surface. A control surface does not
x
have any mass of its own.
So, for a control surface the conservation equation takes a form that in is equal to out and
therefore, at the control surface the amount of heat which comes in is q s” and the heat
that gets transformed, that gets transported by the conduction is simply going to be a
 T 
 k 
 x  x 0
38
 T 
qs "    k 
 x  x 0
So, how would it look like? Let us say this is my q s”, this is the temperature profile and if
I draw a tangent to this at x = 0.
So, this temperature could be a function of x and time. So, this tangent is essentially your
T at x equals 0. Multiply that with minus k and what you get is q ”. So, this is by a
s
x
conservation by app invoking conservation at the control surface of the in input heat flux,
and the heat which gets transported in the x direction. The third condition could be an
adiabatic wall or insulated wall. So, if it is an adiabatic or an insulated wall what you
have then here is that, this is the surface on this side its perfectly insulated. So, if it is
perfectly insulated then qs” must be equal to 0. If it is insulated then no heat can cross
this surface.
So, if no heat can cross this surface, I am again going to write the equality of the heat
that goes out and the equality of and the heat that reaches the surface by conduction. So,
this side is solid and I on the on the other side, I have perfectly insulated this. So, q s” due
to insulation is equal to 0. So,
 T 
qs "    k 
 x  x 0
 T 
So, this would give you that   = 0.
 x  x 0
So, perfect insulation or an adiabatic surface would simply show you that the, gradient of
temperature the spatial gradient of temperature at that plane is equal to 0; which comes
directly again by a heat balance on the control surface since no heat can cross the control
dT
surface, this k on the other side has to be equal to 0 which would give rise to since k
dx
dT
cannot be 0 since is 0. So, the temperature gradient vanishes, temperature gradient
dx
becomes equal to 0 on a surface which is perfectly insulated or which is adiabatic and I
39
will give you examples of adiabatic surface later on, when we discuss about how what is
the what is the amount of heat for a system in which you have heat generation.
A system in which you have symmetric generation of heat everywhere then, you would
see that at certain point inside the inside the solid object, the temperature reaches the
dT
maximum. And when temperature reaches maximum then; obviously, at that point
dx
dT
would be equal to 0. So, if is 0 in that plane, then no heat can cross this plane in
dx
either direction and that such a surface is known as the adiabatic surface and the required
boundary condition, the prerequisite for having no heat crossing this adiabatic plane from
dT dT
either side is to 0. So, to be 0 can be a valid boundary condition, if we know that
dx dx
at certain location the its either an adiabatic surface or it is a surface which is perfectly
insulated not allowing any heat to escape so, that could be the third boundary condition.
And the fourth boundary condition we would see that its a case of convection, surface
condition. Convection surface condition it would tell you that if we have a solid on this
surface side and let us say a fluid on this side and what you have is the, fluid is moving
with some velocity and let us say its temperature is T∞ and the thermal and the convective
heat transfer coefficient is h and over here the temperature varies as a function of x and
time.
So, if I take this as my control surface, and let us say that T ∞is more than T. So, if the
solid is exposed to a hot environment. So, heat is going to move in this direction from the
fluid to the interface by convection, from the interface to the inside of the solid by a
conduction process. And at the interface the convection must be equal to the conduction.
 T 
So, what is conduction? Conduction is  k  . So, this is the conductive flow of
 x  x 0
heat conductive heat flux which must be equal to the convective heat flux, which is
T (at x = 0) - T∞.
40
This left the solid side, this is the fluid, the fluid is moving with a velocity with some
velocity having a temperature T∞ and with a convection heat transfer coefficient as h, and
let us say that the fluid temperature is more than that of the solid. So, this is how the
temperature inside the solid would change with time.
So, if we take this to be our control surface, then the amount of heat which reaches by
convection the same amount has to be transported by conduction since this is a control
surface and therefore, conservation demands that convection and conduction are equal,
and if we write that conduction equals convection at x equals 0. This is let us say this is
the location of x equals 0 then what you have is
 T  = h (T -T(0,t))
 k  ∞
 x  x 0
where T∞ is the temperature of the fluid at a point far from it. So, this is at x = 0 at any
given time t. So, these kinds of conduction, convection, mixed boundary conditions are
also prevalent in the study of heat transfer. So, once again if you look at the possible
boundary conditions, one is the temperature that could be one condition, the heat flux is
known where the heat flux is equated to the conductive heat flow. So, that could be
another boundary condition, the third could be an adiabatic or insulated wall condition,
T
in both cases qs”is 0 which would tell you that is equal to 0.
x
41
And the fourth condition is this which is a convection surface condition where the
equality between conduction and convection at x = 0 would simply give you that
 T  is this the convective flux. Now depending on what is the condition what is
 k 
 x  x 0
the specific problem that you are dealing with, one or more of these conditions can be
used.
So, any solution of any heat transfer problem is first to identify what is the coordinate
system that needs to be used. Is it a cylindrical system or is it a planar system or a
spherical system. Then write the governing equation for that situation and based on the
physics of the problem, you cancel the terms which are not relevant. If it is one
dimensional problem with heat generation, but at steady state then your equation would
be if you look at the equation once again you would say that temperature is a function
only of x, its not a function of y, its not a function of z and it is not a function of time.
d 2T
And q is not 0; if that is the case then the equation would be the first among the left
dx 2
hand side, the second and the third term would be 0 since T is not a function of y and T
is not a function of z. The fourth term is q dot by k where q dot is the generated heat per
unit volume. And the right hand side temperature is not a function of time. So, therefore,
the only term on the right hand side can also be equated to 0.
So, your governing equation in that case would be
d 2T q
 0
dx 2 k
So, this way looking at the problem at hand, you would be able to simplify the governing
equation to arrive at the final form of the governing equation. Then three out of the four
boundary conditions that I have discussed so far, which are relevant in for this specific
case. Is temperature at a given location known, is the heat flux known or does the
physics of the problem tell you that the flux at some x location such some location would
be 0 because either it is perfectly insulated or it is an adiabatic surface.
So, you know which boundary condition to use or is it that at some location, this solid is
exposed to a fluid and the solid is losing or gaining heat by convection. So, using
42
Newton’s law and the equality of conduction and convection at that specific location,
you should be in a position to write the mixed boundary condition which is the boundary
condition 4 that I have discussed.
So, that is how a conduction problem is to be attempted and if you can solve the
governing equation with the boundary conditions, what you get as a result is temperature
as a function of location as well a function of time. And once I have temperature known
at any point inside the control volume, then I can do all other possible analysis to obtain
the flux at a point and so on. So, what I would do now is, I will give you just a problem
for you to practice and then the in the tutorial part of it the problem can be analyzed in
greater detail. So, very quickly the problem that I would like to give you is about a solid
wall, which is generating some amount of heat.
So, this solid wall there is some amount of heat which is generated. So, E g and it is not a
steady state. So, there is some E in it, some amount of heat comes into the system, some
amount of heat goes out of the system. The x starts from here and x this is from 0 to L,
and it is known that the temperature inside this is as a function of x,
T(x)= a+bx+cx2
So, this L is equal to 1 meter and this is the temperature profile at a certain instant of
time.
43
So, that is why you would see that the temperature function, the temperature expression
does not contain any time, because it is measured and expressed in this form at a given
instant. And the values of a, b, c which are constant these are, a is 900 ºC, b is -300 ºC/m
and c is -50 ºC/m2.
The value of q which is the amount of heat generated inside the system is 1000 W/m 3.
So, this is heat generation per unit volume and the area of the wall is 10 m 2. The material
of the wall, ρ is 1600 kg/m 3, the thermal conductivity k is 40 W/m K and Cp the heat
capacity is 4 KJ/Kg K.
What you need to find out is; what is the rate of heat transfer entering the wall entering
the wall which is at x = 0 and leaving the wall which is at x = L. What is the rate of
energy storage in the system at this instant. So, what is the value of E st and third one.
So, you have to evaluate what is the rate of energy stored in the system at the given
instant of instant, where the temperature can be expressed in this form. And the last part
is; what is the time rate of temperature change at x = 0, x = 0.25 and x = 0.5 m.
So, once again the problem I described, where we have a wall in which some amount of
heat is generated, some amount of heat comes into the system and it leaves out of the
system, and as a result of imbalance of these three quantities the energy stored in the wall
will also change. However, the temperature at a given instant of time can be expressed in
the polynomial form of a+bx+cx2. The relevant quantities for example, L, a, b, c are
provided as well as the amount of heat generation and the area through which the wall is
the area of the wall that is also provided.
So, this area cross sectional area is 10 m 2, the density, thermal conductivity and specific
heat of the wall material is also provided. What you have to find out that what is the heat
transfer rate of heat transfer entering the wall that is at x = 0, leaving the wall which is at
x = L and what is the rate of energy stored at this for this system at that instant of time
T
and time rate of temperature change. So, at these three locations.
t
So, how do you solve the problem? As I mentioned the first thing that you need to
identify is, when you look at x = 0 plane, the amount of energy which enters the wall as a
44
result of which the temperature profile is provided. So, if I take my this as my control
surface, then
 dT 
q in   kA 
 dx  x 0
So, at this location the amount of heat that comes in is going to be taken up by
 dT 
conduction on this side, and the conductive heat flow is  kA  So, by equating q in
 dx  x 0
with the conductive heat flux and knowing temperature as a function of time, I can find
dT dT
out what is and evaluate what is at x = 0 and therefore, find out what is what is
dx dx
the value of qin.
 dT 
So, q in   kA  and the temperature profile is known to me. Similarly
 dx  x 0
 dT 
q out   kA 
 dx  x  L
So, the T as a function of x is known to me.
dT dT
So, I can find out what is , what is at x equals L, the value of k and the value of
dx dx
A are known to me. So, I would be able to find now what is q out . So, these two it should
be able to evaluate. Now what is E stored . So, from my energy equation
E in  E out  E g  E st
So, E in is nothing, q in and my E out is nothing, but my q out , what is E g ? The value of E g
equals where the q is provided as 1000 W/m3. So, when everything is expressed on a per
unit volume basis, my E in is known, E g is known, E out is known. So, I should be able to
find out what is E stored . And I leave this part for you to find out for you to think and solve
45
T
and what I can tell you is that this , is you will see that it is not a function of x , value
t
of temperature change with time that you would determine, it will not be a function of x.
So, the time rate of temperature change at x equals 0.25 or 0.5, all are going to be equal.
What I am what I am going to do is I will you the values the answers to this problem and
you can try it on your own.
So, q in is going to be 120 kilo watts, the q out is going to be 160 kilo watts and the energy
stored in the system is this part is going to be equal to - 30 kilo watt so; that means, the
system is losing this much of energy at that instant of time and you would say that del T
divide the temperature change with time, the value of this would be - 4.69 * 10 ^ ?(-4 )
ºC/ s.
So, the answer to part the first one the in is 120, the out is 160, the stored is minus 30 and
the time rate of temperature change is going to be - 4.69 * 10 ^ ?(-4 ) ºC/ s.
46
And the way this part is going to be solved, I will simply write it once again the
governing equation for this case would be
k  2T q T
2
 
 Cp x Cp t
So, this is the form of the governing equation, which you would obtain by writing the
energy equation and cancelling the terms since T is a function of x and time.
 2T  2T
So, all the terms that those are and will be of will be set to 0, temperature as a
y 2 z 2
 2T
function of x is known to you. So, you should be able to find out what is . The
x 2
T
values of k, ρ, Cp, q are known to you. So, of only unknown here . And looking at
t
 2T
the expression of T as a function of x you would see that is going to be a constant
x 2
that simply going to be 2 C.
 2T
So, is simply going to be 2 C. So, the entire left hand side which is time rate of
x 2
change of temperature is going to be a constant. So, it is immaterial at which value of x
you are measuring the time rate of temperature change, it is not going to be a function of
47
x and when you plug in the values you would see that, the time rate of change of
temperature would be - 4.7 * 10^ (-4) ºC/ s.
So, this completes our today’s lecture, where we have seen what could be the relevant
boundary conditions and we have we have discussed a problem where a direct
application of Fourier’s law and the conservation of energy can be used to arrive at
numbers for the heat, which comes into the system going out of the system and time rate
of change of temperature and so on.
So, I believe if you go through this your concept of conductive heat transfer must be
clear right now and the use of the heat diffusion equation. In the next classes we will see
how this heat diffusion equation can be modified to obtain special cases, solutions for
special cases and what form would it take for, not only for Cartesian coordinate systems,
but also for cylindrical and spherical systems.
48
Heat Transfer
Lecture - 05
One Dimensional Steady State Conduction
Previously, we have derived what is the conduction equation for Cartesian co-ordinate
systems, cylindrical systems, and spherical systems. In the development of those
equations, we have assumed the temperature can vary in all three possible directions and
it can vary as a function of time as well. And there may be a situation in which some heat
is going to be generated inside the control volume.
So, the equation that we have that is the generalized governing equation for conductive
heat transfer in a solid medium. But, we are going to see some of the special cases which
arise out of this equation which has practical applications; for example, in many of the
situations we see that a plane wall has layers of insulation on it, to reduce the heat loss
during winter or heat gain during the summer.
So, when a plane wall has several layers of insulation attached on one side of it and if we
know the temperatures of each of the junctions between the insulator and the wall, then it
is possible to predict what is the total heat loss from the wall to the outside at steady state
conditions. Because, in practical applications we mostly are concerned with what is
going to happen at steady state. So, it is important therefore, to simplify the heat
diffusion equation for special conditions when you have only a plane wall and at steady
state. We would also add another restriction that there is no heat generation in the control
volume that we are considering. Again, going back to the example of heat loss from the
walls of a room at steady state with insulations on the outside of the wall; obviously,
there is no heat generation in the insulation.
We would also need to appreciate the fact that at steady state, no matter whatever be the
thickness of the insulation or the material of the insulation; the same heat goes through
all the layers. So, the heat rate is constant the amount of heat which will be lost by a hot
room with layers of insulation on the outside is a fixed quantity. And secondly, when you
think about a plane wall and insulation on top of it, then you do not have any variation in
the heat transfer area. So, the same area is exposed to heat transfer from inside the room;
49
to the outside, and the cross-sectional area does not change. So, therefore, the special
characteristics of a plane wall is that the area remains constant, and if the area remains
constant we understand from physical principles; that since there is no heat generation
and at steady state, the same amount of heat must travel through all these layers of
insulation.
So, for planar systems the heat; both the heat rate and the heat flux are constant; which
may not be the case for cylindrical systems or spherical systems. Because as you move
away from the center of a cylindrical or a spherical surface, the radius increases and
since the radius increases the area available for heat transfer also increases. So, the
quantity of heat will remain constant, but the flux will not be constant. So, that is the
major difference of conduction in planar systems, at steady state and conduction in radial
or spherical systems.
The fact that heat flux may vary for non-planar systems will also lead to interesting
observations, that it will also affect the temperature profile. In one case (for planar
systems), the temperature distribution is going to be linear; whereas, in the case of
cylindrical and spherical systems, they are going to be non-linear.
So, our starting point for analysis of conductive heat transfer in plane walls starts with
the heat diffusion equation. And then we are going to simplify the equation based on its
one-dimensional heat transfer; no generation of heat and at steady state.
50
So, we start first with this equation which we have derived previously in the class:

 2T  2T  2T q 1 T . So, this is temperature as a function of x, y, z, as well as
   
x 2 y 2 z 2 k  t
time.

So, there is a transient component also attached to this and q is the heat generated per
unit volume. So, therefore if it is 1 dimensional and steady state condition then I can say
that T is a function only of x; let us assume that x is the direction in which the heat flow
takes place and T is not a function of time.
So, when I talk about steady state what I mean is that the temperature can be a function
of position, but at a given position the temperature is not a function of time. So, for the 1-
dimensional case temperature is a function of x, but temperature is not a function of time

and we will also assume that there is no heat generation in the system therefore q , is 0.
So, if we apply these 3 conditions to the above equation, the governing equation reduces
 2T
from the heat diffusion equation as  0.
x 2
Integrating the above equation would simply give you the temperature as a function of x
as: T  C1 x  C2 ; where C1 and C2 are constants of integration. So, now to solve this I
need boundary conditions. Specifically, two boundary conditions are required and what
you would consider is that the temperatures at the two sides of the wall are known.
Let us consider that T at x = 0; that means, at one edge of the wall the temperature let us
say it is a T1 and T at the other edge of the wall which is at x = L is equal to T 2. And we
will assume that T1 is greater than T2 though the derivation will work fine; even if you do
it in the other way that T2 is greater than T1; only the direction of heat transfer will be
different.
So, when you use these two boundary conditions, the solution would turn out to be:
51
x
T  x    T2  T1   T1 . So, this tells me that temperature varies linearly with x. So, this
L
clearly again shows that the temperature at any position is a function of x only and it
depends on the imposed temperature difference which is T 1 minus T 2; it also depends on
the geometric parameter, this L is the thickness of the wall, and the imposed
temperatures.
From Fourier’s law we know that q, the heat flow in the x direction is equal to minus k
times A dT/ dx where k is the thermal conductivity. So, one can write looking at this
dT T1  T2 q k
expression and finding out  and the heat flux is simply x   T1  T2  .
dx L A L
The important point to note here is that the heat flow; the total rate of heat flow with
units of Watt and heat flux with units of Watt per meter square both are independent of
x. So, our derivation simply shows what we have started with the basic premise that both
the heat rate and heat flux are constant in planar systems; which would not be the case
for a spherical or radial system, where the cross sectional area keeps on changing as you
move in the r direction.
Now let us assume that we have a system of a plane wall with three materials which are
A, B and C. There would be four possible junctions between these three, materials
(denoted by 1, 2, 3, and 4 respectively) and let us assume that the temperature profile is
52
linearly decreasing. And as you can see that as you move from 1 towards 4 the
temperature decreases, and we know that heat always flows in the direction of decreasing
temperature.
So, at steady state the same q will flow through all the materials - A, B and C. So, in a
 T T 
multi layered since q is constant we can write: q A   k A AA  2 1  for the first
 x A 
 T T 
material; and qB   k B AB  3 2  for the second material and finally, for the third
 xB 
 T T 
material - qC  kC AC  4 3  .
 xC 
Rearranging the above equations, we get:
q A xA
T2  T1  
AA k A
qB xB
T3  T2  
AB k B
qC xC
T4  T3  
AC kC
The k A , k B , and kC are the conductive thermal conductivities of material A, material B,
and material C respectively; and x A , xB , and xC are the thicknesses of each of these
materials.
Adding the above three equations and rearranging them by taking q as a common
parameter would result in:
T4  T1
q
x A xB xC
 
k A AA k B AB kC AC
53
It is to be noted for the case of materials with same cross-sectional area, the area term in
the denominator could also be taken to be a constant value.
So, what I can think of this as T 1 is one of my potential and let us say I have a
resistance, then I have T 2 another resistance and I have T 3 and another resistance, and I
have T 4.
x A
So, temperatures can be viewed as potentials and the heat transfer resistances are: ;
k A AA
xB xC
; and . So, what you have then is the overall potential difference divided by
k B AB kC AC
the sum of resistance. So, only thing is the potential difference to be replaced in Ohm’s
law to arrive at the heat transfer law is overall temperature difference by sum of all
conductive resistances.
So, you have 3 or 4 materials sandwiched one after the other and if do know what the
temperatures at the 2 end points are, then you can find out what is the flow of heat
through this combination materials; through this multilayered wall. And the result that
you would get is something like what you have obtained in the case of Ohm’s law.
x A
One thing to be noted here is that the conduction resistance is , whereas the
k A AA
convection resistance (from Newton’s law of cooling) is equal to 1/h A. So, if a wall is
connected with the convection environment then I am going to have 3 conduction
resistances in series and a convection resistance being added to it.
We would now go to a situation which is more practical, where these walls through
which conduction takes place they are exposed to an environment outside air. So, the
room that I am sitting in it is an air-conditioned room; it has multiple layers of insulation
on the outside and then it is exposed to the outside ambient temperature. So, the amount
of heat loss from this wall is going to be combination of all the conduction resistances
due to the insulation and the wall that we have for this room; as well as what is the
convection resistance on the outside of the building which connects it to the ambient
temperature.
54
We will quickly see how the result would look like for a system in which I have both
conduction and convection; however, it is a planar system such that both the heat rate
and the heat flux are going to be constant.
So, I have conduction, convection both are combined, and the diagram probably looks
something like this, where this side is T infinity 1 and this side let us say the temperature
is T infinity 2.
So, I have a hot fluid which is moving up in this direction on this side of the wall and I
have a cold fluid which is moving on the outside of this; so, this my wall. And the
temperature here is T infinity 1, the temperature here is T infinity 2 and the temperature
is going to sharply fall because of the thermal boundary layer that we have discussed
before.
So, let us say this is the surface temperature of 1 and then I have a T S 2 which is the
surface temperature of 2 and then it falls to T infinity 2. So, over here the fall is also
going to be like this and in between T S 1 and T S 2; since it is a plane wall, at steady
state without any heat generation; the temperature profile must be linear. The
temperature T infinity 1 is more than T infinity 2 and therefore, the flow of heat is going
to be from the left to the right.
55
So, this is a system in which I have conduction in the plane wall, convection on the
inside as well as on the outside of the plane wall. So, if I draw the resistance diagram for
this where the outside potential is T infinity 1 and the potential over here is T infinity 2;
these are the 2 temperatures then I have some sort of a convection resistance which
brings me to T S 1 and then the conduction resistance which brings me to T S 2 as per
the diagram and then I have a convection resistance which then results in asymptotically
and smoothly merges with T infinity 2.
So, this is my convection resistance which must be equal to and let us assume that this
hot fluid maintains a convection coefficient h1 over here and a convection coefficient of
1
h2 at this point. So, therefore, the resistance is simply going to be ; as we have
h1 A
L
discussed before. The conduction resistance is going to be ; where L is the thickness
kA
of the wall, A is the area, k is the thermal conductivity and over here it is going to be
1
.
h2 A
So, that is what we would get in this case therefore, the total flow of heat q is simply
going to be T infinity 1 minus T infinity 2 divided by R total; where R total refers to the
algebraic sum of these 3 resistances; since they are in series. So, this is R total would
1 L 1
simply be equal to   that is the total resistance to heat transfer in this
h1 A kA h2 A
system.
It is sometimes advisable to express these 3 resistances is in terms of an overall heat

transfer coefficient. And we would see the reference to overall heat transfer coefficient
coming many times in convective heat transfer not only in convective heat transfer, but
more commonly in the case of designing of heat exchange equipments, where instead of
each of these individual resistances we express them in terms of an overall resistance and
this is generally denoted by the symbol U and therefore, the q the heat flow is going to be
U A, delta T overall as I mentioned the U appears mostly in the case of conduct
convective heat transfer.
56
So, the total amount of heat transfer is simply going to be according to Newton’s law of
cooling as q equals U A delta T, where U is the overall coefficient which may contain
both conduction and convection as we have seen in the example that we are discussing.
The A is the cross-sectional area through which the heat transfer takes place and delta T
is the overall temperature difference.
And it; obviously, does not assume that it is a linear or anything, but we understand that
it is going to be linear in the solid, but it is going to have a shape which is where the
temperature from the ambient comes to the temperature of the surface in a fashion which
is consistent with our explanation based on thermal boundary layers.
Now, let us look at composite walls which are slightly different. Let us say you have a
system in which up to this point, the wall is made of some material and from this point
onwards it is of another material, but in between you have 2 different materials.
So, you have a material E with a thermal conductivity k E of length L E; for this one its
length is going to be LF, the thermal conductivity is going to be the k F and this is
material F. And I have k G in here, this is material G and this is material H where it is
going to be L H and k H and the temperature here on this side is going to be the T 1 on
this side it is going to be T 2.
57
So, as you can see that the heat is going to flow; if T 1 is greater than T 2 heat is going to
flow from 1 to 2, but in here after it reaches this surface; part of the heat is going to move
through this, part of the heat is going to move through this and they are going to emerge
in material H and flow. So, this is the overall direction of q, but at this location at this
plane, it is going to get divided and at this plane it going to combine; so that you get the
same q and here you also going to get q.
This is a 2 D condition where the temperature is not a going to be a function of x alone, it

could also be a function of y. So, there are 2 ways by which you can represent this in
terms of an equivalent electric circuit. So, you have T 1 on one side T 2 on the other side
LF
LE
and here the resistance is simply going to be and here it is going to be A . We
kE A kF
2
write A/2 because we assume that half of the area is taken up by the material F and half
of the area is taken by material G.
So, that is why I am putting it A by 2; it can be any fraction and it does not matter one
third of the area can be made of F and 2 third will be made of G. So, appropriately the
fraction of area is to be is to be replaced over here and, in this case, it is going to be
LG
LH
A ; as I said A by 2 is just for this specific case. And over here it is going to be .
kG kH A
2
So, you have 3 resistances in series for the top part and so, this is my T 2 and this is my
T 1 and heat is going to flow through this and through this. So, q 1 is flowing through
this and q 2 is flowing through this where the sum of q 1 plus q 2 is going to be equal to
q.
So, whenever you come across a composite walls; it is customary to express them either
like this or in either like this or in this fashion, but in both cases we have to appreciate
that near the junctions there would be 2 dimensional effects, 2 dimensional conduction
where T is going to can be a function of both x and y where y is this direction and x is at
this direction.
58
Explaining these complicated flows of heat through composite walls by a simple
resistance in series or resistance in parallel mechanism would only give you an
approximate value of the temperature, the junction temperatures and the temperatures the
total flow of heat. So, this is an approximation and higher the difference between the
thermal conductivities between the adjacent layers the error is going to be enhanced.
So, one should be careful about expressing the results in this specific form and most of
the time, it is going to be dictated by how fast you want a result; whether you can tolerate
approximations and what is the physical condition of difference in thermal conductivities
such that the difference in temperature across a fixed x is not that significant. So, that
must be considered before you start solving it.
59
Heat Transfer
Lecture - 06
Temperature Distribution in Radial Systems
In the last class we discussed about what would be the resistance for heat flow through a
planar wall or a combination of several plane walls. And, there we saw that the
temperature difference between two ends of the wall can be viewed as the potential
difference which causes the heat to flow from high temperature to low temperature and
the resistance, if the heat flow can be expressed in the form which is similar to that of
ohm’s law where the potential difference is to be replaced by the temperature difference
and the current is to be replaced by the heat flow. And whatever we have in the
denominator of q equals δT by r, this r is equivalent to thermal resistance.
And the thermal resistance we have evaluated for a plane wall. And also if we have
several such walls sandwiched one after the other then the overall temperature difference
between the outer edge of wall 1 and the outer edge of wall n, then, this temperature
difference is the overall temperature difference i.e. overall potential gradient. And the
denominator is going to be a sum of all the individual resistances of each of these walls.
So, in essence the heat flow through a composite wall where cross sectional areas remain
the same at steady state can be expressed as heat flow equal to the overall temperature
difference by the sum of all resistances. And therefore, all these resistances it has been
shown that they are in series.
So, based on that we now clearly understood that the resistance of the plane wall is going
to depend on the geometry; that means, what is a cross sectional area, what is the
thickness and will also going to depend on the thermo physical property which is the
thermal conductivity of the specific wall that we are considering. So, the resistance is
simply going to be L, where L is the thickness divided by k times A, where, k is the
thermal conductivity of the material and A is the area which is perpendicular to the
direction of heat flow.
60
L
Resistance =
k. A
We have also seen that for composite walls of different type in which you have wall
number 1 whose area is , then another wall of different thermal conductivity whose area
is let us say A by 2 and another wall, wall number 3 whose area is a by 2, but it is of
another material. So when such walls, composite walls of unequal cross sectional area
are put together then there are two ways by which you can represent the heat flow; one is
you can assume that there is going to be a series, that is the top 3 are going to be in series
whereas the top and the bottom series are going to in parallel. So, we discussed about
that as well.
One of the characteristic features of heat flow through plane wall is that the heat flow is
constant, not only heat flow at steady state is constant, the heat flux is also constant since
the cross sectional area which is perpendicular to the direction of heat flow does not
change. So, as in planar systems the cross sectional area does not change with x, with the
direction of heat flow therefore, both the heat rate and the heat flux are going to be
constant.
However, I have mentioned in the previous class that for radial systems and for spherical
systems in which as you move in the direction of r, the cross sectional area keeps on
changing. And therefore, even though the conservation equation demands that heat rate
has to be a constant, the flux may not be a constant. And therefore, we are going to have
a different expression for the resistance to heat flow for the case of radial systems and for
spherical systems. So, that is what we are going to analyze in today’s class.
So, it is going to be radial temperature distribution at steady state and the mode of heat
transfer is conduction, but you can always have convection at the outer or the inner edges
of a radial of a wall. And therefore, it in that case it is going to become a convection
conduction problem, but for the time being let us start first with conduction only case in
a radial system.
61
So, the figure that you see over here is that of a that of a hollow cylinder whose inner
radius is equal to ri and the outer radius is r0 and the temperature over here in at the inside
is maintained at Ti whereas, the temperature at the outside is maintained at T o. So, this is
a situation the length of the hollow cylinder is L and we are trying to find out what is the
flow of heat through this area to outside. So, if we assume that Ti is greater than To then
this is the radial direction in which heat will flow through the solid material of the
cylinder and we are trying to find out what is going to be the resistance for heat flow.
For this case for steady state condition with no heat generation and the case where T is as
you can see is going to be a function of r only it is definitely not a function of θ and it is
not a function of z. So, T is a function of r only therefore, the heat diffusion equation for
radial systems reduces to only this much, because we do not have to consider the
variation of temperature with r or θ. So, T is not a function of either z or θ. So therefore,
all the terms containing gradient of T with respect to z or with respect to θ can be
cancelled as it is a steady state one. So, T is not a function of time and there is no heat
generation. So, therefore, if you look at the heat diffusion equation this is the form it is
going to take ok.
1 d  dT 
 kr 0
r dr  dr 
And one can integrate this equation and what you get is
62
T (r )  C1 ln r  C2
Boundary conditions:
T(ri) = Ti; T (r0)= To
So, this is your governing equation and in order to solve the governing equation you
need 2 boundary conditions which are the 2 known temperatures in this case. So, T at r i
is equal to Ti and the second is T at r0 is equal to To. So, when you put these 2 boundary
conditions to evaluate C 1 and C 2, the final expression, that you are going to get is
Ti  To r
T (r )  ln  To
r r2
ln i
ro
So here, what I think we should appreciate is that the temperature distribution unlike the
case of planar system is not a linear function of position, but it is going to be a
logarithmic function of the radial location of the plate.
So, that is very important to note that it is no longer going to be linear distribution, it is
going to be logarithmic distribution, and as you move in the direction towards the outer
radius the area of heat flow, the area available for heat flow continuously increases. So,
since it increases the temperature, it is not going to be linear and temperature here you
can see it is going to be linear function of position.
Now, what is remaining is since we are trying to correlate cause and effect, the cause
being the temperature difference and the effect being the heat flow and we are trying to
find a relation or express our results in the form of something similar to ohm’s law
between cause and effect. So, if you do that to this equation then automatically the
resistance to conductive heat transfer in radial systems will come out. So, that would be
our next exercise to see what form the resistance would take in a radial system.
So, if you look at this equation once again then I can write this as
dT Ti  To 1

dr r
ln i r
ro
63
Using Fourier’s law,
 k 2 L(Ti  To ) 2 kL(Ti  To )
qr  
ri r
ln ln o
ro ri
So, if I take this to the next logical step and here you see this is the effect and this is the
cause. So, if I bring twice by k L in the denominator what you have is something similar
to ohms law
(Ti  To )
qr 
r
ln o
ri
2 kL
where the cause is the potential difference is a thermal potential difference which is
expressed as temperature difference, and the effect is the radial flow of heat. And what I
ro
ln
have in the denominator is ri that simply is the resistance to heat transfer for radial
2 kL
systems.
64
So, I am going to write in the next page as temperature difference. So, what you have
here in the denominator is your resistance the thermal resistance which; obviously, is
this.
ro
ln
ri
Rth 
2 kL
So, this is an important result which gives us what would be the resistance to heat
transfer in such case. Let us assume that you have a composite situation in which you
have a number of walls like this, so this is r2. Similarly have r3 and r4 with each one
having thermal conductivity of k1, k2 and k3 and you also have convection inside as well
as convection outside, with the inside condition given as the temperature of the fluid in
here is T∞1 which is creating a convection coefficient of h1 at the inner surface of the
composite wall.
On the other side, you have T∞2 as the temperature of the fluid which is made to flow on
the outside which is also producing heat thermal convective heat transfer coefficient of
h2. This is quite common in many of the practical situations for example, in heat
exchangers which we would see, where the hot fluid let us assume that it is flowing
through a tube and somehow it is going to come in thermal contact with a cold fluid with
which it will exchange heat. And therefore, the cold heat is going to be heated, going to
absorb, going to gain energy out of this hot fluid.
And let us see hot fluid, fluid is flowing from point a to point b from where it is
produced to where let us say it is going to a reactor. So, in the transit you want to
maintain the heat loss to a minimum you want to keep the heat loss to a minimum. So,
how do you do that, you put insulation on top of the pipe top of that tube. So, this
insulation the pipe and the insulation they are may be 2 or 3 different types of insulation,
one is an insulation which is thermally which is going to protect heat loss which is going
to minimize heat lost and in order to protect the insulation you may have another outer
cover.
So, as the heat travels from the inside to the outside it is going to experience different
materials as it travels, and it is going to also see materials of different thermal
conductivity. And as it travels it will see that the heat transfer area will keep on
65
changing. So, under such circumstances what you are going to get is the picture that I
have drawn where you have a composite wall where at the in the inside you have some
convection coefficient. And at the outside you have a different convection coefficient
which is probably provided by the ambient air which flows over the over the lagged pipe.
So, what would happen to such a case is in is depicted over here and I am simply going
to write what is the form of the final equation going to be.
T1  T1
qr 
r r r
ln 2 ln 3 ln 4
r1 r2 r3
 
2 k A L 2 k B L 2 kC L
So, the radial flow of heat is going to be due to the temperature difference, the
temperature difference inside minus temperature difference outside.
So, these are the 3 conduction resistance through the material A B and C, I also have a
convection which is taking place over here. So, what is the convection the convective
heat transfer will be
T1  T1
qr 
1
 r1 Lh1
2
This comes from Newton’s law of cooling which simply says that
qr  hAT  h1 2 r1L1  T1  T1 
So, from here I have written it in the form of temperature difference by some sort of a
resistance. So, this is the expression for convective heat transfer resistance in the case of
radial systems. So, this is must be added to this overall resistance. Hence,
T1  T1
qr 
r r r
ln 2 ln 3 ln 4
1 r1 r2 r3 1
   
2 r1 Lh1 2 k A L 2 k B L 2 kC L 2 r4 Lh2
So, I think it is clear to you now is that what happens in this case where you have
multiple walls radial walls one after the other of different thermal conductivities, the
66
inside is exposed to a convection environment the outside is also exposed to another
convection environment with 2 different temperatures and 2 different heat transfer
coefficients.
So, the heat flow which has to remain constant in order to maintain equation of
conservation of energy is the overall temperature difference which is T1  T 2 by the

sum of all resistances and by that I mean by the sum of all resistances is the convection
resistance on the inside, all the possible conduction resistances that the heat flow is going
to face. So, if there are 3 walls there is going to be 3 different conduction resistances and
when it reaches the outer edge of the outer wall it experiences a convection once again.
And from Newton’s law of cooling we have seen how we can express the heat flow as
temperature difference divided by a resistance and the convective heat transfer resistance
1
is expressed as . And the area simply is going to be 2 rL that is the cross section
hA
that is the inner area of the inner cylinder.
So, you this way you can quite easily write what would be the form of the equation for
radial flow in systems and as I mentioned they are very common in many industrial
situations. Now what I would do next is something very interesting is normally what we
feel is that, if I am feeling cold I just wear a sweater which is nothing but an insulation
and this reduces the loss of heat from my body to the ambient and I do not feel cold
anymore.
So, in the way the purpose of the sweater or the jacket is to ensure that the heat flow
from my body to the ambient get slowered that is what is, that is what insulations do,
they will use the flow of heat from the hot object to the cold object, but is it ever possible
that by adding an insulation you are making more or higher flow of heat from the hot
object to the cold object.
So, this is counter intuitive, but it may happen and I am going to show how under what
conditions by adding insulations you simply increase the rate of flow of heat through that
insulation. So, that is a very interesting concept it is known as the critical thickness of
insulation. So, if your thickness of insulation is below a certain limit then by adding
insulation you increase heat flow, if it is above the critical insulation thickness by adding
insulation what you get is what we commonly expect that the heat flow rate reduces, but
67
the interesting part is that region in which your sizes are is your size is below that of
critical thickness of insulation so, by adding insulation you are increasing the loss of
heat.
So, you would first mathematically see what needs to happen for you to decide; what is
this critical thickness of insulation. And then we will talk about some of the practical
uses, practical situations in which you may expect to encounter critical thickness of
insulation and it has certain advantages as well and what is the scientific reason for such
thing to happen. So, let us look at the derivation once again. So, next wego to critical
thickness of insulation. Let us say I have a cylinder and a solid cylinder and an
insulation. So, this is ri and let us assume that the temperature here is maintained at T i,
temperature at the junction between the solid rod and the insulation. So, this is your
insulation and the insulation radius is r0 and on the outside it is exposed to a convection
environment with temperature and heat transfer coefficient as h and T∞.
So, it is a solid rod of radius r i, the junction temperature is at Ti, the insulation radius is r0
and it is exposed to h and T∞. So, as before from here I can write that the heat flow is
simply going to be
Ti  T
qr 
ro
ln
ri 1

2 kL h2 ro L
So, there is nothing new I have simply used this for a system in for this case. So, if this is
q then I am going to see is it possible to mathematically get at which point my heat flow
is going to be maximum. So, if I can do that then I will probably get an idea of what is
the concept of critical thickness of insulation.
68
So, starting with this
Ti  T
qr 
ro
ln
ri 1

2 kL h2 ro L
I am going to differentiate this and set this to be equal to 0. So, I am trying to find out is
there a radius r0 where r0 is the insulation thickness which would maximize the flow of
heat which is denoted by q.
 1 1 
2 L(Ti  T )   2
dq
0  kro hro 
2
dro  ro 
 ln r 1 
i
  
 k ro h 
 
 
1 1
 2
kro hro
k
ro 
h
69
This r0 is the critical thickness of the insulation which would maximize heat flow is
known as the critical insulation thickness.
So, this critical insulation thickness therefore, it tells you that if you have a solid cylinder
and you are putting insulations around it. So, this is your first level of insulation this is a
and then at the next instant you increase the thickness of insulation some more and you
are going to increase the thickness even more as long as your radius of this as you are
adding. So, if this let say this is r0 which is equal to k/h as long as you are below this
value you are less than k0/h your q will increase. So, if r0 is less than k/h, add insulation q
will increase, if r0 is greater than k/h, add insulation and q will decrease.
So, this is the concept of critical insulation thickness. So, there exists thickness when if
you add insulation the heat flow will increase, go beyond that and you add insulation and
the expected things will take place, that is the heat flow will decrease then why it should
happen this can only happen when the thickness is very small.
So, when the thickness is very small by adding another layer of insulation you are
increasing the resistance for flow of heat through that added layer of insulation; that
means, you are increasing the conduction resistance by making the layer thicker;
however, with putting the layer of insulation on the outside you are making more area
available for conduction because the area available for conduction is simply 2πr
multiplied by length, length is a constant. So, as r increases your area available for
convection increases. So, with increase in r the conduction resistance will increase ok.
So, the convective heat transfer is helped by adding insulation because of the additional
area conduction heat transfer is going to be reduced by adding the insulation. So, these
are 2 parallel mechanisms which compete with each other and for certain value of r as
we have been previously equal to k/h the result is you get the maximum heat transfer.
So, if you are below this value of k/h, increase the insulation thickness, increase the area
available for convection and that more than offsets the additional conduction resistance
that you put in the system that is what the concept of critical insulation thickness is. So,
let us just workout with those since r is equal to k/h, what are the typical values of k/h
and that would tell us something about when do you expect this concept of critical
insulation thickness to be when you are going to encounter such concepts.
70
So, let us write what are the typical values of k and the typical values of h and then we
will know what this is. So, k for insulation material is roughly about 0.0 3 watt per meter
Kelvin and the h the convection over here is mostly going to be the convection in air
which is the order of 10 watt per metre square per Kelvin. So, r c the critical thickness of
insulation is going to be about 3 millimeters.
So, your size has to be less than 3 millimeter for the concept for the occurrence of the
phenomena associated with critical insulation thickness, most of the normal conditions
you do not deal with 3 millimeter thin wire or 3 millimeter thin material. So, for most
cases your r is going to be more than r c and therefore, the concept of critical thickness of
insulation is not relevant.
So, concept of critical thickness you do not encounter the phenomenon associated with
critical insulation thickness since the length scale involved is only about 3 millimeters.
So, what we say is that I increase insulation and I reduce heat transfer, but one example I
can give you which would show the role critical thickness insulation thickness may play
for system.
So, when you think of very thin wires which conduct electricity, the diameter can be less
than 3 millimeters. So, when current passes the heat is generated the ohmic heat is the
joule, joule heat due to joule heating some amount of heat is generated and you want to
dissipate that ok, but at the same time we do not want to live wire without any electrical
insulation put on it.
So, what you do is on the thin wire you put a layer of insulation, but what you get is
something very interesting you not only make it safe for the wire to be safe since you
have put an electrical insulation this electrical insulation is going to act as a thermal
insulation as well, because the heat that is generated is going to dissipate to the
atmosphere by means of conduction through the electrical insulation and convection to
the outside.
Now, if your radius of the electricity carrying wire is less than 3 millimeter, then you are
enhancing the heat loss from the system by putting an electrical insulation on top of it.
So, you serve 2 purposes, you cover the electrical wire with an insulation and by doing,
so, you are increasing the heat transfer and therefore, the wire itself now can be at a
lower temperature which will probably prolong its life and will be less hazardous.
71
So, critical thickness of insulation are mostly relevant in very thin electrical wires where
you put an insulation on top of it, but for most of the practical systems since the
dimension that we deal with are and the dimensions are more than 3 millimeters you do
not get that in most of the applications, but concept wise it is very interesting that it is
difficult to it is counter intuitive when I say that by adding insulation I increase the flow
of heat.
So, that is something which in this class we have seen that it is a peculiar nature of the
phenomena which is a direct result of change in the heat transfer area with radius. As you
go outward the area increases and therefore the heat rate remains the same, but heat flux
will change.
So, this class we have seen what is radial system, what are the resistances associated with
radial systems, what is the peculiar phenomena of critical thickness of insulation. Then in
next class we will very quickly go through what is going to be the equivalent form for
spherical systems, and something which is known as the overall heat transfer coefficient.
And then we will go into systems in which there is going to be generation of heat at
steady state and what would be the form of temperature distribution for systems which
are generating heat on their own.
So, that would be the topic of the next class.
72
Heat Transfer
Lecture - 07
Tutorial Problem on Critical Insulation Thickness
We have already been introduced to the concept of critical insulation thickness. What we
have seen is an interesting result, which shows that when you add an insulation to a
conductor which is generating heat. Normally, we would expect that since we had
insulation, the heat flux would decrease, but in some special cases we looked at the
equation of heat transfer between a current carrying conductor that is generating heat,
with an insulation on top of it. And, we have seen that the heat flow from the conductor
to the ambient can be maximized for a specific thickness of the insulation, which is
called the critical insulation thickness. We understood that this anomaly can be explained
by the simultaneous increase in the heat transfer resistance when you apply the heat, as
well as the increase in the surface area available for convection such that the convection
resistance would decrease. So, this increase in the resistance to conduction through the
insulator, and the decrease in resistance by the availability of enlarged surface area for
convective heat transfer, these two we will provide a situation wherein the heat loss from
the system would be a maximum. And the thickness where it takes place is known as a
critical thickness of insulation. We have also seen that in most of the practical
applications, the concept of critical thickness of insulation does not appear. Since the
magnitude of critical thickness insulation thickness is less than the order of millimeters.
So, therefore, in practical systems we do not come across the critical thickness of
insulation, but in some special cases for example, in the case of a current carrying
conductor this could be of relevance.
We are going to solve problem in the in this class that would clarify our concept more
and so, it is going to be tutorial class based on the concept of critical thickness of
insulation, but before we move on to that there is another type of resistance to heat
transfer that we have not discussed so, far.
When a composite system consisting of two different materials are brought together, you
are never going to get a perfect contact in between the two surfaces, due to the
73
imperfections present i.e. the roughness; on the surfaces> Hene the two surfaces coming
onto contact with each other will never have a 100 percent point to point contact at every
location. There would exist an area of very low thickness which would be filled by air;
and, this air is going to give rise to a significant increase in the heat transfer resistance.
So, whenever you bring two surfaces together, due to the roughness present in them, the
actual heat transfer resistance is going to be much more than the individual conduction
resistances of the two solid blocks. So, this enhanced heat transfer resistance is called
sometimes called contact resistance.
If we look at the figure over here what you see is that the two materials A and B and an
at a molecular level they are; obviously, not very not going to be smooth. So, therefore,
there would be some amount of air, which is going to be entrapped in the spaces in
between these two solid surfaces.
And therefore, if you look at the temperature profile and if T A is the temperature on the
A side of the interface and T B is going to be the temperature on the B side of the
interface. Due to the presence of the low thermal conductivity air in between these two
materials A and B, there is going to be temperature jump across this interface. This
temperature jump is expressed in terms of a contact resistance, which as per our previous
understanding any resistance is cause, divided by effect. So, the cause is a temperature
74
difference T A minus T B, and the same heat travels through this. So, it is going to T A
minus T B by heat flux.
So, when you express the contact resistance in this form; obviously, its units are going to
be Kelvin per Watt. In practical systems most of the times you like to reduce the contact
resistance, some examples are when two surfaces are melted together. So, as to have very
good conductive flow of heat you do not want a thin layer of air existing in between the
two. On the other hand, if you are using material B as in that figure as an insulator then
having an air pocket in between A and B can be beneficial as it provides an additional
resistance. Therefore, depending on your final application the phenomenon of contact
resistance can be helpful, or you may not want to have a contact resistance. So, the
obvious question therefore, is what is going to be the value of contact resistance? The
contact resistance values are significantly more in many cases than the conduction
resistances of the two materials, which are brought together. So, one has to take into
account the contact resistance in most of the real situations. Now, let us say for some
reason you would like to reduce contact resistance. The obvious one is that, you would
like to make the two surfaces that you would like to combine as smooth as possible. So,
one way of reducing conduct resistance would be to make the surface very smooth. The
other option to reduce contact resistance, one may also apply pressure so, as to increase
the area of contact in between the two surfaces. So, the two options to reduce contact
resistances the resistances are to make the surface smooth and to apply pressure at the
contact level. The third option which is also used sometimes is to make sure, that you do
not have air in between the two surfaces which are brought into contact, you rather have
an interstitial fluid that has a higher thermal conductivity than that of air. You may like
to replace the air by a heavy oil with a high thermal conductivity. So, what you do is you
coat the two surfaces initially with a heavy oil with high thermal conductivity and then
bring the two surfaces together.
75
We now move to the problem that we are going to solve in this class on the concept of
critical thickness of insulation. So, the first problem that we are going to do in this,
tutorial problem it is about an electric wire through which 700 Ampere of current flows.
The diameter of the wire is given as a 5 millimeter, the resistance per unit length is
6 104  / m . The electric wire is in an atmosphere where the temperature is 30 oC ,

and the convective heat transfer coefficient from the outside of the wire to the to the
atmosphere is 25 to Watt per meter square per Kelvin.
The first part of the problem is if the cable is bare what is going to be its surface
temperature? So, you just have this cable where this temperature T infinity is 30 degree
centigrade and h is 25 Watt per meter square per Kelvin and you have a current which is
flowing through this.
So, at steady state some amount of current is going to be produced and this amount of
heat must be dissipated to the atmosphere by the convection process. So, in this case
your governing equation heat generated must be equal to the heat that is dissipated by a
convection process.
So, what is the heat generation in the cable? So, heat generation in the cable per unit
length of the cable is simply going to be I square R. So, it is 700 square times R where R
is the R is the resistance per unit length, which is provided in the problem as
76
6 104  / m . So, the unit of this should be Watt per meter. So, it is the heart generation
per unit length in the wire because of this much of current flowing through it.
And you would see that this is going to be 294 Watt per meter heat is generated in here,
this generated heat 294 must be equal to the h A by L where A is the area. So, it is pi D L
by L. So, this is the heat to which is lost by convection per unit length T of the surface of
the wire minus T infinity. So, your value of h is known, which is 25 T infinity known so,
you would be able to obtain T S of the bare cable surface to be equal to 778.6 degree
centigrade.
So, that is the answer to the first part which simply gives you what is going to be the
cable surface temperature, when there is nothing on it.
The part 2 of the problem says that a very thin insulation with a contact resistance of
0.02-meter square Kelvin per Watt is applied on the cable, what are the insulation and
cable surface temperature?
So, this I have the cable and then on top of the cable I have an insulation, this is a
cylindrical system I have this insulation. So, this one is my cable and here I have the
insulation. In between the insulation and the cable, and I have some contact resistance.
So, what is required is what is the temperature of the cable surface T S, that we have
obtained in our previous part of the problem is 778.6, and what is the insulation
77
temperature, what are the insulation and the cable surface temperature, what is the
insulation temperature; that means, what is going to be T i? So, these two we have to
evaluate now the same amount of current is flowing through the wire but, you have an
insulation on top of it right now.
And, the contact resistance is mentioned here. So, we need to consider the contact
resistance of 0.02-meter square Kelvin per Watt. So, the question is what is going to be
the value of T S and what is going to be the value of T i.
If you think of the situation as if two reservoirs are connected by a pipe. So, if you
maintain the level fixed and if you add a pipe in between them there would be definite
flow from this reservoir to this reservoir. Now, what you would like to do is you are
going to construct the pipe in between. So, you add more resistance to flow in between
the two reservoirs. So, the resistance is going to be more, but you would require the same
flow rate. So, with the resistance being higher, you would still like the same flow rate
then the only option available to you is to increase the height difference between the two
reservoirs. So, the potential must be increased if you would like to maintain the same
flow in between these two reservoirs with additional resistances in between them.
The same concept would be applicable here as well; the heat generated is the same. So,
at steady state the same heat has to go through the insulation through the contact
resistance first then through the insulation, and then the heat will be convected out to the
atmosphere, but since you have additional resistance the temperature on the wire surface
must be increased in order to have the same flow rate in between the two.
I am going to write the equation for the heat flow as the cause divided by the resistance.
So, the heat flow I square R remains the same the delta T is unknown to me T of the
cable surface minus T infinity. T infinity is known to me, but T of the cable surface is
not known to me divided by the sum of resistances. And, what are the resistances; we
need to see what resistances we need to consider finding what is the cable surface
temperature.
So, let us write the governing equation in here; where this q prime which would be the
same as in the previous problem divided by T of S minus T of infinity divided by 1 by h
pi d plus Rtc. So, this Rtc is the contact resistance. And, since everything is expressed in
terms of per unit length.
78
The key point is to note in the problem there is a very thin insulation. Since, the
insulation thickness is very small I am neglecting the conduction resistance provided by
the insulation. So, that is why I do not have the conduction resistance of the insulation
included in here. What I only have is the convection resistance at this point and the
contact resistance at this point. So, q double prime would simply be equal to a 294 as it
was before this T S is unknown to me, I do not know what is the cable surface
temperature now, but we realize based on a previous discussion that T S has to be much
more than 778, which we have obtained, for the first part of the problem T infinity is 30
and 1 by h is 25 pi times D the diameter of the cable is 5 into 10 to the power minus 3,
and the contact resistance point 0.02 is to be divided by pi times D. When you work out
this T S is going to be 1153 degree Centigrade; So, compared that with the value of T S,
which we have obtained in part 1 which was 778.6 degree Centigrade.
So, as I as I said before, the potential must be increased to have the same flow rate. So,
the temperature of the cable surface will now become 1153 to have the same amount of
heat flow. So, what is going to be the temperature of the insulation outer surface,
remember the insulation is very thin.
Since the insulation is very thin you simply can assume that it has the same diameter as
that of the wire. Now, if that is the case you are still going to dissipate 294 Watt per
meter of heat from the insulation. So, 294 Watt per meter must be equal to h times pi D,
where D is the diameter of the insulation, times T of the insulation minus T infinity.
This D, which is the insulation diameter, since it is very thin it is going it can be taken to
be equal to that of the wire. And once you do that your temperature of the outside of the
insulation surface will remain identical to it is value of part 1 that is 778.6.
Let us go to the third part of the problem, which is interesting.
79
Now, I am going to have a finite thickness of insulation. So, what thickness of insulation,
and now I am specifying the value of the k thermal conductivity as 0.5 Watt per meter
Kelvin will yield the lowest value of the maximum insulation temperature.
So, what I have now is I have the wire I have the contact resistance in and then I have the
insulation, this insulation thickness is now finite. So, here I have contact resistance, here
I have conduction resistance and here I have convection resistance.
Now, if you look at this figure it is obvious that the maximum insulation temperature no
matter whatever you do is going to be at this point, at the surface where it is connected
with the wire in presence of a contact resistance. So, we are trying to find out what is the
going to be the maximum value of this contact resistance. Let us call it as T prime
maximum value of this temperature. Now, what is going to the minimum value what is
going to be the lowest value of this maximum temperature?
Now, the maximum temperature again I use the concept of 2 reservoirs and somehow
instead of increasing the resistance of the pipe in between the 2 reservoirs and decreasing
the resistance in between the two. So, as I decrease the resistance in between the two in
order to have the same flow rate; now I can reduce the height difference to be provided.
If I increase the resistance I have to provide more as I have explained before, and if by
some means I can reduce the resistance of the connecting pipe then I can bring this
80
down. So, that a less potential difference in this case a less temperature difference has to
be provided.
So, how do you reduce the resistance of heat flow when you have an insulation on a
wire; obviously, the least resistance possible is going to be the at the situation where you
have the critical thickness of insulation. Because based on our previous discussion, we
know that when the critic insulation thickness reaches the critical value the resistance
will become minimum, and when the resistance is minimum then I can work with a
lower potential difference or a lower temperature.
So, coming back to the figure over here, when the thickness of the insulation is equal to
the critical insulation thickness, then this T prime is going to have a minimum value. So,
T prime will be minimum when the resistance to heat transfer is going to be minimum
and the resistance to heat transfer is going to be minimum when your Rc critic insulation
thickness is equal to the critical insulation thickness, which by definition is going to be
given by k by h, which is 0.5 Watt per meter Kelvin divided by 25 Watt per meter Kelvin
is equal to 0.0 2 meter.
So, what this tells is if I can provide the insulation thickness to be equal to 0.02 meter if
Rc is less than 0.02 meter then by increasing Rc I decrease the resistance and enhance
the flow of heat. So, when my flow is kept constant and if I keep on increasing Rc till I
reach the point of 0.02 meter, my resistance will keep on decreasing. And as resistance
will keep on decreasing my T prime will keep on decreasing.
So, if I say the T prime which I know is the maximum insulation temperature, this
maximum insulation temperature will have its lowest value, when Rc is equal to 0.02
meter.
So, if Rc is 0.02 meter, then the thickness of the insulation of insulation would be 0.0175
meter. So, when you provide a thickness of insulation of 0.01 75 meter; your temperature
at the inner surface of the insulation is going to be the least of all other situations and let
us find out what is it what it is going to be? So, what I have then is the wire, and then I
have the insulation around it and with a contact resistance in here ok. So, if I the heat
produced per unit length would still be 294, that is the say Watt per meter that the same
heat which is going through it and I am going to write, it between T insulation inside.
81
So, my 294 would simply be equal to T prime, that is the T insulation inside minus T
infinity the T infinity is still 30 and the R insulation is going to be 0.02 by 0.0025
divided by 2 pi 0.5 plus 1 by h is 25 pi times 0.04. So, this would be your this should be
the equation and T prime you are going to get as 318.2 degree Centigrade. So, this is
very interesting.
I will quickly go through what I have done here, I have written the equation the heat
transfer equation between this point, the insulation, and the atmosphere outside. So,
between this point and T infinity there are 2 resistances - one is the conduction resistance
and the second is the convection resistance. Since I am writing it from this point
outwards, the critical insulation thickness does not come into my equation.
The same amount of heat which is generated will pass through the wire, through the
contact, through the insulation, to the outside. So, this is what the heat is. So, this 294 in
all cases will remain the same. So, what is the, what is the potential it is T insulation
inside minus T infinity what are the resistances one is the conductive resistance of the
insulation and convective resistance at the outside of the insulation.
So, R insulation for a cylindrical system I use the formula and for convective resistance
it simply 1 by h A and since I am expressing it in per unit length it simply going to be 1
by h times pi D 0. This D 0 is now the new D 0, which is based on the critical thickness
of insulation. So, when you work out the number this T prime is going to be 318.2. So,
you can see is that you have achieved a significant reduction in the temperature by
having an insulation on the wire. So, that is the beauty of critical thickness of insulation
you have added insulation, but what you get because of it is a reduction in temperature.
82
Heat Transfer
Lecture - 08
Heat Source Systems
In the last tutorial problem, we have seen how the concept of critical insulation thickness
can give rise to interesting situations. Today, we are going to move to something
different; we are going to analyze systems in which we are going to have conduction in
the solid and maybe conduction and convection both present at the solid fluid boundary.
But most importantly these are cases in which some amount of heat is being generated in
the control volume.
So, one obvious example of that could be the joule heating, when current passes through
a conductor, we all know that some amount of heat is generated. So, how do we find
what is going to be the temperature profile? How would it look like and what are the
other additional observations that one can make and for those case situations including
the relevant boundary conditions.
So, we are first going to start with the planar system in which a plate of thickness twice L
is taken to be the control volume. The cross sectional area be let us say some constant A.
So, the volume of this plane element is going to be L times A and will have the
coordinates, the origin of the coordinate system would be at the mid plane. So, this my x
equal 0 is going to be at the mid plane as you can see in this figure.
83
So, this is my x equal 0 which is at the mid plane of wall whose thickness is twice L and
it has this area which is equal to A. So, the volume of that would simply then be equal to
twice L times A and this is a mid plane. So, you would to like to see how the temperature
would change when we have some amount of heat generation. The amount of heat
generation is simply the joule heating. So, it is I 2 Re . This is the amount of heat
produced in watts, but if you remember in the heat diffusion equation, the q is always
heat generated per unit volume. So, the q to be used in the conduction equation would
simplify
I 2 Re
q 
V
where, v is the volume of the system.
And it is at steady state, we would analyze this only for steady state and if it is at steady
state, the entire right hand side of the equation would be 0 and T is a function of x only.
So, T is not a function of y or of z. So, T is a function of x only. So, if that is the case,
then the heat diffusion equation would simply be simplified as (for steady state)
d 2T q
 0
dx 2 k
84
So, this would be my governing equation which would have to be solved with the help of
appropriate boundary conditions. So, let us say the boundary conditions are
T ( L)  TS 1
’
T ( L)  TS 2
So, these are the two boundary conditions. So, you can integrate this equation; it is a
simple equation to integrate and you would be able to obtain using these boundary
conditions, the final form of the heat diffusion equation; I am not deriving all the steps,
you can simply integrate this equation, use the boundary conditions, find out what are the
what are the integration constants.
 2  x 2  TS 2  TS1 x TS1  TS 2
qL
T ( x)  1    . 
2k  L2  2 L 2
I leave this for you to do this exercise of converting the governing differential equation
with the help of these boundary conditions to the final form of this.
So, heat flux if this is the temperature profile inside the solid between -L to +L; then, the
heat flux at any point can simply be obtained by finding out what is d T/ d x and multiply
it with k by invoking Fourier’s law. If it is so that
TS1  TS 2  TS
Then what you see is that the temperature distribution would be symmetric, would be
symmetric at x equal to 0 and the temperature distribution would simply be as
 2  x2 
qL
T ( x)  1    TS
2k  L2 
So, if the two temperatures at the two sides are kept constant and equivalent; then, the
temperature distribution would simply look like this and you can see its going to be a
parabolic distribution and for this special case where they are equal, this is what the
profile would look like. So, at this plane the temperature as you can see is going to be
maximum and heat is generated inside the system.
85
So, you have a q which is the heat generation per unit volume and this is how the
temperature profile would look like where this, the two temperatures are T S on both sides
of the wall. So, one can also. So, this is I think this is absolute thick layer that when the
two temperatures are equal you are going to get a parabolic distribution that is going to
be symmetric at x equals 0 plane.
So, we can extend this a bit further to find out what is going to be the maximum
temperature and from the profile that I have drawn for the temperature you would, you
can clearly identify that x equals 0 is going to be the mid plane is going to be the plane of
maximum temperature and you can obtain the expression for this temperature by simply
putting x equal to 0 in the equation.
There is one more part to it since the temperature is maximum at x equals 0, dT/dx at that
point would be 0. So, dT/dx at x equals 0 would be 0 which simply tells you that for all
for all practical purposes the mid plane acts as an adiabatic plane.
In physical terms if this is a mid plane, then no heat can cross the mid plane. It is some
sort of a peak of the temperature and you have valleys on both sides, the heat flows in the
direction starting at the mid plane and nothing crosses the mid plane. So, the mid plane
can be treated, can be called as an adiabatic surface. So, let us see how it would look
like.
So, once you have the expression for the temperature in here, the maximum temperature.
86
The maximum temperature would simply be would takes place at x = 0 and this would
be
 2
qL
T0   TS
2k
And the temperature distribution can also be expressed in terms of a dimensionless

temperature,
2
T ( x )  T0  x 
 
TS  T0 L
So, as I said that
dT
x 0
dx
Since, this is 0 the mid plane is an adiabatic surface. So, this is one part of the problem
and let us there are there are certain modifications, we can suggest to this governing
equation which is I write it again as
d 2T q
 0
dx 2 k
87
Let us say that you have x = L, this is the surface and at one point, one side of it is
insulated. The other is maintained at a fixed temperature. Let us call this as T S. So, this is
the plane at which it is kept insulated and this is the plane at which the temperature is
maintained at a constant value.
So, how would the profile look like? And we can see that what I have drawn in this
figure is simply the previous figure; if you look at the previous figure, I have only drawn
half of it. And since, it is an adiabatic surface dT/dx is 0 in this case. So, this is how the
profile would look like.
So, this is the case. So, in order to bring parity between what I have done before and this
one. So, this is at x equals 0 and at x equals L. So, the profile of temperature would
probably look something like this. So, this is the same when you compare with this part.
So, this is an inverted parabola and this is half of the parabola, where the boundary
condition at this point would simply be dT/dx = 0. So, that is the case when you have the
one side of it is insulated. So, you are going to get a parabolic distribution. So, it is half
parabola at this location.
So, in some case, so this is another situation. So, in some cases, it is it would be shown it
would be known that TS is unknown, but T∞ is known. So, what I mean by that is this is
the same surface, where this is x equals L and what you would have here is this is my x
equals 0 and this is at -L.
Now, on this surface this is the temperature T S; TS is known, but T∞ it is in contact with a
fluid whose temperature at a point far from the wall is known. So, how would the
temperature, how do you express TS in terms of T∞ and then, solve the equation such that
the temperature profile inside this can be obtained not in terms of TS, but in terms of T∞.
So, in order to that only thing that you change the governing equation will still be the
same. The only thing that will change is I am going to take this as my control surface and
whatever heat that comes to the control surface by conduction must be taken out by
convection.
So, this is the condition which is going to be maintained at x equals L. So, at x = L, the
conductive heat flow upto the surface must be equal to the heat that is lost from the
88
surface to the surrounding fluid by convection and using Newton’s law of of cooling this
heat flux is simply going to
dT
k xL  h(TS  T )
dx
So, what I have here is nothing new, I am simply using this as the control surface and
equating conduction and convection and I am writing Newton’s law on one side which
takes care of the solid side of the interface and Newton’s law, Fourier’s law and
Newton’s law which is for convection which is which is on the liquid side of the
interface equating these two and knowing that T(x) as we have done in the previous
problem.
 2  x2 
qL
T ( x)  1    TS
2k  L2 
And from the temperature profile, we also understand that we also know that
dT qx


dx k
dT
So,  k xL  qL
  h(TS  T )
dx
89
So, one more time what I have done is from the temperature profile that was obtained in
the previous part, I have I have calculated what is dT/dx. I have multiplied dT/dx with -k.
dT
So,  k is the heat which is coming to the surface which is equal to the heat that is
dx
convected out which is the heat flux which is convected out which by Newton’s law of
cooling which be would be a h(TS  T ) .
Now, of all the heat generated inside the plane wall, half of it is going to travel in this
direction; the other half is going to travel in the reverse direction. So, if the two sides, if
the boundary conditions at two sides are maintained identical. So, half the heat which is
generated in this must be equal to qL
 . So, the heat generated would be total heat
generated with qL
 times area.
Q  qLA

So, that is half of the heat generated in the entire wall which travels in this direction. So,
 is; so, this is the heat generated per unit volume multiplied by L. So, this is simply
qL
the flux. So, by equating conductive heat flux with the convective heat flux and the
amount of heat generation, I can write this expression for this expression. So, what this
simply says that
qL

TS  T 
h
So, simple problem of heat generation inside the plane wall has given as the, we can
make the following observations. If the two end temperatures, two temperatures of the
two sides are identical; then we are going to have as parabolic distribution, symmetric
parabola, symmetric inverted parabola in which describes the temperature distribution.
At the mid plane, since dT/dx is 0 that can be that it can be said that the mid plane is an
adiabatic surface. There are you can make one of the surfaces of the plane wall, one of
the surfaces you can make it perfectly insulated. The moment you make it perfectly
insulated, the relevant boundary condition at that condition would be dT/dx equal to 0
and when you look at that system, it is simply the half of the plane wall that we have
considered previously.
90
So, in half of the plane wall, dT/dx was 0 at x equals 0. In the case when one side is
perfectly insulated, dT/dx is 0 at x equals 0. So, the case of the temperature distribution
for the wall one side of which is insulated would look like an half of an inverted
parabola.
And then, we have also seen that the temperature of the surface may not be known, but
the temperature of the fluid which is contact with the surface is known. So, by the use of
equality of conductive heat flux and convective heat flux at the solid fluid interface, I can
express the temperature of the solid in terms of the temperature of the fluid.
So, that known temperature can then be substituted into the expression for temperature
which contain TS. So, TS can then be substituted by use of the concept of equality of
conduction and convection and through the use of T∞ instead of TS; known T∞ instead of
TS. So, that is the more or less case for the plane wall. Now, let us see how it would look
like if you have a cylindrical wall if you consider a cylindrical wall or a cylindrical
surface, cylindrical volume.
So we, looking at the case heat source systems.
And the heat source system that we are going to do is cylinder with a heat source and
whenever I talk about this heat source for this problem as well as for the case of plane
wall, these heat sources are uniformly distributed. So, all these cases are uniformly
91
distributed. So, this is a simple case of let us say through an wire some current is flowing
and you are going to generate some amount of heat which is going to be simply I 2 R and
this I 2 R is a heat generated divided by the volume would be the q .
So, the heat diffusion equation for a cylindrical system and its steady state and will
assume that temperature is a function of r only. Temperature is not functions of θ as well
as functions of z. So, there is no axial temperature distribution, there is no temperature
distribution with respect to θ, but the temperature is only a function of r. So, it is a one
dimensional steady state conduction in a cylindrical system with a heat source. So, that is
what we are going to analyze. So, the heat diffusion equation for a cylindrical system
undergoing one dimensional steady state conduction.
1 d  dT  q
r   0
r dr  dr  k
All these equations would be provided to you. So, you do not need to memorize any of
these equations, but you should at least know how to cancel out the terms which are not
relevant for a specific case. So, given the heat diffusion equation for cylindrical and
spherical or spherical systems you should be able to use your understanding of the
problem to get rid of the terms which are not relevant in a specific situation.
So, since its one dimensional steady state conduction with heat generation. So, I only
have T as a function of r. So, all the T, all the variation in T with respect to θ and with
q
respect to z, those terms are neglected and I am going to keep in there and the right
k
hand side of the heat diffusion equation which takes into account the transient effects
that term can also be neglected.
So, I have a compact form in which the one of the term represents variation of T with r
and the second term represents the heat generation in the system. So, I am exactly
solving the same problem as in the case of a plane wall, but this time the geometry is
different. So, that is why the form of the governing equation is different. The boundary
conditions will always remain the same. So, let us see how this would look like when we
solve this equation.
So, what you get from here is
92
dT q
r   r 2  C1
dr 4k
q 2
Integrating, T (r )   r  C1 ln r  C2 , 0  r  R
4k
where, R is the radius of this wire.
Since, your temperature cannot be indeterminate at r equal 0 which essentially tells you
that C1 has to be equal to 0. So, that is obvious because since T(r) has to be finite at r = 0.
So, in order for T(r) to be finite at r = 0; C1 will obviously, has to be equal to 0. So,
q 2
T (r )   r  C2
4k
Now, what is the second boundary condition? So, this is BC, boundary condition 1 and
boundary condition 2, let us assume that T at r equals capital R is the wall temperature of
the wall which is known to us.
T (r  R )  TW  known
If the wall temperature of the wire is known to r is known to us which is at r = R. Then,

through the use of this boundary condition one would be able to find out.
93
 2
qR
T ( R )  TW    C2
4k
 2
qR
Or, C2  TW 
4k
So, when you put this back in this equation, what you get is
2
q  2 r 
qR
T  TW  (R2  r 2 )  1 
   
4k 4k   R  
You can also express it in dimensionless form and the dimensionless form of this would
be
2
T  TW r
 1  
T0  TW R
So, this would be the dimensionless form of the temperature profile and this is the
dimensional form of the temperature profile. So, in here T 0 as I said is the centerline
temperature that is,
 2
qR
T0  T (r  0)   TW
4k
Now, there is certain similarity of this equation or this equation with some of the things
that we have done in fluid mechanics. So, let us think of a fluid mechanics problem in
which you have a tube through which a liquid is made to flow either due to gravity or
due to gravity plus pressure difference imposed on the system.
Now, if you remember your fluid mechanics, this is what the velocity profile would look
like. The velocity due to no slip condition would be 0 at the walls; whereas, it is going to
be maximum at r = 0. So, the profile in that case would exactly look something like this,
where if you, I would request you to go back and see what is the velocity profile in such
cases and there you would see that the velocity profile is velocity maximum multiplied
  r 2 
by 1     .
  R  
94
So, it is a parabolic distribution of velocity; where, V max is the maximum velocity which
happens at r = 0. So, when you think of the case of heat transfer, conductive heat transfer
in a cylindrical system with heat generation, there also you see that the temperature
  r 2 
distribution can be expressed in parabolic form with something in front of 1     ,
  R  
which if you look at the expression of the maximum temperature. So, that is similar to
maximum temperature.
So, fundamentally in terms of the governing equation, in terms of the boundary

conditions the similarity between heat and momentum transfer they are apparent. The
similarities are obviously there. The same type of distribution, it is a parabolic form
parabolic form of temperature distribution and the maximum lies at r = 0 be it velocity or
be it temperature. So, r = 0, for the case of cylindrical system is the adiabatic surface at
which your dT/dr would be 0 and which will tell you what is the maximum temperature.
In the same way we have done before if this T W, the temperature of the surface is not
known to us; then but the temperature of the adjoining fluid which is known to us; then, I
can use again conduction convection equality at r = R and relate TW with T∞. So, how do
I relate that? At r = R, the amount of heat which reaches this point, which reaches at r =
R must be by conduction. Since, I am talking about conductive heat transfer in this case
and on the other side of the interface, it is going to be taken up by convection. So, if I
have a situation in which this is the surface, all the heat that are coming in by conduction,
the temperature of this point is TW and the temperature some point over here is T∞.
dT
So,  k 2 RL rR  h(2 RL)(TW  T )
dr
So, this is the way through which one can obtain the unknown TW in terms of the known
T∞. So, exactly the same way that we have done for the case of plane wall, the unknown
temperature of the wire surface can be replaced by the known temperature of the fluid
surrounding it.
So, whatever we have done so far, let me conclude. We have analyzed the systems where
heat generation is possible, heat generation is taking place for a cylindrical system and
for a planar system and we have seen that the plane of symmetry is x equals 0 at for the
95
case of plane walls and the line the central line at r equals 0 that is going to be the point
where the heat where the temperature is going to be maximum. And we also know that
how to relate the temperature of the wall or temperature of the surface to the temperature
of the surrounding fluid through the use of Newton’s law of cooling and Fourier’s law of
conduction taking that as the control surface and importantly, the profile in all cases will
be parabolic.
And in the next class is going to be a tutorial one in which we are going to solve the
problem which will have a plane wall with heat generation and sandwhich and there is
another wall of a different materials being adjacent to it and we are going to have a fluid
with the surrounding fluid. So, you have material 1, which has heat generation; material
2 which does not have a heat generation and then, on the other side of material 2, I have
convection.
So once we solve this problem, I think all the concepts, all the special characteristics of
the temperature distribution that we have discussed so far will be more clear to you. So,
the next class is going to be tutorial one.
96
Heat Transfer
Lecture - 09
Tutorial Problems of Heat Generating Systems
So, this should be a tutorial class based on the concept that we have covered in the
previous class, which is the temperature profiles and special considerations for systems,
which have some heat generation capability in them. We have seen that the temperature
profile in most cases should be a parabolic, it is going to be a symmetric parabola, if both
sides of the plane wall are maintained at a constant temperature. The mid plane at x = 0
is going to be the plane with maximum temperature, this implies that plane is going to be
an adiabatic one.
So, the first case I will draw your attention to is the figure that I have over here. So, this a
plane system of material A of thickness 50 millimeter, and then on the other side of A I
have another material B, whose thickness is 20 millimeters. One side of A is insulated;
that means, at x is equal 0 you have insulation. The temperature at this point is denoted
by T 0, the temperature at the interface between A and B is T 1. And on the other side of
B, I have a liquid which is at a temperature of T infinity far from the wall and the liquid
is flowing along the plane along this B. So, its temperature is 30 degree centigrade and
97
this flow creates a convective heat transfer coefficient of 1000 Watt per meter square per
Kelvin.
So, I hope the picture is clear to you. A wall which generates heat, sandwiched between
an insulated surface, and another wall of different thermal conductivity, but without any
heat generation. Wall B on the other side is exposed to a convection environment where
the temperature and the h are provided.
The thicknesses are different and if you look at the material A, it has a thermal
conductivity of 75 and the amount of heat generation in this case is going to be 1.5 106
Watt per meter cube, and when we consider the wall B it has no heat generation
capability; however, the thermal conductivity of this is twice that of A. So, the first part
is sketching the temperature profile under steady state and we would also assume that T
is a function of x only. So, if that is the sketch we first need to find out what does the
temperature profile look like.
So, it is a sketch, and hence no numerical values are required. You just need to see how
the temperature profile would look like. So, let us first consider the wall A on one side is
insulated. And on the other side, the junction temperature between A and B is known and
we are calling it is T 1.
So, if on one side of a plane wall the temperature is known, the other side is perfectly
insulated, and in this wall if heat is generated then from our previous study we know that
the temperature profile is going to be that of an inverted parabola. In fact, it is going to
be half of an inverted parabola. So, how it would look like in the material A where there
is heat generation. So, the temperature profile in this would probably look something like
half of a parabola with the apex at x = 0. So, this is T 0 and this is T 1 when I come to
material B, the T 1 is known the T 2 which is the interface temperature between the
material and the fluid if T 2 is known and since q dot is 0. So, in this specific case q dot
is not equal to 0 and in this case q dot is equal to 0.
So, what you are going to get is it is a case of 2 temperatures known with no heat
d 2T d 2T
generation. So, your governing equation would simply be  0 . So, if  0 then
dx 2 dx 2
the temperature profile as we have seen before is going to be linear.
98
So, this is what the temperature profile should look like at material B. And obviously,
there is continuity between, so T 1 on the A side of the interface must be equal to T on
the B side of the interface, but from this point onwards it is going to be linear. And of
course, the value of thermal conductivity would also tell us the gradient of this line.
Higher the value of thermal conductivity, lower would be the slope of this line. So, if you
did not have q dot present in material A you only have K A and K B to consider, then the
slope of the temperature profile for material A would be more as compared to K B.
So, a system of higher thermal conductivity simply tells you that you require only a
small delta T to conduct the same amount of heat. So, looking at the temperature profile
at steady state between different surfaces, you would be able to say which one is of
higher thermal conductivity and which is of lower thermal conductivity.
So, the points so far is that it is going to be half of an inverted parabola and in B, it is
going to be linear. Then the question still remains is how the temperature T 2 would,
which is at the interface between the material B and the liquid change from T 2 to T
infinity.
Again, I reiterate that there would be the formation of a boundary layer close to the
surface of the solid, which is in contact with the liquid. So, due to the formation of the
boundary layer the temperature would change from T 2 to that of T infinity in a region,
which is quite thin and that is the concept of thermal boundary layer the same way we
had the momentum boundary layer.
So, this change in temperature from T 2 to T infinity will be very sharp over a thin region
and then it is going to asymptotically reach the value of T 2. So, the profile would look
something like this and this is your T infinity. So, this is the boundary layer in which the
temperature is going to change, and you would be able to see, how would the
temperature profile look like in this in this case.
So, the salient features of the temperature profile are that it is an inverted parabola in A
in B it is simply going to be a straight line and from B to liquid it is a sharp change due
to the presence of the thermal boundary layer. Sharp change and then gradually it will go
to T infinity.
99
The part B of this problem tells us that we need to find out T 0 of the insulated surface
and T 2 of the cooled surface. So, this is A, B and you have a liquid in here. So, this is
your T 0 and this is your T 2 all the numbers are provided to you. So, you must find out
what is going to be the T 0 of the insulated surface and T 2 of the cooled surface. Now, it
is probably easier if you find out T 2 first and then go for T 0. Why so, because you have
some amount of heat which is generated in A, but there is no heat which is generated in
B. So, when steady state will reach then what you would see is that all the heat which is
generated in A has to travel to the right of the figure all the way to the liquid, because
one side of A, at x equals 0 is perfectly insulated.
So, no heat can cross from A to the outside, and all the heat has to travel inward through
B to the liquid. So, the heat flow from A through B up to the interface of the solid and
liquid is going to be by conduction. Beyond that fluid solid interface, the heat is going to
be transported by convection.
So, if you find can find out what is the total amount of heat generated in the system? And
equate that with h A delta T, which is Newton’s law of cooling and this delta T is simply
going to be T S minus T infinity. So, that is going to be the equation which one should
use in order to find, what is the heat, what is the unknown temperature T 2, because T
infinity is known to me. So, h times area times T 2 minus T infinity, this is the heat
which is moving in this direction by Newton’s law of cooling must be equal to the
100
amount of heat generated per unit volume times L times and this length obviously will
have to be LA. So, this is L A and what we have here is L B. So, this is the equality of
heat generation inside and all the heat being convected out at that point.
So, A will cancel and what you would get is that T 2 is equal to T infinity plus q dot L A
by h, which when you put the values at 30 plus 1.5 106 into 0.05, because this is 50
millimeters divided by h is 1000 Watt per meter square Kelvin. So, your T 2 would turn
out to be 105 degree Centigrade. So, that is going to be the temperature at this point.
Now comes the second part; how to evaluate the temperature T 0 over here? Now, if you
write the governing equation for A. So, for material A the governing equation would be

d 2T q
  0 at steady state.
dx 2 k
So, this would give you dT   qx  C1 and the use of first boundary conditions. So, C 1
dx kA
is an integration constant the first boundary condition B C 1, tells me something about
what is going to be the value of slope at x equal 0 the value of slope d T / dx at x equals
0. So, d T / dx at x equal 0 must be 0 since this is an insulated surface since x equal 0 is
an insulated surface. So, this simply tells me that C 1 is going to be equal to 0. So, one of
the integration constants is automatically taken care of. I integrate this once again. So,

2
the temperature which is a function of x would simply be: T  x    qx  C2 , where C 2
2k A
is the other integration constant. So, T at x equals L A that means at this point the

2
temperature we know that it is T 1 and this should would give me C2  T1  qL
2k A
Therefore, the temperature profile for material A would be:
 
qL2 qx 2
T  x   T1  
2k A 2k A
101
So, this would be the temperature profile inside A in terms of T 1, which is the
temperature over here, but my aim is to obtain T 0, I have to evaluate what is T 0? So, I
get an expression of T in terms of T 1, which is yet to be evaluated. So, I am simply
going to first put x equal to 0 in this expression and the moment I put x is equal to 0, then
T x is simply going to is equals T 0.
So, T at x equals 0 from this previous expression T at x equals 0 I call it as T 0 which

would be q dot by twice k A times L A square plus T 1. So, the previous equation
substituting x is equals 0 my T at x equals 0 is T 0 x equals 0. So, this is what the form is
going to be, but the problem remains I do not know what is T 1? How do I find out T 1,
T 1 is unknown to me?
So, in order to find T 1 and I draw the profile once I draw picture once again. So, this is
A, B and you have T infinity over here and you have some q dot in here. So, the same
heat will flow in this direction and also to the fluid. So, the temperature over here is T 1.
So, the potential at this point is T 1, the potential at this point is T 2 and the potential at
this point is T infinity.
So, if I draw something like a thermal circuit, what I get is the temperature at this point is
T 1. Over here is T 2 and over here is T 3, in between T 1 and T 2, I have a conduction
resistance, in between T 2 and T infinity I have a convection resistance. So, what are
102
these are R double prime conduction and double prime simply tells it is per unit area
basis and this is R double prime convection.
So, for example, R double prime convection the R convection is 1 by h A. So, R double
prime convection is simply going to be 1 by h. So, that is going to be the convective
resistance in between these 2 points. And what is the current or equivalent of current,
which is heat flowing through it that must be all the heat which is being generated in here
that is q times A times L A.
So, that is the heat which is flowing through these 2 resistances and since I have
expressed these 2 resistances in terms of per unit area basis. So, I drop the area from
here. So, this is the heat generation per unit length which flows from T 1 to T 3 through
T 2. So, the same amount of heat flows to each one of them. And these 2 resistances are
obviously in series.
So, this heat which flows through this is q dot times L A is equal to the potential
difference, T 1 minus T infinity divided by the sum of all resistances, which R double
T1  T
prime conduction plus, R double prime convection. So, this is going to be LB  1 .
kB h
So, I plug in the values of L B, k B, h, T infinity, and q dot, L A, etcetera what you are
going to get is T 1 as 115 oC . So, this is going to be the junction temperature. And what
we have done over here is we have an expression for T 0 in terms of T 1 where q k and L
A are known to me. So, once I substitute T 1 in this expression q dot L A square by twice
k A plus T 1, when I substitute this in here putting in the values of q dot etcetera what
you would get is T 0 as 140 oC .
So, this problem is a nice example of how the concept of heat generation in a system can
be applied when there is heat generation, when there is no heat generation, when you
have conduction, when you have conduction and convection both present in the system
and so on.
I will just give you 1 more problem with answer for you to practice on based on the
concept of again heat generation infrared systems. So, the problem that we are I am
103
going to discuss, I am going to provide to you as a homework as a problem for you to do
at home is I have a system, which is x this is at x equals 0 and this is x equals L this side
is perfectly insulated.
Ans the thermal conductivity of this is k and you have some sort of a radiation which is
falling on this surface. And the radiation is going to be absorbed by the material based on
how close it is to the surface. So obviously, when you are on the surface you get the
maximum absorption and as you move far away from this surface which is exposed to
radiation, the amount that is absorbed is going to be less.
So, the microwave radiation causes some sort of a heat generation, which is obvious. So,
this heat generation is going to be a function of x where this is the x direction, and it is
 x
going to be q0  1   , where q 0 is just a constant and let us assume that this wall that
 L
the x equal 0 is maintained at a temperature T equals T 0. So, the first part of the
question is deriving the differential equation, for the temperature profile, and the second
is solve with boundary conditions, to obtain the profile. So, that is the problem you have
radiation, which is incident on the wall. The absorption has resulted in a volumetric
source of heat generation, the functional form of which is given over here this q 0 is a
constant and of course, it is it is going to be a decreasing function of x.
104
So, starting with the problem first find out what is the differential equation and then use
the appropriate boundary conditions to solve for the temperature profile. So, I suppose all
of you would be able to see, that the governing equation is simply going to be

d 2T q
  0 . The only difference with the previous problem the ones that we have
dx 2 k
discussed so, far is this q dot is not a constant, but it is a function of x. That is the only
thing which you have to appreciate for solving this problem.
So, the variation of q dot which is heat generated per unit volume with x the functional
 x
form is provided as q0  1   . So, if you look over here then the governing equation
 L
 x
q0 1  
2
would simply be d T L . And what are the boundary conditions? The
2
  0
dx k
boundary conditions will remain the same that at x = 0 T is a temperature which is
known to us and at x equals L; that means, at this point d T / dx is 0 since it is an
insulated wall.
So, the governing equation is obtained in the same way with the understanding that this q
0 is a function of x now and these are the 2 boundary conditions. So, you solve it on your
own I am simply going to give you the final form for you to check. So, the temperature
q0 L2  x x 2 x 3 
profile here would be T  T0      .This is going to be the temperature
2k  L L2 3L3 
profile, which you should get for such a case.
I will give you one more problem very quickly again you have to solve it on your own.
The third tutorial problem is you have a cylindrical heat source at the middle.
105
So, this is a heat source at the middle in the one that I have drawn over here. So, this is a
tube, on the inside of the tube I have a coating, and this is just a radiant heater and here I
have a coating which is to be dried. So, this coating is to be cured by heating. And this is
the cylindrical heater, which is placed over here the purpose of this heater is to provide a
constant heat flux q1ll , to this. The thickness is r 1 from here to here and from this point
to this point is r 2. So, this is the tube this r 2 minus r 1 provides the tube wall thickness.
The first part is finding the temperature distribution in the tube wall. So, how does the
temperature change in the tube wall? It is obviously a cylindrical system. So, if the tube
is cylindrical and the second part is if r 1 is 25 millimeters, r 2 is 38 millimeters, what is
the power required? What is the power required per unit length of the tube? And to
maintain T S 1 to be equals 150 degree Centigrade and T S 2 equals 25 degree
Centigrade, where the thermal conductivity is 10 Watt per meter per Kelvin.
So, how do you solve this? How do you find the temperature distribution in the tube
wall? Now when you see that tube wall when you think of the system, there is no heat
which is being generated in the tube wall. The only thing that this tube wall experiences
is some amount of heat is coming in at this inner surface and it is going to get absorbed
at the inner surface, then the same heat is going to flow through the tube to the other
side.
106
So, the radiant heater at the centre provides the energy, which in incident on the inner
surface gets absorbed, maintains the temperature required to cure the coating at 150
degree, then that heat has to travel through the solid material of the tube and go to the
other side.
So, this temperature is provided as 150, this temperature is provided at 25, the thermal
conductivity is known in this case the only thing you have to find out is what is the
amount of heat, that the heater must produce per unit length of the heater or per unit
length of the tube in order to maintain the condition.
So, any such problem should start with identifying what is the boundary what is the
governing equation and what are the boundary conditions? So, I am going to write the
governing equation for the tube when I am going to write the governing equation for the
1  dT 
tube it is going to be r   0 . So, this is the governing equation, and what you
r  dr 
need to identify is that I did not add any heat generation. Because, no heat is generated,
whereas heat is absorbed in one of the boundaries.
So, it should come as a boundary condition not in the governing equation. Since it is only
at the boundary I am going to use it as a boundary condition. So, what my boundary
conditions are that q double prime the heat flux is equal to q 1 double prime, which is
this known one, q 1 double prime at r equals r 1 which is at this location. So, this can
then be expressed as minus k d T / dr at r equals r 1 is q 1 double prime. This is your
boundary condition one, this is your boundary condition one. And what is the boundary
condition BC 2, the boundary condition 2 simply tells you that the temperature at the
other end of the surface is known to you, that is T is equal to T S 2 at r = r 2. So, this is
my second boundary condition, which gives me what is the temperature at this point?
q1ll r1  r2 
When you solve this you should get T r  Ts 2   ln  . So, this is the temperature
k  r
profile, which you would get. So, this is T which is a function of r. So, if I at r equals r 1
this expression then becomes T S 1, because at r equals r 1 T is T S 1 minus T S 2 as q 1
double prime r 1 by k ln r 2 by r 1. So, this is your solution of temperature profile this
part and in here I identify the T is T r.
107
So, at r equals r 1 T r becomes equal to T S 1; T S 1 minus T S 2 this part will remain
unchanged except r is going to be equal to r 1, all quantities in here are known except q 1
double prime. So, this equation would provide you a value of q 1 double prime to be q 1
double prime is the flux. So, I have to find out what is the power required per unit length.
So, power required per unit length would be twice pi r 1 q 1 double prime and the
numerical value would be 18.7 6 kilo Watt per meter.
So, this is the value that you should get by working out the value of q 1 double prime, q
1 double prime is the flux, but you are asked to calculate what is the power required per
unit length? So, that is why you multiply it with twice pi r 1. And therefore, you get this
to be the value of power to be supplied per unit length of the tube.
108
Heat Transfer
Lecture – 10
Transient Conduction
In this class we are going to start a new topic. So, far we have restricted our analysis to
steady state situations; that means, when the temperature can be a function of location,
so, for example, x y and z it was not a function of time. So, those all steady state
equations, relations, concepts that we have developed so far needs modification when we
try to apply it for situations where the steady state has not been reached and therefore,
the temperature can in addition to the space coordinates can also be functions of time.
Now the transient process is prevalent in nature wherever you see, wherever you change
one of the boundary conditions allow across let us see a solid object whenever you
change the boundary conditions the solid material has to go through a transient state
before a new steady state is reached. So, you quench a steel ball or when you expose a
rod at a different temperature essentially what you are doing is, you are changing the
boundary condition of the ball or that of the rod.
In both cases you change T ∞, the temperature of the surrounding medium which would
impart a change in temperature inside the solid object and it would take some time before
it reaches a new steady state value, when it reaches a steady state value from that point
onwards the temperature would be function of space coordinate so, once again,
incorporation of this time variation of temperature makes the process makes the equation
more complicated to solve. So, the heat diffusion equation that we have used so, far the
right hand side of the heat diffusion equation contains the term 1/α, α being the thermal
diffusivity times δT/ δt.
So, this right hand side is no longer 0. So, in our governing equation not only the
relevant terms of the left hand side or the heat generation term are to be kept on the left
hand side, the right hand side also now contains temperature as a function of time which
makes the process, which makes the solution more difficult to obtain, more complicated
and you require additional initial conditions on temperature. So, you at least need one
condition on temperature with respect to time to start solving that process. In many
109
transit processes we have to resort to numerical solution of the partial differential
equation.
But in some special case you could see that the situation can be resolved by making
certain assumptions. So, these assumptions fortunately are valid in many situations and
once you make this assumption the entire transient heat transfer problem can be resolved
in a compact analytical form of solution is possible. So, we would like to see what would
be that simplifying assumption that one can make in order to make the process, in order
to make the governing equation much more simpler for which analytical solution is
possible.
So, therefore, we understand 2 things that transient conduction is something which is

there everywhere. We encounter transient conduction in many everyday processes,
industrial processes situations and this arises mostly by the change of one of the
boundary conditions and secondly, the variation of temperature with time makes the
problem difficult to solve. So, first we are going to look at the simplifying assumptions
and before we do the simplifying assumptions, I am going to show conceptually what
can happen when a steel ball or a copper ball of high thermal conductivity is quenched in
cold water. So, a copper ball of 200 ºC is suddenly put in a container which contains
water at room temperature.
So, the temperature of the copper ball will decrease with time as the copper ball is going
to lose its energy, lose its heat and thereby reducing its temperature. So, if I think
conceptually what is going to happen inside the copper ball, its temperature is decreasing
with time, but since the thermal conductivity of copper is very high it is probably safe to
assume that the variation in temperature inside the copper ball is not going to be
significant, that is the assumption based on which all our subsequent analysis would be
based would depend.
So, a copper ball of this big size is put in water and this was initially at a temperature of
200 and so, when it comes in contact with water at 30 ºC, the sides are going to cool very
fast. So, that side temperature, the temperature at the periphery at the outer surface of the
copper will decrease rapidly with time, but what is going to happen to the temperature let
us say at the centre of the copper ball. Can we assume that the entire copper ball during
its transient process remains spatially isothermal that means there is no change in its
110
temperature with respect to x, y, z or since its a spherical system r, θ and Φ it is a
function only of temperature.
If I can assume that then the entire governing equation can be reduced to a simple
ordinary differential equation and this kind of an assumption where the temperature of
the solid is assumed to be invariant with position is known as the lumped capacitance
model of transient conduction and that is what we are going to look at first, but let me
first draw the diagram of what happens when you quench a steel ball in a liquid at a
different temperature.
So, let us see that we have this as the container which contains some liquid and you have
an object whose temperature initially is Ti, the initial temperature and this is the situation
at when t is less than 0. So, your temperature everywhere inside the solid object is going
to be equal to Ti and then you drop it in the liquid. So, it comes in here; however, in this
case the temperature of the solid object is going to be a function of time. So, for any t
greater than or equal to 0 the temperature is going to be a function of time.
And the temperature of this liquid is let us say it is T ∞ and temperature T∞ is less than Ti.
So, this is essentially a cooling of the solid object T i is greater than T∞. You put it in
water and you see that the temperature changes with time. So, what is the lumped
capacitance model, the lumped capacitance model or lumped capacitance assumption
111
tells you that the temperature of the solid is spatially uniform at any instant of the
transient process.
So, this one is important that the temperature is uniform, now what do you think when
this is possible when you are going to have a solid object whose temperature is going to
remain constant with respect to x, y, z or r, θ, Φ, but it is going to decrease since it is in
contact with a fluid whose temperature T∞ is less than the temperature of the object.
Now, common sense tells us that this is only possible theoretically, only possible when
the solid object has very high thermal conductivity or to be 100 percent accurate the
thermal conductivity of the solid must be infinite. So, if the thermal conductivity of the
solid is infinite then there is no resistance of conduction of heat inside the solid because
we understand that, let us say for a plane wall the thermal resistance to conduction is L/k
so, this k is the thermal conductivity.
So, when the thermal conductivity is very large; then L/k can be approximated to be
equal to 0 and therefore, the resistance to conduction inside the solid is 0. So, if you do
not have any resistance to conduction or very little resistance to conduction inside the
solid then at any instant of time the temperature of the solid object would remain
uniform, at the next instant the temperature is going to be less as compared to the
previous instant; however, it will still be uniform temperature throughout the solid
object.
So, this is what lumped capacitance assumption or lumped capacitance model for
transient conduction is all about that you assume the temperature of the object will not
change with space, but it will change with time. There are certain assumptions, certain
advantages or major advantage of this lumped capacitance assumption is that in your
heat diffusion equation it is heat in minus heat out plus or minus heat generated is equal
to the energy rate of energy storage. So
E IN  E OUT  E G  E STORED
So, all these dots refer to the time rate of change of energy coming into the system which
is 0 for the case that we are dealing with where we have just dropped a solid object in a
liquid. So, there is no energy and the solid object temperature is higher than that of the
112
liquid so no heat enters the solid. So, E IN is 0. You will have E OUT definitely because
the solid object is losing heat. Some energy is going out of the solid object and let us also
assume that no energy is generated inside the solid object. So, the third term on the left
hand side of the equation E IN  E OUT  E G , the only remaining term would be  E OUT , the
time rate of energy lost by the solid object.
And on the right hand side it is rate of energy stored so, as a result of the solid object
losing some energy due to maybe convection in this case the total energy content of the
solid will keep on decreasing. So, the non zero term of this equation would be  E OUT is
equal to the time rate of energy stored in the system. So, that is the advantage, that is the
form of the equation. Now we are going to write this and see how the lumped
capacitance assumption makes the entire process simpler so, let us proceed with this.
So, we understand that as I mentioned
E IN  E OUT  E G  E STORED
So, this ( E IN ) is 0, this ( E G ) is 0. So,
dT
 hAs (T  T )  VC
dt
Mark what I have done over here and simply see T, I have used a normal ordinary
differential.
So, what I am seeing is that T is a function of time only and T is not a function of for this
case r, θ, Φ. So, if this is the case then if I write this equation at any instant this equation
would be true, so this temperature is going to be a function of time, but it is not going to
be function of r, θ, Φ. So, that is why I get a simple equation to deal with for a much
more complicated process of transient conduction.
So, the beauty of lumped capacitance is essentially giving us a compact ordinary

differential equation for a very complex process and we would later on see what is the
quantitative criteria that must be satisfied before we can invoke lumped capacitance
assumption, but before that let us see what this leads to,
113
Let,   T  T
where the theta is essentially a temperature difference defined as the temperature of the
solid object which is a function of time minus T∞.
VC d
So,  
hAs dt
Upon this solid state straight forward equation which upon integration in the using the
initial condition that at t  0, T (0)  Ti , as it was the initial temperature of the solid. So,
this is the initial condition and if I integrate this equation with the help of this initial
condition what I am going to get is that.
t
VC  d
   dt
hAs i  0
so, i  Ti  T which is simply the initial temperature minus the temperature of the fluid
surrounding it. So, this can be integrated. So, this would give me the variation of theta
the temperature as a function of time.
VC i
ln  t
hAs 
114
Also the thermo physical properties of interest would be ρ, the density, C, the heat
capacity, the geometric parameters are going to be volume of the solid, surface area of
the solid and the operational parameter would be the convective heat transfer coefficient
which among other things depend on the conditions imposed by an external agency on
the solid.
So, let us see if you have a stirrer present in the fluid where you are going to put the
solid, then the speed of the stirrer would dictate what is the value of this h is going to be.
So, externally imposed conditions in many cases dictate what would be the value of the
convective heat transfer coefficient. So, in that expression there are some thermo
physical properties, some geometrical properties and some operational parameters
together with those parameters or the clubbing of those parameters would give us the
variation in temperature of a solid object as a function of time. We could see that the
temperature is going to change in a logarithmic fashion.
So, it is a very compact relation which you have obtained which has significant industrial
usage because you can quickly estimate what is going to be the temperature of an object
when it is quenched with something else. So, how the temperature of a solid object is
going to decrease, how much time would it require for the solid to come to some
desirable temperature such that it is ready for further processing. So, in many industry
the formation of a specific material goes through discrete steps and some of the steps
may require high temperature, some of the steps required curing the solid objects with a
given temperature difference over a certain amount of time. So, and then repeating the
process with different variables. So, with this variation in the property, with this
variation in the time may give rise to difference in properties.
So, the desirable properties of a solid can be fine tuned by exposing it to a temperature
environment and letting it heat or cool in a specific fashion. So, it is very important to
know how long would it take for the solid to attain certain temperature and an easy way
to estimate that is to use the lumped capacitance model which would exactly tell you that
after how much time you were going to reach this temperature or in order to obtain your
if you leave for x amount of time what is going to be its temperature.
115
So, these are important calculations which are relevant in many industrial applications.
So, let us prove it a little further and see what additional information we can obtain from
this. So,
 T  T   hAs  
  exp    t
i Ti  T   VC  
Now, will you see hAs you can recall that the heat transport is given by convection is
given by hAs times δt. So, this hAs is nothing, but an inverse of a resistance. So, if I try
to express this whenever you have exponential some constant times t. So, inverse of this
entire group is nothing, but a thermal constant of the system. So, if I write it in that form,
thermal time constant let us call it as  t
1
So,  t  .VC  Rt .Ct
hAs
this VC is the thermal capacitance of the system. So, they are expressed as Rt .Ct
where Rt is the thermal convective resistance and Ct is the lumped thermal capacitance of
the solid. So, this Rt is nothing, but the thermal convective resistance of the solid and C t
is the lumped thermal capacitance of the solid. Now once I obtain the temperature the
next thing that is of importance to me is to find out how much heat is lost by the solid in
a given amount of time.
So, the expression that I have obtained gives me the temperature variation with time.
What I am more interested or I am also interested to know is how much of heat is lost or
gained by the solid in this transient process so, which I am going to find out next.
So, what I do then is in order to obtain the amount of heat lost or the total energy
transfer.
116
t t
So, Q   qdt  hAs   dt

0 0
This Q is the amount of energy lost or gained by the physical process of convection so,
this is heat by convection. So, this is the heat that is lost or gained by the solid object due
to convection.
So, if you look at the previous slide,   T  T , and we have a complete expression
t t
  t  
Q   qdt  hAs   dt   VC   i 1  exp   
0 0    t 
I would bring in the previous slide one more time so, that there is no confusion in here,
heat lost by convection must be equal to hAT and   T  T .
So, this hA(T  T ) gives you the amount of heat lost at any given instant when you
integrate it over time, this gives you the total amount of heat lost during the period from
0 to t. So, this θ is to be substituted using the expression of it as obtained from the
lumped capacitance model. So, what is that, this θ multiplied by exponential the entire
thing, the integration of that. So, once you put the expression of theta from this in here
and integrate what you are going to get is this form of the expression.
117
t t
  t  
Q   qdt  hAs   dt   VC   i 1  exp   
0 0    t 
So, the total heat lost over a period t can be related to the thermo physical properties, the
thermal time constant of the process, the thermal capacitance, the initial temperature
difference and finally, with time.
So, this Q, the total energy transfer must be equal to ESt such that Q  ESt . So, this Q
can be substituted in here such that you can obtain the total energy stored in the system.
Now for quenching; that means, when you are cooling the solid object this Q is positive
and for heating; that means, when you are dropping the solid object in a liquid whose
temperature is more than that of the solid then Q is negative so, these 2 which you have
to keep that in mind.
So, for quenching; obviously, the total energy content of the solid will decrease and
when you are heating the solid; that means, when you are dropping the solid in a liquid
whose temperature is more than that of the solid, then the solid is going to gain energy.
So, this way the convention to follow is that for quenching it is going to be positive the
amount of heat transfer and for heating it is going to be negative since it moves in the
other direction. So, one of the prerequisites of the lumped capacitance model or the use
of the lumped capacitance model is that the temperature of the object is going to be
spatially uniform.
So, let us try to see physically what would have to happen in order for the solid to behave
in that specific way or in another words let us try to put a quantitative criterion which
would enable us to use the lumped capacitance model in many practical situations and in
order to do that I am going to look at a solid object, let us say a solid wall which is
initially at a constant temperature and suddenly the solid wall is exposed to a
environment where the temperature is less than that.
So, these solid objects let us say it is at 100 ºC and suddenly it brought into this room
where the temperature is 25 ºC. So, from both the walls the solid is going to lose heat and
its temperature is going to decrease. Now how would that profile look like in realistically
when you consider not only the thermal conductivity of the material, but also the
geometric parameter.
118
And the geometric parameter of interest is obviously you can see is the thickness of the
solid. So, instead of this being the solid if we are considering this to be the solid only the
cover of the mobile is the solid then it is pretty thin as compared to the overall one.
So, the chances of the solid remaining at a constant temperature at any given instant of
time would be more for this cover as compared to the whole item itself. So, as we reduce
the thickness of the solid it is likely that the constancy of temperature with position will
most probably be valid. So, as this thing becomes thicker and thicker the centre
temperature is going to be substantially different from the temperature of the side. So, it
is not only the thermal conductivity that plays a critical role in determining whether
lumped capacitance is valid, it is also the geometric dimensions of the object that play a
role as well.
So, let us see pictorially how would it look like and come up with a number. So, what we
are looking at is validity of lumped capacitance LC method. Let us say we have a wall
whose temperature at this end is TS1 and here it is T∞ and TS1 is definitely greater than T∞
and you also have a flow of the fluid which maintains a convective heat transfer
coefficient of h. So, and this is let us say T S2. So, the amount of heat which gets
transferred through this is going to go out of this if I consider this one.
119
So, this is by conduction and this heat will go out by convection. So, at steady state the
amount of heat which comes to this point must be equal to the amount of heat which
goes out of it. So,
kA
(TS 1  TS 2 )  hA(TS 2  T )
L
where L is the thickness of the solid must be equal to the convective heat loss from this
side of the interface. So,
TS 1  TS 2 hL L / kA RCOND
  
TS 2  T k 1/ hA RCONV
So, whether or not TS1 is going to be equal to or close to TS2 would depend on the ratio of
these 2. So, TS1 would be equal to TS2, if this part the conductive resistance is very small
is 0 or the convective resistance is very large. So, whether or not T S1 would be equal to
TS2 which is the fundamental assumption of lumped capacitance it would depend on the
resistance of these two and the resistance of these two is simply hL/k and if you look at
hL/k, it is a dimensionless number and it is known as Biot number
hL
Bi 
k
The biot number therefore denotes that whether or not which one is important,
conductive resistance or convective resistance and from this expression we see that the
convective resistance has to be large as compared to conductive resistance in order for
TS1 to become equal or close to T S2. So, if R convection is large as compared to R
conduction then the value of biot number must be quite small as compared to one for
lumped capacitance to be valid.
So, the simple expressions simple thought experiment tells you that what is going to be
the requirement of the conduction and convective resistances and everything is expressed
in a compact dimensionless number hL/k, what is h, it is a convective resistance, L, the
thickness over which the temperature changes from T S1 to TS2. So, it is the geometric
dimension of the solid across which the maximum temperature drop is taking place and
what is k, k is the thermal conductivity of it.
120
So, this biot number has to be less than significantly smaller than 1 to ensure that T S1 is
equal to or close to TS2, when can that happen either h has to be small or L has to be
small, if you look at the definition of biot number here in order for biot number to be less
than 1 either h has to be small, L has to be small or k has to be large. So, under this
condition the temperature profile inside the solid will more likely the very small or no
variation in the temperature and then it is going to go to T ∞. So, this is what it is going to
look like. So, most of the resistance is going to be the convective resistance and since the
temperature here does not vary, the conduction resistance is almost going to be very very
small.
So, the biot the prerequisite for using the lumped capacitance model is to ensure is to
calculate what is the value of biot number first and see if it is less than 1, the common
criteria for assuming lumped capacitance is biot number has to be less than 0.1 for using
the lumped capacitance model. So, one can use biot number, LC is valid for biot number
less than 0.1. This is the criteria generally used to ascertain whether or not lumped
capacitance is valid in a system and if this is the case then it looks something the profile
would look something like this and we can use the first equation that I have written over
here
 E OUT = E STORED
dT
Or,  hAs (T  T )  VC
dt
So, now since temperature is not a function of r, θ and Φ, at any instant this equation is
true which can then be integrated to obtain the total amount of heat which is lost or
gained by the subject and not only that, what is the temperature of the object as a
function of all these parameters and time. So, it is the thermal time constant and which is
this is inverse of the thermal time constant. So, a very convenient method for describing
natural convection for describing the transient conduction is assuming it is to be space
wise isothermal, but varying with time.
The moment I make space wise isothermal assumption the equation is an ordinary
differential equation, which can be integrated with the help of an initial condition and the
space wise isothermal condition can be used quantitatively when a newly defined
dimensionless group known as Biot number is found to be less than 0.1. That check must
121
first be done to ensure the validity of lumped capacitance, once that is done it is a simple
equation to solve which would give you important information about how an object
would cool which would with time has significant industrial applications in various
sectors. So, we are going to go slightly different in the next class and solve a tutorial
problem on transient heat transfer.
122
Heat Transfer
Lecture - 11
Lumped Capacitance (Contd.) and Tutorial Problem
So, we have seen how Lumped Capacitance model can be used to analyze the transient
heat transfer process. And therein we have also seen the importance of Biot number to
decide whether the temperature inside the solid can remain space-wise isothermal. And if
it is so, then there it is a compact expression, which can be obtained starting with the
  
fundamental energy equation E in  E out  E stored . So, we get a compact expression, with a
time constant. And the time constant would contain h which is the convective heat
transfer coefficient, A which is the surface area the thermal conductivity, the heat
capacity etc.
So, we understood that Biot number has to be less than 1, in order for the lumped
capacitance model to be valid. And we have put a number which has been observed
experimentally to correctly reflect the transient process that is if Biot number is less than
0.1, then the lumped capacitance is valid. So, quantitatively one has to make sure that the
Biot number for the specific system under consideration is less than 0.1, before one can
proceed with the equation derived with lumped capacitance model. Now what happens
when Biot number is greater than 0.1?
In that case the conduction resistance can be significant or as compared to the convection
resistance and this simplification is not valid and temperature inside the object is going
to be a function of its location not only of the time and therefore, it is more complicated
and we would quickly see towards in the next class, how to solve problems in which case
the Biot number is greater than 0.1. But there is one more thing which we must ascertain
before we move on is the Biot number contains h which is the convective heat transfer
coefficient. There are several relations or correlations available by which one would be
able to calculate what is the heat transfer coefficient depending on which fluid is being
used, what is the temperature range of operation, what is the velocity with which the
fluid is moving over the solid or the fluid is made to move over the solid to obtain the
value of the convective heat transfer coefficient.
123
The value of k is a thermo physical property, which is fixed the moment you fix the
solid, which is to be subjected to transient conduction. So, h and k are of known to us
now then what about Lc? So, we call Lc as characteristic length, if you look at the
definition L is nothing but V by As where V is the volume of the system and As is the
surface area which is available.
124
Lc is defined as V by A S in when V is the volume of the solid; A S is the surface area.

