Heat Transfer 1. Conduction 2021

MECHANICAL ENGINEERING AND SOLID MECHANICS SERIES
MATHEMATICAL AND MECHANICAL ENGINEERING SET
Volume 9
Heat Transfer 1
Conduction
Michel Ledoux
Abdelkhalak El Kami
Wiley
Heat Transfer 1
Mathematical and Mechanical Engineering Set
coordinated by
Abdelkhalak El Hami
Volume 9
Heat Transfer 1
Conduction
Michel Ledoux
Abdelkhalak El Hami
Wiley
First published 2021 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as
permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced,
stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms and licenses issued by the
CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the
undermentioned address:
ISTE Ltd John Wiley & Sons, Inc.

27-37 St George’s Road 111 River Street
London SW19 4EU Hoboken, NJ 07030
UK USA
www.iste.co.uk www.wiley.com
© ISTE Ltd 2021

The rights of Michel Ledoux and Abdelkhalak El Hami to be identified as the authors of this work have
been asserted by them in accordance with the Copyright, Designs and Patents Act 1988.
Library of Congress Control Number: 2020949611
British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library
ISBN 978-1-78630-516-9
Contents
Preface ................................................................................................. ix
Introduction .......................................................................................... xiii
Chapter 1. The Problem of Thermal Conduction: General

Comments ............................................................................................. 1
1.1. The fundamental problem of thermal conduction ............................. 1
1.2. Definitions ......................................................................................... 2
1.2.1. Temperature, isothermal surface and gradient ............................. 2
1.2.2. Flow anddensityofflow............................................................ 4
1.3. Relation tothermodynamics............................................................... 5
1.3.1. Calorimetry................................................................................. 5
1.3.2. Thefirstprinciple......................................................................... 6
1.3.3. Thesecondprinciple.................................................................... 6
Chapter 2. The Physics of Conduction ........................................... 9

2.1. Introduction......................................................................................... 9
2.2. Fourier’slaw...................................................................................... 9
2.2.1. Experiment.................................................................................... 9
2.2.2. Temperature profile .................................................................... 12
2.2.3. General expression of theFourierlaw.......................................... 14
2.3. Heatequation...................................................................................... 16
2.3.1. General problem ......................................................................... 16
2.3.2. Mono-dimensionalplaneproblem............................................... 18
vi Heat Transfer 1
2.3.3. Case of the axisymmetric system ............................................... 24

2.3.4. Caseofthesphericalsystem....................................................... 25
2.4. Resolutionofaproblem .................................................................... 26
2.5. Examplesofapplication.................................................................... 29
2.5.1. Problemsinvolvingsphericalsymmetry ..................................... 40
Chapter 3. Conduction in a Stationary Regime ............................... 53

3.1. Thermalresistance............................................................................... 53
3.1.1. Thermalresistance:planegeometry............................................ 53
3.1.2. Thermal resistance: axisymmetric geometry.
The case of a cylindrical wall ................................................................. 62
3.1.3. Thermal resistance to convection ............................................... 65
3.1.4. Criticalradius............................................................................... 67
3.2. Examples of the application of thermal resistance in plane geometry . 69
3.3. Examples of the application of the thermal resistance
incylindricalgeometry................................................................................. 85
3.4. Problem of thecriticaldiameter.......................................................... 92
3.5. Problemwiththeheatbalance............................................................ 99
Chapter 4. Quasi-stationary Model .................................................. 103

4.1. We can perform a simplified calculation, adopting the following
hypotheses................................................................................................... 103
4.2. Method:instantaneousthermalbalance............................................ 104
4.3. Resolution............................................................................................ 106
4.4. Applicationsforplanesystems.......................................................... 107
4.5. Applications for axisymmetric systems ............................................ 152
Chapter 5. Non-stationary Conduction ........................................... 183

5.1. Single-dimensionalproblem............................................................... 183
5.1.1. Temperature imposed at the interface at instant t =0.................. 184
5.2. Non-stationaryconductionwithconstantflowdensity..................... 190
5.3. Temperature imposed on the wall: sinusoidal variation ..................... 193
5.4. Problemwithtwowallsstucktogether............................................... 200
5.5. Applicationexamples......................................................................... 204
5.5. 1.Simpleapplications...................................................................... 204
5.5. 2.Somescenesfromdailylife......................................................... 213
Contents vii
Chapter 6. Fin Theory: Notions and Examples .............................. 237

6.1. Notions regarding thetheoryoffins.................................................. 237
6.1.1. Principleoffins............................................................................ 237
6.1.2. Elementaryfintheory................................................................. 238
6.1.3. Parallelepipedfin......................................................................... 242
6.2. Examplesofapplication.................................................................... 249
Appendices .......................................................................................... 263
Appendix 1. Heat Equation of a Three-dimensional System ....... 265
Appendix 2. Heat Equation: Writing in the Main

Coordinate Systems ............................................................................ 273
Appendix 3. One-dimensional Heat Equation .................................. 283
Appendix 4. Conduction of the Heat in a Non-stationary Regime:

Solutions to Classic Problems ........................................................... 291
Appendix 5. Table of erf (x), erfc (x) and ierfc (x) Functions .......... 295
Appendix 6. Complementary Information Regarding Fins ............ 297
Appendix 7. The LaplaceTransform ................................................. 301
Appendix 8. Reminders RegardingHyperbolic Functions ............... 309
References ............................................................................................. 313
Index ...................................................................................................... 315

Preface
Thermal science is to thermodynamics as decree is to law. It answers the

following question - which all good leaders must (or should) ask themselves
whenever they have an “idea”: “How would this work in practice?”. In a
way, thermal science “implements” thermodynamics, of which it is a branch.
A thermodynamics specialist is a kind of energy economist. Applying the
first principle, they create a “grocery store”. With the second principle, they
talk about the quality of their products. I add or remove heat from a source
or work from a system. And the temperature, among other things, defines the
quality of the energy for me.
But by what means do I take or do I give? Even calculations of

elementary reversible transformations do not tell us by what process heat
passes from a source to a system.
Thermal science specifies how, but “evacuates” the work. If in a given

problem related to, for example, a convector where electrical energy
(therefore in the “work” category) appears, it is immediately dissipated into
heat by the Joule effect.
Three heat transfer modes can be identified: conduction and radiation -

which can be seen separately, although they are often paired up - and
convection, which is by nature an interaction of fluid mechanics and
conduction.
Dividing the study of thermal science into three is the result of logic.
Presenting this work in three volumes is somewhat arbitrary; in our opinion,
x Heat Transfer 1
however, this split was necessary in order to keep the volumes in the
collection a reasonable size.
This book is Volume 1 of a collection of problems on heat transfer,

devoted to thermal conduction and numerical approaches to such transfers.
Despite being a collection of exercises a priori, a large part is given over to
recalling the practice. To a large extent, the book constitutes a first
introduction to the thermal calculation of practical devices, which may be
stand-alone. For the subsequent calculations, the reader will not be spared
the use of specialist textbooks or encylopedias available in the field of
thermal engineering.
The book is intended to reach a wide audience, from technicians to

engineers, to researchers in many disciplines, whether physicists or not, who
have a one-off transfer problem to resolve in a laboratory context. With this
in mind, the theoretical developments in the text itself are as direct as
possible. Specialist readers, or those who are simply curious about further
theoretical developments (general equations, specific problems,
mathematical tools, etc.), may refer to the Appendices.
Volume 1, dedicated to “classic” approaches (analytical treatment) to

conduction, will be of interest primarily to readers who are looking for
“simple” prediction methods.
After the generalities outlined in Chapter 1, Chapter 2 presents the

physical laws of conduction, describing the Fourier law and setting out the
heat equation. We then pinpoint the fundamental content of the problems
found in conduction and their approach. At this point, we will need to
distinguish between stationary (we also refer to permanent regimes) and
non-stationary problems.
Chapter 3 deals with conduction in a stationary regime. Emphasis is

logically placed on plane and cylindrical geometries. Importance is placed
on the concept of thermal resistance, an essential tool for thermal scientists,
in both of these geometries. This chapter provides many examples of
application of this concept.
Chapter 4 presents the notion of a quasi-stationary regime; this method,

although approximate, does in fact have undeniable practical scope. Valid
for relatively slow transfers - presuming the temperature environment is
Preface xi
homogeneous - instantaneous balances can be obtained, in addition to the

laws of thermal evolution that allow valid approximations to the problems
that, in principle, relate to the variable regime. Again, in this context, this
chapter deals with the plane and cylindrical problems.
Chapter 5 deals with variable conduction regimes. The most classic

monodimensional problems are tackled: fixed temperature at the interface at
the instant t =0, non-stationary conduction at constant flow density,
temperature with sinusoidal variation fixed at the interface and the problem
of two adjoining walls.
Many examples are provided to illustrate this important aspect of

conductive transfers.
Chapter 6 presents the notion of fin theory, which is associated with a

few simple examples.
Within the Appendices, there are tables of error functions and their
offshoots: erf, erfc, ierfc, as well as reminders that are often essential for
hyperbolic functions. They also provide information on the notion of treating
certain non-stationary problems using Laplace transformations. Again, many
examples are given to illustrate another important aspect of conductive
transfers.
November 2020
Introduction
I.1. Preamble
Thermal energy was probably first perceived (if not identified) by

humanity, through the Sun. The themes of night and day are found at the
center of most ancient myths. Humanity’s greatest fear was probably that the
Sun would not return again in the morning. Fire became controlled in
approximately 400,000 BP. Thermal transfer was therefore a companion of
Homo ergaster, long before Homo sapiens sapiens.
However, it took a few hundred thousand years before so-called “modern”

science was born. Newtonian mechanics dates from three centuries ago.
Paradoxically, another century and a half passed by before energy was
correctly perceived by scientists, in terms of the new field of thermodynamics.
Furthermore, a systematic study of heat transfer mechanisms was carried out
at the end of the 19th century, and even later for the study of limit layers, the
basis of convection.
Heating, lighting and operating the steam engines of the 19th century
were all very prosaic concerns. Yet this is where revolutions in the history of
physics began: the explosion of statistical thermodynamics driven by
Boltzmann’s genius, and quantum mechanics erupted with Planck, again
with Boltzmann’s invovlement.
Advances in radiation science, particularly in sensor technology, have

enabled us to push back our “vision” of the universe by a considerable
number of light years. To these advances we owe, in particular, the renewed
interest in general relativity that quantum mechanics had slightly eclipsed,
xiv Heat Transfer 1
through demonstration of black holes, the physics of which may still hold
further surprises for us.
Closer to home, fundamental thermal science, whether it is conduction,

convection or radiation, contributes to the improvement of our daily lives.
This is particularly true in the field of housing where it contributes, under
pressure from environmental questions, to the evolution of new concepts
such as the active house.
The physics that we describe in this way, and to which we will perhaps
introduce some readers, is therefore related both to the pinnacles of
knowledge and the banality of our daily lives. Modestly, we will place our
ambition in this latter area.
There are numerous heat transfer textbooks in different formats:

“handbooks” attempting to be exhaustive are an irreplaceable collection of
correlations. High-level courses, at universities or engineering schools, are
also quite exhaustive, but they remain demanding for the listener or the
reader. Specialist, more empirical thematic manuals are still focused on
specialists in spite of all this.
So why do we need another book?
The authors have taught at university level and in prestigious French

engineering schools, and have been involved in the training of engineers on
block-release courses. This last method of teaching, which has been gaining
popularity in recent years, particularly in Europe, incorporates a distinctive
feature from an educational point of view. Its practice has, in part, inspired
this book.
The aim is to help learners who have not had high-level mathematical
training in their first years following the French Baccalaureate (therefore
accessible to apprentices), and pupils with more traditional profiles. At the
same time, we would like to show this broad audience the very new
possibilities in the field of digital processing of complex problems.
When a miner wants to detach a block of coal or precious mineral from a

wall, they pick up a pneumatic drill. If we want to construct a tunnel, we
must use dynamite. The same is true for physicists.
Introduction xv
Whether they are researchers, engineers or simply teachers, scientists

have two tools in their hands: a calculator and a computer (with very
variable power). Since both authors are teacher researchers, they know they
owe everything to the invention of the computer. From the point of view of
teaching, however, each one of the two authors has remained specialist, one
holding out for the calculator and “back-of-the-napkin” calculations, and the
other one using digital calculations.
The revolution that digital tools has generated in the world of “science”
and “technical” fields, aside from the context of our daily lives, no longer
needs to be proved. We are a “has been” nowadays if we do not talk about
Industry 4.0. The “digital divide” is bigger than the social divide, unless it is
part of it...
Indeed, the memory of this revolution is now fading. Have students today
ever had a “slide rule” in their hands? Do they even know what it is? Yet, all
the physicists behind the laws of thermal science had only this tool in hand,
giving three significant figures (four with good visibility and tenacity),
leaving the user to find the power of ten of the result. It goes without saying
that a simple calculation of a reversible adiabatic expansion became an
ordeal, which played a part in degrading the already negative image of
thermodynamics held by the average student.
This reminder will seem useless to some; slide rules are at best sleeping
in drawers. But there is a moral to this story: no matter what type of
keyboard we type on, a calculator or a computer, our head must have control
over our fingers. This book has been written on the basis of this moral.
A good physicist must have a perfect understanding of the idea of an

“order of magnitude”. For this, the tool is a calculator. We always do a rough
sizing of a project before moving on to detailed modeling and numerical
calculations.
The two authors belong to the world of engineering sciences, meaning

most of their PhD students have entered the private sector. One of them,
having moved into the aerospace sector, came back to see us very surprised
by the recurrence of “back-of-the-napkin” calculations in his day-to-day
work.
xvi Heat Transfer 1
Fundamental or “basic physics” concepts are taken from a type of manual

that is resolutely different from those dedicated to the numerical approach.
In this case, the authors allow themselves to believe that it is no bad thing to
collect them all together in a single book, for once. This is a significant
difference that will surprise some and, without doubt, be criticized by others.
Nevertheless, when reading this book, an “average” student will be initiated
to a field that teaching models generally promised “for later on” (or never if
he/she never goes beyond a certain level of education). It is also true that
fully immersed in equations and complex calculations, specialist readers will
be able to “be refreshed” when faced with the short exercises, which can
sometimes surprise and encourage them - why not - to go back to their roots
(assuming they had indeed been there).
Another significant difference is that this book is directed at a large

scientific audience, which covers possibly the entire field: researchers, PhD
students or those who have obtained Confirmation and are just starting out in
the field, technicians, students or professionals, engineers. This last type of
scientist is perhaps the main target of this book.
So, what is this book for?
Above all, it contains problems to be worked on, of which most are

accessible to all, from the level of an apprentice technician upwards, either
one or two years after the Baccalaureate. This book was written in France,
where scientific teaching is structured around universities, engineering
Grandes Ecoles, engineering training through apprenticeships and two types
of technician training sections at high schools or universities. In countries
with simpler models, readers should also find it useful.
It seems necessary to surround these problems with strong reminders of

past learning, so that the reader does not need to permanently refer back to
their manuals. We see two advantages in this: a presentation of the scientific
material focusing on the problems, and a second chance for readers to
integrate notions that perhaps had not been well understood in the initial
teaching.
Lastly, upon rereading, the authors also recommend this book as an

introduction to the taught disciplines.
Introduction xvii
I.2. Introduction
Thermal science is to thermodynamics as decree means is to law. It

answers the following question - which all good leaders must (or should)
ask themselves whenever they have an “idea”: “How would this work in
practice?”.
In a way, thermal science “implements” thermodynamics, of which it is a

branch.
A thermodynamics specialist is a kind of energy economist. Applying the

first principle, they create a “grocery store”. With the second principle, they
talk about the quality of their products. I add or remove heat from a source
or work from a system. And the temperature, among other things, defines the
quality of the energy for me.
But by what means do I take or do I give? Even calculations of

elementary reversible transformations do not tell us by what process heat
passes from a source to a system.
Thermal science specifies how, but “evacuates” the work. If in a given

problem related to, for example, a convector where electrical energy
(therefore in the “work” category) appears, it is immediately dissipated into
heat by the Joule effect.
Three heat transfer modes can be identified: conduction and radiation -

which can be seen separately, although they are often paired up - and
convection, which is by nature an interaction of fluid mechanics and
conduction.
Dividing the study of thermal science into three volumes is the result of
logic. Presenting this work in three volumes is somewhat arbitrary; in our
opinion, however, this split was necessary in order to keep the volumes in
the collection a reasonable size.
The first volume, entitled Heat Transfer 1, is dedicated to “classic”

approaches (analytical treatment) to conduction, which will be of greater
interest to readers who are looking for “simple” prediction methods.
The second volume, entitled Heat Transfer 2, is dedicated to “classic”

approaches (analytical treatment) of radiation, and assembles digital
xviii Heat Transfer 1
approaches of these various transfer modes. It is aimed at engineers or

researchers who want to resolve more complex problems.
The third volume, entitled Heat Transfer 3, is focused on convection

transfers. As we have already pointed out, all of these transport operations
are rarely pure and lead to problems that involve three inter-connecting
transfer modes, conduction, convection and radiation.
Before our readers immerse themselves in a text that, despite our best
efforts, remains intellectually demanding, we propose a short text that is a
little lighter.
I.3. Interlude
Let us imagine, in a “B movie” context, a somber hostel in the gray fog of

a port in the middle of nowhere. Sailors from a faraway marina come and
drink away their troubles. And as always, the drink helping them along, they
turn to fighting.
Let us entrust Ludwig Boltzmann to direct the film. Our B movie heroes
are getting agitated, delivering blows to one another. Each one of them has
moderate kinetic energy, distributed heterogeneously among them in the
room. For some reason, they get involved in a general brawl. Their average
kinetic energy becomes much greater.
In everyday language, we would say that things are hotting up.
This would bring us right into line with a fundamental concept of

Boltzmann, who was the first to hypothesize that heat is made up of
molecular agitation. The temperature in a gas is proportional to the
average quadratic energy of the molecules that make it up:
E = 1 kT
C2
Using this model, we will return to the physical basis for all transport
phenomena.
On the way, we rarely escape from the explosion of a door or a window,

giving in under the repeated beatings of the brawlers.
Introduction xix
We have just modeled the pressure, due to the transfer of the quantity of
movement on the surface, by the impact of molecules.
Let us now imagine that the altercation is initially located in the corner of
the room: a smaller group starts fighting between themselves.
From kicks to punches, after multiple impacts within the group and its
immediate neighbors, the agitation will spread: we have just seen the
mechanism of heat propagation by transfer of impacts.
Let us place an imaginary separation (geometrically but immaterially

defined) at the center of the room. Let us count the sailors that cross through
it within a unit of time.
This wall is now crossed by kinetic energy: we have defined a flow of

heat.
Let us put a metal ring with a surface area of S = 1 m2 in the room. On

both sides of this ring, the blows exchanged constitute a transfer of kinetic
energy - we have just defined the heat flow density.
And we have just understood the nature of the propagation of thermal

flows by impacts.
Let us suppose that the great majority of the brawlers come from a ship
with a white uniform. Let us suppose that another boat in the port has
uniforms that are red. The red ones are initially all united. We will then
quickly see that the red mariners, as they receive and return blows, spread
out across the room. We have just shown the mechanism of diffusion of
matter, of a component within a mixture.
We will have a better qualitative understanding that the fundamental law

of conduction (Fourier Law) is formally identical to the law for the diffusion
of mass (Fick Law).
Let us put our agitated sailors in the compartments of a flatcar train,

where they continue to fight. And let us start the train moving. The kinetic
energy that they contain is transported from one point to another.
We have just invented thermal convection.

xx Heat Transfer 1
We can go further.
Let us imagine a series of flatcar trains on a set of parallel tracks. The

train furthest to the side is fixed to a platform. All of these trains are full of
sailors. Let us suppose that our train follows the outside, parallel rail tracks.
No brakes will prevent these trains from moving. Only the last train, at the
platform, is stuck.
For a reason we do not need to analyze (cinema allows all kinds of

fantasy), “clusters” of fighting sailors jump from one wagon to the next.
These “clusters” contain a component of speed that is parallel to the train,
which will communicate information about the quantity of movement to the
adjacent train. These trains will then start to move, more quickly the closer
they are to the outside train. And the same occurs up to the train at the
platform. This train will not move, but a force will be applied to its brakes.
We have just discovered the mechanism of dynamic viscosity. At the same

time, the parallel trains in relative movement give us a picture of the notion
of boundary layers.
At the same time, these agitated clusters carry their disordered kinetic
energy, “thermal” agitation. We have just seen the mechanism of the thermal
boundary layer.
Finally, let us include a few red mariners in the crowd of white. They will
be carried with the clusters, and we have just invented the limit layer of
diffusion of a species.
We are in a fantasy, and let us benefit from it as far as we can. To finish,

let us suppose that this is carnival day; each sailor has a belt equipped with
bells.
All the individuals have a different speed, and the impacts are random, all
the bells start to jingle, each with a different frequency. The distribution of
frequencies will depend on the statistical distribution of speeds
(Boltzmann statistics), and the intensity of noise produced will depend on
the total agitation energy of the sailors.
We have just understood the basic mechanism of radiation. We have

just realized why the theory of radiation needed to use the concepts of
Introduction xxi
statistics derived from the work of Boltzmann - a brilliant pupil ofPlanck -

to produce the emissions spectrum of a black body, for example.
NOTE.- the model is certainly simplistic. The emission comes from quantum
transitions in the gas atoms.
Here, we have already deviated from the pure substance of the book, but
we could go even further.
Let us suppose that our agitated sailors are in a room with one mobile
wall (a nightmare scenario frequently seen on the silver screen).
The incessant impacts of the fighters on this wall create a force that
pushes it. This force, reduced to a surface unit, explains the notion of
pressure.
By pushing against this wall, our crowd applies work that is greater than
the resistance.
Here we see an equivalence spring up between work and heat that, at a

fundamental level, are simply two mechanical energies: one ordered and the
other disordered. The first principle of thermodynamics is illustrated by
this.
We can see that the incidence of an average blow on the wall is rarely
normal.
Therefore, an average fighter will have a trajectory that will be reflected

off the wall. And only the normal component of its speed will be able to
push (or transfer work to) the wall.
Thus, we see that it will be impossible for the crowd (taken to mean a
gas) to give all its energy to a mobile wall.
The fundamental mechanism that leads to the second principle of

thermodynamics has just been demonstrated.
These “light-hearted” images, which will perhaps not please everyone,

were an oral support for the presentation of different transport phenomena by
one of the authors. We hope that the reader, once they have studied this
xxii Heat Transfer 1
book, will want to return to this text. They will then have understood, we
hope, the images that lead to the development of thermodynamics.
And if this text has a moral, it would be: Writing down

thermodynamics, just like thermal science, is based on continuous equations.
The fundamentals ofphysics that determine these phenomena arise from the
field of the discontinuous: discontinuity of matter, divided into particles;
discontinuity of light, divided into photons.
1
The Problem of Thermal

Conduction: General Comments
1.1. The fundamental problem of thermal conduction
The fundamental problem of thermal conduction involves the

determination of temperature domains and flows across particular surfaces,
for a given physical situation, in one or several given environments.
The resolution of a thermal conduction problem involves:

a ) for all problems, a heat equation is considered, resulting from a local
heat balance;
b ) for specific conditions of the problem, the conditions are constituted at
the limits that are applied to the heat equation.
In the most general case, these temperature domains T are not

homogeneous. They can be three-dimensional and variable over time:
T=T(x,y,z,t)
In other systems, we note that T = T (r,d,z,t) or T = T(r,в,Ф,t).
Temperature is expressed in degrees Celsius (°C), formerly degrees

centigrade, or in Kelvin (K).
We know that the temperature, expressed in Kelvin, is measured from

absolute zero. The conversion rule is known as: TK=TC+273.15 .
Heat Transfer 1: Conduction, First Edition. Michel Ledoux and Abdelkhalak El Hami.
© ISTE Ltd 2021. Published by ISTE Ltd and John Wiley & Sons, Inc.
2 Heat Transfer 1
The most general problem presented in thermal conduction is therefore

extremely complex, since the heat equation has four dimensions (three in
space and one in time).
Fortunately, significant simplifications are possible for many problems.
The temperature field can only depend on a spatial variable. We say that
conduction is monodimensional or bidimensional.
The temperature can remain fixed at each point as time progresses.
This last remark divides the approach that we will adopt into two parts:
- stationary conductive heat transfer;

- non-stationary conductive heat transfer.
Two important categories will be examined in this chapter:
- problems of stationary conduction with a single dimension: T= T(x)

or T= T(r)
- problems of non-stationary conduction with a single dimension:

T=T(x,t) or T=T(r,t)
For many problems, the analytical approach is possible. This will be the
case, in particular, for the stationary or non-stationary problems with a single
dimension. In (most) other cases, a numerical approach is required.
1.2. Definitions
1.2.1. Temperature, isothermal surface and gradient
The temperature T is a parameter defined in all thermodynamics classes.
As of now, we can define isothermal surfaces in space. An isothermal

surface is a surface on which the temperature is constant:
Isothermal surface S ^ T = Cte for an entire surface. [1.1]

The Problem of Thermal Conduction: General Comments 3
Furthermore, we will look at the important relationship between this

surface and heat flow.
Let us note that in a non-stationary problem, the isothermal surface is

mobile in time.
The temperature field also allows a very important vector to be defined,

whose use will be explained later on: temperature gradient.
Initially, we define this vector in a Cartesian system. Later on, we will

see that it is quite easy to determine its components in other systems when
the symmetries in a problem make that a pertinent option.
In a Cartesian system, this gradient is constructed by taking, for each of

its components, the partial derivative of the temperature with respect to the
relevant coordinate axis. In other words:
dT
dx
—— dT
grad T [1.2]
дУ
dT
dz
For other systems, it will be useful to refer to Appendix 2.
This mathematical operation is known to all physicists. With a minus

sign, it is the derivation of a force from a potential.
We should note that in a stationary problem, this gradient is a vector

attached to each point in space and fixed in time. In the case of a
non-stationary problem, this gradient is a vector attached to each point in
space and variable in time.
The temperature gradient possesses a fundamental property, which is

demonstrated in Appendix 2: the gradient vector is normal at each point in
space to the isothermal surface that passes through this point.
This property will be familiar to those who have studied electric fields.
4 Heat Transfer 1
1.2.2. Flow and density of flow
These two essential parameters will be used throughout this book.
Thermal flow Ф is the quantity of heat Q that passes a given surface in

a unit of time:
ф=Q [1.3]
This flow is expressed in Joules per second (J.s-1 ) or even better, in

Watts (W).
Density of the thermal flow ф, which can be even greater, is the flow
that passes through a unit of surface:
ф= Q
[1.4]
S tS
This flow density is expressed in Watt per m2 (W.m-2).
It is important not to confuse these two parameters, which also do not

have the same units.
In this book, we will systematically note the flow with a capital Ф and
the flow densities with a lowercase ф.
These flows and flow densities may not be uniform through a surface,
which can also vary rapidly over time. We then deal with a definition on a
differential basis, through small surface areas and with small time intervals.
Through a small surface d S :
d Ф( t )= dQ [1.5]
dt
с!Ф d 2 Q
Ф = ^ = ^^ [1.6]
dS dt dS
The quantity of heat d2Q is measured on a small surface d S and on a

small time interval d t, which becomes twice differential. The flow d Ф (t)
becomes differential since it is the relationship of a twice differential
quantity by a small surface d S at a given time t. The flow density ф is the
relationship between two differential quantities, which remains finite
(expressed in flow per unit of surface and time).
1.3. Relation to thermodynamics
We have already pointed out that when we talk about thermal science,
thermodynamics is not far off. More precisely, it is its main principle.
Before the first principle, there was calorimetry.
We should be aware that each time that we draw up a balance for an

environment, which will be particularly true in Chapter 4, we link a
variation in temperature to an exchange of heat by a calorimetric operation.
1.3.1. Calorimetry
Calorimetry was, in fact, born before the first principle. We (mainly

chemists) noted that the relative variations in temperature due to an
exchange of what we had then designated as “heat”, depended on the mass
m of the environments and a coefficient that is today denoted as the mass
calorific capacity c . In other words, when a body transfers heat to another
body, the relationship (with an obvious notation) is:
m1 c1 AT=-m2 c2 AT2 [1.7]
The minus sign indicates that one body is cooling and the other is
heating.
This leads to the relationship between the quantity of heat Q given to a

body and its rise in temperature Д T (or the inverse in the case of cooling),
which can be written as:
Q = mc A T [1.8]
6 Heat Transfer 1
1.3.2. The first principle
Later on, Joule will demonstrate the equivalence of two forms of energy:
work and heat.
His experiment can be reproduced on a daily basis. It is sufficient to turn

a small paint mixer in water contained in a Thermos flask. Why a Thermos?
So that the heat given to the water does not escape.
We measure the electrical energy provided to an engine for a certain time,

with energy measuring the work “given” to the water. The heating of the
water whose mass and calorific capacity are known measures the quantity of
heat brought to the water. This quantity of heat can only come from the
transformation of work provided through heat. In Joule’s era, heat was
measured in calories.
One calorie raised the temperature of 1 gram of water by 1 degree

Celsius. Work is evaluated in Joules. Modestly, this author only had a
kilogram meter.
Hence, two conclusions will always be in the background of our study:
- qualitatively: work and heat are two equivalent and interchangeable

forms of energy;
- quantitatively: the equivalence between a calorie and (today) a Joule is

written as:
1 calorie = 4.18 Joules [1.9]
In principle, this is part of the repertoire of all high school pupils, but it is
not a bad thing to give a reminder of it.
1.3.3. The second principle
This principle is about quality and reversibility. It is easy to transform

work into heat. Simple friction is sufficient. This is what Joule experienced.
It is more complicated to transform heat into work. Several sources are

required, at least one hot and one cold, and the temperature of the sources
becomes important. Moreover, we never recover all the heat that we provide
with the hot source in work. We always “throw” some into the cold source.
This will be the large Carnot work, among others.
Those who have closely read the text at the end of our preamble will
understand why. The blades of the mixer transform a “directed” energy into
a chaotic thermal agitation. Chaos will also never spontaneously return to
order.
All this comes from the genius (we defend our choice of word) of
Boltzmann, who will also be mentioned several times in the following
chapters.
To conclude, let us say that in all our transfers, we are dipping into the
realm of irreversibility and the permanent creation of entropy, but this will
be self-evident for the reader.
2
The Physics of Conduction
2.1. Introduction
As we have known from time immemorial, from a qualitative point of

view, heat only moves “spontaneously” from hot(ter) zones to cold(er)
zones. This is the phenomenon of conduction. In thermodynamics,
spontaneously means without the intervention of “work”.
Thermodynamics reveals that heat is made up of the kinetic energy of the

constituents of matter, molecules or atoms that move with a certain agitation.
The temperature has indeed been linked to the average kinetic energy of these
molecules by Boltzmann. The molecules are immobile at absolute zero,
which therefore becomes the minimum physically accessible temperature.
From a quantitative point of view, transfers by conduction are governed

by an apparently simple general rule: Fourier’s law.
2.2. Fourier’s law
2.2.1. Experiment
In order to set up this law, we imagine a simple device, which is

represented in Figure 2.1.
A material of thickness e and cross-section S is placed between two

isothermal plates. The temperatures T1 and T2 of the left-hand and
right-hand faces of the test material are measured using thermocouples, for
10 Heat Transfer 1
example. The left-hand plate is equipped with an electrical resistance that

V2 . , , , , .
delivers a power of P = —. The right-hand plate is cooled.
The whole apparatus is placed within a very thick layer of material that is
a very poor conductor of heat. Here, we make the hypothesis that the heat
dissipated by the electrical resistance passes all the way through the material
by thermal conduction.
The density of thermal flow is constant in the material, from left to right,
and is equal to:
P
Ф=- [2.1]
S
We therefore observe, after many tests (in particular, with variation of P

and e ), that for a given material, the flow density is related to the thickness
and temperature by:
? = ATTT
—T [2.2]
e
where A is a characteristic coefficient of the material being tested.
Surface S
Thermocouples
Figure 2.1. Diagram of a simplified device for determining Fourier’s law.

For a color version of this figure, see www.iste.co.uk/ledoux/heat1.zip
The Physics of Conduction 11
This coefficient, a property of the material, is known as the thermal

conductibility of the material. It is expressed in J s-1m-1K -1 or much more
frequently in W s-1K-1
In many situations, we can consider this conductibility to be constant.

The experiment demonstrates that this conductibility depends on the
temperature. In certain problems, in particular in the case of high
temperatures and heterogeneous thermal fields, we must take into account
the spatial variation of the thermal conductibility within the material, which
can be expressed explicitly as:
1 = 1( x, y, z) [2.3]
or more often by a relationship between the conductibility and the

temperature:
1=1(T) [2.4]
Materials with high values of 1 are known as conductors. Materials with

low values of 1 are known as insulators.
Here are a few values to set out the general idea:
Metals are conductors of heat:
1Copper =386 W s-1K-1
1Al = 204 W s-1K-1
Many solids (including refractory ones) are insulators:
For example, for a light resinous wood:
1 = 0.12 W s-1K-1
Liquids are relatively poor conductors of heat. For example, for water at
20°C:
1Water =0.597 W s-1K-1

12 Heat Transfer 1
Gases are poor conductors of heat. For example, for air at 20°C :
hair = 0.0257 Ws- K-
REMARK.- Air is thus one of the best thermal insulators. However, between
two faces, separated by a layer of air of thickness of the order of 2 cm,
natural convection phenomena leads to a “parasitic” thermal transfer. Our
solution is thus to trap air in alveolar structures, such as networks of mineral
threads (glass fiber or rock “wool”). The solid material then leads to parasitic
thermal conduction, known as a “thermal bridge”, which explains how rock
wool presents a thermal conductibility of the order of h = 0.4 Ws-1K-1,
higher than that of air. If the faces are made of glass (double glazing), the
distance between the two panes must be small.
We can also observe that strictly speaking, a vacuum is an absolute

insulator, since there are no molecules that are likely to propagate impacts.
We will see later on however, that a vacuum is ideal for the propagation of
radiation and in that sense does not present an obstacle to the transmission of
heat. This explains the structure of a thermos flask, whose internal walls,
positioned face to face, are made of reflective glass.
The metrology described here is used in practice. It turns out to be

difficult to use, insofar as any three-dimensionality of heat flows is a source
of error. It therefore requires the use of a material that is a better insulator
than the one being tested, or failing this, very large volumes.
Other “dynamic” methods using non-stationary thermal transfer

properties (impulse response methods) have been developed.
2.2.2. Temperature profile
T1- T2
The law ф = h------ - implies that, in the case of the experiment, in other
words, a mono-dimensional case, the profile of speed T(x) is linear. Here,

we have defined x as the x-axis of a plane that is parallel to the walls of the
material (plates) located on an axis Ox, which is perpendicular to the
heating and cooling plates. T(x) is then the temperature of a plane of the
material, perpendicular to Ox .
T1- T2
Indeed, the law ф = Л------- is valid for all distances e between two
planes perpendicular to Ox . This remains true for two planes that are
extremely close together. e= dx is then an infinitely small distance (as
small as we want), and the difference in temperature T1- T2 becomes a very
small difference dT. Fourier’s law is then written, in differential form, as:
dT
ф = л---- [2.5]
dx
e =dx is as small as we want, and ^S— is then the derivative of T (x)

dx
with respect to x .
ф is constant, so the function T(x) is linear.
Figure 2.2. Linear temperature profile in a homogeneous material. For

a color version of this figure, see www.iste.co.uk/ledoux/heat1.zip
14 Heat Transfer 1
2.2.3. General expression of the Fourier law
The above explanations are sufficient for mono-dimensional problems.

