Everyday Heat Transfer

EVERYDAY HEAT
TRANSFER
PROBLEMS
Sensitivities To
Governing Variables
by M. Kemal Atesmen
© 2009 by ASME, Three Park Avenue, New York, NY 10016, USA (www.asme.org)
All rights reserved. Printed in the United States of America. Except as permitted
under the United States Copyright Act of 1976, no part of this publication may be
reproduced or distributed in any form or by any means, or stored in a database or
retrieval system, without the prior written permission of the publisher.
INFORMATION CONTAINED IN THIS WORK HAS BEEN OBTAINED BY THE

AMERICAN SOCIETY OF MECHANICAL ENGINEERS FROM SOURCES BELIEVED
TO BE RELIABLE. HOWEVER, NEITHER ASME NOR ITS AUTHORS OR EDITORS
GUARANTEE THE ACCURACY OR COMPLETENESS OF ANY INFORMATION
PUBLISHED IN THIS WORK. NEITHER ASME NOR ITS AUTHORS AND EDITORS
SHALL BE RESPONSIBLE FOR ANY ERRORS, OMISSIONS, OR DAMAGES ARISING
OUT OF THE USE OF THIS INFORMATION. THE WORK IS PUBLISHED WITH
THE UNDERSTANDING THAT ASME AND ITS AUTHORS AND EDITORS ARE
SUPPLYING INFORMATION BUT ARE NOT ATTEMPTING TO RENDER
ENGINEERING OR OTHER PROFESSIONAL SERVICES. IF SUCH ENGINEERING
OR PROFESSIONAL SERVICES ARE REQUIRED, THE ASSISTANCE OF AN
APPROPRIATE PROFESSIONAL SHOULD BE SOUGHT.
ASME shall not be responsible for statements or opinions advanced in papers

or ... printed in its publications (B7.1.3). Statement from the Bylaws.
For authorization to photocopy material for internal or personal use under those
circumstances not falling within the fair use provisions of the Copyright Act, contact
the Copyright Clearance Center (CCC), 222 Rosewood Drive, Danvers, MA 01923,
tel: 978-750-8400, www.copyright.com.
Library of Congress Cataloging-in-Publication Data
Atesmen, M. Kemal.
Everyday heat transfer problems : sensitivities to governing variables / by M. Kemal
Atesmen.
p. cm.
Includes bibliographical references and index.
ISBN 978-0-7918-0283-0
1. Heat–Transmission–Problems, exercises, etc. 2. Materials–Thermal properties–
Problems, exercises, etc. 3. Thermal conductivity–Problems, exercises, etc.
4. Engineering mathematics–Problems, exercises, etc. I. Title.
TA418.54.A47 2009
621.402’2–dc22 2008047423
TABLE OF CONTENTS
Introduction ................................................................................................... 1
Chapter 1 Heat Loss from Walls in a Typical House ............................. 5
Chapter 2 Conduction Heat Transfer in a Printed Circuit Board .... 13
Chapter 3 Heat Transfer from Combustion Chamber Walls.............. 25
Chapter 4 Heat Transfer from a Human Body During

Solar Tanning ............................................................................ 33
Chapter 5 Efficiency of Rectangular Fins .............................................. 41
Chapter 6 Heat Transfer from a Hot Drawn Bar .................................. 51
Chapter 7 Maximum Current in an Open-Air Electrical Wire .......... 65
Chapter 8 Evaporation of Liquid Nitrogen in a

Cryogenic Bottle ...................................................................... 77
Chapter 9 Thermal Stress in a Pipe ........................................................ 85
Chapter 10 Heat Transfer in a Pipe with Uniform Heat

Generation in its Walls ......................................................... 93
Chapter 11 Heat Transfer in an Active Infrared Sensor .................. 103
Chapter 12 Cooling of a Chip ................................................................. 113
iii
Everyday Heat Transfer Problems
Chapter 13 Cooling of a Chip Utilizing a Heat Sink

with Rectangular Fins ......................................................... 121
Chapter 14 Heat Transfer Analysis for Cooking in a Pot ................ 131
Chapter 15 Insulating a Water Pipe from Freezing ........................... 139
Chapter 16 Quenching of Steel Balls in Air Flow .............................. 147
Chapter 17 Quenching of Steel Balls in Oil......................................... 155
Chapter 18 Cooking Time for Turkey in an Oven .............................. 161
Chapter 19 Heat Generated in Pipe Flows due to Friction ............. 169
Chapter 20 Sizing an Active Solar Collector for a Pool ................... 179
Chapter 21 Heat Transfer in a Heat Exchanger ................................. 195
Chapter 22 Ice Formation on a Lake .................................................... 203
Chapter 23 Solidification in a Casting Mold ....................................... 213
Chapter 24 Average Temperature Rise in Sliding Surfaces

in Contact .............................................................................. 221
References .................................................................................................... 233
Index .............................................................................................................. 235
iv
I NTRODUCTION
Everyday engineering problems in heat transfer can be very

complicated and may require solutions using finite element or finite
difference techniques in transient mode and in multiple dimensions.
These engineering problems might cover conduction, convection and
radiation energy transfer mechanisms. The thermophysical properties
that govern a particular heat transfer problem can be challenging to
discover, to say the least.
Some of the standard thermophysical properties needed to
solve a heat transfer problem are density, specific heat at constant
pressure, thermal conductivity, viscosity, volumetric thermal
expansion coefficient, heat of vaporization, surface tension, emissivity,
absorptivity, and transmissivity. These thermophysical properties
can be strong functions of temperature, pressure, surface roughness,
wavelength and other properties. in the region of interest.
Once a heat transfer problem's assumptions are made, equations
set up and boundary conditions determined, one should investigate
the sensitivities of desired outputs to all the governing independent
variables. Since these sensitivities are mostly non-linear, one should
1
analyze them in the region of interest. The results of such sensitivity

analyses will provide important information as to which independent
variables should be researched thoroughly, determined accurately,
and focused on. The sensitivity analysis will also provide insight into
uncertainty analysis for the dependent variable, (Reference S. J. Kline
and F. A. McClintock [9]). If the dependent variable y is defined as a
function of independent variables x1, x2, x3, … xn as follows:
y = f(x1, x2, x3, … xn)
then the uncertainty U for the dependent variable can be

written as:
U = [(∂y/∂x1 u1)2 + (∂y/∂x2 u2)2 + (∂y/∂x3 u3)2 + … + (∂y/∂xn un)2]0.5
where ∂y/∂x1, ∂y/∂x2, ∂y/∂x3, …, ∂y/∂xn are the sensitivities of the

dependent variable to each independent variable and u1, u2, u3, …,
un are the uncertainties in each independent variable for a desired
confidence limit.
In this book, I will provide sensitivity analyses to well-known
everyday heat transfer problems, determining ∂y/∂x1, ∂y/∂x2,
∂y/∂x3, …, ∂y/∂xn for each case. The analysis for each problem will
narrow the field of independent variables that should be focused
on during the design process. Since most heat transfer problems
are non-linear, the results presented here would be applicable only
in the region of values assumed for independent variables. For the
uncertainties of independent variables—for example, experimental
measurements of thermophysical properties—the reader can find the
appropriate uncertainty value for a desired confidence limit within
existing literature on the topic.
Each chapter will analyze a different one-dimensional heat transfer
problem. These problems will vary from determining the maximum
allowable current in an open-air electrical wire to cooking a turkey
in a convection oven. The equations and boundary conditions for
each problem will be provided, but the focus will be on the sensitivity
of the governing dependant variable on the changing independent
2
Introduction
variables. For the derivation of the fundamental heat transfer

equations and for insight into the appropriate boundary conditions,
the reader should refer to the heat transfer fundamentals books listed
in the references.
Problems in Chapters 1 through 6 deal with steady-state and
one-dimensional heat transfer mechanisms in rectangular coordinates.
Chapters 7 through 10 deal with steady-state and one-dimensional
heat transfer mechanisms in cylindrical coordinates. Unsteady-state
problems in one-dimensional rectangular coordinates will be tackled
in Chapters 11 through 14, cylindrical coordinates in Chapter 15, and
spherical coordinates in Chapters 16 through 18.
The following six chapters are allocated to special heat transfer
problems. Chapters 19 and 20 deal with momentum, mass and heat
transfer analogies used to solve the problems. Chapter 21 analyzes
a counterflow heat exchanger using the log mean temperature
difference method. Chapters 22 and 23 solve heat transfer problems
of ice formation and solidification with moving boundary conditions.
Chapter 24 analyzes the problem of frictional heating of materials in
contact with moving sources of heat.
I would like to thank my engineering colleagues G. W. Hodge,
A. Z. Basbuyuk, E. O. Atesmen, and S. S. Tukel for reviewing some of
the chapters. I would also like to dedicate this book to my excellent
teachers and mentors in heat transfer at several universities and
organizations. Some of the names at the top of a long list are
Prof. W. M. Kays, Prof. A. L. London, Prof. R. D. Haberstroh,
Prof. L. V. Baldwin, and Prof. T. N. Veziroglu.
M. Kemal Atesmen
Ph. D. Mechanical Engineering
Santa Barbara, California
3
CHAPTER
HEAT LOSS
FROM WALLS
1
IN A TYPICAL HOUSE
H eat loss from the vertical walls of a house is analyzed under

steady-state conditions. Walls are assumed to be large and
built in a planar fashion, so that one-dimensional heat transfer
rate equations in rectangular coordinates may be used, and only
conduction and convection heat transfer mechanisms are considered.
In this analysis, radiation heat transfer effects are neglected. No air
leakage through the wall was assumed. Also, the wall material thermal
conductivities are assumed to be independent of temperature in the
region of operation.
Assuming winter conditions—the temperature inside the house is
higher than the temperature outside the house—the convection heat
transferred from the inside of the house to the inner surface of the
inner wall is:
Q/A = hin (Tin – Tinner wall inside surface) (1-1)
Most walls are constructed from three types of materials: inner wall
board, insulation and outer wall board. The heat transfer from these
wall layers will occur by conduction, and is presented by the following
rate Eqs., (1-2) through (1-4):
5
Q/A = (kinner wall/tinner wall) (Tinner wall inside surface – Tinner wall outside surface) (1-2)
Q/A = (kinsulation/tinsulation) (Tinner wall outside surface – Touter wall inside surface) (1-3)
Q/A = (kouter wall/touter wall) (Touter wall inside surface – Touter wall outside surface) (1-4)
The heat transfer from the outer surface of the outer wall to the
atmosphere is by convection and can be expressed by the following
rate Eq. (1-5):
Q/A = hout (Tout – Touter wall outer surface) (1-5)
Eliminating all the wall temperatures from Eqs. (1-1) through (1-5),
the heat loss from a house wall can be rewritten as:
Q/A = (Tin – Tout)/[(1/hin) + (tinner wall/kinner wall) + (tinsulation/kinsulation)

+ (touter wall/kouter wall) + (1/hout)] (1-6)
The denominator in Eq. (1-6) represents all the thermal resistances

between the inside of the house and the atmosphere, and they are
in series.
In the construction industry, wall materials are rated with their
R-value, namely the thermal conduction resistance of one-inch
material. R-value dimensions are given as (hr-ft2-F/BTU)(1/in). The
sensitivity analysis will be done in the English system of units rather
than the International System (SI units). The governing Eq. (1-6)
for heat loss from a house wall can be rewritten in terms of R-values
as follows:
Q/A = (Tin – Tout)/[(1/hin) + Rinner wall tinner wall + Rinsulation tinsulation

+ Router wall touter wall + (1/hout)] (1-7)
where the definitions of the variables with their assumed nominal

values for the present sensitivity analysis are given as:
Q/A = heat loss through the wall due to convection and conduction
in Btu/hr-ft2
6
Heat Loss From Walls In A Typical House
Tin = 68ºF (inside temperature)

Tout = 32ºF (outside temperature)
hin = 5 BTU/hr-ft2-F (inside convection heat transfer coefficient)
Rinner wall = 0.85 hr-ft2-F/BTU-in (wall inside board R-value)
tinner wall = 1 in (wall inside board thickness)
Rinsulation = 3.5 hr-ft2-F/BTU-in (insulation layer R-value)
tinsulation = 4 in (insulation layer thickness)
Router wall = 5 hr-ft2-F/BTU-in (wall outside board R-value)
touter wall = 1 in (wall outside board thickness)
hout = 10 BTU/hr-ft2-F (outside convection heat transfer coefficient).
The heat loss through a wall due to changes in convection heat

transfer is presented in Figures 1-1 and 1-2. Changes in the convection
heat transfer coefficient affect the heat loss mainly in the natural
convection regime. As the convection heat transfer coefficient increases
into the forced convection regime, heat loss value asymptotes.
Resistances from both inside and outside convection heat transfer are
too small to cause any change in heat loss through the wall.
The heat loss through a wall due to changes in insulation material
R-value is presented in Figures 1-3 and 1-4. Higher R-value insulation
1.8
1.79
Q/A, Btu/hr-ft2
1.78
1.77
1.76
1.75
5 10 15 20 25
Outside Convection Heat Transfer Coefficient, Btu/hr-ft2-F
Figure 1-1 Wall heat loss versus outside convection heat transfer coefficient
7
1.8
1.78
Q/A, Btu/hr-ft2
1.76
1.74
1.72
1 3 5 7 9
Inside Convection Heat Transfer Coefficient, Btu/hr-ft2-F
Figure 1-2 Wall heat loss versus inside convection heat transfer coefficient
material is definitely the way to go, depending upon the cost and
benefit analysis results. The thickness of the insulation material is also
very crucial. Thicker insulation material is definitely the best choice,
depending upon the cost and benefit analysis results.
1.8
Q/A, Btu/hr-ft2
1.6
1.4
1.2
3 3.5 4 4.5 5
Insulation "R" Value, hr-ft2-F/Btu-in
Figure 1-3 Wall heat loss versus insulation R-value
8
6
Q/A, Btu/hr-ft2
0
0 2 4 6 8 10 12
Insulation Thickness, in
Figure 1-4 Wall heat loss versus insulation thickness
The effects on heat loss of inner and outer wall board R-values
and thicknesses are similar to the effects of insulation R-value and
thickness, but to a lesser extent. Sensitivities of heat loss to all the
governing variables around the nominal values given above will be
analyzed later.
Sensitivity of heat loss to the outside convection heat transfer
coefficient can be determined in a closed form by differentiating the
heat loss Eq. (1-7) with respect to hout:
∂(Q/A)/∂hout = (Tin – Tout)/{h2out [(1/hin) + Rinner wall tinner wall

+ Rinsulation tinsulation + Router wall touter wall + (1/hout)]2} (1-8)
Sensitivity of heat loss to the outside convection heat transfer

coefficient is given in Figure 1-5. Similar sensitivity is experienced
for the inside convection heat transfer coefficient. The sensitivity
of heat loss to the convection heat transfer coefficient is high in
the natural convection regime, and it diminishes in the forced
convection regime.
Sensitivities of heat loss to insulation material R-value and
insulation thickness are given in Figures 1-6 and 1-7 respectively.
9
0.09
∂ (Q/A) / ∂hout, F
0.06
0.03
0
0 5 10 15 20
Outside Convection Heat Transfer Coefficient,
Btu/hr-ft2-F
Figure 1-5 Sensitivity of house wall heat loss per unit area to the outside
convection heat transfer coefficient
These two sensitivities are similar, as can be expected, since the

linear product of insulation material R-value and insulation thickness
affects the heat loss, as shown in the governing heat loss Eq. (1-7).
Absolute sensitivity values are high at the low values of insulation
0
∂(Q/A) / ∂ Rinsulation,
−0.5
(BTU/hr-ft2)2(in/F)
−1
−1.5
−2
−2.5
0 1 2 3 4 5 6
Insulation Material R-Value, hr-ft2-F/BTU-in
Figure 1-6 Sensitivity of house wall heat loss per unit area to insulation
material R-value
10
∂ (Q/A) / ∂t insulation, BTU/hr-ft2-in

0
−0.5
−1
−1.5
0 2 4 8 6 10 12
Insulation Thickness, in
Figure 1-7 Sensitivity of house wall heat loss per unit area to insulation
thickness
material R-value and insulation thickness. Sensitivities approach

zero as insulation material R-value and insulation thickness values
increase.
A ten-percent variation in independent variables around the
nominal values given above produces the sensitivity results given in
Table 1-1 The sensitivity results are given in a descending order and
they are applicable only in the region of assigned nominal values,
due to their non-linear effect to heat loss. The one exception is
temperature potential, (Tin − Tout), which will always be ±10% due
to its linear effect on heat loss. Material R-value and its thickness
change have the same sensitivity, since their linear product affects the
governing heat loss equation.
Heat loss through the wall is most sensitive to the temperature
potential between the inside and outside of the house. Changes in
wall insulation R-value and thickness affect heat loss as much as
the temperature potential. Continuing in order of sensitivity, wall
outer board R-value and thickness changes affect heat loss the most,
followed by wall inside board R-value and thickness. Wall heat loss is
11
Table 1-1 House wall heat loss change per unit area due to a
10% change in variables nominal values
House Wall House Wall

Heat Loss Heat Loss
Change Due To Change Due To
A 10% A 10%
Nominal Decrease In Increase In
Variable Value Nominal Value Nominal Value
Tin − Tout 36°F −10% +10%
Rinsulation 3.5 hr-ft -F/BTU-in
2
+7.467% −6.497%
tinsulation 4 in +7.467% −6.497%
Router wall 5 hr-ft2-F/BTU-in +2.545% −2.545%
touter wall 1 in +2.545% −2.545%
Rinner wall 0.85 hr-ft -F/BTU-in
2
+0.424% −0.424%
tinner wall 1 in +0.424% −0.424%
hin 5 BTU/hr-ft -F
2
−0.110% +0.090%
hout 10 BTU/hr-ft -F
2
−0.055% +0.045%
least sensitive to both the inside and outside heat transfer coefficient
changes. Wall heat loss sensitivity to both the inside and outside
heat transfer coefficient changes is an order of magnitude less than
sensitivity to temperature potential changes.
12
CHAPTER
CONDUCTION
HEAT TRANSFER
2
IN A PRINTED
CIRCUIT BOARD
C onduction heat transfer in printed circuit boards (PCBs) has been

studied extensively in literature i.e., B. Guenin [4]. The layered
structure of a printed circuit board is treated using two different
thermal conductivities; one is in-plane thermal conductivity and the
other is through-thickness thermal conductivity. One-dimensional
conduction heat transfers in in-plane direction and through-thickness
direction are treated independently. Since the significant portion of
the conduction heat transfer in a PCB occurs in the in-plane direction
in the conductor layers, this is a valid assumption. Under steady-state
conditions and with constant thermophysical properties, the in-plane
(i-p) conduction heat transfer equation for a PCB can be written as:
Qin-plane = Q1i-p + Q2i-p + … + Qni-p (2-1)
where the subscript refers to the layers of the PCB. Using the
conduction rate equation in rectangular coordinates for a PCB with
a width of W, a length of L, layer thicknesses ti and layer thermal
conductivities ki, Eq. (2-1) can be rewritten as:
13
W Σti kin-plane (TL=0 – TL=L)/L = W t1 k1 (TL=0 – TL=L)/L

+ W t2 k2 (TL=0 – TL=L)/L
+ W tn kn (TL=0 – TL=L)/L (2-2)
In-plane conduction heat transfer in a PCB represents a parallel

thermal resistance circuit which can be written as:
(1/Rin-plane) = (1/R1) + (1/R2) + … + (1/Rn) where Ri = L/(kitiW) (2-3)
where
kin-plane = Σ(kiti)/Σti (2-4)
Through-thickness (t-t) conduction heat transfer in a PCB represents

a series thermal resistance circuit, and the through-thickness conduction
heat transfer equation for a PCB can be written as:
Qthrough-thickness = Q1t-t = Q2t-t = … = Qnt-t (2-5)
which can be expanded into following equations:
W L kthrough-thickness (Tt=0 – Tt=Σti)/Σti = W L k1 (Tt=0 − Tt=t1)/t1

= W L k2 (Tt=t1 − Tt=t2)/t2 = … = W L kn (Tt=tn-1 − Tt=tn)/tn (2-6)
Inter-layer temperatures can be eliminated from Eqs. (2-6), and a

series thermal resistance equation extracted as follows:
Rthrough-thickness = R1 + R2 + … + Rn where Ri = ti/ki (2-7)
where
kthrough-thickness = Σti/Σ(ti/ki). (2-8)
A printed circuit board is commonly built as layers of conductors

separated by layers of insulators. The conductors are mostly alloys
of copper, silver or gold, while the insulators are mostly a variety of
14
Conduction Heat Transfer In A Printed Circuit Board
epoxy resins. Therefore the in-plane thermal conductivity Eq. (2-4)

for a PCB can be rewritten as:
kin-plane = [kconductor Σtconductor + kinsulator (ttotal − Σtconductor)]/ttotal (2-9)
Similarly, the through-thickness thermal conductivity Eq. (2-8) for

a PCB can be rewritten as:
kthrough-thickness
= ttotal/[(Σtconductor/kconductor) + (ttotal − Σtconductor)/kinsulator] (2-10)
The sensitivities of these two PCB thermal conductivities are

analyzed for a 500 µm-thick printed circuit board, with the assumed
nominal values for thermal conductivity of the conductor and
insulator layers given below:
kconductor = 377 W/m-C for copper conductor layers and
kinsulator = 0.3 W/m-C for glass reinforced polymer layers.
In-plane thermal conductivity versus percent of conductor layers

to total printed circuit board thickness is given in Figure 2-1. In-plane
thermal conductivity starts at the all-insulator thermal conductivity
value of 0.3 W/m-C, and increases linearly to conductor thermal
conductivity at no insulator layers.
Sensitivities of in-plane thermal conductivity to changes in kconductor
and kinsulator are represented in Figure 2-2. As you can see, the two
sensitivities are opposite.
The sensitivity of in-plane thermal conductivity to conductor
thickness is a constant, 0.75 W/m-C-µm. The sensitivity of in-plane
thermal conductivity to insulator thickness is also a constant,
and it is the opposite of sensitivity to conductor thickness, namely
–0.75 W/m-C-µm.
A ten percent variation in variables around the nominal values
given above produce the sensitivity results in Table 2-1 for in-plane
thermal conductivity. For these nominal values, in-plane thermal
15
400
300
kin-plane, W/m-C
200
100
0
0 20 40 60 80 100
% Conductor Layers Thickness
Figure 2-1 In-plane thermal conductivity versus percent of conductor layers

thickness to total printed circuit board thickness
1
∂kin-plane / ∂kinsulator
∂ kin-plane / ∂kconductor &
∂ kin-plane / ∂ kinsulator
0.8
0.6
0.4
0.2
∂kin-plane / ∂kinsulator
0
0 20 40 60 80 100
Figure 2-2 Sensitivity of in-plane thermal conductivity to kconductor and

to kinsulator
16
Table 2-1 In-plane thermal conductivity change due to a 10%

change in variables around nominal values for a
500 micron thick PCB
In-Plane Thermal In-Plane Thermal

Conductivity Change Conductivity Change
Due To A 10% Due To A 10%
kconductor 377 W/m-C −9.99% +9.99%
Σtconductor 250 µm −9.98% +9.98%
Σtinsulator 250 µm +9.98% −9.98%
kinsulator 0.3 W/m-C −0.01% +0.01%
conductivity is 188.65 W/m-C. Insulator thermal conductivity is

the least effective independent variable in this case due to its low
value. Variations in the sum of conductor thicknesses and the sum of
insulator thicknesses affect in-plane thermal conductivity in opposing
directions, but with the same magnitude.
Through-thickness thermal conductivity versus percent of
conductor thickness has a non-linear behavior, and it is given
for a 500-micron PCB in Figure 2-3. Through-thickness thermal
conductivity is similar to insulator layer thermal conductivity for up
to 80% conductor layer thickness of the total printed circuit board,
and therefore is not a good conduction heat transfer path for printed
circuit boards.
Sensitivities of through-thickness thermal conductivity to
kconductor and to kinsulator are given in Figure 2-4. The sensitivity of
through-thickness thermal conductivity to changes in conductor
thermal conductivity is negligible throughout the percent conductor
layer thickness. The sensitivity of through-thickness thermal
conductivity to changes in insulator thermal conductivity increases
and becomes significant as the thickness percentage of the insulator
layers decreases.
17
20
kthrough-thickness, W/m-C
15
10
0
0 20 40 60 80 100
Figure 2-3 Through-thickness thermal conductivity versus percent

conductor layer thickness
∂kthrough-thickness / ∂ kconductor &
∂ kthrough-thickness / ∂ kinsulator
20
∂ kthrough-thickness /
∂kinsulator
16
12
0 ∂ kthrough-thickness /
0 50 100 ∂ kconductor
Figure 2-4 Sensitivity of through-thickness thermal conductivity to

conductor and insulator thermal conductivities versus percent
conductor layer thickness
18
∂kthrough-thickness / ∂ kconductor &
0.05
∂ kthrough-thickness / ∂kinsulator,
0 ∂ kthrough-thickness /
∂ kconductor
W/m-C-um
−0.05
−0.1
−0.15
−0.2 ∂ kinsulator
0 20 40 60 80 100

conductor and insulator thickness
Sensitivities of through-thickness thermal conductivity to

conductor and insulator thickness are given in Figure 2-5. The
sensitivities become significant as conductor thickness
approaches 100%.
A ten-percent variation in independent variables around the
nominal values given above produce the sensitivity results given in
Table 2-2 for through-thickness thermal conductivity, which has a
nominal value of 0.6 W/m-C. In this case, through-thickness thermal
conductivity is very resistant to conductor thermal conductivity
variations in the region of interest.
A second and a similar analysis can be performed for a plated or
sputtered thinner circuit. A 10-µm thick circuit is considered with the
assumed nominal thermal conductivities below:
kconductor = 377 W/m-C for copper conductor layers and
kinsulator = 36 W/m-C for aluminum oxide insulating layers.
In-plane thermal conductivity versus percent of conductor layers

to total thickness is given in Figure 2-6. In-plane thermal conductivity
19
Table 2-2 Through-thickness thermal conductivity change

due to a 10% change in variables around nominal
values for a 500 micron thick PCB
Through-Thickness Through-Thickness
Thermal Conductivity Thermal Conductivity
Nominal A 10% Decrease In A 10% Increase In
Σtinsulator 250 µm +11.09% −9.08%
kinsulator 0.3 W/m-C −9.99% +9.99%
Σtconductor 250 µm −9.08% +11.09%
kconductor 377 W/m-C −0.01% +0.01%
400
300
kin-plane, W/m-C
200
100
0
0 20 40 60 80 100
Figure 2-6 In-plane thermal conductivity versus percent of conductor layers

to total thickness
starts at the all insulator thermal conductivity value of 36 W/m-C and

increases linearly to a conductor thermal conductivity of 377 W/m-C.
Sensitivites of in-plane thermal conductivity to kconductor and kinsulator are
given in Figure 2-7. As you can see, the two sensitivities are opposite.
20
1
∂ kin-plane / ∂ kconductor
∂kin-plane / ∂kconductor &
0.8
0.6
0.4
0.2
0
0 20 40 60 80 100
Figure 2-7 Sensitivity of in-plane thermal conductivity to kconductor and

to kinsulator
The sensitivity of in-plane thermal conductivity to conductor thickness

is a constant at 34.1 W/m-C-µm. The sensitivity of in-plane thermal
conductivity to insulator thickness is a constant, and it is the opposite of
sensitivity to in-plane thermal conductivity, namely –34.1 W/m-C-µm.
A ten percent variation in variables around the nominal values given
above produce the sensitivity results in Table 2-3 for in-plane thermal
conductivity, which has a nominal value of 87.15 W/m-C. Insulator
thickness is the dominant independent variable in this region of interest.
Through-thickness thermal conductivity versus percent of
conductor thickness has a non-linear behavior, and it is given in
Figure 2-8. The percentage of the conductor layers thickness to total
circuit thickness affects the through-thickness thermal conductivity
at all conductor layer thicknesses. Through-thickness conduction heat
transfer is much more prominent in thin-plated or sputtered circuits.
Sensitivities of through-thickness thermal conductivity to kconductor
and to kinsulator are given in Figure 2-9. The sensitivity of through-
thickness thermal conductivity to changes in conductor thermal
conductivity is negligible at low percentages of conductor layer
thickness. On the other hand, the sensitivity of through-thickness
21
Table 2-3 In-plane thermal conductivity change due to a 10%

change in variables around nominal values for a
10-micron thick circuit
In-Plane Thermal In-Plane Thermal

Conductivity Conductivity
Nominal A 10% Decrease A 10% Increase
Variable Value In Nominal Value In Nominal Value
Σtinsulator 8.5 µm +33.3% −33.3%
kconductor 377 W/m-C −6.5% +6.5%
Σtconductor 1.5 µm −5.9% +5.9%
kinsulator 36 W/m-C −3.5% +3.5
thermal conductivity to changes in insulator thermal conductivity

is significant; it increases to a maximum at around 90% conductor
layer thickness, and finally decreases sharply as conductor thermal
conductivity starts to dominate.
400
kthrough-thickness, W/m-C
300
200
100
0
0 20 40 60 80 100
Figure 2-8 Through-thickness conduction heat transfer coefficient versus

percentage of conductor layer thickness
22
∂ kthrough-thickness / ∂ kconductor &

∂ kthrough-thickness / ∂kinsulator
3.5
3
∂kthrough-thickness /
2.5
∂ kinsulator
2
1.5
1
0.5
0
0 20 40 60 80 100 ∂ kconductor

conductor and insulator thermal conductivities versus
percentage of conductor layer thickness
∂ kthrough-thickness / ∂kinsulator &
∂kthrough-thickness / ∂kconductor,
10
0 ∂kconductor
−10
W/m-C-µm
−20
−30
−40
−50
0 20 60 ∂kinsulator
40 80 100

conductor thickness and to insulator thickness
23
Table 2-4 Through-thickness thermal conductivity change

due to a 10% change in variables around nominal
values for a 10-micron thick circuit
Through-Thickness Through-Thickness
Thermal Conductivity Thermal Conductivity
Change Due To A 10% Change Due To A 10%
kinsulator 36 W/m-C −9.85% +9.18%
Σtinsulator 8.5 µm +9.76% −8.17%
Σtconductor 1.5 µm −1.55% +1.59%
kconductor 377 W/m-C −0.18% +0.15%
Sensitivities of the through-thickness heat transfer coefficient

to conductor and insulator thickness are given in Figure 2-10. The
sensitivity to insulator thickness becomes significant as the percent of
conductor thickness approaches 100%.
A ten percent variation in independent variables around the
nominal values given above produce the sensitivity results in
Table 2-4 for through-thickness thermal conductivity, which has a
nominal value of 41.65 W/m-C. In this case, through-thickness thermal
conductivity is most sensitive to insulator thermal conductivity and
insulator thickness variations.
24
CHAPTER
HEAT TRANSFER 3
FROM COMBUSTION
CHAMBER WALLS
C ooling the walls of a combustion chamber containing gases at

high temperatures, i.e., 1000°C, results in parallel modes of heat
transfer, convection and radiation. This problem can be approached
by assuming one-dimensional steady-state heat transfer in rectangular
coordinates and with constant thermophysical properties.
Convection heat transfer per unit area, from hot gases to the hot
side of a wall that separates the cold medium and the hot gases, can
be written as:
(Q/A)convection = hcg (Tg – Twh) (3-1)
Radiation heat transfer per unit area from hot gases which are
assumed to behave as gray bodies to the hot side of a wall can be
written as:
(Q/A)radiation = hrg (Tg – Twh) = = εg σ(Tg4 – Twh4) (3-2)
And so the total heat transfer from the hot gases to a combustion
chamber wall is:
25
(Q/A)total = (Q/A)convection + (Q/A)radiation = (hcg + hrg) (Tg – Twh) (3-3)
These two heat transfer mechanisms act as parallel thermal

resistances, namely:
(1/Rtotal) = (1/Rconvection) + (1/Rradiation) (3-4)
where hcg is the convection heat transfer coefficient between gas

and the hot side of a wall, hrg is the radiation heat transfer coefficient
between gas and the hot side of a wall, Tg is average gas temperature
and Twh is average hot side wall temperature. The radiation heat
transfer coefficient hrg is defined as:
hrg = εg σ(Tg4 – Twh4)/(Tg – Twh) (3-5)
where εg is emissivity of gas and σ is the Stefan Boltzmann constant.