So, for a plane wall of thickness 2 L with convection from both sides, Lc is equal to L.
So, it is half the thickness. If you think of a long cylinder if you use this formula L c
r0
equal to V by A S you would see that L c is going to be equal to where r0 is radius of
2
the cylinder and for a sphere again; if you use this formula, your L c is simply going to
r0
be . So, the expression therefore, gives you what length scale to be used to evaluate
3
the Biot number. It is either half the thickness where both the halves of a plane wall are
r0
experiencing conviction, for a long cylinder it is and for a sphere it is going to be
2
r0
equal to .
3
However, for the conservative estimate of Biot number L c is defined as the length scale
over which the maximum temperature change is taking place. So, what is L c once again
it is the length scale over which the maximum temperature change is taking place. So,
when you think of a cylinder, in case a long cylinder which is exposed to a convection
environment, the maximum temperature is going to be at the center line and the
125
minimum temperature is going to be on the surface of the cylinder; that means, over a
distance equal to r.
So, even though the theory suggests that L c equals V by A S, it suggests that the L c
r0
would be equals , in order to be remain conservative on the estimate of Biot number,
2
L c is taken to be equal to r0 ; the distance over which you would expect the maximum
change in temperature.
Similarly, for a sphere; the maximum change in temperature is going to be over the
radius of the sphere. So, if you consider the centre point and if you considered the
surface, that is the length scale over which maximum change in temperature is expected.
r0
So, instead of taking L c to be equal to be for a sphere, the L c is taken to be equal to
3
r. So, what we are doing here is that, we are making the Biot number calculation and its
applicability for the lumped capacitance approximation more stringent.
r0
So, if Biot number with L c equals is less than 0.1, it may happen that Biot number
3
with L c equals r is not less than 0.1, but to ensure the validity and accuracy of lumped
capacitance model, it is always the thickness, the length scale over which maximum
temperature drop is taking place is taken to be the length scale to be used for the
calculation of Biot number. Now we are going to do a little bit of more jugglery with the
expression of Q or with expression of temperature and to see what the role of
temperature is and if we could express the role of temperature front moving inside a
solid, in a more coherent form. So, we know how the temperature is changing, but let us
say a solid is suddenly put in a liquid with a lower temperature.
So, how fast the solid is losing heat or how fast the temperature front is progressing
inside the solid, we would probably be able to see a slightly more insight into the process
if you rearrange the terms a little bit more. So, what I have then is my starting point,
which is the heat which is getting into the system with Lc equals V by As the expression
126
 T  T   hAs   hAs hAs hL k t
is   exp     t  . This can be expressed as  c 
i Ti  T   VC   VC VC k  C L2c
where L c is the characteristic length.
So, I have not done anything except that in order to bring Biot number in here I put an L
c in the numerator. So, k by rho c is nothing, but the thermal diffusivity alpha. So, k by
rho c is substituted by alpha, where alpha is as I have told you before is the thermal
diffusivity; and therefore, h A S t by rho v c this whole thing is equal to
 T  T
  exp   Bi  Fo  . This entire thing is dimensionless, and F o is the short form
i Ti  T
t
of Fourier number this Fourier number which is defined as .
L2c
t
So, this Fourier number compares the penetration depth of the temperature front
L2c
inside the solid. And therefore, together with Biot and Fourier number one would be able
to obtain the variation of temperature inside a solid undergoing transient conduction and
it somewhat decouples two effects. One is the Biot number which is the resistance to
conduction by resistance to convection and the second is Fourier number, which tells us
127
something about the time dependent part something about how a length-scale can be
compared with the penetration depth of the temperature front. It gives you some idea
about how the entire temperature change in transient process can be expressed in terms
of physically explainable phenomena, which can be expressed in terms of
thermophysical properties; the dimension of the system that is L c and conditions
imposed by external agency on the outside of the solid, which would dictate what is
going to be the value of the convection coefficient.
Now let us see sample problem on the use of lumped capacitance model in the transient
conduction. So, let us look at a problem now. The problem that we are looking at is that
of a thermocouple.
So, we have a thermocouple, which has leads. So, this thermocouple let us assume that it
is spherical in shape. And this thermocouple is kept in a gas stream to measure what is
the temperature of this. So, a thermocouple junction which is approximated as this sphere
is used for temperature measurement in a gas stream. The properties which are given are:
h is 400 Watt per meter square Kelvin, the thermal conductivity of the material of the
thermocouple is 20 Watt per meter Kelvin, the thermal capacity is 400 Joule per kg
Kelvin that of the thermocouple material. The density is 8500 kg per meter cube. The
first part of the problem is to find the thermocouple junction diameter.
128
So, essentially the diameter of this needed for the thermo couple I write it as T c to have
a time constant of one second ok. So, the problem here is that, I am not sure if the
lumped capacitance model can be used in this case since the junction diameter is not
known to me I cannot calculate Biot number, which is h L c by k. So, L c is unknown.
I cannot calculate that Biot number but let us assume that the Biot number is less than
point one, and the lumped capacitance approximation is valid; and then calculate the
diameter of the thermocouple junction and then re-calculate what is the Biot number to
check whether the Biot number that you obtain by this method is less than 0.1.
So, in this way we will be able to obtain what is the diameter as well as check the
validity of lumped capacitance model so as to ensure that the method that we have
adopted is correct. So, if I assume that lumped capacitance model is valid, then I know
exactly what the expression for the thermal constant is. So, we put the thermal constant
expression and equate that to one second to obtain what is the diameter. So, that is what I
am going to do now.
So, the surface area is simply going to be pi D square the volume is going to be pi D
VC
cube by 6. So,  t  is the thermal time constant of the process as we have derived
hAs
before. Substituting relevant values, the diameter of the thermocouple junction would be
equal to 7.06 104 meters. So, that is a diameter once I have the diameter, I must check
r0
what is Biot number. So, with L c to be equal to what we get is, Biot number to be
3
2.35 104 , which is definitely less than 0.1.
So, if it is less than point one then whatever we have decided for the validity of lumped
capacitance is to be correct since the value of Biot number that we calculate comes out to
be less than 0.1. Now I could have used the more conservative calculation procedure to
see what the Biot number would turn out to be. So, that means, instead of the
r0
characteristic length to be at , which follows directly from V by A S, I should
3
probably use the length scale over which you get the maximum temperature drop. (Refer
129
Time: 25:10) The maximum temperature change takes place between the centre and the
periphery; that means, over the radius r which in this case over the radius r0.
So, let us see what the value of Biot number would be if I use the conservative estimate
r0
of the characteristic length L c. So, if I use this L c to be equals not , but simply by r
3
this should give to a Biot number value equal to 7.05 104 . So, in both cases you can
see that the value of Biot number is less than 0.1, and once that happens you know that L
c is valid. So, once L c is valid. So, whatever you have done over here must is correct.
The second part of the problem tells us to find out if the junction is at 25 C and is placed
in a gas stream, how long will it take.
So, the problem is the junction is at a temperature of 25 degree Centigrade. In any

measurement device that time required to sense the new value should be very less; i.e.
the measuring device should be very fast. Such that once you place the thermocouple in
the gas stream, it should very quickly find out what is the temperature of the gas stream.
So, the response of a measurement device plays an important role in deciding which
thermocouple to choose and therefore, it’s a parameter which detects the performance.
So, the thermocouple initially is at 25 suddenly you place it to measure the gas
temperature, which is 200 degree Centigrade and you need to find out how long would it
130
take for the thermocouple to reach very close to the temperature of the gas because if we
look at the expression it is an exponential expression. So, the temperature of the
thermocouple is going to reach the gas stream temperature asymptotically.
But we would like to have within a plus/minus 1-degree accuracy, how long would it
take for the thermocouple to record a temperature of 199 degree Centigrade, which is
within one degree of the actual temperature. And in the previous part we have already
established that the lumped capacitance approximation is valid and therefore, we now
can proceed to find out what is the time required. Since the lumped capacitance is valid,
 T  T
  exp   Bi  Fo  . Once you put the values of this rho D C h T i etcetera, the
i Ti  T
time you are going to get is 5.2 seconds, which is approximately equal to if you see the
expression of the thermal constant. So, its roughly about 5 times the thermal constant of
the system. So, from the time you place the place the thermocouple junction into the gas
stream, it would require roughly about 5 seconds for it to correctly measure the
temperature of the stream with an accuracy of plus/minus 1 degree Centigrade.
So, we know that we have an approximation. So, we would be aware of the value that
you have obtained is subject to the following assumptions. Forgetting about lumped
capacitance which you have shown to be valid, the other thing that we have neglected
here is the radiative losses. So, if you think of the thermocouple junction, its temperature
is rising and its going towards 200 degree Centigrade, at that temperature it is this
possible that some amount of heat would be lost from the thermocouple bead apart from
the convective losses, but we did not consider the radiative losses from the thermocouple
junction; first assumption. The second assumption the thermocouple junction is not
floating in the air, there are some leads which are made of solid materials that places the
thermocouple in the flowing stream of the gases. These leads of the thermocouple which
as you can see from this figure, they are made of solids and they will also conduct some
heat away from the thermocouple.
So, not only we have not considered the radiative losses from the thermocouple bead we
also did not take into account the conductive heat flow through the lead of the
thermocouple. So, the value that we have obtained as 5.2 seconds, it is subjected to these
two assumptions. But as the first estimate it is safe to see that the response time of the
131
thermocouple to reach the final temperature is roughly going to be about 5 seconds. So,
this shows you one of the clear examples of the utility of transient heat transfer and the
way the entire transient process can be simplified by assuming it as a lumped capacitance
system.
132
Heat Transfer
Lecture - 12
Transient Heat Conduction
So, far we have been studying Transient Conduction and several assumptions associated
with it, we have seen that incorporating lumped capacitance model, that is if we can
assume that the temperature is going to be spacewise isothermal, as the object is getting
cold; that means, temperature inside the solid will remain in variant which position, it is
a function of time only, in that case significant simplification of the entire problem is
possible. And we are going to get a closed form solution, for the variation of temperature
with time.
We have also seen that this assumption can only be used, when the Biot number which
gives us some idea about the convection, the conduction resistance and convection
resistance. It is the ratio of convection resistance to conduction resistance, when the Biot
number is small ideally when it is less than 0.1, then we can safely use the lumped
capacitance model in transient conduction.
So, the significance of Biot number and that of another dimensionless number which is
called Fourier number is clear to all of us. So, what I would do in this class is I am going
to solve another problem on transient conduction, where you would see that the lumped
capacitance model is valid, but it is slightly different than the one that we have attempted
so far. So, far what we have done is when I write the energy equation, it is rate of energy
in minus rate of energy out plus rate of any generation of energy inside the control
volume, must be equal to the rate of energy stored in the control volume.
The formula, the methodology, the modeling which we have adopted so far assumes that
there is no energy which comes in it is let us say it is a Quenching problem. So, there is
no energy in to the control volume. Energy goes out principally by a convective process
and, there is also no energy generated in the system. And therefore, the energy going out
is simply going to be equal to the rate at which the energy is stored, inside or stored or is
depleted inside the control volume. And the equations which we have obtained
133
essentially assumes that only two terms, one term on the left hand side and the term on
the right hand side, they will remain in the physical description of the situation.
But what happens in the case, where we have energy generation as well in the control
volume. So, whenever you are going to use any formula be careful to identify whether,
the basis on which the physical system based on which the formula has been derived,
whether or not its identical with the problem that you are solving. So, the first tutorial
problem of today is when an electric current passes through an electric wire and the
temperature obviously will rise as a result of the heat generation and, let us also assume
that the wire is exposed to a convection environment.
So, as the temperature of the wire rises, it is going to lose heat by convection and, part of
the heat which is generated in the wire is going to be used to raise the energy stored, the
capacity energy stored in the system. So, rate of energy in minus rate of energy out, plus
rate of energy generation must be equal to the rate of energy stored. So, this way the
inclusion of the generation term makes the problem different than those which we have
already discussed and solved in this.
So, the first problem today and I hope you can read this which it says is that a long wire
of diameter D equals 1 millimeter. So, you have a diameter a wire of diameter 1
millimeter is submerged in an oil bath. So, it has oil everywhere and the oil bath is at a
temperature of 25 ºC.
134
And the wire has electrical resistance per unit length. So, I call it as Re to be 0.01 ohm
per meter and the current which flows through this, that I is equal to 100 Ampere. The oil
bath creates a convective environment, where value of h is equal to 500 watt per meter
square Kelvin. So, there are three parts, or few parts of the problem, the first one is what
is the steady state temperature of the wire. The second part is from the time the current is
applied, how long does it take for the wire to reach a temperature, which is within 1 ºC of
the steady state value.
So, these are the some properties of the wire are given for example, ρ, C and k thermal
conductivity all these are provided. But essentially the problem is that of heat transfer by
convection to the oil heat generation, because of the flow of current unit. And since it is a
transient process it would require some amount of heat, it would also denote that some
amount of heat is going to be stored in it.
So, the equation if I write that
rate of heat in – rate of heat out ± rate of heat generation = rate of energy stored in the
system.
So, the first part of the problem, we can write that at steady state, when the steady state is
there, the right hand side of this equation is 0 because, at steady state it is the temperature
is not going to be a function of time. Since the temperature it could still be a function of
135
position, but at steady state there would be no change in the energy stored inside the
system. There is no heat which comes into the system. So, out should be equal to
generation. So, at steady state the amount of heat generated because of the resistance of
the wire and the current which is flowing through it, must be equal to the energy which is
going out.
And this out is by a convective process, we know the value of h and so, on. So, the
energy generated must be equal to energy out. So, if I is a current which is flowing
through it and R is the resistance of the entire wire, then it should
I 2 R  h DL(Tw  T )
So, at steady state this equality must hold, in order to obtain what is the steady state
temperature. Now, it is a simple equation which can be or where you can substitute the
values and, get the final value of the temperature of the wire, which is going to be a
constant and which will not change with location.
And therefore, the value of Tw can be obtained from this relation. One point to note here
is that this value of the resistance per unit length is provided in the problem. So, if I
divide the left hand side this side by Re/L, then I have all the way very all the numbers
with me so,
(100) 2 *0.01
Tw   25  88.7C
 *1*103 *500
So, this should be about 88.7 ºC. So, the steady state temperature of the wire is 88.7 ºC.
The next part of the problem tells us that we need to find out, what time does it take for
the wire to reach the temperature, which is within 1 ºC of this steady state value. So,
starting at 25 ºC, how much time has to, after how much time the temperature will be 1
ºC of the steady state value; that means, what is the time required for the wire to reach a
temperature of 87.7 ºC.
One thing we have to keep in mind here is that, we are using the lumped capacitance
assumption. Whenever we use lumped capacity assumption, or lumped capacitance
model is imperative that we find out whatever we have done, whether or not it is going to
136
be valid. So, the way to check the validity of lumped capacitance model is to find out
what is the value of Biot number.
So, once the value of the Biot number is calculated and, if it is less than 0.1, then
whatever we have done so far or whatever we are going to do by incorporating the
lumped capacitance model would be correct. So, the next step or the one of the major
steps of solving transient conduction problem, using lumped capacitance is to show that
your assumption is valid by calculating; what is the value of the Biot number. So, that is
what we should do next.
So, this value of the Biot number, I can calculate right now which would be
hro 500*1*103
Bi    0.005  0.1
k 20* 2
I have used ro ideally this length scale is simply volume by area. So, if it is volume by
area for a cylindrical one it is not going to be r o, it is going to be most likely ro/2, but as I
have mentioned in order to be on the conservative side of calculations, on the
conservative side the length scale is always chosen as the dimension across which we are
getting the maximum temperature drop. So, for the case of a cylindrical system, you
would expect the maximum drop in temperature in the solid between the centre line and
that of the surface.
137
So, to be really to be conservative about the number the magnitude of the Biot number
we will use always ro for the case of cylinders and, for the case of spheres as well
whereas, for the case of a plane wall, we will use half the thickness of the wall, when it is
heated from both sides, uniformly from both sides heated or cooled. So, that is the length
scale that we should use for calculating the value of the Biot number. So, we put the
values in here what you would get is the value of h is 500, the value of r o can be
substituted in here, in the value of so, this is by 2 and the value of k which would give
you the value of Biot number to be really really small definitely less than 0.1.
And therefore, the major this thing is LC lumped capacitance as a model is valid in here.
So, once we know that lumped capacitance is valid, then we should be able to able to
calculate, what is the time required for this thing to reach within 1 º of the steady state
temperature. And we know that the steady state temperature is 88.7 ºC.
So, essentially we have to find out how long does it take for the wire to reach a
temperature of 87.7 ºC, which is 1 ºC of the steady state temperature. In order to do that I
am going to first write physically what is happening in this case.
So, in this case the governing equation takes the form
E g  E out  E st
138
So, this E g is the additional term, which we are getting for this specific problem. Now,
what is E g if we express everything in terms of energy only.
E g  I 2 Re L ,
So, this is Re which is a resistance per unit length
E out  h DL(T  T )
This T is the time variation time varying temperature of the system, but since Biot
number is less than 0.1, the T inside the wire will remain a constant. So, it is not going to
vary in space, but it is going to vary in time.
d  D2 dT
E st   VCT    LCp
dt 4 dt
 D 2 dT
So, I 2 Re L  h DL(T  T )   Cp L
4 dt
And the length can be cancelled from all sides. So, what we end up with is the variation
of temperature with time.
dT h D 4 4 I 2 Re
 (T  T ) 
dt  Cp D 2  Cp D 2

4h 4 I 2 Re
C1  
 CpD  Cp D 2
So, that is straight forward so, whatever we got is a straight forward expression in terms
of the variation of temperature with time which is so, this is the governing equation,
which we need to solve and it requires 1 boundary condition. So, but before that I think
we can also define that   (T  T ) , it is just that I am defining it in this way. So, my
governing equation would therefore be
d
   C1  C2 
dt
139
And at the boundary condition would be, t  0,   i , which is the initial temperature
difference between the wire and the surrounding medium. So, this equation has to be
solved with this initial condition to obtain the variation of theta with time. Now, the
solution of this is pretty simple and I would only show you some of the steps.
And, I think you should do it on your own and what you get this thing that is
140
d  C 
 C1   2 
dt  C1 
 C 
ln   2   C1t  C3
 C1 
C2 I 2 Re
Where we have we can see that from our previous expression for  . And this
C1  DL
C3 is the constant of integration which can be evaluated, if we supply the appropriate
boundary conditions. So, the appropriate boundary condition as I said at t  0,   i ,

which would give rise to
 C 
C3  ln i  2 
 C1 
So, after everything is said and done what you get is
C2
T  T 
C1
 exp[C1t ]
C2
Ti  T 
C1
So, if you look once carefully to this equation once again what you should notice is that
that we are getting a similar form to that, what we have obtained, when we did not have
any energy generation.
But the presence of the energy generation term because of the joule heating of the
present in the wire creates, additional or introduces additional terms into the governing
equation, so which must be taken into account. So, the entire purpose of any modeling is
for you to ensure, that you did not miss out on any of the physical processes which are
taking place in the system.
So, once you identify the physical processes, then it would be easy to substitute the
corresponding mathematical form of the physical process in your equation, and when
you do the algebraic sum of the contribution of all physical processes, which are
contributing to the total amount of energy which is stored in the system, then what you
have is your governing equation that is all that there to is, just identify the processes put
141
the terms in the conservation equation. And, the sum must be equal to the right hand side
which is the time rate of change of energy stored inside the system, it is that simple.
And once you do that then it is going to be a case of integration of an ordinary first order
ordinary differential equation, and you require a boundary condition in this case an initial
condition at time t = 0.
So, what you see is if the time at t = 0, that is initially you know what is the temperature
of the solid to start with, then that provides you with the boundary condition as is the
case in this specific problem. So, what remains here is to put the values of T i the values
of the thermophysical properties for example, ρ, Cp, k.
And you use the value of convective heat transfer coefficient h, which is a parameter
which depends on many things including the operating conditions and, what you are
going to get is the unknown temperature. And we need to calculate the time required for
the wire to reach 1 ºc of the steady state temperature.
So, your T the temperature that you are that you are shooting for is one º less than the
steady state temperature. And if that is known since the steady state temperature is
known the only unknown in that equation is the time. So, that time you can calculate
easily. So, when you put the values in here what you would get is that your T∞ is 25, your
the other values that you get the values of C 1,
4h C
C1   0.5s 1 , 2  63.7
 CpD C1
Where Ti is the initial temperature.
When there is no current flowing through the wire, the initial temperature must also be
equal to T∞. So, Ti is T∞ and T is within 1 ºC of steady state temperature. So, this must be
equal to 87.7 ºC, so, when you put all these values in there what you should get as the
time required is 8.31 second. So, this gives you some idea about how to look at a
problem, which is which is different than what is given, what you have done, what you
have seen in your text book. So, you should be prepared to modify the equation, given in
your text depending on what you have what you have in the present scenario. So, this is
all about lumped capacitance that I wanted to teach in this class.
142
What there is still a vast number of situations, in which the lumped capacitance would
not be valid. So, how do you treat such a system the fundamental equation the
fundamental conduction equation the differential equation will be valid, but how you are
going to solve them is the question. So, the problems shifts from that of physical
understanding and heat transfer to the problem of solving partial differential equations,
where you may have been in the simplest case, with temperature as a function of time
and it is also a function of one of the dimensions so, let us say x. So, your governing
equation would be
 2T T
k 2
  Cp
x t
So, this kind of equation can be solved in a number of ways the pd one of the ways of
solving the pd is the method of combination of variables the method of separation of
variables. So, these different types of methods are available in your text, which I would
not cover in this specific course, but in most of the cases when you solve a two
dimensional heat transfer problem, you are going to end up with a series solution. And
that series solution can be truncated for some special cases and, those truncated solution
of the series solution would can be expressed in graphical form. So, that is one way of
analytical finding out what is the how does the temperature vary as a function of x, y and
time.
That is another method, which is the numerical method and you have probably heard
about the finite difference method, if not you are going to read it in your numerical the in
your course on numerical computation, where each equation can the differential equation
can be approximated as a difference equation. So, you can define definite nodes in the
control volume and write the difference equation across that node, and what you would
end up with is a series of equation series of algebraic equations, which can then be
solved to obtain what is the temperature of each of these nodes.
So, the finite difference is a powerful tool of numerically solving differential equations,
the differential the heat diffusion equation in order to obtain the temperature at every
node, which essentially is a numerical computation problem. So, I will not teach that in
this in this course, but it would it should remain, you should be familiar with how to
143
convert a governing, how to convert a differential equation to a difference equation. And,
then solve the resulting algebraic equation to obtain the temperature at every node.
Now, in your textbook if you are interested, you can see how the difference equation can
be written there are difference different ways by which, you can write the difference
equations, what you are going to what would be the form of the equation, if one node is
surrounded by other nodes in the free space or what if the node is essentially sitting on
the surface. Based on that your difference equation would be slightly different, but there
is a specific methodology which one can adopt which is there in the in the textbook and,
if you are interested you can take a look at that by which you can convert the difference
equation the differential equation into difference equations.
Now, what would be the spacing between the nodes or so, to say what is going to be the
size of the grid and, other numerical considerations in order to increase the accuracy of
your method, in order to ensure that your solution remains stable. All these are part of a
numerical solution of differential equations course, which either you have done or you
are going to do.
So, if you are interested in take a look at the finite difference method of solution of heat
diffusion equations. So, the major part which we have covered is the lumped capacitance
model, which gives you a closed form solution, but there is a vast literature available for
analytical solution, as well as for numerical solution.
In some special cases these numerical solutions, or the reduced form of the analytical
solutions can be expressed in graphical form, which is very convenient to use from an
engineering stand point. So, in the next class we are going to see what is going to
happen, if we can truncate the series solution arising from the analytical solution of the
heat diffusion equation and express them graphically. And use them to solve for
temperature variation with time and, temperature variation with space for systems in
which the Biot number is not less than 0.1 such that the lumped capacitance model is not
valid.
So, we will see the graphical solution of transient conduction problems, where lumped
capacitance assumption cannot be made in the next class.
144
Heat Transfer
Lecture - 13
Transient Conduction - Heisler Chart
The last part of Transient Conduction is to use the graphical form of the series solution.
In some of the cases the series solution can be truncated and the results for many
situations are expressed in a graphical form. So, that gives us a powerful tool for to find
out what would be the temperature as a function of position. Remember in this case the
Biot number is not less than 0.1. And therefore, the temperature is a function of the
spatial coordinates. Unlike the case where you have assumed that the temperature inside
the solid remains constant with respect to space it only varies with time.
But this is a case in which the temperature inside the solid can be a function both of
space and of time. So, let us look at how it is done, and you should look at your textbook
to see how the different charts and graphs are provided. These type of series solutions are
commonly known as Heisler charts.
We are looking at graphical representations of unsteady heat conduction. The results are
provided in graphical form which are known as Heisler charts. The graphical
representations are valid for three of the most common cases of transient heat transfer
145
that is: a plane wall of thickness twice L. It is also there for radial systems where r 0 is
the radius of the wire and it is also there for a spherical system where r 0 is the radius the
sphere.
For the cases of a plane wall and a cylindrical wire, T 0 is the centerline temperature that
means, the temperature along the axis. Whereas, for the case of a spherical system T 0
simply refers to the center temperature. It is to be noted that all of these; the plane wall,
the cylinder and the sphere are exposed to convection environment at a temperature T
infinity. So, as time progress, based on the difference between the initial and the
temperature to which it is exposed, the temperature inside the solid will vary with time.
So, we will say that at time at time (t = 0); T of the solid would simply be equal to a
constant temperature Ti which is uniform throughout. Initially, the temperature of each
of these solids is at a constant value equal to T i and the time varying temperature is
given as T which is a function of both x and time minus T infinity. So, this is some sort
of a difference temperature this is for a plane wall whereas, for the case of a cylinder it is
going to be T which could be a function of the radial position and time.
So, this is what we have for theta i which is the initial temperature difference: T i minus
T infinity. And finally, what we have as theta 0 to be equals T 0 minus T infinity. So,
theta the difference temperature denotes T as a function of x and t minus T infinity for
the case of plane wall whereas, for the case of cylinders or spheres this theta is simply T
(r, t) minus T infinity. So, for the case of a cylinder this is simply the central line
temperature as compared to the T infinity.
And for the case of sphere it is the center temperature with respect to the T infinity. So,
theta i is T i minus T infinity, the solid object could be hotter or cooler; and theta naught
which is the central line temperature minus T infinity. So, I guess the three nomenclature
of theta i and theta 0 are clear, theta is the temperature as compared to T infinity, theta i
is the initial temperature as compared to T infinity and theta 0 is the centerline
temperature in comparison to T infinity. The results of these are expressed in graphical
form and I will show you one or two of representative graphs how they look like.
146
0
So, let us take the plane wall of thickness 2L and you would see that this is plotted as
i
a function of Fourier number between 0 to 1. So, at t = 0; the theta 0 must be equal to
theta i. So, all points in the solid will be at a constant temperature at time t = 0. As the
time progresses you are going to get a family of curves like this. So, this is one figure
which is provided in the text. So, this essentially gives you the theta 0 which is the mid
plane temperature for an infinite plate (the plate is wide) of thickness 2L. However, its
thickness is finite, and that thickness is equal to 2L. So, this curve provides you with
what is the centerline temperature and how it varies with time for different values of Biot
number.
So, having this graph for the plane wall would let you find out what is the centerline
temperature as a function of time. But, but you also need the temperature at every point
in between. So, what is to be done if you require not just the centerline temperature, but
the temperature at every point in between. So, for that there is a second curve. The
k 
second curve is plotted as a function and ; again from 0 to 1.   T  x, t   T ;
hL 0
 0  T0  T ; and i  Ti  T . The two curves that you see are for plane wall are given in
your text. You would also see that similar curves are there for the case of spheres as well
as cylinders. So, let us say you would like to find out what is going to be the temperature
147
of a plane wall which is experiencing convective heat loss or gain from two sides of the
plate of thickness 2L and it is really wide in other directions. So, how do you find the
temperature knowing that lumped capacitance model cannot be used. So, use Heisler
chart.
The first Heisler chart where the x-axis is going to be to be Fourier number. The x axis is
Fourier number which contains the time and the y axis is the centerline temperature so,
for different values of inverse of Biot number. So, what you need to do is first find out
what is the Biot number for your system, then find out what is the value of Fourier
number corresponding to the time that you require the temperature.
So, let us say that at, t = 10 minutes you know the value of alpha, you know the
characteristic length. So, you know the value of Fourier number you go all the way up to
the Biot number which relates to your material and then come to this side find out what
is the value of theta 0. The theta 0 is nothing but T 0 you understand that this T 0 is also
a function of time whereas, T infinity remains constant this T i, the initial temperature is
not a function of time. So, therefore, this has a fixed value. So, in this theta 0 is a
function of time. So, through the use of this curve at a given t for a given value of inverse
of Biot number you find out how the central line temperature is changing with time.
Then you come to curve two, wherein you see that inverse of Biot number in the x axis
and the family of curves that you see are at different locations in the solid. Therefore, to
obtain the centerline temperature, you require just one curve. To evaluate the
temperature at points in between the central line and the surface you require the second
curve as well. So, this is the way by which one can obtain for simple geometries, plane
wall, cylinders or spheres what would be the value of temperature as a function of
position and as a function of time. All these are valid for Biot number greater than 0.1
where the lumped capacitance model is not valid.
Now, there is a second part to it: in some cases, you would also like to know how much
of heat the control volume has gained while it is being heated or if it is being cooled then
how much heat it has lost. Since, the control-volume is not space-wise isothermal you
would exactly like to know what the energy content of the body is at a given instant of
time and that brings us to a third Heisler chart which will give us the value of the energy
content of the object as a function of time. So, the first two charts are to find out what is
the central-line temperature and what is the temperature at any point in the control
148
volume. The third chart of Heisler refers to the energy content of the body as a function
of time.
Q
So, the curve looks something like this, this is from 0 to 1 which looks something
Q0
h 2 t
like this. So, this is which is nothing but Fourier number times Biot number square
k2
hL
and these are .
k
Q refers to heat loses for the infinite plane or infinite cylinder or infinite sphere. And Q 0
is the initial internal energy of the body, we know that initially, the entire object is at a
temperature of T i which is a constant. So, Q 0 can be written as  CV  Ti  T  which is
nothing but  CV  i . So, in this figure Q is the actual heat lost or gained depending on
whether it is heating problem or a cooling problem by the body in time t.
So, if you know what the time is, if you know all the thermophysical properties.
Essentially, if you know the Fourier number and the Biot number then using these graphs
corresponding to a different value of Biot number you should be able to calculate what is
Q. Since, your Q 0 is known to you.
149
So, since you know the initial internal energy content of the object you would be able to
find out what is the energy that the body has lost. So, through this chart you then would
be able to calculate the total amount of energy gained or lost in the system. So, this is the
third Heisler chart which can be used for many of the engineering calculations provided
you are dealing with the plane wall or you are dealing with an infinite cylinder or you are
dealing with a sphere.
But there is one word of caution here, this truncation of the infinite series solution which
has been used to obtain to generate these curves has a limitation. It is only valid when the
necessary condition for Heisler chart to be valid, that is Fourier number must be greater
than 0.2. This condition has to be kept in mind while using the Heisler chart.
So, the problem that we are going to deal with is on transient conduction. You consider a
steel pipe line which is 1 meter in diameter. The pipeline has a wall thickness of 40
millimeter and it is heavily insulated on the outside. The walls of the pipe are at a
uniform temperature of minus 20 degree C.
150
So, what we have initially is that T at any x but 0 time is equal to T i is equal to minus 20
degree C and then we have hot oil which is pumped through this pipe with a convective
heat transfer coefficient corresponding to 500. So, h is equal to 500 Watt per meter
square at the inner at the inner surface.
Now, you must find out what is going to be the appropriate Biot and the Fourier numbers
8 minutes after you start the flow of the hot oil. So, initially everything was at a
temperature of minus 20 and then suddenly you start passing hot oil through this and you
would like to find out how the temperature will change at different points.
So, what we can do then is first calculate the Biot number after t = 8 minutes. So, when
hL
you go for t = 8 minutes, the Biot number is and by putting the values in there you
k
would see that the Biot number is equal to 0.313 so, it is greater than 0.1. So, LC is not
valid.
That is one observation and you can also obtain the value of Fourier number and you
should check yourself that is this is going to come to 5.64. So, this is the first part of the
problem. The second part of the problem is after t = 8 minutes, what is the temperature of
the exterior pipe surface covered by the insulation; that means, what is the temperature of
this surface after 8 minutes.
151
Now, the interesting thing what you need to realize or need to see is that this is exactly
like half of a pipe of thickness 2 L. Now, if you remember what happens at x equal to 0
that means, at the centerline of the planar system, it acts like an adiabatic wall. The
temperature is maximum at that point and therefore, x equals 0 from x 0 to L is identical
to that from x 0 to minus L.
So, having the pipe insulated at x equal to 0 ensures that I can the solution of the plane
wall for this case as well. So, the outside of the pipe wall is nothing but the centerline
temperature of Heisler chart.
So, since this is an adiabatic plane and this is an insulated plane this part and this part are
identical and therefore, Heisler chart is going to be valid. So, part two of this is what is
going to be the temperature of the exterior pipe surface. So, if you need to find that out
then what you do need to do is you are going to use Heisler chart and for an inverse of
Biot number to be equal to 3.2 you find out what is theta 0 by theta i from Heisler chart.
So, what do you have this here is essentially you are using the first curve. So, for t equals
8 minutes that means, you are calculating the value of Fourier number, you know the
Biot number. You go all the way up to the value of the corresponding Biot number for
this plane wall, come to the side and find out what is the value of theta 0 where theta 0 is
nothing but the central line temperature minus T infinity.
152
So, with this from Heisler chart number 1 you should be able to calculate this theta 0 by
theta i to be equal to 0.22 which would give you the centerline temperature at t equals 8
minutes to be equal to 42 degree Centigrade. The third part of this problem asks you
what is the heat transfer to the inner surface; i.e. what is the heat flux to the pipe from the
oil at 8 minutes?
So, how much of heat gets transferred from the oil to the pipe. So, that is essentially the
problem. The temperature of the hot oil which I think is not mentioned over here, the
temperature of the oil is 60 degree. So, when it says what is the heat flux to the pipe from
the oil at t equals 8 minutes then I am essentially trying to find out what is the
temperature at x equals L and t equals 8 minutes. If I can find this, then the heat flux at x
equals L will simply be equal to h times temperature at x equal to L at 8 minutes minus T
infinity where T infinity is that of the oil.
In order to find the temperature at x equal to L what you need to do is use the second
chart and in the second you know what is the inverse of Biot number. You go all the way
x by L to be equal to 1 and read what is the value of the x axis. So, here theta by theta
naught, theta naught you have already calculated from here. So, you can calculate theta
which is corresponding to x equal to L. So, this value would give you the temperature of
the inner wall of the pipe after at a time equals to 8 minutes.
So, when you use Heisler chart number 2; you should be able to see that your theta at L
corresponding to a time of 8 minutes which is 480 seconds by theta naught to be equal to
be equal to 0.86, which would give as T L 480 to be 45 degree Centigrade. And
therefore, q double prime L would be 500 times 45 minus 60 equals minus 7500 Watt per
meter square. So that means this much of heat flux is going to come from the oil to the
inner surface after 8 minutes.
153
So, the last part is you have to find out what is the value of Q till 8 minutes. What is the
value of heat that has that has been transferred from the hot oil to the pipe over the time
of 8 minutes. So, obviously here you are going to use the 3rd chart where this is the
amount of heat which has been transferred. This is the amount of heat which was which
was initially contained in the solid wall and here you have the Fourier number which
contains time and Biot number square.
So, at time equal to 8 minutes or 480 seconds you calculate this one, you go all the way
up to curve corresponding to the Biot number in this specific case, come over here and
find the value Q by Q naught. When you do that for this specific problem, you would see
that for a Biot number of the present system to be equal to 0.313 and Biot number square
times Fourier number equal to 0.55 your Q by Q naught from the curve you will read as
0.78.
So, your Q is 0.78 Q naught and this should be Q naught should be equal to rho C V T i
minus T infinity. So, this Q then is 0.78 rho C and V is V is pi D times, times L so, this T
i minus T infinity.
So, this is that will be volume is pi DL where L is the thickness of the pipe times length
of the pipe. So, when you do this, this is Q prime and Q prime is simply Q by length of
the pipe and when you put the numbers you would see Q prime to be equal to 2.7 107
Joule per meter. The minus sign in the total amount of heat that is transferred from the
154
oil to the pipe. So, the only significance of the minus sign is the direction of heat
transfer.
So, now at the end what we have in transient conduction is that we know there are two
principle methods for solving transient conduction problems. One is the lumped
capacitance which is easy to use, which gives us the compact form, but it has certain
limitations. It can only be used if the value of Biot number which is conduction
resistance by convection resistance is less than 0.1.
The second one is when the lumped capacitance model cannot be used, what is the
alternative. The alternatives are either go for an analytic solution, go for a numerical
solution or in a in some cases this series solution arising out of the analytical solution,
that have been presented in the form of curves for geometric geometries like pipe, the
plane wall, the cylinder or the sphere.
And these charts can be successfully and effectively used to obtain not only the
centerline temperature, but temperature at any point in between. Apart from that the
Heisler chart can also be used to find out what is the total amount of energy transferred
from the object to the outside ambient or vice versa and as a function of the initial energy
content of the object. So, using all three curves one would be able to obtain what is the
temperature and what is the energy that that gets transferred over a certain period of time
for a specific solid, for a specific convective heat transfer coefficient and so on.
155
Heat Transfer
Lecture - 14
Heat Transfer from Extended Surface
So, today we are going to have our last formal class on conductive heat transfer, and we
are going to end the study of conduction with a topic, which is the very common in
everyday applications. We are going to talk about extended surfaces. So, what is an
extended surface? An extended surface generally refers to a solid, where inside the solid
the conductive heat transfer takes place and from the exposed surface of the solid, we are
going to have convective heat transfer.
So, through the solid we have conduction and from the surfaces of the solid, exposed to a
fluid environment we have convective heat transfer. Now one of the major examples of
extended surface heat transfer is the case of fins. Now you probably have heard about
this term before, what a fin generally refers to which is the thin solid surface through
which is used for enhanced heat transfer. You would see those fins in various
applications previously when you could look at the back of the refrigerator you would
see those extended surfaces.
If you look at the radiator of a car or the radiator of the engine block of a motorbike, you
would see that there are protrusions of solid, which looks like thin plates and those thin
plates provide the additional area available for convection. So, generally the heat transfer
coefficient from solid to air is quite low. So, in order to; so we cannot change h without
expanding more energy; that means, making the fluid flow over the solid at a faster
velocity. What we can do is we can enhance the area, increase the available area which is
to be utilized for convective heat transfer.
So, q  hAT , this A is the area available for convection if it can be increased by putting
certain fitments to the hot surface, which are to be used for additional heat transfer and
they are collectively called as fins. Fins come in different shapes, sizes and in any case
you have to justify the use of fins. Because if you have seen the radiator of a car their
significant amount of fabrication cost involved in making the fin and also there is going
to be some cost related to attachment of the fin to the hot surface. So, unless those fins
156
meet certain fundamental criteria of additional or enhanced heat transfer, the use of a fin
is never going to be prescribed.
So, one must know what are the benefits that one is going to get out of having the on the
hot surface, what is its efficiency, what is its performance and based on these 2
parameters not only you decide whether to attach a fin, you also decide what kind of a
shape, what are the design parameters of an ideal fin should be. So, we will in this class
have a quick look of the fins, the governing equations, solution of the governing equation
for some specific boundary conditions, which would give rise to an idea of what is the
performance index of a fin and what is the efficiency of a fin. So, let us start with the
very fundamental case of when a solid surface is experiencing convective heat transfer,
and we would like to see the different shapes of fins which are commonly used in
everyday applications.
So, let us first take a look at what a fin would look like, what an extended surface would
look like.
So, this is a solid, whose temperature is let us say at T S and we have some air which or
some fluid which is flowing over it. So, this is a TS the temperature and the area exposed
for conductive for the convective heat transfer is A. So, the heat loss from the surface is
simply going to be
157
q  hA(Ts  T )
So, as I said this h is more or less fixed, the moment you fix T ∞, the moment you fix
what is the velocity of the air which is moving what is the shape and size of the solid and
so on.
So, I do not have any control on h, but I can enhance this area to increase the amount of
heat transfer by the application of extended surface. So, these extended surfaces are
commonly known as fins. So, you can have the fins generally look something like this,
the shape of the fin with which we are mostly familiar with.
So, this is what the fins look like example of extended surface area. So, this is 2 fins
which are connected to a base and you have flow of let us say air at T ∞ with some
velocity, which is going to give rise to a convective heat transfer coefficient of h and the
temperature of the fin let us denote is denoted this as T S and if this is my x direction then
my TS is going to be a function of x.
How far you are from the base and let us see the temperature of the base is T b. So, the
problem that we deal with is to how to extract more heat from the surface using
convective heat transfer. And also having these attachments the fins on top of the
surface, which enhances the area available for convection, but making these fins
fabricating these fins, attaching it to the base obviously would involved some cost and
we would look at what is the ideal situation, what is the basic requirement that you need
to fulfill in order to prescribe the use of the fin.
So, there are several types of fins which are commonly you would see commonly. So,
this is example of a straight fin, the one that I have drawn over here. So, this is an
example of a fin and its width is let us say w, it has a thickness equal to t and this is the x
direction. So, this is a straight fin of uniform cross section.
So, here you can see the cross sectional area of the fin is equal to w times t, which
remains fixed and the temperature of the base is T b whereas, the temperature of the fin is
let us call it is TS and this area is available for convection. The area at the top and the
area at the bottom they are available for convection whereas, you are going to have
conductive heat transfer through the fin. So, as the heat comes towards the tip of the fin,
158
part of it is going to leave the surface by convection and so, there would be diminishing
amount of conductive heat, which is moving towards the tip of the fin.
So, it is the example of a straight fin with uniform cross section. On the other hand you
would sometimes also have fins which look like this, where the cross sectional area does
not remain constant and you have fin whose cross sectional area is going to be different.
So, if you look at this is what you would see is the cross sectional area of the fin keeps
on decreasing in the convection area, the area available for convection that is more or
less fixed, but the area available for convection conduction keeps on decreasing.
So, it is an example of a straight fin of non uniform cross section this is another example
of fin some of the fins which are commonly seen in applications. So, these are known as
annular fins. So, these annular fins are connected to. So, this is the cylinder which is hot,
whose temperature you would like to reduce. So, you attach an annular fin and this is the
surface which is additional surface, which is available for convective heat transfer. So,
what we see that the purpose of the fin is to provide additional surface area for
convective heat transfer.
The fins can be of different shape, different size, its cross section which the area which is
which is to be used for conduction may or may not remain constant. So, this kind of
various variations in fin size and shape are possible. So, now, what we are going to do is
we are going to find out we are going to derive a general formulation for the heat transfer
simultaneous conduction and convection from the surface of a fin and using those that
governing equation with appropriate boundary conditions, we would like to obtain what
is the temperature profile inside the fin.
So, because if we can find out what is the temperature profile of the fin, then if I if I
considered x at x = 0 to be the base, where the fin is attached to the base. So, x = 0 refers
to the point where the fin is attached to the base, from where you would like to extract
dT
the heat then  kAb x 0 , would give you the amount of heat that the fin extracts from
dx
the base.
159
So, if you look at this figure once again, if this area of the fin over here; let us say this is
the area is Ab. So, the amount of heat which is taken up by the fin at steady state must be
dT
equal to  kAb x 0 where x starts from here.
dx
dT
So,  kAb x 0 , gives you the amount of heat taken up by each of these fins ok. So, this
dx
would give you an idea if you are making any additional benefit by having the fin. So, if
you do not have the fin, then this area at the base would still be available for convective
heat transfer. So, in that case the whole of A b would be exposed to whatever is flowing
here with h and T∞.
So, if I did not have the fin attached to this point, the amount of heat transfer in absence
of fin , the amount of heat transfer would simply be equal to hAb (Ts  T ) . So, this is the
amount of heat we should have been taken up by the flowing air, if the fin is not there.
The moment I attach a fin the amount of heat transfer must be equal to the conductive
dT
heat transfer at the base. So, its  kAb . So, if you look at these 2 quantities, this is
dx
when the fin is not present and this is when the fin is present. So, the performance the
effectiveness all these definitions should have these 2 quantities either in the numerator
or in the denominator.
So, what would be the effectiveness of a fin? If it is there how much of heat it is
extracting as compared to if it is not there. So, these 2 quantities over here would give
you some idea about what is the performance of the fin.
160
So, the efficiency of a fin; the efficiency which is lets call it as  f is defined as the actual
heat transferred divided by heat which would be transferred if the entire fin was at the
base temperature. I will write it once again, just to make sure that the concept is clear to
all of us.
There are 2 concepts which are to be defined here. So, when I talk about effectiveness
and efficiency, what exactly do I mean? In the case of effectiveness, I have to decide
whether having the fin makes any sense should I have the fin or should not I have the fin.
161
When I look at the fin efficiency, I am trying to find out how efficient is the fin. So, let
us first talk about efficiency. When I have decided that yes I am going to use fin for a
specific geometry for a specific operation of a specific application now, I am going to
think when do you think the fin is going to be maximum effective.
Now the fin is the gradient which is responsible for a heat transfer from the fin to the
atmosphere, is simply temperature of the surface of the fin at any location, which we
denote as TS and temperature of the fluid which is a constant which we denote as T∞.
So, the amount of heat transfer from the fin is dependent on T S- T∞. This temperature of
the fin TS is a function of x, it is equal to T b; that means, temperature of the base and as I
move if this is a fin and this is my base, this is at T b and as I move towards the tip of the
fin, the temperatures gradually decrease. So, since the temperature gradually decreases.
So, you start at 100 ºC, in the air is blowing at 25 ºC. So, as you move towards the tip of
the fin, the temperature gradient between the surface of the fin and the air will
continuously decrease.
Since it decreases, you are going to have lesser amount of heat transfer from the fin as
you move in the direction of +x. Now how do you make the process maximum efficient,
how would you achieve maximum efficiency in the heat transfer process? That should
only be possible if by some means or if by the nature of the material of the fin that you
have used, the entire fin is at the temperature of the base.
So, instead of temperature of the fin wearing at points decreasing as you move away
from the base, if the situation is such that the entire fin is it at T b, then the heat transfer
would be Tb-T∞, which would be a constant and you are going to get the maximum
amount of heat transfer from a given fin. So, that is how the efficiency of the fin is
defined which is actual heat transferred divided by the heat that would have been
transferred, if the entire fin is at its base temperature.
So, I am going to write the definition of the efficiency as actual heat transferred divided
by heat which would be transferred, if the entire fin was at base temperature. So, this
would give you an idea of what is the efficiency. Because we always defined efficiency
in such a way, it has a maximum value equal to 1 or 100 percent. So, the actual heat
transferred where the temperature is a decreasing function of x divided by the heat which
162
would have been transferred if the entire fin was at Tb, the best temperature that would be
the logical choice to define the efficiency of a fin.
So, now let us think about what is fin performance and this would allow us to decide
whether or not fin use. So, the fin effectiveness, what  f is defined as fin heat transfer
rate over heat transfer rate that would exist without the fin. So, if you do not have the fin,
this entire area is at Tb and of this much is the area where the fin was attached to. So, let
us call it as area at the base.
dT
So, the fin heat transfer rate would simply be  kAb x 0 , where this is the point where
dx
your x where this is the direction of x. Now the heat transfer rate that would exist
without the fin, it would simply be hAb (Ts x0 T )
So, this is the definition of the fin effectiveness, where the fin heat transfer rate is given
obviously, by the numerator and the heat transfer rate that should exist without the fin is
given by this is simply hAb (Tb  T ) . Now the rule of thumb is use of fin is rarely for
rarely prescribe if  f is less than 2. So, one can only prescribe the use of fin if its
effectiveness is greater than or equal to 2. So, if this condition is met, then you go for
attachment of a fin to the surface and once you do that then you look at what is going to
be the efficiency of the fin.
So, I will very quickly tell you what would be the form of the fin efficiency, the
efficiency of the fin and then we will see how that would give us an idea of what is the
what should be the shape of the fin. So, for a straight fin, straight fin uniform cross
section the fin effectiveness will be
1/2
 kP 
f  
 hAc 
So, what you see here that is  f the fin effectiveness is a function of thermal
conductivity is a function of P/Ac and it is also going to be a function of h. You would
like to have the largest value of  f possible. The larger is the value of  f as compared
to 2, the better is going to be its performance. So, what would be the nature of the fin
163
material? If you look at the expression for  f once again, the material of construction of
the fin should have a high value of thermal conductivity; that means, k has to be large,
fin the value of h should be small. The value of h small means the use of fins are only
prescribed when we have heat transfer from a solid to a gas, which traditionally has low
values of convective heat transfer coefficient.
So, the fin material should be should be of large k such that the entire base entire fin is
going to be at a temperature close to that of T b not exactly at Tb, but close to that of Tb.
So, the material of construction of the fin should be large, the fin is prescribed when you
have heat transfer from the solid to the vapor to the gas. So, that is why you would see
use of fins mostly for the case of air not for when a solid is in contact with a liquid which
has the high value of heat transfer coefficient. And finally, P/A c should be large; that
means, for every for the ratio of the perimeter to the cross sectional area should be large;
that means you are going to have a large number of thin fins attached to the solid surface
from where you would like to extract the heat. So, it gives us idea about the material of
construction of the fin, the shape of the fin and the situation where fins are to be used.
So, together they would give you a complete picture of when and why and how fins are
to be used. As very quickly go through the governing equation the derivation of the
governing equation and then we can move on to the next part.
164
So, the governing equation this is a conduction analysis, where you have this being the
fin, which is attached to a surface. So, this is the fin. So, this is my x direction and from
here I have the conductive heat transfer and let us see this is the surface from where I
have convective heat transfer. So, if I enlarge this, it is going to look something like this.
So, this distance is dx, this area available for convective heat transfer is dA S, which is
going to lead to the convective flow of heat and this area would be a function of x as well
this is a cross sectional area.
So, you have qx by convection coming into the control volume, qx  dx which is the
conductive heat going out of the control volume. So, by conservation I can write
whatever heat is coming at this point, must be equal to whatever heat that is going out
through this surface and whatever heat is going to be lost by the convection process.
qx  qx  dx  dqconv
So, this is just a conservation equation, the heat coming in must be equal to the heat
going out by conduction and heat going out by convection.
dT
Also, qx   kAc ,
dx
dqx
qx  dx  qx  dx (by using Taylor series expansion)
dx
dqconv  hdAs  T  T 
Where, Ac being the cross sectional area and T∞ is the temperature of the fluid
surrounding it.
So the governing equation becomes
d  dT  h dAs
 Ac   T  T   0
dx  dx  k dx
So, this becomes the governing equation and if I expand this, the final equation that I am
going to get is
165
d 2T  1 dAc  dT  1 h dAs 
     T  T   0
dx 2  Ac dx  dx  Ac k dx 
This is the final form of the equation that I am going to get. So, this would be the
governing equation for heat transfer utilizing a fin. Now instead of this fin having a
variable cross sectional area, if I assume that the fin is just like this, a fin with constant
cross section, then this term would disappear and what I am going to have is only the
first term and the second term to be equal to 0.
So, if I define   (T  T ) , and if I assume that Ac  f ( x) , then the same governing

equation can be rewritten .
dAs
Also,  P ( Perimeter ), As  Px
dx
So, this would be the reduced governing equation I call this as theta as well; so,
d 2  1 h 
 P   0
dx 2  Ac k 
And if I assume this to be equal to m2, then its
166
d 2
2
 m 2  0
dx
2 hP
Where, m 
kAc
So, this is going to be the governing equation for a fin of constant area. And the solution
of this equation solution of such kind of equation, these are homogeneous second order
differential equation. So,
 ( x)  C1e mx  C2e  mx
That is the standard solution which you can obtain. So, you need two boundary
conditions for this, one is  (0)  Tb  T  b , So, the first boundary condition is that the
temperature at the base is known to me which is Tb. So, Tb  T  b
The second boundary condition can have different variations like A, B and C. So, what
would I get in as the second boundary condition can vary? Let us say I have a fin which
is very long. So, what is going to happen if the fin is very long in this direction, then the
temperature of the tip most likely is going to be very close to that of the air. So, for a
very long fin, the boundary condition could be that as the x tends to ∞, temperature of
the fin is equal to T∞, that is one condition what which one can use ok.
If the fin is short what is its thermal conductivity is very high, then what is going to
happen is the entire fin is going to be at a constant temperature, which would be very
close to that of the base. So, a short fin of very high thermal conductivity, the boundary
condition at x = l; that means, at the edge is going to be T is equal to T b so, that is
another possible boundary condition what that can happen.
The third type of boundary conditions that can happen is that, whatever heat comes to the
tip by conduction must be convected out through the cross sectional area, which is
exposed to the ambient environment. So, what is the amount of heat which comes to the
dT
tip? It must be  kAc at x equal to l, that is the amount of heat which comes to the tip
dx
by conduction through the fin and if all of that at steady state.
167
Since it has to dissipate out of the cross sectional area by a convective process, the
amount of heat loss by convection from the tip of the fin would be hAc (TL  T ) . So,
these different types of boundary conditions are possible and based on that the fin
question has been solved, you can look at the textbook for the various forms of the
equations, various forms of the temperature variation, the temperature formed that form
of the temperature for a fin of infinite length of short length, but very high thermal
conductivity or for a fin whose tip is active. That means, the amount of heat coming to
this to the fin end must be dissipated out.
So, these combinations the first condition is always fixed. That the temperature at the
base is known it is that is at x = 0 the boundary condition is known, what happens at x=
l? We have three or four different types of boundary conditions and for each of those
boundary conditions, these solutions are available in your text. So, you should take a
look at them, but you should also be very careful about the alternate conduction analysis
and realize that the assumption of constant cross sectional area may not be valid in all
cases. For example, in the tapered fin which I have drawn the area is not area available
for conduction is not a constant so, which could give rise to interesting problems. And
the fin material should be of high thermal conductivity, the shape of the fin should be
thin and closely spaced and you are going to use fins only when you have heat transfer
between the solid and the gas generally not between the solid and the liquid.
And there are 2 parameters which define whether or not a fin is to be used or what is the
efficiency of the fin. The fin should can be used only when its performance is the index
is greater than equal to 2, and it is going to be most effective when the entire fin is at its
base temperature. So, that is all I wanted to cover in this class.
I guess you have had some idea about and what is involved in heat transfer from an
extended surface. There are vast literature and information available in many textbooks
and many hand books about how to design fin, when to add fin, the fabrication
techniques the way you attach the fins to a surface and so on. So, if you are interested
you can take a look at them, but always remember the fins are useful way and the
attachment of fins is a useful way to enhance heat transfer and this simply provide an
extended surface. Then therefore extended surface heat transfer is an important topic in
conduction heat transfer.
168
Heat Transfer
Lecture - 15
Fins and General Conduction Analysis
I will start with the derivation of the generalized conduction analysis and then see how
they can be applied for specific boundary conditions, all of which are realistically
possible depending on what is the size of the fin, its material of construction, and the
conditions at the tip of the fin. As mentioned previously, an extended surface is one
which is added to a hot surface in order to dissipate more heat from that surface. So, fins
can come in different shapes and size, but no matter what we need to justify the
effectiveness of the fin, the performance of the fin whether we should go for attaching
fins which are costly and requires additional fabrication to the hot surface. So, for that
we have defined the two quantities, effectiveness and the performance and, we have seen
that there is a certain numerical value of the effectiveness, which you have to cross in
order to prescribe the use of the fin.
But let us look at the mathematical side of it, how we can develop an expression that
would give us an idea of the temperature distribution. The one-dimensional temperature
distribution in a solid fin where we are going to have conduction in let us say the x
direction through the cross section of the fin and through the periphery of the fin we are
going to lose heat by convection. So, it is a case in which both conduction and
convection are present, and we will have the more general situation in which the cross-
sectional area can vary with x. So, we will not assume that is a constant cross-sectional
area fin, where the A c the cross-sectional area can also be a function of location. And
similarly, the peripheral area which is available for heat transfer can also be different.
169
So, let us look at the figure which we have drawn in the last class; we have a circular fin
of variable cross section. So, this A c is a function of x and this area of the fin is
available for conduction. So, what I have done is, I have taken a small section of length
dx and enlarged it over here, identifying the terms the by which heat can come into the
control volume and leave the control volume.
So, what you see here is that q x is the heat flux which is coming at x = 0, and qx plus dx
is the heat conductive heat which is leaving the surface at x plus dx. So, this is
conductive heat in and conductive heat out, we realize as we have stated before that the
cross-sectional area could be a function of x, this area denoted by dAs. So, if I write a
heat balance across this volume element, what I am going to get is that the heat that
comes in by conduction is equal to the heat that goes out by conduction and the heat that
goes out by convection. So, qx must be equal to q at x plus dx plus dq convection. So,
this is simply from energy balance for this volume element, we also know that from
dT
Fourier’s law, the heat flow by conduction can be expressed as  kAc , remember that
dx
A c in here can be a function of x. So, using a Taylor series expansion of q x and
dqx
neglecting higher order terms, I can write that qx  dx  qx   dx . So, what we have left
dx
with is, what is the heat that goes out by convection and I invoke Newton’s law of
cooling, which tells me that convective heat transfer is h area, surface area, times the
170
temperature difference i.e. hAs  T  T  . So, this T is the temperature of the surface at the
point where you are calculating the convective heat loss.
We realize that this T can also be a function of location or it can be a function of time.
So, now, I have identified each of these terms in equation 1 which is the conservation
equation. So, when I substitute the expressions in here, what I get from equation 1 is this
form. So, this is A c and this is A s and, once I expand this and identifying or
recognizing that A c can also be a function of x. The general form of the energy equation
for a, for one dimensional conduction in an extended surface is given as:
d 2T 1  dAc  dT 1  h dAs 
   T  T   0
dx 2 Ac  dx  dx Ac  k dx 
This equation is an energy equation, and if it can be solved it is going to give us the
variation of temperature with location and here we have correctly identified the
conductive heat transfer and convective heat transfer. The limitation of this equation is
its valid for one dimensional conduction only, but as I have mentioned before, in most of
the fins, the cross-sectional area is generally small, they can be long with very small
cross-sectional area. One of the key requirements of a fin material is that they should
have high thermal conductivity. So, if you have high thermal conductivity with a small
cross section, then it can safely be assumed that the temperature is going to be a function
of x, that is its axial location, perpendicular direction to the heat flow.
But at any location since the fin is thin its temperature is not going to vary with, let us
say y. So, one dimensional conduction is a good approximation to express the heat
transfer and the temperature distribution in such situations. Since this general equation is
a second order equation, it may require two boundary conditions. So, what are the two
possible boundary conditions? One boundary condition for the fin would be where it is
attached to the base and the temperature at the base is known, temperature at the base is
let us say it del x, denoted as T b. So, the temperature at the base of the fin is known, that
is one condition which is fixed and then I must think of what is going to happen to the
other end of the fin.
The case of what is happening at the other end of the fin could give us different
expressions for temperature distribution in the fin and we will just quickly go through
171
some of those other possibilities and the resulting expressions. I am not going to write all
the expressions, or we will not talk about how to solve these, you should refer to your
text where all these has been provided. What I am going to give you, is a glimpse of how
depending on the boundary condition what are the realistic boundary conditions, and
what can they tell us about the material of construction of the fin, the shape and size of
the fin and so on.
So, let us look at a simple case in which the cross-sectional area does not vary with x. So,
if the cross-sectional area does not vary with x and I am talking about let us say a
rectangular fin, a rectangular fin which is attached to the hot surface. So, if it is a
rectangular fin of constant cross section, its cross-sectional area is simply going to be
perimeter times x, where x is the distance in the direction of heat flow or it could be a
cylindrical thin fin, where the cross-sectional area is also kept constant.
So, if we have the cross-sectional area to be a constant then let us see what are the
simplifications that we can make to the general equation which we have derived. So, the
governing equation for a constant area constant cross-sectional area fin would consists of
the first term of the equation and the third term of the equation, the second term can be
safely withdrawn. So, the expression for a rectangular fin or for a cylindrical fin of
d 2T hP
constant cross-section would look like:   T  T   0
dx 2 kAc
172
So, the first cases is rectangular fin, which looks something like this ok, and this is a
direction of heat which is been picked up by the fin from the hot surface, which is at a
temperature of T b, temperature of the base the convection is from here. So, this is q
convection, and which is because of, let us say air flowing over the fin at a temperature
of T infinity, with a convective heat transfer of h. This is my x direction and the entire
length of the fin is L, the width of this is W, the thickness of the film let us say its t and
so therefore, the A c, the cross-sectional area which is available for conduction A c is
simply going to be W times t.
So, A c is equal to W times t, the cross-sectional area which is available for conduction
and if I talk about the perimeter of this fin, which would simply be equal to 2w plus 2t.
So, that is the perimeter of the fin and this is the cross-sectional area of the fin, you can
similarly have a situation where you have a cylindrical fin of constant cross section. So,
here the heat goes, picked up by the fin we call it as q f the same over here, this is q f and
we also have the same T infinity and h and let us assume that the diameter of this
constant cross section pin fin is D. This is again x and the entire length of the fin is L. So,
here the perimeter is going to be pi times D and the cross-sectional area which is utilized
for conductive heat transfer is going to be pi D square by 4.
Now, what we see is that A c is a constant, the surface area A s is simply going to P
times x. So, if I take a slice of the area of length x, then the surface area available for
conduction would simply be equal to P times x. So, what are the implications of that,
then d A c by dx would be equal to 0 and d A s by d x is going to be equal to P. So, if
you look at the governing equation which was there in the previous slide, I simply write
instead of d A s by dx, I replace that with P, where P is the perimeter of the fin. So, the
d 2T hP
governing equation therefore, becomes   T  T   0 . So, this is the governing
dx 2 kAc
equation where the A c is constant and of course, if A c is constant the area, the cross the
surface area can be expressed as P times x. So, if I define theta, which is an excess
temperature, which is T temperature at any location minus T infinity, then this governing
d 2 hP
equation should simply become just a more compact form    0 . So, this is
dx 2 kAc
173
going to be the governing equation, same governing equation but in a in a slightly more
compact form.
Let us define h P by k A c to be some other constant, because here h, P, k, and A c all are
constant. So, if I define this h P by k A c to be another constant to be equal to m square,
d 2
then the governing equation becomes 2
 m 2  0 . So, this is a linear homogenous
dx
second order differential equation with constant coefficient and the solution of this is
  x   C1e mx  C2 e  mx , these c 1 and c 2 are constants of integration.
So, I require 2 boundary conditions, boundary condition 1 is fixed, that is theta at x

equals to 0 should be equal to T at the base, minus T infinity and let us call it as theta b.
So, that is the concept that is the temperature which is known at x equals to b and for
boundary condition 2, you have you have several cases. So, I will discuss but not write
what are the solutions for this. So, the first case is known as the active tip, tip means at
the other edge of the fin. So, what it is the other edge of the fin, you have the heat
coming in to the tip by conduction, the heat that goes out of it is by convection.
To maintain steady state, the conductive flow of heat up to this point must be equal to the
convective heat which is taken out from the tip of the fin. So, that is to maintain steady
state, otherwise the temperature will change with time. So, what is the conductive heat
174
dT
which comes in? That must be equal to  kAc at x = L. What is the convective heat:
dx
hAc  Tx  L  T  . So, this is going to be the boundary condition for this case where the tip
is active and therefore, the amount of heat which comes up to the tip by conduction must
be taken out by convection. So, that is the standard way of looking at things and for these
two boundary conditions in your text, you would see what the temperature distribution
for these two boundary conditions would be.
So, I will not write that, you look at your text, the other condition what you can do is let
us assume that the fin is very long. If the fin is very long, then by the time you will reach
to the other end of the fin its temperature has decreased from its original starting value of
T b and now it lies very close to the ambient air itself. So, if the base temperature is T b
and I have a very large fin which is attached to it then as I progress the temperature of
the fin is going to be higher than the temperature of the surrounding air, but it is going to
be lower than the base temperature.
So, if I provide sufficient length and knowing that the fins is, of sufficient length then
what you would expect is that at the very end there is not going to be any difference
between the temperature of the fin and the temperature of the surrounding air. So, that is
the case as L tends to infinity, the temperature of the fin at that location tends to be equal
to the temperature of the air. So, that is known as the long fin approximation.
175
So, if I write the expression for this is as, for a long fin as x tends to infinity then you are
going to get T of x equals L is equal to T infinity. So, that is that is the second condition,
similarly one can one can think that the tip of the fin is adiabatic. The third condition is
the tip of the fin is adiabatic. So, if the tip of the fin is adiabatic, then the corresponding
d
boundary condition would be  0 at x = L. So, that is a standard condition, that you
dx
would get for any adiabatic surface and the fourth boundary condition is temperature is
specified at x equals L. That means, theta at x equals L is known, which is let us call it as
theta L.
The total heat if you look at the figure which I have drawn over here, the total heat which
is transferred by the fin at steady state, must be equal to the heat that it picked up from
the base at x equals to 0. So, which I have denoted as qf. So, I would like to know how
much of heat that the fin can dissipate when it is attached to the base.
So, that must be equal to the conductive flow of heat into the fin from the base which,
which is denoted by q f for these 2 cases. So, what is q f? Since, I have mentioned that q
f is nothing, but the amount of heat taken up by the fin, at the base at steady state by
dT
conduction. So, this must be equal to  kAc at x = 0.
dx
176
So, knowing the profile of T, I would be able to obtain what is dT/ dx and what is the
value of d T/ dx at x equals 0. Put it back in here and what you would get is the total heat
to be taken up by the fin under different conditions. and I have also mentioned that, this
temperature distribution will; obviously, vary depending on what is the expression for
the second boundary condition that you have taken in this case. But no matter what, once
you have the temperature distribution you can find out its gradient and once you have the
gradient you can put x = 0 to obtain what is the value of the gradient at the beginning, at
the point where the fin is joined to the solid surface. And thereby you should be able to
calculate: what is the total amount of heat to be taken up by the fin? Which has some
direct bearing on fin performance. So, this fin performance is defined as the fin heat
transfer rate, this is the amount of heat taken up by the fin and the heat transfer rate
without the fin. So, what is heat transfer rate, when you have the fin is q f which is this
one, divided by what is the heat transfer rate without the fin it should be h, area would be
A c, area cross sectional area at the base divided by theta at the base.
So, theta b is nothing, but T b minus T infinity, A c b is A c b is nothing, but A c at x, x

equals to 0 and h is the heat transfer coefficient. So, obviously, to justify the use of the
fin, the amount of heat taken up by the fin must be more than, when the fin was not
there. So, this fin performance must be greater, must be greater than one and it is denoted
as epsilon f. And the as I said the rule of thumb is epsilon f must be greater than 2 the fin
should at least be able to dissipate twice the amount of heat which is going to be
dissipated when the fin is not there. So, how does the expression of fin performance tell
us something about the shape of the fin, the material of construction of the fin, what
should be the situation in which use of a fin is justified. So, you have k in the numerator,
since k the thermal conductivity is in the numerator in order to enhance the effectiveness
of the fin k should be large. So, the material of construction of a fin should always be of
high thermal conductivity, you have P by A c should be large, that is p the perimeter A c
is the cross-sectional area.
So, thin fins are always preferred, and you have h is, h in the denominator. The use the
fin is justified only for those cases where the value of h is low, h is small and when h is
small the most likely condition of h being small is when you are going to have heat
transfer from the solid to air, which typically have very low heat transfer coefficient. So,
this completes my discussion on fins, their performance, their efficiencies, the governing
177
equations, the modified form of the governing equation, if you have a constant cross
section fin. And what are the different types of boundary condition one can have to
obtain the expression of the temperature distribution in the fin and how that expression
can be used to obtained the total heat up-take.
So, the problem that I am going to give you which is the practice problem, there is a fin
of this shape, where the diameter of the fin is a function of x and the temperature at x
equals to 0. So, this is x and the entire length is L. So, T 0 is the temperature at this end.
So, this is nothing, but a truncated solid cone, whose diameter is a changing with x and
1.8 x
D  ae L . So, this is the functional form of how diameter changes with x and a is a
constant, the value of a is 0.8-meter, x in meters and L is the length of the fin. The length
of the fin is 1.8 meter, it has a thermal conductivity of 8 Watt per meter Kelvin.
There is a uniform heat generation in the cone as 1989 Watt per meter cube. So, the
amount of heat generated in the truncated solid cone is 1989 Watt per meter cube, the
lateral surface of the fin, these surfaces of the fin are insulated. So, heat can enter or
leave only from these 2 at x equal to 0 and x equal to x equal to L, the temperature at x
equal to 0 which I call as T 0 is known as 300 degree Centigrade and the heat rate that is
the amount of heat which comes into the cone this is q x at x equals 0 this q x at x equals
0 is 500 Watt.
178
Find, 1 what is the temperature at x equals L that is what is the value of T L? And
second, the heat transfer rate at x equals L at the right-hand surface. So, the things that
you must find out is what is the value of T L? And what is the value of q L; that means, q
at x equals L? So, at one end the temperature and the heat rate are known, some amount
of heat is generated in the solid cone all the sides are perfectly insulated.
So, whatever heat that is generated must travel in axial direction only there is no radial
flow of heat and we will assume the T is a function only of x T is not a function of r. So,
T is a function of x only, and the temperature at this end is given as 300 and the heat
enters at x equals 0, the amount of heat that enters at x equals 0 is 500 watt, what you
have to find out is what is the temperature at x equals L and the and what is the heat rate.
I will quickly give you some pointers on this, the total heat which enters at x, minus the
total heat that leaves at x plus del x, plus q dot the amount of heat which is generated,
multiplied by pi r square delta x should be equal to 0. So, the conservation equation, if I
take a thin strip of this, as my control volume, some amount of heat is coming in which I
call it as Qx, the amount of heat which goes out of this is Qx plus dx and the amount of
heat generated in here per unit volume.
So, I multiply q dot with the volume which is pi r square times delta x. So, that is going
to be my conservation equation. And once you divide both sides by delta x; that means,
Qx at x minus Qx at x plus delta x divided by delta x should be equal to minus q dot pi r
square. And when I take delta x tends to 0; that means, I am using the formula for the
d 
first derivative, what I should get out of here is Qx  Q   r 2  .
dx
179
So, your final form of the governing equation Qx is q dot pi instead of r, I am going to
use this expression for diameter, I converted to r and, what you get over here is pi a
square e to the power 3.6 x by L times 4, pi D square, pi D square by 4. So, this is my
governing equation, now this governing equation can now be integrated, Qx is going to
be L by 3.6 q dot pi a square e to the power 3.6 x by L and I am going to have here as 4
plus c 1. c 1 is a constant of integration, what is the boundary condition? I know what is
the value of Qx at x equals 0 to be equals 500 find out what is the x value for c 1 and
secondly, you can also this is a total rate of heat.
So, Qx is equal to minus k A dT/ dx from Fourier’s law. So, if you substitute that in here
what you should get is another expression in terms of temperature. Integrate that to
obtain the temperature as a function of x, but you need another boundary condition
because there would be one more integration constant which would come and that is T at
x equals 0 is 300 degree Centigrade.
So, substituting Fourier’s law into my governing equation, after I evaluate c 1, I should
be able to obtain a differential equation for T in terms of x; integrate that expression to
obtain T as a function of x and the integration constant can be evaluated with a known
temperature at x equals 0.
180
So, the answer to the problem that you should get that T at x equals L to be equals 76.2
and Q at x equals L to be 18.3 kilo Watt. So, this is a nice example to show that the, it is
always better to start from first principles. Write the conservation equation, derive what
would be the temperature distribution or heat flux distribution, identify the appropriate
boundary conditions, solve for the integration constants and arrive at the desired value of
either the temperature or the heat flux.
So, this concludes our study of conductive heat transfer in extended surfaces and from
next class onwards we should start one of the very important chapters in heat transfer
which is convective heat transfer.
181
Heat Transfer
Lecture - 16
Fundamentals of Convection
We are going to start with another mode of Heat Transfer namely convection, which is
almost omnipresent whenever there is a contact between a solid and a fluid and the solid
temperature and fluid temperature are different. Now, in addition we may have the fluid,
have a velocity with which it flows over the solid. So, in that case the heat transfer, the
major heat transfer, the maximum amount of heat transfer is going to take place through
convection. So, convection requires the presence of a medium and in most of the cases
there would be an imposed flow of the fluid over the solid.
So, convection is characterized by the presence of a moving fluid stream, in contact with
a solid. In some special cases, there maybe this motion of the fluid can be induced by a
density difference by a, by a buoyant force without the presence of an external agency
that drives the fluid over the solid surface, which is known as natural convection or free
convection. So, an object which is or a hot object which is placed in, let us say water, it
is going to lose its heat by convection and unless there is a flow imposed flow of the
water stream. Then it is going to be natural or free convection, where the liquid near the
solid, its temperature will increase, its density will decrease and due to the buoyant force
that hot liquid will rise up along the solid and to be replaced by cold water from the
surrounding.
So, a natural current would therefore, set in which is known as natural convection. So,
we will treat natural convection separately towards the end of our discussion on
convective heat transfer. But right now we are going to mostly concentrate on situations
in which there is an imposed motion of the fluid in contact with the solid, the
applications of convective heat transfer or the occurrence of convective heat transfer is
everywhere in industrial processes. So, wherever you have to heat up a fluid which is
entering a reactor or you would like to cool a stream of liquid before you discharge it to
somewhere, it requires the exchange of heat in between two fluid streams. It is also very
182
important to ensure that the chemical plant is going to operate at its highest efficiency, if
you can regenerate some of the heat which would otherwise be lost.
So, convection plays an important role in how you can effectively design an equipment,
which is going to have the maximum amount of convective heat transfer between a solid
and the fluid or between two fluids streams separated by a solid barrier. These
considerations play an integral role in evaluating, in determining the efficiency of the
process. So, the heat exchangers that we would again consider towards, on in this course
the heat exchangers, the design operating principle of heat exchangers they rely heavily
on convective heat transfer.
So, it is important that we understand and learn convection, but in order to have a formal
mechanism to study convective heat transfer one has to start with the simplest possible
case. Where, we have a solid plate which is at a higher temperature in relative to the fluid
which, with which it is in contact. And as I said we are going to concentrate mostly on
forced convective heat transfer; that means, the fluid which is in contact with the fluid
with the solid is moving with a certain velocity.
Now whenever a fluid comes in contact with a solid which is stationary, then there is
going to be a change in the hydrodynamic pattern of the fluid which is flowing over it.
So, we will discuss that and the concept of boundary layers would be relevant in
describing the convective heat transfer, that is expected out of the solid plate, out of the
flat solid plate. So, we have chosen flat solid plate because it is the simplest possible
geometry that you can think of. Any change any curvature in the solid with which the
liquid is in contact will give rise to additional complexity, which is slightly more
complex. So, our starting point would be, we would like to see how the heat transfer
from a solid plate would take place when it is in contact with water. Let us say the fluid
is water, which is very common which we see almost every day.
So, what is going to happen to it? So, we start with the fundamental concepts of
boundary layer, some of you I think most of you already are aware of the concept of
boundary layer, when we considered the momentum transfer part of it.
183
So, let us say this is a solid, which we have, I have said it is a temperature of T s and this
stream of fluid is approaching the solid with a temperature of T ∞ and let us assume that
Ts is greater than T∞. So, the fluid is going to come in contact with the solid, extract heat
by the convection process from the solid and as a result of which the temperature of the
solid, temperature of the liquid in the vicinity of the solid will rise.
So, this increase in the temperature of the fluid very close to the solid plate is something
that we need to model in order to get an idea of what is the heat transfer coefficient. It
was the principle equation, the relation that we are going to follow, that we would we are
going to use extensively in describing convective heat transfer is Newton’s law of
cooling, which simply tells us the amount of heat, which is lost from the solid in this
case, is can be written as q  hAT , where h is the convective heat transfer coefficient,
A is the area in contact and T is the temperature difference between the solid and that
of the liquid, which is at a point far from the solid.
So, when I look at this one, this the amount of heat lost from the solid would simply be
equal to hAT , where T is defined as Ts  T and if there is a mechanism, if, let us say
there is a heater which maintains, which maintains the temperature of this, of the solid at
constant at Ts then Ts is a constant the temperature with which the liquid is coming in
contact or approaching the solid that is also a constant. So, T is simply going to be a
constant and if T is a constant the area which is in contact with the fluid that is also a
184
constant. So, essentially the heat q that is there is lost from the solid is going to be a
function of h.
So, how can we manipulate h, the convective heat transfer coefficient or how can we
evaluate the convective heat transfer coefficient that becomes the study of convective
heat transfer. So, how do I relate h with other parameters, now what are the parameters
on which h would depend on. So, if you think heuristically, let us see the fluid if it is
moving at a higher velocity, all of us realize that when you are outside and a cold wind
blows on a winter day, you feel more cold as compared to the case where the air velocity
has significantly reduced.
So, the temperature in both cases will remain the same; however, the velocity will be
able to extract, because of the velocity the cold air would be able to extract more energy
from your body and thereby you would feel cooler. So, h, the convective heat transfer
coefficient is going to be a strong function of velocity, convective heat transfer
coefficient would also depend on the thermo physical properties of the fluid which is
flowing over the solid.
So, what are the thermo physical properties on which it would depend on, since we talk
about flow, the two properties which come to our mind automatically, one is what is its
viscosity, that is an important that will play an important role in what is going to be its
density. So, ρ and μ would come into any expression of h, specially when we concentrate
only on the momentum transfer part of it; that means, only fluid flow part of it.
Now, this heat is being taken up by the fluid by as sensible heat. So, the temperature of
the fluid in contact with the solid will start to rise and whenever you are going to have an
increase in temperature of the fluid, the capacity that heat, the thermal capacity of the
fluid must be taken into account. And the one of the parameters, one of the properties
which define the thermal capacity of the fluid in terms of extracting heat from the solid
has to be Cp which is the specific heat.
So, Cp is going to play a major role and how is heat going to transfer between the solid
and the liquid. So, the molecules of the fluid which are flowing over the solid, when they
come in contact with the solid due to the no slip condition. Which I am sure you are
aware of from your fluid mechanics study; that means, the molecules of the fluid,
molecules of the moving fluid, which are in contact with the solid they do not move so,
185
they become static. So, there would be a decrease in velocity as we approach the solid
and on the solid the velocity of the fluid would be equal to 0, which is the no slip
condition.
So, at the interface, at the solid liquid interface there would be solid in one side and static
molecules of liquid on the other side. So, if I draw it, if this is my solid part of it, if this is
the solid and this is the liquid. Then even if the liquid is flowing some the molecules of
the liquid, which are in contact of with the solid due to no slip condition, they have a 0
velocity. Now, when they have a 0 velocity, so the heat is going to get transferred from
the solid to this static liquid molecules by conduction, because, if you remember the
conduction is prevalent when there is no motion of the molecules in contact.
So, the mechanism by which heat gets transferred from the solid to the liquid molecules,
solid to the static liquid molecules is by conduction, but the molecules out of this,
molecules beyond this static layer they have a velocity. As they have a velocity now, the
heat transfer between two liquid molecules one which is static and one which is moving
the heat transfer here is going to be by convection. So, conduction and convection both
will exist in order to have convective heat transfer from the solid to the liquid.
So, we understand that an important part here is that you can never have convection
without conduction. So, you have to have conduction through the layer, through the
static layer of the fluid molecules which are clinging to the side of the solid and thereby
having gaining heat by from the solid by conduction. On the other hand, the other side
they are exposed to mobile liquid molecules and these, the interaction between the static
liquid and the mobile liquid the heat transfer is going to be by convection. So, when we
say that a convection from the solid takes place, we need to understand that the
convection is going to be preceded by conduction so the layer of the immobile
molecules. So, we can never have convection without conduction.
So, let us now concentrate on what is going to happen to the liquid or to the fluid which
is coming in contact with the solid. So, the molecules which are over here, there going to
get heated and, but the molecule at this point they do not know that a hot plate exists. So,
as I move to some more distance, the molecules at a slightly higher depth if this I call as
my depth, at a slightly higher depth would know that there is a hot plate which exists.
And the extent of the influence of the solid plate would propagate more and more into
186
the liquid as I move in the direction, in the x direction. So, this is my x direction and this
is my y direction. So, for small values of x, the effect of the plate in terms of a change in
temperature is going to be limited to a point very close to the surface, as I move further
with x the depth the penetration of the temperature front, will be more and more into the
liquid.
So, if I approximately join them together, I am going to get a layer which more or less
demarcates the range where you would expect a change in temperature. So, the
temperature over here is simply going to be T∞, the temperature over here is Ts. So, in
between this point and this point there is going to be sharp change in temperature from
T∞ to Ts and if you go beyond this, beyond this point the temperature everywhere is
going to be equal to T∞. So, the temperature profile probably would look something like
this, vertical because T∞ is not a function of y for the region beyond this imaginary layer
and here T the temperature is going to be a function of both x and y. So, inside this layer,
I would write it clearly again T is a function both of x and y.
So, further you are in terms of x, the temperature is going to be more further you go
away from the solid, the temperature will reduce. So, therefore, t is a function both of x
and y, but in here T∞ is not a function of either x or y. So, this line imaginary, line which
demarcates the change in temperature field and a constant temperature field due to the
effect of convective heat transfer from a solid to a flowing fluid is known as the
boundary layer, thermal boundary layer. You are aware of what is a hydrodynamic
boundary layer, which is defined in the same way so if I, just to recap some of these fluid
mechanics part of it which is going to be required for our study of convective heat
transfer. If this is my plate, where I have flow, which let us say it is the velocity is V ∞
then there would be an imaginary layer like this there, would be a layer like this in which
the velocity this is my x and this is y.
So, in here the velocity is going to be a function of both x and y, out here the velocity
this is a x component of velocity V x is going to be a constant and for a flat plate this
should be equal to V∞. So, for a so this layer which is in here, the velocity varies from 0
due to no slip condition at the solid it increases asymptotically and then becomes a
constant. So, this is a Vx over here which is a constant, but in here the velocity is a
function of both x and y. So, all the effects of convective or flow is confined within this
187
layer and most of these layers are, if these layers which are called boundary layers, why
they called boundary layers because they demarcate between two different types of flow.
When we considered the flow over a flat plate, whose temperature is equal to the
temperature of the fluid, no heat transfer is taking place, but the only thing that is
happening is that due to no slip condition, on the solid the velocity is 0. And as you
move away from the solid the effect of the solid will be felt lesser and lesser as you
move away from the solid and after a certain point, the velocity or the moving fluid will
not realize that there exists a stagnant solid plate below it.
So, the layer up to which the effect of the solid is filled by the moving fluid is known as
the hydrodynamic boundary layer and of course, the motion of the fluid molecules
slipping past one another near the solid plate, the property of relevance is; obviously,
viscosity. So, the viscosity is the one which transfers the presence, the effect of the
presence of the plate into greater depths of the fluid. So, viscosity in ρ they play a very
important role in defining what is the hydrodynamic boundary layer thickness, which are
going to, which are generally very small of the order of a few millimeters for normal
sized plates.
So, within this few millimeters, the effect of viscosity is important, outside of this
boundary layer the viscous effects are unimportant and the entire flow can be treated as
inviscid flow. A flow in which the viscosity can be assumed to be equal to 0 because
there is no momentum transfer in a direction perpendicular to the flow and therefore, it is
inviscid in nature. However, inside the thin layer close to the solid surface the effect of
viscosity cannot be neglected.
So, we have viscous flow inside the boundary layer, hydrodynamic boundary layer and
inviscid flow outside of the boundary layer. Now, from your fluid mechanics you
probably also remember that in an inviscid flow, can be explained by Euler’s equation.
In Euler’s equation is the one which is the simplified form of Navier stokes equation,
which Euler’s equation is for inviscid fluid. Where viscosity can be set equal to 0 in
Navier stokes equation, in order to obtain the Euler’s equation and we get Bernoulli’s
equation starting from Euler’s equation, but that is a separate story.
So, inside the boundary, inside the hydrodynamic boundary layer the flow is viscous. So,
the viscous transport of momentum as well as the convective transport of momentum,
188
both will have to taken, will have to be taken into account. Similar to hydrodynamic
boundary layer what we are discussing right now is thermal boundary layer. So, thermal
boundary layer is the region in which the temperature changes, temperature varies with y
from a value equal to Ts, all the way to the constant value of T∞.
Here I have the velocity changing from 0 to V ∞, here the temperature changes from Ts
which is the temperature of the substrate to infinity, beyond this point the velocity is
constant, beyond this point the temperature is constant. So, there is a similarity between
the thermal boundary layer, boundary layer and this is known as the hydrodynamic
boundary layer. So, in most in most of the cases, the thickness of the hydrodynamic
boundary layer which is generally denoted by δ and the thermal boundary layer which is
denoted by δt they are not equal.
So, in one hand you have δt and over here you have δ. So, this δ the thickness of the
boundary layer; obviously, as you can see from this figure is a function of x, here δt is
also a function of x; however, δ and δt may not be equal.
So, we make it situation in which, if this is the solid plate; you have the thermal
boundary layer and you have the hydrodynamic boundary layer. Or in you may have
situations in which the thermal boundary layer would be below and the hydrodynamic
boundary layer would be above.
189
So, let us call this is hydrodynamic boundary layer, and this is the thermal boundary
layer so here it is going to be TBL and HBL on for some very special cases you discuss
them in detail later on these two would coincide. So, when these two coincide here I
have for hydrodynamic boundary layer, this is delta thermal boundary layer thickness at
any location is given by is denoted by δt. So, in this case δ would be equal to δt. So,
when we would see that this special condition only appears when Prandtl number which
 Cp
is defined as Pr  =1
k
Now, why would that happen, let us let us expand this a little bit more, Prandtl number is
 Cp / 
Pr   
k k /  Cp 
I can write it as dividing both the numerator and the denominator by the ρ, the density I
can write it as k /  Cp and if you remember your fluid mechanics this  /  is known as
the kinematic viscosity and this k /  Cp which you have just seen is denoted by α. So,
this is known as the momentum diffusivity and this α is known as the thermal diffusivity.
Both have units both α and ν have units of m 2/s, which is the same as the diffusion
coefficient that you probably have heard of which is the diffusion coefficient of A and B
this also have units of m2/s.
So, another important observation which would not probably with related to heat
transfer, but you are going to come across this many times is that conceptually α, ν and
DAB are all similar and all will have units of m 2/s. In one case this refers to momentum
transfer, this refers to heat transfer and this refers to mass transfer. So, conceptually there
is not much difference between these three transport processes, heat, mass and
momentum transfer and they in there at some point of time would be the base on which
the unified treatment of heat, mass and momentum transfer can be undertaken. That you
are going to study in a separate course which is transport phenomena that looks at the
fundamentals of all these transport processes.
But coming back to convective heat transfer or coming back to the nature, to the growth
of these boundary layers, what we see is that the value of Prandtl number which is  / 
they simply tell us about momentum diffusivity and thermal diffusivity. So, the growth
190
of these layers they are strongly dependent the TBL, the Thermal Boundary Layer would
strongly depend on the thermal diffusivity. How fast, how easily the temperature front is
getting into the moving fluid and the for the case of hydrodynamic boundary layer its α
which is defining, how the hydrodynamic boundary layer is growing.
So, when numerically these two are equal both the HBL and the TBL would grow
together, would grow at the same rate therefore, the value of Prandtl number equal to
unity which appears only when momentum diffusivity and thermal diffusivity are equal.
This is a special case which would let us take the value of the thickness of these two
layers to be identical. So, this is the background which we are going to, which we are
going to utilize in deriving some of the equations of convective heat transfer.
So, we understand here that conduction and convection both exist for the case of
convection, but conduction can be a standalone process, which do not require the
presence of convection. Conduction mostly, conduction when it happens inside a solid
that is no question of any movement of the molecules, no net movement of the molecules
and therefore, conduction is specified by 0 velocity or no velocity. Whereas, convection
you have to have a velocity which could be imposed, that occurs in most of the industrial
processes or it could be without the imposition of a velocity it is there because of the
presence of a velocity.
Because, in the presence of a temperature gradient which induces a difference in

buoyancy a change in the value of the density are so therefore, have buoyant force would
make the fluid in contact with the solid, hot solid rise and that is what is known as a
natural convection.
Before we take this, slightly further and do of mathematical treatment of convection.

Next another concept which I would like to introduce, a new I will follow it up with the
in the next class with further details. In many of these cases the equations can be
obtained, the governing equations can be obtained if you assume a small control volume
through the faces of which heat mass and momentum can enter the control volume. So, I
can assume in a, in a free space a cuboid of side, size xy z . So, through so;
obviously, this cuboid will have 6 faces and truth is 6 faces the, let us say the mass is
allowed to come and heat is allowed to come as well. So, I am going to first write or
191
draw this cuboid and try to identify mentally, what are the process, what are the ways by
which, let us see heat can enter into this control volume.
So, I am going to write the physics of flow into the control volume and the associated
energy which comes into the control volume. So, this difference equation when I write
and when I divide all sides by xy z what I do is I convert the difference equation,
which is a statement of the physics of the situation to a differential equation. In this
differential equation can then be integrated with appropriate boundary conditions to
obtain either the velocity profile or the temperature profile.
So, this kind of approach where a small cuboid is assumed in the flow space, the
difference equation written convert it to differential equation and then solved is known as
the shell momentum balance or shell heat balance or in the case of mass transfer it is
known as the shell species balance. So, the first thing that one should do in order to
derive all these equations or all these concepts is defined as a shell. So, let us try to
define a shell and identify through the faces what is going to come in terms of energy
into the space, into the space that I have defined over here.
So, let us just draw, so this is my coordinate system and I am going to draw the shell. So,
this is the one which I have, let us say this is my x, this is my y and this is the z. So, as
you can clearly see this is x , this is z and this is y , this is the space which I have
defined and I have flow of a liquid in all possible directions which approach in this.
192
So, this point is, the coordinate of this point is x, y and z and coordinate of this point is
x  x , y  y and z  z . So, I have 6 faces in this case, the face which is

perpendicular to the x direction; that means, this face, which you do not see, this face is
known as the x face. So, the x face is perpendicular to the x direction, its area as you can
clearly see the area is going to be yz . Similarly, y face will have area, y face is the
face below the bottom face which is this one, the y face will have area x z and this z
face which is the one which is see over here, this z face will have an area of xy .
So, what we would assume is that the fluid is going to enter through x, y. So, the fluid
enters through x, y and z faces and leaves through faces at x  x , y  y and z  z .
So, this is what the fundamentals of shell moment, shell heat balance is. So, you have
this shell of dimensions xy z which is situated in a flow field and the flow is going to
come, it would enter through all these faces in a when and leave through the other sides.
That means, they what when it can come through x it leaves at x  x , y and y  y , z
and z  z .
It is a three dimensional flow so, there is going to be components of velocity as Vx , Vy
and Vz and the temperature is going to be a function of x, y and z as well. Whenever

liquid enters into the control volume, it will carry with it some amount of energy and
when it leaves it is going to carry some amount of energy. So, some amount of energy is
coming into the control volume, some of it is going to live in the control volume, the
coming to that of the energy to the control volume is through three faces, it is going to
leave through the other three faces.
So, there could be as a result of this process a net amount of energy, which is added to
the control volume this net amount of energy could also reduce. So, in that case I am
going to simply use a minus sign, but let us assume that some amount of energy is added
to the control volume, now this control volume can also do some work or some work
maybe done on it.
So, that is possible since we are considering, taking into considerations all possibilities
we should also consider that the control volume can do some work or some work can be
done on it. So, the net amount of heat which you add and the amount of work that this
193
system does or it is being done on the system, this sum total of this must be equal to the
time rate of change of internal and kinetic energy of the system.
So, if you recall first law of thermodynamics what I have stated. So, far in terms of the
control volume is nothing, but the statement of first law of thermodynamics for an open
system. Where all effects are considered, the energy which comes with the flowing fluid,
it will have a thermal energy component; it will also come with a velocity. So, there will
be some kinetic energy component and an internal energy component, kinetic energy is
because of the velocity of the fluid stream, internal energy is because of its temperature,
whatever be its temperature.
So, we would like to write all those terms containing kinetic energy and internal energy
through all the 6 faces, they are going to give me the net heat being net energy, both
kinetic and internal being added to the control volume. We will also have to take into
account whether the control volume does any work or some work is being done on the
control volume, that is going to be another component of the difference equation. As a
result of this the total energy content, internal and kinetic, total energy content of the
control volume will change with time.
So, we are not restricting our self to steady state, we also allow the energy and the total
energy can change which time inside the control volume. So, when I express in that
terms, what I am stating is nothing, but the first law of thermodynamics, where all effects
are considered and from this difference, difference equation when we think about, when
we take all the appropriate terms into account for example, let us see work done on the
system or by the system. So, what are the forces against to which what can be done, one
obviously, is a body force for example, a gravity which acts on the entire volume of the
control volume.
So, gravity is a body force which for against which the control volume may do some
work, the other forces are not everywhere, not acting everywhere on the control volume,
but acting on the surfaces, so surface forces will also have to be taken into account. So,
what are the, for most what is the most common surface force is pressure, so the pressure
force is acting on the control volume. So, these are two examples of the forces which can
operate on the control volume, one is a body force the other is a surface force. So, we
have to identify the most common body forces and surface forces plugged into the
194
equation that I have just described, and then try to see mathematically; what is the end
result of it.
So, the end result of it should give us an equation which is an energy equation, which
should have embedded into it both conduction, convection the work done and as a result
of which the time rate of change of temperature of the control volume. So, that would
give me the complete energy equation, which I will be able to use for a specific
application, cancelling the terms which are not relevant for the problem that we are
dealing with and get to a simplified form of equation. The same way we have obtained
the equation of conduction in the case of conductive heat transfer. So, I would stop here
today, but what I have discussed is about conduction, the thermal boundary layer the
velocity boundary layer and the concept of shell heat balance.
And while describing shell heat balance I have use the concept of first law of
thermodynamics for an open system and the shell that I have defined of size xy z
which is fixed in space and through all the faces the fluid is coming with some internal
energy and some kinetic energy. And some work is being done on the system or by the
system and the work is being done either against gravity, which is the example of a body
force or some pressure forces which are acting only on the surfaces of the define control
volume. When I sum all of them together what I get is the total, the time rate of change
of internal and kinetic energy is of the control volume of the shell that we have to just
define.
So, we will write down the terms in the next class and see how beautifully this concept
would result in a compact differential equation for temperature and this for the spatial;
that means, x, y, z variation of temperature, as well as the temporal variation of
temperature, that means how the temperature changes with time. So, my aim is to obtain
an expression of t as a function of x, y, z and time that is what I would like to get out of
this energy equation. And once I have that, then I am going to transform this equation for
the case of convective heat transfer to see whether or not I get a compact expression for
h, the convective heat transfer coefficient the engineering parameter of interest that
engineers would like to find out before designing any equipments.
So, starting from fundamentals through some, little bit of mathematics I would like to get
a differential equation, from the differential equation and expression of each which can
195
be used by practicing engineers. So, that is the whole chain which I would like to cover
in the, in some of the future classes.
196
Heat Transfer
Lecture - 17
Equations of Change for Non-isothermal Systems
I would like to introduce a concept which we started discussing in the last class, which is
called the shell balance. Now, this shell balance can be of momentum, it can be of energy
it can also be of a species. When we do a shell momentum balance we get a governing
equation that describes the change in velocity as a function of x, y, z and time. When we
write the shell heat balance we should get what is known as the equation of energy.
Similarly, if I write it for a species, which is let us say reacting with another species in a
flowing fluid field, then we would get the species conservation equation which is going
to be very important in mass transfer. Let us try to concentrate on how we can write a
shell heat balance, the trick is to define a shell of let us say some size del x, del y, del z
which is fixed in space.
So, a cuboid shape of size del x, del y, del z we have 6 faces: one is the x face which is
perpendicular to the x direction, one is the y face which is perpendicular to the y
direction and the other is going to be that z face which is perpendicular to the z direction.
Each of these faces will have areas associated with them and through these areas heat
energy can come into the control volume. Because of which the internal energy of the
box would change. When we talk about energy will we are not going to talk only about
the internal energy, we also must consider the kinetic energy. So, a fluid may come
through the x face with certain velocity and therefore, certain kinetic energy, at a
temperature which is different form the temperature of the fluid contained in the box.
Therefore, the entering fluid will have some internal energy and some kinetic energy
associated with it. So, it will enter the x face and will leave the x plus delta x face,
similarly it would come through y, leave at y plus delta y, come at z and leave at z plus
delta z.
197
Let us once again go through the equation of change for a non-isothermal system, here I
have drawn this box which is del x, del z and del y. These are the x, y and z directions
and the x face which is this one is perpendicular to the x direction and therefore, its area
is y z . Similarly, the y face which is perpendicular to the y direction would should
have area of xz ; and the z face would have an area of xy . We are going to now
write what would be the form of the energy equation for a system of size del x, del y, del
z which is which is fixed in space and this is our coordinate system.
198
So, let us see how the equation of change would look like for a non-isothermal system.
When there is a flow, some amount of energy comes with the flow and when I talk about
energy I speak about both the internal and the kinetic energy. So, some internal and
kinetic energy can come into this volume element by convection and it is going to go out
again by convection, from the x plus del x, y plus del y and z plus delta z faces. Let us
say I have a temperature difference, which exists in the x direction; then obviously, I am
going to have some flow of heat through conduction through this x face. So, a difference
in temperature either in x or in y or in z even if the fluid is still, would give rise to
conductive heat transfer. So, if there is a temperature difference, there is going to be a
conductive heat transfer. I have combined the terms together to write it in the form of net
rate of heat addition to the volume element by conduction. So, this takes care of all the
heat that comes to the system by conduction or by convection. However, there is one
missing term that I should consider at this point, that is the rate of work done by the
system on the surrounding; and hence, the total energy of the system should reduce, and
that is why we have a minus sign. Had this been a case of work being done on the
system, then this sign should be positive. So, what I have written over here is nothing,
but the first law of thermodynamics, for an x and since I am allowing fluid to enter and
leave, this must be for an open system.
Now, from this generalized energy equation one should be able to deduct the commonly
available equation for kinetic energy of a system. And therefore, what you would left out
with is the energy equation where we are only considering internal energy which is
manifested by a change in temperature.
199
So, first of all let us see what is going to be the rate of accumulation of internal energy
and kinetic energy within the system, which is defined as del x del y del z. So, this must
 1 2
be  U   v  where, U is the internal energy per unit mass and v is the velocity.
t  2 
This internal energy per unit mass, here I have multiplied it with rho. So, this becomes
internal energy per unit volume. So, if I would like to find out what is the total rate of
accumulation of internal and kinetic energy within x, y, z this must be multiplied by del
x del y del z, which makes it the rate of change of internal energy for a system, whose
dimensions are del x del y and del z. As long as my U is defined as internal energy per
unit mass, rho is the density.
 