The general heat equation implies extending this law to the
three-dimensional case. To do so, we will use the following reasoning. A
reminder of this calculation will be given in the Appendix.
Figure 2.3. Case of two infinitely close isotherms; relationship with the gradient.
Let us consider two isothermal surfaces with respective temperatures T

and T+ dT, positioned very close to each other. In other words, M , a point
on isotherm T . We will now consider two infinitely small surfaces dS1 and
dS2 , neighbors on each of the isotherms, at point M . They have practically
the same normal n, directed conventionally from the surface with
temperature T towards the surface of T+ dT. Locally, we observe the
situation found in a plane mono-dimensional system.
The thermal flow that traverses dS1 and dS2 will be expressed as:
dT
d Ф = - 2---- d S1 [2.6]
dx
where x is an axis counted on the axis for n. Let us recall that the minus
sign indicates that heat propagates in the direction of decreasing
temperatures. In this expression, the sign implies a positive flow when it

flows in the direction of the normal.
Let us consider a surface dS' based on the same section of cylinder as

d S1 and d S2, whose normal n' makes an angle в with the normal n that
is common to dS1 and dS2 .
The same thermal heat flow evidently traverses dS', dS1 and dS2
We have a known geometrical relationship:
dS1 = dS 'cose [2.7]
And the traversing flow dS' is written as:
dT dT
d Ф = - A---- dS1 ==-A----- dS cos в [2.8]
dx dx
Near M , the temperature gradient is reduced to a component along the

axis of x :
(grad -) = A^T [2.9]
which we can re-write as:
grad- T = Ad- [2.10]

dx
We note that the normal for a unit vector is one:
IIй 1=1 [2.11]
In fact, the flow will take the form of a scalar product:
d Ф= - A^dS 'cos в = \grad —1| In' cos в d S' [2.12]
dФ = ||grac? —||||n *|| cos (grad -, n') d S' [2.13]

16 Heat Transfer 1
Taking into account the orientations of the normal described above, we

obtain the general form of the Fourier law:
d Ф = - grad T. n' d S' [2.14]
REMARK.- Thus, the thermal flow that traverses a closed surface will be
written as:
Ф= jjAgradT.n dS [2.15]
S
2.3. Heat equation
2.3.1. General problem
All thermal conduction problems go through several stages, for any

selected approach:
a ) setting out a thermal balance. This balance can be written explicitly for
each problem, and this will be the case, in particular, in Chapter 4 concerning
quasi-stationary regimes.
We can also use an assessment written once and for all for a small region:
this will be the heat equation;
b ) writing special conditions for the problem, which will constitute the
conditions at the limits of the equation. This point will be examined in
greater detail in section 2.4, which follows.
This paragraph is dedicated to establishing the heat equation.
To describe the most general problem, this equation is a partial

differential equation whose solution is the local and instantaneous
temperature, in other words a variable with four dimensions: three spatial
variables and one time variable. We could, depending on the symmetries of
the problem in question, work in different coordinate systems. Throughout
the problems posed in this book, we will frequently work in Cartesian
coordinates (problems known as “plane problems”), in cylindrical
coordinates, or more rarely, in spherical coordinates.
The temperature will therefore be expressed in different ways:
- Cartesian coordinates:
T= T(x,y,z,t) [2.16a]
- cylindrical coordinates:
T = T (r ,0, z, t) [2.16b]
- spherical coordinates:
T = T (r ,0, Ф, t) [2.16c]
A general problem of this kind is expressed in four dimensions; it is one

of the most complex problems to deal with.
Often the device in question has symmetries that allow the problem to be
reduced to one spatial dimension. The temperature now only depends on two
variables, space and time. We can distinguish between:
- plane problems:
T= T(x,t) [2.17a]
- cylindrical symmetry:
T= T(r, t) [2.17b]
- spherical symmetry:
T= T(r, t) [2.17c]
In these cases, the heat equation, becoming an equation of two

dimensions, will take on a simplified form.
The problems dealt with in this guide will almost systematically be

related to this single-dimensional hypothesis.
18 Heat Transfer 1
For this reason, we will initially establish the heat equation in two
dimensions. This construction will be based on a direct expression of the
thermal balance.
As a matter of general culture, we consider it important for thermal

scientists to know the full equation, which is often a starting point for
various scientific publications that they will need to consult over the course
of their practice.
For this purpose, we will establish the heat equation in four dimensions,
using the tools of vector geometry. For readers who are not accustomed to
these methods and to lighten the explanations a little, we have put this
establishment in the Appendix, simply giving its result below.
2.3.2. Mono-dimensional plane problem
2.З.2.1. The terms of conduction per m2 of wall

The problem is mono-dimensional: T only depends on x . Therefore,
any plane perpendicular to Ox is the isotherm T= T(x).
In other words, an axis Ox which is used as a reference framework

for the direction of heat propagation; this axis is orientated from left to
right in Figure 2.4.
We write the instantaneous heat balance for a given volume determined
by two planes normal to Ox , infinite neighbors of axes x and x + ds and
surface S with a unit surface area for simplification purposes: The flows that
traverse it are therefore equal to the densities of flow ф.
x + dx
Figure 2.4. Establishment of the mono-dimensional heat equation

In other words, the axis plane x , perpendicular to Ox ; let us consider the

quantity of heat that traverses it per m2, or density of flow ф:
Let us write
dT
ф( x ) = - A— [2.18]
dx
What is the role of the minus sign in this expression?

dT dT
—< 0 when the heat moves from left to right, and is positive
л----
dx dx
when the heat moves in the direction of increasing x .
dT dT
— > 0 when the heat moves from right to left, and is negative
Л----
dx dx
when the heat moves in the direction of increasing x .
dT
The expression - A— is written at point x of the axis Ox . How is
dx
dT
- Л— written when we go from the x -axis plane to the x + dx -axis plane,
dx
with dx infinitely small?
General case of any function f (x): when x goes from x to x + dx , a

function f (x) goes from f (x) to f (x) + ^fdx.
dx
dT
ф( x) = - Ad— is a specific f (x) function.
dT
Therefore, when we go from x to x + dx, -A— becomes
dx
ф( x) + d^xd=_ AdT+dL -AdT dx [2.19]
dx dx dx dx
20 Heat Transfer 1
Let us now consider a plane located at x + dx and the quantity of heat

that traverses it per m2.
We write:
dT d dT
^( x + dx ) = - A----- I----- ----- A---- dx [2.20]
dx dx dx
Using the same reasoning made for x,
ddT d ( .TT 1
- A----- 1---- I -A----- I dx is positive when the heat moves in the direction
dx dx I dx )
of increasing x .
And - A^-+— I -A^-1 dx is negative when the heat moves in the

dx dx I dx j
opposite direction to increasing x .
2.3.2.2. Balance of conduction terms

Let us now consider the thin section of unit surface S, of thickness dx ,
limited by two planes located respectively in x and x + dx .
We will set up the balance of heat entering and exiting the section by
conduction.
When we set up this balance, we do not know1 what the direction of the
variation of T(x) is.
Ox is orientated from left to right, x <x+dx
On the left-hand plane, if the heat goes from left to right, it is entering;
entering heat must be counted positively in the balance.
1 Since we are in the process of establishing the differential equation that will allow the
function T(x) itself to be determined!
dT
Therefore, this term will be written - 2— per m2 because it is positive
dx
if the heat moves in the direction of x (from left to right)
Through the plane x + dx, the entering heat must also be counted
positively in the balance.
Yet -2^—+— I -2^— I dx is positive when the heat moves in the

dx dx I dx j
direction of x, in other words if the heat is exiting the section.
Therefore, this is the opposite term, which will intervene in the balance,
in other words
d dT d L, dT L
2 +1 2I dx [2.21]
dx dx I dx j
REMARK.- The manipulation of the signs in front of the terms can surprise
or worry certain readers. It is fundamental to understand what is happening.
dT
It must be understood that the negative sign in the expression -2----
dx
indicates that the heat moves physically from hot to cold and that the flow is
positive if it moves in the direction of Ox , whereas the sign used in writing
a balance indicates whether the heat enters into the volume (necessarily a
positive term) or exits the volume (necessarily a negative term). Readers are
invited to take the time that is required to understand this correctly. This will
be of great benefit.
The overall heat balance in the small volume will therefore be:
ddT ddT d (d dT V d (d dT V
-2----- +- 2----- +---- 1 2 I dx = — I 2-----I dx [2.22]
dx dx dx ^ dx j dx ^ dx j
22 Heat Transfer 1
2.3.2.3. Variation of the temperature with time

At this stage, we note that thermal conductibility is associated with the
first derivative. The hypothesis of invariable conductibility has not been
made.
The term — I A^-T I dx is positive if more heat enters than exits. The
dx I dx )
small volume heats up.
The term — I AT I dx is negative if more heat enters than exits. The

dx I dx j
small volume cools down.
— I A^- I dx is exchanged per unit of time. During the time dt, the
dx I dx j
quantity of heat exchanged with the outside is d I A^- I dxdt. It therefore
dx I dx j
modifies the temperature of the small volume of dT .
By simply applying the principles of calorimetry, we can write that a

change of temperature dT of the small volume is due to an exchange of
heat dQ equal to:
dQ = pcSdxdT [2.23]
Obviously,
d b
dQ = — ---- I1,
I AdT dx dt [2.24]
dx I dx j
and we have finished our balance:
dQ = pcSdxdT = dI A— Idxdt [2.25a]

dx I dx )
and we should recall:
S=1 [2.25b]
This gives the single-dimensional expression for the heat equation:
d T d (^ dT
Pc—= d x I d x [2.26]
dx
Let us note that we are now using d, because T = T (x, t) and the
derivatives are partial.
Two specific cases will intervene in our applications:
a) often the thermal conductibility will be a constant:
d T d2 T
---- = a----- [2.27]
dx d x2
This is where thermal diffusivity will come in:
2
a =---- [2.28]
Pc
b) another important case is stationary thermal phenomena. As much heat

enters the volume as exits it. Conduction is known as stationary, and the
equation becomes:
if the thermal conductibility remains variable:
± (2dT 1 = 0
[2.29]
dx ^ dx )
We note that T=T(x) and that we have therefore moved to the straight
'<” ’ that replaces the “ d ”, in particular partial derivatives.
24 Heat Transfer 1
and for a constant thermal conductibility:
д T д2 T
----= a----- [2.30]
д x д x2
This equation is a specific case of the complete equation as it is

established in the Appendix:
dT A div gradrr
= ---- T [2.31a]
д t Pc
or, for constant physical properties:
d T = a Дat
---- T [2.31b]
дt
2.3.3. Case of the axisymmetric system
Here, we give the forms of the heat equation in axisymmetric coordinates

in a mono-dimensional case. To lighten the text, we have included details of
the calculation in the Appendix, which is based on an approach that is
analogous to the previous one.
In this case, the axis becomes an Oz axis and the heat (therefore the
variation in temperature) follows the vector radius r; positive orientation of
each radius from the cylinder axis to the outside of the cylinder. The
same reasoning given previously leads to the following conclusions.
In other words, a single-dimensional field. The temperature depends only

on r, the distance of a point from the axis. Isotherms are therefore
cylinders with radius r: T=T(r)
We no longer think on the basis of a m2 of surface area but instead in

terms of a length of 1 m of cylinder.
д T 1 d {, д т
Pc a7= -----
r дr l
1 Ar —
дr
[2.32]
Figure 2.5. Establishment of the single-dimensional

heat equation Axisymmetric geometry.
With the two specific cases:

- constant physical properties:
d T a д ( дT
---- =-------| r— [2.33]
дt r дr l дr
- stationary conduction:
d ( dT^ л
— [2.34]
drIl r dr)I = 0
We note that r is maintained under the first derivative, which is

characteristic of equations of cylindrical symmetry.
2.3.4. Case of the spherical system
As previously seen, here, we will provide the forms of the heat equation
in an axisymmetric coordinate system in a mono-dimensional case. To
26 Heat Transfer 1
lighten the text, we have included the details of the calculation in the
Appendix, based on the same approach as the one used above.
„ д T 1 д Г 2д T
p c---- =------- I A r ---- [2.35]
д t r2 д r l дr
Here, r is the distance from a point at the origin of the spherical

coordinate system.
With the two specific cases:

- constant physical properties:
д T a д ( 7d T
=I r 2 [2.36]
д t r2 д r l д r
- stationary conduction:
d ( dTT
--- I r 2---- =0 [2.37]
drl dr
We note that r2 is maintained under the first derivative, which is

characteristic of equations with spherical symmetry.
2.4. Resolution of a problem
Heat conduction leads to a very wide range of problems, in terms of their

type, structure and difficulty. The problems can be mono-dimensional; the
temperature then depends on a single coordinate. In Cartesian coordinates,
we have a plane model. In cylindrical coordinates, we have a problem with
axial symmetry. These two types will often be encountered in the examples
that readers will need to process in this book.
In the general case, the problem is bi- or tri-dimensional.
The problems can also be dependent on time, in which case, they are
non-stationary. It is often the case that temperature is independent of time,
we then have a stationary problem.
We can see that there will be significant variation in the difficulty of

resolving these problems, from a mono-dimensional stationary problem,
where T= T(x) or T= T(r) to a three-dimensional non-stationary
problem, where, for example, T = T(x,y,z,t) or T = T(r,0,z,t). In the
first case, we will have analytical solutions that are part of the fundamental
culture of a thermal scientist. In the second case, a numerical approach is
often required.
Regardless of its type or difficulty, a thermal conduction problem comes

down to two things:
A space is defined, in general, in a “region” D defined by a closed

surface area S . Through this surface S , heat can be exchanged with space
outside the domain. Here, several modes of heat transfer can play a role
(conduction, convection, radiation).
The objective of a conduction problem is more often to find the heat

flows exchanged between the region and the space outside of it. For this
purpose, the temperature distribution near the interface must be known.
Beyond this, for example, for reasons of resistance of materials, we are
seeking the internal temperature distribution of the region D .
In all cases, and no matter how sophisticated the problem is, there will be
two stages in resolving the problem:
1) determination of the temperature domain, generally T= T(x,y,z,t)
in the region D ;
2) determination of the flows at the interfaces, in general, from local
temperature gradients grad T (x,y,z,t) at these interfaces.
Mathematically, the problem will be posed as the resolution of a heat

equation, in a more or less complete form, associated with limit conditions.
The heat equation can be set up by dividing the region into an infinite
number of small elements and by writing a heat balance for each of these
small elements. We will proceed in this way in the explanations given
immediately below. It will then be useful to set up this balance in two simple
cases, in the plane problem and in the axisymmetric problem.
28 Heat Transfer 1
A different method, faster and more elegant, consists of using the general
methods of differential geometry, which is associated with the previous
procedure in a way that is not immediately obvious.
Let us note that this heat equation in the most general case will include
time dependence. In addition, if necessary, it should take into account the
heterogeneity of the material in the region D .
The conditions at the limits are given by the physical constraints

imposed on the region D from the outside. These constraints determine
several types of problem, which are most often described as:
- distribution of temperature TW imposed on the surface S.
Mathematically, we have what is known as a Dirichlet problem;
- distribution of the thermal flow density ф imposed on the face. This
problem can be summarized as imposing the value of the gradient in the
material of D on the surface S. Mathematically, we obtain a Neumann
problem.
Data concerning relationships between temperature and thermal flow

density ф on the surface S .
This is particularly the case when the thermal exchange occurs with the
space outside the region by conduction, convection or radiation. In this way,
we can obtain various examples of the forms seen, since Te is a reference
temperature (distant temperature, external temperature of another material,
atmospheric temperature outside D , etc.):
,d T WT -eTe
T
Ф = ЛТ~ [2.38]
dn S R,uth
d> t h(TW - [2.39]

v = л^~
dn S
d> —
ф = Л- t
= £&\ [2.40]
dn S
or a combination of these different expressions. In the presence of

convection associated with radiation, we will have, for example:
дТт = h(Tw - Te)+ еа( TW - те4) [2.41]

ф = л^~
дn
where h is the convection coefficient, Rth is the thermal resistance and

еа is the product of the Stefan constant and emissivity.
Coupled transfers must also be envisaged:
д> т Tw-T + h(Tw - Te)+ ev(TW - Te4) [2.42]

ф= л^~
dn S Rth
We will see further examples of this later on. Some will be processed in
the section of the book on radiation, in particular, in Chapter 5.
2.5. Examples of application
NOTE.- The following examples have been placed at this point in this guide,
in line with the logical progression of the book and its presentations.
We have just examined the various ways of writing the heat equation.
The problems in this section are direct applications and solutions of this
equation. We show, in particular, the effect of a variation of the thermal
conductibility with temperature, the effects of the internal generation of heat,
and we examine some problems in spherical symmetry.
However, we no longer comply with a logical progression of

explanations, which would require us to propose increasingly difficult
exercises to readers. Indeed, the problems in this chapter can appear, for a
reader with little experience of the field of differential equations, more
difficult than the exercises in Chapter 3, for example.
In this case, readers can go on without prejudice to the exercises in the

next few chapters, and when they have acquired a little more agility in
resolving the other questions, and have finished reading, they can return to
the examples in this chapter.
30 Heat Transfer 1
EXAMPLE 2.1.- Face with variable conductivity
A plane face with thickness e is made up of a homogeneous material

whose thermal conductivity can be represented by:
2 = 20 (1 + aT) [2.43]
where 20 is the thermal conductivity at T = 0 °C . T is therefore expressed

in Celsius.
The faces are subject to temperatures T1 and T2 .

1) Write out the equation which will allow us to determine the
distribution of temperature T=T(x) on the face.
2) How does the temperature vary as a function of x ? Determine the

flow density ф which traverses the face. Deduce from this, the thermal
resistance Rth of this face.
3) Is the flow lower or higher than what would have been calculated with
2= 20 ?
We will use the following data:
T1 = 35°C ; T2 = 20°C ; a = 5.10-3 C-1; 20 = 0.03kcal.hr-1Ce = 20cm ;
Solution.-
1) To solve this problem, we must solve the mono-dimensional heat

equation in the stationary regime, written taking into account the variation of
the thermal conductibility with temperature.
This equation is written as:
—2—T = 0 [2.44]
dx dx
In this equation:
A = A(T) = A0 (1 + aT) [2.45]
and T= T(x) is the unknown.
This second-order linear equation requires two limit conditions, which

are given by temperatures on the two sides of the face. Choosing the origin
on the left-hand face:
x =0 ; T=T1 [2.46a]
x =e; T=T2 [2.46b]
2) The differential equation can be re-written as:
— A0 (1 + aT)—T = 0 [2.47]
dx dx
The constant A0 can be eliminated.
The equation will be solved by integrating twice by x:
— (1 + aT )—T = 0 [2.48]
dx dx
The first integration leads to a first constant C1 :
(1 + aT )—T = C [2.49]
dx
We note that (1 + aT)^— T can be written as the derivative of a

dx
function:
d dT dT d T2 d ( T2A
(d 1 + aTd)— dT
T =—+ aT— = — + — a— =—I T + a— I [2.50]
dx dx dx dx dx 2 dx I 2 )
32 Heat Transfer 1
The equation can be re-written as:
t2 1 ,
d I T + a— I = C,dx [2.51]
I 2 J 1
which is integrated into:
(_ T2 1 _
I T + a— I = C x + C7 [2.52]
I 2 J 1 2
And determined by the limit conditions:
( T2 1
x=0 ; T = T1 therefore I T1 + a— I = C1 *0 + C2 [2.53]
( T2 1
x=e ; T = T2 therefore I T2 + a— I = C1 * e + C2 [2.54]
It results that:
( T2 1
C7 = I T + a — I [2.55]
2 I 1 2 J
1 \( T2 1 ( T2 1
C, = — I T2 + a — I — I T + a — I [2.56]
1 e LL 2 2 JI 1 2 JJ
The numerical values of C1 and C2 are:
C2 = 38.06 [2.57a]
C1 = — 85.3 [2.57b]
To calculate the flow densities, it is not necessary to express T (x)

explicitly.
In fact, the first integration has given us:
(1 + аТ) d-T = C1 [2.58]
which is the flow density at the nearest 20.
Therefore,
ф = 2 (1 + ат )dx
—T = C1 2 [2.59]
ф = 85.3*0.0348 = 2.97 W. m-2 [2.60]
The thermal resistance of the wall can be deduced by:
т-т
Rth = T-T [2.61]
Rh = — = 5.05 m2 KW [2.62]
2.97
3) To compare the real flows to those calculated with 2= 20 , two

methods can be used.
a) Compare the flow density found in 2 directly to the flow density
ф0, which traversed a wall of thermal conductibility 20.
т-т
% = 2,-1---- 2 [2.63]
e
In an SI coordinate system:
20 = 0.03 kcal.hr .K = 0.03*4180 = 0.0348w.m-.K- [2.64]

0 3600
^0 = 2.61W
34 Heat Transfer 1
The real flow is higher than this value. The real wall is less insulating
than the homogeneous wall.
We note that we can write
T-T e
?с = [, with Rth0 = V [2.65]
Rth 0 Л)
$0 = 0.0348 02 = 2.61W.m-2 [2.66]
b) Compare the thermal resistance found in 2 to Rth0 , thermal

resistance of a wall of thermal conductibility 20
Rh 0 = el. = 0.2 = 5.75 m 2. K W-1 [2.67]

th0 20 0.0348
The resistance of the real wall Rth = 5.05 m2 KW 1 comes out lower
than this value. We confirm that the real wall is less insulating than the
homogeneous wall.
EXAMPLE 2.2.- Problems with generation of internal heat. Evacuation of

heat in a bar of uranium
A bar of uranium has a diameter D = 29 mm .
The nuclear reactions that take place here produce a volumetric power q
expressed in qW.m-3 .
The thermal conductivity of uranium is 2= 27W.m-1.K-1.

1) Determine the distribution in the bar under the stationary regime. At
the external surface, the temperature is Te =200 °C. What is the maximum
temperature Tmax obtained? Where do we observe this?
2) Uranium melts at Tf =1132 °C. Determine the maximum volumetric

power qmax that can be extracted from the bar if we do not want to make the
uranium melt in the center of the power plant, which is always best to
avoid?
Solution.—
Here, we must resolve the heat equation, written in cylindrical

coordinates with a volumetric production term.
The problem has rotational symmetry; therefore, T=T(r)
2 d ( dT^
----- 1 r-----1 + q = 0 [2.68a]
r dr ^ dr )
We know one limit condition.
In principle, this is insufficient, but we will see another piece of

information come to light later on.
R = — = 14.5 mm [2.68b]
2
r=R; T=Te [2.69]
The equation can be solved by integrating twice:
d ( dT^ -qr
— 1 r---- 1 = ------ [2.70]
dr ^ dr ) 2
dT -qr2
r = + C1 [2.71]
dr 22
dT - qr C1
[2.72]
dr 22 r
T (r )= -qr- + CLnr + C [2.73]

4 4 42 1 2
The temperature must remain finite everywhere in the tube, which is not
the case in the center,
36 Heat Transfer 1
unless we use
C1 = 0 [2.74]
C2 is then easily deduced from the external temperature of the tube:
-qR 2
Te= +C2 [2.75]
42
qR 2
2e [2.76]
42
Finally, we can put the temperature in the form:
qR 2 2
T(r)-Te = 1 RJ [2.77]
42
We obtain a parabolic temperature profile. The maximum temperature

Tmax is obtained for r = 0 , in other words in the center of the tube. The
maximum temperature reached is:
qR2
t t t + q— [2.78]
max e 42
2) If we do not want the temperature of the bar to exceed 1132° C , we

find the maximum value of the volumetric power to be:
T = T + qmax R2 [2.79]
42
In other words
42(TF- T )
qmax [2.80]
4*27*(1132-200)
qmax ----- —i------- ------ -’ = 4.79.108 W.m-3 [2.81]
(14.5.10-3)2
EXAMPLE 2.3.- Maximum dissipation of an electrical resistance
An electrical resistance takes the form of a metal tube with diameter

d= 2 cm surrounded by an insulating duct of thickness e= 5 mm. The
thermal conductibility of this duct is 2 = 0.06 m-1 K-1.
In other words, hi = 5 W m-2K-1 the convection coefficient outside the

system.
We do not want the metal part of the resistance to exceed Tmax =150 °C.
The atmospheric temperature is Te =20 °C.
What power can dissipate this resistance per meter?
Solution.-
Let us calculate the resultant thermal resistance of the insulation and

convection:
d
=—
- the internal radius of the insulator is ri2
- the external radius is re=ri+e
Ln I - ----- I T (0.01 + 0.005

Ln I
I ri ) . 1 I 0.01 1
Rth +----------------- [2.82]
2 n2 2 n (r + e) h 2 n 0.06 2 n *0.015*5
=1.075+2.12=3.197mKW-1 [2.83]
In other words, the flow per meter that leads to a temperature

Tmax =150°C:
Д T 150 - 20
Ф =----=---------- = 40.66 W [2.84]
Rh 3.197
38 Heat Transfer 1
EXAMPLE 2.4.- Temperature in steel wire, which an electric current is

passed through.
We consider a steel wire with a diameter d= 2 mm. Its thermal

conductibility is 2 = 54 m-1 K-1. An electric current is passed through it,
which dissipates a power per ml of wire by the Joule effect P= 100 Wm-1 .
This power is dissipated homogeneously in the mass of the wire. In other
words, q, the resulting volumetric power, is expressed in Wm-3 .
The external temperature of the wire is maintained at Te =60°C.
1) What is the value of q?

2) Write the balance equation in a small circular volume of unit length
and thickness dr, included between cylinders of radius r and r+ dr.
3) Solve this equation and give the expression for T(r).
4) What is the maximum value of temperature Tmax in the wire. Where is

it located?
Solution.-
1) The volume of a meter of wire is:
VOL = 1*ndL = n4410 6 = 3.14.10-6 m3 [2.85]
and is therefore equal to:
q = — = 1°^- = 3.18.106 Wm~3 [2.86]

VOL 3.14.10-6
2) We write a classic heat balance for L= 1 m of wire
dT dT d dT7dT \
—2n 2— —2n 2----- l —2n 2 — I + 2nr dr q = 0 [2.87]
dr dr dr dr )
Dividing the two terms by 2n2
1 ddT q
r + =0 [2.88]
r dr dr X
1 ddT q
----- r---- + —= 0 [2.89]
rdr dr X
associated with a limit condition that we have imposed at the surface of the
wire:
r=R; T=TW [2.90]
3) Resolution of this equation is of a type that is often found physically:
ddT qr
— r----=------ [2.91]
dr dr X
Integrating the two terms
dT qr2
r — = --—+ C1 [2.92]
dr 2X
dT _ qr C1
[2.93]
dr 22 r
T = - qr2 + C Lnr + C [2.94]

42 1 2
T must remain finite at the center of the wire, therefore: C1 = 0
At the surface,
r= R, we find T= Tw [2.95]
Therefore:
[2.96]
40 Heat Transfer 1
Finally:
T - T = qR- 1 - (r 1
[2.97]
w 4Л LRJ
4) We see that the temperature will be a maximum for x = 0 , in other

words, the center of the wire:
For r = 0
Tmax -Tw =qR- [2.98]

4Л
= qR- = 3.18.106 *10-^14710-2 c

Tmax -Tw ^= ^= .. [2.99]
4Л 4*54
Hence, the maximum temperature at the center:
Tmax -Tw =60+1.47.10-2 = 61.47°C [2.100]
2.5.1. Problems involving spherical symmetry
EXAMPLE 2.5.- Thermal resistance of two concentric spheres
We consider two concentric spheres with radius r1 and r2 , maintained at

constant temperatures T1 and T2 .
The solid environment that separates them is homogeneous, isotropic and

presents a thermal conductivity Л that is independent of the temperature.
1) Calculate the thermal power dissipated in the stationary regime.
2) What is the thermal resistance of the system?
Solution.-
1) The thermal power sought is equal to the incoming flow in the sphere
of radius r1 . This flow is equal to the flow traversing through each
intermediate sphere between r1 and r2 . It is also equal to the flow that

traverses the sphere of radius r2 .
For a value of r between r1 and r2 , the general expression of this flow is:
dT
Ф = 4п Ar2---- [2.101]
dr
And in particular, at the interfaces of the region contained between the

two spheres of radius r1 and r2 :
dT
Ф = -4nAr 2---- = -4nAr 2 dT [2.102]
dr dr
The distribution of T(r) must therefore be determined in advance.
It must be deduced by writing a thermal balance that will lead to a

differential equation.
By writing that the incoming and outgoing flows of a thin spherical

section of thickness dr are equal, we obtain:
. dT 2 . dT 2 d (. 2 dT ^ d (. 2 dT ^
-A—4nr2 + A—4nr2 + — \ A 4nr2 — I dr = — \ A4nr2 — I = 0 [2.103]
dr dr dx ^ dr J dr ^ dr J
In other words, for a homogeneous body of constant thermal

conductibility:
d ( 2 dT
-^\ r -г =0 [2.104]
dr ^ dr
With the limit conditions
r=r1; T=T1 [2.105a]
r=r2; T=T2 [2.105b]

42 Heat Transfer 1
We note the term included below the derivative symbol. With this written
expression, we take into account the variation of the spherical surface
traversed by the flow when r varies from r to r+ dr.
Resolution of this equation is obtained by integrating twice
d I 2 dT 1 [2.106]
dr 1^ r ----
— dr 1)= 0
2 dT r [2.107]
r T~ = C
dr
dT _C1
[2.108]
dr r2
C
T (r ) = - C1 + C2 [2.109]
r
The constants will be determined by the limit conditions
C
r=r ; T = T1 ; T =---- + C2 [2.110]
r1
C
r = r2 ; T = T2 ; T2 =-------- -- + C2 [2.111]
r 2
Subtracting the two expressions:
T1 - T2 = - C1 I - - —1 [2.112a]
1 ri r2 )
c1 =-TT
1 -T\ [2.112b]
1 I111
I r1 r2 )
We note that if T1> T2, T1-T2>0 and C1 is negative.

C2 can then be written in two different ways:
C
C2 = T + -L [2.113]
r1
C
—2 = T2 + _L [2.114]
r2
The flow in r1 or in r2 is written:
dr dr
Ф = - 4 nAr 2— r=r = -4 nAr 2 — 1 r=r1
[2.115]
dT d f- 1
dr 1 dr
Ф = -4nAr 2---- r=r = - 4nAr2---- I----L + C2 I r-r
I r 2) r=
-
= -4nAr 2— = -4 nAr 2 —1 [2.116]
r2 r=r1
dT -
Ф = -4 n Ar 2---- = = - 4nAr2r—
r=r2 2 r=rr2 = - 4n Ar2 —1
= [2.117]
dr
We verify that the two values are indeed identical
Ф= 4 nAr2 —1 = 4 nAr 2 [2.118]

f1 11
I-------- I
I r1 r2 )
2) By definition of the thermal resistance Rth per m2:
T-T
ф = 4 nr 2 T---- 2 [2.119]
Rth
44 Heat Transfer 1
By identification:
Rh =LT-1 [2.120]
^r1 r2
EXAMPLE 2.6.- Conduction in a sphere: Whittaker model
Whittaker proposes the following general expression for the Nusselt

number, which allows the thermal transfer between a full isothermal sphere
of radius R and diameter D, at a constant temperature Tw and an
atmosphere at temperature Te :
1 2
Nu=2+ 0.4 Re + 0.06 R] [2.121]
where Re is the Reynolds number based on the diameter of the sphere and
Д is the dynamic viscosity.
This equation will be valid for 3.5 <Re <80 000 and 0.7 <Pr <380
Justify term 2 in this equation.
Solution.-
The second term gives the value of the Nusselt number when the
Reynolds numbers are zero. This corresponds to the absence of flow around
the sphere. Therefore, there is no longer any convective transport and term 2
corresponds to a purely conductive transfer, which in all cases coexists with
convection when the latter is present.
Remark.- We will see in the following that this transfer by conduction, in

this spherical symmetry system, will be stationary. This would not be the
case with a plane face; a non-stationary transfer would then be necessary.
We calculate the exchange, by conduction, between a full sphere of radius

R at constant temperature Tw and an atmosphere at temperature Te .
We are in a single-dimensional stationary regime: T=T(r). The

isotherms are spheres with radius r.
We will now write the balance for the space found between two spheres
of radii r and r+ dr:
d dT
Л—nr 2---- = 0 [2.122]
dr dr
which is reduced to the differential equation:
d dT
л—nr 2---- = 0 [2.123]
dr dr
with the limit conditions
r=R; T=TW [2.124a]
r ^ж ; T^T [2.124b]
We will consider the new function:
0( r )= T - Te [2.125]
Defining:
00n = we
T -T [2.126]
The system to be resolved is transformed into:
d dT
--- r 2---- = 0 [2.127]
dr dr
with the conditions:
r =R ; 0 = 00 [2.128a]
r^^ ; 0^0 [2.128b]

46 Heat Transfer 1
Integrating the equation twice, two constants C1 and C2 appear:
rd 00 = Cr
1---- [2.129]
dr
1 0
d 1 _ C
~11 [2.130]
dr r 1
-C
0( r ) = ——1 + C2 [2.131]
r
Applying the limit conditions:
r =R ; =—
00,0R 1 + C22 [2.132]
r^^ ; 0 = C2 [2.133]
which gives:
C1=-R00 [2.134]
0, r )=R0 [2.135]
r
Therefore, the thermal flow density for the sphere:
dTT .00 .- R0o ,R0o 20

m=-X =- 2— =- 2-------- 0 = 2 —0 = — [2.136]
dr r=R dr r=R r1 r=R R1R o
which gives a convection coefficient
h= ^ = ^ = 2 = 22 [2.137]
Ta - T~ 0o~ R~ D
We deduce the classic expression for the Nusselt number from this:
Nu=D [2.138]
Nu = 2 [2.139]
This Nusselt number is only valid for an exclusively conductive transfer.

In general, convective terms are added to it, which take the form of
polynomials with powers of the Reynolds numbers of the sphere.
EXAMPLE 2.7.- Evaporation of a drop: Godsave model
A spherical drop of liquid, radius R and diameter D , isotherm at a

temperature Tw , is immobile in an atmosphere at temperature Te , at which it
vaporizes.
In other words, pL is the density of the liquid, LV is the heat of

vaporization of the mass and D is the thermal conductibility of air.
Te>TL>Ta, where TV is the vaporization temperature of the liquid
(saturating vapor temperature at atmospheric pressure).
We suppose that the surface of the drop remains equal to TV throughout

the vaporization.
1) Considering that all of the heat brought to the drop is used to vaporize
the liquid at the surface, write the variation in time of the mass of the drop.
2) Deduce from this the differential equation that the diameter of the drop
obeys.
3) Show that the reduction of the diameter of the drop is linear:
D(t)=D0 (t)-Kt
where D0 is the initial diameter of the drop and K is a constant for which
the expression will be given.
Deduce the vaporization time TV for the drop.

48 Heat Transfer 1
We will use a quasi-stationary calculation, supposing that, even for slow

changes to the diameter, the Nusselt number taken from the Whittaker
equation remains valid:
2
Nu = 2 + 0.4 Re + 0.06 Re p0.4 I Ue [2.140]

r I
I Uw
This equation would be valid for 3.5 <Re <80 000 and 0,7 <Pr <380
Solution.—
1) The thermal flow ФW provided to the drop will, at instant t, have the
form:
ФW = S (t) h(Ta - Tv ) [2.141]
where S(t) is the surface area of the surface of the drop:
S(t)= 4nR2 = 2nD2 [2.142]
This flow, used to vaporize the liquid, allows a mass mv to be

transformed, per unit of time, into vapor:
m v = Pl ф w Lv [2.143]
mv = PLLV 2nD2h(T - Tv ) [2.144]
The instantaneous mass of the drop is m(t):
4 3 D3
m (t )= —nR =n— [2.145]
36
If the drop is immobile, the Whittaker equation is reduced to
Nu = 2 [2.146]
This result, which can be deduced directly (see previous exer cise), results
from a written theory in a stationary regime. When R (t) varies slowly with
temperature, we suppose that this Nusselt value remains valid.
Here, we will continue the development, maintaining the n otation Nu ,

even if we know the value for an immobile fluid. The explanation for this
will be given at the end of the solution.
h will be deduced from the Nusselt definition:
Nu=D [2.147]
N Nu D
h = —— [2.148]
D
Finally:
ND
mv = PlLv 2nD*~D~(Ta - Tv ) = Pl Lv 2nDNu D(Ta - Tv) [2.149]
The balance for the mass of the drop is therefore written as:
dm d nD3
[2.150]
mv = dt ~ Pl dt 6
D* dD
vLL =- P
2 dt [2.151]
2) By deduction, combining the two expressions for mv:
tm
v = PlLv 2nDNu D(Ta TV )= Pl 2 dt [2.152]
I)!17 = -Lv 4 П Nu D Ta - Tv ) [2.153]

50 Heat Transfer 1
That we can re-write, by making the time derivative for D2 appear:
1 dD2
--------- = -L 4n N Л(T - T) [2.154]
2 dt Vr u \ a V)
where:
dD. = _ Lv 8 n Nu Л(Ta - Tv ) [2.155]
In other words:
dD = К
k [2.156]
dt
Writing:
K = Lv 8П Nu Л(Ta - T )
3) The linear equation can easily be resolved into
D2 =-Kt+C [2.157]
with the initial condition:
t=0;D(0)=D0 [2.158]
Hence the law:
D2 =D02 -Kt [2.159]
The vaporization time TV is reached when the diameter of the drop is

zero:
0 = D02 - К TV [2.160]
TV = D7 [2.161]
K
with
K = Lv 8П Nu Л(Ta - Tv ) [2.162]
This law is known under the name of the law expressed in D2. We
effectively note that the vaporization time is proportional to the square of the
initial diameter and proportional to the temperature differential between
atmosphere and saturation point.
When Nu = 2, we have:
K =16n LV Л(Ta - Tv ) [2.163]
D02
T [2.164]
V 16П Lv Л(Ta - Tv )
REMARK.- We have kept Nu until the end of the calculation because this
model can still be applied in the presence of wind blowing across the drop.
In this case, use the form proposed by Whittaker:
i 2
Nu=2+ 0.4 Re+0.06 Re p0.4 I Re [2.165]
r II Rw
We see that in this way, a vaporization drop continues to satisfy a law for
D 2. Convection obviously has the effect of increasing the constant K ,
therefore reducing the vaporization time. This result can be applied in
various practical situations: reactors in chemical engineering, agriculture
(spraying of pesticides), injection of petrol or diesel into a cylinder, rocket
engines, etc.
3
Conduction in a Stationary Regime
3.1. Thermal resistance
This practical concept is applicable to plane and cylindrical geometries. We

will demonstrate that, depending on the composition or geometry of an
insulating assembly, several definitions for thermal resistance will be
generated, and we must be careful not to confuse them.
3.1.1. Thermal resistance: plane geometry
In some practical calculations, we obtain the notion of thermal resistance

using the “simplified” model of a plane wall.
It is important, at this point, to define the resistance that we apply more

carefully.
In fact, there are two thermal resistances that are very different:
a) the first is defined per unit of area of a wall. We will refer to it as
“thermal resistance”, which is denoted by Rth . We will use this almost
systematically in this chapter. We prefer to use this term because for a
homogeneous wall, this resistance is also a characteristic of the material it is
made of. It is generally related to the calculation of flow density.
b) the second is defined for a given surface of the wall. We will refer to it
as “global resistance”, which is denoted by RG .
In general, it is related to the calculation of a flow.
54 Heat Transfer 1
З.1.1.1. Resistance per unit of surface area (per m2)

We know that the density of thermal flow ф that passes through a plane
wall of thickness e made of a material with thermal conductibility 2, to
which we apply a temperature difference T1- T2, is given by:
[3.1]
e
If S is the surface area of the wall, the flow Ф that traverses it is then
T-T
ф = S ф = 2ST---- T— [3.2]
e
We can rewrite the thermal flow density as:
T1-T2 =Zr-Tt
ф= -1------- 2 [3.3]
e Rth
2
We have therefore defined the thermal resistance per unit of surface area,
Rh, or thermal resistance per m2:
e
Rth = [3.4]
2
We note that for a wall made of a unique homogeneous material, this

thermal resistance thus defined is a characteristic of the material. Moreover,
this characteristic is often used for commercial purposes.
With the same idea in mind, we also define a conductance U:
U=R1h ■ 2 [3.5]
Conductance can also be used as commercial product data. For the sake
of simplicity, we will not use this parameter in examples proposed in this
book:
Conduction in a Stationary Regime 55
e -1
Rth = — is expressed in m2KW [3.6a]
U = — is expressed in W m 2 K 1 [3.6b]
e
3.1.1.2. Overall resistance

Here, we start with the expression for a flow Ф that traverses an area S
of a plane wall with thickness e that is made of a material with thermal
conductibility 2, to which a temperature difference T1 - T2 is applied, which
is given by:
T-T
ф = S у = 2ST-----T [3.7]
e
which can be rewritten as:
ф = T -T2 = Ti -T2 [3.8]

e Rg
2S
We have thus defined the overall thermal resistance, RG , or the thermal

resistance per m2:
RG = 2S [3.9]
We note that this resistance can no longer be characteristic of a material,

but depends on the surface area of the wall.
Similarly, we can also define a conductance UG :
U = R- = — [3.10]
56 Heat Transfer 1
Rth = 22 is expressed in m2 KW 1 [3.11a]
U = — is expressed in W mz K 1 [3.11b]
e
3.1.1.3. Practical application of thermal resistance

The concept of thermal resistance becomes useful when considering
practical cases of walls made up of superimposed materials of different types
and thicknesses.
We then distinguish between walls made up of superimposed layers and

those made up of composite materials.
a) Plane walls of superimposed layers: addition of thermal

resistances
In this common case, the wall, for which we want to determine the
insulation performance, is made up of layers. We can still reason in terms of
density of thermal flow, because this is the same density that passes through
all materials.
The notion of thermal resistance per unit of surface area will therefore be
given preference. We can therefore obviously define a thermal resistance
e
Rh = 2 [ ]
for each component of the wall and a “resultant” thermal resistance Rth of
the wall as a whole.
We will demonstrate this in an example.
EXAMPLE 3.1.- Resultant thermal resistance of a plane wall
What is the resultant thermal resistance of a plane wall composed of three

layers of materials with respective thermal conductibilities 2 , 22 and 2,,
and respective thicknesses e1 , e2 and e3 ?
Solution.—
We can set out the thermal resistance (per unit of surface area) for each
of the layers:
Rthi = eY [3.13a]
4
e
Rh 2 = ± [3.13b]
Z
e
Rh 3 = ± [3.13c]
Z3
We will denote the extreme temperatures applied to the two faces of the
wall as T1 and T4 , and the temperatures of the intermediate interfaces
between the layers of the wall as T2 and T3 .
The thermal flow density T1 can be written for each of the materials as:
71 - 72
9= [3.14]
Rh 1
72 — 73 [3.15]
9=
Rth 2
73 — 74 [3.16]
9=
Rth3
We express the differences in temperature and add the three expressions

obtained:
71-72= ^Rth1 [3.17]
72-73= VRth 2 [3.18]
73 -T4 = yRth3 [3.19]

58 Heat Transfer 1
(T1 - T2) + (T2 - T3) + (T3 - T4 )= m( Rth 1 + Rth 2 + Rth 3) [3.20]
T1-T4 T-T
m =------------- 1--------- 4----------- = —---------- 4 [3.21]
Rth1+Rth2+Rth3Rth
Finally, we obtain:
T1-T4 T-T
m =------------- 1--------- 4----------- = —---------- 4 [3.22]
Rth1+Rth2+Rth3Rth
which leads us to the relationship between the overall thermal resistance Rth
and the thermal resistances of each wall:
Rth = Rth1+Rth2+Rth3 [3.23]
We also note here a thermal analogy with Ohm’s law, assimilating the
density of thermal flow to an intensity and the temperatures to potentials.
b) Composite wall
In this case, the wall is no longer made up of layers that are simply
superimposed: it is heterogeneous. A problem of this kind, tackled from a
rigorous point of view, involves resorting to a numerical method. We can
find an approximate solution (more or less) to this problem using the notion
of overall thermal resistance.
As an example, let us consider the case of a wall made up of a concrete

base with an upper part made up of bricks, covered by two layers of
insulating material.
EXAMPLE 3.2.- Composite wall
More precisely, a wall of thickness e2 is made up of a concrete base with

thermal conductibility XC and a surface area 51, and a face (of the same
thickness e2) with thermal conductibility XB and a surface area 52. Each of
the two faces of the entire wall is covered with a layer of insulation of
thermal conductibility XI and thickness e1 .
We apply a temperature difference T1- T2 between the two external faces

of this composite wall. What is the flow passing through the wall?
Solution.—
We divide the composite wall into two plane walls made up of plane
layers:
- for the first wall, the concrete assembly and the two layers of insulation;
- for the second wall, the brick wall and its two layers of insulation.
Each of these walls can be treated using an analogous approach to that

previously applied for the plane multilayer wall; but this time by writing the
expression for the two flows Ф1 and Ф2 passing through the concrete and
the brick respectively, in terms of overall thermal resistances. This is made
possible by the surface area that the three layers have in common.
We can define various overall thermal resistances for each layer of each
of these multilayer walls.
For the base wall, with a surface area S1
concrete:
ei [3.24]
RG =
XcS1
each layer of insulation:
ei [3.25]
RGI =
XiS1
For the brick wall, with a surface area S2
brick:
e1 [3.26]
G
RGB =
XbS2
60 Heat Transfer 1
each layer of insulation:
R = 1 e [3.27]
GI2 0 c
AS
I 2
For each of the two sections of the composite wall, we therefore find:
RG1= RGI1+RGC +RGI1 [3.28]
RGI 2=RGI 2+RGB +RGI 2 [3.29]
The two flows Ф1 and Ф 2 can be written as:
1 -2T
ф = JT 2 T-T
2 2________
= ________ J1
[3.30]
RG1RGI1+RGC+RGI1
ф = JT T2
1 -2 =_______ J1T-2T2________
[3.31]
RG2RGI1+RGB+RGI1
The overall flow Ф traversing the complete wall is then given by:
T-T T-T T-T

ф=ф1 +Ф 2 - +-1—2-_-1—2-
_-1—2 [3.32]
12RGG1 RG2RG
1G2G
A simple calculation shows that:
R RG1 +Rg 2
[3.33]
G~ RG1RG2
3.1.1.4. Electrical analogy

We note that RG1 and RG2 are related to the overall resistances that make
them up, like resistances placed in series. Furthermore, the two expressions
for Ф1 and Ф2 are related by Ohm’s law analogy to temperatures.
№3
i■ 1I 2 Ii 3 ■3
Figure 3.1 . Thermal resistance. Electrical analogy for three superimposed walls.
We can then imagine an equivalent electrical diagram, which may induce

more complex reasoning when the number of layers in the wall increases.
REMARK.- This method is an approximation. We effectively neglect the fact

that the temperature profiles are different in the two superimposed wall systems
1 and 2. The approximation becomes rough if the thermal conductivities of
the materials in these two walls become very different.
Figure 3.2 . Thermal resistance. Electrical analogy for a composite wall.