Heat transfer occurs through the wall by conduction and is
defined as:
(Q/A)total = (kwall/L)(Twh – Twc) (3-6)
where kwall is wall material thermal conductivity, L is thickness of the

wall and Twc is the wall temperature at the cold medium side of the wall.
Heat transfer between the cold medium side of the wall and the
cold medium occurs by convection and is defined as:
(Q/A)total = hc (Twc – Tc) (3-7)
where hc is the convection heat transfer coefficient between the cold

medium side of the wall and the cold medium, and Tc is the average
temperature of cold medium.
In this example, Twh is going to be the dependent variable, and it
will be solved by iterating a function using a combination of above
Eqs. (3-3), (3-5), (3-6) and (3-7), as follows:
(Tg – Tc)/{[1/(hcg + hrg)] + (L/kwall) + (1/hc)}

− (hcg + hrg)(Tg – Twh) = 0 (3-8)
26
Heat Transfer From Combustion Chamber Walls
The above governing equation can be rewritten as an iteration

function K as follows:
K = −(Tg – Tc) + {[1/(hcg + hrg)]

+ (L/kwall) + (1/hc)}(hcg + hrg)(Tg – Twh) (3-9)
During iteration to determine Twh, all temperatures should be in

degrees Kelvin because of the fourth power behavior of radiation
heat transfer. Also, all thermophysical properties are assumed to
be constants. Nominal values of these variables for the sensitivity
analysis are assumed to be as follows:
Tg = 1000ºC (1273 K)
Tc = 100ºC (373 K)
kwall = 20 W/m-K
L = 0.01m
hcg = 100 W/m2-K
hc = 50 W/m2-K
εg = 0.2
σ = 5.67×10−8 W/m2-K4
For these nominal variables, the iteration function crosses zero at

1075.93 K as shown in Figure 3-1, and 42.5% of the total heat transfer
from hot gases to the hot side of the wall comes from radiation mode;
the rest, 57.5%, comes from convection mode.
The effects of hot gas temperature and cold medium temperature
to hot side wall temperature are shown respectively in Figures 3-2
and 3-3. Hot gas temperature affects hot side wall temperature almost
one-to-one, namely a slope of 0.925. However, cold side medium
temperature affects hot side wall temperature almost five-to-one,
namely a slope of 0.221.
The effects of wall parameters—thermal conductivity and wall
thickness—on hot side wall temperature are shown in Figures 3-4
and 3-5. Changes in wall thermal conductivities below 10 W/m-K
are more effective on hot side wall temperature. Hot side wall
temperature sensitivity to wall thickness is pretty much a constant,
3.5 C/cm.
27
2000
Iteration Function, K
1000
−1000
400 600 800 1000 1200 1400
Twh, K
Figure 3-1 Iteration function versus Twh
The convection heat transfer coefficients on both sides

of the wall have opposite effects on hot side wall temperature,
as shown in Figures 3-6 and 3-7. As the hot gas side convection
heat transfer coefficient increases, hot side wall temperature
Hot Side Wall Temperature, °C
2000
1600
1200
800
400
0
0 500 1000 1500 2000
Gas Temperature, °C
Figure 3-2 Hot side wall temperature versus gas temperature
28
1000
900
800
700
0 200 400 600 800 1000
Cold Medium Temperature, °C
Figure 3-3 Hot side wall temperature versus cold medium temperature
increases as well. Sensitivity of hot side wall temperature to

variations in the hot gas side convection heat transfer coefficient
is more prominent at lower convection heat transfer coefficient
values.
860
840
820
800
0 10 20 30 40 50
Wall Thermal Conductivity, W/m-K
Figure 3-4 Hot side wall temperature versus wall thermal conductivity
29
840
830
820
810
800
0 0.02 0.04 0.06 0.08 0.1
Wall Thickness, m
Figure 3-5 Hot side wall temperature versus wall thickness
The variation of hot side wall temperature at different hot

gas emissivities is given in Figure 3-8. Hot side wall temperature
is more sensitive to changes in lower values of hot gas
emissivity.
1000
900
800
700
600
0 100 200 300 400 500
Hot Gas Side Convection Heat Transfer Coefficient, W/m2-K
Figure 3-6 Hot side wall temperature versus hot gas side convection heat
transfer coefficient
30

1000
800
600
400
200
0 100 200 300 400 500
Cold Medium Side Convection Heat Transfer Coefficient,
W/m2-K
Figure 3-7 Hot side wall temperature versus cold medium side convection
heat transfer coefficient
When the nominal values of the variables given above are

varied ±10%, the results shown in Table 3-1 are obtained. Hot
side wall temperature sensitivities to a ±10% change in the
governing variables are given in descending order of importance,
950
900
850
800
750
700
0 0.2 0.4 0.6 0.8 1
Hot Gas Emissivity
Figure 3-8 Hot side wall temperature versus hot gas emissivity
31
Table 3-1 Effects of ±10% change in nominal values of

variables to hot side wall temperature
Change In Hot Change In Hot

Side Wall Side Wall
Temperature For Temperature For
Tg 1273 K −11.493% +11.981%
hc 50 W/m2-K +2.070% −1.991%
hcg 100 W/m2-K −1.262% +1.136%
εg 0.2 −0.912% +0.860%
Tc 373 K −0.295% +0.295%
kwall 20 W/m-K +0.056% −0.046%
L 0.01 m −0.051% +0.051%
and they are applicable around the nominal values assumed for
this study.
Hot side wall temperature is most sensitive to variations in hot
gas temperature. Next in order of sensitivity are the convection heat
transfer coefficients on both sides of the wall. Changes to emissivity
of hot gases affect the dependent variable at the same level as the
convection heat transfer coefficients. Next in order of sensitivity is
the cold medium temperature. Hot side wall temperature is least
sensitive to variations in wall thermal conductivity and wall thickness.
This variable order of sensitivity is applicable around the nominal
values assumed for this case, due to the nonlinear relationship
between the dependent variable and the independent variables.
32
CHAPTER
HEAT TRANSFER
FROM A HUMAN
4
BODY DURING
SOLAR TANNING
T he solar tanning of a human body was analyzed under

steady-state conditions with one-dimensional rate equations
in rectangular coordinates, and using temperature-independent
thermophysical properties. Human skin that is exposed to direct solar
radiation is considered to be in an energy balance. Energy goes into
the skin from both direct solar radiation and solar radiation scattered
throughout the atmosphere. Energy leaves the skin through a variety
of means and routes: by convection heat transfer and radiation heat
transfer (into the atmosphere), by conduction (to the inner portions
of the body), by perspiration, and by body basal metabolism. Other
energy gains and losses, such as those due to terrestrial radiation,
breathing and urination, are negligible.
Energy balance at the human skin gives the following heat transfer
equation:
Qsolar radiation absorbed + Qatmospheric radiation absorbed – Qconvection

– Qconduction to body – Qradiation emitted – Qperspiration − Qbasal metabolism = 0 (4-1)
33
For the present sensitivity analysis, the heat transfer rate equations
that will be used, and the nominal values that will be assumed for
energies outlined in Eq. (4-1), are given below.
Qsolar radiation absorbed = 851 W/m2 (4-2)
which assumes a gray body skin with an absorptivity, α = ε, of 0.8, on

a clear summer day at noon, with full sun exposure
Qatmospheric radiation absorbed = 85 W/m2 (4-3)
which is assumed to be about 10% of Qsolar radiation absorbed.
Qconvection = h(Tskin – Tenvironment) (4-4)
where h is the heat transfer coefficient between the skin surface that
is being tanned and the environment. In the present analysis, h is
assumed to be 28.4 W/m2-K and Tenvironment is 30°C.
Qconduction to body = (kbody/tbody)(Tskin – Tbody) (4-5)
where kbody = 0.2 W/m-K, tbody = 0.1 m, and Tbody = 37°C.
Qradiation emitted = εσ(T4skin – T4environment) (4-6)
where emissivity of skin surface ε = 0.8 and σ = 5.67 × 10−8 W/m2-K4.
Qperspiration = 337.5 W/m2 (4-7)
which corresponds to a 1 liter/hr perspiration rate for a human body

with a perspiration area of 2 m2.
Qbasal metabolism = 45 W/m2 (4-8)
which represents a 30-year-old male at rest.

There are ten independent variables that govern the dependent
variable Tskin in this heat transfer problem. Sensitivities to these ten
34
Heat Transfer From A Human Body During Solar Tanning
variables are analyzed in the region of the nominal values given above.
The governing Eq. (4-1) takes the following form, and can be solved
for Tskin by trial and error.
C1T4skin + C2Tskin = C3 (4-9)
where C1 = εσ, C2 = h + kbody/tbody, and
C3 = Qsolar radiation absorbed + Qatmospheric radiation absorbed – Qperspiration

− Qbasal metabolism + εσT4environment + hTenvironment + kbody/tbodyTbody
All the calculations are performed in degrees Kelvin for

temperature, since the governing equation is non-linear in
temperature. These sensitivities are presented in Table 4-1 below
in the order of their significance.
The most effective variable on the skin temperature is
Tenvironment and the least effective is the thermal conductivity of
human tissue. The skin temperature is an order of magnitude less
sensitive to changes in the thermal conductivity of human tissue—
skin-to-body conduction heat transfer length, heat transfer due to
basal metabolism, body temperature, atmospheric radiation absorbed,
and emissivity of skin surface—than changes in the temperature of
the environment, solar radiation absorbed, convection heat transfer
coefficient, and heat lost due to perspiration. Changes in some
variables, such as the environmental temperature and heat transfer
due to perspiration, behave linearly in the region of interest, and give
equal percentage changes to the dependent variable on both sides of
the variable's nominal value.
It is important to remind the reader that the order shown in
Table 4-1 is only useful in this region of the application due to
non-linear behavior of the sensitivities. The non-linear affects of
variables such as the convection heat transfer coefficient are given
in Figure 4-1 for Qsolar radiation absorbed = 851 W/m2 and for three different
perspiration rates.
The sensitivity of skin temperature to the convection heat transfer
coefficient is significant up to 50 W/m2-K. The heat transfer coefficient
35

variables to skin temperature
Skin Skin
Temperature, Temperature,
Tskin, Change Tskin, Change
Tenvironment 30°C −6.03% +6.03%
Qsolar radiation absorbed 851 W/m 2
−5.11% +5.13%
h 28.4 W/m -K2
+2.93% −2.51%
Qperspiration 337.5 W/m 2
+2.03% −2.03%
Emissivity of
0.8 +0.53 −0.51
skin surface, ε
Qatmospheric radiation absorbed 85 W/m2 −0.51% +0.51%
Tbody 37°C −0.45% +0.45%
Qbasal metabolism 45 W/m 2
+0.27% −0.27%
Skin-to-body
conduction 0.1 m −0.12% +0.10%
length, tbody
ktissue 0.2 W/m-K +0.11% −0.11%
from the skin surface to the environment can be determined from

appropriate empirical relationships found in References [6] and [10].
The heat transfer coefficient in the natural convection regime is
around 5 W/m2-K. If the wind picks up to, say, 8.9 m/s (20 mph),
then the heat transfer coefficient is in the turbulent flow regime, and
it increases to 20 W/m2-K. As the perspiration rate goes down, this
sensitivity increases. The sensitivity curves are given in Figure 4-2.
Similar results are obtained for an afternoon solar radiation
by assuming half the noon solar radiation, i.e., Qsolar radiation absorbed =
425.5 W/m2, and they are given in Figures 3-3 and 3-4.
As the solar radiation goes down, the skin temperature and its
sensitivity to the convection heat transfer coefficient decreases.
36
Tenvironment=30°C, Tbody=37°C, Qsolar radiation absorbed=851 W/m2

70
65
Skin Temperature, °C
60 Perspiration=0.5
liters/hr
55
Perspiration=1.0
50 liter/hr
45 Perspiration=1.5
liter/hr
40
35
30
0 50 100 150 200
Convection Heat Transfer Coefficient, W/m2-K
Figure 4-1 Skin temperature versus convection heat transfer coefficient

for Qsolar radiation absorbed = 851 W/m2 and for three different
perspiration rates
0
∂ Tskin / ∂h, m2-K2/W
−0.2
Perspiration=0.5
−0.4 liters/hr
Perspiration=1.0
−0.6 liters/hr
Perspiration=1.5
−0.8 liters/hr
−1
0 50 100 150 200
Figure 4-2 Skin temperature sensitivity to convection heat transfer

coefficient versus convection heat transfer coefficient for
Qsolar radiation absorbed = 851 W/m2 and for three different perspiration
rates
37
Tenvironment=30°C, Tbody=37°C, Qsolar radiation absorbed=425.5 W/m2
45
Skin Temperature, °C
40 Perspiration=0.5
liters/hr
Perspiration=1.0
35
liter/hr
Perspiration=1.5
30 liter/hr
25
0 50 100 150 200
Convective Heat Transfer Coefficient, W/m2-K
Figure 4-3 Skin temperature versus convection heat transfer coefficient

for Qsolar radiation absorbed = 425.5 W/m2 and for three different
perspiration rates
0.2
0.1
∂ Tskin / ∂h, m2-K2/W
0 Perspiration=0.5
−0.1 liters/hr
Perspiration=1.0
−0.2
liters/hr
−0.3 Perspiration=1.5
liters/hr
−0.4
−0.5
0 50 100 150 200
Figure 4-4 Skin temperature sensitivity to convection heat transfer

coefficient versus convection heat transfer coefficient for
Qsolar radiation absorbed = 425.5 W/m2 and for three different
perspiration rates
38
60
55
50
Tskin, C
45
40
35
30
20 24 28 32 36 40
Tenvironment, C
Figure 4-5 Skin temperature versus environment temperature
In high perspiration rates, the skin temperature is below

the environmental temperature, and it approaches the
environmental temperature as the heat transfer coefficient
increases.
50
45
Tskin, C
40
35
0.5 0.6 0.7 0.8 0.9 1
Emissivity Of Human Skin Surface, ε
Figure 4-6 Skin temperature versus emissivity of human skin surface
39
30
∂ Tskin / ∂ε, C
28
26
24
0.5 0.6 0.7 0.8 0.9 1
Emissivity Of Human Skin Surface, ε
Figure 4-7 Skin temperature sensitivity to emissivity of human skin surface

versus emissivity of human skin surface
Temperature of the skin behaves as shown in Figure 4-5 with

a constant sensitivity to the changes in the temperature of the
environment, in the domain of interest. The sensitivity behavior is
almost one-to-one; namely, ∂Tskin/∂Tenvironment equals 0.92.
Another variable that is analyzed in detail in the region of interest
is the emissivity of human skin surface. Skin temperature increases
with an increase in emissivity as shown in Figure 4-6.
Figure 4-7 provides the skin temperature sensitivity to emissivity
of human skin surface. The sensitivity decreases in a linear fashion as
the emissivity increases.
40
CHAPTER
EFFICIENCY OF
RECTANGULAR
5
FINS
H eat transfer from a surface can be enhanced by using fins.

Heat transfer from surfaces with different types of fins has been
studied extensively, as seen in References by Incropera, F. P. and
D. P. DeWitt [6] and by F. Kreith [10].
The present sensitivity analysis represents rectangular fins under
steady-state, one-dimensional, constant thermophysical property
conditions without radiation heat transfer. Energy balance to a
cross-sectional element of a rectangular fin gives the following second
order and linear differential equation for the temperature distribution
along the length of the fin.
d2T/dx2 – (hP/kA) (T – Tenvironment) = 0 (5-1)
where h is the convection heat transfer coefficient between the

surface of the fin and the environment in W/m2-C, k is the thermal
conductivity of the fin material in W/m-C, P is the fin cross-sectional
perimeter in meters, and A is the fin cross-sectional area in m2.
There can be different solutions to Eq. (5-1) depending upon the
boundary condition that is used at the tip of the fin. If the heat loss
41
to environment from the tip of the fin is neglected, the following

boundary conditions can be used:
T = Tbase at x = 0 and (dT/dx) = 0 at x = L (5-2)
The solution to Eqs. (5-1) and (5-2) can be written as
(T – Tenvironment) = (Tbase – Tenvironment) [cosh m(L-x)/cosh (mL)] (5-3)
where L is the length of the rectangular fin in meters and m =

(hP/kA)0.5 in 1/m.
The heat transfer from the rectangular fin can be determined from
Eq. (5-3) by finding the temperature slope at the base of the fin,
namely
Qfin = −kA(dT/dx) at x = 0 or (5-4)
Qfin = (Tbase – Tenvironment) sqrt(hkPA) tanh(mL) (5-5)
Here the sensitivities of variables that affect the efficiency of a

rectangular fin will be analyzed. Fin efficiency is generally defined by
comparing the fin heat transfer to the environment with a maximum
heat transfer case to the environment, where the whole fin is at the
fin base temperature, namely η = Qfin/Qmax where Qmax = hAfin(Tbase −
Tenvironment). For a rectangular fin, the fin heat transfer efficiency
is approximated by using Eq. (5-5), and by adding a corrected fin
length, Lc, for the heat lost from the tip of the fin.
η = tanh(mLc)/(mLc) (5-6)
where m = [h 2(w + t)/k wt ]0.5 and Lc = L + 0.5t.

For cases where the fin width, w, is much greater than its
thickness, t, m becomes
m = (2h/kt)1/2 (5-7)
42
Efficiency Of Rectangular Fins
There are four independent variables that affect the rectangular

fin heat transfer efficiency. These are the convection heat transfer
coefficient, h; the thermal conductivity of the fin material, k; length of
the fin, l; and thickness of the fin, t.
The sensitivity of efficiency to these four independent variables can
be obtained in closed forms by differentiating the efficiency equation
with respect to the desired independent variable. For example:
∂η/∂h = (0.5/h)[(1/cosh2(mLc)) – (tanh(mLc)/(mLc))] (5-8)
Fin efficiency as a function of the convection heat transfer

coefficient for two different thermal conductivities—aluminum and
copper—is given in Figure 5-1. The sensitivity of rectangular fin
efficiency with respect to the convection heat transfer coefficient is
given in Figure 5-2.
Fin efficiency is good in the natural convection regime and
degrades as high forced convection regimes are used. Sensitivity of
fin efficiency to the convection heat transfer coefficient is high in the
0.9
Fin Efficiency
kcu=377.2
0.8
W/m-C
0.7 kal=206
0.6 W/m-C
0.5
L=0.0508 m
0.4 t=0.002 m
0 100 200 300 400
Convection Heat Transfer Coefficient, W/m2-C
Figure 5-1 Rectangular fin efficiency versus convection heat transfer

coefficient for two different fin materials with L = 0.0508 m and
t = 0.002 m
43
−0.001
kcu=377.2
∂h/∂ h, m2-C/W
W/m-C
−0.002 kal=206
W/m-C
−0.003
L=0.0508 m
−0.004 t=0.002 m
−0.005
0 100 200 300 400
Figure 5-2 Sensitivity of rectangular fin efficiency to convection heat

transfer coefficient versus convection heat transfer coefficient
for two different fin materials with L = 0.0508 m and t = 0.002 m
natural convection regime and decreases as the forced convection

heat transfer coefficient increases.
Fin efficiency as a function of fin material thermal conductivity
for two different convection heat transfer coefficients—natural
convection regime and forced convection regime—is given in
Figure 5-3. The sensitivity of rectangular fin efficiency with respect
to fin material thermal conductivity is given in Figure 5-4.
Fin material thermal conductivity does not affect fin efficiency in
the natural convection regime except in the region of low thermal
conductivity materials. However, in the forced convection regime
the behavior is quite different. Fin material thermal conductivity
affects fin efficiency, and high thermal conductivity materials have to
be used in order to achieve high fin efficiency. The sensitivity of fin
efficiency to fin material thermal conductivity is high for low thermal
conductivities. The sensitivity diminishes as high fin material thermal
conductivities are utilized.
44
1
Fin Efficiency, η
h=5
0.8
W/m2-C
h=100
0.6 W/m2-C
L=0.0508 m
0.4 t=0.002 m
0 100 200 300 400 500
Fin Thermal Conductivity, W/m-C
Figure 5-3 Rectangular fin efficiency versus fin material thermal

conductivity for two different convection heat transfer
coefficients with L = 0.0508 m and t = 0.002 m
0.0084
h=5
0.0063 W/m2-C
∂ η/∂ k, m-C/W
0.0042 h=100
W/m2-C
0.0021
L=0.0508 m
t=0.002 m
0
0 100 200 300 400 500
Fin Thermal Conductivity, W/m-C
Figure 5-4 Sensitivity of rectangular fin efficiency to fin material thermal

conductivity versus fin material thermal conductivity for two
different convection heat transfer coefficients with L = 0.0508 m
and t = 0.002 m
45
1
Fin Efficiency, η
0.9 h=5
W/m2-C
0.8
h=100
W/m2-C
0.7
t=0.002 m
0.6 k=377.2 W/m-C
0 0.02 0.04 0.06 0.08 0.1
Fin Length, m
Figure 5-5 Rectangular fin efficiency versus fin length for two different

convection heat transfer coefficients with t = 0.002 m and
k = 377.2 W/m-C
Fin efficiency as a function of fin length for two different convection

heat transfer coefficients—natural convection regime and forced
convection regime—is given in Figure 5-5. The sensitivity of rectangular
fin efficiency with respect to fin length is given in Figure 5-6. Figure 5-6
shows sensitivities for combinations of two different convection heat
transfer coefficients and two different thermal conductivities.
Figure 5-5 shows that fin efficiency is a weak function fin length in
the natural convection regime, but this weakness becomes a strong
function of fin length in the forced convection regime. These results
can also be seen in Figure 5-6. In the natural convection regime,
sensitivity of fin efficiency to fin length is low, but increases as the fin
length increases. In the forced convection regime, sensitivity of fin
efficiency to fin length starts low, goes through a maximum as the fin
length increases, and decreases as the fin length increases further.
Fin efficiency as a function of fin thickness for two different
convection heat transfer coefficients—natural convection regime
and forced convection regime—is given in Figure 5-7. The sensitivity
of rectangular fin efficiency with respect to fin thickness is given in
Figure 5-8. Figure 5-8 shows sensitivities for combinations of two
46
8
7 h=5 W/m2-C & @
6 k=377.2 W/m-C
h=100 W/m2-C & @
∂ η/∂ L, 1/m
5 k=377.2 W/m-C
4 h=5 W/m2-C & @
3 k=206 W/m-C
2 h=100 W/m2-C & @
k=206 W/m-C
1
t=0.002 m
0
0 0.05 0.1
Fin Length, m
Figure 5-6 Sensitivity of rectangular fin efficiency to fin length versus fin

length for combinations of two different convection heat transfer
coefficients and two different thermal conductivities with t = 0.002 m
0.9
Fin Efficiency, η
0.8 h=5
W/m2-C
0.7 h=100
W/m2-C
0.6
0.5 k=377.2 W/m-C

L=0.0508 m
0.4
0 0.001 0.002 0.003 0.004 0.005
Fin Thickness, m
Figure 5-7 Rectangular fin efficiency versus fin thickness for two different

convection heat transfer coefficients with k = 377.2 W/m-C and
L = 0.0508 m
47
9
8
7
h=5 W/m2-C & @
6 k=377.2 W/m-C
∂η/∂ t, 1/m
5 h=100 W/m2-C & @

k=377.2 W/m-C
4 h=5 W/m2-C &
3 k=206 W/m-C
h=100 W/m2-C & @
2 k=206 W/m-C
1
L=0.0508 m
0
0 0.001 0.002 0.003 0.004 0.005
Fin Thickness, m
Figure 5-8 Sensitivity of rectangular fin efficiency to fin thickness versus

fin thickness for combinations of two different convection heat
transfer coefficients and two different thermal conductivities
with L = 0.0508 m
different convection heat transfer coefficients and two different

thermal conductivities.
Figure 5-7 shows that fin efficiency is a weak function of fin
thickness in the natural convection regime, but this weakness
becomes a strong function of fin thickness in the forced convection
regime. In Figure 5-8, in the natural convection regime, sensitivity of
fin efficiency to fin thickness starts high at low fin thickness values,
but decreases as the fin thickness increases. In the forced convection
regime, sensitivity of fin efficiency to fin thickness starts low, goes
through a maximum as the fin thickness increases, and decreases as
the fin thickness increases further.
A ten percent variation in independent variables around the
nominal values produces the following sensitivity results (Table 5-1)
for fin efficiency. The results are given for the natural convection
regime in descending order, from the most sensitive variable to the
least. Table 5-2 gives similar results for the forced convection regime.
48
Table 5-1 Rectangular fin efficiency change due to a 10%

change in variables nominal values for the natural
convection regime
Rectangular Rectangular
Fin Efficiency Fin Efficiency
L 0.0508 m +0.218% −0.239%
k 377.2 W/m-C −0.129% +0.106%
t 0.002 m −0.124% +0.102%
h 5 W/m -C
2
+0.117% −0.116%
The order of significance in fin efficiency change follows the same

pattern with respect to variables in both the natural convection and
forced convection regimes. However, in the forced convection regime,
fin efficiency changes are an order of magnitude higher than the
natural convection regime.
Table 5-2 Rectangular fin efficiency change due to a

10% change in variables nominal values for the
forced convection regime
Rectangular Rectangular
Fin Efficiency Fin Efficiency
L 0.0508 m +3.436% −3.478%
k 377.2 W/m-C −1.916% +1.639%
t 0.002 m −1.844% +1.574%
h 100 W/m -C 2
+1.806% −1.729%
49
CHAPTER
HEAT
TRANSFER FROM
6
A HOT DRAWN BAR
A hot drawn bar, assumed to be moving at a constant velocity out of

a die at constant temperature, will be treated as a one-dimension-
al heat transfer problem. The Biot numberfor the bar, htD/2k, will be
assumed to be less than 0.1, to assure no radial variation of tempera-
ture in the bar. Here ht is the total heat transfer coefficient from the
bar surface in W/m2-K, the sum of the convection heat transfer coef-
ficient and radiation heat transfer coefficient, D is the bar diameter,
and k is the bar thermal conductivity in W/m-K. Conduction, convec-
tion, and radiation heat transfer mechanisms affect the temperature
of the drawn bar.
Energy balance can be applied to a small element of the bar with a
width of dx:
Conduction heat transfer into the element – Conduction heat

transfer out of the element – Convection heat transfer out of the
element to the environment – Radiation heat transfer out of the
element to the environment = Rate of change of internal energy
of the element
51
The energy balance on the element can be written as follows:
Qconduction at x – Qconduction at x+dx – Qconvection from dx – Qradiation from dx

= ρcpAdx(dT/dθ) (6-1)
where ρ is the density of the bar in kg/m3, cp is the specific heat of the
bar at constant pressure in J-kg/K, A is the bar cross-sectional area
in m2, T is the temperature of the bar element in K, and dx/dθ is the
drawn bar velocity in m/s.
Assuming that all the bar thermophysical and geometrical
properties are constants, the following one-dimensional, second-order
and non-linear differential equation is obtained:
d2T/dx2 – (hP/kA)(T-Tenvironment) – (σεP/kA)(T4 – T4environment)

= (ρcpU/k)dT/dx (6-2)
where P is the bar perimeter in m, σ is Stefan Boltzmann constant,

5.67×10-8 W/m2-K4, ε is the bar surface emissivity, and U=dx/dθ the
speed of the hot drawn bar. The differential equation (6-2) reduces
to steady-state heat transfer from fins with a uniform cross-sectional
area; if the radiation heat transfer and the rate of change of internal
energy are neglected, see References by F. Kreith [10] and by
Incropera, F. P. and D. P. DeWitt [6].
The boundary condition for this heat transfer problem can be
specified as follows:
T = Tx=0 (temperature of the drawn bar at the die location)

at x = 0 (6-3)
and
T = Tenvironment as x goes to 4 (6-4)
The governing differential equation (6-2), along with boundary

conditions (6-3) and (6-4), can be solved by finite difference methods
and iteration, in order to determine the temperature at the i'th
location along the bar.
52
Heat Transfer From A Hot Drawn Bar
Another method to solve this non-linear heat transfer problem is to

define a radiation heat transfer coefficient utilizing the temperature of
the previous bar element, i-1, as follows:
hradiation = σε(T4i-1 – T4environment)/(Ti-1 – Tenvironment) (6-5)
and write the differential equation for location i along the bar as
d2Ti/dx2 – [(hconvection + hradiation) P/kA](Ti -Tenvironment)

= (ρcpU/k)dTi/dx (6-6)
The solution reached by linear differential equation (6-6) for

location i along the bar is valid for small x increments along the bar,
i.e. < 0.01 m, since the radiation heat transfer coefficient is calculated
using the temperature of the previous element i-1.
The solution to the above second order linear differential equation
(6-6) which satisfies both boundary conditions, (6-3) and (6-4), is:
(Ti – Tenvironment)/(Tx=0 – Tenvironment)