Therefore,   U   xy z  would simply give us the rate of accumulation of
 t 
internal energy within del x del y del z and similarly the same logic will also be applied
1
to  v 2 . The next one is, I am going to write would be the convection of internal energy
2
and kinetic energy into the element.
200
In this figure, some amount of let us, this is my x face, which has an area of y z , and
some amount of fluid is going to enter through this x face. Let us say, this velocity with
1 2
which it comes into the control volume is vx . So, when you multiply  v , with
2
 y z  vx it is going to give you the kinetic energy per unit time. So, together this
whole term gives is evaluated at x on this face. So, it comes in through the x face and
goes out through the face at x plus del x.
 1 2
So, the entire thing is going to be at  y z   U   v vx , all evaluated at x plus
 2 
delta x. So, the amount of energy, internal and kinetic energy in due to convection. So,
when I consider the net energy balance due to convection, the equation would be:
 
 1 2  1 2  
y z  U   v vx   U   v vx 
   2        2     
 at x at x x 
So, this is what the expression for in minus out by convection through the x face would
look like. Now, I can write the same thing for the y face, the only thing which will be
different here is the y face has an area equal to xz and since we are talking about y
201
then the vx must be replaced with v y and instead of evaluating it at x, it is going to be
evaluated at y and y plus delta y, similarly for z face this area is going to be xy , vx is
to be replaced by vz and everything else will remain same.
So, in total to consider the convective flow of internal and kinetic energy into the volume
element, del x del y del z, I will have 6 terms. Two terms each for x, y, and z face would
give me total of 6 terms that would signify what is the total amount of heat which comes
into the system by convection.
As I have explained before, we must think about net rate of heat addition by conduction,
I would not say heat here, energy because I also have the kinetic energy to take care of,
net rate of energy input by conduction. And here I am going to express it in terms of the
component of heat flux in the x direction.
So, qx is the heat in per unit area per unit time. So, I must multiply it with the
appropriate area since it is x face this must be equal to y z . So, this denotes the heat
that comes to the control volume through the x face and the one that goes out through the
x plus delta x face. So, these two terms together one at x and one at x plus del x
multiplied by y z , together they tell us about the net rate of heat addition by
conduction through the x face.
202
Similarly, I am going to have the y face which will have xz , the area and the heat
flux is are going to be q y , at y minus q y at y plus delta y and for the z face it is going to
be xy times qz , evaluated at z minus qz at z plus delta z. So, these 6 terms again, 6
terms for each of the faces would tell me about the net energy input by conduction and so
I have taken care of the convection and I have taken care of conduction. So, what is left
is work done by the system on the surrounding.
Now, work done as we all know it simply, it can be against volumetric forces, volumetric
forces that which are acting on the entire volume of the volume element. The common
example would be gravity and the second one is against surface forces, surface forces
which could be against pressure, which could be against viscous forces. So, these two are
again the common examples of surface forces.
Now, I am not interested in work done, because work done is, force times distance in the
direction of the force. So, what would be rate of work done that is the time rate of work
done, it would be force times distance by time. So, the work done is given by:
  xy z   vx g x  v y g y  vz g z  . We know that, distance by time; obviously, would
give you the velocity. So, this is what you are going to get for the rate of work done. So,
rate of work done would simply be expressed as force times velocity. So, let us quickly
203
write the expression for the force against gravity forces, so rate of doing work against
gravity would simply be equal to minus since, it is against work is done against gravity.
So, v x times g x velocity and this is acceleration due to gravity plus v y g y plus v z g z.
So, when you, when you see this equation you would be able to see that it is the force in
the x direction which is v x, del x del y del z times rho and this is this is multiplied by g
x. So, this totally gives you the rho del x del y del z times g x is the force, because this is
mass, this is mass per unit volume this is volume, and this is the acceleration. So, this
gives you the force and force multiplied by the velocity in the appropriate direction. So,
v x, v y, and v z would give you the rate of work, rate of doing work against gravity.
So, again what is going to be the form for against pressure, it should be the area on
which let u say the x face del z times P v x, evaluated at x plus delta x minus P v x at x.
So, this is one term, plus del x del z, p v y at y plus del y minus p v y at y, this is going to
be the second term plus del x del y, p v z times z plus del z minus p v z. Look at these
terms one more time and see what the mean, pressure is force per unit area. Whatever be
the pressure at x plus del x is multiplied by the appropriate area which is del y del z, to
give to give us the force in the x direction acting on the control volume, acting on the
volume element at x plus del x and we understand that the rate of work done is force
times velocity. So, the pressure-work is given by:
     
 y z Pv Pv
  x x  
  x z   Pv 
 y y   xy   
Pv Pvz  Pvz 
   
 x x x   y y y   z z z 
So, these 6 terms together would give us the work done against the pressure forces by the
volume element del x del y del z. What is remaining here is the work done against
viscous forces, now work done against viscous forces this I am going to neglect for the
time-being because work done against viscous forces is something like solid friction. So,
what happens when you work against friction forces, you are pulling an object over a
rough surface. So, you have to overcome the frictional forces exerted by the rough
surface, as a result of which there is going to be heat generation in it and any work that
you do that in order to make that block move over a rough surface, is going to be
converted into heat and it will change the energy of the of the system. Similarly, when
fluid flows specially at high speed through a small duct, there is going to be tremendous
velocity gradient which is present.
204
So, let us say I have a jet which very thin, and the fluid is coming at a very high velocity.
So, the velocity is large and if the velocity is large and the gap is small, then the velocity
gradient would be very large, and we understand that the viscous force is related to
velocity gradient. The shear stress is mu times velocity gradient. So, if the velocity
gradient is large or if the viscosity is large, in that case you will have a strong force that
you need to overcome in order to make the fluid flow through that thin conduit at a very
high velocity. If that happens then you do substantial work against the viscous forces and
whenever you do that kind of work against viscous forces the temperature will increase.
And that increase in temperature which is obtained at the expense of work done by the
system must be considered for any form of energy equation.
However, this is only relevant in some special situations, we do not get the heat
generation due to viscosity in many of the practical problems. As you would see it
requires high velocity gradient and very high viscosity. So, what are the places in which
they become relevant, when a rocket reentered earth’s atmosphere its velocity is very
large, the atmosphere is still, but the rocket is coming down with a very high velocity.
So, near the boundary layer, formed close to the rocket, the velocity changes from that of
the rocket which is very large, to velocity equal to 0 which is the velocity of the
atmosphere. So, this thinness of the boundary layer and the very high speed of the rocket
at reentry would ensure that the frictional heat generated is tremendous. And that is why
you would see that the rocket comes almost like a red-hot object and there has to be
special protective arrangements to ensure the safety of the astronauts inside the rocket.
So, that is an extreme example, in some cases viscous polymer is extruded by making it
flow through a very thin gap, if that is the case then the viscosity is high, the velocity is
large as you would like to have higher throughput of the polymer when you are making a
sheet out of it. So, the velocity combined with the high viscosity of the polymer ensures
that you cannot neglect viscous dissipation. However, we will neglect viscous heat
dissipation, wherever the viscous heat dissipation is irrelevant; but I will tell you that
how to incorporate additional terms into the energy equation which would take into
account the viscous heat generation.
So, if you look at your textbook and look at the full form of the energy equation, you
would see there are a bunch of terms which have which are multiplied with mu. So, the
easiest way to identify which term of your energy equation in your text relates to be,
205
relate to viscous heat generation, is to look for terms containing mu. If in your problem
the viscous heat generation is negligible drop the entire set of terms containing mu and
what you would have is the energy equation that we are going to use for most of the
realistic applications. So, since those terms are complicated, I am dropping them for the
time being, but making you aware that in some special situations you need to add them to
ensure that your energy equation is complete.
206
Heat Transfer
Lecture - 18
Equations of Change for Non-isothermal Systems (Contd.)
We will continue with our study of the Derivation of the Equation for Non Isothermal
Systems. This we have started in the last class and we have proceeded by defining a
control volume of dimensions xy z . And we have identified through each of the six
faces that is at face at x and at x  x , y and y  y , z and z  z , some amount of

energy comes into this defined volume by means of convection and by means of
conduction. So, whenever there is convective flow into the volume element that we have
defined, that flow carries with it some amount of internal and kinetic energy and its
going to come at x and it is going to leave at x  x .
So, we have identified three pairs of terms, one showing the amount of energy both
kinetic and internal, that comes in at some face and leaving at x  x , leaving at y  y
and leaving at z  z . So, these 6 terms together would give us the total amount of
energy that is being added to the volume element due to convection. Similarly, we have
also identified the conductive energy which flows into the defined volume element
because of a difference in temperature. So, that is no velocity is involved since
conduction does not require any movement of the medium.
So, we have expressed the heat flux at each of these locations, such as qx , q y , qz those
three are the in terms that is the flux and the one that goes out at x  x is simply denoted
by qx x x , qy y y and qz z z . So, the net amount of thermal energy which is being
added to the defined control volume, the volume element that you have defined should
be, if you just consider the x face is the amount of energy that comes in per unit, per unit
time would be qx which is the flux, multiplied by the area.
And the area of the x face is simply yz . So, the amount of energy, thermal energy that
is be internal energy that is being added to the volume element would simply be
qx x yz . And the going out term it will be  qx x x yz . So, these two terms would
207
give us the conductive energy which is being added to the volume element through the
face at x and x  x .
Similarly, I should be able to write what is the energy be added through y and y  y
and z and z  z . So, the algebraic sum of these 6 terms should give us the total amount
of internal energy which is being added by conduction to the volume element. So, I have
6 terms for convective flow of energy and 6 terms for conductive flow of energy, that
together would give me the total amount of energy being added to the system.
In this system we have to identify that it is an open system such that the fluid is allowed
to enter and leave to the volume element. So, that gives us the total amount of energy
being added to the system and if we think about the first law of thermodynamics we
would also have to take into account the work done by the system or work done on the
system. Work done by the system is going to result in a reduction of its energy.
So, it would come with a minus sign in front of it and work being done on the system,
since it increases the energy of the system, so, it is going to come with a plus sign and
the work done can be against different forces. The forces can be categorized into two
distinct groups; one is a volumetric force, the force which is acting on every point inside
the volume element.
So, a common example of that would be gravity and it could also be done against surface
forces and one of the surface forces which we can easily identify is the pressure. So,
when we talk about the work done against volumetric forces or against surface forces,
the work done would be simply force times to distance and the force, the rate of work
done would therefore, be force times distance by time and since distance by time is
velocity, so, the rate of work done against or by external forces be volumetric or be its
surface forces, would simply be the force multiplied by the velocity in that direction.
So, when you talk about the force being exerted due to pressure on the x face, would
simply be p at x pressure evaluated at x multiplied by the area, the area of the x face
being yz . So, this is the force p at x multiplied by yz , this is the force and it has to
be multiplied by the component of velocity in the x direction which is vx evaluated at x.
208
So, the entire work done against pressure forces would simply be px evaluated at x
multiplied by yz , multiplied by the component of velocity which is vx , also evaluated

at x. And the work done on the other side at x  x would still be the same, except the
pressure is now evaluated at p at x  x and the velocity is evaluated at x  x .
In a similar fashion I should be able to find out what are the other 4 terms for example,
the y term would be p y multiplied by the area of the y face which is xz multiplied by
the component of velocity in the y direction which is v y . And the one that is on the
y  y face would simply be p y at y  y , multiplied by xz multiplied by v y ,
evaluated at y  y .
So, these 6 terms together give us; will give us the force, the work done against the
surface forces. Similarly I have shown in last class the expression for the work done
against volumetric forces, I have purposefully omitted the viscous dissipation term. That
is the force, that is the work to be done against viscous forces, as I have said the viscous
forces, that viscous dissipation or the work done against viscous forces, they become
prominent as compared to the other terms under some specialized conditions.
The specialized conditions generally refer to, if the viscosity is large or if the velocity
gradient is large. So, the velocity has to be large and the length scale over which the
 vx
velocity changes is small, such that , if that is a velocity gradient we are talking
y
about this has to be large and μ has to be large, the viscosity has to be large.
So, if those two conditions are satisfied then the viscous work done against viscous
forces, which generally is manifested by a change in temperature; since it is a dissipation
function, it is going to give rise to a change in, give rise to energy released and therefore,
the temperature will also change.
And this term which takes into account the work done against viscous forces would only
be important in two or three cases as I have mentioned it could be for reentry of a rocket
or for very high viscous fluid when it is flowing through a small conduit, a thin conduit.
Such that, the velocity may not be high, but the distance over which the velocity changes
209
is very small, which would be which relevant in an upcoming area of fluid mechanics,
heat transfer and application which is known as microfluidics.
Where the structure of the of the system in which this change in velocity is occurring is
very small, they are of the order of tens of microns or maybe hundreds of microns. So,
even though the velocity is small since, the area since the length scale over which the
velocity change is of the order of microns, the velocity gradient itself is large.
So, in some very high speed extrusion of polymers reentry of rockets or in some micro
fluidic systems, this viscous dissipation can lead to significant changes in energy and
those terms must be included in the energy equation. But, I would not derive those terms,
I will show you from the textbook what would the energy equation look like and you can
easily identify the terms which correspond to viscous dissipation that we are not
considering at this point of time.
So, the result of all these heat input, all these energy coming in by convection and
conduction and the energy change as a result of work done by the system against
volumetric forces or against surface forces. When you combine all of them together and
if the system is at unsteady condition, if the system is not at steady state, then the
algebraic sum of all these terms should result in and change of the net energy content of
the defined volume element. So, the change in energy content both internal energy and
kinetic energy of a volume element in an open system is a result of several factors, the
conductive flow of energy the convective flow of energy and all terms that refer to work
done by or on the system.
210
So, this much we have covered in the last class and when you see this equation, the one
that I have shown over here is what I have described.
Rate of accumulation of internal and kinetic energy= (Rate of IE and KE in by

convection) – (Rate of IE and KE out by convection) + (Net rate of heat addition by
conduction) – (Net rate of work done by the system on the surrounding)
And since it is by the system that is why it comes with its preceded by a negative sign so
which is nothing, but the first law of thermodynamics for an open system.
211
So, after that I have defined the volume element in here, as this is the volume element
and I have identified what are the terms for convection (through x face)
  1   1  
yz vx  U   v 2  x  vx  U   v 2  x x 
  2   2  
I have also identified what is the energy input by conduction, so these are the 6 terms
that I have referred to.
Net rate of energy input by conduction:
yz  qx x  qx x x   xz  q y y  qy y y   xy  q z z  qz z z 
212
And then I have also identified the work done against volumetric forces for example,
against gravity and against pressure, which would be these terms.
Rate of work done against gravity =    xyz   vx g x  v y g y  vz g z 
Rate of work done against pressure=
y z  pvx  x x   pvx  x   xz  pv y  y y   pv y  y   xy  pvz  z z   pvz  z 
213
So, when I combine all these terms together and divide the, divide both sides by xy z
and then, take the limit when all of them approach 0, what I get is a reduced form of the
equation which is
DT  p
C  k  2T  T    v   v
Dt  t 
I have omitted large number of steps in between, which would be available in your
textbooks so I am not going through all the steps in the class. What I would like to do is I
would like to clarify the genesis of each of these terms as I have described as a result of
conduction, convection work done and unsteady state effects.
So, what you get out of that is an energy equation, now from that generalized energy
equation I subtract the mechanical energy equation and what I get is known as the heat
transfer equation, the energy equation where only internal energies are taken into
account.
D
So, if you look at the terms over here, the first term this is the unsteady term, this
Dt
has a special meaning and is known as the substantial derivative. This is a special
mathematical function which simply tells you this is going to be
DT  T T T T
  vx  vy  vz
Dt  t x y z
D
So, this is the expanded form of the , this one substantial derivative which is defined
Dt
in this way. So, here if you see this term is definitely the unsteady term, what about these
terms vx , v y , vz and with the temperature gradient in there. So, whenever you have
velocity associated with any term in energy equation, that must refer to convective flow
of heat, because only in convection you have the velocity associated.
So, conduction takes place only when the medium has a velocity of its own. So, if you
T
concentrate on, if you look at the first term on the left hand side, the first part of that
t
214
is the unsteady term. The other three terms all contain the components of velocity, which
are vx , v y , vz in a rectilinear coordinate system.
So, since they have velocity in that in its form so these three terms refer to the convective
heat transfer process. So, therefore, this whole term is unsteady, plus convection where
this term is due to the unsteady behavior of the system and the convection is manifested,
represented by these three terms. Now, what is the first term over here, I will call it a
second and this is the third term, the first term on the right hand side.
If I concentrate on the first term on the right hand side, it is going to be
  2T  2T  2T 
k 2  2  2 
x  y z 
These terms have the k, thermal conductivity and they talk about temperature gradient.
 2T
So, k is nothing, but, the gradient in the heat flux.
 x2
 2T   T 
Also, k k
x 2
 x   x 
  T 
= k
 x   x 
 q x  q y  qz
=  
x y z
 T 
Since by Fourier’s law  k   qx
 x 
 qx
So, this term gives me the  , the scalar component of the heat flux vector in the x
x
 2T  qy q
direction. Similarly, 2
would simply be  and this term would be  z . So, this
y y z
three terms on the right hand side, which I denoted by one is nothing, but it represent
conduction heat transfer or conduction energy transfer.
215
So, the left hand side is unsteady and convective heat transfer, the first term on the right
hand side is conduction heat transfer as I have shown here, the second term, the term 2 is
due to expansion effects. And if this is expansion effects it essentially tells you that what
would be the effect of this expansion effect, what would be the work done against special
forces and so on. And the third term which is v , this is known as the work done
against the surface forces which is the frictional effects, this is known as the dissipation.
And this dissipation is due to friction, this v is known as the dissipation function and the
form of this v is quite complicated and I did not derive it. So, therefore, the equation,
the energy equation, the complete form of the energy equation I did not derive all the
terms, but what I have done is, I have shown you the significance of each of these terms
in the overall scheme of things.
So, for most of the practical cases this third term can be neglected, the v , the
dissipation function or the energy change of a system due to viscous dissipation is
neglected and you can clearly identify the term by noting that you have a μ in front of
this.
So, all the terms which have the viscosity in front of them in the energy equation they
referred to the dissipation function and therefore, for most of the practical situations this
third term can be neglected. The second term may or may not be relevant, so the second
term will have different connotation in terms of its significance in the energy equation.
And let us see what it what would be relevant and when it is not relevant and we can get
back to an equation that we are more familiar with.
216
So, I am writing this equation once again, it is
DT  p
C  k  2T  T    v   v
Dt  t 
Where, C is the heat capacity and v = 0 for this case.
Now, if it is for fluids with constant pressure, which is a reasonable assumption in many
situations or if ρ, the density is a constant. Then a fluid is with constant pressure then this
part is going to be 0 and if ρ is a constant, then this part is going to be 0.
So, if ρ is a constant then δv would be equal to 0 which is nothing, but one form of
equation of continuity that you must have studied in fluid mechanics. So, if ρ is constant,
this part would be 0, if fluid is at constant pressure then this term would be 0 and this is
anyway to be equal to 0. If you use any of these conditions, then the equation would be,
this is the form of the energy equation which is mostly used for convection as well as
conduction.
Now, let us see what is going to happen if it is just if you have, if its convection plus
conduction both are present.
For both conduction and convection
217
T T T T k   2T  2T  2T 
 vx  vy  vz     
t x y  z C   x2  y 2  z 2 
If it is a conduction only case, then,
vx  v y  vz  0 (since in conduction, there is no velocity of the medium)
T k   2T  2T  2T 
So,     
 t  Cp   x 2  y 2  z 2 
So, I think you can recognize this equation now; this is the equation which is also known
as the heat diffusion equation that we have studied before
So, I will write it once again, the heat diffusion equation.
T k   2T  2T  2T 
    
 t  Cp   x 2  y 2  z 2 
One of the things one should always do is whenever you develop a new equation you
should always try to see if it reverts to the known form for limiting cases.
So, the equation that we have derived is a general equation, is energy equation where we
have taken into account the unsteady state, unsteady effect we have taken the effect, we
218
are considering the effect of convection, conduction work done and viscous dissipation
that we are not considering.
So, to this equation, now I am going to impose the condition that there is no expansion
effects or the density is constant, so I take care of the second term on the right hand side.
So, I am left with one term on the left hand side and one term on the right hand side, the
term on the left hand side is the unsteady state effect and convection together, the term
on the right hand side is only the conduction term.
What I have done after that is I will set convection to be equal to 0 by setting the velocity
is to be equal to 0, the velocity components to be equal to 0. And therefore, what I have
is the pure heat transfer due to pure conduction.
So, heat transfer by pure conduction we have seen before, we have derived before in
conduction which is the heat diffusion equation. So, in the limiting case of no convection
and no work done what we get is the heat diffusion equation. So, if you see this equation
once again, if this is for a one dimensional conduction only situation. Let us say the
temperature is a function of x only, it is not a function of y or z, then this term this
k  2T
equation would be .
 Cp  x 2
T k  2T
So, 
 t  Cp  x 2
This would be the equation for this and if I simplified it further it is one dimensional
conduction only situation, but at steady state. So, if it is a steady state case, then this
temperature is a function only of x, it is not a function of time. So, what you have then is
k d 2T
 0 . Temperature is a function of x only and temperature is also a function of
 Cp dx 2
time in this case, here temperature is a function of x only.
So, this can be cancelled and what I have then is
d 2T
0
dx 2
219
So, this must look very familiar to you now since, you have studied conduction. If you
have heat generation due to some electrical sources heat generation in the system, then
q
you are going to have +  0 . So this form of the equation you have seen before. So,
k
which is the one dimensional conduction only steady state situation with heat generation
and if you set heat generation to be equal to 0, this is what you are going to get and the
result of the solution of this would be T is a linear function of x.
So, one you can clearly see that starting with the most general and slightly complicated
expression, you can see this is the expression for where you have conduction, you have
convection and conduction both present. Which is this form, if you set the convection to
be equal to 0 what you get is the heat diffusion equation. If you simplify the heat
diffusion equation, by assuming it is a one dimensional conduction only case you get the
equation which is the equation for transient one dimensional conduction.
So, this is the case of transient, one dimensional conduction case and when you drop the
d 2T
transient make it a steady state you simply get this  0 , no need to use a partial
dx 2
differential since temperature is a function of x only. So, T is a linear function of x and if
it is a system with heat generation, you simply at the heat generation term where q is the
energy generated in the system per unit volume.
So, with this, that concludes the development of the heat diffusion equation for most of
the cases. And if you look at your textbook Incropera and David or any textbook will
have this energy equation in, I have to discuss the Cartesian coordinate equation, you
would see that in cylindrical coordinates and you would see that in spherical coordinates
as well.
So, the trick is to first identify whether you have cylindrical system or a Cartesian system
or a spherical system. Then look at the equation, look at the full form of the energy
equation and then depending on the situation at hand cancel the terms which are not
relevant.
So, if it is a conduction only case cancel all terms which contains the velocity, velocity if
it is a steady state case cancel the term which has time derivative of temperature. If it is a
220
situation in which viscous dissipation is not relevant drop all terms that contain the
viscosity μ. So, I will just show you an example which is s the full form of the energy
equation in Cartesian and in cylindrical and so systems. So, this is the equation that you
would see and it is available in your textbook.
So, if you look at carefully over here, this is
 T T T T 
 Cp   vx  vy  vz
 t x y  z 
So, you have a transient term and 3 terms which contain velocity, so they must refer to
convection. As I have shown you the entire left hand side is a combination of transient
effects and convective effects, when you come to the right hand side you have
  2T  2T  2T 
k 2  2  2 
x  y z 
This terms contains k, the thermal conductivity and I have shown you that these three
terms refer to conductive heat, the heat transfer due to conduction only and finally, there
2 2
v  v 
is a complex set of terms which are nothing, but the gradient square,  x  ,  y 
 x  y 
221
and so on. But the important point to note here is that all these terms they contain μ
explicitly.
So, since they contained μ without even knowing the full derivation, you know that these
terms that contain μ referred to viscous dissipation, the dissipation of energy due to
viscosity of the solid and as I said it becomes relevant only for high speed, high viscosity
flow or flow in small micro fluidic systems. So, for most of the practical purposes you do
not need to consider this term, these terms at all, you are going to consider only on the
left hand side and on the right hand side for convection, the left hand side for transient
and for convection.
So, that is the governing equation which we are going to use for all our treatment of
convection subsequently, similar to the rectangular coordinates we have the equation in
cylindrical coordinates as well as in spherical coordinates.
The significance remains the same, the left hand side contains a transient term end terms
which contains vr , v , vz so they must refer to convection. The right hand side, there
would be three terms, these three terms which contain k. So, they refer to the conductive
transport of energy, all the other terms that you see contain μ.
So, this entire group of terms they refer to viscous dissipation in cylindrical systems, that
can be neglected for most of the applications. And when you come to spherical
222
coordinates again the same thing transient, convection, convection, convection,
conduction, conduction, conduction and the rest of the terms contain μ. So, this entire set
of terms is nothing, but the viscous dissipation which can be neglected.
So, what we have seen in this class is the derivation, the simplified derivation of the
energy equation where all effects are considered. The energy equation that you have
derived starts from a very fundamental law, which is the first law of thermodynamics for
an open system, we have taken care of convection, conduction the work effects and
everything else.
So, this would be the starting point for our studies in convection. So, this is slightly
complicated concept, for I think once you go through the textbook and read and look at
the equations carefully. I am sure the concepts would be clear to you and if there are any
questions I will be more than glad to answer, answer them and once you master these
equations atleast the significance of the equations then the rest can move in a much
smoother fashion.
223
Heat Transfer
Lecture - 19
Tutorial on the Application of energy Equation
This part of the class is going to be a tutorial, where I would solve a problem and also
give you a problem to do it on your own. So, the problem that we are dealing with is a
very interesting one.
Let us consider 2 cylinders which are concentric to each other. So, we have one cylinder
and another cylinder enclosed in between and the gap between the 2 cylinders is very
small and there is a thin film of lubricant in between. You get this in practical situations,
whenever you have a bearing where one of the surfaces is moving with a high speed and
the other surface is stationary. So, to reduce friction whenever there is high relative
speed in between the two surfaces we use lubricants. Most of these lubricants are oils
which have high viscosity and they act in a way to reduce or minimize the wear and tear
of the moving system. Now, when you think of this situation it is an ideal case for
viscous heat generation, because we have a moving part at a high velocity a stationary
part and a very thin gap in between. So, the velocity changes from the velocity of the
moving part to 0 over distance which is very small. So, if you think of the gradient of
velocity it is v of the moving part minus 0, which is that of the velocity of the stationary
part divided by the distance between the two, which is very small. Therefore, this ratio is
going to have a large value, and I have told you that the viscous heat generation is
generally gradient squared multiplied by viscosity. So, this is potentially an ideal
candidate for viscous heat generation. So, fine we have viscous heat generation, but what
is; what are the implications?
Let us say the surface which is moving has some temperature Tb the base temperature
and the surface which is stationary has another temperature. So, whenever you are
choosing a lubricant for this specific purpose, you must make sure that the temperature
of the entire system is not going to exceed the rated temperature for a specific lubricant.
Every lubricant has a maximum temperature that it can withstand. If you go beyond that
224
temperature, the ingredients of the lubricating oil start disintegrating; thereby reducing
the efficiency of the lubricant.
So, we are going to analyze the situation to find out if we are correct in assuming that the
maximum temperature is going to be the temperature of the hot surface. And therefore, if
you know the temperature of the hot surface I can choose a lubricant that can withstand
that temperature.
So, the system looks something like this as I have drawn over here, these are cylindrical
systems. One inside the other and you are going to have the inner cylinder which is
stationary the separation between the 2 cylinders is b, and the radius of the cylinder is
equal to R and we are what we are saying is that this b the separation between the 2
cylinders which is very small as compared to the radius; and the intervening space is
filled with the lubricants. The temperature of the inner cylinder is T 0 and the
temperature of the outer cylinder is T b. And, we will assume that T 0 is greater than T b.
What, we need to find out is what is the temperature distribution that we have in the
lubricant? Am I safe to assume that the maximum temperature of the lubricant is going to
be T 0 and of course, the minimum is going T b. And therefore, my selection of lubricant
should be based on T 0.
225
The first thing that one can do which you probably have studied in your fluid mechanics,
that if the gap between two systems in a cylindrical coordinate is very small, then you
can convert the cylindrical system into a Cartesian coordinate system without
introducing any appreciable error. Therefore, this system can be analyzed without any
loss of accuracy by opening them up and making them as two flat plates. So, the
cylindrical system can be converted to a flat plate system where one of the plates is
moving, and the other one is stationary. We can only do that when the gap is very small
compared to the radius or in other words the curvature is going to be very small. For a
system with a large R and small gap this is a perfectly valid assumption for converting a
cylindrical coordinate system to a rectangular coordinate system, without any loss of
accuracy. This helps us because, we would always prefer a Cartesian coordinate system
for its ease of handling.
Let us say this is the one which is the top plate. So, this has a temperature as in the figure
equal to T b and it is moving with a velocity, which must be equal to R times omega
where omega is the angular velocity. So, this linear velocity here is simply V is equal to
R omega, and the temperature over here is T 0 temperature over here is T b.
So, you would expect this T b is more than T 0. So, you would like to see if we have the
maximum over here or that is not the case. Now, since the top plate is moving, and the
bottom plate is stationary, in your fluid mechanics you have probably read this to be as
known as the Couette flow. Couette flow is a one in which you do not have any pressure
gradient imposed on it, the top plate is moving and therefore, it is going to drag the liquid
below and this layer of liquid is going to drag the layer below it and so on.
So, the motion of the top plate is being transmitted through the liquid due to the viscosity
of the liquid. So, the velocity profile for a system like this looks something. So, this
velocity let us say this is my Z direction and this is my x direction. So, the velocity is
going to be only in the Z direction and it is going to be a function of x alone.
So, the motion of the top plate is going to induce velocities at different layers of the
lubricant present in between the moving part and the stationery part, and it is going to be
a linear distribution of velocity and the non-zero component of velocity is going to be in
one dimensional flow. So, v Z in the Z direction is going to be a function of x and is
going to be linear function of x.
226
This type of flow is known as the Couette flow and starting with an equation of motion
x
you would be able to obtain this velocity in the Z direction to be equal to V .
b
The z-component Navier Stokes equation in the absence of any pressure effects or
  2v  2v  2v 
gravity is simply going to be   2z  2z  2z   0 . Since vz is a function of x
 x y z 
 2 v z  2 vz  2 vz
alone. So,   0 ; and what you have than is  0 , then this is going to
y 2 z 2 x 2
give us vz  C1 x  C2 ; where C 1, and C 2 are the integration constants and the boundary
conditions are at x equals 0. That means, on the plate the vz  0 which is known as the
no-slip condition and at x equals b; that means, on this plate vz  V  R
So, these are the 2 boundary conditions with which you would be able to solve for C 1
and C 2 it is a linear distribution, and this would give rise to the linear variation of the
x
non-zero component of velocity which is v z, as V . If there is an imposed pressure
b
present in the system, then this straight-line behavior is no longer going to be valid, and
you will have a different type of velocity profile. So, this is the relation, which is going
to be useful to us when we solve for the energy equation. So, in this entire thing what we
have done is we have converted the lubricant system, from cylindrical to a Cartesian
coordinate and we identify that the velocity profile. So, let us see what is going to be the
energy equation for this case. You can also see that the motion is going to be in this
direction, and the energy is going to be transported in a different direction. And, this is a
case where viscous dissipation is significant. So, if viscous dissipation is significant in
here then the viscous dissipation form of the equation of energy must be used for a
rectangular system
227
T
First, it is a steady state system. So,  0 . Once again if you look at this is x and this
t
is z. So, I only have v z I do not have a v x. So, in this case v x and v y all are equal to 0.
So, if that is the case then in this equation this v x term would be 0 v y term would be 0
and you have v z times del T del z.
228
T
So, what I can write then from the left-hand side of the equation is that  C p vz would
z
remain on the left-hand side. Let us see what I have on the right hand what will remain
 2T
on the right-hand side. The right-hand side I have k , if you look at this, the
x 2
temperature is a function of x; so, the first term would remain in the expression. Since,
 2T  2T
temperature is not a function of y and z; we have: k 2  0  k 2 . So, the only
y z
 2T
remaining term over here is going to be k , then comes the viscous dissipation terms.
x 2
So, what are the viscous dissipation terms? Since, there is no x component of velocity,
there is no y component of velocity, there is a z component of velocity del v z, but that is
not a function of z. So, what I then have on this side from my energy equation is
2
 v 
 z  .
 x 
2
d 2T  dv 
Therefore, the entire equation then can be written as k 2
    z  . This is the
dx  dx 
dvz x
governing equation for the system that we are analyzing, what is ? My v z is V . So,
dx b
2
dv V d 2T V
my z  . Therefore, my equation becomes further simplified: k 2      .
dx b dx b
So, you integrate this simple ODE twice and you are going to get
2
  V  x 2 C1 x
T      C2 ; where C 1 and C 2 are the constants of integration. In order
kb 2 k
to evaluate this, you need two boundary conditions. The boundary conditions are: the
temperature at x = 0 is T 0, temperature at x = b is T b. Use these two boundary
conditions to evaluate C 1 and C 2.
What you would get is this form of the temperature distribution:
229
T  T0 x 1  x   x 
  Br   1 
Tb  T0 b 2  b   b 
Now, this B r is very interesting, this B r is known as Brinkman number. And, you would
V 2
see that Brinkman number is nothing but . Brinkman number essentially tells
k  Tb  T0 
you the extent to which viscous heating is important relative to the heat flow from the
imposed temperature difference, which in this case is simply equals to T b minus T 0.
V 2
So, it is clear that , you can divide this side by L square and the denominator
k  Tb  T0 
by L square and from the definition of viscous heat you can see that the top refers to the
viscous heat which is generated, where as the bottom is the imposed temperature
difference. So, they refer to the conductive flow of heat.
Now, let us see if we get back to something which we would expect. Let us say there is
absolutely no viscous heat regeneration, and only conductive heat flows from the high
temperature to the low temperature. So, for a case in which the two end temperatures are
known, it is steady state and there is no heat generation, the temperature distribution will
be linear. So, how do I see if our analysis is correct as I mentioned to you previously
always look at the limiting form of the solution and see if it reverts to a form which is
230
known to be correct. So, how do I make sure that there is no viscous heat generation? Let
us say that there is no velocity. So, if there is no velocity then this second term on the
right-hand side is equal to 0 and you get a linear distribution of temperature in the
medium. So, this part is for linear distribution, which is caused by conduction only and
the second one is due to viscous generation of heat. So, this Brinkman number therefore,
refers to viscous generation of heat in a system where the thickness is very small. There
is something more interesting to it; if Brinkman number is greater than 2, there is a
maximum temperature at a position intermediate between the 2 walls. So, find out what
is d T/ d x set it equal to 0. And, from this expression you would see that Brinkman
number if it is greater than 2, then you are going to get a maximum temperature at a
position in between these 2 walls.
So, if you do not have viscous dissipation then the maximum temperature is going to be
at T b, and the profile in that case would look like a linear distribution, but if you get a
Brinkman number greater than 2 then the profile may look like this. And therefore, the
maximum temperature would lie at a point in between the top plate and the bottom plate.
This is the interesting result, why is this interesting result, because you have chosen you
lubricant based on T b the maximum temperature that you think is present in the system
which is the temperature of one of the plates. But, due to viscous heat generation there is
a point where the temperature exceeds that of T b. So, the intermediate position
temperature if it exceeds the temperature of the upper plate, then it would be wrong for
you to choose the lubricant based on T b, rather you have to see what that temperature
would be from this expression and find out and specify the lubricant. So, in any
lubrication system, where there is viscous heat generation you must find out what is the
value of the Brinkman number. If the Brinkman number is greater than 2, you must do
further calculation to find out what you where is the maximum and what is the value of
the maximum temperature for such a system
So, a simple problem can give you so many new ideas about how to treat conduction.
How to how to use first of all how to use equation of energy? How to cancel the terms
which are not relevant? See the importance of conduction; see the importance of
conduction when there is flow present, but in a direction perpendicular to the heat flow.
Therefore, you are not considering convection explicitly; however, the presence of the
velocity in your equation or rather the presence of the velocity gradient in your energy
231
equation, tells you that you need to probe further in order to obtain how would the shape
of the temperature profile look like, more importantly, are you going to get a maximum
in the profile due to generation of viscous heat in between the top plate and the bottom
plate. And, the dimensionless number which tells you the importance of viscous heating
relative to the heat flow due to the imposed temperature gradient, which is known as
Brinkman number that plays a critical role, and you must find out what is the value of
Brinkman number before you choose a lubricant. So, from concepts to practical
applications, this is one example which I think would be interesting to anyone who is
studying convective heat transfer.
232
Heat Transfer
Lecture - 20
Nusselt Number of a Heated Sphere in Stagnant Air
So, we have covered the energy equation, its derivation, the significance of each of these
terms in the equations. And, I have shown you how to use that equation for a specific
example, in which the viscous heating was considered to be important for a planar
system. So, I would show you the energy equation once again. And, then solve a few
problems on it, after I explain the significance once again.
So, let us look at this energy equation, which I think now you would be able to see, and
as you can see the energy equation is presented for all 3 coordinate systems. So, the first
one that you see this is for Cartesian coordinate systems, this one is for cylindrical
coordinate systems, and the last one is for spherical coordinate systems.
233
So, the first one is in x, y, z and the temperature is a function of x, y and z and it could
also be a function of time. So, the transient behavior is also included in it. And, if you
look at the equation I have added this v term in all 3 equations, this v refers to the
viscous dissipation, the energy generation due to viscous heat, which in most of the cases
as I have mentioned before may be neglected. And it is easy to identify the viscous
dissipation term in energy equation by the appearance of viscosity μ with all such terms.
So, in the expanded form of the equation, where the complete form of v is included, you
should be able to identify them as referring to viscous dissipation energy generation by
viscous dissipation, because of the appearance of μ in all those terms. What you see here
as Q is the energy generated by some other means in the system. So, it could be
electrical, nuclear or some other form of energy, which is generated inside the system.
So, the significance of v should now be clear to all of you and the expanded form of v
for all these systems would be provided to you. And, they are also available in the text
book, which I did not include in these equations. One more point here is I never expect
you to remember this equation. So, whenever you need to use the energy equation, the
equation in this specific form would be provided in the question papers.
234
So, do not memorize this equation, but rather if the equation is given you should be able
to identify what are the significance of each of these terms, and how to resolve them.
How to neglect or take them into consideration based on the physics of the problem?
So, if I take the first equation, which is the energy equation in the Cartesian coordinate,
T
what you see as the first term on the left hand side this simply refers to the transient
t
nature of the process. So, had this been a steady state process this first term would
obviously, be equal to 0; so this first term would be equal to 0 for a steady state system
all the other terms the remaining 3 terms on the left hand side, they have velocities the
component of velocity vx , v y and vz associated with them.
So, they collectively refer to the convective transport of heat into the system. So, the
appearances of vx , v y and vz in their origin and their component form they refer to the
transport of energy by convection in the x direction, in the y direction and in the z
direction.
So, the entire left hand side of the energy equation refers to transient term, the first term
being the transient term and the second, third and fourth are the convective terms. When
you go to the right hand side the appearance of thermal conductivity k, essentially
signifies or identifies these terms that are associated with conductive heat transfer, heat
 2T  2T  2T
transfer due to conduction. So or or they refer to conduction.
x 2 y 2 z 2
So, the right hand side contains the conductive terms, the viscous dissipation term and
the energy generation inside the system. So together this provides the complete form of
the energy equation, which does not take into account the mechanical energy, because
this equation was derived from the first law of thermodynamics considering both the
internal and the kinetic energy. From that equation the kinetic energy equation was
subtracted to obtain the energy equation in this form, which does not take into account
the kinetic energy or the mechanical energy part of the equation.
So, these 3 are the equations, which we would use in all our problem solution. So, if you
look at the second and the third equation, the cylindrical and the spherical coordinates; in
235
the cylindrical coordinates you have the independent variable, independent space
variable as r ,  , and z. So, the radial direction, the angular direction and the axial
direction. So, they are associated with velocity in the r direction, velocity in the 
direction, and velocity in the actual direction. So, transient and convective terms in a
cylindrical coordinate systems, the right hand side refers to conductive heat transfer in
the r direction, conductive heat transfer in the  direction and conductive heat transfer in
the z direction.
And similarly, we have the dissipation function and the heat generation. One point to
note here is that obviously, since the coordinate systems are different, the expression of
v here and over here are going to be different and the complete expanded form of v for
different coordinate systems are available in your textbook.
And, when we come to the cylindrical coordinate systems we have r ,  , and  . These
three terms refer to convection, transient, the conductive heat transfer in the r direction,
conductive heat transfer in the  and in the  direction. Again, we have this v and Q .
So, I guess the significance of the terms in energy equation is now clear to you. So, what
you need to do what you should do in solving problems of combined conduction and
convection or convection alone is to see, which of these terms is significant. And, for the
problem at hand and then you can cancel out the terms which are not present which are
not relevant in that case.
So, if it is a steady state; obviously, this term would be equal to 0, if it is a one
dimensional conduction only case, then vx , v y and vz are 0. So, the entire left hand side
would be 0 for a conduction only case no convection. And, only these 3 terms would
remain. And, normally v is neglect v is neglected and if we may or may not have Q the
heat generated in the system.
So, for a one dimensional conduction only case, the only the first term will remain. One
dimensional steady state, conduction case, with no heat generation; the governing
equation would simply be
d 2T
k 0
dx 2
236
And, you know that this is going to give rise to a linear, temperature, distribution as you
have done in your conductional analysis. If in addition you also have Q to be present,
then the form of this equation be
d 2T 
k Q  0
dx 2
And, you should be able to solve the equation with appropriate boundary conditions as
defined or as mentioned or as you can see in form from the statement of the problem.
So, this once you follow a methodical approach in identifying the equation, in identifying
and cancelling terms which are not relevant in the equation. It should give rise to a
compact equation describing the temperature distribution in the system, either in
presence or absence of conduction or convection. And, mostly in absence of viscous heat
dissipation and without heat generation by some other means.
So, this is how you would solve any problem in heat transfer, where both conduction and
convection are present. So, it should not be difficult, the equation looks too long and
complicated, but once you understand their significance the solution of the problem
should be in theory very simple. At least you would be able to obtain the governing
equation and identify the boundary conditions.
In some cases the analytical solution is possible, you would be able to integrate the
governing differential equation without much of a trouble and then you get an analytic
expression for temperature as a function of position and or time. In some cases the
analytical solution would not be possible. So, we may have to go for a semi analytics
solution and in many cases a numerical solution.
So, there are several packages available for numerical solution of the heat diffusion
equation and we will discuss them as we move along. One more point which is should be
apparent when you look at the energy equation carefully is the appearance of vx , v y and
vz , the 3 components of velocity. Since, you have the velocity component embedded into
the energy equation it requires the complete expression of vx , v y and vz .
237
In other words you need to solve the Navier-Stokes equation or you need to solve the
momentum equation, as you have done in fluid mechanics to obtain an expression of vx ,
v y or vz . So, once you plug the functional form of vx into the energy equation, then the
only dependent variable would be temperature and the independent variable should be x,
y, z and possibly time.
So, the presence of the velocity terms in the energy equation mix the energy equation
coupled with the momentum transfer equation. So, a pre requisite for the solution of
energy equation, in cases where convection is important; that means, where you have
these velocity terms is that you need to know the velocity expression from your
momentum equation solution.
So, that way heat transfer is always coupled to the momentum transfer, but the reverse
may or may not be true, if you look at your Navier-Stokes equation or equation of
motion the momentum equation, in your fluid mechanics textbook you would see that
there are no terms containing T in there. So, that equation is independent of temperature
provided the thermophysical properties that appear in the momentum equation, namely
the density and the viscosity, as long as they are they are not affected by a change in
temperature, as long as they are constant for the temperature range of operation, the
momentum equation can be solved independent of the energy equation.
So, that way momentum equation is uncoupled from the energy equation, but the energy
equation is coupled to the momentum equation. So, there is a one way coupling between
momentum equation and energy equation as long as the thermophysical properties
viscosity and density remain unaltered during the process.
So, it is important to identify and appreciate that there is a one way coupling between the
momentum equation and the energy equation. Solution of energy equation requires the
solution of momentum equation, but the reverse is not true. If the reverse is also true;
that means, for cases in which the temperature variation is such that viscosity and density
would also vary based on the temperature, then there is a two way coupling between the
momentum equation and the energy equation.
So, if their is a two way coupling between these 2 equations then both of them the
situation is more complicated and you need to solve both the energy equation and the
238
momentum equation simultaneously. So, that is very important to realize the importance
of the assumption of constancy in viscosity and density to make the momentum equation
uncoupled from the energy equation, which may not be true in all cases for which
combined solution at the both of these equations will have to be simultaneously solved.
So, we would see some of the examples when a mostly when they are uncoupled and you
would be able to solve them separately. As we would see in the next class when we talk
about the thermal boundary layer and so on. But, in today’s class we would see a very
simple problem and which would establish the relation between convection and
conduction. And, for that we would choose a system, which is very simple to visualize.
Let us say a solid ball; spherical ball is at a temperature T which T 0 , which is higher than
the temperature of the room in which the ball is suspended in still air.
So, in this room if there is a steel ball whose temperature is 100 ºC let us say and it is
suspended in this room, where there is no forced flow of air over the surface. So, if all
the fans etcetera in the room is switched off all the windows are closed, then it is safe to
assume that there is no flow of heat due to convection. Since, there is no external agency
which forces the fluid to move past the solid sphere, which is at a higher temperature
than that of the air surrounding it. So, in room full or room of still air a ball a spherical
ball is being cold. So, you can think of it as the limiting case in which the convective
flow is slowly brought to 0.
So, we would like to see what is going to be the convective heat transfer coefficient, the
limiting value of the convective heat transfer coefficient when there is no flow. So, it is a
hypothetical situation. All convections require the presence of a flow, but if I am slowly
reducing the flow velocity the flow past the cylinder. Then if I can make the velocity
very small then all the heat is going to be transferred from the solid to the air by a
conduction process.
And, if I equate that conduction by which heat is lost from the sphere to a hypothetical
heat transfer coefficient at 0 velocity, then the value of h convective heat transfer
coefficient or the value of the dimensionless number, which in convection is Nusselt
hl
number, which is defined as . H is the convective heat transfer coefficient, L is the
k
239
hl
length scale and k is the thermal conductivity of the fluid surrounding it, then this
k
refers to Nusselt number. And, you would see that in almost all of the relations for
convective heat transfer coefficient, the relation or correlation is expressed in terms of

Nusselt number.
So, that way Nusselt number plays a very important role in quantifying convective heat
transfer from any object, it could be a flat surface, it could be a spherical ball or it could
be a bank of tubes through which a hot fluid is flowing and cold air is pumped over
them, cold air is made to flow over them to cool the liquid which is passing through the
tubes. So, in all cases the heat loss or gain by the fluid is expressed in terms of a
dimensionless number, which is called Nusselt number.
So, the problem that we are going to solve in this class looks at the value of Nusselt
number when the value of the imposed velocity is reduced and it is slowly brought to a
value equal to 0. So, at that point convection stops and conduction takes over, but we
would like to pinpoint to find out the asymptotic value of the heat transfer coefficient or
the Nusselt number for the case of 0 velocity.
So, that is the problem, which we are going to look, which we are going to examine in
this class.
240
So, the problem that we have now is then find out the numerical value of Nusselt
hL
number, which is denoted by Nu  , where h is a convective heat transfer coefficient,
k
L is the length scale of the system that we are discussing and k is a thermal conductivity
of the surrounding fluid in stagnant air. Let us assume the fluid is air and important one
is stagnant. So, there is no flow in the air.
So, I draw the system as this spherical ball has radius equal to R. And the surface
temperature is maintained at Tr, whereas the temperature at a point far from it is T ∞ and
the spherical ball is going to lose heat by a combination of conduction and convection,
and then you would like to see how it is going to lose its heat?
So, for this let us assume a thin shell of air surrounding the hot sphere. And, I am going
to write the energy equation for the air which is situated in this thin shell of air
completely surrounding the sphere. So, if you look over here for the energy equation and
if you see there is no imposed velocity and it is a steady state case.
So, the first term would be 0. vr , v and v all are 0 since it is stagnant air. So, therefore,
the entire left hand side of this equation would be 0. Since, there is no flow or very little
flow to talk about and their is no question of viscous dissipation and therefore, this term
would also be 0. And, there is no heat generation in the thin shell, that I have drawn over
here. So, therefore, the Q would be 0. So, the terms which would remain are the terms
that represent conductive heat flow.
Now, let us look at what is the variation of temperature with r ,  and  . Obviously, the
shell that we have drawn over here, there is going to be no dependence of temperature
with  and no dependence of temperature with  . Temperature is only going to depend
on how far this shell is from the centre of the sphere? Or in other words temperature is
going to be a function only of r and nothing else.
So, this equation even though it looks long and complicated, once you understand the
significance of the terms then you would be able to cancel the whole of left hand side the
second term on the right hand side, the third term, the fourth term and fifth term. And
therefore, the governing equation for a spherical system, which is dissipating its energy
into a room full of stagnant air can simply be written as
241
1 d  2 dT 
 kr 0
r 2 dr  dr 
So, this becomes your governing equation.
So, you can again, it is so simple to start with the general equation and obtain your
governing equation without just by looking at the equation looking at the system and
cancelling the terms. So, if my thermal conductivity is constant, I am simply going to
write
dT
r2  C1
dr
and after integration it would be minus
C1
T   C2
r
So this requires two boundary conditions. So, what could be the boundary conditions;
obviously, you can see that the temperatures at 2 locations are provided to you. The
temperature on the surface of the sphere is known is equal to T r and temperature at a
point far from the sphere is the temperature of the air, which is T ∞. So, the first boundary
condition would be as
r  , T  T , C2  T
Therefore, the equation would become
C1
T  T  
r
And, the second boundary condition would tell you at r = R; that means, on the surface
of it T is equal to TR, which would give you
C1   T  TR  R
So, these are the 2 integration constants which you can evaluate by using appropriate
boundary conditions. So, therefore, using the expressions of C 1 and C2 you would be able
to write the temperature profile
242
in air that surrounds the hot sphere as
R
T  T   TR  T 
r
Or, you can rearrange it a little bit more to express it in a more elegant fashion as
T  T R

TR  T r
So, this is what the temperature distribution in the air is known. So, if you look at the
interface of the solid and the gas which is air so; that means, at this point, at the interface
between the solid and the gas the heat which comes in by conduction must be equal to
the heat which goes out by convection.
So, at the interface between the solid and the liquid, the solid and the gas is immobile
due to the no slip condition. So, all the heat transfers through this immobile gas layer
must take place by conduction and on the outside of this gas layer the air is mobile. So,
therefore, you can have convection from this immobile layer. So, if I think of this as my
immobile layer of the gas molecules, then heat is going to get transferred by conduction
over here and over here the molecules are mobile. So, heat is going to transfer by
convection for the layer of molecules which are situated just outside of the interface and
are free to move.
243
So, in order to maintain steady state this conductive heat must be equal to the convective
heat. So, whatever be the amount of heat that goes through the molecules by a
conduction process must be equal to the convective process by, which this heat is being
picked up by the mobile gas molecules. So, if we understand that I am simply going to
write that at the interface qconduction is equal to qconvection.
So, what is the expression for qconduction? By Fourier’s law it is going to be
dT
k r R  h(TR  T )
dr
 1  hR
or,  k (TR  T ) R   2   h(TR  T )  1
 R  k
hD
or,  2 = Nu
k
So, this is a very important relation which tells you what would be the value of Nusselt
number. The limiting value of Nusselt number, when there is no flow of air surrounding
the hot sphere. So, therefore, you get a numerical value of Nusselt number in the limiting
case when there is no convection. Or this is a very interesting result as the expression of
h for velocity equals 0 is something, which is proven experimentally this is an
asymptotic value which would give you the limiting value of Nusselt number for the case
of heat loss from a sphere.
And, further calculations, further analysis showed that in presence of velocity; obviously,
the value of h would increase, because more the velocity more would be the value of
convective heat transfer h will increase and as h increases the value of Nusselt number
will also increase.
So, the expression of Nusselt number which was obtained in a semi analytical fashion
and using certain correlations, that relation is Nusselt number based on diameter the
length scale is taken as the diameter of the sphere is
1/4
1/2  
2/3 0.4
NuD  2  (0.4 Re  0.06 Re ) Pr  
 s 
244
So, what this tells this is this is a correlation, which was obtained you; obviously, do not
need to remember this, but what it shows is when the velocity is equal to 0, velocity is
set to 0, the Reynolds number would be 0. And the entire second term on the right hand
side would be 0 and the value of Nusselt number would approach to and this is what you
have obtained over here. That the limiting value of Nusselt number at steady state for a
sphere which is losing heat in a stagnant air is equal to 2.
So, that there is a simple, but elegant example of the use of the energy equation to obtain
the governing equation, use of appropriate boundary conditions known temperatures in
this case. To obtain the temperature profile in air and define a Nusselt number for the
case when there is no flow of air, no flow of the surrounding air which is you can assume
that it is the limiting case you can see that is it can be expressed as the case where you
have convection with 0 velocity.
That is a hypothetical limiting value of Nusselt number to be equal to 2 and numerous

experiments and analysis has proved this to be correct. And the expression of Nusselt
number generally used for flow of air over a hot sphere we say, which is equal to 2 plus
some function of Reynolds number. So, when Reynolds number is 0 Nusselt number
asymptotically, reaches a value equal to 2. So, this is one example of use of energy
equation. We will see some more examples of that in the coming classes.
But, in the next class we will start with flow over a flat plate and how to obtain the
Nusselt number for that case in laminar flow.
245
Heat Transfer
Lecture - 21
Momentum and Thermal Boundary Layers
We have seen the use of energy equation. And, now we will go back to our original aim
to obtain an expression for h, the convective heat transfer coefficient. And, we are going
to choose the simplest possible flow geometry, which is flow over a flat plate. So, a flat
plate is kept stationary in a flowing air stream, whose temperature is different from the
temperature of the flat plate. Let us assume that the flow is laminar and at steady state
and you would like to see what the expression for Nusselt number would be, which
contains the convective heat transfer coefficient “h” for such case. But, before we do
that, I would recall once again that there would be the formation of a hydrodynamic
boundary layer, and the formation of a thermal boundary layer on the surface of the solid
and let us say that air is approaching it as with a constant velocity.
So, the velocity of the air, which is coming is constant and is not a function x or y or z. It
is a 1-dimensional flow of air, which is approaching a flat plate. When this air meets the
solid due to viscosity and due to no slip condition, the flow very close to the plate would
be minimal. And, the velocity of air which is flowing over the plate will increase from 0
on the plate due to no slip to the velocity of the air, which is approaching it beyond a
certain thickness of air layer, which is known as the boundary layer.
So, a hydrodynamic boundary layer is that thin layer of air in contact with the stationary
plate inside which, you would expect a variation in velocity with the y direction. So, the
velocity varies with the y direction and it asymptotically reaches the value of the
approach velocity beyond that layer. So, inside the layer since the flow varies with y,
looking at the equation of continuity you would realize, that the flow inside the boundary
layer is 2-dimensional. We will for the time being assume that the depth of the field is
very large compared to the thickness of the plate. Therefore, the velocity or the
temperature is not a function of z, it is a function only of x and y, but it is not a function
of z. There is going to be a variation in velocity with x and variation in velocity with y.
246
Similarly, when we think of the temperature boundary layer or the thermal boundary
layer, the temperature of the fluid which is approaching the solid plate at a different
temperature this is constant, this is invariant this does not depend on either x or y or z.
But, the moment it meets a heated plate, the temperature of the immovable air molecules
due to no slip condition, is going to be equal to the temperature of the solid plate. As, we
move up, the temperature gradually changes till it asymptotically becomes equal to the
temperature of the incoming fluid. So, the thickness of the layer in which the temperature
varies from the temperature of the base plate, to the temperature of the free stream is
known as the thermal boundary layer. And, as again the temperature inside the thermal
boundary layer is going to be a function of the axial position; that means, it is going to be
a function of x and it is going to be a function of y.
So, we would have to see how to solve the equation for laminar flow, inside the thermal
boundary layer, but as I have told you before since the momentum transfer and the heat
transfer equations are coupled, we probably first need to solve what is the expression for
velocity inside the hydrodynamic boundary layer and then use that knowledge for the
solution of the thermal boundary layer to obtain an expression for Nusselt number.
But, before we go up to that objective, we need to simplify the equation of continuity,

equation of motion, and equation of energy. So, it seems that there is a way to simplify
these equations to make it more compact such that we can attempt an analytics solution,
if not analytic a semi analytic solution, but the goal is to obtain the heat transfer
coefficient. Now, to obtain the heat transfer coefficient, we first need to see what those
approximations are which are collectively known as the boundary layer approximation.
247
So, I start with boundary layer approximation in this class and I have drawn this
substrate which is stationary. Let us say there is a flow of air towards this plate with a
velocity U, which is a constant and the temperature of this is also a constant. And, it is 2-
dimensional flow, which simply tells you that the non-zero component of velocity are v x
and v y but v z is equal to 0. The concept of boundary layer and the associated analysis,
they together play a very important role in evaluating any transport process taking place
near a surface when it interacts with a moving fluid.
vx v y
The equation of continuity would simply give you   0 . The x-component of
x y
vx v    2v  2v 
momentum equation is vx  v y x   2x  2x  . So, here I have assumed that
x y   x y 
the there is no imposed pressure gradient and no body force; and the flow is at a steady
T T k   2T  2T 
state. The x-component of the equation of energy is vx  vy   2  2 .
x y  C p  x y 
So, these 3 are our governing equations, which we need to solve, but before we solve that
let us see if you can reduce this equation. If, you can get rid of some of the terms based
on the physical understanding.
248
So, let us first heuristically try to see what is happening? This is my x direction, and this
is the y direction. So, as you can see most of the flow is in the x direction, there is very
little flow in the y direction. So, which tells you that this v x must be very large in
comparison to vy. So, can we neglect the neglect the second term with respect to the first
we cannot do that, because vy maybe small, but the variation of vx with y is going to be
very large in comparison to variation of vx with x. Therefore, this part is going to be very
large in comparison to this term. So, we need to keep in mind that the thickness of the
boundary layer is very small. And, the velocity varies from varies from 0 at this point to
the free stream velocity capital U, over here. So, since the dy over which the velocity
changes from 0 to U is very small this term, the gradient with respect to y is going to be
very large compared to the gradient with respect to x, which tells us that none of the
terms on the left-hand side of the x component of equation of motion can be neglected.
We need to keep both the terms on the left-hand side in our equation. The second
observation is that, both the terms on the left-hand side refers to convective transport of x
momentum inside the boundary layer. However, the terms on the right-hand side which
contain µ or k are collectively referred to as diffusive transport terms. So, what is
diffusive transport of momentum? This is the transport of momentum because of the
presence of viscosity, such that the velocities are transmitted through layers which
moving at different velocities.
Now, of these two since the variation in velocity vx with x is small as compared to
variation in vx with y; the overall contribution of the diffusive component of the x
momentum is very small. And therefore, it can be equated to 0. I would explain it one
more time. Both terms on the right-hand side refer to diffusive transport of momentum.
They are diffusive transport which can easily be identified by the appearance of the
viscosity in front of them, and the variation of vx with x is small. Now, the diffusive
transport depends on the velocity gradient. So, the diffusive transport of x momentum in
the x direction is very small, as compared to the diffusive transport of the x momentum
in the y direction. So, this would allow us to drop the first term on the right-hand side.
And therefore, the reduced form of the x component of momentum transfer inside the
boundary layer is two terms on the left-hand side and only one term on the right-hand
side.
249
Now, let us apply the same logic to my energy equation. As, before this v x is very large
in comparison to vy; however, the change in temperature with respect to y is going to be
very large in comparison to the change in temperature with respect to x. So, T changes
from T S, which is the surface temperature of the substrate to T infinity, where T infinity
is the temperature of the fluid, which is flowing over the flat plate. Since, the thermal
T
boundary layer is small in thickness compared to the overall length scale. So, would
y
T
be much more than ; which leads to the conclusion like in equation of motion, that
x
none of the terms on the left-hand side of equation energy can be made equal to 0.
Now, we come to the right-hand side. As, before the appearance of k the significance of
these 2 terms are the diffusive transport of energy in the y and in the x direction, or these
are the conductive transport of energy in the x and in the y direction. The conduction
depends on temperature gradient. Now, since the temperature changes less in the x
direction, as compared to the large change in the y direction, the contribution of this
would be small and it can be equated to 0.
Therefore, you can see the similarity between these 2 equations, the x component of
equation of motion inside the boundary layer; and, the equation of energy inside the
boundary layer can be simplified by dropping the x component of diffusive transport,
and the conductive transport of energy in the x direction in comparison to the y direction
and none of the terms on the left hand side can be neglected. So, this is purely from
heuristics, that is from our understanding of what happens to the flow inside the
boundary layer. Both in terms of momentum transfer as well as in terms of heat transfer.
These approximations or these simplifications together are known as boundary layer
approximations.
I have also written the equation of energy and here we did not consider the viscous
dissipation, nor did we consider any amount of heat which is generated inside the
system. But, if you again if you look at the energy equation, the presence of v x and vy
tells us that we need to solve these two equations first: equation of continuity and
equation of motion to obtain expressions of vx and vy which could then be substituted in
the energy equation and then this equation can be solved to obtain T as a function of x
250
and y. Only, when we obtain T as function of x and y, we should be able to obtain what
is the expression of convective heat transfer coefficient, or what is the expression of
Nusselt number?
251
Heat Transfer
Lecture - 22
The Flat Plate in Parallel Flow – Hydrodynamics and Momentum Transfer
We have already seen what kind of approximations one can make to the equation of
motion and equation of energy, when these two equations are applied for the case, where
we have flow over a flat plate. So, which is like flow parallel to a flat plate and we know
that the two boundary layers hydrodynamic and the thermal boundary layer will grow on
the plate, whose thickness will slowly increase with the axial distance x.
And, inside the boundary layers the flow is going to be two dimensional outside of which
the flow can be treated as one dimensional. So, we are going to start using the reduced
form of these equations for the case of flow over a flat plate, when the plate temperature
and the temperature of the fluid which is approaching the plate are different.
So, there would be heat transfer from the plate to the fluid, as well as formation of a
thermal and momentum boundary layers on top of the plate, within which most of the
transport processes are going to be concentrated. Outside of the boundary layer, there is
going to be no exchange of either momentum or energy.
So, it is extremely important to have the concept of boundary layer and the fact that all
the transport processes are confined within a thin layer close to a surface and is very
important from an engineering stand point. Because, whatever we do we need to
maximize heat transfer from a surface.
So, our point of interest or rather our zone of interest, will always be the region very
close to the surface, within the thermal boundary layer. And, we should make design
modifications to ensure that we are going to get the maximum heat transfer from the
surface by tweaking or by manipulating the flow conditions inside the thermal or the
momentum boundary layer.
So, we would first see the form of the equation, which we have derived in the previous
class. So, if you see what we have over here.
252
This is a flat plate, which is placed parallel to the flow where the flow is coming with a
velocity equals to U  and its temperature is at T .
So, these fluid when it comes in contact which the surface, which is let us say at a
temperature of Ts, there is going to be exchange of heat between the surface and the
fluid. And, this denotes the thermal boundary layer the edge of the thermal boundary
layer. So, temperature changes from Ts to T within this thin boundary layer. For clarity
this thickness is greatly exaggerated. So, therefore the real thickness of a boundary layer
would be very small of the order of millimeters and so, it will lie very close to the solid
surface.
But, it is shown like this over here. Initially the flow inside the boundary layer even
though it is two dimensional it will remain laminar, but as we move more and more into
the x direction, after certain point there is going to be turbulence present or the presence
of turbulence inside the boundary layer cannot be neglected.
And, this initiation of turbulence inside the boundary layer will change the thickness of
the boundary layer in a more drastic fashion and in a more irregular pattern. So, what we
have on the right hand side of this dotted line is the existence of turbulent flow, in the
boundary layer and the artificial demarcation, where it takes place. it is denoted as x c or
the transition length.
253
So, beyond xc all the flow inside the boundary layer is treated as turbulent whereas,
before this xc, the flow inside the boundary layer is treated as laminar. As, with the case
of laminar and turbulent flow, while flow is taking place through a pipe. There is no
distinct number. That is change from laminar to turbulent is not sudden, rather it is a
gradual process in which the fluctuations present inside the flow cannot be contained,
cannot be dampened enough by viscous forces. And therefore, they start to grow and
create a turbulent condition.
So, there is no magic number beyond which the flow is treated as laminar and beyond
which the flow is treated as turbulent, but for convenience as we have done for the case
of pipe flow, Reynolds number equals 2100 and beyond is taken to be turbulent flow.
Similarly, for the flow over a flat plate, when the flow is parallel to the flat plate the
transition Reynolds number, which beyond which the flow will be turbulent is taken to
be as 5×105.
So, the Reynolds number transition is taken to be equal to 5×10 5and beyond which the
flow is going to be turbulent, and this Reynolds number is defined as
xU  
 5×105

Where  is the viscosity of the fluid and I have a length scale over here this length scale
is taken as x.
So, this is the length scale so value of Reynolds number will be different at different
points, and when the Reynolds number transition 5×105 this x is going to be equal to xc.
So, this when it is equal to 5×105 the flow is taken to be turbulent. So, the difference with
the pipe flow Reynolds number is in the length scale. So, this x keeps on increasing. So,
beyond a certain point we are going to get turbulent flow in the system.
So, this is a universally accepted value of Reynolds number value before which it is
laminar and beyond which it is turbulent. So, if we look at these are the 3 equations
which we have derived using the boundary layer approximation. So, this is the equation
of continuity, this is the equation of motion, kinetic equation of motion in the x
254

component this is the kinematic viscosity = and when you look at the energy equation,

it is also using boundary layer approximations and this  is the thermal diffusivity
k
which is denoted as .
C p
So, the presence of vx and v y in the energy equation makes this equation coupled with the
momentum equation, the x component of momentum equation, whereas the absence of
temperature in these 2 equations make them decoupled from the energy equation.
So, there is a one way coupling between the momentum equation and the energy
equation, but the reverse coupling is not true. There would be a coupling as well if the
parameter here the physical property  and  start to vary significantly with the change
in temperature. If, that does not happen, if the  and  can be treated to be a constant
within the temperature zone of our operation, then there is going to be only one way
coupling between the two.
So, it is important even before we start thinking about the solution. So, thinking about
solving these two obtain T as a function of x and y. We, need to first solve these 2
equations, which was done in your fluid mechanics class. So I will very quickly briefly
go through this without solving it.
First of all these are 2 equation, which will have to be solved simultaneously. So, the one
way of reducing the number of equations that we need to solve, we can define a stream
function as

vx 
y

And, we can define the other one as v y   . So, when we do that and since  is an
x
exact differential, therefore, the order of the differentiation is unimportant. So, when you
put this over here and this one over here, the equation of continuity gets satisfied
automatically. So, we do not need to solve equation of continuity separately, if we define
our velocity in terms of a stream function  and utilizing the definition of vx and v y in
255
terms of  , we can see that this equation is automatically satisfied. So, we will just have
this equation to solve.
Now, this equation being a partial differential equation, we can try to solve it using a
method of combination of variables. So, what is done in method of combination of
variables is a new independent variable  is defined, which is for this specific case, is
U
defined as   y , where, U is a free stream velocity,  is the kinematic viscosity and
x
x and y are the independent variables.
v v y  2 vx
So, this  contains both y and  and when you evaluate this vx , v y , x , and
x y y 2
in terms of  then something interesting would at would automatically you can see.
And, the dimensionless stream function which is f ( ) , where

f ( ) 
 xU
So, these 2 are the definitions one is about a variable, which is a combination variable of
the 2 independent variables that appear in this equation and this is simply non
dimensionalizing the stream function by this.
So, therefore, utilizing these two our aim would be to express this equation not in terms
of x y and vx or v y , but everything should be expressed in terms of  , where  would

become the new independent variable and f would become the new dependent variable.
So, if we can do that if we can show that f is a function of  only, which is from this
equation then the partial differential equation can be converted to an ordinary differential
equation.
So, that is the whole beauty of this approach where you combine two independent
vx v y
variables to a new independent variable. And, all these vx , v y , , etcetera are
x y
expressed in terms f or its derivative. I will just show you one example and then will
present the solution.
256
So, what is vx , is by definition of a stream function expressed as
   U
vx   . , where   y ,
y  y x
df U df
So, vx   xU . . U
d  x d
The point to note here is that f is a function of  only. Similarly,
1 U  df 
vy   d  f 
2 x  
vx vx  2 vx
In a similar fashion one can find out what is the expression for , and , all
x y y 2
the other terms present in the x component of the momentum equation. When, you
substitute all of this and expressions for this in the equation what you get is this equation.
d3 f d2 f
2  f 0
d 3 d 2
So, this becomes the new governing equation for flow inside the hydrodynamic boundary
layer. What is the advantage that we see that this is now an ordinary differential which is
257
non-linear, but we can possibly try to solve it using the methods which are available to us
and the best method to solve such an equation would probably be the numerical method.
So, the all these vx v y etcetera are now expressed in terms of the dimensionless stream
function f and the independent variables x and y are clubbed together in f. So, now, let us
see in order to solve this equation, if you look at the solution we need to have 3 boundary
conditions, if we need to solve the equation.
So, let us try to find out what are those boundary conditions. So, I have a plate and the
flow is taking flow is taking place over the plate. So, what are the conditions that you can
see about vx and v y on the plate. So, on the plate means that is y equals 0 for any values
of x. For y  0 , vx , v y  0 due to no slip condition.
So, the no slip condition dictates that both vx and v y would be equal to 0. So, that is
those are the 2 conditions what one can write based on the location at y equals 0. Let us
see what is going to happen on the outer edge of the other edge of the boundary layer,
which is the edge of the boundary layer beyond which the flow is one dimensional. What
happens to the velocity of the fluid, the x component of the velocity of the fluid, the
velocity of the fluid x component is 0 on the plate as we move away from the plate
progressively the velocity in the x direction increases. And, from the edge of the
boundary layer and beyond the velocity or in the x direction would be equal to the free
stream velocity of the fluid which is coming towards the plate.
So, the other condition could be that
y  , v x  0
as y tends to infinity the mathematical speaking, the velocity in the x direction

approaches that of the free stream velocity. So, those are the 3 conditions. What happens
at y equals 0 and what happens as y tends to infinity? At, y equals 0 no slip condition
dictates that the velocity both components of velocity would be 0 and at as y tends to 0,
at a point far from the solid plate the velocity, if for the fluid stream would be equal to
the free stream velocity. That means, vx is simply going to be equal to U. These 3 are the
boundary conditions, physical boundary conditions, which we need to express in terms
258
of f and  and try to solve the non-linear ordinary differential equation that we have
obtained over here.
So, I am going to write those boundary those equations over here first. So,
For y  0 , vx , v y  0
  0, f   0  f
df
So, vx = U
d
So, it should give f  to be equal to 0, which comes from vx to be equal to 0. And, then
comes v y to be equal to 0, we understand that from here f  is 0, so, in order for v y to be
0, f must be 0 as well. So, the other condition is f  and both f  and f will be 0, since
velocity would be 0 at y equals 0.
So, y equals 0 corresponds to  equals 0, vx is 0 corresponds to f  is equal to 0, v y to

be equal to 0 then it must be f has to be equal to 0.
So, these are 2 boundary conditions from which one can write. And, the second condition
is as y tends to infinity vx would be equal to vx would be equal to the free stream

velocity. So, what is y equal tends to infinity. So, as y tends to infinity
  , f   1
So, vx tends to U. So, therefore, f  must be equal to one as  tends to infinity. So, these
3 are the boundary conditions that one needs to have in order to solve for this. So, I think
the physical nature of the equation and the boundary conditions are very clear to you.
Now, the next thing is even with these simplifications, even with the identification of the
boundary conditions an analytical solution for this equation is not possible. One has to
result to numerical solution of this equation and see what those results tell us to explain
the flow and momentum transfer inside the hydrodynamic boundary layer. I am spending
time on this, because as we have seen without a knowledge of the hydrodynamics inside
the boundary layer, one would not be able to solve the thermal boundary layer.
259
So, let us quickly see how would a numerical solution? What I am going to give you just
the table containing the values of  and the values of f, f  and so on. And, some of the
values I will list. Those values would be sufficient to tell us give us some idea of what is
going to be the important parameters? For example, the thickness of the boundary layer
or the shear stress coefficient for flow over a flat plate. If, we can get these 2
informations they would be sufficient to correlate the results with those of the thermal
boundary layer.
So, let us look at how would the result look like in the case of hydrodynamic boundary
layer.
(Refer Slide Time: 21:54).
So, the table that we have over here, this is from the numerical results. So, the first one is
U
y . The second one quantity is f, the third one is f  and the fourth one is f  .
x
So, at  equals 0, we know that vx is equal to 0. So, f  is 0 and f is also 0 from v y . So,
both f  , f would be 0 and the numerical solution of the double derivative of f which is
d2 f
is 0.332.
d 2
260
So, there are several numbers which are listed over here, I will I not write everything
what you would a interesting thing is when this  is 5 the value of f is about 3.28329,
the value of f  is 0.99155 and is f  is 0.01591. There are other numbers as well which
I do not need right now.
So, what is the significance of this table? This table which is obtained from the
numerical solution of the governing equation using the boundary conditions that I have
already discussed. So, from the table it is clear that for  equals 5, f  is 0.99, f  is
vx v
equals 0.992. And, what is f  , if you again look at f  this f  is . So, x is 0.992,
U U
v
what they tell us this is interesting; that means, for a value of  equals 5 the value of x
U
or vx reaches 99 percent of the free stream velocity.
So, which is unique because this tells us that at a value of  equals 5 the velocity inside
the boundary layer is almost equal to 99 and the working definition of boundary layer
thickness is that point, that point in y where the velocity reaches 99 percent of the free
stream velocity. So, this point then refers to the edge of the boundary layer. Now, if this
U
is edge of the boundary layer and if since I know   y then when I put the value
x
of  to be equal to 5 from here this y must be equal to the  , which is the local
thickness of the boundary layer.
So, this  is a function of x , as we move along the plate the value of  will increase,
but for a value of  equals 5 then y must be equal to  . So, when you reorganize this
equation what you are going to get is a very important relation,
5.0 x xv 
 , Re x 
Re x 
Where, Rex is the local value of the Reynolds number.
So, this gives you a compact expression for the thickness of the boundary layer at
different axial locations. Now, this equation since we have assumed it is only laminar
261
flow this is valid for a Reynolds number transition Reynolds number of 5 ×10 5. So, as
long as your Reynolds number is less than 5 ×10 5 , the thickness of the boundary layer at
any location , thickness of the hydrodynamic boundary layer at any location, can be
expressed by this formula.
On a similar note one can write the wall shear stress that is the shear stress felt by the
wall by the plate over which this flow is taking place. So,
vx U d2 f
w   y 0  U  0
y  x d 2
So, this is the stress at y equals 0. So, that is why we have this subscript w signifying
wall with  and this after a bit of simplification by converting this y to  and so on.
So, the wall shear stress this is Newton’s law of viscosity and then after converting the
vx to f and y to  this is the expression that you should get.
So, the wall shear stress that is knowledge of wall shear stress therefore, requires the
value of f  at  equals 0 and from the table over here you know that what is the value
of f  at  equals 0. So, therefore, one can write
U
 w  0.332U
x
0.332 U 2
w 
Re x
So, this is another important result as it gives you the value of the wall shear stress for
flow over for flow over a flat plate and a corollary of this is the wall shear stress
coefficient, which is expressed as Cf, which is defined as
w 0.664
Cf  
1 Re x
U 2
2
Let us spend some more time on this wall shear stress coefficient. So, first of all why we
would like to have the concept of wall shear? The wall shear essentially is the
262
engineering parameter of interest, because whenever you are trying to design something
whenever you are trying to evaluate, the force exerted by a moving fluid on a stationary
platform. You need to find out what is the wall shear stress.
The wall shear stress then you will have to be multiplied by the area in order to calculate
the total force.
So, therefore, the wall shear stress plays a very important role, in the design of the
surface evaluation of the force on a surface and so on. And, they contain they are
function of the local value of the Reynolds number, the properties of the liquid ρ and μ
and the imposed condition which is the velocity of the fluid, which is flowing over the
plate.
And, another engineering parameter of interest is wall shear stress coefficient, which is
1
expressed as wall shear divided by U 2 , which is also known as the dynamic pressure.
2
So, the wall shear stress coefficient is a function is the ratio of the wall shear by the
0.664
dynamic pressure and which is x which can be obtained directly from this as .
Re x
I think this much of fluid mechanics should be sufficient for us to progress to find out
what is a find out how to handle the heat transfer thermal boundary layer the heat
transfer part of it? Since thermal and the fluid mechanics is linked to heat transfer by the
appearance of velocities in the expressions one first need to know, what is the velocity?
What are the components of velocity? What is the thickness of the boundary layer? What
is the shear stress coefficient and then try to get into the thermal boundary layer in order
to obtain the solution.
So, what would be the equivalent one for the case of heat transfer, the C f in heat transfer
is related is similar in concept to the heat transfer coefficient or the Nusselt number. So,
our aim is to obtain Nusselt number. So, what we would do is based on the knowledge of
the flow inside the hydrodynamic boundary layer, we would convert this knowledge to
our thermal boundary layer and try to derive the quantities which are of relevance in heat
transfer. Namely the temperature and the heat transfer coefficient. So, we are spending
so, much time on the fluid mechanics just to make sure that we understand the next step,
263
when we have to use these results in order to obtain the Nusselt number which we would
do in the next class.
264
Heat Transfer
Lecture - 23
The Flat Plate in Parallel Flow – Heat Transfer
In the last class we have seen how the growth of the hydrodynamic boundary layer on a
flat plate can be analyzed analytically. Though towards the end we had to resort to
numerical solution because we could convert the PDE to an ODE by combining the
independent variables by defining a stream function. However, the resulting ordinary
differential equation that we obtain was non-linear and an analytical solution was not
possible.
However, the numerical solutions give us some very interesting results, looking at the
value of the dimensional stream function and its relationship with vx and vy and noting
they would be 0 at y equal 0; due to the no slip condition. We could also obtain the
expression for the growth of hydrodynamic boundary layer as a function of operational
parameter, which is the velocity with which the fluid approaches the plate. And the
length scale which is the axial length in the direction of flow and thermo-physical
properties which are µ and ρ. Therefore, the expression for δ, the thickness of the
boundary layer was expressed in terms of Reynolds number as delta to be equal to
5x
.
Re x
Refer Slide Time: 02:00)
265
For the case of the wall shear stress, we obtained the expression; and another quantity of
engineering importance is the wall shear stress coefficient or Cf, which turned out to be
0.664
. And these two rows in the numerical solution where important which gives us
Re x
some idea of η = 0 which corresponds to no slip condition. So, from there we obtained
the expression for the thickness of the boundary layer. We would now like to use this
knowledge to analyze flat plate in parallel flow, but this time it is going to be the heat
transfer case.
266
So, in the heat transfer when there is going to be a difference in temperature between the
plate and the fluid which is approaching. So, let us say T S is the surface temperature, and
the fluid is approaching at a temperature of T infinity with its velocity equals to U and
then there is going to be growth of the thermal boundary layer. It may or may not
coincide and may lie above or below the hydrodynamic boundary layer.
But this is the thermal boundary layer, let us call it as TBL and the energy equation
T T  2T
corresponding energy equation would be vx  vy   2 . This is the equation we
x y x
would have to solve, but we already have an idea of the value of vx and vy at different
values of x. So, our hydrodynamic solution has already given us the value of f and f
prime which are required to calculate the vx and vy.
T  Ts
So, we would now define a dimensionless temperature as: T *  ; where, TS is the
T  Ts
temperature of the solid surface, T infinity is the temperature of the fluid outside of the
thermal boundary layer. And we also assume the T star, is a function of both x and y.
The value of T star would depend where we are in terms of the axial location as well as
in terms of the vertical location. Therefore, this T star is a function of eta. So, this is what
we have assumed and when we make the substitutions, in this equation the governing
d 2T * Pr dT *
equation turns out to be  f  0 . The point to note here is first of all the
d 2 2 d
appearance of f prime in the energy equation coupled with the hydrodynamic part of the
solution.
So, the presence of f denotes that the energy equation and the momentum equation are
coupled. You would not be able to solve the energy equation unless you solve the
momentum equation. So, you need to know the value of f apriori before you start solving
this equation; the second one has to note is the appearance of Prandtl number. So, this
Cp
Prandtl number is defined as you know as where, Cp is heat capacity, µ is the
k
viscosity and k is the thermal conductivity. The value of η and f are already known to
you from the solution of the hydrodynamic part of the momentum transfer equation. So,
267
if this is known then I should be able to numerically solve for T* as a function of η,
provided the value of Prandtl number is known to me.
So, to numerically solve this equation, the only thing I need to do is specify the value of
Prandtl number, the moment I specify the Prandtl number to be let us say 0.7. Then I
should be able to numerically solve it because, if dependence of f on η is already known
to me and therefore, I should be able to obtain T star as a function of eta for a specific
value of Prandtl number. So, that is the only thing we have need to realize in here that,
T* as a function of η can be obtained at a specific value of Prandtl number.
It is to be clarified here and it has been found that for this range of Prandtl number 0.6 to
dT *
50, the results for the surface temperature gradient which is d , can be expressed as

at   0
to be equal to 0.332 Pr 13 . This is a very important result, as we have seen we need to

know the numerical value of Prandtl number to solve it. So, we assign different values of
Prandtl number and start solving this equation 0.5, 0.75, 1, 1.25, 10, 20 and so on. What
dT *
is interesting, the observation that was obtained is that d , which you can obtain from

at   0
the solution of this, provided you assume a value of Prandtl number, can be expressed as
a function of Prandtl number over a wide range of Prandtl number.
dT *
Now, why is it important, why I am at all interested in d . The same reason I was

at   0
interested in trying to find out what is the velocity gradient at y = 0. That means, this
value at the solid surface gives me what is the force exerted by the moving fluid on the
solid, in other words I am defining an important transport coefficient related to the force
exerted by a moving fluid on a stationary solid.
268
dT *
Similarly, d should give me some idea of the heat transfer process that is taking

at   0
place in the situation, when a cold fluid flows over a hot plate. So, the temperature
gradient at the interface should give us an idea of what is the convective heat transfer
coefficient in this process and we understand the entire study of convection is trying to
find out what is the convective heat transfer coefficient. What is the relation or
correlation that can be used to obtain the value of the heat transfer coefficient, and in
dT *
order to do that, the knowledge of d will play a very significant role.

at   0
dT *
So, next what we are going to see is how this d can be related to the convective heat

at   0
transfer coefficient, but to do that I need to have an idea of what is the Prandtl number.
So, what we have seen is that for a large range of Prandtl number, the temperature
gradient at the interface can be fitted to the variation in Prandtl number. So, I will start
writing it and then I will show you how it would give us an expression for the heat
dT *
transfer coefficient. So, let us start with that, so this d is equal to 0.332 Pr 13 this is

at   0
our starting point, the heat transfer coefficient hx.
269
That means, a local value of the heat transfer coefficient can be defined from Newton’s
qsll
law of cooling as: hx  . T is the temperature of the solid surface; T infinity is
 Ts  T  S
the temperature of the liquid. So, if this is the flux the Newton’s law of cooling connects
the convective heat transfer coefficient with qsll in this way. So, I can write it as
 
1  T 
hx   k .
Ts  T  y 
 y 0
 
So, this is Newton’s law of cooling and this is substituting Fourier’s law of conduction
for the surface heat flux because, as we have discussed many times before on the surface,
the fluid molecules are immovable. And therefore, the transport of heat through this
immovable layer of molecules of the fluid does take place only by conduction and that is
why I am substituting the Fourier’s law of conduction in here. So, this now can be
270
T  Ts
modified little bit, it is T, I am trying to use the definition of T*. Where the
Ts  T
T  Ts T *
k
definition of T *  these times y .
T  Ts 
y0
k T *
hx 
So, these two cancels, the negative sign will disappear so, Ly * because, I am
y*  0
y
converting this into dimensionless one as well. So, y*  , where L is the entire length
L
of the plate over which this heat transfer is taking place. Look at this equation carefully,
what you have in here is, if I bring this L on the other side and k over here an offshoot of
this equation is hx the local value of heat transfer coefficient.; x denotes the location L by
hx L
k is equal to del T star by del y star, at y star equal to 0. This is dimensionless and it
k
is denoted by Nusselt number, Nux (the local value of Nusselt number).
This is a unique definition of Nusselt number, which simply tells you that the Nusselt
number is nothing, but the dimensionless temperature gradient at the interface. So, the
hL
proper definition of Nusselt number is not simply , the correct way to express
k
Nusselt number is: It is physically the dimensionless temperature gradient at the solid
fluid interface.
So, that is one way to look at the genesis or the physical significance of Nusselt number.
dT *
So, now, let us proceed with this and see that with our knowledge of d , can

at   0
somehow be plugged into this expression to obtain a more compact and useful
expression for Nusselt number in laminar flow for flow over a flat plate.
271
So, I continue with this once again and after simple re-arrangement, we see that the final
1 1
expression the local value of Nusselt number as Nu x  0.332 Re 2 Pr 3 . The limitations
is that, the Prandtl number has to lie between 0.6 and 50, which was used to obtain this
expression. So, what I see here then, that a compact expression for Nusselt number can
be obtained by simply invoking the result over this and realizing that T* is a function of
η. And this expression is going to be extremely useful for analyzing situations in which
we have flow over a flat plate where the temperature of the plate and the temperature of
the fluid are different.
So, we are going to have convective heat transfer taking place and this is the first time
that you see an expression for h is obtained in terms of Reynolds number which
characterizes the flow, which brings in the hydrodynamic part of it into the expression.
Cp
And a function of Prandtl number which is and it gives you some idea of the
k
relative importance of convection, conduction, momentum diffusivity, thermal
diffusivity into our discussion.
But as an engineer, you are probably more interested in finding out the average value of
the heat transfer coefficient or the average value of the friction coefficient over the entire
plate surface. You do not want to know most likely what is the value of the heat transfer
coefficient at a specific location, you would rather try to find out what is the average
value of heat transfer coefficient. Because, that would give you some idea of the total
amount of heat transferred.
272
So, this is about average coefficients. So, Cf and I put a bar over it to say that this is the
s
Cf 
average value, must be 1 . But here I am using the average one so,
U 2
2
 s  1 0  sx dx .
x
therefore, the average value of wall-shear stress can be expressed as:
x
 s  1 0 0.332
x U 2
So, this can be further modified as dx . Upon integration and
x Re x
1
simplification, we get: C f  1.328 .
Re x
Similar to the momentum transfer the value of the average heat transfer coefficient
1 x
would simply be h h dx , which is the definition of lengthwise average.
x 0 x
1
k U 2 x 1
Multiplying both sides by k/x would give rise to: h  0.332 Pr 0.5    0 x dx .
x   
0.5
273
So, once you do this your h, the average heat transfer coefficient is going to be equal to
twice of the local heat transfer coefficient. Therefore, your average value of Nusselt
number, the expression for Nusselt number would be average h times x by k and this
1 1
would turn out to be Nu  0.664 Re x 2 Pr 3 . So, this gives you the average value which
is more important in engineering situations than the local value of the heat transfer
coefficient. So, what we say then is C f average is equals twice C f local and h average is
twice local.
So, we have two equations, one equation is for the local value of the Nusselt number or
local value of the heat transfer coefficient and simply finding out the average. And the
averaging is done by integrating it over a certain length and dividing it by the length,
which is the standard definition for average based on the length. What you would see is
the local value and the average value differs only by a factor of 2. This expression is also
valid for a Prandtl number range: between 0.6 to 50.
274
Heat Transfer
Lecture - 24
The Effects of Turbulence
Previously we have seen, how the velocity and thermal boundary layer, grows over a
surface over a flat plate when there is a relative motion between the solid flat plate and
the fluid in contact with it. So, the system that we have analyzed, which is the simplest
possible situation is where we have a flat plate and the flow is taking place over it, where
the temperature of the plate and that of the fluid are different. As a result of which we are
going to have momentum transfer as well as heat transfer from the plate to the fluid.
Now, the momentum transfer is going to be principally governed by the viscosity of the
fluid and the density of the fluid. Whereas, heat transfer being a combination of
convection and conduction is going to be a function of k, the thermal conductivity of the
liquid; ρ, the density; Cp, μ, as well as the aim is the entire exercise was to connect, the
heat transfer coefficient or in order to obtain an expression of heat transfer coefficient for
simultaneous flow and heat transfer on a flat plate.
Now, if you recall we have said categorically that the analysis that we have done is
restricted to laminar flow only. Now, when it is laminar flow then the entire transport,
whether it is momentum or it is heat is going to be governed by the mostly by the
molecular transport of momentum or molecular transport of heat.
So, we have used the approximations, which for flow inside the boundary layer both the
hydrodynamic boundary layer, as well as the thermal boundary layer. And, we have seen
that the left hand side of the energy equation or left hand side of the momentum
equation, they contain only the advection terms. And, since based on an order of
magnitude analysis, we have demonstrated that none of the terms for a 2 D flow inside
the boundary layer can be cancelled.
So, I am going to have an
275
T T
u v is equal to the molecular transport of momentum. And, then molecular
x y
transport of momentum we have seen that due to the prevalence of a large temperature
gradient across a thin boundary layer, most of the heat is going to get transferred by
diffusion by thermal conduction in a direction perpendicular to the flow; that means, in
the y direction. Whereas, since the temperature is not changing appreciably as the flow
takes place in the x direction, the molecular transport of energy in the x direction, that is
the conductive heat transfer can be neglected.
So, with that and with our knowledge of with our previous knowledge of the solution of
the hydrodynamic boundary layer, we did obtain a closed form solution for the heat
transfer coefficient or more specifically a relation for Nusselt number, which is h d
where d is the length scale. In this case it is the length of the plate it is a variable length
of the plate, the location of the plates. So, the local Nusselt number be (h×x)/k and we
have seen that it is a function of Reynolds number, and it is a function of Prandtl number.
So, the compact expression which we have obtained for Nusselt number which is valid
for a specific range of Prandtl number and which is also valid only for laminar flow
characterized by a local value of Reynolds number less than 5 ×10 5. These two
constraints on Reynolds number and Prandtl number gave us a relation for heat transfer,
the convective heat transfer from a flat plate when it is placed in parallel flow over a
fluid stream, which has a temperature different from that of the plate.
So, upto this point I think it is it is perfectly fine, we know how what is the genesis? How
the, how the heat is getting transferred and so on? The situation gets more complicated as
we move into turbulent flow. We understand that in laminar flow the streamline motion
the plate like motion of adjacent fluid layers are mostly maintained. So, layers of fluid
slip past one another and the invisible string, which tries to slow down the faster moving
fluid or tries to accelerate the slower moving fluid, the viscosity, which is the molecular
transport property, that controls the molecular transport of heat and momentum control
everything.
But, in turbulent flow something else happens. The temperature at steady state at a given
location at a given x and y in laminar flow is a constant. So, if you could measure the
temperature at an axial location at some value of y; that means, at a specific point in the
276
flow stream inside the boundary layer, that temperature or that velocity will remain
invariant with time.
So, you are going to read only one value of it. However, in turbulent flow the situation is
more chaotic. So, most of the that is going to flow cross flow everywhere there is going
to be well mixed behavior of the flow inside the boundary layer and everywhere else as
well.
So, the momentum or heat gets of the transfer of momentum and heat they will get
augmented by the presence of this inter mixing this chaotic motion, which is a hallmark
of any turbulent flow. So, the entire fluid is going to be very well mixed and therefore,
the temperature that you measure at a given point is going to be a function of time. In
other words you would see local fluctuation in the value of the temperature that you
measure, and the value of the velocity that you measure.
So, the velocity is not going to be like in laminar flow if you measure the either the
velocity or the temperature.
So, the velocity with this thing which time this it does not really change, it is a straight
line and let us say the temperature is also going to be a straight line without any
perceptible changes with time.
277
However so, this is the case for laminar flow at steady state, of course, this is going to be
at steady state. But, if you do not measure and if your measurements system is very
accurate very sensitive and if you measure the velocity or the temperature what you
would see is local variations like this? It is not just going to be a constant value. And,
similarly for temperature you would see the value of the temperature is going to be
something like this it is not just one value.
So, this is the time. However, if you take an average of that average of these values both
for temperature and temperature as well as for velocity, what you would get this can be
called as v or time average velocity.
v  v  v
And, this is time average temperature can be denoted as T*.
So, when we say that the velocity of the fluid in any in a specific situation is this much 5
meter per second. What essentially we see specially, for turbulent flow is this is the time
average velocity or the time average temperature.
So, these are average temperatures and let us say I would like to know, what is the
temperature? What is the velocity at this point and what is the temperature at this point.
So, the local value of temperature would simply be T  T  T  this v velocity is going to
be v  v  v .
So, this v’ and T’ are the fluctuating components of velocity or temperature. Of course, in
laminar flow you do not have a fluctuating component and v, the velocity is essentially
same as the time average velocity. So, they are constant so, that time average value
would; obviously, be equal to the same velocity. Similarly, for temperature variation and
in turbulent flow can be expressed as the sum of a time average temperature plus a
fluctuating temperature.
So, these are the fluctuating components and there are certain properties, which one has
to keep in mind is that this T , the fluctuating component or v , the fluctuating
component of velocity, if you take the time average of that this is going to be equal to at
0.
278
T , v  0
T   0, vT   0
So, you have up and down and it is completely an arbitrary function and it is so, if you
take a large enough time and take the time average of the fluctuating component
individual component should be 0, but the same cannot be said about their product. So,
this may or may not be equal to 0 ok. So, T   0, vT   0
So, the fact that the time average of the fluctuating component is 0, but the time average
of the product of the fluctuating components is not 0, has given rise to certain interesting
changes in the equation of motion and the equation of energy. The temperature is a scalar
so, I only have to deal with T’, but velocity is a vector. So, I may have 3 components of
the flux the fluctuating components of velocity as vx , vy , vz .
So, therefore, the time average of vx is 0, but vx vy  0 . So, whatever I said about
velocity and time is also true for velocity component and velocity component product of
2 velocity components. So, the time average of vx is 0, but vx vy  0 . Now, this non zero
part vx vy  0 has given rise to additional terms in the equation of energy and in the
equation of motion.
So, what physically, what I mean here is that the presence of the fluctuating components
essentially mean, additional transport of momentum or additional transport of heat. Over
and above that specified by the laminar heat transfer, laminar heat transfer or laminar
momentum transfer. We all understand that when you mix something mix a hot fluid and
if you increase the velocity, if you increase chaos in the system, disorder in the system,
then there is going to be more heat transfer, which is taking place.
So, this additional heat transfer or additional momentum transfer they are direct results
of the non-zero nature of the fluctuating components of velocity and temperature or
velocity component in two directions. I will not go into very much details of it, but you
would probably know that the equation of motion can be can be expressed in terms of the
not for just the velocity but the instantaneous velocity. By instantaneous velocity I mean
this v. So, the instantaneous velocity is a sum of the time average velocity and the
279
fluctuating component instantaneous temperature similarly is the time average
temperature and the fluctuating temperature.
So, when you use this expression of velocity in equation of motion or this expression of
temperature in equation of energy two identical equations, which are fundamentally,
conceptually same as that of the equation of motion.
but being different because of the presence of this they appear and I simply write them
as, let us say I am writing it for 2 dimensional flow which is the flow inside a boundary
layer, everything now is in terms of the times smooth value. So, that is why the bars
should appear over here.
So, this is the x component of equation of motion is
 v v  dp   vx 
  vx x  v y x       vx vy 
 x y  dx y  y 
So, this is the equation of motion
Similarly, the equation of energy would be
 T T    T 
 C p  vx  vy   k   C p vyT  
 x y  y  y 
280
So, the extra terms these 2 extra terms on the right hand side, they account for the effect
of turbulent fluctuations in momentum and turbulent fluctuations in temperature.
So, if you look at only this part μ times velocity gradient this is nothing, but the shear
stress this is nothing, but the shear stress in laminar flow and this is nothing, but the
conductive heat transfer in laminar flow. So, this  tot when I express this entire thing and
call it some sort of a shear stress. So,
vx
 tot      vx vy
y
So, this is the total shear stress exerted by a moving fluid which as 2 components one is
due to laminar, so, this is due to laminar part of it which with which we are concerned so,
far and this is the turbulent part of it ok.
Similarly, the total amount of heat transfer is going to be
T
qtot
  k   C p vyT 
y
So, this is again this is a laminar one this is the laminar part this is the turbulent part and
this is the turbulent heat transfer. And, this term has a special name which accounts for
the shear stress experienced by a turbulent fluid by a fluid moving in turbulent flow, it is
called as the Reynolds stress.
So, Reynolds stress accounts for the additional amount of momentum transferred, that
one would expect in turbulent flow and the similar term for the case of heat transfer is
given by the this term. So, these 2 are the laminar counter parts of momentum and heat
transfer these are the turbulent contributions to the momentum transfer and to the heat
transfer.
Now, when it is expressed in this way sometimes it is more convenient to express it in

terms of something which we the term that we use very commonly, we call it as eddies.
So, what are eddies you can visualize you can conceptually say that eddies are nothing,
but packets of fluid, which has a very short length before which it disintegrates and
mixes with the fluid stream once again. So, eddy is hypothetically a packet of fluid
which flows cross stream to carry with it more momentum and more energy than that
281
prescribed in laminar flow. So, the more the eddies more would be the heat transfer, and
more would be the momentum transfer.
So, eddy is something which transports energy and momentum in turbulent flow. So, in
many times the turbulent flow is characterized by the presence of eddies, the turbulent
flow gets enhanced momentum transfer, heat transfer as well as mass transfer, because of
the formation of eddies. Now, there has been considerable attempts to define to
characterize or to put a correlation of the eddy transport of heat , eddy transport of
momentum and that of mass. So, we would quickly see what how eddy can be combined
with our traditional understanding of laminar transport and get some idea of the
enhanced transport encountered in turbulent flow.
So, let us look at what this eddy concept is and we are going to look at eddy. So, these
are this eddy the model it is a conceptual model so, for turbulent flow in momentum and
in heat. So, the turbulent flow in momentum and heat they are expressed in terms of this
the presence of these eddies.
And, generally they are expressed as I will write the defining equation of the eddy
vx
 M    vx vy
y
282
So, this if you look at this equation and then you would see why we are trying to express
the eddy in this specific form. I have a k over here or I have a ρ μ over here, which is
expressed in terms of velocity gradient. So, I have the additional transport of momentum
over here. So, if I could express this additional term in terms of something that we know,
vx
that is  . So, what μ does in laminar transport in transport of momentum in laminar
y
flow, this  M if you look at the similarities between these 2 equations  M plays
apparently the same part as that of μ/ρ, in the case of laminar transport of momentum.
vx
So, this entire term is expressed in terms of  M . So, therefore, the total shear stress
y
vx
the  tot which we have written before as     vx vy in the previous page now,
y
therefore, can be expressed as
vx
 tot    (   M )
y
And, similarly one can write so, this is the new expression for  in turbulent flow and
you can clearly see the additional transport of momentum due to the presence of this  M .
So, if this is called the kinematic viscosity, the other one is called the eddy viscosity. So,
the same way we have done it for momentum transfer we can also do it in the case of
heat transfer as
T
H  vyT 
y
I am replacing k , I am not going to use k instead I use  H epsilon H, which is the eddy
transport coefficient for turbulent flow heat transfer.
So, therefore,
T
qtot
    C p (   H )
y
283
So, these 2 equations when you see you can clearly see so, this is the thermal diffusivity
and this is known as the eddy thermal diffusivity.
So, these 2 the value of the thermal diffusivity tells you what would be the laminar flow
contribution of heat transfer, and this is the eddy transport of heat transfer for the case of
turbulent flow.
Now, the situation is apparently simple, if we know what are the values of the for
momentum and  H for heat transfer then my job is complete. Then, I should be able to
find out what is the total amount of momentum transfer i in turbulent flow, or the total
heat transfer in turbulent flow. But, it is easier said than done even the people have tried
to express the epsilons, the eddy diffusivities be it mass momentum or heat. In terms of
known fluctuations in terms of fluctuations present, which are also a function of the
degree of turbulence present in the system it is extremely difficult.
So, far no generalized expression for these eddy diffusivities are available, which is
going to be valid over a wide range of turbulence present in the system. So, what we
have we do understand the role eddies play in the transport process. It gives us a very
good pictorial description of the intermixing between different layers and additional
transport in turbulent flow as compared to laminar flow, but qualitatively it gives a good
clear picture, but it is so difficult to express them quantitatively.
So, we have to adopt other approaches, which we will discuss in the next class other
approaches of heat transfer, other approaches have expressing turbulent flow heat
transfer and turbulent flow momentum transfer, which rely heavily on empiricism.
So, even though it relies on empiricism there has been sufficient theoretical background
people have tried to put sufficient theoretical support for the relations in turbulent flow,
but at some point of time you need to have experimental results to quantify the transport
the additional transport of heat and momentum as well as mass, in the case of turbulent
flows. So, what we are we are going to do then is we are going to look at the how would
the profile look like when you would see what is the, how does the temperature profile?
and the velocity profile vary in here.
284
So, let us say you have this is a y direction the flow is taking place like this and this is
your axial direction x, and you are trying to see what is the velocity profile?
So, the velocity profile is going to be something like this. So, the velocity profile;
obviously, is 0 the velocity is 0 over here and it keeps on increasing as you move in this
direction, and ultimately the velocity here is going to be equal to U. So, this is the
velocity which is velocity in the x direction which is being plotted.
Similarly, if you do plot the same thing for the case of turbulent flow, you would expect
a profile of velocity like this it took almost up to be with this point. So, this is the
hydrodynamic boundary layer thickness and over this thickness the velocity varies from
0 to roughly about U or to be specific to about 0.99 U.
But, in this so, this is the case for laminar delta for the case of laminar, but in this case
this is also velocity in the x direction and I am plotting it with y. In this case also in this
case over a very small distance of δ, this is called delta turbulent, the velocity changes
from the 0 on the surface of the solid due to no slip condition to the velocity of the free
stream. So, one can clearly see that  TURB   LAM .
So, what can you say about the velocity if you velocity gradient. So, if you find out
vx vx
y0 and y0 . So, this is for the case of turbulent and this is get for the case of
y y
285
vx
laminar. What you would see is that, this y0 for the case of turbulent flow will
y
vx
definitely be more than y0 . Since the y over which the velocity changes from 0 to
y
vx
U ,this, what is much smaller as compared to this? So, your y0 for the case of
y
vx
turbulent flow is going to be more than y0 for the case of laminar flow.
y
So, this conceptually explains, the additional transport or the enhance transport of
momentum and heat in turbulent flow. So, I have given you an idea I have given you a
heuristically, I have shown you that why would you expect more heat transfer more
momentum transfer in turbulent flow as compared to laminar flow. Pictorially you can
think of the formation of eddies; which aids in the additional transport of heat and
momentum.
You can also express this something similar to Newton’s law or something similar to
something similar to Fourier’s law, where the term eddy viscosity or eddy viscosity or
the eddy transport eddy thermal conductivity would give you the additional transport. So,
this is very good to understand the effect of turbulence, but as I have said before there
are many statistical models, which try to explain, which try to quantify the eddy in
turbulent flow the contribution of eddy in turbulent flow.
So, even though it qualitatively gives a very good idea very good picture of the whole
situation, we still do not know how to have a model based on statistical fluctuations,
based on a study of the fluctuation, based on a statistical study of the fluctuation in a
flow field both in terms of velocity fluctuation and temperature fluctuation. And to come
to a compact expression of the heat transfer or of momentum transfer.
So, we have to resort to semi empirical analysis of these processes and couple the results
that, we have from our hydrodynamic treatment of turbulent flow and couple it with that
of the thermal treatment, the treatment or the quantification of the heat transfer in
turbulent flow.
286
So, in tomorrow in the next class, we will look at turbulent heat transfer for from a flat
plate in contact with the flowing fluid, where the temperature of the flowing fluid is
different from that of the solid plate. So, when we have turbulent flow what would be the
expression for Nusselt number for such cases and, interestingly as I have said before the
Reynolds number for flow over flat plate the length scale that is defined is the axial
x v
position. So, the Reynolds number is defined as and the x keeps on increasing as

you move along the plate.
So, therefore, it is likely it is possible, that you are going to have laminar flow to begin
with and at certain position on the plate the x becomes. So, large the Reynolds number
become so, large that it changes from laminar to turbulent. Even though the transition
from laminar to turbulent is going to take place over a length not at a length, you
realistically cannot have a situation where at this point anywhere on this side is laminar
and anywhere on the other side is turbulent, it never happens it always takes place over a
range, but we put for convenience we put a value of 5 × 10 5 a Reynolds number a value
of 5 × 105 to quantify presence or absence of turbulence.
In our previous classes we have seen what is the Nusselt number relation for the first
part, from 0 to xc over which the flow is laminar? In the next class we would see using
semi empirical relations already available to us from momentum transfer what is going to
be the expression for Nusselt number, when the flow switches from laminar to turbulent.
Such that together it would give me a complete quantitative idea of heat transfer over a
flat plate from a flat plate, when there is you have laminar flow and you have turbulent
flow. Or in certain cases we can trip the flow from the very beginning at x equal to 0
such that the entire flow becomes turbulent.
So, these kind of different situations, we will derive and discuss in next in the subsequent
class and there will be lots of problems, which we will solve over here to give you an
idea not only the concepts behind this equations, but how to use this equations for
practical calculations of heat transfer.
287
Heat Transfer
Lecture - 25
Turbulent External Flow
In the previous class we were discussing about the concept of eddies and, how eddies
transport additional momentum and heat in turbulent flow. So, whenever there is
turbulent flow, the small packets of fluids, otherwise known as eddies are automatically
formed in the moving fluid. These eddies transport additional momentum and energy
from one point to the other, as they travel in the moving stream. They are the reason why
we encounter higher values of heat transfer, and momentum transfer in turbulent flow.
We have also discussed that how the growth of the boundary layer, which was
continuous and slow for the case of laminar flow, suddenly starts to increase at a rapid
rate. There is no clear demarcation about where laminar flow ends and turbulent flow
begins, but just for convenience we assign a specific value of axial Reynolds number to
the flow. We define this Reynolds number using x, that is the axial length scale of the
system. So, as I move into move along the plate the value of x increases, the value of
Reynolds number also increases. And therefore, there would be a transition number
assigned as 5 105 beyond which it is assumed that the flow is entirely turbulent.
So, based on this and based on the concept of eddies there are two terms which we have
discussed in the last class, eddy diffusivity and eddy thermal conductivity. Now both
these terms are to be evaluated experimentally, and there are lots of theories which have
been proposed to estimate the value of these eddy diffusivities. However, a clear-cut
expression for the eddies is yet to be universally accepted. So, people are using different
techniques to decide about what would be the momentum transfer in laminar flo, as well
as in turbulent flow and what would be the value of heat transfer. And the parameter of
interest in convective heat transfer is the heat transfer coefficient, denoted by h or the
dimensionless form of this is known as the Nusselt number.
So, the entire study of convective heat transfer is to is to obtain the correct expression for
the Nusselt number for a variety of flow conditions. Similarly, the for the case of
momentum transfer the engineering parameter of interest is C f or the friction coefficient.
288
We have already derived the expression for Cf in laminar flow; but, the question still
remains: how we extend this expression for Cf in turbulent flow and is there a way
through which once I know the expression for Cf, the expression for Nusselt number can
be automatically deduced. In other words, we are trying to see that if there exists a
similarity between momentum transfer or Cf and heat transfer there is Nusselt number.
So, the relation between Cf and Nusselt number and under what condition this relation of
equivalence between Cf and Nusselt number can be used is going to be the topic of this
class and 2 or 3 subsequent classes. Because, what I have told you earlier is that the heat
transfer is coupled to momentum transfer, but momentum transfer is not couple to heat
transfer provided the thermo physical properties do not change. So, we can solve fluid
mechanics part of the transport phenomena independent of heat transfer. But the heat
transfer part cannot be solved without solving the momentum transfer part. So, our
efforts are directed to solve the problem which is simpler of these two, that is the
momentum transfer in turbulent flow. I will not discuss about the technique or the
method that was used to obtain an expression for C f. You probably have read it in your
fluid mechanics class, if you have not there is an excellent treatment of turbulent flow
expression for Cf, in the text book fluid mechanics by Fox and McDonald; wherein they
have introduced something called the momentum integral equation. It is to be noted that
any quantities that are obtained out of the solution of momentum integral equation, are
not point values, but they are length averaged values. So, momentum integral equation
provides a powerful tool to examine what is happening for flow inside a boundary layer.
This momentum integral equation does not have restrictions in terms of its applicability,
i.e. it could be used for both laminar and turbulent flows. Whereas, the Navier stokes
equation is valid only for laminar flow, whenever there is turbulence present in the
system the additional terms would appear to account for the enhanced momentum
transfer. And those additional terms collectively are known as Reynolds stresses. So, the
Navier stokes equation as such is not valid for turbulent flow. And these Reynolds
stresses are extremely difficult to evaluate and get a specific value. Therefore, the
integral momentum integral equation gives us a way to address turbulent flow inside the
boundary layer. Another advantage of this is that, the solution of the momentum integral
equation for Newtonian fluids would give rise to an ordinary differential equation, which
can simply be integrated to obtain the parameter of interest, which is δ, the thickness of
289
the boundary layer as a function of Reynolds number. The same way it was done for the
case of laminar flow.
And, once you have δ, then it is also possible to obtain the expression of C f, but in there
we have to invoke the Blasius friction factor or in other words the f that you see in
Moody’s diagram of friction factor versus Reynolds number. So, with the knowledge of
friction factor versus Reynolds number as given in Moody diagram and using
momentum integral equation it is possible to obtain an expression for C f in turbulent
flow. So, I would not derive that expression since it is something which probably has
been covered in fluid mechanics.
So, I leave that part and will directly provide you with the expression for C f in the case of
turbulent flow or the expression for delta with x or with Reynolds number for the case of
turbulent flow. The important thing that I would like to show you in this class is once
this Cf and δ they are known to us, we can determine the expression for Nusselt number,
when we have heat transfer as well inside the thermal boundary layer.
So, our analysis with the a priori knowledge of the expression of C f that was obtained
from the fluid mechanics, momentum integral equation, Moody’s diagram and so, on.
So, if we come to see over here you initially have laminar flow, and at x c (the transition
point), the flow changes from laminar to turbulent. And as you can see the growth of the
290
turbulent boundary layer is much more rapid as compare to the growth of the laminar
boundary layer.
So, the expression for Cfx,. from experiment and theory, as a function of the axial location
1
can be expressed as C fx  0.058 Re x 5 . This expression is valid for a Reynolds number of
5 105 to 107 . And it can also be shown that δ, the hydrodynamic boundary layer
 1
thickness by x where x is the axial location is:  0.37 Re x 5 . If you wish, you can see it
x
in the reference, fluid mechanics by Fox and McDonald there they have shown how
these two expressions are obtained using momentum integral equation. So, we will not
solve this will take these two and proceed to see, how our knowledge of momentum
transfer in turbulent flow can be extended expanded to take into account the heat transfer
in turbulent flow. In order to do that, I start with writing the continuity equation which is
a statement of the conservation of mass; and the momentum equation for flow inside the
boundary layer. I do not neglect the diffusive transport of x momentum in the x direction.
u v
The equation of continuity is:   0 ; and the x-component of the momentum
x y
u v 1 P  2u
equation is: u  v    2 .
x y  x y
291
We understand from our previous discussion, that the thermal transport in the x direction
by conduction is small as compared to the transport of heat in the y direction, because of
the small length scale that is the boundary layer thickness, which is associated in the y
direction. And for most of the practical cases most of the cases we encounter, the viscous
dissipation is small and can be set equal to 0. This is a heat which is generated by joule
heating or by nuclear sources etc. Therefore, the energy equation can be written as:
T T  2T
u v   2 and I call this as equation 3. So, we are going to work with these 3
x y y
equations, the continuity equation, the x component of the momentum and the energy
equations.
We start with non-dimensionalizing the equations. So, if I non-dimensionalize these