62 Heat Transfer 1
3.1.2. Thermal resistance: axisymmetric geometry. The case of a

cylindrical wall
This problem corresponds to the common practical case of a cylindrical

duct surrounding a tube that needs to be thermally insulated.
We have a ring-shaped duct of length L with an internal and external

radius, made of a material with thermal conductibility 2. We denote Tt as
the temperature of the internal face with radius ri , and Te as the temperature
of the external face with radius re .
The problem is a little more complex than the previous one. The
single-dimensional heat equation must be solved: we work in cylindrical
coordinates (r, в, z) and the temperatures vary only with the radius r. The
equation was given in Chapter 2, section 2.2.2:
d b
— dT I =I 0n
I Ar---- [3.34]
dr I dr )
and for a homogeneous material (which does not vary):
d ( dT I л [3.35]
— I r---- I = 0
dr I dr )
This second-order equation must be solved with two limit conditions. The
temperatures on the two internal and external faces of the ring shape are
given by:
r=ri; T=Ti [3.36a]
r=r; T=T
ee [3.36b]
We will integrate the differential equation twice in succession, where two

constants C1 and C2 will emerge:
first integration:
dT r
r—=C [3.37]
dr
dT C1
So — = — [3.38]
dr r
second integration:
T(r)=C1Lnr+C2 [3.39]
The two conditions at the limits will give the value of the constants:
Ti=C1Lnri+C2 [3.40]
Te=C1Lnre+C2 [3.41]
T-T T-T
=-i----rr
C11 e- = —i-----e- [3.42]
Ln — Ln —
reri
The value of C2 is not very important because to obtain the flow density,
we derive the flow per meter of length and the thermal resistance:
Ti- Te1
^ = - ^dT =—
C1
^ = - A----- - [3.43]
dr r r r
Ln—
ri
As we note, the flow density is not the same on both faces (the flow is the
same and the lateral surface is different):
internal face:
Ti- Te1
Pi = - e [3.44]
rr
Lnr ri
ri
external face:
T
i- e T 1
pe = - ^^—e~ [3.45]
rr
Ln r re
ri
64 Heat Transfer 1
The flow on a unit sleeve length will give (fortunately) the same result on
both faces:
T-T1 T-T
Ф i = 2п- yi
iii = 2i
nrik—rr = 2nk—r [3.46]
Ln- -i Ln—
rr
T-T1 T-T
= 2nr yi
Ф eei = 2nre
k—----
-- -— = 2nk—----
- - [3.47]
Ln— -e Ln —
We therefore introduce here a thermal resistance “per unit length”, a

concept that obviously must not be confused with the definitions in plane
problems:
T-T T-T T-T

=ф = 2 nk----r - = ir
Ф = Ф ir ir
[3.48]
rR
Ln — Ln— th /m
ri ri
2 nk
r
Ln —
ri
R„,
th/m = [3.49]
2nk
This expression must be remembered while remaining fully aware of the

definition of this resistance. We note that its dimension is the inverse of
thermal conductibility, which is expressed in mKW-1 .
For a homogeneous sleeve of insulation of length L, in the presence of

pure conduction, we end up with a flow given by:
T-T
Ф= L —---- r [3.50]
Rth/m
3.1.3. Thermal resistance to convection
In practice, we note a temperature imposed on the face of a wall less

often than forced or natural convection. For a plane problem, we can express
the density of thermal flow using an expression of the form:
Vw = h (Tw - T) [3.51]
where index w denotes the face and index e an “atmospheric” temperature,

which is often known as a “distant” temperature in theories of forced
convection.
For a cylindrical problem, the radius r of the exchange surface must be

taken into account. We therefore work in flow Фw/m per unit length (and not
in flow density, per m2):
Ф w / m = 2 nrh (Tw - Te) [3.52]
3.1.3.1. Resistance to convection in plane geometry

We can rewrite Vw in the form:
T-TT-T
Vw = h (Tw - Te ) = TVT = TT—- [3.53]
1 RthC
h
We then note a thermal resistance to convection that is given by (per m2):
Rhe = h [3-54]
Obviously, this resistance will be additive to the resistances of each layer

of a plane wall.
3.1.3.2. Resistance to convection in axisymmetric geometry

We can rewrite Vw in the form:
T-T T-T
Фw/m = 2nrh (Tw -Te) = 1 = -^ [3.55]
1 RthCL
2 nrh
66 Heat Transfer 1
we have thus defined a thermal resistance per meter of ring:
1
RthCL = [3.56]
2 nr h
We note that the flow ФwL that escapes by convection (or that enters or
exits from the inside) from a length L of tube will thus be expressed by
T-T
ФwL = L -^-^
R
[3.57]
A calculation with an analogous form to the plane case will lead to an

additivity of these resistances to transfer, both conductive and convective:
LU] ex I
1 +у +—L_
I rint J
Rth/m = [3.58]
2 nri ht f-gs 2 nX 2 nrehe
In this expression, we take into account the fact that the internal and
external convection coefficients can differ (possibly significantly if we have
a transfer to a liquid on the inside and to the air on the outside).
REMARK.- in the above formula, it goes without saying that, in the sum
of all the rings, we attribute the value of internal and external radii of each
element to re and ri (see examples for a better understanding).
Thus, a flow that passes through a multilayer sleeve of length L , with

internal and external convection, will be expressed by:
T-T
Ф = L —---- e- [3.59]
Rth/m
This will obviously be the same on the internal and external faces.
3.1.4. Critical radius
Let us consider a cylinder of radius ri surrounded by an insulating sleeve.

We denote 2 as the conductibility of the sleeve insulation and r its external
radius. The part of the sleeve with this radius is subject to the convection of
coefficient h .
In other words, Rth is the thermal resistance of the sleeve, resulting from
the resistance of the insulation and the resistance of convection.
We would naturally think that increasing radius r would lead to an

increase in insulation. However, the lateral surface of the assembly is also a
possible factor of an increase in losses. To study the competition between
these two phenomena, we will study the sign of the derivative with respect to
r for the resistance Rth .
We know that Rth has the expression for the case described here:
T fr1
Ln\ - I
r1
Rthth/. m =—— +--------- [3.60]
2nA 2nrh
and
T ' r1
_ h / m = JL
.d R . LnУ r-J I + 1 _ 1________ 1
dr dr 2nA 2nrh = 2nAr 2nr2 h
[3.61]
_
= 1 Г_1__ 1_’
2 nr A rh
68 Heat Transfer 1
In terms of signs:
1 1 1
sign = sign
2п2 2nrh 2 nr 2 ~~h
[3.62]
2
r-
rh-2 ___ h_
= sign Mi = sign 2r
by introducing the “critical” radius rC = —, we note that:
r- rC
sign
2п2 2nrh = gn-jrC = sign (r -rC ) [3.63]
In conclusion, we note that:
- when r < rC, the derivative th îh is negative. Increasing r leads to a

dr
reduction in Rth . Increasing the thickness of insulation leads to an
increase in losses. This is only paradoxical if we forget that it is in
competition with the increase in the lateral surface area;
- when r > rC , the derivative th îs is positive. Increasing r leads to an
dr
increase in Rth . Increasing the thickness of insulation leads to a
reduction in losses. The insulation takes precedence over the increase in
lateral surface area.
This phenomenon is important in practice when we set up the external
diameter of an insulating sleeve.
3.2. Examples of the application of thermal resistance in plane

geometry
EXAMPLE 3.3.- Heating an igloo
A polar explorer is invited by the Inuit into their igloo. The explorer, a
trained physicist then ponders the thermal behavior of this type of building.
REMARK.- The Inuit is the name given to those commonly referred to as

Eskimos. This translates simply to ‘people’.
The igloo is built of ice. The thickness of the walls is e= 40 cm. The
total surface area of the walls of the igloo is 5 = 25 m2
The external temperature is Te =-40°C.
We will take into account a convection coefficient h for both inside and
outside the igloo.
Figure 3.3. An igloo (source: chalettime.com). For a color

version of this figure, see www.iste.co.uk/ledoux/heat1.zip
We want to maintain an internal temperature equal to Ti =8 °C.
Remark.- The Inuit coat themselves in seal fat for better thermal protection.
They also sleep fully clothed. Polar exploration means having to change
Western habits.
70 Heat Transfer 1
What thermal power P must the wood fire inside the igloo produce?
We will make the following approximations:
- all calculations assume that the walls have a practically plane geometry;
- losses through the ground (ice) and the smoke hole will be neglected
(the Inuit do not like to suffocate);
- the thermal conductibility of the ice is: AG = 2.1 W. m 1. K 1;
- the convection coefficient (both internal and external) is:
h= 10W.m-2.K-1.
Solution.-
We will evaluate the thermal resistance per m2 of wall. This includes a

component due to the material and a component due to convection.
R = - + - + - = 0,4 + — = 0.39 Km2W [3.64]

th A h h 2.1 10
The energetic power of the fire, that is, the energy that it liberates per unit
of time, must compensate for the thermal losses, meaning the thermal flow
resulting from the thermal resistance that has previously been calculated.
Thus:
25 8+40
P =Ф =S T -T [3.65]
Rth 0.39
P=3077W [3.66]
Example 3.4.- Thermal resistance of a wall
1) Calculate the thermal resistance per m2 of a 70 cm-thick stone wall.
2) Calculate the thermal resistance per m2 of a 70 cm-thick stone wall,

insulated by a 25 cm-thick layer of rock wool.
3) We impose a temperature difference of 11°C between the two faces;

calculate the thermal flow traversing through 85 m2 of uninsulated wall and
insulated wall.
We give:
for the stone:
e =70.10-2m
2 = 3 W. m -2. K 1
for the rock wool:
e =25.10-2m
2LV = 0.045 W.m-2.K-1
Solution.—
1) We must first calculate the overall thermal resistance Rth of the wall.
The wall is simply made up of a layer of stone. The thermal resistance of

e
each layer is calculated by the equation Rth = ^, which is applicable to a
homogeneous plane wall.
The resistance to transfer of each convection coefficient will be: Rth conv = —
h
For the stone:
e = 70.10-2 m
2=3W.m-2.K-1
Rwaii = 2 = 0.233 K. m W-1 [3.67]

72 Heat Transfer 1
2) For the rock wool:
e = 25.10-2 m
ALV = 0.045 W.m-2.K-1
= - = 5.55 K.mW [3.68]

A
Using the additive nature of the resistances to the transfer, the resulting
thermal resistance is therefore:
Rth =0.233+5.55=5.78K.m2.W -1 [3.69]
3) The flow of outgoing heat for a surface area S of the wall can be
calculated by:
T-T
ф = S ф = S-----e- [3.70]
Rth
Ф= 85^^ = 161.8 W [3.71]

5.78
EXAMPLE 3.5.- Great Wall of China
The Great Wall of China has, on average, a height of 7 m and a width

of 5 m.
Figure 3.4. The Great Wall of China. For a color version

of this figure, see www.iste.co.uk/ledoux/heat1.zip
We will consider it to be constructed entirely from stone, with thermal

conductivity 2 = 3 W. m 1. K 1.
1) What is the thermal resistance per m2*if we do not take the wind into
account?
2) Rework the calculation, considering the wind on each side to be
represented by a convection coefficient h= 16 W.m-2.K-1.
3) What is the interest of a calculation of this type, in your opinion?
Solution.—
1) By definition of the thermal resistance of a plane wall:
Rh = 4 = 5 = L67W4 m2°K [3.72]

23
2) By definition of the thermal resistance:
Rth = - + - = 1.67 + — = 1.79 W m2°K [3.73]

th 4h 16
3) If there is one, for learning purposes, it is to show that a stone wall is a

very bad insulator.
EXAMPLE 3.6.- Study of losses by conduction through double glazing
Double glazing is made up of two glass panes separated by a layer of

immobile dry air.
The thickness of each pane of the glass is 3.5 mm and the layer of air is
12 mm thick. The thermal conductivity of the glass is 0.7 W.m-1.K -1and for
the air, it is 0.024W.m-1.K-1.
1) For a drop in temperature A T = 5 °C between the two outside faces of

the double glazing, calculate the thermal losses for a pane of dimension
1 m2.
74 Heat Transfer 1
For this evaluation, we neglect the effect of the convection coefficient on

either side of each pane of glass.
2) Compare these thermal losses to those obtained with a single pane of

thickness 3.5 mm.
Solution.—
Stationary problem: Thermal balances
1) We will calculate the thermal resistances of the panes
The thermal resistance of a plane wall is defined by
Rth = J [3.74]
For window B, we associate two types of face, two panes, of thermal

resistance RthvV and a layer of air with thermal resistance Rthair .
That is, for the glass:
e = 3.5.10-3 m
J= 0.7W.m-2.K -1
R = - = 5.10-3 K.m2W [3.75]

J
and for the air:
- =1.2.10-2 m
J= 0.024W.m-2.K -1
Rar = - = 0.5 K. m 2.W-1 [3.76]

J
Applying the additive nature of the resistances to transfer, the thermal

resistance of the window is given by:
Rth =(2*5.10-3)+0,5=0.51K.m2.W-1 [3.77]
The thermal losses per m2, that is, the density of flow of losses, is given
by:
ДT
Ю =----- [3.78]
Rh
ДT = 5C [3.79]
ю = — = 9.8 W.m-2 [3.80]

0.51
2) A single-glazed pane would lead to a loss per square meter of:
ю =— = —^ = 1000 W. m-2
[3.81]
Rthv 5.10—3
The advantage of double glazing is obvious here. However, we note that

the thickness of the glass chosen for this example is relatively weak from a
practical point of view.
EXAMPLE 3.7.- Study of the thermal quality of a window
To construct a French window, we consider three glazing options:

A) A 6 mm-thick single-glazed window;
B) A double-glazed window made up of a layer of air of 1.5 cm,
bordered by two 6 mm-thick glass panes;
C) A triple-glazed window, made up of three 6 mm-thick glass
panes separating two layers of air, each with a thickness of 1.5 cm.
76 Heat Transfer 1
Figure 3.5. Different types of glazing
1) Calculate the thermal resistances for 1 m2 of each of these types of

window.
2) Compare these thermal resistances with those for a 20 cm-thick
concrete wall (D) insulated by a 15 cm-thick layer of rock wool and limited
by a layer of BA13 (135 mm-thick plasterboard).
3) Framing of the French window
The triple glazing C is framed in a French window with dimensions given

in the following diagram. The frame is made of 4 cm-thick, medium-weight
deciduous wood. The dimensions in the diagram are given in cm.
«—90-►
Figure 3.6. Dimensional drawing of the French window
The French window is installed in a house. The external temperature is

Te =5 °C and the internal temperature is Ti =20 °C.
Calculate the thermal flow Ф that passes all the way through the French
window.
This produces the following thermal conductibilities:
Jair = 0.026W.m-1.K -1 Jconcr-t- = 2W.m-1.K -
Jglass =1.05W.m-1.K -1 JBA13 = 0.25W.m-1.K-1
JLV = 0.041W.m-1.K -1 Jwood =0.23W.m-1.K-
Solution.—
1) Calculation of thermal resistances
The thermal resistance of a plane wall is defined by:
Rth = J [3.82]
For window A, we have a single wall with
e= 6.10-3 m
J= 1.05W.m-2.K -1
RM = - = 5.71.10-3 K.m2W-1 [3.83]

J
For window B, we associate two types of wall.
-
Two glass panes A, with the thermal resistance RhA = j , and a layer of
air with thermal resistance Rthair :

78 Heat Transfer 1
e =1.5.10-2 m
2 = 0.026 W. m -2. K 1
Rr = - = 0.577K.m2W [3.84]
2
Applying the additive nature of the thermal resistances “in series”:
RthB=2*RthA+Rthair=0.588K.m2.W-1 [3.85]
For window C, we associate two types of wall.
Three glass panes A, with thermal resistance RthA , and two layers of air
with thermal resistance Rthair :
RthA = 5.71.10-3K.m2.W-1 [3.86]
Rhar = - = 0.577 K. m 2.W-1 [3.87]

2
By applying the additive nature of the thermal resistances “in series”:
RthC=3*RthA+2*Rthair=1.171K.m2.W-1 [3.88]
2) Concrete wall
For the concrete wall, we associate three types of wall.
-
A concrete wall, with thermal resistance Rthconcrete = -j, a wall of rock
wool with thermal resistance RthLV , and a wall of BA13 plasterboard with
the thermal resistance:
RBA !3 = 2 [3.89]
For the concrete:
e = 20.10-2 m
2 = 2 W. m -2. K 1
Rhconcrete = = 0.1 K.ШW’* [3.90]

2
For the rock wool:
e =15.10-2m
2=0.041W.m-2.K-1
RthLV = e = 3.66K.m2W [3.91]

2
For the BA13:
e =1.3.10-2m
2= 0.25W.m-2.K -1
R№ai3 = 2 = 5.2.10-2 K.m2.W-1
By applying the additive nature of the thermal resistances “in series”:
RthB=Rthconcrete+RthLV+RthBA13=0.1+3.66+5.2.10-2=3.81K.m2.W-1 [3.92]
3) French window
This flow will be equal to the sum of the flows traversing the wood and
the window.
The total surface area of the door is:
S=2.1*0.9=1.89m2 [3.93]
80 Heat Transfer 1
The surface area of the glass SV through which the flow passes is:
SV = (2.1 - 0.6)* (0.9 - 0.2 ) = 1.05 m2 [3.94]
The surface area of wood SB through which the flow passes is equal to:
SB =1.89-1.05=0.84m2 [3.95]
The density of flow pBois passing through the wood will be:
Pwood = R' [3.96]

thwood
with A T = 20 - 5 = 15 °C
and
e=4.10-2m
2 = 0.23 W. m K 1
R = | = 0.174 K. m 2.W-1 [3.97]

2
Thus:
15
^wood = 0T77 = 0.8621W. m ’2 [3.98]
.
The density of flow фV passing through the glass will be:
Vgla.. = P-99]
thglass
with
AT=20-5=15°C [3.100]
and (window C):
R,hgiass = 1.171 K. m W " [3.101]
Thus:
15
<P^ = — = 12.81 W. m [3.102]
.
The total flow will be:
Ф= Sb Vwood + Sv Vglass [3.103]
ф = ( 0.84* 86.2) +(1.05 *12.81) = 85.86 W [3.104]
EXAMPLE 3.8.- Thermal balance of an uninsulated roof
A slate has a thickness of 3 mm.

1) What is its thermal resistance per m2?
2) This slate is mounted on battens, meaning that it is placed on a plank
of resinous wood with thickness e = 15 mm.
Slate
Batten
Figure 3.7. Slate mounted on a batten
What is the thermal resistance per m2 of the slate-batten assembly?

3) A roof is made up of two panes of 20*7 m2. This roof is made up of
battens covered with slates over an attic that is not thermally insulated.
The temperature of the upper face of the slates is Te =30°C, and the
temperature of the attic is Ti =10°C.
82 Heat Transfer 1
Figure 3.8. Dimensional drawing of the roof
What is the flow of heat Ф passing through the roof?
The thermal conductibility of the slate is Aslate = 2.2 W. m "1. K 1
The thermal conductibility of the batten is Abatt = 0.12 W. m 1.K 1
Solution.—
1) The calculation is classic
eslate =3.10-3m
Aslate =2.2W.m-2.K-1
Relate = e = 1.363.10-3 K.mW-1 [3.105]

A
2) Slate on battens
For window B, we associate two types of wall.
The slate, with thermal resistance Rtha = 1.363.10-3 K.m2.W 1 and the
batten with thermal resistance Rthvol :
e =1.5.10-2 m
A= 0.12W.m-2.K -1
Rhbat, = J = 0.125 K.m W" [3.106]
By applying the additivity of the thermal resistances “in series”:
RhB = Rhsate + Rhbat = 0.1264K. m 2W—1 [3.107]
3) Thermal leaks through the roof
The surface area S through which the heat passes is:
S=2*(20*7)=280m2 [3.108]
The flow density is
AT
?b = T~
[3.109]
thB
A T = 30 -10 = 20 °C
rnK = 20 = 158.23 W. m z [3.110]

B 0.1264
The total flux is:
Ф = SVb [3.111]
Ф = 280*158.23 = 44304 W [3.112]
EXAMPLE 3.9.- Heating a detached house
A detached house has dimensions of10 m x 7 m on the ground and a

height of H= 3 m.
Its walls are made of agglomerated concrete with a thickness e= 20 cm;

they are insulated by 15 cm of rock wool.
84 Heat Transfer 1
The internal hi and external H= 3 m convection coefficients are both

equal to hi=he=10W.m-2.K-1.
The internal temperature is Ti =21°C, and the external temperature is

Te = 3°C.
We consider the roof to be very well insulated.

1) What is the total thermal resistance per m2?
2) What is the total flow passing through the walls?
3) What is the power of the radiator heating this detached house?
Solution.—
1) For the concrete:
e = 20.10-2 m
2 = 1.4 W. m K '
Rconcree = 2 = 0.143 K.m2.W-1 [3.113]
For the rock wool:
e =15.10-2m
2LV = 0.045 W.m-2.K-1
RLV = -e = 3.33K.m2.W [3.114]
Using the additive nature of transfer resistances, the resulting thermal

resistance is therefore:
Rth = R ■ + Rlv + 1- + 1- = 0.143 + 3.33 + -2 = 3.67 K. m W-1 [3.115]

hehi 10
2) The surface area of the walls is:
S = 2*17*3 = 102 m2 [3.116]
The flow of the outgoing heat for a surface area S of the wall can be
calculated by:
T-T
Ф= S-i---- e- [3.117]
Rth
21 - 3
Ф= 102----- = 500.3 W [3.118]
3.67
3) The heating power required to maintain the temperature will be

equal to the quantity of heat to introduce into the detached house in units of
seconds, in order to compensate for the leaks. It will therefore be equal to the
flow of the losses:
P = ф= 500.3 W [3.119]
3.3. Examples of the application of the thermal resistance in

cylindrical geometry
EXAMPLE 3.10.- Insulating sleeve
1) What is the thermal resistance per meter, Rth1 , of a rock wool sleeve
with an internal diameter di = 6cm and an external diameter de = 10cm.
The thermal conductibility of the rock wool is A = 0.045 W. m 1. K 1 .
2) We subject this to a temperature difference of A T = 25 °C and do not
take into account the convection. What is the flow Ф1 passing through a
length L = 15 m of the sleeve?
3) We take into account the two convection coefficients: an internal
coefficient hi = 5 W m-2K-1 and an external coefficient he = 8 Wm-2K-1.
Given these conditions, recalculate the flow Ф2 through 15 m of this
sleeve.
86 Heat Transfer 1
Solution.—
1) The resistance (per meter) of the sleeve is given by the following

equation:
d i —10
Ln -e Ln
Rth1
dt 6 1.807mKW-1 [3.120]
2 nA 2 п* 0.045
2) Similarly, the flow is calculated by:
Ф1 = L — = 15-25- = 207.5 W [3.121]

1 Rh i 1.807
3) We will recalculate the thermal resistance per meter of length:
т d-e
Ln
d 11
R th 2 =------
_ 1— +-------------- +
2 nA 2 nrtht 2п( rt + e) he
1 10
Ln
------- 6----+--------12------- + 1
6.316mKW-1 [3.122]
2п*0.045 2п*0.01*5 2 n*0.015*8
Ф2 = L — = 15-25- = 59.37 W [3.123]

2 Rth2 6.316
EXAMPLE 3.11.- Insulation of a metal tube
A horizontal metal tube with diameter dt =3 cm is maintained at a

temperature Tt = 200°C. The atmospheric temperature is Te = 20°C.
We want to reduce the flow of heat transferred by this tube to the room in
which it is found.
To do this, we surround it with a sleeve of a rock wool insulator of

diameter dm = 50 cm and thermal conductibility A= 0.045W.m-1.K -1.
We consider the natural convection coefficient at the external surface of

the insulator to be h= 10 Wm-2K-1. We consider that this convection
coefficient maintains a very similar value when the sleeve is removed, and
that the thermal exchange takes place on the external surface of the tube.
1) Suppose that the transfers are reduced to convection only, what flow of
heat is transferred to the environment by a meter of tube if we do not install
the insulating sleeve?
2) What is the thermal resistance per meter of length of the sleeve/
convection assembly?
3) What is the flow of thermal loss Ф 2 per meter of length of the tube if
the sleeve is installed? Does the insulation solution seem efficient to you?
4) Taking into account the relationship between the flow Ф2, the
atmospheric temperature and the external temperature of the insulating
sleeve and of the flow determined in question 3, what is the temperature TW
of the external surface of the sleeve?
Solution.—
1) We have a simple convection flow, whose thermal resistance is (per

meter):
R„ = ---- 1---- = 1-------- 1-------- = 1.061 mKW [3.124]

t 2 n r h 2 n*0.015*10
The flow is calculated as:
AT 200 - 20
Ф = =170W [3.125]
Rth1 1.061
REMARK.- We could have calculated this flow directly using the definition
of the convection coefficient. Calculating the flow through a lateral surface
of the tube of 1 m in length:
Ф1 = Sh A T = ndih AT
88 Heat Transfer 1
This obviously gives the same result, since the thermal resistance is
defined using this last expression!
2) The outside radius changes and becomes:
d
re2 = 0.25 m
=— [3.126]
The thermal resistance takes into account the insulator and the
convection:
T (r ^ < 0.25 ^
Ln 1 — 1 Ln 1------ 1
„ IrJ 1 I 0.015 J 1
R^7 = + = + [3.127]
th2 2 nA 2 nreh 2 n *0.045 2 n 0.25*10
Rth =10.01mKW-1 [3.128]
3) The flow becomes:
ДT
Ф2 = — = 10.01 = 17.98 W [3.129]
In other words, there is a reduction in thermal losses of 89%. The

efficiency is very satisfactory.
4) The temperature at the surface of the sleeve is obtained from the

thermal convection resistance, which is already calculated:
11
Rt,v = ---------- =------------------- = 6.37.10-2 mKW [3.130]
thconv 2nr h 2n* 0.25 *10
On the two sides of the convection limit layer, we then have:
T-T T-T
Ф2 = TW------- e- = - WL-------- e = 17.98 w [3.131]
2 Rthconv 6.37.10-2
Therefore:
TW -Te =1.14°C [3.132]

Moreover:
TWe = 21.4°C [3.133]
EXAMPLE 3.12.- Insulation of a tube with convection
We insulate a tube with an external diameter de = 6 cm and an internal

diameter di = 4cm through which the fluid flows, and which we consider to
be uniform and equal to T0 =170 °C.
The external temperature is Te =20 °C.
We insulate the tube with a sleeve of rock wool of thickness e= 5cm.
1) First, we consider the steel tube to be a perfect conductor and the

convection is perfect on the outside (with the convection coefficient h being
infinite).
90 Heat Transfer 1
a) Calculate the thermal resistance per meter between the fluid and the
external atmosphere.
b) Calculate the thermal flow leaving the fluid and transferring to the
outside, for a length of 10 m of tube.
2) Second, we take into account an external convection coefficient of

h = 5 W.m-2.K 1 and the thermal conductivity Asteel of the tube.
a) Calculate the new value of thermal resistance per meter between the
fluid and the external atmosphere.
b) Calculate the new value of the thermal flow leaving the fluid and
transferring to the outside over a length of 10 m of tube.
Numerical data
Density of air: p = 1.3 kg. m 3.
Specific heat capacity of air: cV = J. kg-1 .
The thermal conductibility of various elements and materials is given as

follows:
Air: Aair = 0.026W.m-1.K-1 Rock wool: Awool = 0.047W. m-1. K-1
Glass: Aglass = 1.05 W.m-1.K-1 Hard stone: Astone = 2.4 W.m-1.K-1
Steel: Asteel = 54 W.m-1.K-1 Wood: Awood = 0.12W.m-1.K-1
Solution.—
1) The steel tube is a perfect conductor and convection is perfect on

the outside.
a) The thermal resistance RthLR per meter of a cylindrical sleeve with an

internal radius of r1 and an external radius of r2 and with thermal
conductibility A is calculated using the expression:
Rth [3.134]
2 пл
d
For the rock wool sleeve, the internal radius is r, = -2- and the external
radius is re + e:
. \ 3+5 1
Ln\ I
R^ ooi =----
thwool 2п 0.047- = 3.321 m2.KW
- ------- [3.135]
b) The resistance is calculated per meter of tube.
Therefore:
170 - 20
Ф = =10* = 451.7W [3.136]
Rth 3.321
2) We take into account an external convection coefficient of

h = 5 W.m-2.K-1 and thermal conductivity of Asteel for the tube.
a) The thermal resistance of the tube alone is
Ln\ Ln (3 I
Rh = - ri =^-^ = 1.24.10-3 m2.KW-1
±±= [3.137]
th 2пХ 2*п*53
The thermal resistance (resistance to convective transfer), due to the

convection coefficient, is:
Rthconv --------------— = 0.398 m2.KW-1 [3.138]

2 п (r + е) he
92 Heat Transfer 1
The resulting thermal resistance is:
Rth = Rthube + Rthwool + Rthconv [3.139]
Rh = 1.24.10-3 + 3.321 + 0.398 = 3.72 m \ KW-1 [3.140]
b) The thermal flow leaving the fluid and moving to the outside for a
10 m-long tube is calculated in the same way as above with the new thermal
resistance:
AT 170 - 20
Ф=----- = 10*----------- = 403.3 W [3.141]
Rh 3.72
3.4. Problem of the critical diameter
EXAMPLE 3.13.- Optimization of the insulation of a tube
We have a cylindrical tube with internal radius r1 and external radius r2 ,

which is made of a material with thermal conductivity 2 . Let us assume
that we want to insulate with a sleeve with external radius r3 and thermal
conductivity 22. Then, h and he are, respectively, the internal and external
convection coefficients.
We will calculate the resistances for a length of 1 m.

1) Give the expression for the thermal resistance Rthtot of the tube alone.
2) Give the expression for the thermal resistance Rthtot of the assembly
tube + sleeve.
3) Determine the conditions for which the addition of a sleeve does
indeed lead to a reduction in thermal losses.
Data:
r2=1.5cm; 22=0.1W.m-1.C-1; h2=6W.m-2.C-1

Solution.—
1) Let us calculate the thermal resistance of the tube alone. We will deal
with the problem for 1 m of tube.
The thermal resistance Rth per meter of a cylindrical sleeve, with internal
radius r1, external radius r2 and thermal conductibility 2, is calculated using
the expression:
Ln\
D _ I
\ r1 [3.142]
h = _ 0
2 П2
NOTE.- We should remember that we can deduce the thermal resistance

Rth (L ) for a sleeve of length L :
Ln\
Rth, (L) = thR.L = L [3.143]
2 ПЛ
Applied to the tube alone, we find:
Ln I Ln I -2-
-e- I
rr
Rh (L = 1 m ) =--------— =--------^-1- [3.144]
2 n2 2 n21
2) The resistance of the sleeve Rthsleeve is calculated as above:
Ln I — I Ln I —
rr
R., (L = 1 m ) =-------- — =-------^2- [3.145]
hsleee ( ) 2 n2 2 n21
The resistance of a convective layer Rthconv is calculated by:
R =—1— [3.146]
thconv
2nrh
94 Heat Transfer 1
The total resistance results from this:
Ln \ — I Ln \ — I
,, l Г1 J lrJ 1 1
h = ~--------------------- +------------ +------------ +----------
Rthtot [3.147]
2 n11 2 n11 2 n r1 hi 2 n r3 —
3) We will examine the effect of adding a sleeve
The parameters r1 , r2 , 11 and 12 are fixed. Therefore, we can only

change the values of r3 and 12. We will presume in a first approximation
that he remains invariable when r3 varies to a reasonable degree.
The optimization will therefore be related to the component of Rthtot that

contains r3 , which can be written as:
Ln\r- I
+J [3.148]
2 n11 2 n r3 he
It is therefore necessary to examine the variation of R (r3 ) when r3

varies. We refer to the case studied in section 3.1.4.
We then define the critical radius:
1
rc =± [3.149]
dR
- when r3 < rc, the derivative — is negative. An increase in r leads
dr
to a reduction in Rth . Increasing the thickness of insulation leads to an
increase in losses;
1
- when r3 becomes greater than the critical radius r3c = — , it becomes
useful to increase the thickness of insulation.

The numerical value of this critical radius is written as:
r2 = 1.5 cm ; 22 = 0.1 W.m-1C-1; h2 = 6 W.m —2 C-1
r33c = —
6 = 1.67 cm [3.150]
We verify that r3c is indeed greater than r2. In addition, the increase in
the thickness of the sleeve insulation becomes effective when it is greater
than 1.67-1.5=0.17cm; in other words, 1.7 mm is low. We see that, in
practice, the condition is not very constraining.
EXAMPLE 3.14.- Optimization of the insulation of a steel bar
A steel bar of diameter 5 cm is surrounded by a sleeve of glass with an

external diameter of 12 cm. The conductibility of the glass is
2 = 1.05 W. m-1. K 1. On the outside of the sleeve, the convection coefficient
is h=8W.m-2.K-1.
The tube is maintained at a temperature TB =70 °C. The external

temperature is Te = 21°C.
1) Calculate the flow of heat between the tube and the external
atmosphere for a tube of length 1 m.
2) We replace the glass sleeve of external diameter 12cm by a new glass
sleeve with an external diameter of 26.3cm . What happens to the heat flow?
3) We replace the glass sleeve of external diameter 26.3cm by a new
glass sleeve with an external diameter of 36cm . What happens to the heat
flow?
4) Compare the flows found in questions 1-3. What can you observe?
How do you explain this result?
96 Heat Transfer 1
Solution.—
1) Sleeve of D= 12cm
We calculate the thermal resistance Rth of the sides of the bottle, using
the equation:
rr
Ln — Ln—
r= r_ [3.151]
2n A 2n A ’
which is applicable to a homogeneous cylinder.
Di = 5.10-2 m
De =12.10-2 m
A=1.05 W.m-2.K-1
he = 8W.m-2.K -1
Ln D Ln 12
R = ------- i- =------------ 5 = 0.133 W. m 2.K [3.152]
th 2nA 2* n*1.05
The total thermal resistance, including the convection coefficient, is:
L
Rtot
---- D- +--- 1 = 0.133 + 1 ,
2 nA 2 nrehe 2* n*6.102*8
=0.133+0.3316=0.464W.m-2.K-1 [3.153]
The flow of heat Ф exiting the system per meter for Tt = TB = 70 °C and
Te = 21°C is:
*, = T^T = 70-21 = 105.6 W

[3.154]
Rot 0.464
2) Sleeve of D= 26.3cm
We calculate the thermal resistance Rth of the sides of the bottle, using
the equation:
rr
Ln — Ln—
Rth = =----- - [3.155]
2nA 2nA
Di = 5.10-2 m
De = 26.3.10-2 m
A = 1.05 W.m-2kK4
he = 8W.m-2.K -1
LnD Ln —
Rh = ------- i- =-------- 5----- = 0.251 W.m-2kK [3.156]
th 2 nA 2* n*1.05
LnD
R = D. + 1 = 0.133 +------------- 1------------
tot 2nA 2nrehe 2*n*13.1510-2 *8
=0.251+0.151=0.402W.m-2.K-1 [3.157]
The flow of heat Ф exiting the system per meter for T = TB = 70 °C and
Te = 21°C is:
*, = T^T = 70-21 = 121.9 W

[3.158]
0 Rtot 0.402
98 Heat Transfer 1
3) Sleeve of D= 36cm
We calculate the thermal resistance Rth of the bottle walls, using the
equation:
rr
Ln — Ln—
r =__ r_ [3.159]
2пЛ 2пЛ
Di = 5.10-2 m
De=36.10-2 m
Л= 1.05 W.m-2.K-1
he= 8W.m-2.K -1
Ln D Ln 36
Rth = ------- l- =-------5 = 0.299 W.mz.K [3.160]
th 2пЛ 2*п*1.05
LnD
1
Rtot ------ — +----------- = 0.299 +
2 пЛ 2 nrehe 2* п*18.10-2 *8
=0.299+0.1105=0.409W.m-2.K -1 [3.161]
The flow of heat Ф exiting the system per meter for Tt = TB = 70 °C and
Te = 21°C is:
*, = T^ = 70-21 = 119.8 W
[3.162]
0 tot
0.409
4) Here, we represent the effect of the critical radius on the insulation

with a cylindrical sleeve.
The critical radius is
rC = ^ =1.05 = 0.13 m [3.163]

he 8
r = 6cm < rC ; Ф = 105.6 W [3.164]
We can then compare with different radii:
r = 13.15 cm > rC ; Ф = 121.9W, the initial radius is smaller than the

critical radius, r=6 cm<rC , the thickness of the glass increases, and the
flow of losses increases.
r = 18cm > rC ; Ф = 119.8W, the initial radius is larger than the critical
radius, r=13.15cm>rC , the thickness of glass increases, and the flow of
losses decreases.
3.5. Problem with the heat balance
Example 3.15. Insulation for a network of tubes
A heat network is made up of steel tubes with an internal diameter

d =2ri =18cm and an external diameter de=2re=20cm. The thermal
conductibility of the steel is AS = 54 W m 1 K 1. This tube is lagged with
foam of thickness e= 5cm. The conductibility of this foam is
Am = 0.04W m-1K-1.
Water flowing around the circuit at temperature Twat will be considered,

for simplicity, as constant, which is equal to 70°C. The temperature of the
ground TG is homogeneous, which is equal to 6°C. The internal convection
coefficient is given by hi = 700 Wm-2K-1.
100 Heat Transfer 1
1) Initially, we assume that the conductibility of the steel is infinite. We

also presume that the convection does not involve thermal resistance
(infinite convection coefficient).
a) Give the expression and the value of the thermal resistance Rth1 per
meter of foam.
b) What then is the thermal loss for 5 km of pipe?
2) We now take into account the thermal conductibility of the steel.
Moreover, we consider the influence of the convection coefficient h
between the flow of water and the tube.
a) Give the expression and the value of the thermal resistance per meter
Rth2 for the foam, steel and convection elements.
b) What does the thermal loss then become for 5 km of pipe?
3) The volume of water is qV = 10 m3hr-1. Does the hypothesis of a
constant water temperature for 5 km seem reasonable to you for an
evaluation of the losses?
Solution.—
1) Using the hypotheses in this question, the thermal resistances of the

steel duct and convection are considered to be zero.
a) We have a simple problem of conduction between the water at the
given temperature and the air, via a thermal resistance Rth1 of the foam. This
thermal resistance, which is, of course, a resistance per meter of pipe, is
calculated using the known relationship:
re +e 0.1+0.05
n------ Ln------------
r = 0,1 [3.165]
2 nAm 2 П0.04
Rth/m =1.613KmW-1 [3.166]
b) This resistance is expressed per meter. The flow loss will therefore be:
_ , AT
Ф= L---- [3.167]
with A T = 70 - 6 = 64 C, and L = 5000 m
64
ф = 5000------ = 1.98.105 W [3.168]
1.613
2) We will therefore need to add the resistances per meter for the foam,
the steel (which is calculated using an identical equation) and the convection
coefficient.
a) There will be a resulting resistance, taking into careful account the

various values of the radii, which are different for each sleeve:
r+e r
Ln —----- Ln —
riri1
Rth2 — + + [3.169]
2 п ЛmSi 2 п Л„ 2 nh
Rth2 —1.613+3.1.10-4+2.27.10-4 [3.170]
We obtain an almost identical value:
Rth2 —1.613KmW-1 [3.171]
b) The thermal loss will therefore be identical to that previously

calculated.
3) We will write a thermal balance between the input and the output
of the tube: the difference between the incoming heat flow and the outgoing
heat flow, with the flow, is equal to the thermal losses.
This gives us:
Ф = pcqv (T - ToUl) [3.172]
where p = 1000 kg.m-ъ is the density of water and c = 4180 J.kg 1 is its
specific heat capacity (values which should be known in principle). Tin and
Tout are, respectively, the input and output temperatures of water in the tube:
qv — J°- = 2.78 m3 5-1 [3.173]

V 3600
102 Heat Transfer 1
1.98.105 =1000*4180*(Tin -Tout) [3.174]
Tin - Tout =16.97C [3.175]
The loss of 17°C represents 24% of the input temperature. This is quite a
rough approximation.
4
Quasi-stationary Model
We often want to know how the temperature of a body, that is subject to

an outside transfer (either heating or cooling), will change.
4.1. We can perform a simplified calculation, adopting the

following hypotheses
a) The body with volume VOL is spatially homogeneous: in particular,

its temperature T (t), volumetric mass p and specific heat capacity c are
spatially constant.
b) The temperature varies slowly over time: we then suppose that the
transfer to the exterior can be calculated using relationships taken from the
stationary theory. In addition, the thermal balance, as established, assumes
that the temperature is constant.
We can observe that these hypotheses are not sufficient, in several ways:
The temperature of the body cannot be rigorously constant, otherwise,

transfer to the exterior would not be possible (conduction or convection on
internal faces, for example). We then presume that the required temperature
gradients are located on a small zone near the internal limits of the body. We
could take this into account with a convection coefficient, which is
expressed with a thermal resistance (internal and/or external convection).
104 Heat Transfer 1
The transfer should, strictly speaking, correspond to the heat equation in

the transitory regime. Here, we effectively neglect the effects of this
transitory nature. The timescales applied here must therefore be large.
4.2. Method: instantaneous thermal balance
We resolve the problem with a thermal balance. We write that the

variation of the temperature of the body in a very short time dt is due to the
transfer with the exterior.
We will give the basis for the calculation here, by way of an example.
REMARK.- Be careful, we cannot learn a “recipe” here, some aspects of the

model can vary from one problem to the next. We need to understand a
method.
So, for a homogeneous body with volume VOL, density p and specific
heat capacity c, its mass will be:
m=pVOL [4.1]
During the time dt , the quantity of heat exchanged by the body and the
exterior is given by:
dQ=pcVOLdT [4.2]
This quantity has the same sign as dT . dQ is therefore positive for

heating, and negative for cooling (here, we observe the usual conventions of
thermodynamics).
This quantity of heat given or extracted from the body will be due to the
thermal transfer, which can generally be written as:
ф=SH-T
[4.3]
where Te is the external outside temperature (“distant” from the body).