= exp{[(U/2α) – sqrt((U/2α)2 + m2)]x} (6-7)
where α = k/ρcp is the thermal diffusivity of the bar in m2/s and
m2 = (hconvection + hradiation)P/kA in 1/m2. (6-8)
This temperature distribution solution reduces to steady-state heat

transfer from rectangular fins with a uniform cross-sectional area
and the above applied boundary conditions, (6-3) and (6-4), if the
radiation heat transfer and the hot drawn bar velocity are neglected,
i.e., hradiation=0 and U=0 (see References by F. Kreith [10] and by
Incropera, F. P. and D. P. DeWitt [6]).
Sensitivity to governing variables is analyzed by fixing the drawn bar
temperature at the die, i.e., Tx=0=1273 K. There are eight independent
variables that affect the temperature distribution of the hot drawn bar.
The sensitivities of bar temperature to these variables are analyzed by
assuming the following nominal values for a special steel bar:
53
D = 0.01 m
ρ = 8000 kg/m3
cp = 450 J/kg-K
k = 40 W/m-K
ε = 0.5
U = 0.02 m/s
Tenvironment = 298 K
The last independent variable is the convention heat transfer

coefficient between the surface of the bar and the environment.
The convection heat transfer coefficient can be determined from a
drawn bar temperature requirement at a distance from the die. In
the present analysis, Tx=10 is specified to be 373 K. At approximately
x=10 m the radiation heat transfer contribution almost diminishes.
The convection heat transfer coefficient that meets the Tx=10 = 373 K
requirement is determined from the above solution to be:
hconvection = 46.17 W/m2-K
which is in the turbulent region of forced cooling air over the

cylindrical bar, i.e., ReD = 1083 where ReD= VairD/νair. The empirical
relationship for the convection heat transfer coefficient for air flowing
over cylinders is given in Reference [10] by F. Kreith as:
hconvectionD/kair = 0.615 (VairD/νair)0.466 for 40 < ReD < 4000 (6-9)
where Vair is mean air speed over the cylinder (2.18 m/s in this case),
kair is air thermal conductivity, and νair is air kinematic viscosity.
Air thermophysical properties are calculated at film temperature,
namely the average of bar surface temperature and environmental
temperature.
A comparison of convection and radiation heat transfer coefficients
as a function of distance from the die is given in Figure 6-1.
Heat transfer due to radiation is at the same order of magnitude
around the die. As the bar travels away from the die, radiation heat
transfer diminishes rapidly.
54
50
Heat Transfer Coeffieicents
hconvection
40
W/m2-K
30
20
10
hradiation
0
0 2 4 6 8 10
Distance From Die, m
Figure 6-1 Radiation and convection heat transfer coefficients as a function

of distance from die
Hot drawn bar temperature distributions, both with and without

radiation heat transfer, are shown in Figure 6-2. Radiation heat
transfer effects on bar temperature cannot be neglected below x=4
meters from the die.
In the initial sensitivity analysis, radiation heat transfer effects
will be neglected, namely hradiation=0. Hot drawn bar temperatures
as a function of distance from the die for different convection heat
transfer coefficients are given in Figure 6-3. Temperatures are very
sensitive to low convection heat transfer coefficients.
The sensitivity of bar temperature at x=10 m to the convection heat
transfer coefficient is given in Figure 6-4. Bar temperature sensitivity
is high at natural convection and at low forced-convection heat
transfer regions. As the forced-convection heat transfer coefficient
increases, bar temperature sensitivity to the convection heat transfer
coefficient decreases.
Bar temperature at x=10 m as a function of the convection heat
transfer coefficient is shown in Figure 6-5. Increasing the convection
heat transfer coefficient reduces its effects on bar temperature.
55
1400
1200
Temperature, K
1000
Without Radiation
800
With Radiation
600
400
200
0
0 2 4 6 8 10
Figure 6-2 Hot drawn bar temperature with and without radiation heat
transfer effects
1400
1200 hconvection=5
W/m2-K
Temperature, K
1000 hconvection=20
W/m2-K
800 hconvection=40
W/m2-K
600
hconvection=60
400 W/m2-K
200
0 2 4 6 8 10
Figure 6-3 Hot drawn bar temperature for different convection heat trans-
fer coefficients with hradiation=0
56
∂T/∂ hconvection, m2-K2/W 0
−10
−20
−30
−40
−50
0 10 20 30 40 50
hconvection, W/m2-K
Figure 6-4 Bar temperature sensitivity @ x=10 m to convection heat trans-

fer coefficient
Bar Temperature @ x=10 m, K
1000
800
600
400
200
0 10 20 30 40 50 60
hconvection, W/m2-K
Figure 6-5 Bar temperature @ x=10 m as a function of convection heat

transfer coefficient
57
400
380
360
340
7000 7500 8000 8500 9000
Bar Density, kg/m3
Figure 6-6 Bar temperature @ x=10 m versus bar density
Bar temperature at x=10 m varies close to a linear behavior with

bar density, as shown in Figure 6-6. ?T/?ρ is 0.0239 K-m3/kg in the
7000 to 8000 kg/m3 bar density range.
Bar temperature at x=10 m also varies close to a linear behavior
with bar-specific heat at constant pressure, as shown in Figure 6-7.
?T/?cp is 0.4254 K2-kg/J in the 400 to 500 J/kg-K bar specific heat
range.
Bar temperature at x=10 m varies linearly with bar thermal
conductivity, as shown in Figure 6-8. Bar temperature is a weak
function of bar thermal conductivity in this problem. ?T/?k is 0.0007
K2-m/W in the 20 to 60 W/m-K bar thermal conductivity range.
Bar temperature at x=10 meters versus bar velocity is given in
Figure 6-9.
Hot drawn bar velocity does not affect bar temperature at low
velocities, i.e., U<0.01 m/s. As bar velocity increases above 0.01 m/s,
bar temperature increases rapidly. Bar temperature at x=10 meters
has a sensitivity of about 10,000 K-s/m to changes in bar velocity
around 0.03 m/s. As bar velocity increases further, bar temperature
sensitivity at x=10 meters starts to decrease.
58

400
380
360
340
400 420 440 460 480 500
Bar Specific Heat, J/kg-K
Figure 6-7 Bar temperature @ x=10 m versus bar specific heat at constant
pressure
373.04
373.03
373.02
373.01
373
20 30 40 50 60
Bar Thermal Conductivity, W/m-K
Figure 6-8 Bar temperature @ x=10 m versus bar thermal conductivity
Bar temperature at x=10 meters versus bar diameter is given

in Figure 6-10. Bar temperature at x=10 meters is not sensitive to
bar diameter changes in small diameter values, i.e., D<0.005 m. As
bar diameter increases, bar temperature sensitivity goes through a
59
700
600
500
400
300
200
0 0.01 0.02 0.03 0.04 0.05
Bar Velocity, m/s
Figure 6-9 Bar temperature @ x=10 m versus bar velocity

1000
800
600
400
200
0 0.01 0.02 0.03 0.04 0.05
Bar Diameter, m
Figure 6-10 Bar temperature @ x=10 m versus bar diameter
maximum of about 19,500 K/m at around D=0.02 meters, and starts to

decrease as the bar diameter increases.
Bar temperature at x=10 meters versus environmental temperature
is given in Figure 6-11. The relationship is linear as expected.
60

380
375
370
365
360
15 20 25 30
Environment Temperature, C
Figure 6-11 Bar temperature @ x=10 m versus environment temperature
The slope of the curve is 0.9231 K/C under the assumed nominal
conditions.
When the nominal values of the independent variables given above
are varied +-10%, the results shown in Table 6-1 are obtained. The
sensitivity analysis is conducted by neglecting radiation heat transfer
at x=10 meters.
The convection heat transfer coefficient, bar diameter, bar velocity,
bar density and bar-specific heat at constant pressure have the same
order of magnitude sensitivity on bar temperature at x=10 m. Changes
in the temperature of the environment affect bar temperature at x=10
m, at an order of magnitude less. Changes to the thermal conductivity
of the bar have the least effect on bar temperature at x=10 m. The
sensitivity magnitudes and order that are shown in Table 6-1 are only
valid around the nominal values that are assumed for the independent
variables for this analysis. Bar velocity, bar density and bar-specific
heat at constant pressure have the same sensitivity effects on the
temperature of the bar as can be seen in Eq. (6-7).
Another interesting sensitivity analysis can be performed around
x=0.5 m, where both the convection and the radiation heat transfers
are in effect.
61
Table 6-1 Effects of ±10% change in nominal values of vari-

ables to bar temperature @ x=10 m
Bar Temperature Bar Temperature

@ x=10 m Change @ x=10 m Change
hconvection 46.17 W/m2-K +5.879% −4.549%
D 0.01 m −4.986% +5.280%
U 0.02 m/s −4.986% +5.279%
ρ 8000 kg/m 3
−4.986% +5.279%
cp 450 J/kg-K −4.986% +5.279%
Tenvironment 25ºC (298 K) −0.619% +0.619%
k 40 W/m-K −0.0007% +0.0007%

variables to bar temperature @ x=0.5 m, including
radiation heat transfer
Bar Temperature Bar Temperature

@ x=0.5 m Change @ x=0.5 m Change
D 0.01 m −1.386% +1.183%
U 0.02 m/s −1.386% +1.183%
ρ 8000 kg/m 3
−1.386% +1.183%
cp 450 J/kg-K −1.386% +1.183%
hconvection 46.17 W/m -K2
+0.788% −0.783%
ε 0.5 +0.500% −0.482%
Tenvironment 25ºC (298 K) −0.046% +0.046%
k 40 W/m-K −0.0003% +0.0003%
62
Sensitivities to a +-10% change in independent variables are given

in descending order of importance in Table 6-2.
All the sensitivities to governing independent variables at x=0.5
meters are at the same order of magnitude, except Tenvironment and k.
Variations in radiation and convection heat transfer losses from the
bar have similar effects on bar temperature close to the die.
63
CHAPTER
MAXIMUM
CURRENT IN
7
AN OPEN-AIR
ELECTRICAL WIRE
M aximum current in an open-air single electrical wire (not in a

bundle), will be analyzed under steady-state and one-dimensional
cylindrical coordinates, with constant property conditions. The wire is
assumed to be a cylindrical conductor with a certain diameter and with
certain material characteristics; i.e., resistivity. The wire conductor is
insulated with concentric layers of insulation material that can stand
up to a certain wire conductor temperature, Tc, which will be the
temperature rating of the wire. Heat generated in the wire conductor
is I2R and the temperature within the wire conductor is assumed to be
uniform—the conductor temperature does not vary radially from the
center to the outer radius of the wire conductor.
Heat is transferred by conduction through the wire insulator and by
convection from the surface of the wire insulator to the environment.
Radiation heat transfer from the surface of the wire insulator is
neglected. The conduction heat transfer from the conductor to the
outer radius of the wire insulator can be written by the rate equation
in cylindrical coordinates:
Q = 2πLkins (Tc – Tinsulation outer radius)/ln(rw/rc) (7-1)
65
The convection heat transfer from the outer surface of the wire
insulator to the environment can be written by the rate equation in
cylindrical coordinates:
Q = 2πrwLh (Tinsulation outer radius – Tenv) (7-2)
The heat transfer mechanisms in Eqs. (7-1) and (7-2) are in a series
thermal resistance path. The energy balance for this heat transfer
problem, heat generated by the conductor equals heat lost to the
environment, can be written as follows, by eliminating Tinsulation outer radius
from Eqs. (7-1) and (7-2):
I2R = (Tc − Tenv)/{[ln(rw/rc)/(2πLkins)] + (1/2πrwLh)} (7-3)
where the first term in the denominator is the conduction heat

transfer resistance in the wire insulation, and the second term is
the convection heat transfer resistance at the insulated wire's outer
surface.
Definitions of variables:
I = Current through the conductor in amps
R = Resistance of the wire can also be written as ρL/πrc2 where
ρ = Resistivity of the conductor material in Ω-m
L = Length of the conductor in meters
rc = Conductor radius in meters
Tc = Temperature rating of the wire in C
Tenv = Temperature of the environment in C
rw = Radius of insulated wire in meters
kins = Thermal conductivity of wire insulation in W/m-C
h = Convection heat transfer coefficient in W/m2-C
The maximum current that a wire can stand can be written from
Eq. (7-3) by replacing resistance of the wire with resistivity of the
conductor material:
Imax = πrc {(2/ρ)(Tc − Tenv)/[ln(rw/rc)/kins + (1/rwh)]}1/2 (7-4)
66
Maximum Current In An Open-Air Electrical Wire
There are six independent variables that govern the maximum

current allowed in the wire. These are the conductor radius, rc, the
conductor resistivity, ρ, the temperature potential (the temperature
rating of the wire minus the temperature of the environment), Tc˛−
Tenv, the radius of the insulated wire, rw, the thermal conductivity of
the insulating material, kins, and the open air convention heat transfer
coefficient, h.
The sensitivities of maximum current allowed with respect to
these six variables will be analyzed as a function of wire conductor
diameter. Since wire gauges are used in the industry, the sensitivity
results are presented for different wire gauges. Table 7-1 provides the
standard wire gauges and their conductor diameters.
In this problem it is fortunate that the sensitivities can be obtained
in a closed form by differentiating the above maximum current
Table 7-1 Wire conductor diameter versus AWG
Wire Conductor Diameter, mm American Wire gauge (AWG)

0.8128 20
1.0236 18
1.2900 16
1.6280 14
2.0520 12
2.5900 10
3.2640 8
4.1150 6
5.1890 4
6.5430 2
7.3406 1
8.2550 1/0
9.2710 2/0
10.4140 3/0
11.6840 4/0
67
40
∂Imax/∂ (2rc), A/mm
30
20
AWG20 AWG6 AWG2 AWG1/0 AWG3/0

10
0 2 4 6 8 10 12
Wire Conductor Diameter, mm
Figure 7-1 Maximum wire current to wire conductor diameter sensitivity as

a function of wire diameter
equation with respect to a desired variable while holding the other

variables constant, namely ∂Imax/∂(2rc), ∂Imax/∂Tenv, ∂Imax/∂ρ, ∂Imax/∂rw,
∂Imax/∂kins, and ∂Imax/∂h.
For example,
∂Imax/∂Tenv = πrc {(2/ρ)(Tc − Tenv)/[ln(rw/rc)/kins + (1/rwh)]}−1/2

{−(1/ρ)/[ln(rw/rc)/kins + (1/rwh)]} (7-5)
The sensitivity of maximum wire current to conductor radius is

given in Figure 7-1.
Change in maximum wire current with respect to wire conductor
diameter is high at small diameters, and decreases as the wire
diameter increases.
Figure 7-1 was obtained by using the following values for the other
variables:
ρ = 1.60E-08 Ω-m
Tc = 60ºC
Tenv = 30ºC
68
400
Maximum Current
300 NEC
Current
Allowed, A
Allowed, A
200
Maximum
Current
100
Allowed, A
0
0 2 4 6 8 10 12
Figure 7-2 Maximum wire current allowed versus wire conductor diameter
@ a wire temperature rating of 60ºC
rw = 3rc
kins = 0.3 W/m-C
h = 9 W/m2-C
Under the above conditions the maximum wire current calculated

is compared to National Electric Code (NEC) recommendations in
Figure 7-2.
The calculated maximum wire currents are very much in line with
the NEC-recommended values up to AWG6. The variable assumptions
given above start to predict higher maximum current values for large
diameter wires, and therefore have to be adjusted for large diameter
wire applications.
The sensitivity of maximum wire current to environmental
temperature, ∂Imax/∂Tenv, is given in Figure 7-3.
Sensitivities in Figure 7-3 are for a wire temperature rating of
60ºC. As the environmental temperature approaches the wire rating
temperature, the maximum wire current allowed becomes more
sensitive to environmental temperature changes.
69
0
–2
∂ Imax/∂ Tenv, A/C
–4
–6 Tenv=0°C
Tenv=20°C
–8 Tenv=40°C
–10 Tenv=50°C
–12
0 2 4 6 8 10 12
Figure 7-3 Maximum wire current to environment temperature sensitivity as

a function of wire conductor diameter at different environmental
temperatures for a wire temperature rating of 60°C
The sensitivities of Maximum wire current allowed to wire

insulation thickness, rw − rc, are given in Figures 7-4 and 7-5.
For a given insulation thickness, the change in maximum allowable
wire current reaches its peak at a certain wire conductor diameter
and then starts to decrease, as seen in Figure 7-4. As the insulation
thickness increases, the maximum sensitivity point moves towards
smaller diameter conductors. Another way to look at the maximum
allowable wire current sensitivity to insulation thickness is shown
in Figure 7-5, where the sensitivities decrease as the wire insulation
thickness increases for different gauge wires.
The sensitivities of maximum wire current allowed to
wire insulation thermal conductivity, ∂Imax/∂kins, are given in
Figure 7-6.
Maximum wire current is more sensitive to wire insulation thermal
conductivity variations at lower thermal conductivity values and in
large gauge wires. Maximum wire current is almost insensitive to
wire insulation thermal conductivity variations below 3 mm diameter
(AWG8) conductors.
70
10000
Ins Thickness=
0.5*Wire Dia
∂ Imax/∂ (rw-rc), A/m
8000
Ins Thickness=
0.75*Wire Dia
6000
Ins Thickness=
1.0*Wire Dia
4000
Ins Thickness=
1.25*Wire Dia
2000
Ins Thickness=
1.5*Wire Dia
0
0.000 0.002 0.004 0.006 0.008 0.010 0.012
Wire Conductor Diameter, m
Figure 7-4 Maximum wire current to insulation thickness sensitivity as

a function of wire conductor diameter for different insulation
thicknesses
8000
∂ Imax/∂ (rw-rc), Amp/m
7000 AWG20
AWG16
6000
AWG12
5000 AWG8
4000 AWG4
AWG1
3000
AWG2/0
2000 AWG4/0
1000
0.5*d 0.75*d d 1.25*d 1.5*d
Insulation Thickness
Figure 7-5 Maximum wire current to insulation thickness sensitivity as a

function of insulation thickness for different wire diameters
71
300
AWG20
∂ Imax/∂ kins, A-m-K/W
250
AWG16
200 AWG12
AWG8
150 AWG4
100 AWG1
AWG2/0
50 AWG4/0
0
0.2 0.25 0.3 0.35 0.4
Thermal Conductivity Of Wire Insulation, W/m-K
Figure 7-6 Maximum wire current to thermal conductivity of wire insulation

sensitivity for different wire gauges
500
∂ Imax/∂ h, A-m2-K/W
400 AWG20
AWG16
300 AWG12
AWG8
200 AWG4
AWG1
100 AWG2/0
AWG4/0
0
8 12 16 20
2
Convection Heat Transfer Coefficient, W/m -K
Figure 7-7 Maximum wire current to convection heat transfer coefficient

sensitivity for different wire gauges
72
600
∂ Imax/∂ h, A-m2-K/W
500
h=7 W/m2-K
400
h=8 W/m2-K
300 h=9 W/m2-K
h=12 W/m2-K
200 h=15 W/m2-K
h=20 W/m2-K
100
0
0 0.002 0.004 0.006 0.008 0.01 0.012
Wire Diameter, m
Figure 7-8 Maximum wire current to convection heat transfer coefficient

sensitivity as a function of wire diameter for different convection
heat transfer coefficients
The sensitivities of maximum wire current allowed to convection

heat transfer coefficient, ∂Imax/∂h, are given in Figures 7-7 and 7-8.
Maximum wire current is more sensitive to convection heat transfer
coefficient variations at high convection heat transfer coefficient
values and in large gauge wires. As the wire surface area increases
in large gauge wires, changes in the heat transfer coefficient gain
importance.
The sensitivities of maximum wire current allowed to wire
conductor resistivity, ∂Imax/∂ρ, are given in Figure 7-9.
The value of ∂Imax/∂ρ is always negative; when wire resistivity
increases maximum wire current decreases, and vice versa. Maximum
wire current is more sensitive to wire conductor resistivity variations
at low wire conductor resistivity values and in large gauge wires.
The above sensitivity graphs show that maximum wire current
sensitivity to all the governing variables behaves non-linearly.
Sometimes it is more appropriate to analyze these sensitivities in
the region of interest, and rank them according to their effects on
maximum wire current.
73
0
∂ Imax/∂ ρ x 10-8, A/Ohm-m
AWG 20
–50 AWG 16
AWG 12
AWG 8
–100 AWG 4
AWG 1
–150 AWG 2/0
AWG 4/0
–200
1.00E-08 1.50E-08 2.00E-08 2.50E-08 3.00E-08
Wire Resistivity, ρ, Ohm-m
Figure 7-9 Maximum wire current to wire conductor resistivity sensitivity as

a function of wire conductor resistivity for different wire gauges
For example, maximum wire current sensitivities are analyzed for

a 4.115 mm diameter, AWG6, conductor with a temperature rating of
60ºC for the following nominal values of the variables:
Conductor diameter: 2rc = 4.115 ± 0.4115 mm

Conductor resistivity: ρ = 1.60E-08 ± 0.16E-08 Ω-m
Temperature Rating Of Wire − Temperature of the environment:
Tc − Tenv = 30 ± 3ºC
Wire insulation thickness: rw − rc = 2rc ± 0.2rc = 4.115 ± 0.4115 mm
Wire insulation thermal conductivity: kins = 0.3 ± 0.03 W/m-C
Convection heat transfer coefficient: h = 9 ± 0.9 W/m2-C
The nominal values are varied ±10%, and the effects of these
variations are presented in Table 7-2.
The sensitivity effects to maximum wire current shown in
Table 7-2 are given in descending order. The most effective variable
is the wire conductor diameter, and the least effective is the thermal
conductivity of the wire insulation material. There is an order of
magnitude difference in the effects of these two variables. In this
region of interest, it would be more advisable to focus on the accuracy
74

variables to maximum wire current for a 4.115 mm
diameter, AWG6, conductor with a temperature
rating of 60ºC
Maximum Wire Maximum Wire

Current Change Current Change
Conductor
4.115 mm −13.9% +14.5%
Diameter, 2rc
Conductor 1.60E-08
+5.4% −4.7%
Resistivity, ρ Ω-m
Temperature Rating Of
Wire Tc – Temperature 30ºC −5.1% +4.9%
Of Environment, Tenv
Convection Heat
9 W/m2-C −4.32% +4.01%
Transfer Coefficient, h
Insulation Thickness,
2rc −2.35% +2.17%
rw − rc
Thermal Conductivity
0.3 W/m-C −0.93% +0.78%
Of Wire insulation, kins
of the wire conductor diameter rather than the accuracy of the wire
insulation thermal conductivity. It is important to remember that the
maximum wire current change order shown in Table 7-2 is good only
in this region of the application, due to non-linear behavior of the
sensitivities.
75
CHAPTER
EVAPORATION
OF LIQUID
8
NITROGEN IN A
CRYOGENIC BOTTLE
H eat transfer in cryogenic bottles involves conduction, convection

and radiation modes. Dewar invented the vacuum flask at the
beginning of the twentieth century to minimize heat transfer and
contain low or high-temperature fluids in it.
In this example, a cryogenic bottle with a stainless steel inner tube,
a vacuum gap and an outer insulation layer will be utilized to store
liquid nitrogen. The bottle will have a venting system to release the
evaporating nitrogen. Heat transfer from the sides of the cryogenic
bottle will be considered. The top and the bottom surfaces of the
bottle are assumed to be well-insulated. The temperature of the inner
wall of the inner tube is assumed to be that of liquid nitrogen, namely
a negligible convection heat transfer resistance between the liquid
nitrogen and the inner wall of the inner tube.
Heat transfer occurs from the environment to the nitrogen
under steady-state conditions and in one-dimensional cylindrical
coordinates. The heat transfer from the sides of the tube can be
calculated by using the following series circuit:
Q = (Tenvironment – Tliquid nitrogen)/Σ Rij (8-1)
77
where Q is the steady-state heat transferred from the environment

to the liquid nitrogen in Watts, T is the temperature in K, and Rij
represents the thermal resistances on the thermal path in series
between Tenvironment and Tliquid nitrogen.
R21 is the conduction heat transfer thermal resistance between
the inner surface of the inner tube, radial location (1), to the outer
surface of the inner tube, radial location (2); in other words, thermal
resistance through the thickness of the inner tube.
R21 = ln(r2/r1)/(2πLkss) (8-2)
where
ln is natural logarithm of argument (r2/r1),

r1 = inner radius of inner tube in m,
r2 = outer radius of inner tube in m,
L = height of the tube in m, and
kss = thermal conductivity of stainless steel inner tube in W/m-K.
R32 is the radiation heat transfer thermal resistance between the

inner surface of the outer insulation tube, radial location (3), and the
outer surface of the inner tube, radial location (2); in other words,
through the vacuum gap.
R32 = 1/(2πr3Lhr) (8-3)
where
r3 = inner radius of outer insulation tube in m,

hr = εσ(T34 – T24)/(T3 – T2) in W/m2-K,
ε = emissivity of inner surface of outer insulation tube,
σ = Stefan-Boltzmann constant, namely 5.67 × 10−8 W/m2-K4
T3 = temperature of inner surface of outer insulation tube in K,
T2 = temperature of outer surface of inner tube in K
R43 is the conduction heat transfer thermal resistance between the

outer surface of the outer insulation tube, radial location (4) and the
inner surface of the outer insulation tube, radial location (3).
78
Evaporation Of Liquid Nitrogen In A Cryogenic Bottle
R43 = ln(r4/r3)/(2πLkinsulation) (8-4)
where
r4 = outer radius of outer insulation tube in m and

kinsulation = thermal conductivity of insulation material of the outer
tube in W/m-K.
Renvironment-4 is the convection heat transfer thermal resistance between

the environment and the outer surface of outer insulation tube, radial
location (4). Radiation heat transfer between the environment and
the outer surface of outer insulation tube is assumed to be negligible.
Renvironment-4 = 1/(2πr4ho) (8-5)
where
ho = convection heat transfer coefficient between the environment

and the outer surface of the outer tube in W/m2-K.
Since hr depends on unknown temperatures T3 and T2, an iterative

method is used to calculate these temperatures, and thereafter hr
(i.e., the radiation heat transfer coefficient between the inner surface
of the outer insulation tube and the outer surface of the inner tube, or
through the vacuum gap). All calculations are done using the Kelvin
temperature scale due to fourth power calculations in the radiation
heat transfer coefficient. The thermophysical properties used in these
calculations are assumed to be constants.
Nominal values of the variables used in this problem are as follows:
Tenvironment = 293 K
Tliquid nitrogen = 77 K
r1 = 0.1 m
r2 = 0.106 m, namely inner tube stainless steel wall thickness is 6 mm
r3 = 0.116 m, namely vacuum gap is 0.01 m
r4 = 0.166 m, namely outer tube insulation thickness is 0.05 m
L = 0.8 m
kss = 10 W/m-K
79
kinsulation = 0.002 W/m-K

ho = 10 W/m2-K, namely in natural convection regime
ε = 0.5, for inner surface of outer insulation tube
Hfg = 2 × 105 J/kg, latent heat of evaporation for liquid nitrogen
ρliquid nitrogen = 800 kg/m3
It is assumed that the initial volume of liquid nitrogen in the bottle

is 5 liters, or that it contains 4 kg, Mliquid nitrogen, of liquid nitrogen. The
dependent variable in the calculations is the time, t in seconds, that
it takes to completely evaporate the liquid nitrogen. Heat losses,
Q, from the cryogenic bottle, times time, t, should equal the total
latent heat of vaporization for the liquid nitrogen, Mliquid nitrogen Hfg, for
complete evaporation.
t = Mliquid nitrogen Hfg/Q (8-6)
It takes 198 hours, or 8.25 days, for 5 liters of liquid nitrogen to

evaporate under the above nominal conditions.
Unknown temperature T3, temperature of the inner surface of the
outer insulation tube, is obtained by iteration, as shown in Figure 8-1.
180
160
Temperature, K
140
120
100
80
11
13
15
17
19
21
1
Number Of Iterations
Figure 8-1 Convergence of temperature, T3, at the inner surface of an outer

insulation tube
80
Temperature calculations converge to a final value after ten iterations,

with less than 1% error.
Liquid nitrogen evaporation time is very sensitive to the cryogenic
bottle's insulation layer thermal conductivity, kinsulation. Five liters
of liquid nitrogen evaporation time versus insulation layer thermal
conductivity is given in Figure 8-2. The sensitivity of evaporation
time increases rapidly as thermal conductivity decreases below
0.001 W/m-K, as shown in Figure 8-3.
The second dominant heat transfer mechanism is the radiation
heat transfer R32. The sensitivity of evaporation time to the emissivity,
ε, of the inner surface of the outer insulation tube, is shown in
Figure 8-4. Evaporation time increases as emissivity of the insulation
inner surface, and therefore the radiation heat transfer, decreases.
When the nominal values of the variables given above are varied
±10%, the results shown in Table 8-1 are obtained. Sensitivities of
the time that it takes for five liters of liquid nitrogen to evaporate to a
±10% change in independent variables are given in descending order
1500
Evaporation Time, hr
1000
500
0
0 0.002 0.004 0.006 0.008 0.01
Insulation Layer Thermal Conductivity, kinsulation, W/m-K
Figure 8-2 All-liquid nitrogen evaporation time versus bottle insulation

layer thermal conductivity
81
0.0E+00
∂ t/∂ kinsulation, hr-m-K/W
–1.0E+06
–2.0E+06
–3.0E+06
0 0.001 0.002 0.003 0.004 0.005
Insulation Layer Thermal Conductivity, kinsulation, W/m–K
Figure 8-3 Sensitivity of all-liquid nitrogen evaporation time to bottle

insulation layer thermal conductivity
–100
∂ t/∂ ε
–200
–300
–400
0 0.2 0.4 0.6 0.8 1
Emissivity, ε, Of Inner Surface Of Outer Insulation Tube
Figure 8-4 Sensitivity of all liquid nitrogen evaporation time to emissivity of

the inner surface of an outer insulation tube
82
Table 8-1 Effects of ±10% change in the nominal values of

variables to time for 5 liters of liquid nitrogen to
evaporate
Change in Time Change in Time

For 5 Liters Of For 5 Liters Of
Liquid Nitrogen Liquid Nitrogen
To Evaporate For To Evaporate For
L 0.8 m +11.111% −9.091%
kinsulation 0.002 W/m-K +9.323% −7.694%
r1 0.1 m +8.062% −6.931%
r4 − r3 (outer
tube insulation 0.05 m −7.212% +6.947%
thickness)
ε 0.5 +1.671% −1.434%
Tenvironment 293 K +1.064% −1.043%
r3 − r2
0.01 m +0.751% −0.740%
(vacuum gap)
r2 − r1 (stainless
steel inner tube 0.006 m +0.449% −0.445%
wall thickness)
ho 10 W/m2-K +0.0315% −0.0258%
kss 10 W/m-K +0.00006% −0.000049%
of importance, and they are applicable only around the nominal values
assumed for this study.
The time required for five liters of liquid nitrogen to evaporate
is most sensitive to cryogenic tube height, insulation layer thermal
conductivity, inside radius of the inner tube, and outer tube insulation
thickness. The second tier of sensitivities are an order of magnitude
less; emissivity of inner surface of the outer insulation tube,
temperature of environment, vacuum gap thickness, and stainless
steel inner tube wall thickness. Time for five liters of liquid nitrogen to
83
evaporate is least sensitive to the convection heat transfer coefficient

between the environment and the outer surface of the outer tube, and
the thermal conductivity of the stainless steel inner tube.
R43, conduction heat transfer thermal resistance between the
outer surface of the outer insulation tube and the inner surface of the
outer insulation tube, dominates the results in this case. The second
dominant heat transfer mechanism is R32, radiation heat transfer
thermal resistance between the inner surface of the outer insulation
tube and the outer surface of the inner tube; in other words, the
vacuum gap.
It is important to remember that these calculated sensitivities are
good only around the nominal values assumed for this case. Changing
these nominal values will change the magnitudes and orders of these
sensitivities, a result of the non-linear form of governing heat transfer
equations.
84
CHAPTER
THERMAL
STRESS IN A PIPE
9
T hermal stresses generated by temperature variations in the

wall of a pipe have been studied extensively in Reference by
Timoshenko, S. and J. N. Goodier [17]. The stress, strain, radial
displacement relationships in cylindrical coordinates are treated
in detail in Reference [17]. To calculate the thermal stresses in
a pipe wall, the temperature distribution in the pipe wall has to
be known. The temperature distribution in the radial direction,
R, can be obtained from a steady-state, one-dimensional heat
conduction equation in cylindrical coordinates. By assuming constant
thermophysical properties and no heat sources in the pipe wall, the
heat conduction equation for the temperature distribution, T, is:
d2T/dR2 + (1/R)dT/dR = 0 (9-1)
If the temperatures at the inner surface, Ti, and the outer surface,
To, of the pipe wall are known, Eq. (9-1) can be solved by using the
following boundary conditions:
T = Ti at R = Ri (9-2)
85
and
T = To at R = Ro (9-3)
Equation (9-1) can be integrated and the following temperature

distribution in the pipe wall can be obtained by applying the boundary
conditions in Eqs. (9-2) and (9-3):
(T − To) = (Ti – To) [ln(Ro /R) / ln(Ro /Ri)] (9-4)
Derivation of thermal stress integral equations for σR, σθ, σZ, thermal
stresses in three cylindrical coordinates, are detailed in Reference [17].
The thermal stress integral equations for σR, σθ and σZ were developed
by applying three conditions to the stress-strain relationships. One of
the conditions results from the fact that the strain along the length of
a long pipe is zero. The integration constants are determined from two
boundary conditions. These boundary conditions come from the radial
stress, σR, being zero at the inner and outer surfaces of the pipe wall.
The thermal stress integral equations are:
σR = [αE/(1 – ν)](1/R2){[(R2 – Ri2)/(Ro2 – Ri2)]