equations it would give us something like the similarity of the transport processes. And,
the similarity of the transport processes is going to be valid for certain conditions. So, we
have to identify the conditions, that one needs to check before they can use this similarity
between momentum transfer and heat transfer.
Now, whenever you try to do something an exercise like that, the first thing one should
do is to non-dimensionalize the equations. The advantage of non-dimensionalizing any
governing equation is that at times you would see the automatic emergence of certain
dimensionless groups, which would give you a clearer picture of the transport
292
phenomena taking place inside the system. In simpler terms, if you have Reynolds
number to be equal to 5000, you know what is the flow condition that exists inside the
boundary layer. So, it does not matter whether you have flow of air or flow of water or
flow of glycerin over a flat plate. If this specific value of Reynolds number is identical,
then you can get a solution of the dimensionless film thickness or the friction coefficient,
which can be compared for flow between three dissimilar fluids. So, the behavior of the
flow or behavior of heat transfer inside the boundary layer will be identical if the
Reynolds number is the same. But, would I need to specify some other dimensionless
groups or some other conditions such that I can say it is not only the flow the heat
transfer would also be the same. So, that is what we are trying to do. So, the first step of
doing that is non-dimensional the governing equations and see, which groups come out
of that automatically.
So, the parameters that we use for non dimensionalization are all dimensionless
x
quantities: x*  where L is the length of the plate over which the flow takes place;
L
y u
y*  ; u*  , where V this is the x component of velocity, the y component of
L V
v * T  Ts
velocity is v*  ; the dimensionless temperature is defined as T  , where TS
V T  Ts
is the temperature of the solid surface of the plate, and T is the temperature at a point far
from the plate in the y direction.
293
First, I would write what is the going to be the governing equation? So, the governing
u * v * P *   2u
equation for the case of flow is going to be: u *  v*   * and
x * y * x * VL y *2
this is for the x component of the momentum equation. And, if I write the energy
T * T *   2T *
equation it would be u *  v*  . Upon inspection, we observe that the
x * y * VL y *2
coefficient on the right-hand side of the dimensionless momentum equation turns out to
1
be , which is nothing, but the Reynold’s number based on the entire length of the
Re L
plate. So, non dimensionalizing the equation automatically projects up a similarity
parameter, which that we know from our experience a priori knowledge that it has to be
Reynolds number. Similarly, the coefficient on the right-hand side of the dimensionless
1
energy equation turns out to be: . Therefore, we note that, the similarity
Re L Pr
parameter for heat transfer is Reynolds number as well as Prandtl number.
Now, let us look at what is going to be the boundary conditions. So, the first boundary
condition that I am going to write is at the wall. At the wall, we have the no-slip
condition; that means, both the x component velocity, which is u* at any value of x*, but
294
since it is at the walls y* = 0; would be equal to 0. Similarly, the y component of velocity
at any value of x star, but at y* = 0 should also be equal to 0. So, these two are nothing,
but the no-slip condition. We are also aware of what is going to happen at the free
stream; at the free stream; that means, the value of u* for any value of x star, but at y to
u
infinity, which is simply going to be at this value of x*. It is to be noted that the
V
velocity with which the fluid approaches the flat plate is V, and the constant velocity
outside of the boundary layer is denoted by u .
So, if it is a curved plate then the v and u will differ in values, but for flow over a flat
plate this V is equal to u . Therefore, in order to keep the generality of the solution I
have written it as u*, which is defined as u star at infinite value of y; that means, u*
u
when y is somewhere over here it simply going to be , but we understand that for a
V
u
flat plate is going to be equal to 1.
V
295
Similarly, let us try to see what is going to be the condition T* at x* = 0. Remember, that
* T  Ts
T* is defined as T  . So, when y is equal to 0 this T would simply be equal to T
T  Ts
S. So, this is going to be equal to 0. At the free stream, the value of T* at any x*, but at
infinite distance from the plate then at that point T simply becomes equal to T infinity.
So, the temperature of the fluid flowing over a flat plate at a point far from the plate; that
means, mathematical speaking y equal infinity, this T is simply going to be equal to T
infinity and I am going to have this to be equal to 1.
So, this would give us the starting point to derive, when and how the momentum transfer
part of it is going to be like that of the heat transfer. So, let us talk something about the
governing equations, the dimensionless forms and that the boundary conditions. Now,
these two equations one governing momentum transfer and one governing heat transfer,
they would be identical provided two parameters are equal. If your Reynolds number and
Prandtl number is equal to 1, then in dimensionless form there is no difference between
equation 1 and equation 2. I repeat once again. If the Reynolds number values are same i
for the case of momentum transfer and for the case of heat transfer. And, if in addition
Prandtl number is equal to 1, then these 2 equations become identical. In one case we are
talking about the x component of velocity, in the other case we are talking about the
dimensionless temperature.
So, the dimensionless form of u and the dimensionless form of T would be identical
provided these similarity parameters are equal and the value of Prandtl number is equal
to 1. And, if in addition when you look at the boundary conditions, if in addition the flow
is taking place over a flat plate such that u* at x* and y* equals infinity, if this is on a flat
plate, then this boundary condition and this boundary condition would be identical.
These 2 are identical anyway so, for a flat plate the second boundary conditions would
also be identical. So, therefore, it clearly gives us an idea of what is going to be the
similarity? The equivalence between heat transfer and momentum transfer. So, the
dimensionless form of the equation simply tells me that, if the Reynolds number
numbers are kept constant and if Prandtl number is equal to 1, and if the flow takes place
over a flat plate then the two governing equations and the boundary conditions are all
identical.
296
So, when that happens when the governing equations and the boundary conditions are
the same, the two systems become dynamically similar. And, for two dynamically
similar systems the expression of the dependent variable can simply be replaced by the
expression of another dimensionless variable. Or in other words what I am trying to say
is that, if the dynamic similarity between heat transfer and fluid mechanics is achieved,
then any expression of u* can be modified as an expression of T* by simply replacing
the relevant quantities with the relevant dimensionless quantities in the second transport
case.
So, 2 dynamically similar systems I know what u* is going to be, if I know that then any
derivative of u* can also be substituted to obtain what would be the expression for T*.
So, the if I know the velocity expression and if I know the velocity gradient expression,
then through dynamic similarity I should be able to obtain what is the temperature
expression, and the temperature gradient expression.
And this approach when the two systems are made dynamically similar through the
choice of equal Reynolds number and through the peculiar choice of Prandtl number
equal to 1, because Prandtl number may not be equal to 1 for many of the fluids, but if it
is equal to 1 then complete equivalence between momentum transfer and heat transfer
will exist. And, the analogy between these two for a special condition of equality of
Reynolds number and unit value of Prandtl number is known as Reynolds analogy.
297
Heat Transfer
Lecture - 26
Heat and Momentum Transfer Analogy
So, we were discussing about the dimensionless form of the momentum equation,
dimensionless form of the energy equation, the boundary conditions again in
dimensionless forms; essentially for the case of fluid flow, the no slip velocity and what
would be the condition on what will be the condition of velocity at a point far from the
plate such that at that point the velocity is going to be equal to the local free stream
velocity outside of the boundary layer. And similarly, we are also looking at the energy
equation in what would be the form of the boundary conditions?
For example, what is going to be T* that is the dimensionless temperature at any axial
location; but on the plate itself? So, T * ( x* , 0)  0 , that means, y* would be equal to 0
* T  Ts
because of the way we have defined the dimensionless temperature T*. But T  ,
T  Ts
so, on the plate T is equal to Ts; therefore, T* would be equal to 0. At a point far from the
plate, the temperature of the fluid would simply be equal to T ∞ and the value of T* in that
case would be equal to 1.
So, we were looking at 2 equations to governing equations for two processes; one for
heat transfer and the other for momentum transfer. And we saw that the ones that
separate, the combination of terms that separate these 2 equations that differentiates
between these 2 equations are the presence of similarity parameters. One is the Reynolds
number for the case of momentum transfer and the second is Reynolds times Prandtl
number for the case of heat transfer.
So, these are the only difference between Heat transfer and Momentum transfer. So,
what we would like to do is we have proposed then that if we could keep the Reynolds
number to be the same for heat transfer as well as for momentum transfer and if we can
choose a hypothetical fluid with a Prandtl number to be equal to 1, then these two
equations dimensionless form of these 2 transfer equations are identical.
298
And if in addition, we assume that the flow is taking place over a flat plate, then the
governing equation the boundary conditions of the governing equations are also going to
be identical. So, that is the case of dynamic similarity which they tell us that for a
dynamic similar system, the expression of dependent variable for the case of momentum
transfer which is u* can be replaced by the dependent variable of the other equation
which is T*.
And therefore, an analogy, a similarity and equivalence between the momentum transfer
and heat transfer can be established to obtain expressions, known expressions of one
dependent variable from the known expression of another dependent variable. So, will be
look at that it will be very clear towards the end of this class, how it is done.
So, let us look at this slide which was the last slide of the previous class, where I have
identified the governing equations, the similarity parameters, Reynolds number and
Prandtl number. This is for momentum; this is for energy and the boundary conditions
using no slip and a point far from the flat plate, what would be the velocity condition, the
temperature at y = 0 and temperature at y = ∞.
So, with this knowledge when we by keeping the Prandtl number to be equal to 1 and
keeping the Reynolds number to be the same and assuming that the flow is taking place
over a flat plate; everything in this equation in between these equations and the boundary
conditions are identical, so we have a similar system, dynamically similar system.
299
So, what I am going to what we going to do is find out what is Reynolds analogy and
modified Reynolds analogy. So, for that I am going to look at the function what could be
the functional form of u*. I do not know what would be the exact form of it; but I know
that if I could write the functional form of u*, it should contain the independent variable
x*, independent variable y*, the similarity parameter Reynolds number and the pressure
dp*
gradient present in the system which is .
dx*
*  * * dp* 
So, u  f1  x , y , Re L , *  .
 dx 
I do not know how u is going to be connected with x, y or Reynolds number, but I know
that a functional form like this would exist for the case of flow. Now in terms of
engineering interest we would like to find out what is the shear stress at the surface? That
means, at this by at the surface, I mean at y* to be equal to 0 that is at the surface.
So, which I let us say call it as  s , the shear stress which would be
u V u *
s   y 0  y*  0
y L y *
300
V u *
It is simply going to be y*  0
after non-dimensionalization.
L y*
So, that would give me the expression for shear stress and shear stress coefficient, we
understand that by definition its
*
s 2 u
Cf  
1 Re L y * y*  0
V 2
2
Where, V is the approach velocity, ρ is the density. So, that is a definition of C f. The
dimensionless form was obtained by putting the value of  s over here and absorbing this
1
V 2 in it .
2
*
u
So, this if I would write take write to find what is the, what is the * y*  0
.
y
So, if you look at the expression the functional form, the hypothetical functional form of
*
u
u*, I am trying to find out y*  0
. Since, I am assigning a specific value of y * to be
y*
*
u  dp* 
equal to 0; this must be a *  f 2  x* , * , Re L  . Since, I have specified the value
y* y 0
 dx 
of y* to be equal to 0. So, therefore, the y*does not appear over here.
Now, this is the flow; this is a flat plate over which the flow is taking place and this side
is the turbulent flow. Now, if the geometry is prescribed, then you would be able to
dp*
obtain separately. So, this for a prescribed geometry, I will explain on it in a
dx*
moment. Remember that what I have told you before is between in inside the boundary
layer, the flow is viscous; outside of the boundary layer, the flow is inviscid. So, there is
no effect of viscosity in here. Since, you have viscosity present in effect of viscosity
present inside the boundary layer, you cannot use known equations which are available
to give which are there to provide what is the pressure drop as a function of distance.
301
Now, if someone tells you that what is the equation that provides the pressure drop in a
flow? The name that comes to your mind is Bernoulli’s equation because Bernoulli’s
equation would relate the pressure gradient the pressure head, the velocity head and the
gravity head. Now, if I assume the plate to be horizontal which is the case in this case.
So, therefore, the it is going to be the summation of pressure head and velocity head to
be constant. So, if I know this velocity or I can express the change in pressure in terms of
change in velocity head that is what Bernoulli’s equation is all about. Now, there is a
catch though; the Bernoulli’s equation is strictly valid for inviscid flow for flow where
the effect of viscosity is absent.
So, inside the boundary layer, technically I cannot use Bernoulli’s equation as the flow is
viscous there. So, this solution; but the observation is outside of the boundary layer the
flow is inviscid. So, if the geometry is known to me then I would be able to use
Bernoulli’s equation in the flow domain outside of the boundary layer to obtain an
dp dp*
expression for or * independent of everything.
dx dx
dp*
So, if someone gives me the geometry I should be able to obtain, what is outside of
dx*
the boundary layer through the use of Bernoulli’s equation and since the thickness of the
boundary layer is very small, there is no change in pressure with y. That is an assumption
which is a valid assumption considering the small thickness of the boundary layer. So, I
dp* dp*
use Bernoulli’s equation to find out what is . So, can be obtained and for a
dx* dx*
dp*
prescribed geometry is a constant; for that reason the from the expression of
dx*
*
u dp*
* which had otherwise contained , I can drop that. Since for a given
y* y 0
dx*
geometry this pressure gradient is known to me apriory and is a constant.
So, in terms of functional form of the equation whatever I have written over here can
simply be written once again as
302
*
u
y*  0
 f 2  x* , Re L 
y*
Now, now if use
*
2 u
Cf = y*  0
Re L y *
This is where I have obtained the expression for Cf. So, my Cf is simply going to be
*
2 u 2
Cf  y*  0
 f 2  x* , Re L 
Re L y* Re L
So, these are the two equations that one needs to take a look at. First of all, u is a
function of all the independent variables, the operational parameter and the pressure
gradient. From there I obtained the shear stress; from the shear stress, I obtained C f and
*
u
for y*  0
, I obtained the functional form for this special case when the geometry is
y*
known to me. So, this should give me the expression for C f for flow momentum transport
inside a boundary layer. Now, let us see what is going to happen to the temperature
profile? So, if I look at the temperature in expression over here that we have obtained.
303
My temperature profile T* would be a function of u*, x*, v*, y*, Reynolds number and
Prandtl number; but this u*and v*are already a function are already known function of x *
and y*and so on. For example, in this expression itself what we have seen is that u *is a
function of once; you specify x,y, Reynolds and dp/dx, u*is specified.
So, in the governing equation here you do not need to write T * is a function of u* because
the moment you write T*is a function of x*, y*and Reynolds number, you essentially
specify u*. So, by incorporating u*once again in your functional form that would be
simply a repetition.
304
So, therefore, based on the knowledge of the of this governing equation, one should be
able to write the functional form
 dp* 
T *  f 3  x* , y* , Re, Pr, * 
 dx 
dp*
This I am keeping it just to make it complete.
dx*
dp*
But we understand that for a prescribed geometry, I can drop this . So, the same way
dx*
I have done it for the case of shear stress. I am going to write the same thing for the case
of surface heat flux which I call it as qs. So, this is a solid plate, you have the profile and
have flow taking place; I am trying to find out what is the surface heat flux at y * equal to
0. So, the surface heat flux is
T
qs   k f y 0
y
Where k is thermal conductivity of the fluid .
So, that is the Fourier law equivalent.
That is the Fourier’s law in which can be expressed as
305
T
k f y 0
T y
qs   k f y 0  h
y Ts  T
Because my qs, this is the equality of conduction and convection at this point, at the point
where the liquid molecules by due to no slip are stuck to the solid.
So, the heat transfer from the immobile liquid molecules to the mobile liquid molecules,
there you have the conduction and convection equality. So, this q s can be expressed in
terms of Fourier’s law; this q s can also be expressed in terms of Newton’s law which is
h times Ts  T . So, h times Ts  T is also equal to, so , these two are simultaneously
valid at y equals to 0 and therefore, the expression for h can be obtained in this fashion.
So, when you express it in dimensionless form this becomes
 k f  T  Ts  T *
h y * 0
L  Ts  T  y *
So, this gives I am slowly moving towards the dimensionless form of the expression over
here.
So, when you do that when you cancel that the numerator and the denominator what you
would get is
306
k f T *
h y * 0
L y*
hL T *
Or,  Nu  y * 0
kf y *
So, what is hL/k, this is nothing but the Nusselt number. So, we it in convection, we
always try to find what is h or what is the expression for Nusselt number? So, now, I
write a Nusselt number is used for f1 f2 and f3 over here. So, Nusselt number is
T *
Nu  y * 0 .
y*
So, Nu x  f 4 ( x* , Re L , Pr)
T *
when I say its y * 0 , that function should be a function of x* , Re L , Pr provided the
y*
geometry is known to us.
So thus, this Nusselt number expression would be some function f4; I do not know what
this f4 would be? But, some function of x*, Reynolds number and Prandtl number. So,
this is obviously, for a prescribe geometry and if you would like to find out what is the
average value of Nusselt number, length average value of Nusselt number; the moment
you do that, the length average value of Nusselt number; then, x * is obviously dropped it
should be another function f 5  (Re L , Pr) .
So, this is the local value of Nusselt number, this is Nu x this is the, so, this is local value
of Nusselt number and this is the average value of Nusselt number and the bar over Nu
simply denotes it is the average value which is the function of this and for length average
value, it would simply be a function of Reynolds number and Prandtl number.
Now when we use the Reynolds condition, Reynolds analogy; what is I see that dp/dx is
0 and Prandtl number is equal to 1 and if that is the case, then the expression of u * and T*
star must be identical. This is what we were discussing so far. So, the expressions of u *
and T* must be identical. So, what is expression of T*and u*? So, u*is f1 and T*is f3. So, if
307
your if your Prandtl number is equal to 1. So, the equation becomes dynamically similar;
dp/dx is the dependence of dp/dx is not there.
So, therefore, f1 must a f1 must be equal to f1 and f3; f1 and f3 are going to be identical ok.
So, f1 and f3 are identical. It is also then true that the expression for the friction
coefficient which is this f2 must also be equal to the f4 which is the relation for this case.
So, expression for u* and T* must be identical would simply give you that f1 is equal to f3.
And true also for friction coefficient and Nusselt number; so, if it is true for friction
coefficient and Nusselt number what you would get is f 2 is equal to f4. So, these are
collectively known as the Reynolds analogy. The important point here is the major
problem that you one would face in the practical application of Reynolds analogy is the
requirement that Prandtl number has to be equal to 1.
Where are you going to get a fluid whose Prandtl number is equal to 1 and if it is equal
to 1, how are you going to use this analogy for other cases? So, if f 2 is equal to f4; how
does how does that help us? f4 is this, f4 and f2 if these 2 are identical; if f2 and f4 are
identical, I will write these 2 equations once again to show how we can use them in this
case.
So,
308
*
2 u 2
Cf  y*  0
 f 2  x* , Re L 
Re L y* Re L
Nusselt number is simply equal to Nu x  f 4 ( x* , Re L , Pr) .
Re L
So, if f2 and f4 are equivalent, then what we can say is that C f .  Nu .
2
So, this is known as Reynolds Analogy. This is to some extent in some cases, it is
modified in a slightly different way; where it is written that
Cf Nu
  St
2 Re L .Pr
And since the value of Prandtl number is equal to 1, there is no harm in adding a Prandtl
number in this case. I can do that since Prandtl number in Reynolds analogy is equal to 1.
So, this Nusselt by Reynolds into Prandtl has a special name which is called Stanton
Number. So, I can use Stanton number the I can introduce Stanton number, since, the
value of Prandtl number is equal to 1.
So, the more general form of Reynolds analogy is
Cf
 St
2
This is the commonly used form of Reynolds analogy. So, this connects the key
engineering parameter of Cf in fluid friction with h on Nusselt number in convective heat
transfer. So, I would also like to draw your attention to the previous slide that I was
T *
showing Nusselt number is equal to Nu  y * 0 .
y*
This again reinforces my statement that the significance of Nusselt number is nothing but
the dimensionless temperature gradient at the solid liquid interface.
So, that would be the definition of Nusselt number. The more important one is a Nusselt
number contains h; this is an engineering parameter and here I connect Nusselt number
with Cf friction coefficient which also is an engineering parameter. So, through the use of
309
this analogy, I connect the heat transfer with momentum transfer; but there is as I
understand, there is a problem that is only valid for the case when Prandtl number is
equal to 1. So, therefore, in order to extend the validity of Reynolds analogy two
situations; 2 fluids whose Prandtl number may not be equal to 1; a correction factor is
added to this analogy and then, it takes the is called the modified Reynolds analogy.
and is also known as the Chilton Coulburn Analogy to extend the Reynolds analogy. A
2
correction factor is added to this as St.Pr 3 . So, this is the correction factor which is
added
Cf 2
Nu 2
 St.Pr 3
.Pr 3
2 Re.Pr
This extends the Prandtl number to a large range of Prandtl number. So, what you get
then is
Nu  13 C f
.Pr 
Re 2
2
This whole thing ( St.Pr 3 ) is called the Coulburn “j” factor.
310
So, this is the expression for modified Reynolds analogy or Chilton Coulburn analogy
and the validity of this is extended in most of the real systems real fluids, they have
Prandtl number in this range; except for heavy oils which has Prandtl number more than
60 and the other extreme is liquid metals which as Prandtl number way below 0.6. So,
for liquid metals and heavy oils, if we exclude these 2 special type of fluids most of the
liquids the most of the fluids that you ordinarily use, ordinarily come across would be in
this range. And therefore, the Chilton Coulburn analogy extends for a wide range of
Prandtl number.
The advantage, what is the advantage? The advantage is as I mentioned C f expression is
1
already known to us C  0.058 Re  5 ; put it in here and what you get is an expression
f x
for Nusselt number as
4 1
Nu  0.029 Re 5 .Pr 3
The range of validity between Prandtl number 0.6 and 60. See the beauty of it. This is
something which is really really interesting. You have got an expression for Nusselt
number, you have got an expression for h by simply using an analogy which has solid
foundation. So, you the expression for Cf is known to you; you are looking at the
governing equations, non-dimensionalizing the governing equations; the similarity
parameters clearly obtained out of this exercise.
You look at the dimensionless boundary conditions; see under which condition these 2
governing equations become dynamically similar. The moment they become
dynamically similar, the solution of one can be used as the solution of the other. So,
*
u T *
y*  0
which is connected with C f can be substituted by y * 0 which is connected
y* y*
with Nusselt number.
So, the gradient of velocity or the gradient of temperature, all in dimensionless form; one
related to Cf, the other is related to Nusselt number. The momentum with them
dynamically similar, these 2 gradients are identical and what you have then is an
expression for Cf and an expression for Nusselt number. The expression for Cf is already
311
known to you. Therefore, you obtain an expression for the Nusselt number in turbulent
flow.
So, without getting into the complicated statistical analysis of eddy formation, velocity
distribution, unknown velocity distribution, the fluctuations in temperature and in
velocity; you have a tool now through the use of an analogy and an extended analogy by
incorporating Prandtl number corrections, you now have the expression for convective
heat transfer coefficient in turbulent flow. That is the beauty of this analysis or this
analogy.
So, Reynolds analogy or modified Reynolds analogy which is also known as Chilton
Coulburn analogy is a powerful tool which lets you find out the expression for h in
highly turbulent flow. So, now, I have the complete picture in heat transfer; external heat
transfer, flow the heat transfer in external flow simplest possible example flow over a flat
plate. I have an expression for h in the early part where the flow is laminar up to a value
of Reynolds number 5 ×105. And through the use of analogy, I have an expression for the
Nusselt number beyond Reynolds number 5 ×105; that means, when the flow is turbulent.
So, together they give me a complete picture of what would be the heat transfer
coefficient in laminar flow and what would be the heat transfer coefficient in turbulent
flow? More importantly, this I would show you the next class a corollary of that is the
flow is never fully turbulent and the flow can change from laminar to turbulent. So, in
most cases any flow has a laminar portion to begin with and then, it becomes turbulent.
So, those kind of flows are commonly encountered they are known as Mixed flow. The
early part its laminar later part it turns turbulent. So, how these relations can be modified
to express the average heat transfer coefficient for the case of mixed flow. But that is
there is no new concepts and involved there. What is important is again, I would bring
your attention to this equation which simply gives you the Nusselt number for the case of
turbulent flow as a function of Reynolds number and as a function of Prandtl number.
I should mention as I was telling you this is when the flow is turbulent from the
beginning. So, when the flow is turbulent from the beginning. This expression can be
used to obtain to obtain the value of h and so on. But in most of the cases the flow is
laminar to start with and then it turns turbulent those kind of flows are known as Mixed
flow.
312
So, I will give you the expressions for mixed flow based on the expression of Nusselt
number in laminar flow and in turbulent flow in the next class. But however, I would
once again write the Nusselt number for the case of laminar flow which here just to
1 1
compare them is Nu  0.332 Re 2 .Pr 3
So, this is for laminar flow and this one is for turbulent flow. So, together if you combine
this and this together what I get is mixed flow. But this is obtained almost completely
analytically, this has some approximation built into it; but it gives us the analogy is give
us a powerful tool to convert heat transfer data from the momentum transfer obtain an
expression for heat transfer and vice versa.
So, will solve here quite a few problems on this to clarify the concepts and to show you
how this analogy can be effectively employed in problem solving.
313
Heat Transfer
Lecture - 27
Mixed Boundary Layers
We are going to look at what is mixed boundary layer; that means, we would understand
that in real situations, a boundary layer should always start as laminar and after you cross
a certain value of Reynolds number which we have decided to be equal to 5 105 , it
changes from laminar to turbulent. But we understand that the situation is not going to be
purely turbulent from the beginning. It starts to become turbulent from some point which
is denoted by xc in the figure. So, if I have a flow over a flat plate and if I have
simultaneous heat and momentum transfer, the expression for turbulent flow heat
transfer coefficient is valid only when the flow is turbulent from the very beginning. That
can happen if there is some sort of turbulence promoter at the edge of the plate which
disturbs the flow in such a way that it becomes turbulent even though the value of
Reynolds number at that point is very small. So, in the absence of turbulence promoter
you are always going to have a laminar part first and then, it becomes turbulent.
xc
Therefore, if we consider  0.95 i.e. if the transition occurs at the very beginning,
L
314
then one can use this equation (refer to the equations in the previous lecture) for the
entire length even though we understand that there must be some sort laminar flow to
start with.
However, to remain more in-tune with what happens in the real situation the expression
for hL should consider the value of the laminar heat transfer coefficient and the value of
the turbulent heat transfer coefficient. And if I am making an average of the heat transfer
coefficient, the length average heat transfer coefficient, I need to integrate this from 0 to
xc; where xc is the distance at which the flow changes from laminar to turbulent. The
expression for h turbulent will have to be substituted from here. So, from 0 to xc and xc to
L, if we integrate with the appropriate expression of the convective heat transfer
coefficient and dividing the whole thing by L; I should be able to obtain what is the
1  xc L
average value of the heat transfer coefficient: hL   hlam dx   hturb dx  .
L  0 xc 
Substituting the relevant expression would result in:
1 4
k U  2 xc 1 U  5 L 1 
hL   0.332     dx  0.029     1
hturb dx  .
L    0
x    xc
x 5 

We have discussed previously that V would be equal to U infinity for flow over a flat
plate which is also known as the Zero pressure gradient flow. But to keep the generality
of the solution intact, we are using U infinity over here. I haven’t done anything
additional here, I just identified that up to x c, the flow remains laminar, and beyond xc it
becomes turbulent. So, to obtain the heat that is lost from a hot surface when it is in
contact with a flowing fluid, you simply use the average value of a heat transfer
coefficient. So, the heat loss from the plate would be hL A  Ts  T  . If you want to find
out the heat loss from a slice on the plate, you must use the local value of heat transfer
coefficient at that point; but, if you want to find out what is the total heat loss from the
substrate, then you have to use the average value of heat transfer coefficient.
315
So, that is the whole idea the aim a objective of this exercise. So, this is what the
complete expression is and once you integrate this what you are going to get here is
Nusselt number based on the length.
The average value of Nusselt number would be:
1
 4 4
 1
NuL   0.664 Re x ,2c  0.037 Re L5  Re x ,5c  Pr 3 . So, over here I can put Re x, c that is the
 
Reynolds number at transition to be equal to 5 105 as per our consideration, then the

length averaged value of Nusselt number would simply be Nu L  0.037 Re 5  871 Pr
4
 1
3
. This is valid for conditions at 0.6  Pr  60 ; and 5 105  Re L  108 , with Rex, c =
5 105 .
Similarly, the friction coefficient, Cf L when you have both laminar and turbulent can be
0.074 1742
written as: C fL  1
 ; and this is valid for 5 105  Re L  108 with Rex, c =
Re 5 Re L
L
5 105 .
So, these two relations together are for mixed flow. So, if you have mixed flow on a
plate, where heat transfer or momentum transfer is taking place; then the corresponding
316
relations to be used are these two. It is to be noted that all the fluid properties for
example, in Reynolds number the ρ, the µ; in Nusselt number the thermal conductivity
(k) etc., are to be evaluated at the arithmetic average of the temperature at the infinite
position and the temperature of the surface. This average is known as the Film
T  Ts
Temperature ( ). And the property which is most affected by a change in
2
temperature is the viscosity (µ). The thermal conductivity and other properties are not a
strong function of temperature. So, for special cases with very large variation in the
n
 
temperature from T s to T infinity a viscosity correction factor (    ) is sometimes
 s 
added to the expression that we have we have derived so far.
So, this concludes our discussion on what is going to happen when you have laminar
over a flat plate; when you have turbulent flow over a flat plate and the most likely
scenario of when you have mixed flow on the flat plate. So, the first thing that you do
while solving a problem is first find out the Reynolds number based on the entire length
of the plate and see if this is greater than 5 105 . If it is greater than 5 105 , then you are
going to have laminar flow at the beginning and turbulent flow thereafter.
So, your relation to be used would be the relation for mixed flow. On the other hand, if
you do not reach the transition Reynolds number that would mean that the entire plate is
going to be under laminar flow and you choose your equation appropriately. In some
problems, it would be mentioned that the flow is disturbed from the very beginning in
such a way to induce turbulence in the system. So, if that is mentioned then you can
safely use the turbulent flow equation without considering the presence of a laminar
region at the beginning.
Now, comes the application. We will start with the which one problem today and will
solve other problems of different types all involving external flow and when that is done,
we will start our study of internal flow. So, the first problem that we are going to look at
is about the Analogy.
317
The first problem is about the modified Reynolds analogy or Chilton Colburn analogy.
What it mentioned is that for flow over a flat plate with extreme roughness when I
mention extreme roughness, it simply tells me that the relations that we obtained for the
case of laminar flow cannot be used for the problem that we are discussing because the
Nusselt number relations are valid for a relatively smooth surface.
So, if you have extreme roughness present in it, the convective heat transfer coefficient is
now given by this empirical formula. So, this is something new which has been
experimental evaluated for surfaces with very large value of roughness for air flow at 50
meter per second which is taking place over the plate. So, the air is coming at 50 meter
per second and what is required is: what is the surface shear stress? The surface shear
stress let us denote it as  s at x equals to 1 meter. The only information that you have is
the local value of Nusselt number for the convective heat transfer can be correlated to
1
Reynolds and Prandtl number by this relation: Nu x  0.04 Re0.9 Pr 3 .
Nothing has been mentioned about what would be the expression for friction factor for
the hydrodynamic transport case. Only the heat transfer correlation is provided to you. It
is mentioned that the air temperature which is coming onto the plate is 300 Kelvin. The
density of air is 1.1614 kg per meter cube. The kinetic viscosity which is which is µ by ρ
is 1.589 105 meter square per second and for air the Prandtl number is 0.7. So, this is
318
the problem, what you have to find out is: what is the surface shear stress at a specific
location?
So, the first thing we realize that this problem deals with heat transfer. Heat transfer
relation is provided to you; but the question that was asked to you is about momentum
transfer. So, the Nusselt number relation is provided. However, the shear stress is the one
that you would have to provide. And it is a surface which is extremely rough. So, the
relation that you have for laminar or turbulent flow or mixed flow for C f would not be
valid because all those relations are for flat plates. The moment significant roughness is
introduced, none of the relations or correlations that we have obtained so, far would be
valid.
Now, the only way this problem can tackled is if you go for the analogy between heat
transfer and momentum transfer. But in this analogy, we cannot say that Prandtl number
is equal to 1. So, Reynolds analogy is out of question. The value of Prandtl number for
this problem has been provided to be equal to 0.7.
So, the Reynolds analogy which assumes that Prandtl number is equal to 1 cannot be
used here. We must use the modified Reynolds or the Chilton Colburn analogy. With
that and with the knowledge of the heat transfer relation, I should be able to obtain what
is the expression or the numerical value of Cf? The moment I know the value of Cf, I
s
Cf 
understand that, by definition, Cf is shear stress by dynamic pressure i.e. 1 .
V 2
2
Since the free stream velocity is known to me, the only unknown  s can be evaluated
using the modified Reynolds analogy. Let us quickly work it out and see, what is the
value of shear stress that one can expect for this kind of situation i.e. for flow over an
extremely rough surface where the heat transfer relation is provided.
So, we start with our Chilton Colburn analogy, it simply tells us that C f by 2 is Stanton
number-based times Prandtl number power 2 by 3 and Stanton is the local value of
Nusselt number divided by Reynolds number into Prandtl to the power 2 by 3. So, that is
the definition of Stanton number we can obtain the expression of Nusselt number using
the relation provided to us.
319
1
So, it would be Nu x  0.04 Re0.9 Pr 3 and I have the Reynolds times Prandtl and over
Cf
here, I have Prandtl to the power 2 by 3. So,  0.04 Re x 0.1 . The point to be noted here
2
is that from our knowledge of Nusselt number, I have obtained an expression for C f.
Since Cf deals with momentum transfer alone, it should not contain any Prandtl number.
Because Prandtl number only appears in heat transfer and not in momentum transfer.
Now what is remaining is to find out the value of Reynolds number at the specified
location; that is at, x = 1. So, at x equals to 1 meter; what is the value of Reynolds
number? Plug the value in here and find out what is C f.
xV
We know that: Re x  ; where ϑ is the kinematic viscosity. The value of V is 50 meter

per second and x is 1 meter; the ϑ has been provided as 1.589 105 meter square per
Cf
second. So, the value of Reynolds number is 3.15 106 . So, the  0.04 Re x 0.1
2
Cf 0.1
  0.04  3.15 106  and the numerical value would be 0.0179 which by
2
320
s
Cf 
definition 1 . I know the value of V infinity I know the value of ρ. So, I
V 2
2
should be able to find out the surface shear stress. Therefore,  s is 25.96 Newton per
meter square. This is a perfect example of using the analogy to obtain the heat transfer
coefficient.
We need to solve a few more problems and to see what are the interesting it is an
extreme interesting part of the course, where you can do several calculations with real-
life applications.
321
Heat Transfer
Lecture – 28
Tutorial Problem on External Flow and Behavior of Heat Transfer Coefficient
The next two classes would be tutorial ones where I am going to solve some problems on
convective heat transfer when the flow is taking over a flat plate. In the last class, we
solved one problem in which the flow was taking place over an extremely rough surface
and we have used the analogy to obtain what is the value of the heat transfer coefficient
and so on. But this class is going to be more on problems in which we would directly
identify the flow pattern, whether it is laminar or turbulent and calculate the value of the
heat transfer coefficient, the local value of the heat transfer coefficient and for this
specific problem, we would see what is going to be the heat transfer between two
streams of air at different temperature but separated by a thin metal wall of high
conductivity.
So, it is a flat sheet of metal, thin, with high thermal conductivity and flow is taking
place over the plate as well as under the plate and the two streams of air which are
following on the top at and through and at the bottom of the plate, they are at different
temperatures. So, it which we have to find out what is the value of the local heat transfer
in between the two air streams through the thin plate which is the separating plate in
between the two types of flows.
So, it is more or less straight forward problem, where we need to find out what is the
local value of heat transfer coefficient at the specified location and we know that the
inverse of the heat transfer coefficient would give us the resistance to convective heat
transfer.
So, the total amount of heat transfer, the heat flux, the heat flux between two streams
would be provided by the temperature difference between the two streams of air which
are flowing at the top and below the plate divided by the sum of the resistances, sum of
the convective resistances. Since the plate is thin, we may not have to consider the
conductive heat transfer resistance for heat flow through the plate.
322
So, let us look at the diagram which depicts the flow situation. So, this is a thin flat plate
and we have an air flow at 200º C, but at 60 meter/second which is flowing over the plate
and a temperature of 25 another air flow, at a temperature of 25 ºC and velocity of 10
meter/second which is flowing below the plate.
So, this plate separates the two streams of air flowing at different temperatures. And
since it is thin, so the key word here is thin, since it is thin we do not have to consider the
resistance of the plate, two heat transfer from the hot stream to the relatively cooler
stream. And the heat transfer, there is going to be two heat transfer resistances; one is
here which is the convective heat transfer resistance from the hot air to the plate and
again convective heat transfer resistance which is from the plate to the cooler air.
So, we need to find out what are these two resistances. And we also know that the heat
1
transfer resistance for convective heat transfer is simply , if I call it is for the stream
h1
1
one and this is going to be . So, if we are, so this is when I we are obtaining the heat
h2
transfer, the total amount the heat transfer in terms of the heat flux. Since the area is the
same. So, I can safely drop area and express everything in terms of flux.
323
So, the entire problem boils down to finding out the values of h 1 and h2 from the
available correlation for heat transfer that we have in that we have studied so far. And
some of the properties of the air which are flowing is given as the kinematic viscosity is
20.92 ×10-6 m2/s. The thermal conductivity of air is 0.03 W/m.K and the Prandtl number
for air both the cases can be assumed to be for this is 0.7.
So, this is for stream 1. For stream 2, it is at a different temperature, so, its kinematic
viscosity is going definitely going to be different. So, it is at this is going to be it is
kinematic viscosity, the thermal conductivity is slightly different W/m.K, the Prandtl
number though it will remain more or less the same. So, this is one of the reasons why
for flow for a heat transfer involving air, the Prandtl number is taken to be equal to 0.7 in
most of the cases. As you can see, a difference in temperature between 200 to 25,
substantially alter the value of the kinematic viscosity, it changes the value of the thermal
conductivity as well; however, the Prandtl number more or less remains some sort of a
constant.
So, therefore, it is common practice to take the Prandtl number for situations involving
air to be equal to 0.7. And what is the, statement of the problem is find the heat flux
between the two surfaces, the two streams of air at the midpoint of the plate and it is
given that the length of the plate is equal to 1 meter. So, from here to here is 1 meter. So,
we need to find out the heat flux between the two streams at the midpoint of the plate.
Whenever you come across such a problem, the first thing one needs to find out is what
is the type of flow that is taking place in this cases; in other words, you find out what is
the value of the local value of the local Reynolds number to evaluate whether you have
turbulent flow, laminar flow or maybe the occurrence of a mixed flow.
So, in one case you see this problem the velocity is substantially higher than the other
situation. So, we need to ensure that the flow types are evaluated apriory before choosing
the appropriate correlation for that type of flow. So, the first step is to evaluate what is
the value of the Reynolds number at the specified location for both the streams. So, that
is what we are going to do next, so what we do is will take the fluid stream 1. So, this is
we call as stream 1, in which case, you can find out that the Reynolds number based on
location, the local value of the Reynolds number is
324
60m / s  0.5m
Re x1  6 2
=1.434×106
20.92  10 m / s
And the moment it is 10-6, greater than 5 × 105, so we know that this is going to be a
turbulent flow, turbulent boundary layer which exists for the flow over the flat plate.
Now, next we are going to find out what is the situation for stream 2. We again find out
the Reynolds number which in this case is the velocity is significantly lower gives
10m / s  0.5m
Re x1  6 2
=3.147×105
15.89 10 m / s
So, it is 105, but it is less than 5. So, therefore, this one is going to be in laminar flow
condition. So, the by looking at the values of the Reynolds number, local Reynolds
number we can see that one of the stream is in turbulent flow, the other stream is laminar
flow. Once we ascertain what is the flow of regime that is that exists over these two
plates,
the next one would be again if I calculate what is stream 1, we would see that for the
case of stream 1, the value of the heat transfer efficient at a location of L/2 which is at
the midpoint of the plates would be
k
hL /2,1  0.029 Re L /2 4/5 =133 W/m2.K
L/2
325
Similarly, for stream 2 which is in laminar condition, you can choose the appropriate
expression for corresponding to laminar flow and this is
k
hL /2,2  0.332 Re L /21/2 .Pr1/3 =8.73 W/m2.K
L/2
Now, you may ask that why did I choose the relation corresponding to turbulent flow
when I know that the flow is definitely going to be laminar in the beginning and then it is
going to change to turbulent.
So, the correct expression to be used in this case when we calculate the flow Reynolds
number to be more than 5 ×10 5, the correct relation to be used should be the mixed flow
relations, absolutely right. I should have used the mixed correlation; however, if you find
out if you evaluate what is the transition length for the flow to change from laminar to
turbulent, you would probably find out that the transition length is small as compared to
L/2.
So, if the transition length for change over from laminar to turbulent that length is small
very small less than 10% in fact, correct figure is 0.95. If the length of the transition
region, where the flow changes from laminar to turbulent to the total length if these two
ratios are about 0.95, then you can use the turbulent flow relations for to evaluate what is
the local value of heat transfer coefficient.
So, that you can definitely do; moreover at this for this problem, we are trying to find out
what is the local value of heat transfer coefficient not what is the average value of heat
transfer coefficient. So, if you look at them mixed flow relation, the mixed flow relation
would give you the value of the average heat transfer coefficient. So, considering these
two cases, these two constraints I am compelled here to use the expression for the local
value of heat transfer coefficient based on the flow being entirely turbulent from the
beginning.
So, this is an approximation and as long as you identify that it is an approximation, your
understanding to the entire process would be complete. So, I am using an approximation
to find the local value of heat transfer coefficient assuming the flow to be turbulent from
the very beginning which lets me use the expression corresponding to turbulent flow to
find out what is the value of h.
326
So, what I have got now is the value of the heat transfer coefficient both at the same
location; one for above the plate and the other for below the plate. Now, when you
consider the heat transfer which is taking place, the three resistances are in series here;
one is the convective heat transfer resistance for flow above the plate which is given by
1 1
; the second one what is below the plate convective resistance what is and the
h1 h2
third is the resistance corresponding to the solid plate itself which is the conduction
L
resistance in which is given by , since I am expressing everything in terms of flux. So,
k
L
I am using it is .
k
So, the three resistances in series are one at what the total resistance is going to be
1 L 1
  . Now, it has been mentioned that this is a thin plate. So, the value of L is
h1 k h2
1 1
very small, if the value of L is small, then this resistance can be equated as  ; that
h1 h2
means, I am neglecting the conductive heat transfer resistance since the solid plate which
presumably would have a high value of thermal conductivity to be small in comparison
to in comparison to h1 and h2.
So, therefore, this is the formula for conductive, the total heat transfer resistance and the
heat flux in between the two streams separated by the solid plate would simply be
T1  T 2 200  25
q"  
1 1 1 1 =1434 W/m2
 
h1 h2 133 8.73
So, this problem provides you with an idea of how to calculate the heat transfer
coefficient based on the regime in which the flow is, what is the flow regime, choose the
right value of the heat relation between the Nusselt number, Reynolds number and
Prandtl number and from that calculate the value of the local heat transfer coefficient.
327
Next, I am going to discuss something slightly different is how would the heat transfer
co-efficient behave when we have a flow which starts as laminar and slowly it changes
from laminar to turbulent. So, how would the heat transfer coefficient would look like
and what is going to be the heat transfer, the total amount of heat transfer for flow over a
plate. And for that, I invoke an interesting experiment which probably some of you are
going to do later on in natural convection.
So, what is a natural convection? A natural convection is the case in which you have, let
us say, a rod which is suspended in still air, stagnant air, the rod is heated electrically
such that it is going to lose some amount of heat is going to be generated if this is the
rod, some amount of heat is going to be generated in this rod and the rod at steady state
whatever be the heat generated in this rod by electrical means has to be dissipated to the
outside air. And since the air is not moving, the only way it can dissipate the heat is by
natural convection.
So, what is going to happen is that the air which is surrounding the rod, its temperature is
going to rise and as the temperature rises, its density will decrease as compared to the air
which is as compared to the bulk air. So, there is going to be difference in density in
between the two and the light air would start to rise up to be replaced by cold air from
the surrounding. So, a circulatory flow like this would start for situations involving
natural convection. Now initially, at the tip where it is in contact with the cold air from
where the flow starts it is definitely going to be in laminar flow, but as you move
upwards the situation would come such that the laminar flow will change to turbulent
flow.
So, for part of the rod where the flow is parallel to the rod, this is going to be in laminar
flow and some portion is going to be in turbulent flow. The question is what is going to
happen to the value of the heat transfer coefficient? So, if you can plot the value of the
heat transfer coefficient as the function of axial length in the direction of flow, you
would see some interesting trend. So, in the remaining few minutes, I am going to talk
about that trend and try to explain it that what you would expect for the heat transfer
coefficient for a situation in which either you have flow along a vertical plate whose
temperature vertical cylinder whose temperature is more than that of the air such that a
flow current will set in. So, this circulatory flow or the case of a flat plate over which
you have flow of air and let us say this plate is electrically heated, so that you can
328
calculate what is the amount of heat transfer at every place and calculate experimentally
what is the heat transfer coefficient.
So, how would heat transfer coefficient vary with position that would strongly depend on
what kind of flow regime you have for such a situation? So, let us look at how it would
look like.
So, the example that I have discussed is a solid rod whose temperature, let us it is being
maintained at a constant temperature of T s. So, this is heat transfer coefficient the
behavior of heat transfer coefficient for flow when it changes from laminar to turbulent
and so on.
So, what is going to happen is air in here is going to get heated and it will start to rise
towards the top and cold air is going to come like this and replacing the air which is
moving up along the plate. The same situation you would see if you have just a flat plate
and you have flow of air over it and you are trying to find out how does h vary as a
function of x. So, how does h vary as a function of x where this direction is x and in this
case this is a x direction.
So, when you look at the experimental results where h can simply be calculated as
q
h
AT
329
So, that is a definition of q. Now if you have thermocouples which are embedded on the
plate, so these thermocouples would give you the temperatures as a function of x. So,
these thermocouples would give you what is the local temperature along this. So,
q  q 
h  
AT  A  Ts ( x )  T  
where T∞ is the temperature, the constant temperature of the fluid surrounding it. So, this
is T∞ and this Ts is a function of x which is experimentally measured by a number of
thermocouples which are embedded on it, on the horizontal cylinder or on this plate
itself. So, one you one would be obtained to by putting in the experimental value of T s
and by knowing the heat that it is dissipating, you should be able to obtain h as a function
of x. So, h as a function of x can be obtained by measuring the value of T s as a function
of x through the use of all these thermocouples. When you do that, you see an interesting
trend. Let us say this is the location at x equals 0 and my x keeps on increasing.
So, this is x equals 0 and this is x equals L. So, this is the direction in which x is
changing. And I am plotting the value of h on this axis ok, what you would see the
behavior of the heat transfer coefficient would look something like this. So, what does it
say that at this point over here the value of heat transfer coefficient is a maximum, but as
you move in this direction, the heat transfer coefficient decreases rapidly, the convective
heat transfer coefficient decreases rapidly. It reaches the plateau and then it starts to
increase slowly it starts to increase and which is more or less some sort of an asymptotic
value in over here.
So, what is happening in here? You can clearly mark three regions in this case, the first
region where the flow is essentially laminar through in which the heat transfer coefficient
monotonically decreases with length and till it reaches a value like this, value over here.
In from this point onwards the rate of decrease of h is somewhat arrested and you get a
plateau where it is more or less remaining some sort of a constant. So, do where the
transition from laminar to turbulent is going to take place and once the flow hits
turbulent, the heat transfer coefficient starts to increase because of the formation of
eddies which are going to be responsible for most of the heat transfer in turbulent flow
situations.
330
So, this is a very useful curve which we are going to use in our subsequent problem
solving and for our understanding as well. So, what I have then is, initially the flow is
going to be laminar, the as the flow takes place the thickness of the laminar boundary
layer will keep on increasing. And the thickness of the boundary layer is essentially
directly proportional to the resistance of heat transfer in such a case. So, as the thickness
of the boundary layer increases, the resistance to heat transfer increases and since a
resistance to heat transfer is inversely proportional to the heat transfer coefficient, the
value of heat transfer coefficient starts to decrease.
So, the heat transfer coefficient will start to decrease as the thickness of the film under
laminar flow will keep on increasing in the actual direction. Now, we say that there is a
sharp transition from laminar to turbulent, but that does not happen. The change over
from laminar to turbulent would take place over a significant length in which first small
eddies are going to form and the heat transfer is going to be due to convection, due to the
laminar flow as well as the eddy imposed additional convection which is a hallmark of
turbulent flow.
As the length increases, the Reynolds number keeps on increasing and more and more
eddies are found. So, slowly the reduction in heat transfer coefficient in laminar flow
will be more than offset by the formation of more and more eddies. So, the reduction in
the value of h is arrested and it will start to show the reverse strain that is with increase in
distance with increase in axial location, the value of h is stabilizing and then it will start
increasing.
So, that region where the rate of decrease of h is arrested and reversed is known as the
transition region. Once the eddies take over as the principal mode of heat transfer,
convective heat transfer, any increase in length will increase the number of eddy, the size
of the eddies, there will be more turbulence, more convection and the value of h will start
to increase. But this increase cannot go indefinitely as the thickness of the turbulent
boundary layer is also increasing very rapidly.
So, again after a period of sharp increase in the value of h when turbulence sets in a
plateau will again be reached where the increase in the formation of eddies will be offset
by the increase in the boundary layer thickness which would give us a profile like the
one I have drawn over here. The value of the heat transfer coefficient therefore, starts at
331
the high value since the thickness is the least here, decreases as thickness increases,
eddies start to form, eddies take over they increase the value of the heat transfer
coefficient, the turbulent sets in and then it increases and it reaches somewhat of a
constant value.
So, this picture is important which shows the behavior of the heat transfer coefficient for
flow when you have a laminar to turbulent transition. And this is an experiment which
many of you are going to do in the heat transfer laboratory where you have to calculate
the value of the heat transfer coefficient by noting the value of the axial location
dependence of Ts and generate a curve like this, verify that this is the behavior of the heat
transfer coefficient. So, we will solve few more problems on convective heat transfer for
external flow when the flow is taking place over a plate under different situations and
that should clarify any doubts that you may have.
332
Heat Transfer
Lecture – 29
Tutorial Problem in External Flow and Convection
We are going to solve another very interesting problem on convective heat transfer
where air flow is taking place over a flat plate. Now, this flat plate is slightly interesting,
because this flat plate consists of several flat plates one after the other. So, they are of the
same size, and a slice of a flat plate is put next to the slice of another flat plate so on. Let
us consider that all my fingers essentially form the flat plate with each section thermally
isolated from the next, but together they form a flat plate of uniform width.
I have drawn the figure over here, where you can see that this is one strip of 50
millimeter in length next strip is also of the same length, same and so on; and the width
of the plate is 1 meter. Each small part is insulated from the next small part. Each of
these small parts will have a heater connected to it which is going to provide some heat
which is equal to q1 for strip one, q2 for strip two and so on. The purpose of these strip
heaters is to maintain the temperature of this entire plate at a constant value of 230
degree Centigrade. So, each heater would probably have to provide different amounts of
heat to these sections of the plate. Since, the value of heat transfer from over here is
333
going to be different from the value of heat transfer over here and so on. So, each of
these heaters will then have to provide different amounts of heat to these small sections
to maintain the temperature of all such sections at a constant value of 230 degree
Centigrade and air at a temperature of 25 degree Centigrade and with a velocity of 60
meter per second flows over this.
So, there is going to be convective heat transfer and since the temperature of the plate is
more than the temperature of the air each of them will lose some amount of heat to the
flowing air stream. But the amount of heat lost to the flowing air stream would depend
on the value of the local convective heat transfer resistance. So, the problem then is
identifying the heat transfer resistance for each of these strips, which will be different
from each other. The amount of heat to be provided by each of these heaters would have
to be regulated in such a way that all these strips will be maintained at a constant
temperature of 230 degree Centigrade. The question that is asked is at what heater is the
electrical input a maximum? It is to be noted that these strips are not limited to two to six
strips; there maybe hundred such strips. There are hundreds of heaters connected to
them. So, we need to find out which heater among all these would have to deliver the
maximum amount of heat to maintain the temperature of the entire flat-plate assembly at
a constant of 230 degree Centigrade.
Firstly, we need to understand how the heat delivered by the heater related to the heat
transfer coefficient. We understand that higher the value of heat transfer coefficient,
higher would be the value of value of heat which would be dissipated from the hot plate
to the flowing air and higher is the heat dissipation from the plate to the air, more and
more heat would have to be provided by the heater which is connected to that specific
strip of 50-millimeter length.
So, the question then boils down to how we calculate the variation of heat transfer
coefficient with length on each of these heaters and find out where the value of h is
maximum; because, at that specific value of the heat transfer coefficient, more heat
would be dissipated, and more heat would have to be supplied by the electrical heater.
But as I have mentioned before there are hundreds of strips like that. Are you going to
calculate each one of them separately to find out where your heat transfer coefficient is
the most or is there any shortcut to find out logically based on your heat transfer studies
so far?
334
You would be able to say that, no, I do not have to calculate the heat to be provided to
each of these heaters. I only must take a look at three or four possibilities if I take into
account the heat to be provided to those three situations and find out which is the
maximum, then definitely I have the overall maximum of heat to be provided by the
heater at that specific location.
So, how do I find out the possibilities that one must examine to arrive at the solution for
this problem that is at which heater the heat input has to be the maximum? So, in order to
do that, I will again go back and show you the variation of heat transfer coefficient with
actual distance that I have plotted in the last class.
The value of heat transfer coefficient is very high in the beginning it reduces, it goes to
transition, then turbulent and keeps on increasing and then at some point of time it will
start to decrease again, as the value of the thickness of the turbulent boundary layer will
keep on changing.
Let us consider the plates 1 and 2, when I look at the amount of heat transfer between 1
and 2, I can clearly see that the heat transfer coefficient over the first region is going to
be more than that of the second region and more than that of the third region and so on.
So, I can safely say that q1 is going to be greater than q 2, since h1 is greater than h2. Since
we are talking about the average value of heat transfer coefficient over the entire region I
335
should write h1 which is the length averaged value of heat transfer coefficient over
region 1 and h2 is the length average heat transfer coefficient over plate 2 and since h1 is
greater than h2 more heat will have to be supplied by the heater to strip 1 as compared to
strip 2. Therefore, based on our understanding so far, we understand that q 1 is going to be
greater than q2 and q2 is going to be greater than q3 and so on, but up to which plate
should it extend? If you look at the behavior of heat transfer coefficient with length, we
observe that as long as the flow is laminar the value of h will continually decrease. So,
for the entire range of laminar flow, you only must find out what is the value of q 1, since
you know that q2, q3, etcetera are all going to be smaller than q1. So, of all these that I
have written so, far I need to calculate: what is the heat that is to be supplied to the first
heater. So, this first heater is going to be one of the contenders of my calculation.
On the side, I can calculate: what is the value of Reynolds number based on the first
plate and just to ensure that I have laminar flow in this case. So, Reynolds number for the
U1 L1
case of plate-1 would be and the numerical value turns out to be 1.14 105 . So, the

entire first strip is under laminar flow. This way it will continue if I find out what is Re 2,
what is Re 3, Re 4 and so on till the point when I hit 5 105 ; because I understand that at
a Reynolds number of 5 105 the flow becomes turbulent. So, whatever we were
discussing so far would be valid till the flow remains laminar. Now, let us see what is
going to happen when I have the changeover from laminar to turbulent.
336
We know that, the value of xc, the length at which the transition takes place can be
Re x , c 
simply written as the xc  . Substituting the relevant values in the equation, the
U
transition length turns out to be 0.22 meter. So, if it is 0.22 meter, and each strip is 50
millimeter in length. Therefore, the transition takes place on the 5th heater. So, fifth
heater is the one over where the flow starts to become turbulent.
Since the changeover from laminar to turbulent takes place on the fifth heater, the value
of the heat transfer coefficient on this plate is going to significantly increase from the
value of the heat transfer coefficient which existed prior to the point where it changes
from laminar to turbulent. In other words, value of the heat transfer coefficient from this
point onwards is going to be very high as compared to the heat transfer coefficient that
existed before this.
So, if I need to calculate the value of q1 to ascertain which is going to give me the highest
value of heat be applied, I should also calculate q5 because, q1 is where the flow is
entirely laminar, the film thickness is the least and therefore, there is a possibility that
this q1 is not only the maximum amount of heat that is taking place in the laminar region,
but it may also be a global maximum when we consider all the heaters subsequently.
337
So, of course, between 1 to 4, q1 is going to be the maximum, but I think that there is a
possibility that q1 is going to be the highest among all. So, I need to calculate the value of
q1. Similarly, q5 is the one in which the flow is going to be partially laminar and then it is
going to become turbulent and we know that heat transfer in turbulent flow is higher.
Because, q5 is the one in which the first turbulent flow appears. So, the chances of q 5
being the global maximum cannot be discounted so, I must find out what is the value of
q5 as well. So, in between the first and the fifth heater I need to calculate the value on the
first heater and the value on the fifth heater.
Now, let us see what is going to happen on the sixth heater. Sixth heater is the first heater
where the flow is entirely turbulent, fifth is partly laminar then turbulent, sixth is entirely
turbulent; and we understand that turbulence increases the value of the heat transfer
coefficient then q6 is also a potential contender for being the maximum in terms of the
heat to be supplied to the heater.
What happens to seventh heater and on? As we move into seventh heater, eighth heater,
ninth, and tenth the thickness of the thermal boundary layer thickness will keep on
increasing. And, as it starts to increase the value of the heat transfer coefficient will be
decreasing. So, the highest value of heat transfer coefficient in turbulent flow will remain
on fifth partially and on sixth. Any point beyond 6, the value of the heat transfer
coefficient will be less than that of 6.
Therefore, of all the conditions that we can think of, for flow over a flat plate, where the
flow changes from laminar to turbulent there are only three heaters, where we have to
calculate the amount of heat to be supplied by each heater in order to ascertain which is
going to be the highest. The first heater where the flow is entirely laminar, the fifth
heater where the transition from laminar to turbulent takes place and the next which is
the 6th heater where the flow is entirely turbulent.
338
So, let us start with the heater 1, I need to find out what is q convection for one which
would obviously be qconv1  hL1W  Ts  T  . So, the average value of Nusselt number for 1
1
from our relation for laminar flow is given as: Nu L  0.664 Re0.5 Pr 3 . Substituting the
relevant values, the average value of the heat transfer coefficient turns out to be: 134
Watt per meter square Kelvin. Therefore, qconv1 would be 1370 Watt.
Now, I must find out what happens on heater 5. So, q 5 is simply the amount of heat
which is lost from the fifth heater. Therefore, qconv 5 should be the average value of heat
transfer coefficient between 1 to 5 between times the total area which is 5 times W. So, it
is 5 times the length times  Ts  T  . I need to find out exclusively what is the heat to be
supplied to the fifth plate. Now, in order to do that I need to find out what is the heat
transfer coefficient over here; that means, the value of the average heat transfer
coefficient from plate 1 to this point, but we do not have any such correlation. Whatever
relation we have would give me the value of the heat transfer coefficient from the very
beginning. So, I can find out what is the average value of heat transfer coefficient from
this point to the beginning of the fifth heater and from this point to the end of the fifth
heater. In order to circumvent that problem I am going to find out the total amount of
heat lost from this point to the end which is the average heat transfer coefficient
considering the length from the very beginning up to the fifth heater;, so, it is L 5 times
339
W times delta T and subtracting from this total heat which is lost by strip 1 to 4. So, from
the total heat between 0 to 5 and I subtract that from the total heat between 0 to 4 this,
result would give me the heat which is lost from the fifth strip only.
So, that is why I have written it in this way and I understand that this is going to be 5
times the length of each strip this is going to be the 4 times that of each strip and this
allows me to find out what is q convection on 5. Now, we also understand that 1 to 4 is
under laminar condition, but when I go to the 5 it is under mixed condition. So, this 1 to
4 is to be evaluated using laminar correlation and 1 to 5 is to be is to be evaluated using
the mixed flow relation.
So, what is the average Nusselt number for 4 since it is laminar, I am simply going to
find out the value of the Reynolds number. So, Reynolds number based on plate 4 is
4.56 105 . From this I calculate the average value of heat transfer coefficient considering
the length up to the fourth heater to be equal to 67 Watt per meter square Kelvin. In
contrast the length average value of Nusselt number considering the fifth heater, since it
is in mixed flow I am going to use the mix flow relation which we have obtained before.
So, this is Reynolds number based on the 5 heaters to be 5 strips together minus 871
Prandtl to the power 1 by 3 which when you put the values would come to be 546 and
this would give you h bar 1 to 5 to be equals 74 Watt per meter square Kelvin. So, just
look at this once again the when I am finding out the average heat transfer coefficient up
340
to the fourth heater, everything is under laminar flow. So, I use the laminar relation
except the Reynolds number here is going to be Reynolds number up to 4 which is 4
times Reynolds number up to one since these strips are of the identical length.
So, I can calculate the Reynolds number and then I can calculate: what is the heat
transfer coefficient, when I consider the entire region from 1 to 4. When I consider the
entire region from one to 5 it is in the mixed region and so, I use the mixed flow relation
to find out what is the value of this. So, the q conversion for fifth heater would simply be
equal qconv 5   hL1 to 5 L5  hL1 to 4 L4  W  Ts  T  . Substituting the relevant value would
result in qconv 5 = 1050 Watt
For the six one, qconv 6 can be calculated as: qconv 6   hL1 to 6 L6  hL1 to 5 L5  W  Ts  T  using
the same logic as before. Now, in order to find h bar 1 to 6 and h bar 1 to 5 both are in
mixed flow condition, since the transition has taken place over here, you can use the
appropriate correlations and the value of h bar 1 would turn out to be 85 Watt per meter
square per Kelvin. Substituting the relevant values in the above equation, qconv 6 = 1440.
341
So, what you would then see as the final result is: qconv 6 = 1440; qconv 5 = 1050 Watt; and
qconv1 = 1370 Watt. So, when you compare amongst these three you see that
qconv 6  qconv1  qconv 5 q; therefore, the sixth plate has the largest power requirement.
So, I guess you understood the problem that we have just now solved. We have discussed
in-depth; how would the heat transfer coefficient vary based on the flow regime. So, this
is a direct example of how the relations and the concept that we have developed in the
first few classes for external flow heat transfer can be utilized for a practical problem; in
finding out the amount of heat to be provided in order to maintain a surface at a constant
temperature, which is a requirement in many practical applications. We will solve few
more problems in the coming classes, before we move on from external flow to internal
flow.
342
Heat Transfer
Lecture - 30
Tutorial Problem in External Flow and Convection
Before we conclude the External Convection I would like to solve two more problems, to
show you some more concepts to clarify some of the concepts that we have already
introduced.
And once we are done with the external flow, then we will start to concentrate on
internal flow. For example, the flow which takes place when there is a heat transfer in a
flowing fluid through a tube or a pipe, which is a very common occurrence. There are
many such applications, where the fluid flows through the pipe and another fluid flows
through outside of the pipe; in order to initiate heat exchange between a cold fluid and a
hot fluid.
So, since we have already covered external flow. That is how to treat flow outside of a
surface and then if we can find out the relations or correlations which are applicable for
flow through a conduit, then that would complete, that would enable for the design of
heat exchange equipments that is going to be; that is an indirect heat transfer between a
cold fluid and a hot fluid with something in between which is the first step for the design
of heat exchangers. So, as I said before we conclude our external flow part, I am going to
solve two more problems and explain some of the intricacies of external flow heat
transfer.
So, the problem that I have chosen for this class is about a steel plate which is coming
out of the mill at high temperature. So, the steel plate is being pulled while it is been very
hot at a certain speed. So, the steel plate comes out of the hot rolling mill at a very high
temperature and it is being drawn at certain speeds.
So, the length of the steel plate, the thickness of the steel plate, the temperature, etcetera
all are provided. Now as it is being pulled; obviously, there is going to be a heat transfer
from the very hot steel plate, to the ambient air which is outside. And considering only
convection, that is neglecting the effect of radiation which ideally should have to be
343
considered since this steel plate is coming at a very high temperature, but since we have
not studied radiation yet in this course.
So, we will for the time being neglect radiation and we will only consider the interaction
of the steel plate with the ambient by the convection process. So, as the steel plate is
being pulled it is going to lose heat because of convection. Now whenever it loses heat
by convection, the first thing that we need to evaluate is; what is going to be the value of
the convective heat transfer coefficient of edge. And from our study we know that this
edge is a function of the axial position where exactly you are measuring this edge? What
is most specifically, what is the value of Reynolds number based on length at that point
that would dictate the value of the heat transfer coefficient.
So, if I can find the heat transfer coefficient at a specific location on the moving plate.
Then by applying or by creating a small shell and making the heat balance we should be
able to correlate the heat which is lost and as a result what is going to be the rate of
change of temperature at that point. Because as the plate is moving away from the mill
its temperature will obviously, start to reduce.
So, the rate of reduction of temperature of this steel plate would definitely be a function
of what is the value of the local value of the heat transfer coefficient. So, the problems
the specific problem that we are going to solve in this class, considers the,
344
if you look at the following figure over here, you have a hot roll section from which the
steel plate emerges and it moves with a velocity equal to 20 m/s and initially it is at a
temperature of 1200 Kelvin.
So, the temperature is over here is 1200 Kelvin. The length of the plate; that means, this
entire length is 100 meters and the thickness of the plate is 0.003 meter. So, it is only a 3
millimeter thick steel strip that moves with a velocity equal to 20 m/s.
And the property of the steel is given as ρ and C p, the values of which are given and this
is exposed to air, the ambient is air. And for air, the kinematic viscosity is given as 76.4
× 10-6 m2/s. And the value of k the thermal conductivity is 0.0549 W/m.K. And Prandtl
number for air will generally be very close to 0.7. So, for this specific problem it is given
as 0.702.
So, the first assumption that we are making is that radiation and conduction effects, these
two effects are neglected. And secondly, we would assume that transition from Reynolds
number for the flow to become from laminar flow to turbulent flow is equal to 5 ×10 5
which we have assumed for most of the cases.
So, the questions the things that you have to evaluate are the following; you have to
determine the initial rate of change of strip temperature, strip temperature at 1 meter
from the leading edge, 1 meter from the leading edge; that means, over here which is 1
meter and at the trailing edge. The trailing edge being at the end of the entire length so,
which is at equal to 100 meters.
So, find out what is the temperature, rate of temperature at this point and at this point?
So, this is at a distance of 1 meter and of course, this is at a distance of 100 meters. In the
second part, which is also equally important; determine the location at which the cooling
rate is minimum. At which minimum cooling rate is achieved?
So, these are the two problems, these are the two questions that we need to answer, one
is at this point which is 1 meter from the leading edge, what is going to be the rate of
change of temperature at that point. So, we would like to when I said rate it is the time
rate.
345
dT
So, I would like to find out what is at x equals 1 meter, which is 1 meter from the
dx
leading edge and at the trailing edge which is at x equals 100 meter. So, these two, the
dT
two values of at these two locations are to be evaluated. And the second part of the
dx
problem is find out the location at which minimum cooling rate is achieved.
So, in order to answer this we would refer to the behavior of the heat transfer coefficient
with distance as we have analyzed before. And the second problem that I am going to do
in today’s class would also require the knowledge of how heat transfer coefficient varies
with distance and what kind of changes would you expect when the flow changes from
laminar to turbulent.
So, the initial sharp decrease of the value of edge coming to a plateau and then changing
to turbulence, turbulent mode of heat transfer beyond a certain point. That is something
which you would like to discuss, but let us concentrate on the first part of the problem.
So, the first part of the problem in order to find out the initial rate of change of
dT
temperature; that means, we need to write a heat balance; the way we are going to
dx
write the heat balance is the following.
346
So, will take some portion of the strip enlarged view of that and let us say I am taking a
thin slice of it. The area available for; the area available for the top area so, if you think
of this one. Then this area, where heat transfer is taking place we are going to assume
that this area is As that is the surface area since the thickness of the strip, so, this is the
strip since the thickness of the strip is very small. We need not take the heat loss from the
sides as compared to the width of the plate. So, we are neglecting the heat losses from
the side. So, in here there is no heat input only heat is going out of it. So, if I write my
basic conservation equation it would be time rate of energy coming into the system,
minus time rate of coming going out of the system plus or minus generation or depletion
is going to be equal to the rate of energy storage.
E&IN  E&OUT  GEN / DEP  RATE OF ENERGY STORAGE
So, this is a fundamental energy conservation equation in terms of rate. So, rate of
energy which is coming in minus rate of energy that is going out plus or minus
generation or depletion at is going to result in a change in temperature of the object from
which these processes are taking place. The result of which will manifest itself as a
change in energy stored in the material and then we would simply express the energy
storage in terms of temperature. So, I would, I am going to get a time rate of change of
temperature of the material from which energy is coming in going out generated etcetera.
So, for the problem at hand there is no energy in term so E&IN =0. The only energy and of
course, there is no energy generation or depletion in the system. So the only, this is going
dE
to be my new equation which is  E&OUT  ST .
dt
Now, what is energy out? The energy out is only by convection. So, it is taking place
from the top as well as from the below. So, the energy out in Newton’s law note that I
have written hx, a suffix x denotes that we will have the variation of h with axial location.
So, this is my x, so this takes into account for heat losses from both top and below and
this is simply Newton’s law of cooling. This must be equal to the energy stored, the
& p T .
energy stored is simply mC
dE d
 E&OUT  ST  (mC
& p T )
dt dt
347
Where delta T is T (of the strip) - some T ref; where this Tref is a difference temperature
dT
& p
in order to explicitly express in terms of energy. So, it should simply be mC . See
dt
& p can also be
this is a constant and we will call this as the rate of cooling. Now this mC
expressed as  As C p where, δ is the thickness of the strip, A s is the surface area. So, my
governing equation would simply be
dT   dT 
& p 
2hx As (T  T )  mC    As C p  
 dt i  dt i
So, the time rate of change of the strip temperature
 dT  2hx (T  T )
   = -0.119 hx K/s
 dt i  C p
So, when you put the values of T ∞. So, T is equal to 1200, T∞ is at 300 K. That is the
temperature of the ambient air, ρ is provided as 7900, δ is 0.003 and C p is given as 640
all in appropriate SI units. This should simply be equals 0.119 hx K/s.
So, this is then from your fundamental modelling of the process. This is the expression
and the only unknown that you have to evaluate then is what is the value of h x. And since
it depends on the location axial location of the point at which you are measuring the
temperature variation with time. The value of h x at that location is to be evaluated. So,
the problem then boils down to finding the value of h x. We just now have to choose the
right expression for evaluating hx and we choose the right expression for h x based on
whether it is laminar flow or in the turbulent flow.
Therefore the next step of the problem would be to evaluate the value of the Reynolds
number at x = 1 meter which is case 1 and at x = 100 meters which is case 2 and see
whether they are in laminar zone or in turbulent zone. So, the next step of the problem is
to evaluate the value of Reynolds number which we are going to do next. So,
348
Vx
At x = 1 meter, the axial value of Reynolds number which is , where ν is the

kinematic viscosity of air. So, this is
Vx 20m / s  1m
Re x   =2.62×105
 76.4 106 m 2 / s
Of course, as you can see from the value of Reynolds number this is less than the
Reynolds number at transition, which is taken to be as 5 × 105.
So, therefore, at x = 1 meter, the flow inside the boundary layer is in the laminar board.
So, therefore, you need to choose the correct expression, the correct expression for
Nusselt number or in that way the value of h x based on the laminar correlation of Nusselt
number.
So, the laminar correlation of Nusselt number in this case would simply be
k 0.0549 1/2
hx  0.332 Re x1/2 .Pr1/3   0.332  2.62 105  (0.702)1/2
x 1
where k is the thermal conductivity and x is the axial location. So, the Reynolds number
you have already calculated to be 2.62×105. You know the value of Prandtl number to be
equals 0.702 and you know the value of k. So, hx = 8.29 W/m2K.
349
So, therefore, at x = 1 meter and we are simply going to use the relation that we have
obtained before that
 dT 
   0.987 K / s
 dt i
So, this expression we simply substitute the value of hx and obtain that the time rate of
variation of temperature at x = 1 meter, would simply be equal to 0.987 K/s. So, that is
for part one.
Now let us concentrate on part 2 where, it says at the trailing edge. If you calculate the
value of that is at x = 100 meters, if you calculate the value of Reynolds number; you
would see that the Reynolds number would turn out to be 2.67 ×107. I am not doing this
it is you are going to do it on your own which is definitely greater than the transition
value of Reynolds numbers that is 5 × 105.
So, therefore, flow is definitely turbulent at x = 100 meters. Next step is simple you
simply have to choose the expression for Nusselt number which is valid for turbulent
flow Reynolds to the power 4/5 and Prandtl to the power 1/3. And since, it is Nusselt
number I am going to in order to obtain hx, I am simply going to write k/ x.
k
hx  0.029 Re x 4/5 .Pr1/3
x
350
And then put this as before put the values, put the same values over here one more time.
k 0.0549 4/5
hx  0.029 Re x 4/5 .Pr1/3   0.029  2.67  107  (0.707)1/3
x 100
hx = 12.4 W/m2K
So, if you compare this value of hx which is in turbulent flow with that which you have
obtained at 1 meter, which is in laminar flow. So, you can compare and see that in one
case the value is 8.3, the other case it is 12.4. Therefore, it is consistent with our
understanding that with increase in turbulence the value of h x is going to be more and
dT
once you do that with the relation that we have already obtained this  1.47 K / s
dt
So, again if you compare this value, with the value that you have obtained at x = 1 meter
since it is laminar the heat transfer is less and therefore, the rate of drop of the
temperature of the steel strip is going to be less as compared to at x = 100 meters which
is turbulent leading to a higher value of hx and a higher value of the temperature drop
with time.
So, this again tells us how to solve problems of this. And one of the most important steps
in this type of problems is to find out what is the value of Reynolds number? Based on
the Reynolds number you decide about the flow pattern is it laminar or is it turbulent.
So, based on that whether it is laminar or turbulent you choose the appropriate relations
for Nusselt number and then calculate the value of the heat transfer coefficient or any
other value which are related to heat transfer coefficient. For example, here the time rate
of drop of temperature at a specific location. The second part of the problem tells us to
find out what is the minimum cooling rate that is, that where do you think the minimum
cooling rate is going to take place.
So, if you see in the first part of the problem we have got a value of h x and from that hx
then convective heat transfer coefficient. We obtained an x2 and numerical value of
dT/dt at that specific location. And for 100 meters also we got another value of dT/dt.
The difference between the two only being that in one case you have laminar flow in the
other case you have turbulent flow.
351
So, the laminar heat transfer coefficient is generally smaller than the turbulent heat
transfer coefficient and if you again think what we have done before, how the heat
transfer coefficient in laminar flow would behave, how the transition is going to take
place, and how the onset of turbulence increase sharply the value of the heat transfer
coefficient? But beyond that point the heat transfer coefficient is going to slowly
decrease with increase in thickness; that means, with increase in Reynolds number.
So, roughly the value of the variation of the heat transfer coefficient can simply be
expressed in this form, where you have the heat transfer coefficient hx as a function of x
and it is going to come down sharply and then it increases and then slowly decreases.
So, this is the point where you have the onset of turbulence. On this side you have
laminar flow over here you have turbulent flow. You can clearly see over here the
minimum cooling rate would therefore; minimum cooling rate is associated with the
minimum value of hx. So, minimum cooling rate is associated with the minimum value of
hx. So, the minimum value of hx is going to be somewhere in this region where the
transition from laminar to turbulent takes place.
So, just before the transition the; that means, for a value of Reynolds number, Re xc at 5 ×
105. This is the point, where Reynolds number is 5 × 105 , the point of transition. This is
the one which is going to give rise to the minimum value of heat transfer coefficient and
therefore, it is going to give us the minimum cooling rate.
So, for Rexc equal to 105 you can calculate what is the value of xc,
Re xc  5 105

xc  5  105.
V
5 105  76.4  106

 =1.91m
20
So, at a distance of 1.91 meter from the edge from the leading edge; that means, the point
at which the steel strip is coming out of the hot mill that is the point where you are going
to get the minimum value of edge and therefore, the minimum value of the heat transfer
coefficient.
352
So if we go back to our original figure. We have calculated at 1 meter it is laminar and
we got some value of dT/d x at 100 meters it is turbulent. So, we got some other value of
dT/dx, but this is the point which is which corresponds to let us say this point
corresponds to x equals 1.91 meter.
This is the point which corresponds to the transition from laminar to turbulent and
therefore, the transition from laminar to turbulent is associated with the lowest possible
value of heat transfer coefficient. And therefore, the lowest possible value of dT /dt .
So, the summary of this problem tells us first, we need to find out what is the value of
Reynolds number, what is the flow pattern, choose your relation or correlation
accordingly, in some cases write a simple heat balance, starting with the most
fundamental conservation of energy, and identify the terms which are relevant, express
the terms in terms of, for example, in this case the heat losses were expressed by
Newton’s law of cooling and you get an expression to obtain either the time rate of
change of temperature or any various other quantities that are of interest in practical
applications.
So, in the next class will see another problem which is slightly more conceptual than this
one and that would be the end of external convection and for this course.
353
Heat Transfer
Lecture – 31
Tutorial Problem in External Flow and Convection (Contd.)
So, the last problem on external convection before we switch to internal convection. The
last problem of external convection is somewhat conceptual. It assumes that we have a
plate, whose size is L on one side and twice L on the other side.
The question is, whether the amount of heat transfer would remain independent if I have
the L face facing the wind or the twice L face facing the wind? And is there a way, by
which you can calculate that for what value of Reynolds number, it is really not going to
matter. That means whether the flow is along L or it is along 2 L. So, we all understand
that the heat transfer coefficient is a strong function of the length of the plate. In one case
it sees L as the length of the plate and in another case, its sees twice L as the
characteristic length of the plate.
So, the question once again is what the conditions are, that are to be meet such that, the
total amount of heat transfer from the plate is independent of whether it faces, whether
its flow is over L or 2 L. So, it boils down to what is the condition at which, the overall
heat transfer coefficients, since the area remains the constant; area is always L times
354
twice L. So, the temperature T infinity is also a constant; the temperature of the plate is
also a constant. So, it is the value of h which would dictate what is going to be the total
amount of heat loss, from the surface.
So, hL and h2L, the heat transfer coefficient for the effective length to be L and the heat
transfer coefficient with the effective length to be twice L, are they equal? If they are
equal, then you would see that they are going to be equal only for a very restrictive
special condition. So, that is the last problem that we are going to analyze before we
move to the next subsection.
So, think about the this is the problem statement which I have written for you over here,
that explain under what conditions the total rate of heat transfer from an isothermal flat
plate, whose size is L times twice L be independent of whether parallel flow over the
plate is directed along the length L or its going to be along twice L. So, this is something
which we must evaluate.
But to evaluate that, the first part of this evaluation would necessitate that, we again
recall how the heat transfer coefficient varies with length and then try looking at the
graph of the heat transfer coefficient to see if it is possible to identify the special
condition under which, this is going to be possible.
355
So, let us look at how the heat transfer coefficient varies with location. Since, I am
talking about the total rate of heat transfer, I must deal with the average value of heat
transfer coefficient.
I am plotting the variation of the heat transfer coefficient with the Reynolds number,
because Reynolds number is a function of x. Since, hL is a function of x, therefore, I plot
the variation of hL as a function of Reynolds number. Since heat transfer coefficient is a
function of Reynolds number, I just have to find out some value of the Reynolds number
where, heat transfer coefficient is going to be the same. Now let see, is it possible that
the flow is laminar on both. So, whether you have flow along L or flow along 2L, in both
cases they are going to be laminar.
So, the Reynolds number L would simply be half of Reynolds number twice L which is
universally true since I have L on one side and the size is 2 L on the other side. So, the
Reynolds number based on 2 L would simply be twice that of the Reynolds number
based on L, everything else remaining same.
So, of course, here would be some point, which is going to be R e twice L which would
simply be twice of Re L. Since heat transfer coefficient is a monotonically decreasing
function of Reynolds number in the laminar zone you will never get the relation that for
two values of Reynolds number, where one is the double of the other the value of the
average value of the heat transfer coefficient in both cases would turn out to be the same.
This is very important; I will explain it one more time. The heat transfer coefficient
decreases as Reynolds number is increases.
So, if I fix my Reynolds number based on L. So, Reynolds number based on twice L
would be some point after that and because of the dimensions involved in this case, R e
L is simply going to be half of R e 2L, and the heat transfer coefficient keeps on
decreasing. So, it is never possible to have the same heat transfer coefficient, when the
curve is monotonically decreasing. So, we will never have a situation in which the flow
is laminar in both the orientations still it is going to give you an equal value of hL and
h2L . So, coming back to this figure it is not possible to satisfy the equality of hL when
the flow is laminar. Let us call this as case I.
356
Now case 2 is something where, the flow is laminar on one side and it is turbulent on the
other side or mixed on the other side. So, if this is my R e L and if this is my R e twice L
there may be a possibility that, both would give rise to the same value of hL . So,
pictorially the situation may look something like this (refer to the image above), which
would probably give us the same value of hL . So, my case II is laminar on the shorter
plate, and turbulent or mixed flow on longer. So, as the figures suggests, it is possible I
am going to get a situation in which, this is half of this, but the nature of the curve would
dictate that the value of hL in both cases are the same, which was not possible when both
were laminar.
Case III then remains mixed on both sides of the plate. So, this is something which is
possible, mixed flow on both. If it is mixed flow on both then what it is like this? So, if
both are mixed for R e L as well as, at R e 2 L the nature of hL variation with Reynolds
number would allow the possibility that the value of h would be the same in both cases.
So, I think I have considered all possibilities. The first possibility is that, the flow is
laminar on both which would not allow the equality of hL . The second possibility is
laminar on one side and mixed on the other. The shape of the average h versus Reynolds
number curve would allow that. The third is when both are in the mixed flow range. In
the third case, the sudden increase and the gradual decrease of the convective heat
transfer coefficient in the turbulent region will allow the possibility of equality of the
value of heat transfer coefficients.
So, in the laminar region, the heat transfer coefficient is monotonically decreasing. When
it hits turbulence, the value of h shoots up and then with increase in the value of
Reynolds number, the boundary layer thickness increases. As a result, the resistance
keeps on increasing and slowly the value of heat transfer coefficient in turbulent flow
will start to decrease with increase in the axial distance; that means with increasing
Reynolds number.
So, the fall of h, in the laminar zone is sharp and the rise of h due of the onset of
turbulence is also sharp and thereafter, the fall of the h with Reynolds number or with x
is gradual. That is, the nature of the h versus x or h versus Reynolds number curve. So,
this specific nature of Reynolds number curve would allow the equality between this L
357
and 2 L for the case of laminar on one, mixed on the other and mixed on both. These are
the only two valid viable possibilities for which, the value of the average heat transfer
coefficient will be identical and therefore, the total amount of heat loss will be the same.
Now, we have to physically do the problem and find out what is the value of the
Reynolds number for which, this is possible. So, I will take case II and case III and find
out the value of the Reynolds number and the value of h for which, this is possible and
you do not need to consider case I. So, let us start with case II first.
The case 2 simply tells us that it is going to be laminar on one side and mixed on the
other. We are looking for hL  h2 L and we also understand that Re 2 L  2 Re L . Since this
is a laminar flow, I am simply going to use the Sherwood number relation for laminar
k
flow; and that would be hL 
L 1
0.664 Re0.5 Pr 3  and for 2L, it is going to be
k
h2 L 
2L  4
 1
0.037 Re 5  871 Pr 3 . I am sure you recognize these two; the first equation is
for the laminar and the second one is for the mixed flow regime. We identify that Re L
has to be less than 5 105 because it is in the laminar zone and Re 2 L which is twice Re L
must be greater than 5 105 .
358
So, this is the range in which my Reynolds number should reside and since this equation
cannot be solved directly, we resort to a trial and error solution, by assuming a Reynolds
number in this range you would be able to see that, Reynolds number is to be equal to
3.2  105 when hL  h2 L .
Now, we come to case 3, where you mixed flow on both. If it is mixed flow on both, then
k k
hL 
L 4 1

0.037 Re 5  871 Pr 3 ; and h2 L 
2L  4
1
0.037 Re 5  871 Pr 3 .
And we understand that Re 2 L  2 Re L . Substituting the relevant values and upon
simplification, we would get Reynolds number to be equal to 1.5 106 . So, if your value
of the Reynolds number is 1.5 106 when both are in the mixed flow range then it
becomes immaterial whether L faces the flow or twice L faces the flow.
When you have laminar on one side and mixed on the other side, the value of that
Reynolds number turned out to be 3.2  105 , then you would get equality of heat transfer
irrespective of which side faces the flow. Secondly, for a Reynolds number of 1.5 106 ,
you get the same thing; whether the L side faces the flow or 2L faces; and therefore, the
total rate of heat transfer from the object would be identical as well.
359
So, the beauty of this specific problem is that, it tests your concepts of the variation of
heat transfer coefficient with Reynolds number i.e. how would the heat transfer
coefficient change with Reynolds number in laminar flow, how it increases sharply in
turbulent flow and how it gradually decreases thereafter. So, from a sharp decrease in
Reynolds number for the case of laminar, sudden increase and then gradual decrease;
these characterize the behavior of heat transfer coefficient in laminar and in mixed flow.
So, once you have this picture of heat transfer coefficient in mind, you will never do any
mistake in selecting the ideal relation while you are calculating the heat transfer
coefficient.
And finally, the last part of the question was, is this valid for any type of fluid? That
means, does you results depend on the type of fluid that you have? If you considered the
development of the problem, nowhere we have assumed what kind of liquid is this.
Whether it is air, whether it is water flowing over the plate, glycerol flowing over the
plate; the fundamental result in terms of Reynolds number will be the same.
So, this underscores the importance of Reynolds number as a similarity parameter. A

similarity parameter is one, which allows you to compare results using different fluids. I
need to tell you that if you keep the Reynolds number the same, the nature of everything
else, including flow pattern and in this case heat transfer will fall in line. So, this result
that we have obtained is independent of the physical properties of the fluid. Any fluid
that has a Reynolds number of 3.2  105 would give you the same result. Directional
independence of whether L faces the flow or whether twice L faces the flow or in the
case of mixed flow if the Reynolds number is 1.5 106 you will get the same result
irrespective of the nature of the fluid.
360
Heat Transfer
Lecture - 32
Internal Flow Heat Transfer
Good morning. We are going to start with new chapter today. So far we have
concentrated on external flow that is the flow which is taking place outside of a flat plate
outside of an object and so on. But there exist very important heat transfer which takes
place when a fluid flows through let us say for example, a tube or a pipe. So, that there is
collectively called as internal flow. And many of the heat exchanger designs require a
thorough knowledge of what happens, for flow and simultaneous flow and heat transfer
inside a pipe or a tube.
So, these internal flows appear in the design of many heat exchange equipments let us
say for example, which we would so later on as something which is called a shell and
tube heat exchanger. We would discuss shell and tube heat exchanger in detail in the
coming classes, but a what a shell and heat a shell and tube heat exchanger looks like is
there would be a number of tubes, which are going to be connected to a plate.
So, to these tubes one fluid will flow from one side to the other side.
361
If I draw this, as I said we will do in more detail later on, but a shell and tube heat
exchanger would look something like this. So, there would be a head which is attached at
this end, and there would be something which is called a tube sheet, which would form
the region where the tubes are going to be attached.
So, this is another head at the end, and we will have tubes that are connected from one
end to the other. So, let us draw just 2 tubes to keep the figure simple. So, these are 2
tubes and through which the liquid, one liquid will flow in this direction.
So, this fluid is going to come to the head, I have drawn this head too large, but what is
going to happen is once it enters the shell, then it is going to get divided equally in the
ideal design into all these tubes and these tubes can be numerous depending on your
application, there can be hundreds of such tubes attached to it. So, this liquid will come
outside and will again be collected from over here.
So the tube side liquid let us call it as a liquid tube. This is going to be taken out over
here. And there is going to be another source of liquid, which would flow through this
and will come out from the top, and it is going to take a tortuous path, because there
would be certain barriers in here, which would force the liquid to take a tortuous path.
So, the path of the liquid which is coming into the larger body which is called the shell, it
is going to go like this, interacting with the tube and ultimately come out through this
outlet.
So, this region is called the shell, and as you can imagine these are called the tubes. So,
since you have a large number of tubes, and maybe one or two shells which I will show
you later. So, these are called shell and tube heat exchangers.
Now what happens in shell and tube heat exchangers? This is one of the most efficient
way to transfer heat from one fluid to the other. So, let us say we have a hot fluid with
which is this is shell, shell side liquid and this is a liquid which comes out of the shell.
And let us say the temperature of the shell side fluid is greater than the temperature of
the tube side fluid.
So, the shell side fluid as it flows over the tube there is going to be no direct contact
between the liquid, between the tube side liquid and the shell side liquid. So, the heat
transfer is going to be external flow over the outside of the tube surface, conduction
362
through the tube material and finally, it is going to be internal heat transfer between the
hot wall and the liquid which is flowing in. So, the liquid which is in internal flow inside
the tube is going to exchange heat with the shell side fluid, which is in external flow.
So, if you consider the tube material, the wall of the tube as the deciding surface, then
the tube sidewall confines the liquid which is flowing through the tubes. So, it is going to
be an internal flow, and on the outside of the tube material the flow is external. We have
studied external flow so far in our treatment.
So, what we are going to do is, we need to know what would be the relations or
correlations which would give me hi. hi being the convective heat transfer coefficient on
the inside. So, if you again think of the tube in contact with a fluid on the outside that is
on the shell side then there are 3 resistances.
The 3 resistances which would dictate the total flow of heat from the hot shell side to the
cold tube side, they will consist of h 0; where ho is the convective heat transfer coefficient
on the outside of the tube; hi which is the convective heat transfer coefficient based on
the inside of the tube, and the conduction resistance which is the resistance to conductive
heat transfer due to the material of the tube.
These 3 resistances are in series. So, 1/h in 1/ hi + 1/ ho + the conduction resistance would
give me what is the overall heat transfer resistance for flow when we have this shell and
tube heat exchanger. So, it is extremely important to know how to express h i in terms of
known parameters. So, what would be the value? What would be the expression for h i in
laminar flow and in turbulent flow?
So, depending on the flow rate of the fluid which is flowing through each of these tubes,
I can calculate the Reynolds number. And when I calculate the Reynolds number, and if I
have a relation or correlation, that should give me the value of h i. Then hi with ho and the
conduction resistance which would depend on the material of construction of the tube
and the thickness of the tube I should be able to find out what is the overall resistance
and from that overall resistance we would be able to obtain what is the overall heat
So, it is not hi, it is not ho, it is or it is going to be a combination of all these 3 resistances
inverse of which would be the overall heat transfer coefficient. So, this simple example
363
and the application of these heat exchangers are in various areas from the very large
process plants to small units; where you would see this type of shell and tube heat
exchangers.
So, we will talk about the design of shell and tube heat exchangers and the calculation
procedure and so on. But I think the importance of knowing internal flow is very clear
based on the discussion so far. So, we would start with internal flow. Now when we start
with internal flow, the first thing that one needs to see is when a flow approaches a tube
or a pipe the flow is uniform.
That means it is coming with a velocity which is a constant, and the moment it enters the
tube the viscous interaction of the moving fluid with the stationary solid wall would start
to reduce the velocity near the wall, while near the core this region still does not know
that a solid plate solid wall exists over here.
So, they will follow with this with a constant velocity over here. So, that means, inside a
certain distance inside the tube, the flow is going to vary with r; where r is the radius. So,
this is your r direction. And the total distance total overall the total radius is capital R; so
in this region, the velocity if this is my x direction, then my v x is a function of r. But o
between this point and this point the velocity is a constant. As you move progressively
inside the tube this region, where the velocity is constant will keep on decreasing and the
364
effect the viscous effect of the wall would be felt at greater depths of the fluid inside the
tube.
So, a boundary layer is going to form, inside the boundary layer the flow is going to be
viscous, outside the boundary low boundary layer as per our definition the flow is
inviscid. So, the thickness of the boundary layer will keep on increasing. And at some
point we call it as the hydrodynamic entrance length, at some point from the entrance
these 2 boundary layers will coincide. And from that point onwards if it is laminar flow
the velocity profile is simply going to follow the laminar velocity profile.
So, previously you had a variation and a flat region, but over here as you know from
your fluid mechanics it is going to enter in it and is going to be a parabolic distribution of
velocity. This hydrodynamic entrance length is therefore, very important. Because once
you cross this hydrodynamic boundary hydrodynamic will entrance length, from that
point it is called fully developed flow.
So, in fully developed flow the entire fluid mass is experiencing the viscous forces which
allows you to derive the functional that the dependence of the velocity with R and we all
know that it is going to be parabolic in nature. These and of course, what would be this
hydrodynamic entrance length is something very important. You would like to know
what is this entrance length. This entrance length is definitely going to be a function of
Reynolds number.
So, if you keep on changing the value of the Reynolds number, the hydrodynamic
entrance length would also vary. Because more the value of Reynolds number you are
going to have better transport more transport of momentum through the fluid towards the
core of it. So, there is definitely should be a dependence of the entrance length on the
flow Reynolds number.
And the flow Reynolds number is defined in terms of a length scale which is the
diameter of the tube at the properties which are μ and ρ; viscosity and density and some
sort of a velocity. So, what is the velocity to be used in this case? The Reynolds number
for flow in a pipe is always expressed in terms of the mean velocity, let us say we
denoted as um.
So, your Reynolds number expression should be
365
 um D
Re 

where d is the diameter of the tube and μ is the viscosity. So, we also know that the
where the expression for the entrance length for laminar and turbulent flow are different.
So, I am simply for completeness though you have studied these in your fluid mechanics
course, I will simply write those relations which were obtained for laminar and turbulent
flow, what would be the expression for the entrance length. So, for laminar flow this
entrance length, which is called xfd; the length which is required for the flow to become
fully developed.
So, that is the significance of the fd term divided by D where D is the diameter.
 x fd 
   0.05 Re D
 D  LAM
For the case of laminar is about 0.05 Reynolds number based on diameter. So, the
entrance length can be expressed as a function of Reynolds number and when this
entrance length is non dimensionalized with the diameter of the pipe through which the
flow is taking place and for turbulent for x/D greater than 10, it is assumed flow is fully
developed.
So, the 2 relations that are generally used this one is for laminar. That is
 x fd 
   0.05 Re D
 D  LAM
and for turbulent if x is 10 times the diameter, the point where you are your finding out
whether the flow is fully developed or not if it is 10 times the diameter from the entrance
it is assumed to be fully developed. And the we all know that is what is the expression
for velocity, the velocity in fully developed condition is a function of r only. It is not a
function of x.
So, it is a function of r, but it will remain the same as you move along this. So, this
would be
366
  r 2 
u (r )  2um 1    
  ro  
where um is the average velocity, ro is the radius of the pipe. So, this is the fully
developed velocity profile in here. There is one more important thing in internal flow,
and internal flow of a liquid through a pipe, you are more interested in finding out what
is the pressure drop for such cases.
So, the pressure drop for such cases can be expressed in terms of moody or Darcy
friction factor; which is denoted as
 dp 
 
dx
f   2 
 um / 2
dp
Where , is the pressure gradient, um is the average velocity. So, this is some sort of a
dx
dynamic pressure. And this is how you obtained the friction Darcy friction factor. And
when you talk about the friction coefficient or fanning friction factor, this fanning
friction factor is Cf, which is
s f
Cf  2 =
 um / 2 4
Where  s is the wall shear stress. And you would see that this C f is going to be equal to
f /4.
So, there are 2 different types of friction factor; which are which are used in fluid
mechanics literature one is f the other is C f. So, the relation between f and C f is Cf is
going to be f/4. There is no other difference. So, when you use moody diagram, what you
get is the Darcy friction factor which is f, when you talk about the friction factor from
fanning friction factor, its Cf and the relation between them if you look at the definition
carefully its simply going to be C f = f/4. There is one more thing which one I should
write in the space just make sure that everything is going to be in one page. So, this f is
going to be 64/Re for the case of fully developed laminar flow.
367
64
f=
Re
So, this is the relation that you are familiar with for laminar flow. And for the case of
turbulent flow, the mostly it is to be used by the moody diagram and for this case of
turbulent flow, this is from moody diagram; with which I am sure you are familiar with.
This f is going to be
f  0.316 Re1/4 D , Re  2 104
Or the friction factor can be expressed as
f  0.184 Re 1/5 D , Re  2  10 4
I am simply putting all of them in one page for the sake of completeness and so, what we
have shown here is that what would be the expression for f in laminar flow, and what
would be the corresponding expression for f which are approximated from the moody
diagram.
So, we can either use moody diagram directly; so, you understand the moody you realize
you probably the remember the moody diagram is nothing but the value of f, the friction
factor on the left hand side and on the x axis you have different values of Reynolds
number. And there would be a family of curve for different values of the thickness of the
inside of the pipe.
So, higher the roughness of the surface inside surface of the pipe, for the same Reynolds
number you get a high value of f. So, recollect that the moody diagram, where friction
factor, f is plotted against Reynolds number and at higher values of Reynolds number the
variation of f with Reynolds number becomes almost negligible.
So, for a high value of Reynolds number which is also known as for a highly turbulent
for a completely turbulent flow the value of the friction factor is independent of
Reynolds number. That is one of the observations. The second observation is the inside
roughness of the pipe dictates the value of the friction factor. More the roughness higher
would be the value of the friction factor.
368
So, the same moody diagram can be expressed in terms of the 2 correlations which I
have written over here as this for some value of Reynolds number, and this for other
values of the Reynolds number, for the case when the flow is for the case when the flow
is taking place inside a smooth pipe. So, the results which I have drawn the correlations
that I am proposing, they are only valid for smooth pipes and nothing else. The problem
comes is what we are going to do when we are dealing with thermal situations.
We understand it is we can mentally visualize that the flow as it takes place; it is going to
become independent of x. It is going to be only a function of r. And from that point
onwards, the velocity at any given r location is going to remain constant. If you change r
the velocity will vary, but if you keep the r fixed, the velocity is going to be the same
wherever you go axially inside the pipe.
So, the profile and the value of the velocity remain unchanged once the flow becomes
hydrodynamically fully developed. Now let us think of the same thing for the case of
heat transfer. So, I have a tube which is being heated externally so as to maintain a
constant heat flux all along the pipe length. So, as I move in this direction when the flow
comes in the velocity is uniform and the temperature is uniform as well; as it enters, both
the velocity and the temperature would start to become function of r, and as I move
further and further at some point the velocity is fully developed. So, the velocity will not
change with position at a fixed r anymore. So, let us look at this figure once again. At
any given r, the velocity at all these points are same when the flow is fully developed.
The velocity at this point will remain same, though different from this point, but it will
not be a function of x. But over here the velocity could be a function of x. What happens
in the case of thermal cases? For thermal conditions, since I have constant heat flux all
along the wall of the pipe, the temperature is going to change with x.
So, the way we have defined fully developed flow from an hydrodynamic point of view,
that the velocity at any r is independent of x cannot be said for the case of thermally fully
developed flow because the velocity will always that the temperature the velocity at any
r will remain fixed, but the temperature at that r will keep on increasing since I am
constantly getting some amount of heat from the side walls of the pipe. So, the constancy
of temperature at a given r cannot be used to define a thermally fully developed flow.
369
So, what is the definition that is generally used for thermally fully developed flow. We
say that we are not interested in the value of the temperature, we are more interested in
the shape the profile of the temperature. So, if the temperature profile or more correctly
if the dimensionless temperature profile is not going to be a function of x, when the
dimensionless temperature profile is invariant with x, when that condition is reached, we
say that we have reached thermally fully developed flow.
So, instead of a velocity, instead of a specific temperature we say that it is the shape of
the temperature profile, shape of the dimensionless temperature profile, which is relevant
in defining what is thermally fully developed flow. So, I am going to write the
expression for thermally fully developed flow, and see how that can lead to some of the
interesting results for heat transfer. So, let us start with the definition of thermally fully
developed flow.
So, thermally developed flow it is defined as when
  Ts ( x)  T ( r , x) 
 0
x  Ts ( x )  Tm ( x) 
Where, Ts is a surface temperature.
So, this becomes equal to 0, know when you talk about when you see these 3
temperatures, this is the surface temperature of the wall which could be a function of x.
370
This is the temperature of the fluid, which is flowing fluid which is the function of r; that
means, how close it is to the wall, and it is also a function of x where it x denotes the
axial position. What is Tm(x)? Tm is the mean temperature, Tm is the mean temperature of
the flowing fluid and the way Tm is defined it is total energy which is transported by the
fluid.
So, this is the total energy transported by the fluid by the fluid which; obviously, would
be equal to
E&t    uc TdA
v c
& vTm
 mc
Ac
So, this is if you look at this clearly ρ times you would give you the mass flow rate the
& vTm , where T
mass flow rate per unit area, integrating it over the entire area. So, this is mc
is the temperature. So, this is the total energy content of the flowing fluid. Now this total
energy content of the flowing fluid if we equate that with some mean temperature T m,
then the definition of the mean temperature would be
r
2 o
Tm  uTrdr
um r 2 o 0
So, the only thing that you have to do is the cross sectional for the case of cross sectional
area dAc, you simply have to write 2πrdr because that is the cross sectional area through
which the liquid is flowing. So, the cross sectional area which through which it is
flowing is πr2. So, obviously dAc would be 2πrdr, and when you put that 2πrdr in here,
and you take everything this thing out the ρ and cv which are out of the integration sign.
So, this is what you are going to get. So, this a this are this gives you the definition of T m
And this Tm is a function of x. Because in mean temperature of the flowing fluid will
keep on changing. The let us say I have flow from this side, I find out what is the energy
that is being transported by the fluid over here. And all the time it is receiving energy
from the pipe walls. I find out what is the energy, which is being transported by a liquid
at this point.
So, the energy here is definitely going to be different than energy at this point, because
some amount of energy has already been added by the walls in between these 2 cross
371
sections. So, the Tm the area average temperature, the flow average temperature that you
would get over here, and the Tm that you would get over here they must be different.
So, Tm the mean temperature the flow mean temperature is a function of x; which was
not happening for the case of velocity. The average velocity due to equation of continuity
would remain the same between location 1 and location 2. But the average temperature
between these 2 points would be different. That is what make heat transfer more
complicated, and that is why we cannot say that the velocity in variant being a condition
for fully developed flow, we cannot see the temperature in very the no variation in
temperature is the condition for fully thermally developed flow.
That is why we have to bring in that the shape the dimensionless shape of the profile will
not be a function of x. That is the necessary condition for thermally fully developed flow.
So, let us expand the, this expression and see what additional information we can obtain
out of this. So, this is the all this is the definition, and this being 0 so, the temperature
ratio therefore, is independent of x and the derivative. So, if this temperature ratio that is
the made the expression inside the third bracket since it is independent of x, then it must
also be independent of r. The temperature ratio is independent of x.
So, the derivative of the ratio with respect to r must also be independent of x. So, what I
mean to say is that, if I get if I write it in this way, there is (T s – T), I am just dropping x,
r etcetera, but it they are the same as before by (Ts – Tm). I simply dropped these things to
make it more compact this; obviously, can be expressed as
  Ts  T  T / r r  ro
 
r  Ts  Tm  Ts  Tm
This cannot be a function of x. The ratio the x derivative of this is equal to 0. So,
therefore the r derivative of the same ratio will cannot be a function of x as well.
So, if this is not a function of x, if the since the temperature ratio is independent of x. So,
if I take the r derivative of the temperature ratio, that will also not be a function of x. I
Ts  T
will repeat it once again. I have defined dimensionless temperature; which is .I
Ts  Tm
take the derivative of that with respect to x and that is equal to 0 as per our definition.
372
So, that means, the derivative the dimensionless temperature which I have defined is
independent of x. That is why its derivative with respect to x is 0. Now if I take the r
derivative of the same ratio, that r derivative of that would also not be a function of x
since the ratio itself is not a function of x so, if I take the derivative of that ratio with
respect to r a b c etcetera, it does not matter, the r derivative or any other derivative of
that will not be a function of x. That is what I have written over here.
So, let us look once again to the expression that I have written at this point, which is
simply this, and in order to obtain over here I am using Fourier’s law which is
T
qs   k y 0
y
When you transform the coordinates to r, it is,
T T
qs   k y 0 k r  ro
y r
So, if this is the pipe wall, then this is your r and this is your y.
So, at y equal to 0 transforms to r equals r0, y equals 0 is r equals r0. And since y and r are
in opposite directions so, I can I absorb the minus sign, when I change the variable from
y to r. And
qs  h(Ts  Tm )
which is the heat given to this wall that is Newton’s law of cooling.
So, this is Fourier’s law and this is Newton’s law, this is in terms of the conduction
which is taking place, this is in terms of the convection which is taking place from the
pipe wall to the inside. So, when you put this over here, what you would get the
T
important relation is h/k is not a function of x. So, when you put in here, and from
r
the Fourier’s law and (Ts - Tm) from the Newton’s law, the qs will cancel out from the
numerator and the denominator; what you are going to left with is h/k is not a function of
x.
373
So, an extension of this is h is not a function of x. This is extremely important; k is a
constant physical property. Since k is a constant, what I see from here that h is not a
function of x. So, one of the important statement of the thermally fully developed flow is
that the natural convection heat transfer coefficient h is not a function of x in the
thermally fully developed region.
So, what we have now then obtain is a way to define what is thermally fully developed
flow. And in order to do that, we have expressed it in terms of a dimensionless

temperature, and we took the x derivative of that of that ratio equal to 0, we say that
x
this is the condition for thermally fully developed flow; that means, the flow the shape of
the temperature profile will not change with x, and a direct corollary of that is that the
variation of this ratio with r can also not be a function of x.
When you do that, and when you substitute Fourier’s law and Newton’s law into that
expression; what you get is that the heat transfer coefficient, the convective heat transfer
coefficient is not going to be a function of x.
So, the important property of thermally fully developed flow is that, the convective heat
transfer coefficient remains invariant with the axial position once you establish thermally
fully developed condition. It would not be true in the entrance region, but once you are in
the thermally fully developed region h is not a function of x.
So, in the next class we will see what other interesting results one can obtain ultimately
leading to the relation or correlation, which would give me the value of the heat transfer
coefficient for internal flow based on whether it is laminar flow or turbulent flow and so
on, such that, I can make the complete calculation of overall heat transfer coefficient in
such heat exchanger devices.
374
Heat Transfer
Lecture - 33
Internal Flow Heat Transfer (contd.)
We have seen is that, for the case of internal flow it is difficult to define what is a
thermally fully developed flow. Because, let us say that the tube wall is maintained at a
constant temperature, or a constant heat flux is being applied to the outside of the tube
and the fluid is flowing through the tube. Now when a fluid flows, it is gaining energy
because of either the elevated temperature or the applied heat flux, and the temperature
will vary with axial position. So, the question then comes how do we define thermally
fully developed flow? If you remember there are two criteria, which were to be met if the
flow is to be categorized as hydrodynamically fully developed: one is the independence
of the velocity with axial position, and second, the shape of the profile will remain
unchanged after the entrance length is crossed by the flowing fluid. So, for the case of
thermally fully developed flow we revert to the second. We say that the dimensionless
temperature profile will not change with axial position. But we understand that the
temperature at any given location “r” will vary with position if I have a constant surface
temperature or if the fluid is receiving energy from the side walls due to a heat flux or
due to an elevated temperature. So, you do not talk about the individual values anymore.
What we say is dimensionless profile shape will not be a function of location.
So, that is why we use the expression del del x of a dimensionless temperature profile is
equal to 0. And since the dimensionless temperature profile is not a function of r, so, if
you take del del r of that dimensionless number it is also going to be independent of r
and therefore, we have obtained certain relations. The most important one which
characterizes the thermally fully developed flow is that, the convective heat transfer
coefficient in the thermally fully developed zone is independent of axial position.
So, if you think of the entrance length, there the heat transfer coefficient is a function of
x. But once you go beyond that thermal entry length; which of course, is different when
the flow is laminar or whether the flow is turbulent, but once you cross that thermal entry
length, the h is not going to be a function of axial position. We have also defined T m, the
375
mean temperature of the fluid by considering the total amount of energy which is
flowing in a cross section.
What we are going to do in today’s class is pick two limiting cases. In one case the heat
flux is going to remain a constant; that means, the inner surface is going to receive a
fixed quantity of heat. And in the second case, the temperature of the tube-wall will be
kept constant.
So, in the first case as we understand in order to drive the same amount of heat to the
fluid from the walls of the tube, as and the heat flux is identical, therefore, the
temperature will keep on changing. As the fluid flows in the tube, it’s mean temperature
will increase. In other words, the Ts which is the surface temperature of the tube will
keep on changing with x. The other condition is at the surface temperature of the tube is
kept constant; which is a very common occurrence when saturated steam is condensing
on the outer wall of the tube. So, when the saturated steam at a certain pressure is
condensing then it is going to be at a specific temperature. So, a condensing steam on the
outside of the surface is a perfect example of the temperature being kept constant while
the flow is taking place. So, we will investigate these two conditions, and our aim is to
obtain an expression of h, the convective heat transfer coefficient for flow inside a tube.
So, these are the 2 cases in which you would see that the value of the heat transfer
coefficient is constant and not only constant, the Nusselt number will be equal to a
numerical value based on whether you have constant heat flux or constant surface
temperature.
The first thing that we do is, we are going to start with uniform surface heat flux and I
am restricting to thermally fully developed flow which means not only q s is constant my
h is also constant as I have expressed before.
376
So, for completeness, if I plot the value of h with axial position, the qualitative nature of
the profile would be like this. So, this is h fd, the heat transfer coefficient for fully
developed flow, and the x for which this happens is x fd, that is the entrance length for
fully developed flow. So, this is how it would behave. For the case of uniform surface
heat flux, both h and qs are constant, and the Newton’s law would connect the constant
heat flux with h as qs  h  Ts  Tm  .
dTs dTm
Since both h and qs are constant, I can simply write  ; and we realize that Ts is
dx dx
the surface temperature, which for the case of uniform surface heat flux is going to
change with x; and Tm which is the the area averaged mean temperature based on energy
consideration, is obviously, not a function of r, but it will also change with x. So, as the
fluid progresses into the tube, the Tm will be different. And if it is heat flux which is
being added to the fluid, then Tm will increase with x. So, we will start with the
  Ts  x   T  r , x  
definition of fully thermally fully developed flow which is  0 .
x  Ts  x   Tm  x  
377
T
So, if you solve for , from here what you get is
x
T dTs  Ts  T  dTs  Ts  T  dTm . So, it is a simple derivative taken to find out

    
x dx  Ts  Tm  dx  Ts  Tm  dx
T dTs dTm
what is . Now if qs is constant, then  . So, if you look at these 2 terms,
x dx dx
since these 2 are identical, they can therefore be cancelled; and what you get is
T dTm
 for qsll to be a constant.
x dx
So, if you have a constant surface heat flux condition, then the axial temperature gradient
is going to be independent of the radial location. If you look carefully once again, this T
is a function of r and x, but when it is fully developed and for the case of qsll to be
constant, the gradient in the temperature is going to be independent of the radial location,
dTm
since your is not a function of r. That is a very important statement which we are
dx
going to use later on.
378
So, the second case is for constant surface temperature. Since it is a constant surface
dTs
temperature so,  0 . If you look at the expression that we have obtained by solving
dx
the from the definition of fully developed flow, i.e.
T dTs  Ts  T  dTs  Ts  T  dTm ; since dTs