Quasi-stationary Model 105
Written in this way, Ф is positive for a heat transfer that is directed

towards the exterior, therefore a loss of heat. Close attention must be paid to
this when the balance is established.
During this time, the heat exchange will be Ф dt
The balance will therefore be written as follows, taking into account the
above equations:
dQ =-Ф dt [4.4]
The minus sign is a result of the remark made above. We note that its
presence is a result of the choice made when writing dQ (choice of the
thermodynamic sign convention).
This balance therefore leads us to a differential equation for T(t)
T(t)-T
dQ = p cVOL dT = -Ф dt = - -dt [4.5]
dT T(t)-T-
[4.6]
dt pcVOLRth
This equation must be associated with an initial condition. Often, we write:
t=0; T=T0 [4.7]
We note that writing Rth characterizes a problem. Writing this can, in

effect, take into account all or some of the following parameters:
- existence of an insulating layer;
- convection coefficients, external and possibly internal (e.g. the natural
convection in a room);
- possible radiation phenomena (which, as we will see in another chapter,
can be expressed in certain specific cases by a linear relationship that is
analogous to the use of a convection coefficient).
106 Heat Transfer 1
NOTE.- In this type of approximate treatment, we make sure that we are

choosing constant components for thermal resistance in convection; this is a
usual approximation, which is not necessarily always rigorous.
4.3. Resolution
To solve the differential equation, a new unknown function is often

introduced:
O( t )= T (t)-Te [4.8]
The equation and the initial condition become:
d0 _ 0
[4.9]
dt PcVOLRth
t=0;0 =T0 -Te
The equation is transformed into the form:
d0 dt
------- = - adt [4.10]
~0 PScRth
with
1
a =------------ [4.11]
psc VOL Rth
as the coefficient, which makes the written equation lighter. A source of

error that is frequently seen in “copies” in fact comes from errors in the
systematic rewriting of slightly complex expressions.
Integrating the two terms of the equation:
Ln 0 = - at + LnC [4.12]
A constant Ln C appears, which will be given by the initial condition.

The choice of the constant in the form of a logarithm is a clever trick, which
allows a change to the exponential of the previous expression to be

simplified. Taking the exponential of each term:
exp (Ln0)= exp (-at) exp (LnC) [4.13]
0 = C exp (-at) [4.14]
and
t= 0 T0 -Te =C [4.15]
which results in the final expression for the temperature of the bar:
T(t)-Te=(T0-Te)exp(-at) [4.16]
This expression is general.
If T(t)-Te >0 , then cooling takes place, so T(t) decreases [4.17a]
If T(t)-Te <0, then heating takes place, so T(t) increases [4.17b]
As we will repeat, learning this result by heart must be avoided because it

can lead to errors, but we must instead be able to carry out the reasoning
each time. The following examples will allow the reader to become used to
this.
4.4. Applications for plane systems
EXAMPLE 4.1.- Heating of a room by radiation
REMARK.- the radiation manifests itself here by an imposed heat flow term.
This example has a natural place in a chapter that is dedicated to conduction,
and not to radiation, where the latter field does not require any specialized
calculations.
A room in the form of a parallelepiped with dimensions 3mx4m on the

ground and a height of 2.5 m is lit by a glass window with dimensions
2mx2m.
108 Heat Transfer 1
We consider that the room is perfectly insulated.
The glass is lit from the instant t = 0 by a flow of Ф= 300 Wm-2. We

consider that this flow received by the glass is used entirely to heat the air in
the room.
At time t= 0 , the temperature is Ti =18 °C.
1) Write the differential equation which represents the temperature Ti (t)

of the room.
2) Solve this equation and give Ti (t)
3) How long will the temperature take to reach Ti =25°C? How long will
the temperature take to reach Ti =70°C?
4) Here, the model is very simplified. Under real conditions, do you think
that a temperature of 70°C can be reached in the room? Why?
Solution.—
1) A simple thermal balance gives the differential equation for Ti (t)
pVOLcdT = Ф Sdt [4.18]
the notation and values associated with this are classic pi = 1.3 kg.m-1,
c= 1000 J.kg-1.K-1
VOL = 3*4* 2.5 = 30 m3 and S = 4 m2 is the glass surface [4.19]
dT_ Ф S
[4.20]
dt pVOLc
with the initial condition
t=0; T=Ti =18°C [4.21]

2) Resolution is immediate.
We obtain a linear solution over time:
ФS
T=--------------- t + Cst [4.22]
P VOL c
The constant Cst is given by the initial condition
ФS
T (0) =-------- 0 + Cst = Tt [4.23]
p VOL c
Therefore:
t=T + Ф S t [4.24]
PVOLc
Numerically:
ф' = 300*4 = 3.08.10-2 [4.25]

PVOLc 1.30*30*1000
T =18+3.08.10-2 t [4.26]
3) We will reach T=25°C by increasing the temperature by 8°C , in

other words, time:
125 =--------- - —= = 260 s = 4 mn 20 s [4.27]

25 3.08.10-2
We will reach T=70°C by increasing the temperature by 52°C , in other

words, time:
1„ = 5^- = 1688 s = 28mn 8 s [4.28]

25 3.08.10-2
110 Heat Transfer 1
4) The numbers are not realistic, above all the second, because at a
temperature of T=70°C, it will not be easy to neglect the thermal losses in
the building!
EXAMPLE 4.2.- Thermal evolution of a garden hut
Where necessary, we will use the numerical data that is also given at the
end of the text.
A small garden hut is made up of a single room with internal dimensions

5*7 m2 on the ground and a ceiling height of 2.5 m .
The walls are made of concrete with a thickness of 20cm . The roof is
made of a concrete terrace area of the same thickness.
We insulate all the walls and the ceiling with a 25cm layer of expanded
polystyrene and BA13 plasterboard (13mm thick).
That is, hi and he are the internal and external convection coefficients
respectively (valid for the walls and ceiling).
1) Calculate the thermal resistance per m2 of the insulated wall.
2) The garden hut is heated by an electric radiator.
The external temperature is Te =6°C. What minimum power Pmin must

this radiator have in order to maintain a constant temperature in the garden
hut of Ti =20°C?
We will neglect the thermal losses through the ground. This is a

calculation for an order of magnitude; we can presume that the doors and
windows have the same thermal resistance as the walls.
3) Radiator breakdown.
Initially, this question will be resolved literally; only the numerical

applications requested in question four will be completed.
The radiator breaks down. We presume that the thermal losses can still be
evaluated, as they can in the stationary regime, when the internal
temperature of the garden hut Ti (t) varies with time. The external
temperature Te remains constant.
We will call Rth the total thermal resistance of the garden hut, S the total
surface area of the walls and Vol the internal volume of the garden hut.
a) Write the expression for the instantaneous flow of thermal losses from
the garden hut.
b) Writing the thermal balance for the garden hut, give the differential
equation for the behavior of Ti (t). It will be useful to introduce the function
#( t )= T (t)-Te
4) Numerical application.
Te =6 °C, as above
T,0 =20°C
After the breakdown of the radiator, how long did the temperature take to
fall to T, =10 °C inside the garden hut? Give this time in hours and minutes.
5) Another insulation solution.
Consider a DIY enthusiast who thinks it would be a good idea to replace

the polystyrene and plasterboard on the walls and ceiling, with a sandwich of
five BA13 plasterboards, separated by four air blasts of thickness 2cm .
a) Calculate the new thermal resistance per m2.
b) Te and T,0 maintain the previous values of 6°C and 20°C. How long,
in hours and minutes, will it take for the temperature of the interior of the
garden hut to descend to T, = 10°C with this new insulation?
c) Does this second solution appear to be more satisfactory than the

previous one?
112 Heat Transfer 1
d) Would it have been more useful to replace the polystyrene by glass

wool?
Additional data
Thermal conductibility of air: hair = 0.026 W. m-1. K-1
Thermal conductibility of concrete: ^concrete = 1.4 W. m-1.K-1
Thermal conductibility of polystyrene: hpoly = 0.035W.m-1.K -1
Thermal conductibility of plasterboard: hBA13 = 0.25W.m-1.K -1
Density of air: pair = 1.3 kg. m -3
Specific heat capacity at constant pressure of air: Cp = 1007J.kg -1
Specific heat capacity at constant volume of air: Cv = 719J.kg-1
hi =he =5W.m-2.K-1
Solution.—
1) Thermal resistance of the wall
For the wall, we associate three different types of walls:

- a concrete wall with thermal resistance R
thconcrete ;
- a wall of expanded polystyrene with thermal resistance RthPoly ;
- a wall of BA13 plasterboard with resistance RthBA13
We will calculate the thermal resistances.

For the concrete:
e =20.10-2m
2 = 1.4 W. m к-1
R^ = e = 0.143 K. mW-1 [4.29]

2
For the polystyrene:
e =25.10-2m
2 = 0.035 W. m K-1
R^y? = e = 7.143K. m W-1

2
For the BA13 plasterboard:
e =1.3.10-2 m
2= 0.25W.m-2.K-1
R 3 = e = 5.2.10-2 K. mW-1 [4.30]

2
Applying the additive nature of the thermal resistances “in series”, also
taking into account the transfer resistances that result from the convection
coefficients:
Rb = RMconcree + Rlv + RBA13 + hn~ + h' = 7-74K. m2 W-1 [4.31]
2) Calculation of the leaks in static
The surface area which the leaks are distributed over is:
S=2*(7+5)*2.5+7*5=95m2 [4.32]
114 Heat Transfer 1
The volume of the room is:
VOL =7*5*2.5=87.5m3 [4.33a]
The difference in temperature is:

A T = 20 - 6 = 14 C [4.33b]
The flow of the leaks will be:
AT 14
Ф = S<p = S----- = 95* = 171.83 W [4.34]
R RhB 7,17
3) Radiator breakdown
a) We will give a general expression for the instantaneous flow of
thermal losses from the garden hut.
Ф=ST^
[4.35]
b) Equation giving the evolution of Ti (t)
During a time period dt, the flow of heat exiting the garden hut Ф dt
leads to cooling dTi such that:
Ф dt = - pCVVOLdTi [4.36]
The minus sign expresses that the outgoing heat leads to the cooling of
the room.
Replacing Ф by its expression and introducing 0 , we obtain the
differential equation:
PairCvVoLdT = — S° [4.37]
dt R
This can be rewritten, noting that:
dTt = d(T - Te) = do

[4.38]
dt dt dt
= -а0 [4.39]
dt
with
S
а =------------------------------ [4.40]
PairCvVoLRthB
c) Calculation of Ti (t)
This is a first-order linear differential equation. A single limit condition

will be necessary.
In other words:
00=Ti0-Te [4.41]
At t =0; 0=00 [4.42]
d0
= -a0 [4.43]
dt
d0
—=-adt [4.44]
0
dLn0 = -at + LnC [4.45]
C is a constant. The choice of a constant in Ln C means we can

immediately write the following as:
0( t ) = C exp-at [4.46]
The condition for initial time immediately gives C = 00
0( t ) = 00 exp-at [4.47]
116 Heat Transfer 1
4) Numerical application
Te = 6°C, as previously seen
Ti0 = 20°C
We are trying to find how long it takes after the radiator breakdown for
the temperature to drop to 10°C inside the garden hut.
For 10°C , we have
0 = 10 - 6 = 4°C ; 00 = 20 - 6 = 14°C [4.48]
Calculation of a gives
a=1.5.10-4 [4.49]
0(t) = 00exp-at gives t =—Ln— [4.50]

0 ae0
-14
t =---- Ln— = 8352 5 [4.51]
1.5.10-4 14
t=8352s=10h19mn12s
5) Another insulation solution
a) With polystyrene, the new thermal resistance per m2 is calculated using

the same procedure as above. The thermal resistance per m2 of a wall
becomes
R = 0.2 + 1.3.10-2 + 4*2-0— +1 + - = 4.737 m2,KW [4.52]

2
rhB 0.2 0.25 0.026 5 5
b) With this new insulation, time is proportional to — and therefore

proportional to the thermal resistance Rth , with all of the other parameters
remaining constant.
The new time taken to fall to 10°C will therefore be:
t = 9232 4737 = 50885 = 1 h 24mn 385 [4.53]

8.595
c) With this second solution, the cooling time is twice as weak as for the
initial insulation! Obviously, this solution is less satisfactory than the
previous one.
d) It would not have been more advantageous to replace the polystyrene

with glass wool, since the thermal conductivity A = 0.041 W.m-1.K-1 of the
glass wool (0.041) is greater than that for polystyrene A = 0.035 W. m-1.K-1
EXAMPLE 4.3.- Temperature of a block of metal
We advise you to process questions 1 to 3 literally. Numerical

calculations will only be carried out for question 4.
A steel block has a rectangular base with sides a and b and height H. Its
specific heat capacity is denoted c.
This parallelepiped is placed on its base. We will estimate that no thermal

transfer is possible through the base.
The cube can only transfer energy to the exterior via its other faces, for
which a convection coefficient h applies.
In addition, the density of convection flow will be considered

homogeneous on all exposed faces of the block.
The initial temperature of the block is T0 , higher than the atmospheric

temperature Te .
1) Give the expression of the flow density ф1, then of the flow Ф1
coming out of the block at the initial moment in time.
118 Heat Transfer 1
2) The block cools down. We will note T(t) as its instantaneous

temperature that we will presume to be homogeneous. Write the expression
for the instantaneous flow Ф( t) that exits the block as a function of T(t) and
the relevant parameters.
Deduce from this, by writing the instantaneous heat balance in the block,
the differential equation that describes the evolution of the temperature of
the block.
3) Find the expression for T(t). To do so, it will be useful to define a

variable 0( t )= T (t)- Te
4) Numerical applications
We give
a=30cm; b=20cm; H =15cm
Density of steel:
Psteel = 7864 kg. m -3
Specific heat capacity of steel, c= 460J.kg -1.K -1
h= 5W.m-2.K-1
Te =20°C
T0 =150°C
a) Give the value of ^1 and Ф1
b) How long will it take for the difference between the temperature of the
block and the outside to descend to 10% of its initial value?
Solution.—
1) The expression of the flow density q>1, then of the flow Ф1 is

calculated using the definition of the convection coefficient:
5 = 2 (a + b) H + ab = 0.15 + 0.06 = 0.21 m2 [4.54]
VOL =abH=9.10-3m3 [4.55]
(p = h(T0 - Te ) = 5.130 = 650W.m-2 [4.56]
Ф1 = Sh(T0 - T ) = 0.21*5*130 = 136.5 W [4.57]
2) The outgoing heat over time dt leads to cooling dTi :
Ф( t ) = Sh [ T (t)-Te ] [4.58]
PsteeiVoiCdT =-Ф dt [4.59]
PsteeiVoicdT = —Ф = — Sh [T (t) — Te ] [4.60]
3) This equation is solved immediately.
We define a variable
0( t )=t (t)-T [4.61]
0 = ( Tt - Te) [4.62]
do
= -a0 [4.63]
dt
Sh
a == P.e.Vo.c = 3225'10 [4.64]
120 Heat Transfer 1
d0
—= -a0 [4.65]
dt
d0
— = -adt [4.66]
0
Integrating the two terms:
Ln0= -at +Ln C [4.67]
0=Cexp-at [4.68]
t=0;0=00 [4.69]
0=00 exp -at [4.70]
4) Numerical applications
a) Values of ф1 and Ф1
S = 2 (a + b) H + ab = 0.15 + 0.06 = 0.21 m2 [4.71]
VOL =abH=9.10-3m3 [4.72]
(p = h(T0 - Te ) = 5.130 = 650W.m-2 [4.73]
Ф1 = Sh(T0 - Te ) = 0.21*5*130 = 136.5 W [4.74]
b) We want to determine how long it will take to drop to 10% of the

initial value
0 = 00 exp - at [4.75]
We will calculate the corresponding value of
00=150-20=130° [4.76]
a= 3.225.10-5 [4.77]
t =-1ln— [4.78]
a —0
— = 0.9 [4.79]
—0
t =—ln — [4.80]
a—0
EXAMPLE 4.4.- Life in a castle is not always ideal
An 18th-century castle is built of hard stone. The walls are 75 cm thick.
The plane is rectangular, the facades (see drawing) are respectively

100 m and 40 m in length. The height of the facades is uniform and equal
to 20 m.
Each of the two facades of length 100 m are pierced by 100 m2 of

windows. These windows are made of glass that is 8 mm thick.
The total surface area of the walls (except the windows of course) is
covered with wood of thickness 8 cm.
-------------100 m
40 m
20 m
Windows : 100 m2 for each facade
Figure 4.1. Representation of the castle. For a color version

of this figure, see www.iste.co.uk/ledoux/heat1.zip
122 Heat Transfer 1
We consider two convection coefficients, interior hi = 10W.m-2.K -1 and

exterior he =10W.m-2.K-1; these coefficients apply to the walls, the terrace
and the windows.
In this highly simplified model, we will neglect the thermal losses

through the roof.
1) Calculate the thermal resistance (per m2) of the walls and windows
RthWall and RthWindows , respectively, taking into account the wooden
surrounding and the convection. Calculate the respective surface areas of the
wall, the terrace and the window.
2) We are in winter. The outside temperature is Te =5°C. We wish to

maintain a temperature of Ti0 = 14°C inside the castle.
What heating power is required to maintain this temperature of 14°C

inside the castle?
3) We have managed to achieve a temperature of Ti0 = 14°C inside

the castle, but there is no more wood to put in the chimneys.
The outside temperature stays at Te =5°C.
We are going to calculate the reduction in temperature Ti (t) inside the

castle.
To do so, we will presume that the conductive transfer is quasi-stationary.
The outside temperature stays at Te =5°C.
We are going to calculate the reduction in temperature Ti (t) inside the

castle.
To do so, we are going to presume that the conductive transfer is

quasi-stationary.
a) Write the relationship between the instantaneous thermal loss rate and
the instantaneous internal and external temperatures Ti (t) and Te(t)
respectively.
b) Write the thermal balance equation that connects the derivative of the
internal temperature with respect to time and this thermal loss.
c) Resolve the equation and give the expression for в (t).
We will define 00 = T.0 - Te
d) How long will it take for the temperature to become 6°C inside the
castle? Give this time in hours and minutes.
It was cold in Versailles during Louis XIV’s era. Can you understand
him?
Additional data: Some properties for various materials
Aair = 0.026 W.m -1.K-1 ^rockwool = 0.047W. m -1.K-1
\iass = 1.05 W. m-1. K " Aardstone = 2.4 W.mK
Asteel = 54 W.m-1.K-1 Awood = 0.12W.m-1.K-1
Density of air:
p = 1.3 kg. m -3
Specific heat capacity of air:
cv =718J.kg-1.K-1
Solution.—
1) We will use the additive nature of resistances to transfer.
For the wall, the thermal resistance per m2 is:
Rthmur = ef^ + ew22^ + 1- + 1- [4.81]

AA
stone wood hh
i e
124 Heat Transfer 1
Rthwall =0.312+0.667+0.1+0.1=1.179°C [4.82]
For glass, the thermal resistance per m2 is:
=e_gasL + 11
+.1
tV A. h h [4.83]
glass i _
R = 810- + 0.2 = 0.207 m 2. KW- [4.84]

thV 1.05
The surface areas are:
STotal =5600m2 [4.85a]
wall = 5400m2
S [4.85b]
Sglass =200m2 [4.85c]
2) Heating power is necessary to maintain the temperature of 14°C inside

the castle.
This power is equal to the thermal flow that escapes by conduction:
ф= V T___
_J -TL + V T-T
_>___ L
wall glass [4.86]
thwall thglass
99
ф = 5400------ + 200-------- = 4.992.104 W [4.87]
1.179 0.207
In other words, a power of approximately 50 kW.
3) There is no more wood to burn in the chimneys.
a) We will determine the relationship between the flow of instantaneous

heat loss and the instantaneous internal and external temperatures Ti (t) and
Te = Ct respectively.
Application of a quasi-stationary process is equivalent to maintaining the

relationship between flows and temperature written in the stationary regime:
Ф_ о T (t)- T. + о T (t)- Te
ф °wall p 1 о glass & [4.88]
thwall thglass
b) Balance equation between the outgoing flow and cooling of the air in
the castle:
Ф dt = - pCVVOLdTi [4.89]
the room.
Replacing Ф by its expression, introducing the variable в , we find the

в=T. - T (t) [4.90]
p cvdT=-о +v W—l
pair~^V O OL w wall glass [4.91]
dt Rthwall Rthglass
which can be rewritten, noting that:
dT = d (T — Te) = de
[4.92]
dt dt dt
de
= -ав [4.93]
dt
with
оо
Owall + gljslass
a R,i ,,
thwall R„,
thglass
[4.94]
pair CV Vni
OL R.thh
126 Heat Transfer 1
The volume of the castle is:
VOL =100*40*20=80000m3 [4.95]
a gives
SS
S wall + glass
Rthwall Rthslass 5546 7 43 10-5

a= [4.96]
Pair CV VR
OL th 1.3*718*8.104
This equation, associated with the limit condition:
At the initial instant
t=0 ; e = 00 = Tinto - Te [4.97]
c) The equation is classic and gives the expression for 0 (t).
We define
00=Ti0-Te [4.98]
d0
— -~a0 [4.99]
dt
d0
---- =-ddt [4.100]
0
dLn0=-at+LnC

immediately write:
0(t)=Cexp -at [4.101]
The condition in initial time immediately gives C = 00
0( t)- 00 exp-at2 [4.102]

d) After how long will it be 6°C in the castle?
The time we are seeking is such that:
t = tf; 0 = Of = Tf - Te = 6 - 5 = 1 [4.103]
Of = 00 exp-atf gives tf =— Ln— [4.104]

a 00
-1 1
tf =---------- - Ln = 29572 5 [4.105]
f 7.43.10-5 9
t=8h12mn525 [4.106]
Taking into account the powers to be applied, the phenomenon must have
occurred often.
EXAMPLE 4.5.- Thermal considerations for a farmhouse
An old farmhouse in the form of a parallelepiped has a length on the

ground of 16 m and a width of 7 m, over a height of 4 m. That is, S , the
total surface area of the vertical walls, and VOL the volume that needs to be
heated, will be assimilated to a parallelepiped with dimensions 16*7*4m3 .
We are in winter. The temperature outside is Te =5°C. We want to

maintain a homogeneous internal temperature of Ti0 =19°C.
1) In the original construction of the farm, the walls of the farm are made
of granite of thickness e= 80 cm. The walls are not insulated.
a) Give the expression and the value of the thermal resistance per m2
Rth11 of the farm wall. Take into account an internal convection coefficient
hi = 9 W m-2K-1 hi and an external convection coefficient he = 16 W m-2K-1.
b) To estimate the losses through the roof, we will consider that they are
added to the losses through the walls and are equal to 40% of the losses
through the vertical walls.
128 Heat Transfer 1
For the calculation, we will assimilate the door and windows (small
surface areas) to sections of wall.
What heating power Ф1 is required to maintain a temperature T 0 = 19° C

in these conditions in the farm?
2) We insulate the inside of the vertical walls with a layer of e2 = 15 cm

of rock wool. We insulate the roof so that the losses through the roof still
represent 40% of the losses through the vertical walls.
a) What happens to the thermal resistance per m2 Rh 2 of the wall of the

farm? (Expression and value). We will still take into account an internal
convection coefficient hi = 9 W m-2K-1 hi and an external convection
coefficient he = 16 W m-2K-1
b) What heating power Ф2 is now necessary to maintain a temperature

Ti0 = 19°C in these conditions in the farm?
3) The heating breaks down when it is Ti0 = 19°C inside the farm, and
the external temperature is still Te =5°C.
We are obviously in a case where the house is insulated.
Give the expression for the total loss per unit of time Ф as a function of
the internal temperature difference Ti (t) and the external temperature Te .
We will introduce the variable в = Tt ()) - Te
a) Write the differential equation for в()) and a, a carefully-selected

unique coefficient.
b) Resolve this equation and give the expressions for в()) and T i ())
c) Numerical application. How long will the temperature take to reach

T=8°C?
Numerical data:
Density of air: p = 1.3 kg. m -3
Specific heat capacity of air at a constant volume: cV = 716 J.kg -1.K
Thermal conductibility of air: Xair = 0.026 W m-1K-1
Thermal conductibility of rock wool: XW = 0.045 W m-1K-1
Thermal conductibility of granite: XG = 3 Wm-1K-1
Solution.—
1) The walls are not insulated
a) The thermal resistances of insulation and convection are:
R _ e 1 1
1 = \ + h + ~he [4.107]
Rth1 = 0.267 + 0.111 + 6.25.10-2 = 0.44 K. m2.W-1 [4.108]
b) The exchange surface is:
S=184m2 [4.109]
The flow gives:
T-T
Ф1 = S^----e- *1.4, [4.110]
where the coefficient 1,4 takes into account the losses through the roof.
That is:
Ф1 = 8196 W [4.111]
This is a high value for heating a house!

130 Heat Transfer 1
2) The walls are insulated
The thermal resistance is increased by the resistance of the insulating

material:
_ e 1 1 e2
1 = 1Q + h + h, +1 [4.112a]
Rth1 = 0.267 + 0.111 + 6.25.10-2 + 3.33 =3.77 K. m2 W-1 [4.112b]
The flow of losses will then be:
T-T
Ф1 = S^---- — *1.4 [4.113]
1 Rth 2
Ф1 = 956.6 W [4.114]
This is much more reasonable!
3) The equation for the temperature results from a classic thermal

balance:
PMdT ==—STPT 1.4

[4.115]
dt Rth2
with the initial condition
t=0;Ti=Ti0=19°C: [4.116]
a) Introducing 0( t)
- = -S----- ee-------- 1.4 = -ae [4.117]

dt PaircVVOLRth 2
with:
[4.118]
Pair cV VOLRth 2
the initial condition becomes:
t=0 ; во = Tt0 - Te = 14°C : [4.119]
b) The solution is classic:
Ln в = - at + LnC [4.120]
в = C exp (-at)
at
t= 0 в0 =C =14 [4.121]
Hence, the final expression for the temperature:

Ti(t)-Te=(Ti0-Te)exp(-at) [4.122]
4) We will calculate the time required to drop to 8°C.
Then,
в = 3° C [4.123]
We find:
a=1.63.10-4 [4.124]
the time we are looking for is:
-at=L^tT
[4.125]
T -T
i0 -
1 T-T
tS = aLnTi (t)-Te [4.126]
132 Heat Transfer 1
114
ts = Ln [4.127]
a 3
tS =9450s =2hr37mn [4.128]
EXAMPLE 4.6.- A DIY enthusiast is getting ready for a picnic
A DIY enthusiast has built an insulated box to protect their food for their
picnic.
This box takes the form of a rectangular parallelepiped. It is constructed

from plywood boards of thickness 18 mm. The thermal insulation is
completed by polystyrene sheets of thickness 5 cm, which are stuck to the
inside of the box. Assembled in this way, the inside of the box is a
parallelepiped with a base of 50 * 30 cm2 and a height of 40 cm.
The box has been filled in a house where the atmospheric temperature
was Tint0 =16°C.
This temperature is maintained in the box until it is taken out for the
picnic; it is forgotten outside during a family photo session; the atmospheric
temperature is Te =25°C. The external and internal convection coefficients
are he = 5W.m-2.K-1 and hi = 5W.m-2.K -1 respectively. In addition, no
thermal flow will pass through the base during the problem.
1) Calculate the thermal resistance per m2 of the walls of the box.
2) When the box is taken out, the atmospheric temperature is still

Tint 0 = 16°C in the box. What is the thermal flow entering the box at that
moment?
3) Since heat enters the box, the internal temperature Tint (t) increases.
What is the expression for the total instantaneous flow entering the box?
4) Writing the thermal balance for the air inside the box. Deduce the
differential equation from this, which gives the evolution of Tint (t).
Resolve this equation: to do this, it will be practical to introduce the

variable в = Te - Tint (t). Give the expression for в(t).
5) We consider that certain foods are in danger above 23°C. What is the
maximum authorized time for the photography session? Will the DIY
enthusiast be pleased by this first test?
6) The DIY enthusiast then uses their coolbox: they introduce a liter of
ice into the box.
We will then apply the following:
- the internal temperature of the air in the box becomes Tint = 7°C ;
- this temperature is maintained as long as all the ice has not melted;
- all of the heat entering the box by conduction is used to melt the ice,
whose fusion temperature is L .
a) Since the external temperature is still Te =25°C, calculate the thermal

flow passing through the walls of the box.
b) For how long can the temperature remain at Tint =7°C in the box?
Is this system more satisfactory than the previous one?
We give:
Density of air: p = 1.3 kg. m-3
Density of ice: pice = 917 kg. m“3
Specific heat capacity of air: cV = 716kg.m-3
Thermal conductivities of plywood and polystyrene:
Polystyrene = 0.035 W. m -1.K’1; Plywood = 0.15 W. m-1.K-1
Latent heat of the ice: L= 333kJ is necessary to make 1 kg of ice melt.

134 Heat Transfer 1
Solution.—
1) Thermal resistance of the walls of the box
We are opposite a composite wall. The thermal resistance per unit of

surface Rth will be the sum of the thermal resistances, which can be
calculated using Rth =— for each of the components of this wall.

л
Convection will be present on each side in the form of an equivalent thermal
resistance calculated with Rh = —, where h is a convection coefficient.
Using the above data, we therefore obtain:
_ 18.10-3 + 510— + - + - = 1.949 m2 KW

[4.129]
= 0.15 0.035 5 5
2) Thermal flow entering the box at the initial time instant
The area which the leaks are spread out over is:
S=2*(0.5+0.3)*0.4+0.5*0.3=0.79m2 [4.130]
The volume of the box is:
VOL =0.5*0.3*0.4=6.10-2m3 [4.131]
The difference in initial temperature is:
AT = 25 -16 = 9°C [4.132]
The initial flow of leakage will be:
AT 9
ф = So = S---- = 0.79*------- = 3.648 W [4.133]
R Rh 1.949
3) The total instantaneous flow entering the box is expressed by:
ф=ST-u t)
[4.134]
Rth
4) Differential equation for the evolution of Tint (t)
For a time dt, the flow of heat entering the box Ф dt leads to heating dTt
such that:
Ф dt = + pCVVOLdTi [4.135]
The plus sign expresses that the incoming heat leads to the heating of the
room.
Replacing Ф by its expression, introducing в , we find the differential

equation:
PairCyVOJdTn- = - S — [4.136]
air V OL dt Rth
dT- = -dd-U =-d— [4.137]

dt dt dt
d—
= -a— [4.138]
dt
with
S
a = p CVR [4-139]
Pair V OL th
a gives
a =-----------07^---------- = 7.26.10-3 [4.140]

1.3*716*6.10-2 *1.949
Solving this first-order linear equation requires a limit condition.

136 Heat Transfer 1
At the initial instant in time:
t=0 ; в = 00 = Te - Tinto [4.141]
d0
= -a0 [4.142]
dt
d0 _
— =-a dt [4.143]
в
dLn0 = -at + LnC [4.144]
where C is a constant. The choice of a constant in Ln C means we can

immediately write:
0( t) = C exp-at [4.145]
The condition at the initial time immediately leads to
C = в0 [4.146]
0( t ) = 00 exp-at [4.147]
5) Maximum authorized length of the photography session:
For Tint =23°C, we have 0= 2°C ; 00 =25 -16 =9°C [4.148]
Calculation of a gives a= 7.26.10-3 [4.149]
0(t) = 00 exp-at gives t =—Ln— [4.150]

0 a00
-12
t =-------- LLn— = 207s = 3mn 27s [4.151]
7.26.10-3 9
Such a short duration is obviously unacceptable.

6) Use in an icebox
a) The outside temperature is still 25°C.
The thermal flow passing through the box is calculated for a constant
temperature:
T- T 25 7
Ф = £_j---- inL = 0.79*-------- = 7.296 W [4.152]
Rth 1.949
b) As long as the internal temperature is constant, heat makes the ice

melt.
There is a mass of:
m = piceVolice = 917*10-3 = 0.917 [4.153]
to be melted, which will require an energy of E=mL for this fusion. A

heating time tfusion is therefore necessary, such that:
. = mL
ф t,fusion [4.154]
In other words:
Ф tfi^n = mL = °.917*333.10 = 41.85.103 5 = 11 h 37 mn 30 5 [4.155]

fusion ф 7.296
This time is obviously more acceptable, which easily leaves time for
transportation and photographs!
EXAMPLE 4.7.- Before fridges, there were iceboxes
In the second half of the 19th century, before refrigerators were

invented, iceboxe5 were an item of furniture that came in two part5. The
upper part held a large block of ice that was regularly changed. Food was
kept cool in the lower part. Equipment of this kind was still used in the
1950s.
138 Heat Transfer 1
A cube of ice measuring a0 = 20 cm is introduced into an icebox. The

cube is of course at a temperature of Ti =0°C, and the atmospheric
temperature is Te =10°C. The heat is brought to the cube by a convection
coefficient h= 10W.m-2.K-1.
The heat for the fusion of the ice is Lf = 333.55kJ.kg -1
Figure 4.2. The icebox, the ancestor of the fridge. For a color
1) What is the flow of heat Ф that is transmitted to the cube (neglecting

the transfer through the shelf which the ice is placed on)?
2) Taking this value of the flow, calculate a time tf for the fusion of the
ice cube.
3) Will this time be the observed fusion time? If not, will the exact result
be higher or lower than that seen here? We will not provide an answer by
calculation, but instead through reasoning.
Do you think that the order of magnitude found validates the concept of
an icebox?
Solution.—
1) The exchange surface is made up of five faces; there is no transfer

through the floor:
5 = 5 a2 = 5* (0,2)2 = 0.2 m2 [4.156]
Ф = 5h(T -Tt) = 0.2*10*10 = 20W [4.157]
2) If this flow is maintained throughout the fusion, then we will have a

fusion time such that
Ф tf = mLf [4.158]
where m = pa0 is the mass of the cube and Lf is the fusion heat of the ice.
This gives
pa03 Lf
tff =------- [4.159]
Ф -
917*(0.2)3*333.55.103
t, = ------ i---- ----------------- = 12.24.104 = 34 h [4.160]
f 20
3) This time cannot be exact. Effectively, the exchange surface between
the air and the ice reduces as the ice melts. The fusion time will therefore be
greater than the time found.
We can attempt to calculate this time; we will take the unknown a(t).
For a time dt, the flow of heat entering the ice cube Ф dt leads to a fUsion
of the mass dm
Ф dt = - dmLf [4.161]
dm is connected to da by:
m=pa3 [4.162]
140 Heat Transfer 1
dm = 2 pa2 da [4.163]
Ф dt = - 2 pa 2 daLf [4.164]
The differential equation for this is therefore:
da _ -Ф
[4.165]
dt 2pa2 Lf
with
Ф= 5a2h(Te -Ti) [4.166]
In other words:
da_ -5a2h(Te - Tt) -5 h(Te - Tt)

[4.167]
dt 2pa2Lf 2pLf
And again
da
[4.168]
dt
with:
5 h(T- Ti) 5*10*10 7

a =---- =---------------------- =8 8.17.10 7 [4.169]
2p Lf 2*917*333.55.103
With the limit condition
t=0; a=a0 [4.170]
The solution is obvious and gives:
a (t) = a0 - at [4.171]
The fusion is complete when a(tf )= 0 , in other words, the time tf :
= ц = —0,2 = 2.45.105 ^ [4.172]

f a 1.63.10-6
tf =2.45.105s=67h58mn [4.173]
The orders of magnitude found here do indeed validate the concept.
Numerical data:
Densities
Density of air: p = 1.3 kg. m -3
Density of agglomerated concrete: p = 1,900 kg. m -3
Density of stone: p= 2,600kg.m-3
Density of steel: p= 7,833kg.m-3
Density of ice: p= 917kg.m-3
Density of air: p= 1.3kg.m-3
Specific heat capacities
Specific heat capacity of agglomerated concrete: c= 879 J.kg-1
Specific heat capacity of stone: c= 881 J.kg-1
Specific heat capacity of steel: c= 465J.kg-1.K-1
Specific heat capacity of air: cV = 716J.kg -1.K-1

142 Heat Transfer 1
Thermal conductibilities
Thermal conductibility of glass wool: A = 0.045 W. m-1. K-1
Thermal conductibility of rock wool: A = 0.045 W. m-1.K-1
Thermal conductibility of agglomerated concrete: A= 1.4W.m-1.K-1
Thermal conductibility of stone: A= 3W.m-1.K-1
Thermal conductibility of steel: A= 54W.m-1.K-1
Thermal conductibility of glass: A= 1.05W.m-1.K-1
Thermal conductibility of air: A= 0.026W.m-1.K-1
Fusion heat of ice: Lf = 333.55kJ.kg -1
EXAMPLE 4.8.- 1812. Retreat from Russia where “war is always cold”!
One month after defeating Kutuzov’s Russian army in Moscow, in

September, Napoleon decided to leave Moscow and bring the “Great Army”
back to Europe. This “retreat” was a disaster due to the harsh winter
conditions and decimated almost all of the imperial army. Today’s problems
will be looked at in this context.
Episode 1. In the isba
A group of retreating grenadiers came across an isba that was still

inhabited. After chasing away its Russian occupants (moujiks), throwing
them out into the cold and misery, they settled into the isba. They discovered
wood and a good number of bottles of Vodka.
Let us recall that an isba (in Russian: изба) is a Russian house

constructed from wood, which was the traditional dwelling of Russian
countrymen for a long time.
The isba is heated by a wood fire that maintains a homogeneous

temperature Ti0 =12°C, which is very pleasant given that the temperature
outside is Te = -30°C. The grenadiers therefore decide to stay in this
dwelling as long as the temperature is maintained.
The walls and the roof of an isba are built from wooden logs. The walls
and roof here will be assimilated to homogeneous walls with a thickness of
e= 15 cm. The geometry and dimensions of this construction are given in the
figure below. From the point of view of thermal losses, we will assimilate
the doors and windows to walls. We will consider that no flow will escape
through the ground.
Figure 4.3. An example of an isba and a model of it for the problem.