} (9-5)
σθ = [αE/(1 – ν)](1/R2){[(R2 + Ri2)/(Ro2 – Ri2)]

– TR2} (9-6)
σZ = [αE/(1 – ν)]{[(2/(Ro2 – Ri2)] -T} (9-7)
where
σR, σθ, and σZ are the thermal stress distributions at the pipe wall in
MPa (Mega Pascals),
α is the coefficient of thermal expansion of the pipe wall material in
m/m-C,
E is modulus of elasticity for the pipe wall material in MPa, and
ν is Poisson's ratio for pipe wall material.
86
Thermal Stress In A Pipe
The temperature distribution in Eq. (9-4) is substituted into

Eqs. (9-5) through (9-7) to determine thermal stress distributions in
three directions in cylindrical coordinates. If the heat is flowing from
inside the pipe to the environment, maximum compressive stresses
occur at the inside radius of the pipe wall, and the maximum tensile
stresses occur at the outside radius of the pipe wall.
In the present sensitivity analysis, the maximum tensile stress
condition that occurs at the outside radius of the pipe wall will be
investigated.
In the heat transfer analysis, only convection and conduction heat
transfer mechanisms are considered. Radiation heat transfer from
the outer surface of the pipe to the environment is neglected. The
convection heat transfer per unit length from the fluid in the pipe to
the inner surface of the pipe wall is:
Q = 2πRi hi (Tfluid – Ti) (9-8)
The conduction heat transfer per unit length through the pipe wall
can be written as:
Q = 2πk (Ti – To)/ln(Ro/Ri) (9-9)
The convection heat transfer per unit length from the outer surface
of the pipe wall to the environment is:
Q = 2πRo ho (To – Tenvironment) (9-10)
The heat is transferred in a series thermal circuit. Combining

Eqs. (9-8) through (9-10) demonstrates that heat transferred from
the fluid in the pipe to the environment is:
Q = (Tfluid – Tenvironment)/[(1/2πRihi) + ln(Ro/Ri)/2πk + (1/2πRoho)] (9-11)
where
Q = heat transfer through pipe wall in W/m
Tfluid = Mean temperature of fluid in the pipe in C
87
Tenvironment = Environment temperature outside the pipe in C

Ri = Pipe inside radius in meters
hi = Heat transfer coefficient between inner surface of pipe wall
and fluid in pipe in W/m2-C
Ro = Pipe outside radius in meters
ho = Heat transfer coefficient between outer surface of pipe wall
and environment in W/m2-C
k = Thermal conductivity of pipe wall material in W/m-C
Maximum tensile stress due to the temperature distribution given

by Eq. (9-4) occurs at the outer radius of the pipe. At the outer radius
Ro, the maximum axial stress is the same as in the circumferential
direction, σZ = σθ. Eqs. (9-6) and (9-7) produce the same results
at R = Ri and at R = Ro. Also, it is important to remember from the
boundary conditions that σR is zero at R = Ri and at R = Ro. Eq. (9-11)
can be combined with Eq. (9-6) to provide the thermal stress at R = Ro.
σθ at R=Ro = {α E Q/[4π(1 − ν)k]}{1 – [2Ri2/(Ro2 − Ri2)]ln(Ro/Ri)} (9-12)
Assumed nominal values for the independent variables for the

sensitivity analysis in Eqs. (9-11) and (9-12) are as follows:
Tfluid = 50ºC
Tenvironment = −50ºC
Ri = 0.2 m
Ro = 0.3 m
hi = 500 W/m2-C
ho = 100 W/m2-C
α = 2 × 10−5 m/m-C
E = 210000 MPa
ν = 0.3
k = 20 W/m-C
Thermal stress at the outside radius of the pipe wall is a linear

function of heat transferred from the fluid to the environment, or
(Tfluid − Tenvironment), as shown in Figure 9-1. The slope of this curve is
0.828 MPa/C.
88
120
Thermal Stress At Ro,
100
80
MPa
60
Tenvironment= –50°C
40
20
0
0 20 40 60 80
Temperature Of Fluid Flowing In Pipe, °C
Figure 9-1 Thermal stress at outside radius of pipe versus temperature of

fluid flowing in pipe
As the pipe outside radius, Ro, increases for a constant wall

thickness of 0.1 m, the heat transfer increases, and therefore the
thermal stress increases, as shown in Figure 9-2. Thermal stress
sensitivity to pipe outside radius decreases as the pipe gets larger.
The effects of pipe wall thickness on thermal stress are studied
for a constant outside radius pipe, where Ro = 0.3 m, in Figure 9-3.
Thermal stress increases as pipe wall thickness increases, reaching
a maximum at around 0.15 to 0.2 meters of wall thickness. Thermal
stress then starts to decrease as pipe wall thickness further increases,
due to decreasing heat transfer.
The inside and outside heat transfer coefficients, hi and ho, affect
thermal stress similarly, as shown in Figures 9-4 and 9-5. Thermal
stress sensitivity to changes in the heat transfer coefficient are
significant at smaller values of hi and ho.
±10%, the results shown in Table 9-1 are obtained. Maximum thermal
stress at Ro sensitivities to a ±10% change in the governing independent
variables are given in descending order of importance, and they are
applicable only around the nominal values assumed for this study.
89
86
Thermal Stress At Ro, MPa
84
82
80
78
76
0.2 0.25 0.3 0.35 0.4 0.45 0.5
Pipe Outside Radius, Ro, m
Figure 9-2 Thermal stress at Ro versus pipe outside radius for a constant
wall thickness of 0.1 meters
120
80
MPa
40
0
0 0.05 0.1 0.15 0.2 0.25 0.3
Pipe Wall Thickness, m
Figure 9-3 Thermal stress at Ro versus pipe wall thickness for a pipe outside
radius of 0.3 meters
90
100
80
60
MPa
40
20
0
0 200 400 600 800 1000
Heat Transfer Coefficient, hi, W/m2-C
Figure 9-4 Thermal stress at Ro versus the heat transfer coefficient at inside
surface of pipe wall
160
120
MPa
80
40
0
0 200 400 600 800 1000
Heat Transfer Coefficient, ho, W/m2-C
Figure 9-5 Thermal stress at Ro versus the heat transfer coefficient at

outside surface of pipe wall
91

variables to maximum thermal stress at Ro
Change In Change In
Thermal Stress Thermal Stress
At Ro For A 10% At Ro For A 10%
Pipe Wall
0.1 m +10.049% −13.011%
Thickness, Ro-Ri
α 2 × 10−5 m/m-C −10% +10%
E 210000 MPa −10% +10%
Tfluid-Tenvironment 100ºC −10% +10%
k 20 W/m-C +7.311% −6.378%
ho 100 W/m2-C −5.502% +5.002%
ν 0.3 −4.110% +4.478%
hi 500 W/m -C 2
−1.717% +1.450%
Ro 0.3 m +0.920% −1.248%
Thermal stress at the pipe outside radius is most sensitive to

changes in pipe wall thickness, the coefficient of thermal expansion,
modulus of elasticity, temperature potential between the fluid and
the environment, and thermal conductivity of pipe material. The
second tier of independent variables affecting thermal stress at
the pipe outside radius are the heat transfer coefficient between
outside surface of pipe wall and environment, and Poisson's ratio
of pipe material. Thermal stress at the pipe outside radius is least
affected by changes to the heat transfer coefficient between inside
surface of pipe wall and fluid in pipe, and outer radius of pipe. This
variable sensitivity order is applicable only around the nominal values
assumed for this case.
92
CHAPTER
HEAT
TRANSFER IN A
10
PIPE WITH UNIFORM
HEAT GENERATION IN
ITS WALLS
U nder steady-state conditions, constant thermophysical properties
and uniform heat generation in the walls, the one-dimensional
conduction heat transfer equation in radial direction of a pipe can be
written as (see Reference by Carslaw, H. S. and J. C. Jaeger [17]):
d2T/dR2 + (1/R)dT/dR + Q/k = 0 (10-1)
where the radial heat flux is positive in the negative radial direction
(towards the center of the pipe), Q is uniform heat generation in the
pipe walls in W/m3, and k is pipe wall thermal conductivity in W/m-C.
This differential Eq. (10-1) can be solved for the radial temperature
distribution in the pipe wall by specifying the pipe wall temperatures
with the inner and outer wall radii as boundary conditions.
T = Ti at R = Ri (10-2)
and
T = To at R = Ro (10-3)
93
The radial temperature distribution in the pipe wall which satisfies

boundary condition Eqs (10-2) and (10-3) is:
T(R) = To + (QRo2/4k)[1 − (R2/Ro2)] − {(To − Ti) + (QRo2/4k)

[1 − (Ri2/Ro2)]} [ln(Ro/R)/ln(Ro/Ri)] (10-4)
In this sensitivity study, convection heat transfer and conduction

heat transfer equivalence at the pipe wall surfaces at R = Ri and
at R = Ro will be used to determine the thermal energy generation
requirements in the pipe wall. Energy balances at the inside and
outside surfaces of the pipe wall provide:
(2πRiL) k dT/dR = (2πRiL) hi (Tfluid − Ti) at R = Ri (10-5)
and
(2πRoL) k dT/dR = (2πRoL) ho (To − Tenvironment) at R = Ro (10-6)
where hi is the convection heat transfer coefficient at inside surface

of the pipe in W/m2-C, Tfluid is mean fluid temperature in C in the pipe,
ho is the convection heat transfer coefficient at outside surface of the
pipe in W/m2-C, and Tenvironment is environmental temperature in C.
L is the length of the pipe in meters and it cancels out from the
energy balance equations.
Obtaining dT/dR at R = Ri from Eq. (10-4), the governing energy
balance Eq. (10-5) at R = Ri becomes:
hi (Tfluid − Ti) = 0.5 Q Ri

− [k/Ro ln(Ro/Ri)][(To − Ti) + (Q/4k)(Ro2 − Ri2)] (10-7)
Obtaining dT/dR at R = Ro from Eq. (10-4), the governing energy

balance Eq. (10-6) at R = Ro becomes:
ho (To − Tenvironment)
= 0.5 Q Ro − [k/Ro ln(Ro /Ri)][(To − Ti) + (Q/4k)(Ro2 − Ri2)] (10-8)
94
Heat Transfer In A Pipe With Uniform Heat Generation
The uniform thermal energy generation required per unit length

in the pipe wall, Q, is treated as the dependent variable, and its
sensitivity to other independent variables is analyzed for a specified
inner surface temperature, Ti. Eqs. (10-7) and (10-8) can be
combined to eliminate To and obtain a relationship for thermal energy
requirement as a function of eight independent variables.
Nominal values of independent variables for the present sensitivity
analysis are as follows:
Ri = 0.10 m
Ro = 0.11 m
k = 15 W/m-C
hi = 30 W/m2-C
Tfluid = 50°C
ho = 10 W/m2-C
Tenvironment = −10°C
Ti = 40°C
With these nominal values, the uniform thermal energy generation

required is 23,530 W/m3. Uniform thermal energy generation
requirements behave non-linearly with pipe inner radius variations,
as shown in Figure 10-1. The uniform thermal energy generation
requirement increases with decreasing pipe wall thickness. As the
wall thickness decreases, the conduction heat transfer resistance
decreases, and it gets difficult to keep the inner radius pipe wall
temperature at a constant 40°C.
Uniform thermal energy generation requirements increase as pipe
outer radius increases, as shown in Figure 10-2. Uniform thermal
energy generation requirements decrease with increasing pipe wall
thickness. As wall thickness increases, conduction heat transfer
resistance increases, and it becomes easier to keep the inner radius
pipe wall temperature at a constant 40°C.
Thermal energy requirement versus pipe wall thermal conductivity
is shown in Figure 10-3. Thermal energy generation requirement is
mostly a constant for thermal conductivities above 20 W/m-C, and it
begins rapidly to decrease as the wall material becomes an insulator.
95
250000
Energy Generation, W/m3
200000
150000
100000
50000
0
0.09 0.095 0.1 0.105 0.11
Pipe Inner Radius, m
Figure 10-1 Thermal energy generation requirement versus pipe inner

radius
200000
Energy Generated, W/m3
150000
100000
50000
0
0.1 0.105 0.11 0.115 0.12
Pipe Outer Radius, m
Figure 10-2 Thermal energy generation requirement versus pipe outer

radius
96
24000
23000
22000
21000
20000
0 20 40 60 80 100
Pipe Wall Thermal Conductivity, W/m-C
Figure 10-3 Thermal energy generation requirement versus pipe wall

thermal conductivity
Thermal energy generation requirement decreases as the

convection heat transfer coefficient at pipe inner surface increases,
as shown in Figure 10-4. The slope of the curve is −955.8 C/m, and
no thermal energy generation is required at hi = 54.6 W/m2-C. At
higher convection heat transfer coefficients at the pipe inner surface,
thermal energy has to be taken out of the pipe walls in order to keep
the pipe inner walls at 40°C.
Thermal energy generation requirement increases as the
convection heat transfer coefficient at pipe wall outside surface
increases, as shown in Figure 10-5. The slope of the curve is 5062.1 C/m,
and no thermal energy generation is required at ho = 5.27 W/m2-C,
namely in the natural convection region. At lower convection heat
transfer coefficients at the pipe wall outside surface, thermal energy
has to be taken out of the pipe walls in order to be able to keep the
pipe inner walls at 40°C.
The present case assumes a fully developed flow inside the
pipe. Thermal energy generation requirement versus mean fluid
temperature inside the pipe is shown in Figure 10-6. Thermal energy
97
60000
40000
20000
–20000
–40000
0 20 40 60 80 100
Convection Heat Transfer Coefficient At Pipe Wall
Inside Surface, W/m2-C
Figure 10-4 Thermal energy generation requirement versus convection heat

transfer coefficient at pipe wall inside surface
500000
400000
300000
200000
100000
0
–100000
0 20 40 60 80 100
Convection Heat Transfer Coefficient At Pipe Wall
Outside Surface, W/m2-C
Figure 10-5 Thermal energy generation requirement versus convection heat

transfer coefficient at pipe wall outside surface
98
150000
100000
50000
–50000
–100000
–150000
0 20 40 60 80 100
Mean Fluid Temperature Inside The Pipe, °C
Figure 10-6 Thermal energy generation requirement versus mean fluid

temperature inside the pipe
generation requirement decreases as the mean fluid temperature

increases. The slope of the curve is −2867.4 W/m3-C, and no thermal
energy generation is required at Tfluid = 58.2°C. For higher mean fluid
temperatures inside the pipe, thermal energy has to be taken out from
the pipe walls in order to be able to keep the pipe inner walls at 40°C.
Thermal energy generation requirement versus environmental
temperature is shown in Figure 10-7. Thermal energy generation
requirement increases as the environment temperature decreases. The
slope of the curve is −1044.1 W/m3-C, and no thermal energy generation
is required at Tenvironment = 12.5°C. Pipe walls have to be cooled above
the 12.5°C environmental temperature in order to keep Ti at 40°C.
Thermal energy generation requirement versus pipe inside surface
temperature, Ti, is shown in Figure 10-8. Thermal energy generation
requirement increases as the pipe inner surface temperature
requirement increases. The slope of the curve is 3911.5 W/m3-C,
and no thermal energy generation is required at Ti = 34°C.
When the nominal values of the independent variables given
above are varied ±10%, the results shown in Table 10-1 are obtained.
99
60000
40000
20000
–20000
–40000
–40 –20 0 20 40 60
Environment Temperature, °C
Figure 10-7 Thermal energy generation requirement versus environmental

temperature
250000
200000
150000
100000
50000
0
–50000
–100000
–150000
0 30 60 90
Pipe Inner Surface Temperature, °C
Figure 10-8 Thermal energy generation requirement versus pipe inner

surface temperature
100
Table 10-1 Effects of a ±10% change in nominal values of

variables to thermal energy generation requirement
Change In Change In
Required Required
Thermal Energy Thermal Energy
Generation For Generation For
Ti 40°C −66.49% +66.49%
Tfluid 50°C +60.93% −60.93%
+104.35%
Ro 0.11 m @ 5% decrease −45.38%
in nominal value
+84.65%
Ri 0.10 m −41.75% @ 5% increase
in nominal value
ho 10 W/m2-C −22.08% +22.06%
hi 2
30 W/m -C +12.19% −12.19%
Tenvironment −10°C +4.44% −4.44%
k 15 W/m-C −0.13% +0.11%
Thermal energy requirement sensitivities to a ±10% change in

the governing variables are given in descending order of importance,
and they are applicable only around the nominal values assumed for
this study.
Thermal energy generation required is most sensitive to pipe
inner surface temperature, mean fluid temperature inside the pipe,
pipe outer radius and pipe inner radius. Convection heat transfer
coefficients at the outside and inside surfaces of the pipe wall, and
environmental temperature, are the next independent variables in
order of sensitivity. Thermal energy generation requirement has
the lowest sensitivity to pipe wall thermal conductivity around the
assigned nominal values for this case.
101
CHAPTER
HEAT
TRANSFER
11
IN AN ACTIVE
INFRARED SENSOR
I n an active infrared sensor, the surface temperature is held

constant during the measurement process by providing controlled
energy to the sensor's surface. The sensor surface receives radiation
heat transfer energy from the surface of the object being measured.
The sensor also loses thermal energy to its environment. The
temperature at the sensor's surface can be analyzed by using
unsteady-state and one-dimensional heat transfer rate equations in
rectangular coordinates.
Unsteady-state heat transfer in an active infrared sensor is detailed
in the Reference by J. Fraden [3]. The energy balance of a sensor
element can be written as follows:
Change in internal energy of the sensor with respect to time =

Control energy supplied to regulate the surface temperature of
the sensor − Energy lost from the sensor to the environment by
conduction and by convection heat transfer + Net radiation heat
transferred from the object being measured to the sensor
Net radiation heat transfer between the object and the sensor
is assumed to occur between two gray bodies that are opaque to
103
radiation. If that is true, then the emissivity, ε, and reflectivity, ρ,

characteristics of a gray surface have the following relationship:
ε+ρ=1 (11-1)
In this analysis, it is assumed that the emissivity and reflectivity of

the object and the sensor surfaces are constants in the infrared region
of the electromagnetic radiation spectrum.
The radiation emitted by the object to the sensor can be written as:
Qradiation emitted by object = Aσεobject T4object (11-2)
Some of the irradiation reaching the surface of the sensor is

reflected due to the reflectivity of the sensor surface. The reflected
portion of the radiation emitted by the object can be written as:
Qradiation reflected by sensor = ρsensor (Aσεobject T4object) (11-3)
Eqs. (11-1) through (11-3) can be combined to get the net

radiation emitted by the object to the sensor:
Qnet radiation emitted by object = Aσεsensor εobject T4object (11-4)
In a similar fashion, the net radiation emitted by the sensor to the

object can be obtained:
Qnet radiation emitted by sensor = Aσεobject εsensor T4sensor (11-5)
The net radiation heat transfer between the object and the sensor
is determined by combining Eqs. (11-4) and (11-5):
Qnet radiation = Aσεsensor εobject (T4object − T4sensor) (11-6)
Control energy supplied to regulate the surface temperature of

the sensor is assumed to be in the form of I2R. Energy lost from
the sensor to the environment by conduction and convection heat
104
Heat Transfer In An Active Infrared Sensor
transfer can be written as a rate equation, where the two heat transfer
mechanisms act in series:
Qconduction + convection = (1/RT) (Tsensor − Tenvironment) (11-7)
Energy balance for the sensor can be written as a first order and
non-linear differential equation:
ρcpV(dTsensor/dθ) = I2R − (Tsensor − Tenvironment)/Rtotal

+ Aσεobjectεsensor (T4object − T4sensor) (11-8)
where
ρcpV is the sensor thermal capacitance in W-s/K

dTsensor/dθ is the time rate of change of sensor temperature in K/s
I2R is control power supplied to regulate the surface temperature of
the sensor in W
Tenvironment is the environment temperature in K
Rtotal is the total heat transfer resistance between the sensor and
the environment due to conduction and convection heat transfer
in K/W
A is the sensor area in m2
σ = 5.67×10-8 W/m2-K4 is the Stefan-Boltzmann constant
εobject is the surface emissivity of the object being measured
εsensor is the surface emissivity of the sensor
Tobject is the temperature of the object being measured in K
Tsensor is the temperature of the sensor surface in K
For the steady-state sensor temperature, Eq. (11-8) can be
rewritten by eliminating the left-hand side of the equation, which
means that the change in internal energy of the sensor with respect to
time becomes negligible. Tsensor is calculated from the following quartic
equation, by trial and error:
T4sensor + C1 Tsensor = C2 (11-9)
where
C1 = 1/(RT Aσεobjectεsensor) (11-10)
105
and
C2 = (I2R/Aσεobjectεsensor) + (Tenvironment/Rtotal Aσεobjectεsensor) + T4object

(11-11)
The transient solutions to Eq. (11-8) are obtained from the following
explicit finite difference equation, using small time intervals Δθ:
ρcpV(Tsensor @ (i+1) − Tsensor @ i /Δθ) = I2R − (Tsensor @ i − Tenvironment)/Rtotal

+ Aσεobjectεsensor (T4object − T4sensor @ i) (11-12)
For the present sensitivity analysis, nominal values of the

independent variables are assumed to be as follows:
ρcpV = 0.014 W-s/K

I2R = 0.1 W
Tenvironment = 20°C (293 K)
Rtotal = 100 K/W
A = 0.0001 m2 (a 1 cm × 1 cm sensor surface area)
εobject = 0.9
εsensor = 0.8
The sensitivities are analyzed for three different object temperatures,
namely 100°C (373 K), 500°C (773 K) and 1000°C (1273 K).
Sensor temperature as a function of time for three different
object temperatures is given in Figures 11-1, 11-2, and 11-3. As the
temperature of the object increases, the sensor time constant, or
the time the sensor reaches 63.2% of its steady-state temperature,
decreases. Also as the sensor emissivity increases, the sensor time
constant decreases.
As the object temperature approaches the environmental
temperature, the sensor temperature deviates from the object
temperature because fixed nominal values are used for the analysis
of this heat transfer problem. The control circuit of the sensor
has to respond and change control energy, I2R, in order to achieve
accurate results. The details of the control circuit are explained in the
Reference by J. Fraden [3].
106
200
150
Emissivity
Tsensor, C
Sensor=0.9
100 Emissivity
Sensor=0.8
50 Emissivity
Sensor=0.7
0
0 1 2 3 4
Time, s
Figure 11-1 Sensor temperature versus time for different sensor emissivities
and for an object at 1000°C
200
150 Emissivity
Tsensor, C
Sensor=0.9
100 Emissivity
Sensor=0.8
50 Emissivity
Sensor=0.7
0
0 1 2 3 4
Time, s
107
40
35
30 Emissivity
Tsensor, C
25 Sensor=0.9
20 Emissivity
15 Sensor=0.8
10 Emissivity
Sensor=0.7
5
0
0 1 2 3 4
Time, s

Figure 11-4 Sensor temperature versus time for different total heat transfer
resistances between the sensor and the environment, due to
conduction and convection heat transfer for an object at 1000°C
temperature
108
Table 11-1 Thermal time constant change due to a 10% change

in variables around the nominal values for an
object temperature of 1000°C
Thermal Time Thermal Time

Constant Change Constant Change
Sensor Thermal
0.014 W-s/K −20.39% +10.38%
Capacitance, ρcpV
1000°C
Tobject +16.73% −15.38%
(1273 K)
A 0.0001 m2 +8.12% −7.44%
εobject 0.9 +8.12% −7.44%
εsensor 0.8 +8.12% −7.44%
Rtotal 100 K/W −3.04% +2.31%
20°C
Tenvironment −0.15% +0.15%
(293 K)
I 2R 0.1 W +0.049% −0.049%
Another variable affecting the performance of the sensor is the

heat loss from the sensor to the environment due to conduction
and convection heat transfer. The sensitivity of the sensor response
to total heat transfer resistance between the sensor and the
environment, due to conduction and convection heat transfer, is given
in Figure 11-4 for an object at 1000°C temperature.
As the total thermal resistance between the sensor and the
environment increases, as seen in Figure 11-4, the sensor is better
insulated for losses due to conduction and convection heat transfer. In
such cases, the sensor temperature approaches the object temperature.
A ten percent variation in variables around the nominal values
given above produces the sensitivity results given in Tables 11-1 and
11-2, for thermal time constant and for steady-state temperature,
respectively, for an object that is at 1000°C.
109
Table 11-2 Steady-state sensor temperature change due to

a 10% change in variables around the nominal
values for an object temperature of 1000°C
Steady-State Steady-State
Sensor Sensor
Temperature Temperature
1000°C
Tobject −18.33% +18.22%
(1273 K)
Rtotal 100 K/W −4.13% +3.48%
2
A 0.0001 m −4.07% +3.44%
εobject 0.9 −4.07% +3.44%
εsensor 0.8 −4.07% +3.44%
20°C
(293 K)
I 2R 0.1 W −0.054% +0.054%
Sensor Thermal
0.014 W-s/K 0% 0%
Capacitance, ρcpV
A change in sensor thermal capacitance has the most significant

effect on the thermal time constant in this region of operation.
The thermal time constant experiencees non-linear sensitivity
behaviors from all the variables in the present region of operation,
with the exception of the environmental temperature and the I2R
power input to the sensor. The thermal time constant experiencees
similar magnitude sensitivity behaviors from the sensor area, sensor
emissivity, and object emissivity variables, as expected.
A change in the object temperature has the most significant effect
on the steady-state sensor temperature in this region of operation.
The next set of independent variables that most affect the sensor
temperature are the total heat transfer resistance between the sensor
110
Table 11-3 Thermal time constant change due to a 10%

change in variables around the nominal values for
an object temperature of 100°C
Thermal Time Thermal Time

Constant Constant
Change Due Change Due
To A 10% To A 10%
20°C
(293 K)
Sensor Thermal
0.014 W-s/K −10.33% +10.39%
Capacitance, ρcpV
RT 100 K/W −9.90% +9.83
2
A 0.0001 m −0.45% +0.45%
εobject 0.9 −0.45% +0.45%
εsensor 0.8 −0.45% +0.45%
2
IR 0.1 W −0.028% +0.028%
100°C
Tobject +0.023% −0.023%
(373 K)
and the environment, the sensor area, and surface emissivities of the
sensor and the object. Sensor thermal capacitance does not affect
the steady-state sensor temperature, as expected. The steady-state
sensor temperature experiences non-linear sensitivity behaviors
from all the variables except the environmental temperature and
the I2R power input to the sensor, in the present region of operation.
The steady-state sensor temperature experiences similar magnitude
sensitivity behaviors from the sensor area, sensor emissivity, and
object emissivity variables, as expected.
A similar sensitivity analysis is performed for a low object
temperature case, namely Tobject = 100°C, and the results are given in
Tables 11-3 and 11-4.
111
Table 11-4 Steady-state sensor temperature change due to a

10% change in variables around the nominal values
for an object temperature of 100°C
Sensor Sensor
A 10% Decrease A 10% Increase
Nominal In Nominal In Nominal
Variable Value Value Value
20°C
(293 K)
RT 100 K/W −3.99% +3.95%
2
IR 0.1 W −2.79% +2.79%
100°C
Tobject −2.27% +2.46%
(373 K)
A 0.0001 m2 +1.18% −1.19%
εobject 0.9 +1.18% −1.19%
εsensor 0.8 +1.18% −1.19%
Sensor Thermal
0.014 W-s/K 0% 0%
Capacitance, ρcpV
Temperature of the environment and thermal resistance to heat

loss from the sensor by conduction and convection to the environment
become the prominent variables in this low object-temperature
application. Both the thermal time constant and the steady-
state temperature of the sensor are very sensitive to variations in
the temperature of the environment and the total thermal resistance.
It is important to remember that these calculated sensitivities are
good only around the nominal values assumed for this case. Changing
these nominal values will change magnitudes and orders of these
sensitivities, resulting from the non-linear form of governing heat
transfer equations.
112
CHAPTER
COOLING
OF A CHIP
12
P ower dissipation in electronic chips is a challenging heat transfer

phenomenon, as the chips get smaller and smaller. Most chips
or chip sets use copper or aluminum heat sinks to enhance the heat
transferred out. These heat sinks are attached in a variety of ways to
the chips in order to minimize the thermal resistance between the chip
and the heat sink. In this analysis, a chip in the shape of a rectangular
box, 10 mm × 10 mm × 0.1 mm, is considered. The chip is attached to
its copper heat sink, also in the shape of a rectangular box, 10 mm ×
10 mm × 10 mm, by a thermally conductive epoxy of 10 μm thickness.
Transient heat transfer which occurs during the cooling of a chip
can generally be solved by using the same energy balance equation
as in Chapter 11, without the radiation heat transfer effects. The
temperature of a chip can be investigated by using unsteady-state
and one-dimensional heat transfer rate equations in rectangular
coordinates. Energy balance for a chip can be written as follows:
Change in internal energy of the chip with respect to time =

Power generated by the chip − Energy lost from the chip to
its heat sink and to the environment by conduction and by
convection heat transfer.
113
This energy balance can be written as a first order and linear

differential equation:
ρcpV(dTchip/dθ) = P – (Tchip – Tenvironment)/RTotal (12-1)
where
ρcpV is the chip thermal capacitance in W-s/C
dTchip/dθ is the time rate of change of chip temperature in C/s
P is the energy generated in the chip in W
Tenvironment is the environment temperature in C
RTotal is the total heat transfer resistance between the chip, its heat
sink and the environment due to conduction and convection heat
transfer in C/W
Tchip is the temperature of the chip in C.
Since the governing equation is linear in temperature, centigrade
dimension is used instead of Kelvin. Total heat transfer resistance
between the chip, its heat sink and the environment has two
parallel components. One component is convection heat transfer
resistance between the chip and the environment. The other is the
heat loss in series circuit from the chip through the adhesion layer
by conduction, from the adhesion layer through the heat sink by
conduction, and from the heat sink to the environment by convection.
The heat transfer resistance between the surface of the chip and the
environment can be written as:
Rchip to environment = 1/(hAchip for convection) (12-2)
The series circuit heat transfer resistance between the chip and
the environment going through the adhesive and the heat sink can be
written as:
Rchip through heat sink = [Ladhesive/(kadhesiveAadhesive for conduction)]