      0 , the first and second terms on
x dx  Ts  Tm  dx  Ts  Tm  dx dx
the right hand side of the equation are going to be equal to 0, and the only term
T  Ts  T  dTm
remaining would result in:   . Now if you look carefully at this
x  Ts  Tm  dx
expression, what we have obtained in here is that the axial temperature gradient is now
going to a function of radial coordinate as well. Because the temperature of the fluid is a
function of radial position, as well as axial position. So, that is the difference what we
would get in terms of the fully developed flow, for constant surface heat flux condition,
and for the case where the surface temperature is kept constant.
Now we are going to probe a bit further and will go into energy considerations. We are
going to find out how from energy considerations, we would be able to obtain additional
information of the variation of temperature with position or with r. Ultimately, the goal is
to obtain an expression which is equivalent to that of Newton’s law.
Let us think of a tube and the flow is coming in at some temperature. So, when the fluid
enters the tube, there is a certain temperature difference between the fluid and the tube
walls. So, the surface is at some Ts and the liquid which is coming is at Tm. So, the
gradient is Ts  Tm . As I move along the length of the tube, then my Tm will change, my
Ts may also change if it is a surface condition. So, the gradient which is the cause which
accounts for convective heat transfer which is Ts  Tm can also be a function of x. So, if I
have an inlet and an outlet, at the inlet the temperature difference is Ts at i minus Tm at i.
At the outlet it is Ts at o that is outlet, and the temperature of the fluid is Tm at o. So, the
root cause for convective heat transfer keeps on changing between the inlet and outlet.
So, if I would like to express everything on an overall sense what should be the
expression for something which is similar to the Newton’s law that would give me an
379
average heat transfer coefficient, express the total heat transfer from the wall to the fluid
over a certain length l. I would like to know the total heat transfer q is h A delta T. So,
what this h is going to be and what delta T, I should use since my delta T is a function of
x. That is what we are trying to do next for the case of constant surface heat flux
condition, and for the case of constant surface temperature condition.
So, starting with an energy consideration, we are going to derive these 2 expressions. In
order to do that, I am going to start with an energy balance of the liquid which is coming
into the system.
So, when I talk about the energy balance between the flow, which is coming in and the
flow which is going out, the differential amount of energy dq convection which comes
into this must be equal to mC p dTm . We assume that the thermo physical properties

which are embedded in m and Cp are constant over the temperature range that we are
considering over here. So, this comes directly from conservation of energy. So, if we do
the integration over the length. So, that we would like to find out what is the total amount
of heat q convection that is being added to it. This would result in:

qconv  m C p  Tm ,o  Tm ,i  . So, your energy conservation simply gives you a very basic
relation between the heat that comes in and the heat which leaves the system. Now when
I talk about circular tube with a fluid as it moves in this direction. If I consider between 2
380
points, this gives me whatever be the total energy content over here, and the total energy
content over here. So, this minus this gives me the amount of energy which is being
added from the side walls. But the amount of energy which is being added from the side
walls can also be obtained by Newton’s law of cooling. So, dqconv  qsll Pdx , where P is
the perimeter, and dx is the elemental length. If I consider this to be a circular tube, then
dTm
P is simply going to be equal to pi times D. So, therefore, is simply going to be
dx
qsll P
equal to  . If we represent the qsll in terms of Newton’s law of cooling, we would
m Cp
dTm qsll P P
get: dx      h  Ts  Tm   . Therefore, the axial variation of T can be
m
m Cp m Cp
obtained from this relation. We also should note that generally P, m and C p are

constants. And if I assume that the flow is thermally fully developed, then my h is also a
constant.
381
Now, I will be writing the above equation for the case of constant surface heat flux. I will
dTm P
write this expression once again over here, that is dx    h  Ts  Tm   . This is the
m Cp
expression which we have obtained, and we would like to use it for constant surface heat
flux. So, that means, your qsll is a constant. So, my total q convection would simply be
dqconv  qsll Pdx , since qsll and P are not a function of x anymore. If I integrate on both
sides from x = 0 to x = L, I am going to get qconv  qsll PL . Therefore, from here, I can
dTm qsll P
now write dx   and if you if you look at this expression here, then this is not a
m Cp

function of x. The perimeter is constant, my heat flux is constant, m and Cp all are
constant. Therefore, the axial variation in the mean temperature is not a function of x.
Hence, the mean temperature as a function of x can be obtained from here which is
qsll P
Tm  x   Tm , i  
x . This simply tells me that for the case of constant surface heat
m Cp
flux, the mean temperature of the fluid varies linearly with axial position where x is the
axial position.
What I am going to do in the next class is for condition 2, when the surface temperature
is kept constant and I would like to see how I can express Tm , the temperature mean
temperature of the fluid as a function of x, for the case of constant wall temperature.
382
Heat Transfer
Lecture – 34
Internal Flow Heat Transfers (Contd.)
So, we are going to continue with our discussion on energy considerations in Internal
flow. And what we have seen is that when we do an energy balance that is whatever be
the energy comes in with the fluid and when it goes out it gains some amount of energy.
So, it simply going to be the energy gain is simply going to be mCp  T. We also
understand that something is being added from the side walls by Newton’s side walls the
energy that is being added and it can be expressed in terms of Newton’s law of cooling.
So, the dqs if I consider a differential area element, the amount of energy which is being
added from the side walls by the convection would simply be whatever be the flux
multiplied by Pdx.
So, dq convection the differential amount of energy which is being convected into the
fluid is equal to the flux multiplied by the parameter times dx where dx is the elemental
length of the fluid element.
So, dqconvection is qs multiplied by Pdx. P being the perimeter and for a circular pipe the P
is simply going to be equal to πD where D is the diameter of the pipe. Now, if qs is a
constant which was the previous case condition 1, constant surface heat flux in that case
I can simply integrate this.
And the total amount of heat which comes into the fluid which is q convection can simply be
stated as qs which is the constant surface heat flux multiplied by P times L where, L is
the length that we are considering in fully developed flow. And from there we have
obtained that how does the mean temperature of the fluid vary with axial position and we
saw that for the constant heat flux condition this variation in the mean temperature of the
fluid is going to be linear in nature.
Now, we are going to extend our discussion, include the second condition where the
temperature of the wall is kept constant and we would see how would the average
383
temperature Tm of the fluid change with axial position. And can we obtain something
which is similar to Newton’s law of cooling what would be the form of temperature
difference to be used in that case that is our goal.
So, we are going to start again with the equation that we have derived; we have derived
before where the variation in the mean temperature is simply going to be
dTm P
 h(Ts  Tm ) ,
dx mC& p
T  Ts  Tm
So, this in the, this is going to be the surface temperature which is constant minus T m, the
local value of the mean temperature. So, if I do differentiation on both sides so,
dTm d  T  P
  hT
dx dx & p
mC
So, this is what we can get. Now we can see I am getting a relation; getting a differential
equation in terms of the variation in the temperature difference on the left hand side and
the temperature difference on the right side as well.
384
So, I can integrate this equation; integrate the differential equation that we have obtained
between the, tube outlet and the inlet. So, when I do that what I would get is
To
d T PL 1 L 
    hdx 
Ti
T & p
mC L 0 
So, what I have done anything additional I put the  T over here and took the dx on this
side. So, this is going to be my integration at 0 it is the inlet condition, at L it is the outlet
it is it is simply going to be the outlet condition.
L L
1 1
So,  hdx  hL  hdx  hL =Average value of heat transfer coefficient
L0 L0
Now, when you see this integration, I think something it should strike you as I can
express this as P L and I put a 1 by L over here. Now, when you consider this part. So, I
have simply added L over here and L over here.
Now, when you see what does this tell you? This is the classical definition of length
average quantity. And the quantity in question here is h and therefore, I can write,
L
1
hdx  hL
L 0
So, this is the length average value of heat transfer, convective heat transfer coefficient.
This is the definition of length average length average value and in this case is going to
be the expression for the average heat transfer coefficient.
So, this entire thing now can then be replaced by hL . So, and if I complete the
integration what I get on this side is
T0 PL
ln  .h ,
Ti & p L
mC
PL
& p 
mC h
T0 L
ln
Ti
we realize once again that Ts is a constant.
385
So, this is the relation one can obtain for the case where the T s is constant and I have not
done anything new except bringing in the definition of the average length, average value
T0
of the heat transfer coefficient. So, from here I can find out what is And I am going
Ti
to write that in the next page.
So, what I have then is from the previous expression,
T0 Ts  Tm,o  PL 
  exp   h
 & p 
Ti Ts  Tm ,i  mC 
386
Where Ts is the constant surface temperature and Ti  Ts  Tm at the inlet. So, this is the
temperature difference between the flowing fluid and the tube wall at the outlet and at
the inlet.
 PL 
So, this is simply going to be exp   h  .So, that is the relation is going to be. So,
 &
mC p


the temperature difference it is not the temperature, like the previous case it is the
temperature difference decays exponentially with distance. So, the important observation
from here is that unlike the previous case where your, where the for the case of constant
heat flux condition, constant surface heat flux the mean temperature of the fluid varies
linearly with x.
387
What we get here is that the temperature difference decays exponentially with x.
qsP
Tm ( x)  Tm i  x
& p
mC
So, that is the difference when you have a constant surface heat flux and a constant
surface condition. So, based on this I can find out what is the total convective heat
transfer that is taking place when a fluid flows through while the temperature is
maintained constant at Ts.
So, this
& p (Tm,o  Tm,i )  qconv  mC

mC & p  Ts  Tm ,i    Ts  Tm ,o    mC
& p (Ti  T0 )
while this is the mean temperature at the outlet and this is the mean temperature at the
inlet. The one that I have here is simply reorganized by incorporating a T s term over
here.
So, when look carefully this is the temperature difference between the average, between
the fluid and the surface at location i and this is the temperature difference between the
mean temperature, between the mean temperature and the surface temperature at o. And
also keep in mind that this Ts is a constant in both cases as the entire analysis is being
done for the case of constant surface heat constant heat; constant temperature of the
surface.
388
Now I write qconvection , I just do a little bit of reorganization
T0  Ti
qconv  hPL
T
ln 0
Ti
PL
& p 
mC h
So, T0 L
ln
Ti
& p is being substituted in the expression over here. So, this from
So, this expression for mC
the last page it is substituted as you can see it here. So, once you substitute this in here
this is what you are going to get; this is what you are going to get in look carefully over
here.
T0  Ti
qconv  hPL
T
ln 0
Ti
This is area, this is the heat transfer coefficient, what is this? A s is the surface area and
for a circular tube As is simply going to be equal to P.L and I call this as  Tlm.
Now, what is this  Tlm?
T0  Ti
Tlm 
T
ln 0
Ti
So, this is nothing, but an average temperature difference between the inlet and the outlet
and because of the presence of the logarithmic term in the denominator this temperature
difference with the subscript Lm is termed as the logarithmic temperature difference. So,
this is some another sort of temperature difference where the temperature difference is
expressed in the logarithmic form and for the case of constant surface temperature this
logarithmic difference of temperature is to be used if we want to express the equation in
the form which is similar to that of Newton’s law of cooling.
So, Newton’s law of cooling generally tells us that
389
qconv  hAs Tlm
So, simple temperature difference would not suffice in this case when you are trying to
find out what is the total heat that is to be added, that is added to the fluid as a result of
the elevated temperature of the tube. Then it is simply
q=hA T ,
T =Ts-Tm
but my Tm is a function of x. So, whatever be the T m at the inlet would be different from
the Tm at the outlet even though my T remains a constant.
So, this T at one location and T at the outlet since they are different I need to use
some sort of an average temperature difference in order to obtain what is the total
amount of heat transfer. So, the total amount of heat transfer, q, should be equal to h
where, h is the length average heat transfer coefficient. AS, where AS is the total surface
area and we have just now shown that the average temperature difference at the between
the tube and the inlet temp fluid at the inlet and at the outlet is not arithmetic average.
We have a different kind of average which is called which is the logarithmic average.
So, while expressing the temperature difference to be used to with a with which some
value of h1 in order to obtain the total amount of heat transfer for the case of constant
surface temperature one must use the logarithmic mean temperature. So, this delta TLm
you will see the reference to it this will come again and again when we try to do in the
calculation for total heat transfer in internal flow when the temperature of the surface is
kept constant which you would obtain in many of the situations.
So, never use arithmetic average of the temperature difference between the inlet and
outlet, always use the logarithmic mean temperature difference between the inlet and
outlet. And in conjunction with h and A they would give you the total amount of heat
transferred from the pipe to the fluid. So, this is the new concept which we are
introducing here the relevant temperature difference is logarithmic mean temperature
difference for the case of constant surface heat, constant surface temperature condition.
390
So, will explain the little bit on this and you would see that this is the important relation
over here. And this is what is the definition of the log mean temperature, which is also
called as LMTD, Log Mean Temperature Difference.
So, this is log mean temperature difference. So, you can see this is another form of
Newton’s law another form of Newton’s law. So, we have done these two cases, it is
time for us to now look at apply this concept for the case of laminar flow in circular
tubes.
So, when we talk about laminar flow in circular tubes and we are looking at the thermal
analysis of that. When we are looking at the thermal analysis of laminar flow in circular
tube our starting point should obviously, be the equation of energy.
Now, if you recollect the equation of energy contains velocity terms as well. So, the
expression is simply going to be
T T    T 
u v  r 
x r r r  r 
There will also be heat generation term in there, but we will consider a situation in which
there is this going to be no heat generation.
391
So, I have the on the left hand side I have certain terms which will, which denote
conduction convection, on the right hand side will have the conduction part. So, we are
going to write simply the energy equation try to see which are going to be important and
then try to see is it possible to integrate that expression to obtain a definitive a closed
form solution of Newton’s closed form solution for the heat transfer coefficient for the
case of a circular tube. And of course, if its laminar flow if we can only do it if it is
laminar flow and if it is fully developed flow.
So, why laminar flow because only in laminar flow I know what is going to be the form
of the energy equation. Whenever it is turbulent flow additional terms due to the
formation of eddies which will aid in the transport of energy over and above that of
laminar flow would appear. And it is difficult to consider those eddies or the enhance
transport due to eddies in the form of an equation. And even if you do a closed form
solution of that cannot be obtained. So, you would start with the simplest possible case
where there is laminar flow.
And we will consider only that region from that point onwards when the flow is
thermally and hydro dynamically fully developed and try to see if there is a way to
obtain the value of the Nusselt number; the expression for the Nusselt number for the
case of constant heat flux and for the case of constant temperature. So, these are the two
limiting cases, two ideal cases that we will try to analyze in this part of the class. So, let
us start with this thermal analysis.
If we write the energy equation the energy question is going to be the as we have done
before this is the convective part that is
T    T 
v  r 
x r r  r 
There would be another term over here which would give you the conduction in the x
direction.
So, there would be a term for that as well, but if you consider, it is a case of a tube where
flow is taking place and heat is being added either in the form of a constant heat flux or
in the due to an elevated temperature. So, most of the heat is transported and I have the
flow. So, most of the heat is transported in the r direction by conduction and in the x
392
direction the principal mode of heat transfer is going to be convection. So, conduction in
r direction and convection in x direction these two are going to be the principal mode of
heat transfer therefore, this part is approximated to be 0 the conduction in the x direction
is therefore, neglected only the conduction in the r direction is kept into the energy
equation form. Which is a direct result of the discussion that I have just made.
And since, it is fully developed flow in fully developed flow the v is can be
u
approximated to be 0 ( v  0 ). As well as  0 . And you also understand that the
dx
velocity profile of the fluid which is flowing the nature of this velocity profile is going to
be parabolic in nature. So, we are making certain simplifying assumptions in here that is
its fully developed flow.
So, that is only the flow becomes only one dimensional and the moment the flow
becomes one dimensional the other component of velocity is going to be equal to 0, only
u
the u component is going to be present in here. And  0 , though I cannot say the
dx
T
same thing about , but definitely the axial velocity will not be a function of position
dx
anymore that is the classical definition of hydrodynamically fully developed flow.
And for that flow we also understand that the velocity profile is simply going to be
parabolic in nature. So, this part is already known I can drop this term in over here
instead of u I can express it in terms of the parabolic velocity distribution. So, we are
going to first solve for the case of constant surface heat flux. What we have done today
T dTm
we know that for the constant surface heat flux  .
dx dx
This is what we have done in previous classes for qs to be a constant. So, that is the
special case that we would have. So, I put this term over here.
So,
393
2
1 d  dT  2um dTm   r  
r  1    
r dr  dr   dx   ro  
Just take a minutes time and see if you understand what we have done over here. So, this
is what I have written over here only this part. So, this α comes over here and I am going
to write u which is the point velocity for the case of laminar fully developed flow it is
parabolic in nature. So, this is two times the mean velocity which you remember times
2
r T dTm
1    .And for the case of constant surface heat flux  , where Tm is the mean
 ro  dx dx
temperature of the fluid.
So, starting with the energy equation I have written it for the special case of thermally
fully thermally and hydro dynamically fully developed flow by making the assumption
that V is 0. The principal direction of convective heat transfer is here. So, therefore, there
is no convective heat transfer or a very little heat transfer convection in the r direction.
So, this part is taken to be 0 the there is no conduction very little conduction in the x
direction. So, the only non 0 components these two could be the form of the equation.
So, this equation this is this equation can now directly be obtained as d T m d x is not a
function of r, um is the mean velocity is not a function of r. So, this equation can be
integrated quite easily over here and what you get after integration is T.
2um  dTm   r 2 r4 
T (r )      C ln r  C2
  dx   4 16ro 2  1
394
I would advise you to do this integration on your own and see practice whether you are
getting the same result this C1 and C2 are integration constant which we need to evaluate
by applying certain boundary conditions.
So, the first boundary condition that I write is that T has to be finite at r is 0, T cannot be
infinite at r equals 0. So, in order for temperature to be finite at r equals 0 if you look at
this expression C1 ln r . So, for r equals 0 , T to be finite C 1 must be equal to 0. So, my

first boundary condition finite T at r equals 0 looking at the nature of the expression
simply tells me that is C1 has to be 0.
And the second one I am going to write C2 is going to write
T (ro )  Ts
We keep in mind that Ts is a function of x since it is a constant surface heat flux

condition my Ts is not a constant. So, therefore, when I use these two and I simplify and
find out the expression for C2 the final expression would be
4 2
2um ro 2 dTm  3 1  r  1  r  
T (r )  Ts         
 dx 16 16  ro  4  ro  
 
395
And the inside is going to be, I again please do it on your own to just to practice and to
see you are getting the same result same result this is square.
The mean temperature this T is a function of r. So, if I need to find out what is the mean
temperature which is we are saying as Tm please look at your class notes.
r
2 o
Tm  uTrdr
um ro 2 0
So, one what you have to do is then this T you are going to substitute in here. The
parabolic distribution is going to so, this is going to be parabolic and do the integration.
Once you do this integration what you would get is the mean temperature
11  um ro 2  dTm
Tm  Ts   
48    dx
And since we know that
dTm qsP

& p
dx mC
 D2
Where, P   D, m&  um
4
396
So, this qs is a constant and I am looking at expression for T m which can be used to
dTm
obtain what is . So, when you do that what you would get is
dx
11 qsD
Tm  Ts  
48 k
So, when you use this
qs  h(Ts  Tm )
So, when you put that is that in here. What you get is
48 k
h
11 D
and your Nusselt number based on diameter which is
hD
Nu D   4.36
k
Look at this expression, this is one of the very important relations that can simply be
obtained, that can simply be obtained by doing an analysis which would give you the
Nusselt number to be a constant.
So, the uniform qs constant case fully developed both thermally and hydrodynamically
and if its laminar if you make all these assumptions what you would get is Nusselt
number is a constant independent of Reynolds number based on diameter as long as you
are within the laminar flow range, independent of Prandtl number and independent of
axial location. That is a remarkable result that when you have such a situation the
Nusselt number turns out to be a constant and the in order whenever you sight a result
you have to see what is what are the conditions that you have assumed in order to arrive
at the result. So, the conditions that we are we have assumed is that the surface heat flux
is a constant. It is a thermally and hydrodynamically fully developed flow and it is the
flow is under laminar condition.
When all these conditions are met the Nusselt number has a constant value equal to 4.36
implying that it is not a function of Reynolds number as long as you stay within the
397
laminar flow range. It is not a function of Prandtl number and it is not a function of axial
location.
So, a constant heat flux situation gives you an unique value of heat transfer coefficient.
And, if you know that you are in the laminar flow range you can simply find out what is
the value of heat transfer coefficient for such a situation. Simply equate Nusselt number
to be equal to 4.36 and through the definition of Nusselt number find out what is the
value of convective heat transfer coefficient based on flow inside the tube.
So, what we would do in next class is find out what is going to be the form of Nusselt
number relation when instead of a constant surface heat flux condition I would assume a
constant temperature of the inner surface of the tube and how can I find out this Nusselt
number. So, for two very well defined cases I need to know what is the Nusselt number
expression, in one case I saw Nusselt number to be equals 4.36. We would examine the
Nusselt number for the other case constant surface temperature in the next class.
So, if we do that at least we have some understanding and we know how to handle
laminar flow conditions. And then we have to extrapolate our knowledge of fluid
mechanics and our knowledge of analogy to find out what we can do when we have
turbulent flow inside a pipe. So, what would be the relation or correlations to be used to
obtain hi for the case of turbulent flow on the inside of the tube. If you do that then my,
our knowledge is complete.
I have an h, I know how to evaluate h i for laminar flow for turbulent flow and for
intermediate flows. I already know how to calculate h o then the overall resistance to heat
transfer from the outside fluid to the inside fluid which is flowing through a pipe through
a tube will be known to us which will form an integral part in the design equation of heat
exchanger or heat exchanger devices. So, we will look at the Nusselt number relation for
constant surface temperature in the next class.
398
Heat Transfer
Lecture - 35
Internal Flow Heat Transfer (contd.)
We are discussing internal flow in a pipe or in a tube and we have seen several ways by
which the simultaneous fluid flow and heat transfer inside a tube can be expressed. The
first thing what we have done is we defined what is known as thermally fully developed
flow, and there we have seen that in thermally fully developed flow, though the
temperature will be a function of axial position, the shape of the dimensionless
temperature profile will not change with the axial position. That is how we have defined
what is known as thermally fully developed flow.
We also have seen that the heat transfer coefficient remains a constant for a thermally
fully developed flow. So, h will not be a function of x. We have applied the concept for
the case of laminar flow in a circular tube. Whatever we have discussed so far deals only
with laminar flow; we did not introduce turbulence into our discussion so far. We have
also seen that the 2 cases which appear in most of the times is one where the surface heat
flux is constant, and the second where the temperature of the tube wall remains a
constant. We started with the equation of energy and we have assumed that it is a 2-
dimensional flow. We have written the energy equation and what we saw is that the
some of the simplifying assumptions that we could do for the case of laminar flow; for
example, we have neglected the axial conduction of heat as the fluid moves in the tube.
So, it is a radial conduction which is important. So, we solve that equation for the case of
constant heat flux, and we have seen for that specific case the Nusselt number turned out
to be a constant. It has a specific value which is equal to 4.36.
Therefore, the same type of analysis can also be done for the case of constant wall
temperature. And we would see that similar to the case of constant heat flux, the Nusselt
number for the constant surface temperature will also be a constant. And its numerical
value would be equal to 3.66. So, therefore, it is again independent of the Reynolds
number, it is independent of the type of fluid and so on. And as before the k is evaluated
at the mean temperature of the fluid.
399
So, up to this point it is fine, but how do we expand these results for the case of turbulent
flow? We understand that turbulent flow is much more complex than the laminar flow,
because in turbulent flow there would be other means of transport of heat rather than
simple convection and conduction. Because turbulent flow it is characterized by the
formation of eddies, and those eddies are packets of fluid which carry more energy or
more momentum. And therefore, the equation of energy takes up a much more
complicated form. If you recall from your fluid mechanics the Navier-Stokes equation
which is the equation of motion will have additional terms, which are collectively called
as Reynolds stresses when you have turbulent flow present in the system.
Similar to that, when you have turbulent flow and heat transfer is taking place while the
flow is in the turbulent regime, the resulting governing equations will have additional
terms, due to the augmented heat transfer by the presence of these eddies. So, it is
difficult to express these results for turbulent flow. And therefore, an alternative
approach must be taken. And the tool that we use is the similarity, the analogy between
heat transfer and between momentum transfer. We have the simplest possible case of
Reynolds analogy, when Prandtl number is equal to 1. So, we could say that the Nusselt
number and the friction coefficient will become identical, or the Stanton number based
on heat is going to be equal to Cf; where Cf is the friction coefficient times Rel/ 2; where
400
Rel is the is the Reynolds number. So, that analogy if we can use, and if we know what is
going to be the expression of Cf; or expression of f, the fanning friction factor. If we
know their relations in momentum transfer for the case of turbulent flow, then using the
analogy we can obtain an expression for Nusselt number in turbulent flow for the case of
combined heat and momentum transfer. That is the utility of the analogy which lets us
use expressions obtained in one type of transport process with the other. For example, in
this case, the momentum transport process, and how do I correlate that with the heat
transfer process that is taking place inside the tube.
Now, what is the expression for f, the fanning friction factor, how do I obtain the value
for the case of turbulent flow? All of you have studied fluid mechanics, and you know
that the friction factor for turbulent flow can be evaluated using the Moody's diagram.
Moody's diagram gives you the value of the friction coefficient as a function of Reynolds
number even for the case of turbulent flow, for smooth pipes as well as for rough pipes.
We get a family of curves for different values of smoothness of the pipe, or different
values of roughness of the inner surface of the pipe, and therefore, at any given value of
Reynolds number, you should go to the curve corresponding to the roughness that your
tube has, and read what would be the value of f, that is what Moody's diagram is all
about. So, if I can express this friction factor f as a function of Reynolds number, then it
would be easier for me to extend this result for the case of heat transfer as well. So, for
the turbulent flow inside a pipe we must use the analogy between heat transfer and
momentum transfer, and knowing the expression of f from Moody diagram, I should be
able to obtain an expression of Nusselt number.
401
Let us now concentrate on turbulent flow in a circular tube. And just to recapitulate what
we have already what I have already shown you, the friction factor which you obtained
from Moody diagram is related to the Reynolds number based on the diameter of the
1
tube as: f  0.184 Re D 4 ; and this correlation is valid for Re D  2 104 . This correlation
is known as the smooth pipe approximation. And I have also told you, for the case of
turbulent flow, the flow conditions are less affected when you have roughness in the
pipe. Roughness in the pipe affects laminar flow more than turbulent flow. So, the
microscopic roughness which are present on the inside surface of the tube, affect the
value of the friction factor more in the case of laminar flow as compared to that in the
turbulent flow. And for highly turbulent flow, since the boundary layer thickness is large
as compared to laminar flow, these microscopic imperfections or roughness on the
surface, do not affect turbulent flow that much. So, the relation that I am using is for
smooth pipe, but they can also be used for rough pipes as well. The friction factor as I
have shown in that correlation is directly from the Moody diagram. The Chilton-
Cf f 2 NuD 2
Coulburn analogy essentially tells you that   St Pr 3  Pr 3 where C f is
2 8 Re D Pr
the friction coefficient. Therefore, the friction factor (f) from the Moody diagram is
going to be function of Reynolds number, and here I have Nusselt number Reynolds
number, and Prandtl to the power minus 1 by 3. So, utilizing the expression of f from
402
4 1
here, I can write Nusselt number based on diameter as Nu D  0.023Re D5 Pr 3 , this is
known as the Coulburn equation.
The Coulburn equation simply gives you the expression for Nusselt number when the
flow is turbulent. It highlights the importance of the analogy between heat transfer and
momentum transfer; the Chilten-Coulburn analogy is valid over a wide range of Prandtl
number I have shown before. So, Chilton-Coulburn analogy lets me obtain an expression
for h or Nusselt number in highly turbulent flow by simply using the expression for the
fanning friction factor f obtained from the Moody diagram.
It has been shown experimentally that a slightly better result can be obtained if this
constant 0.023 is changed and therefore, another relation which is mostly shown in the
4
text book is known as the Dittus-Boelter equation, which is: NuD  0.023Re D5 Pr n . This
n will have a value equal to 0.4 for heating; which is when the temperature of the surface
is greater than the temperature mean temperature of the fluid. And n is going to be equal
to 0.3 for cooling; when the temperature of the surface is less than the temperature mean
temperature of the fluid. And the applicable limit of Prandtl number is quite large:
0.7  Pr  160 . 0.7 is the standard Prandtl number for the case of air. And Reynolds
number is more than 10,000, and L/D is greater than equal to 10. So, L/D greater than 10
tells you that it is a thermally fully developed flow, the Reynolds number greater than
10000 tells you that it is in the highly turbulent region, and the range of Prandtl number
ensures that this relation can be used for a wide range of liquids from air on one hand to
a high Prandtl number liquid. So, for all future calculations will stick to Dittus-Boelter
equation which is again just an empiricism added to Coulburn equation to take into
account whether it is a case of heating, or whether it is a case of cooling.
Now if you recall I said that the mean temperature is used to calculate the physical
properties of the fluid. Now when the fluid is flowing through a tube, the temperature
will keep on changing. And as the temperature will keep on changing if it is a case of
heating then the temperature will keep on increasing. So, the value of the thermo
physical property will have different values at different temperatures.
So, if the temperature change is not significant, then the values of the thermal
conductivity will not be substantially different between the entry and the exit because k
403
is not a very strong function of temperature. But there is one property of the fluid which
is a strong function of temperature. And obviously, the property which changes
significantly with a change in temperature is the viscosity.
So, when you calculate the Reynolds number, what is the viscosity that you are going to
use? Is it the entry point viscosity and the Reynolds number is calculated based on that,
or is it somewhere down the line where the temperature is more, the viscosity is less for a
liquid and therefore, the Reynolds number will have a different value?
So, whenever you are using a correlation in order to obtain the value of the Nusselt
number, in order to calculate the Reynolds number, you must be careful about the
temperature change that you expect between the entry and the exit. So, if I have to add a
correction factor to take into account the variation of property, and the property in
question would obviously be viscosity. Because viscosity is the one which changes
significantly between 2 points, then to this Dittus-Boelter equation a correction term for
viscosity must be added.
So, the next correlation that I am going to show you its genesis for this correlation is to
take into account the variation in thermo physical property which is viscosity. So, a
viscosity correction factor is added to the Dittus-Boelter equation in order to extend its
applicability over a wide range of temperature change. So, that equation is known as
Sieder-Tate equation. So, whenever you expect significant changes in temperature along
the length of the tube, do not use Dittus-Boelter equation, but rather use Sieder-Tate
equation that has a viscosity correction factor empirical factor embedded to it.
So, let us see what this looks like.
404
So, when you have a large property variation, the equation to be used is Sieder-Tate
0.14
  4 1
equation. And the Sieder-Tate equation is: Nu D  0.027 Re Pr   5 3
. This
 s 
additional term is added so as to take into account, the possible variation of viscosity,
when the flow is taking place. So, this again has an applicable limit of Prandtl number
between 0.7 and 160, the Reynolds number has to be greater than equal to 10,000 and L
by D is greater than 10. So, properties except  s are evaluated at Tm. And as I said, the
other properties are not a strong function of T m. I must mention that this correlation is
valid for both constant heat flux and constant surface temperature. Whenever you use a
correlation, you have to keep in mind that this is not an analytic expression it is a
correlation. So, there will be some error associated with it, but you should always judge a
correlation with the ease of its use, and how much of error it introduces in your final
calculation. So, judging by these 2 criteria, Sieder-Tate expression has proven to be a
very successful relation to connect the Nusselt number for flow inside a tube, both for
constant surface temperature and constant heat flux conditions for a wide variety of
liquids and over a wide range of temperature.
So, if your temperature of the wall is at 150 and the mean temperature is of the liquid is
30. So, you can have a significant variation of viscosity near the wall as compared to the
rest of the fluid and the correlation of choice is Sieder-Tate. So, if there is a substantial
temperature change, we do not use Dittus-Boelter, but we use the Sieder-Tate relation.
405
There is one more complication. In many of the practical applications you do not have a
circular tube, you have conduits through which the flow is taking place and the cross
section is not circular. So, what do I do for the case of diameter when we have flow
through a non-circular conduit? Can I still use the correlations that I have presented so
far, yes, we can? But we have to be careful to use not just a diameter, because there is no
diameter. It is the flow is taking place through a thin duct. So, what is the diameter in
such a case? So, the diameter to be used for non-circular conduits is the hydraulic
diameter. So, always use hydraulic diameter to calculate the Reynold’s number, and then
use either the Sieder-Tate or the Dittus-Boelter equation.
So, the non-circular nature of the conduit is accounted for through the introduction of the
hydraulic diameter. Please keep that in mind when you are making a calculation for heat
4 Ac
transfer in non-circular conduits, we use the hydraulic diameter which is defined as ,
P
where Ac is the cross sectional area, and P is the perimeter. Now one thing I would also
like to bring to your notice in the case of Sieder-Tate equation or Dittus-Boelter
equation, that this L by D it taken to be greater than 10; that means, the flow is fully
developed. But there would be situations in which the length of the pipe is such that the
entrance region is not insignificant. That means, there is a significant section of the pipe
in which the flow is still developing. So, the correlations which we have obtained for
fully developed flow, namely the Dittus-Boelter equation or the Sieder-Tate equation is
not valid when we have significant portion of the pipe under the developing flow
condition.
So, if you want to have a generalized relation which is valid for turbulent flow which
also takes into account the developing part of the flow, that relation I am going to write
next which can be used over a wide range. So, when I consider entry region; that means,
the region where the flow is still developing, the relation to be used is:
1
 Re Pr  3   0.14
Nu D  1.86     . Here, NuD is the average Nusselt number based on the
 L   s 
 D 
diameter of the conduit. So, the keyword here is when I consider the entry region. This
expression is purely empirical.
406
Therefore, this will conclude our treatment of internal flow and heat transfer when the
flow takes place in a pipe that could be circular or non-circular as well
407
Heat Transfer
Lecture - 36
Internal Flow and Heat Transfer (Tutorial)
Today we are going to discuss and solve a few problems on Internal Flow. So, this would
be a tutorial class where I will try to solve at least two problems both involving Internal
Flow. So, far we have calculated, we have obtained the relationship of heat transfer
coefficient, convective heat transfer coefficient for flow through a tube for the case of
laminar flow, for the case of turbulent flow, in situations where we cannot neglect the
effect of the entry length and also situations in which there is a substantial temperature
change between the mean temperature of the fluid which is flowing through the tube and
the tube surface temperature.
This significant variation in these two temperatures would require a correction term to be
added to any in any relation or correlation that we have obtained for Nusselt number
because, the viscosity is the property which is going to be most affected by a change,
sharp change in temperature between the surface and the bulk. So, a viscosity correction
factor in certain cases has to be introduced in the expression to obtain the correct value
of the heat transfer coefficient. So, we have various relations for laminar flow, it is a
straightforward relation, the Nusselt number involving Reynolds number and Prandtl
number.
For the case of turbulent flow we have used analogy to obtain the relation between
Nusselt, Reynolds and Prandtl and we also have a viscosity correction factor which is
added to it, and on top of that in some cases we have to take into account the entry
length. That means, the length required for the flow to be fully developed thermally is
comparable or is significant such that it cannot be neglected and therefore, a correction
factor for the entry length part will have to be added to the correlation, which was
developed for turbulent flow. So, all those we have covered in our previous classes. So,
today we are going to concentrate on a few problems.
408
So, the first problem the deals with is about a fluid engine oil, which are characterized by
high value of Prandtl numbers. Engine oil is entering at a rate of 0.5 kg/s and at a
temperature of 25 ºC into a tube whose diameter is 25 mm, and the tube surface
temperature is maintained constant at 100 ºC. So, once again engine oil flows through a
tube of diameter 25 mm at a rate of 0.5 kg/s, and the temperature at the point of entry for
the engine oil is 25 ºC. The tube surface temperature is maintained at 100 ºC .
So, what we need to calculate is, what is the oil outlet temperature if the length L of the
tube through which the flow takes place in one case it is 5 m, in another case it is 100 m.
And we are also ask to calculate for each of these cases the Log Mean Temperature
Difference and compare the value of the LMTD with the arithmetic mean temperature,
that is arithmetic mean of the temperature at the entry and temperature at the exit for two
values of L = 5 and 100. The relevant properties for oil are also provided, the heat
capacity is given, the viscosity is provided, the thermal conductivity, the Prandtl number
for engine oil just see that the it has a very high value of Prandtl number.
But, additional information which is provided which I have marked in red is the viscosity
that the viscosity of engine oil at the temperature of the surface, which in this case is 100
ºC. So, μ at the surface temperature is provided and you would see how much of
difference between you have, the μ at the bulk temperature and the μ at the surface
temperature. So, there is a significant variation in viscosity. So, a viscosity correction
409
factor must be employed in order to obtain a correct value of the heat transfer coefficient
or the value of the temperature and so on.
So, the fundamental equation with which we are going to start is
Ts  Tm ,o  PL 
 exp   h
 & p 
Ts  Tmi  mC 
Where, Ts is the temperature of the solid surface of the tube, T mi is the temperature at the
inlet of the fluid, which is flowing through , P being the perimeter , L being the length,
m&being the flow rate .
So, we need to find out what is the value of the mean temperature. So, this is something
which you would like to obtain. The length it could in one case it is 5 meter and in one
case it is 100 meter so, the problem essentially boils down to how to evaluate h .
So, that is the that is all there is to this problem; how would we are going to find the
value of h . Now, whenever we try to find the value of h we need to know what is the
flow pattern; is this laminar flow or is it is going to be turbulent flow. So, that is the first
thing that one has to evaluate and therefore, I need in order to do that I have to find out
what is the value of the Reynolds number.
4m& 4  0.5kg / s
Re D    304.6
 D   0.025  0.0836 Ns / m2
Now if you look at this value this definitely tells you that it is going to be a laminar flow.
Since the value of Reynolds number is only about 300. We also need to consider the case
of entry length, if the entry length is significant. So, therefore, we need to use the relation
that Nusselt number, the average value of Nusselt number based on the diameter this
tube is
1/3 0.14
 Re.Pr    
Nu D  1.86    
 L / D   s 
So, that is the standard relation which we have used for the case of entry length is
significant.
410
Also Sieder Tate equation which takes into account the variation of viscosity between the
bulk and the tube wall temperature. So, this is the relation which we are going to use I
know the value of Reynolds number, the value of Prandtl number is provided in here the
length is either 5 or 100, the diameter, the values of viscosity are provided. So, from here
I can calculate: what is the Nusselt number and from Nusselt number I should be able to
obtain what is the value of the convective heat transfer coefficient.
So, proceeding with the value for the length to be 5 meters I am going to
1/3 0.14
0.141  304.6  1205   0.0836 
h s  1.86       119W / m 2 K
0.025  5 / 0.025   0.1441 
So, therefore from the expression over here I can easily find out what is the value of Tmo.
   0.025  L 
Tm ,o  100  75exp   119  = 28.4ºC
 0.5  2035 
So, this is what it is going to be and this Tmo would turn out to be 28.4ºC.
In a similar fashion I am not doing this again the L equals 100 meters, the h that you are
going to calculate is 40 W/m2K. So, compared to the value of h at 5 and this is h at 100
you see the difference, you see how the average value reduces as the length is increased.
411
This so, you would be able to see the variation of h with length and this when you apply
over here the Tmo for the case of 100 meters it will turn out to be 44.9 ºC.
So, again see the change in the difference in change in the temperature between and
between at the outlet for a smaller length and for a larger length. And obviously, for
larger length the outlet temperature definitely is going to be more than that of the smaller
length. And, when you talk about when you do the second part that is find out the
difference between the between the arithmetic mean temperature and the log mean
temperature. So, for L equals 5 meters you calculate
To  Ti 71.6  75

TLM   =73.3ºC
ln To / Ti ln 71.6 / 75
So, these are the values of the temperature difference at the outlet between the bulk and
the wall, temperature difference at the inlet between the bulk and the tube surface
temperature. So, this you would see this turn out to be at 73.3 ºC and if you do the
To  Ti
arithmetic mean which would simply be , this is also going to be 73.3 ºC.
2
So, for a short length we do not see much of any difference between the log mean
temperature difference and the arithmetic mean temperature difference both are equal to
73. When you do this for 100 meters, this if you would calculate this and see that TLM
to be equals 64.5 ºC and T arithmetic mean comes out to 65.1 ºC.
So, for this specific fluid at two different lengths of the pipe we see almost no difference
for shorter length of the pipe and the differences will start to creep in as the length of the
pipe is increased, but still it is not going to be noteworthy. But, this whatever we have
obtained for heavy oil the same calculations may, if you do for a different fluid then
these values are going to be different. And, we will start to see significant departure
between the values of the log mean temperature difference and arithmetic mean
temperature difference. And, you know in our previous classes I have shown you that
why LMTD is the correct representation of flow; when the tube surface temperature is
maintained constant in internal flow. So, in your subsequent calculations even though for
this specific problem you do not see much of a change, always stick to the log mean
412
temperature difference. Now, in the remaining few minutes I am going to solve one more
problem quickly and then we will move to something else in the next class.
So, in this case the problem 2 which you call is a problem 2, it is an air which is flowing
at a mass flow rate of 0.05 kg/s through a duct whose diameter is 0.15. The inlet
temperature of the air is 103 ºC and after L = 5 m, it cools to 77 ºC. So, this is the duct,
you have flow of air which is m& provided 0.05 kg/s, the Ti is 103 ºC and diameter is 0.15
m. The length is 5m, the h o, the heat transfer coefficient from the outside of the duct to
outside which is ho is provided as 6 W/m 2K and the value of T∞ here, the T infinity is
provided as 0 ºC.
So, air flows at a mass flow rate of 0.05 kg/s with a temperature of 103 ºC. It enters the
duct and the duct is exposed to air at 0 ºC with the heat transfer coefficient between the
outside of the duct to the atmosphere as 6 W/m 2K. The first part is calculate the heat loss
from the duct. So, for the overall length L; so, the for the entire length of the duct how
much is the heat loss from the duct to the outside and the second part is determine the
heat flux and the duct surface temperature at duct surface temperature T at x = L.
So, what is the heat that is total heat that is going out for the entire length and what is the
value of the heat flux at x = L; that means, at 5 m and what is the temperature of the duct
surface at x = L? And here we are assuming that the duct is of small thickness of high
thermal conductivity. So, there is no difference no conductive resistance. So, the
413
temperature on the inside of the duct and the temperature on the outside of the duct both
are equal. So, this is a very thin duct and therefore, no conduction resistance. The first
thing what is the total heat loss from the air to the outside can simply be obtained by
making a heat balance. So, we know what is the temperature at which it is entering, if we
know what is the temperature at which it is leaving and I know what is the mass flow
rate of the air which is flowing through this duct and I know it is C p. So, I can simply
find out what is going to be the temperature, what is going to be the total heat which is
lost from the hot air to the outside. So, let us calculate that first. So, when I do calculate
the first one the from energy balance I can simply find out the total q
& p (TmL  Tmo )  0.05kg / s  1010 J / kgK  (77  103)  1313W

q  mC
So, the air was entering at 103 ºC and it was lead. So, this is whatever be the temperature
at the inlet and its cools to 77 ºC. So, that is why the temperature at the outlet is 77 ºC
and the minus sign simply denotes that heat is lost from the air to the outside. So, that is
why we have the negative sign. So, I think that this part is clear. Now, the second part is
to obtain what is the heat flux at a given location, heat flux at x = L. Now, let us draw the
thermal circuit first. We have the temperature of the air which is higher than the
temperature of the outside. So, this is my temperature of the outside that I denote as T ∞
and I know that this T∞ is at 0 ºC.
So, from 77 ºC to 0 ºC heat transfer is taking place and the heat flux therefore, simply is
qs at a location L. So, this is the heat flux. So, the double prime denotes the flux at x = L
and it travels from high potential to lower thermal potential this T mL is 77 ºC and this T∞
is 0 ºC. In order to go from 77 ºC to outside it encounters three resistances, the three
resistances are one is the resistance between at the inner surface of the tube. So, this is
the tube the first resistance is between is at the inner surface of the tube. The second
resistance is in the tube itself which we are neglecting since, the tube diameter is very
small and the other resistance is from the outside of the tube to the outside atmosphere.
So, this is the convective heat transfer coefficient on the outside and this is the
convective heat transfer coefficient at the inside neglecting whatever we have as the
conduction resistance in the duct. Therefore, if I draw a circuit what I would get is
something like this; this is going to be my temperature of the surface of the duct at
414
location L and from inside to the wall this resistance is going to be 1/h x which we
understand is a function of L. And, from the out from the outside of the tube to the
outside atmosphere the resistance is simply going to be 1/h o and we know what is the
value of ho which has already been provided to be as 6 W/m2K .
So, I have a flow of current through these two resistances which are in series that is all
there is to the problem. So, the only unknown this TmL is known,T∞ is known. So, in
order to evaluate qs ( L) , somehow I need to calculate what is the value of hx; if I can do
that then I should be able to calculate what is the heat flux from the air which is flowing
to the outside. So, the only job that remains is to obtain 1/h o, obtain the heat transfer
coefficient in this case.
So, as with any such situations first we have to find out what is the Reynolds number.
So,
4m& 4  0.05
Re D    20404
 D   0.15  208 107
So, with this high value of Reynolds number this is definitely going to be a turbulent
flow which is taking place inside which is taking place inside the tube. And, when you
calculate this L/D, the length by diameter you would see that for 5 meter length of a
diameter, this turns out to be 33.3 and since it is greater than 10, so, definitely this is
415
going to be a fully developed flow. So, a fully developed flow relation for the case of
turbulent flow maybe used in order to obtain what is going to be the heat transfer
coefficient. We understand that in some cases the flow may not be fully developed and
therefore, a significant portion of the pipe might be under developing flow.
So, therefore, any relation which is valid only for fully developed flow may not be used
in such situations, but here we see that L/D is greater than 10, it is about 30 or so. So,
therefore, the fully developed relations or correlations for turbulent flow can be used in
this specific situation to obtain what is the value of the heat transfer coefficient. So, that
is all there is to the problem. So, let us quickly evaluate what is the value of h and then
from there we find out what is the value of the total heat flow heat flux at that specific
location. So, we are going to use this relation Nusselt number
hx ( L) D
Nu   0.023Re D 4/5 .Pr 0.3  57.9
k
57.9  0.03
So, hx ( L)   11.6W / m 2 K
0.15
So, that is the value of the heat transfer coefficient for this specific situation.
So, my heat flux which is
TmL  T 77  0
qs ( L)    304.5W / m 2
1/ hx ( L)  1/ ho 1/11.6  1/ 6
The surface temperature, Ts L
qs ( L)
Ts L  TmL   77  304.5 /11.6  50.7 C
hx ( L)
When I consider convective heat transfer from the bulk fluid to the inner surface of the
tube and this is the internal heat transfer convective heat transfer resistance.
So, this is the temperature of the air at a distance of 5 meters when it enters when it
enters at. So, the air enters at 103 ºC with the flow rate of 0.05 kg/s, the outside heat
transfer coefficient is given as 6 W/m 2K and when it travels 5 m the air cools down from
103 ºC to about 50 ºC.
416
So, this is another example of how to find the heat transfer coefficient and from there to
find out what is going to be the heat flux at that point everything is done at steady state.
What are the values of the internal heat transfer coefficient, the external heat transfer
coefficient and these two are going to be in series. So, for solving any problem on
internal flow the first thing one should do there are few relations which are available.
One is the overall energy balance, what is the total energy lost or gained by the fluid if
you know the inlet and outlet temperature these would and the flow rate.
& p T would give you the total amount of energy which is lost or gained by the
So, mC
flowing fluid and then. So, that is between the fluid at the inlet and outlet, but you also
need to consider in certain cases, what is the condition, what is the potential for heat
transfer. So, what is the temperature at the outside and what is the temperature at the
inlet and at the outlet what is it is temperature and what is the temperature at the outside
or may get the temperature at the wall of the tube. So, the effective temperature
difference at the inlet and the effective temperature difference at the outlet; so, (Ts-Tm) at
the inlet and (Ts-Tm) at the outlet, these temperature differences are the cause for heat
transfer between the wall of the tube and that of the fluid. But the actual temperature
difference to be used throughout when you consider the entire length of the pipe is not
the arithmetic difference, it is the Logarithmic Mean Temperature Difference or LMTD.
So, when you do that and based on our calculations previous classes the derivations we
know how Tmo and Tmi, the inlet and the outlet they are connected in terms of the
perimeter, in terms of the diameter, in terms of the mass flow rate, in terms of some of
the physical properties of the fluid and most importantly what is the overall heat transfer
coefficient. This heat transfer coefficient, the internal heat transfer coefficient the
convective heat transfer coefficient for the internal flow can be evaluated as a function of
Reynolds and Prandtl number.
Based on whether it is laminar flow, whether it is turbulent flow or whether it is a

turbulent flow where significant variation in viscosity is to be taken into account and the
third and the final one for Sieder Tate relation in which the effect of the entrance length
cannot be neglected. So, for all those cases the relations or the correlations where,
provided I have discussed them in detail. So, the problem essentially boils down to how
to calculate which relation or correlation to take to calculate the value of the heat transfer
coefficient. Once you do that from that point onwards it is simply going to be two
417
resistances in series and similar such problems where, the internal and the external flow
the hi and ho are to be connected in order to obtain what is the total resistance for heat
flow from the inside of inside fluid to the outside or vice versa.
So, that is all that we have discussed; for completeness I will also tell you some of the
relations for external flow which are to be used when steam condenses on one tube or
steam or any or heat transfer is taking place when we have considered one tube, when
you consider a number of tubes or banks of tubes. And the other possibilities when I
have tubes like this the flow of the hot air let us say on the outside can be across the tube
bundle or along the tube bundle. So, all these different scenarios would give rise to
different relations or correlations. You need not remember any of those relations, but you
should be aware that for each of these situations you have a specific relation or a
correlation available in your text and you have to choose the right relation.
In any question of heat transfer in my course, I will provide you with all the relations and
correlations. But, you have to choose which relation to for which situation; that is where
you understanding comes. I do not want you to memorize anything, I would rather want
you to understand what are the limitations, what are the specific situations and which
equation or which relation is to be used and why under what circumstances; if you
understand that that is more than sufficient.
So, in tomorrow’s class we will starts I will give briefly give the some of the relations for
external flow and then we will look into another interesting heat transfer situation which
is known as the natural convection; where the flow takes place because of a difference in
buoyancy and not because of any imposed pressure gradient. So, which is a very
common occurrence in many of the heat transfer situation; so, you need to have some
idea of how to calculate the heat transfer in natural convection.
418
Heat Transfer
Lecture - 37
Free Convection
In previous classes, we have seen the relations and correlations for external flow as well
as for internal flow. In all these cases there was an imposed pressure gradient specially
for internal flow which drives the fluid through the tube. There are several locations in
which there is no externally imposed flow on the system even though a flow will
automatically start because of a difference in buoyancy. So, this buoyancy difference
driven flow will lead to convection as well which is known as the free convection or
natural convection. This is a very important part of the overall transport process around
many structures, buildings and surfaces. So, whenever there is an absence of a forcing
velocity, the heat transfer under that situation which is caused by density difference due
to temperature is termed as free convection or natural convection. The applications of
natural convection as I mentioned are many: they appear in atmospheric process to semi
semiconductor chips, when a semiconductor chip is cooled in absence of a fan which
forces the fluid to move over it, the way it loses heat is natural convection. So, from a
very small length scale to a very large length scale, we will have natural convective heat
transfer under many situations. It is therefore important to know the physics of natural
convection, what are the relations and correlations which are to be used in order to obtain
the heat flux or the heat transfer coefficient in natural convection. And just to have an
idea of the order of magnitude value of the heat transfer coefficient that one would
encounter in natural convection process.
So, let us look at the development of the governing equation and see what additional
information we can obtain by simply writing the equation of energy, equation of motion,
and equation of continuity and see what the similarity parameter would be, or which
dimensionless group would automatically appear in those equations. It is to be noted that
any relation of heat transfer coefficient will therefore, be a function of Nusselt number
besides Reynolds and Prandtl. We would observe that some other number may also
appear for the case of natural convection and the point to start this analysis is by using
the fundamental equations - the energy, the motion, and the continuity equation.
419
So, let us look at the figure over here which shows that I have a plate which is in contact
with air. The plate temperature is T S and the temperature of the surrounding air at a point
far from the plate is T and the corresponding density of air is  . The gravity is acting
vertically downwards; and we will assume that the temperature of the plate is more than
the temperature of the air. Therefore, the temperature of the fluid near the wall will
increase, its density will decrease, and it starts to rise. As it rises, this fluid is going to be
replaced by the cooler fluid from beneath. So, you have a cooler fluid from over here and
then the process would start once again. So, the hot fluid near the wall rises towards the
top to be replaced by cold fluid from the ambient. This is what is known as free
convection and the free convection is classified based on whether the flow is bounded by
a surface or if it is a free boundary flow.
So, let us write the equation; let us assume that the flow due to the density difference in
forced flow inside the boundary layer is laminar. But however, it is also possible that will
have turbulence at some point at a larger value of x. So, this is x = 0, y = 0; that is the
origin that we have and the velocity in the x-direction is u, the velocity in the y direction
is v and we understand that the flow outside of this thin layer is essentially stationary.
So, if I draw the velocity profile of the air which is rising close to the plate, it is going to
be 0 at x = 0 and y = 0; due to the no-slip and no penetration boundary conditions. As we
progress further, we would observe that the profile would be somewhat parabolic with
the maximum velocity occurring somewhere in between. We understand that v is going
420
to be very small, i.e. the velocity in the y-direction is going to be very small as compare
to the velocity in the x direction. Therefore, the x momentum equation is simply going to
u u 1 P  2u
be u v   g y   2 . Note that I have neglected the viscous transport of
x y  x y
momentum in the x direction as the velocity gradient in the x direction is very small as
compared to the velocity gradient in the y direction due to the thinness of the region
u
where this flow is taking place. However, u is large as compared to v, but is very
y
u
large as compared to .
x
So, none of the terms on the left-hand side could be neglected. However, we also know
that if I write the y component of the equation of motion, since there is no force in the y
direction, the gravity is acting downwards. So, y momentum equation would give you
P
 0 . The pressure gradient at any point inside the boundary layer must be equal to
x
the pressure gradient in the region outside the boundary layer. So, the pressure gradient
P
at these 2 points would be identical since  0 and P does not vary in the y direction.
x
If that is the case, then in this region, u is 0. There is no velocity in the x direction, no
velocity in the y direction. The effect of gravity would be there, but the velocity variation
with y is also not going to be present.
So, the equation that I have written over here, the x component of the momentum
equation; if I write it for a region outside of the thermal boundary layer or outside of this
1 P
layer, then all these terms would be 0 except   g y  0 . So, that is the form of the
 x
P
equation if I write it outside of the layer. Therefore,    g y ; and the rho is
x
essentially  .
421
So, if I put this equation back in here, then the left-hand side will remain the same, in the
right-hand side, these 2 terms can therefore be clubbed together to give the equation as
u u g y  2u
u v         2 .
x y  y
Now, this is the term which is the force body force acting on the fluid under natural
convection condition. To express this in a more convenient form, we introduce a
thermodynamic quantity which is the volumetric thermal expansion coefficient. The
volumetric thermal expansion coefficient which is termed as  , and   

  
1 
T P
;
1     
so,  can also be approximated as    . Therefore, this    can be written
  T  T 
as   T  T  . Therefore, the x-direction momentum equation can be written as:
u u  2u
u v  g y   T  T    2 . So, this is the equation of motion where we have
x y y
introduced a thermodynamic parameter  which is known for most of the gases. The
analysis was undertaken because it is better to express the equation in terms of the delta
422
T rather than in terms of delta rho by introducing the concept of volumetric thermal
u v
expansion coefficient. The continuity equation can be written as:   0 ; and the
x y
T T  2T
energy equation can be represented as: u v   2 ; where  is the thermal
x y y
diffusivity.
So, these 3 equations are to be solved simultaneously in order to obtain what is known as
the thermal coefficients for the relations for natural convection. One point I would like to
make here is that, the next step involves non-dimensionalizing the equation. As we have
done before and in order to non-dimensionalize this, we need to define a length scale
which is straight forward. Because, the length scale would most likely be the length of
the plate along which this rise of the liquid is taking place. I also must bring in some
reference velocity which is yet to be decided. I am simply going to write the
dimensionless form of the equation without going through the all the steps, what I would
suggest you is to look at Incropera and Dewitt, the textbook of this course for the
derivations from the dimensional equation to the dimensionless equation, but those steps
are merely substituting the variables in terms of the dimensionless variables and no new
physical concepts are involved. So, we will skip those steps and only write the
dimensionless form of the equation which would give us more insight into the natural
convection process.
423
Therefore, the dimensionless form of the equation is written as:
u * * u* g y   Ts  T  L 1  2u *
u*  v   ; where L is the length scale, u0 is the
x* y* u02 Re L y*2
velocity difference that is yet to be determined (reference velocity). Just see this equation
and compare this equation mentally. What would we have obtained if this is a case where
there is no temperature gradient present in the system? So, if there is no pressure gradient
present and no temperature gradient present in the system then, this equation of motion
would simply revert to the equation that we already know the first term on the right-hand
side would be 0 since Ts  T . So, these 2 terms will remain, and this term will remain as
well, and you can clearly see that the similarity parameter. For such a case, as you have
seen countless times before would be the Reynolds number. So, for a situation in which
there is no temperature difference, natural convection does not take place and the
equation of motion will have the known similarity parameter as Reynolds number only.
However, in presence of a temperature difference between the plate and the surrounding
air I have an additional extra term as the first term on the right-hand side of the of the
momentum equation. So, anything that I define will not only contain the similarity
parameter Reynolds number, but it should also contain something which will consist of
contribution from the first term and again the contribution from the first term is nothing
but the contribution from natural convection.
424
So, let us work a little bit on that term to see if any additional dimensionless terms come
out of the first term on the right-hand side. But in order to do that, if you look at your
class notes once again you have a u02 present in the equation. We know that term u0 is a
velocity that we have not yet defined. So, we need to get rid of this u02 in some way. Let
us multiply both sides of the equation by Re 2 L , which is nothing, but the Reynolds
2
uL
number based on the entire length which would be  0  . Therefore, what you would
  
get is a new dimensionless number which is known as Grashof number, which is denoted
g y   Ts  T  L3
by Gr, and the Grashof number based on the length is defined as: . The
2
Grashof number has the same significance as Reynolds number in forced convection. We
know that for forced convection, the Nusselt number is a function of Reynolds number
and Prandtl number. So, whatever role Reynolds number played in the case of forced
convection, the same role will be played by this Grashof number in the case of natural
convection or free convection. So, the Grashof number can simply be expressed as the
flow due the ratio of 2 forces; the viscous forces in the denominator and the temperature
induced forces in the numerator.
So, the Nusselt number for natural convection should be a function of Reynolds number
a function of Prandtl number and a function of Grashof number. So, unlike forced
convection, we will have an additional term present in any relation or correlation of
natural convection which is Grashof number. So, let us look at Grashof number in a bit
more detail.
425
In natural convection, additionally, this Grashof number if you look at the definition of
Grashof number it is nothing but the ratio of buoyancy forces to the viscous force. So, if
GrL
1 , the combined effects of natural and forced convection are to be considered. On
Re 2L
GrL
the other hand, if  1 then free convection can be neglected. Conversely, if
Re 2L
GrL
 1 then the forced convection effect can be neglected. For such a case, the Nusselt
Re 2L
number would simply be a function of Grashof number and Prandtl number. Many
relations and correlations are available for free convections; I will simply give you some
examples of those relations. For example, the average value of Nusselt number in natural
1
hL 4  GrL  4
convection can be written as: Nu L     g  Pr  . This g(Pr) is a function of
k 3 4 
Prandtl number the values of which are provided in the text. These correlations are quite
complicated, and of course I do not want you to memorize and it is not also possible to
remember all these relations or correlations, but you should know when to use which
relation. So, in any questions in any exam, I am going to provide all these relations to
you, and you must pick the right relation or correlation citing justification and then use it
to solve the problems.
426
I am going to give you some more examples of these relations; so, some other relations
1
1 Pr  4 ; where, Ra is the Rayleigh
for natural convection are: Nu x  Ra 4  
 0.952  Pr 
number, and it is simply Grashof number at x times Prandtl number ( Ra  Grx Pr ).
So, under these situations the turbulence in natural convection is defined when Rayleigh
number at the transition which is Grashof number at the transition times Prandtl number
is about 109 . So, this is the limit for Rayleigh number for the case of turbulent flow. The
most common examples of natural convection is when you have a vertical plate and you
have natural convection taking place from its sides, and the correlation for finding out
what would be the heat transfer coefficient under these conditions are given by Churchill
and Chu where the Nusselt number relation is provided as:
2
 
 
 1 
 0.387 RaL 6 
NuL  0.852  8  . So, this is an impossible correlation to remember
   0.492  916  27 
 1    
   Pr   
  
or to do anything with this. So, what the researchers have done over the years is they
427
have huge number of experimental results under various conditions, and they start with
some logic that Nusselt number has to be a function of Grashof number and Prandtl
number or you can express it as a function of Rayleigh number and Prandtl number, and
the dependence of the Nusselt number on the Rayleigh number or the Prandtl number
these are obtained by fitting the experimental results with a suitable mathematical
function, giving rise to these empirical expressions.
So, there are several such relations available in the text. I am giving you some of the
examples to make sure you use the right correlation to find out the value of the heat
transfer coefficient. All these correlations are valid for certain ranges of Rayleigh
number. For example, turbulent flow and laminar flow the cross over takes place for
Rayleigh number of about 109 and for certain ranges of Prandtl number. So, use those
relations, but use them judiciously for the situation that you have at hand in order to
obtain the value of h. But, the important thing, one of the reassuring things, is that no
matter which correlation you use, the value of h is insensitive to some extent to the
correlation that you have used. So, the value of heat transfer coefficient in natural
convection is very small. For the case of air, it is going to be between 1 to 10 Watt per
meter square per Kelvin. Since the value of heat transfer coefficient is small and it does
not vary too much based on which correlation you use, your results are going to be
approximately correct. But always consult your text or your hand book and choose the
right relation for such cases.
So, the relations that I have shown you so far are valid only when natural conviction is
predominant. But there would be situations in which both the natural convection and
forced convection would be present. So, what would be the expression for such
situations? I know separately how Nusselt number depends on Reynolds and Prandtl, for
the case of forced convection and how Nusselt number depends on Rayleigh number and
Prandtl number for the case of free convection. But, if both are significant, how do I get
the overall Nusselt number for such a situation? So, that is what I am going to write next
the case where both are important.
428
That means, you have both, free and force convection, and both are important. Then the
effective Nusselt number is Nusselt number based on forced convection plus or minus
Nusselt number based on pure natural convection. So, this plus is for the assisting and
transverse flow whereas minus is for opposing flow. So, the Nusselt number for when
both are present is simply an algebraic sum to the power some coefficient n where this n
is generally taken to be equal to 3, and this is only a first approximation to find out what
is the combined heat transfer coefficient when both are present. So, if we have both
acting in the same direction, then it is going to be plus. If you have natural convection
and force convection in the opposite direction, then you are going to have a negative
sign.
The case when there is a constant surface heat flux is a very common situation which one
would encounter in the case of natural convection where the Nusselt number is denoted
1
by a relationship: Nu x  0.6  Grx Pr  5 and this is valid for 105  Grx Pr  1011 , while the
flow is laminar. For the case of turbulent flow, the relationship is going to be:
0.72
Nu x  0.568  Grx Pr  , and this relation is valid for 2 1013  Grx Pr  1016 . My idea is
not to give you too many correlations because these correlations are availability in you
text and you can look it up at any point of time. What I would like you to take away from
this class is that, I need you to understand why natural convection is taking place. And
since it is taking place, the pressure gradient force flow is provided by a difference in
429
density caused by a difference in temperature. So, the body force term will have to be
modified by considering the difference in density. The difference in density is related to
difference in temperature. So, it is more logical to bring in a thermodynamic quantity
beta which is the coefficient of volumetric expansion and write the equation of motion in
a clear fashion which will highlight the role of temperature difference in natural
convective flow along a flat plate.
So, this additional contribution from the buoyancy difference induced flow is clubbed in
terms of a new dimensionless number which is known as Grashof number. So, if it is
purely natural convection, my Nusselt number would be a function of Grashof number
and Prandtl number. If it is a forced convection only then Nusselt number is a function of
Reynolds number and Prandtl number. If both are present, then the effective Nusselt
number is expressed as an algebraic sum of Nusselt number due to forced convection and
Nusselt number due to free convection with an exponent n present in the relation. So, the
values of n are generally taken to be equal to 3 and there are a multitude of relations
which are available for Nusselt number under various conditions. So, the important thing
to know here is that the value of the natural convective heat transfer coefficient is quite
small as compared to its counterparts in forced convection. So, if we have a system in
which forced convection is taking place inside a tube and free convection is taking place
outside of the tube. The rate determining step, which is the one that offers maximum
resistance to heat transfer, is going to be provided by the natural convection. The natural
convective heat transfer coefficient or the inverse of that, because inverse of heat transfer
coefficient is a resistance to heat transfer. Therefore, an idea of the natural convective
heat transfer coefficient is very important in many of the calculations that you would
encounter while designing the heat exchange equipment. So, identifying the flow pattern,
identifying the role of natural convection in that process, and choosing the right
correlation for that specific situation are all very important. The first two I have covered
extensively in this class but choosing the relation or which relation is to be used and
what would be the form of that equation, are all available in the text. So, I would
encourage you to look at the text and see for what condition which correlation is to be
used.
430
Heat Transfer
Lecture – 38
Heat Exchangers
We are going to start something very important in this class. The one of the major
purpose of this course is to know how in an engineering application heat can be
transferred from one fluid to the other. So, the applications of these kind of heat
exchanger are many, when you are trying to cool a process stream before you can
discharge it or when you would like to heat the process stream before it enters the
reactor, and when you have some heat which is going out in a waste stream that you
would like to recover. So, the waste heat recovery all those play a significant part in the
overall design and economics of a chemical plant.
So, efficient design of heat exchange between two process streams is one of the
fundamental requirements of a good design in a chemical plant. The majority of the
equipment that you would see, heat exchange equipment that you would see in a
chemical process plant are heat exchangers.
There the heat exchange is also associated with other unit operations, but in this course
we are concentrating on purely heat exchange equipments. So, two streams how they
come in indirect contact and exchange heat between them. So, if you have a very hot
water available to you, and you would like to heat up cold water which is coming in.
Then what you would do is, instead of mixing them direct mixing you are probably
going to create a barrier in between the hot fluid and the cold fluid and both are moving.
So, if I think of a pipe inside, through the inside of which the cold fluid is moving, and
through the outside the hot fluid is moving with a metal barrier which is a tube wall in
between we have the simplest possible heat exchanger which is aptly called a double
pipe heat exchanger. So, these are two concentric pipes through the annular space one
fluid is flowing, and through the core tube another fluid is flowing both are not mixed,
but they are in thermal contact through the walls of the tube.
431
So, how to design, first of all what kind of heat exchange equipments are available? The
simplest one as I said is a double pipe heat exchanger, but what if I have instead of 1
pipe I would like to have 10 such pipes, what would the flow arrangement look like? So,
there would be one big shell and number of thin pipes small pipes which are going to go
through the shell. One fluid is going to flow on the outside of the tubes in the shell and
another fluid of a different temperature is going to flow through the tubes. So, that is the
shell and tube heat exchanger.
As I said in a double pipe it is just an annular space between two coaxial pipes, go
through them inner pipe some fluid is flowing through the intervening space, annular
space some amount of fluid is flowing. So, both of them enter at one side and leave at the
other side or it may so happen that one is entering from one side, the other is entering
from the other side. So, in one case the two fluids are flowing parallel to each other, in
another case the two fluids are flowing counter to each other.
So, based on flow pattern the heat exchangers can be divided into parallel flow heat
exchanger or counter flow heat exchanger. So, based on the flow direction, we can
categorize the heater exchangers in these two. There would be some situations when you
do not know if it is a pure counter flow or if it is a pure parallel flow. Taking example
from the double pipe heat exchanger, if I have the double pipe, but instead of the double
pipe ending at this point I am going to bend this and then it is going to come in this
direction only. So, the fluid is going to enter through the pipe travel, change its direction
and then will come back.
And the same is going to happen for the annular fluid as well. The annular fluid and the
fluid through the tube both enter at this position and then they flow like this. So, this is
the double pipe heat exchanger with a bend. And I will show you through some of the
pictures that I have what is a heat exchanger, what is the shell and tube heat exchanger,
what are the different components of shell and tube heat exchanger. How can you
classify them based on flow pattern, based on how many tube passes, I will explain what
they are how many tube passes how many shell passes that you have in such a situation.
Designing the heat exchanger is an involved process; there would be a thermal design of
the process and then there would be a mechanical design of the process. So, how much
of pressure head you need to provide in order for the fluid to move from the inlet to the
432
outlet. So, what is the pressure head to be provided by a pump? And when the fluid
enters the shell or tube and when it exits from the shell or the tube there would be some
pressure drop involved. So, the pump that you design should provide you with sufficient
head to overcome the flow resistance inside the heat exchanger at the entry and the exit
points, and to whatever pressure that you want the fluid to be delivered at the exit. So,
sum of all these would let you size your pump or choose the capacity and the power of
the pump. So, that is something which is also very important.
The next thing comes is how are you going to place the tubes, how were the tubes going
to be supported inside the shell, how are they going to be connected to each other
through some common feed line? So, those designs are also going to be very important.
Again when the fluid flows through the shell the outer bigger diameter pipe, you do want
to have turbulence present inside such a system. So, there would be some barriers which
are known as baffles present in the system which would create better mixing, better heat
transfer in this. So, the mechanical design of all these supports where the tubes are
connected, the baffle design etcetera is something which will not be covering in this
course.
In this part of the course we will mostly concentrate on the thermal design part of the
heat exchanger. So, we will talk about what a heat exchanger is, what are the different
types of heat exchanger which are common and how to do the operations of those heat
exchangers and how to design thermally the performance of a heat exchanger.
So, let us look at some of the pictures of heat exchangers that I have obtained from
various sources, those pictures were not drawn by me. So, I have appended at the below
the source from where these pictures were taken, there are many other pictures available
in your text, in the textbook over the internet, and many other sources you can take a
look at it to understand the operation, configuration, flow of fluid in different types of
heat exchanger.
433
So, let us start with the first picture of the heat exchanger that you see over here. So, this
is an U-tube heat exchanger the one which you see on the left. So, the white ones that
you see are the tubes. So, the tubes through this is called a head and this is the tube sheet
where all these tubes are connected. So, you have some amount of fluid, some fluid tube
side fluid which is coming in, it is getting distributed, and will start to flow through these
tubes come at this point change its direction and then flow in the reverse direction, they
will come to the outlet plenum and will be collected from here.
So, this is the in of the tube side fluid and this the out of the tube side fluid. You also
have shell side fluid which is coming into the system and then it is going to change its
direction because of the presence of the baffles. So, baffles are like a disc with one
portion which is cut off. So, the liquid cannot flow through the baffles, they will come
and flow in the small intervening space in between the end of the baffle and the shell.
So, this is the outer one is the shell and the baffles will force the liquid to take a path like
this, take a tortuous path like this, providing good contact, good turbulence and good
values of the heat transfer coefficient. So, the shell side fluid will flow over these tubes
and come out from this side. So, let us say the tube side fluid is at a lower temperature.
So, it enters and the shell side is at a high temperature. So, when it travels, it comes in
contact with the shell side fluid, exchange heat, and you are providing sufficient length
of the tube to ensure that the outlet temperature is significantly higher than the inlet
434
temperature, and when you have a substantial transfer of the energy contained in the
shell side fluid at the inlet, which is transferred to the tube side fluid.
So, the sensible heat increase of the tube side fluid must be equal to the sensible heat
decrease of the shell side fluid. So, this is known as a U-tube heat exchanger. The blue
one shows the path of the shell side fluid and the arrows are showing you the flow path
of the tube side fluid. This is something very common which is known as straight tube
heat exchanger or one pass tube side. One pass tube side means the liquid is going to
pass through the shell and tube through this heat exchanger only once. So, the liquid
comes through the tubes, comes on the other side and so, this is the tube side fluid, in it
goes through the shell only once and are then collected at the exit. The shell side fluid on
the other hand flows in the tortuous path, because of the presence of the baffles and are
collected over here.
So, both the shell side fluid and the tube side fluid, goes through this heat exchanger only
once, they are not recycled to this thing once again which is common what you see in a
two pass tube side. So, what happens in a two pass tube side? You have the inlet plenum
through which the liquid is going to come through the tubes come to this side, and then
will travel once again so, once again through these tubes to the other side.
So, this is the separator which does not allow these two streams mix. So, liquid side, the
tube side liquid starts from here travels in this direction changes this direction, and then
comes to the other half of the tubes over here. The shell again it flows along this path.
So, this is a two pass tubes side shell and tube heat exchanger.
These shell and tube heat exchangers are the most common type of heat exchangers
which you would see in a process industry. So, depending on how many times the tube
side fluid flows through the shell, it can be one pass one tube pass heat exchanger, two
tube pass heat exchanger and so on and the depending on how many times the shell side
fluid flows through the heat exchanger there can be multiple shell pass heat exchangers
as well. So, 2 4; that means, 2 shell pass 4 tube pass. So, these kind of configurations are
also possible.
So, more the number of the passes what you would get is more would be the value of
heat transfer. The value of the temperature of the outlet cold fluid is going to come close
to the outlet hot fluid. So, therefore, it is thermally beneficial to make the residence time
435
longer to make the number of passes available for the fluid to flow such that, the two
temperatures will come very close to each other and more heat transfer can take place,
more total quantity of heat transfer can take place between the hot fluid and the cold
fluid. But everything comes with a price; the price that you pay for this better heat
transfer is the amount of pressure gradient required to or the pressure head required to
make the liquid flow multiple times in a tortuous fashion.
So, the pump that you would have for multi pass heat exchangers, it has to have higher
capacity which is expensive. So, you make a judgment that whether you are going to use
a two pass heat exchanger rather than a four pass heat exchanger and decide, which one
is going to be ultimately, it is going to be governed by economics. So, whether the
additional cost of the pump is offsetting the gain that you get in a better heat transfer, this
is a calculation which you have to make.
So, we will concentrate in this course on the thermal part of the heat, the only the heat
exchanger part, but the mechanical design is also equally important, for which there is
dedicated design course in any chemical engineering curriculum which you are going to
do where you are going to calculate the pump power needed, how to design the tube
sheet, how to design the baffles and how to make a long tube stay more or less
horizontally it should not sag.
So, tie rods are provided to make them to keep them straight. So, all those what would be
the shell diameter, for what to depend on the pressure, what should be the thickness, how
you are going to attach the heads to the shell and these what are the mechanisms which
you have to provide based on the pressure inside the shell or inside the tube those are
important parameters that one has to keep in mind.
Additionally, you will have some other problem to deal with that is the problem of
scaling, the formation of scales during prolonged operation in any heat exchange device.
So, the scales are formed both on the inside and on the outside of the tubes, and they are
a strong function of what kind of liquid you are using and what is your material of
construction of the tube.
So, your choice of material of construction of the tube will be dictated by what fluid at
what temperature, you are going to handle. Is it a corrosive fluid? Is it a fluid that may
form that may react with the tube material? Is it something which may contain suspended
436
particles, which will form scales during prolonged runs. These are the considerations
which you have to make in deciding which fluid is going to be on which side, shell side
or tube side. What is going to be the material of construction?
And whenever such a scale is formed on the outside or on the inside of the tube, they
provide additional resistance to heat transfer. So, the resistances to heat transfer, if it is a
perfectly clean tube it is going to be 1/h i where hi is the convective heat transfer
coefficient based on the inside of the tube, 1/ho, where ho is the outside heat transfer
coefficient and the conduction resistance of the material of construction of the tube
considering the thickness of the tube. But these scales which are going to form they
would provide additional resistances.
So, your design equation should take into account what is the maximum value of scaling
resistance due to scale, it is also known as dirt factor that you have to take into account to
ensure that your designed heat exchanger is going to perform as per these specifications.
So, you need to over design, to take into account the scales which will definitely form
during operation.
And what is generally done in heat exchangers is, after a certain length of time which is
to be decided which is to be made known to the user by the designer or the fabricator of
the heat exchange equipment is what is the time after which you need to service the
equipment. And servicing one major part of servicing the equipment is to clean the inside
and the outside such that this heat exchanger after cleaning will have a performance as if
it is a new tube and there is no resistance to heat transfer due to the formation of these
scales, dots and so on.
So, the run time, the total run time of the heat exchanger before it should undergo
maintenance, must also be specified. So, there are several factors which have to be kept
in mind while designing a heat exchanger. So, let us quickly look at some of the other
types of heat exchangers, we and how the liquid will flow through a heat exchanger.
437
So, if you look at the next one, this is a simulation of the flow of energy through the shell
and the tube. So, you have some hot fluid; some hot fluid coming over here from the
shell side, it changes its direction and then it goes out. Whereas that the core, the another
fluid is flowing through the tube it changes its direction and starts moving in this way.
This picture is probably more clear on this side, where you have the entry of the shell
then it goes like this travels in after finally, it comes out of this. So, these equipments are
quite complex to design, to fabricate, to maintain, to clean and so on, but they form an
essential part of any chemical process plant.
438
The, this is the simplest possible heat exchanger which is a double pipe heat exchanger.
So, you have a flow which is taking place through a double pipe and through the annular
another fluid is moving and therefore, you get a very good heat transfer in such a case.
This is interesting conceptually, if you look at the figure carefully then you have a flow
situation in which this is the direction of flow of the cold fluid, and the direction of flow
for the hot fluid is also the same. The two arrows denote the direction of flow of the two
fluids and the hot fluid temperature is going to come down, this is the exit temperature of
the hot fluid. The cold fluid temperature is going to increase, as it flows along the heat
exchanger and this one is the outlet temperature of the cold fluid.
So, you can easily see that the delta T between the hot fluid and the cold fluid, which is
the prime cause of heat transfer is going to decrease significantly as the outlet is
approached. And one more point to look at is that the outlet temperature of the cold fluid
can never be more than the outlet temperature of the hot fluid. So, there is some sort of a
barrier which the cold fluid can never cross. If you provide infinite length of the heat
exchanger then the cold fluid temperature, outside outlet temperature will be equal to the
hot fluid outlet temperature, but it can never cross it.
So, this type of arrangement in which both the hot and the cold fluids flow parallelly as
has been shown over here in a double pipe heat exchanger are known as parallel flow
heat exchanger. So, these two are parallel flow; when they flow in the opposite direction.
439
So, the cold fluid, this is the cold fluid inlet it flows in this direction, this is the cold fluid
outlet. This is the hot fluid inlet and it moves in the reverse direction and this is the hot
fluid outlet.
Here you see that throughout the length of the heat exchanger you maintain a sizable
significant appreciable T between the hot and the cold fluid, which was not the case
for parallel flow. So, for counter flow a flow in opposite direction the T for the same
inlet and outlet temperatures, the T is going to be more for the case of counter flow as
compared to that of the parallel flow. So, counter flow heat exchangers are definitely
more efficient thermally as compared to the parallel flow.
One more thing that we need to see here is that, due to the nature the direction of the
flow the cold fluid outlet temperature can exceed the hot fluid outlet temperature, which
was not possible for the case of parallel flow. The cold fluid outlet temperature can be
more than the hot fluid outlet temperature therefore; you get better heat transfer for the
case of counter flow as compared to that of parallel flow. So, you will see how to
calculate the effective heat transfer, effective T for the case of parallel and counter
flow and is it going to be some sort of an arithmetic average or is it going to be some
other type of average, which would give me the effective T between the hot and the
cold fluid for these two situations. So, we will look into that as well.
440
This is the another animation of the heat exchanger. So, this as you can see the cold fluid
denoted by the blue, and the hot fluid denoted by the red and how the temperature of the
this blue slowly it turns and it goes towards the red. So, that is another example of how
the temperature of the cold fluid is getting heated by while exchanging heat with the hot
fluid.
This is an actual heat exchanger and you can have an idea of the size of this heat
exchanger how large it can be, and the different parts of the heat exchanger are shown
here. So, you have the nozzles through which the liquid is going to come into the heat
exchanger, this outer one is the shell and since a significant change in temperature
increase or decrease is associated, you need to provide an expansion joint to absorb that
change in length, and there are some vents or drains which are added in the heat
exchanger since it is going to be big and large and it is going to be a lot of liquid will
flow through it. So, there is a significant weight of it and it needs to be supported
appropriately on the factory floor. So, this is just a picture of how it would look like.
And these are the baffles that you see, the baffles with part of it cut out which would
allow the liquid to flow. So, these are the flow direction would be over this in this
direction and so on. So, these are called segmental cut baffles from a circular disc, a part
of it is cut out for passage of liquid, you have the tie rods and spacers which would allow
the tubes to remain straight and so on.
441
So, this is another view of the tube sheet, where all the tubes are connected and these are
the pictures of the head where these are channel and removable covers and different
types of front heads, which are associated which are attached to the heat exchanger to so,
such that the liquid can change its direction. And the shell can the head of the shell, the
front heads, they can they can take the additional pressure and so on.
So, these are the different types of the rear heads. So, you can look at it and you can get
an idea of how it would look like.
442
And these are some of the types of shells which we commonly encounter in the design.
So, this is one pass shell the liquid is going to come in and it is going to move out of this.
This is a two pass shell with longitudinal baffle so the liquid is going to come shell side
in this, it will travel and then it will change its direction and will go out. In some cases a
split flow this baffle what you see here instead of being attached to one end it is right at
the middle.
So, the flow pattern would be here they need to travel in both directions, and again come
and meet the other part of the flow and will go outside. So, this is an example of a split
flow; similarly you may have double split flow, a split flow you may have a divided flow
where the liquid is going to come over here and then it is going to divide into two and
you can have instead of just an one pass you may have a cross flow of the shell side fluid
over the tubes, which are laid perpendicularly to the direction of the flow.
443
So, the large number of the very large number of configurations for these types of flows
are possible in heat exchanger design, and what you see here are the nozzles and flange
which are used to attach the tube and the head to the actual shell and create a heat
exchange equipment. So, it is a very important piece of equipment, and we would like to
see how to express, how to design, how to find out the heat transfer performance, the
total quantity of heat transfer, the length to be provided, the number of tubes to be
provided and so on; based on a comprehensive calculation on out of the understanding of
internal and external flow that we have studied so far.
So, in your next class we would concentrate on the design the actual calculations of heat
exchange in a double pipe as well as in shell and tube heat exchangers, what kind of
temperature difference to take based on what are the characteristics of parallel flow that
of counter flow and in some cases you may have partly counter flow, partly parallel flow
in one section of the heat exchanger when the tube side liquid travels in this direction
and the shell side liquid is coming in this fashion then it is counter flow.
But at the end the tubes turn and that there the tube side liquid then flows in this
direction, the same direction as that of the shell flow. So, for half of the section the flow
is counter flow, in the bottom half the flow is parallel flow. So, how to handle those
situations, what kind of temperature difference we should make. Do we have to take
arithmetic average or some other type of average, and what is the condition for us to take
444
arithmetic average because that would be the simplest possible averaging. All these
issues we will look from next class onwards.
445
Heat Transfer
Lecture - 39
Heat Exchangers (Contd.)
A quick recapitulation of what we have seen in the last class, we have seen the design
and the flow characteristics in a heat exchanger. The simplest possible form of a heat
exchanger would be a double pipe concentric heat exchanger, in which one liquid will
flow through the tube and the other will flow through the annular space. The more
common and slightly complicated design comes with shell and tube heat exchanger. In a
shell and tube heat exchanger, there will be one shell and multiple number of tubes
which are connected to a tube sheet and the fluid is going to be through those tubes and
they will come from the other side and on the shell side, the liquid is going to be at cross
flow over the surface of the tubes. So, the flow through the tube is going to be straight,
whereas the flow in the shell side is made tortuous purposefully; to enhance turbulence
and to increase the heat transfer coefficient. And we have also seen that there can be
multiple tube passes and multiple shell passes. So, one shell, two tube; two shell, four
tube passes are quite common. The purpose of using multiple passes is to ensure that the
residence time for a specific fluid is going to be more inside the heat exchanger, allowing
better heat transfer such that the overall heat transfer efficiency and the overall quantity
of heat transfer between the cold and the hot fluid can be increased.
So, that is why we have multiple passes present in some of the heat exchangers, but any
increase in number of passes will also enhance the pressure drop and pressure drop plays
a very critical role in deciding whether to use multiple passes, the required area, the
diameter of the pipe etc. So, the pressure-drop and the enhanced heat transfer, these two
should be balanced keeping in mind the overall cost of heat transfer. So, increase in heat
transfer by increasing turbulence would result in an increase in pressure drop.
Furthermore, we have also seen that depending on the direction of flow, the flow of the
hot and the cold fluid may be parallel or maybe counter flow. So, we have parallel flow
heat exchangers and counter flow heat exchangers. In some of the cases part of the flow
is going to be parallel, whereas part of the flow is going to be counter. We will have
446
situations when mixed flow also prevails. So, depending on the direction of flow it can
be parallel, counter and mixed flow.
The basic construction of the shell and tube heat exchanger also includes the baffles,
which are obstruction to the flow path, which makes the fluid to move in a zigzag pattern
in a tortuous path inside the shell with the aim of increasing the heat transfer coefficient.
The tubes are connected by a tube sheet and the heat exchanger is closed by different
types of heads and the design of those heads, tube sheet, baffles essentially complete our
understanding of the heat exchange equipment, their efficiency and their performance.
And as mentioned previously, the pressure-drop place a critical role in deciding all these
factors. The other part that I spend some time on is about fouling; with prolonged use in
between maintenance, there is going to be some deposition or some reaction between the
liquid that is flowing and the inside of the tube surface or the outside of the tube surface
which is exposed to the shell side fluid. The formation of these scales is referred to as the
fouling phenomenon, which essentially increases the heat transfer resistance. So, if I did
not have any fouling then there would be three resistances in series; one is the convective
heat transfer coefficient for the inside of the tube, convective heat transfer coefficient for
the outside of the tube, and the thermal resistance of the tube material which also would
depend on the thickness. If we have a fouling which is the realistic situation then, two
more resistances are to be added to these three resistances. One is due to the fouling
inside of the tube and the other is fouling due to fouling on the outside of the tube.
447
So, the overall heat transfer coefficient generally denoted by U is the inverse of the sum
of all these resistances. That is to mean, U is the overall heat transfer coefficient, and the
1
inverse of that is a resistance to heat transfer, so should be the sum of all these five
U
resistances in series. So, in engineering design it is important to know what the overall
heat transfer coefficient of a system is, in order to ascertain the length of the heat
exchanger required to drive a specific amount of heat transfer between the hot and the
cold fluid. So, this overall heat transfer coefficient is very important, and in most of the
design equations or in the thermal design exercise of heat exchange calculations, we end
up finding out what is the value of the heat transfer coefficient.
So, I am going to show you the relationship of the overall heat transfer coefficient with
1 1
, the convective heat transfer resistance based on the inside; , the convective heat
hi ho
l
transfer resistance based on the outside; , the conduction resistance, and the two
kA
ll ll 1
fouling resistances: R f ,i and R f ,o . So, if you look over here this that is this U is the
UA
overall heat transfer coefficient, which you can define based on either the inside area A i,
where Ai   Di L or you can express it in terms of an external area, Ao   Do L . So,

these two are the inside surface area of the tube and the outside surface area of the tube.
So, U can be defined based on the inside or based on the outside area. So, the suffix of A
in this term would determine how do you express U in terms of internal area or in terms
of external surface area of the tube. The first one is the convective resistance, this is the
ll
conductive resistance, and this is the convective resistance on the outside. Whereas, R f ,i
ll
is the contact resistance for the inside surface, and R f ,o is the contact resistance for the
ll
outside surface and this R f is known as the fouling factor. The values of the fouling
factor are known for different types of fluids, and they would definitely depend on the
time for which this heat transfer equipment is in use and so on but, the knowledge of this
R llf for a specific fluid is essential in order to find out the overall heat transfer
coefficient. For many instances a significant part of the heat transfer resistance comes
448
from the contact resistance. I am now going to show you some of the typical values of
the heat transfer coefficients that you would expect in real systems.
So, some of the values of U, which would give some idea of U, that are generally
encountered this U is in unit Watt per meter square Kelvin.
So, from water to water, the value is about 850 to 1700 Watt per meter square, then
water to oil is slightly lesser. The important one here is steam condensing on the outside
of the tube, while water is inside the tube; this is an important value which is about 1000
to 6000.
So, look at this number, this is the largest of the three that, I have shown so far. So,
whenever steam condenses let us say, I have a tube through which water is flowing and if
it is saturated steam at a certain pressure is condensing on the outside. So, when it does,
the temperature of this surface is going to be constant. Not only the temperature of the
surface is going to be constant, which would correlate to the pressure of the saturated
steam that you have it is also going to maintain a constant temperature and a very high
value of heat transfer coefficient.
So, the value of heat transfer coefficient is significantly larger than the others and
therefore, the steam condensing is going to give you a high value of heat transfer
coefficient. So, the other one is finned tube heat exchanger, which is may be water in
449
tubes and air on the outside air in cross flow. So, for that case it is going to be only 25 to
50. As you can see, whenever you have a gas you are going to have heat transfer from a
liquid to a gas or in any process which involves heat transfer to gas, the value of the heat
transfer coefficient is going to be significantly smaller than that between liquid to liquid.
So, between liquid to liquid water to oil or water to water is 1500 to 1000. Whereas for
finned tube heat exchangers, where the tubes itself will have fins like the ones that you
see in the radiator of a car. The heat transfer coefficient is only 25 to 50. Compared to
that the natural convection heat transfer coefficient is going to be of the order of 10 only.
So, looking at the different values of the heat transfer coefficient, you then have some
idea that you would always like to have a situation with a high value of heat transfer
coefficient.
Sometimes you may have to use natural convection, you will encounter natural
convection in many of the applications and whenever you have from water to air the heat
transfer to a gas use of fins may be justified, but even with the use of fins you would
only get from 25 to 50 Watt per meter square Kelvin, significantly lower than that when
two liquids are interacting thermally and exchanging heat in between. So, this is just a
rough idea of the values of U, the overall heat transfer coefficient that you generally
encounter in real systems. Now let us look at how to calculate the overall heat transfer
coefficient and in order to find out the overall heat transfer coefficient, we first must
define a temperature difference. Because when you think of a tube the liquid enters it at a
certain temperature and then goes out at a certain temperature, the total amount of heat

gained, let us say it is simply going to be m C p  Toutlet  Tinlet  .
Similarly, the liquid, which is flowing outside of the tube with different mass flow rate is

also going to gain or lose some heat and m C p  Toutlet  Tinlet  would give you the amount
of heat that the outside fluid transfers to the inside fluid. However, this is based on
overall energy balance of each of the streams separately, but these two streams are also
interacting thermally across the tube wall, where it faces three resistances or five
resistances including the thermal resistance as, we have discussed before. So, the amount
of heat that gets transferred is going to be:
450
 Do 
1 1 1 1 1 Ri',0 Ri'',0 ln  Di 
      
UA U i Ai U o Ao hi Ai ho Ao Ai Ao 2 kL
where, U is the overall heat transfer coefficient, A is the surface area and delta T is some
sort of a temperature difference. So, as you move in a specific direction, the value of
delta T would be different at different points, and whether you have parallel flow, or you
have counter flow the value and the behavior of the change of delta T with position
would also be different. So, a judicious choice for delta T has to be evaluated in order to
use the equation that mC p T of one fluid is equal to mC p T of the second fluid, is
equal to UAT . What that delta T is going to be? That is what we are going to find out
in our next exercise.
So, in order to do that, we are going to start with an enthalpy balance. The enthalpy
 
   
balance would simply give me q  m h ihi  iho  mc ici  ico , where this is the fluid
enthalpy and i is for the inlet. So, this is for the outlet and this i is for the inlet. In other
 
   
words, you can express it as: q  m h C ph Th  Th  m c C pc Tc  Tc . The temperatures
i o i o
that I had used here are mean fluid temperatures at a given location, it is not a point value
it is just the area averaged value and I also know that q  UATm . So, this Tm is the
451
temperature difference. I am going to do this for parallel flow heat exchanger and for
counter flow heat exchanger. So, let us first start with parallel flow heat exchanger.
So, for parallel flow the profile would look something like this is the x and this is T. So,
this is temperature of the hot fluid at the outlet, temperature of the cold fluid at the inlet,
temperature of the cold fluid at the outlet.
So, this is the temperature difference at the inlet and this is the temperature difference at
the at the outlet and, if I take a small slice. So, here the temperature difference let us call
it this as delta T and some amount of heat is going to come from the hot fluid to the cold
fluid as in the direction that is shown, and this difference is the dTc , change in
temperature of the cold fluid and this is dTh change in temperature of the hot fluid as we
move in this direction. So, over this infinitesimal length, the temperature of the cold fluid
increases and the temperature of the hot fluid decreases. So, dq for this can be written as:

dq   m h C ph dTh . Since the energy is reducing I have put a minus sign here and for the

cold fluid, it is going to be dq   mc C pc dTc and this equations you can simply integrate
to obtain the total heat, which is across the entire length of the heat exchanger as

 
qc  mc C pc Tci  Tco and similarly you can write it in terms of the hot fluid also in which
452

 
case it is going to be qh  m h C ph Th  Th . Additionally, you also have d q to be:
i o
dq  UdTdA this is the amount of heat that comes from the hot fluid to the cold fluid.
So, this T at this point is T h minus Tc, the temperature of the hot fluid minus
temperature of the cold fluid and from this I can simply write, d T is dTh minus dTc
since T is Th - Tc. So, d T can be written as:
 1 1   1 1 
d T  dq     U TdA    c.
 C h Cc   Ch Cc 
2 2
d T  1 1 
So, I bring this T over here and therefore, I have   U     dA . Since,
1
T  Ch Cc  1
the heat transfer coefficient, Ch, Cc all these are constant and therefore, they can be kept
outside of the integration sign. So, that is a simple expression which can be obtained and
once you integrate this with the locational values incorporate it in there.
What you are going to get is simply
T2  T T T T  UA
ln  UA  hi ho  co ci   Thi  Tho    Tco  Tci   . For parallel flow
T1  q q  q 
453
T2  T1
q  UA
T1  Thi  Tci and T2  Tho  Tco . Therefore, T  , where this term is
ln  2 
 T1
simply the log mean temperature difference. So, the final form of this equation is going
to be q  UATlm .
So, what I see here is that the overall heat transfer coefficient is related to the total
transfer of energy, this is the surface area and the relevant delta T to be used is not the
arithmetic mean or any other mean the equation simply tells you that you have to use the
logarithmic mean temperature difference. However, if I will simply look at delta T 2 by
delta T 1 and find out what is delta T arithmetic divided by delta T log mean what you
see is when it is 1 obviously, there is no difference when it is 1.2 this is 1.0028, but when
it is 1.5 it starts to affect your results and with increase in this ratio of delta T 2 by delta
T 1 the departure of delta T arithmetic from delta T l m starts to become significant.
So, as long as the ratio of delta T 2 by delta T 1 this ratio is within 1.5, it really does not
matter if you are using an arithmetic mean or a log mean, but for any change in
temperature ratio delta T 2 by delta T 1 beyond 1.5, it is always advisable to use the log
mean temperature difference rather than simple arithmetic temperature difference
between the inlet and the outlet.
So, the heat transfer relations are to be calculated based on log mean temperature
difference, not the simple arithmetic mean of the simple temperature difference. So, this
we have done for the case of parallel flow.
Now, let us look at how it is going to behave for the case of counter flow.
454
In counter flow what you see is the hot fluid flows in this direction the cold fluid flows in
the reverse direction. So, what you have over here is T1  Thi  Tco and T2  Tho  Tci
and as before, we again take a small slice of this heat exchanger where the change in the
hot side temperature is d T h and the cold side temperature is simply d T c and the heat
flows in this direction is d q.
Note that in the case of counter flow heat exchanger the outlet temperature of the cold
fluid can exceed the outlet temperature of the hot fluid, which was not possible for the
case of parallel flow. So, that is a specific advantage characteristics of counter flow. So,
here also as before you can write in the same way that the total amount of heat transfer is
simply going to be q  UATlm , but in this case T1 is simply going to be Thi  Tco
whereas T2 would be Tho  Tci . And as I told you before, counter flow is going to be
more efficient than the parallel flow, when the temperatures and the flow rates are kept
unchanged.
So, three of the special cases that we would examine here is the first case is the capacity
of the hot fluid is very large as compared to the capacity of the cold fluid. So, in this case
this is how the temperature profile would look like. So, this is for the cold fluid and this
is for the hot fluid and the hot fluid temperature does not change at all when C h is very
large as compared to Cc. So, this in this case this straight line is when C h tends to infinity;
that means, let us say for a condensing vapor. So, for a condensing vapor the
455
temperature, as I mentioned before will remain constant which for a condensing vapor
the Ch will tend to infinity.
So, the reverse is when you have Cc is very large as compared to C h. So, this is the hot
fluid this is the cold fluid and this specific situation will arise when it is going to be
straight line will arise when you have an evaporating liquid.
So, for an evaporating liquid as long as its temperature is going to remain constant and
this is how it would look like. And for this special case where C c is going to be equal to
Ch both these lines parallel and counter flow this you would see that T1  T2 . So,
everywhere in the flow path between the hot and the cold fluid the T will remain
constant.
So, the take away from this is for both parallel flow and counter flow this q can be
expressed as q  UAT , whereas the special case is if the capacity Ch is much more than
Cc which is the case for condensing vapors there is almost no change in temperature on
the hot side of the fluid which is obviously, going to depend on what is the temperature
and pressure of the steam which is condensing. On the other hand, if your C c the capacity
of the cold fluid is much more than that of the hot fluid then therefore, you are you are
not going to experience any change in the cold side temperature and for the special case
when the same fluid is flowing inside and outside of the tubes, the C c and Ch are same.
So, the difference between the cold fluid and the hot fluid temperature at any given
location is constant. So, whatever you have a temperature difference you have at the inlet
will be the temperature difference at the outlet and any point in between.
So, we now know the basic fundamentals of what temperature difference to be used and
what kind of behavior we will have depending on the capacity of the fluid in parallel
flow and in counter flow, but not all flows are purely parallel or purely counter flow
there are mixed flows as well. So, in the next class we will quickly take a look at the
mixed flow cases.
456
Heat Transfer
Indian Institute of Technology Kharagpur
Lecture - 40
Heat Exchangers (Contd.)
So, we have seen the cross flow and the parallel flow and how to calculate the log mean
temperature difference. We have also seen at for special cases where the heat where the
capacity of the hot side it tends to infinity, for example, for the case of a condensing
vapor the temperature profile of the hot side fluid remains constant whereas the cold side
fluid it increases. Conversely when the cold side fluid temperature has a very high
capacity, in that case the cold fluid temperature will remain constant whereas the hot
fluid temperature decreases.
So, these 2 separate cases where C h is very large as compared to C c or Cc is very large as
compared to Ch that we have seen how would the profile look like. We have also seen
how the 2 profiles would look like when C h is equal to Cc that is the capacity of the hot
fluid is the same as that of the cold fluid. In which case the temperature profiles for the
hot fluid and for the cold fluid they will simply be parallel lines straight parallel lines
where the  T will remain constant throughout.
Now, as I mentioned that the heat exchangers can be multi pass heat exchangers or the
heat exchangers could be such in which you are going to have cross flow across the tube
bundles. So, when you have a multi pass heat exchanger you may start with parallel
flow, but in the second pass it is going to be counter flow and so on and when in cross
flow the flow is not going to you cannot say that it is parallel flow or counter flow it is
rather the flow is across the flow of the flow, whatever liquid that is flowing through the
tube and the SHELL side fluid is going across.
So, all these different types of flow which cannot be clubbed as purely counter flow or
purely flow in the same direction are called as mixed flow. So, what is to be done to
LMTD in order to take into account the multi for multi pass heat exchangers situations
which are common in multi pass heat exchangers and in cross flow heat exchangers. So,
what is done is a correction factor to LMTD has been introduced which is commonly
457
denoted by F. So, by this factor F the LMTD has to be corrected before you can use it in
the relation, q  UATLM .
So, the revised formula for cross flow and for multipurpose heat multi pass heat
exchangers would simply turned out to be q  UAF TLM where, F is the correction factor
This correction factor F would depend on various things, it is going to depend on what is
the temperature of the SHELL side, the temperature of the tube side both for the hot fluid
and the cold fluid and the results for a large number of situations are expressed
graphically, from which it would be possible for us to find out what is the value of F for
a specific flow situation.
So, the correction factor F evaluated from the graph used in conjunction with LMTD will
give us the corrected LMTD that can then be used in the relation q  UAT . So, I am
going to show you how this F varies with these parameters and the graphs are available
in your text. So, whenever you while designing heat exchangers whenever you come
across this kind of mixed flow situations always look for the correct value of F to be used
in the LMTD relation.
So, let us take a look at what we have over here. So, for the case of correction what we
need to do is we need to correct TLM  F TLM based on pure counter flow. So, this is
important you first calculate TLM as if the flow is taking place in counter flow ok.
458
So, you calculate TLM as if the flow is taking place in counter flow multiply that with F
in order to obtain TLM for the case of mixed flow. So, this is for mixed flow conditions
this is for counter flow and this F is the correction factor which is to be used for this. So,
this is what we have is Graphical Representation of the factor F and it is going to look
something like this.
There will be a family of curves and these the value of P is defined in this way and the
Ti  To
family of curves are for different values of R , where, R  So, this temperature T
to  ti
refers to the shell side temperature and t refers to the tube side temperature, o for the
inlet, i for the outlet.
So, P if you can see P ranges so this is from 0 to 1, P rangers from 0 to 1, so this
represents the thermal effectiveness of the tube side fluid and R ranges from 0 to infinity,
0 corresponds to pure vapor. So, if it is a pure vapor then you can see that t o is going to
be equal to ti. So, it is a pure vapor or it is a condensation on shell side and infinity to
evaporation on the tube side.
So, when you read it carefully once you would see the significances which are obvious
based on how we have defined P and R. So, the important part here is you calculate what
are the different temperatures and then from the different temperatures you find out what
is the value of F and use that value of F to read to correct the log mean temperature
difference that you would have obtained if it is pure counter flow.
So, pure counter flow log mean temperature difference corrected by the factor F would
give you the log mean temperature difference to be used in the relation q  UAT so that
is all in that expression and we are going to solve problems to show you how to use
LMTD in real situations for actual practical problems. But we would do that in the next
class, in the remaining time of this class to discuss I am going to discuss about the
simplest possible heat exchanger that you would come across which is a double pipe heat
exchanger.
So, I will talk about the description and the design considerations of a double pipe heat
exchanger and then once that is over then we are going to solve there will be a tutorial
459
class, where we are going to solve problems on LMTD and problems on double pipe heat
exchanger before we move on to the design of shell and tube heat exchangers.
So, what is the double prime heat exchanger, a double prime heat exchanger consists of 2
concentric tubes of different diameter. So, there would be I mean the tube inside tube is
going to be circular in cross section, whereas the space between the between the inside
tube and the outside tube is going to be annular in nature. So, there would be flow
through the tube and flow through the annular space may be in the same direction,
maybe in opposite direction that is in parallel flow or in counter flow and these tubes are
going to be the going to be attached to the main pipe through certain glands, heads and
so on.
Now, if you think of the design of these they are the easiest to design ok, they are easy to
fabricate. The only limitation is that it is it is going to be expensive if you want to have a
double pipe heat exchanger that bends and comes back once again. So, the tube in the
tube the fluid flows in a one direction and comes back in the reverse comes back in the
reverse direction and you have annular flow in there. So, whenever you have a bend like
that you are going to introduce pressure drops, the other limitation of double pipe heat
exchanger is that it only provides small values of the available area for heat transfer.
So, the heat transfer heat transfer area provided by a double pipe heat exchanger is
generally small. So, wherever you have a small duty heat transfer situation you can
prescribe double pipe heat exchanger, but the moment heat transfer area requirement
increases the double pipe heat exchanger will not be economical anymore.
It is also difficult to clean the double pipe heat exchanger to get rid of the deposits the
dirt and that would deposit that would deposit on the inside and outside of the tube
during elongated periods extended periods of action. Finally, the double pipe heat
exchanger is limited by the length you do not get too big too long a pipe to make a
double pipe heat exchanger.
So, double pipe heat exchangers are of generally of the size of 12 feet 20 feet 25 feet at
the most and those are going to be used for fabricating a double pipe heat exchanger. So,
double pipe heat exchangers will only be cost effective if your heat duty is small and you
also have to think another disadvantage of double pipe heat exchanger is that it requires a
460
large space as compared to a shell and tube heat exchanger which is much more
compact.
So, you space comes at a premium so therefore use a double pipe heat exchanger is not
justified unless the heat load requirement is small and you have enough space to
accommodate a double pipe heat exchanger. However, they are the simplest want to
fabricate simplest type of heat exchangers to analyze.
So, you quickly go through and they are they are used in industry at a number of number
of places number of situations. So, we need to know what is double pipe will quickly go
through the double pipe there are design issues and certain some of the calculations and
then move on to shell and tube. So, let us first draw a picture of the double pipe heat
exchangers.
As I said the length could be between 12, 15 and 20 feet. So, this is what a double pipe
heat exchanger would look like, the one that you see in the green is the inside tube this is
the outside tube and this is the inner area through which the liquid flows. So, the double
pipes are connected to the base by glands. So, this is one gland this is another gland and
you have a T junction here which is where the fluid is going to come, which is going to
flow through the. So, the hot fluid or let us say the shell side fluid is going to come
through this, the shell side is going to flow in this direction and then it is going to be
collected at the other end.
461
So, they are not mixed with what is happening inside the tube side fluid and the tube side
fluid it is a counter of let us assume it is a counter flow heat exchanger it is going to
move in this direction and to outside. So, this is what the simplest heat exchanger that
you can think of.
But it is only used if area required for heat transfer is small of the order of 100 to 200
feet square that is the area which is required and the coefficients the film coefficient that
means, the hi or ho can be calculated. If it is a laminar flow then the Sieder-Tate
correlation that is Nusselt number based on diameter is
1/3 0.14
 Re.Pr    
Nu D  1.86    
 L / D   s 
If it is turbulent , I am not going to use this. So, it is going to be Dittus-Boelter equation,

Dittus-Boelter correlation for Sieder-Tate and we understand that is Sieder-Tate equation
is to be used when you have large property variation
So, when the temperature of the tube at the temperature of the tube if your viscosity
where is significantly then you are going to use Sieder-Tate relation, otherwise you use
the Dittus-Boelter equation that we have described before that we have. But for
completeness I am going to include it here as well, where
Nu D  0.023Re 4/5 Pr1/ n
This n can be 4 if Ts is greater than Tm and n can be 3 if your Ts is less than Tm and for
the case of Sieder-Tate, the Nusselt number is simply going
0.14
4/5  
1/3
Nu D  0.029 Re Pr  
 s 
So, this is when you have a significant variation, in the case of in the case when your
significant variation in the property. But whatever I have described here they refer to
flow inside the tube. So what is to be done when you flow in the annuli? The common
practice for that is you find out what is the what is the hydraulic diameter for the annuli
and then use all these equations depending on whether you have laminar flow, you have
turbulent flow, you have significant variation in property. In all those correlations are
462
equally applicable for the case of annulus annuli as well provided you replace the
diameter by the hydraulic diameter then all these relations can be used.
So, to in order to find the for the case for the annuli, you find out the hydraulic diameter
4 AC
which is Dh and is defined as Dh  . So, Ac is the cross sectional area for flow of
P
the liquid and P is the wetted perimeter. So, this P is wetted perimeter of the annulus
annular region and Ac is a cross sectional area which is available for flow. So, D h is
therefore,

4
4
 Do2  Di 2 
Dh   Do  Di
  Do  Di 
So, for turbulent flow therefore this can be used in terms of it by simply replacing D h Di
in Reynolds number. So, in Reynolds number replace D with the hydraulic diameter for
the Annular region and you take care of the fouling factors as before when you have
fouling present in this
1 1
  Rdi  Rdo
U D UC
463
So, this is the clean coefficient which is which depends on ho and hi, this is the
coefficient while the double pipe heat exchanger is in operation it is sometimes called as
the design or overall coefficient. So, this is the design or the overall coefficient and this
is the clean coefficient and these 2 together are sometimes replaced sometimes expressed
as Rd where Rd is dirt factor. So, this dirt factor the clean coefficient together they would
give us the overall coefficient which is to be used for all our design purposes and one
more thing that we need to need to keep in mind is this clean overall coefficient is simply
1 1
 and the heat transfer coefficient on the in this is simply going to be that is
ho hi
Ai ID
directly from geometry and it is hio  hi  hi .
A OD
And finally what you get once you get all these? You are going to get q  U D AT .
Remember that in the case of double pipe heat exchangers there is no need to calculate
the correction factor F for LMTD. Because for double pipe heat exchangers it is either
purely counter flow or purely parallel flow, there is no question of mixed flow for the
case of double pipe heat exchangers. So, the only thing which remains right now is to
how is to express the pressure drop or to calculate the pressure drop and how it is related
to the flow rate of the fluid flowing in the annulus as well as flowing in the tube. Well
this flowing in the tube we know how to calculate the pressure drop.
So, the same equation, the same value of friction factor that we have obtained the Blasius
friction factor or the friction factor which you have obtained from the Moody diagram
can be used for the tube and the same can be used for the annular region as well. If we
replace the diameter to calculate the Reynolds number with the hydraulic diameter of the
system, so the pressure drops are generally calculated using the standard relation of  P
that we have used for the case of Moody diagram friction factor that we have obtained
from the moody diagram. So, I will simply write those expressions for you, but it is
already known to you.
464
16
So, the pressure drop in pipe and annuli can simply be obtained, if it is f  , if it is
Re
laminar
0.125
If it is turbulent it is, f  0.0014  0.32 .
 Re 
So, this is from so this is for the case of laminar flow, this is for the case turbulent flow,
you can use either this correlation or you can simply use Moody diagram and for annuli.
So, this is for the tube or pipe, for annuli replace D by De ' that is the effective diameter
to obtain the friction factor F.
So, this is this is the hydraulic equivalent diameter that we have obtained and
4 fG 2 L
P 
2 g  2 De '
Where, G is simply the mass velocity. So, G is defined as the mass velocity and
generally a P of 5 to 10 psi is provided in double pipes. So, to once again to calculate
the pressure drop we can use the same relations that we have used for calculating the
pressure drop in a pipe. For the case of annuli we replace the diameter by the hydraulic
diameter use the same relations and the pressure drop can be expressed the standard
465
formula that we have done in terms of the velocity or it can be expressed in terms of the
mass velocity and the you would see later on that the heat transfer coefficient when a use
the relations the heat transfer coefficient is going to be proportional to G to the power
something I think G to the power 0.8.
So, when you increase the value of G the mass velocity of the fluid flowing through the
tube or through the annulus you are simply going to increase h, but when you look at the
pressure drop calculations, the pressure drop calculation increases with G in as G2. So, an
increase in h with G will be compensated by the requirement of higher pressure. So,
there is an economic decision which has to be made on how much of pressure drop you
can sustain you can provide and what is the enhancement in terms of the energy transfer
that you are getting out in the process. So, its in engineering calculation based on the cost
estimates which one has to perform to decide what is the flow rate to be used in a double
pipe heat exchanger. The customary is to provide about 5 to 10 psi pressure drop allow
that much of pressure drop for flow in a double pipe.
Now, whenever you have some entry point, the entry point there would be some losses at
the exit there would be some losses and if you are going to provide a bend at the end. So,
that it is going to look like in U tube heat exchanger instead of a double pipe heat
exchanger you are going to add some more pressure drop.
So, all these calculations are required in order to make a judicious decision in the design
of a double pipe heat exchanger, which is easy to fabricate, expensive to maintain only
economical if the area requirement is small and if you have if you do not have any
problem of space, because these double pipe heat exchangers they take up significant
amount of space if you compare that with the heat transfer area provided by them. So, for
all practical cases whenever you have a high surface area transfer area requirement and
you do not have enough space to put a bulky double pipe heat exchanger one should
always go for shell and tube heat exchanger which we are going to discuss in the next
class.
466
Heat Transfer
Lecture – 41
Tutorial Problems on Heat Exchanger Calculations
This is going to be a tutorial class on what we have studied in the last few classes. What
we have done in those classes? We have defined what is going to be the log mean
temperature difference, what is a counter flow and a parallel flow heat exchanger, what
is a double pipe heat exchanger its advantages and disadvantages.
And while calculating the log mean temperature difference, we understood that if there is
mixed flow that is a flow type which cannot be categorized either as purely parallel flow
or purely counter flow which in many situations. For example, whenever you have
multiple tube passes in a shell and tube heat exchanger or even in a double pipe heat
exchanger, if the flow is across the tubes. So, for all those cases which were somewhere
in between the parallel flow and counter flow which have collectively termed as mixed
flow a collection factor to LMTD has to be used.
The correction factor is generally denoted by f and so, the LMTD for mixed flow would
be f times the LMTD for counter flow. So, the log mean temperature difference was
derived, it came automatically as the correct mean to be used for the temperature
difference between the inlet and the outlet. And we have also seen that if the temperature
difference, the temperature of the change of the hot fluid and the temperature change of
the cold fluid are the difference is not substantial then arithmetic average can be
substituted for LMTD.
However, in order to maintain uniformity throughout our discussion we will always use
LMTD for all our calculations because, that is the correct average logarithmic average
that is a correct average to be used for the temperature difference. We have also seen
what a shell and tube heat exchanger will look like with multiple number of tubes and in
some cases there can be multiple shell passes as well.
So, there can be 1-2 heat exchanger, there is one shell pass that is the shell fluid is going
to flow through the shell only once. And two tubes passes where there will be flow
467
through the shell in the upper half and then through the flow through the tubes in the
upper half of the shell and then it changes its direction and flows back through the shell
once again.
So, the tube side fluid enters the shell, leaves the shell, comes back to the shell and then
leaves once again. So, 1-2, 2-4 and similar such multi-pass heat exchangers are quite
common. And whenever you have those kinds of heat exchangers, the LMTD has to be
corrected by the incorporation of the correction factor denoted by f.
And this f is to be read from a graph which I have shown that the value of f can be
calculated, can be read from the graph where the two non dimensional temperatures are
defined. And as a function of those two non dimensional temperatures, these value of f
can be read from the graph which is provided in any text book on heat transfer.
So, with the background, I think we are in a position right now, to solve for some
specific problems in this class. And then from next class we will go into purely shell and
tube heat exchangers, their efficiency calculation and so on.
But, right now let us concentrate on the first problem; which is a double pipe heat
exchanger it is a thin walled double pipe heat exchanger I have underlined the thin
walled these two words. So, as to mean that, this essentially means that the resistance to
468
heat transfer by the wall of the pipe is negligible as compared to that of the inside
convection coefficient and outside convection coefficient.
I forgot to mention that we have also introduced the overall heat transfer coefficient U
which is just the algebraic sum of the resistances provided by the inside convection
coefficient of the tube. The thermal resistance, conduction resistance of the tube and the
resistance which is provided by the outside convection from the outside of the tube.
So, normally it is a sum the 1/U which is the overall heat transfer coefficient would be
1 1 1
 
U hi ho
Where, hi and ho are the inside and the outside convection coefficients. And another term
which is due to the resistance of conduction through the tube wall. To this we sometimes
have to add the dirt factors or the deposition on the inside as well as on the outside of the
tube which will provide additional resistance to heat transfer. So, those dirt factors are to
be as they act as additional resistances and they are to be added to this relation in order to
obtain the overall heat transfer coefficient which is denoted by U.
So, in this specific problem it is a thin walled double pipe heat exchanger. So, we do not
need to consider the conduction resistance, and nothing has been mentioned about the
dirt factors. So, we will assume that these are relatively new or just serviced and
therefore, the heat transfer resistance due to these dirt factors they do not exist. So, it is
going to be sum of only two resistances; one is the inside convection coefficient and the
other is outside convection coefficient.
So, engine oil is to be cooled from a high temperature to a low temperature and the
coolant used is water, the flow rates of the, if you see the flow rates for each of them.
That means, for the oil as well as the water are provided to be 2 kg per second, the
engine oil temperature is to be reduced from 160 to 60 ºC, the ID is 0.5 meter the OD is
unimportant as we have mentioned its thin walled.
So, had this not been the case then the OD of the tube would have to be provided with
the materials specified such that, we can find out what is the conduction resistance of the
pipe? but that is not relevant over here. It is also mentioned the overall heat transfer
coefficient is 250 (W/m2K), what you need to find out is what is the length of the heat
469
exchanger? The length of the double pipe heat exchanger to be provided so, that this heat
transfer process is possible.
In order to do that first we need to find out what are the properties and the properties can
be found from your text book. So, if I consider oil, the average temperature is the
average of these two. And at that temperature the value of the heat the value of the
specific heat can be read from the table in your text or will be provided for the coolant
water. We do not know what is the cooler, we do not know what is the mean temperature
is going to be.
So, we are going to assume some an average T m average water temperature let us assume
this to be 40 ºC, it is not going to matter much you would see towards the end. And the
coolant water is the coolant water is at 25 ºC so; obviously, the temperature is going to
be more outlet temperature is going to be more than 25. So, I assume some T m which can
later be checked with the actual value that you are going to get. So, this is going to give
me a value of Cp for the cold water as 4179.
First of all I can find out what is the total heat transfer which is taking place.
 