The density of wood for heating is p = 675 k.m 3
The combustion of one kilogram of heating wood provides Q= 14MJ.
The thermal conductibility of wooden logs is A = 0.15 W. m-1. K-1.

144 Heat Transfer 1
We will take into account an internal convection coefficient

hi = 5W.m-1.K-1 and an external coefficient resulting from a strong wind
he =30W.m-2.K-1.
1) Calculate the flow Ф of thermal losses through the walls and the roof
of the isba. Deduce from this, the mass of wood m that must be burned each
hour to maintain the temperature in the room.
2) The grenadiers have discovered a stock of two steres of wood in a

storage area. How long will they be able to stay in the isba before continuing
their retreat?
We should recall that a stere of wood is equivalent to a volume of wood

of 1m3 .
Solution.—
1) We must first calculate the overall thermal resistance Rth of the wall.
The wall is simply made of thick wood. The thermal resistance of each
layer is calculated using the equation: Rth =— , applicable to a homogeneous

л
plane wall.
The resistance to the transfer of each convection coefficient will be:
1
[4.174]
h
—= 15.10-2 m
л= 0.15 W.m-2.K-1
Rmur = Л = 1K. mW-1 [4.175]

Using additivity of resistances to transfer, the resulting thermal resistance

is therefore:
Rh = - + — + — = 15.10 +1 + — = 1.233K.m2W [4.176]

th 2 h he 0.15 5 30
The outgoing flow of heat for a surface area S of the isba will be
calculated by:
T-T
Ф = Sisba <P = STRTT [4.177]
Ti- T-
?= iR e [4.178]
Rth
The surface of walls, to which we add the ceiling:
Surface of the facade wall is:
Sfac =2.5*10
. =25m2 [4.179]
Surface of the background wall is of the form:
Sback.wall =3.5*10=35m2 [4.180]
Surface of a lateral wall is of the form:
Sat =(5*2.5) + (52*1j = 12.5 + 2.5 = 15 m 2 [4.181]
Width of the roof: l = .J52 + 12 = 5.1 m 2 [4.182]
Area of the roof: Sroof = 5.1*10 = 51m2 [4.183]
Total surface area of the isba:
Isba =fac
S S +S
fond lat )+S
+(2*S roof
[4.184]
146 Heat Transfer 1
SIsba = 25 + 35 + (2*15) + 51 = 141 m2 [4.185]
p = 12 -(-30) = 34.06 w.m_2

[4.186]
1.233
ф = Sф = 141*34.06 = 4802.5 W [4.187]
Each hour, the following energy is required
E=4802.5*3600W=1.73.107*
J [4.188]
to compensate for losses.
This energy will be given out by the hourly combustion of a mass m of

wood:
.10 = 1.23 kg. hr-i

m = 1.73* [4.189]
14.106
2) The grenadiers have two steres of wood (VOL = 2m3), in other

words a mass
M = pVOL = 675*2 = 1350 kg [4.190]
They will therefore have
1350
tS =------ = 1098 h [4.191]
S 1.23
in other words:
tS = 45.75 days [4.192]
Episode 2. A strong Vodka
One of the grenadiers discovered a full bottle of Vodka, and he puts it in

his bag to warm himself up when necessary.
He goes out of the isba. The outside temperature is as it was before,

Te = -30°C. The air in the bag is at the same temperature.
At this moment, the temperature of the Vodka in the bottle is Ti0 =12°C.
The Vodka, which contains 60% water is likely to freeze. Its volume
then increases, and the full bottle that contains it explodes.
The Vodka freezes at TF =-17°C.
Figure 4.4. The bottle of Vodka and its model. For a color
The bottle is modeled by a parallelepiped of glass with a thickness of

e= 5mm and internal dimensions of 25 cm*8 cm*8 cm . We will only
take into account the conduction on all four sides and the bottom of the
bottle1.
The conductibility of glass is AV = 1.05 W.m-1.K-1, the volumetric mass

of the Vodka is p = 916 kg.m-3 and its specific heat capacity is
c =3, 500J.kg-1.
1 We therefore neglect the losses through the neck and the cork here.
148 Heat Transfer 1
We will consider that the external wall of the bottle is at the external
atmospheric temperature.
1) We will consider that the Vodka is then cooled by a quasi-stationary

process.
a) Write the relationship between the instantaneous flow ofheat Ф(t)

that enters the bottle and the instantaneous internal and external temperatures
Ti (t) and Te respectively.
b) Deduce from this, the differential equation for Ti (t).
2) Solving the differential equation. How long after the grenadier goes
out will the bottle explode?
3) Is the Vodka completely lost?
4) One of the grenadiers did not follow his companions.
He has found a hidden stock of 50 bottles of Vodka that are identical to

those carried by the other grenadiers. He says to himself that by emptying
them onto the fire, he will be able to stay a little longer in the warm hut. He
does this, emptying the bottles little by little to maintain the internal
temperature, constant at Ti0 =12°C.
The combustion of a kilogram of Vodka provides Q'=11.9MJ.kg-1 .
How much longer will the grenadier who has stayed behind be able to
stay in the isba? Is his idea a good one?
Solution.—
1) We presume a quasi-stationary process
a) We calculate the thermal resistance Rth of the walls of the bottle, via
the formula that is applicable to a homogeneous plane wall.
e= 5.10-3 m
2 = 1.05 W.mK-
Riass = 510- = 4.76.10-3 K.m2.W [4.193]

glass 1.05
The internal surface of the bottle through which losses pass is composed:
of four sides with area SC =8.10-2*25.10-2=2.10-2m2 [4.194]
of a base with surface area SF =8.10-2*8.10-2=6.4.10-3m2 [4.195]
In other words, the total exchange area:
S=4SC+SF=8.64.10-2m2 . [4.196]
Presuming a stationary process, the flow of heat Ф that exits the bottle
for a value Tt (t), for an area S = 40 m2 is calculated by:
T-T
p = Sp = S-e------- i [4.197]
b) For a time interval dt, the flow of heat exiting the bottle Ф dt leads to
heating dTi such that:
Ф dt = -pcVOLdTi [4.198]
the Vodka.
Replacing Ф by its expression, introducing the variable в , we find the

в = Tt (t)- Te [4.199]
P-rCVo^T = - S в [4.200]
dt R
150 Heat Transfer 1
which can be rewritten, noting that:
dT =- d (T — T)■ — de
[4.201]
dt dt dt
de
— -a0 [4.202]
dt
with
S
la [4.203]
pcVOLRth
The volume of liquid within the bottle is
VOL = 8.10-2 *8.10-2 *25.10-2 = 1.6.10-3 m2 = 1.61 [4.204]
a gives
a= 8.64.10-2 =3.54.10-3s-1
[4.205]
916*3500*1.6.10-3*4.76.10-3
This equation is associated with the limit condition:
At the initial moment in time
t—0 ; в — 0o — Tinto -Te —12-(-30) — 42°C [4.206]
2) This equation can be resolved classically
de a
— -a0 [4.207]
dt
de „
— —-a dt [4.208]
e
dLnd — -at + LnC [4.209]

immediately write:
0( t) = C exp-at [4.10]
The condition at the initial time immediately gives C = 60 [4.211]
6( t ) = 60 exp-at [4.212]
The bottle will explode when Ti = -17°C,
that is 6=6G = -17- (-30) =13°C [4.213]
This will occur at a time tG such that:
6=6
G0 exp-atG [4.214]
In other words:
tG = -1 Ln6 [4.215]
G a60
-113
tG =-------------------- L Ln [4.216]
G 3,538.10-3 42
tG =331,5s =5mn31s [4.217]
3) As long as it is frozen, the Vodka remains in a block and can be

recovered. All we need to do is make it melt in a bowl. However, we need a
bowl and some fire! We could also put it in a goatskin and place it in the
stomach of a horse that has recently died.
4) Combustion of the 50 bottles provides an energy EV equal to:
Ev = 50 pVolQ ' [4.218]
EV =50*916*1.6.10-3*11.9.106=8.7.108J [4.219]
4802.5 J is needed to compensate for the losses for a second.

152 Heat Transfer 1
The grenadier who stayed behind will have an extra time of
tL = 8.7.10 = 1.81.105 s [4.220]

L 4802.5
tL=50h26mn [4.221]
This is all right. He could have kept one aside to warm himself up in the
snow.
4.5. Applications for axisymmetric systems
EXAMPLE 4.9.- Steel bar and glass sleeve.
Part one. Static study
A steel bar with a diameter di=2ri=5cm is surrounded by a glass

sleeve with an external diameter de1=2re1=12cm. On the outside of the
sleeve, the convection coefficient is h= 8 Wm-2K-1. The thermal
conductibility of the glass is A, = 1.05 Wm-1K-1
The tube is maintained at a temperature TB =70°C. The external

temperature is Te =21°C.
1) Calculate the flow of heat between the tube and the external
atmosphere for a tube length of 1 meter.
2) We replace the glass sleeve of external diameter 12cm by a new glass

sleeve with external diameter de2=2re2=26.3cm. What happens to the
heat flow?
3) We replace the glass sleeve of external diameter 23.6cm by a new

glass sleeve with external diameter de3=2re3=36cm. What happens to the
heat flow?
4) Compare the flows found in questions 1-3. What do you observe?

How do you explain this result?
For use if needed, we give:
Density of air: Pc 1'3 kg'm [4.222a]
Specific heat capacity of air, at constant volume:
cV =716J.kg-1 [4.222b]
Density of steel: Psteel = 7,833 kg•m [4.222c]
Specific heat capacity of steel: csteel =465 J.kg-1 [4.222d]
Thermal conductibility of steel: Asteel = 54 W. m-1. K-1 [4.222e]
Thermal conductibility of glass: Aglass = 1.05 W.mK 1 [4.221f]
Solution.-
1) We must calculate the thermal resistance of the sleeve (which will

be per meter of length), then the thermal flow.
The thermal resistance is given by adding the thermal resistance of the

glass sleeve and the thermal resistance due to convection:
Ln— Ln—
Rh =—- + — = —— + — [4.223]
2nA 2nh 2nAS 2nh
We obtain:
Rth/m =0.464KmW-1 [4.224]
The difference in temperature between the two sides of the sleeve is:
A T = TBe
- T = 70 - 21 = 49°C [4.225]
154 Heat Transfer 1
In other words, the thermal flow evacuated to the exterior per meter of
bar:
TT
Ф, = L----T [4.226]
Rh 1
with L= 1 m [4.227]
49
Ф=-------= 105.6 W [4.228]
0.464
2) We continue the same calculation with the new value of external

diameter de2=2re2=26.3cm. We note that physically, this value has an
influence on the conductive transfer in the glass, as well as on the convection
to the outside.
We then obtain
Rth/m =0.4029KmW-1 [4.229]
Ф= 121.6 W [4.230]
3) Again, we continue by applying the same calculation with the new

value of the external diameter de2=2re2=26.3cm. We note that physically,
this value influences the conductive transfer in the glass, as well as the
convection to the outside.
We also find
Rth/m =0.4097KmW-1 [4.231]
Ф= 119.6 W [4.232]
4) Here, we must think of the phenomenon of “critical radius”.
We see that:
- from 12 cm to 26.3 cm, the diameter increases and the losses increase;
- from 26.3 cm to 36 cm, the diameter increases and the losses reduce.
We know that when we increase the thickness of the insulating sleeve,

there is competition between the increase in insulation, which tends to
reduce losses, and the increase in the external lateral surface area of the
entire object, which tends to increase losses. When the external radius is less
than a “critical” radius, the competition is to the advantage of the external
surface area and if the thickness of insulation in this zone is increased, this
corresponds paradoxically to an increase in losses. When the external
radius is higher than the critical radius, the trend is inverted and
increasing the thickness of the insulation then contributes to reducing losses.
In this case, the value of the critical radius is:
2 54
rc = - = — = 0.131 m = 13.1 cm [4.233]
Ch8
At de2=2re2=26.3 cm of diameter, we are practically at the critical

radius
Therefore:
- for de < rc = 26.2cm (from 12cm to 26.3 cm), the diameter increases
and the losses increase;
- for de>rc=26.2cm (from 26.3cm to 36 cm ), the diameter increases
and the losses are reduced.
Part two. Steel bar and glass sleeve; evolution over time
We will look again at the steel bar from the previous example,
surrounded by the glass sleeve with diameter de2=2re2=26.3cm. The
external convection coefficient is still h= 8 Wm-2K-1.
In addition, we give:
Volumetric mass of steel: pS = 7,833 kgm-3
Specific heat capacity of steel: cS = 465 J kg-1K-1

156 Heat Transfer 1
In fact, the thermal losses lead to a reduction in the temperature of the

bar, in line with a law TB (t) that will be calculated. To do so, we will use a
quasi-stationary balance calculation.
We will neglect the thermal losses through the extremities of the bar and
the glass sleeve.
1) Write the relationship between the flow exiting the bar per meter of
length, Ф, T— (t) and Te
2) Deduce from this, the differential equation that connects —— to TB

dt
and Te .
Then, deduce the equation for ——.

dt
3) Give the expression for T— (t).
4) After a certain time period t0,5 , will the temperature of the bar have
reduced by half? Give this time in hours, minutes and seconds.
We recall, to be used as needed:
Density of air: pc = 1.3 kg. m -3
Specific heat capacity of air, of constant volume: cV = 716 J.kg-1
Density of steel psteel = 7,833 kg. m-3
Specific heat capacity of steel: csteel = 465 J.kg-1
Thermal conductibility of steel: Asteel = 54 W. m-1. K-1
Thermal conductibility of glass: Aglass = 1.05 W.m-1.K 1

Solution.—
1) By definition of the thermal resistance:
TT
ф= TB—— [4.234]
We had calculated:
Rth/m =0.4029KmW-1 [4.235]
2) A thermal balance for a meter of bar gives us the differential

equation, which is satisfied by TB :
dT T-T
PSC—B = —B------ e- [4.236]
dt Rth
with the initial condition:
t=0;TB=TBi=70°C [4.237]
The sign “ - ” in the equation indicates that the flow of heat (positive)
contributes to the reduction in the temperature; the derivative is therefore a
negative term.
3) Resolution of this equation is classic:
The variable 0 = TB - T is introduced [4.238]
The equation becomes:
d00
Psc OL dt ~ Rth [4.239]
VOL is here the volume of a meter of bar, in other words:
(
VOL = nri2 = п 2.5.10-2 )2 = 1.96.10-3 m3 [4.240]
158 Heat Transfer 1
or:
do dt
— =---------- = - adt [4.241]
0 psCRth
1
with a = [4.242]
psc VOL Rth
with the initial condition that becomes:
t=0 ; О = TBi - Te = 49°c [4.243]
LnO =-at + LnC [4.244]
A constant Ln C appears, which will be given by the initial condition.

The choice of the constant in log form is a trick that allows us to simplify the
change to an exponential for the preceding expression. Taking the
exponential for each term:
exp (LnO) = exp (-at) exp (LnC) [4.245]
О= Cexp(-at) [4.246]
at
t= 0 TBi -Te =C=49 [4.247]
Hence, the final expression for the temperature of the bar:
TB(t)-T=(TBi-T)exp(-at) [4.248]
4) We calculate
a =----- 1----- =------------------ 1------------------= 3.47.10-4 [4.249]

pcVOLRth 7833*465*1.96.10-3*0.4029
We want to change the temperature from TB = 70°C to TB = 35°C
in other words, TBi - T from TBi -T=49°C to TBi -T=35 -21 = 14°C
The time required, tS , corresponds to the relationship:
14 = 49exp(-atS) [4.250]
or even:
14 = 49exp(-atS) [4.251]
-114 1 49
tS =— Ln— = Ln— [4.252]
S a 49 3.47.10-4 14
tS =3610s=1hr10s [4.253]
EXAMPLE 4.10.- Insulation of a laboratory chemical reactor
A chemist wishes to maintain a mixture containing m= 15kg of reagents

for a minimum time of one half-hour at a constant temperature, to the
nearest one and a half degrees.
To do so, he places his reagents into a pyrex cylinder with an internal

diameter di and external diameter de . The tube is horizontal and of length
L . The tube is fixed by thick supports on its right-hand and left-hand faces.
Throughout the problem, we will neglect the transfers through these two
lateral faces.
In the calculations, we will neglect flows of quantities of heat that are

required for the variations in temperature of the pyrex.
The atmospheric temperature in the laboratory is set by the administration

to Te =19°C. The initial temperature of the reagents is Tri =100°C.
We will apply an external convection coefficient h.

160 Heat Transfer 1
1) The chemist is not a great physicist; he begins his experiment with

the simple glass tube.
a) Calculate the thermal resistance per meter of the glass cylinder, Rth1 ,
and deduce from this the flow Ф verre 1 that escapes from the reagents at the
initial instant of the reaction.
b) Give an approximate estimation of the reduction in temperature of the

reagents in a second right at the beginning of the experiment.
Do you intuitively believe, without any other calculation, that the

temperature can be maintained constant to the nearest degree for two hours?
2) On the basis of good advice from one of their physicist colleagues,

they seek to improve the process by encasing its reactor in a rock wool
sheath of thickness e.
We have T(t) as the temperature of the products during cooling.
a) Give the expression2 for the resultant thermal resistance Rth2 for this
new reactor. Calculate its numerical value.
b) Establish that, taking into account this new value Rth2 , the flow
Фisol (t) exiting this new reactor configuration can be written as
Ф = в( T - Te); в is a coefficient whose numerical value will be given.
c) Then, write a differential equation for T (t).
d) Give the expression for T(t). To do so, it will be useful to define a

variable 0( t ) = T (t)- Te
e) By how much will the temperature of the reagents reduce after half an
hour? Who is better, the physicist or the chemist?
2 Since the thickness is not known, this question will be treated literally up to e.
Numerical data:
di = 5 cm
e = 6 cm
d
e= 3cm
L=1.5m
Tri =100°C
Te =19C
h= 5W.m-2.K-1
m= 15 kg
Specific heat capacity of the reagents: c =4, 000J.kg-1
Thermal conductibility of the rock wool: ALR = 0.045 W. m-1.kg-1
Thermal conductibility of pyrex: APyr = 1.13 W. m-1.kg 1
Solution.—
1) We begin the experiment with the simple glass tube
a) The thermal resistance RthLR per meter of a cylindrical sleeve with

internal radius r1 , external radius r2 and thermal conductibility A is
calculated from the expression:
Rth [4.254]
2 nA
to which the resistance to the transfer due to convection is added

162 Heat Transfer 1
Rconv =1T^ [4.255]

2 n re A
Rth1 is calculated using the equation drawn up in the example
r 6
hr 1 Ln 5 1
r
1 " 2nA + 2nreh 2n3 2n 3.10-2, .5
= +
=2.5679.10-2 +1.061=1.0867Wm-1K-1 [4.256]
Ti-Te=100- 19=81° [4.257]
T-T 81
Ф = L^---- e- - = 1.5.-------- = 111.81Ж [4.258]
Rth 1 1.0867
b) To reduce the temperature of the reagents by one degree:
Q = mc =15.4000=60,000J [4.259]
At = 15 ; AT ~ H1!81 = 1.8610-3° [4.260]

60000
1.5 degrees over half an hour represents “on average” 8.33.10-4 °s-1.
Significant doubts remain over the order of magnitude. A more
comprehensive calculation is required.
2) We are attempting to improve the process by encasing its reactor in

a rock wool sheath of thickness e.
a) We will calculate the thermal resistance R

th2 that results from this new
reactor.
Lnr r+e 3
Ln—---- Ln Ln 6
r 1 =2.5 +---------- 3----- + 1
Rth1 = ri + +' e
2nA pyrex 2nALR 2nre h 2n*1.13 2n*0.045 2n*6.10-2*5
=2.5679.10-2 +2.451+0.5305=3.0077 [4.261]

re'=6.10-2m [4.262]
b) The flow Фisol (t) exiting this new configuration of reactor can be
written as:
ф = в(T - Te) [4.263]
TT TT
Ф= L—--- e- = 1.5.—---- e- = 0.499(T -T) [4.264]
Rth1 3.0077
Deducing from this: [4.265]
в = 0.499 [4.266]
c) Heat balance: by writing that the heat exiting per unit of time cools the
reagents, we draw up the differential equation for T(t)
mcdTi =-Ф [4.267]
Ф = в (T - T) [4.268]
with
в = 0.499 [4.269]
mcdT =-в (T - Te) [4.270]
d) We will introduce the variable в = (Tt - Te) [4.271]
Ф = в(T. -Te) = ре [4.272]
в = 0.499 [4.273]
164 Heat Transfer 1
mcdT = -Ф = - в (T — T.) [4.274a]
0 = ( T - t. ) [4.274b]
dO
= -a0 [4.274c]
dt
в 0.499
a=в =---------- = 8.31.10-6 [4.275]
mc 15*4000
t=0 ; 0=00 =100-19=81° [4.276]
d0
= -adt [4.277]
0
Ln0= -at +Ln C [4.278]
0=Cexp-at [4.279]
0=00 exp -at [4.280]
e) By how much will the temperature of the reagents reduce after half an
hour.
0=00 exp -at [4.281]
00 =81° [4.282]
a= 8.31.10-6 [4.283]
t =0.5hr =1,800s [4.284]
0=81exp-8.31.10-6.1,800=0.985.81=79.98 [4.285]
A T = 1.2 ° [4.286]
Mission accomplished. The physicist is indeed better.

EXAMPLE 4.11.- A good beer must be cold
A bottle of beer is assimilated to a glass cylinder3 of height H= 26 cm,

with internal diameter Di = 6cm and thickness e= 5mm.
On a hot summer day, a bottle of beer has been left next to a window.
The temperature of the beverage has reached T0 =25°C.
To drink a cold beer, at TF 1 =6°C, a tourist places the bottle in the main
body of the fridge, which is at a temperature of Te1 =4°C. We will consider
that the natural convection imposes a uniform convection coefficient on the
external surface of the bottle, he = 5W.m-2.K-1. We will also consider that
the temperature Ti (x) of the beer is uniform at all times in the bottle and
that, as a consequence, the cooling process is quasi-stationary.
In addition, we point out that the beer freezes at TG =-2°C and that, in
this case, the beer dilates and the bottle explodes.
Figure 4.5. The bottle and its model
1) The bottle is placed in the fridge
a) Calculate the thermal resistance Rth per meter of length of the glass in
the bottle.
3 Here, we will neglect the thermal effects of the neck and the bottle cap.
166 Heat Transfer 1
b) Calculate the thermal flow Ф0 coming out of the bottle at the initial
moment in time; in other words, when the bottle is placed in the fridge.
2) The beer cools
a) Give the differential equation that satisfies the temperature Ti (x) of

the beer in the bottle.
b) Give the expression for Ti (x).
c) How long will the tourist have to wait before drinking their cold
beer?
3) Since this time is too long, the tourist’s wife, who is not really a
physicist, places a second bottle of beer, also at T0 = 25°C in the freezer
compartment, where the temperature is Te2 =-15°C.
She thinks that the beer should be left for a shorter period in the freezer
and leaves it for a time t2 = 1 hour
a) Will she drink a very cold beer?
b) Estimate the minimum time necessary for the bottle to explode.
c) What do you think of all this?
Figure 4.6. The can and its model

4) Looking again at questions 1 and 2 for a can of beer, it is

assimilated to an aluminum cylinder with height H= 30cm, internal
diameter Di = 6cm and thickness e=1mm.
We will conserve the same convection coefficient he .
Additional data:
Volumetric mass of the beer: p = 1,000 kg. m -3
Specific heat capacity of the beer: c= 4,180J.kg-1
Thermal conductibility of the beer: Л = 0.6 W. m-1. K-1
Thermal conductibility of the glass: Л = 1.05 W.m-1.K-1
Thermal conductibility of aluminum: Л = 204 W.m-1.K-1
Solution.—
1) At the initial moment in time
a) We calculate the thermal resistance Rth of the sides of the bottle,

using the equation:
rr
Lnre Lnr
__ T_ =__ [4.287]
2пЛ 2пЛ
ri = 3.10-2 m
re = 3.5.10-2 m
Di = 6.10-2 m
De = 7.10-2 m
Л=1.05 W.m-2.K
168 Heat Transfer 1
r
Ln Ln7
---- r- =------- 6— = 2.34.10-2 W.m-2.K [4.288]
2nA 2*n*1.05
b) The total thermal resistance including the convection coefficient is:
r
Ln r
tot = --- rr +----1 = 2.34.10-2 +---------- 1----- ------
t 2nA 2nrh 2*n*3.3.10-2 *20
= 2.34.10-2 + 0.241 = 0.264 W. m -\K-1 [4.289]
The flow of heat Ф exiting the bottle for an initial value Tt 25°C is
calculated for a cylinder length H= 26 cm, in other words:
T-T 25-4
Ф0 = H—---- e = 0.26*------- = 20.7W [4.290]
0 Rh 0.264
2) In a quasi-stationary model, the instantaneous flow exiting the bottle is

written as:
V(t)= hT Ro Te ( [4.291]
a) For a time dt , the heat flow exiting the bottle Ф dt leads to heating
dTi such that:
Ф dt = - pcVOLdTi [4.292]
the liquid.
Replacing Ф by its expression and introducing the variable в , we find

the differential equation:
в = T. (t)- Te [4.293]
PbeerCVoLdT =-H /- [4.294]

dt Rtot
dTi d (Ti - Te) de

----- = = [4.295]
dt--------- dt---------- dt
de
=-ae [4.296]
dt
with
H
a =----------- [4.297]
PcVoR
The volume of liquid contained in the bottle is
(
VOL = n* 6.10 ) *26.10-2 = 7.35.10-4m3 [4.298]
a gives
a =------------ 26.10-\---------- = 3.20.10- s -

[4.299]
1000*4180*7.35.10-4 *0.264
This equation is combined with the limit condition:
At the initial instant:
t=0;e=e0=Ti0-Te=25-4=21°C [4.300]
b) This equation can be resolved classically:
de
=-ae [4.301]
dt
170 Heat Transfer 1
dO = -a ,
„ dt [4.302]
O
dLnO = -at + LnC [4.303]
C is a constant. The choice of a constant in Ln C allows us to

immediately write:
O(t)=Cexp-at [4.304]
The condition at the initial time immediately gives C = O0 [4.305]
O(t) =O0 exp-at [4.306]
c) The bottle should be taken out when Ti=Tf=6°C,
so O=Of =6-4=2°C [4.307]
This will happen at a time t1 such that:
Of =O0 exp -at1 [4.308]
In other words:
t1 =-! Ln0-
[4.309]
1 aO0
-12
[4.310]
1 3.2.10-4 21
t1 =7,348 s=2hr2mn [4.311]
3) The bottle is in the freezer compartment
a) The new value of O0 is O0 =25 - (-15) =40°C [4.312]

The temperature of the beer after t2 =1hr =3600s will then be
Of = 00 exp-at2 = 40exp(-3.2.10-4*3600) [4.313]
Of =12.64 [4.314]
Tf =12.64+ (-15) = -2.36°C [4.315]
And no, the bottle will have exploded. Disaster!
b) We can estimate the freezing time
The bottle will explode when Of =- 2 -(-15 ) = 13° C, in other words

after
-131
t =------- l Ln— = 3512 s = 58.5 mn [4.316]
exp 3.2.10-4 40
We are not far off, it is a shame. Disaster again, but only by a small
margin!
c) Unlucky
4) In the case of the metal can, it is necessary to look at the problem

again, by changing the value of the thermal resistance of the container.
ri = 3.10-2 m
re = 3.1.10-2 m
Di = 6.10-2 m
De = 6.2.10-2 m
1 = 204 W.mK—
Lnr-e 6 6.2
Rh = 1= П ^2 = 2.56.10-5 W. m ~2. K4
—
[4.317]
2n1 2*n*204
172 Heat Transfer 1
We see that the thermal resistance of the metal will be negligible in

comparison to that of the convection coefficient
r
Ln r
== rr + 1---- = 2.56.10-5 +---------- 1-----------
tot 2пА 2nrh 2* п*3.1.102*20
=2.56.10-5+0.257=0.257W.m-2.K-1 [4.318]
The initial instantaneous flow will be:
TT 25 4
Ф0 = Н-i------- e- = 0.3*-------- = 24.5 W [4.319]
0 R 0.257
The volume of liquid has been modified slightly:
V = п* (6.10
OL 4
) *0.3 = 8.48.10-4 m3 [4.320]
We will have a value close to the previous one:
a =------------- 30.10----- ---------- = 3.29.10-4 5 - [4.321]

1000*4180*8.48.10-4*0.257
And the cooling time will be modified a little in the end:
-:1---- Ln-2
t =------
[4.322]
1 3.29.10-4 21
t1 = 7147 5 =1hr59mn [4.323]
COMMENT.-
The value chosen for the convection coefficient he = 20W.m-1.K-1 may

appear a little high for pure natural convection.
We could argue that in certain fridges, the air is mixed around.

Put more simply, we admit that this value of he is an “ad hoc” value,
which involves taking into account two phenomena that intervene in two
areas of the box or the bottle:
- natural convection through the top;
- direct contact of the bottle with the shelf, which necessarily imposes the
temperature Te =4°C on part of the external wall of the recipient and which
automatically eliminates all convection effects on this zone.
Example 4.12.- Thermal audit of a yurt
As and when necessary, the numerical data provided at the end of

this problem will be used.
A Mongol yurt is mainly made up of a wall of sheep’s wool felt of

thickness e , topped with a roof, which is also made of felt.
Remarks.- As opposed to a Kyrgyz yurt, a Mongol yurt is always a

permanent habitat. Felt is an unwoven fabric made by pressurizing and
boiling natural fibers, sometimes with a chemical treatment.
A yurt will be assimilated to a cylinder of diameter D on the base and a

height H . The roof will be assimilated to a plane disk.
The convective transfers in effect have an inside convection coefficient

hi and an outside convection coefficient he . These coefficients are identical
for the vertical walls and the roof.
We will neglect the thermal losses through the ground.
To heat the yurt, the head of the family has lit a wood fire that consumes
mkg of wood per hour. A kilogram of wood provides an energy Q by
combustion. We largely consider that only 50% of heat produced passes
through the felt by conduction. The remaining 50% exits directly with the
smoke and has no incidence on the temperature in the yurt.
We consider that when the fire is lit, we are in a stationary regime. The
internal temperature of the yurt Ti is constant. The outside temperature is Te .
174 Heat Transfer 1
Figure 4.7. A yurt and its heating. For a color version of

this figure, see www.iste.co.uk/ledoux/heat1.zip
1) We are attempting to evaluate the conductive thermal losses in the

yurt. To do so, we will use two hypotheses.
a) In an initial model, we consider that the thermal resistance per m2 of

the lateral area of the yurt can be calculated in the same way as a plane felt
wall of the same thickness (and of course of the same length as the central
pillar in the yurt)4. The roof obviously remains a plane felt surface. Deducing
from this, the flow that passes through the walls of the yurt for a temperature
difference between the inside and the outside of 22°C.
b) In a second model, we take into account the cylindrical form of

vertical walls. To do so, we calculate the thermal resistance of the lateral
walls per meter of length. Obviously, the roof remains a plane surface.
Calculate the outgoing conductive flow over the entire felt surface area of
the yurt in the case where the temperature difference between the inside and
the outside is 22°C.
c) Are the values found in a and b so different? Are you surprised?

Which of the two calculations seems to be the most suitable to you?
4 This approximation is inspired by the construction of a “standard” yurt that is often carried
out using five plane vertical wall panels.
2) Given the answer to the previous question, show that the correct
expression for the instantaneous flow exiting the yurt as a function of the
difference between the internal and external temperature takes the form
Ф = в(Tt - T) [4.324]
where в is a coefficient whose numerical value will be given.
3) How much wood will it be necessary to burn in 24 hours to maintain

the internal temperature of the yurt at 14°C?
4) With no stock left, the head of the family went to get some wood in his
old Land Rover. He will take a whole day to do this. Will the family suffer
from the cold?
a) Write the differential equation that is obeyed by the internal

temperature Ti(t).
We will maintain the relationship written in 2 when Ti (t) is variable over

time (quasi-stationary calculation). The variable в = (Tt - Te) will be
introduced to useful effect.
b) Give the literal expression for в (t) .
c) We consider that below 5°C, it is not possible to live in the yurt. How
long will it take for this temperature to be reached?
Will the head of the family come back in time to keep the family
comfortable?
5) Thermal science of sport.
This question can be treated independently from question 4.
We consider that a fit person pedaling vigorously on a home-trainer gives

off heat of power P.
A Mongolian sports club rented a yurt that is identical to the one seen in
previous questions and installed 10 home-trainers there. The club does not
176 Heat Transfer 1
have the means to heat the yurt and counts on the heat given off by the
cyclists to maintain an acceptable temperature in the tent.
Remark.-
There is now no fire, so no losses via smoke going up the “chimney”.
The outside temperature is Te =5°C.
a) What is the inside temperature during the stationary regime when 10

Mongol cyclists pedal inside the yurt?
b) Is it a good idea to not have installed heating?
Additional data:
D=6m
H=2.3m
e= 10mm
Te =-8°C
Thermal conductibility of the felt: AF = 0.035 W. m-1. kg-1
hi = 5W.m-2.K-1
he = 16W.m-2.K-1
Volumetric mass of air: p = 1.3 kg. m-3
Specific heat capacity of the air: cv = 718J.kg -1.K-1
Q= 14MJ.kg-1
Heating per cyclist: P= 400W

Solution.—
1) We are seeking to evaluate the conductive thermal losses in the

yurt, looking at two models: approximation of the plane wall, cylindrical
(sleeve) wall.
a) Plane wall
The thermal resistance per m2 of the walls of the yurt is:
Rh =- + — + — = -10- +1 + — = 0.548 W m2°K [4.325]

th X ht he 0.035 5 16
„ nDD- .
5= + nDH = 28.27 + 43.35 = 71.62 m 2 [4.326]
AT 22
Ф = 5---- = 71.62--------= 2875 W [4.327]
Rth 0.548
b) Cylindrical wall
The thermal resistance per meter of length of the cylinder is:
Ln 1 De-1 T (6.02^
Ln 1 1
Rth = +
1 Di J
2nXF
i 1
2nrihi
+
1
=
1 6 J
2nrehe 2n* 0.035
+
1 —
2n*3*5
+
1
2n*3.01*16
=1.5110-2+1.0610-2+3.3110-3=2.9.10-2W-1m2°K [4.328]
The flow will therefore be as follows, summing up what passes through

the cylinder and what passes through the circle (roof)
nD2
SRoof = = 28.27 m2 [4.329]
4
Д T = 22 ° [4.330]
ДT
Ф = 28.27 +2.3-------- - =1134.9+1744.8=2897.9W [4.331]
0.548 2.9.10-2
178 Heat Transfer 1
c) Are the values found in a and b close?
The thickness e is small in comparison to the radius of the yurt. We can

demonstrate that the two forms of the flow are equivalent if
r ~ re Ln— = Ln 11 + -el = e [4.332]

r l rJ r
The form that takes account of the cylindrical geometry is more suitable.
2) The form that takes account of the cylindrical geometry is more

suitable.
Therefore:
AT
Ф = 28.27 +2.3---------r =в ( T - T) [4.333]
0.548 2.9.10-2
в=130.9 [4.334]
3) What is required to maintain the internal temperature of the yurt

at 14°C?
1 kg of burned wood provides 4.106 J .
Through conduction, in 24 hours, we lose:
2898*3600*24=2.504.108J. [4.335]
In total, we therefore lose (conduction + evacuation of the smoke) double

this, in other words 5.077.108 J .
5.077.108
It is therefore necessary to burn: = 35.77kg of wood every 24
14.106
hours.
4) The family remains without fire for an entire day.
a) We will write the differential equation that is obeyed by the inside

temperature Ti(t).
We define 0 = ( Tt - Te) [4.336]
Ф = в (T — Te) [4.337]
в = 130.9 [4.338]
DD-
Vol = n—H = 65.03m3 [4.339]
4
pVoiCdT=-ф=-в ( t - Te) [4.340]
0=(Ti-Te) [4.341]
d0
= -a0 [4.342]
dt
a = —в— = 2.156.10"3 [4.343]
b) That is the literal expression for 0(t) obtained by resolving the

differential equation.
a = —в— = 2.156.10"3 [4.344]
d0
=-a0 [4.345]
dt
180 Heat Transfer 1
t = о ; о = 00 = 14 -(-8) = 22°C [4.346]
d0 = -a
— „ л
dt [4.347]
о
Ln0 = - at + LnC [4.348]
0 = C exp - at [4.349]
о=о0 exp -at [4.350]
c) Below 5°C, it becomes impossible to live in the yurt.
If Ti = 5C, then 0= 13 [4.351]
This intervenes at a time t such that:
a = —в— = 2.156.10"3 [4.352]
13 =22exp-at [4.353]
113 1
t =—Ln _ =--------- - (-0.526) [4.354]
a 22 2.156.10-3
t= 244.01s [4.355]
The Mongol family will suffer from the cold.
5) Thermal science of sport
a) Internal temperature in the stationary regime when 10 cyclists are

pedaling
Relationship between flow and temperature:
Ф= 130.9 (Tt - Te) [4.356]

Ф= 10x400 = 4,000 W [4.357]
T -t = 4000 = 30.56°c [4.358]

ie130.9
Ti =30.56° -8=22.56°c [4.359]
b) It is a good thing that extra heating was not installed inside the yurt.
But after their efforts, the cyclists should not stay for too long inside the yurt
because it will cool down.
5
Non-stationary Conduction
In non-stationary conduction problems, we must take into account

changes in temperature over time. The problem can be very complex.
Indeed, it is the temperature that is the solution to the complete heat

equation:
дT
(
р c---- = div A grad T
дt
) [5.1]
the temperature field T is therefore a function of four variables, three in

space and one in time, depending, for example, on the geometry of the
problem: T = T(x,y, z, t) or T=T(r,0, z, t). We say that the problem is a
four-dimensional problem.
In addition to this, generally, the physical properties of the material can vary.
We understand that the most complex cases will require a digital

approach. In this chapter, we will focus on an analytical approach. To do
this, we will consider single-dimensional problems in plane geometry.
5.1. Single-dimensional problem
We consider an environment of semi-infinite (undefined) length that has

a plane interface with its exterior. The problem is presumed to be single
dimensional. In particular, on all planes parallel to the interface, the
temperature is considered at all times to be constant. We locate the position
on an Ox axis. The temperature field at an instant t will therefore be given
184 Heat Transfer 1
by an expression T=T(x,t) that must be determined as a function of the

conditions imposed at the entry interface.
5.1.1. Temperature imposed at the interface at instant t =0
This is the prototype problem in non-stationary conduction. For this

reason, we will focus more on it and will provide a full analysis leading to a
classic result that is part of the “basic culture” of all thermal scientists.
REMARK.- A less well-informed reader, or one who is allergic to

mathematics, should not worry. Simply knowing the expression for the
temperature field obtained in the following paragraph, they will understand
many examples proposed below.
The environment is presumed homogeneous with constant physical

properties as the temperature varies, in particular the volumetric mass p,
thermal conductibility 2 and specific heat capacity cP .
Remark.- Again we note that a parameter derived from the previous chapters
has appeared, and that this will have a pivotal role in all non-stationary
conduction problems, thermal diffusivity a, defined by:
a =— [5.2]
a plays a key role and is fundamental in problems with constant

physical properties.
In all of these types of problems, we often introduce a reduced

temperature #( x, t) that can be defined in different ways.
We will only require that this reduced temperature is a linear function of

T , in other words, explicitly in the form
e(x,t) = T(x•t) — T [5.3]

' ' T2 — T3
Non-stationary Conduction 185
where T1 , T2 and T3 are temperatures, and which are not all necessarily
different, selected depending on the problem in question, and, in particular,
on its limit conditions.
We note that the term reduced applies, in principle, to adimensional

variables. We will also find derived variables that will retain their dimension
as a temperature.
In this problem, we consider an environment at the initial temperature Ti .