+ [Lheat sink/(kheat sinkAheat sink for conduction)]
+ [1/(hAheat sink for convection)] (12-3)
114
Cooling Of A Chip
The total heat transfer resistance from the chip is a parallel

combination of Eqs. (12-2) and (12-3):
(1/RTotal) = (1/Rchip to environment) + (1/Rchip through heat sink) (12-4)
The solution to the governing first order differential Eq. (12-1) can
be written as follows by using the initial condition of Tchip = Tenvironment:
Tchip = Tenvironment + PRTotal [1 – exp(−Θ/ρcpVRTotal)] (12-5)
where ρcpVRTotal is the thermal time constant for the chip, which is the
product of its thermal capacitance and its total thermal resistance.
Under steady-state conditions, the solution in Eq. (12-5) provides
the allowable chip power dissipation as follows:
P = (Tchip − Tenvironment)/RTotal (12-6)
For the present sensitivity analysis, the nominal values of the above
ρcpV = 0.0197 W-s/C
(assuming a silicon dioxide chip with 10 mm × 10 mm × 0.1 mm

dimensions)
h = 200 W/m2-C (convection heat transfer coefficient at the chip
and heat sink surfaces)
Tenvironment = 30ºC
Tchip = 90ºC
Achip for convection = 0.000104 m2
Ladhesive = 0.00001 m
Aadhesive for conduction = Aheat sink for conduction = 0.0001 m2
Lheat sink = 0.01 m
Aheat sink for convection = 0.0005 m2 (assuming a copper alloy heat sink
with 10 mm × 10 mm × 10 mm dimensions)
kadhesive = 10 W/m-C (assuming silver epoxy adhesive)
kheat sink = 400 W/m-C
115
200
150
RTotal, C/W
100
50
0
0 100 200 300 400
Figure 12-1 Total chip heat transfer resistance versus the convection heat
transfer coefficient
Total heat transfer resistance between the chip, its sink and the
environment is given as a function of the convection heat transfer
coefficient in Figure 12-1. As the convection mechanism goes into
forced convection, and especially forced convection in liquids,
35
Chip Power Disipation, W
30
25
with heat sink
20
15
without heat sink
10
5
0
0 200 400 600 800 1000
Figure 12-2 Chip power dissipation versus the convection heat transfer
coefficient with and without a heat sink
116
Cooling Of A Chip
total heat transfer resistance decreases dramatically. This decrease

in total heat transfer resistance allows the chip to be operated at
higher power dissipation, as shown in Figure 12-2. Figure 12-2 also
shows the effects of using a heat sink. The slope of the chip's power
dissipation with a heat sink versus the heat transfer coefficient,
(∂P/∂h), is 0.033 m2-C. The slope of the chip's power dissipation
without a heat sink versus the heat transfer coefficient is much lower,
∂P/∂h = 0.012 m2-C.
A ten-percent difference in variables around the nominal values
given above produce the sensitivity results given in Table 12-1 for
the power dissipation capability of the chip. The results are given in
descending order from the most sensitive variable to the least.
Table 12-1 Steady-state chip power dissipation capability

change due to a 10% change in variables around
the nominal values
Chip Power Chip Power
Dissipation Dissipation
Capability Capability
Tchip − Tenvironment 60ºC −10% +10%
h 200 W/m-C −9.81% +9.77%
2
Aheat sink for convection 0.0005 m −8.05% +8.01%
2
Achip for convection 0.000104 m −1.76% +1.76%
Aadhesive for conduction =
0.0001 m2 −0.23% +0.19%
Aheat sink for conduction
kheat sink 400 W/m-C −0.22% +0.18%
Lheat sink 0.01 m +0.20% −0.20%
kadhesive 10 W/m-C −0.009% +0.007%
Ladhesive 0.00001 m +0.008% −0.008%
117
3.5
Chip Time Constant, s
3
2.5
2
1.5
1
0.5
0
0 100 200 300 400
Figure 12-3 Chip thermal time constant versus the convection heat transfer
coefficient
Changes in chip temperature potential, convection heat transfer

coefficient and heat sink surface area for convection heat transfer
affect the chip power dissipation capability the most. Chip surface
area for convection heat transfer comes next in the order of
sensitivity. Changes in independent variables affecting the conduction
heat transfers, Aadhesive for conduction, Aheat sink for conduction, kheat sink Lheat sink,
kadhesive, and Ladhesive, contribute the least to sensitivities in chip power
dissipation. These sensitivity results shown in Table 12-1 are valid
only in the region of analysis for the governing independent variables.
The thermal time constant, ρcpVRTotal , is an important parameter in
chip design and testing. Thermal time constant, which is the product
of thermal capacitance and total thermal resistance, is analyzed as
a function of the convection heat transfer coefficient, and is given
in Figure 12-3. As the convection mechanism goes into forced
convection, and especially forced convection in liquids, the thermal
time constant decreases dramatically.
A ten-percent difference in variables around the nominal values
given above produces the sensitivity results given in Table 12-2 for the
118
Cooling Of A Chip
Table 12-2 Chip thermal time constant change due to a 10%

change in variables around the nominal values
Chip Thermal Chip Thermal

Time Constant Time Constant
h 200 W/m-C +10.88% −8.90%
ρcpV 0.0197 W-s/C −10% +10%
2
Aheat sink for convection 0.0005 m +8.76% −7.42%
Achip for convection 0.000104 m2 +1.79% −1.79%
Aadhesive for conduction =
0.0001 m2 +0.233% −0.190%
Aheat sink for conduction
kheat sink 400 W/m-C +0.223% −0.183%
Lheat sink 0.01 m −0.201% +0.201%
kadhesive 10 W/m-C +0.009% −0.007%
Ladhesive 0.00001 m −0.008% +0.008%
thermal time constant of the chip. The results are given in descending
order, from the most sensitive variable to the least.
Changes in the convection heat transfer coefficient, chip thermal
capacitance and heat sink surface area for convection heat transfer
affect the chip thermal time constant the most. Sensitivities of the
chip thermal time constant to the rest of the independent variables
follow the same order as in Table 12-1. Changes in independent
variables affecting the conduction heat transfer contribute the least to
chip thermal time constant sensitivities.
119
CHAPTER
COOLING OF
A CHIP UTILIZING
13
A HEAT SINK WITH
RECTANGULAR FINS
H eat transfer from a surface can be enhanced by using fins. This

chapter combines ideas from Chapters 5 and 12, and analyzes the
cooling of a chip utilizing a heat sink with rectangular fins.
The temperature of a chip can be analyzed using unsteady-state
and one-dimensional heat transfer rate equations in rectangular
coordinates. Energy balance for transient heat transfer during the
cooling of a chip can be written as follows:
Change in internal energy of the chip with respect to time =

Power generated by the chip − Energy lost from the chip to its
heat sink and to the environment by conduction and convection
heat transfer
This energy balance can be written as a first order and linear

differential equation:
ρcpV(dTchip /dθ) = P − (Tchip − Tenvironment)/RTotal (13-1)
121
where
ρcpV is the chip thermal capacitance in W-s/C
dTchip /dθ is the time rate of change of chip temperature in C/s
P is the heat generated in the chip in W
Tenvironment is the environment temperature in C
RTotal is the total heat transfer resistance between the chip, its
heat sink with fins and the environment, due to conduction and
convection heat transfer in C/W
Tchip is the temperature of the chip in C
Since the governing equation is linear in temperature, centigrade
dimension is used instead of Kelvin. Total heat transfer resistance
between the chip, its sink and the environment has two parallel paths.
One path is the direct convection heat transfer between the chip and
the environment. The other is the heat loss in series from the chip to
the adhesion layer by conduction, from the adhesion layer to the heat
sink body by conduction, and from the heat sink and its fins to the
environment by convection. The convection heat transfer from the
heat sink and its fins to the environment has two parallel components;
one is the convection heat transfer from the sink base surface to the
environment, and the other is the convection heat transfer from the
fins’ surfaces to the environment.
The convection heat transfer resistance between the heat sink’s
un-finned surfaces and the environment can be written as follows:
Rheat sink convection from un-finned surfaces = 1/(hAheat sink convection un-finned surfaces) (13-2)
The convection heat transfer between heat sink fin surfaces and
the environment can be written by combining Eqs. (5-5) and (5-6):
Qfin total = NηhAfin (Tfin base − Tenvironment) (13-3)
The convection heat transfer resistance between the heat sink’s fin
surfaces and the environment can be written as follows:
Rheat sink convection from fin surfaces = 1/(NηhAfin) (13-4)
122
Cooling Of A Chip Utilizing A Heat Sink With Rectangular Fins
N is the number of rectangular fins and η is the rectangular fin

heat transfer efficiency as defined in Chapter 5. “h” represents the
convection heat transfer coefficient at the chip surfaces, the heat
sink fin surfaces and the un-finned surfaces. It is assumed that the
convection heat transfer coefficient does not vary with a change in the
number of fins and fin lengths.
The series circuit heat transfer between the chip and the
environment that goes through the adhesive layer and the heat sink as
conduction mechanisms, and from the heat sink to the environment
as parallel convection mechanisms between the un-finned and finned
heat sink surfaces, can be written as follows:
Rchip through heat sink = [Ladhesive /(kadhesiveAadhesive for conduction)]

+ [Lheat sink /(kheat sinkAheat sink for conduction)]
+ [1/(NηhAfin + hAheat sink convection un-finned surfaces)] (13-5)
The heat transfer resistance between the surface of the chip and
the environment can be written as:
Rchip to environment = 1/(hAchip for convection) (13-6)
The total heat transfer resistance between the chip, its heat
sink with fins and the environment can be written by combining
Eqs. (13-5) and (13-6) in a parallel thermal circuit:
RTotal = {1/[(1/Rchip to environment) + (1/Rchip through heat sink)]} (13-7)
The solution to the governing first order differential Eq. (13-1) can
be written as follows, using the initial condition of Tchip = Tenvironment
and assuming temperature-independent thermophysical properties:
Tchip = Tenvironment + PRT [1 − exp(−Θ/ρcpVRTotal)] (13-8)
where ρcpVRTotal is the thermal time constant for the chip, which is the
product of its thermal capacitance and its total thermal resistance.
123
For the present sensitivity analysis, the nominal values of the above
ρcpV = 2.52 W-s/C
(assuming a silicon dioxide chip in the shape of a rectangular box
with 0.03 m × 0.03 m × 0.002 m dimensions)
h = 50 W/m2-C
Tenvironment = 30°C
Tchip = 90°C
Achip for convection = 0.00114 m2 which can be detailed as [0.03 × 0.03
+ 2 × (0.03 + 0.03) × 0.002]
Ladhesive = 0.002 m
Aadhesive for conduction = Aheat sink for conduction = 0.0009 m2 which can be
detailed as (0.03 × 0.03)
Lheat sink = 0.005 m
N = 7 (seven rectangular fins which are nominally 0.002 m thick
and 0.03 m in length)
Aheat sink convection un-finned surfaces = 0.00108 m2 which can be detailed as
[with 0.002 m thick seven fins, namely 0.03 × (0.03 − 7 × 0.002)
+ 2 × (0.03 + 0.03) × 0.005]
Afin = 0.00186 m2 which can be detailed as [for a 0.002 m thick, 0.03 m
long, and 0.03 m wide fin, namely 2 × 0.03 × (0.03 + 0.5 × 0.002)]
kadhesive = 100 W/m-C (assuming silver epoxy adhesive)
kheat sink = 300 W/m-C
It is assumed that the heat sink and the fins are machined from the
same block material, and the minimum machinable spacing between
the fins is 0.002 m. For seven fins, the resulting fin thickness is 0.002 m.
The heat transfer efficiency for a rectangular fin can be calculated
from Eq. (5-6) given in Chapter 5, η = tanh(mLc)/(mLc) where
m = [2h(w + t)/kwt ]1/2, Lc = L + 0.5t. For cases where the fin width,
w, is much greater than its thickness t, m becomes m = ( 2h /kt )1/2.
The rectangular fin efficiency for this nominal case is 0.95.
Using the above nominal values, and changing the fin length
and the number of fins on the heat sink, the chip heat dissipation
characteristics given in Figure 13-1 are obtained. As fin length
increases, so does chip power dissipation. The chip power dissipation
124
100
Chip Power Dissipation, W
80 Lfin=0.01 m
Lfin=0.03 m
60 Lfin=0.05 m
Lfin=0.07 m
40
Lfin=0.09 m
20 Lfin=0.15 m
0
0 2 4 6 8
Number Of Rectangular Fins
Figure 13-1 Chip power dissipation versus number of fins for different fin
lengths
asymptotes around Lfin = 0.15 meters for an eight-fin heat sink. With
increasing fin length, the efficiency of the fin decreases, even if the
convection heat transfer area increases. Adding the maximum amount
of fins (eight in this case due to machinability constraints), with
lengths of up to 0.15 m per fin, can enhance the chip heat dissipation
by as much as 12-fold, as compared to a non-finned heat sink.
The sensitivity of chip heat dissipation to fin length is given
in Figure 13-2. The sensitivity approaches zero as the fin length
increases. By increasing the fin length, the chip power dissipation
improves less and less, and choosing the right fin length becomes a
cost-benefit issue. For example, using Figure 13-1, if the desired chip
power dissipation is 40 W, choosing a four-fin heat sink design, with a
fin length of 0.05 m, will suffice.
Chip power dissipation versus the convection heat transfer
coefficient is shown in Figure 13-3. Cases with seven fins and with
no fins are compared for a fixed fin length of 0.03 meters. Chip
power dissipation varies linearly with the convection heat transfer
coefficient, since conduction heat transfer resistances through the
125
1400
1200
∂ P/∂ Lfin, W/m
1000 N=4
800 N=5
N=6
600
N=7
400 N=8
200
0
0.01 0.04 0.07 0.1 0.13
Rectangular Fin Length, m
Figure 13-2 Chip heat dissipation sensitivity to fin length versus fin length
160
Chip Power Dissipation, W
140
120
100 N=7
80 N=0
60
40
20
0
0 50 100 150 200
h, W/m2-C
Figure 13-3 Chip power dissipation versus convection heat transfer coefficient
with no fins and with seven fins for fin length = 0.03 m
126
40
Thermal Time Constant, s
30
20
10
0
0 50 100 150 200
Convection Heat transfer Coefficient, W/m2-C
Figure 13-4 Chip thermal time constant versus convection heat transfer
coefficient
adhesive and through the heat sink are negligible for this case
study. The slope of the chip power dissipation versus the convection
heat transfer coefficient line for a seven-fin design, (∂P/∂h), is
0.71 m2-C. The slope of the chip power dissipation versus the
convection heat transfer coefficient line for the finless design is
0.13 m2-C. It is apparent that the present nominal heat sink design,
with seven fins, enhances the chip power dissipation by over five-fold.
The thermal time constant for the chip is the product of thermal
capacitance and total thermal resistance, ρcpVRT, and it is given as a
function of the convection heat transfer coefficient in Figure 13-4.
The thermal time constant for the chip is a strong function of the
convection heat transfer coefficient at low values of forced convection
heat transfer regime; i.e., h < 50 W/m2-C.
In this finned heat sink design, the heat sink base and the fins
are assumed to be the same material. The thermal conductivity of
the heat sink and fin material also starts to affect the chip power
dissipation at lower values of kheat sink , i.e., kheat sink < 100 W/m-C. The
chip power dissipation versus heat sink and fin material thermal
conductivity is given in Figure 13-5 for a seven-fin design with a fin
127
45
Power Dissipation, W
40
35
30
25
20
0 100 200 300 400
Heat Sink & Fin Thermal Conductivity, W/m-C
Figure 13-5 Chip power dissipation versus heat sink and fin thermal
conductivity for Lfin = 0.03 m and N = 7
0.7
0.6
∂ P/∂ ksink, m-C
0.5
0.4
0.3
0.2
0.1
0
0 100 200 300 400
Heat Sink & Fin Thermal Conductivity, W/m-C
Figure 13-6 Chip power dissipation sensitivity to heat sink and fin thermal
conductivity versus heat sink and fin thermal conductivity for
Lfin = 0.03 m and N = 7
128
Table 13-1 Steady-state chip power dissipation capability

change due to a 10% change in variables around
the nominal values for Lfin = 0.03 m and N = 7
Chip Power Chip Power
Dissipation Dissipation
Capability Capability
Tchip − Tenvironment 60°C −10% +10%
h 50 W/m-C −9.41% +9.29%
kheat sink 300 W/m-C −0.574% +0.476%
Lheat sink 0.005 m −0.289% +0.287%
kadhesive 100 W/m-C −0.148% +0.122%
Ladhesive 0.002 m +0.134% −0.134%
length of 0.03 meters. The sensitivity of chip power dissipation to

heat sink and fin thermal conductivity is shown in Figure 13-6, again
for a seven-fin design with a fin length of 0.03 meters. The sensitivity
increases fast for low values of heat sink and fin material thermal
conductivity.
A ten-percent difference in independent variables around the
nominal values given above produces the sensitivity results given
in Table 13-1, shown in descending order for the power dissipation
capability of the chip. The chip’s physical dimensions are assumed
to be constants. This sensitivity analysis is performed for a seven-fin
heat sink design with a fin length of 0.03 meters.
Changes in chip-to-environment temperature potential and the
convection heat transfer coefficient affect the chip power dissipation
capability the most. Changes in conduction heat transfer variables
about the nominal values chosen for this analysis have the least affect
on chip power dissipation capability.
129
CHAPTER
HEAT
TRANSFER
14
ANALYSIS FOR
COOKING IN A POT
F ood cooking in a pot placed on a gas burner can be a very

complicated heat transfer problem, if multi-dimensional transient
heat transfer for the pot and for the food in the pot are considered.
The heat transfer mechanisms can get challenging, if boiling heat
transfer regime is treated. The best way to approach such a heat
transfer problem is to make simplifying assumptions and create
a simple model, then verify the results of the model by reliable
experiments. In the present heat transfer model, to simulate cooking
in a pot, an unsteady-state and a coupled one-dimensional heat
transfer analysis will be utilized with the following assumptions.
A cylindrical pot gets its cooking energy from a gas burner at a
constant rate. The pot is assumed to be made out of copper and
have a uniform temperature throughout, namely negligible internal
conduction resistance. The pot conduction heat transfer resistance,
Lc/k, is small as compared to the pot surface's convection heat
transfer resistance, 1/h, or Biot Number = hLc/k < 0.1, where Lc is
a characteristic length for the pot, which is the volume of copper
divided by the pot's outer surface area.
131
The pot loses energy to food from its bottom and its side by natural
convection heat transfer. The natural convection heat transfer
mechanism between the pot and the food is analyzed up to a food
temperature of 105°C at sea level atmospheric conditions; boiling heat
transfer regime is not considered. The pot also loses energy to the
environment from its bottom and its side by natural convection and
radiation.
The food receives energy from the bottom and the side of the
pot by natural convection. Temperature gradients in the food are
neglected, and the natural convection heat transfer coefficient
between the pot and the food is calculated using an average food
temperature. Conduction and radiation heat transfer mechanisms
between the pot and the food are neglected. Also, the natural
convection and radiation heat transfers from the top of the food to
the environment are neglected. Neglecting these secondary heat
transfer mechanisms introduces an initial error of about 7% in the net
heat transferred to the food, but this error diminishes fast to zero as
cooking time increases.
Food is assumed to have the same thermophysical properties as
water. Temperature-dependent variations for all of the thermophysical
properties are considered. The thermophysical properties are
calculated at a film temperature, which is the average of the surface
and the medium temperatures.
Energy balance for the food can be written as follows:
Change in internal energy of the food with respect to time =

Energy input by convection from the bottom of the pot to the
food + Energy input by convection from the sides of the pot to
the food
Energy balance for the pot can be written as follows:
Change in internal energy of the pot with respect to time =

Energy input from the heater to the bottom of the pot − Energy
output by convection from the bottom of the pot to the food −
Energy output by convection from the sides of the pot to
132
Heat Transfer Analysis For Cooking In A Pot
the food − Energy output by convection from the bottom of

the pot to the environment − Energy output by convection from
the sides of the pot to the environment − Energy output by
radiation from the bottom and sides of the pot to the environment
The governing coupled and transient first-order differential

equations for the food and the pot are detailed as follows:
(ρcpV)food(dTfood/dΘ)
= hpot bottom to food natural convection Apot bottom (Tpot − Tfood)
+ hpot sides to food natural convection Apot sides (Tpot − Tfood) (14-1)
(ρcpV)pot(dTpot/dΘ)
= Qin − hpot bottom to food natural convection Apot bottom (Tpot − Tfood)
− hpot sides to food natural convection Apot sides (Tpot − Tfood)
− hpot bottom to environment natural convection Apot bottom (Tpot − Tenvironment)
− hpot sides to environment natural convection Apot sides (Tpot − Tenvironment)
− hpot to environment radiation (Apot bottom + Apot sides) (Tpot − Tenvironment) (14-2)
All the natural convection heat transfer coefficients can be

obtained from empirical relationships in literature for the appropriate
geometries; for example, for natural convection heat transfer from
heated vertical plates and horizontal plates, see References [6] and
[10]. The natural convection heat transfer coefficient from the bottom
of the pot to the food is:
(hL/k) = 0.5 RaL0.25 for 104 < RaL < 107 (14-3)
(hL/k) = 0.15 RaL0.33 for 107 < RaL < 1011 (14-4)
The natural convection heat transfer coefficients from the sides

of the pot to the food and the environment are obtained from the
empirical relationship:
(hL/k) = {0.825 + 0.387RaL /[1 + (0.492/Pr)9/16]8/27}2 (14-5)
133
The natural convection heat transfer coefficient from the bottom

of the pot to the environment is obtained from the empirical
relationship:
(hL/k) = 0.27 RaL0.25 for 105 < RaL < 1010 (14-6)
where the Rayleigh number, RaL, is the product of the Grashof

number and the Prandtl number, RaL = gβ(Tpot − Tfood)L3/(να). The
Grashof number represents the ratio of buoyancy forces to viscous
forces in a natural convection heat transfer system. The Prandtl
number is the ratio of momentum diffusivity to thermal diffusivity,
namely Pr = ν/α. Tfood in the RaL equation becomes Tenvironment when
the natural convection heat transfer coefficient considered is between
the pot and the environment.
Thermophysical properties, variables and constants in the above
Eqs. (14-1) to (14-6) are defined as:
(ρcpV) = thermal capacitance, namely product of density, specific
heat at constant pressure and volume, in W-hr/C
T = Temperature in C or in K when radiation calculations are used
Θ = Time in hr
h = heat transfer coefficient for natural convection or radiation heat
transfer mechanisms in W/m2-C
A = Heat transfer area in m2
Qin = Heat input from the gas burner to the pot in W
k = Thermal conductivity in W/m-C
L = Characteristic length for the surface area, where natural
convection heat transfer occurs and is defined as the ratio of heat
transfer surface area to its perimeter in m
g = Gravitational acceleration in m/s2
β = Volumetric thermal expansion coefficient in 1/K
ν = Kinematic viscosity in m2/s
α = (k/ρcp) thermal diffusivity in m2/s
Governing Eqs. (14-1) and (14-2) are solved using the explicit
finite difference method with a ten-second time interval.
Temperature-dependent thermophysical properties of water and
134
air are obtained from References [6] and [10]. The radiation heat
transfer coefficient is obtained from the following equation:
hradiation = σε(T4pot − T4environment)/(Tpot − Tenvironment) (14-7)
where σ is the Stefan-Boltzmann constant, 5.67 × 10−8 W/m2-K4, and

ε is the emissivity of the copper pot outer surface. The heat input
from the gas burner to the pot is 750 W. Dimensions of the pot and
the food height are as follows:
Pot inside diameter = 0.127 m

Pot height = 0.076 m
Pot side thickness = 0.0018 m
Pot bottom thickness = 0.0063 m
Food height in the pot = 0.051 m
Food and pot temperatures are given as functions of time in

Figure 14-1. Food temperature is calculated up to 105°C, where
the applicability of the natural convection heat transfer coefficient
equations end and the boiling heat transfer regime starts. In addition,
water heating experiments are performed in a copper pot with the
above geometrical dimensions to verify the present model. The
results of these experiments are also given in Figure 14-1. The results
of the present simple heat transfer model match the experimental
results closely. Deviations between the model and the experiments
are seen at initial cooking times; i.e., less than one minute, and as the
water temperature approaches the boiling temperature. The present
heat transfer model can be improved by including the temperature
gradients in the water, by using a more appropriate empirical
relationship for the natural convection heat transfer coefficient as the
water temperature approaches its boiling temperature, and by taking
smaller time increments during calculations.
The natural convection heat transfer coefficients from the
bottom of the pot and the sides of the pot to the food are given in
Figure 14-2. The heat transfer coefficients increase as the pot and
the food temperatures increase, mainly due to the temperature
dependent properties of water. As the food temperature increases,
135
120
Food and Pot Temperature, C
100
80 Tfood, C
Tpot, C
60 Experiment, C
40
20
0 2 4 6
Time, min
Figure 14-1 Food and pot modeling temperatures, and food experiment
temperatures, as functions of time
Transfer Coefficient, W/m2-C
1600
Natural Convection Heat
1200
hpot bottom to
800 food, W/m2-C
hpot sides to food,
400 W/m2-C
0
0 2 4 6
Time, min
Figure 14-2 Pot bottom to food and pot sides to food natural convection
heat transfer coefficients
136
Table 14-1 Changes in time for the average food temperature

in the pot to reach 105°C due to a 10% change in
variables around the nominal values
Change in Time Change in Time

for the Food in for the Food in
the Pot to Reach the Pot to Reach
105°C Due To 105°C Due To
Qin 750 W +11.16% −9.09%
Thermal Varies
Capacity with time, −8.36% +8.36%
of Food W-hr/C
Apot bottom 0.01267 m2 −2.71% +2.84%
Thermal
Capacity 0.1286 W-hr/C −1.65% +1.65%
of Pot
Apot sides 0.02027 m2 −0.84% +0.84%
Heat Loss Varies
from Pot to with −0.30% +0.30%
Environment time, W
hpot bottom to food Varies with
+0.17% −0.15%
natural convection time, W/m2-C
hpot sides to food Varies with
+0.13% −0.12%
natural convection time, W/m2-C
its volumetric thermal expansion coefficient increases and its

kinematic viscosity decreases. The natural convection heat transfer
between the bottom of the pot and the food is more dominant than
the one between the side of the pot and the food, mainly due to the
difference in characteristic lengths that affect the Rayleigh number.
A ten-percent variation around the nominal values of independent
variables given above produces the sensitivity results given in
Table 14-1, for the time that the average food temperature in the pot
137
reaches 105°C. The sensitivity results are given in descending order,

and they are applicable only in the region of assigned nominal values
due to their non-linear effects.
The time it takes for the food to reach 105°C is most sensitive
to heat input from the gas burner into the pot, and to the thermal
capacity of the food. The sensitivity order continues with the pot
geometry and the pot thermal capacity. The time it takes for the food
to reach 105°C is an order of magnitude less sensitive to changes
in energy loss to the environment, and to changes in the natural
convection heat transfer coefficients between the pot surfaces and
the food.
If all of the 750 W heat input from the gas burner goes into the
food, the time for the food to reach 105°C is 5 minutes. All the
heat losses from the pot to the environment, as well as the thermal
capacity of the pot, lower the efficiency of the food cooking system.
It takes 6.23 minutes for the food to reach 105°C under nominal
conditions.
138
CHAPTER
HEAT
TRANSFER AND
15
INSULATING
A WATER PIPE
P rotecting water pipes from freezing requires transient heat

transfer calculations and extensive knowledge of the thermo
physical properties of insulating materials. Under severe environmental
conditions, insulating the water pipe alone might not be sufficient to
prevent it from freezing. Other protections, such as running water
in the pipe or heating tape around the pipe, might be necessary. In
this heat transfer study, a long pipe with an insulation layer wrapped
around it is considered. The interface between the pipe’s outer
surface and the insulation inner surface is assumed to be in good
contact, with no air gaps or other imperfections to cause any
contact resistance. The pipe is filled with stationary water, and is
only exposed to the environment where there is convection heat
transfer between the outer surface of the insulation and the
environment. The time that it takes the water in the pipe to start
freezing is analyzed, and its sensitivities to governing independent
variables are investigated.
Solutions to the one-dimensional form of the transient heat
conduction equation in cylindrical coordinates, for a long cylinder
and with appropriate boundary conditions, govern this heat transfer
139
problem and are given in the Reference by P. J. Schneider [16]

as follows:
∂2T/∂r2 + (1/r)∂T/∂r = (1/α)∂T/∂θ (15-1)
where T is the temperature in centigrade along the radial position, r,

of the insulated pipe in meters, θ is time in seconds, and α is thermal
diffusivity of pipe or insulation material, k/ρcp, in meter squared
per second.
In literature, Eq. (15-1) has closed form solutions as a series of
Bessel functions for a variety of boundary conditions (see
Reference [1]). In the present analysis, Eq. [15-1] is solved using the
explicit finite difference method, (see References [6] and [16]), with
the appropriate boundary conditions. Eq. (15-1) can be written in a
finite difference form, and a radial node temperature at a new time
step is determined by known temperatures at the surrounding nodes
from previous time step calculations, as follows:
Trθ+1 =[1 − (2+Δr/r)Fo] Trθ + (1+Δr/r)Fo Tr+1θ + Fo Tr−1θ (15-2)
where Fo is the Fourier number defined as αΔθ/(Δr)2, Δθ is the time

interval in seconds used in calculations, and Δr is the radial node
interval in meters used in calculations. In order to achieve stable
step-by-step solutions in time to Eq. (15-2), the coefficient of Trθ
has to be positive (see References [6] and [16]). In the present
calculations, appropriate time and spatial intervals are used so that
the Fourier number is always less than 1/(2 + Δr/r).
The initial condition and the boundary conditions to solve
Eq. (15-2) are as follows—initial condition, θ = 0, for the finite
difference calculations is:
Tr = Twater (15-3)
Boundary conditions for the outer pipe-insulation interface and the

pipe inner radial nodes are obtained from applying an energy balance
to a control volume around that particular node. At the inner-most
node of the pipe, it is assumed that the water and half the pipe wall
140
Heat Transfer And Insulating A Water Pipe
thickness have the same temperature. This assumption is good for

small-diameter pipes and for pipe wall materials with high thermal
conductivities. The pipe inner node energy balance per meter of pipe
is given below. The inner node starts at the center of the pipe and
goes out half way in wall thickness of the copper pipe:
Energy lost due to conduction out from the radial element =

Change in stored thermal energy in nodal volume. [This nodal
volume starts at r = 0 and goes to 0.5(r2 − r1), which includes the
water in the pipe and half the pipe wall thickness.]
or
kcopper 2π (Tr1θ − Tr2θ)/ln(r2/r1)

= {ρwatercp waterπr12 + ρcoppercp copper2π[r1 + 0.25(r2 − r1)] 0.5(r2 − r1)}
(Tr1θ+1 − Tr1θ)/Δθ (15-4)
where
r1 = inner radius of copper pipe in meters

r2 = outer radius of copper pipe in meters
kcopper = thermal conductivity of copper pipe in W/m-C
ρcopper = density of copper pipe in kg/m3
cp copper = specific heat of copper pipe in J/kg-C
ρwater = density of water in kg/m3
cp water = specific heat of water in J/kg-C
It should be noted that thermophysical properties are assumed to

be constant during the present calculations.
The interface node energy balance per meter of pipe for the node
at the interface between pipe outer surface and insulation layer inner
surface is:
Energy gained due to conduction into radial element − Energy

lost due to conduction out from the radial element = Change in
stored thermal energy in nodal volume. [This nodal volume starts
at the middle of the pipe wall thickness, 0.5(r2 − r1), and goes
out to the middle of the first insulation layer increment Δr,
(r2 + 0.5Δr).]
141
or
kcopper 2π (Tr1θ − Tr2θ)/ln(r2/r1) − kinsulation 2π (Tr2θ − Tr2+Δrθ)/ln(r2 + Δr/r2)