& ph Thi  Tho  2kg / s  2260 J / kgK  (160  60) K  4.5  105W
q  mC
So, a simple sensible heat change there is no phase change which is involved. So, this
should give me the value of q.
470
Now, when I do this use this value of key q and apply it for the case of cold fluid, this q
is also going to be

q  m&cC pc Tco  Tci 
where Tco is the temperature of the cold fluid at the outlet and Tci is the temperature of the
cold fluid at the inlet. So, since this is known to be 25, all these values are known as
well.
So, the only unknown Tco can be evaluated from this as and when you plug in the
numbers, they are going to be this we have evaluated just previously, and this would be
79 ºC. And now, one thing I would like to bring to your notice over here is the outlet
temperature of the engine oil is 60 ºC. Whereas, the outlet temperature of water the
coolant water is 79 ºC now, what does this tell us. Here we see that the outlet
temperature of the cold fluid exceeds the outlet temperature of the hot fluid. And in our
previous discussion we have seen that if it is a case of parallel flow, the outlet
temperature of the hot fluid can never exceed the outlet temperature of the hot fluid.
Only if the flow is in parallel mode it is possible to have outlet temperature of the cold
fluid exceeding the outlet temperature of the hot fluid.
I will draw those profiles once again so, that we are perfectly clear in our concept about
why this has to be a parallel a counter flow situation. So, if you remember this, if this is
471
parallel flow the temperature of the hot fluid and if this is x and the cold fluid would be
something like this, both are flowing in the same direction. So, the outlet temperature of
the so, this is the cold and this is the hot fluid outlet temperature of the cold fluid cannot
exceed the outlet temperature of the hot fluid.
However, so, this is parallel flow; however, if you have this kind of a one in which one
flows these two are flowing in counter flow fashion, then the outlet temperature of the
cold fluid may exceed the outlet temperature of the hot fluid. So, this is the hot fluid this
is the cold fluid.
Then here we have the outlet temperature of the cold fluid is 79 whereas, T ho is 60
Kelvin. So, since Tco is greater than Tho, this must be a case of counter flow, this in
information is important since we are going to use this information in finding what is
going to be the Tlm . So, Tlm for this case we are considering this case is going to be
Tlm 
Thi  
 Tco  Tho  Tci 
 Th  Tco 
ln  i 
 Th  Tc
 o i 
This would be always will be different if it is a parallel flow situation. So, that is why it
is important to know whether it is counter flow or parallel flow, and this is simply going
to be the ratio of these two.
So, once you put in the values over here so, this is going to be
81  35
Tlm   55
81 ºC
ln
35
So, your length is going to be
q 4.5 105
L   21m
U  DTlm 250    0.5  55
This comes from the formula that
q  UATlm
A   DL
472
So, we have seen a couple of considerations while making, while solving this problem.
One is that very important is you need to find out whether it is parallel flow or counter
flow and calculate Tlm based on whether it is parallel or in counter flow hence; first is
all the properties are to be evaluated at the mean temperature.
So, I have assumed the mean temperature to be 40 and when I calculate the temperature
of Tco here as about 80 then, I see that we exceed the my assumed mean temperature to
some extent. So if I want to further refine by result I need to use this value of the
temperature and recalculate the value of Cpc and do this calculation once again in. So,
that I get a more accurate result of l, but since the value of Cp of water is not a strong
function of temperature this would give me an acceptable value.
So, always use the property at the average temperature make sure whether you have
laminar flow or counter flow and then use q  UATlm by choosing the right expression
for Tlm and find out the length or anything else which is unknown. So, in some cases
may be L would be known you have to find out U so, that is the first problem which we
have solved in this class.
The second problem is a somewhat different so, this is a shell and tube heat exchanger
furthermore it is 1-2 heat exchanger. So, it is one shell pass and two tube passes. So, if I
just consider the tube, the shell side fluid is going to come in from here, flow in a
473
tortuous path and go outside, but the tube side if they there is a divide over here. So, the
fluid enters through this, passes through the tubes, changes it is direction it and then
travels in this and is collected from here.
So, this is 1-2 heat exchanger whereas, the tube side liquid is going to be in the shell
twice whereas, the shell side fluid passes only once. So, it is a 1-2 heat exchanger, there
are 30000 tubes and the wall is thin and since the wall is thin it means the conduction
resistance can be neglected, the diameter of each tube is 25 millimeter, the steam is
condensing on the outside of the tube. So, here we have steam condensing on the tubes.
And so, there is the entire portion of the tube is going to be maintained at a constant
temperature, since; it is a condensing steam the value of h is 11000 W/m 2. The q, the heat
rate should be 2×109 Watts. So, that much of heat is to be removed, the cooling water
rates through all these tubes is 3×104 kg/s.
So, if you consider water rate through each tube so, therefore, through each tube it would
simply be three into 104 kg/s divided by the number of tubes is equal to 30000. So,
therefore, though each tube it is going to be 1 kg/s, the water enters at 20 ºC and steam
on the outside condenses at 50 ºC.
So, the first thing that we have to find out is temperature of the T co, the outlet
temperature of the coolant liquid; that means, what is the T at water outlet? The inlet is
20 what is the outlet temperature and the second part is what is the length required which
is L let us say per pass. So, what is the length of the tube that is to be provided per pass
so, as to effect this kind of a transport process. The properties which given are for water
4179 J/kgK, the value of viscosity is provided 10 -6N/sm2, the thermal conductivity is
0.613 W/mK and the value of Prandtl number is 5.83.
So, this is a classic problem of 1-2 heat exchanger, large number of tubes thin; that
means, no conduction resistance diameter is provided steam is condensing since, steam is
condensing the temperature on the outside of the tube is going to be a constant which as
mentioned is 50 ºC. The value of the heat transfer is coefficient is quite high on the
outside. The heat to be removed is specified the cooling water rate is provided from the
cooling water rate by dividing the cooling water rate to the by the number of tubes, I can
find out what is the rate of water flow through each of the tubes.
474
I need to find out what is Tco. Temperature of the cooling water at the outlet and what is
the length required per pass? The physical properties the relevant physical properties are
provided here.
The one first part is straight forward; the first part is what is the first part is what is T co
temperature of the outlet? So, this q the heat to be removed is

q  m&cC pc Tco  Tci 
So, therefore,
q& 2  109
Tco  Tci   20   36 ºC
m&c C pc 3 104  4179
Since, we are making an overall heat balance. So, the total flow rate is to be considered
not the flow rate through each of the tubes.
So, when you do that this Tco would come out to be 36 ºC, that is a 1 st part. In order to do
this 2nd part I am going to use this relation
q  UAF TlmCF
475
This is what I was telling you that the situation that we have over here is not pure counter
flow or is not pure parallel flow for this part, if we assume that the flow is taking place
like this. So, this is going to be both of them the water and the steam they are in the same
direction whereas the water changes its direction when coming back.
So, it is not a case of pure counter flow or for a case of pure parallel flow. Therefore, this
factor f is to be used and this factor f when multiplied by Tlm based on counter flow
would give me the actual heat actual temperature difference that is to be used in this
formula. So, one of the task of this problem is to find out the value of F, but before that I
find out what is the heat transfer area
A = N2LπD
N is the number of tubes and LπD would be the area of one tube to be multiplied by N in
order to ensure that I have taken all those 30000 tubes in my calculation.
And since, it is going to be twice its going to go through the shell twice a factor two,
must be added in here. So, LπD is the flow area of one tube N L π D is the flow area of N
number of tubes and since the tube tubes are going through the shell twice that is why the
factor 2 is to be included to correctly account for the heat transfer area. And the overall
heat transfer coefficient is simply going to be
1
U
1 1

hi ho
1 1 1
 
U hi ho
So, this value is provided to be 10 something some value is provided for h o as 11000 W/
m2, I need to find out what is the this value. So, what is h i is to be found out. In order to
do this, the first thing that one needs to do is find out what is the Reynolds number based
on diameter.
476
So, Reynolds number based on diameter to find out whether it is a laminar flow or
turbulent flow
4m& 4 1kg / s
Re    59570
 D   0.025  855 106 Ns / m2
Hence it is definitely a turbulent flow.
So, if it is a turbulent flow then, we can simply use Dittus-Boelter equation for finding
out what is Nusselt number which is
Nu D  0.023Re D 4/5 Pr 0.4  308
This Prandtl number (Pr) is provided as 5.83 in the statement of the problem. So, once
you do all these, the Nusselt term is going to be equal to 308 and hi would therefore be
k
hi  Nu D  7552W / m 2 K
D
So, this hi would turn out to be 7552 W/m 2K. Therefore, the overall heat transfer
coefficient,
477
1
U
1 1

hi ho
1
U  4478W / m 2 K
1 1

7552 11000
So, in the formula that we have just said q  UAF TlmCF , I know what is the value of U
now, the value of A is known, the area is not yet known the value of q is known, the this
part is yet to be obtained. So, the next job is find out what is what is the value of F? Now
if you remember, if you see your class notes the F is expressed as a function of two
dimensionless numbers, two dimensionless temperature differences
t o  ti
P
Ti  ti
Ti  To
R
to  ti
So, this capital T refers to the shell side fluid, small t refers to the tube side fluid and
when you put in the put in the values in here, if you look at the statement of the of the
problem, this the this would be
to  ti 36  20
P   0.53
Ti  ti 50  20
Ti  To
The other factor which is defined as R  where T refers to the shell side fluid and
to  ti
t refers to the tube side fluid. Now if you see for the case of steam condensing on the
outside the steam is going to condense on the tubes at a constant temperature and it is
mentioned that the steam is condensing at 50 ºC.
So, therefore, the value of Ti and T0 are equal to 50 ºC which essentially gives you the
value of R to be equal to 0. Now what you need to do is you have to see what is the
corresponding value of F from the figure.
478
And from the figure you would see that where F is expressed as a function for of P and R
from the from that figure for a value of P equals 0.53, but for a value of R to be equal to
0, this would give your value of F to be equal to 1.
If F is equal to 1, then the length from the expression that q  UAF TlmCF , this A is
simply going to be equal to
A=(2πDL)n
2πDL is the area of one pipe multiplied by the area of one pipe considering it is spending
2 times, it is a 2 pass heat exchanger. So, 2 pass since it is a 2 shell 2 tube pass heat
exchanger that is why these 2 comes in.
So, the total area available for heat transfer is 2πDLn. The value of U is known. So, the
only unknown that we have in here is
TlmCF 
T
hi  
 Tco  Tho  Tci   (50  36)  (50  20)  21
 Th  Tco  (50  36) ºC
ln  i  ln
 Th  Tc (50  20)
 o i 
Whenever you have this problem it is always better to draw the two temperatures. So,
this is the hot side and this is the cold side, but in this specific problem the hot side
479
temperature remains fixed at 50 ºC whereas, the cold side temperature increases from 20
to about 36. So, therefore TlmCF would turn out to be 21 ºC.
So, the value of this is known, the value of F is equal to 1 and the value of q is provided
so, the total length is going to be
q 2 109
L   4.51m
UN 2 DF TlmCF 4478  30000  2  0.025 1 21
So, when you calculate this the value of L is going to be equal to 4.51 meter
So, in order to have this heat transfer process you have to have 30000 pipes in a 1,2 shell
and tube heat exchanger. So that means, 1 shell pass and 2 tube passes with steam
condensing on the outside of the tubes at a temperature of 50 ºC or rather maintaining a
temperature of 50 ºC throughout the length of the tube with the heat transfer coefficients
specified as 11000. Now first you calculate what is a Reynolds number. Through
Reynolds number of flow through each of these tube that being that will turn out to be
turbulent and when this turbulent flow takes place through each of the tube I can use the
standard relation to find out what is the heat Nusselt number and what is h i, the heat
So, the overall heat transfer coefficient can then be calculated neglecting any dirt factor
or neglecting any conduction resistance. So, that U part is done, the next part is to find
out the right value of the temperature difference. The log mean temperature difference
since it is a mixed flow case has to be modified by the incorporation of the factor F. So,
we calculate F using the standard method and we see that the value of F turning out to be
equal to 1.
So, next step is to find out what is the log mean temperature difference based on pure
counter flow. And there we see that the shell side temperature remains constant whereas,
the cooling temperature, cooling water temperature increases from 20 to about 36. find
out the log mean temperature difference.
So, using the formula q  UAF TlmCF you would be able to find out when you express
the area in terms of the diameter, the perimeter and the length you would be able to find
out the length to be used in this shell and tube heat exchanger. So, this problem is a nice
480
example of the use of LMTD of the evaluation of h 0 and then finally, using it through a
simple heat balance to find out the one unknown, in this case the length of the pipes to be
used for this case.
So, in the next class, we will see how we can design to go into more depths of the design
of shell and tube heat exchanger.
481
Heat Transfer
Lecture – 42
Tutorial Problems on LMTD and Dirt Factor
We will continue with more problem, tutorial problem on Log Mean Temperature
Difference, shell and tube heat exchangers, and Dirt Factors. So, the specific problem
that we are going to deal with a heat exchanger with the process streams, that is the inlet
fluid and the outlet fluid are specified. The temperatures at the inlet and outlet are also
known, the flow rates are known and the shell and tube heat exchanger provided is a 1-4
heat exchanger; that means, 1 shell pass and 4 tube passes.
The properties are also provided the heat transfer coefficient on the outside; that means,
on the shell side is provided, and more importantly a value of the dirt factor. There is a
resistance due to the deposition of materials on the inside, as well as on the outside of the
tubes which would provide a significant amount of additional heat transfer resistance that
is mentioned. So, the specific heat exchanger that is designed has some dirt factor, it
must provide a value of a dirt factor. It should be, it should work with a dirt factor that is
provided in the statement of the problem.
What that means, is that if our calculations show, that a dirt factor, that we obtain from
our calculation is more than the dirt factor which is to be provided in the design of it,
then the system should work, because the system can handle a higher value of dirt factor
than the specified one. So, the heat transfer or the heat exchanger would work. So, the
problem I have described it over here and then I will slowly go through the problem, the
solution, and some of the concepts which would automatically appear out of the solution
of this problem.
So, this problem is about log mean temperature difference, multi pass heat exchanger,
the value of dirt factor is provided; we need to calculate the dirt factor for the specific
situation at hand. And, then make a decision about the suitability of the heat exchanger
for the specific purpose.
482
So, let us look at the problem over here. So, it is a problem on log mean temperature
difference and dirt factor. So, what we have is a shell and tube heat exchanger so, the
outside one that we see is the shell and the green line which you see, so, this is the shell
and in the shell side 20000 kg per hour of kerosene at 200 ºC is entering, which is being
cooled to 94 ºC.
So, the mass flow rate of kerosene and the  T, Tinlet and Toutlet that is provided, it is a 1-4
heat exchanger, so, the tube side fluid enters, completes the second pass, this is the third
pass and then comes out after it crosses the shell 4 times. So, that is why it is the 1-4
shell and tube heat exchanger; that means, 1 shell pass and 4 tube passes.
On the tube side we have crude oil, which is a tube side liquid which enters at 38 ºC and
leaves at 77 ºC. So, this is 38 it leaves at 77 and the shell side is at 200, and it leaves at
94. It has been mentioned that the heat transfer coefficient on the shell side which is h o is
920 W/m2K. There are a total of 160 tubes; that means, 40 tubes per pass.
So, that is a number of tubes per pass which is present in the shell and tube heat
exchanger. The, pipes are the tubes are 4.88 meter long, the internal diameter is provided
and the total heat transfer area of these tubes is given as 61.5 m2.
The properties of crude oil, which are provided, the viscosity, the specific heat and the
Prandtl number is given. So, it has been mentioned that this shell and tube heat
483
exchanger should provide a combined dirt factor of this much 5.28 × 10 -4 So, based on
the data which are provided in here we have to make a judgment of is the heat exchanger
suitable for this operation.
That means, these conditions do they require, do they can they handle a dirt factor even
higher than this or a dirt factor lower than this. So, if calculations based on these data
provide an overall dirt factor which is less than the dirt factor which is provided by the
heat exchanger, then we have a problem. Then, it cannot handle it would not be able to
handle the specific operation.
So, we have to first calculate what is the log mean temperature difference. So, once we
calculate the log mean temperature difference we should also be able to calculate, what is
the overall heat transfer coefficient on the tube side based on the Reynolds number and
Prandtl number of the crude oil that is provided.
So, we have the LMTD, we have the heat transfer coefficient. So, on the inside and the
heat transfer coefficient on the outside is provided. So, combining these two together,
which is 1/hi and 1/ho, heat transfer coefficient on the inside as well as on the outside
would give me the value of Uc, or overall heat transfer coefficient, which is called also
known as without any dirt, which is also known as the clean coefficient.
So, we have a clean coefficient now and this clean coefficient would have to be used in
order to find: what is the R d, the dirt factor for this specific case? Now, for the dirt factor
one has to equate q that is the total amount of heat that is
q  UATLM
This U that we would use is the U based on the dirty condition present inside the heat
exchanger.
So, this T first the log mean temperature difference first will have to be modified to
make it applicable for multi pass heat exchangers. So, the LMTD based on pure counter
flow has to be converted into the corrected LMTD by the incorporation of the factor F,
the correction factor that we have discussed before.
484
& p T . The m&for the tube side fluid is known to us, C p of the
So, q  UAT and q  mC
& p T all for the

tube side fluid and T of the tube side fluid is also known to us. So, mC
tube side must be equal to UD, where UD is the value of the dirty condition overall heat
transfer coefficient times A, A being the flow cross section, A being the heat transfer
area. This would give me the value of U D; because that is the only unknown in the
equation, then comparing UC and UD we should be able to arrive at the value of the dirt
factor.
So, let us do the step by step calculation and see how we can arrive at the value of the
dirt factor, with this basic operation demands and whether or not what would be it is
value relative to the dirt factor to be provided by the heat exchanger. And, only then we
can make a judgment of whether this heat exchanger is possible to use this heat
exchanger for the transfer that we have just described. So, let us start with the calculation
of LMTD first.
So, when we look at the solution of this the things that we have provided first is the value
of ho is 920 W/m2K. The thermal conductivity, k of the crude oil would be from the
Prandtl number, from the known value of the Prandtl number the value of k would
simply be 2.0516.
485
C p 2.0516 103  3.596 103
So, k    0.133W / mK
Pr 53.36
Now, I need to calculate LMTD. And, since it is a multi-pass heat exchanger, a corrected
value of LMTD to be is to be used. So, the correct in order to obtain the corrected value I
have to multiply the LMTD by the factor F, which we have described in order to do that
this LMTD is to be calculated based on counter flow. So, if we go back to this figure
once again, based on counter flow then what we have is as if the whole fluid is moving in
this direction.
Instead of these 4 passes I just have 1 pass, where the crude oil enters from this side at 38
and it leaves on the other side at 77. And, the shell side fluid comes at 200 and flows
counter to the fluid and leaves at 94. So, the modified figure for this would be this being
the heat exchanger. The shell side fluid comes at 200 it leaves at 94 and the tube side
fluid comes at 38 and leaves at 77. So, this is a pure counter flow representation of the
previous figure. So, therefore, the LMTD would simply be
(200  77)  (94  38)

( LMTD)CF   85.15
(200  77) ºC
ln
(94  38)
And, since it is a 1-4 exchanger, 1-4 heat exchanger, so correction needed for LMTD.
And, in order to calculate the correction, in order to use the results which are available in
your text, in the form of figures we need to calculate P, which is
t2  t1 77  38
P   0.24
T1  t1 200  38
Similarly, one can calculate the other parameter in the graph for finding out the
correction factor for LMTD.
T1  T2 200  94
R   2.72
t2  t1 77  38
Once, you use these two and use the from graph you would see that the correction factor
for the correction factor for the temperature would be equals 0.85. So, in that case
TACTUAL  ( LMTD)CF  FT  85.15  0.85  72.4 C
486
So, this is going to be the corrected value of the, corrected value of the heat transfer
corrected value of the temperature difference. So, this part is over.
And, now we have to see what is going to be the I know h o to be equals 920 W/m 2K,
what is going to be the value of h i. So, that is my next step. And, in order to do that I am
going to use the relation which I have we have seen many times the Dittus Boelter
equation,
Nu  0.023Re D 0.8 Pr 0.4
So, in order to calculate Reynolds number I understand that I have 40 pipes per pass. So,
if I equate, if I try to find out what is the mass flow rate through all these 40 pipes in a
pass, which is ρ times area cross sectional area of 1 pipe.
So,
 (0.0204)2
  N  u  m&COLD =67600/3600 kg/s
4
Once again the mass flow rate of the cold fluid can be calculated based on cross sectional
area multiplied by u, which would give me volumetric flow rate through 1 tube, multiply
with N the number of tubes per pass, this is going to be the volumetric area of this, is
going to be the volumetric flow rate of the cold fluid entering into the heat exchanger,
487
when I multiply it with ρ this becomes the mass flow rate of the cold fluid entering into
the shell and tube heat exchanger.
And, this we know to be equals 67600/3600 kg/s as in the problem statement. It is

mentioned that the crude oil enter at 600 kg/hour. So, that is the flow rate of the crude oil
which enters the shell and tube heat exchanger since there are 1-4 passes.
So, if I consider there is only be 40 tubes which are available for this much of crude oil
to flow. So, what I am trying to do in this case is trying to find out what is this unknown
velocity u for flow through the pipes? So, everything here is known I will be able to
calculate what is the value of uρ
u   1436.26kg / m 2 s
Du 
why I am keeping it as uρ since my Reynolds number is going to be So,

Du  0.0204 1436.26
Re    8148
 3.596 103
So, therefore, this is definitely in the turbulent range. And, since in the turbulent range
that that is why the Dittus Boelter equation can be used and it is a problem of heating.
So, that is why Prandtl number heating of the tube side fluid. So, therefore, the Prandtl
number has the exponent equal to 0.4, otherwise it should be equal to 0.3. So, with the
value of Reynolds number and with Nusselt number to be equals this from Dittus Boelter
equation this would give you the value of Nusselt number to be equals 154.
Nu  0.023Re D 0.8 Pr 0.4  154
And, this would give you the value of the heat transfer coefficient on the inside to be
154  0.133
hi   1006W / m 2 K
0.0204
So, this is this is another important result that I know what is the value of my hi.
So,
488
1 1 1
 
U c hi ho
U c  480.5W / m2 k
So, this is nothing, but the clean heat transfer coefficient.
So, once I know the heat transfer clean heat transfer coefficient, then the next step is to
find what is the heat transfer coefficient for a heat exchanger, which has been in use for
quite some time such that the dirt factor resistance, resistance due to the dirt deposited on
the tube and on the outside of the tube must be taken into account.
So, in order to calculate the dirty coefficient I know that my value of the heat transfer
area provided is 61.5. So, UD the dirt heat transfer coefficient is
67600
 2.05 103  (77  38)
Q =337.4 W/m2K
UD   3600
AT 61.5  72.38
& p T for the cold as well as for the hot, but we are going to use the cold
Q is simply mC
& p T for the hot side as well.

over here it does not matter. So, this is must be equal to mC
So, we are we are using the cold side in this specific case, but we could have used the hot
side as well.
489
So, from our previous calculation we have found UC to be equals 480.5 W/m2K the point
to note here. Obviously, for obvious reasons the heat transfer coefficient for the cold side
for the clean operation is going to be significantly more than the heat transfer coefficient
when the entire system is in operation for some time. So, the clean coefficient, when
everything is clean and there is no dirt, no dirt factor, no resistance due to reposition is
going to be more than that of the dirty coefficient when we have to take this into account.
So, the relative values of the dirt factor, relative values of the clean coefficient and that
of the dirty coefficient would give us some idea of what this dirt factor is going to be.
We also understand that this dirt factor is can be treated as a resistance, which simply
gets added to the overall resistance due to the clean coefficient. So, the clean coefficient
consists only of hi and ho provided we assume that the conduction resistance through the
tubes is small.
So,
1 1
  Rd
U D UC
So, knowing the 2 value of the clean coefficient and the dirty coefficient one should be
able to find out, what is the value of the dirt factor for this specific operation and then
compare that with whatever dirt factor the exchangers must provide as per it is design.
So, our next step them simply is to find out what is the value of the dirt factor. So, let us
see what this value would be and your Rdcalculated for this specific purpose.
UC  U D
Rd CALCULATED 
U CU D
Rd CALCULATED  8.8 104
Rd REQD  5.2 104

Rd CALCULATED  Rd REQD
You can see that this Rdcalculated is greater than the Rdrequired, which is the required one. So,
therefore, what it shows is that the calculated Rd, which comes from the stable operation
of the heat exchanger is more than the designed one.
490
So, this heat exchanger can then definitely handle a situation, which is more than the R d
which has been provided for this heat exchanger. So, that means, in other words the heat
exchanger can then handle dirtier fluids as compared to the designed one. So, we have
obtained that we have calculated the dirt factor based on the stable operation of the heat
pipe, heat exchanger, including kerosene and crude oil and what we see is the combined
dirt factor for that operation is more than the designed one.
So, the heat exchanger operates perfectly even when its dirt factor is more than the
designed one. So, the conclusion of this problem would be yes this is heat exchanger can
be used for the designed heat exchange heat designed dirt factor and beyond that.
So, the problem then therefore, gives us an idea of the log mean temperature difference,
the corrected values of the log mean temperature difference, evaluation of the Reynolds
number, choosing of the correlation to be used for calculation of h i, the inside heat
transfer coefficient, combining ho and hi to give us the value of the clean coefficient.
From the process data calculate, what should the Rd be the dirt factor would be the dirt
factor is simply going to be a combination of the clean coefficient and the dirty
coefficient and these then can be used to decide about the applicability of the heat
exchanger for a specific operation.
So, this concludes our study on log mean temperature difference and it is used in heat
exchanger relate heat exchanger calculations, but the log mean temperature difference if
you think once again requires the 4 temperatures to be known. Specially, the inlet and the
outlet temperature of the hot fluid and the inlet and the outlet temperature of the cold
fluid must be known to us.
If one is unknown then through heat balance we will probably be able to reduce that, but
all 4 temperatures need to be known in order to proceed for heat exchanger design
utilizing LMTD, because the first step in heat exchanger design is to find the log mean
temperature difference. So, if I do not know the 4 temperatures I cannot proceed. So, this
under many circumstances, we would know only the inlet temperatures the outlet
temperatures are not known to us.
So, the procedure based on LMTD would not be straight forward to use would not be
able to use LMTD method. So, we need to find an alternative method to go for the design
491
of heat exchangers, where the temperatures of the 4 streams are not known, what that
would be, what that approach would be we will see in the next class.
492
Heat Transfer
Lecture – 43
Epsilon – NTU Method – 1
So far, we have used the LMTD method for heat exchanger calculations and heat
exchanger designs, but LMTD would require the knowledge of all the four temperatures
if the inlet temperatures are on are only known and the outlet temperatures are not known
use of LMTD would require an iterative procedure.
So, in order to avoid the iterative procedure a separate method a different method that
does not require the knowledge of the outlet temperatures as well is proposed and is
being used which is known as the effectiveness NTU method. So, we have to first define
what is effectiveness and what is NTU? NTU stands for Number of Transfer Units. So,
what this number of transfer units mean we first have to define.
So, when you think of the effectiveness first, let us try to visualize a system in which a
hot fluid is entering and its temperature is changing from T h i which is the temperature
of the hot fluid on the at the inlet and T h o where the it is the temperature of the hot
fluid at the outlet. So, when you think of the overall heat balance the q which the hot
fluid has lost and which has been transferred to the cold fluid the hot fluid the loss of
energy of the hot fluid can simply be expressed as m dot h C p h multiplied by T h inlet
minus T h outlet.
Similarly, the same amount of heat since it is being used up by the cold fluid. So, the
cold the heat balance written on the cold fluid would be m dot C the cold fluid mass flow
rate multiplied by C p C and T h T c outlet minus T c inlet. So, that is a simple heat
balance which expresses the rise in temperature of the cold fluid and drop in temperature
of the hot fluid and the simple mass balance, simple energy balance would give us that
whatever be the heat lost by the cold fluid gained by the cold fluid is equal to the heat
lost by the hot fluid.
Now, when you think a little bit more I will write those expressions and when you think
a little bit more you would see that there is a maximum temperature drop which is going
493
which is will be possible for different two different cases, two different scenarios based
on what is m dot c C p c and m dot h C p h what these the product of these two m dot c C
p c and m dot h C p h which are the capacities of the hot fluid to release heat and the cold
fluid to absorb heat they together would give us some idea of the maximum temperature
drop or jump which is possible for a specific situation and this can then be manipulated
to obtain and idea of what is the effectiveness of heat transfer in a effectiveness of heat
transfer in a heat exchanger.
So, let us start from the beginning writing a simple heat balance between the hot fluid
and the cold fluid.
So, we will start over here that the effectiveness or NTU method where NTU simply
stands for number of transfer units. So, this is to be used if only the inlet temperatures are
known the use of LMTD requires an iterative procedure and an alternative methodology
which is known as the effective NTU method is to be adopted. So, what we are going to
find out is the maximum possible whenever we talk about effectiveness we have to
express it as a ratio that is the denominator should be the maximum and the numerator
should be the actual.
Any effectiveness is defined as the actual amount divided by the ideal or the maximum
amount. So, in order to define effectiveness of a heat exchanger we must find out what is
the maximum possible heat transfer rate that is possible that is maximum possible heat
494
transfer rate of a heat exchanger and this maximum possible heat transfer rate let us call
it as q max and see what is going to be the value or expression for the heat exchanger.
So, if I provide the following thing if I provide an infinite length of the heat exchanger
then the delta T should approach T h i minus T c i. So, the temperature of the hot fluid at
the inlet and temperature of the cold fluid at the inlet in a heat exchanger that is the
maximum temperature drop that is possible
So, hot fluid enters at T h i and leaves at T h o whereas, the cold fluid enters at T c i and
leaves at T c o. So, the maximum difference in temperature which exists in a heat
exchanger is the temperature of the hot fluid and the temperature of the cold fluid both at
the inlet. So, if we can provide an infinite length then that is the heat transferred that is
what is going to be the maximum change in temperature which can be achieved by heat
exchanged between the cold fluid and the hot fluid.
But, what is going to happen, which liquid is going to attain the maximum temperature
drop first, is it going to be the hot fluid whose temperature is going to decrease from T h
i to T c i or is it going to be the cold fluid whose temperature whose temperature is going
to rise from T c i to T h i? Obviously, which liquid is going to experience the maximum
possible temperature drop provided an infinite length of the heat exchanger is provided
would depend on the value of m dot c C p c and m dot h C p h. These capacities which
are the product of the mass flow rate and the specific heat whichever the total amount of
heat transfer remains the same remains equal the amount of heat lost by the hot fluid
must be the equal to the amount of heat gained by the cold fluid.
But, which liquid would face the maximum possible temperature drop would strongly
depend on the value of the m dot C p. So, the m dot C p whichever is lower between the
hot fluid and the cold fluid the value of m dot C p if it is lower for let us say for the hot
fluid then the temperature drop the maximum temperature drop would reach faster for
the hot fluid in comparison to that of the cold fluid.
So, since the value of q which is m dot C p multiplied del by delta T is the same both for
the hot fluid and for the cold fluid whosoever which roots which so ever fluid has
smaller value of m dot C p the delta T drop of rise would be faster for that fluid. So, if
for example, if the m dot C p for the hot fluid is lesser as compared to the m dot C p of
495
the cold fluid it is the hot fluid which is going to experience higher drop of its
temperature.
So, the hot fluid in that case would reach the highest possible temperature drop in a heat
exchanger which is T h i minus T c i. So, when the hot fluid has reached that temperature
difference your simple heat balance would tell you that the cold fluid is yet to reach that
delta T which would give us some idea of the heat transferred under that situation. So, let
us write it and see what it gives rise to.
So, for cold fluid temperature capacity if it is less than the capacity of the hot fluid and C
is simply m C p m dot C p if it is h this is going to be h for the C it is going to be C. So,
if C c is less than C h then q max would be capacity of the cold fluid times T h i minus T
c i. So, T h my i minus T c i as I have explained before it is a maximum possible
temperature drop in the system. So, if C c is less than C h then this q max will be this if it
is other way round. So, please note that in both cases the temperature drop remains the
same whichever is whichever is a so, is less than whichever is smaller it would reach the
maximum temperature drop faster; so, the effectiveness so, I guess these two are clear to
all of you.
And, then the effectiveness for anything which is denoted by epsilon in this case is the
actual heat transferred by the maximum heat which is transferred. So, this is actual heat
transferred by maximum possible heat transferred. So, epsilon is therefore, C of h the
496
actual heat transferred the drop of the hot fluid from the inlet to the outlet. So, this must
be the heat which is transferred actually by the hot fluid which should also be equal to C
of C T c out minus T c i. So, it can be expressed either the actual heat transfer can either
be expressed in terms of the hot fluid losing energy or the cold fluid gaining energy the
denominator is both cases would be the minimum of the C h and C c whichever is
smaller is going to reach the maximum possible temperature difference faster.
So, therefore, the maximum possible heat transfer must be equal to C minimum that is
the lesser of C h or C c multiplied by the maximum possible drop the denominator
remains the same for both cases and you can express whichever you can express in
whichever way you want whichever is convenient either in terms of the hot fluid or in
terms of the cold fluid. So, I am going through this part once again so, as to ensure that
there is no confusion in your mind a fluid which has a lesser capacity its temperature
drop is going to be more as compared to a fluid of higher capacity when the heat
transferred between the two is identical.
So, for a cold fluid with a lesser value lesser value of the capacity would have a higher
temperature drop and this higher temperature drop would ensure that it is going to reach
the maximum possible temperature drop faster as compared to that of the hot fluid. So,
the minimum of C h and C c the capacity of the hot fluid and the capacity of the cold
fluid multiplied by the maximum possible temperature drop which is T h i minus T c i
would provide the maximum amount of heat transfer which is possible. So, the
effectiveness which is always defined by actual divided by maximum, so, the actual can
be bit expressed based on either the cold fluid or the hot fluid and the maximum would
simply be equal to minimum C minimum of C h and C c multiplied by the allowable
maximum allowable temperature jump which is T c T h i minus T c i.
So, therefore, the effectiveness that the way we have defined must lie within a range of 0
to 1, ok. So, that is what the effectiveness of a system should be. I mean generally
whatever efficiency effectiveness that we define would vary between 0 and 1. So, the
effectiveness the way we have defined ensures that the effectiveness of a heat exchanger
will be between 0 and 1. So, your epsilon should be within this value between 0 to 1 and
it appears only when you provide maximum this maximum possible heat transferred is
when L tends to infinity; that means, you provide a very larger length for the for the two
streams to come to thermal equilibrium perfect thermal equilibrium at the exit.
497
So, if epsilon T h i and T c i are known the actual heat transfer can be expressed as q as
epsilon C minimum T h i minus T c i. So, that is the purpose of the whole exercise if the
inlet temperatures are known this is this is based on process conditions and process
conditions and the physical property because your C is simply defined as m dot times C
p. So, this C minimum is the minimum of this then you can find out knowing the value
of epsilon what is the actual amount of heat transfer. Sometimes for it is also expressed
for any heat exchanger epsilon is a function of number of transfer units C minimum by C
maximum. So, this NTU is defined as UA by C minimum and NTU is nothing but
number of transfer units I will explain it slowly.
We understand that if I know epsilon I do not need to know what is the outlet
temperature only a knowledge of the inlet temperature would be sufficient to find out the
value of q provided I know epsilon. So, the our effort from now on would be to find out
how one can obtain the value of epsilon be it analytically or through the through the use
of a series of graphs which are nothing, but a solution of this, but a representation of the
analytical solutions which are available.
While doing so, we also understand that this epsilon the effectiveness is a function of the
ratio of C minimum by C maximum. So, if your C minimum the capacity minimum is
corresponds to that of the cold liquid then f then this simply would be the C of the cold
fluid divided by the C of the hot fluid. And, another parameter NTU which is the number
of transfer units is added in here that is an effect of the process the effect of the heat
transfer coefficients effect of the heat transfer area on this. So, which NTU is defined as
U times A by C minimum and the concept of number of transfer units would also appear
subsequently as we would see in your studies of heat transfer.
So, number of transfer units would give you some sort of an idea of the overall
performance of the system based on the some overall coefficient. In this case it is heat
transfer coefficients the area involved in heat transfer and the capacity factor which is C
minimum. So, the if I know my epsilon as a function of NTU and C min by C max then
from this relation knowing the value of epsilon, knowing the value of inlet and outlet
temperatures of the inlet temperatures of the hot fluid and inlet temperatures of the cold
fluid I would be able to obtain what is the total amount of heat transfer in the heat
exchanger. So, the entire emphasis from now on would be to find out this relation and I
498
would show you just one example of how this epsilon can be calculated for parallel flow
heat exchangers.
Similar example similar solutions are also possible are also available for counter flow it
heat exchangers and various other flow configurations. So, if you know the flow
configuration if you know the other relevant parameters, then you should be able to
obtain what is the value of epsilon either from a series of relations which are provided or
the graphical representation of those relations. So, it becomes quite easy as compared to
the LMTD method from those references which are provided in any text book to find out
the total or overall heat transfer the total amount of heat transfer and then go for the
further design of the heat exchangers.
So, let us start with one simple case of evaluating epsilon for a specific flow
configuration.
The one which we are going to do is to find out the relations between effectiveness and
NTU and we are doing it for a parallel flow heat exchanger with C minimum equals C h;
that means, the minimum of the capacity is due to the for the hot fluid only. So, what is
the definition of epsilon it is the actual amount of heat which has been transferred. So,
capacity of the hot fluid T h i minus T h o which is the temperature drop of the hot fluid
multiplied by the capacity and the maximum amount of heat transfer possible which is C
499
minimum the maximum amount of temperature drop or jump which is possible in this
case.
And, since my C minimum is equal to C h in this case that is what we have assumed. So,
these two will simply cancel out and therefore, the effectiveness can be written as T h i
minus T h o as C h is equal to C minimum this has been provided. So, this is parallel
flow heat exchanger. So, a pictorial representation of this would simply be this is T h i, T
h o this is T c out and T c i. So, this one is delta T we call it as delta T 1 this is delta T 2
it is a difference in between the temperature between these sorry, this is parallel flow.
This is not the one for parallel flow it should look something like this both are travelling
in the same direction. So, this is T h i this is T c i this is T c o and this is T h o. T 1 that I
have now drawn is parallel flow, this is for counter flow that is not the one since we have
assumed parallel flow only in this case. So, the difference between the inlet is delta T 1
and this is delta T 2, ok. So, that is how it should look like.
And, what is C minimum by C maximum? C minimum times delta T minimum must be

equal to C maximum times delta T m C p delta c. So, it is m C p this by m C p cold fluid
must be equal to T c o minus T c i divided by T h i minus T h o. So, if you see here m
dot C p h multiplied by delta T on the hot side must be for heat balance must be equal to
m dot C based on the cold side multiplied by the temperature drop in the cold slide cold
side. So, from a simple heat balance it would give you that C minimum by C max is
simply the temperature simply expressed as the ratio of the temperature difference of the
two cases.
So, if I think of q from here the heat that is transferred C h T h i minus T h o which is C c
of the cold fluid T c o minus T c i and when I think it in terms of log mean temperature
difference UA for the case of parallel flow LMTD which is T h o minus T c o that is
delta T 2 minus delta T 1 which is. So, this is nothing, but delta T 2 this is delta T 1 and
the denominator would be ln of delta T 2 by delta T 1.
So, this is the basic fundamental relation which connects the heat transfer is equal to the
heat loss by the hot fluid, heat gained by the cold fluid and in terms of transfer between
the heat if I include a heat transfer coefficient, a heat transfer area then heat transfer
coefficient times heat transfer area times a temperature difference which is nothing but
we which where the temperature difference as we have seen before is expressed in terms
500
of a log mean temperature difference and not the average temperature arithmetic average
temperature difference.
So, the temperature difference of choice to be used for this case is the log mean
temperature difference which is this. So, as you can see this delta T 1 keeps on varying in
a non-linear fashion tilt which is delta T 2. So, a log mean temperature difference is the
appropriate one. So, this is from heat balance this is also from heat balance where we are
talking about the transfer of heat between the hot and the cold hot and the cold fluid.
So, this when I write C h when we take it a bit further then q would simply be equal to
minus UA by ln delta T 2 by delta T 1 is equal to multiplied by q by C h plus q by C c.
How did you get into this how did you get this one? So, we know that this is q so, T h i
minus T h o T h i minus T h o is q by C h and T c o from here T c o minus T c i is q by C
c so, this is what we have in here.
So, when you have T h o minus T h i, T h o minus T h i in here T h o minus T h i must

be equal to minus of q by Ch. So, taking this one and this one I can write then the
difference as q by C h as q by C h with a minus sign and taking into the consideration
that T c i minus T c o, T c i minus T c o is minus q by C c which is minus q by C c the
negative sign taken outside this is what we have done.
501
So, the log mean temperature difference definition T h i minus T c o and T h i minus T c
i for the case of a parallel flow can also be written as T h o minus T h i which is minus of
q by C h and T c i minus T c o can be written as minus q by C c. So, this is what one can
obtain for the case of the heat transfer between this see the main purpose of this is to get
rid of the outlet temperature. So, wherever T c o or T h o appears through the use of the
heat balance we try to get rid of this the outlet temperatures and see what we can obtain
as an expression involving epsilon and NTU to express the heat transfer for such
situations.
So, I am not doing anything new not bringing any new concept except the maximum
possible temperature difference and which of the fluid is going to reach that maximum
faster as compared to the other. Whatever else I am going is simply an algebraic
manipulation of the expression in order to bring in order to get rid of T c o and T h o and
express everything in terms of C c, C h, epsilon, U, A and q that is all I am doing. So,
that is the only thing which we are doing in this specific exercise.
So, when you come back to this once again then ln delta T 2 by delta T 1 can be written I
mean delta T 2 by delta T 1 therefore, can be written as minus UA 1 by C h plus 1 by C
c. So, if you look at this one I will bring this over to this side. So, this becomes ln delta T
2 by delta T 1 and the since q will. So, when I bring it to this side the q will cancel from
all sides and I have minus UA 1 by C h plus 1 by C c.
For the parallel flow when I express it in terms of the differences I understand my C h is
equal to the C minimum as we have defined here and C c is the C maximum. So, when I
take C h outside it is simply going to be 1 plus C minimum by C maximum. So, Ch is
assumed as the minimum and this obviously, is therefore, the maximum. So, this is the
expression that you would update.
And, now I bring this NTU which is defined as the one which is on the outside of this
bracket this is the expression. So, therefore, what I would get is T h o minus T c o by T h
i minus T c i which is this one, since it is the logarithmic form it is simply going to equal
to exponential minus NTU into 1 plus C minimum by C maximum. So, what then I have
is a relation which brings in the new concept NTU which is UA by C minimum and
slowly I am transforming everything in terms of NTU either NTU and epsilon C
minimum and C maximum.
502
So, when we see the last expression that I have written it still contains some of the outlet
temperatures T c o and T h o. So, if I can now get rid of this T c o and T h o then
everything is going to be either in terms of T h i, T c i, epsilon C minimum by C max and
U that is our objective that is what we would like to do. So, I will continue this in my
next class when we derive a compact expression for epsilon in parallel flow, where only
the inlet temperatures are known the outlet temperatures are not known to me.
So, once we do that we can have other relations for other flow geometries, for multi path
systems, for cross flows and so so on for which in your text you would see the results are
presented in a graphical form and it becomes very easy. Trust me on this, it becomes
very easy to solve for the unknown temperatures in such cases and then proceed towards
a design of heat exchanger treatments.
So, I will stop here today and I will pick it up from this point in the next class to
complete our discussion on effectiveness NTU method.
503
Heat Transfer
Lecture – 44
Epsilon – NTU Method -1 (Continued)
We have been discussing with the effectiveness NTU Method for the design of heat
exchangers. And, I have told you in the previous class, that this method is useful when
we do not know any of the outlet temperatures. So, the inlet temperature of the hot fluid
and the inlet temperature of the cold fluid are known, but we do not know the outlet
temperatures. So, use of LMTD approach for calculation or for design of such systems
becomes tedious as we have to go for an iterative solution.
So, the effectiveness method takes care of these unknown values of the outlet
temperature and expresses the effectiveness which was denoted by ϵ in terms of NTU,
where NTU is defined as the number of transfer units. So, in the last class we have
defined what effectiveness is and how effectiveness and NTU are connected for a simple
situation of a parallel flow 1-1 heat exchanger.
But, we did not complete it in that class. So, I will continue the derivation of the
pertinent relation between ϵ, NTU and so on. For the case of a parallel flow heat for a
parallel flow heat exchanger, and we would see what would happen if it is a counter flow
if it is a system which contains multiple tube passes, and possibly multiple shell passes
as well, and what is going to happen if it is a boiler, or a condenser. Such that in one of
the streams the temperature will not change, what would be the value of ϵ, or what
expression would result between ϵ and NTU, when one of the liquids, one of the fluids is
undergoing phase change as in boilers or condensers.
And, then towards the end we would see that all these results for the if ϵ- NTU relation,
the relation between effectiveness and number of transfer units are provided in the text in
the form of tables and graphs. So, in the next class we would solve a problem to see, how
to calculate the ϵ, the NTU, the unknown temperatures and so on through the application
of an example. But, today again as I start with the basic definition of the effectiveness
and how it is going to be connected with NTU and so on. So, I will quickly go through
504
the slides, which we have discussed in the last class and then move on to derive the final
relation.
So, what we have done in the previous class is we started to work on ϵ - NTU method,
and the in order the maximum amount of heat transfer, the maximum possible heat
transfer can take place, when the temperature drop goes from Thi, that is the temperature
of the hot fluid to Tci, temperature of the cold fluid. So, that is the maximum possible
temperature difference that exists in a heat exchanger, and if the temperature of the hot
fluid reduces to the temperature of the inlet temperature of the cold fluid or the
temperature of the so, this is going to give rise to the maximum possible amount of heat
transfer.
& p . So, capacity of the

However, the if the capacity where the capacity is defined as mC
hot fluid would be
Ch   mC
& p
h
and for this cold fluid (Cc) it simply going to be  mC

& p  So, if Cc is less than Ch, then the
c
maximum amount of heat that can get transferred would be Thi  Tci .
505
And, thus similarly the simple logic would tell me that whichever is the lower of the
capacity between the cold and the hot. The qmax is simply going to be the capacity of that
one which is the minimum. So, I can write

qmax  Cc Thi  Tci 
So, this is the maximum temperature drop that can take place qmax is simply going to be
related by this.

qmax  Ch Thi  Tci 

qmax  Cmin Thi  Tci 
The effectiveness was defined as the actual heat which has been transferred divided by
the maximum possible heat transfer, and the maximum possible heat transfer
corresponds to a length of infinite length of the exchanger. So, the actual heat transfer
can either be Ch times the temperature difference between the hot fluid inlet and the
outlet, or Cc and the temperature difference between the outlet and inlet for the cold
fluid. So, these 2 are the numerators are the actual heat that has been transferred and;
obviously, through heat balance these 2 are equal.
Whereas, the denominator is going to be as I said C minimum times the maximum

possible temperature drop that is possible in a heat exchanger. So,
506
q

qmax


Ch Thi  Tho  

Cc Tco  Tci 

Cmin Thi  Tci  
Cmin Thi  Tci 
As you can see from the definition the epsilon will be 0  1 . And, if we know the
value of ϵ, the value of Thi and the value of Tci, then the actual heat transfer is simply
going to be q  Cmin Thi  Tci . 

Since, ϵ-NTU method is used when the inlet temperatures are known. So, therefore T hi
and Tci are known to me. And therefore, it is possible to calculate the actual heat transfer,
based on the difference of the inlet temperatures provided ϵ is known. So, the entire
exercise that we are going to undertake in this part of the course is to find a relation
between ϵ. And, for any heat exchanger ϵ is going to be a function of NTU. Where NTU
UA
is defined as the number of transfer in which it is, NTU  , and the ratio of Cminimum
Cmin
 C 
by Cmax, which is sometimes also referred to as Cr . So,  f  NTU , min 
 Cmax 
507
So, then when we went into this effectiveness-NTU relation and we started to derive the
expression for a parallel flow as I mentioned before, parallel flow heat exchanger with
Cminimum equals Ch. So, Ch   mC

& p  for the hot fluid is the lower one.
h
So, therefore, ϵ would be the total heat actual heat transferred divided by the maximum
heat transfer that could take place. And, in this case Ch  Cmin . So, Ch and Ch would
cancel from the numerator and the denominator. So, your ϵ should simply be the
temperature change of the hot fluid divided by the maximum change that can take place
in a heat exchanger of infinite length. So, it is Thi  Tci .
The Cminimum capacity of Cminimum by Cmaximum is simply this, in this specific case h is the
minimum and cold fluid is the maximum and for a parallel flow this is how the
temperature would change? So, the hot fluid would decrease from Thi to Tho whereas, the
cold fluid temperature would increase from T ci to Tco and the T1 , which is the
temperature difference existing at the inlet is simply Thi  Tci . And, the temperature
difference between at the outlet is this difference whereas, in between it is going to vary
in a non-linear fashion. And, we already have seen that the pertinent T to be used for
such cases is the log mean temperature difference.
So, the overall temperature difference, the average temperature difference between the
inlet and the outlet is simply the log mean temperature difference. And, since it is a 1-1
heat exchanger so, the it is a parallel flow simple 1-1 heat exchanger. So, no correction
factor to LMTD needs to be provided. So, whatever be the LMTD that we calculate
based on the inlet temperature difference, and the outlet temperature difference can
simply be used to express the total heat transfer in the form of q  UATLM , where, U is
the overall heat transfer coefficient, the overall heat transfer coefficient consists of the
convective resistance at the inside of the tube, the convective resistance at the outside of
the tube, and the thermal conductive resistance of the pipe wall if the pipe is not thin. So,
if the pipe is thin the overall heat transfer coefficient has 2 components, what one is h i,
the heat transfer coefficient based on the inside of the inside area inside of the tube and
ho. To this sometimes we add the dirt factor, which is simply the deposition and scaling
508
of the inside as well as the outside of the tube that can lead to significant resistances. So,
all these resistances are going to be in series. So,
Cmin  mC

& p
h

Tc  Tci
 o

Cmax  mC
& p 
Thi  Tho 
c
Tho  
 Tco  Thi  Tci 
   
q  Ch Thi  Tho  Cc Tco  Tci  UA
T2
ln
T1
So, U is defined. Therefore, q  UATLM and with no correction factor for this specific
case. This q is also equal to the Ch the capacity on the heat side, now capacity of the hot
fluid, temperature drop of the hot fluid, capacity of the cold fluid and temperature drop
of the cold fluid. And, this is simply the definition of LMTD which we have seen before.
So, therefore, from these 2 relations I can write T, the temperature difference of the hot
side is simply q/Ch, temperature difference of the cold side is simply q/Cc, and since q is
equal to this.
509
q
Thi  Tho 
Ch
q
Tco  Tci 
Cc
UA q q
q   
T  C h Cc 
ln 2
T1
T2 1 1
ln  UA   
T1  Ch Cc 
And therefore, upon simplification what you get is simply this factor. And therefore, the
heat the temperature jump at the temperature change at the hot outlet temperature and
cold outlet temperature, divided by the maximum temperature drop that can take place is
going to be
UA
NTU 
Cmin
Tho  Tco   C 
 exp   NTU 1  min  
Thi  Tci   Cmax  
We cannot deal with outlet temperatures, since they are unknown to us and the whole
point of doing the NTU ϵ analysis is to get rid of the outlet temperatures in the final
expression such that we can use that.
510
So, we start from this point and write the expression once again over here
Tho  Tco   C 
 exp   NTU 1  min  
Thi  Tci   Cmax  
So, this is the expression which we are going to start with and which we would like to
solve.
So, starting from this point the sorry that is the starting point of the exercise in today’s
class. So, we will start with this and try to get rid of the outlet temperatures. In order to
do that I express
Tho  Tco Tho  Thi  Thi  Tco


Thi  Tci Thi  Tci
If you look at our previous expression, which is which we have done at the at the
beginning, which is
Cmin  mC

& p
h

Tc  Tci
 o

Cmax  mC
& p 
Thi  Tho 
c
So, that is the expression which we have obtained this is simply a heat balance.
So, from here
511
Cmin
Tco  
Th  Tho  Tci
Cmax i

So, the first express first expression for Tco which I have obtained is from a heat balance.
So, therefore,
Cmin
Tho  Tco T ho  
 Thi  Thi  Tci  Th  Tho
Cmax i


Thi  Tci Thi  Tci 
Tho  Tco C 
Again,    1   min 
Thi  Tci  Cmax 
So, everything is in the form of ϵ.
So, I simply reorganize it a little bit so, taking the ϵ count it simply going to be
Tho  Tco  C 
 1   1  min 
Thi  Tci  Cmax 
Tho  Tco   C   C 
 exp   NTU 1  min    1   1  min 
Thi  Tci   Cmax    Cmax 
512
So, what you would see here is that this temp this temperature difference containing 2
temperatures 2 outlet temperatures is now replaced by an expression containing ϵ and
NTU, but it does not contain any of the outlet temperatures. So, this is the beauty of the
ϵ- NTU method starting with a simple definition of ϵ. And, at and for applying it for the
case of a parallel flow heat exchanger 1-1 heat exchanger in which Cminimum is taken to be
equal to Ch.
And, for making simple suggestions statements heat balance, definition of ϵ, definition of
NTU, we have arrived at an expression that contains only ϵ and NTU. And what is NTU?
NTU contains the overall heat transfer coefficient U. So, the overall heat transfer
coefficient of the heat exchanger can be obtained if through the relation between ϵ and
NTU. So, this expression that we have obtained for ϵ-NTU for parallel flow 1-1 heat
exchanger in which C h is the minimum capacity; that means, Ch is less than Cc can also
be obtained when for in a similar fashion when C c is the minimum, that is Cc is less than
Ch.
So, this is the simplest possible form of ϵ- NTU relation and I will just simply complete
this to give you an idea of what is the final form of epsilon going to be. So, epsilon from
here would be equal to
  C 
1  exp   NTU 1  min  
  Cmax  

C
1  min
Cmax
Similar same result can also be obtained for the case when C minimum is Cc. So, the above
equation applies to this equation, applies to any parallel flow heat exchanger that is what
is important. Similar expressions have been developed for a variety of heat exchangers
and I will show you some of some of the expressions over here.
For 2 or more shell passes when I have 2 or more shell pass NTU is to be replaced by
NTU by n. Now, for a boiler or condenser let us say that the temperature of one of the
phases will remain constant. So, if it is a case of stream condensing on the outside then
the temperature of the temperature of the hot stream is not going to change, it will remain
constant at that value. On the other hand if you have evaporation taking place then the
temperature of that phase will also remain constant. So, it is not going to be any
513
temperature change for the case of boilers and condensers. So, let us see what expression
epsilon would take for the case of boilers and condensers.
So, for boilers and condensers, so only phase change is taking place, no temperature
variation. So, what is going to happen is this
Cmin
Cr  0
Cmax
The important thing is the heat exchanger behavior is independent of flow arrangement.
So, it does not matter if it is in parallel flow or in counter flow because that concept does
not apply when the temperature of one of the phase is remaining constant.
So, essentially what it gives you the heat exchanger behavior, heat exchange behavior is
independent of the flow arrangement and from the previous one you can clearly see that
this would lead to where this is 0, this would lead to
 1  exp( NTU )
So, this is the expression for between ϵ and NTU, when you have heat either a boiler or a
condenser and heat exchanger behavior is important and is independent of the flow
arrangement.
514
Sometimes for design it is more convenient to have not ϵ as a function of NTU, but
 C 
NTU  f , min  . And, for this has been solved for a number of flow geometries for a
 Cmax 
number of situations and there are graphs available, which would give you the value the
Cmin
relation between NTU and ϵ for different values of .
Cmax
So, it from I will show you some of the graphs and if NTU is less than equal to 0.25 all
heat exchangers have the same effectiveness. The same effectiveness and regardless of
what is the value of Cr. And, it is also shown that for C r greater than 0, NTU greater than
0.25 counter flow is definitely more effective than compared to parallel flow. So, these
are some of the observations which you can make when you look at the heat exchange
this curve.
So, what I show you here is the from the text there are certain relations and correlations
which are which can be seen, for the case of parallel flow, for the case of counter flow,
shell and tube with 1 shell pass and multiples of 2 in terms of tube passes, and shell
passes and if you have 2 shell passes then there must be 4 tube passes and so on. And,
cross flow then all exchangers for which Cr is equal to 0 which is the case for condensers
and for boilers the relation between epsilon and NTU are provided in the text.
515
So, they have already been solved I have already showed you the simplest example of
parallel flow heat exchanger, but here in the text in your text in the book of Incropera
and Dewitt the text book you would see this kind of relations for ϵ and NTU, which are
available for most of the cases that we can think of. As I mentioned sometimes it is for
the design purposes it is better to have NTU in terms of ϵ
So, the next table in the in your text, gives you the same thing, but right over here the
NTU is on the left hand side, they are related explicit relation of NTU with ϵ which are
provided for parallel flow counter flow shell and tube and for all types of exchangers
where Cr is equal to 0.
516
So, these are the relations where NTU is expressed in terms of ϵ. And, the figures that I
was talking about it is also from your text, what where you see the effectiveness ϵ of a
parallel flow heat exchanger, which is the equation number 11.28 in the previous page.
And, you would see that the values what you can do is you can you can find out the if
you know the value of NTU, and then is find out what is the Cminimum/Cmaximum read that go
to up to that point and find out what is ϵ or effectiveness of the heat exchanger.
Once, we solve a problem in the next class it would be even more clear to you. So, this is
for a parallel flow heat exchanger, this is for a counter flow heat exchanger, where the
effectiveness is expressed against number of transfer units for different values of Cr,
which is Cminimum/Cmaximum, 0 corresponds to the case of a condensers and boilers and so
on. And, the last one that you see is when Cr is equal to 1; that means, Cmax is equal to
C minimum and this is the so, therefore, let us say when water flows in both the shell and
tube with the same flow rate. So,  mC

& p  is equal to  mC
c
& p  and this is the relation
h
which you would get.
517
And, all these ones for this is for shell and tube heat exchanger with 1 shell and any
multiple of 2 tube passes that is 2 4 6 etcetera. Again, this is the value for C r equals to 0
and this is the value for Cr equals 1, the values of ϵ are plotted against NTU and you also
have the shell and tube heat exchanger with 2 shell passes and any multiple or 4 tube
passes. So, that is a relation which has been provided.
So, this is for cross flow heat exchanger which one can see, but we are going to restrict
ourselves to these 4 curves. And, when we solve the problem we would see how the
518
knowledge of ϵ can tell us what is NTU or if I know the NTU for a specific process can
we calculate ϵ. And, from the fundamental definition of ϵ the moment I know what ϵ is
then I should be able to find out, what is the inlet temperature and what are the unknown
outlet temperatures of the hot and the cold fluid.
So, there are 2 approaches as I mentioned for the design of heat exchangers, if you know
all if or if you can find out from heat balance all the temperatures of the inlet and the
outlet streams probably use of LMTD is better suited for that condition, but if you do not
know the outlet temperatures and iteration is required to solve to go to adopt the LMTD,
in such a case it is easier to use the ϵ - NTU approach, where the significant number of
graphical solutions or graphs available for different flow situations. And, different heat
exchangers would enable you to do the calculation in a much more faster and direct way
as compared to the iterative process of LMTD for that specific cases.
So, both approaches can be adopted you decide based on the data that you have which
one is going to be more suitable for faster calculation and easier evaluation of all the
temperatures. And, once you have all the temperatures then you would be able to find out
other design parameters.
So, we will talk a bit more about this epsilon NTU method in subsequent classes, but I
think we have had enough information to solve a problem and clarify about epsilon NTU
method and clarify some of the doubts. And, to enable you how to use these graphs
charts relations and arrive at a solution. So, the next class is going to a tutorial on ϵ-NTU
method.
519
Heat Transfer
Lecture – 45
Tutorial Problems on Epsilon – NTU Methods
So, we have a fair idea of what is effectiveness NTU Method and when this can be used.
This is principally used under those circumstances when we do not know what the outlet
temperatures are. And instead of going through an iterative procedure for LMTD type of
calculation the ϵ-NTU is a direct method which with the help of the relations derived for
many flow configurations specially their graphical representations would allow us to
calculate the unknown temperatures, the heat transfer coefficients fairly quickly and
directly from the knowledge of ϵ-NTU methods. So, this class is going to be tutorial one,
we will solve one problem and see how effective this Epsilon NTU method is.
So, for that we take the problem where a steam at a specific pressure is being condensed
and this problem is about saturated steam which is at a pressure of 0.14 bar is condensed
in a shell and tube heat exchanger. The shell and tube heat exchanger is 1, 2- heat
exchanger. So, one shell pass, 2 tube passes each pass each tube pass would contain 130
tubes, which are made of brass the length of the tubes is 2 meters.
520
The tube internal ID is 13.4 millimeter, the OD is 15.9 millimeter, the cooling water
enters at 20 ºC, at a cooling water enters the tubes at 20 ºC with a velocity equals 1.25 m/
s. The value of the overall heat transfer coefficient on the outside of the tubes applicable
for the outside of the tubes is 13500 W/m2K.
What we have to find out is: what is the overall heat transfer coefficient? What is the
temperature of the cooling water at the outlet? And what is the rate at which steam gets
condensed in this heat exchanger. So, needless to say here we do not know we know
what is going to be the temperature of the steam since it is saturated, and the pressure is
provided. So, you would be able to know, what is the temperature of the steam?
However, the one of the outlet temperatures is unknown and the steam condensation rate
is also unknown. So, there is no way use of simple LMTD would give us the temperature
of the cooling water, it would require an iterative procedure.
On the other hand, this is the case where the steam is condensing and there is no change
in temperature in that phase. So, Cr is 0 and therefore, a straight forward relation
between ϵ and NTU may be available. So, this is a perfect example for the use of the ϵ-
NTU method. So, we would do that, but in order to do this we need to know some of the
properties. So, for the steam which is saturated at 0.14 bar and you can ever it you can
get the temperature from the steam table which is T SAT is the temperature of the hot fluid
which is the same at the inlet and outlet is 327 Kelvin, the value of the latent heat for
condensation is 2373 KJ/kg, the value of the heat capacity is 1898 J/kgK.
Whereas, for he cooling water the specific volume is 1.005 × 10 -3 m3/kg the value of Cp,
the viscosity, and the thermal conductivity, the Prandtl number and brass which is the
material of construction of the tube. And, there is significant thickness of the brass it is
not a very thin tube the thermal conductivity of the brass is 114 W/mK.
In order to evaluate the property of water we only know: what is the inlet temperature of
the water. So, in order to find out the average temperature at which all these properties
are to be calculated, we can just assume a temperature for example, 60 º the temperature
of the outlet water has let us say 50 º. So, the average temperature would be 35 and we
read all these values at 35.
If after calculation you see that there is significant difference between the temperature
that you have assumed and the temperature which you have obtained. So, instead of 50 if
521
you get 90 then the properties are going to be significantly affected and you need to do
one more iteration, just one more iteration to ensure that you get the right values of the
physical properties.
But since it is not going to be a very strong function of temperature so, one logical
logical choice of the outlet temperature is in the right range would give you sufficient
accuracy for all design calculations. So, you need not have to take any arbitrary value,
just logically take some guess value of the outlet temperature. And, if your final value is
not too different then stick to it. If instead of the outlet temperature your assumed outlet
temperature of 50 º, if you get 55 or 60 even 60 it will not make any difference in your
physical properties to the extents such that your final result is going to be affected. So,
that is what you have to do.
So, we will start this problem of this specific case utilizing ϵ-NTU and realizing that now
I have 3 resistances to consider, one is the internal the convection resistance based on the
inside of the tube, the conduction resistance of the brass tube and the convection
resistance based on the outside area of the tube for which the value of h 0 is given, but the
value of hi is not given. So, therefore, we need to start from the whole water side
calculate what is its Reynolds number, use the appropriate relation, try to find out what is
the value of hi. And once I have hi since I know h0 and the resistance because of the brass
plate brass wall is to be included then I would get an overall heat transfer coefficient U.
So, in the expression of ϵ and NTU, the value of U would be known to me, and the value
of area heat transfer area that is also known to me since I know how many tubes are
present and what would be the outside area for 2 meter long pipes. So, we have 132 pipes
2 meter long with known outer diameter. So, we should be able to find out what is the
total heat transfer area.
So, once we have the total heat transfer area and the value of the overall heat transfer
coefficient then the one side of the equation the NTU part is known to me, the only
unknown is ϵ. So, we would see how this problem can be solved.
522
So, let us start the problem solution of the problem and the first thing that one has to do
is to find out what is the overall heat transfer coefficient based on the outside tube area.
So, this U with suffix 0, which U0 which simply denotes is based on the outside tube
area. So, this is
1
1 r r r 1
U 0    0 ln 0  0 
 h0 k ri ri hi 
This is the say straight forward conduction where
r0
ln
1 1 ri 1
  
2 r0U 0 2 ro ho 2 k 2 ri hi
where U0 is the overall heat transfer coefficient.
So, this is the relation the overall heat transfer coefficient is the sum of the external
convective resistance and that is why r0 is here. We know that Reynolds number based
on diameter and this is the inner diameter would be
(1.005 103 )1 1.25 13.4 103

Re Di   21673
769  106
So, when you plug in the values this would turn out to be 21673.
523
So; obviously, it is a turbulent flow and it is fully developed based on whatever we have
studied for that case. So, I am going to use Dittus-Boelter equation and the Dittus-
Boelter equation would simply give you Nusselt number.
hi Di
NuDi   0.023Re0.8 Pr 0.4  131
kf
And from the Nusselt number you should be able to calculate what is the value of hi
0.62
hi  3
131  6057W / m 2 K
13.4 10
So, what is U0 putting into the value of the only unknown here is h i, because the thermal
conductivity is known, the geometric parameters are known, h0 value is known. So,
1
 1 15.9 103 15.9 15.9 1 
U0    ln    3549W / m 2 K
13500 115 13.4 13.4 6057 
So, now I have obtained what is the average value, what is the overall heat transfer
coefficient based on the outside area. Now to find the outlet temperatures one part of the
problem is over to find the outlet temperature of water we need to use ϵ-NTU method.
So, how am I going to use ϵ-NTU method, the total amount of heat transfer would be
524

q  Cc Tco  Tci 
q
So, Tco  Tci 
Cc
So, that is the relation which we have to find out. Now this q is to be obtained so, this is
the outlet temperature, but I do not know the outlet temperature I only know the inlet
temperature. So, Tci and Thi which is equal to Tho this is known to me. So, how do I do
this?
q, the total amount of heat transfer would be effectiveness times q max.
q= ϵ×qmax= ϵ×Cmin× (Thi - Tho)
So, this is the definition of the effectiveness epsilon, where q, the actual heat transferred
is ϵ times the maximum heat which is transferred. So, that is the way you calculate this.
So, therefore, the minimum capacity rate is that of the cold water as Ch tends to ∞.
So, capacity of the hot fluid since it is steam condensing on the outside will obviously
tend to infinity. So, therefore, we are dealing with the situation in which Cr is equal to 0.
So, that would allow us to use a simple relation between ϵ and NTU and since NTU is
known we should be able to calculate ϵ and from ϵ we can calculate q and from q we can
525
calculate what is Tco. So, that is the those are the steps which I am going to follow one
after the other.
So, Ch tends to ∞ and
Cminimum  Cc   mC
& p
c
kg  m
So, m& (  AU m ) N  995 3
 (0.0134) 2 m 2 1.25  130  22.8kg / s
m 4 s
Hence, Cminimum = 22.8 kg/s ×4178 J/kgK=95270 W/K
So, this is the value of C minimum that we have obtained. So, now what is NTU, I have
the all the values required for NTU because NTU is defined as
NTU = U0A0/ Cminimum
I have calculated what is U0 in the previous calculation and therefore, I can find out what
is the NTU.
NTU being defined as U0A0/ Cminimum which in this case would be 3549 is U 0. The area
would not be the cross-sectional area, this is the area available for convective heat
526
transfer. The area available for convective heat transfer would be π×D×L where, D is the
outside diameter because that is the area which is available for heat transfer. So, that is
the area multiplied by the number of tubes. So, this is N × 2 this is the number of passes
ok.
So, just be careful about using while calculating the overall area for heat transfer. The
total area for heat transfer would be the outside area of one tube multiplied by the
number of tubes per pass and the total number of passes. Because when you do that only
then you get the total heat transfer area in the heat exchanger considering all tubes,
considering one tube, considering all tubes and considering all number of all passes. So,
the total would give you the total surface area available for heat transfer which is by
definition A0 in NTU.
So, I have already calculated U0, now know what is A0 and the Cminimum I have calculated
in the previous slide. So, this Cminimum is simply going to be the one which we have
calculated earlier. So, this NTU would be 0.968 so, from the figure of ϵ versus NTU. So,
let us now look at the figure that we have over here.
If you look at the numbers over here we are dealing with a shell and tube heat exchanger
with 1 2 arrangement. So, this is the parallel flow heat exchanger.
527
This is a counter flow heat exchanger; this is a shell and tube heat exchanger with N with
1 shell and any multiple of 2 tube passes. So, we have the first case which is 1 shell 2
tube passes and the value of NTU we just calculated is about 0.968 which is about which
is about one and we need to we need to find out what is the value of ϵ for such a case.
So, from NTU equal to 1, since my Cr is 0 we go all the way up to this point and come
over here and find ϵ. So, with a known value of NTU to be equal to 0.968 we go up to
the curve which is Cminimum/Cmaximum=0 which should be the case for condensers and
boilers and we read what is the value of ϵ. So, this value of ϵ is going to be slightly more
than 0.6. Let us take the ϵ to be equal to point 0.62. So, knowing the value of NTU
knowing the ratio of Cminimum/Cmaximum we find out what is the value of ϵ.
So, the value of epsilon that we obtain from here from the figure for or from the graph
for ϵ versus NTU, NTU is equal to 0.968 and C minimum/Cmaximum since it is condensing to be
equal to 0, ϵ would be equal to 0.62. So, the next part is straight forward.
 
Tco  Tci   Cmin imum Thi  Tci / Cc
So, when you put the values in here your Tco to be equal to 41.1 ºC.
So, our assumption of the temperature to be equal to 50, the outlet temperature to be
equal to 50 is not a bad one which we have used to calculate all the physical properties.
528
So, since it is ok it is not too different from our assumed values. So, we will stick to this
value and proceed ahead, instead of assumed 50 if it comes to be may be 75 or. So, then
we have to go do one more iteration, but 50 and 41 is not that different and all our
properties that we have taken for water are not that sensitive to temperature change and
therefore, this is a reasonably good assumption and we can proceed with this.
So, the second part of the problem is also over we have calculated what is the
temperature of the cooling water outlet, which was not possible using LMTD without a
without substantial iteration. So, ϵ-NTU methods clearly gave us a simple way to tackle
the problem of knowing the outlet temperature of the cold fluid.
The third one which is remaining is, what is the condensation rate of the steam; so, let us
do the last part over here. The condensation rate of the steam which is let us say
W

Cc Tco  Tci  95270 (41.1  20) K
q K
m&h  
h fg h fg 2373  103 J / kgK
This would result in a value of m&to be equal 0.85 kg/s. So, that is the condensation rate
of steam. So, this problem which the tutorial problem, which we have solved in the class
clearly shows the advantage of the ϵ-NTU method. So, the idea would be to obtain NTU
and then read the value of ϵ. In order to obtain the NTU in order to calculate the NTU
you need to calculate the overall heat transfer coefficient which is U.
529
You need to know, what is the total area available for heat transfer which is A 0 and you
also have to have an idea of what is the value of C minimum. Once you do that the value of
NTU is known to you and depending on what kind of heat exchanger you are dealing
with, is it a parallel flow 1 1 heat exchanger, counter flow 1 1 heat exchanger, if it is a
single cell multiple tube passes or multiple shell multiple tube passes, for all of them
there are curves available in your text book which would be provided to you provided
these kind of questions come. So, these graphs would be provided to you.
You and then from the calculating the value of NTU and reading the corresponding value
of ϵ, corresponding to a specific ratio of Cr you would be able to obtain the effectiveness
the value of the effectiveness ϵ of a heat exchanger. And once ϵ is known then you can
use simply the definition of ϵ to evaluate, what are the outlet temperatures, what are the
temperatures of the 2 outlet streams. And once you have that then the total amount of
heat transfer, the calculation of total amount of heat transfer is fairly straight forward.
So, I hope you understood the concept of effectiveness how ϵ-NTU method can be used
for the design, specially for the evaluation of the unknown values of the outlet streams
and so on. We would solve a few one more problem on ϵ-NTU method as tutorial and
then move on to the next topic.
530
Heat Transfer
Lecture - 46
Tutorial Problems on Epsilon - NTU Methods
We are going to continue on our study of ϵ-NTU method. I have solved one problem on
ϵ-NTU in the previous class. This class is also going to be a tutorial in which I will solve
2 more problems on ϵ-NTU method. So, hopefully by the end of today, you will have a
fair idea of how to solve problems involving ϵ- NTU method. So, the first problem that
we are going to deal with is a concentric pipe heat exchanger. So, it is a double pipe heat
exchanger. So, we do not know whether it should be operated in counter flow or in
parallel flow. Some of the properties are given for example, what is the mass flow rate of
the of one fluid, the mass flow rate of the other fluid, the C p heat capacity of one and the
heat capacity of the other.
Some of the temperatures are also provided, in fact 3 of the temperatures are provided.
We need to find out what is the maximum possible heat transfer rate which is possible,
which can be achieved in this heat exchanger also what is the value of ϵ. And secondly,
we have to decide whether this should be operated in counter flow or in parallel flow. So,
that is in a nutshell is a problem, first find the maximum possible heat transfer rate and
secondly, whether it should be run in counter flow orientation or in parallel flow.
531
So, now come to the numbers, over here what you see in this problem is it is a concentric
& p
mC & p
mC
tube heat exchanger, the cold fluid , the hot fluid both are given. And
temperature of the fluid which is entering, the cold fluid which is entering is 40º, the
temperature of the outlet of the cold fluid is 95 and for the hot fluid the inlet temperature
is mentioned as 210. So, if you see these 3 temperatures are given so; obviously, by
simple heat balance the fourth temperature can be obtained. However, in order to find the
maximum possible heat transfer rate and finally, to find out the ϵ we need to make our
calculations based on ϵ- NTU method and secondly, we need to find whether it should be
operated in counter flow or in parallel flow. So, I started the at the first part of the
problem which is to find out what is the maximum possible heat transfer rate.
So, we know that from our previous studies the maximum amount of heat transfer which
is possible in any heat exchanger is

qmax  Cmin Thi  Tci 
& p
Cmin  mC
Where,
So, the maximum possible temperature change that can operate provided you give
enough length of it is going to be T hi, which is the temperature of the hot fluid at the
& p
mC
inlet and the temperature of the hot fluid at the outlet. So, all these etcetera are
given. So, here if you look at the cold fluid and the hot fluid; obviously, the minimum of
& p
mC
has to be for the hot fluid.
532
& p
mC
So, for the hot fluid is less than that of the cold fluid therefore,
Cmin  m&h Ch
And therefore, q max can simply be calculated as
 
qmax  m&hCh Thi  Tci  0.125  2100(210  40)  44625W
So, when you calculate the numbers it is going to be 44625 watts. So, this is what we
have as the maximum heat maximum temperature that can happen and in fact, this
should be Tci. So, therefore, this is the maximum temperature drop. So, the maximum
& p
mC
temperature drop multiplied by Cmin, which is the minimum of and when you put
this numbers, this is the maximum amount of heat which can be transferred in the heat
exchanger. And of course, this is an idealized situation in order for the temperature of the
hot fluid this temperature difference to be achieved in any heat exchanger, you have to
provide infinite length of the heat exchanger pipes.
So, as we have discussed, the ϵ, the maximum possible heat transfer that is achieved will
take place, when the temperature of the hot the maximum temperature difference in
existing in the system is achieved. So, therefore, in counter flow that is only possible in
counter flow when the temperature of the hot fluid at one end, and the temperature of the
cold fluid at the other end at the inlet condition they both will change. And, the
temperature of the cold fluid and the temperature of the hot fluid at the inlet this
difference if, we can exchange all the heat between these two fluids that is the maximum
&
possible heat transfer which is given by the formula that I have just written it is mCmin
multiplied by the temperature difference maximum temperature difference that is
possible in the system.
533
So, therefore, the next part is rather straight forward. The ϵ by definition is
q 
m&cCc Tco  Tci   m& C

qmax

qmax
h h T
hi  Tho 
But since, the T (hot) at the outlet is not known to me at this point. So, therefore, I am
simply using these values since they are known to me. So, what I do next is I will simply
put in the numbers over here. So, the ϵ would simply
0.125  4200  (95  40)
  0.65
44625
So, that is the value of ϵ, which can which is obtained in the present scenario. In order to
do the second part that is whether it should be run in counter flow or in parallel flow.
534
You would prefer a heat exchanger which would require the least amount of heat
exchange area. So, the requirement of heat exchange area is the prime factor in deciding
whether, to go for counter flow or to parallel flow. And of course, as an engineer in order
to reduce the cost you would go for that type of flow, which would result in the
minimum heat transfer area for a specific operation.
So, our next job is to find out what is the ratio of the 2 areas, required for counter flow
and required for parallel flow and see which one is the smaller one. The one which
would give the smaller area of heat transfer, smaller heat transfer area requirement would
be the preferred mode of operation. So, we would calculate: what is the heat exchange
area requirement for parallel flow, as well as for counter flow. So that is what we would
do next and to decide which one it should be. So, let us start with this area requirement
calculation. We understand that
q  UATLM
So, this is the log mean temperature difference, this is the heat transfer area which is
required. So, therefore,
q  UAPF TLM PF  UACF TLM CF
Where the subscript PF refers to parallel flow and CF refers to counter flow.
So, this is the relation which we are going to use and therefore, I need to see which one
of these two is going to be the smaller one. In order to do that, so I find out what is a
counter flow divided by a parallel flow and what I see is that this U will cancel. And
therefore, it is going to be the inverse ratio of the log mean temperature difference for
these two cases. Therefore, this is simply going to be
ACF TLM PF

APF TLM CF
So, this area if I could evaluate it would give us the idea of which one is going to be the
more efficient one. Now the unknown in here is I still do not know what is the value of
Tho is unless I calculate what is Tho, I will not be able to evaluate these log mean
temperature differences. And the Tho the heat transfer the temperature of the hot fluid at
the outlet would simply be
m&cCc 0.125  4200
Tho  Thi 
m&hCh
 
Tco  Tci  210 
0.125  2100
(95  40)
Tho  100o C
Therefore, I have I now know what is the temperature of the hot fluid at the inlet,
temperature of the hot the cold fluid at the outlet, temperature of the cold fluid at the
inlet and temperature of the hot fluid at the
535 outlet. Since four temperatures are known to
me, I would be able to calculate what is the log mean temperature difference when it is in
parallel flow and when it is in counter flow.
So, therefore, to write this the area required for counter flow divided by the area required
for parallel flow would simply be
ACF

 (210  40)  (100  95) /  ln170 / 5 
APF  (210  95)  (100  40)  /  ln115 / 60 
ACF
 0.55
APF
So, when you do this when you evaluate this ratio it would turn out to be 0.55. So, which
simply shows, shows that the area required in counter flow is less than the area required
for parallel flow. So this is the one which should be the preferred mode of operation is
so, the preferred mode of operation is therefore, going to be the counter flow.
So, this simple problem would clarify some of the idea about the maximum heat transfer
rate which can be achieved, when the temperature of the cold fluid outlet reaches the
temperature of the hot fluid at the outlet. And then, what is the what is so that is the that
is the maximum possible heat transfer. Then you find out what is the actual heat transfer,
the ratio of these two would give you the idea of the effectiveness, ϵ of the exchanger.
Secondly, you are also finding out what is the area requirement in two cases, one in
counter flow the other in parallel flow. And you would see that they are; obviously, that
they are going to be in the inverse ratios of the log mean temperatures for these two log
mean temperature differences for these two cases.
536
So, quickly draw the profile for parallel flow and counter flow as I have shown you. Find
out what is  T1, what is  T1, calculate what is the ratio of the LMTD and that ratio
would tell you, what is the area which one would require the smaller heat transfer area
weather in parallel flow or in counter flow. And you would see that, the counter flow for
this specific case is going to require the lesser amount of heat transfer area. So, that is
going to be the preferred mode of operation for this heat exchanger. So, I guess this
concept is clear to you now, I will quickly solve one more problem and then we will
move on to the next topic. So, therefore, the next problem
537
which we are going to deal with is again it is a double pipe heat exchanger and it uses
hot and cold water.
The Thi the inlet temperature of the hot fluid is 200 ºC and the inlet temperature of the
&
cold fluid is 35 ºC, the flow rates of the hot fluid, so therefore, mh for the hot fluid is 42
&
kg/hour, mc is going to be equal to 84 kg/hour and the value of overall heat transfer
coefficient is provided as 180 W/m 2K. So, the first one is find out what is the maximum
&
heat transfer rate. Now if you look at the values of the mh , this one is going to be less
half of whatever be the cold water flow rate. So, therefore, the maximum heat transfer
rate, qmax, as we understand is going to be
42

qmax  Cmin Thi  Tci  3600
 4302  (200  35)
qmax  8281W
So, these two temperatures are known to me and note here that both the outlet
temperatures unknown. So, this problem is a prime candidate for solution using ϵ-NTU
&
method. So, and this C minimum must be equal to mh since its flow rate is half of that
and the value of Cp is going to be roughly the same. In fact, the value of C p for the hot
fluid is equal to 4302 J/Kg-K and Cp for the cold fluid is 4186 J/Kg-K So, therefore, the
product of these two is definitely going to be less than the product of these two. So, C
minimum is the hot one and therefore, if we put
538
this in there,
42
 
qmax  Cmin Thi  Tci 
3600
 4302  (200  35)
qmax  8281W
This qmax is equal to 828 W. So, that is straight forward, the second one is if the heat
exchanger is being operated in counter flow with heat transfer area to be is equal to 0.33
m2, find the outlet temperatures. So, the problem states that the above heat exchanger is
operated in counter flow with an area of 0.33 m 2, the value of U has already been
provided find out the outlet temperatures. Now we can we can clearly see that, this is
again ϵ-NTU method should be the preferred one because, both the outlet temperatures
are unknown to us, I do not know what is T ho and also what is T co.
So, we have to use ϵ- NTU method for this case. So, in order to find out the ϵ- NTU
method, first we have to find out, what is the value of Cminimum and what is the value of
Cmaximum. Thus,
42
Cmin   4302  50.19W / K
3600
84
Cmax   4186
3600
Cmin
 0.514
Cmax
539
So, that is a straight forward substitution of the values. The next is, we also know that
from definition the NTU the number of transfer unit is defined as
UA 180  0.33
NTU    1.184
Cmin 50.19
Therefore, I know my values of NTU, I know that this is being operated in counter flow
Cmin
 Cr
and C max is also known to me. So, when all these are known to me, either you can
read from the graph or you can use the relation which is provided for counter flow
operation in a double pipe heat exchanger.
So, the expression that I am writing from the text and I am again mentioning that, if such
a problem comes, the expression connecting ϵ and an NTU would be provided. You do
not have to memorize this and you only have to identify that this is a counter flow
situation and this request you need to find out what you have the value of NTU, you
know what is the value of Cr, that is the ratio of Cminimum and Cmaximum. So, choose the
relation which represents this situation and try to use it to find either the unknown ϵ or
the unknown NTU. In this case the unknown is ϵ, so I am using a relation
1  exp[ NTU (1  Cr )]
  0.616
1  Cr exp[ NTU (1  Cr )]
Please look at these relations in your text and be sure to use the right relation for the
system at hand. So, when you put these values in here I am not doing all the calculations,
you would you would see that this ϵ would turn out to be 0.616. So, that is the value of
epsilon which you should get for such for after you put in the values in there. We also
know that, ϵ is the actual heat transfer which could be


Ch Thi  Tho 

Cmin Thi  Tci 
I will repeat once again, I have now the value of ϵ evaluated. I also know the definition

Ch Thi  Tho 
of ϵ, which is

Cmin Thi  Tci .
So, this in here, you would see that
 
Tho  Thi   Thi  Tci  200  0.616(200  35)  98 C
540
So, the temperature of the hot fluid can be evaluated as simply 98 ºC. And from energy
balance the other unknown is what is T co, the temperature of the cold fluid at the outlet,
so from energy balance you can write that
  