We impose at the initial moment in time a temperature Te at the interface
between the environments and the exterior.
Determination of the temperature field T= T(x, t)is a question of

solving the single-dimensional heat equation, which is written:
dT d2 T
= a------ [5.4]
d t-----d x2
with which we associate the limit conditions:
T(0,t)=Te [5.5]
T(x,0)=Ti [5.6]
x^^ ; T^T [5.7]
REMARK.- We should recall that the heat equation results from a thermal
balance and the limit conditions from the specific conditions of an
experiment.
We will introduce an unknown function that takes the form of a reduced

temperature, which is linear with unknown temperature, defined by:
e( x,, )=T (x,t)-Te [5.8]

Tie-T
186 Heat Transfer 1
The problem is rewritten as:
d0 d2 0
---- = a----- [5.9]
дt д X2
with which the following limit conditions are associated:
0(0,t)=0 [5.10]
0(x,0)=1 [5.11]
x^^ ; 0^1 [5.12]
We are looking for a solution where the reduced temperature depends

exclusively on a composite variable
£ = ~^= = ~F^= [5.13]

2y[Ot 4^ at
Of the two ways of writing it, we generally prefer the second.
In other words,
0(x, t ) = 0(£) [5.14]
REMARK.- This variable does not appear intuitively. We can look for its
expression in the form 0(x, t) = axm tn. Here, we do not give the
calculation and advise curious readers to work through it themselves.
The heat equation, which initially has two variables, will be transformed
into an equation with a single variable. We will therefore go from a written
expression in the form д to a written expression in d.
The derivatives are transformed as follows:
д0(x, t) _ d 0(^) д ^
[5.15]
дt ~ d^ дt
Э0( x, t) _ d0(£)d d
[5.16]
dx d2 dx
With:
d__<Ц I t-1 _ I (-11-21 _ -I I_-d

[5.17]
dt dt jaa Jaa ^ 2 J 2jaat t 21
d_ _д[ x _ 1
[5.18]
d x dx 4 4at 41at
The terms of the equation are transformed into:
Э0(x, t) -d d 0(d)
[5.19]
dt ~ 21 d
d0( x, t) 1 d0(d)
[5.20]
dx 4 a at dd
Э2^(x, t )_ d [ d0(d) ~ 1 d Г d0(d) 1

d x2 d x _4 4at d 2 4 a at d x L d ^ J
1 1 d d0(d)
[5.21]
4 а at 4 aat d d dd
d20( x, t)_ 1 d 20(d)

[5.22]
d x2 4 at d d 2
And the equation becomes:
- d d0(d) _ 1 d 20(d)
[5.23]
21 dd 4 at d 222
188 Heat Transfer 1
We will note, for the purposes of simplifying the expression:
d d°^
° = d{ [5.24]
°._ d2°(Я
° = d? [5.25]
which can be written as:
- %°' = a— °" [5.26]

4at
or even:
°" + 2 $0' = 0 [5.27]
The solution to this equation will be obtained by integrating twice, which

leads to the appearance of two constants. These constants are determined, as
a function of the problem in question, by the limit conditions.
°" - 2$0' = 0 [5.28]
integrates into:
Ln °1 = - ? + LnC1 [5.29]
We have chosen to express the constant by a logarithm to simplify the

following expression:
°1 = C1 exp- ^2 [5.30]
° = C1 j^ exp - u2 du + C2 [5.31]
REMARK.- For practical reasons, we have chosen to express the solution

using a defined integral with a lower limit that is arbitrarily zero. This lower
limit will not be zero for all problems and is corrected by the constant.
The expression found for 0(Q) encourages us to remind ourselves of the

definition of the error function
erf Q = ^= j^ exp- u2 du [5.32]
in addition to the property:
j exp- u2du =—n [5.33]
which ensures that the error function is normalized to 1.
The limit conditions are written using the variable
x
i = ,----- [5.34]
4 at
Application of the limit conditions indicates to us that:
0(0) = 0 [5.35]
0(7 = 1 [5.36]
Hence:
Q
0=C1jO exp-u2du+C2 [5.37]
0
0 = C1 j^ exp- u2 du + C2 ^ C2 = 0 [5.38]
1 = C1 exp- u 2 du + C2 [5.39]
Hence:
1=C' =Т—1-------- ="Г [5.40]

jo exp- u2 du n
190 Heat Transfer 1
The reduced temperature is therefore given by:
u2 du = erf (£) [5.41]
We can, if required, re-express T(x, t) as:
( x \
T(x,t)=(Ti-Te)erf +Te [5.42]
To illustrate this result, we show the evolution of the temperature in

Figure 5.1 in a material with thermal diffusivity a= 5.5.10-6 m.s-2
Figure 5.1. Evolution of a “thermal ladder”. Representation of the reduced

temperature. 0( x, t) = T Txx—)T T1 For a color version of this figure,
see www.iste.co.uk/ledoux/heat1.zip
5.2. Non-stationary conduction with constant flow density
This problem calls into question a wall that has a plane face. The physical
properties of the material are presumed invariable, that is, following the
usual notation of this book, p, c, 2.
We impose a constant thermal flow density ф0 on the plane face. The

initial temperature of the material, homogeneous throughout, is Ti.
t=0
Figure 5.2. Non-stationary conduction: constant flow density imposed

on a face at the initial instant. For a color version of this figure, see
www.iste.co.uk/ledoux/heat1.zip
The transfer is presumed to be single-dimensional (the plane surface is

large in comparison with the dimensions of the problem). The temperature
will therefore be distributed into isothermal planes, parallel to the contact
surface.
We locate these planes on an axis Ox, with origin O on the surface of

the wall where the temperature is imposed. The wall is of indefinite length.
Its local temperature (on a given plane) is T =T(x,t).
The problem is characterized, as always, by a local balance equation

expressed by the heat equation. Here it is in a single-dimensional
non-stationary form, associated with the limit conditions that determine the
specific conditions.
192 Heat Transfer 1
Hence, the equation:
dT d2 T
---- = a------ [5.43]
dt d x2
where a is the thermal diffusivity of the material,
X
a =----- [5.44]
p cP
The limit conditions express that the temperature away from the surface
of the wall is not affected by the modulation; the second expresses the
temperature condition imposed.
In other words:
T (~, t ) =T [5.45]
dT T [5.46]
- ^dx = ^°’
Here, we give the solution, which is, this time, in dimensional form:
у/
T ■ x
2 ^0 at1 t Zierfc
(x, 2t\) - T2 =------ x [5.47]
X ya4at7
Here, we see a function ierfc % appear, where % = - ---- , defined by:

4 at
ierfc % = —;= exp (-%)- %erfc % [5.48]

П
where erfc % is the complementary function of erf %, defined by:
erfc % = 1 - erf % [5.49]

The functions erf %, erf % and ierfc% are tabulated in Appendix 5. These
three functions are represented in Figure 5.3 as well as the Appendix:
Figure 5.3. The functions erf %, erf % and ierfc%
5.3. Temperature imposed on the wall: sinusoidal variation
This problem involves a wall with a plane face. The physical properties
of the material are presumed invariable, that is, following the usual notations
of this book, p, c, Л.
Here, we do not consider a problem beginning at an initial instant in

time, but instead the results of a temperature variable on the transfer
between the outside and the material.
This temperature will have a fixed value to which a sinusoidal

variation is added.
In this type of problem, when we impose this type of condition, we have

first a transitory solution, then an established regime. Here, we are interested
in this established regime.
194 Heat Transfer 1

large with regard to the dimensions of the problem). The temperature will
therefore be distributed in isothermal planes, parallel to the contact surface.
We locate these planes on an axis Ox, with origin O on the surface of

the wall where the temperature is imposed. The wall is of an undefined
length. Its local temperature (on a given plane) is T= T (x,t).

expressed by the heat equation. Here it is in the single-dimensional non-
stationary form, associated with the limit conditions that determine the
In other words, the equation:
dT dT
---- = a----- [5.50]
dt d x2
where a is the thermal diffusivity of the material,
A
a =----- [5.51]
p cp
The limit conditions express that the temperature away from the surface
of the wall is not affected by modulation; the second expresses the imposed
temperature condition.
In other words:
t (~, t )=T [5.52]
T (0, t) = T + T0 cos at, [5.53]
where T0 is a modulation amplitude and a is a pulsatance.

Figure 5.4. Non-stationary conduction: sinusoidal temperature on a face
A new relative temperature function is defined:
0=T-Ti [5.54]
The problem becomes
d0 d2 0
— = a----- [5.55]
дt д X2
associated with
0(~, t ) = 0 [5.56]
0(0, t) = T0 cos at [5.57]
We will use a complex formulation to solve this: we look for the reduced
temperature in a complex form; the solution will be the real value of this
complex function.
REMARK.- This process is currently used in classes on electricity. Readers

who are not familiar with this method should refer to these.
Therefore, we write:
0(0,t)=T0 eiat [5.58]
And we look for 0 (x, t) in the form:
0(x,t)=f(x)eiat [5.59]
196 Heat Transfer 1
This form, separating terms in x from terms in t , expresses that we are

in an “established regime”.
Along the same lines:
в(,х, t) = 0 [5.60]
0(0, t) = T0 eiat [5.61]
The solution will be the real part of the result that is found.
The terms of the equation can easily be deduced:
— = iafeiat [5.62]
dt J
d^l = f" ea [5.63]

d x2 J
Hence the equation:
ia f eiat = af" eim [5.64]
( f"+l^L) ea = о
[5.65]
Leading to:
f"+iaL=о [5.66]
a
Which has the solution:
f (x)=Aexp [5.67]
where A and B are two constants.

Here, we see the square root of i appear, an imaginary number that is

already equal to i = V—1 .
REMARK.- iji is not necessarily familiar to all readers. We will easily find
its value by writing it in the form of a complex number where a and b are
real numbers to be found:
i/i = a + i b [5.68]
)
i = (y/i 2 = (a + ib)2 = a2 + b2 + 2i ab [5.69]
by identification:
0+i= a2-b2+2iab [5.70]
a = b = —1=
22
Therefore:
г 1+i
i/i =---- [5.71]
2
The limit conditions allow the following to be written:
at x = 0: [5.72]
f= 0 ^ A + B = T0 [5.73]
at infinity:
x ^^ ^ f^ 0 [5.74]
The function therefore becomes:
irn i® 1 + i I® 1 + i I®
f (x) = A exp x + B exp — x = A exp—— J —x + B exp---- =-. —x
a a a a
198 Heat Transfer 1
> ( ггт X ^
= A exp x +В exp 2a
a x 7 exp - i. .-x 7 [5.75]
V 7 V V
which can only be zero at infinity for: A = 0.
Therefore:
В = To [5.76]
Finally:
a .
e( x, t ) = f (x) eat = To exp - '— x expi at [5.77]
a
Again explaining Vi:
0( x, t )= To exp - x expi at = To exp

[5.78]
(
0( x, t) = T
jo at
V
(
= To exp - cos at - [5.79]
Taking the real part, the solution to our problem:
( a л
cos at - iaxJ [5.8o]
V
The final expression of T(x,t) comes out as:
( ( — 1
T(x,t)= Ti +T0 exp cos —t — 2 aa x [5.81]
к к 7
We see that the temperature at a point of the material is the sum of the
initial constant temperature and of a modulation.
This modulation has an amplitude and a phase that depends on the

considered plane:
on the x axis plane, the amplitude is:
T0 exp [5.82]
The phase is in this same plane:
Ф= —x [5.83]
2 2a
This problem has various practical applications; it allows us to

understand how the variation of the external temperature can act on the
temperature of a wall, in line with alternating day-night temperatures for a
thin wall, or seasonal variations for a thick wall (in the case of the old walls
of farms or thatched cottages, barns and so on etc.).
Another application is in the metrology of thermal materials, in particular,

insulating materials, for which the method presented at the beginning of the
chapter may prove to be tricky. We apply a modulated temperature on the
surface of a material and we place thin probes in different planes at a
distance from this surface. By measuring the temperature in different
locations, we can measure the thermal diffusivity, using the amplitude, or
even better, using the phase shift.
200 Heat Transfer 1
5.4. Problem with two walls stuck together
This problem involves two materials that have a plane face.
These two materials are denoted by 1 and 2 ; they are initially at the
respective initial temperatures T1i and T2i. Their respective physical
properties, which are supposedly invariable, are p 1, cP 1, X 1 and p2, cP2, X2
when written using the usual notations of this book.
At the instant t = 0, these two materials are stuck together by a plane

face. The transfer of non-stationary heat is studied through this contact
surface.
t=0
Figure 5.5. Non-stationary conduction: contact of two bodies at different temperatures.


large in comparison to the dimensions of the problem). The temperature will
therefore be distributed in isothermal planes, parallel to the contact surface.
We locate these planes on an axis Ox, with origin O on the separation

surface. Material 1 will correspond to the negative axes (conventionally to
the left of the origin) and material 2 to the positive axes (conventionally to
the right of the origin).
To structure this in an equation, we will presume that these environments

extend infinitely to the left and right. The respective temperatures of these
two materials will be indicated by:
T1=T1(x,t) and T2=T2(x,t). [5.84]
Therefore, contrary to the other problems seen here, the solution lies in
two temperature functions.

expressed by the heat equation. Here it is in the non-stationary
single-dimensional form, associated with limit conditions that determine the
These two systems need to be written down for each environment, where
each differential equation is only valid for half of the space.
So we have the system of equations:
In environment 1 ,
д д T1 д T1
x<0 —= a [5.85]
дt ’ дx
In environment 2 ,
д дT
- д T2
x > 0 —-■ a. [5.86]
дt дx
where a is the thermal diffusivity of the materials,
A
a =----- [5.87]
p cp
We will define two temperature variables, which are not reduced here:
01 = T1 - T1 i and 02 = T2 - T2 i [5.88]
The system to be resolved becomes:
In environment 1 ,
д01 д01
x<0 —b = a —1 [5.89]
дt 1 дx
202 Heat Transfer 1
In environment 2,
д 02 д 02
x>0 —- = а^ —- [5.90]
дt 2 дx
where a is the thermal diffusivity of the materials,
Л
а =------ [5.91]
p cP
associated with the limit conditions:
In environment 1 ,
x<0
01 (x,0)=0 [5.92]
01 (-ж, t ) = 0
In environment 2,
x>0
02 (x,0)=0 [5.93]
0i (+~, t ) = 0 [5.94]
In addition, at the interface (x = 0)), we have a continuity of flows and

identity of the temperatures in the two environments:
x=0 Л^-T- = Л2 —
[5.95]
дx дx
x=0 T1(0,t)=T2(0,t) [5.96]

hence, further to this:
d 61 d 62
x=0 4^ = ^23 [5.97]
dx dx
x=0 61-62=T2i-T1i [5.98]
Here, we will not give the details of this quite complex solution, the
perspective being to provide an applicable relationship.
The temperatures in the two environments will be given in a

non-dimensional form. Logically, we find two elements that appear in the
previous problems:
the variable
[5.99]
2jOt 4^ at
the error function, in its complementary form
erfct = 1 - erf t [5.100]
where
erf ( = -n lexp- u ■ du [5.101]
In environment 1 “on the left”:
T1 (x, t)— T1 E2 r |x|

---------------- =----------- erfc [5.102]
T. 2 — Ti1 Ei + E2 4\^-
Here, we note the use of an absolute value. Indeed, in this environment, all
axes are negative.
204 Heat Transfer 1
In environment 2 “on the right”, we have:
T2 (x, t)- T 2 E x x
--------------- =---------- erfc .__ - [5.103]
T1 — T2 E + E2 4a^t
At this point, note that the denominator of the reduced temperature is

different in the two cases (influence on the sign).
In these two equations, we see a new parameter E appear, characteristic

to each material. It is called effusivity and is defined by:
E = Xpc [5.104]
In other words:
E1 = 4 P1 c1 [5.105a]
E2 = A2 p2 c2 [5.105b]
5.5. Application examples
5.5.1. Simple applications
EXAMPLE 5.1.- Thermal response of a concrete block
A concrete block with dimensions 60 * 30 * 20 cm3 rests at the initial

moment in time on its largest face on a metal plate that is maintained at a
constant temperature of 60°C. The initial temperature of the concrete block
is 20°C.
The heat transfer is presumed to be single-dimensional, therefore

perpendicular to the metal plate.
How long will it take for the center of the concrete block to reach the
temperature of 40°C?
The volumetric mass of the concrete is p = 2300kg.m\ its specific

heat capacity is c= 878 J.kg-1.K-1 and its thermal conductibility is
AB = 1.4 W.m-1.K-1.
Solution.—
We are faced with a classic problem of non-stationary single-dimensional

transfer, with a temperature of Te =60°C imposed on a face of a solid,
initially at a homogeneous temperature
Ti =20°C. [5.106]
We make the change unknown:
g = T ( X. t)-Te [5.107]
T, - T
The solution is classic.
We define the composite variable:
= = (( X. t )=-^= = -X= [5.108]

24lat 4 at
we then know that:
0 = erf (?) [5.109]
where
erf (^) = ~^n & u2 du [5.110]
is the error function, for which a table is provided in Appendix 5.

206 Heat Transfer 1
Let us calculate the thermal diffusivity of the block:
a = — =-----—-----= 6.93.10-7 m\s [5.111]

pc 2300*878
If the temperature is T=40 °C at this point, we have a value of
в = 40-60 = -20 = 0.5

[5.112]
20 -60 -40
This value is obtained for
£ = 0.477 [5.113]
This value must be found in the middle of the block, in other words, at
x = 0.1m from the heating face.
We easily deduce the corresponding time:
xx
£( x, t )= =------ [5.114]
2^1at 4 a at
-1 f x Y
4 a ^%) [5.115]
2
1 f 0.1 Y
t= , ,L 1 [5.116]
4*6.93.10-7 ^0.477)
t=15855s=4h24mn15s [5.117]
EXAMPLE 5.2.- Thermal response of a steel cube
We plunge a large cube of stainless steel into a bath of molten salts. The
initial temperature of the metal is 25°C. The temperature of the salt bath is
500°C. This temperature is imposed on the cube faces. Given the size of the
cube, we estimate that the transfer of non-stationary heat for each face is
single-dimensional (the flow of heat is locally perpendicular to each face).
How long will it take for the temperature to reach 400°C at 2 cm from the
surface of each face of the cube?
Data for stainless steel: a= 5.53.10-6 m.s-2.
Solution.—
We are facing a classic problem of single-dimensional non-stationary

transfer, with a temperature Te =60 °C imposed on a face of a solid, initially
at a homogeneous temperature Ti =20 °C.
We implement the change unknown:
T(X ( x, t)-Te [5.118]

в= T - T
ie
xx
= = (x x, t ) = [5.119]
we then know that:
0 = erf (?) [5.120]
where erf (g) = -^ê u2 du is the error function, for which a table is
provided in Appendix 5.
If the temperature is T=400°C at this point, we have a value of
400 - 500
о= = 0.21 [5.121]
25 - 500
208 Heat Transfer 1
This value is obtained for
% = 0.188 [5.122]
This value must be found at x = 2.10-2 m from the wall in contact with
the molten salt.
We can easily deduce the corresponding time:
%(x, t ) = ~X= = ~i^= [5.123]

2»at J4 a
2
1 (x^
t =—1 - 1 [5.124]
4a L %J
2
-0^211 [5.125]
t= 1 .(
4*5.53.10-6 ^ 0.188 J
t=512s=8mn3s [5.126]
EXAMPLE 5.3.- Thermal response to a concrete wall
We consider the external layer of a wall. This layer is made of

agglomerated concrete with a thickness of 20 cm. The thermal diffusivity of
this concrete is a= 8.38.10-7 m.s-2.
The wall is initially at a homogeneous temperature of 15°C.
The external temperature is suddenly brought down to 0°C.
How long will it take for the temperature to fall to 5°C at a distance of
5 cm from the internal face of the concrete wall? Express this time in hours
and minutes.
Solution.—
We are facing a classic problem of single-dimensional non-stationary

transfer, with a temperature Te imposed on a face of a solid, initially at a
homogeneous temperature Ti .
The change unknown is made:
T (x,t)-T
в = -(x-)— e [5.127]
T-T
ie
xx
£ = £(x, t ) = [5.128]
we then know that:
в = erf (?) [5.129]
2 f£ ,
where erf (g) = ^= £ e du is the error function, for which a table is
provided in Appendix 5.
If the temperature is T= 5°C, we have, at this point, a value of
5-0
в= = 0.333 [5.130]
15 - 0
This value is obtained for % = 0.305.
This value must be found at x =20 -5 =15cm = 15.10-2 m from the face
of the wall in contact with the molten salt.
210 Heat Transfer 1
We easily deduce the corresponding time from this:
xx
£( x, t ) = —= = ^= [5.131]
lyj a t 4 a at
-1 f x Y
4 a ^%) [5.132]
1 f 0.15 У
t= , 7। । [5.133]
4*8.38.10-7 ^0.305)
t =72157 s=20h2mn37s [5.134]
EXAMPLE 5.4.- Thermal response of an immersed cube
We dip a cube of stainless steel (18%Cr / 8%Ni) with sides of 14 cm,

initially at an atmospheric pressure of 21°C, into a container filled with
boiling water at 95°C.
We will suppose (approximation) that the non-stationary thermal transfer

observed is single-dimensional with respect to each of the faces.
We will also presume that over the course of the entire transfer, the water
will impose a temperature of 95°C onto each face of the cube.
We provide the thermal properties of the steel grade used:
Density:
p = 7816 kg.m3
Thermal conductibility:
X = 16.3 kg.m3 Wm-1 K 1
Specific heat capacity:
c= 460 J.kg-1
1) Evaluate in the context of this approximation how long it will take for
the temperature to reach 55°C at x = 4cm from each of the faces.
2) Evaluate by the same process (which then becomes quite large) how
long it will take for the temperature to reach 80°C at the center of the cube.
SOLUTION.- 1) Here, we see a problem of variable conduction at an

imposed wall temperature. We therefore know that the reduced
temperature will be of the form:
в = erf (£) [5.135]
with
T (x,t)-T
в = -(-,-) [5.136]
T -T
ie
where Te is the imposed temperature and Ti is the initial temperature of the

cube.
Prior to all problems of this type, we will calculate the thermal diffusivity
of the steel
a = _^_ =--- 16.3----- = 4.53.10-6 ms-i

[5.137]
pc 7816*460
we fix
T=55 °C at - = 4.10-2 m [5.138]
the time, which is the value that we are looking for, will be deduced from the
value of giving в as prescribed.
In other words:
в = 55 95 = 0.54
[5.139]
21 - 95
212 Heat Transfer 1
Consulting the table of erf (£), by interpolation, we find:
0.54 = erf (0.5633) [5.140]
From this, we deduce the time that we are looking for
£( x, t )=^= = 0.523 = , 4.10 == [5.141]

2< at 74*4.53.10-6* t
^ 2 --------------- в"
t = ( ------2 [5.142]
^ 0.523 J 4*4.53.10-6
t=322.8s=5mn23s [5.143]
2) The approach followed in the previous question will be used again

here
At the center of the cube
x=7cm=7.10-2 m [5.144]
For 80°C at the center of the cube, 0 has the value:
0 = 80 95 = 0.2027 [5.145]
21 - 95
So, in the same way as before:
0.2027 =erf(0.182) [5.146]
So, the time:
^ 2 -----------------
t = (-- ------2 [5.147]
I 0.182 J 4*4.53.10-6
t=8164s=2hr16mn4s [5.148]
5.5.2. Some scenes from daily life
EXAMPLE 5.5.- Problem of thermal phenomena on a daily basis: ironing day
A housewife is ironing her laundry. She inadvertently places her iron on a

pile of laundry. The iron has a power of P= 1000W, and the surface of the
plate (surface in contact with the laundry) is S = 100 cm2. The atmospheric
temperature at which the laundry is initially found is Ti = 20°C.
Figure 5.6. The iron and its pile of laundry
The pile of laundry will be assimilated to a homogeneous, semi-infinite

environment whose physical characteristics are:
Thermal conductibility of the laundry:
2 = 0.46 W. m-1.kg 1
Density:
p = 929.kg. m-i
Specific heat capacity of the laundry:
c=1830J.kg-1
214 Heat Transfer 1
We consider that the polyester material starts to melt at temperatures of

TF =240°C:
1) What is the flow density ф0 imposed by the iron at the top of the pile
of laundry?
2) Give the expression as a function of the time taken by the top of the
pile of laundry to heat up: T(x=0, t)-Ti .
How long will it take for the laundry at the top of the pile to begin to
melt?
3) What is the temperature at x = 5 cm under the surface of the pile 30 s

after the iron has been placed on the pile? (We ignore the effects of the
surface fusion on the textile properties).
SOLUTION.- 1) The density of flow ф0 imposed by the iron on the top of the
pile of laundry is obtained by dividing the flow P= 1000 W by the heating
surface S = 100 cm2 = 10-2 m2:
1000
P0 = =105W m-2 [5.149]
10-2
2) Expression as a function of the time taken for the top of the pile of
laundry to heat up:
We are facing a non-stationary problem with imposed constant flow. The

temperature is then expressed by:
О = ^^-yOîerC (£) [5.150]

Л
with the usual unknown and variable:
0( x, t ) = T (x, t)-T [5.151]
xx
i = —/= = “/= [5.152]
2 at 4 a at
We are looking for the value of
T(x = 0, t)-T = 0(0,t) [5.153]
In this case,
S=-0= = 0 [5.154]
2yJat
0 (0, t) = -^-~yJat ierfc(0) [5.155]

7 7 2
We know that
ierfc(0) =0.5642 [5.156]
We calculate
a = — = 2.7.10 [5.157]
We will reach the temperature T(0,t)=220 °C at the top of the pile for
the time tF such that:
0 (0, t) = V0 a t ierfc (0) [5.158]

4 7 2
0 (0, t ) = 240 - 20 = 220 = 2.10-J2.7.10-7 * t„ *0.5642 [5.159]

V 0 0.46 F
tF0,5 =220 =1.72 [5.160]
tF = 2.9 s [5.161]
It is highly advisable to be very careful!

216 Heat Transfer 1
3) The temperature at x = 5 cm below the surface of the pile 30 s after the

iron is placed on the pile will be calculated using the expression в for
x = 5.10-2 m and t= 30 s.
Then:
5 10-2
c = —. • - = 0 878 [5.162]
2V2.7.10-7 *30
ierfc(0.878) =0.073 [5.163]
в = 2^'at~ierfc (£) [5.164a]
( )
в 5.10-2,30 = ^.10-J2.7.10-7 *30 * 0.073 [5.164b]
0.46
в(5.10-2,30) = 90.33 [5.165]
T (5.10-2,30) = 110.33 [5.166]
Some of the laundry can be saved...
EXAMPLE 5.6.- The physicist is going to cook an egg
An egg, when it is laid, is made up of a shell, egg white and yolk. The
white of the egg is a part of the egg, made essentially of an albumin,
ovalbumin, which protects the yolk or the zygote.
The albumin presents advantages due to its properties as a coagulant and

surfactant (this allows beaten egg whites to hold their shape).
It is coagulation properties that we are interested in here.
This coagulation occurs at 57°C.
In the egg yolk, the proteins also coagulate, but at a slightly higher
temperature of 65°C.
In the simplified model that follows, it will be considered that an egg is

made up of
1) A central cylindrical zone of radius 0.5 cm, known as the egg yolk.
2) An annular cylindrical zone surrounding the first, of thickness 5 mm,
called the egg white.
3) An extremely thin shell that will have no effect on the thermal
phenomena.
Water
Figure 5.7. Diagram of an egg
We will consider that the white and the yolk of the egg have the same
thermal properties, which are very close to those of albumin. We will
therefore consider that the egg white and the yolk constitute a single
environment, which is also strictly immobile.
1) Initial calculation
A very thin plane wall, for which we will consider thermal resistance to
be zero, separates two spaces containing:
- On the left, water at 100°C, which will only have the effect of
maintaining this temperature at the separation for the duration of the
problem.
218 Heat Transfer 1
- On the right, albumin, composed mainly of water containing

macromolecules. This environment will be considered perfectly immobile,
thus excluding all appearance of convection coefficients at the interface.
Water Albumin
Te = 100C Ti = 20C
Figure 5.8. Initial calculation
The thermal characteristics of the albumin are evaluated to be:

Thermal conductivity:
2 = 0.6 Wm-2 K -2
Density:
p = 1000 kg. m -3
c= 4180 Jkg-1K-1.
Remark.- Well-informed readers will certainly not miss the fact that these
are the thermal properties of water. It is not always easy to find precise data
in terms of the physics and chemistry of food items. Egg white contains 88%
water. Among the other constituents of albumin, we find 10.6% of globular
proteins, the main one being known as ovalbumin. For this simple model, we
have chosen data related to water, which are easier to access.
At an initial instant in time, the temperature of the albumin, in the right

hand compartment, is uniformly equal to Ti = 20°C.
a) How long will it take for the temperature to reach T1 =57 °C in the
albumin at a distance e = 0.5cm from the separation surface with water at
100°C.
b) How long will it take for the temperature to be equal to T2 =65 °C in

the albumin at a distance e= 1 cm from the separation surface with water at
100°C.
2) Cooking
This question, requiring no additional calculation, only requires a little

physical reasoning, which is a highly desirable quality of all thermal
scientists.
We will admit that if the part at the center of the egg white has reached a
temperature of 57°C, the white will be coagulated and the yolk will remain
soft. We will have a soft-boiled egg.
In the same way, if we reach the minimum temperature of 65°C at the

center of the yolk, the white and the yolk will coagulate and we will have
cooked a hard-boiled egg.
a) Soft-boiled egg
Remembering that both a “soft-boiled” and “hard-boiled” egg are cooked

by plunging the egg into boiling water (here presumed to be 100°C), it is
only the cooking time that makes the difference.
Modeling the egg geometrically, as we did at the beginning of this text,

we will examine how much time is required, using our model, to cook a soft-
boiled egg.
REMARK.- The highly non-stationary aspect of the operation explains why it

is recommended to take the eggs out of the refrigerator for a sufficient
amount of time before cooking, in order to preserve the shell. Here, we have
a material problem (thermal shock) that falls outside the remit of our
considerations.
b) Hard-boiled egg
In the same way, estimate the order of magnitude of the time required to
cook a hard-boiled egg.
220 Heat Transfer 1
c) Are these results coherent with your personal experience (if any)?
If not, in your opinion, why?
Solution.—
1) Initial calculation
We are facing a single-dimensional non-stationary problem with an

imposed temperature on a marker in space.
We will determine the thermal diffusivity of the albumin
—
a =--- = 1.43.10-7 m25-1 [5.167a]
Pc
a) We are looking for the time t required so that equal T1 reaches 57°C
in the albumin at a distance e= 0.5cm from the separation surface with
water at 100°C.
This is a problem with an imposed wall temperature.
We will firstly calculate the thermal diffusivity of the albumin
a = — = 1.43.10-7 m25 [5.167b]

Pc
We define a reduced variable:
T-T
e=------ - [5.168]
Ti- Te
with
Ti = 20°C [5.169]
Te =100°C [5.170]
We know that 0 will be of the form
0 = erf Q [5.171]
with:
x
Q = ^= [5.172]
4 a at
We write
T(x,t)=57°C [5.173]
then:
0= 0.537 [5.174]
and the table gives us:
Q= 0.519 [5.175]
the following is then deduced:
Q= 5.10-3 = 6.61
[5.176]
^4*1.43.10-7 t tit
/= = 12.7 [5.177]
t=162s=2.7mn [5.178]
b) We now look for the moment when the temperature T2 reaches 65°C in
the albumin at a distance of e= 1 cm from the separation surface with the
water at 100°C.
The calculation will be based on the previous one:
T-T
0 =------- [5.179]
Ti- Te
222 Heat Transfer 1
i =20°C
T [5.180]
Te =100°C [5.181]
в = erf Q [5.182]
x
i = ,----- [5.183]
4 a at
T(x,t)=57°C [5.184]
в = 0.437 ; Q = 0.413
^= 10-2 = 13.22
[5.185]
^4*1.43.10-7 t t[t
4= = 32.01 [5.186]
t=1026s=17mn [5.187]
2) Cooking an egg
a) Soft-boiled egg
The preliminary calculations give us the result: using the modeling in the
albumin, looking for a temperature of 57°C at 5 mm from the shell, we
estimate the time taken for coagulation, therefore for cooking, as
2.7 minutes.
b) Hard-boiled egg
In the same way, using the modeling seen above in the albumin, looking
for a temperature of 65°C at 1 cm from the shell, we estimate the
coagulation time, therefore the cooking time, to be 17 minutes.
c) These results are relatively coherent with personal experience
Cooking a soft-boiled egg is of the order of 3.5 minutes. Modeling is

correct here.
Cooking a hard-boiled egg is of the order of 10 minutes. This model

gives a higher value.
Over and above the approximation made on the thermal properties of the
white and the yolk, here, we use the results of a plane model in a revolving
system. The geometry selected here is particularly schematic.
In addition, no influence has been attributed to the natural convection in

the egg.
EXAMPLE 5.7.- Barbecue evening. Cooking meat on a plancha
A system of cooking on a plancha is made up of a rectangular heating

plate which emits heat uniformly through its upper surface, of dimensions
50cm * 30cm and with a power of P= 3600W.
We place a piece of meat that is presumed to be perfectly plane onto the

plancha. The heating surface is treated with a non-stick coating, and no fat or
oil is required for cooking.
The useful thermal properties of the meat will be taken to be:
Thermal conductivity:
2 = 0.47 Wm-2 K-2
Density:
p = 1080 kg.m '
c= 4000 Jkg-1K-1.
The slice of meat, which has just been taken out of the fridge, is at
Ti = 15°C. We will consider that the meat will be cooked “rare” at a
temperature of TCS = 50°C.
We will also consider that, although the slice has a finite thickness, the
evolution of temperatures within the meat is calculated as it is for a body of
infinite thickness.
224 Heat Transfer 1
We also consider that the physical properties of the meat are modified
very little by the cooking1 (which is a debatable approximation):
1) We are seeking to determine the thickness of the meat cooked after a
specific duration of time.
a) What is the density of the flow ф0 imposed by the plate on the meat?
b) What thickness of meat will be cooked after 2 mn?
2) After these two minutes, what will the temperature TS be at the surface
of the meat placed on the plate? All good cooks know that meat must be
“sealed” at the surface (meaning that the surface layer must be “pyrolyzed”.
Is this the case here?
3) A slice of meat with a thickness of 1.5 cm is cooked by placing it for
two minutes each side on the plancha.
Will the meat be:
- cooked “rare”?
- cooked “medium rare”?
- grilled all the way through?
Solution.-
1) We attempt to determine the thickness of the meat cooked after a

determined amount of time.
a) The flow density ф0 imposed by the plate on the meat is deduced

from the flow, which is equal to the power dissipated by the plate:
ф = p = 3600 W [5.188]
The surface of the plate is
S = ^°|81 = 6.16.10-2 [5.189]
1 We will admit it even for “pyrolyzed” meat at the surface (effect sought for “sealed” meat).
This approximation is obviously a subject of discussion. We will not discuss this here.
ft Ф 36' ' = 2.4.104 wm [5.190]

0 S 6.16.10-2
b) What thickness of meat will be cooked after 2 mn?
Let us firstly calculate the thermal diffusivity of the meat:
A 0.47 -7 —2
a =---- =-----------------= 1.09.10 ms [5.191]
pc 1080*4000
We will define the usual reduced variable:
xx
i= f~ = l----- [5.192]
2< lat 4 at
We need to resolve a problem of non-stationary conductive transfer with

a flow density imposed on the input surface (x = 0). We need to calculate
the temperature on this input face. To do so, we use the known result of the
difference in temperature в between the temperature at time t and the initial
temperature:
в = T (x, t)-Ti [5.193]
в = ^Jaîerfc (£) [5.194]
At time t=2 mn=120 s, we are looking for at what point the temperature
is equal to TCS =50°C, in other words,
в = T (x, t)-Ti = 50 -15 = 35°C [5.195a]
35 = 2* 2.4.10 ^1.09.10-7 *120 ierfc(£) [5.195b]

0.47
from this, we deduce:
35 * 0.47 2
ierfc (%) =-------------------------------- = = 9.47.10-2 [5.196]
2*2.4.104 * J1.09.10"7 *120
226 Heat Transfer 1
Hence,
4 = 0.782 [5.197]
4 = 0.782 = . x = = -Д= [5.198]

^/1.09.10-7*120 4a at
x = 5.65.10-3 m=5.65mm [5.199]
2) What will the temperature TS be after these two minutes, at the surface
of the meat placed on the plate?
We look again at the solution giving 0( t)
0 = TS - t [5.200]
0 = 2Y'O~ierfc (4) [5.201]
0 = 2*2.4.10 ^1.09.ю-7* t ierfc(4) [5.202]

0.47
0 = 33.7 Jtierfc (4) [5.203]
4 = ~F= =------ X---- = = 0 [5.204]

4aat 6.6.104 *4t
ierfc(0) =0.5642 [5.205]
t = 2 mn = 120 5 2 [5.187]
0=33.7*7120 * 0.5642 = 208°C [5.206]
TS = 233°C [5.207]
And of course: the meat will be “sealed”.
3) We cook a slice of meat of thickness e= 1.5cm by placing it on the

plancha for two minutes on each side.
The penetration distance at 50°C is 5.65 mm.
We will have “rare” cooked meat for 2 * 0.565 = 1.13cm and the center
will be very “raw” for 1.5-1.13=0.37cm =3.7 mm, with a well-sealed
surface.
The meat will be rare and cooked well.
EXAMPLE 5.8.- Fusion of a weld
A cube with sides a= 10cm is made up of two parallelepipeds A and

B, each with sides a * a * a welded with tin on its large faces.
We estimate that the welding is a very thin film of tin that does not act on
the temperature distribution in the copper block with sides a.
Figure. 5.9. A block for welding

228 Heat Transfer 1
Figure 5.10. The welding
The copper cube is balanced with the atmospheric temperature

Ti =21 °C.
We fix the copper cube below a horizontal plate P at the temperature of

310°C.
The fixing is sufficiently rapid for us to estimate that the upper face of the
cube is rapidly brought to a temperature of T0 =310°C, which it will
retain.
We give the thermal properties of copper:
Density:
р = 8954 kg.m '
X = 386 kg.m-i Wm 1 K 1
c= 383 J.kg-1 .
We presume that the transfer of non-stationary heat in the copper cube is

practically single-dimensional.
So, Ox is the vertical axis, with origin O at the center of the upper face
of the cube.
0
р
Figure 5.11. Diagram of the device
1) Give the general expression for the temperature T(x, t) of any

horizontal plane in the cube.
2) Tin melts at 232°C. How long after the cube and the plate are put into
contact will the cube B fall?
Solution.—
1) We are in the presence of a non-stationary conductive transfer at

an imposed wall temperature. The solution is known:
Defining:
T(x,t)-Te
О= [5.208]
230 Heat Transfer 1
we have:
0 = erf (?) [5.09]
where:
x [5.—10]
Q = I-----
4 a at
Writing T(x, t) explicitly: [5.—11]
T(x,t)=(Ti-Te)erf(Q)+Te [5.—1—]
2) The tube will fall when the tin has melted. This will intervene at
time tF where the temperature will reach TF =232 °C in the plane that
a -2
separates the two half-cubes, in other words at x = — = 5 cm = 5.10 m .
We calculate the thermal diffusivity of the copper:
Л 386 -a -I
[5.—13]
a =----= = 1.126.10 m 2s s
pc 8954*383
The desired value of 0 is:
in other words:
0 = 232 ~ 310 = 0.27 [5.21a]

—1- 310
Consulting the table for erf (Q), by interpolation, we find that:
0.27 = erf (0.2aa) [5.215]
From this, we deduce the time that we are looking for
x 5.10-2
Q (x, t) = —= = 0.244 = [5.216]
2Jat J4*1.126.10-4* t
2
I 5.10-2 ] 1
t = I--------I ---------------- T [5.217]
^ 0.244 J 4*1.126.10-4
t= 93.23 s [5.218]
EXAMPLE 5.9.- Welding on copper
We wish to weld a plate of copper of thickness e= 5 mm onto a base that

has already been covered with a thin layer of tin. Initially, the plate, tin and
its base are all at a homogeneous temperature of Ti =21 °C.
For the plate of zinc, we place it on the layer of tin. We presume that the
contact is perfect and that no layer of air can pass between the copper and
the tin.
On the upper part of the copper, we place a soldering iron tip with a
surface that is presumed to be much larger than the soldering zone. The
problem of conduction will thus remain purely single-dimensional.
Soldering iron tip (Fixed temperature)
x=0
Soldering plate
Tin layer
Base
x Axis
Figure 5.12. Diagram of a tin weld
232 Heat Transfer 1
We will try to approach the problem with two different hypotheses

(questions 1 and 2)
We give:
For the copper
Density:
р = 8954 kg.m-3
X = 386 Wm-1K-1
c= 383 J.kg-1
For the zinc
Density:
р= 7144 kg.m-3
X= 112 Wm-2K -2
c= 384 J.kg-1
1) We presume that the welding iron imposes a temperature

Te =400 °C on the upper surface of the copper plate.
a) The tin melts at TF =232°C. What is the minimum time that the iron tip
must be left on the plate of copper so that the weld can be performed?
b) Look again at the problem for a zinc plate of thickness 3 mm.