= {ρcoppercp copper2π[r2 − 0.25(r2 − r1)] 0.5(r2 − r1)
+ ρinsulationcp insulation2π[r2 + 0.25(r2 + Δr)] 0.5Δr} (Tr2θ+1 − Tr2θ)/Δθ
(15-5)
where
kinsulation = thermal conductivity of insulation material in W/m-C

ρinsulation = density of insulation material in kg/m3
cp insulation = specific heat of insulation material in J/kg-C
The outer node energy balance per meter of pipe—the node at

the interface between the insulation layer outer surface and the
environment is:
Energy gained due to conduction into the radial element −

Energy lost to environment due to convection out from the radial
element = Change in stored thermal energy in nodal volume
[This nodal volume starts at −0.5Δr from the outer radius of the
insulated pipe and goes out to the outer radius, r3.]
or
kinsulation 2π (Tr3-Δrθ − Tr3θ)/ln(r3/r3 − Δr) − ho 2πr3(Tr3θ − Tenvironment)

= ρinsulationcp insulation2πr3 0.5Δr (Tr3θ+1 − Tr3θ)/Δθ (15-6)
where
ho = Convection heat transfer coefficient between the outer surface

of insulation and the environment in W/m2-C
r3 = Outer radius of insulated pipe in meters
Nominal values of independent variables used in the present
sensitivity analysis are as follows:
r1 = 0.023 m
r2 = 0.025 m
r3 = 0.035 m
142
Twater = 12°C
Tenvironment = −20°C
ρwater = 1000 kg/m3
cp water = 4200 J/kg-C
kcopper = 50 W/m-C (Note: water pipe is assumed to be copper alloy)
ρcopper = 8800 kg/m3
cp copper = 400 J/kg-C
kinsulation = 0.04 W/m-C
ρinsulation = 26 kg/m3
cp insulation = 835 J/kg-C
ho = 10 W/m2−C
Δrcopper pipe = 0.001 m
Δrinsulation 0.001 m
Δθ = 0.2 s
The dependent variable in the calculations is the time it takes

the water in the pipe to reach 0°C. After water reaches 0°C at the
inner radius of the copper pipe, it will take extra time for it to freeze
completely. This freezing time can be calculated from the energy
required to compensate for the total latent heat of fusion, Hfg,
required for the water in the pipe:
Hfg water = 334800 W-s/kg
After the temperature of the inner node reaches 0°C, the time
required for complete freezing is as follows:
θfor complete freezing

= ρwaterπr12 Hfg /[kcopper2π(0 − Tr2θ=time @ r1=0°C)/ln(r2/r1)] (15-7)
From Eq. (15-7), the flow rate required to prevent complete

freezing also can be calculated:
Flow rate to prevent complete freezing in kg/s

= [kcopper2π(0 − Tr2θ=T @ r1=0°C)/ln(r2/r1)] L /Hfg (15-8)
143
10
Time To Start Freezing, hr
8
k insulation=0.02
6 W/m-C
k insulation=0.04
4 W/m-C
k insulation=0.06
2 W/m-C
0
0 10 20 30 40
Insulation Thickness, mm
Figure 15-1 Time for water to start freezing versus insulation thickness for
three different insulation thermal conductivities
12
Time For Water To Start
10
Freezing, hr
8
6
4
2
0
–50 –40 –30 –20 –10 0
Environment Temperature, C
Figure 15-2 Time for water to start freezing versus environmental

temperature
144

variables to time for water to start freezing
Change In Time Change In Time

For Water To For Water To
Start Freezing Start Freezing
For A 10% For A 10%
ρwater 1000 kg/m3 −8.645% +8.645%
cp water 4200 J/kg-C −8.645% +8.645%
kinsulation 0.04 W/m-C +8.282% −6.774%
Twater 12°C −8.115% +7.816%
Tenvironment −20°C −7.363% +8.669%
Insulation
0.01 m −5.766% +5.640%
thickness, r3 − r2
ho 10 W/m2-C +2.829% −2.314%
ρcopper 8800 kg/m3 −1.315% +1.315%
cp copper 400 J/kg-C −1.315% +1.315%
Copper pipe
0.002 m +0.243% −0.240%
thickness, r2 − r1
ρinsulation 26 kg/m3 −0.041% +0.041%
cp insulation 835 J/kg-C −0.041% +0.041%
kcopper 50 W/m-C +0.003% −0.003%
where Tr2 is the temperature at r2 at the time when r1 reaches 0°C and
L is the length of the insulated pipe in meters.
Under the above given conditions, water reaches 0°C in 3.31 hours
and complete freezing will occur in an additional 14.94 hours.
Time for the water to reach 0°C is calculated for different insulation
thicknesses and for three different insulation thermal conductivities.
The results are given in Figure 15-1. At lower insulation thermal
conductivities, time for the water to reach 0°C becomes more
sensitive to insulation thickness.
145
Another significant variable affecting the time for water to start

freezing is the environmental temperature. Figure 15-2 shows
the effects of environmental temperature on time for water to
start freezing under the nominal conditions. Time sensitivity to
environmental temperature increases rapidly above −10°C.
are varied ±10%, the results shown in Table 15-1 are obtained.
Sensitivities to a ±10% change in the governing independent variables
of time for water to start freezing are given in descending order of
importance, and these results are applicable only around the nominal
values assumed for this study.
The changes in variables most affecting the time for water to start
freezing are water thermal capacitance, insulation material thermal
conductivity, and water and environmental temperatures. Insulation
thickness, the convection heat transfer coefficient at the outer
surface of insulation layer, and the copper pipe thermal capacitance
properties compose the middle of the pack in order of sensitivity. The
time for water to start freezing is least sensitive to changes in copper
pipe thickness, insulation material thermal capacitance, and copper
pipe thermal conductivity.
146
CHAPTER
QUENCHING
OF STEEL
16
BALLS IN AIR
FLOW
Q uenching, or rapid and controlled cooling, has been used for

centuries to harden steel and increase the toughness of metal
alloys. Quenching poses a difficult heat transfer problem, as it deals
with high-temperature materials being cooled in a controlled way, in
mediums such as air, water, oil, liquid nitrogen, or some other
special quenching fluid. The material properties and the fluid
properties change fast, and during the quenching process, heat
transfer between the material body and the medium can go through
several regimes. The medium surrounding the hot body can have
film boiling, transition from film boiling to nucleate boiling, nucleate
boiling, and natural convection as quenching time progresses.
In this chapter, the quenching of a small spherical ball made out of
steel is analyzed in a forced convection air flow. The conduction heat
transfer resistance within the ball is assumed to be much less than
the heat transfer resistance between the surface of the ball and the
quenching medium—the Biot number that is defined in Eq. (16-2) is
less than 0.1. This assumption allows the use of the following energy
balance in the present unsteady-state heat transfer application (see
Reference by J. P. Holman [5]).
147
Rate of decrease of internal energy in spherical steel ball =

Heat lost to surrounding air from the surface of the spherical ball
Since the air properties, convection heat transfer coefficient, and

radiation heat transfer coefficient change by time, the finite difference
form of the energy balance can be written as follows:
(ρVcp) (Tsphere (i+1) − Tsphere (i))/dt = −hi A (Tsphere (i) − Tair) (16-1)
Tsphere (i+1) is temperature of the ball at time (i + 1) in C,

Tsphere (i) is temperature of the ball at time (i) in C,
Tair is temperature of air away from the ball in C,
h is the total heat transfer coefficient in W/m2-C, between the
surface of the steel ball and air, which is the sum of the convection
and radiation heat transfer mechanisms, and varies with time. It is
always calculated at the i’th time in order to be able to determine
the temperature at i + 1’th time.
A is surface area of the steel ball in m2,
dt is time interval used in calculations in seconds,
ρ is density of the steel ball in kg/m3,
cp is specific heat of the steel ball in J/kg-C, and
V is volume of the steel ball in m3.
ρcpV/hA is defined as the thermal response time for the quenching
process in seconds.
Tsphere initial is initial temperature of the ball in C when the quenching
process starts.
The Biot number is defined as
Biot number = hmax R/ksteel (16-2)
where hmax is the maximum heat transfer coefficient, the sum

of convection and radiation heat transfer mechanisms between
the surface of the steel ball and the quenching medium encountered
during the quenching process, R is the radius of the sphere ball
in meters, and ksteel is thermal conductivity of the steel ball in
W/m-C.
148
Quenching of Steel Balls in Air Flow
The convection heat transfer coefficient is calculated from an

empirical relationship, given in literature, for forced convection heat
transfer from the surface of a sphere to air (see Reference [5]).
hforced convection = 0.37 (kair /D) (UmD/νair)0.6 for 17 < ReD < 70000 (16-3)
where
kair is the temperature-dependent thermal conductivity of air in

W/m-C,
D is diameter of sphere ball in m,
Um is mean velocity of air away from the ball in m/s,
νair is temperature dependent kinematic viscosity of air in
m2/s, and
ReD is the Reynolds number defined as (UmD/νair).
The radiation heat transfer coefficient is calculated from
hradiation = ε σ (T4sphere − T4air)/(Tsphere − Tair) (16-4)
where
ε is emissivity of steel ball surface and

σ is Stefan-Boltzmann constant, namely 5.67 × 10−8 W/m2-K4.
During the calculations, instantaneous thermophysical properties

of air, kair and νair, are calculated at mean temperatures of Tsphere and
Tair, namely at (Tsphere + Tair)/2.
Steel ball properties ρ, cp, ε, and D are assumed to be constants.
The nominal values of governing independent variables for the
present sensitivity calculations are as follows:
D = 0.01 m,
ρ = 7800 kg/m3
cp = 430 J/kg-C
ksteel = 40 W/m-C
ε = 0.2
Um = 5 m/s
149
Tair = 27°C
Tsphere initial = 800°C
dt = 1 s
The present case is analyzed below using different time intervals.

As the time interval is reduced to around one second, the resulting
temperatures-time profiles are overlaid, and therefore dt = 1 second is
used for the results presented in the present sensitivity calculations. For
the nominal case, the Biot number is 0.022 for a hmax of 174 W/m2-C. Steel
ball temperature versus time is given in Figure 16-1. For the nominal
case, the steel ball reaches 36.8% of (Tsphere initial − Tair), the thermal
response time for the quenching process in 57.0 s, and it reaches 50°C
in 153.4 s.
Initially radiation heat transfer is 18% of the total heat transfer.
Radiation heat transfer effect diminishes fast and convection heat
transfer dominates the cooling process, as shown in Figure 16-2.
The effect of mean air velocity is analyzed without violating
the convection heat transfer coefficient’s range of application in
Eq. (16-3), 17 < ReD < 70000, and the Biot number being less than
0.1. The time for the steel ball to reach 50°C and the time for it to
reach 36.8% of (Tsphere initial − Tair) are shown as a function of mean
800
Temperature, °C
600
400
200
0
0 50 100 150 200
Time, s
Figure 16-1 Steel ball temperature versus time
150
200
150
h, W/m2-C
h convection
100 h radiation
50
0
0 50 100 150 200 250
Time, s
Figure 16-2 Convection and radiation heat transfer coefficients versus time
air velocity in Figure 16-3. Changes in mean air velocity below 5 m/s
affect the quenching process time significantly.
Figure 16-4 shows that for the present analyses, the internal
conduction resistance for the steel ball is much less than the
400
Time To
300 Reach
50°C, s
Time, s
Time To
200
Reach
36.8% of
100 (Tsphere
initial-Tair), s
0
0 5 10 15 20
Mean Air Velocity, m/s
Figure 16-3 Time for the steel ball to reach 50°C and to reach 36.8% of
(Tsphere initial − Tair) versus mean air velocity
151
0.06
Biot Number
0.04
0.02
0
0 5 10 15 20
Mean Air Velocity, m/s
Figure 16-4 Biot number versus mean air velocity

independent variables to time for the steel ball
temperature to reach 50°C, t50°C
t50°C t50°C
Change For Change For
D 0.01 m −14.25% +14.79%
varies with
kair +11.07% −9.18%
temperature
ρ 7800 kg/m3 −10.51% +10.44%
cp 430 J/kg-C −10.51% +10.44%
Um 5 m/s +6.53% −5.66%
varies with
νair −6.17% +5.90%
temperature
Tsphere initial − Tair 773°C −3.70% +3.02%
ε 0.2 +0.47% −0.47%
152
convection heat transfer resistance between the surface of the

steel ball and the environment, for the range of mean air velocities
considered in the present calculations, i.e., Biot number < 0.1.
Time for the steel ball to reach 50°C, t50°C and sensitivities to a
±10% change in the governing independent variables are given in
descending order of importance, and they are applicable only around
the nominal values assumed for this study. Cooling time to 50°C
is most sensitive to independent variables such as ball diameter,
density, and specific heat which make up the steel ball’s thermal
capacitance and the medium’s thermal conductivity. Next in the order
of sensitivity is the mean air velocity and the air kinematic viscosity.
Changes in the initial temperature potential, (Tsphere initial − Tair), affect
t50°C about ± 3% in this case study. Steel ball surface emissivity is the
least effective of independent variables to t50°C.
153
CHAPTER
QUENCHING 17
OF STEEL
BALLS IN OIL
I n Chapter 16, the quenching medium analyzed was air. When water,
oil and other similar mediums are used for quenching, heat transfer
gets more complicated because the quenching medium goes into
different heat transfer regimes, such as film boiling, transition
between film boiling and nucleate boiling, nucleate boiling, and free
convection, as the surface temperature of the material being
quenched decreases.
Heat transfer in the boiling regime depends on material surface,
quenching medium and material surface combination, and on the
quenching medium's saturated liquid and vapor thermophysical
characteristics. The best way to achieve heat transfer coefficients
for boiling heat transfer in a quenching process is by experimental
measurements (see Reference by Lee, W. J., Kim, Y. and,
E. D. Case [12]).
In order to assume a uniform temperature in the steel ball, namely
Biot number < 0.1, (see Reference by J. P. Holman [5]), oil quenching
instead of water quenching is considered. Oil quenching heat transfer
coefficients are in the order of 1000 W/m2-C
155
If water is used as a quenching medium, the heat transfer

coefficients can easily exceed 10,000 W/m2-C. To satisfy the Biot
number criteria, a steel ball with a diameter of 0.0006 m or less
has to be used. In small surface areas, boiling heat transfer gets
more complicated as the ratio of buoyant and capillary forces play
significant roles in boiling characteristics. For Biot number > 0.1
and for time varying heat transfer coefficients, finite difference or
finite element methods in multi-dimensional and unsteady-state heat
transfer have to be utilized.
The energy balance for the steel ball in oil quenching medium can
be written as:
Rate of decrease of internal energy in spherical steel ball =

Heat lost to surrounding oil from the surface of the spherical ball
Since the heat transfer coefficient will change by time, the finite
difference form of the energy equation is the same as Eq. (16-1):
(ρVcp) (Tsphere (i+1) − Tsphere (i) )/dt = − hi A (Tsphere (i) − Toil) (17-1)
Tsphere (i+1) is temperature of the ball at time (i+1) in C,

Tsphere (i) is temperature of the ball at time (i) in C,
Toil is temperature of oil in C,
hi is the experimentally obtained total heat transfer coefficient in
W/m2-C, between the surface of the steel ball and oil, which is the
sum of the convection and radiation heat transfer mechanisms, and
varies with time.
A is surface area of the steel ball in m2,
dt is time interval used in calculations in seconds,
ρ is density of the steel ball in kg/m3,
cp is specific heat of the steel ball in J/kg-C, and
V is volume of the steel ball in m3.
In the present calculations, experimentally obtained heat transfer

coefficients for oil quenching are used (see Reference [12]). These are
assumed to be total heat transfer coefficients; namely, they account
156
Quenching of Steel Balls in Oil
3000
B
nucleate
boiling transition
htotal, W/m2-C
2000
free
1000
convection film boiling
C
(vapor
A
blanket)
0
0 200 400 600 800
Steel Ball Temperature, °C
Figure 17-1 Total heat transfer coefficient versus steel ball temperature for
oil quenching
for both boiling and radiation heat transfer mechanisms. For oil
quenching, the total heat transfer coefficient as a function of steel ball
surface temperature is shown in Figure 17-1.
In temperatures above 750°C, point C in Figure 17-1, the steel
ball encounters film boiling heat transfer and is covered with a vapor
blanket. The temperatures between points B and C, 600°C to 750°C,
constitute the transition region from boiling heat transfer to film
boiling heat transfer. The temperatures between points A and B, 200°C
to 600°C, are in the boiling heat transfer regime. Temperatures below
200°C are considered to be in the free convection heat transfer regime.
Steel ball properties ρ, cp, ksteel, and D are assumed to be constants
during the oil quenching process. The nominal values of governing
independent variables for the present sensitivity calculations are as
follows:
D = 0.012 m,
ρ = 7800 kg/m3
cp = 430 J/kg-C
157
ksteel = 60 W/m-C
Toil = 70°C
Tsphere initial = 900°C
dt = 0.1 s
For the present calculations, a time interval of 0.2 seconds is used
because the thermal time constant for quenching in oil has dropped
an order of magnitude as compared to quenching in air (see Chapter
16). Calculations are also repeated with a 0.1 second time interval,
which improves the results by only one percent. The Biot number
for this case, hR/ksteel, is 0.1 for an average heat transfer coefficient
of 1000 W/m2-C, where R = D/2. Steel ball temperatures versus time
are shown in Figure 17-2 for nominal heat transfer coefficients, and
for a ±10% variation about the nominal heat transfer coefficient
values given in Figure 17-1. Figure 17.2 shows the initial seconds
of the quenching process to emphasize the effects of heat transfer
coefficient variations. Boiling heat transfer phases occur during the
initial seconds of the oil quenching process. For the nominal case,
after 8.6 seconds, the natural convection heat transfer regime starts.
800
Temperature, °C
Nominal h
(Figure 17-1)
600
10% higher h
than nominal
400 10% lower h

than nominal
200
0 2 4 6 8 10
Time, s
Figure 17-2 Steel ball temperature versus time for oil quenching for three
different heat transfer coefficient distributions
158
Quenching of Steel Balls in Oil
Table 17-1 Effects of a ±10% change in nominal values, of

heat transfer coefficients given in figure 17-1, to
time for the steel ball to reach 100°C, and to time
for it to reach 36.8% of (Tsphere initial − Toil)
Change in Change in
Time to Reach Time to Reach
100°C For A 100°C For A
10% Decrease 10% Increase
Nominal In h Nominal In h Nominal
h Figure 17-1 +11.3% −9.6%
Change in Change in
Time to Reach Time to Reach
36.8% of 36.8% of
(Tsphere initial − Toil) (Tsphere initial − Toil)
For A 10% For A 10%
Decrease In Increase In
Nominal h Nominal h Nominal
h Figure 17-1 +12.3% −9.6%
Sensitivities of time for the steel ball to reach a temperature of

100°C and time for the steel ball to reach 36.8% of (Tsphere initial − Toil),
thermal response time for the quenching process, are analyzed as
the dependent variables to the changing heat transfer coefficients.
The results are provided in Table 17-1. Oil quenching time
sensitivities to changes in heat transfer coefficients are significant,
and they are one-to-one. These results emphasize the importance of
process controls during quenching.
159
CHAPTER
COOKING
TIME FOR
118
8
TURKEY IN
AN OVEN
E very year when Thanksgiving comes around, the question of a

turkey’s cooking time in an oven looms. This is an unsteady-state
conduction heat transfer problem in spherical coordinates, whose
solution can be found in References [1], [6], [11], and [16]. The turkey
is assumed to be stuffed and in a spherical shape in which the spatial
variations of temperature are only in the radial direction. Turkey also
has constant thermophysical and physical properties in space and in
time. Energy balance in the radial direction, r, for a small increment
of time, t, gives the following partial differential equation for the
temperature-time distribution for a turkey:
∂2Tturkey /∂r2 + (2/r)∂Tturkey /∂r = (1/α)∂Tturkey /∂t (18-1)
A closed form solution to Eq. (18-1) can be obtained by applying

the following initial condition:
Tturkey = Tturkey initial at time t = 0 (18-2)
and the following two boundary conditions, one at the center
161
at radius r = 0, r(∂Tturkey /∂r) = 0 (18-3)
the other at the outer surface
at radius r = R, − k(∂Tturkey /∂r) = h(Tturkey − Toven) (18-4)
where
α = (k/ρcp) is the thermal diffusivity of turkey in m2/s,

k is the thermal conductivity of turkey in W/m-C,
ρ is the density of turkey in kg/m3,
cp is the specific heat of turkey at constant pressure in
J/kg-C, and
h is the convection heat transfer coefficient between the surface
of the turkey and the oven environment in W/m2-C.
The solution to Eqs. (18-1) through (18-4) can be found in

References [1], [6], [11], and [16], and it is as follows:
(Toven − Tturkey @ r)/(Toven − Tturkey initial) =
(18-5)
where δn is the n’th positive root of the following transcendental

equation:
1 − δn cot(δn) = Biot number = hR/k (18-6)
The center temperature of the turkey is of particular interest for

cooking. The USDA recommends that for a cooked turkey, the center
of the stuffing should reach a temperature of 74°C. In the present
sensitivity analysis, the temperature time history at the center of the
bird is analyzed. Eq. (18-5) reduces to the following relationship at
the center, r = 0, of the turkey:
162
Cooking Time For Turkey In An Oven
(Toven − Tturkey @ r=0)/(Toven − Tturkey initial) =
(18-7)
There are eight independent variables that govern the present

sensitivity analysis. They are:
W = Weight of turkey in kg
ρ = Density of turkey in kg/m3
k = Thermal conductivity of turkey in W/m-C
cp = Specific heat of turkey at constant pressure in kJ/kg-C
Tturkey initial = Initial temperature of turkey in C
Tturkey final = Final temperature desired at the center, r = 0, of
turkey in C
Toven = Oven temperature in C
h = Convection heat transfer coefficient in W/m2-C
The radius, R, for the turkey is obtained from its weight and its
spherical assumption, using R = (3W/4πρ)1/3. There are three
non-dimensional variables that can capture all the temperature time
distributions resulting from Eq. (18-7):
The dependant variable is a non-dimensional temperature, Θ, namely
Θ = (Toven − Tturkey @ r=0)/(Toven − Tturkey initial). (18-8)
The independent variable is the dimensionless time, namely the

Fourier number,
Fo = αt/R2. (18-9)
The dimensionless parameter is the Biot number,
Bi = hR/k, (18-10)
which is the ratio of internal to external thermal resistance for

the turkey. Most solutions to unsteady-state heat transfer problems
163
shown in literature, i.e., References [10] and [16], are shown in

graphical form by using the non-dimensional variables given in
Eqs. (18-8), (18-9) and (18-10).
To find these solutions, five terms of Eq. (18-5) are utilized:
Θ = C1 exp(−δ12Fo) + C2 exp(−δ22Fo) + C3 exp(−δ32Fo)

+ C4 exp(−δ42Fo) + C5 exp(−δ52Fo) + … (18-11)
The five-term expansion solution in Eq (18-11) gives very accurate

results. The fifth term only contributes at very small Fo numbers,
i.e., contribution to Θ for Fo = 0.05 is less than 10−5. First the δ1, δ2,
δ3, δ4, and δ5 have to be determined from the positive roots of the
transcendental Eq. (18-6) for different Biot numbers.
The ovens that will be considered in the present sensitivity
analysis will have a convection heat transfer coefficient range of 1 to
60 W/m2-K. The turkeys will have a weight range of 2 to 20 kg. The
turkey’s thermophysical properties are assumed to be ρ = 1000 kg/m3,
cp = 3.33 kJ/kg-K, and k = 0.5 W/m-C. With these assumptions, the
diameter of turkey ranges from 16 cm to 34 cm. The Biot number
range will be 0.13 to 16.
The temperature time history of interest for cooking the turkey
falls in a time range between one and ten hours. For this cooking time
range, the Fourier number range is 0.1 to 0.2.
The temperature time history for the turkey can be presented by
using the three non-dimensional variables in Eqs. (18-8), (18-9) and
(18-10), as shown in Figure 18-1.
A typical turkey is cooked at an oven temperature of 190°C, with
an initial turkey temperature of 5°C and a final desired temperature at
the center of the cooked turkey to be 80°C. These temperature values
provide a region of interest for the dimensionless temperature Θ to be
at 0.43. The present sensitivity analysis is for a 7.3 kg stuffed turkey
with a 24 cm diameter.
There are two variables that characterize the oven that is being
used, oven temperature and the convection heat transfer coefficient
on the surface of the turkey inside the oven. For a fixed oven
temperature, Toven = 190.5°C , the cooking times can be obtained
164
1
Dimensionless Temperature
0.9 Bi=0.13
At Center Of Turkey
0.8 Bi=0.5
0.7 Bi=1
0.6 Bi=2
0.5
Bi=3
0.4
0.3 Bi=5
0.2 Bi=7
0.1 Bi=10
0 Bi=16
0.1 0.2 0.3 0.4 0.5
Fourier Number, Fo
Figure 18-1 Dimensionless temperature at the center of a stuffed turkey

versus Fourier number for different Biot numbers in the region
of interest for cooking a turkey
from Figure 18-1 for different convection heat transfer coefficients

or Biot numbers. The cooking times versus the convection heat
transfer coefficient for three different size turkeys are given in Figure
18-2. The results in Figure 18-2 are for a final desired temperature of
73.9°C at the center of the turkey. The USDA recommended cooking
times for different size turkeys are also shown in Figure 18-2. USDA
recommended cooking times require a good convection oven with a
convection heat transfer coefficient ranging from 15 to 20 W/m2-C.
As the convection heat transfer coefficient increases cooking
time decreases, and this decrease in cooking time is very sensitive
to changes in the convection heat transfer coefficients, especially in
low values. The behaviors of these curves for different stuffed turkey
weights are similar. The slopes of the curves in Figure 18-2 are given
in Figure 18-3.
Figure 18-3 shows that ∂(Cooking Time)/∂h decreases sharply
as the forced convection heat transfer coefficient values approach
natural convection heat transfer coefficient values. The other variable
that is governed by the oven is oven temperature. The sensitivities
165
4.5 kg (10 lb)

15 stuffed turkey
14
13 7.3 kg (16 lb)
Cooking Time, hr
12 stuffed turkey
11 10 kg (22 lb)
10
9 stuffed turkey
8 USDA recommended
7
6 cooking times:
5 for a 4.5 kg (10 lb)
4 stuffed turkey 3-3.5 hrs,
3 for a 7.3 kg (16 lb)
2
0 5 10 15 20 stuffed turkey 4-4.5 hrs
for a 10 kg (22 lb)
stuffed turkey 4.75-5.25
hrs
Figure 18-2 Stuffed turkey cooking time as a function of the convection

heat transfer coefficient for stuffed turkeys of different weights
for Toven = 190.5°C and for Tturkey final = 73.9°C
∂(Cooking Time) / ∂h, hr-m2-C/W
−0.1
W=4.5 kg
(10 lb)
−0.2
W=7.3 kg
(16 lb)
−0.3
W=10 kg
(22 lb)
−0.4
−0.5
0 10 20 30
Convection Heat Transfer Coefficient,W/m2-C
Figure 18-3 Stuffed turkey cooking time to convection heat transfer

coefficient sensitivity as a function of the convection heat
transfer coefficient for stuffed turkeys of different weights, for
Toven = 190.5°C, and for Tturkey final = 73.9°C
166
∂ (Cooking Time) / ∂ (Toven), min/C
−0.8
−1 Turkey Center Final

Temperature=73.9°C
−1.2
Turkey Center Final
Temperature=79.4°C
−1.4
Turkey Center Final
−1.6 Temperature=85.0°C
−1.8
180 190 200 210 220
Oven Temperature, °C
Figure 18-4 A 7.3 kg stuffed turkey cooking time to oven temperature

sensitivity versus oven temperature for different turkey center
final temperatures
of the cooking time to oven temperature for different stuffed turkey

centerline temperatures are given in Figure 18-4.
The absolute value of the cooking time sensitivity to oven
temperature decreases as the oven temperature increases. Sensitivity
curves are similar in this application region for different final turkey
center temperatures.
The above sensitivity graphs show that the cooking time sensitivity
to governing independent variables behave non-linearly. Therefore,
it is more appropriate to analyze these sensitivities in the region
of interest and rank them according to their effects on the cooking
time. The results in Table 18-1 are obtained from Figure 18-1 by
interpolating between appropriate non-dimensional temperatures,
Biot numbers and Fourier numbers.
The effects shown in Table 18-1 to cooking time are given in
descending order. The most effective variable is the specific heat of
the stuffed turkey at constant pressure. The cooking time sensitivities
to all the independent variables are at the same order of magnitude.
167

variables to cooking time
Cooking Cooking
Time Change Time Change
Decrease In Decrease In
Nominal Nominal Nominal
Specific Heat @
Constant 3.33 kJ/kg-C −10% +10%
Pressure, cp
Tturkey final @ r=0 73.9°C −8.12% +8.12%
Toven − Tturkey initial 186.1°C +8.10% −6.41%
Thermal
0.5 W/m-C +7.10% −6.14%
Conductivity, k
Weight, W 7.3 kg −6.02% +6.02%
Convection
Heat Transfer 10 W/m2-C +5.13% −4.42%
Coefficient, h
Density, ρ 1000 kg/m3 −4.32% +4.63%
Some variables, such as the turkey’s specific heat at constant

pressure, final temperature desired at the center of turkey, and
weight of the turkey behave linearly in the region of interest, and give
the same magnitude cooking time change percentages on both sides
of their nominal values. It is important to remember that the ranking
order shown in Table 18-1 is good only in the region of nominal
values for the present calculations, due to non-linear behavior of the
sensitivities.
168
CHAPTER
HEAT
GENERATED IN
19
PIPE FLOWS DUE
TO FRICTION
H eat generated in pipes or in orifices due to fluid friction in

high-viscosity fluids can be substantial. In the present analysis,
the heat generated in steady-state and fully developed pipe flows is
investigated for fluids of different viscosities. The pressure drop in a
pipe due to fluid friction, ∆P, is generally defined as the product of the
fluid friction factor, f, non-dimensional length of the pipe, L/D, and
kinetic energy of the fluid, ρVm2/2, flowing in it.
∆P = f (L/D) (ρVm2/2) (19-1)
Eq. (19-1) has been determined by dimensional analysis (see

Reference [15]). The friction factor is a function of Reynolds
number, ReD = ρVmD/µ, and pipe inner surface roughness, e/D. The
friction factors, f, for fully developed pipe flows have been obtained
experimentally for different Reynolds numbers and pipe inner surface
roughness conditions, and they are given in References [6] and [10] as
graphs, called Moody Diagrams, or as empirical equations.
169
The pressure loss in the pipe is related to the heat generated in the
pipe by the first law of thermodynamics:
Q = (VmA) ∆P (19-2)
where
∆P is the pressure drop in pipe in N/m2,

L is the length of pipe in m,
D is the internal diameter of pipe in m,
ρ is the density of fluid in kg/m3,
Vm is the average fluid velocity in pipe in m/s,
Q is heat generated due to friction in W, and
A is cross sectional area of pipe in m2.
The first example analyzes a high-viscosity engine oil, and flow

rates are in the laminar flow region; i.e., ReD < 2000. The friction
factor for laminar flow is given as follows, (see Reference [10]), and it
is independent of pipe surface roughness:
f = 64/ReD (19-3)
Then the heat generated due to friction, after combining Eqs. (19-1),
(19-2) and (19-3), becomes:
Q = (128/π) µL (VmA)2/D4 (19-4)
Nominal values of the independent variables for this example

are assumed to be as follows; the density of oil cancels out of the
governing Eq. (19-4).
L = 400 m,
D = 0.1 m,
VmA = 1000 liters/min (0.0167 m3/s), and
µ = 0.486 N-s/m2 @ 27°C.
Temperature effects on the kinematic viscosity of engine oil, µ/ρ,

from 5°C to 55°C, are given in Figure 19-1. The kinematic viscosity of
170
Heat Generated In Pipe Flows Due To Friction
0.003
Kinematic Viscosity, m2/s
0.002
0.001
0.000
5 15 25 35 45 55
Mean Oil Temperature, °C
Figure 19-1 Kinematic viscosity of oil as a function of mean oil

temperature
oil decreases as its temperature increases. Similar behavior is seen in

Figure 19-2 for the heat generated due to friction.
The mean temperature increase in oil behaves the same way as shown
in Figure 19-3. ∆Tmean is calculated from Q/(ρVmAcp), assuming that all
150
Heat Generated, kW
100
50
0
5 15 25 35 45 55
Figure 19-2 Heat generated by friction for oil flowing in a pipe
171
5
Temperature Increase, °C
0
5 15 25 35 45 55
Figure 19-3 Temperature increase in oil due to friction
the heat generated due to fluid friction goes to increase the temperature
of oil. cp is the specific heat of oil at constant pressure, in J/kg-C.
Sensitivities of heat, generated by oil flowing in a pipe in the
laminar flow region, to the governing independent variables can be
obtained by differentiating Eq. (19-4):
∂Q/∂L = (128/π) µ (VmA)2/D4 (19-5)
∂Q/∂µ = (128/π) L (VmA)2/D4 (19-6)
∂Q/∂D = (−512/π) µ L (VmA)2/D5 (19-7)
∂Q/∂(VmA) = (256/π) µ L (VmA)/D4 (19-8)
The sensitivity of heat generated to pipe internal diameter, Eq.