Ch Thi  Tho  Cc Tco  Tci 
Tco  87.2 C
So, this again shows you the utility of ϵ- NTU method. In this specific case you do not
know either of the outlet temperatures, what you know are the mass flow rates, the
specific heat capacity, the overall heat transfer coefficient and the heat transfer area for
heat transfer area available for such operations.
So, using an LMTD approach to solve for this would require a long iterative procedure.
So, ϵ-NTU provides an alternate, elegant and shortcut method to find out what are these
two unknown temperatures. So, all the problems that I have solved in the last class and in
this class and underscores the utility of ϵ -NTU method and the concept of ϵ is very
important which is just the ratio of the actual heat transfer divided by the maximum heat
transfer possible in an exchanger provided you give enough length; mathematical
speaking ideally infinite length such that the maximum achievable temperature drop is
possible. Once you have the value of ϵ known to you and the other parameters are known
to you, then you should be able to calculate: what are the outlet temperatures for the hot
fluid, as well as for the cold fluids.
So, I hope that with these solved problems you are in a position now to solve problems
on your own and I would request you to look at your text book or any other text book in
heat transfer, see some of the problems which have been solved in the book, as well as
do exercise problems on your own. And, also do they are the tutorial sheets which would
be provided and the assignments which would be provided. And, if there are any
questions the TA and I would be available to answer them online and if there are any
confusion
541
please do not hesitate to ask us and we will try to help you to master this specific
technique which is commonly used when both outlet temperatures are not known to us.
So, in the next class we will discuss some other modes of some other types of heat
exchange equipments which are also very important in industry. For example,
evaporators, the forward feed evaporators, the backward feed evaporators, their
advantages and disadvantages and types and so on. And after that is over we will then
move on from convection to the last part of this course, which is going to be the radiative
heat exchange between surfaces.
542
Heat Transfer
Lecture – 47
Boiling, Evaporation and Evaporators
This would be the last class on Convective Heat Transfer. What we have covered in
convective heat transfer so far are when we have flow outside of a plate or a tube which
is external flow, internal flow, the concept of fully developed flow the relations and
correlations which are applicable both in laminar flow and in turbulent flow for flow
outside as well as flow inside. And, we have seen what is, how to design a heat
exchanger based on the log mean temperature difference method and if both the outlet
temperatures are not known then a better method known as epsilon NTU method or
effectiveness NTU method can be adopted.
And, we have solved a number of problems and you would be it be given more problems
in the tutorial sheet which you try should try to solve and if there is any difficulty we will
discuss about it, but we will close the convective heat transfer today with something that
we have not discussed that much in our previous studies this is about evaporation and
boiling heat transfer. The utility, the importance of evaporation and boiling heat transfer
would be more apparent in the next half of this class when we talk about the evaporators.
As you are probably aware of evaporators is one of the major equipment, major unit in
many heat transfer, many chemical plants were a solution a dilute solution is going to the
concentrated by evaporating part of the solvent thereby making it more concentrate. So,
the applications of evaporators, the types of evaporators and the reason for concentrating
a specific product in a dilute solution we would discuss about it.
So, the first starting point is to know a little bit about evaporation and boiling in we are I
am also going refer to a specific experiment which is a very famous experiment where
we would see how the boiling curve that is what the heat flux removed from a wire
which is submerged in water how does it vary when you change the temperature
difference between the wire which can be heated electrically and that of the surrounding
water which is assumed to be saturated.
543
So, at any given point of time you can control the amount of heat which passes through
the which is generated due to the passage of current through the wire and at steady state
the amount of heat which is provided to the thin wire most of the cases it is a nichrome
wire of high resistance. So, the joule heating is we would see how the joule heating or
how the dissipation of heat from the wire follows the specific trend with delta T which is
the temperature difference between the wire and the surrounding liquid that has been
imposed on the system and will get some interesting or interesting idea out of that
boiling curve.
The evaporation is also known as quiet boiling in which case most, let us say you take a
beaker of water and you place it on a hot plate. So, near the near the interface with the
hot plate the temperature is going to be almost close to that of the hot plate, but this
temperature would change drastically if you just move up a little bit. So, a large change
in T can be observed in a region very close to that of the heater. Beyond that point the
temperature more or less remains constant and when you come close to the interface it is
assumed that the interface is at equilibrium with its vapour.
So, interface temperature is for the case of evaporation the interface temperature most of
the times is approximated as the saturation temperature of water or the liquid that we are
using or for the given conditions. So, the temperature profile if you could draw it as a
function of the distance from the heater in the beaker then it is going to be very short
near the region of the hot plate then it more or less follows a constant pattern and there is
a small temperature jump which most of the times it is neglected. So, T sat is going to be
equal; T sat is going to be equal to T naught that is the temperature of the on the liquid
side of the interface.
So, this is how the evaporation is perceived and evaporation is a slow, as you can
understand evaporation is a slow process it is totally different from that of the boiling,
where in the case of boiling you have phase change taking place everywhere in the liquid
column that we were just discussing. But, in the case of evaporation phase change is
taking place only from the top of the liquid and there are numerous relations and
correlations which connect the heat which is dissipated from the heater as a function of
delta T.
544
So, these relations and correlations are available in the text that is the book Fundamentals
of Heat and Mass Transfer by Incropera and Dewitt and also on any other book or text
book on heat transfer. So, we can take look at the relations and whenever you have to
find the heat transfer coefficient for such processes go to the go to the reference, find out
which relation or correlation is appropriate for the specific situation and use it.
But, we are going to now talk about the case of boiling curve where a wire which is
submerged in water the current is being passed through the wire which heats up and you
can measure what is the, you can find out how much of heat is being lost by the thin wire
to the surrounding fluid. And you can calculate, you can measure the variation of this q
as a function of delta T and we would like to see how it changes and this has given us the
boiling curve of what of boiling curve for the case of water.
So, near the vicinity of the wire, if this is the wire then near the vicinity of the wire what
are the different steps that we can encounter, that we will encounter as the temperature of
the wire is progressively increased. This is what we is going to give rise to the boiling
curve.
545
So, let us look take a quick look at the boiling curve for saturated water at atmospheric
pressure. So, we can see on the left-hand side I have plotted the heat flux which is q
double prime s in Watt per metre square and delta T is the x axis the independent
variable, where delta T is defined as T s minus T sat. So, temperature of the surface
minus T sat. So, initially up to a delta T of above 5 degree the heat flux gradually
increases; slowly increases with delta T. So, this kind of relaxed increase of heat flux
with delta t this region is mostly governed by free convection. So, up to about 5 degree
Centigrade temperature difference it is a free convection which is which takes which is
responsible for heat transfer.
So, you have the setup could be like this it is just a schematic of the setup. So, you have a
specific wire of high resistivity high resistance which is connected to a power source and
this is submerged in water. So, you pass a current through it and from the voltage and the
current V times A is going to give you the amount of heat that is being dissipated by the
wire in the water. And the T sat is the saturation temperature and the q this divided by
area would give you can also find out what is Q double prime that which has been
plotted over here.
546
So, when the temperature difference is about 5 degree it is free convection, but as the
temperature difference is increased the first bubbles will start to appear on the wire and it
is going to be the onset of the nuclear boiling nucleate boiling. So, over here what you
are going to get is onset of nucleate boiling. So, this point is where the first bubbles will
appear, but the bubbles are going to be detached from one another. So, bubbles are going
to form and the bubbles will get released and they will try to move towards the top
because of (Refer Time: 09:48).
But, since they heat super heat content of the bubbles are not that great specially at low
values of delta T they will disintegrate and collapse back to become the liquid once
again. So, this formation of bubbles and the movement of the bubbles will ensure that
your q, the slope of the q double prime s versus delta T increases more rapidly with delta
T as compared to the free convection. So, as you increase it beyond that as you increase
the current through this even more.Then what you are going to have is bubbles which are
going to form and the bubbles are going to get detached and there is going to be a lot of
mixing in over here. So, disturbances near the interface should increase and as the
disturbance is mixing or and the interference between the bubbles increase at this point.
So, you are going to get higher and higher values of heat flux as a function of delta T. So,
it is going to increase rapidly, but as it is doing so a point would come where the wire
will see, will be covered by almost like a blanket will start to get covered with a blanket
of the vapour. So, this is your wire which is going to be covered going to start to getting
covered with a blanket of the vapour.
So, therefore, there would be the problem of the liquid reaching the surface changing its
phase and forming the bubble. So, the additional resistance provided by the low thermal
conductivity vapour bubbles clinging to the surface accumulating on the surface. So,
there are two competing mechanism; one is the bubbles are taking the vapour bubbles are
change I mean it is liquid is changing phase creating the liquid bubble. The liquid
bubbles will take the latent heat and will rise in the liquid there by creating the keeping
the surface cool.
The other competing mechanism would be, the competing mechanism would be the
bubbles forming an additional layer of resistance on the wire. Therefore, these two are
547
acting at cross purposes formation and release of the bubbles and formation of a blanket
of the vapour over the solid which increases the heat transfer. These two are the
competing mechanism and they will balance out resulting at some point in the maximum
heat flux which you can obtain for such cases and the point where this happens is known
as the critical heat flux.
So, in critical heat flux is the which takes place when these two more or less cancel out
each other and this is what is known as the as the critical heat flux and from the nucleate
boiling you start. So, this is the onset of nucleate boiling and over here in critical heat
flux you are going to go from nucleate boiling nucleate towards film boiling. So, critical
is the transition between the nucleate and the film boiling.
If you go beyond that the this region is marked as dotted because in most of the cases it
is very difficult to get this region which is an unstable region and there as the
temperature difference increases the thickness of this vapour layer will simply be more.
It will keep on increasing and the resistance due to the entire wire is going to get
blanketed by a layer of the vapour with very low thermal conductivity. So, with increase
in delta T more bubbles are formed and the heat flux keeps on decreasing.
So, the point where you are going to get the minimum heat flux for such a situation is
known as the Leidenfrost point. So, Leidenfrost point which simply says that this q
double prime is going to be minimum. Once you cross this the mode of heat transfer is
going to be I am mean that radiation would start to play a role. So, the increase in delta T
the heat flux will start to rise again because your T delta T is more. So, therefore, the
change in delta t the increase in delta T will start to compensate for the lower heat
transfer coefficient resulted because of the formation of this. So, it will start to increase
and beyond certain point the radiation is going to take place, take over and it is going to
be very very steep curve from there.
So, somewhere around this point; somewhere around this point you are going to get the
burn out of the nichrome wire which is this one the wire over here. So, when that
happens this is roughly a measure of what is the heat flux. So, this is connected to a
voltmeter to a power source. So, you slowly increase the power and you see how this is
changing and when at certain point the wire brakes that is the point which is known as
the critical which is known as the critical heat flux.
548
So, the knowledge of critical heat flux is extremely important in the design of
equipments which involve boiling. So, whenever you have a boiling situation for a in a
heat exchange equipment for example, in an evaporator you need to know what is the
maximum heat flux that this surface can handle without going to a going to a point where
which change if increase in delta T your heat flux will decrease. So, we are going to be
on the other side of the boiling curve rather than with change in delta T heat flux is
increasing if you are at or near the critical heat flux a change in delta T can be counter
productive instead of getting more heat flux you are going to get less heat flux.
So, the limitation of your boiling equipment in order to in order to evaluate the point up
to which your system, your equipment, your designed equipment evaporator can work
you really need to know what is the critical heat flux for such as such a point. So, the
important points here are that through this experiment you have an idea of the different
modes of heat transfer starting from natural convection to forced convection. And, in
forced convection you are going to get the onset of nucleate boiling, where the bubble or
vapour bubbles are going to form, detached from the surface, rise towards the top, but
will probably assimilate once again by a face change process with water.
So, more number of bubbles means more heat transfer. So, with the change in delta t in
the nucleate boiling region your temperature your heat flux keeps on increasing, but has
that happens there is going to be formation slowly that is going to be formation of a
vapour blanket on the wire. Therefore, these two competing mechanisms would try to
make it try to come to come to a balance. And, the point where the maximum occurs is
known as the critical heat flux beyond that with change in delta T the heat flux will keep
on decreasing or at the most will remain more or less a constant. So, you are not utilizing
anything, getting anything new out of your equipment by having it operating near at or
near the critical heat flux region.
So, designing of heat exchange equipments specially which those which involve boiling
would require that you have a fair knowledge accurate idea almost what is going to be
the critical heat flux for such situations. And again, there are relations and there are
correlations available which would give you the value of the critical heat flux for some
of the substrates liquid combination and you can look at your text to find out what they
are I am not going to write all of them down. What I am going to do next is to show you
some quickly little bit about the evaporators their designs and so on.
549
So, let us move on to evaporators and we will just deal with chemical evaporators and in
chemical operators are of two types one is natural circulation, it can be a single or a
multiple effect; I will talk about what is single and multiple effects are. And the second
obviously, is going to be forced circulation where you require higher throughput, higher
values of heat transfer you are going to get you are going to use this.
So, the natural convection type I, if we start working with natural convictions some of
the possible exchanges are going to be this is the simplest one, where you have a liquid
up to this point and then there are steam which is coming in and you have pipes or tubes
like this. So, steam is going to pass through these tubes come to the other side and
whatever is remaining it goes out of the vent, whatever condense is the condensing steam
is going to come out as drips. The feed comes in here at the top and the concentrated
product which is known as the thick liquor come can be collected at the end.
So, this is a horizontal tube this is the horizontal tube heat evaporator you can also have
as a name suggests you can also have tubes like this where steam is going to be go in and
out and so, the same thing is on this side. So, this is filled with liquid, it is for the vapour
to go come out of this. So, you have steam which is going so, I will write this the steam
will is going to pass through them and you have the same thing on the right hand side
and the liquid is going to come down rise due to natural convection reach over here
change its direction and come back again.
550
So, this kind of circulation will go on in the intervening space between the tubes and
therefore, you have the feed which is coming in here and the thick liquor would be
collected from the bottom and you have steam which is in here. So, steam comes in here
and this is obviously, going to be the drips and any remaining vapour can come out of
this. So, these are two examples of natural circulation evaporators and in many of the
cases the vapour which comes out from these streams, these vapours can then be used in
another evaporator which is in series which is after this one.
So, instead of instead of having just one evaporator which is also known as the single
effect evaporator you can have multiple effect evaporators, but more number of
evaporators 2 3 4 are going to be in series and the vapour from the evaporator 1 is going
to be used to heat that of that of 2 and so on which I will draw, I will show you.
So, the forced circulation; when we look at the forced circulation one you simply have
the design construction would approximately remain the same. So, in a forced circulation
type of evaporator you still have the steam and the liquid is forced by a pump which goes
through the space between them, goes over to over here to another chamber where the
vapour is collected.
So, you have the liquid over here and you keep on circulating; you keep on circulating;
you keep on circulating this and it gets more and more heated. The vapours are going to
be collected this is a drip from the feed or any you can give some you can provide the
551
feed from over here and these kind of systems are used where the concentration problem
involves a solution with poor flow. And it may have a it may form scales it may have a
thermal undesirable thermal characteristics or it is highly viscous. So, you really need to
force the liquid through the spacing where it is in connect in contact with the steam
through these tubes.
So, for highly viscous fluids this kind of poor circulation would be an ideal evaporator
where you force the feed to go through this collect the vapour and you can continuously
reduce the solvent content of the material and therefore, you can have a system in which
it will automatically start reducing the water content of this. So, as I said you can have
multiple effect evaporators as well. So, one reactor, one evaporator after the other, so,
you get a feed and the feed comes in contact with steam. The feed temperature has also
increased the vapour which is generated from the feed contains some significant amount
of energy. So, vapour from the feed is then going to go as the heating material into
another evaporator which follows the first one. So, where the feed from the where the
vapour from the first evaporator is going to come and heat the material and so on.
So, in that way you can utilize more and more heat the heat efficiency of the entire
system will enhance. And, you can have a system in which both the feed and the steam
will come in the same direction will travel in the same direction or we can have a system
of the evaporators. So, series of evaporators where the feed is going to come from one
the feed is going to come to the first evaporator where the steam is going to come to the
last evaporator. So, you can have a forward feed or a backward feed.
So, I will quickly draw the pictures of the figures of these two forward feed evaporator,
backward feed evaporators and talk about their merits and demerits and once I draw the
pictures I think the concept will be very clear to you what is known as what is the
multiple effect evaporator and what is going to be forward feed and what is going to be
the backward feed. So, let us quickly draw these pictures and try to see what we get out
of this.
552
So, this one is as you can see it is the forward feed and this is the backward feed. So, let
us see what happens over here the feed and the steam both come in evaporator 1. The
concentrated product which is slightly heated because since it is come in contact with
this steam goes to the next evaporator. The vapour generated out of this contains lot of
energy. So, it goes to heat up whatever is coming at this point and it goes into number 2.
The same way the product is going to go as the feed to the third evaporator and the
vapour which is generated in 2 goes as the as the heating stream into 3.
So, finally, what you get out of 3 at the bottom is the final product and any vapour which
leaves 3 is going to be cooled using cooling water and then it can be either it can be
brought back to the product or it can be stored elsewhere. So, once again in forward feed
both the in what in forward feed both of them travel in the same direction and the liquid
feed flows in the same direction as that of the vapour it is a forward feed.
So, the vapour and the liquid; the vapour and the liquid are travelling in the same
direction. So, it is known as the backward feed the forward feed. In the backward feed
this steam comes at 1, the feed comes at 3 and there can be more such evaporators in one
after the other. So, when the feed comes it is cold it is first going to get it is first going to
encounter the vapour which is coming from. So, this is the vapour which is coming from
2. So, the feed gets slightly concentrated, you are going to it is going to lose some
amount of some amount of vapour which can be condensed and collected.
553
So, the feed then slightly heated now it is pumped through 2, 2 as the new feed. The
heating one comes from the vapour of 1. So, into the concentrated part from 3 is the feed
and the vapour from 1 is the heating medium even more concentration being done at 2
and then it is pumped to 1. So, when it goes into 1, the new feed it has the maximum
concentration and there in it comes in contact with live steam it gets further concentrated
and it is this is going to be the final product.
So, let us see the relative advantages and disadvantages. First, backward feed: the feed
which is coming is the feed which is coming is at all it is talk about the forward feed
first. The forward feed a vacuum is maintained in the last effect this affect and the liquid
flows itself from 1 to 2, 2 to 3 and so on in backward feed you need a pump to make the
liquid flow towards from 3 to 1, and in order to remove this feed at the final, at the
product you probably need a pump at this point, whereas you required pumps after every
effect in the case of backward reactor.
If the feed is at a higher temperature is higher than the, if the feed liquid itself is higher
than the saturation temperature then some evaporation some flashing will take place
automatically as in 1 and this flashing will take place in every effect thereby reducing the
steam requirement. So, if your feed is at a higher temperature than the saturation
temperature the moment it enters the first effect it is going to flash and that is going to be
part of the vapour which is going to evaporate and then the rest is going to condense and
the same thing will happen in 2 and 3 and since the feed is at a higher temperature than
that of that saturation temperature your steam requirement would be less the problem that
we. So, that is advantages.
The problem that you see in forward feed is as the product from 1 becomes the feed of 2
the product of 2 becomes the feed of 3 and then ultimately you are going to get the final
product, the concentration keeps on increasing between 1, 2 and 3. So, the liquid starts to
become more and more viscous therefore, it is difficult to make it flow. So, the last ones
is it has the highest viscosity liquid the it will have the least overall heat coefficient since
the liquid is going to be very concentrated at this point.
On the other hand in backward feed the dilute liquor enters at the last at the last end and
leaves concentrated from the first which is the hottest one which is at the highest
temperature. So, here the liquid must be heated in each effect and you would require
554
more steam in backward feed, but if the feed is initially hot therefore, if the feed is
initially hot then it is there is no point in letting it enter into 3 in backward feed. So, if the
feed is initially hot use forward feed, if the feed is initially cold and or if the fluid is
viscous introduce it here.
So, as the liquid becomes more and more concentrated the temperature is also increasing
so, therefore, the effect of increase in viscosity, the detrimental effect of increase in
viscosity is compensated in backward flow as compared to that in forward flow. So,
steam cost will be less for backward feed if the feed is cold and less for forward feed if
the feed liquor is approximately at the operating temperature of the first effect of higher.
So, depending on the condition of the feed you have to choose whether you are going to
use a backward feed or a forward feed. So, if your feed is hot at a high temperature
relatively high temperature then by all means use forward feed, where flashing is going
to take place inside evaporator one and that vapour is going to be utilized in 2 and so on.
Therefore, the steam requirement at each of these effects will be lower because of vapour
flashing which can take place only when the liquid feed liquid is at high temperature.
But, as the in forward feed as you move from 1 to 2, 2 to 3 it is becoming more and more
concentrated viscosity increases and therefore, your heat it heat transfer will keep on
decreasing. So, if you are dealing with dealing with a liquid which when it is
concentrated, when it is when the solvent is evaporated it becomes highly viscous do not
use forward feed. In that case if it use of a backward feed would be justified because as it
goes from 1 to 2 to 3 in the reverse direction the temperature keeps on increasing and
therefore, and the concentration keeps on increasing as well.
So, the increase of viscosity due to concentration increase will be offset by an increase in
temperature. So, use of a backward feed would be justified in such a case. So, this is in a
very the nutshell is about evaporated, we spoke about boiling, we spoke about
evaporators, the type of operations the natural convection the evaporators which are
which depend on natural flow, natural convective flow. And in some cases you have to
have forced flow forced convection effects forced convection convective evaporators and
the arrangement of multiple evaporators either in forward flow or in backward flow with
their advantages and disadvantages. The relations and correlations are available in the
555
text which I am not reproducing over here, so, you can simply see those in your text and
use them whenever appropriate.
So, this concludes our discussion on convection and from next class we will move over
move over to Radiative Heat Transfer.
556
Heat Transfer
Lecture – 48
Radiation - Fundamental Concepts
We will start Radiation for in this class and follow it towards the end of this course
which would complete our study on Heat Transfer. So, far we have dealt with two modes
of heat transfer, conduction and convection both of which require the presence of a
medium. However, radiation is the one as you are all aware of does not require the
presence of a medium. So, radiation plays a major role in many of the process industries
and in significantly varied applications as we know of. So, we would concentrate on
radiation, the radiator the radiation flux, the intensity and flux, the radiative properties.
How do we characterize the subject based on its radiative properties? What is an ideal
surface as far as radiation is concerned? We are talking about black body over here. And
then there is any surface resistance to radiation that by which you can express, since,
radiation depends on temperature and on the nature of the surface. So, we will see what
is Stefan Boltzmann law that; obviously, you know about. But, is there any surface
resistance to radiation since all these surfaces and non ideal, they are not black bodies.
So, can some resistance be ascribed for exchange of radiative heat transfer between the
surface and its surroundings.
So, what is the surface potential? What is the surface resistance? What is the potential
just outside of the surface when we consider the surface resistance? So, the body as a
whole if I consider it as a black body we will have some sort of radiative potential. But,
since its surface is not acting like a black body. So, there will be some resistance. So, just
outside the surface we can think of another potential with two potentials connected by
resistance as in a circuit, electrical circuit where the resistance in here is going to be
surface resistance to radiation. So, those concepts we would also discuss in this part of
the course.
Also two bodies can exchange radiation with in between them, that they can also
exchange radiation with itself and they can they can also exchange radiative heat transfer
with the surroundings. So, if the palm of my hand, if the palms of my hand are two
surfaces
557
are two different temperatures. And if I press them like this, there would be radiative
heat exchange in between these two surfaces. Not only that they are also going to
exchange heat with the walls of this room through the space in between this. So, if I
bring them close to each other, the amount of radiative heat which they are going to
exchange with the surrounding will keep on decreasing.
So, in ideally, if they are touching each other then all the heat which is released by one is
going to be intercepted by the other. But, if I separate them out together then it is going
to be, then the heat that gets transferred from one the fraction of the heat which gets
transferred from one reaching two that fraction will decrease. So, the how much of
surface 1 is visible from surface 2 would definitely depend on what is the distance
between the two or if we talk in mathematical terms what is the solid angle subtended by
that would by this by surface 2 onto 1. So, we will see the concept of solid angle as well
in this course.
And then let us say we have 3 surfaces which form an enclosure. So, you can think of it
as a triangle, where the 3 sides are exchanging radiative heat with one another ok. So,
part of the heat released by 2 is going to go to 1, part going to 3 and maybe it may so,
happen that part of the heat released by 1 is going to be absorbed by 1 itself. So, if it is a
curved surface like this then the part of the energy which is emitted by 1 is going to
incident on 1 itself right. So, that is possible, but if it is a flat surface then none of the
radiative energy which leaves 1 is going to come back to 1.
So, the nature of the surface, the shape of the surface will also play a role; let us say we
have 3 flat surfaces which are forming a triangle. And these 3 sides are at 3 different
temperatures and there is going to be radiative exchange of heat in between them ok.
What is going to be the net heat to be supplied or extracted to each of these triangles,
each of these sides of the triangle so, as to maintain thermal equilibrium? These
informations will be extremely useful for the efficient designs of furnaces, because as
you know furnaces operate at a very high temperature.
So, how much of heat you need to supply to the furnace to maintain the temperature of
one of the surfaces at a desired level; you need this kind of calculations. So, as you can
see we are always bringing in electrical analogy for radiative calculations specially when
we talk about the potential, the resistance and the heat flow is something which is similar
to
558
that of current. So, there has to be a network method, see if you remember your electrical
technology this star delta connections the problems on star and delta connections which
we have done in electrical technology, we are going to do similar calculations for the
case of radiation as well which would give, which would allow us to effectively design a
furnace.
So, that is more or less what we would like to cover in this part of the course. But, let us
start with radiation first and radiation is the energy emitted by an object which does not
require the presence of a medium and all surfaces above absolute 0 would emit radiation.
They will also receive radiation so, there is going to be a process of emission and
absorption that are going on that will be going on.
The concept of radiation can be explained by 2 theories, the first one is Maxwell’s theory
and the second one is Planck’s theory. So, Maxwell’s theory of electromagnetism can be
used to express some of the radiation phenomena where Max Planck’s theory is also
applicable for the case of radiative heat exchange. So, apart from the properties of the
surface we should also be aware from the very class first class itself that the radiation can
be explained either in Maxwell either using Maxwell’s theory of electromagnetic,
Maxwell’s electromagnetic theory where the radiation is treated as electromagnetic
waves.
559
So, radiations are treated as electromagnetic waves in Maxwell’s electromagnetic theory
wherein max Planck’s theory the radiation is treated as photons or in other words they
are Quantas, quanta of energy. So, both we will find its use, both concepts have been
utilised. So, the results from the electromagnetic theory are used to predict the radiative
properties of materials, while the results from Planck’s concept has been used to predict
the magnitude of radiation energy emitted by a body at a given radiation.
Now, radiations can happen over a large wavelength range and view on one side we have
the gamma rays and one on the other side we have radio waves. So, whenever an object
is emitting radiation the chances are that it is going to emit radiation over all the
wavelengths. So, radiation is spectral in nature by spectral I mean that there is going to
be a wavelength dependence of radiation and not all the energy coming out of a surface
is going to be equally spaced over all wavelengths. So, there will be certain wavelengths,
there will be a wavelength range in which most of the energy would be concentrated and
it may not be presented other wavelengths. So, the spectral nature of radiation is
extremely important and we will have to take that into account.
So, when you think of this spectrum of radiation and the way it is divided into different
different sections. So, there is a specific wavelength range where we can see that is the
visible wavelength range. There is a specific wavelength range we in which most of the
radiative heat transfer is taking place. And you have an infrared range and you have the
gamma rays range where you have energy associated with radiation. But, the thermal
energy is concentrated somewhere in the middle and that is going to be our zone of
interest for most of the topics that we would cover over here.
So, let me draw what kind of waves, what kind of radiation we would get and as a
function of the wavelength and see whether we can identify and demarcate the region in
which the radiation is going to be important. So, if I draw that profile over here.
560
What it looks like then is. So, this is the spectrum of electromagnetic radiation. So, any
wavelength, any radiation where the corresponding wavelength, where the associated
wavelength is greater than 102, these are in the microwave region. Anything which is
smaller than 10-4 these are gamma rays. In between 10-4 and roughly somewhere slightly
more than 10-2 these are the x rays. So, between 10 -1 to about 102, this is the region which
is of most importance to this part of the course which is thermal radiation; so, between
10-1 to about 102.
And this region this one is the visible range where you have the red on one side and you
have the violet on the other side. So, this is more or less the visible spectra the more or
less the spectrum that you would get in radiation, but between 0.1 and 100 this is the
range in which most of the thermal radiation is going to be concentrated. And this would
be the region which we would study in this part of the course. So, as well as I was telling
you the spectral nature and the so called specular nature. So, this is monochromatic light,
monochromatic radiation emission and if I plot it as a function of λ which is wavelength.
There is going to be a variation like this some arbitrary variation. And since, it depends
on wavelength this is called spectral distribution.
561
What it essentially tells is that you are getting a continuous non uniform distribution of
monochromatic components, by monochromatic I mean single wavelength components
and the magnitude of the radiation at any wavelength and the spectral distribution vary
with the nature and temperature of the surface. So, this kind of distributions what you are
getting out of the surface, out of the object, would depend on whatever be the
temperature of the object and whatever be the properties of the object.
So, the monochromatic distribution of radiative emission from a surface is a continuous

function of wavelength, but the value of the emission is going to be different at different
wavelengths. So, that is what happens, that is what would happen in most of the real
surfaces. Whereas in some cases your and there is no directional distribution on top of it.
But, if you look at this surface then from a point it is this one will have a directional
distribution which is something different from over here.
So, here the distribution depends only on wavelength, here the distribution depends on
the direction as well ok. So, this is the difference between these two surfaces. So, this is a
spectral distribution and the spectral is used to refer to the nature of the dependence
where it depends on the wavelength. And the thick the wave nature of the thermal
radiation tells you that this λ which is the wavelength is going to be (c)/ (frequency of
radiation). So, this is the speed of light that is the standard definition speed of light.
The speed of propagation in the medium and λ is the wavelength and this is the
frequency. So, if the medium through which the propagation takes place is vacuum, then
that c is going to be speed of light in vacuum. So, the dependence between the
wavelength and frequency is simply an inverse relationship with a constant which is the
speed of light in vacuum when we are talking about the vacuum. So, now, let us talk
about the radiative properties, what are the radiative properties that we have? When a
light is incident on a surface part of it is going to get reflected, part of it is going to be
absorbed and part of it is going to get transmitted.
So, the fraction of the energy which gets reflected as compared to what is the overall
energy incident on it is known as the property is the reflectivity. So, the reflectivity is
defined as the fraction the numerator is going to be the fraction which is
562
reflected fraction of energy which is reflected and the denominator is the total energy
incident on it. So, the other option, other possibility is that part of the energy which does
not get reflected and enters through the surface is going to get absorbed. So, the
absorptivity, the factor absorptivity is defined as the amount of energy which is absorbed
in the material divided by the amount of energy which is incident on it.
Now, let us say that part is reflected; part gets into the system and then goes out of the
system or out of the material after part of it is going to be absorbed. So, what can happen
on incident energy on a surface is one it may get reflected, one it may get absorbed and
the third is it may get transmitted through the surface. So, the three properties which are
defined to denote these natures of the surface are reflectivity, absorptivity and
transmissivity. As the name suggests the denominator in all cases would be the amount
of energy incident on the surface.
So, in the case of reflectivity it is a fraction which gets reflected. Absorptivity is the
fraction which gets absorbed and transmissivity is the fraction which gets transmitted.
So, the reflectivity, the absorptivity and transmissivity, transmissivity all are fractions
having a value from 0 to 1. And since the denominator in all cases is the same, the sum
of alpha plus beta plus gamma tau is going to be equal to 1. So, if I define the properties,
ρ+α+τ=1
563
So, this is the reflectivity, this is the absorptivity and this is the transmissivity; so,
fraction of fraction which gets reflected, fraction which gets absorbed and fraction which
gets transmitted. So, that is why sum of these 3 would be equal to 1. Now, if I go back
one slide and show you this one here this one is this kind of surfaces where the emission
depends on the direction they are known as specular. So, specular is a surface where
there is a directional dependence of radiation from a surface.
So, wherever there is a directional dependence or directional distribution of energy out of

a surface it is known as specular. And when there is no such dependence it is called a
diffuse. A diffuse is a surface where the incident beam incident beam of energy is
distributed uniformly in all directions. So, apart from these properties of radiation, this
property of the surface is also important. Specular when there is a directional distribution
to the energy emitted by a surface. The second type of surface is diffuse where the
incident beam is distributed uniformly in all directions.
So, we will see what are the examples and the special features and the simplifying
situations where assuming a surface specular or assuming a surface as diffuse we will
provide. The other concept that I would like to introduce before I close this class is some
sort of an ideal. Whenever we talk about real surfaces the properties the absorptivity,
reflectivity and the transmissivity there must be
564
something which can be taken as an ideal surface ok. So, we are going to compare the
performance or the characteristics of a real surface against that of the ideal.
So, such ideal surface is in radiative heat transfer is assumed is termed as the black body.
So, what is a black body? A black body is something it is an idealized concept; a black
body is something which absorbs everything that is incident on it. So, therefore, the
absorptivity of a black body is equal to 1. On the other hand the second characteristics of
a black body is that at a given temperature no other object emits more energy than that of
a blackbody. So, blackbody emits the emissive power of the black body is maximum
compared to all other surfaces. So, black body absorbs anything and it emits the
maximum amount of energy.
The concept of black body is extremely important in radiative heat transfer. Because, it
lets you set a standard surface against which the performance of all other surfaces are to
be evaluated. So, let us note down the special nature characteristics of a black body. It is
an idealized surface, idealized concept and the first one is absorbs everything as I said.
The second one is maximum radiation at a given temperature that is the second one and
the third one that black body is a diffuse emitter. As I said the radiation from a black
body is a function radiation is a function of temperature and wavelength. But it is not a
function of direction.
If you again compare this, what we have done in this case the radiation takes place at all
wavelengths and the amount of radiation is going to be a function of temperature. So, if I
increase the temperature I am going to get a second curve which would show that it is
going to provide higher amount of energy so, emission. So, the radiation emission is
definitely a function of λ it is also a function of temperature. So, common sense tells us
that more the temperature higher is going to be the emission. Apart from that there could
be directional dependence of radiation what we call as specular, where there is the from
the point there is going to be different in different directions it would emit different
amounts of energy.
So, ideally the radiation is a function of temperature, it is a function of wavelength and it

is function of direction. But, in for case of a black body if the radiation emission is not a
function of direction, it is a function only of temperature and that of λ. So, when a
surface is not a function of. When the emission from a surface is not a function of the
direction it is
565
called diffuse emitter. So, in order to be a blackbody these 3 conditions are to be satisfied
that it should observe all incident radiation coming at all possible coming from surfaces
at all possible temperatures and all possible wavelengths.
And at a given temperature it is going to emit the maximum amount of radiation that you
can think of and it is going to be a diffuse emitter. So, examples of there are some
surfaces which closely mimic the behaviour of a black body for example, the closest
approximation of a blackbody. The closest approximation of a black body can be
obtained when you think of a small cavity. So, this is a small cavity which you can think
of it with a very small opening over here. So, any ray which comes inside is going to be
reflected multiple times and the chances of multiple time it is going to reflected and the
chances of its going back outside through the narrow pore is extremely small.
So, what it tells us is it gives us the situation in which any energy which comes inside is
going to be absorbed which it is going to be absorbed multiple times inside the system
and the fraction of energy which may go out is going to be extremely small. So, it obeys
or it confirms to the first approximation of a black body; that means, anything which is
incident on it is going to be absorbed over here. Now, let us think of the surface once
again the chances of the emission that you are going to get out of this is same in all
direction. So, if something leaves through this pore, it can be in any possible direction.
So, therefore, this is one of the closest approximations that you can get of a black body
and let us assume that you have placed an object over here.
In the one which enters is going to reflect multiple times and come to it from all possible
directions right. So, energy is going to be incident on any object placed inside the cavity.
So, therefore, it is going to be a diffuse irradiation of interior surfaces. Since, it is coming
all surfaces all points on the surface. So, the body placed inside the cavity will
experience evaporation from all possible directions. So, therefore, the black body is
going to be a diffuse emitter, since it is going to irradiate at the object inside the cavity
equally with equal probability from all directions.
So, normally a cavity is the closest approximation of a black body that one can get. So,
the properties of blackbody, the radiative properties of any real material, how they differ
from that of a black body and what is the closest approximation of a black body we
566
have discussed that. But, what is left to be said is, what is the black body radiation
intensity. Is there a formula, is there some way by which we can find out what is the
spectral and once again spectral means wavelength dependance. What is the spectral
radiation intensity of a black body? As we understand it is going to be a function of
temperature, it is going to be a function of wavelength. There it is going to be some other
constants which are going to be involved in that expression.
So, in the next class will see: what is the spectral radiative intensity of a black body and
how this can be connected with the radiation flux from a point source. If I put a point
source of radiation on a given surface, how much of total radiation I am going to get out
of the surface.
So, that is what we will cover in the next class.
567
Heat Transfer
Lecture - 49
Spectral Blackbody Radiation Intesity and Emissive Power
We will discuss more about radiation in this class. In the last class we just introduced you
to radiation, the requirement of not having any medium; that means, a radiation can take
place without any medium being present. The radiative properties of a substrate in terms
of reflectivity, absorptivity and transmittivity and the concept of an ideal body in terms
of radiation, which is termed as a black body which can emit; whose emission at a given
temperature is the maximum, it is going to absorb everything that falls into it. And, it is a
diffuse emitter; that means, the emission coming out of a black body does not have any
directional dependence.
We have also introduced the concept spectral; that means, anything which depends on the
wavelength at which we are considering at which the emission is taking place. So, all
these quantities, all these properties namely the reflectivity, the transmittivity and the
absorptivity are spectral in nature; that means, the absorption of a radiation is going to be
different at different values of the wavelength. So, the spectral nature of radiation
concept of a black body and that the radiation can be treated in 2 ways, the one is based
on the wave nature and the second is based on the quantum concept these were
introduced in the last class. In this class we are going to know more about two
fundamental properties of a black body.
That means, first is going to be: what is the black body radiation intensity, when a black
body is at a given temperature. So, how much of energy per unit area per unit wavelength
and per unit solid angle we will introduced the concept later on that this black body is
emitting. So, that is the spectral black body radiation intensity.
And secondly, we will see what is going to be the total emissive power of a black body at
a given temperature; that means, if a black body having unit area is placed at the center
of a hemisphere, how much would be the total energy emitted by this black body of unit
area at a given length and at a given temperature. So, these are the two main concepts
which would cover in today’s class.
568
So, let us first start with black body radiation intensity in which is going to so, we denote
that as I the intensity, b stands for the black body, lambda denotes the wavelength
dependence and of course, it is going to be different and different temperature. So, the
temperature has to be specified while we are discussing about the black body radiation
intensity. And in this case, we are going to rely on the Planck’s distribution. So, what is
black body radiation intensity? It is defined as the magnitude of radiation energy which
is emitted by a black body at an absolute temperature T at any wavelength lambda in any
given direction.
So, it gives you the idea the quantum of radiation energy which is emitted at in any given
direction and at a specific wavelength. So, the radiation energy of course, would
therefore, depend on the temperature, it is going to depend on the lambda, the
wavelength and it will have a directional property.
So, using Planck’s distribution the magnitude of the black body radiation intensity can be
expressed for the case of vacuum. So, the magnitude of the black body radiation
intensity, spectral black body radiation intensity at a given temperature into vacuum is
expressed in this form. Where h is the well-known Planck’s constant. So, h is the
Planck’s constant, the value of which is available in any text book I am not writing it
over here; c is the, if c is the velocity of light and the k the k that you see over here is the
Boltzmann constant. So, that is also the value of which is also available in your text.
569
So, if you if you think about what would be the units of the black body radiation
intensity from the definition we understand that it is going to be energy per unit area per
unit wavelength per unit solid angle because, we are talking about in any given
directions. So, I will tell you what this solid angle is all about, but the definition got the
units of the black body radiation intensity therefore, can be expressed in terms of energy
which is Watts, area, wavelength is customarily expressed in terms of micron. So, that is
why the micron is there in the denominator and the solid angle is the unit of solid angle is
steradian. So, steradian comes into the denominator.
So, the units of the black body radiation intensity I b lambda T, the spectral blackbody
radiation intensity is Watt per meter square per unit per micron per steradian. So, what is
a solid angle let us since we are going to use the concept of solid angle.
You are probably aware of this solid angle, but I will still do this one more time. So, let
us see this is the over here therefore, it is sort of a cone which is my area of interest let us
say the area is dA. The distance from the center from this point which I denote it as O let
it be r. So, this is the solid angle subtended by the area dA at the center and this dA is
normal perpendicular to the direction.
So, therefore, this dA perpendicular to the dotted line and the solid angle is defined as the
area times r square. Had this area been at an angle with the line which comes out of this
origin the specifying the direction then the projection of this area perpendicular to the
570
direction has to be provided. That means, dA has to be substituted by dA cos theta if this
is the area through which the radiation is passing through. So, therefore, the projection of
this area on to this side has to be substituted for dA. So, dA is therefore, perpendicular to
the direction and this has to be kept in mind.
So, if you extend this so, for a hemisphere this if the from the center, from its center the
solid angle is simply going to be equal to twice pi. And when you consider as full sphere
this is simply going to be equal to 4 pi. So, these are obvious; that means, what is going
to be, what is going to be this value of the solid angle for the case of a hemisphere and
for the case of a sphere. So, as before if let us say as I was telling you if this is the origin
and I have some area which is not perpendicular to which is not perpendicular to this and
therefore, let us say the area vector makes an angle of 45 degree with this direction, with
this direction.
And let us say if this d area, the area is about 4-centimeter square. So, this is the area
vector which is always perpendicular to this area. So, therefore, using the definition d
omega the solid angle as dA by r square and this dA is simply going to be the area which
is let us dA 1. So, this must be equal to dA 1 the projection of dA 1 in this direction. So,
this is going to be dA 1 cos theta 1 divided by r 1 square. So, this is how the solid angle
is evaluated.
So, if it is just part of a circle and if this is dL at a distance of r then the plane angle this
alpha let us call it as d alpha. So, d alpha is simply going to be dL by by r, that is this
standard relation when we were talking about a plane angle. And, when we are talking
about a solid angle like this, if this is the distance is r and this area is dA n then in that
case the d omega is simply going to be dA n by r square, where A n is the area normal to
this direction. So, that is what we call is that plane angle and what is known as the solid
angle. So, next thing is important because, next going to give us some more insights into
the whole process.
571
And it is a very important quantity which is known as the black body emissive power.
So, what is a black body emissive power? The expression for which we need to find out
so, what is black body emissive power? It is the radiation energy per unit area of a black
body at T at an absolute temperature T in all directions in all directions in a
hemispherical space. So, this is what is known as the black body emissive power and we
need to find out what would be it is expression based on our knowledge of blackbody
radiation intensity.
So, whereas, black body radiation intensity is in a given direction the black body
emissive power is in all directions in a hemispherical space. So, the difference between
these 2 must be kept in mind in one case the radiation intensity specifies a direction. So,
it is the direction which is specified for the case of radiation intensity whereas, for the
case of emissive power it is assumed that the black body is placed at the center of a
hemisphere. And, we are trying to see what would be the total energy emitted by this
black body of unit area which is placed at the center of the hemisphere in all possible
directions. So, we need to find out we understand that this I b lambda T which we have
defined previously it has a directional dependence. So, the solid angle subtended by the
area at a distance from the center at a distance from the center is used to obtain what
is the black body radiation intensity. Now, what we have to do is we have to make this
area which is subtending a solid angle over here travel in such a way that it defines a
hemisphere.
So, if we can integrate the area in such a way that the entire hemispherical dome over
this unit surface area can be specified then
572
the total energy which passes in all directions
in the hemispherical space at a given temperature will be known as the black body
emissive power. And, that is what we are going to evaluate we are going to derive in this
in this part of the class.
So, from the intensity now we are going towards emissive power. In one case the area is
fixed it is dA 1 and we are trying to see what is the solid angle in the other case this area
is essentially the area of the hemisphere. So; obviously, the solid angle will change
depending on where this unit area is placed. So, this has to be taken into account while
evaluating this. So, let us draw the draw this first and see what we get out of this. The
first one let us say we have this is the area vector, this is the area, we would like to find
out what is going to be the intensity in this specific direction.
So, this is the preferred direction let us call it is omega and in this direction the intensity
is I b lambda T. The area over here is this is the area we are talking about and initially
this is the area dA at a temperature T. But, as you can see it is not perpendicular it is not
perpendicular to the direction of propagation of to the direction in which I would like to
find out I b lambda T.
So, what I do is I try to see: what is the azimuth angle of this area over here and this
being the angle theta. So, what I do is I will draw the projection of the green one dA and
therefore, this area is simply going to be dA cos theta. So, my dA cos theta is now
perpendicular to the direction in which I would like to find out what is I b lambda T.
Once again, the I b lambda T is the black body radiation intensity. In fact, spectral black
body radiation intensity at a given temperature in this at a specific direction which is a
this direction.
The object from where this intensity is coming is dA. So, I am going to take a projection
of this make the area of perpendicular to the direction. So, therefore, this area is to going
to be dA cos theta. So, if I write the intensity in this case which is I b lambda T should
equal to energy per unit area per unit wavelength and per unit solid angle. So, this is my
definition of the energy.
573
So, the spectral radiation energy let us the spectral radiation energy emitted by emitted
by dA by dA the surface element dA in which which through an elemental solid angle d
omega. So, this angle solid angle is d omega. So, this d omega I would like to find out
how much of how much of spectral radiation energy emitted by dA passes through the
solid angle d omega. So, this is the energy which is contained within this tube that I
would like to find out ok.
So, this must be equal to I b lambda T then cos theta d omega times dA d omega. So, this
energy; obviously, would be in order to obtain the energy I need to multiply the intensity
with area. So, this is going to be my area, the solid angle is this. So, this is the solid
angle, this is this is what is going to be the energy the spectral the spectral energy. So, if I
looked like to find out the energy by unit if I want to do it with in terms of unit area, it
should simply be equal to I b lambda T cos theta d omega.
I will go through it once again; my intensity is defined as energy per unit area per unit
wavelength per unit solid angle ok. So, the area is dA which makes an angle of theta with
the area vector. This d this dA and I would like to find out how much of energy is going
through this tube which forms a solid angle dA at this point. So, in order to do that the
first thing is I need to make sure that this area is placed in a direction perpendicular to
this.
So, which is going to be dA cos theta and this is this is this is the azimuth angle which
we will discuss later on. So, the energy per unit wavelength or in other words the spectral
energy would therefore, be the product of intensity times area which is perpendicular to
the direction times the solid angle. So, that in then it should be the spectral radiation
energy emitted by dA would be the intensity times area which is dA cos theta times solid
angle which is d omega. If you like to find out what is the spectral radiation energy
emitted by an unit area I simply divided by dA and this is the expression of the spectral
radiation energy emitted by and unit area instead of dA an unit area.
So, this is the quantity which I am going to use, but in order to effectively use this
quantity I need another factor ok. So, another figure so, the figure that I am going to use
going to draw is this one.
574
So, this is a hemisphere which I am drawing of base radius r and this is a slice of that
which where the angle is d phi. And once I draw this it would be clear to clear to you.
The angle over here is theta and this angle, this small angle is d theta ok. And this is d
phi so, this one must be equal to r times d phi because this is a plane angle where the
radius is r and the angle is d phi. So, the length of the chord must be equal to d phi. This
when we go all the way up to 90 minus theta I get this line, if I go all the way to 90 I get
a point. If I go up to this point this is simply going to be r d phi sin theta ok.
So, when sin theta becomes equal to 90 which is at the top then would this would be
equal to 0. So, the length of the line will vanish, and it will become a point. When you
think of this one, this is d theta and the length the this thing the radius is r. So, therefore,
this one is simply going simply going to be equal to r times d theta.
So, if I project it slightly in a better way this is what you get as r d phi. And when you go
all the way up to this point so, this is d phi, this is d theta. So, what you get like this the
area this length is r d theta and this length has to be r d phi sin theta. So, my area which
is defined by the angles phi and theta is r d theta sin theta is 1 one dimension, the other
dimension is r times d theta.
So, therefore, from this figure my d omega the solid angle is simply going to be equal to
dA 1 divided by r square where dA 1 is this area. So, this area is dA 1, the area which is
denoted by the red crosses and this dA 1 is simply the product of these 2 length scales.
575
So, it is going to be r d theta is one length one side the other side is r d phi sin theta
divided by r square. So, this d omega is simply sin theta d theta d phi that is going to be
that is going to be the solid angle.
So, when we go back to this figure once again I have my spectral radiation energy
emitted by unit area is I b lambda T cos theta times d omega. So, I need to put instead of
d omega in this expression the d value of d omega that I have obtain. So, therefore, I b
lambda T cos theta d omega would simply be equal to I b lambda at a constant
temperature times cos theta sin theta d theta d phi. So, this is this is going to be the
spectral radiation energy emitted by an unit surface area element through which
substance a solid angle equal to phi over here now.
Now, I would like to make integrate this expression in such a way such that this area is
going to represent the entire hemispherical area. So, if I can do the integration in such a
way that this area is going to encompass the entire hemispherical area. So, I can see that
in that integration my theta is going to vary from 0 to 90 degree, 0 to 90 degree and my
phi is going to vary from 0 to 2 pi that is what the variation is going to be. Let us look at
it to another way this is what I am trying to do.
So, in order to create the hemispherical space my phi is going to be from 0 to 2 pi

whereas, my theta is going to be from 0 to phi. So, if I can let this area travel in terms of
theta from 0 to 90 degree and in terms of phi from 0 to 2 pi then this area encompasses
the entire hemispherical area available to exposed when the black body is placed at the
center.
So, the black body the spectral skill depends on the wavelength the spectral blackbody
radiation which is emitted per unit surface area in the hemispherical space in the
hemispherical space. Hemispherical space which is denoted as E b lambda T this is what
is the spectral black body radiation emitted per unit surface area into the hemispherical
space would simply be I b lambda T which is outside of the integration sign. And phi
would be from 0 to 2 pi and theta would be from 0 to pi by 2 and inside would be cos
theta sin theta d theta d phi.
One more time since it is black body radiation intensity so, it is independent of the
direction, since the black body radiation is diffuse. So, I b lambda T does not depend on
the direction and therefore, it can be taken out of the integration sign. So, what is left is
576
cos theta sin theta d theta d phi and as I have explained the theta the phi is going to vary
from 0 to 2 pi and theta is going to vary from 0 to pi by 2. If I perform this integration I
have an expression of E b lambda T in terms of I b lambda T. So, what that expression
would be? Once you perform this one once you perform this integration it is there in
your text.
I am not going to do it over here it is a very simple integration. This is the expression for
body radiation the spectral blackbody emissive power. So, since it is so, it is energy per
unit area per unit wavelength that is going to be it is unit energy per unit area per unit
wavelength. So, the spectral black body radiation intensity is related to spectral emissive
power of an unit based on an unit area. In a hemispherical space is denoted by specific
relation.
So, when you use function for I b lambda T E b lambda T would be c 1. So, the emissive
power the spectral emissive power of a black body of unit area at a given temperature is
provided as a function of wavelength and as a function of temperature. So, this is the
important part.
So, E b lambda T is a function of lambda and is a function of temperature. This specific

expression will be utilized in the next class to show how the emissive power of a
blackbody spectral emissive power of a black body depends on the wavelength and
depends on the temperature.
577
So, what we have done in this class is we have placed a black body of unit area inside
hemispherical dome. And, we have found out what is the total emissive power of this
black body having unit area at a given temperature, what is the spectral power emissive
power of black this black body into the hemispherical space.
So, this is what it looks like, in this hemispherical space how much of radiative energy a
black body of unit area at a given temperature is providing. So, this is related to the
intensity of radiation and we know the intensity of radiation through the use of Planck’s
function.
And then we can find out E b lambda T as a function of the wavelength and as a function
of temperature. So, this functional relation we will explore a bit further and then you
would see that it is going to give rise to certain relations that we know of; for example,
the most common example or relation of radiation the Stefan Boltzmann law can be
directly derived from this blackbody emissive power.
So, we will do that in the next class.
578
Heat Transfer
Lecture – 50
Weins’s Law, Stephen Boltzmann Law, Blackbody Radiation Function, Tutorial
Problem
In the previous class, we spoke about the Black body Emissive Power and we wanted to
see what is the relation between the Black Body Emissive Power commonly denoted by
Eb at a given temperature with the black body radiation intensity. And the relation which
we have obtained considering a black body of unit area placed at the centre of a
hemisphere and measuring what is the total energy emitted per unit time per unit area per
unit wavelength by a black body into the hemispherical space. We have obtained the
relation between the emissive power and the intensity of blackbody radiation.
So, the relation that we have obtained is the following
Eb (T )   I b (T ) W/m2μm
where Eb (T ) is the spectral black body radiation flux which is essentially the energy
emitted by a black body at any given temperature T. We need to remember that this T, the
temperature in radiation always does refer to the absolute temperature per unit area per
unit wavelength about that wavelength λ per unit time.
579
And this micron refers to the wavelength of the radiation, that is wavelength of the
radiation.
If you remember the units of I b (T ) that has units of W/m2μmsr, but since we have
integrated this over the entire hemispherical space taking into account the variation of the
Eb (T ) does not contain any steradian, any solid angle. So, if
angle. So, therefore, this
this is the formula and we already know through use of Planck’s function; what is the
I b (T ) .
expression for
So, using the Planck’s function for I b (T ) , I can write

C1
Eb (T )  5
  exp[C2 / KT ]  1
W/m2μm
The only thing that we have done in this, to this expression is substituted the expression
for the intensity from Planck’s function which we have discussed in the previous class.
So, when you do that, then you get this expression for the black body radiation flux and
the C1 and C2 are constants. So,
C1  2 hc 2  3.743 108
hc
C2   1.4387  108
k
So, what this expression tells us is that at any given temperature and at any given
wavelength Eb (T ) can be computed. So, not only you can and so, Eb (T ) is a function
both of temperature and of wavelength and therefore, if you know the wavelength and
the temperature, the black body radiation flux can be computed.
So, provided we know what is the temperature, then the spectral blackbody emissive
power, I use the word spectral here purposefully to stress that the black body radiation
intensity will also depend on the wavelength at which we are measuring the radiation
flux.
So, all these are strongly dependent on the wavelength. Therefore, since we know now
what is the relation between Eb (T ) , which is the black body radiation flux with λ and T,
we can plot the black body radiation function as a function of λ, the wavelength at
different values of temperature.
580
Once again, Eb (T ) is a function of wavelength and is a function of temperature. So,
keeping one constant in this case the temperature constant, I can compute the variation of
Eb (T ) with λ and then, I can choose another value of temperature and again compute the
same variation of Eb (T ) as a function of λ.

So, I get a series of curves, the series of curves at different values of temperature which
shows the variation of the black body radiation flux as a function of wavelength given by
Eb (T ) as a function of λ
the expression which is shown over here. So, I fix T and plot
and then, I choose a different value of temperature and plot this again. Therefore, this
would result in a family of curves which would look something like this.
So, the variation of Eb (T ) as a function of wavelength will show variations like this.
These are at different values of temperature T1, T2, T3 etcetera and this T1 is greater than
T2 greater than T3 greater than T4 and so on. So, the top one, top curve corresponds to the
highest temperature, the bottom curve corresponds to the lowest temperature.
So, if these maxima the point at which the Eb (T ) is a maximum; that means, the
blackboard emissive power is maximum, if these points are joined together, the locus of
this line is very interesting. It is λT corresponding to the maximum value of Eb (T )
581
is simply a constant 2897.6 μmK. The always in radiation the wavelength is referred to
as a wavelength is expressed in microns and T is in Kelvin. So, I will go through it once
again.
Using the formulation that what we have written in the previous slide, let me write it
once again here.
C1
Eb (T )  5
  exp[C2 / KT ]  1
W/m2μm
Using this formulation, I fix temperature. Let us say I have taken the temperature to be
equal to T1 here and this is T2, T3 and T4.
So, if I fix T1, I can plot Eb (T ) as a function of λ. This is its variation. Then, I fix the
temperature at T2 slightly lower than T1 and compute this function once again and this is
what I am going to get corresponding to T 2. Then, T3 which is less than T2 this is the
distribution of Eb (T ) as a function of λ. Interestingly, when you look at the maxima of
Eb (T )
these curves, what you see is that the product of λT corresponding to the max of
is a constant.
 T  MAX  2897.6
μmK
So, if you find out what is the corresponding λand what is the value of T 4; if you multiply
them together and you find out what is the corresponding λ and what is the
corresponding temperature T1; so, essentially it will tell you that
 1T1  MAX   2T2  MAX

This relation that λ corresponding to maximum Eb (T ) is a constant is known as Wein’s
Displacement Law.
So, what it states is that if you know the temperature, if you know the temperature, then
the corresponding value of the wavelength at which Eb (T ) is a maximum can be simply
obtained by this relation. We will see its use in subsequent our subsequent analysis, but
let us do something which is even more which we know even better.
582
So, the radiation energy that is emitted by a black body, by a black body at T over all
wavelengths so, the expression that I have this is the emission radiation emission by a
black body at a specific wavelength.
But we probably are more interested in to finding out what is the total energy emitted by
the black body over all possible wavelengths. So, mathematically speaking, I would like
to know what is the total energy emitted by the black body over a wavelength range of 0
to ∞. So, if I could find out the total if I would like to find out what is the total energy
emitted by the black body.
Then, the spectral black body radiation must be integrated over all possible values of
wavelength. So,
 
C1
Eb (T )   5
d
0
  exp[C2 / T ]  1
So, that is how you obtain the total amount of energy emitted by a black body at an
absolute temperature of T, which is what we are going to do next.
Since, what we are just saying that this must be equal to integration from 0 to ∞.
 E  (T )d  "
b
0 that is a standard definition I have put the expression incorporating
Planck’s distribution.
583
So,
x  T
x 
4 C1
Eb (T )  T  5
dx
x 0
x  exp[C2 / x]  1
  5.67 108 W/m2K4 =Stephan Boltzmann Constant
Eb (T )   T 4 W/m2
Eb (T ) is W/m2μm. Since, you have integrated it over all
So, the expression the unit for
possible wavelength. So, therefore, this one will not have this micron over here it is
simply W/m2. And this sigma, you are well aware of this is known as the Stephen
8
Boltzmann Constant. Stephen Boltzmann Constant and the value of this σ is 5.67 10
W/m2K4 this relation obviously, then is known as the Stephen Boltzmann Law.
So, this is one of the unique laws of radiation which tells you that the total energy
emitted by a black body at an absolute temperature T per unit area per unit time is a
function of absolute temperature to the power 4 and the constant of proportionality is
simply σ is which is a constant which can be evaluated by evaluating this definite
8
integral and it has a numerical value of 5.67 10 W/m2K4 and this σ is known as the
Stephen Boltzmann Constant.
So, what we then see is there are two interesting observations from here. The first of all
that the black body radiation function, the black body that emissive power of a black
body is a function of wavelength and is a function of temperature. Since it is a function
of wavelength so that is why the emissive power of a blackbody, we use the adjective
spectral before it to underscore the importance of wavelength while specifying the
emissive power of a black body.
Now, when you integrate it over all possible wavelengths, what you get is the total
emissive power of the black body at that given temperature. So, it is the amount of
energy emitted by a black body per unit area per unit time over all wavelengths possible.
So, that is why its total and I can drop the spectral from the
584
description of the emissive power. The moment I integrate over it and the result of
disintegration is the Stephen Boltzmann Law.
In some applications, you are not interested to know what is the total power emitted by
the black body. You are more interested to find out what is the power emitted by the
black body within a certain wavelength range let us say you would like to find out
between 2 to 4 μm, how much of the energy is going to be concentrated.
So, the entire blackbody emits radiation starting at λ=0 to λ=∞, but you would not want
that; you want within a specific range of λ how much of energy is going to be released,
how much energy is going to be emitted by the black body as a fraction of the total
energy emitted by the black body.
So, this is an important parameter which we will keep on using in our subsequent
discussion. That means, the fraction of energy emitted by a black body within a specific
wavelength range. So, that is that is known as the Black Body Radiation Function.
So, we would see what is a Blackbody Radiation Function? So, Blackbody Radiation
Function as I have expressed to you; so, it is between a wavelength from 0 to λ at a given
temperature is the amount of energy emitted by the black body between 0 to
585
λ divided by 0 to ∞; that means, the total energy emitted by the black body at that a
given temperature.
So, once again, the fraction the black body radiation function, the fraction of energy
which is emitted by the black body within a wavelength range from 0 to λ is nothing but
the amount of energy emitted by the black body between 0 to δ divided by the amount of
energy emitted by the black body over 0 to ∞. So, this is the total energy and this is the
amount of energy within a fixed wavelength range.
 
 Eb (T )d 
0
 E  (T )d 
0
b
f 0 (T )  

T 4
 E  (T )d 
0
b
C1
Eb (T )  5
  exp[C2 / T ]  1
T
C dx
f 0  1  5  f   ,T 
 x 0 x  exp[C2 / x ]  1
So, this integral can be evaluated and f 0 (T ) can be calculated for a given λT. So, you
would I guess in the idea behind this is very clear to you. I am trying to find out the
fraction of energy which is emitted by the black body that is going to be concentrated
within a wavelength range of 0 to λ.
So, if this is a fraction, then this is going to be the energy which is going to be within 0 to
λby the total energy emitted by the black body. Total energy emitted by the black body is
nothing but is σT4 and this can be substituted by the Planck’s function as we have done
before. When you incorporate that you get the fraction in this form. This is again an
integral equation which can be evaluated provided you specify λ and T. So, once you
specify λ and T, this fraction can be evaluated.
586
So, this fraction is calculated for a given λ times T and you would see that the results of
the fraction is provided in a table.
So, in that table the total emissive power of a black body within which is contained
within the wavelength range is provided as a function of T and as a function of λ. So, one
you must keep in mind that this T is always going to be in Kelvin and λ is going to be in
micron.
587
So, if you look at the table now, the table here shows you; the table here shows you the
Eb
5
value of λT in micron and this is T . This is how the fraction is shown. So, you know
what is λT. Let us say you pick a value of λT to be over here. λT is 2777.8. So, any
combination of λ and T which gives the products to be about 2777, it would give you a
value of the f as 0.22.
So, let us write it over here if I write it in this so, for a value of λT equal to 2777, the
value of the fraction is 0.2285. So, for λT equals 2777.8, the value of f is going to be
equal 0.22. So, which means that you choose some value of λ and T, the product comes
close to 2777.8. So, if that is the case, then within that wavelength range at that given
temperature, you can find out what is the fraction of energy that is emitted by the black
body is contained within the wavelength.
And I think it would be more clear to you once we solve a problem using the table and
this f the black body radiation function, the table of blackbody radiation function you can
refer to and find out that given the value of λ and T what is the fraction that is going to
be, what is a black body radiation function. So, we will quickly solve 1 problem and I
think that would clarify any remaining doubts that you may have. But fundamentally, we
are finding out the fraction of energy which is contained within 0 to
588
some λ at a given temperature for the blackbody emissions. So, let us quickly solve 1
problem and then we will move on.
So, what the problem says is that the emission is from a surface, from a black body at
T=1000 K. The first part is what fraction of the total energy is emitted below λ=5μm. So,
what we need to do then is this my λ=5μm, T is 1000 K. So, λT is 5000 μm.K.
So, when we would say fraction. So, I would like to find out 0 to 5 μm. The value of the
black body radiation function between 0 to 5 μm, fundamentally what this tells us is f 0
to 5 is the fraction of the total energy that is emitted by a black body at a given
temperature between the range 0 to 5 μm. So, from the adjoining table, I need to find out
what is the value of f of 0 to 5 μm.
589
So, for which case λT is 5000 μm.K. So, if you look at the value over here for a value of
λT to be equal to 5000 μm.K , the value of f is 0.66371. So, from the table, I read that
this is equal to 0.6637 for λT equals 5000 μm.K which simply tells me that 66.4 % or
0.664; the fraction 66.4 % of the total energy is emitted by the black body at 1000 K,
below λ=5 μm; that means, this is the value of the of the fraction.
Second question is what is the wavelength below which the emission is 10.5 % of the
total emission at 1000 K? So, it is reverse of the previous problem. In this case f 0-λ at a
given temperature of 1000 K is provided to be 0.105. We need to find out what is the
value of λ. So, once again the emission is
10.5 % of the total emission at 1000 K. So the fraction which of energy which is
contained between 0 to the unknown λ at 1000 K is specified to be
0.105. We need to find out what is the value of λ for such a case.
Once again, we look at the table and when we look at the table, now we have to look at λ
to be the fraction to be 0.105. So, when you look at the fraction to be 0.105, you read the
value of λT to be 2222 μm.K. So, this refers to the value of λT. And since, you know the
temperature to be equals 1000 K. So, therefore, your λis simply going to be 2.222 μm.
590
So, at 1000 K, the emission is going to be
10.5 % of the total emission at and wavelength of 2.2 μm and the third one is. So, what is
the wavelength? What is the λ at which the maximum spectral emission occurs at
T=1000K .
So, the third question is what is the wavelength at which you are going to get the
maximum spectral emission? The moment I say spectral that means, its wavelength
dependent so, the value of the temperature is provided to you; you have to find out what
is the wavelength at which the emission is going to be maximum.
So, of course the law that is that we need to use for this is Wein’s Displacement Law.
The Wein’s Displacement Law simply states that λT is a constant when we are talking
about the maximum spectral radiation intensity at a given temperature. Since the
temperature over here is provided we need to find out what is the λ in this case. So, this
problem can simply be solved using Wein’s Law as λT corresponding to the maximum
emission is 2897.6 μm.K. So, for T= 1000K, this λ would simply be equals to 2.8976
μm. So, therefore, if your temperature is at 1000 K, then the maximum the spectral
maximum of emission from a black body at 1000 K will take place at a wavelength of
2.8 μm.
So, what we have discussed, if I would like to do a summary of this is that looking at the
black body emission when incorporating the Planck’s formula in it, we have derived the
Wein’s Displacement Law which shows that the product of λT corresponding to
maximum spectral emission from a black body is a constant which is 2897.6 μm.K.
Secondly, the spectral emission if you would like to convert it to the total emission from
a black body over all possible wavelengths that would give rise to the well known
Stephen Boltzmann Law where the proportionality between the emission and the
temperature is E is proportional to T4 and the proportionality constant is Planck’s
constant.
Next, we wanted to know the fraction of energy which is emitted between a certain
wavelength range as compared to the total energy emitted by the black body overall
wavelength range. So, this black body radiation function is evaluated for different values
591
of the product of λ and T different values of the product of the wavelength and
temperature.
So, these black body radiation functions provide us with the knowledge of how much of
energy is going to be emitted by a black body at a given temperature within a specific
range of wavelengths. And we have solved the problem numerical problem which I think
would clarify any doubts that you may have in the use of Wein’s Displacement Law, the
Stephen Boltzmann Law and the Blackbody Radiation Function.
592
Heat Transfer
Lecture - 51
Kirchhoff’s Law
In this class, we are going to establish the relation between two of the very important
properties related to radiation from earth to a surface or to radiative properties of a
surface which are known as the emissivity and the absorptivity. Emissivity gives us the
idea of how close a surface is to a blackbody for a blackbody, its emissive power is
maximum and the value of emissivity is equal to 1. But for most of the real surfaces, the
it is their emissive powers are not as good as that of a blackbody. So, the value of
emissivity for a real surface varies between 0 to 1 for maximum for the case of a
blackbody.
On the other hand, most real surface can observe every say everything all energy which
is incidental on it, unlike a blackbody a blackbody absorbs all energy that are incident on
it. But for a real surface, it is going to reflect part of the incident energy, it may transmit
a fraction of the incident energy as well the rest is going to be absorbed by it. So, the
absorptivity of a real surface is less than 1, it is equal to 1 for a blackbody.
So, what is the relation between these two the emissivity and the blackbody and
emissivity and the absorptivity of real surfaces. Now, for that we are going to do a
thought experiment, what we would think is that an enclosure which is maintained at a
temperature of some absolute temperature T and a small object is going to be placed
inside the enclosure. So, the small object is going to receive energy from all directions
because it is inside the enclosure and initially the temperature of the small object could
be different than that of the enclosure temperature which is assumed to be very large the
size of which is going to be very large.
So, you let it inside let it receive energy from the enclosure, let it emit energy as well and
throughout this process the the temperature of the object will slowly come towards the
temperature of the enclosure. We will assume that it is only radiation which is taking
place. So, when perfect thermal when thermal equilibrium has reached the temperature
593
of the body and the temperature of the enclosure they are going to be the same and we
are going to start our analysis from that point onwards.
So, if you look at the what I have said, if you look at the pictorial description of this. So,
this is an enclosure which is a blackbody it is emitting radiation which is emitting
radiation and this is the object which is placed inside the enclosure and the amount of
energy which is incident on it is q i lambda T; i refers to the incident, lambda shows the
spectral nature and T is the temperature of the object. So, q i lambda T is the spectral
radiative heat flux from the walls which are at a temperature T and which are incident on
the body.
So, the this what it reaches on the body is essentially q i lambda T. Now, part of it is
going to be absorbed by the body. So, what is going to be the amount which is absorbed
by the body, let us denote it as q lambda T. Again, the lambda is kept here to underscore
to highlight the spectral dependence of radiation phenomena and that is why we keep
lambda over here. So, this is alpha which is the spectral absorptivity multiplied by the
incident radiation on the object.
So, whatever be the incident radiation on the object which is denoted by q lambda i T
multiplied by the spectral emissivity would give you the amount of energy which is
absorbed by the body. But when you think of the thermal equilibrium, the temperature
has the temperature is steady now so, therefore, whatever amount of radiation it is
594
getting from the enclosure walls and what whatever be the fraction that is absorbed by
the object, the same amount must be emitted by the object in order to maintain the
constant temperature.
So, the energy which is actually absorbed by the surface must be equal to the energy
which is emitted by the surface that is what thermal equilibrium demands. So, if we will
look at the previous expression now that we have written this q lambda T which is the
heat flux absorbed by the body must also be equal to the heat flux which is released
which is emitted by the object. Furthermore, we note that this incident radiation is
coming from a blackbody ok, the incident radiation is coming from a from the enclosure
walls which in itself is is is a black body.
So, therefore, this q lambda i which is at some temperature this is q lambda b at T so, this
is the spectral blackbody emissive flux. This we are familiar with two observations we
have made over here. The first observation is this is coming from a blackbody so; this
must be equal to the spectral blackbody emissive flux. So, this can simply be replaced by
this. Secondly, your thermal equilibrium our thermal equilibrium demands that whatever
is absorbed must be emitted in order to maintain thermal equilibrium.
So, if I work on this little bit more, what I would get is that q lambda T by q lambda b T
is equal to alpha lambda T. I have only replaced this by the spectral blackbody emissive
flux. Now, if you think about the definition of epsilon then q lambda T whatever is
emitted divided by whatever is emitted by a blackbody at that specific temperature is the
emissivity. Emissivity is the amount emitted by an object divided by amount emitted by
a blackbody at the same temperature that is the definition of spectral emissivity.
So, if you look at these two now, since q lambda T is equal to the amount of heat flux
absorbed by the body and to maintain thermal equilibrium, it is also the amount of
energy emitted by the body. So, this in order to maintain equilibrium, this must be equal
to the heat flux emitted by the body. So, if that is the case and if you look at these two
definitions now, what you can simply say is that epsilon lambda is equal at a given
temperature is equal to alpha lambda at a given temperature. This is what is known as the
Kirchhoff’s Law.
So, the Kirchhoff’s Law, simply tells that the spectral emissivity is equal to the spectral
absorptivity provided and that is a big one provided the emission is coming from a
595
blackbody at the same temperature. Remember, we have assumed that this has reached
thermal equilibrium with this. So, whatever be the temperature of the blackbody since
this is inside the enclosure for a long time it has also reached the same temperature as
this one. So, when we when these conditions are met we can say that the spectral
emissivity is equal to the spectral absorptivity.
So, the so, the assumptions involved are that the radiation is coming from a blackbody as
the same temperature as that of the object itself. Only when these two conditions are met
we can say that Kirchhoff’s Law is valid that is the radiation is coming from a blackbody
and the temperature of the blackbody from where the radiation is coming is the same as
that of the body itself. So, these are two major restrictions on the use of Kirchhoff’s Law.
It may so happen also that if we integrate this relation over all the wavelengths then, it
may happen that it may you epsilon is going to be epsilon at a given temperature is going
to be alpha at a given temperature. So, not only spectral emissivity is equal to the spectral
absorptivity, the overall emissivity is also equal to the overall absorptivity provided and
this is a big one provided that the incident and the emitted radiation are have same
spectral distribution.
This would require some more analysis on our part before you appreciate the whole
meaning of the incident and radiation incident and the emission are having the same
596
spectral distribution, what exactly we do mean by that. So, I will I will go into this in the
next slide.
But this simple thought experiment tells us that if two objects are in thermal equilibrium
such that their temperatures are the same and if the emission to the object is coming from
a blackbody. Then the spectral emissivity would be equal to the spectral absorptivity or
epsilon lambda at a given temperature is going to be alpha equal to alpha lambda. But the
two major constraints for this are the emission has to come from a blackbody and the
temperature of the blackbody must be the same as that of the object.
So, of course, you can see that these two are quite restrictive and we we really have to
think is there any other way by which we can make a statement about the values of
epsilon and the values of alpha. Now, since these are spectral quantities one would be
tempted to integrate them over the entire wavelength to see what is the overall what is
the relation between the overall emissivity and overall absorptivity? So, that is something
which we are going to do next in this part of the class.
So, what we start with is a relation which is spectral directional emissivity. I will call it
as simply emissivity since it is spectral and it has it is directional. So, lambda and theta
are given and are that these two subscripts must appear with emissivity and the other one
is spectral directional absorptivity which is alpha, since it is spectral I have lambda; since
its direction I have theta.
597
So, what is we I we do not we cannot prove that in this course, but what statistical
thermodynamics tells us tells us that these two epsilon lambda theta will always be equal
to alpha lambda theta. So, this comes from statistical thermo and this is this follows from
something which is called principles of detailed balancing, which is beyond the scope of
this course.
I will not introduce it here but, this is what you can obtain from the statistical
thermodynamics and this is known as the most fundamental form of Kirchhoff’s Law
which is always valid. Irrespective of what you have whether where the radiation is
coming from what it is emitting what is the temperature, this relation will always be valid
ok.
Now, what we are going to do is we are going to put on restrictions to this relation and
try to see the more restrictive form of the Kirchhoff’s Law. So, I am from this I am going
to find out what is the spectral emissivity and what is the spectral absorptivity using just
the definition of the spectral from the expression for spectral and directional.
So, since it is directional in order to get rid of the direction, I need to integrate it over the
entire hemispherical space. So, the same way we have done it for the blackbody when it
is placed inside a hemisphere, I am simply going to find out this is as cos theta sin theta d
theta d phi. This is what we would the expression and 0 to 2 pi, 0 to pi by 2, then the
value this is for the blackbody and for the blackbody the emissivity has a value equal to
1.
So, what I will have then is cos theta, sin theta, d theta and d phi. The this one this is the
directional one so, the definition of absorptivity is again 0 to 2 pi as before 0 to pi by 2,
then alpha lambda times theta q lambda i cos theta sin theta d theta d phi; this is the
incident radiation and if we do it for the case of a blackbody, then this absorptivity will
have a value equal to 1 therefore, I will only have this as cos theta sin theta d theta d phi.
When these two are going to be equal that is the big question? So, we need to find out
under what condition epsilon lambda is going to be equal to alpha lambda.
So, let us look at the possibilities here. The first one is what we can say is that the
irradiation is diffuse, if the irradiation which is q lambda i which is this one, if this is
diffuse then, it will have no directional dependence. It will be a constant with respect to
598
the direction that is what diffused is all about a property is diffused a radiative property
is diffused when it does not depend on the direction.
So, if the irradiation is diffused, then it is not going to depend either on theta or on phi.
So, therefore, this can be taken outside, this can be taken outside, they will cancel and
after that if you look at these two. Let us say this q i terms are not there and you are
looking at only this and this except the q and since epsilon lambda theta is always equal
to alpha lambda theta in absence of q this epsilon lambda must be equal to alpha lambda.
So, that is the first condition which needs to be maintained. Second is the surface is
diffused surface itself is diffused.
So, what is the advantage of making this assumption? If the surface is diffused, then
epsilon lambda theta does not depend on theta. So, this can be taken outside, and alpha
lambda theta does not depend on theta so, that can also be taken outside. So, since
epsilon lambda theta and epsilon lambda theta and alpha lambda theta are equal and if
they do not depend on this, then they would simply give you these terms will cancel what
is left is epsilon lambda theta and alpha lambda theta. So, epsilon lambda is equal to
epsilon lambda theta and alpha lambda is equal to alpha lambda theta and since these two
are equal, you are going to get epsilon lambda is equal to alpha lambda.
I think you should look at the expression carefully and imagine what would happen when
the directional dependence of the property is not there. So, these two can be taken
outside alpha lambda theta and epsilon lambda theta. Whatever is left inside the
integration sign are simply equal so, they can be cancelled out from numerator and
denominator for both cases.
And therefore, you would get alpha lambda is equal to epsilon lambda; that means, the
spectral values of emissivity is equal to the spectral values of absorptivity. So, this is
known as the restrictive condition of Kirchhoff’s Law what are the restrictions? It can
either be that the surfaces are diffuse or the incident ray the incident radiation is diffused.
In both cases what you are going to get is that the spectral emissivity is equal to the
spectral absorptivity.
So, once again if you look at this is known as the restrictive form restrictive form of
Kirchhoff’s Law. Now, one of the advantages is the this one is reasonable for many
engineering situations. That means, it is safe to assume that the irradiation is diffuse and
599
number 2 is the it is reasonable for many surfaces that the surface itself can be assumed
as a diffused surface. So, it tells us that the restrictive form of Kirchhoff’s Law is not too
restrictive a condition which is followed by many surfaces.
So, therefore, what I write is that epsilon lambda is equal to alpha lambda. This is the
relation which we have obtained subject to these 2 conditions. So, from this pic from the
spectral directional, I come to spectral and from spectral, now I would go to the overall.
So, what is the definition for the case of overall?
The overall emissivity is from the spectral one, 0 to infinity epsilon lambda q lambda b T
d lambda divided by 0 to infinity q lambda b T d lambda; are they going to be equal?
That is the question mark. On the other side, I have alpha, the absorptivity which is equal
to 0 to infinity, alpha lambda the incident radiation divided by 0 to infinity q I lambda T
d lambda.
So, these are just the definitions of overall emissivity from spectral emissivity. So, the
spectral absorptivity and this is the overall absorptivity. So, when these two are going to
be when I am going to put this equality sign. So, the first one is irradiation corresponds
to emission from a blackbody at the surface temperature T. So, if that is the case, then q
lambda i at T is equal to q lambda b from a blackbody at T and therefore, 0 to infinity q i
q lambda i at T is equal to q b T which is nothing but saying that this is equal to 0 to
infinity q lambda b T d lambda.
600
So, if the radiation corresponds to emission from a blackbody at the surface temperature,
then this is to be replaced by q lambda b T d lambda. So, once you replace that these two
would become identical because, we have already assumed that epsilon lambda is equal
to alpha lambda. Since, we have assumed the spectral properties are the same and now,
we are assuming that the irradiation is also coming from a blackbody at the same
temperature. Therefore, this is going to be same as this and this part as I shown here is
going to be equal to this part. So, if this is the condition even though it is a very
restrictive condition, if this is the condition then alpha is going to be equal to epsilon or
this is a major statement or if the surfaces are gray if the surfaces are gray.
So, what is a gray surface? A gray surface is the one for which this epsilon and alpha
lambda are independent of lambda let us see what I am what we mean here. If a surface
is gray then they do not depend on alpha anymore. See, if they do not depend all on alpha
anymore, they can be taken out of the integration sign and here alpha can be taken out of
the integration sign. So, what you have if you do not have epsilon, if you do not have this
one, then the numerator and the denominator are the same. If you do not have alpha
inside the integration, then the numerator and the denominator are the same.
So, what you have in that specific case is that epsilon lambda is equal to alpha lambda
and this is equal to epsilon and this is equal to alpha. This is true since our starting point
was the spectral emissivity is equal to the spectral absorptivity and what we are saying
here is that since the surfaces are gray, then epsilon is equal to epsilon lambda and alpha
is equal to alpha lambda which shows that these two must be equal that means, epsilon is
equal to alpha.
So, even though this is a very restrictive condition which says that the this the equality of
emissivity and absorptivity overall not the spectral overall can be obtained only the
emission is coming from a blackbody at the surface temperature. There is a different way
of expressing the same thing which tells us that the overall values would be equal, if we
assume a surface for which epsilon lambda and alpha lambda are independent of lambda.
If you do assume that which defines a special class of substrate as the gray substrate, for
a gray substrate since these two are independent of directions. So, these two can be taken
outside of the integration sign numerator and denominator in that case will cancel out.
601
So, what I would left with is epsilon equals epsilon lambda equals alpha lambda equals
alpha which essentially means epsilon is equal to alpha, the overall emissivity is equal to
the overall absorptivity.
So, this is the most restrictive form of a Kirchhoff’s Law. But this is for the gray surface
gray surface is so; this is the most restrictive form of Kirchhoff’s Law which says that for
a gray surface, these two are going to be equal. So, the total hemispherical emissivity is
equal to the total hemispherical absorptivity. So, calculations for radiation exchange
between surfaces are greatly simplified if this equation can be if this can be assumed to
be true.
So, many calculations most of the calculations of radiation between two or three surfaces
when they are exchanging radiation among themselves should be greatly simplified, if
we can assume that they are gray surfaces. Fortunately, many of the surfaces many of the
some of the many of the engineering surfaces the assumption of a gray behavior,
assumption of gray behavior this does not incorporate significant errors.
So, the moment we can so, if we can we should always try to see if gray behavior can be
assumed for the engineering surfaces for the surfaces which are exchanging radiative
heat among themselves. So, if we can do that we would show next is how this would
simplify the calculations significantly.
But, if it is not if the spectral values the spectral the if the values of the alpha and epsilon,
if they depend on wavelength then this assumption cannot be used, and I would just give
you an example which would show that why we cannot use that approximations for
surfaces which have spectral distribution of emissivity and absorptivity.
602
So, for that I am going to draw 2 curves which would show how would the values of
alpha and epsilon are varying with lambda. This is alpha which is a function of lambda
and let us say this is equal to 1 and this is lambda 1. So, the value the behavior of lambda
alpha is something like this it is close to 1 from this one, it sharply falls and becomes
almost like 0 so, this is the value of lambda 1.
On the other hand, the amount incident 1, this is my lambda 1; the incident 1 is such that
the spectral distribution of this 2 incident rays. So, this is lambda, this is lambda and let
us call this as q lambda type of emission 1 this is incident type of incident 1 which is a
function of lambda and this is q lambda 2 incident 1 times lambda. So, this is case 1 and
this is case 2. So, what we would see from here is alpha is almost equal to 1 in this case
and the reflectivity is 0. These are opaque substrates so, reflectivity is 0, alpha 1 is equal
to 1, rho 1 is equal to 0 and in this case alpha 2 is about 0 so, rho 2 would be equal to 1.
So, what is the above surface, if this is the surface the surface which we are talking
about; this surface has the property where the absorptivity varies from this is roughly 1
up to a wavelength of lambda 1 and beyond lambda 1, it comes to a value equal to 0,
almost close to 0 and 2 different incident radiations are falling on the surface. So, the 2
radiations are q lambda 1 incident radiation and q lambda 2 incident radiation. One is
concentrated in the range 0 to lambda 1, the other is concentrated in the region beyond
lambda 1.
603
So, since in this region, it is absorptivity is equal to 1, almost all of the incident radiation
is going to be absorbed by the surface. Whereas, over here since the spectral distribution
of the incident radiation is such that since the value of alpha 2 here is about 0, none of
the absorption, none of the incident radiation is going to be absorbed on the surface ok.
So, the values of alpha, they are drastically different in these 2 cases ok. So, the based on
the irradiation field, you your values of alpha lambda are drastically different and the
value of epsilon on the other hand is independent of irradiation.
So, what I mean here that the value of lambda if you calculate based on an irradiation
like this and the value of alpha when you calculate your value of alpha based on
irradiation like this this two are drastically different. On the other hand, the value of
epsilon is independent of irradiation therefore, there is no basis to say that alpha will
always be equal to alpha is always going to be equal to epsilon.
So, absorptivity and emissivity; hemispherical absorptivity and hemispherical emissivity

may not be equal. So, we have to be careful while using emissivity equal to absorptivity,
this relation to know what is the irradiation field? Has it been mentioned that the surfaces
that are we are talking about the surfaces and the surfaces are gray. So, they do not have
any directional, they do not have any spectral dependence of the properties. So, using
Kirchhoff’s Law, one has to be very careful to use it with the most restrictive condition
that is alpha is equal to epsilon or the with a restrictive condition in which epsilon
lambda is equal to alpha lambda. So, for these two cases one has to be careful.
But, this spectral dimensional emissivity will always be equal to spectral directional
absorptivity. So, if you look at the development that I have done in today’s class, starting
with epsilon lambda theta equal to alpha lambda theta; how I have obtained alpha equals
epsilon and what are the approximations or assumptions that I had to make in order to
get alpha equals epsilon.
So, this should be very clear in your mind, I am going to solve few problems on this
Kirchhoff’s Law concept. So, that it is going to be very clear to all of you that is when
and how you can use Kirchhoff’s Law and what are the simplifications that we can
obtain once if we can assume that it is their gray surfaces that is overall emissivity
overall hemispherical emissivity is equal to overall hemispherical absorptivity. Once you
604
solve the problems, it will be more clear to you which we would take up in the next
class.
605
Heat Transfer
Lecture – 52
Tutorial on Emissivity, Absorptivity and Blackbody Radiation Functions
We were discussing about the Emissivity and Absorptivity. So, when these two can be
taken to be equal. So, in order to do that, we started with the spectral directional
emissivity and spectral directional absorptivity and we understand that from statistical
 ,   ,
thermodynamics, these two will always be equal. So, therefore, = , λ denotes the
spectral nature of emissivity; whereas, θ denotes the directional dependence of the
emissivity. So, the spectral directional quantities emissivity and absorptivity are equal.
Starting with that, we wanted to get rid of the directional dependence. So, in order to get
the directional dependence part, I mean in order to remove the directional dependence we
had to make certain assumptions. So, that is once we do that we get the form that
 ;
 ,    ,
Through another set of assumptions, we arrive at the overall hemispherical emissivity, ϵ

is equal to the overall absorptivity, α
So, this is the most restrictive form of Kirchhoff’s Law, but which is also the most useful
as well. However, one thing one has to keep in mind is that the ϵ depends on the
emission from the surface itself; whereas, α depends on the irradiation falling on to the
surface so, these two are different things. So, even so, it is not easy to say that the α is
going to be equal to ϵ, since they refer to emission and irradiation from different sources.
So, I will quickly go through whatever we have done in the previous class and then, try
to solve the problem which would clarify your doubts.
606
So, we start first with this one that we have done in the last class that
 ,    ,
i
This is the most fundamental form of Kirchhoff’s Law. In order to q obtain the spectral
emissivity; we need to integrate it over all possible direction in a hemispherical space.
So, we already understand that this is going to be the form of that.
And the absorptivity, if you would like to find out what is the spectral absorptivity from
the spectral directional absorptivity, we have to again integrate it over the space.
However, here I am going to have the incident spectral, incident radiation on the object
in the numerator as well as in the denominator. So, the question is when these two are
going to be equal. So, if the irradiation is diffuse; that means, if this is a diffused
i
irradiation then this q is going to be independent of direction.
So, both can be taken outside from numerator and denominator and they will simply
cancel out.
2  /2 2  /2
i
    cos  sin  d d      q
0 0
,
0 0
, cos  sin  d d
  2  /2
 2  /2
 
i
  cos sin  d d
0 0
  q
0 0
cos  sin  d d
i i
So, that is the first assumption that is if the irradiation is diffused, I get rid of q and q
 ,    ,
both from the numerator and denominator. Then, obviously,
 ,   ,
Secondly, if the surface itself is diffused;
607
then, these and are independent of
 ,   ,
the direction. So, they can be taken outside and since and are same; what you
608
would see is that the numerator and denominator will simply cancelled, will cancel out
 ,   ,
leaving in the outside and on the outside and they will simply cancel out.
So, that is also a case in which we would get

  to be equal to  . So, this relation is
true,
 is equal to   is to if the irradiation is diffused and if the surface itself is
diffused. Now, these irradiation is diffused for many surfaces; whereas, the surface to be
diffuse is reasonable for many engineering situations. So, it is customary it is not unusual
to take this relation, slightly restricted form of Kirchhoff’s law that

 is equal to  
The question becomes more critical if you would like to go from the spectral emissivity
to overall emissivity or spectral absorptivity to the absorptivity itself. By definition, the
overall emissivity is simply this is the irradiation coming from a blackbody. This is also
irradiation from a blackbody, this is the incident radiation and this also is the spectral
incident radiation.
So, we would like to see when these two are going to be equal. So, if it so happens that
the incident radiation that is falling on the surface is coming from a blackbody at the
same temperature as that of the surface.
609
i
Then, q (T )  qb (T ) .
So, the incident radiation is coming from a black body at the same temperature. So,
therefore, this would be
 
i
 q (T ) qb (T )   qb (T )d 
0 0
(since     )
So, therefore,  alpha would be equal to  , since qb is equal to q so, you are going to
have qb (T ) on both sides. So, that is the most restrictive form of Kirchhoff’s Law
which assumes that the surface from where the irradiation is coming from is a black body
and its temperature is equal to the surface on which the radiation is falling.
So, that is the most restrictive form of Kirchhoff’s Law. Secondly, if we can see that if
 and   are independent of

the surfaces are gray that is a new kind of surface in which
λ; that means, the spectral and the spectral emissivity and spectral absorptivity are
independent of α.
Then, they can simply be taken outside and once you take outside, then the numerator
and the denominator will simply cancel out leaving  on this side and   on this side.
And since   is equal to  so, therefore, we can say that  is equal to λ; that means the
overall emissivity of a surface is equal to the overall absorptivity of the surface.
So, this is what I was talking about and therefore, these two are equal so, this is another
form of Kirchhoff’s Law. This is the most useful form of Kirchhoff’s Law, but it defines
a new kind of surface; new class of surfaces which are known as gray surfaces. Now, for
many surfaces the assumption of gray behaviour is justified so, what it says is that the
irradiative properties, mostly the emissivity and absorptivity are independent of
wavelength.
So, that means, spectral emissivity is equal to the overall emissivity and this gives us the
desired relation that the emissivity is equal to absorptivity. So, if you know one, then you
know the other and this simplifies many of the calculations specially radiation between
surfaces in a closed container.
So, if we can assume gray behaviour for 610

such surfaces, then our calculations of radiation
exchange between surfaces in an enclosure becomes a lot simpler. So, we would always
611
try to see if this is the case and proceed from there. However, one must use caution while
using the relation α is equal to  because, we still have to keep in mind is that the 
depends on emission from the surface; whereas, α depends on irradiation from a source.
So, we need to see that whether the emission and the incident radiation, they are in the
same spectral range at which   alpha lambda is equal to  . So, whether the spectral
emissivities and spectral absorptivities are same in the range, in the operating range that
we need to carefully observe.
So, I have given you an example which I will show you once again. So, this is a case in
which you have the absorptivity which is roughly about 1 upto a wavelength of λ 1 and
then, it falls to a very low value equal to 0. And you have 2 irradiations; the first is
irradiation is like this. It is between 0 to λ 1 and the second irradiation is this where it is λ 1
and beyond. So, if you consider this case, this irradiation between 0 to λ1, α is about 1.
So, the reflectivity is 0, we are considering opaque substrates. However, if the irradiation
in the second part, then irradiation if it irradiation falls in this wavelength range. Then, α
to the absorptivity is 0 as suggested by this figure and the reflectivity is about 1.
612
So, therefore, the value of α, the amount of energy absorbed by the surface strongly
depends on the spectral distribution of the incident radiation as well as the value of α in
i
that region. So, you can see that the, but; however, the ϵ is independent of q , ϵ does not
depend on the irradiation rather it depends on what is the temperature and what is the
   is equal to  depending on the irradiation,

nature of this substrate? This is also
you get a value of α equal to 1 or α equal to 2.
However, your value of ϵ will remain unchanged and therefore, it is not safe to say that α
is always going to be equal to ϵ. So, if radiation were in a spectral region which is which
the λ is less than equal to λ1 or λ is greater than λ4. I will draw these figures once again.
So, I have figures like this, let say this is λ1 and this is
q i and I have λ over here. Let us
say I have an irradiation which looks like this is at λ 3 and this is at λ4. Second case is I
have another range in which I simply have qλ which is from a blackbody, this is as before
λ1 and my distribution is in here.
So, this is λ2 and here, I have λ4. So, if the radiation is in a spectral region λ less then λ 1
or lambda greater than λ4, gray surface behaviour cannot be assumed. I will go through
these figures once again, this is a case in which the values are close to 1 and then, it falls
sharply beyond λ1. You have two incident radiation, one is within λ1, the two is beyond
λ1 so, if it is within λ1; then α is going to be equal to 1, the reflectivity is going to be 0.
However, if the irradiation is beyond λ1, then α2 this value is going to be about 0 and the
value of the reflectivity is going to be equal to 1. So, the value of α strongly depends on
the irradiation, on the spectral nature of irradiation; whereas, ϵ is independent of
radiation that is one case.
i
Let us say we have we have another situation in which the q falls over here and it is
also falling in this region and the corresponding distribution is
613
shown as  and   . This is λ1 and this is λ4 so, the curve looks something like this and
this is ϵλ0 equal to α λ0.
So, as you can see over this wavelength range λ1 to λ4, the values of emissivity is equal to
the values of absorptivity. So, if you have radiation within λ 1 and λ4, if it is within that;
then, we can say that overall emissivity and overall absorptivity are going to be the same.
However, if you have irradiation that is before α1 or after α4; that means, in this or in this
region, then the gray surface behaviour cannot be assumed. So, this is something which
we have to keep in mind what I think the issue would be more clear when we solve a
    .
problem and therefore, you where we would see that even though
That means the spectral emissivity and the spectral absorptivity are the same, but the
overall values will not be the same; that means, gray surface behaviour cannot be
assumed principally, primarily because the irradiation is from a source which has a
different condition as compared to that of the substrate. So, let us solve that problem and
see how we can arrive at such a such a case.
So, the problem that we are going to deal with, first of all let me draw the emissivity
curve as a function of λ in micron. So, this is about 0.5, this is about 0.8, 1.5
614
μm and 10 μm obviously, it is not as per scale in the value and this is 0.1. So, if value of
emissivity would be like this was all the way to 0.5. Up to 10, its value is going to be 0.5
and then, it goes all the way up to 0.8 and it remains at 0.8. So, this is the distribution the
spectral distribution of emissivity as a function of wavelength.
So, let us call this is ϵ λ1; this is ϵλ2 and this is ϵλ3. So, this is what is provided in this. The
problem that we have is a diffuse fire brick wall of T equals 500 Kelvin has ϵ λ as shown.
So, this is the figure of ϵλ versus λ and is exposed to a bed of coals at 2000 Kelvin.
So, we have a surface whose spectral absorptivity is shown, but it is also exposed to a
bed of coals at 2000 Kelvin. So, at the onset you can see that the emission is coming
from a bed of coals which is which can be treated as a blackbody at 2000 Kelvin. So, the
temperature from where the emission is coming from is different from the temperature of
the substrate itself.
So, we cannot use the most restrictive condition of Kirchhoff’s Law. So, we do not know
whether this ϵ is going to be equal to α. So, what you have to find is total hemispherical
emissivity and emissive power of the brick wall and second is total absorptivity of the
wall due to irradiation from the coal.
So, these are the two, I need to find out ϵ, I need to find out the emissive power of the
brick wall and total absorptivity of the brick wall due to irradiation from coal. So, we
first start with the definition of ϵ; where, Ts is the temperature of the surface. So, by
definition it is going to be

  ( )q
0
,b ( , Ts )d 
 (Ts )  
 q (T )d 
0
b s
So, this can now be simplified, this is nothing, but this one is nothing but this qb, the total
emissive power of a black body which is σT4. So, this what is known, but I can break this
into 3 parts from 0 to 1.5 from 1.5 to 10 and from 10 to infinity because between 0 to 1.5,
the value of ϵλis known and is a constant. Between 1.5 and 10 is also known and between
10 and infinity, it is also known to us.
615
So, I am going to break the numerator into 3 parts in the following way that my ϵ which
1 1 
is at Ts is going to be ; the constant value of epsilon between 0 to 1.5; 2 between
1.5 to 10 and so on.
So,
1 2 
 q d 
b  q d 
1
b  q d 
2
b
0
 (Ts ) 1  2  3
qb qb qb
where, qb is the radiation emissive power of a blackbody over the entire
So, this is simply dividing it into 3 parts and adding them together now, if you look at
this part only it is nothing but the black body radiation function.
So, it simply tells you that what fraction of energy is going to be contained between 0 to
λ1 for emission from a blackbody which is the definition of blackbody radiation function.
So, this can be written as
 (Ts ) = 1 F0 1  2 [ F02  F0 1 ] 3 [ F0  F02 ]
So, this gives me fraction of energy which is contained between λ1 and λ2.
616
We understand that F0 is simply equal to 1. So, the fraction of energy contained
between 0 to λ is simply going to be equal to 1. Now, I have to evaluate all this, in order
to evaluate these, I need to first find out what is λ 1Ts. So, the λ1 in this case is 1.5 μm, if
λ2 is 10 μm and so on.
So, Ts is 500 Kelvin. So, λ1Ts is 750 μmK and the corresponding value of F which you
can see from these tables that we have provided in the last class for ϵ T s to be equal to
F0  is approximately equal to 0. On the other hand, when you do this;

750, the value of
this is going to be λ2Ts. So, λ2Ts is 10 μm, Ts is 500 Kelvin. So, this is 5000 μm.K and the
F02
corresponding value of from the same table is about 0.634. So, this from here is 0.1,
2 3
is 0.5 and is 0.8.
So, I am going to write this as 0.1, this as 0.5 and this is equal to 0.8. So, the value of
 (Ts ) would simply be equal to 0.61. So, the hemispherical emissivity of the surface at
T equals Ts that means, at 500 Kelvin would simply be equal to 0.61. So, given the
spectral distribution of ϵ using the black body radiation function, one can obtain the
overall emissivity from the spectral emissivities which are provided in the experimental
results and what is going to be its emissive power?
4
If this is a blackbody, then its emissive power would simply be equal to  T or in this
4
case  Ts , but it is not a black body. So, therefore, its emissive power is going to be
  Ts 4 and we have already evaluated epsilon, the hemispherical emissivity of the
4
surface over all the wavelengths to be equal to 0.61. So, 0.61 times  T would be the
emissive power of this wall which is at 500 Kelvin. So, when you plug in the values the
value that you are going to get is going to be 2161W/m2.
617
So, that is the emissive power of the wall, 2161W/m 2 is the emissive power of the wall.
Now, comes the calculation of the absorptivity.
So, how do I calculate α? The alpha is defined as


  q ( ) d 

l
0
 
i
 q d 
0
So, this is the spectral incident radiation and this is a spectral absorptivity. Now, since
the surface is diffused, this surface is diffused, it was mentioned in the problem itself and
the keyword is diffused if it is diffuse; then,    . So, whatever figure that we had for
ϵ would be the same thing for this case as well.
So, this was the value of  which was provided. Since, the surface is diffused the same
figure; the same values can be taken as that of the spectral absorptivity. The question that
remains is what am I going to do with this remember that this incident radiation is from
coal which is approximated as a blackbody at T equals 2000 Kelvin.
So, therefore, the value of α can simply be used as

1 2 
   q (T )
1 b c   2 qbT (c)
1
   q T (c)
2
3 b
0
 
 
 
 qb (Tc )
0
 qb (Tc )
0
 q (T )
0
b c
618
Now, these are constants; we know the values to be equal to 0.1. This value is equal to
0.2 and this value is equal to 0.8. However, this part is simply the black body radiation
function. So, this alpha then can be written as
   1 F0 1   2 [ F02  F0  1 ]   3 [1  F0 2 ]