2) We look again at the problem of the copper plate of 5 mm in another

hypothesis. We will presume, this time, that the welding iron imposes a
thermal flow density of ф = 5.105 W.m 2 on the upper surface of the plate.
After time t found in 1a, what will the temperature be in the tin?
What conclusions do you draw from this concerning the validity of our
previous calculation? In order to reason this answer as best possible, it will
be useful for us to be able to calculate the temperature of the upper surface
of the copper plate at time t calculated in 1a.
Solution.—
1) We have a non-stationary transfer problem at an imposed

temperature.
a) We will calculate the thermal diffusivity of the copper:
A 386 -4 -I
a =---- = = 1.125.10 m25 [5.219]
pc 8954*383
We will define a reduced variable:
T-T
в =----------- e- [5.220]
T - Te
with
Ti =21°C [5.221a]
Te =400°C [5.221b]
We know that в will be of the form
в = erf Q [5.222]
with:
x
i = ,----- [5.223]
4 at
234 Heat Transfer 1
We want to reach
T (x,t) =232°C [5.224]
at the distance from the iron tip equal to x =e=5 mm
We have
0 = 0.443 [5.225]
and the table gives us, following interpolation:
% = 0.415 [5.226]
from this, we then deduce:
0.415 - 5Л0-3 — 0.235

[5.227]
74*1.125.10-4t t/t
t[t - 0.566 [5.228]
t -0.32s [5.229]
b) With the zinc plate, we take the same approach
A 112
a —---- - =-------------- — 4.08.10-5 m25 [5.230]
Pc 7144*384
We will define a reduced variable:
T-T
0 ------ e- [5.231]
T - Te
with
T. - 21°C [5.232]
Te - 400°C [5.233]
We know that 0 will be of the form
0 = erf Q [5.234]
with:
x
Q = ,— [5.235]
4 at
We wish to reach
T (x,t) =232°C [5.236]
at the distance from the iron tip equal to
x =e=5 mm [5.237]
We have
0= 0.443 [5.238]
and the table gives us, after interpolation:
Q= 0.415 [5.239]
from this, we then deduce:
0.415 - 3.10-3 =
[5.240]
74*4.08.10-5t tit
t[t - 0.566 [5.241]
t -0.32s [5.242]
We obtain the same time for this thinner zinc plate.

236 Heat Transfer 1
2) In the context of this second hypothesis, we have a problem of an

imposed flow density
е = T (x, t)-T [5.243]
We know that the difference of temperature between T(x, t) and the

reduced temperature obeys
е = -J^'Otêrfc (£) [5.244]
We are looking for the temperature at x = 5 mm and t= 0.32s
£ will have the value previously seen of £ = 0.415 [5.245]
The corresponding value of ifrf (£) is given by the table:
ifrf (£) = 0.2429 [5.246]
We therefore find
е= 3.77°C [5.247]
which leads to
T =24.77°C [5.248]
which is obviously highly insufficient for a good weld. The welding time has
been significantly underestimated by the first hypothesis, which must
therefore be pushed back in time.
We can understand this by calculating the temperature at the point of

contact of the iron tip, so, for x = 0
е (0, 0.32) = 2j^jat~ierfc(0) = 8.49°C [5.249]
The temperature on contact with the welding iron is
T =29.49°C [5.250]
6
Fin Theory: Notions and Examples
6.1. Notions regarding the theory of fins
6.1.1. Principle of fins
In principle, the use of a fin to dissipate heat is based on a simple idea.
Let us look at an example. A cylinder with radius r= 5 mm and height

h= 3 cm generates an internal power of P = 3W , which will be equal to the
flow Ф0 that must be dissipated. Its temperature must not exceed T0 = 85°C.
We suppose an ambient temperature of Te = 20°C.
If we assume natural convection is cooling this cylinder, for example

with a convection coefficient h = 10 W. m2 кK-1, the exchange surface will be
S = 9.4.10-4m2 and the temperature of the cylinder will be such that
P = Ф0 = Sh(T01 - Te); in other words, T01 = 339°C, which is much higher
than T0 = 70°C. We can consider increasing the convection coefficient that
should then be equal to h'= 63.8 W.m2.K-1 . Perhaps this could be imagined
for forced convection, but at the price of a significant technical complication.
We can more simply imagine welding eight parallelepiped plates, or

wings, on the initial cylinder, with a cross-section of 0.3 *3 cm2 and length
of L= 1cm. Neglecting the surface of the body that has remained free
238 Heat Transfer 1
(the wings occupy 2.4 cm of 3.14 cm of the outside), the exchange surface of
the eight wings will become 5' = 8*6.6.10-2 *1.10-2 = 5.28.10-3 m2.
We will choose, for example, aluminum to construct these fins.

Supposing that the plate conducts heat infinitely, the surface of the fin will
be uniform and equal to that of the cylinder that requires cooling, in other
words T02. According to P = Ф0 = Sh(T02 -Te), this new temperature is
denoted as T02 = 77°C, which now fulfills the condition T02 < T0 .
However, in reality, things do not happen like this.

a) The fin is not an infinite conductor, and a longitudinal temperature
gradient is required.
b) Consequently, the flow evacuated by the sides of the wing reduces the
cylinder towards the free end of the wing. The flow evacuated in reality Ф
will therefore necessarily be lower than the “ideal” flow found for an
isothermal wing.
Furthermore, we will see that the relation between the two flows defines
the efficiency n of the fin:
Ф
n = Ф0
6.1.2. Elementary fin theory
A wing takes the form of an object, generally metallic, with one end
welded to a body that has a temperature T0 . This object will have a cross
section S that can vary as a function of the distance from one of its cross
sections with respect to the body.
Two main geometries can exist: a linear fin and a circular fin.
The temperature T0 will constitute a limit condition in this problem. It

may be that it is determined in the body to be cooled by other phenomena, in
particular the generation of a quantity of heat in the body.
Fin Theory: Notions and Examples 239
Cylindrical, simple or multiple fins
Figure 6.1. Two types of fin: linear and cylindrical
The basic problem that we have to resolve each time we study a wing is
as follows.
Given the temperature T0 of the body to which a fin is welded and

the ambient temperature Te , what is the total flow of heat that escapes
from the fin? This flow will correspond to the quantity of heat removed
from the body that needs to be cooled.
To resolve this problem, we will divide the fin into infinitely thin
sections, and write a thermal balance for each of these sections:
Flow entering by conduction = Flow exiting by conduction + Flow

exiting by convection.
We thus obtain a differential equation (a special form of the general

equation of stationary thermal conduction) that will be resolved by applying
the limit conditions, which will vary from one problem to another.
240 Heat Transfer 1
Figure 6.2. Model of the parallelepiped linear fin
We will locate the various sections of the fin on the Ox axis that is
perpendicular to the body considered, where the origin O is on the body in
question. At this stage of the equation, the fin is not a parallelepiped, and
therefore S= S(x).
We will set out a fundamental hypothesis, which is common when

dealing with these problems: the field of temperatures will be single
dimensional. In other words, the isotherms of the fin are perpendicular
planes to the Ox axis; or T= T (x).
This hypothesis presumes that the lateral thermal gradients are negligible
in comparison with the longitudinal gradients. Physically, this comes down
to assuming that the lateral convective flows are weak in comparison with
the axial conductive flow.
In addition, the fin will be made up of a homogeneous material of

constant conductibility A.
On the lateral surface of the wing, the convective transfer is determined

by a convection coefficient h . We will require an additional local parameter;
the perimeter of the section for which we will write the balance as:
p = p(x).
We will consider a section of the fin of thickness dx located at a

distance x from the body.
We will write the flow. First, we note that the flows are directed from the
body towards the end of the wing; therefore, the temperature decreases
dT
according to Ox and---- < 0. The flow entering by conduction is given by:
dx
Ф, = -AS (x)— [6.1]

dx
This quantity is positive.
The flow exiting by conduction is given by:
Ф S =-AS (x) dT + — ( - AS (x)—1 dx [6.2]

ax dx ^ <x x J
This quantity is also positive.
The flow exiting by convection:
Фconv = [P(x)dx]h(T(x) - Te ) [6.3]
In this expression, p(x)dx is the lateral surface through which the flow
taken out by convection flows.
The incoming flow is divided into exiting conductive flow and

convective flow to the outside. The balance is therefore written as:
Ф e =Ф S +Ф Conv [6.4]
dT dT d ( dTV
- AS (x)--- = - AS (x)----- 1----- 1 - AS (x)---- I d x
ax 'dx dx ^ 'dxJ [6 5]
+ [p(x)dx]h(T(x) -T)
After simplification:
d AS (x) dT 1
-hp(x)(T(x)-Te)=0 [6.6]
dx dxJ
242 Heat Transfer 1
We note that in the most general problem, the term in S(x) remains in
the derivative, which can lead to equations that are complex to resolve.
6.1.3. Parallelepiped fin
6.1.3.1. Solution
To illustrate the type of solutions found, we consider here the
simplest case, that of a parallelepiped fin. We also provide a few other
solutions in Appendix 6.
First, a point about vocabulary and notation must be specified. We call S
the face of the wing that is parallel to the surface of the body (at the
extremity origin ( x = 0 )); a surface S is fixed to the body to be cooled.
In the case of the parallelepiped wing, S is a rectangle with sides l and

e . We will call the largest of these sides the width l of the wing, and the
smallest of these sides the thickness e . The third dimension of the
parallelepiped will be defined as the length L of the wing. It should be noted
that for a short wing, the “length” of the wing can be smaller than its
“width”. For example, we can have: l=3cm,e=3m,L=1cm.
Then, for this parallelepiped wing:
S=le=Cte [6.7]
and
p =2(l+e)=Cte [6.8]
The balance equation will therefore simplify to:
d2 T
(
''.Sdx - hp T (x)— Te = 0 ) [6.9]
where, again:
(
d2T - ^ T (X)— T = 0 ) [6.10]
Here, this has led to the introduction of a parameter ю, a classic for these
fin problems. This parameter is defined by:
hp
Ю = —^~ [6.11]
AS
For ease of calculation, we will introduce the reduced variable:
0(x)=T(x)-Te [6.12]
Then:
d0( x) dT (x)
[6.13]
dx dx
The equation to resolve becomes:
dО
------- Ю 0 = 0 [6.14]
d xг
Various conditions can then be envisaged:

a) The fin is infinitely long. In terms of limit conditions, we therefore
have:
x=0;0=T0-Te=00 [6.15]
-
eeT=0 [6.16]
b) The fin has a finite length of T0 - Te = 00. We assume that the

dissipated flow at its extremity is negligible. We therefore have the limit
conditions:
x=0;0=T0-Te=00 [6.17]
x=L - A — = 0 ^ do = 0 [6.18]
dx dx
Here, we will look at these two geometries.

244 Heat Transfer 1
We can also take into account the flow at the extremity. Keeping the
same convection coefficient to simplify, we will have:
x=0 ; 0 = T0 - T = 00 [6.19]
x =L ; - A — = h (T - T) ^ — =-h 0 [6.20]
dx e dx A
The general solution to the differential equation
d 20
— Q a2 0 = 0 [6.21]
dx 2
is known.
We can write it in two forms:
the most common one is:
0= Aeax + Be-ax [6.22]
or a form that uses hyperbolic functions:
0=Ashax+Bchax [6.23]
This last form will most often be preferred for practical reasons.
Moreover, these two expressions are equivalent, since the sine and cosine
hyperbolics shax and chax are linear forms of the exponentials eax and
e-ax.
NOTE.- Readers who are not familiar with hyperbolic functions can refer to
Appendix 8, which contains a few reminders that are essential for these
functions.
We will distinguish two cases, as indicated above:

- the wing is long enough to be assimilated to an infinite length;
- the wing has a finite length.
6.1.3.2. Problem of an infinite fin

We need to resolve the equation:
de
— - &■e = о [6.24]
dx
with the limit conditions:
x=0 ; e = Tо - Te = e0 [6.25]
x >° ; e ^ Te - Te = 0 [6.26]
In this specific case, we will use the solution of the sum of two
exponentials:
e =Ae" + Be-ax [6.27]
The two constants are determined by the limit conditions:
x=0;e=e0 =Ae0+Be0=A+B [6.28]
x ^^ ; e = Ae~ + Be-° ^ 0 [6.29]
where A placed in front of an infinite term is necessarily zero. The

following holds:
B = e0 [6.30]
We therefore obtain:
e=e0 e [6.31]
and then:
T (x )=( To - Te) e- *x + Te [6.32]

246 Heat Transfer 1
The parameter of interest in this type of problem is the flow Ф that is

evacuated, which will be the flow observed at the connection between the
wing and the body, that is, at x = 0:
Ф = -ShdT
dx
(
=-SЛ(To -Te) -me-mx )x0 =
[6.33]
x=0
Ф = -S dT ( )
=—S Л( To — T) — me ~°,X x=0 [6.34]
x=0
Ф= S Лю( T0 - T ) = 00 S Лю [6.35]
which can then be written by the definition of:
m=. Ihp
[6.36]
ЛЛ8
= S лЛ ( T0 - Te )
Ф= SЛю(T0 -Te) [6.37]
Ф= S S Лкр (T0 --T) ) = 00JS ^hp [6.38]
We can compare this flow to Ф 0, the flow that we would have with an
ideal wing; in other words, the isotherm at the temperature of the body needs
to be cooled. This flow Ф0 is calculated by considering:
Ф0 = pLh(T0 -T ) = pLh00 [6.39]
6.1.3.3. Problem of a fin of finite length

We now consider a fin of the same geometry as previously examined,
with a length that is finite and equal to L . We will neglect the thermal flow
at the end of the fin.
We need to resolve the equation:
dв
——- a в = 0 [6.40]
dx
with the limit conditions:
x=0 ; в = Tо - T = 00 [6.41]
d^ d0
x=L ; - 2---- = 0 ^---- = 0 [6.42]
dx dx
Here, we will use hyperbolic functions to write the general solution to

this equation:
в = Ash ax + Bchax [6.43]
The two constants are determined by the limit conditions:
x=0 ; в = T0 - T = в0 = Ash0 + Bch0 = B [6.44]
x=L ; -2 — = 0 ^— = AachaL + BashaL = 0 [6.45]

dx dx
We obtain:
B = во [6.46]
We therefore obtain:
BashaL в0 shaL
A =------------- = - 0------------------ = -в0 thaL [6.47]
achaL------ chaL
and then:
в( x )= в01 —s1 ^L shax + ch ax I [6.48]

I ch aL I
248 Heat Transfer 1
After a few transformations:
- shaLshax + chaLchax
0( x )= 00 chaL
[6.49]
- shaLshax + chaLchax
chaL
We know that:
ch(x-y)=chxchy-shxshy [6.50]
0(x ) can be rewritten as:
0( x )= = 0 (- shaLxhax + chaLchax |=0 i c^L - x) 'I [6.51]

v 0 0| chaL I 01 chaL I
Or even:
T(x)= 00 Ic^' .'\ + T. =(To -Te) ich^11 + T. [6.52]

^ ch O)L J ^ ch O)L J
The parameter of interest in this type of problem is the evacuated flow Ф

that will be the observed flow at the connection between the wing and the
body, so at x = 0 :
dT -mshm(L - x) 1
Ф = -S2----
dx
----ch------
aL
L\ + T. [6.53]
x=0 x=0
Ф= S 2( T - T ) I ashaL ' = S 2a( T - T ) thaL [6.54]

v 0 I chaL J v 0 e
which can be written by the definition of:
a=,\SS [6.55]
Ф= SAco(T -T)thvL = SA hp(T -Te)thvL [6.56]

AS
Ф = 7SAhp(T0 -Te)thrnL = 00 yjSAhpthrnL [6.57]
6.2. Examples of application
In the above fin theory, the transfers involved are reduced to conduction
and convection. A wing can also have radiation. Examples of this case have
been attributed to the section dedicated to radiation in Chapter 5.
EXAMPLE 6.1.- Rectangular fin
A rectangular wing is mounted on a body at constant temperature

T0 =75 °C . The ambient temperature is Te = 25°C.
The convection coefficient between the fin and the ambient air is
h =10W.m-2.K-1.
The surface area of the fin (not including the part welded to the body to
be cooled) is S = 20 cm2.
1) If the entire surface of the wing was at T0, what would the flow Ф0
evacuated by the wing be?
2) The flow evacuated by the wing is, in fact, Ф = 0.9 W. What is the
efficiency n of the fin?
Solution.-
1) The flow that will escape by convection of an isothermal wing at

T (x) = T0 over the entire surface area S can be simply written as:
Ф0 = Sh(T - Te) [6.58]
Ф0 = 20.10-410(75 - 25) = 1W [6.59]

250 Heat Transfer 1
2) By definition of the efficiency:

ф
П = ^= 0.9 [6.60]
EXAMPLE 6.2.- Calculation of a fin
An aluminum fin of width l= 5cm, length L= 10cm and thickness

e= 3 mm is embedded in a wall. The conductibility of the aluminum is
2 = 204W. m-1.K-1.
The base of the wall is maintained at T0 =300°C, the ambient

temperature is Te =30°C and the coefficient of convective transfer is
h =10W.m-2.K-1.
1) Determine the temperature TF at the extremity of the fin and the flow
ф extracted by the fin if we neglect the thermal gradients in the width and
thickness directions.
2) The efficiency n of the fin is the ratio of the flow extracted Ф, in

comparison with the flow ф0 that would be extracted by the wing of the
same geometry, whose temperature will be uniform and equal to the
temperature of its base. Calculate this efficiency.
3) Calculate the relative error that would be made by considering the
temperature of the end of the fin is equal to the ambient temperature.
Solution.-
1) We define
The reduced variable:
e=T (x)-T [6.61]
Oo = To - T [6.62]
The perimeter pe of the fin is:
pe=2(l+e) [6.63]
The parameter a is defined as:
hp
O)2 = _ê. [6.64]
AS
For a wing of finite length:
Ф= ASa00 th( aL)=Jhpe AS 00 th( aL) [6.65]
For an infinitely long fin, we will have:
Ф = AS a00 =y[hpe AS 00 [6.66]
We obtain:
pe=2(l+e)=2(5.10-2+3.10-3)=0.106m [6.67]
S=le=5.10-2*3.10-3=15.10-5m2 [6.68]
a I I0*0,!06 =^64 = 5.88 m.,

[6.69]
204*1.5.10—4
Ф = 7hpe AS 00 th ( aL)
= 710*0.106*204*1.5.10-4 (300-30)th (0.588) = 25.72W
th (0.588)= 1.8004 0.555 = 0.529 [6.70]

1.8004 + 0.555
252 Heat Transfer 1
2) Efficiency of the fin

The flow that would escape by convection of an isothermal fin at
T(x) = T0 over the full lateral surface area Ss can be simply written as:
SS= peL [6.71]
Ф0 = Ssh(To -T)e = PeLh(To — T)e = PeLh#o [6.72]
So, the general expression for the ratio:
_ф _ hppe AS 00 th ( aL) _7( hpe)2 « 00 th («L)_ th ( aL)

п = Фо = PeLh0o = PeLh0o = aL
[6.73]
n = ,ы^ = 0529 = 089, _ 0.9

[6.74]
aL o.588
3) The balance equation for the fin must be resolved with the
imposed limit condition. In other words:
d2T
ASdx; + hpe (T - Te ) = 0 [6.75]
Introducing the variable:
0=T(x)-Te [6.76]
00=T0-Te [6.77]
and the parameter a such that:
a=_ê hp
[6.78]
AS
The equation becomes:
d^0 + «20 = 0 [6.79]

dx2
The two limit conditions become:
x=о ; e = t0 - Te = e0 [6.80]
x=L ; e = Te - Te = 0 [6.81]
The general solution can be expressed by a sum of sinusoidal functions or

hyperbolic functions. The second way of writing it is more suitable:
0 = Ash ax + Bchax [6.82]
The constants A and B are determined by the limit conditions:
x=0;e=e0 =Ash0+Bch0=B [6.83]
x=L;e=0=AshaL+BchaL [6.84]
B = e0 [6.85]
A shaL =-BchaL [6.86]
A =---- — [6.87]
thaL
For the expression 0( x) and the flow Ф exiting the wing (calculated at
x = 0), we have:
e1
— =-------- sh ax + ch ax [6.88]
e0 thaL
ф =- S A— - S A—
dx dx
[6.89]
. ( a
=-s a— e0 I---------- shax + ch ax
d x 01 thaL
x=0
254 Heat Transfer 1
a( -chax + shax) -ach (0)

Ф = - S A 00 —i -S A 00 [6.90]
0 thaL x=0
thaL
_ . 1 r—— 1
Ф = SЛО0 = JS2hpe00- [6.91]
thaL thaL
^ppe AS 00 ^/10*0,106*204*1,5.10-4 (300 - 30)

Ф= t=90,92W [6.92]
th( aL) = th (0,588)
with:
th (0.588) = 0.529 [6.93]
We overestimate the evacuated flow of:
90.92 - 25.72
=253% [6.94]
25.72
This is explained by the fact that the actual temperature found, presuming
that the flow at the end of the wing is zero, can be written as:
О
= -thaLshax + ch ax [6.95]
О0
0(L) = -00(thaLshaL + ch ax) = 00 ch*ax — sh*aL = -°^- [6.96]

ch ax ch ax
0( L ) = -00 (thaLshax + chax) [6.97]
This gives:
270
0( L )= T (L)-T =
e
270
ch 0.588
= 229.2 [6.98]
T(L)=229.2+30=259.2C [6.99]
We note that this temperature is much higher than T0 . The flow of losses
from the edges is necessarily much lower than would be found if the
temperature descended to Te at the end of the wing.
EXAMPLE 6.3.- Cooling of an IT component
We are intending to cool an IT circuit. The heat can only be evacuated

from this circuit through its upper section, whether it is plane or horizontal,
which takes the form of a square with sides a= 5 cm (see Figure 6.3).
The ambient temperature around the circuit is Te =25 °C.
Figure 6.3. Diagram of the component to be cooled
1) First, we allow the heat to dissipate by natural convection. The natural

convection coefficient is then h1 = 8 Wm-1K-1.
Knowing that the circuit must not exceed the temperature TC =70 °C,
what is the maximum flow Ф1 that can be evacuated by natural convection?
2) Second, a radiator is installed on the circuit, which is made of a square

plate A with sides a= 5 cm and five copper fins, vertical and with a height
L= 4 cm and thickness e= 3 mm (see Figures 6.4 and 6.5).
We will consider that the flow evacuated by each fin is calculated

independently of the presence of the other fins (no interactions between the
fins). Moreover, the foot of each wing will be at a temperature TC.
256 Heat Transfer 1
The limit temperature of the circuit is still TC =70 °C.
The natural convection coefficient maintains the same value

h2 = 8Wm-1K-1.
We set the thermal conductibility of copper to be AC = 386 W.m-1.K-1.
We consider the thermal contact between the circuit and plate A to be

perfect.
a) Calculate the flow of heat Ф A that is evacuated by each fin. To do so,

we will neglect the flow of heat emitted by the horizontal surface of each fin.
Figure 6.5. Geometrical characteristics of the fin

b) Calculate the total flow Ф 2 evacuated by the radiator.
3) We add a ventilator to the system. The convection coefficient then

increases to h3 = 30 W m-1K-1. Calculate the new value Ф3 of the total flow
that is evacuated by the radiator.
4) The circuit dissipates P = 19 W by the Joule effect. The limit

temperature of the circuit remains TC =70 °C.
Among the cooling solutions found in questions 1-3, which are

acceptable?
5) We return to the case in question 1, where the flow is evacuated by

natural convection. Through a flow Ф4 = 19 W that is dissipated by the
circuit, what is the superficial temperature Tmax of this circuit? Does this
seem acceptable to you?
6) Plate A is, by its very nature, connected to the upper part of the circuit
by a thin layer of x = 0.1 mm of a paste known as “thermal”, based on
silicon with thermal conductibility AP = 0.9 W.m-1.K-1. This paste avoids
the presence of air inclusions between the element and the plate which, even
if it is thin, would be a disturbing thermal resistance. We will denote TS2 as
the new temperature taken at the foot of each wing.
a) Establish two relations between the flow Ф 5 evacuated and the

temperatures TC and TS2 , on the one hand, and the temperatures Te and TS 2 ,
on the other hand, on either side of the layer.
b) Deduce from this the flow evacuated by the radiator in the case where
the ventilator is used. Does this value pose a problem?
Solution.-
1) Natural convection only

The exchange surface is:
SC = 25 cm2 = 25.10—4 m2 [6.100]

258 Heat Transfer 1
The flow evacuated by natural convection will therefore be determined

by the maximum temperature and the ambient temperature:
Ф1 = Sh1( TC - Te )= 25.10-4*8* (70 - 25) [6.101]
Ф1 = 0.9 W [6.102]
2) Radiator equipped with fins
a) Evacuation via a single fin
We consider the case of a parallelepiped wing of length L . The flow

will be neglected at the end.
The distribution of the temperature along the fin will be given by:
e = e0 th ( aL) [6.103]
The flow at the origin of the fin (flow removed from the body to be
cooled) will be given by:
Ф= aSoO th( aL) = hPpe AS e0 th( aL) [6.104]
where:
e0 =TC-Te=45°C [6.105]
We calculate the characteristics of the fin:
Perimeter:
pe =5+5+0.3+0.3=10.6cm=0.106m [6.106]
Surface area:
5 = 5*0.3 = 1.5cm2 = 1.5.10-4 m [6.107]

Parameter:
hp 8 8*0.106p
a = -^- =----------------------- — = 3.83 [6.108]
AAcs у 386*1.5.10-4
aL=3.83*4.10-2 =0.153 [6.109]
and
th(aL) = 0.152 [6.110]
hhpe As = 78*0.106*386*1.5.10-4 = 0.222 [6.111]
We deduce from this the flow evacuated by the fin:
ФA = Jhpe AS 00 th( aL) = 0.222*45*0.152 = 1.518 W [6.112]
b) We will easily deduce from this the total flow Ф2 evacuated by the
radiator.
Overall, the five fins will evacuate:
5 ФA = 1.518*5 = 7.59 W [6.113]
The flow to which the convection of the plane face must be added
becomes:
Ф2 = 7.59 + 0.9 = 8.49 W [6.114]
3) We add a forced convection device

We recalculate the characteristics of the fin:
Perimeter:
pe =5+5+0.3+0.3=10.6cm=0.106m [6.115]
260 Heat Transfer 1
Surface area:
5 = 5*0.3 = 1.5cmi = 1.5.10-4 m2 [6.116]
Parameter:
hp 30*0.106 „
a= =J---------------- = = 7.41 [6.117]
A^ 3 386*1.5.10-4
aL = 7.41*4.10-2 = 0.296 [6.118]
and
th(aL) = 0.288 [6.119]
hppe As = 330*0.106*386*1.5.10-4 = 0.429 [6.120]
We deduce from this the flow evacuated by the fin:
ФA = Vhpe AS 00 th( aL) = 0.429*45*0.288 = 5.56 W [6.121]
Overall, the five fins will evacuate:
5 ФA = 5.56*5 = 27.8W [6.122]
The flow to which the convection of the plane face must be added
becomes:
Sh1(TC-Te)=25.10-4 *30*(70-25)=3.37W [6.123]
Ф3 = 27.8 + 3.37 = 31.17 W [6.124]
4) The only acceptable solution is the one that evacuates more than
19W; that is, the one that includes the ventilator (question 3) is acceptable.
5) The flow evacuated by natural convection must now be Ф 4 = 19 W,

which will determine Tmax via the relationship:
Ф 4 = 19 = Sh1( TC - T )= 25.10-4*8* (Tmax - 25) [6.125]
19
Tm- = K1F4 + 25 = 975 °C [6'126]
This temperature is obviously unacceptable for a semiconductor

component.
6) An intermediate layer of thermal paste is applied

a) We establish two relationships between the flow ф 5 evacuated and
the temperatures TC and TS2 , on the one hand, and the temperatures Te and
TS2 , on the other hand, on either side of the layer.
Passing through the paste layer, the flow creates a difference in the
temperature TC- TS2 :
T-T 09
Ф5 = ' x = 10-4 = 9000 (TC — Ts 2) [6.127]
In addition, the relationship between the flow and Te , and TS2 remains
unmodified. We consider again the previous results:
Ф5 = 5 Jhpe AS (T - Ts2)th( ®L) + Sh1(TC - T) [6.128]
Ф5 = 0.429*(T -TS2)*0.288 + 25.10-4 *30*(T -TS2)
= [0.618 + 7.5.10-2 ](T - TS2) [6.129]
Ф5 = 0.693(T -Ts2) [6.130]
b) The flow evacuated by the radiator when the ventilator is used is then
re-evaluated as:
Ф5 = 9000 (TC - Ts2) [6.131]

262 Heat Transfer 1
So:
_ Ф
T - T =—5- [6.132]
C S2 9000
Ф5 = 0.693(Te - Ts2 ) [6.133]
So:
Ф
T - TS 2 =—— [6.134]
es2 0.693
Subtracting the two temperature differences:
Ф5
(T - T„,)-(T - T„ ) = — [6.135]
Cs2es20.693 9000
1 1
TC - Te =Ф 5 =1.443Ф5 [6.136]
0.693 9000
‘c__ j 45
Ф5 = ----- = 31.2 W [6.137]
1.443 1.443
This result is identical (with the calculation rounded, it should be slightly

less) to the result of question 3. The thermal paste fulfills its purpose well.
Appendices
Appendix 1
Heat Equation of a
Three-dimensional System
In this appendix, we present the heat equation in the general case of a

three-dimensional system.
A1.1. Reminder: writing the Fourier law
Figure A1.1. Case of two infinitely close isotherms; relationship with the gradient.
Let us consider two isothermal surfaces with respective temperatures T

and T+ dT, which are very close to each other. We have M as a point on
the isotherm T . We will consider two infinitely small surfaces dS1 and dS2
that are next to each other on each of the isotherms, on M . They have
266 Heat Transfer 1
practically the same normal й, conventionally directed from the surface of

temperature T towards the surface of T+ dT. Locally, we have the situation
found in a single-dimensional plane system.
The thermal flow that passes through dS1 and dS2 will be expressed by:
dT
dФ = -A d S1 [A1.1]
dx
where x is a measured axis on the axis bearing >7. We should recall that the
minus sign indicates that the heat propagates in the direction of decreasing
temperatures. In this expression, the sign implies a positive flow when it
flows in the direction of the normal.
Let us consider a surface dS' supported on the same section of cylinder

as dS1 and dS2, including the normal n', make an angle в with the
normal >7 that is common to dS1 and dS2 .
The same flow of thermal heat obviously passes through dS', dS1 and
dS2 .
We have the known geometric relationship:
dS1 = dS 'cose [A1.2]
and the flow passing through dS' is written as:
d ф = - AdTdS1 ==-A—dS 'cose [A1.3]

dx dx
Near M, the temperature gradient is reduced to a component along the

axis of x:
dT
grad7 T ) x
A----
dx
[A1.4]
Appendix 1 267
which can be rewritten as:
ggaaT T|| = AdT [A1.5]

dx
We note that:
I" fl=1 [A1.6]
In the end, the flow will take the form of a scalar product:
d ф= - ATxdS 'cosd = |\gaad т|||п '||costf TS1
'| ( )
\rrad т||||и cos grad T, n' d S' [A1.7]
Taking into account the orientations of the normal that are specified
above:
dф= -gaadT.n' TS1 [A1.8]
Therefore, the flow of heat that passes through a closed surface will be
written as:
ф= jj A grad T. n dS [A1.9]
S
By convention, the normals are always orientated towards the external

side of the closed surface.
REMARK.- This convention is imposed in order to correctly write the

integral theorems of vectorial geometry: the Stokes equation, Ostrogradsky’s
equation, etc.
All outgoing flow will thus be counted positively in the integral, and all
flows entering will be counted negatively. The integral therefore represents a
resulting flow balance, positive if the volume contained in the surface loses
heat.
268 Heat Transfer 1
The expressions drawn up here now allow us to write different

expressions for the thermal balance in a closed surface. The various forms
given to these balances will constitute the various forms of the heat equation,
the starting point for all thermal conduction problems.
A1.2. Heat equation
Let us consider a region given by a surface S. In the most general case,

the thermal conductibility of the material 2 is variable in D.
The balance of flows through S is given by
ф= jj2gradT.n dS . [A1.10]
S
This integral is positive if more heat exits than enters, negative if more
enters than exits.
In the former case, the flow will contribute, overall, to “cooling” the region
D. The sensitive quantity of heat integrated into D will diminish over a time
dt. In the latter case, the flow will contribute, overall, to heating the region
D. The sensitive quantity of heat integrated into D will increase over a
time dt.
Ф= jj2gradT.n dS [A1.11]
S
expresses heat transfer per unit of time.
The sensitive quantity of heat integrated into D is written as:
Q = ill D PcTda [A1.12]
where p is the (local) density of the material and c is its (local) specific heat
capacity. The balance equation can therefore be written as:
dQ
= -Ф [A1.13]
dt
Appendix 1 269
or
d- JJJ pcTda =- jj A grad T. n dS [A1.14]
Integrating the derivation with respect to the time in the integral and
transforming the integral of the surface into an integral for the volume, we
find
jjj pc^—da =-jjAgradT.n dS = jjj divAgradT a [A1.15]

dt 4
Rewritten as:
jjj
D Pc df-~ divAgradTda = 0 [A1.16]
This integral is written on a region that can be of arbitrary shape. It can

therefore only be zero if the integrant is zero everywhere. This gives the
equation for the point, general and classic shapes of the heat equation:
pc dL = div(AgradT) [A1.17]
In the case of a stationary problem, the derivative with respect to time

disappears, and we obtain:
(
div Agrad T = 0 ) [A1.18]
Case of constant physical properties
A significantly important case is found in homogeneous environments,

with constant physical properties, both in space and in time:
A = Cte, p = Cte, c = Cte, [A1.19]

270 Heat Transfer 1
The previous expressions are simplified and give forms in which the heat
equation is most generally known:
In the stationary regime:
2 div ( grad T ) = 0 [A1.20]
Therefore, in the end:
div grad T = 0 [A1.21]
or, more often:
ДT = 0 [A 1.22]
using the relationship:
div grad f = Д f [A 1.23]
In the stationary regime, the temperature is a harmonic function, where

Д is the Laplace operator.
In the non-stationary regime:
---- 2 div graddr

dT =---- T [A1.24]
d t pc
d T = a at
-- ДT [A1.25]
dt
Here, we see the fundamental parameter of the properties of the material

for non-stationary problems, the thermal diffusivity
2
a =--- [A1.26]
Pc
Therefore, generally, in problems with constant properties of the material,

the coefficient characterizing the material is only the thermal conductibility
Appendix 1 271
for stationary problems and the thermal diffusivity for non-stationary

problems.
Use of the Laplace operator

—*
Д T = div grad T [A1.27]
is very practical. Writing equations therefore remains valid for all coordinate
systems. It is therefore necessary to use the known form of the Laplace
operator for each system, in order to have the heat equations in this coordinate
system.
Therefore, we can easily rewrite the heat equation in all coordinate

systems, in particular, Cartesian, cylindrical or spherical. Obviously, we will
find the expressions established in the case of the simplified systems studied
in the main text of this book.
Appendix 2
Heat Equation: Writing in

the Main Coordinate Systems
The heat equation results from a balance between the heat that, by
conduction, enters, exits and accumulates in an elementary volume.
A2.1. The elementary volume
The geometry of this elementary volume depends on the coordinate

system used.
In an orthonormalized system (x,y,z), the volume is a parallelepiped

with sides (dx,dy,dz). Its volume is dx.dy.dz.
In a cylindrical system (r,0,z), the elementary volume is an annular

sector, with length dz , included between two cylinders with radius r and
r+ dr, and on a sector with angle d0. Its volume is r.dr.d0.dz.
In a spherical system (r, Ф,0), the elementary volume is determined by

two spheres of radius r and r+ dr, longitudes 0 and 0+ d0, and
colatitudes Ф and Ф + dФ. Its volume is r2sinФdrd0dФ .
274 Heat Transfer 1
Orthonormalized
Cylindrical system Spherical system
system
Reference and
Reference and coordinates Reference and coordinates
coordinates
Orthonormalized
Cylindrical system Spherical system
system
Elementary volume Elementary volume Elementary volume
A2.2. Problem with variable physical properties
In the most general case, the volumetric mass p, the specific heat
capacity °C and, in particular, the thermal conductibility A can vary with
temperature. The heat equation is then written as follows:
Appendix 2 275
In the orthonormalized system
^3T д К dT I д К dT I д К dT
pC=I AI + I AI+I A [A2.1]
dt dx 1 dxJ Эy 1 dyJ dz 1 dz
In the cylindrical system

dT 1 d | dT | 1 d | dT | д | dT
pC--- =-------- 1 A r I +----------1 A-I +—I A---- [A2.2]
dt r dr 1 dx J r2 00[ d0 J dz 1 dz
In the spherical system

nr3 T [A2.3]
pC Э7
1 d П 2 - dT T | 1 d | д dT | d ( Q . d T |ГЛП/|-.
—I A r2sinФ— I + —---------- 1 A— I + 1 AsinФ I [A2.4]
r 2 sin Ф Э r 1 Э x J sin ФЭ01 d0 J ЭФ1 ЭФ J
A2.3. Problem with constant physical properties
A2.3.1. The problem can often be simplified
If the density p, the specific heat capacity °C and, in particular, the

thermal conductibility A are independent of the temperature, we then define
the thermal diffusivity a:
a=A [A2.5]
pC
and the heat equation becomes
^T = a Д T [A2.6]
dt
where Д T is the Laplace operator for the temperature T (x, y, z, t).