(19-7), is shown in Figure 19-4. The sensitivity is prominent at
low pipe internal diameters, and approaches zero fast as the pipe
diameter increases for this case. The sensitivity of heat generated
to pipe flow rate, Eq. (19-8), is given in Figure 19.5. The sensitivity,
172
0.E+ 00
∂ Q/∂ D, W/m
−1.E+ 07
−2.E+ 07
−3.E+ 07
0.05 0.07 0.09 0.11 0.13 0.15
Pipe Inside Diameter, m
Figure 19-4 Sensitivity of heat generated due to friction to pipe internal

diameter for fully developed laminar flows
∂Q/∂(VmA) increases linearly with the increasing flow rate. These

sensitivities are calculated at a mean oil temperature of 27°C.
are varied ±10%, the results in Table 19-1 are obtained. These heat
generation sensitivities are given in descending order of importance,
and they are applicable only around the nominal values assumed for this
case. Heat generated due to friction is most sensitive to pipe internal
diameter and to flow rate. Heat generated due to friction has a one-to-one
sensitivity relation to the length of the pipe and to fluid viscosity.
Similar sensitivity calculations are done for water, which has a
three order of magnitude lower viscosity than the engine oil analyzed
previously. Nominal values of the independent variables for this
example are assumed to be as follows:
L = 400 m,
D = 0.1 m,
VmA = 1000 liters/min (0.0167 m3/s), and
µ = 0.855 × 10−3 N-s/m2 @ 27°C.
ρ = 997 kg/m3 @ 27°C
173
1.5E+07
∂ Q/∂ (VmA), W-s/m3
1.0E+07
5.0E+06
0.0E+00
0.01 0.03 0.05 0.07 0.09
Flow Rate, m3/s
Figure 19-5 Sensitivity of heat generated due to friction to pipe flow rate for
fully developed laminar flows
In this case friction factors are lower, ReD is higher in the

order of 105, and this case falls into a steady-state and fully
developed turbulent flow region in a pipe. For the friction
coefficient in smooth pipes in the turbulent flow region, the

variables to heat generated due to friction for
steady-state and fully developed laminar flow of
engine oil in a pipe
Heat Generated Heat Generated

Due To Friction Due To Friction
Change For A 10% Change For A 10%
D 0.1 m +52.42% −31.70%
VmA 1000 L/min −19% +21%
L 400 m −10% +10%
µ @ 27°C 0.486 N-s/m 2
−10% +10%
174
following empirical equation is used (see Reference [6]):
f = (0.790 ln(ReD) − 1.64)−2 for 3000 < ReD < 5 × 106 (19-9)
The heat generated due to friction for fully developed turbulent

flow in a smooth pipe can be obtained by combining Eqs. (19-1),
(19-2) and (19-9). Heat generated by water flow in a pipe is
analyzed between the mean water temperatures of 5°C and 55°C.
The kinematic viscosity of water is three orders of magnitude less
than the engine oil kinematic viscosity, as shown in Figure 19-6.
Heat generated due to friction, and therefore the mean temperature
increase for water, is two orders of magnitude less than the previous
oil flow case, as shown in Figures 19-7 and 19-8.
±10%, the results in Table 19-2 are obtained. These heat generation
sensitivities are given in descending order of importance, and they are
applicable only around the nominal values assumed for this case. Heat
generated due to friction is most sensitive to pipe internal diameter
and then to the flow rate, as in the previous case. A change in the
1.50E– 06
Kinematic Viscosity, m2/s
1.30E– 06
1.10E– 06
9.00E– 07
7.00E– 07
5.00E– 07
5 15 25 35 45 55
Mean Water Temperature, °C
Figure 19-6 Kinematic viscosity of water as a function of mean water

temperature
175
2.6
Heat Generated, kW
2.4
2.2
2
5 15 25 35 45 55
Figure 19-7 Heat generated by friction for fully developed and turbulent
water flow in a pipe
viscosity of water is the least effective independent variable to heat

generated due to friction.
Changes in pipe inner surface roughness also effect the heat
generated due to friction in a pipe. The effects of pipe inner surface
0.038
Temperature Increase, °C
0.036
0.034
0.032
0.030
0.028
5 15 25 35 45 55
Figure 19-8 Temperature increase in mean water temperature due to friction
176

variables to heat generated due to friction for a
steady-state and fully developed turbulent flow of
water in a smooth pipe
Heat Generated Heat Generated

Due To Friction Due To Friction
Change For A 10% Change For A 10%
D 0.1 m +65.95% −36.75%
VmA 1000 L/min −25.60% +30.67%
L 400 m −10% +10%
ρ 997 kg/m 3
−8.15% +7.99%
µ @ 27°C 0.855 × 10 N-s/m -3 2
−2.02% +1.86%
% Increase in Heat Generated
160
120
80
40
0
0 200 400 600 800 1000
Pipe Inner Surface Roughness, µm
Figure 19-9 Percent increase in heat generated due to friction as a

function of pipe inner surface roughness for water flowing in a
pipe, ReD = 2.47 × 10+5
177
roughness are analyzed for ReD = 2.47 × 10+5 using the Moody
Diagram (see Reference [10]). The results are given in Figure 19-9
for steady-state, fully developed, and turbulent water flowing in a
pipe. The effect of pipe inner surface roughness to heat generated
due to friction increases linearly above e = 100 µm. A pipe with an
inner surface roughness of 100 µm generates 38% more heat than
a smooth one.
178
CHAPTER
SIZING AN 20
ACTIVE SOLAR
COLLECTOR
FOR A POOL
S olar collectors used to heat water have been improving steadily in

solar energy conversion efficiency as the demand to go to greener
energy sources increases. To be able to design a pool solar collector
system, a good knowledge of year-around environmental conditions
has to be gathered. Average daily incident solar radiation, average
daily daytime and nighttime environment temperatures, and average
daily daytime and nighttime relative humidity must be known.
Sizing an active (directly circulating the pool water) solar
collector to heat a pool's water requires knowledge of the average
solar insolation at the location of the collector, and the heat transfer
efficiency of the solar collector.
A pool's physical and thermophysical properties must be known
to deal with radiation heat transfer, convection heat transfer, and
evaporative cooling from the water's surface during the day and night.
Conduction heat transfer from the pool's structure to the earth also
has to be treated. A plastic pool surface cover will be considered in
order to minimize heat losses to the environment when the pool is
not in use. Therefore, a plastic cover's physical and thermophysical
properties must also be known.
179
The present heat transfer model for the pool solar collector
system neglects temperature gradients in pool water. The model
is for unsteady-state heat transfer in one-dimensional rectangular
coordinates. There are 25 independent variables that govern the
dependent variable, size of the solar collector, and are considered
for sensitivity analysis.
The present heat transfer model has the following assumptions:
The pool is covered at night for 14 hours and open for swimming
during the day for ten hours. Energy lost from the water in the solar
collector at night and energy lost in piping between the pool and the
solar collector are considered to be lumped into the solar collector
efficiency. Average heat transfer properties are used during both
day and night hours. When a pool cover is used at night, there is no
evaporation from the pool's surface.
The following energy balances can be constructed for the water in
the pool for the day-time and night-time hours.
Energy balance during the day for the pool:
Qin solar collector + Qin convection from environment during day − Qout evaporation during day
+ Qin net radiation on pool surface − Qout conduction to earth
= mw cpw (dTpool/dθ) (20-1)
Energy balance during the night for the pool with a cover:
− Qout convection to environment at night − Qout conduction to earth

− Qout radiation from pool cover surface = mw cpw (dTpool/dθ) (20-2)
Energy balance during the night for the pool without a cover:
− Qout convection to environment at night − Qout conduction to earth

− Qout radiation from pool surface − Qout evaporation at night
= mw cpw (dTpool/dθ) (20-3)
where mw is weight of water in the pool in kg, cpw is specific heat

of water at constant pressure in W-hr/kg-C, Tpool is average pool
temperature in C, and θ is time in hours.
180
Sizing An Active Solar Collector For A Pool
Each term in energy balance Eqs. (20-1), (20-2) and (20-3)

is defined as follows. Net solar energy the water receives while
circulating in the solar collector is:
Qin solar collector = η q Ac (20-4)
where η is the solar collector heat transfer efficiency, or the

percentage of solar insolation that can be converted to heating the
pool water by the collector and its connecting pipes, which can vary
between 0.4 and 0.7.
q is the average solar insolation on the collector in W/m2. Ac is the
area of the solar collector in m2.
Convection heat transferred from the environment to the water
surface during the day is:
Qin convection from environment during day

= hday Apool water surface (Tenvironment during day − Tpool) (20-5)
where hday is the average convection heat transfer coefficient between

pool water surface and environment during the day, Apool water surface
is pool water surface area (pool length times pool width in m2) and
Tenvironment during day is average environmental temperature during ten
hours of day in C.
Heat lost from the water due to evaporation to the environment
during the day is:
Qout evaporation during day = M Hfg (20-6)
where M is evaporation rate in kg/hr-m2 and Hfg is latent heat of

evaporation for water in air in W-hr/kg.
The mass transfer rate equation can be defined similar to the
convection heat transfer rate equation. By applying the perfect gas
law, the mass transfer rate equation can be written as follows:
M = (hdiffusion day/R Tpool )(pw − pa) (20-7)
181
By assuming thermal diffusivity of air, α in m2/hr, being equal to the

diffusion coefficient of water in air, D in m2/hr, and by assuming the
water in air concentration gradient at the water-air interface to
be the same as the temperature gradient—Nusselt number (hday L/ka) =
Sherwood number (hdiffusion day L/D), the mass transfer coefficient can
be obtained from the heat transfer coefficient (see Reference [2]):
hdiffusion day = hday/ρa cpa (20-8)
hdiffusion day is the mass transfer coefficient between pool water surface
and air in m/hr, R is the universal gas constant which is equal to
0.08205 m3-atm/kmol-K, pw is the saturation pressure of water vapor on
pool water surface at pool temperature, pa is partial pressure of water
vapor in air at average environmental temperature (calculated from
pa = φpsaturation, where φ is relative humidity in the air and psaturation is
the saturation pressure of water vapor in air at average environmental
temperature), L is a characteristic length of pool surface, namely pool
length, ka is thermal conductivity of air in W/m-K, ρa is density of air in
kg/m3, and cpa is specific heat of air at constant pressure in W-hr/kg-C.
The net solar radiation that is absorbed by water is:
Qin net radiation on pool surface = αw G Apool water surface (20-9)
where αw is average solar radiation absorptivity at water surface

and G is average solar insolation on pool water surface in W/m2, that
includes both direct and diffused solar energy.
Conduction heat transfer from the walls of the pool to earth is:
Qout conduction to earth

= (kpool wall Apool-earth surfaces/tpool wall)(Tpool − Tearth) (20-10)
where kpool wall is thermal conductivity of pool walls in W/m-K,

Apool-earth surfaces are surface areas between the pool and earth in m2,
defined as [(2 × pool length + 2 × pool width) × pool water height +
pool length × pool width], tpool wall is average pool wall thickness in m,
and Tearth is average earth temperature.
182
Heat lost during the night from the pool with a cover to the
environment can be written as a series thermal circuit of convection
and conduction heat transfer:
Qout convection to environment at night

= Apool water surface (Tpool − Tenvironment during night)/[(1/hnight )
+ (tcover /kcover)] (20-11)
where Tenvironment during night is average environmental temperature at

night in C, hnight is the average convection heat transfer coefficient at
night integrated over the length of the pool, given by the following
empirical relationship (see Reference [6]), in W/m2-K:
(hnight L/ka) = (0.037 ReL0.8 − 871) Pr0.333
for 0.6 < Pr < 60, 5 × 105 < ReL < 1 × 108, and Recritical = 5 × 105 (20-12)
where ReL = VL/ν is Reynolds number, with V being average air

speed in m/hr over the length of the pool L in m and ν being
kinematic viscosity of air in m2/hr at film temperature, the average
of water surface and environmental temperature. Pr = ν/αt is the
Prandtl number, where αt is thermal diffusivity of air in m2/hr
at film temperature, the average of surface and environmental
temperature. Recritical is the transitional Reynolds number from
laminar to turbulent flow on a flat plate. tcover is thickness of pool
cover in meters and kcover is thermal conductivity of pool cover at
pool temperature.
Heat radiated from the pool cover surface to the night sky is:
Qout radiation from pool cover surface = εpool cover surface σT4pool cover surface (20-13)
where εpool cover surface is emissivity of pool cover surface and σ the is
Stefan-Boltzmann constant, namely 5.67 × 10−8 W/m2-K4.
Energy balance during the night for the pool without
a cover has convection and radiation heat transfers defined
as follows:
183
Qout convection to environment at night = hnight Apool water surface

(Tenvironment during night − Tpool) (20-14)
Qout radiation from pool surface = εpool water surface σT4pool (20-15)
where εpool water surface is emissivity of pool water surface.

Qout evaporation at night is calculated the same way as Qout evaporation during day
except that hdiffusion night is calculated from the heat transfer coefficient
integrated over the length of the pool; Eq. (20-12), using an analogy
between heat and mass transfer.
(hdiffusion night L/D) = (0.037 ReL0.8 − 871) Sc0.333
for 0.6 < Sc < 3000, 5 × 105 < ReL < 1 × 108, and Recritical = 5 × 105 (20-16)
where D is the diffusion coefficient for water in air in m2/hr and Sc is

the Schmidt number defined as ν/D.
Governing Eqs. (20-1) and (20-2) are solved by explicit finite
difference method and by iteration, to determine the required solar
collector area that will heat the pool water to a minimum of 25°C
in two days, specifically one ten-hour day followed by a 14-hour
night, followed by another ten-hour day. Nominal conditions for the
Solar collector variables:

η = 0.7
q = 500 W/m2
Pool variables:
L = 50 m (pool length)
W = 25 m (pool width)
H = 2 m (pool water height)
G = 350 W/m2
αw = 0.96
εpool water surface = 0.96
ρw = 1000 kg/m3
184
cpw = 1.162 W-hr/kg-K

kpool wall = 0.7 W/m-K
tpool wall = 0.1 m (average wall thickness)
kcover = 0.04 W/m-K
tcover = 0.02 m
εpool cover surface = 0.1
Tpool initial = 20°C (pool water temperature at θ = 0)
Environmental variables:
Tenvironment during day = 30°C (average environmental temperature
during the ten-hour day)
Tenvironment during night = 15°C (average environmental temperature
during the fourteen-hour night)
Tearth = 20°C (average earth temperature around pool walls)
Other properties during the day:

hday = 2 W/m2-K
ρa = 1.1614 kg/m3
cpa = 0.28 W-hr/kg-K
φ = 40% (relative humidity of air during the day)
pw = saturation pressure of water vapor on pool water surface at
pool temperature is obtained from dry saturated steam temperature
tables (see Reference [8]).
psaturation = saturation pressure of water vapor in air at average
environmental temperature is obtained from dry saturated steam
temperature tables (see Reference [8]).
Other properties at night:

D = 0.0936 m2/hr
ν = 0.0572 m2/hr
Pr = 0.707
ka = 0.0263 W/m-K
V = 3600 m/hr
φ = 40% (relative humidity of air during the night)
pw = saturation pressure of water vapor on pool water surface at
pool temperature is obtained from dry saturated steam temperature
tables (see Reference [8]).
185
psaturation = saturation pressure of water vapor in air at average

environmental temperature is obtained from dry saturated steam
temperature tables (see Reference [8]).
Under these nominal conditions, and with a pool cover at night, a

solar collector of 1192 m2 (95% of the pool surface area) is required
to heat the pool water from the initial temperature of 20°C to 25°C
in two days. Without a pool cover at night, the solar collector area
requirement almost doubles to 2054.5 m2. Radiation heat loss to night
sky and evaporation heat loss to environment cause the required
collector size to double without a pool cover at night.
Solar collector efficiency and average solar insolation are the two
dominant variables that affect similarly the required solar collector area.
Required solar collector area is very sensitive to the product of solar
collector efficiency and the average solar insolation. Solar collector area
versus its efficiency is shown in Figure 20-1. The sensitivity of solar
collector area to its efficiency is shown in Figure 20-2. Sensitivity of
required solar collector area to these variables increases fast as collector
efficiency and solar insolation values decrease.
8000
Collector Area, m2
6000
4000
2000
0
0 0.2 0.4 0.6 0.8 1
Collector Efficiency
Figure 20-1 Solar collector area versus solar collector efficiency at average
solar insolation of 500 W/m2
186
0
∂ (Collector Area) / ∂ η, m2
−10000
−20000
−30000
0.2 0.4 0.6 0.8 1
Collector Efficiency, η
Figure 20-2 Sensitivity of solar collector area to collector efficiency at

average solar insolation of 500 W/m2
Initial pool water temperature, average environmental temperature

during the day, and the earth's temperature under the pool have
significant effects on sizing the solar collector area. These effects are
shown in Figures 20-3, 20-4 and 20-5, respectively, and their behaviors
3000
Collector Area, m2
2000
1000
0
15 17 19 21 23
Pool Water Initial Temperature, °C
Figure 20-3 Solar collector area versus pool water initial temperature
187
1250
Collector Area, m2
1220
1190
1160
25 27 29 31 33 35
Average Environment Temperature During The Day, °C
Figure 20-4 Solar collector area versus average environmental temperature

during the day
2000
Collector Area, m2
1500
1000
500
5 10 15 20 25
Average Earth Temperature Under The Pool, °C
Figure 20-5 Solar collector area versus average earth temperature under
the pool
188
1500
Collector Area, m2
1300
1100
900
700
30 35 40 45 50 55 60
Pool Length, m
Figure 20-6 Solar collector area versus pool length
are almost linear. Required solar collector area decreases 384 m2

per degree centigrade increase in the initial pool water temperature.
Similarly, required solar collector area decreases at a much smaller
rate, −7.2 m2, per degree centigrade increase in average environmental
2000
Collector Area, m2
1600
1200
800
400
10 15 20 25 30 35 40
Pool Width, m
Figure 20-7 Solar collector area versus pool width
189
2000
Collector Area, m2
1600
1200
800
400
0
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
Pool Water Height, m
Figure 20-8 Solar collector area versus pool water height
temperature during the day. Required solar collector area sensitivity to

average earth temperature under the pool is −52.7 m2/C.
Another set of variables significantly affecting required solar collector
area are the pool's physical dimensions. Sensitivities for pool length,
width and water height are shown in Figures 20-6, 20-7 and 20-8,
respectively. These sensitivities are constants. A 23.7 m2 increase in
solar collector area is required for every meter of increase in pool length,
a 47 m2 increase in solar collector area is required for every meter of
increase in pool width, and a 1049 m2 increase in solar collector area is
required for every meter of increase in pool water height.
Solar insolation on the pool water surface during the day also
affects required solar collector area significantly. The effects on
required solar collector area of average solar insolation on pool water
surface, and of water absorptivity, are depicted in Figures 20-9 and
20-10, respectively. Required solar collector area decreases with
increasing average solar insolation on pool water surface by 3.4 m4/W.
Required solar collector area decreases 12.5 m2 per 1% increase in
water surface absorptivity of incoming solar radiation.
Relative humidity also has an effect on sizing the solar collector.
This effect is compared in Figure 20-11 for a pool with and without
190
2000
Collector Area, m2
1500
1000
500
200 250 300 350 400 450 500
Average Solar Insolation On Pool Surface, W/m2
Figure 20-9 Solar collector area versus average solar insolation on pool
surface
a cover at night. Required solar collector area decreases 4.8 m2 for

every percent increase in relative humidity for a pool that is covered
at night. The sensitivity for a pool without a cover at night is a 6.5 m2
collector area decrease per 1% increase in relative humidity. This
1800
Collector Area, m2
1600
1400
1200
1000
0.5 0.6 0.7 0.8 0.9 1
Absorptivity Of Pool Water Surface
Figure 20-10 Solar collector area versus absorptivity of pool water surface
191
2800
Collector Area, m2
2300 Without
Cover At
Night
1800
With Cover
At Night
1300
800
0 0.2 0.4 0.6 0.8 1
Relative Humidity
Figure 20-11 Solar collector area versus relative humidity
higher sensitivity comes from higher evaporation heat loss to the cold
night atmosphere.
If a region with a relative humidity of 10% is considered to build a
solar heated pool, as compared to a region with a relative humidity
of 60%, with everything else being the same and considering a pool
cover at night, the dry region requires 240 m2 more solar collector
area because more heat is lost due to evaporation to the dry
atmosphere.
There are over 25 independent variables that govern this heat
transfer problem. When the nominal values of the variables given
above are varied ±10%, the results shown in Table 20-1 are obtained
for a pool that is covered at night. Required solar collector area
sensitivities to a ±10% change in the governing variables are given in
the nominal values assumed for this study.
Required solar collector area is most sensitive to changes in initial
pool water temperature, followed by pool water height and pool water
specific heat at constant pressure. The next set of variables in order
of sensitivity are solar collector efficiency, average solar insolation on
collector, average solar insolation on pool water surface, average solar
192

variables to required solar collector area for a
pool covered at night (Note: Effects less than 1%
are not included in Table 20.1)
Change In Change In
Required Solar Required Solar
Collector Area Collector Area
For A 10% For A 10%
Tpool initial 20°C +64.35% −64.35%
H, pool water
2m −17.53% +17.62%
height
1.162
cpw −17.37% +17.45%
W-hr/kg-K
η, solar collector
0.7 +11.16% −9.06%
efficiency
q, average solar
insolation on 500 W/m2 +11.16% −9.06%
collector
G, average solar
insolation on 350 W/m2 +10.07% −10.07%
pool water surface
αw, average −4.19% with
solar radiation only a possible
0.96 +10.07%
absorptivity at 4.2% increase
water surface to 1
L, pool length 50 m −9.90% +9.98%
W, pool width 25 m −9.82% +9.90%
Tearth 20°C +8.89% −8.81%
Tenvironment during day 30°C +1.85% −1.76%
φ, relative humidity 0.40 +1.60% −1.59%
tpool wall 0.1 m +1.26% −1.01%
kpool wall 0.7 W/m-K −1.09% +1.17%
193
radiation absorptivity at water surface, pool length, pool width, and

average earth temperature around pool walls.
The sensitivities then drop an order of magnitude to less than 2%.
The average environmental temperature during the day continues in
the order of sensitivity, followed by relative humidity, average pool
wall thickness, and pool wall thermal conductivity.
The rest of the independent variables have less than 1%
sensitivities on the required solar collector area. These variables,
in order of descending sensitivity, are the emissivity of the pool
cover, temperature of the environment at night, the convection heat
transfer coefficient during the day, thermal conductivity of pool cover,
thickness of pool cover, mean air speed over the pool at night, and
thermophysical properties of air.
194
CHAPTER
HEAT
TRANSFER
21
IN A HEAT
EXCHANGER
T he design of a heat exchanger can be a good challenge for

engineers. The design methods cover a vast variety of engineering
disciplines, such as heat transfer, fluid mechanics, stress analysis,
corrosion, materials, economics, etc. There are two popular heat
transfer design methods covered in literature; see References [5], [6],
[10] and [15].
One heat transfer design method is called the log mean
temperature difference method. If the hot and cold fluid inlet and/or
outlet temperatures are specified, the heat exchanger design can be
performed by using this method. On the other hand, if the hot and
cold fluid inlet and/or outlet temperatures are not specified, and a
comparison between various types of heat exchangers is required,
then the effectiveness method is preferred.
In this chapter, the sensitivity of a heat exchanger design to
governing independent variables is analyzed. Heat transfer in a
counterflow concentric pipe liquid-to-liquid heat exchanger is
considered using the log mean temperature difference design method.
Hot engine oil flows in the inner pipe and the cold water flows in
the outer pipe. Under steady-state conditions, by neglecting heat
195
transfer to the environment, and assuming constant thermophysical

properties, the heat transferred through a small element of the heat
exchanger from the hot oil equals the heat received by the cold water.
dQ = −ρhVmhAh cph dTh = ρcVmcAc cpc dTc (21-1)
Heat transferred through a small element of the heat exchanger in

Eq. (21-1) can also be expressed in terms of an overall heat transfer
coefficient between the hot fluid and the cold fluid, through the
separating wall:
dQ = U (Th − Tc) dAwall (21-2)
Eqs. (21-1) and (21-2) can be combined to eliminate dQ, and then
integrated between the inlet and outlet temperatures of the heat
exchanger to give the following log mean temperature difference
design method equations for a counterflow heat exchanger:
Q = ρhVmhAh cph (Th in − Th out) (21-3)
Q = ρcVmcAc cpc (Tc in − Tc out) (21-4)
UAwall [(Th out − Tc in) − (Th in − Tc out)]

Q= (21-5)
ln[(Th out − Tc in)/(Th in − Tc out)]
where
UAwall = UhAwall hot fluid side = UcAwall cold fluid side (21-6)
and Awall hot fluid side represents inside surface area of the inner tube and
Awall cold fluid side represents outside surface area of the inner tube. The
overall heat transfer coefficient from the hot fluid side is:
Uh = 1/[(1/hh) + Rh foul + (Dh ln(Dc/Dh)/2kss)