What remains is then to calculate 0 to λ 1, but we have to keep that, in this case λ 1Tc is
equal to 1.5 μm × 2000 Kelvin. This is what we are different from the previous one,
while we are calculating λ1Ts; now we are calculating λ1Tc. Ts was 500 so, when we were
evaluating the emissivity, it is the temperature of the substrate which is important.
When we are calculating the absorptivity, it is the temperature of the source which is
important. So, this is ϵ for the surface substrate and α is the temperature of the source
which is 2000 Kelvin. So, when you do that the value of F 0-λ would turn out to this you
can calculate.
F0 1
So, this would turn out to be 0.273. So, would be 0.273and the value of would
come out to be 0.986 and so on. And again, it is going to be λ2Tc; λ2 is going to be equal
to 10 and Tc is going to be equal to 2000. So, you should be able to find out what is the
value of this.
So, once you do this, the value of α; the final value of α after you put these things. So,
this is equal to if I look at the profile again, αλ1 is 0.1, αλ2 is 0.5 and αλ3 is 0.8. So, this is
0.1, this is 0.5 and this is 0.8. So, when you plug in all these numbers, the value of α that
you are going to get is 0.395. So, the value of the hemispherical absorptivity is equal to
0.395 ,ok compare that with the value of the hemispherical emissivity which is 0.61.
619
So, even though    ; the ϵ is not equal to α, that is the message which I would like
you to take from this class. So, the spectral values of emissivity and absorptivity maybe
the same, but the hemispherical absorptivity may not be equal to hemispherical
emissivity. The primary reason for that is the emission is calculated based on the
temperature of the substrate itself which is at 500 Kelvin; whereas, the alpha the
absorptivity is calculated based on the temperature of the source from where the
emission is coming.
So, the surface therefore, we can say that the surface is not gray; ϵ is not equal to α. The
emission was associated with 500 Kelvin; whereas the absorption is related to 2000
Kelvin. So, this is a nice exercise which would show you how to evaluate the overall
emissivity and the overall absorptivity with the help of black body radiation functions.
And to appreciate that even though the spectral values maybe same, the overall values
may not be equal; stressing the fact that it is not a gray body.
So, this I considered to be very useful example and there are many such examples both
solved and unsolved in the textbook and I would advise you to take a look at them and
see how these quantities are calculated using the table of blackbody radiation functions.
620
Heat Transfer
Lecture – 53
Solar Radiation and the Concept of View Factors
So, therefore, we have now a fair idea of what is emissivity and what is absorptivity. The
same way we have obtained the expressions for absorptivity, the expressions for
reflectivity and expressions for transmittivity can be obtained. Exactly the same relation
except that the spectral absorptivity values are to be replaced by the spectral reflective
spectral transmittivity values and so on.
In all 3 cases; that means the absorptivity, the transmittivity, and the reflectivity in all
these cases the incident radiation from a source is to be taken into account, unlike the
case of emissivity whereas, the surface temperature itself is of primary concern. So, that
is the only difference that we have in the definition of emissivity and the 3 quantities
which characterizes the which are the absorptivity, reflectivity, and transmissivity these 3
depend on incident radiation whereas, the emissivity depends on the temperature of the
surface itself.
There is a vast literature available on the values of these properties for a number of
surfaces under different conditions. So, if you ever come across a problem it is almost
sure that the reflect the radiative properties of that surface of that material will be
available in the literature. And, we also have seen when and how to make the
assumption, that the grey surface behavior can be assumed, and if we can make the grey
surface assumption, then the hemispherical emissivity would be equal to hemispherical
absorptivity, which will reduce complexities in our subsequent calculations as we would
see later.
But, there is one more part of radiation which is very useful, which has practical
applications and which we receive almost every day we I am talking about the solar
radiation. So, on the earth’s surface we have an atmosphere, and through the atmosphere
the radiation the sun’s radiative energy comes to the earth surface without which we any
life form on earth will not be able to sustain.
621
So, it is important to know; what is the value of this radiation, which is coming from the
sun just before it enters the atmosphere. And, once it enters the atmosphere there is going
to be some absorption into the sun’s rays, the absorption would be spectral in nature so;
that means, some of them would be absorbed at a specific wavelength more as compared
to another wavelength.
And, it is also going to get reflected; re-reflected and come will come back to the surface
again. So, these factors will have to be kept in mind while designing many of the useful
devices for example, a solar collector. So, how much of energy you are going to get
based on the angle of the sun, based on how much of diffused sunlight you are going to
get from other sources, which are reflecting the sun’s radiation and so on.
But, fundamental quantity of importance is how much of radiation you are going to get,
if you have place a plate just outside the earth’s atmosphere in a direction perpendicular
to the ray of the sun, that is a constant.
So, when you place an a surface just outside the earth’s atmosphere and measure how
much of radiative energy, it is getting from sun that that is a constant and that is
generally denoted by G S which is a constant. And, if you look at the figure, the way it is
so, this is where the radiation from the sun.
622
Is coming from and it is hitting the earth’s atmosphere and I have placed the surface
which is perpendicular to the direction of the sun’s rays. And, this G S it is also known
as the solar constant the solar constant is measured and the value of G S is 13 53 Watt
per meter square ok. So, this is the solar constant.
But, the solar energy which is falling perpendicularly normal to the outer surface of
earth’s atmosphere G 0, which I have denoted here must be equal to G S times cos theta
also in Watt per meter square. So, this G 0 is the extra-terrestrial solar radiation. So, this
is what it is and the total radiation radiative flux. This is the total that you are going to
that is the total radiation, that reaches the earth’s surface from all directions over the sky
is called over the sky is a sum of what we get over here G S times cos theta plus q df all
in Watt per meter square.
So, this is the direct solar radiative flux per unit area normal to the direction of the solar
beam and so, this is this is G S is this one and G S cos theta is on this surface, but this q
df is the diffuse solar radiation incident on a horizontal surface. So, that is all about the
solar radiation and the other radiative properties.
Next, we are what we are going to do is something very interesting it is known as the
concept or the of view factors. So, what is a view factor? But, the first thing is we what
we are going to assume is that the, whenever we are talking about the surfaces are diffuse
emitters and diffuse reflectors.
623
So, view factor is a very interesting property, very interesting concept in radiation. Now,
when a body is when radiation is coming out of a body and you place another one close
by not all radiation coming out of body one would reach body 2 ok. So, if you keep them
close together the fraction of energy emitted by surface 1 which is intercepted, which
falls directly on body 2 would keep on increasing. As, I move the surface to away from
surface 1, the fraction of energy emitted from surface 1, that directly reaches surface 2
will decrease. The part of the energy is going to go out in the space in between in
between the 2 objects.
But, as I bring them close together the fraction increases. So, it is directly a function of
how much of surface 2 is visible to surface 1. So, that is known as the view factor. So,
view factor is the fraction of energy emitted by a surface that is going to fall directly on
surface 2. So, it is expressed as capital F with a subscript 1 2. So, F 1 2 denotes the
fraction of energy emitted by surface 1, that is directly incident on directly intercepted by
surface 2.
And, as I said let us think about these as 2 plates, if I bring the 2 plates close together this
fraction would increase, if they touch each other then this fraction has a value equal to 1.
That means, all energy emitted by 1 is going to be intercepted by 2. If, I move them away
this fraction the value of this fraction decreases.
So, if you think a bit more in this then you can clearly see that it all depends if we if you
are an observer on surface 1, then whatever be the solid angle subtended by 2 on 1, 2 as
seen from 1 is going to denote what fraction of energy from surface 1 is going to reach
surface 2. If, I bring them to close together the solid angle increases, if I put them back
away the solid angle decreases.
So, therefore, the view factor is going to be a strong function of solid angle. Now, these 2
surfaces besides that these 2 surfaces are parallel to each other. Let us say I have one
surface at an angle and the second surface is also at an angle. So, what I need to do is if
this is the case then the view factor would be different.
So, the view factor not only depends on the solid angle subtended by 2 on 1 or 1 on 2, it
also depends on what is the what is the angle with which they are inclined with respect to
the this parallel these parallel lines. So, in order to do that first have to take a projection
624
of this surface on to this plane, and a projection of this surface onto this plane. And, then
I evaluate what is the solid angle subtended by 2 on 1.
So, it is not only the solid angle the angle of these surfaces will also play an important
role. And, you know that the solid angle is expressed in terms of area divided by r square
and we can extend that to find out what is the view factor from 1 to 2, when both of them
are at an angle and the distance between the 2 is known to us.
So, whatever be the energy that is going out of one what fraction of it is going to reach 2
is the view factor. And, this is very important industrially, because if you think of let us
say you have a cylindrical furnace ok. Where the 2 ends are one end is heated the other
sides are simply working or the let us say the top is heated circular top is heated and you
are keeping the keeping the material, which you want to dry, which you want to expose
to high temperature at the back.
So, you would like to know, what is the amount of energy provided at the top is going to
hit the material directly. Because some of it is going to hit the cylindrical side walls, and
some of it is going to go opposite and hit the material which you would like to treat.
So, the concept of view factor is extremely important especially when you are dealing
with enclosures ok. Special enclosures for example, furnaces. What is the fraction of
energy? That you are actually going to get based on your direction, based on your
distance and so on. So, we would like to go a bit deeper into the view factor and get to
and try to find out, what is the, what there are literature available on the values of view
factors for common geometries. We can utilize and manipulate them in order to obtain
view factors for unknown or view factors for complex geometries, view factors for
complex orientations is a method which is known as view factor algebra. And, we would
see; what are the relations, which exist for view factors between surfaces and so on.
So, let us first start with the simplest possible case is view factor between elemental
surfaces. So, you have a surface, which is A 1, on that I have a surface which is d A 1
and this is the direction of the area vector. Over here I have another surface which is A 2,
this surface of the elemental surface over here is d A 2, this is the elemental surface over
d A 2.
625
And, let us say this is the area vector direction of area vector over here. So, first of all I
connect these 2 and when I connect these 2 let us say the distance between them is r, this
angle is theta 1, this angle is theta 2 and you have emission coming out of these surfaces
the intensity I call this as I 1 and this call as I 2.
So, what I need to do is let us first say that d omega 1 2 is the solid angle under which an
observer at d A 1 sees the surface element d A 2 ok. And, I 1 is the intensity of radiation
from intensity of radiation from surface 1 ok. And, it is going out diffusely in all
directions in the hemispherical space. So, the rate of radiative energy the rate of radiative
energy let us call it as d Q 1, which leaving which leaves d A 1 and strikes d A 2 would
be d Q 1 is equal to d A 1, I 1 cos theta 1 d omega 1 2.
So, let us see let us understand this a bit more carefully. This is the radiation intensity.
Radiation intensity is energy per unit area per unit solid angle ok. So, d Q 1 is the rate of
radiative energy, then per unit are would be d A 1 cos theta d A 1 cos theta. And, the
solid angle subtended by 2 on 1 is d omega 1 2. So, if you consider I once again, it is the
amount it is the intensity is defined as the energy as the energy, which go which goes out
of a surface per unit area per unit per unit area which is d A 1 cos theta per unit per unit
solid angle and so on.
And therefore, since it is a diffuse surface, so there we do not need to invoke the we do
not need to invoke the spectral nature of the surface. So, the only thing which remains to
be evaluated is what is d omega 1 2, d omega 1 2 is area divided by r square. So, what is
the area d omega 1 2 is the d A 2 what is the solid angle d A 2 is making on is creating
on d A 1. So, I first have to take a projection of d A 2, which is perpendicular to the
distance vector between d A 1 and d A 2. So, d A 2 I take it like this. So, d A 2 cos theta
2 is the area of d A 2, which is perpendicular to the area vector r and the distance
between them is r.
So, by definition of a solid angle it is area by distance square. So, d omega 1 2 is simply
this. So, therefore, d Q 1 would be equal to d A 1 I 1 cos theta 1 cos theta 2 d A 2 by r
square. So, this is the this is the radiative energy which is leaving d A 1, that strikes d A
2 in terms of the areas, the in terms of the elemental areas, the distance between them
and the 2 angles the areas make with the corresponding normals theta 1 and theta 2.
626
Now, so, this in order to find the view factor this the rate of radiation energy d Q 1
leaving d A 1 that strikes d A 2 must be divided by the total rate of energy which is
which is originating from A 1 into the d A 1 into the hemispherical space. So, in order to
obtain the fraction this is the numerator. So, the denominator would be the total energy
emitted by the d A 1 into the hemispherical space. So, how do I find that from our earlier
discussion, we know that the rate of radiation, rate of radiative energy let us call it as Q
1, on leaving the surface element, d A 1 in all directions over the hemispherical space
hemispherical space. This we have done before this Q 1 would be d A 1 phi the azimuth
angle from 0 to 2 pi theta 1 to be 0 from 0 to pi by 2, the intensity cos theta 1 sin theta 1
d theta 1 times d phi.
So, this expression we have obtained before as well ok. So, for a diffused so, there is
nothing new in here. For a diffusely emitting and diffusely reflecting surface I 1 is not
going to be a function either of theta and phi. So, I 1 can be taken outside and the
integration without I 1 can be performed. So, once you perform that it simply going to be
phi times I 1 d A 1.
So, this Q 1 is known to me I already have from my previous slide, the expression for d
Q as d A 1 I 1 cos theta 1, cos theta 2 d A 2 by r square. So, therefore, the view factor d
F, it is an elemental it is a view factor between elemental area d A 1 to d A 2 would be
the
627
amount of energy released by emitted by 1, that strikes d A 2 directly which is d Q 1
divided by the total energy emitted by surface 1 in the hemispherical space.
So, if you divide this by this what you are going to get is cos theta 1, cos theta 2, d A 2
by pi r square. So, this is a very important formula which tells me tells us, what is going
to be the elemental view factor between 2 surfaces A 1 to A 2. So, the amount of energy
released amount of energy emitted by 1 that strikes 2 directly, which is this part and the
total energy emitted by surface d A 2 which is Q 1 that we have evaluated. So, this is the
1. If I have d F d A 1 to d A 2, then d F d A 2 to d A 1 can simply be obtained by
changing the 1 to 2. So, this is simply going to be cos theta 1 cos theta 2 d A 1 by pi r
square.
So, the elemental view factor from 1 to 2 has this expression from 2 to 1 can simply be
obtained by interchanging the subscripts 1 and 2. So, what this if you look at these 2
expressions? They are the same except d A 2 appears over here and d A 1 appears over
here. So, if you divide this by this, then what you get here is d A 1 the view factor the
elemental view factor between d A 1 to d A 2 would be equal to the area, d A 2 d F d A 2
to d A 1.
This is a very important relation in view factor this is a very important relation involving
view factor, it is known as the reciprocity relation. What this reciprocity relation tells us
is that product of the area, multiplied by a product of the area and the view factor from
the area to another area is equal to the product of the other area multiplied by the view
factor from the second area to the first area.
So, this is the reciprocity relation related to view factors. The same type of analysis can
also be done for areas of finite size A 1 and A 2. If, that is the case then the view factor
from 1 to 2 is expressed as F 1 2 and the view factor from 2 to 1 is expressed as F 2 1.
So, F 1 2 stands for the fraction of energy emitted by 1 that strikes 2 directly. And,
similarly for F 2 1, we have to keep in mind that all these surfaces are assumed to be
diffuse emitters and reflectors.
So, there is no directional dependence of the intensity of radiation coming out of these
surfaces. So, once you extend these relations for finite surfaces the reciprocity relation
which we have obtained for the elemental surface can simply be written in following
way.
628
So, the reciprocity relation for a finite surface reciprocity relation for a finite surface can
simply be written as A 1 F A 1 to A 2 is equal to F 2 F 2 F A 2 to A 1. So, this is the
reciprocity relation which is important in view factor geometry. So, these are essentially
the properties of view factors.
So, the first property of view factor is the reciprocity relation. The second property of
view factor is known as the summation relation. So, what the summation relation tells us
is that, let us say a surface a surface 1 is enclosed like this where this is 2 and this is 3.
So, there would be view factor from between 1 and 2, which is F 1 2 that is the fraction
of energy emitted by 1 which directly falls on 2.
There would be another 1 which is F 1 3 the fraction of energy emitted by 1 which falls
directly on 3. And, there could be because of the nature of this; this area some amount of
energy some amount of emission which is coming out of 1 may fall on 1 it itself, in
which case the view factor of that can be denoted as F 1 1. So, F 1 2 for A 3 zone
enclosure so, to say the some energy from 1 can fall on 2, some energy fraction can fall
on 3, and some can fall on itself.
If this is a curved surface, some amount of energy emitted by the surface may fall on the
surface itself, in which case it has a view factor to itself. So, it has a view factor with
respect to surface 2 A, view factor to surface 3, but it may also have a view factor to
itself.
629
So, if you add all these view factors add all these fractions then they must be equal to 1.
So, the fraction of energy emitted by 1 which falls on 2, fraction of energy emitted by 1
which falls on 3, and fraction of energy emitted by 1 which falls on 1 if you sum them all
together what you are going to get is simply the new simply equal to 1. So, which is
known as the summation relation and which is expressed as F 1 1 plus F 1 2 plus F 1 3 is
equal to 1. So, that is the second relation of view factor which would be useful while
dealing with such dealing with such cases.
And, there are other things which are the third type of relation is F A i to A I is equal to 0
if A i is plane or convex F A i to A i is not equal to 0, if A i is concave. Let us think
about this for some time ok. What it tells us is that, if a surface is a flat surface, if a flat
surface is like this page.
Then, in any emission which is coming out of this page of this surface, there is no chance
of it coming on to the surface directly. So, if it is it can go to the wall gets get it may get
reflected and come may come back onto the surface, but there is no chance of emission
from surface directly falling on surface 1 itself. Which is true, if it is a plane 1 or if it is a
convex 1 If, it is a convex surface then of course, there is no chance of emission coming
out of 1 falling on 1 itself.
But, on the other hand, if this is a concave surface, then the emission from this point
may hit the other points on the surface. So, for a concave surface you are not going to
630
you cannot say, that F 1 1 or F A I A I is equal to 0. So, the third condition on the view
factor would be that it is equal to 0 for a plane or a convex surface, but it will not be 0, if
it is a concave surface. So, we have 3 relations then for evaluating the view factors, one
is the summation rule, the second is reciprocity rule and the third which depends on the
nature of the surface itself. And, you can be you can you can find out different-different
values I mean you can find out the values of the view factors for several such cases.
Now, the other part is there are a vast majority of literature, which is which deals with
the values of the view factor, which are evaluated analytically ok. The analytical
expressions are quite complex I will not go into those expressions in this course, but
there are graphical methods. There are graphical there are many graphs which are
available based on different geometries what would be the view factor.
For example, you can consider of an enclosed cylinder, what is the view factor of the top
wall with respect to the bottom the top circular plate with respect to the bottom circular
plate which; obviously, depends on what is the distance between the cylinder and what is
the radius of the cylinder?
So, there are there are charts available which would tell you what is the view factor for
such a case. You may have a cube in you may have a cube or a cuboid and the view
factor from 1 of the flat surfaces to any one of the other surfaces can all are also
documented in the form of charts. So, you would be you would be able to obtain you
would be able to use that as well.
So, these various geometries and their corresponding view factors are provided in many
in many of the books in your text book itself, you would see that the view factors for
such surfaces or combination of such surfaces in an enclosure are provided. So, while
solving for while trying to find out the unknown view factor, first try to see if
information about such a configuration is available in the literature, in your text book.
And, if it is so, then directly use the values of those view factors for your calculations. If
not then maybe sufficient inputs given in the problem itself, which when you visualize
the system appropriately, the value of the view factor would be apparent. And, if one
value of the view factor can be obtained, the other values of the unknown view factors
can also be obtained either by the use of summation rules, reciprocity relations and so on.
631
So, in the next class we will solve some problems on finding out the view factors either
using chart or using any of the relations that I have just described.
632
Heat Transfer
Lecture – 54
Determination of View Factors
Whenever we have emission from a surface and there is another surface nearby what is
normally going to happen is that all the emission which is coming from surface 1 will not
strike surface 2. Therefore, the energy emitted by surface 1 may not be intercepted by
surface 2, how much of energy what fraction of energy that is emitted by surface 1,
which is going to fall directly on surface 2 is known as the view factor. So, if the two
surfaces are very close together then the view factor of 1 to 2 would be obviously, equal
to 1; that means, all the energy emitted by surface 1 is going to be incident on surface 2
directly.
So, this concept of view factor, how much of the emitting surface is visible from another
surface nearby is important, because it would give us a tool during our calculation of
emission and interaction radiative interaction between surfaces which will form part of
an enclosure. So, whenever we would like to design an enclosure and we would like to
find out, how much and the three surfaces is forming, let us say the three surfaces
forming the enclosure are at different temperatures, one is receiving the heat, the other is
emitting the heat and maybe the third surface is insulated.
So, how much of interaction and how much of energy is going to be incident on one
surface is going to come out of the other surface, how much of energy is to be extracted
from a third surface to keep it isothermal, all these are important in industrial
calculations for example, the design of a furnace. So, the concept of view factor is
therefore, extremely important in the designing of radiative heat exchange processes in
an enclosure for example, in a furnace. And we know what are the uses of furnace from
baking a surface at a specific temperature to exposing another surface at a very high
temperature? The third application could be drying of a surface through radiative heat
and so on.
So, the examples of radiative heat exchanger in a furnace are many and from different
fields, but the first thing that one needs to know is what is the fraction of energy emitted
633
by one which is going to be intercepted by the second directly, without considering
reflection of energy from the adjoining surfaces. So, this is view factor which we have
discussed in the previous class.
And we have also seen some of the properties of view factors for example, if you take a
view factor and you find out the view factor of that surface to all the surfaces nearby
which forms an enclosure. So, whichever room that you are sitting right now, the let us
say the floor of that room is your surface 1. And then you have 5 more surfaces which
are forming the enclosure the four walls and the roof which forms so, these 6 surfaces
will form the enclosure. Now when you consider the surfaces on the floor, then the
possible view factors are F1 where 1 refers to the floor, F 12 from the floor to the roof F13
F14 F15 are all the walls which we have.
And interestingly there is a view factor of the floor to itself; that means, it is possible so,
that some of the energy emitted by the floor is going to be absorbed by the floor directly.
Now when you considered the floor obviously, if the floor is plain and level there is no
chance of energy emitted by the floor absorbed by the floor itself. So, F11 which is the
view factor of the floor to itself 1 1 that would obviously, be equal to 0 and the sum of all
other view factors would be equal to 1.
So, the first rule is the rule of summation rule which tells you that the sum of all possible
view factors from a surface would be equal to 1 which obviously, is intuitively correct as
well, since you have since a view factor gives you the fraction. So, the sum of all
fractions of energy coming from a surface has to be equal to 1.
The second one which we have derived in the last class is the concept of reciprocity, the
reciprocity relations where we have shown that A1F12, A1 being the area of surface F12 is
the view factor of one with 2 should be equal to A2 that is the area of the second surface
times F21. So, the reciprocity would give another set of relations between surfaces which
are forming an enclosure.
So, the first thing is summation the second rule is the reciprocity relation and the third
rule, which applies to convex and plane surfaces for which you can see that A 11 is equal
to 0. So, for a convex or a plane surface no fraction of energy emitted by the surface is
going to be absorbed by the same surface, which we cannot say for the case of
634
concave, because when you think of a concave surface, part of the energy emitted by the
surface will probably be absorbed by the surface itself.
So, that is fine, but and we have also seen how one can analytically calculate, the view
factor of different surfaces and the view factor of obviously, when you talk about the two
surfaces the view factor would depend on the magnitude of the area. That means,
whatever be the length scale associated with area 1, whatever be the length scale
associated with area 2 and obviously, the distance between the 2. So, the view factors are
normally expressed graphically based on the relevant dimensions and the separation
between the two surfaces.
So, in your text and in any text you would find that these view factors are calculated for
some of the standard surfaces and those standard relations for view factors the results for
the view factors are expressed in graphical form. And it could be very useful to use those
graphs for a number of applications, where at least you would be able to find one of the
view factors.
Now, if you find one of the view factors, then the next thing one should do is try to see,
if you can use summation rule, reciprocity rule or the nature whether its curved or a flat
surface rule to see if another view factors which are forming the enclosure if the other
view factors can also be evaluated. So, you would solve we would see some of the
problems towards the end of this class, but I will give you some idea of how the how the
results of view factors for common surfaces are expressed in the graphical form.
635
So, the first one that you would see over here, this is the view factor of an elemental area
dA1 to a surface A2 with dimensions of L1 and L2 and d is the distance between them. So,
dA1 is the elemental surface, d is the distance between the surface and the tip this edge is
the corner of the surface A2, which has dimensions of L1 and L2.
So, you will number of curves like this where the x axis is L 2/D, D is the separation and
the family of curves are for different values of L 1/D. So, this curve is elemental this is
FdA1-A2 from an elemental surface to a rectangular surface, which is having an area A 2 and
dimensions L1 and L2. So, the curve one is for an elemental surface area. Please note the
location of the elemental surface with respect to the surface so, that is the corner is
directly above dA1. So, if you know this the values of L2, D and L1, you should be able to
find out what is the value of the view factor.
Now, you come to this one so, this is going to give you FA1-A2 from a rectangular surface
A1. So, this a is rectangular surface A1 to another rectangular surface A1. So, this top
surface is A1 which are parallel to each other and directly opposite. So, here you see that
this is a cuboid and we are talking about the base area which is denoted by A1 the area
which is parallel and directly opposite to A 1 is; obviously, the top of the rectangle. So, I
am trying to find out what is FA1-A2? So, this is from a rectangular surface A1 to another
rectangular surface A2 which is parallel to A1 and directly above or directly opposite to
it.
636
So, these are also expressed in graphical form the L 1 and L2 are the dimensions of the
rectangular surface, D is the distance between the between them. So, the x axis as before
L2/D and the family of curves are for different values of L 1/D. So, one would be able to
obtain what is going to be the view factor for such a case. And if you find out FA1-A2, FA1-
A3, FA1-A4, FA1-A5, FA1-A6 can simply be calculated using any of the known relations like
reciprocity and or the summation rule.
The 3rd one again it gives you the value of A 1 to A2 from a rectangular surface A 1. So,
this is your A1 to a rectangular surface A2 which are touching each other, so that means,
which are adjacent and perpendicular to each other. So, you can think of you can
imagine A1 and A2 are the two sides of this rectangle or if you open up like this if you
open a book like this where they are perpendicular to each other, the two surfaces are
perpendicular to each other, then the view factor of this surface to this surface, which is
adjacent; that means, they are touching each other and which are perpendicular to each
other it is that view factor that we are trying to find out in here. So, inside a rectangle the
base and the adjoining surface is the value which we are trying to get.
So, again it is going to depend on L 1, L2; that means the length scales which define the
area. So, if you look over here then you have different value so, this is L 2/W and this is
L1/W, where W is the width of these two plates, L 1 and L2 are the other length scales
which are present here. So, again you can find what is the view factor from here?
Then last one that you see is F A1-A2 between two coaxial parallel disks. So, these two
coaxial parallel disks they are so, therefore, the centre of this and the centre of this one,
the they are coaxial and you should be able to find out what is the view factor of 1, with
respect to the other provided you know what are the values of r 1 r2; that means, the r is
radius of disk one, radius of disk two and the separation between the two. So, the x axis
is L/ r1, the y axis the different values of the curves are r2/L and you can find out what is
FA1-A2.
So, using the mostly these are the four fundamental figures, that can be used to find out
what is the view factor of a surface to the other. In some cases a combination of these
would be required, in most of the cases the view factor obtained from these graphs are to
be combined with the summation rule or the reciprocity rule relation to
637
obtain the unknown view factor the other unknown view factors. In some cases you
really have to do some sort of an algebra in order to find out what is the view factor for
such cases. So, quickly see some of the examples of the use of these charts to obtain the
view factors of unknown surfaces.
So, the first one that we are going to do is let us think of a rectangular surface like this.
So, this is the rectangular surface and you have but this is a cube. So, all sides are same
and you have to first find out what is the view factor from the base of a cube to each of
these surfaces. So, we will denote the base surface as 1, the top surface as 2 and the
others are 3,4, 5 and 6.
So, from this figure you are going to find out what is the view factor your L 1 and L2 are
the same here and it is also equal to D since it is a cube. So, your L 1/D is L2/D is 1 so,
from the corresponding point where L2/D is equal to 1, you go all the way to the curve
which corresponds to L1/D equals 1, and you read the value of FA1-A2.
Once again A1 is the base of the cube A2 is the top of the cube, since it is a cube L 1 is
equal to A2 is equal to D, so that means, both these ratios would have value equal 1. So,
you go up to the right curve and find out what is F A1-A2. So, that is the use first use of this
curve. So, from the figure what you are going to see that you are going to check on your
own is F12 is equal to 0.2, that is from the figure.
638
You can also write that Fij is equal to 1 and in this case j varies from 1 to 6. And F11; that
means, the view factor of 1 to itself would obviously be equal to 0 since it is a flat
surface. F12, we have just found out to be equal to 0.2. And by symmetry since it is a
cube F13 would be equal to F 4 would be equal to F15 and F16, Fij, for j = 2 to 6 should be
equal to Fij would be equal to 0.2.
So, F13, F14, F15 and F16 using the summation rule and knowing the value of F 12 to be
equal to 0 and knowing the symmetry involved in a cube, your F ij should be equal to 0.2.
So, all from 1, all view factors except to itself would be equal to 0.2. So, F 12 is 0.2 , F13,
F14, F15 and F16 all are going to be equal to 0.2.
The second one is a cylinder like this and the question that you have we have to answer
is from the what is the view factor so, all these are what are the view factors? So, this is
for 1 from the this and this is from the base of a cylinder of radius r and height L to it is
top surface and the cylindrical surface is for r equals L. So, we need to find out what is
the view factor from the base of a cylinder of radius r so, this is radius r and length L. So,
from the base of the surface to the top of the surface for the case, when r is equal to L so,
what you need to do is you need to use obviously, this figure, but in this case r 1 would be
equal to r2 and the value of L is assumed to be for the problem given r 1 is equal to r2 is
equal to L. So, therefore, your L/ r1 and r2/L will have values equal to 1.
So, you can read what is the value of F12 so, if the radius r and L if both are same, then
from the graph from figure, you would see that you should check it on your own F 12, so,
this is my surface 1, this is my surface 2 and the curved surface is curved surface is 3.
So, F12 is 0.2 and then I invoke summation rule which says F11 is 0, F12 is equal to
0.2 and therefore, sorry this is 0.4, F 11 so, this is 0.4 and therefore, F 13 must be equal to
0.6. So, this is definitely the value that we are looking for since F11 is 0, F12 is equal to 0.
When you use reciprocity relation, then A3F31 should be equal to A1F13. So, we need to
find out what is the view factor of 3 from 3 to 1. So, F 31 is A1/A 3F13, we know what is
the value of F1 F, we know the value of what is F13 so, your F31 would be equal to
F31= (A1/A3) F13
And when you put the values in there you are going to get equal to 0.3 for r=L.
So, the value of F13, we have already obtained it to be 0.6 so,
639
A1/A3 ratio for r=L, you would give you the value of F 31 to be equal to 0.3. So, what you
have obtained is F12, F13 and F31, so, all values are obtained in this fashion.
The third one is more interesting so, the third one simply tells you that this is the
between the surfaces of two concentric spheres with A1 and A2 being the inner and outer
surface. So, this is A1 and this surface is A2, so obviously, what we can see from here is
that for F11, this is concentric spheres. And we have something in between the
intervening space F11 has to be equal to 0, because F11 constitutes a convex surface. So,
for a convex surface the view factor to itself of the surface to itself must be 0 because
any emission coming out of A1 is not going to fall directly on A1 again, it may get
reflected from the inner surface and may incident on surface 1 again.
But, the definition of view factor is that it is the emission the fraction of emission, which
falls directly which is emitted by a surface and false directly on itself. So, since for a
convex surface, there is no chance of an emission being incident again on the same
surface therefore, F11 where 1 refers to the inside the surface area of the inside sphere F 11
must be equal to 0. And you can also say that the summation rule and physically you can
imagine that all emission emitted from 1, since it is inside the larger sphere all emission
from 1 must fall on 2 therefore, the view factor F 11 maybe 0, but F12 will have a value
equal to 1.
640
So, what we can write from here is that the summation rule and common sense that will
tell you that F12 will have a value equal to 1 so, this is summation rule. On the other hand
you can see that some of the energy which is emitted from 2 is going to strike 1, but
some of the energy emitted from 2 may strike to itself once again. So, this since A 2 the
inside area of the outer sphere is a concave one, I cannot say that F22 is equal to 0. So,
this saying this would be incorrect, since that is not the case.
So, the two view factor so, this is for surface 1 and for surface 2, I need to find out what
is F22 and what is F21, I cannot say F22 to be equal to 0 which is incorrect so, and rather I
should concentrate on F21. Now, how can I find F21, I know what is the value of F 12. So, if
I know F12 and if I know F12, then F21 can be calculated from reciprocity relation, that is
the other relation which we have, so, reciprocity gives us A 1F12 must be equal to A2F21,
that is the reciprocity relation.
So, F21 is equal to A1/A2 since F12 is equal to 1. So, that is the other relation and you can
then therefore, calculate F22 from summation as F21 + F22 = 1. So, this must give you
F22 = 1- (A1/A2). So, that is the other relation which I would like to highlight a little bit
more.
So, simple use of reciprocity and summation would give you the value F 22. Now, A1 is
the surface area of the inner sphere and A 2 is the inside surface area of the outer sphere.
Now, let us think of a situation in which the a very small sphere is placed inside a very
large sphere. So, the sphere inside is very small therefore, A 1 is small and the sphere
outside is very large and therefore, A2 is large.
If that is the case then all emission coming out from surface 2 is going to fall directly on
itself, the small area the small sphere which is located at the centre it really does not
matter. So, common sense tells us that when A 2 becomes very small, sorry A1 becomes
very small the inside sphere becomes very small, we can in the limiting case we can
assume that it is not going to absorb anything which is coming out of 2. And therefore,
the fraction of energy emitted by 2 which is going to be incident on 2 itself directly will
have to be equal to 1.
So, whatever we thought to be correct from an intuitive point of view, if you look at the
equation this is going to corroborate that. So, when A1 becomes small and A2 becomes
641
large, F22 will have a value equal to 1. So, what I would suggest is the if you look at your
text book and some of the other text books on radiative heat transfer you would see that
there are several configurations which are provided as examples or as exercise is for you
to practice to see whether or not you can correctly evaluate what is the view factor for
such cases.
So, I urge you to practice the concept is very simple, the two relations are also very
simple the summation rule and the reciprocity relation, I have shown you three examples
of how to calculate the unknown view factor, there are examples, which you should solve
yourself to see if you have understood the concept clearly. And as I mentioned before the
concept of view factor is going to be extremely important, when we are going to
calculate radiation exchange between surfaces which form an enclosure for example, a
furnace. So, we will see how this is to be how this is to be done in the subsequent classes.
642
Heat Transfer
Lecture - 55
Radiosity Blackbody Radiation Exchanges, Relevant Problem
We are going to start with new concept in this class which is known as Radiosity. So,
what is radiosity? When you see a surface, let us say when you are looking at a hot
surface which is in an enclosure, but the side walls are also hot. So, when you are
looking at the surface, if you do not have the enclosed side walls, you are going to get
some amount of energy which is coming out of the hot surface. This is due to emission
and emission is a volumetric phenomenon.
So, because of its temperature there would be some amount of heat flux, some amount of
radiative heat, which is coming from the surface and that is the emissive power of the
surface, depending on whether it is a blackbody, or not a blackbody by the introduction
of emissivity you can predict what is the heat flux emissive heat flux coming out of the
surface.
So, using emissivity and the Stefan-Boltzmann law you can say that the emissive power
of the surface, which is at an absolute temperature of T would simply be equal to ϵσT 4, if
this is a blackbody, ϵ is equal to 1 and the emissive power is simply σT 4. Now, let us say
this hot surface is enclosed by another surface which is even hotter than the surface that,
we were considering. So, that is going to be some amount of the second surface is also
going to emit the emission will take place from the second surface. Part of which is
going to fall on the first surface and we know how to calculate that part of energy which
comes from surface 2 that falls directly on 1 by the incorporation of view factor, but that
is beside the point.
Let us say that we have the surface 2 which is emitting and this emission is going to
strike surface 1, which is emitting anyway when that happens, if this is a real surface
then and let us assume that it is an opaque surface. So, part of the energy which is
incident from 2 on 1 is going to be absorbed and part of it is going to be reflected. So, if
you are an observer and you are standing just outside of surface 1, what you see coming
643
out as radiative energy from surface 1 is not the emissive power alone, you also have to
consider whether or not the surface is reflecting some amount of incident radiation.
So, what you see while standing outside of the surface is different from the emissive
power or emissive power of the surface. So, what you see on the outside, just outside of
the area which is emitting and reflecting all at the same time is called the radiosity. So,
the radiosity refers to emissive power of the body and how much of incident emission it
is reflecting.
So, it is going to be a function of what is the value of the incident radiation and what is
the value of the reflectivity of the surface together, they would give you an idea of the
emissive potential of the body, that is the emission potential of the surface, which is a
sum of its inherent emission and the reflection of incident radiation that is what is known
as radiosity. So, whatever have described so far.
If you look at the figure I think it is it will become even more clear to you is I have a real
surface which due to its temperature is emitting so, this is the emission and there is some
irradiation, which is coming on to the surface from another nearby surfaces and its going
to reflect portion of the irradiation. When you combine this together if you are the person
who is standing over here, you not only sense the emission you also sense the reflected
portion of the irradiation and this is what is called the radiosity.
644
So, radiosity generally is not equal to the emissive power, for real surfaces. But if the
case is such that the surface is not reflecting anything, whatever is irradiated on the
surface it is going to absorb that without reflecting any portion of it. So, which;
obviously, I am not describing a blackbody in this case. So, for a blackbody for the
special case of a blackbody there would not be any reflected portion of radiation and
emissivity will solely constitute the radiosity of the surface.
So, for a blackbody the emissive power and the radiosity are equal, but for real surfaces
the blackbody real surfaces the radiosity and emissive powers are different. So, the
radiation flux which is radiosity is based on the actual surface area. And radiosity is
generally denoted by J, if it is spectral it is Watt per meter square per micron (W/m 2μm)
and from the spectral, you can find out what is the total value of radiosity, since this λ
this denotes the spectral property. So, that is why its emissive power watt per meter
square power per unit area per micron, since I have the spectral property over here.
2  /2
J  ( )    I , e r cos  sin  d d
0 0
So, this Jλ for any λ is equal to 0 it is connected with the intensity, as the intensity the
spectral and this contains both emission and radiation that is the significance of e and r
over here. So, I am trying to relate the spectral radiosity to the spectral radiative in
intensity and since it is the radiative intensity it should contain both the emissive part and
the reflected part. So, when we are trying to find out the so, this is directional and
spectral; this is directional and spectral intensity.
So, now, I we know how to do this for; how to find out how to perform this, somehow, I
need to take this out of this and once I have this Jλ this the total emissivity is going to be

J   J  ( ) d 
0
So, once again I will explain what I have written over here, this is the intensity of
radiation, I refers to the intensity. Since I have the λ over here, this is the spectral
intensity of radiation coming from a surface. And since it is coming from a real surface I
need to consider both the emissive intensity and the radiative intensity of this. So, this is
the spectral directional radiation intensity from a surface.
From this I would like to convert this to spectral only. So, I need to take care of the
direction part over here and as we have done before this J  ( ) should denote the
645
emissive the radiate radiosity of a surface, when of a surface of unit area, when it is
placed at the centre of a hemisphere. So, how do I do that? I
646
integrate over all possible θ and Φ. And therefore, this should give me the emissive
power per unit wavelength that is why I have the spectral.
Once I have the spectral one, then I should be able to find out by integrating it from 0 to
∞, what is the overall radiosity of a surface by integrating over all possible wavelengths.
The first integration becomes easy to perform, if we assume that its diffuse reflector and
diffuse emitter, if the surface that we are talking about is a diffuse reflector and if it is a
diffuse emitter; that means, after whatever comes out of the surface it does not have any
directional property and whatever is incident on it after reflection, the radiation has no
directional property, then I  ,e  r does not depend on θ and Φ.
So, this can be taken outside of the integration sign and the integration can be performed
for cos  sin  d d between the limit 0 to 2π and 0 to π/2. So, once we do that we know
that the result for that we have said before the result for that would be equal to π. So,
therefore, Jλ the spectral radiosity is simply going
J  ( )   I  ,e  r
This is first relation and the second relation; obviously, is going to be
J   I e r
This is the overall radiosity and this is the spectral radiosity, but more importantly is the
concept that what radiosity is that it is sum of these two.
The next step which we are going to do is blackbody radiation exchange, what is
blackbody radiation exchange? Let us say I have two blackbodies. So, these are all
647
blackbodies and this has an area Ai, which is at a temperature of Ti, this has an area Aj
which has a temperature of Tj. So, this is let us say the area vector and the area vector
here is nj, it does not matter whatever will be the area vector.
So, when an observer sees the blackbody it sees not only its emissive power, it sees also
whatever emission coming from J getting reflected from b and therefore, it is what it see
is coming out as emissions as Ji, which should be equal to, we are going to find out what
this Ji is similarly the emission which is coming out from over here is going to be the
radiosity of the black of the object.
For the special case where the body is a blackbody it is not going to reflect anything. So,
for this special case of a blackbody the radiosity and emissive power are the same. Now,
these two are blackbodies so they are only going to emit energy based on their inherent
temperature, not on any reflection from incident from any nearby from any object from
any other surface.
So, for surface i and for surface j the two blackbodies, who are interacting with each
other, we know that we have to take into account the radiosity, because that is the actual
potential of the surface in terms of its radiative power. So, it is not just the emission it is
a reflection as well. So, if you think of the emission, emissive power of an object of a
surface, it is the radiosity and not the emissivity, but since we are dealing with a
blackbody the emissive power and the total radiative power which is the radiosity they
are equal, because the absorption is going to be total and reflection is going to be equal
to 0.
So, for this case only since we are dealing with blackbody this J i is simply Ji is simply
going to be equal to Ebi and Jj is going to be equal to E bj. So, which tells us that the
radiosities are going to be; radiosities are going to be replaced by E bi and Ebj so, qi-j; that
means, rate at which radiation leaves i surface i and is intercepted by j, what is this going
to be? It is first of all the amount of energy which comes out is A i × Ji, because Ji is unit,
Ji has units of energy per unit area so, I multiply it with corresponding area and I get the
total amount of energy which comes out of it, comes out of surface i or these the area of
the object having area Ai, but not all of it is going to strike j the fraction which is going
to strike j directly is Fij.
qi  j  ( Ai J i ) Fij
648
So, this entire thing; obviously, gives me the rate at which radiation which leaves i is
intercepted by j. Similarly, you can write
q j i  ( Aj J j ) Fji
So, this would be the radiation at the rate at which radiation which leaves j and is
intercepted by i. So, this is the view factor of j to i and I think these are must be clear to
you now. So, the net exchange between 1 and 2, net rate of exchange between i and j, if I
qi j
call it as would be
qi j  qi  j  q j i  ( Ai J i ) Fij  ( Aj J j ) F ji
Now AiFij is equal to Aj, Fji which is nothing, but the reciprocity relation and we also
understand that Ji is simply equal to Ebi and Jj is simply equal to Ebj. Since this is a
blackbody and we also understand that E b for a blackbody is simply equal to σT 4, this is
Stefan-Boltzmann law. So, if Eb is equal to σT4 . So, when you put all of them together
what you are going to get is
qi j  Ai Fij (Ti 4  T j 4 )
So, if you look at this I have simply substituted Aj, Fji by Ai, Fj, Fij which I can write from
the reciprocity relation Ji and Jj I have taken to be Ebi and Ebj and I know that Eb is equal
to σT4. So, this is a very important relation which tells us what is the net rate at which the
ith surface, the surface i is going to receive energy from the net energy exchange
between surface i and surface j, when both of them are blackbodies.
So, if I apply this for an enclosure the energy which we i surface i is going to exchange is
going to be j equal to 1 to n. So, this is what; obviously, you can see this rule to be an
extension of this one. So, this is extremely important relation which we have done it for
the blackbody, but we will see how real surfaces can be handled in the remaining time I
will quickly solve a problem.
649
So, this is a cylindrical furnace whose length is 0.15 meter, diameter is 0 point this is
open to the surrounding temperature the curved area is 1, this is your A 1 then which has a
temperature 1350 ºC and A 2 which is this surface its temperature T 2 is 1650 ºC and this
is 300.
So; obviously, some energy is to be supplied to wall to maintain them at these two
temperatures and they are going to lose energy through this enclosure. Let us say I am
putting as hypothetical surface which is A3, I need to find this is what we have to find
out. So, hypothetical surface A3, this is the one that, we have obtained, the heat loss this
is clearly I am going to lose the some heat from 1 to 3 and some heat from 2 to 3. So, this
is the heat loss and then
q  A1F13 (T14  T34 )  A2 F23 (T2 4  T34 )
650
Now, from the figure that we have done in the last class, it is a case like this, where r 1
and r2 are same and you know what is the length, what is the distance between the them.
So, here r2 and r3 are the same and you know what is the distance between these 2 as
0.075 meter so, you should be able to obtain using the figure from figure for r j/L to be
equal 0.0375/0.15 and L/ri to be equal to 0.15/0.0375 you find out F23 to be equal to 0.06.
And F23+F21+F22=1, since it is a plane surface this is flat surface, this is equal to 0. So,
you can find out F21 to be equal to 0.94. So, F21 is known to you now, from reciprocity
relation A1F12 is equal to A2F21. So, therefore, F12 can be calculated after you put the
values of A1 and A2, this to be calculated as 0.118. And from symmetry from symmetry
F13 must be equal to F12. So, when you do all these so therefore, you know what is F13
from here F13 is F12, F12 is 0.118, F23, you have calculated from here F21 you have also
calculated.
So, therefore, the total q would simply be the area over here.
q= (π×0.075×0.15)×0.118×σ[(1623)4-(300)4]+π/4(0.075)2×0.06× σ[(1923)4-(300)4]
q= 1844 W
The value of sigma; obviously, you can you can see its 5.67 ×10-8 W/m2K4
651
So, when you do this the value of q turns out to be 1844 Watt. So, this is a simple
example which shows how much of heat is lost by a cylindrical furnace, but if you think
of the utility or application of this, this is very significant so, any cylindrical furnace let
us say this side is not properly insulated it is going to lose some energy. And at this high
temperature the energy the mode of exchange of energy is going to be radiation. So,
using the formula and we have assumed all the surfaces the assumption here is all
surfaces are treated as blackbody.
So, all surfaces are treated as blackbodies. So, I have in the first part of the class I have
found out what is the exchange of heat between two blackbodies at different temperature.
So, when you introduce the fraction the fraction energy; that means, a view factor
concept, this is what you have get we have obtained, then from the figure of view factors
calculate what is F23, and once you have F23using the summation relation, you can find
out what is F21, because that is going to be important not only F21 is going to be important
F12 is also going to be important F13 is also going to be important as well.
So, using reciprocity relation and the known value of F 21, I calculate what is F12 and from
symmetry I find out what is F 13. Therefore, my q, I have obtained all the parameters
required for finding out the total loss of energy A 1 and A2 from geometry F13 and F23, F13
and F23, we have obtained from these relations the last thing I need to mention stress
again and again that all temperatures to be used must be in Kelvin in radiative heat
exchange. So, you plug in the values and what you get is the total heat loss from the
furnace, through the opening, through the top opening which is exposed to atmosphere at
300 Kelvin.
So, this is a very simple case of an enclosure problem in which all surfaces are black and
therefore, you do not need to introduce the concept of radiosity, but always keep in mind
that radiosity is the correct potential; correct radiative potential true radiative potential of
a surface, because it takes into account the emissive power as well as the part of the
energy which is reflected from the surface itself.
So, one emission which is a volumetric phenomenon, depends on the temperature

inherent temperature of the object. And the second is a surface phenomenon which
depends on the radiative property reflectivity of the surface, together they form the
radiosity. Calculate what is the radiative heat exchange between two blackbodies at
different temperature? Invoke the view factor concept you get a compact relation.
So, to apply those relations into real problems, you need to know the view factor which
652
you can read from the graph for such enclosures. Once you have that view factor and if
once you have used the summation and the reciprocity relations, then you will see that
you could find all the view factors which are to be used to calculate the net heat
exchange from one or more of the surfaces to another. And that is precisely what we
have done in this specific problem, where all the enclosed areas are assumed surfaces are
assumed to be black bodies such that radiosity is equal to the emissive power.
What we would take up next class from the in the remaining two classes or three classes
is that, when we are dealing with real surface, where the surface potential or the radiosity
is different from the inherent emissive potential of the surface due to its temperature. So,
we will take the cases of non-black bodies to be more precise, we will take the cases
assuming the bodies are gray, such that epsilon is equal to the alpha this this relation the
Kirchhoff’s law, the most relaxed form of Kirchhoff’s law is valid.
And from there we try to form a network of potentials and resistances and try to see what
is the current; that means, what is the heat which flows between these nodes, each node
representing an area of an enclosed surface. So, those are very interesting calculations
and you would see lot of similarity with electrical technology, the way these radiation
problems for an enclosure are treated which we will do in we will, which we will start
from the next class.
653
Heat Transfer
Lecture - 56
Network Method for Radiation Exchange in an Enclosure
We already know what radiosity is and what is the connection between the radiosity of a
surface and that of the emissive power of a black body. Now, this radiosity is going to
play a major role when we are going to do calculations about radiative heat exchange in
an enclosure. This is the practical application of whatever we have studied so far.
For example, if you are designing a furnace and you would like to maintain a surface
exposed to a constant temperature or maybe a constant heat flux, so what is the power
that you need to provide through the other surfaces which are forming the enclosure, so
that the temperature of the test surface, temperature of the surface of interest can be
maintained let us say at a constant temperature.
So, whatever we are going to going to study in the next few classes would tell us about
the radiative heat exchange in an enclosure, how much power is required for such cases
and we have to also think of the situation in which the walls of the enclosure are going to
interact with each other in terms of radiative heat exchange. Not only they are going to
emit some radiation, they are also going to absorb some incident radiation, which are
falling on it. So, this kind of emission as well as absorption, a fraction of the radiation
incident on the surface is going to be absorbed because of its inherent temperature its
going to emit.
So, the balance between these two whatever is being emitted and whatever is being
incident or rather the fraction of the incident energy which is absorbed by the surface, a
heat balance between the two should give us what is the amount of heat to be supplied to
that surface to either maintain it at a constant temperature. So, the enclosure therefore
takes into account the interaction between the surfaces which are forming the enclosure.
So, you can clearly see, imagine the role the view factor is going to play in these
calculations. So, we need to be concerned about what is the view factor of one surface to
the other surfaces which are forming the enclosure because the
654
radiative energy which is emitted by the surface is going to be intercepted by the other
surfaces and the view factor F12F13 and so on, they will play a vital role in it.
Therefore, these are all the real surfaces and they may not be black bodies. So, if they are
real surfaces, what is the relation between the black body emissive power and the
radiosity, if I consider that we are already aware of. So what is going to be the emissive
power of a real surface or real object to that of the radiosity of the surface. This is some
relation that we are need to find out. So, may be we can think of it as if there is maybe a
resistance which is going to be in between the black body emissive power corresponding
to the temperature of the substrate and the radiosity of the substrate.
So, these are the two potentials; the potential of the body in terms of its temperature had
this been a black body and the radiosity, the actual heat flux which is being observed by
someone standing just outside of the surface. What is the relation between the two
because that is the amount of heat which is going to flow from one surface to just outside
of the surface. So, what is the resistance that connects the two? We will discuss that in
this class itself. So, there has been, it has been proposed that following a network method
like the same type of network method which you have done in electrical circuits
something like that can be adopted for the case of radiative heat exchange between the
surfaces in an enclosure.
But in order to do that we need to make certain assumptions. Some of these assumptions
are obvious and they may not introduce much error to it, much error to the final results,
but there are two assumptions which are slightly restrictive in nature. So, we have to be
careful about what is the source of inaccuracy in our final calculation and what is the
contribution of each of these assumptions to the final answer.
So, we start in this class with the calculation procedure for radiative exchange between
surfaces forming an enclosure and these are the following assumptions we shall go
through one by one.
655
So, the first is we are going to start the network method for the radiation exchange in an
enclosure. We would assume that the radiative properties are uniform. That means they
are not a function of wavelength or of θ. So, this is something which is followed by many
of the engineering materials. It is a diffuse emitter of the surface. The surfaces forming
the enclosure are diffuse emitters and diffuse reflectors.
This is also something which is followed in many realistic situations. Radiative heat flux
leaving a surface is uniform. So, if you have the surface, then the radiation that is being
emitted by the surface is irrespective of whether you are measuring it here or anywhere
else, so its radiation is uniform over a surface. Secondly, if you have a surface and you
have incident radiation falling on it, then the radiation is going, we are assuming that the
radiation is uniform over the entire surface.
So, that radiation leaving the surface and the radiation which is incident on the surface,
both are same. Both remain constant at all points on the surface. These two are the
problematic cases. They may or may not be valid in all cases and therefore, these two are
the source of errors, but in order to simplify our calculation, there is that we need to
assume these two, but we have to understand that these two may not be followed in some
situations. The other assumption which we need to make is it is an opaque surface. So,
therefore α + ρ = 1. That means, the reflectivity and absorptivity is equal to 1 and the
transmittivity is equal to 0.
656
We would also assume that it is either at uniform temperature or an uniform heat flux is
provided over the surface. So, we can either say that the temperature of the surface is
known or the heat flux to be added to the surface or extracted from the surface is known
to us. The last assumption is that the enclosure which is let us say the enclosure which is
formed by this, the inside of this enclosure is filled with a non-participating medium.
This is important because as we would see towards the end of this course is that there are
certain, most of the gases absorb and scatter some of the radiation which is passing
through it.
So, when that happens depending on what is the thickness of the gas layer through which
this radiation is propagating. The amount of radiation which you would get at the end of
the gas film is obviously going to be different from the radiation which enters the gas
film. So, these are known as non-participating medium. There are specific laws which
would tell you what is the decrease in radiation intensity as you move along in the gas.
So, we will take, we will consider that later, but for the time being for network exchange
method what we are assuming is that the enclosure is filled with a non- participating
medium. So, radiation coming out of one surface is going to reach the other surface
without any loss due to absorption in the part in the gases or air which is present inside
the surface.
So, all these are some of the assumptions which are made to calculate, but using
radiation using the network method what is the radiation exchange between surfaces, but
before we proceed to that as I said before we need to find out what is the relation
between the black body emissive power, let us say the temperature of the surface is T.
So, the corresponding black body emissive power would be σT 4, however it is not a
blackbody.
The surface is a real surface; it is not a blackbody. So, therefore the emissive power of
the surface had this been a black body and the radiosity which is the radiative flux that is
perceived just outside of the surface, what is the relation between the two? So, if these
two are different and these two obviously are different since the surface is not a black
body. So, the difference in the potential between these two points must be connected by
some sort of a resistance and this resistance would account for the drop in potential
between the surface which is not a black body and the
657
radiosity of the surface. So, our next job is to find out what is the relation between
radiosity and emissive power of a surface. So, let us try to find that out.
(Refer Slide Time: 11:1
So, the concept that we are going to introduce here is known as the surface resistance to
radiation and you can imagine I think you can follow that this surface resistance would
obviously be 0 for the case of the imaginary black body, but we would come to that later.
Think of the surface and the properties of these are this it is A i being the area, Ti being
the temperature, ϵi being the emissivity and ρi being the reflectivity and as I have said
that τ=0, so, it is an opaque surface. This is my zone i, and let us assume this is the thin
layer at which we are going to measure the radiosity.
So, the flux which is coming out of it is the radiosity and there is some amount of
incident radiation which is given as Gi. So, Gi in this case is the irradiation at zone i
which is nothing, but the radiation flux which is incident on the surface and the surface
has area equal to Ai and the unit of this will be in W/m 2. Ji is the radiosity at zone i. So,
this is nothing, but the radiation flux which is leaving surface Ai unit also would be in
watt per meter square.
Now, the amount of heat, amount of radiative energy which is leaving the surface is J i in
W/m2, the amount of incident energy which is falling on it is Gi. So, therefore, the net
amount of heat if I call qi as the rate at which the energy would have to be transferred to
the surface Ai to maintain it at a constant temperature of Ti. So, these
658
are the three things as you can see from here. Gi is the irradiation at zone i, Ji is the
radiosity at zone i. So, it is a net, it is a radiative flux leaving the surface and q i is the rate
at which energy would have to be transferred to the surface because something is
coming, something is leaving. So, in order to maintain a thermal equilibrium, some
amount of energy will have to be added to this or extract depending on which one is
higher.
But if I say that qi is the rate at which energy would have to be transferred to the surface
or from the surface, it all depends on the values of Gi and Ji, then qi is the rate at which
energy would have to be transferred to the surface A i to maintain it at a constant
temperature Ti. Then, you can clearly see this qi must be equal to
qi  J i  Gi in W/m2. So, this J is the radiative energy which is leaving the surface as
i
observed by an observer just outside of the surface and qi is essentially based on the net
energy balance in the interior of the surface.
So, I guess this is a fairly straightforward explanation of the radiosity, the irradiation and
the net amount of heat which is to be added or extracted from the surface to maintain it at
a constant temperature. So, qi, the relation between qi the net energy, Ji the radiosity
which is the radiation flux observed just outside of the surface and G i which is the
radiative heat flux incident on the surface. The balance between the two would give you
the relation which I have just written as
qi  J i  Gi
So, let us expand this a little further and see what we would get out of this.
659
First of all this Ji is radiation which is emitted by the surface. The definition of J i by the
surface, this is the first contribution plus the second contribution is what is the radiation
which is reflected by the surface. So, this is the definition of radiosity which is two parts;
radiation emitted by the surface and the radiation reflected by the surface. So, Ji the
radiosity for a real surface which is not a black body must be
J i i Ebi  i Gi
where ϵ is the emissivity, Ebi which is the black body emissive power corresponding to
the temperature of the surface.
So, had this been a black body, then the radiosity and Ji and Ebi could be same, but we
understand that the real surfaces will have an ϵ. So, therefore, these two may not be equal
plus some amount of energy is incident on it, incident on the surface a fraction of it
denoted by the reflectivity is going to be reflected out of this. So, this is the emission
which is coming out and this is part of the reflection which is coming which is being
reflected by the surface Ji is a sum of these two. So, the amount of emission which is
coming out is i Ebi , the amount of reflection which you would get is i Gi .
So, if you are the person standing over here to you, the total amount of energy which is
coming out of this surface would be the sum of these two and we also understand that for
this specific surface i  i  1 , rather i   i  1 . Since your  i  0 is equal to 0, there

is no transmittivity, but it is also a grey surface.
Therefore, what you can write is i  i  1 . So, these two are equal absorptivity and
660
emissivity are equal. Therefore, the relation
J i i Ebi  i Gi
can be written as because my aim to get rid of this G i from this relation
and try to see if I can connect Ji with Ebi.
So, in order to do that first I find out what is Gi which is from this equation is
J i  i Ebi
Gi 
i
So, this can be therefore G i can still be written as putting the value of ρ i from here as
1 i . So,
J i  i Ebi
Gi 
1 i
So, if you now see q which we have obtained as
qi  J i  Gi
this is the fundamental balance equation which we have written I am now going to get
rid of this Gi. So, my total heat, the heat which gets transferred, heat which has to be
applied is
J i  i Ebi
qi  J i 
1 i
So, the two relations which are we are using is this one and the other is this one and this
one, the aim of this is using these 3 equations. We are trying to get rid of Gi and express
the q in terms of Ji and Ebi. So, we would continue with this, this relation which I am
going to write once again.
I get 661
J i  i Ebi
qi  J i 
1 i
So, when you do this, you are you are simply going to get
i
qi  [ Ebi  J i ]
1 i
So, this is the relation which would give you the amount of heat that needs to be
transferred, net heat that needs to be transferred to a surface which is having irradiation
as well as
662
emission. So, therefore this is something which one has to find out and if you also look
carefully, this Ebi and Ji can be treated as something of a potential. Let us expand this little
further and Qi, the total amount of heat would be
Ai i
Qi  qi Ai  [ Ebi  J i ]
1 i
Ebi  J i Ebi  J i
Qi  
1 i Ri
So, Ai i
1 i
Ri 
where Ai i
So, this is extremely important what we have obtained here. What we see in this is I have a
potential which is the black body. The potential of the surface potential of the substrate if it
was a black body, Ji is the radiosity of the surface and the radiosity of a real surface is
different from that of a black body. So, it is the potential that an observer standing just
outside of the surface would have seen, would have sensed had this been a black body,
these two would be equal. This is also evident from the definition of R i, the resistance
where ϵi for a black body would be equal to 0 and therefore, this resistance would vanish
and therefore, in for a black body Ebi would be equal to Ji for a black body ok, but for a real
surface, this is some sort of looks like an Ohm's law.
The Ohm's law is basically potential difference divided by a resistance which gives you a
current. So, in this case, the potential differences are the black body radiative potentials,
potential minus the radiosity divided by some sort of a resistance which is equal to the
current and the current in this case is equivalent to that of heat flow. So, this R i is known as
the surface resistance to radiation, ok.
So, we started with the expression that J i i Ebi  i Gi that was my starting point. So,
what would happen for a black body let us see once again. My ϵ i would be equal to 1 and
my reflectivity because this since it is a black body, it is going to absorb everything and
therefore, ρi would be 0 and therefore, Ji would be equal to Ebi. So, this is for a black body,
this is what you would get. So, for a black body there is no resistance, no surface resistance
to radiation, but for any real object the potential inside the body, the emissive potential
inside the body had this been a black body and the potential measured just outside of the
surface, the difference between these two potentials is the cause which makes the heat flow
from one direction to the other.
663
So, we may lose some heat from the surface which has to be again supplied to main from
outside to maintain it at a constant temperature or the surface may gain some heat because
of this exchange process and again in order to maintain the constant temperature, that
amount of heat will have to be extracted from the surface. So, the concept of surface
resistance to radiation is extremely important as it would connect the black body radiation
power of a surface to that of a real surface. So, what is the surface resistance to radiation
which takes into account not only the emissive power, emissive nature of the surface, but
also takes into account the reflectivity.
So, if a surface forms one part of the enclosure for radiative heat exchange, it is not only
going to emit, it is also going to get some irradiation from other surfaces. So, the algebraic
sum of the emission and that of absorption due to radiation coming from another object.
The algebraic sum would tell you whether heat has to be extracted or heat is to be supplied
to maintain the surface, to maintain the surface at a constant temperature.
The concept which relates these two potentials, one inside the object and the other just
outside the surface of the object, the emissive power of a black body at the same
temperature as that of the surface and the radiosity of the surface. These two are the
potentials. The difference in these two which was non-existent for the case of a black body,
so, the difference between these two potentials causes heat to move either from the surface
or to the surface.
So, in order to calculate these two resistances, I need to know these two potentials and
these two potentials taking an analogy from Ohm's law must be connected by some sort of
a resistance and that resistance which we have evaluated that resistance R i is called the
surface resistance to radiation. So, if I knew the two potentials in here and if I knew the
resistance in there and as you could see the resistance depends only on emissivity, it is
1 i
Ai i which depends only on the emissivity. So, if the emissivity of the surface is known,
then the surface resistance to radiation will also be known.
So, if the two potentials are known, then the total amount of heat flow to or from the
surface that is required to maintain the temperature of the surface at a constant value can
664
easily be found out and while doing this, I am allowing the exchange of radiative energy
with from which are coming from other surfaces forming the enclosure.
So, this is the first tip to realize to understand how the radiosity and the black body
emissive power are connected by the concept of resistance, by this resistance the very
surface resistance to radiation. So, when we move in the next class to the network method,
it would be really easy to draw this circuit with the blackbody radiative power and the
radiosity as two nodes, two potentials connected by a resistance Ri. We will take this up in
the next class.
665
Heat Transfer
Lecture – 57
Network Method – Two and Three Zone Enclosure
In the previous class, what we have seen is that there exists something called the surface
resistance radiation which connects between the radiative potential of the surface had this
been a blackbody and the radiosity of the surface which is the radiative flux just which is
can be observed just outside of the surface. Based on these concepts we are going to start
our first network method for a very simple case which is a two-zone enclosure.
So, you have two surfaces which are forming an enclosure and we would like to find out
what is the net heat exchange between these two surfaces which are forming the
enclosure.
So, let us look at how this would look like. So, it is a two-zone enclosure and this is my
zone – 1, where the area is A1 it is at a temperature T1 and the value of ϵ is ϵ 1. The rest is
zone 2 which again has area A2 the temperature T2 and the emissivity is ϵ2. So, zone – 1
and 2 are maintained at constant temperatures and we would like to find out what is
the net radiative heat exchange
666
between the two. So, let us call the net head radiative heat exchanged between 1 and 2 as
Q1-2 which is the net radiation heat transfer from zone 1 to 2.
So, the amount of heat transfer from 1 to 2 is called as Q 1-2, since we make an energy
balance. So, from an energy balance from an energy balance I can write that
Q1-2 = (Radiation energy which is leaving A1 that strikes A2) – (Radiation energy living
A2 that strikes A1)
So, it is the radiative energy which is leaving A 1 that strikes A2 and the other one is this
is also the 2 is also emitting energy and a fraction of that energy may strike zone 1.
So, Q1-2 which is the net radiative heat transfer from 1 to 2 must be equal to this. So, let
us try to express this in terms of known quantities. So, what is the radiation energy
leaving A1? This is the radiosity; because radiosity is the potential of zone one which is
seen by an observer who is standing just outside of A 1. So, therefore, the surface 1 ,has
potential which the radiation energy which is coming out of surface one must be equal to
J1 now a fraction of that is going to strike the second surface. So, what is the fraction of
that surface? It is going to be A 1F1-2 that is the view factor. So, A 1F1-2 multiplied by J1
would give you the radiation energy leaving A 1 that strikes A2.
Similarly, the radiation energy leaving A2 is J2 and what a fraction of that is going to
strike surface 1. So, this must be equal to A 2 F2-1, I will go through it once again. This is
the net radiation heat transfer from 1 to 2. So, 1 to 2 must be equal to whatever be the
radiation energy leaving A1 that strikes A2. The radiation energy which is leaving A1 is
this which is Watt per meter square times meter square. So, this is the total radiation
energy which is leaving surface 1. A fraction of that is going to strike 2, what would that
fraction be? From the definition of view factor, we know that F 12 which is from 1 to 2
would give us the amount of the fraction of energy emitted by 1 that strikes 2.
So, the amount of energy is J 1 A 1. So, the fraction that strikes 2 is (J1 A1 )F1-2. Similarly
the radiation energy which is leaving A2 is J2 A2. So, of this amount of energy this
fraction is going to strike 1. So, the radiative energy leaving A2 that strikes A1 would
simply be J2 A2 F2-1. So,
Q1-2 = (J1 A1)F1-2 – J2 A2 F2-1
we can simply used the reciprocity relation that A1F1-2 = A2 F2-1

So, using the reciprocity relation I can express it in this way.
So, therefore, this is 667

Q1-2 = A1F1-2 (J1– J2)
So,
1
R12 
A1 F1 2
this A1F1-2 is nothing, but the resistance. So, J 1 and J2 are the radiosities of surface 1 and
surface 2. These two are the potentials and from the fundamental equation we can see
1
that these 2 potentials are to be divided by A1 F1 2 to obtain what is the net heat flow
from 1 to 2. So, if this is the current, this is the potential difference then this denominator
1
must be equal to some sort of a resistance which is defined as A1 F1 2 and this is the
R12 .
resistance between 1 and 2 is
Q1 2 
 J1 – J2 
1 / A1 F12
So, when we think of the surface when we think of this surface. Once again this is my 1
and this is my 2 which in between I have R1-2, the potential at this point is J1 the potential
at this point is J2 and if I have assumed these to be blackbodies with potentials at E b1 and
Eb2 then the resistance in between these two is what we have what we have defined
before. So, if this is R1 and this is R2 then I can write this draw the circuit diagram E b1,
this is my J1, this is my J2 and this is Eb2. 668
They are connected by a resistance which is the
surface resistance to radiation for surface 1, this is R2 which is
669
the surface resistance to radiation for surface 2 and in between I have R 1-2 with and we
already know what is what is R1 and what is R2.
So, I can write this as
Eb1  Eb 2  T14   T2 4
Q1 2  
R1  R1 2  R2 1  1  1  1   2
A11 A1 F1 2 A2 2
So, the heat flow from 1 to 2 must be equal to the potential difference by the sum of
these three resistances which are in series.
So, this gives you the amount of heat which is to be supplied from this is from 1 to 2
which is Q1-2. This concept this is so, this is for the relation for a two-zone enclosure. The
concept of two zone enclosure which is straight forward can now be extended for more
realistic situations like three zone enclosure. So, what is the three-zone enclosure and
how do I express it in terms of the concepts that we have developed so far; so, instead of
two surfaces, as was done previously, now I have three surfaces. So, previously one
surface was interacting with only ones other surface in this case one surface for a three
zone enclosure one surface would be interacting with two more surfaces.
So, there will be three radiosities to deal with J 1, J2 and J 3 three blackbody radiation
power is to be is to be considered which is Eb1, Eb2, Eb3. Eb1 and J1, are connected Eb2and
J2 are connected Eb3 and J3 are connected by the relation (1 - ϵi)/Aiϵi that is the resistance
in between. But, what happens in between let us quickly take a look at it and then we
will be in a position to do all these furnace calculations and so on.
So, our next topic which we are going to consider now is how to formulate the circuit
diagram for a three-zone enclosure. The enclosure is being formed by three zones.
670
So, let us see what three zone enclosure look like and in three zone enclosure the Q i-j
would be
Ji  J j
Qi  j 
Ri  j
1 1
Ri  j  
Ai Fi  j Aj Fj i
Again,
Ebi  J i
Qi 
Ri
1 i
Ri 
Ai i
So, these are the relations which we are going to use for this case.
So, let us think of a three-zone enclosure like this where this is A 1 maintained at a
temperature T1 with an emissivity equal to ϵ1, this is A2 maintained at a temperature T2 ϵ2
and A3, T3, ϵ3. I would like to find out what is the net radiative heat exchange between
the two.
So, the first thing that I would do is for 1, my potential is going to be Eb1 which is inside
the surface, but it is a blackbody radiative potential. So, it is a real surface, it is not a
blackbody. So, its potential over here is going to be J 1 here it is going to be J 3 and here it
is going to be J2. So, this is going to be my J 1. On the other hand over here, I have this as
Eb2 which is this surface it is connected to J2 in here I have Eb3 with the radiative potential
being J3. 671
So, these are the three points or the six points that I have done for each of the surface the
potential if it is a blackbody the radiosity. Since it is not a blackbody it is the radiative
amount of radiation energy which is coming out of the surface as observed by an
observer standing outside of the surface. Similarly, Eb2 J2 Eb3 and J3; these are connected
by radiation this resistance which is this. So, what this resistance should be? This
1  1
R1 
A11
This one and these two are
1 2
R2 
A2 2 and
1 3
R3 
A3 3
So, that is straight forward from our previous discussion.
Now, this is now going to interact with 2 as well as it is going to interact with 1. These
two are this is the potential of 1 which is going to interact with 2 and it will also interact
with 3. So, let us say it is interacting with 3 and of course, there is a resistance in here
and J1 is interacting with J2. So, there would be resistance in here as well. So, what is the
1
formula for interaction between J1 and J2? It is R1-2. So, it should be A1 F1 2 .
1
So, the resistance R1-2 would simply be equal to A1 F1 2 . You can see clearly that it can
also be equal to A2 F21 , because of the reciprocity relation. So, whether it is R1-2 is
1 1
expressed as A1 F1 2 or A2 F21 they mean the same thing, right. F1-2 is the view factor of 2
from 1 in this case this resistance is going to be R 1-3 which would be only be equal to
1
A1 F13 and this can as I said this can also be written as A3 F31 .
Now, one thing that is remaining in this to complete this circuit is between J 3 and J2. So,
what is in between J3 and J2? There must be some resistances in here as well whereas,
1
where this R2-3 can be expressed as A2 F2672
3 . So, the connection
1
between J2 and J3 the resistance R1-2, R1-3, R 2-3 would simply be equal to A2 F23 or it can
1
also be written as A3 F3 2 using the reciprocity relation.
So, this essentially takes into account all the entire circuit. And because of the difference
in values between Eb1 and J1 Eb3 and J3 Eb2 and J2 some amount of heat is entering over
here which is which I call it as Q 1 some is entering which we call it as Q 2 and some is
entering over here which is Q3. Now, the potential difference between Eb1
673
and J1 arises due to the nonblack nature of the body. Had this been a blackbody E b1
would simply be equal to J 1. And therefore, Q1 would be equal to 0.
In our in the most general case we do not assume that the body from where radiation is
coming or the radiation is taking place is not a blackbody. So, there is a difference
between the blackbody emissive power and the radiosity. So, a current that is an
equivalent heat must flow from Eb1 or J1 Eb1 to J1 or J1 to Eb1. So, Q1 can be positive or
negative depending on whether J 1 is less than E b1 or J1 is more than Eb1. Another point
which I would like you to remember is that the value of J 1 is affected by the presence of
other surfaces emitting radiation. So, whatever is the value of J 1 is going to be dependent
is going to be going to be dependent on whatever radiation it receives from the other
surfaces. So, 1 2 and 3 these three surfaces are therefore connected any change in the
value of Q3 or Q2 would affect the value of J 1 as well.
So, this is the network method for three zone enclosures and you would be able to find
out what is the current at each of these nodes, what is the amount of heat that is to be
supplied or to be extracted from node 1. The node 1 is surface 1 from this surface to
maintain its temperature at a constant value. So, that is extremely important, and we will
solve a number of problems on these. But, let us think of a surface which is perfectly
insulated. So, this if one of the surfaces is perfectly insulated that is known as the re-
radiating surface.
674
And, a simplification can be made for the case of a reradiating surface. So, in many in
many engineering applications a zone can be thermally insulated and in such a case if in
such a case the net radiative heat flux in that particular zone is 0 and such a surface is
called a reradiating surface.
So, let us say this is a reradiating surface and in a reradiating surface obviously, the value
of Q3 is equal to 0, whatever it gets it radiates the same amount. So, we have A 1, T1, ϵ1,
A2, T2, ϵ2, but this surface 3. So, this is 1, 2 and 3 surface 3 is a reradiating surface. So,
how would the how would the circuit look like in that case? You have E b1 you have J1
this is J2 this is Eb2 over here. However, J3 is equal to Eb3 because it is a reradiating
surface. So, if it is a reradiating surface then there is no flow of heat through this surface
and therefore, it in such a case J3 is equal to Eb3.
1  1
Let us complete the circuit over here. So, I have resistance which is A11 , this is R 2 is
1 2 1 1
A2 2 . So, in this case R should be equal to A1 F13 , this is R would be A2 F23 and
1-3 2-3
1
R 1-2 would be A1 F1 2 . So, for this reradiating surface since Q is 0 there is no heat that is
to be extracted or to be supplied in for a reradiating surface. So, it is perfectly insulated
let us say in which case since Q is equal to 0, E b3 is equal to J3. So, for a perfectly
insulated perfectly insulated surface is a closed approximation to a reradiating surface
reradiating surface and Q would be equal 0. So, Eb3 is equal to J3.
So, if I now consider between Eb1 and Eb2, if I am trying to find out what is the value of
the heat to be supplied to Q 1 the same heat in order to maintain thermal equilibrium the
same heat will have to be extracted from 2. So, if this is Q 1 and you have Q2 in here, so,
Q1 must be equal to Q2, and this is simply going to be
Eb1  Eb 2
Q1  Q2 
R
So, what is the effective resistance between these two points? This resistance R 1 and R2
are going to be in series and the equivalent resistance of these three where these two are
in series and which this resistance in parallel with the sum of these two. The equivalent
resistance between Eb1 and Eb2 can be substituted and this would simply be where the
effective resistance R would simply be equal to R1 plus R2.
675
So,
1
 1 1 
R  R1      R2
R
 12 R13  R23 
When you take it further this R would simply be

1
 
1  1  1 
R   A1 F1 2   1 2
A11  1 1 

 A1 F13 A2 F23 
So, we have covered a number of interesting important observations in this class starting
with the concept of radiosity and the concept of surface resistance to radiation. We went
ahead and to we found out what is the resistance to heat transfer between two surfaces
1 1
AF
which we found to be equal to A1 F1 2 or i i  j . Through the use of this surface
resistance to radiation and the resistance to radiation between surfaces which are which
are functions of the view factor I could create a circuit the same way we have done it in
electrical technology so to say.
So, a three-zone enclosure forming by three surfaces would have 6 resistances, three
1  1 1   2
resistances each for each of the surfaces. So, A11 , A2 2 and so on would form the
surface resistance to radiation for these three surfaces. Each of these surfaces will
interact with the other two surfaces and the resistance in that case would simply be equal
1 1
AF
to A1 F1 2 or i i  j . So, the circuit which we have developed then on one side I have Eb1
surface resistance to resistance the surface resistance to radiation J 1 Eb2 surface resistance
to radiation J2 similarly for Eb3.
So, I have the three nodes of a triangle where the potentials are J 1, J2 and J3. J1, J2, J3 are
1 1 1
connected by A1 F1 2 , A2 F23 and A1 F13 that completes the circuit. So, I should be able
to find out what is the net heat which is entering 1, which is entering through 2 and
which is entering through 3. In order to maintain the thermal equilibrium, the algebraic
sum of Q1, Q2 and Q3 must be equal to 0. We took it one step further we define what is
676
the reradiating surface where it emits whatever it gets. So, it is a perfect insulated surface
is an example of a reradiating surface.
And, I have shown you how what a reradiating surface a three-zone enclosure circuit
diagram can be drawn, where there is going to be two surface resistances in series and
the third resistance which when you convert it gives a simple expression for the net heat
flow between these two surfaces. So, we are going to do number of problems on this to
677
clarify any doubts that you may have, and these are very interesting applications; real
applications in furnace calculations.
So, we will take those up in the next two classes.
678
Heat Transfer
Lecture - 58
Tutorial Problem on Radiation Exchange using the Network Method
We have introduced the concept of network method for Radiation Exchange in an

enclosure. And there we have seen that there are two resistances to deal with when we
are considering the radiative Heat Transfer from a surface. One is the inherent resistance
to radiation of the subject of the surface itself, which gives us the difference between the
radiosity of the surface and the emissive potential had this surface been a blackbody.
So, the emissive power of the surface if it were a black body, and the radiosity which is
the radiation flux which would be observed by a person who is standing just outside of
the surface; these two are connected by the emissivity of the surface itself. And after that
from that read from the radiosity, the energy is going to be exchanged with the
surrounding surfaces, with the surfaces which other surfaces which form the enclosure.
And there the fraction of energy emitted by the surface which is intercepted by the other
surfaces, namely the concept of view factor came into play.
So, we had 2 resistances to consider, one is the surface resistance to radiative emission
1 i
and the second is the exchange between two surfaces. So, one was Ai i , where ϵ1 is the
emissivity of surface 1 and the other one the A1 is the area of the surface. And in between
two surfaces it is going to be 1/A1 F12, which is the resistance which gives you the
resistance of radiative heat transferred between two surfaces, where the view factor for
the surface 2 from 1 is denoted by F12.
And we can use the summation rule and the reciprocity rule to obtain the unknown
values of the view factors for these enclosures. So, with that then we have proceeded to
write, to draw, what would be the circuit diagram for 3 surfaces forming an enclosure. 2
surfaces forming an enclosure is straightforward, but the 3 surface one we have to think
of 3 potentials at 3 ends, which are the black body, which are the emissive potential of
these 3 surfaces if they were black bodies, then the radiosity at the 3 apex of the triangle
which forms the circuit.
679
So, the radiosity of surface 1 that of 2 and of 3 and then these radiosities are connected
through the radiation exchange resistance between 1 and 2, which is 1/A1 F12,
between 1 and 3, which is 1/A 1 F13 and between 2 and 3, which is 1/A 2 F23. So, that
completes the circuit and one should be able to obtain the heat flow rate into this
enclosure by the standard techniques which I have discussed in the previous class.
In some cases we have something called a reradiating surface. So, whatever it gets, it
radiates back the same amount, which would be true at steady state for the; if a surface is
perfectly insulated. So, if a surface is, if the back side of the surface is perfectly insulated
then, whatever it gets by the by the means of radiation from other surfaces it must radiate
it back in order to maintain the constant temperature of the surface. Such a surface is
known as the reradiating surface.
For such a surface, the surface emitting emissivity does not play any role. That means,
the resistance between the black body emissive potential of the surface, if it is a black
body and the radiosity the these two are equal and therefore, E b3 if the surface 3 is the
reradiating surface, then Eb3 would be equal to j3.
So, the radiosity and the black body emissive potential for a radiating surface are equal
moreover, if you can calculate what is j3, which is equal to Eb3, then through the use of
Stefan Boltzmann equation that is Eb3 is equal to σT34, you can obtain what is the value of
the unknown temperature of the surface.
So, for a reradiating surface, ϵ is equal to 1. There is no resistance, no surface resistance

to radiation and the unknown temperature of the surface can be obtained simply through
the use of Stefan Boltzmann equation and the substantial simplification of the entire
assembly can be obtained. So, I think we have covered this much and if we now solve a
problem then it would be even more clear to all of us.
680
So, the problem that I have chosen is this for solving the network method radiation
exchange. We have a paint baking oven, which consists of a long triangular duct. So, it
looks something like this. It is an equilateral triangle. So, these three sides are equal and
it is extended in this direction. So, this is a paint baking oven and this side one of the side
a heated surface is maintained at 1200 Kelvin.
So, let us call this as surface 1. So, T this is 1; so T1 is 1200 Kelvin and another surface is
insulated. So, this surface is insulated, and let us call this as; and obviously if it is
insulated as per our previous discussion, this has to be a reradiating surface. So, this is
the reradiating surface and the ϵ value for this is provided to be equal to 0.8.
The value of epsilon for the other surface which is at 1200 Kelvin is also mentioned at
0.8. So, ϵ is equal to.08 for both the walls and the paint baking, the paint panels, the
panels are kept like this. So, this is my other surface, this is the surface 2 which is the
painted panels, are kept in here. And the ϵ2 for this case is 0.4 and the temperature that
needs to be maintained for proper baking of the paint T 2 must be kept at a 500 Kelvin.
So, as you can clearly see in order to maintain the temperature over here at 500 Kelvin
also temperature over here at 1200 Kelvin, some heat must be supplied to the to the
surface 1. So, this is an equilateral paint baking paint baking oven, where the painted
panels are kept at the bottom that I call as surface 2, the value of ϵ2 T2 are known.
681
This is the heated surface which is 1 surface 1 at 1200 degree, ϵ is 0.8. This is perfectly
insulated. So, as per our understanding, this is a reradiating surface. So, whatever it gets,
it must emit the same thing back. And therefore, the formula or the concept of radiating
surface is applicable here.
So, what we need to know, the problem asks us to find out what is the rate at which heat
must be supplied to surface 1 to maintain its temperature at 1200 degree Kelvin. So, that
is the first part of the problem. So, what is the amount of heat that needs to be supplied to
the surface at 1200 degree Kelvin to maintain its temperature at that value. So, that is the
first part of the problem.
The second part of the problem tells us to find out, what is going to be the temperature of
the insulated surface; that means, what is going to be TR. So, therefore, I need to find out
what is q’, which is the heat and the rate of energy to be supplied to the heated site per
unit length. So, this must be Q by L, where Q is the total energy supplied and L is the
length of it. So, amount of heat to be supplied per unit length, let us call it as q’ and we
would also like to find out what is going to be the value of TR.
So, the first thing that we need to do is convert this to the resistance diagram which we
have shown before.
682
So, once again I will quickly draw this oven, where this is my surface 1. This is surface
2, which is the painted panel and this is the reradiating surface. The temperature here is
given as 1200, the ϵ is 0.8, here ϵ is 0.4 and T 2 is 500 Kelvin, let us call the ϵ R over here
is also 0.8, I do not know what is the value of TR in this.
So, let us convert this to the circuit diagram that we that we generally use. So, over here,
this is the reradiating surface. Since it is a reradiating surface, JR must be equal to EBR
that is the definition of reradiating surface. And since it is insulated, q R would be equal to
0, so we start from here and we realize that, these 2 relations are going to be valid. Then
this is going to exchange radiation with 1, this is also going to exchange radiation with
surface 2.
So, I have 2 resistances to consider here ok. So, let us call it as 1 and this is my 2, which
is the painted panel. So, as per our previous discussion, this is simply going to be 1/A1F1R
and between 1 and 2, I also have radiation resistance. So, this is going to be 1/A 1F12
between 1 and 2, and this 1 is going to be 1/A 2F2R. So, these are the three resistances
inside the enclosure, but both 1 and 2 are not reradiating surfaces. So, they must have
surface resistance to radiation.
So, this is going to be E b2. The emissive potential of surface 2 had this been a blackbody
at the same temperature and this the resistance over here, the surface resistance to
radiation is simply going to be 1/ϵ 2A2, in a similar fashion this is going to be E b1 and the
1 i
resistance over here is going to be Ai i .
So, that completes the circuit. And let us assume, let some amount of heat is entering
through this which is q2, the amount of heat which is to be supplied or extracted. I do not
know at this point what whether it is extracted or it is going to be supplied. But if you
look at the 3, will look at the figure then obviously, 2 is going to receive some amount of
energy from one. Since the temperature of 1 is significantly more than that of 2. So,
obviously heat is to be extracted, but that then the way we have done it, if we are going
to calculate the value of q2, this should turn out to be negative such that heat is going to
be taken out.
So, we are doing it in this way, it still does not matter because, we will get a negative
value of q2. And let us say the amount of heat which is to be supplied at 1 is q1. So, at
683
any point of time in order to maintain steady state, the algebraic sum of these 2 would be
0 and we understand that qR is therefore 0.
So, the system that we have over here, the system can be modeled as a 3 zone enclosure
ok. So, the one that is to be supplied over here we said q 1, the heat that is supplied to
surface 1 q1 would simply be this potential, this is a potential difference. I have these 2
resistances in series. That means this and this resistance in series and these two are going
to be in series with the equivalent of this resistance.
So, when we try to find out the equivalent of this resistance, so these two are in series,
which in turn are parallel to this one. So, this has to be replaced by the equivalent
resistance of this circuit, which is going to be in series with these two. So, if you do a
little bit of, if you do it on your own, what you would see that the heat flow which is
similar to that of current flow is going to be
Eb1  Eb2
q1 
1 1 1 1 2
 1

A1 1  1 1  A2 2
A1 F12    
 A1 F1R A2 F2 R 
So, that is the standard formula. Then you are going to use ϵ 1 to be equal to 0.8 and this
we would also be able to and ϵ 2 is equal to 0.4. And from symmetry, one can write that
F12 is equal to F1R is equal to F2R is 0.5. So, that is what, so whatever heat since it is an
equilateral triangle, whatever heat comes out of this is half of it the view factor for this
one is going to be 0.5 and the view factor is going to be 0.5 over here as well, the view
factor to itself since it is a plain surface is 0.
So therefore, F11 is equal to F22 is equal to FRR, these are 0, since these are plane surfaces.
So, by symmetry F12 and F1R and F2R all are going to be equal to 0.5 and we also can say
that A1 is equal to A2 is equal to W × L where, W is the width and the width of this the
W, the width of this is equal to 1 meter ok. So, W the width of this is equal to 1 meter.
So, this is 1 meter, this one is 1 meter as well as this one. So, this is W times L, where L
is equal to the duct length. And we have to find out what is q1/L? This is the one which
we have to find out.
684
So, from this expression which is straightforward. We can proceed to obtain that q1’,
q1 5.64  108W / m 2 K 4 (12004  5004 ) K 4
q1 '  
L 1  0.8  1  1  0.4
 1 

0.8 1m  1m  0.5  (2  2)  0.4 1m
So, A1F12 and when I take L out of this it is going to be W times F 12 plus the equivalent
resistance of the middle 1, which is this. And again A 1 is simply equal to 1 meter. F 1R is
0.5. So, 1/0.5 is equal to 2. So, (2 + 2)-1, and that is what I have written over here.
685
kW
q1 '  37
m
q1 ' q2 '  0
kW
q2 '  37
m
And of course, if q1 in this figure to maintain steady state, if q1 is 37 kilowatt per meter,
that amount of heat is to be supplied at surface 1. And since this is insulated, the same
amount of heat is to be extracted out of the surface which is forming the panels or
surface 2.
q2’, the heat that is to be extracted would be equal to 37 kilowatt per meter with a minus
sign then the second part. So, what is the second part tells us how to find out what is the
value of the TR, the temperature of the reradiating surface. This has to be evaluated. Now
in order to do that, what you have to do is first of all you have to find out what is the
value of J1, the value of J2, then at this node, the total heat flow the sum of the algebraic
sum of the heat flow must be equal to 0.
This is the same way we have we have done it in electrical technology. So, if I could find
what is J1 and what is J2, then the heat flow from here to here must be equal to the heat
flow from here to here in the opposite direction. So, the heat flow from 1 to R would be
(J1-JR) divided by the resistance which is this. The heat flow from 2 to J R would be (J2-JR)
divided by this resistance.
So, these two must be equal and opposite in order to, in order to ensure that there is no
net heat flow at this point. So, the first step for the second part of the problem is to find
out the unknown values of J1 and J2. Now we know what is the heat flowing from here to
here, what is the value of the resistance, what is the value of E b1, so J1 can simply be
calculated and that is what I am going to do next. So, for this case at surfaces 1 and 2, J 1
would simply be equal to
1 1
J1  Eb1  .q1 '
1 W
So, this is the heat flow. So, the heat flow would be the potential difference. So, what is I
am saying is that
Eb1  J1
q1 ' 
1 1
GW
So, this is what we are going to do and when you put the value of Eb1, which is
5.67×10-8.
686
Putting the values
W 1  0.8 W
J1  5.67 108 2 4
(1200) 4  37000  108323W / m 2
mK 0.8  1m m
In a similar fashion, you should be able to find out
687
J2
1 2
J 2  Eb2  .q2 '  59043W / m 2
2 W
So, we have now evaluated what is the value of what is the value of J 1, what is the value
of J2. So, the unknown in this case is now, JR.
So, for that we are simply going to say that from energy balance at the reradiating
surface
J1  J R J  JR
 2 0
1/ A1 F1R 1/ A2 F2 R
The sum of these two in order to maintain steady state must be equal to 0.
So, when you when you expand we expand this what we are going to get is the value of
J1 - JR and this is
108323  J R J R  59043
 0
1 1
W  L  0.5 W  L  0.5
J R  83683W / m2  EBR   TR 4
TR  1102 K
And this is just an equation in JR you should be able to obtain the value of JR to be 83683
W/m2watt per meter square. And for a reradiating surface JR is equal to EBR.
And EBR = σTR4. It is a blackbody so obviously, there is no emissivity, emissivity is equal

to 1. The only unknown in this is TR, which should be equal to 1102 Kelvin. So, this
688
problem is a nice example of how to calculate the radiosity, how to calculate the overall
resistance for heat transfer between surfaces.
689
What to do if we have a what simplification can we make if we have a real reradiating
surface and if we can find out what is the J R radiosity of the reradiating surface which
would be equal to the black body emissive potential for the for the if for the reradiating
surface at the same temperature. So, if we have a blackbody at the same temperature as the
reradiating surface. Then we from the since the radiosity and emissive power are equal for
a reradiating surface. Therefore, EB3 in this case the EBR that we have done for in this case
would simply be equal to σTR4. So, the unknown temperature of the reradiating surface can
be quickly obtained following this method. So, this is one complete example tutorial
problem on network method for radiative heat exchange in an enclosure and in the next
class I would solve one more problem and that should and once you practice these
problems from your text, it should clarify any doubts that you have. Otherwise I encourage
you to contact me and the teaching assistants such that you have a clear idea of how to do,
how to analyze the exchange of radiation between surfaces in an enclosure.
690
Heat Transfer
Lecture - 59
Radiation Shields
In many practical applications you would see that you would like to prevent a surface
from receiving radiation, or a surface losing its energy through radiation. Let us take the
first point and you do not want a surface to receive energy by radiation. One of the
practical examples where you encountered this situation is for cryogenic storage of
liquids which have a very low boiling point.
Let us see you have liquefied air or you have liquefied nitrogen ammonia and so on. And
you are storing them in containers you try to insulate them as much as possible from the
outside, but in order to reduce the radiation which is coming from the ambient to the tank
which is stored liquid nitrogen. Sometimes it is advisable to put a shield around the
container.
So, if you have a spherical container for storing liquid nitrogen what you do is, these
spheres are never I mean the walls are not solid. You may have on the inner sphere which
holds the liquid nitrogen another layer of a material which surrounds the sphere. And this
outside sphere the outside covering of the main tank this significant, this significantly
reduce the radiative heat that comes to the sphere from the ambient and thereby heating
up the liquid nitrogen inside.
So, these kind of protective shields are known as radiation shields, so in order to choose
the material of construction for a radiation shield which is going to protect the inner core
which you would like to keep at a low temperature that is very important. We need to
know what should be the radiative property of the radiation shield.
So, radiation shield is something which hinders the flow of heat through it, hinders the
flow of radiative heat through this. Thereby protecting the cooler temperature, the cold
temperature, the cold storage inside and we are going to find out what is the property that
needs to be: what is the property of the material of this radiation shields. One property is
691
obvious; that means, it should be opaque; that means, the radiation there is not going to
be any transmission of incident radiation through the shield to the other side where you
would like to keep the temperature cold. So, the very first property of a radiation shield is
that it must be opaque in nature.
So, we take care of the surface property that it is to be opaque transmitivity should be 0.
But what about the emissive property does it have to have a high emissive emissivity or a
low emissivity, which one is going to be preferred for a radiation material. So, that is the
one which we are going to study now, what would be the emissive property of the
material that is to be used for radiation shield. And what kind of modification to the
equation we need to have in order to incorporate the presence of radiation shield in a heat
radiation problem.
So, let us look at the case in which you have a surface over here which is the plate, I call
it as plate 1 which is maintained at a temperature T1, its property is ϵ1, area is A. And
then you place another one which is the radiation shield and we I call it as surface 3. So,
the temperature is T3 on this side and T3 on the other side. Let us for generalization sake
assume that the emissivity of the material 3 radiations shield which is facing 1 is ϵ 31, and
the emissivity of 3 which is facing 2 is ϵ 32. And this is let us say the plate 2, where the
temperature over here is T2, the emissivity is ϵ2 and the area is A.
692
So, this is we are considering radiation heat transfer between two so, we are considering
radiative heat transfer between two very large opaque plates parallel to each other. And
we put a radiation shield in between so this is more or less what we have in this. So, the
heat flow from 1 to 2; therefore,
Eb1  Eb 2
Q1 2 
R1  R1 2  R2
Q1-2 must be equal to whatever be the potential of this and whatever be the potential of
plate 2 by the sum of all the resistances.
So, what are the resistances? One is R 1, the second one is R 1-2 and the third one is R2 ok.
So, if there is no radiation shield let us first consider this and then we will consider the
radiation shield in between 1 and 2. So, if there is no radiation shield in between 1 and 2
what you would get as
 T14   T2 4
Q1 2 
1  1 1 1 2
 
A11 A1 F12 A2 2
These three are going to be the resistances and for parallel plates A1 would be equal to A2
let us say this is equal to A. And of course, if the parallel plates are close to each other in
that case F12 would be equal to 1. So, therefore, all energy emitted by 1 is going to strike
2 if these two are very close to each other. So, no energy will escape through this if there
very close to each other.
So, F 12 is equal to 1. So, what you would get?

A (T14  T2 4 )
Q1 2 
1 1
 1
1  2 W
These many Watts that is the heat transfer the total amount of heat transfer radiative heat
exchange between 1 and 2, but in absence of a shield.
Now let us say we have a shield in between these two now. So, I will draw this circuit
diagram over here. So, for this one is going to be E b1, this one is going to be J 1 and the
1  1
resistance here is A1 , so I take the A to be the same. Then between J 1 and J31, so this
1
is J3 facing 1 this should be AF13 this is the resistance. Then we have between J 3 and Eb3,
693
1   31
so this is Eb3 and this resistance; obviously, is going to be A 31 . And then from Eb3 I
will have another resistance for J32 which is just over here. So, this is J31, this is J32. And
1   32
what is J3? The resistance that connects Eb3 and J32 must be equal to A 32
1
Now, J32 is going to be over here and this is J2 and this is simply going to be AF32
1
So, F13 and this is AF32 and this J2 is connected to Eb2 where the resistance in this case
1 2
would be A 2 . So, what you see here is once again you start at this point which is E b1,
over here this is J1, J1 to J31 that is the radiosity of surface 3 facing 1. Then you have for this
one you have Eb3 and J32 is the radiosity of surface 3 facing 2; then J32 and you have to find
out what is J2 and J2 and inside it is Eb2, so the number of nodes that you have are 1, 2, 3, 4,
5, 6.
So, what you have then is 1, 2, 3, 4, 5, 6 you have an extra.

So, 1, 2, 3, 4, 5, 6 and 7 nodes and in each of these nodes between E b1 and J1 you have the
surface resistance to radiation between J1 and J31. You have the resistance formula for the
enclosure J31 and Eb3 surface resistance and so on. So, you get the complete picture for this.
694
Now if I write the same Q the same formula like this one for this. For radiation shields
the formula would simply be
A (T14  T2 4 )
Q1 
1  1 1   31 1   32 1 2
1  1
1  31  32 2
So, here we know where we assume that F 13, since the plates are long and parallel they
are going to be equal to 1 everything is in series.
So, the resistances, so this is the first resistance since F 13 is equal to 1 and I have taken A
to the denominator. So, your aim through the use of this shield is to reduce the amount of
heat that one is going to lose or one is going to gain depending on what application you
have in mind.
So, when you are choosing the material of construction for the radiation shield the only
value that you have to consider is the emissivity ok. So, if you look at the expression
once again then you would be able to find out what is the property that you want. So, let
us take a look at the expression once again your ϵ1 and ϵ2 are known are known to you.
So, definitely you would like to use a material of low emissivity for a shield, so this is
what you would prefer. It becomes even more apparent if you simply write this Q1 as if
695
1 1
(   1)
you take this as  1 2 . So, just a bit of reorganization would give you that the heat
flow is going to be this.
A (T14  T2 4 )
Q1 
1 1 1 1
(   1)(   1)
1  2  31  32
Compare that with what we have obtained for the case of Q 12 when we did not have any
shield. So, because of the presence of the shield we have an extra term in the
denominator and additional resistance provided by the shield.
And looking at the expression you can clearly see you would like to have as small a
value of emissivity as possible for the shield material such that this resistance becomes
significant. So, this justifies our previous statement that we would like to have for the
shield a material of very low ϵ. So, we want low values of ϵ 31 and ϵ32 to make our shield
and this term if you compare it with this one this term provides the additional resistance
due to the presence of the shield material. So, for this special case and we can make a
special case when if ϵ for all the surfaces are equal. Then
A (T14  T2 4 )
Q
2
2(  1)

So, if there are N parallel shields, then this one can be generalized as Q for N number of
parallel shields is
A (T14  T2 4 )
QN 
2
( N  1)(  1)

And here you are simply going to get 2 and therefore, the QN in the case of N shields the
formula is going to be this.
So, this is more or less what I wanted to cover in radiation shields in your text you would
see that if the radiation shields are also common for tubes as well. So, you have a tube
which you would like to protect. So, you put another shield over here and the same
concept would also be applicable that is the radiation shield is going to provide an
additional resistance to radiation for the case of shield.
But there you just have to keep in mind that since the geometry is a cylindrical one. So,
your two areas, the area of the tube side and area of the protection radiation shield may
696
not be equal and while writing the resistance you have to use 1/A tube F12 in one case. And
in the other case instead of Atube you have to write A shield.
697
So, if these are very close to each other you can make an approximation that A shield is
equal to Atube and you will get back to the same result that we have obtained just now.
Otherwise just draw the circuit diagram in all these cases draw the circuit diagram
identify what is ϵ1? What is A1? If are they equal and not ϵ1 ϵ2 etcetera write the
resistance to heat transfer in an enclosure and also write the resistance for heat transfer
1  1
the resistance. The surface resistance to irradiation what is A11 ?
So, and see whether they are connected in parallel or they are connected in series, are
those shields, in the case of shields they are going to be connected in series one after the
other. And you simply have to add the resistances in order to find out what is the total
flow of heat in presence or absence of one or a number of shields. So, what I would do is,
I will give you one more problem to practice on for the case of radiative heat exchange in
an enclosure. Discuss some of the salient features and in the rest you have to solve that
problem.
So, for this problem what we do is, we have chosen a cylindrical furnace. So, this is a
cylindrical furnace in which let us call this as my surface 1. So, it is area is A 1,
temperature is T1, and it is a black body. So, Fϵ1is equal to 1.
698
So, it is a black body for the case the surface 2 is A 2 and its temperature is T2 and this is
also a black body. So, ϵ2 is equal to 1, the surface A 3 is insulated. The length, so this is a
cylindrical furnace, the length is 0.3 meter, and diameter is also 0.3 meter. So, this is 0.3,
and this diameter is also 0.3 meter ok. The surface A 1 and A2 the end surface and the
lateral surface are black. As you can see the way I have drawn it ϵ 1 and ϵ2 are both equal
to 1. Since they are black bodies and they are insulated as well ok.
The temperature of 1 is maintained is at 500 Kelvin, temperature at 2 is 400 Kelvin. I

will change this A3 is insulated as I have drawn over here. So, let me state the problem
once again it is a cylindrical furnace whose length and diameter are equal. So, this is 0.3
meter, this is also 0.3 meter, the surface 1 and the lateral surface 2 are both black; A1 is
to be maintained at 500 Kelvin, A2 has to be maintained at 400 Kelvin, the last surface A3
is insulated. You have to find the net radiation heat transfer from each of the surfaces.
And find the temperature of A3 that is you have to find: what is the value of T 3? It has
been given that for such a system F13 is equal to 0.172.
So, F1-3 is 0.172 the first thing that you have to do is you have to find out the unknown
view factors. So, F 13 is 0.172. So,
F13 + F11 + F12 = 1. Since it is a plane surface F 11 is equal to 0. So, you get F12 is equal to
0.828. You can also write from reciprocity relation that
A1F12 = A2F21. And therefore,
F21 = (A1/A2) ×F12
 (0.3) 2 / 4
F21  0.828  0.207
  0.3  0.3
So, from symmetry from symmetry, we can write F21 = F23 = 0.207.
699
So, if I draw the circuit diagram for this case ok. Now E b1 = J1, Eb2 = J2, both 1 and 2 are
black bodies. Additionally, Eb3 = J3, but for a different reason not for a blackbody since
A3 is insulated. However, we have some heat which is coming in as Q 1, some heat where
is coming as Q2. And of course, in this case Q3 is 0, since it is insulated.
So, the problem is pretty straightforward now. So, in order to find the net radiative heat
transfer from each of these surfaces let us find Q1. Q 1 would be
Eb1  Eb 2
Q1  1
 
 1 1 
  
 A1 F12 1 1 

 A1F13 A2 F23 
And, when you put all the values in there you are going should get Q1 to be equal to
143.46 Watt, so that is what Q3 is.
The next one is temperature of A3 or in other words what is the temperature T 3 in this
case. In all such cases you need to find out what is J 3 and since J3 is equal to Eb3, I need
to know: what is the numerical value of Eb3. Because the moment I know the
700
value of numerical value of Eb3; Eb3 is simply equal to σT34. So, I should be able to find
out what is T3, so, the requirement for this specific type of problem is to find out what is
J3. And in order to obtain J3 what I am going to say is that to find J3
J 3  J1 J 3  J 2
 0
1 1
A1 F13 A2 F2 3
That means, the algebraic sum of the current or in this case the heat at any node is equal
to 0.
So, if you do this when you put all these values in here you should be able to see J 3 to be
equal to 1811.4 W/m2. This is equal to Eb3 and Eb3 is equal to σT34. When you put the
values in there you should get the value of T 3 to be 422.7 Kelvin. So, this is another
example of how do you, how you convert the complex radiation exchange geometry to
something which now have the analogy from electrical science and find out what is
surface resistance to radiation.
What is a radiation exchange between these between the enclosures and then see which
resistance is in parallel, which resistances in series. And the flow of heat is simply going
to be the potential difference based on the black body emissive potential divided by the
effective resistance between those two surfaces. And everything else follows from there.
And in some cases, you have to use the view factor algebra the relations of the view
factor and you also have to remember that for reradiating surfaces the black body
emissive power is equal to the radiosity for that surface. So, if you keep all these in mind
then the problem on these can be tackled without much of a problem.
So, we have one more class left and in that class we are going to I am going to mostly
talk about what happens, if the system the enclosure that we are talking about is filled
with a participating medium which is a very common occurrence where the gases present
in the enclosure would start to participate in the radiation process. It would start to
absorb some of the radiation and therefore, it is going to violate one of the major
assumptions of the network method that the gases are not participating in the radiative
exchange process.
So, that would conclude our study on radiation as well as our this course on heat transfer.
701
Heat Transfer
Lecture – 60
Gaseous Radiation (Participating Medium)
This is going to be the last class of heat transfer. I hope you have learned something
some basic fundamentals about heat transfer, its applications, the design of heat
exchangers, the different modes of heat transfer both steady state and at unsteady state.
And for the last few classes, we were learning radiation in how and where radiation is
important and how can you make furnace calculations, which are consisting of more than
one surfaces forming an enclosure.
So, we saw the fundamental nature of radiation the emissivity, absorptivity, reflectivity,
transmittivity, the relation between emissivity and absorptivity and what is Kirchhoff’s
law, when and under what conditions, the different forms of Kirchhoff’s law are going to
be valid that. We also studied the concept of view factors and based on view factors, how
much of energy coming out of an object at a specific temperature is going to be
intercepted by another object, which is situated nearby and the view factor plays a very
important role in it, that is how much of the first surface, how much of the second
surface is directly visible from the first surface, which is going to be a function of the
solid angle and so on.
And after that we introduced the concept of a radiosity. So, it is the potential the
radiative potential of a surface, which is observed by someone standing just outside of
the surface. So, it not only takes into account the emission coming out of the object, it is
also going to account for the reflected part of any irradiation on the surface. So, if any
irradiation is falling on the surface part of it is going to get reflected based on the
reflectivity of the surface and therefore, what we see as the radiate, true radiative
potential of the surface is not just the emissive power, but also the part of the reflection
part of the incident energy which gets reflected. So, it is the sum of the two.
And the difference between the emissive power of the surface had this been a black body
and the radiosity of the surface is expressed in the form of a circuit, where E bλ T that is a
radiative potential the emissive power of the surface if it is a
702
blackbody at a specific temperature, the tempering the same temperature as that of the
surface and j being the radiosity of the object these 2 potentials are related by a
resistance, which is known as the surface resistance to radiation and the difference
between the potential of these 2 and based on the resistance for radiation of the surface,
we can find out what’s the heat flowing from the surface to the outside? Or how much of
heat is to be supplied to the surface to maintain its temperature constant?
Based on this, the enclosure method has been developed and we have seen what it is
going to look like for a 2 zone enclosure? What it is going to like for a 3 zone enclosure?
How to convert it into star delta? And so on. So, more or less the concepts, which you
have already which you already know based on your based on, your concepts of
electrical technology would be equally applicable here for example, the sum of all heat
coming to a node would be algebraic sum would be equal to 0 and so on. So, when you
think of the the triangular type of resistance circuit curve then, you know that 1
resistance is going to be in parallel with the sum of the 2. So, whether it is a resistance in
series already in parallel based on that, you would be able to connect the potentials of the
2 surfaces and find out, how much of heat gets transported from one surface to the other
by radiation and as I said before that the main application of this method is in trying to
find out, what is going to be the heat to be supplied to a furnace to maintain its hot
surface at a constant temperature or how much heat is going to be received by the
surface, which you would like to treat if it is inserted in if it forms one of the surfaces of
the furnace? So, if you put something in the furnace, how much of heat its going to
receive by radiation from the adjoining hot walls.
703
So, those are very important calculations and which with the applications with the
examples that I have solved in the class, I think you do probably know, how it is going to
be, how the radiation is going to be treated for an enclosure as well. In all the discussion
that we have had so far, we have assumed something very important for example, the
case of an enclosure, an enclosure is usually never under vacuum ok. So, the enclosure
has some gases present in it, it could be air, it could be nitrogen, inert atmosphere or
something else in all our assumptions in all our calculations. So far, we assume that the
gas or the vapor present inside the enclosure, the gases they do not take part into the
radiation process.
So, as far as the radiative heat transfer is concerned, all these gases are truly transparent.
So, they either they do not absorb anything, they do not emit anything and so on, which
this assumption is mostly true for non-polar gases, for example, nitrogen, oxygen and so
on, but the moment you have polar gases or vapor present in the enclosure, they start to
absorb radiation and they will also depending on their temperature can emit radiation as
well.
Whenever such gases or vapors these gases and vapors, who do participate in the
radiative process, it is called radiation under participating media, radiation in the
presence of participating media. Now many of the common, common gases or vapors
that we deal with which are expected to be present in an enclosure are polar are non polar
in nature. So, whatever we have done so far will still be valid, but let us take the case,
when we have humid air present in the furnace or the furnace
704
is filled up with carbon dioxide or something else, which starts to participate in the
radiation process then, we whatever be the emission coming out of a hot surface and you
have another hot surface nearby and the intervening space is filled with water vapour
then, all these emission are not going to come directly to the surface, part of it is going to
be absorbed and; obviously, as the thickness of this film, thickness of this vapour film
increases more is going to be the absorption.
So, if you have absorption on one side of the film as the gas as the radiation proceeds
deeper and deeper into the gas film, it is going to be absorbed and its intensity is going to
reduce. Now this reduction of intensity as a function of the thickness of the film or as a
function of the distance traveled by the by the incident radiation into the participating
media, that is going to be important in our calculations. Secondly, the gas itself may start
emitting radiation as well. So, there are two aspects to the participating media that, we
have to be careful about, one is the absorption and second is the emission. So, we are
going first were going to start with the absorption of radiation as absorption of radiation
as a function of distance; that means, as a function of thickness of the gas.
So, that is the first thing, we are going to start with. So, our first thing is the important
point here is the participating medium. So, the gas that present in it is going to be going
to participate in it. So, examples for this are carbon dioxide, this also H 2O vapor or a
combination of these two. So, these are polar molecules, the polar molecules the polar
molecules start to participate in it and they absorb over a wide range. And the
complication to this if even more is that radiation is concentrated, radiation for these
molecules is concentrated at particular wavelengths, these particular wavelengths are
called bands.
So, that is a further complication, which we will also have to keep in mind, but this
coming back to the volumetric absorption of radiation in participating gases. This is a
figure which so this is the starting of the gas film, which is participating is of thickness L
and you have some spectral radiation, which is falling on the surface and it starts to
propagate as it propagates some amount of radiation is going to get is going to get
I  ,0
absorbed by it. So, if you have a monochromatic beam of , which is incident on the
medium the intensity gets on reducing. And let us take an infinitesimal thickness of the
participant medium denoted by dx then the amount of absorption or change in the
radiative intensity due to the passage of radiation through a thickness dx is given as
705
dI  ( x)   k I  ( x) dx
706
This K λ is a property, which is known as the absorption coefficient, which is going to be
different for different gases.
So, for a non participating medium the value of K λ would be 0, but for participating
medium; For example, CO2, H20 depending on what whatever be, the participating
medium, the value of the absorption coefficient would be different and they are known as
well. So, if we assume that k  f ( x) , then this equation can simply be integrated with
the boundary condition that at x equal to 0. The value of the monochromatic irradiation is
I  ,0
equal to , then integrating this equation, the intensity the spectral intensity after a
distance L as compared to the spectral intensity at the beginning at x = 0 would simply
be equal to
I ,L
 e  k L
I  ,0
where L is the thickness of the film and this equation is known as Beer’s law.
So, Beer’s law it states that the intensity of radiation after the incident radiation passes
through a distance L is going to decay exponentially and the constant is known is a
property of the gas, which is the absorption, which is the absorption coefficient. So, what
are going to be the use of this equation, the obvious use of this equation is it gives you as
some idea of the transmittivity. The transmittivity, the spectral transmittivity τ λ, which
would be equal to
I ,L
   e  k L
I  ,0
and the absorptivity, which is denoted by   is simply

   1     1  e  k L
So, if we assume that the reflectivity to be equal to 0 and this would give you
   1     1  e  k L
And if Kirchhoff’s law is valid, if we assume that Kirchhoff’s law is valid in that case,
  
707
So, Beer’s law provides the spectral emissivity of the medium. So, the important thing
that I would like to here is that it gives the ϵ for the medium. So, what you see over here
is that rather to several applications of Beer’s law, first of all it gives you an idea of how
much of radiation you are going to get at a distance L, if the incident radiation has a
I  ,0
value . So, that is what is Beer’s law, and this is a traditional definition of
transmittivity. So, transmittivity is nothing, transmittivity is going is simply going to be
e  k L .
And since for non reflecting surface, which is a valid approximation for a gas,
absorptivity and transmittivity are related by this expression. So, what you get is the
absorptivity as 1 minus this and if Kirchhoff’s law is valid. So, this is if Kirchhoff’s law
is valid, then we can write    . So, Beer’s law gives you an estimate of the value of
the emissivity for the medium. So, that is what the now there is few more things we also
know. So, we have now an idea of what is going to be the absorptivity in gas mass. But
we would also like to have some idea of what is the emissive potential of a participating
gas participating gas or a vapor, because not only it is going to absorb since it absorbs, it
is going to emit as well.
708
So, how do I express the emissive power of a participating medium which is a gas or a
vapor. So, our next analysis starts with the gaseous emission and you understand that for
in many cases a hot gas is going to emit radiation, which is substantial gaseous emission
and absorption. So, the aim is to the aim is to evolve to determine, the radiant heat flux
from a gas to an adjoining surface. So, this is what the aim of the present analysis is and
the method, which is used here is known as the Hottel method. This is method is a
graphical method, which involves determining the radiation emission from a
hemispherical gas. So, the first thing that is assumed here that it is a hemispherical gas
mass, gas mass with a temperature Tg and it is emitting a to a surface element dA 1, which
is at the center of the hemisphere.
So, the Hottel method gives us first it assumes that there is a hemispherical mass of hot
gas and whatever be the emission, it is going to be it is going to be concentrated at its
center. So, if I have a hemispherical mass of hot gas, the Hottel method gives you the
emission from the hemispherical mass of gas concentrated on the unit area, which is
placed at its center.
So, that is what Hottel method tells you and the expression that, he has used in order to
express this irradiation on the surface is given as this
Eg  g  Tg 4
This ϵg, we have we already know what this ϵ g is the gas emissivity and it is correlated it
is, obviously, it is going to be a function of temperature and T g is and it is going to be a
function of temperature, it is going to be a function of the temperature of the gas, it is
going to be a function of pressure of the gas and so on.
So, what he has shown is that, the in a graphical

709 way, he has calculated what is this value
of ϵg going to be and the this is a function of gas temperature, which is Tg and he has got
a family of curves and this is in increasing order of p w times L. So, pw this is for water.
So, for water the value of the value of the emissivity is obtained by a curve like this,
where this pw is equal to the partial pressure of the water vapor, how much of water
vapor is present in air partial pressure of water vapor.
So, which is logical right because, the more the water more vapor is present in air the
more it is going to absorb. So, whatever expression that he has provided whatever values
that he has calculated for ϵg, it is definitely a function of f will be the function of T g,
which is the temperature of the gas, but it is also a function of the amount of water
present into it at a given point of time. So, the figure that you see over here is for 1
atmosphere. So similarly, pressures the value of ϵ at other pressures can also be
calculated.
So, there are 2 major issues in here, the first one is this is for 1 atmosphere. So, similar
curves must be obtained for higher or lower pressures and secondly, it is assumed that
the gas forms a hemispherical enclosure around an object, which is placed at the center
of the hemisphere. So, these 2 are limitations of Hottels approach or Hottels method. So,
for each of the each of the pressure different from 1 atmosphere, you have to have a
similar such curve for the water vapor, I will come to that hemispherical part later on, not
only that if you have instead of water, you have carbon dioxide present in air then, you
are going to have a different curve for the value of ϵg.
So in other words, a pressure correction must be made, if the pressure is different from
that of the atmospheric pressure. Secondly, if it is something not water vapor, but some
other participating medium is present another family of curves would have to be would
have to be available in order for us to make the calculation. There is a third part as well,
what if two participating medium are simultaneously present in the mixture of gas at
temperature Tg. For example, you may have air in, in air both water vapor and carbon
dioxide present simultaneously which can happen. So, if that happens and if that air mass
is at a certain temperature then, how do you calculate the value of ϵg? The
710
emissivity of the gas mass containing different quantities, different partial pressures of
water and carbon dioxide present simultaneously.
So, you need furthermore charts and so on, in order to get an idea of what is the value of
ϵg going to be. So, quickly give you an example and more examples of these are available
in your text, which if you can refer to whenever you need to calculate the value of ϵg for
a participating medium, which is present at a given partial pressure and when the system
pressure is 1 atmosphere or maybe different from 1 atmosphere. So, the first thing is the
effect of pressure then, we are going to look at the effect of the presence of more than
one participating medium. And finally, at the last part that we are going to clarify is how
to take care of this hemispherical assumption and how to take, how to express, how to
obtain the proper length scale to be used to evaluate, what is going to be the value of ϵ g?
But first let us look at the effect of pressure.
So, the for pressures different from 1 atmosphere, the this is the figure of a pressure
correction, which is Cw and you get curves like this and these are in increasing order of
the partial pressure of water times L where L is the characteristic length and this is p w the
partial pressure plus P, the total pressure divided by 2. So, this is the partial pressure of
water vapor, this is the total pressure divided by 2 and ϵ of water for a pressure, which is
not equal to 1 atmosphere is equal to C w times ϵ of water, when the pressure is equal to 1
atmosphere.
w, p 1atm  Cw w, p 1atm
711
So, this is a correction factor which must be must be taken into account, when your
pressure is different from that of that of 1 atmosphere. And you get similar results that
are available in literature for other gases and when you have, when the both let us say
both CO2 and H2O vapors are present. If that is the case then you have in this method,
you have the collection factor, this is partial pressure of water divided by partial pressure
of carbon dioxide plus partial pressure of water. Partial pressure of carbon dioxide and
partial pressure of water and you have another family of curves, which these are in
increasing order of L, which is the length scale that we are going to talk about next.
So, this and the value of ϵ g when both CO2 and water vapor are present is going to be
equal to
g w  c  
So, this   value can be read from these curves from these graphs and together they
would give you some idea of what is the emission from a gas mixture, consisting of
water vapour as well as carbon dioxide or any other participating medium. So, the last
thing what is remaining in our discussion is how do we get, how do we get an idea of
what is going to be the L the length scale. So, what is the length scale to be used, because
you are not going to get you know assuming a hemispherical mass of participating gas
radiating and your surface of interest is at the center that is something, which is very
inconvenient to use and may not be the may not be the right one, right calculation
method.
712
So, we need to get an idea of an equivalent length that can safely be used with whatever
we have discussed so far. So, this characteristic length L, this is what I am going to
describe to you next.
So, the next one is about the next point is about length. So, what is going to be the length
scale for to be used in this case? So, the concept that I am going to introduce is called the
mean beam length, what is mean beam length? It mean beam length is let us call it MBL.
So, MBL, mean beam length is the, mean beam length is the emissivity the radius of a
hemispherical gas mass, whose emissivity is equivalent to the gas mass geometry of
interest.
So, this MBL is denoted by Le. So, Le is the radius of a hemispherical gas,
hemispherical gas mass whose emissivity is equivalent to the gas mass geometry of
interest. So, there is a table, which tells you about the geometry, what is the characteristic
length and how the beam length mean beam length is related to the characteristic length?
So, these are straightforward, the first one let us say it is a sphere ok. So, the
characteristic length is, obviously, D which is the diameter and the mean beam length
can be calculated as 0.65 D. If it is infinite circular cylinder and you have radiation to
curved surfaces again, this is going to be D and this is going to be 0.95 D infinite parallel
planes.
So, it is going to be the characteristic length is going to be spacing between the 2 plates,
let us call it as L and the main beam length is going to be 1.8 L. If it is a cube and you
713
are considering radiation to any of the surfaces then; obviously, the characteristic length
is going to be L and this is going to be 0.66 L.
The important one, these are standard standards standard geometries a sphere, an infinite
circular cylinder, when you are calculating when you are trying to find out, what is the
radiation to the curved faces, an infinite parallel planes separated by a distance L or if it
is a cube of side L then, there analytically you can find out, what is going to be the value
of the Le, but if you have an arbitrary shape of volume V and surface area A. So, it is a
volume V surface area A then, the characteristic length as we have done before is volume
by area, that is your characteristic length or V/A and the beam length is going to be 3.6
V/A and this is just an approx value.
So, the mean beam length as is shown here gives you. If you look at the table once again,
that the concept of mean beam length is the radius of a hemispherical gas mass, whose
emissivity is equivalent to the gas mass of geometry that, we have with us. So, the gas
mass of gas mass of any geometry that is available to us. If we can find out what is the
corresponding beam length mean beam length Le, then that Le would be used in all the
relations and correlations that we have developed, so far for the case of gaseous
emission.
So, the point therefore, is how to find out Le for any arbitrary geometry? The what we
have seen is that for specific geometries like a cube, a cylinder, parallel plates and
spheres, what is going to be the relation between characteristic length and mean beam
length and ultimately, what we have seen is that for the case of any arbitrary shape, the
characteristic length is simply going to be V/A and what is the relation between the beam
length and V/A for any arbitrary shape.
The aim is to obtain Le because, if we can get Le for the enclosure surface and the
volume of the volume of the participating medium inside an enclosure then, the
corresponding Le can be obtained. The moment, I have Le then using the methods that I
have described at the beginning. You can find out what is the value of ϵ? What is the
value of emissivity of the participating medium as a function of temperature and as a
function of partial pressure and so on; and what is the value of the absorptivity? What is
the value of transmittivity? And everything else then can be calculated and you will be in
a position to find out, what is the net amount of radiative heat transfer in between a hot
714
gas and an adjoining surface which is something very important that, we would like to
know for many of the practical gases that we come across.
So, that is the end of the chapter radiation, which I wanted to wanted to tell you, I
wanted to introduce you to in this course and the radiation is therefore, one of the very
interesting, very practical branch of heat transfer that we must be aware of that we know
how to calculate the radiative heat exchange, especially for equipments involving
elevated temperatures, what is the radiative heat loss or a gain.
And many of the natural phenomena that, we see around us is also due to a due to
radiation, how much of heat we receive from the sun and so on. So, radiation plays a
critical role in engineering in everyday life and through the discussion of the past few
classes, I think you would be in a position now to make a calculation of the heat
exchange between objects between solid objects forming an enclosure, and between an
solid object and an adjoining gas, if the adjoining gas is a participating medium.
So, let us brings us to the end of this course. I hope you have enjoyed it, if there are any
questions, I would encourage you to post them at the forum and they would I and that the
TAs would try to answer them. And hopefully this course has given you a foundation for
calculating heat transfer in a variety of ways both conduction, convection, radiation
steady state, unsteady state, internal flow, external flow, fully developed flow,the heat
exchanger calculations, radiation and so on; so best of luck.
Thank you.
715
THIS BOOK
IS NOT FOR
SALE
NOR COMMERCIAL USE
(044) 2257 5905/08

nptel.ac.in
swayam.gov.in

Heat Transfer

Uploaded by

Document Information

Original Title

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

Heat Transfer

Uploaded by

Copyright:

Available Formats

INDEX

S. No Topic Page No.

(Refer Slide Time: 09:22)

2) The second book that we would occasionally refer to is transport phenomena by

As discussed previously, heat transfer is energy in transit due to a temperature difference.

(Refer Slide Time: 13:00)

own the conservation equation would simply be E in = E out .

(Refer Slide Time: 06:07)

Similarly, the equation fo the y and z-directions can be represented as:

So, the final conservation equation can be represented as:

the time rate of change in energy storage inside the system.

If k is a constant; that means, it is not a function of x, y, and z; then the system is

(Refer Slide Time: 21:39)

A slight rearrangement gives another form of this equation:

Or in other words it is expressed as

So, therefore, your governing equation would be

So, in convection it is customary to define a convective heat transfer coefficient and it

(Refer Slide Time: 18:28)

So, qs” which is the heat flux at x = 0, must be equal to,

So, your governing equation in that case would be

(Refer Slide Time: 33:54)

So, the T as a function of x is known to me.

be able to evaluate. Now what is E stored . So, from my energy equation

(Refer Slide Time: 42:38)

(Refer Slide Time: 43:56)

(Refer Slide Time: 06:17)

(Refer Slide Time: 12:18)

Rearranging the above equations, we get:

The k A , k B , and kC are the conductive thermal conductivities of material A, material B,

(Refer Slide Time: 23:02)

It is sometimes advisable to express these 3 resistances is in terms of an overall heat

(Refer Slide Time: 30:34)

This is a 2 D condition where the temperature is not a going to be a function of x alone, it

T(ri) = Ti; T (r0)= To

(Refer Slide Time: 12:15)

qr  hAT  h1 2 r1L1  T1  T1 

conservation of energy is the overall temperature difference which is T1  T 2 by the

So, starting with this

So, that would be the topic of the next class.

(Refer Slide Time: 04:27)

6 104  / m . The electric wire is in an atmosphere where the temperature is 30 oC ,

(Refer Slide Time: 14:02)

Let us go to the third part of the problem, which is interesting.

where, v is the volume of the system.

And the temperature distribution can also be expressed in terms of a dimensionless

So, as I said that

(Refer Slide Time: 14:40)

cooling which be would be a h(TS  T ) .

So we, looking at the case heat source systems.

(Refer Slide Time: 19:38)

So, what you get from here is

where, R is the radius of this wire.

If the wall temperature of the wire is known to r is known to us which is at r = R. Then,

(Refer Slide Time: 25:17)

So, fundamentally in terms of the governing equation, in terms of the boundary

(Refer Slide Time: 01:08)

Therefore, the temperature profile for material A would be:

(Refer Slide Time: 16:51)

So, T at x equals 0 from this previous expression T at x equals 0 I call it as T 0 which

(Refer Slide Time: 23:05)

profile, which you should get for such a case.

So, therefore, we understand 2 things that transient conduction is something which is

(Refer Slide Time: 07:30)