276 Heat Transfer 1
A2.3.2. This Laplace operator is written, in each of the

coordinate systems

д 2 t d2 T д2 T
ДT ----- 1------ 1----- [A2.7]
д x2 д y2 д z2

1 д д T 1 д2 T д2 T
Д T =------ r +--------+---- [A2.8]
r д r д x r 2 Э02 д z 2

1 д ддTT 1 д2 T д . д дT
ДT = — r 2sin Ф-----I ' + sin Ф [A2.9]
r 2sin Ф д r д x sin Ф д02 дФ дФ
A2.3.3. Hence the expressions for the heat equation in a

time-dependent regime

дT д2 T д2 T д2 T
— = a------ I-------- 1------- [A2.10]
дt ^ д x2 д y2 д z 2

дT 1 д д T 1 д2 T д2 T
---- = a------ r------- 1------------ 1------- [A2.11]
дt r д r д x r 2 д02 д z2
дT a д 1 д2 T
д T +, ------------- ^дT
r 2sin Ф----- 1 ---- sin Ф----- [A2.12]
дt r 2sin Ф д r д x sin Ф д 02 дФ дФ
Appendix 2 277
A2.4. Time-independent regime
In numerous cases, the temperature does not depend on the time:
дТ=о
[A2.13]
t
A2.4.1. In the case of variable physical properties
The heat equation then becomes:

д ( TT i д ( TT i д ( TT
---- I A---- I +---- I A---- I +---- I A---- [A2.14]
д x ^ д x ) дy l дy ) дy l дy

1 д К дt i 1 д ( . дT | д ( дT |
A r---- I + --------I A I + —I A---- I = 0 [A2.15]
r дr l д x I r2 д0\ д0 I дz l дz I
д ( дT i 1 д ( д дT i д Г . л дT
Ar 2sin Ф— I +------------ 1 A— I +------ 1 Asin Ф-----
=0 [A2.16]
дrl д x I sin Фд#1 д0 I дФ l дФ
A2.4.2. In the case of constant physical properties
In this case, the Laplace operator of the temperature is zero. The value of
the physical properties no longer intervenes in the general expression for the
temperature. Note that the thermal conductivity intervenes in the calculations
for flows
AT = 0 [A2.17]
278 Heat Transfer 1

d2 T d2 T d2 T
------ 1------- 1------ [A2.18]
d x2 д y 2 д z 2

d T 1------------
1 d r-------
----- d2 T
1 d 2 T 1-------
[A2.19]
r д r д x r 2 д02 d z2

д d T +, 1 d2T . ^dT
r 2sin Ф----- +Sin Ф =0 [A2.20]
дr d x sin Фд02 дФ дФ
A2.5. Writing the Fourier law. How can the expression for the
gradient be found?
The Fourier law is written using the temperature gradient:

^
—>
dp = ZgradT. ndS [A2.21]
It is therefore useful to know how to find the expression for the gradient
in various coordinate systems. To do so, we need to know:
a) How to write the expression for the components of the small
displacement vector dl due to an elementary modification of the
coordinates.
This gives a vector of components.

dx
dl dy [A2.22]
dz
Appendix 2 279

dr
—*
dl rd0 [A2.23]
dz

dr
—
dl rd Ф [A2.24]
r sin Ф d0
b) Recall that the temperature gradient is normal at the isotherms.
If dl is a vector that is a tangent to the surface isotherm T= Cte, the

scalar product of the gradient and of dl will be zero.
In the displacement dl, T therefore does not vary. The exact total
differential dT is zero.
Writing in each system of coordinates, we will find gradT leading to the

components of dl in the differential dT.
Figure A2.1. Relationship of the gradient to the isotherm, a particular

case of the relationship of the gradient to the equipotential
280 Heat Transfer 1

_ дT TT dT - ,T n
d T =--- dx +------ dy +------ dz = grad T . dl = 0 . [A2.25]
дx дy дz
Therefore, the components of the gradient are:
дT
дx
-
дT
grad T [A2.26]
дУ
дT
дz

,t дT TT дT
dT =--- dr +------ dU +------- dz [A2.27]
дr дU дz
= ^dr + 1 ÛrrclO + д~^ = gradT . dl = 0 [A2.28]
Therefore, the components of the gradient in the cylindrical system are:
дT
дr
-
grad T 1 IT [A2.29]
r ди
дT
дz

д T. дT дT
dT =--- dr +------- d Ф+------ dd [A2.30]
дr дФ дU
Appendix 2 281
дT 1 дT 1 дT ^ -
=---- dr +-------- rdФ+------------- rsinФdd = gradT. dl = 0 [A2.31]
дr r дФ rsinФ дz
Therefore, the components of the gradient in the spherical system are:
дT
- i"t a
gradT - д— [A2.32]
1 дT
r sin Ф д 0
Appendix 3
One-dimensional Heat Equation
In this appendix, we present the single-dimensional heat equation in the

case of cylindrical and spherical geometries. As specified in section 2.3, we
have included the details of deriving this heat equation for these two cases.
A3.1. Case of an axisymmetric system
In this case, we will no longer talk about an Ox axis, but we direct each
radius positively from the axis of the cylinder towards the outside of the
cylinder. Reasoning in the same way as before, we obtain the following
conclusions.
Therefore, we have a single-dimensional field. The temperature depends

only on r, the distance from a point at the axis. The isotherms are therefore
cylinders with radius r : T= T (r).
We no longer reason in m2 of surface area, but we reason for a length of

1 meter of cylinder1.
The flow per meter of cylinder through a cylinder of radius r and of

a unit length is then
dT
- A--- 2nrL, with L = 1, that we will no longer write out in the following.
dr
1 Indeed, in the rest of the calculation, we see that the surface through which the conduction
takes place is proportional to r and is therefore going to remain within the derivative.
284 Heat Transfer 1
dT < 0 when the heat goes from the axis towards the outside, and
dr
dT
- A---- is positive when the heat goes in the direction of increasing r.
dr
dT
— > 0 when the heat goes from the outside towards the axis, and
dx
dT
- A---- is negative when the heat goes in the direction of decreasing r.
dr
dT
Calculate the change in - A--- 2nr when we go from r to r + dr, with
dr
dr being infinitely small:
dT dT dd \^ dT L
- A- 2nr becomes A 2nr +----- 1 A2nr— I dr . [A3.1]
dr dr dr 1 dr I
And, changing the sign,
-
dT
A A—2nr +-- 1 I -A2nrd------
2nr becomes -dT^d dTIU dr . [A3.2]
dr dr dr I dr I
We note that an “r” remains below the second derivative of the last
expression.
Hence the need to work on a linear meter of cylinder and not a m2.
We will now consider a cylinder located at r+ dr. The quantity of heat

that passes through it per meter of length of cylinder is:
dT d ( dT к
- A-2nr +----- 1 -A2nr---- I dr [A3.3a]
dr dr 1 dr I
Appendix 3 285
Taking up the reasoning set out in r :
dT d
-A 2nr +— dr is positive when the heat goes in the
dr dr
direction of increasing r (s moving away from the axis).
dT d ( dT)
If -A----2nr +---- I -A2nr— I dr is negative, then the heat goes in the
dr dr 1 dr I
direction of the decreasing r (moving closer to the axis).
By a reasoning that is analogous to the plane problem, we write the

thermal balance in the small volume 2nr dr in a time dt:
dTI dT d ( dT V
dQ = pc 2nr drdT = —A 2nr + A 2nr +----- I +A2nr I dr dt [A3.3b]
dr dr dr 1dr J
,
p. crdrdT d I +Ar dT I,
= — „
I dr dt [A3.4]
dr 1dr J
д T 1 d Г dT
---- =------ I Ar---- [A3.5]
дt r дr 1 dr
with the two specific cases:
Constant physical properties:
дT a д ( дT ^
-r- = -V-1 I [A3.6]
дt r дr 1 дr I
Stationary conduction:
d ( dT | ..
— I r---- I = 0 [A3.7]
dr1 dr I
We note that r is still under the first derivative, which is characteristic of

equations with cylindrical symmetry.
286 Heat Transfer 1
A3.2. Case of a spherical system
In this case, we no longer talk about an axis Ox, but we positively direct
each radius of the axis of the cylinder to the outside of the sphere.
Reasoning in the same way as before, we reach the following conclusions.
We have a single-dimensional field. The temperature only depends on r,

the distance from a point at the origin of the coordinates, the center of the
considered spherical system. The isotherms are therefore spheres with
radius r : T= T (r).
We now reason in terms of flow passing through a sphere.
The balance will be established for a space between two spheres that are
infinitely close to radius r and r+ dr.
We recall that the surface of a sphere of diameter d= 2 r is:
S = nd2 = 4nr2 [A3.8]
and that its volume is:
V = 4nr3 = nd3
[A.9]
^ 6
A3.2.1. The flow through a sphere of radius r
dT
The flow through a sphere with radius r is then - 2----4nr2.
dr
dT dT
If —< 0, the heat goes from the origin towards the outside, and — 2----
dr dr
is positive when the heat goes in the direction of increasing r.
If dT > 0, the heat goes from the outside towards the axis, and — 2ddT- is
negative when the heat goes in the direction of decreasing r.

Appendix 3 287
When we go from r to r + dr, dr is infinitely small, - A----4nr2

dr
becomes:
-A—4nr2 +— f-A4nr2 — 1 dr [A3.10]

dr dr ^ dr J
We note that an “r2” remains under the second derivative of the last
expression.
Let us now consider a sphere of radius r+ dr. The quantity of heat that
passes through it per meter of length of the cylinder is:
-A—4nr2 +d f-A4nr2 — 1 dr [A3.11]

dr dr ^ dr J
Using the reasoning made for x
-A^— 4nr2 +— f -A4nr 2 I dr is positive when the heat goes in the

dr dr ^ dr J
direction of increasing r (s moving away from the axis).
Therefore, -A----4nr2 +---- f -A4nr2----- I dr is negative when the heat

dr dr ^ dr J
moves in the direction of decreasing r (moving closer to the axis).
We will set up the balance of incoming and outgoing heat by conduction

in the space between two spheres of radius r and r+ dr.
Through the “interior” sphere:
dT
- A----4nr2 is positive if the heat goes in the direction of increasing r.
dr
In this case, the heat enters the space and the term is counted positively in
the balance.
288 Heat Transfer 1
Through the plane r + dr, the incoming heat must also be counted
positively in the balance.
ddT d I _ dT )
Now -X—2nr +---- 1 -X2nr— I dr is positive when the heat goes in
dr dr 1dr J
the direction of increasing r, in other words when the heat exits the annular
space.
This is therefore the opposite term, which will intervene in the

balance, in other words A^—4nr2 +— I X4nr 2^1— I dr .
dr dr 1 dr J
Hence the resulting term for conduction in the balance:
ddT . 2 ddT . 2 d I 2 dT | , d I 2 dT |
-X 4nr2 + X--- 4nr2 +— I X4nr2— I dr = — I X4nr2 r I [A3.12]
dr dr dx 1dr Jdr 1dr J
By reasoning in the same way as in the previous problem, we write the

thermal balance in the small volume 4nr2 dr on a time dt:
dT 2dT 2 d 2 dT \
dQ = pc 4nr 2 drdT = -X 4nr + X 4nr X4nr ---- I dr dt [A3.13]
dr dr dx dr J
d 2 dT L
pc 4nr 2 drdT = X4nr ---- I dr dt [A3.14]
dx dr J
д dT 1 d Г 2dT
pc=1 Xr2 [A3.15]
d t r2 d r 1 dr
with the two specific cases:
Constant physical properties:
dT a d ( 9dT^
-:;-=—I r I [A3.16]
d t r2 d r 1 dr J
Appendix 3 289
Stationary conduction:
d
[A3.17]
dr dr J
We note that r2 is maintained within the first derivative, which is

characteristic of equations with spherical symmetry.
Appendix 4
Conduction of the Heat in a

Non-stationary Regime:
Solutions to Classic Problems
Hypotheses:
The media are homogeneous.
Their properties are constant.
Specific heat capacity c, density p, thermal conductibility 2 and
consequently, thermal diffusivity a = are constant.

pC
The problems are spatially single-dimensional: the temperature only

depends on a variable in space x or r and on the time t.
The definition of the variable в (re-dimensioned or reduced) varies from

one problem to the next.
The solutions use a composite variable % and various special functions.
xx
% = ^= = ~n= [A4.1]
at 4a at
292 Heat Transfer 1
Error function and extensions, tabulated in Appendix 5:
erf (П-T f'e - ud [A4.2]

П 0
erfc (g)-1 —2= f *e - u2 du [A4.3]

П 0
ierfc (g) = iexp (-g2)- gerfc (g) [A4.4]

П
Problem 1: medium with temperature imposed on a wall
Initial temperature of the environment Ti
Imposed temperature Te at x = 0, at time t = 0
g = T (x. t)-T [A4.5]

T - Te
g = er (g)
f [A4.6]
Problem 2: medium with flow density <p0 imposed on a wall
Constant density of flows ^0 imposed at x = 0, at time t = 0
g-T(x,t)-Ti [A4.7]
3 = ~'J^yiat~ierfc (g) [A4.8]

Appendix 4 293
Problem 3: medium with sinusoidal variation of temperature

imposed on a wall
Temperature T(t) imposed at x = 0
T (0, t ) = Tt cos (mt) [A4.9]
“Stationary regime” solution:
0 = T (x, t)-T [A4.10]
( Im } ( x A
0 =T i exP -J— x cos mt -J— x
I 2aa J I 2aa J
[A4.11]
amplitude:
( ,.
m
exp - — x [A4.1a]
I 2aa J
phase shift:
mx x
[A4.13]
a2 a
Problem 4: abrupt contact between tw o media with respective

homogeneous temperature T1 and T2
Border between the two environments at x = 0 ; contact at time t = 0
0 = T1 (x, t)-T i [A4.14]

1 T 2 - T1
294 Heat Transfer 1
q
T(x, t)- Ti 2
[A4.i5]
2 - Ti1
Ti2-
Effusivity of media 1and 2
Ei 4*i Pi C [A4.i6]
E2 /
— у ^2 P2 C2 [A4.i7]
medium i for x<0
Qi = E1 +1 E2 erf (I^) [A4.i8]
medium 2 for x>0
E
Q = E1+E2 erfc(^)
Appendix 5
Table of erf (x), erfc(x)

and ierfc(x) Functions
These three functions erf (x), erfc ( x) and ierfc(x), are fundamental in
processing single-dimensional non-stationary conduction problems.
They are defined by:
erf (x), erfc(x) and ierfc ( x) functions
erf (x) = П Jo exp(-^2)d^ [A5.1]
erfc (x ) = 1 — П Jx exp (—£2) d^ [A5.2]
ierfc(x) = —L- exp(-x2) -xerfc(x) [A5.3]

П
x erf (x) erfc (x) ierfc (x) x erf (x) erfc (x) ierfc (x)
0 0 1 0.56418958 1 0.84270079 0.15729921 0.05025454
0.05 0.05637198 0.94362802 0.51559947 1.1 0.88020507 0.11979493 0.03646538
0.1 0.11246292 0.88753708 0.46982209 1.2 0.91031398 0.08968602 0.02604895
0.15 0.16799597 0.83200403 0.42683646 1.3 0.93400794 0.06599206 0.01831432
296 Heat Transfer 1
0.2 0.22270259 0.77729741 0.38660791 1.4 0.95228512 0.04771488 0.01267002
0.25 0.27632639 0.72367361 0.34908866 1.5 0.96610515 0.03389485 0.00862286
0.3 0.32862676 0.67137324 0.31421848 1.6 0.97634838 0.02365162 0.00577194
0.35 0.37938205 0.62061795 0.28192557 1.7 0.98379046 0.01620954 0.0037993
0.4 0.42839236 0.57160764 0.25212759 1.8 0.9890905 0.0109095 0.00245876
0.45 0.47548172 0.52451828 0.22473278 1.9 0.99279043 0.00720957 0.00156419
0.5 0.52049988 0.47950012 0.19964123 2 0.99532227 0.00467773 0.00097802
0.55 0.56332337 0.43667663 0.17674618 2.1 0.99702053 0.00297947 0.00060095
0.6 0.60385609 0.39614391 0.15593537 2.2 0.99813715 0.00186285 0.00036282
0.66 0.64937669 0.35062331 0.13354955 2.3 0.99885682 0.00114318 0.0002152
0.7 0.67780119 0.32219881 0.12009827 2.4 0.99931149 0.00068851 0.00012539
0.75 0.71115563 0.28884437 0.10483226 2.5 0.99959305 0.00040695 7.1762E-05
0.8 0.74210096 0.25789904 0.09117366 2.6 0.99976397 0.00023603 4.0336E-05
0.85 0.77066806 0.22933194 0.07900271 2.7 0.99986567 0.00013433 2.2264E-05
0.9 0.79690821 0.20309179 0.06820168 2.8 0.99992499 7.5013E-05 1.2067E-05
0.95 0.82089081 0.17910919 0.05865589 2.9 0.9999589 4.1098E-05 6.4216E-06
1 0.84270079 0.15729921 0.05025454 3 0.99997791 2.209E-05 3.355E-06
These functions are represented in Figure A5.1.
Figure A5.1. erf %, erf % and ierfc^ functions. For a color

Appendix 6
Complementary Information
Regarding Fins
Here, we provide the solutions that result from a single-dimensional

conduction model
T =T(x)
A6.1. Rectangular wings. Solutions to classic problems
Fin with cross-section e* l and length L
S: surface of the cross-section of a fin; S= e*l
pe : perimeter of fin; p =2(l+e)
X: thermal conductibility of the material that makes up the fin
h : convection coefficient between the fin and the surrounding

atmosphere
T0 : temperature of the fin at its origin (temperature of the body to cool

down)
Te : ambient temperature
298 Heat Transfer 1
We define a parameter a:
hp>
a=,-?■ [A6.1]
AS
Relative temperatures:
0 = T ( x)-T [A6.2a]
00 = To - T [A6.2b]
Case 1: Infinitely long fin
Distribution of the temperature along the fin:
в = 0oexp (-ax) [A6.3]
Flow at the origin of the fin (flow taken from the body to be cooled):
Ф= ASa0o =Jhpe AS 0o [A6.4]
Case 2: Fin of length L. Zero flow at the extremity
Distribution of the temperature along the fin:
0 = 0o th ( aL) [A6.5]
Ф= ASa0o th( aL)=y/hpe AS 0o th( aL) [A6.6]

Appendix 6 299
Case 3: Wing of length L. Convection coefficient h at the end
Distribution of the temperature along the wing:
cosh [m(L -x)] + ^l sinh [m(L -x)]

0 = 00 h [A6.7]
cosh [mL ] +----- sinh [mL ]
ml
th [mL }+-J ,_______ th [mL }+-J

Ф = lSm00------ ----- mi=Jhpe IS 00---- - ----- ^^ [A6.8]
1+ —th [mL ] 1+ —th [mL ]
mi L J ml L J
Appendix 7
The Laplace Transform
The Laplace transform is a precious tool for the resolution of differential

equations with partial derivatives, in particular, with two variables. Indeed, it
allows this equation to be transformed into an algebraic equation, where only
derivatives with one variable continue to be present.
In problems that depend on time, the Laplace transform is often used with
respect to this variable.
It was therefore widely used to establish the main solutions for

non-stationary conduction, like in this book, or even in signal theory.
We give little basic information about this method. We have chosen to

shorten this for our readers. For most complex problems, numerical methods
are applied more often.
The inverse transform, or Mellin-Fourier transform, is not used in practice

due to the mathematical difficulties posed by the inversion of the expressions
found in physics.
In practice, we compare the result found with the Laplace transform with
a table of transforms that were previously established. We find many in the
literature.
302 Heat Transfer 1
A7.1. Definition
The Laplace transform transforms a function of a variable, for example,

f (t) in another function F(p). We note that F is a function of another
variable p .
The function f (t) must be real, locally integrable, defined for t > 0 . It
must be limited:
|f (t)|<Kea [A7.1a]
It is defined by:
f (tH F( p) = f-e-ptf (t)dt [A7.1b]
The notation
^ F ( p ) = £ f (t), [A7.2]
or an analogue, is often adopted. We often write the origin function with a

lowercase letter and the destination function in capital letters, whether the
alphabet used is Roman or Greek. This is not an absolute law, but a practical
convention that we will adopt here.
The properties of the transforms of derivatives and primitives of the

origin function make the transform particularly useful.
A7.2. Derivatives and integrals
df(t)
Transform of a derivative with respect to time-------- :
dt
£ d (t)= Ь ■" d (t) dt = e ' "f (t )|0-1- pe ’ "f (t) dt [A7.3]
£ ^dtr) =— f (0)+pГ e ~ "f ()) dt [A7.4]
after an integration by parts.

Appendix 7 303
We therefore have an important property:
df(t)
d
£ ( = -f (0) + pF (p) [A7.5]
We see the presence of the term - f (0). In many cases, f (0)= 0, but
we must always keep the existence of this term in mind in each new
problem, in the case where this condition of nullity at the origin of the signal
time is not fulfilled. We also see why it is important that the function is
limited, beyond the fact that this property guarantees the convergence of the
definition integral.
Transform of an integral in time I f (t)d t: [A7.6]
£ | f (t) dt = J 0~ e - pt | f (t) dtdt = F(p) [A7.7]
A7.3. We will give two examples
A7.3.1. Transform of the Heaviside step function
The Heaviside function is defined by:
t<0 ; H(t) = 0 [A7.8]
t>0; H(t)=1 [A7.9]
Its transform will be:
- e
-pt
TO
H(t)^ F( p) = [~e-ptdt = [A7.10]

p 0 p
In the same way, all multiples of the Heaviside function, such that
t<0;f(t)=0 [A7.11]
304 Heat Transfer 1
t>0; f(t)=a [A7.12]
will be transformed into:
-ae-pt a
f(t) ^F(p) = 10 ae-ptdt = [A7.13]
p p
A7.3.2. Transform of the erfc function
We will simply give the result with a view to a later calculation
The function f(t, x) = erfc axr

Л/ t
exp-axdp
is transformed into F (p ) =------------ — [A7.14]
p
A7.4. Resolution of a problem of single-dimensional non-stationary

transfer
The interest of the transform appears for a differential equation with

partial derivatives. For example, for the heat equation:
''^ = afx) ^ pF( p,x) = aiFpx,

[A7.15]
dt d x2 d x2
We see that F(p, x), the transform for f (t, x), responds to a differential
equation with one variable.
It will then be necessary to also transform the limit conditions.
We will apply this method to resolving a problem to which we have

already found the solution in section 5.1.1, in other words the non-stationary
conductive transfer at a fixed wall temperature at the initial instant.
Appendix 7 305
This time we are looking for the solution via a new reduced temperature:
T-T
в( x, t ) = T - T [A7.16]
ei
The notations for the temperatures remain as those in section 5.1.1.
This new definition is motivated by the desire to have an unknown that is

canceled out at the time origin.
Answers the equation
d f (x, t) Ъ 2 f (x, t)
[A7.17]
dt a d x2
with the conditions to fulfill:
в( x,0) = 0 [A7.18]
в( 0, t ) = 1 [A7.19]
0(™, t ) = 0 [A7.20]
в( x, t) is transformed into 0( x, p)
£в( x, t) = 0( x) [A7.21]
0(<™, t ) = 0 [A7.22]
We apply the Fourier transform to the equation:
-^— - p 0(x) = 0 [A7.23]

d x2 a
taking into account:
£ d^ = p Q-в (t = 0) = p 0 [A7.24]
dt
306 Heat Transfer 1
The integration provides a general solution:
0( x, p ) = A exp px + B exp - px [A7.25]

a a
The conditions are transformed into:
0( 0, t ) = 1 gives:
£0 (0, t ) = 0( 0, t ) = - [A7.26]
p
At infinity, 0( x, p) must remain finite.
The constants are determined by these conditions.
At infinity, the positive exponential is divergent; therefore: A = 0
0 (0, p) = —, therefore:
0 (0, p) = B exp - p0 0 = 1 [A7.27]

a p
and:
B =- [A7.28]
p
Therefore, the following remains:
fp
exp - x
0( x, p ) =
a [A7.29]
p
which is of the form F(p)= [A7.30]

p
Appendix 7 307
1
with a = —;= [A7.31]
fa
The solution to the problem will therefore be:
0( x, t ) = erfc x [A7.32]
4 at
where we recognize the variable £( x, t) = .----- [A7.33]

4a at
which is very coherent with the solution found in section 5.1.1:
T-T T-Ti+Ti-T T-Ti

------ =------- '■ \---- - = 1-------- - = 1 - erfc £ = erf £ [A7.34]
Ti-Te Te-Ti Te-Ti
Appendix 8
Reminders Regarding
Hyperbolic Functions
Hyperbolic functions are defined by:
exx - e--xx
hyperbolic sinus shx =---- ^---- [A8.1]
x -x
e +e [A8.2]
hyperbolic cosinus ch x =---- 2----
shx ex - e-x
hyperbolic tangent thx =----- =---------- [A8.3]
ch x ex + e-x
We can also define a hyperbolic cotangent
1 ch x ex + e-x
coth x =---- =------=---------- [A8.4]
th x sh x ex - e-x
The denomination of hyperbolic sinus, cosinus and tangent comes from

the relative similarity of the relationships found between these functions,
with the same relationships found for sinusoidal functions.
It must be noted that the similarity stops at these relationships and that
the form of the hyperbolic functions is totally different from that of the
sinusoidal functions.
310 Heat Transfer 1
Figure A8.1. Representation of the hyperbolic functions. For a

color version of this figure, see www.iste.co.uk/ledoux/heat1.zip
The main relationships to know are:
ch2x -sh2x =1 [A8.5]
sh(-x)=-shx [A8.6]
ch (-x) =ch x [A8.7]
th (-x)=-th x [A8.8]
dshx
------- = chs [A8.9]
dx
dchx
-------- = shx [A8.10]
dx
dthx 1
[A8.11]
dx ch 2 x
Appendix 8 311
sh(x+y)=shxchy+chxshy [A8.12]
ch (x +y)=ch x ch y +sh x sh y [A8.13]
which implies:
sh(x-y)=shxchy-chxshy [A8.14]
ch (x -y)=ch xch y -sh xsh y [A8.15]

References
[BYR 66] Byron Bird R., Stewart W.E., Lightfoot, E.N., Transport Phenomena,
John Wiley & Sons, New York, 1966.
[COS 97] Cosar P., Aide memoire du thermicien, Amicale des anciens eleves de 1’Ecole de
thermique (ed.), Elsevier, Paris, 1997.
[GOS 93] GOSSE J., Guide technique de thermique, Dunod, Paris, 1993.
[HOL 86] HOLMAN J.P., Heat Transfer, McGraw Hill, New York, 1986.
[JAN 16] JANNOT Y., MOYNE C., Transferts thermiques, Edilivres, Paris, 2016.
[KNU 58] KNUDSEN J.G., KATZ D.L., Fluid Dynamics and Heat Transfer, McGraw Hill,
New York, 1958.
[LEO 79] LEONTIEV A., Theorie des echanges de chaleur et de masse, Editions MIR,
Moscow, 1979.
[PER 97] PEREZ J.P., Thermodynamique, fondements et applications, Masson, Paris, 1997.
[ROH 98] Rohsenow, W.M., Hartnett, J.P., Cho, Y.I., Handbook of Heat Transfer,
McGraw Hill, New York, 1998.
[SAC 15] SACCADURA J.F., Transferts thermiques, initiation et approfondissement, Lavoisier
Editeur, Paris, 2015.
[VAN 76] Van Wylen G., Sonntag R.E., Fundamentals of Classical Thermodynamics,
SI Vers 2, John Wiley & Sons, New York, 1976.
Index
C, E F, G, H
conduction fin, 237-240, 242, 243, 245, 246,
non-sationary, 2, 183, 184, 190, 249-252, 255, 256, 258-260,
191, 195, 200, 225, 229, 295, 297-299
301, 304 flow, 1, 3-5, 10, 12-16, 18, 19, 21,
stationary, 2, 25, 26, 285, 289 27, 28, 30, 32-34, 37, 40-44, 46,
convection, 12, 27-29, 37, 44, 46, 51, 48, 53-60, 63-66, 70-72, 75, 77,
65-67, 69-71, 73, 74, 84, 85, 79-90, 92, 95-101, 107, 108, 111,
87-92, 95-101, 103, 105, 106, 110, 114, 117-119, 124, 125, 129, 130,
113, 117, 119, 122, 127-129, 132, 132-135, 137-139, 143-145, 148,
134, 138, 144, 152-155, 159, 162, 149, 152, 153, 156, 157, 159, 160,
165, 167, 168, 172, 173, 218, 223, 163, 166, 168, 172, 174, 175, 177,
237, 239-241, 244, 249, 252, 178, 180, 190, 191, 202, 207, 214,
255-261, 297, 299 224, 225, 233, 235, 237-241, 243,
coefficient, 5, 10, 11, 29, 37, 46, 244, 246, 248-250, 252-261,
66, 67, 69-71, 73, 74, 84, 85, 87, 266-268, 277, 283, 286, 292, 298,
89-92, 95-101, 103, 105, 106, 299
110, 113, 117, 119, 122, 127, density, 4, 5, 10, 19, 28, 30, 32, 33,
128, 132, 134, 138, 144, 152, 46, 53, 54, 56-58, 63, 65, 75, 80,
155, 159, 160, 165, 167, 168, 83, 117, 119, 190, 191, 214, 224,
172, 173, 175, 218, 237, 240, 225, 233, 235, 292
244, 249, 250, 255-257, 271, Fourier’s law, 9, 10, 13, 14, 16, 265,
297, 299 278, 301, 305
error function, 189, 203, 205, 207, gradient, 2, 3, 14, 15, 27, 28, 103,
209, 292 238, 240, 250, 265, 266, 278-281
316 Heat Transfer 1
heat thermal
equation, 1, 2, 14, 16-18, 23-25, conductibility, 11, 12, 22-24, 29,
27-30, 35, 62, 104, 183, 185, 30, 33, 34, 37, 38, 41, 47, 54-56,
186, 191, 194, 201, 265, 58, 62, 64, 70, 77, 82, 85, 86, 90,
268-271, 273-277, 283, 304 91, 93, 99, 100, 112, 129, 142,
propagation, 14, 18, 266 143, 152, 153, 156, 161, 167,
176, 184, 205, 210, 213, 228,
I, R, T 232, 256, 257, 268, 271, 274,
275, 291, 297
isothermal, 2, 3, 9, 14, 18, 24, 44, 47,
conduction, 1, 2, 10, 12, 16, 27,
191, 194, 200, 238, 240, 246, 249,
239, 268
252, 265, 279, 280, 283, 286
diffusivity, 23, 184, 190, 192, 194,
surface, 2, 3, 14, 265
199, 201, 202, 206, 208, 211,
radiation, 12, 27-29, 105, 107, 249
220, 225, 230, 233, 270, 271,
temperature, 1-3, 5, 6, 9-17, 22, 24,
275, 291
26-31, 34-38, 40, 44, 47, 49, 51,
flow, 4, 10, 14, 16, 28, 46, 48,
54, 55, 57-62, 65, 69, 71, 73, 77,
54-58, 65, 70, 71, 77, 90, 92,
81, 84-89, 95, 99, 100-104,
124, 132-134, 137, 153, 154,
107-111, 114, 116-118, 122-125,
166, 191, 233, 246, 266
127, 128, 130-134, 137, 138, 143,
ladder, 190
144, 147, 148, 152, 153, 156-160,
resistance, 29, 30, 33, 34, 37, 40,
162, 164-166, 171, 173-176,
43, 53-59, 61-67, 69-79, 81-88,
178-180, 183-186, 190-195,
90-93, 96-98, 100, 103, 106,
199-211, 213-233, 236-241, 246,
110-113, 116, 122-124,
249, 250, 254-258, 261, 262, 265,
127-130, 132, 134, 144, 145,
266, 270, 274, 275, 277-279, 283,
148, 153, 157, 160-162, 165,
286, 291-293, 297-299, 304, 305
167, 168, 171, 172, 174, 177,
217, 257
Other titles from
in
Mechanical Engineering and Solid Mechanics
2021
DAHOO Pierre Richard, POUGNET Philippe, EL HAMI Abdelkhalak
Applications and Metrology at Nanometer Scale 1: Smart Materials,
Electromagnetic Waves and Uncertainties
(Reliability ofMultiphysical Systems Set - Volume 9)
2020
SALENQON Jean
Elastoplastic Modeling
2019
BAYLE Franck
Reliability ofMaintained Systems Subjected to Wear Failure Mechanisms:
Theory and Applications
BEN KAHLA Rabeb, BARKAOUI Abdelwahed, MERZOUKI Tarek
Finite Element Method and Medical Imaging Techniques in Bone
Biomechanics
(Mathematical and Mechanical Engineering Set - Volume 8)
IONESCU Ioan R., QUEYREAU Sylvain, PICU Catalin R., SALMAN Oguz Umut
Mechanics and Physics of Solids at Micro- and Nano-Scales
LE VAN Anh, BOUZIDI Rabah
Lagrangian Mechanics: An Advanced Analytical Approach
MICHELITSCH Thomas, PEREZ RIASCOS Alejandro, COLLET Bernard,
NOWAKOWSKI Andrzej, NICOLLEAU Franck
Fractional Dynamics on Networks and Lattices
SALENQON Jean
Viscoelastic Modeling for Structural Analysis
VENIZELOS Georges, EL HAMI Abdelkhalak

Movement Equations 5: Dynamics of a Set of Solids
(Non-deformable Solid Mechanics Set - Volume 5)
2018
BOREL Michel, VENIZELOS Georges
Movement Equations 4: Equilibriums and Small Movements
FROSSARD Etienne
Granular Geomaterials Dissipative Mechanics: Theory and Applications in
Civil Engineering
RADI Bouchaib, EL HAMI Abdelkhalak
Advanced Numerical Methods with Matlab® 1: Function Approximation
and System Resolution
(Mathematical and Mechanical Engineering SET - Volume 6)
Advanced Numerical Methods with Matlab® 2: Resolution ofNonlinear,
Differential and Partial Differential Equations
(Mathematical and Mechanical Engineering SET - Volume 7)
SALENQON Jean
Virtual Work Approach to Mechanical Modeling
2017
Movement Equations 2: Mathematical and Methodological Supplements
Movement Equations 3: Dynamics and Fundamental Principle
BOUVET Christophe
Mechanics ofAeronautical Solids, Materials and Structures
Mechanics ofAeronautical Composite Materials
BRANCHERIE Delphine, FEISSEL Pierre, BOUVIER Salima,
IBRAHIMBEGOVIC Adnan
From Microstructure Investigations to Multiscale Modeling:
Bridging the Gap
Chebel-Morello Brigitte, NICOD Jean-Marc, VARNIER Christophe
From Prognostics and Health Systems Management to Predictive
Maintenance 2: Knowledge, Traceability and Decision
EL HAMI Abdelkhalak, RADI Bouchaib
Dynamics ofLarge Structures and Inverse Problems
Fluid-Structure Interactions and Uncertainties: Ansys and Fluent Tools
KHARMANDA Ghias, EL HAMI Abdelkhalak
Biomechanics: Optimization, Uncertainties and Reliability
LEDOUX Michel, EL HAMI Abdelkhalak
Compressible Flow Propulsion and Digital Approaches in Fluid Mechanics
Fluid Mechanics: Analytical Methods
MORI Yvon
Mechanical Vibrations: Applications to Equipment
2016
Movement Equations 1: Location, Kinematics and Kinetics
BOYARD Nicolas
Heat Transfer in Polymer Composite Materials
CARDON Alain, ITMI Mhamed
New Autonomous Systems
DAHOO Pierre Richard, POUGNET Philippe, EL HAMI Abdelkhalak
Nanometer-scale Defect Detection Using Polarized Light
de Saxce Gery, Vallee Claude
Galilean Mechanics and Thermodynamics of Continua
DORMIEUX Luc, KONDO Djimedo
Micromechanics ofFracture and Damage
(Micromechanics Set - Volume 1)
Stochastic Dynamics ofStructures
GOURIVEAU Rafael, MEDJAHER Kamal, ZERHOUNI Noureddine
From Prognostics and Health Systems Management to Predictive
Maintenance 1: Monitoring and Prognostics
KHARMANDA Ghias, EL HAMI Abdelkhalak
Reliability in Biomechanics
(Reliability ofMultiphysical Systems Set-Volume 3)
MOLIMARD Jerome
Experimental Mechanics ofSolids and Structures
RADI Bouchaib, EL HAMI Abdelkhalak
Material Forming Processes: Simulation, Drawing, Hydroforming and
Additive Manufacturing
2015
KARLICIC Danilo, MURMU Tony, ADHIKARI Sondipon, MCCARTHY Michael
Non-local Structural Mechanics
SAB Karam, LEBEE Arthur
Homogenization ofHeterogeneous Thin and Thick Plates
2014
ATANACKOVIC M. Teodor, PILIPOVIC Stevan, STANKOVIC Bogoljub,
ZORICA Dusan
Fractional Calculus with Applications in Mechanics: Vibrations and
Diffusion Processes
Fractional Calculus with Applications in Mechanics: Wave Propagation,
Impact and Variational Principles
CIBLAC Thierry, MOREL Jean-Claude
Sustainable Masonry: Stability and Behavior of Structures
ILANKO Sinniah, MONTERRUBIO Luis E., MOCHIDA Yusuke
The Rayleigh-Ritz Methodfor Structural Analysis
LALANNE Christian
Mechanical Vibration and Shock Analysis - 5-volume series - 3rd edition
Sinusoidal Vibration - Volume 1
Mechanical Shock - Volume 2
Random Vibration - Volume 3
Fatigue Damage - Volume 4
Specification Development - Volume 5
LEMAIRE Maurice
Uncertainty and Mechanics
2013
ADHIKARI Sondipon
Structural Dynamic Analysis with Generalized Damping Models: Analysis
ADHIKARI Sondipon
Structural Dynamic Analysis with Generalized Damping Models:
Identification
BAILLY Patrice
Materials and Structures under Shock and Impact
BASTIEN Jerome, BERNARDIN Frederic, LAMARQUE Claude-Henri
Non-smooth Deterministic or Stochastic Discrete Dynamical Systems:
Applications to Models with Friction or Impact
Uncertainty and Optimization in Structural Mechanics
KIRILLOV Oleg N., PELINOVSKY Dmitry E.
Nonlinear Physical Systems: Spectral Analysis, Stability and Bifurcations
LUONGO Angelo, ZULLI Daniele
Mathematical Models ofBeams and Cables
SALENQON Jean
Yield Design
2012
DAVIM J. Paulo
Mechanical Engineering Education
DUPEUX Michel, BRACCINI Muriel
Mechanics of Solid Interfaces
ELISHAKOFF Isaac et al.
Carbon Nanotubes and Nanosensors: Vibration, Buckling
and Ballistic Impact
GREDIAC Michel, HILD Francois
Full-Field Measurements and Identification in Solid Mechanics
GROUS Ammar
Fracture Mechanics - 3-volume series
Analysis ofReliability and Quality Control - Volume 1
Applied Reliability - Volume 2
Applied Quality Control - Volume 3
RECHO Naman
Fracture Mechanics and Crack Growth
2011
KRYSINSKI Tomasz, MALBURET Francois
Mechanical Instability
SOUSTELLE Michel
An Introduction to Chemical Kinetics
2010
BREITKOPF Piotr, FILOMENO COELHO Rajan
Multidisciplinary Design Optimization in Computational Mechanics
DAVIM J. Paulo
Biotribolgy
PAULTRE Patrick
Dynamics ofStructures
SOUSTELLE Michel
Handbook ofHeterogenous Kinetics
2009
BERLIOZ Alain, TROMPETTE Philippe
Solid Mechanics using the Finite Element Method
LEMAIRE Maurice
Structural Reliability
2007
GIRARD Alain, ROY Nicolas
Structural Dynamics in Industry
Guinebretiere Rene
X-ray Diffraction by Polycrystalline Materials
KRYSINSKI Tomasz, MALBURET Francois
Mechanical Vibrations
KUNDU Tribikram
Advanced Ultrasonic Methods for Material and Structure Inspection
SIH George C. et al.
Particle and Continuum Aspects ofMesomechanics

Heat Transfer 1. Conduction 2021

Uploaded by

Copyright:

Available Formats

Heat Transfer 1. Conduction 2021

Uploaded by

Document Information

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 1. Conduction 2021

Uploaded by

Copyright:

Available Formats

What is the book about?

What is the book about?

What information is provided on the book covers/title pages?

What information is provided on the book covers/title pages?

MECHANICAL ENGINEERING AND SOLID MECHANICS SERIES

MATHEMATICAL AND MECHANICAL ENGINEERING SET

ISTE Ltd John Wiley & Sons, Inc.

© ISTE Ltd 2021

Library of Congress Control Number: 2020949611

British Library Cataloguing-in-Publication Data

Introduction .......................................................................................... xiii

Chapter 1. The Problem of Thermal Conduction: General

Chapter 2. The Physics of Conduction ........................................... 9

2.3.3. Case of the axisymmetric system ............................................... 24

Chapter 3. Conduction in a Stationary Regime ............................... 53

Chapter 4. Quasi-stationary Model .................................................. 103

Chapter 5. Non-stationary Conduction ........................................... 183

Chapter 6. Fin Theory: Notions and Examples .............................. 237

Appendices .......................................................................................... 263

Appendix 1. Heat Equation of a Three-dimensional System ....... 265

Appendix 2. Heat Equation: Writing in the Main

Appendix 3. One-dimensional Heat Equation .................................. 283

Appendix 4. Conduction of the Heat in a Non-stationary Regime:

Appendix 6. Complementary Information Regarding Fins ............ 297

Appendix 7. The LaplaceTransform ................................................. 301

Appendix 8. Reminders RegardingHyperbolic Functions ............... 309

References ............................................................................................. 313

Index ...................................................................................................... 315

Thermal science is to thermodynamics as decree is to law. It answers the

But by what means do I take or do I give? Even calculations of

Thermal science specifies how, but “evacuates” the work. If in a given

Three heat transfer modes can be identified: conduction and radiation -

This book is Volume 1 of a collection of problems on heat transfer,

The book is intended to reach a wide audience, from technicians to

Volume 1, dedicated to “classic” approaches (analytical treatment) to

After the generalities outlined in Chapter 1, Chapter 2 presents the

Chapter 3 deals with conduction in a stationary regime. Emphasis is

Chapter 4 presents the notion of a quasi-stationary regime; this method,

homogeneous - instantaneous balances can be obtained, in addition to the

Chapter 5 deals with variable conduction regimes. The most classic

Many examples are provided to illustrate this important aspect of

Chapter 6 presents the notion of fin theory, which is associated with a

Thermal energy was probably first perceived (if not identified) by

However, it took a few hundred thousand years before so-called “modern”

Advances in radiation science, particularly in sensor technology, have

Closer to home, fundamental thermal science, whether it is conduction,

There are numerous heat transfer textbooks in different formats:

So why do we need another book?

The authors have taught at university level and in prestigious French

When a miner wants to detach a block of coal or precious mineral from a

Whether they are researchers, engineers or simply teachers, scientists

A good physicist must have a perfect understanding of the idea of an

The two authors belong to the world of engineering sciences, meaning

Fundamental or “basic physics” concepts are taken from a type of manual

Another significant difference is that this book is directed at a large

So, what is this book for?

Above all, it contains problems to be worked on, of which most are

It seems necessary to surround these problems with strong reminders of

Lastly, upon rereading, the authors also recommend this book as an

Thermal science is to thermodynamics as decree means is to law. It