+ (Awall hot fluid side Rc foul/Awall cold fluid side)
+ (Awall hot fluid side/Awall cold fluid sidehc)] (21-7)
196
Heat Transfer In A Heat Exchanger
All the variables in these governing Eqs., (21-1) through (21-7), are
defined as follows:
ρh = Hot oil density in kg/m3
VmhAh = Hot oil flow rate in liters/minute
cph = Hot oil specific heat at constant pressure in J-kg/C
Th in = Hot oil inlet temperature to the heat exchanger in C
Th out = Hot oil outlet temperature from the heat exchanger in C
ρc = Cold water density in kg/m3
VmcAc = Cold water flow rate in liters/minute
cpc = Cold water specific heat at constant pressure in J-kg/C
Tc in = Cold water inlet temperature to the heat exchanger in C
Tc out = Cold water outlet temperature from the heat exchanger in C
Uh = Overall heat transfer coefficient based on the inside surface of
the inner tube in W/m2-C
Awall hot fluid side = Inside surface area of the inner tube, πDhL, in m2
Uc = Overall heat transfer coefficient based on the outside surface
of the inner tube in W/m2-C
Awall cold fluid side = Outside surface area of the inner tube, π(Dh + 2t)L =
πDcL, in m2
Dh = Inside diameter of the inner tube in m
Dc = Outside diameter of the inner tube, Dh + 2t, in m, and where t
is the thickness of the inner tube
kss = Wall tube material thermal conductivity in W/m-C
hh = Convection heat transfer coefficient between the hot oil and
the inner tube inside surface in W/m2-C
hc = Convection heat transfer coefficient between the cold water
and the inner tube outside surface in W/m2-C
Rh foul = Fouling resistance for the inside surface of the inner tube in
m2-C/W
Rc foul = Fouling resistance for the outside surface of the inner tube
in m2-C/W
In a heat exchanger, fouling resistances affecting heat transfer

on surfaces of walls are caused by corrosion or by foreign material
deposited over time, and they are determined by experimental
methods during heat exchanger life tests.
197
In order to be able to determine the convection heat transfer

coefficients, first the Reynolds numbers for inner and outer tubes
have to be determined.
For the inner tube the Reynolds number definition is given by
ReDinner tube = ρhVmhDh/µh , where µh is the viscosity of hot oil in N-s/m2.
For the outer tube the Reynolds number definition is given by
ReDouter tube = ρcVmcDc-hydraulic/µc , where µc is the viscosity of cold
water in N-s/m2. Dc-hydraulic is the hydraulic diameter of the outer tube
given as Dc-hydraulic = 4 flow cross-sectional area/flow wetted perimeter,
or Dc-hydraulic = Dinside diameter of outer tube − Dc in meters.
The nominal values for the independent variables of the present
heat exchanger sensitivity analysis are assumed to be as follows:
For hot oil variables:
VmhAh = 5 liters/minute
Th in = 70°C
Th out = 40°C
ρh = 870 kg/m3
cph = 2000 J-kg/C
µh = 0.1 N-s/m2
kh = 0.14 W/m-C
For cold water variables:
VmcAc = 25 liters/minute
Tc in = 20°C
ρc = 1000 kg/m3
cpc = 41.8 J-kg/C
µc = 0.00096 N-s/m2
kc = 0.6 W/m-C
For geometry and other variables:
Dh = 0.02 m
t = 0.001 m
Douter tube = 0.025 m
kss = 15 W/m-C
Rh foul = 0.0008 m2-C/W
Rc foul = 0.0002 m2-C/W
198
ReDinner tube is calculated to be 46, which is less than 2000, and

therefore the hot oil flows in the laminar region. It is assumed that the
laminar flow is fully developed and has constant heat flux at the wall.
Therefore the Nusselt number, NuDh, is a constant (see Reference [6]):
NuDh = (hhDh)/kh = 4.36 (21-8)
ReDouter tube is calculated to be 11,758, which is greater than 4000, and

therefore the cold water flows in the turbulent flow region. It is assumed
that the turbulent flow is fully developed and the Nusselt number, NuDc,
is determined from an empirical relationship (see Reference [6]):
NuDc = (hcDc-hydraulic)/kc = 0.023 (ReDouter tube)0.8 Prc0.4 (21-9)
where Prc is the Prandtl number, the ratio of momentum to thermal

energy diffusion, defined as (µc/ρc)/(kc/ρccpc) or (µccpc/kc). µc is the
viscosity of cold fluid and kc is the thermal conductivity of cold fluid.
The counterflow heat exchanger length is considered as the main
dependent variable. Heat exchanger length versus hot oil flow rate is
given in Figure 21-1. The behavior is linear since the hot oil is flowing
300
Heat Exchanger Length, m
200
100
0
0 5 10 15
Hot Oil Flow Rate, L/min
Figure 21-1 Heat exchanger length versus hot oil flow rate
199
160
Length of Counterflow Heat
120
Exchanger, m
80
40
0
30 35 40 45 50
Hot Oil Outlet Temperature, °C
Figure 21-2 Heat exchanger length versus hot oil outlet temperature for
Th in = 70°C and Tc in = 20°C
300
Length of Counterflow Heat
Exchanger, m
200
100
0
0 10 20 30 40
Cold Water Inlet Temperature, °C
Figure 21-3 Heat exchanger length versus cold water inlet temperature for
Th in = 70°C and Th out = 40°C
200

independent variables to heat exchanger length
for a concentric tube counterflow heat exchanger
Heat Heat
Exchanger Exchanger
Length Length
Change For A Change For A
10% Decrease 10% Increase
Th out 40°C +24.98% −20.27%
Th in 70°C −16.78% +14.65%
kh 0.14 W/m-C +10.74% −8.79%
VmhAh 5 liters/minute −10.27% +10.33%
ρh 870 kg/m 3
−10.27% +10.33%
cph 2000 J-kg/C −10.27% +10.33%
Tc in 20°C −6.25% +7.21%
Dh 0.02 m +0.448% −0.377%
VmcAc 25 liters/minute +0.347% −0.282%
ρc 1000 kg/m 3
+0.347% −0.282%
cpc 41.8 J-kg/C +0.341% −0.277%
Rh foul 0.0008 m -C/W
2
−0.235% +0.236%
Douter tube 0.025 m −0.116% +0.126%
Rc foul 0.0002 m -C/W
2
−0.049% +0.049%
kss 15 W/m-C +0.040% −0.033%
t 0.001 m +0.013% −0.014%
kc 0.6 W/m-C +0.0090% −0.0077%
µc 0.00096 N-s/m 2
−0.0057% +0.0054%
µh 0.1 N-s/m2 0% 0%
201
in the laminar flow region and the convection heat transfer coefficient
is independent of ReDinner tube.
Hot oil inlet and outlet temperatures, and cold water inlet
temperature, have dominant effects on the heat exchanger length.
Heat exchanger length as a function of hot oil outlet temperature
is given in Figure 21-2. The heat exchanger length decreases
logarithmically as the requirement for the hot oil outlet temperature
increases.
Heat exchanger length is also a strong function of cold water inlet
temperature, and is shown in Figure 21-3. The heat exchanger length
increases as the cold water inlet temperature increases, and the
length increase behaves exponentially as the required hot oil outlet
temperature is approached.
are varied ±10%, the results shown in Table 21-1 are obtained. These
heat exchanger length sensitivities are given in descending order of
importance, and they are applicable only around the nominal values
assumed for this case study.
Heat exchanger length is most sensitive to hot oil outlet and
inlet temperature requirements. Hot oil flow rate and hot oil
thermophysical properties also affect the heat exchanger length at the
same order of magnitude as the hot oil outlet and inlet temperatures.
From the cold water variables, the heat exchanger length is most
sensitive to the cold water inlet temperature. The sensitivities of
the heat exchanger length to the rest of the independent variables
diminish fast in the present region of application. The heat exchanger
length is least sensitive to changes in fouling resistances, inside
diameter of the outer tube, thickness of the inner tube, thermal
conductivity of the tube wall material, viscosity of water, and thermal
conductivity of water.
202
CHAPTER
I CE
FORMATION
22
ON A LAKE
H eat transfer during ice formation has been studied in detail, in

papers by Lin, S. and Z. Jiang (Reference [13]), and by London,
A. L. and R. A. Seban (Reference [14]). In the present sensitivity
study, planar ice formation is considered. References [13] and [14]
also treat ice formation in cylindrical and spherical shapes. In the
present heat transfer analysis, thermal capacitance of ice is neglected;
in other words, a linear temperature profile through the ice thickness
is assumed. Ice thermophysical properties are assumed to be
constants and are considered at sea level conditions. Lake water is
assumed to be fresh. For proper ice formation heat transfer modeling,
lake water temperature is assumed to be greater than zero degrees
centigrade, and the atmosphere’s temperature is assumed to be less
than zero degrees centigrade:
Tlake water > Tfreezing > Tatmosphere .
Heat transfer from the lake water, through the ice layer, to
the atmosphere is assumed to be in an unsteady state and in
203
one-dimension rectangular coordinates. Governing heat transfer rate

equations are as follows:
Convection heat transfer from the lake water to the growing ice
surface is:
Qlake water to growing ice layer = hlake water-growing ice surface (Tlake water − Tfreezing) (22-1)
Heat transfer from the growing ice layer to the atmosphere is:
Qgrowing ice layer to atmosphere

= (Tfreezing − Tatmosphere)/[(1/hice upper surface-atmosphere) + (x/kice)] (22-2)
Energy balance at the growing ice layer for the latent heat of fusion
required per unit area of ice layer formation is:
ρiceHice (dx/dt) = Qgrowing ice layer to atmosphere

− Qlake water to growing ice layer (22-3)
T represents temperature in C. h represents the convection heat

transfer coefficient in W/m2-C. kice is the thermal conductivity of ice
in W/m-C. ρice is the density of ice in kg/m3. Hice is the latent heat of
fusion for water in W-hr/kg. x is the ice thickness in meters at time, t,
in hours.
The governing differential Eq. (22-3) can be rewritten after
separating the time and the space variables in the following
non-dimensional form:
dt* = (1 + x*)dx*/[1 − Z(1 + x*)] (22-4)
where the non-dimensional time is
t* = [h2ice upper surface-atmosphere (Tfreezing − Tatmosphere) t/(ρice Hice kice)] (22-5)
where the non-dimensional ice thickness is
x* = hice upper surface-atmosphere x/kice (22-6)
204
Ice Formation On A Lake
and where Z is a non-dimensional parameter which is the ratio of two

convection heat transfer mechanisms that govern this heat transfer
problem:
[hlake water-growing ice surface × (Tlake water − Tfreezing)]

Z= (22-7)
[hice upper surface-atmosphere × (Tfreezing − Tatmosphere)]
Eq. (22-4) is integrated using the initial condition
x* = 0 at t* = 0 (22-8)
to obtain the following ice formation rate equation:
t* = (1/Z2) ln[(Z − 1)/(Z(x* + 1) − 1)] − (x*/Z) (22-9)
Nominal values of the independent variables used in the present

sensitivity analysis are as follows:
hice upper surface-atmosphere = 30 W/m2-C

hlake water-growing ice surface = 10 W/m2-C
Tlake water = 6°C
Tatmosphere = −20°C
Tfreezing = 0°C
ρice = 920 kg/m3
kice = 1.88 W/m-C
Hice = 93 W-hr/kg
The time to form a certain thickness of ice under the nominal

conditions given above is shown in Figure 22-1. The ice thickness
limit under the above given nominal conditions that will support the
latent heat of fusion required for ice formation is 0.564 m.
Maximum ice thickness versus the convection heat transfer
coefficient between ice upper surface and atmosphere is given in
Figure 22-2. This condition occurs when the convection heat transfer
from the upper surface of the ice to the atmosphere starts to reach
the same level as the convection heat transfer from the lake water to
205
Ice Thickness Limit = 0.564 m

Under Given Nominal Conditions
2000
Time To Form Ice, hr
1500
1000
500
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Thickness Of Ice, m
Figure 22-1 Time to form ice versus thickness
the growing ice surface. Then there is no more energy left to support
the latent heat of fusion of water to keep forming the ice layer.
Under this condition, the natural logarithm term in the denominator
of the ice formation rate Eq. (22-9) approaches zero, or:
0.6
Maximum Ice Thickness, m
0.4
0.2
0
0 10 20 30 40 50
Heat Transfer Coefficient, hice upper surface-atmosphere, W/m2-C
Figure 22-2 Maximum ice thickness versus heat transfer coefficient

between ice upper surface and atmosphere
206
x* = (1/Z) − 1 (22-10)
Maximum ice thickness is very sensitive to variations in the

convection heat transfer coefficient between ice upper surface and
atmosphere at low values.
The time to achieve a certain thickness of ice depends on the values
of seven independent variables. In this study, sensitivities of time to
form 0.2 meters of ice thickness are investigated. Sensitivities of time
to form 0.2 meters of ice thickness to temperatures of atmosphere and
of lake water are shown in Figures 22-3 and 22-4, respectively.
In Figure 22-3, as the temperature of the atmosphere increases, the
sensitivity of time to form 0.2 meters thick of ice increases. Under the
present conditions, the temperature of the atmosphere has a limit of
−8.4°C for the ice to be able to reach a thickness of 0.2 meters
and support the energy requirement for freezing. Similarly, in
Figure 22-4, as the temperature of the lake water increases, the
sensitivity of time to form 0.2 meters thick of ice increases. Under the
present conditions, the temperature of the lake water has a limit of
14.3°C for the ice to be able to reach a thickness of 0.2 meters.
300
∂ t0.2 m / ∂ Tatmosphere, hr/C
250
200
150
100
50
0
–50 –40 –30 –20 –10
Temperature Of Atmosphere, °C
Figure 22-3 Sensitivity of time to form 0.2 m thick of ice to temperature of

atmosphere
207
80
∂ t0.2 m / ∂ Tlakewater, hr/C
60
40
20
0
2 4 6 8 10 12 14
Temperature Of Lake Water, °C
Figure 22-4 Sensitivity of time to form 0.2 m thick of ice to temperature of

lake water
Sensitivities of time to form 0.2 meters thick of ice to

hice upper surface-atmosphere and hlake water-growing ice surface are shown in
Figures 22-5 and 22-6, respectively.
The convection heat transfer coefficients on both sides of the ice
have opposite effects to the growth of ice. In Figure 22-5, as the
convection heat transfer coefficient between the ice upper surface
and the atmosphere decreases, the absolute value of ice formation
time sensitivity increases, or the time for formation of 0.2 meters of
ice increases. On the other hand, in Figure 22-6, as the convection
heat transfer coefficient between the lake water and the growing ice
surface increases, the ice formation time and its sensitivity increase.
The ice formation time sensitivities to ice density and to the
latent heat of fusion of water are constants. The ice formation time
sensitivity to ice density is positive, 0.115 hr-m3/kg. The ice formation
time sensitivity to latent heat of fusion of water is also positive,
1.14 kg/W. The ice formation time sensitivity to thermal conductivity
of ice is non-linear, and is given in Figure 22-7.
As the thermal conductivity of ice increases, the time to form 0.2
meters thick of ice decreases, as does the absolute value of its sensitivity.
208
∂ t0.2 m / ∂ hice upper surface–atmosphere,
–5
hr-m2-C/W
–10
–15
–20
–25
0 20 40 60 80 100
hice upper surface–atmosphere, W/m2-C
Figure 22-5 Sensitivity of time to form 0.2 m thick of ice to the convection
heat transfer coefficient between ice upper surface and
atmosphere
∂ t0.2 m / ∂ hlake water-growing ice surface,
30
20
hr-m2-C/W
10
0
5 10 15 20 25
hlake water-growing ice surface, W/m2-C
Figure 22-6 Sensitivity of time to form 0.2 m thick of ice to the convection
heat transfer coefficient between lake water and growing ice
surface
209
0
∂ t0.2 m / ∂ kice, hr-m-C/W
–50
–100
–150
–200
–250
1 1.5 2 2.5 3
Thermal Conductivity Of Ice, W/m-C
Figure 22-7 Sensitivity of time to form 0.2 m thick of ice to thermal

conductivity of ice

variables to time for ice thickness to reach 0.2 m
Change in Change in
Time For Ice Time For Ice
Thickness to Thickness to
Reach 0.2 m For Reach 0.2 m For
Tatmosphere −20°C −12.67% +17.05%
kice 1.88 W/m-C +10.87% −8.28%
ρice 920 kg/m 3
−10% +10%
Hice 93 W-hr/kg −10% +10%
hice upper surface-atmosphere 30 W/m -C2
+5.80% −4.61%
hlake water-growing ice surface 10 W/m -C2
−4.31% +4.78%
Tlake water 6°C −4.31% +4.78%
210

Sensitivities of time for the ice thickness to reach 0.2 meters to a
±10% change in the governing independent variables are given in
the nominal values assumed for this study.
Ice thickness formation time is most sensitive to the temperature of
the atmosphere. Ice thermophysical properties are next in the order
of sensitivity. Convection heat transfer coefficients on both sides of
the ice surfaces, and the lake water temperature, round up the ice
formation time sensitivity. Ice formation time sensitivities to all the
independent variables are non-linear, except for ice density and latent
heat of fusion for water. All the ice formation time sensitivities for
the nominal independent variables values assumed for the present
analysis are at the same order of magnitude.
211
CHAPTER
SOLIDIFICATION 23
IN A CASTING
MOLD
E xact temperature distribution solutions to unsteady-state one-

dimensional heat conduction equations in rectangular, cylindrical,
and spherical coordinates, for initial condition and different boundary
conditions, have been provided in literature (see References [1], [6],
[10] and [16]). The reference by Carslaw, H. S. and J. C. Jaeger [1]
extended these exact solutions to moving boundary conditions with
phase change.
In this chapter, sensitivity analysis for the solidification front
of a semi-infinite liquid in a semi-infinite mold is considered. The
unsteady-state one-dimensional heat conduction equation in
rectangular coordinates for the mold region, solidified cast material
region and liquid cast material region, with constant property
assumptions, is given by:
∂2T/∂x2 = (1/α) ∂T/∂θ (23-1)
where temperature T varies with space x and with time θ. α is

thermal diffusivity, k/ρcp, of the region in m2/s, where k is the thermal
213
conductivity of the region in W/m-C, ρ is the density of the region in

kg/m3, and cp is the specific heat of the region at constant pressure in
J/kg-C. Initial and boundary conditions for each region are as follows:
Minus infinity to x = 0 is assumed to be the mold region. In this
case, it is assumed that the mold is made out of sand. x > 0 to plus
infinity is the cast material’s region, which is initially filled with liquid
silver. The solidification front for the liquid silver starts initially at
x = 0 and grows in the +x direction.
Eq. (23-1) is applied to three different regions, in order to obtain
the temperature distributions in each region with the following initial
and boundary conditions:
At zero time, all the mold region from x = 0 to x = −∞ is at the
initial temperature of the sand, Tsand:
T(−x, 0) = Tsand (23-2)
The temperature of the mold region far away from x = 0 always

stays at the initial temperature of the sand, Tsand:
T(−∞, θ) = Tsand (23-3)
The interface between the mold region and the solidified material
at x = 0 has the following energy balance boundary condition:
ksand ∂Tsand/∂x = ksolid ∂Tsolid/∂x (23-4)
The solidified region initial condition is:
L = 0 at θ = 0 (23-5)
where L is the solidification front in meters.

The solidification region boundary at x = L is always at the
solidification temperature of the liquid cast material:
T(L, θ) = Tsolidification (23-6)
214
Solidification In A Casting Mold
At zero time, all the liquid cast material region from x = 0 to x = +∞

is at the initial temperature of the liquid cast material, Tliquid:
T(+x, 0) = Tliquid (23-7)
The temperature of the liquid cast material region far away from
x = 0 always stays at the initial temperature of the liquid cast
material, Tliquid:
T(+∞, θ) = Tliquid (23-8)
The interface between the liquid cast material region and the
solidified cast material region at x=L has the following boundary
condition, which is the result of an energy balance at the solidification
front that supports the latent heat of fusion for the liquid cast
material:
kliquid ∂Tliquid/∂x − ksolid ∂Tsolid/(x = ρsolid H∂dL/dθ) (23-9)
where H is the latent heat of fusion in W-s/kg. Exact solutions

to Eq. (23-1) have been obtained using the conditions in
Eq. (23-2) through Eq. (23-9) for three regions, and they can be
found in Reference [1].
In the present sensitivity analysis, propagation of the solidification
front is analyzed instead of temperature distributions for each region.
Reference [1] shows that in order to satisfy the temperature
continuity at the solidification front all the time, the solidification
front L has to move proportionately to the square root of time,
namely:
L = aθ0.5 (23-10)
where “a” is the proportionality constant, has the dimensions of

m/s0.5, and can be obtained by satisfying the energy balance,
Eq. (23-9), at the interface of the solidified cast material and liquid
cast material (see Reference [1]):
215
exp(−a2/4αsolid)/{[(ρsolidcp solidksolid)/(ρsandcp sandksand)]0.5

+ erf(a/2α0.5solid)} − [(ρliquidcp liquidkliquid)/(ρsolidcp solidksolid)]0.5
[(Tliquid − Tsolidification)/(Tsolidification −Tsand)]
exp[−a2(ρsolid/ρliquid)2/4αliquid]/erfc[a(ρsolid/ρliquid)/2α0.5liquid]
− aH(πρsolid/ksolidcp solid)0.5/2(Tsolidification − Tsand) = 0 (23-11)
where the error function “erf” is defined as

z
erf(z) = (2/π0.5) ∫0 exp(−u2) du (23-12)
and the complementary error function “erfc” is defined as
erfc(z) = 1 − erf(z) (23-13)
Eq. (23-11) is solved for “a” by trial and error. For the present
analysis, liquid silver is considered as the cast material, and sand is
considered as the mold material. For solidification to occur there are
two conditions that have to be met. One condition is that the liquid
cast material’s temperature cannot exc eed the melting temperature
of the mold material:
Tliquid < Tsand melting point (23-14)
The second condition is that the liquid cast material has to have
enough heat loss initially through the mold by conduction heat
transfer to be able to start solidifying. Eq. (23-11) provides this
condition by setting a equal to zero:
(Tliquid − Tsolidification)/(Tsolidification − Tsand)

< [(ρsandcp sandksand)/(ρliquidcp liquidkliquid)]0.5 (23-15)
For the present sensitivity analysis, the following nominal values

are assumed for the independent variables that also satisfy Eqs.
(23-14) and (23-15):
Sand values for the mold region:
216
Tsand = 100°C
Tsand melting point = 1430°C
ρsand = 2330 kg/m3
cp sand = 712 J/kg-C
ksand = 42 W/m-C
Solidified silver values for the solidified cast region:
Tsolidification = 960°C
ρsolid = 10500 kg/m3
cp solid = 235 J/kg-C
ksolid = 400 W/m-C
Liquid silver values for the liquid cast region:

Tliquid = 1000°C which is less than Tsand melting point of 1430°C to satisfy
Eq. (23-14), and is less than 1180°C, which is the “a” equals zero
condition in Eq. (23-15) to start solidification.
ρliquid = 9300 kg/m3
cp liquid = 318 J/kg-C
kliquid = 360 W/m-C
H = 104400 W-s/kg
0.4
Solidification Front L, m
0.3
0.2
0.1
0
0 5 10 15 20 25 30
Time, min
Figure 23-1 Cast solidification front L versus time
217
“a” is obtained by iteration for these nominal values, and is 0.008137

meters/second0.5. The solidification front versus time for the nominal
case is given in Figure 23-1.
±10%, the results shown in Table 23-1 are obtained. “a” sensitivities to
a ±10% change in the governing variables are given in descending order
of importance, and they are applicable only around the nominal values
assumed for this study. The variable Tliquid is varied on the negative
side −4% in order not to fall below Tsolidification. The variable Tsolidification is
varied on the positive side +4.2% in order not to exceed Tliquid.

variables to “a” in solidification front Eq. (23-10),
L = aθ0.5
Change in “a” Change in “a”

For A 10% For A 10%
−20.738% for
Tliquid 1000° C −4% change +69.208%
−17.677% for
Tsolidification 960° C +58.119% −4.2% change
H 104400W-s/kg +9.745% −8.206%
ρsolid 10500 kg/m 3
+9.641% −8.062%
ρsand 2330 kg/m 3
−3.192% +2.933%
cp sand 712 J/kg-C −3.192% +2.933%
ksand 42 W/m-C −3.192% +2.933%
cp liquid 318 J/kg-C −1.520% +1.517%
ksolid 400 W/m-C −1.324% +1.143%
Tsand 100° C +0.773% −0.777%
ρliquid 9300 kg/m3 −0.717% +0.686%
kliquid 360 W/m-C −0.717% +0.686%
cp solid 235 J/kg-C +0.673% −0.652%
218
Solidification front propagation parameter “a” is most sensitive

to the temperature of the liquid cast material and the solidification
temperature of the cast material. The second tier of “a” sensitivity
belongs to the latent heat of fusion and to density of solid region. In
order of sensitivity, the next set of independent variables are mold
region thermophysical properties, namely ρsand, cp sand, and ksand.
“a” is least sensitive to variations in independent variables cp liquid ,
ksolid , Tsand , ρliquid , kliquid and cp solid. Changes to these least sensitive
independent variables affect “a” two orders of magnitude less than
changes to temperature of the liquid cast material and changes of
solidification temperature of the cast material.
219
CHAPTER
AVERAGE
TEMPERATURE
24
RISE IN SLIDING
SURFACES IN
CONTACT
F rictional heating of materials in contact has been studied
thoroughly in heat transfer literature. In this case, an approximate
method developed by J. C. Jaeger (see Reference [7]) for frictional
temperature rise on a sliding square contact area will be analyzed for
sensitivity.
In order to determine the steady-state average temperature
at contacting surfaces, Jaeger [7] used the temperature solution
resulting from an instantaneous point source in an infinite solid for
the following unsteady-state and three-dimensional conduction
heat transfer differential equation, in rectangular coordinates
(see Reference [1]):
∂2T/∂x2 + ∂2T/∂y2 + ∂2T/∂z2 = (1/α)(∂T/∂θ) (24-1)
The temperature distribution and history resulting from an

instantaneous point source at (x1, y1, z1) of strength Q that satisfies
Eq. (24-1) can be expressed as:
T = [Q/8(παθ)0.5] exp{[(x − x1)2 + (y − y1)2 + (z − z1)2]/4αθ} (24-2)
221
T in Eq. (24-2) represents the temperature distribution and

history in an infinite medium at location (x, y, z) and at time θ. α is
the constant thermal diffusivity of the medium defined as (k/ρcp),
where k is the medium’s thermal conductivity, ρ is medium’s density,
and cp is medium’s specific heat at constant pressure. Eq. (24-2) can
be integrated to obtain steady-state temperatures for moving and
stationary line and square contact areas, where heat transfer was
restricted only to contact area (see References [1] and [7]).
In the present sensitivity analysis, a square contact area, 2L × 2L,
that is sliding with a constant velocity, V, over a semi-infinite body is
considered. For LV/2α > 5, the average temperature at the surface of
the semi-infinite medium that is under the sliding square contact area,
designated as area “1,” is approximated as follows (see Reference [7]):
Taverage at surface of contacted semi-infinite medium = (1.064 Q/k1) (α1L/V)0.5 (24-3)
The sliding square contact area also has a conducting semi-infinite

body behind it, and its temperature rises due to the heat generated
during contact. The average temperature of the square sliding contact
area designated as “2” is (see Reference [7]):
Taverage for sliding square contact area = 0.946 L Q/k2 (24-4)
Eq. (24-3) assumes that all the heat generated during the contact
goes to the semi-infinite medium under the sliding contact area, but
in reality this is not the case. A portion, mQ, of the heat generated
during the sliding contact goes into the semi-infinite body and the
rest, (1-m)Q, goes into the square sliding contact area. Equating the
average temperatures in Eqs. (24-3) and (24-4), the proportionality
constant “m” can be determined:
m = k1 (LV)0.5/[1.125 k2 α10.5 + k1 (LV)0.5] (24-5)
The heat flow proportionality constant, Eq. (24-5), is inserted

into Eq. (24-3) to determine the sliding contact area average
temperature:
222
Average Temperature Rise In Sliding Surfaces In Contact
Tsliding contact area average temperature

= (1.064 Q L α10.5)/[1.125 k2 α10.5 + k1 (LV)0.5] (24-6)
The heat generated during the contact comes from the mechanical
energy that is dissipated during the contact:
Q = fd W g V/(4 L2 J) (24-7)
where fd is the dynamic coefficient of friction during the contact,

W is the load at the contact area in kg, g is the gravitational constant,
9.8 m/s2, V is the relative velocity of the sliding contact area over
the semi-infinite medium in m/s, L is the half width of square sliding
contact area in m, and J = 1 kg-m2/s3-W, which is a proportionality
constant that equates mechanical work to heat from the first law
of thermodynamics. Combining Eqs. (24-6) and (24-7) gives the
final form of the sliding contact area average temperature for
(LV)/(2 α1) > 5:
Taverage = (0.266 fd W g V α10.5)/{L J [1.125 k2 α10.5 + k1 (LV)0.5]} (24-8)
where
α1 = thermal diffusivity of semi-infinite medium in m2/s,

k1 = thermal conductivity of semi-infinite medium in W/m-C,
k2 = thermal conductivity of the square sliding contact area in W/m-C.
The following nominal values for the independent variables are

assumed for the present sensitivity analysis. Also both the
semi-infinite medium and the square sliding contact area are assumed
to be stainless steel.
fd = 0.5,
W = 1000 kg,
V = 10 m/s,
α1 = 4 × 10−6 m2/s,
k1 = 15 W/m-C,
k2 = 15 W/m-C,
L = 0.01 m
223
With these nominal values, sliding contact area average

temperature rise is 546°C. Average temperature rises linearly with
load at the sliding contact area and with the dynamic coefficient of
friction. Average temperature sensitivity to load at the sliding contact
area is 0.546 C/kg, and to the dynamic coefficient of friction is 109.1°C
per 0.1 change in fd.
Sliding contact area average temperature rise versus relative
velocity of two bodies is shown in Figure 24-1. Average temperature
increases with increasing relative velocity as a V0.5 function. Sliding
contact area average temperature rise sensitivity to relative velocity
of two bodies is shown in Figure 24-2. Average temperature sensitivity
decreases as V-0.5 as relative velocity increases.
Sliding contact area average temperature versus thermal diffusivity
of semi-infinite body under sliding contact area is shown in Figure
24-3; average temperature increases as α10.5. Average temperature
sensitivity to α1 is given in Figure 24-4, and it behaves as α1−0.5.
Sliding contact area average temperature versus thermal
conductivity of semi-infinite body under the sliding contact area is
shown in Figure 24-5; average temperature decreases as k1−1.
1000
Average Temperature, °C
Sliding Contact Area
800
600
400
200
0
0 5 10 15 20 25 30
Relative Velocity Of Bodies, m/s
Figure 24-1 Sliding contact area average temperature versus relative

velocity of bodies
224
400
∂ Taverage / ∂ V, C-s/m
300
200
100
0
0 5 10 15 20 25 30
Relative Velocity Of Bodies, m/s
Figure 24-2 Sliding contact area average temperature sensitivity to relative

velocity of bodies
2000
1600
1200
800
400
0
0.E+00 1.E−05 2.E−05 3.E−05 4.E−05 5.E−05
Thermal Diffusivity, α1, Of Semi-Infinite Body Under Sliding
Contact Area, m2/s
Figure 24-3 Sliding contact area average temperature versus thermal

diffusivity, α1, of semi-infinite body under sliding contact area
225
1.2E+08
∂ Taverage / ∂α1, C-s/m2
1.0E+08
8.0E+07
6.0E+07
4.0E+07
2.0E+07
0.E+00 1.E−05 2.E−05 3.E−05 4.E−05 5.E−05
Thermal Diffusivity, α1, Of Semi-Infinite Body Under Sliding
Contact Area, m2/s
Figure 24-4 Sliding contact area average temperature sensitivity to thermal

diffusivity, α1, of semi-infinite body under sliding contact area
1600
1200
800
400
0
0 50 100 150 200 250 300
Thermal Conductivity, k1, W/m-C
Figure 24-5 Slider contact area average temperature versus thermal

conductivity of semi-infinite body under sliding contact area
226
Average temperature sensitivity to thermal conductivity of semi-infinite

body under the sliding contact area is given in Figure 24-6, and it
behaves as k1−2.
Sliding contact area average temperature versus thermal
conductivity, k2, of sliding square contacting body is shown in
Figure 24-7; average temperature decreases slightly, and almost
linearly, with increasing thermal conductivity of the sliding square
contacting body. Average temperature sensitivity to k2 is given in
Figure 24-8, and it can be considered a constant at −0.25 C2-m/W.
Under the given nominal conditions, slider contact area average
temperature is most sensitive to the size of the square contact area.
Sliding contact area average temperature versus half width, L, of
square contact area is depicted in Figure 24-9. Average temperature
decreases as L−1.5 as side of contact area increases. Average
temperature sensitivity to half width of square contact area is given in
Figure 24-10, and it behaves as L−2.5.
are varied ±10%, the results shown in Table 24-1 are obtained. Sliding
contact area average temperature sensitivities to a ±10% change in
10
∂ Taverage / ∂ k1, C2-m/W
–30
–70
–110
–150
0 50 100 150 200 250 300
Thermal Conductivity, k1, W/m-C

conductivity of semi-infinite body under sliding contact area
227
550
540
530
520
0 20 40 60 80 100
Thermal Conductivity Of Sliding Square Contacting body, k2,
W/m-C
Figure 24-7 Sliding contact area average temperature versus thermal

conductivity of sliding square contacting body
–0.2
∂ Taverage / ∂ k2, C2-m/W
–0.25
–0.3
0 20 40 60 80 100
Thermal Conductivity Of Sliding Square Contacting Body,
k2, W/m-C

conductivity of sliding square contacting body
228
3000
2500
2000
1500
1000
500
0
0 0.02 0.04 0.06 0.08 0.1
Half Width, L, Of Square Contact Area, m
Figure 24-9 Sliding contact area average temperature versus half width, L,
of square contact area
–50000
∂ Taverage / ∂ L, C/m
–100000
–150000
–200000
–250000
–300000
0 0.02 0.04 0.06 0.08 0.1
Half Width, L, Of Square Contact Area, m
Figure 24-10 Sliding contact area average temperature sensitivity to half

width, L, of square contact area
229

independent variables to sliding contact area
average temperature
Change In Change In
Sliding Contact Sliding Contact
Area Average Area Average
For A 10% For A 10%
Decrease Increase
L, half width of
sliding square
contact area 0.01 m +17.08% −13.30%
k1, thermal
conductivity of
semi-infinite body 15 W/m-C +11.02% −9.03%
fd, dynamic
coefficient of
friction0.5 −10% +10%
W, load at the
sliding contact
area 1000 kg −10% +10%
V, relative velocity
of sliding square
contact area and
the semi-infinite
body 10 m/s −5.17% +4.92%
α1, thermal
diffusivity of
semi-infinite body 4 × 10−6 m2/s −5.10% +4.84%
k2, thermal
conductivity of
sliding square
contact area 15 W/m-C +0.07% −0.07%
230
the governing independent variables are given in descending order of

importance, and they are applicable only around the nominal values
assumed for this analysis.
Sliding contact area average temperature is most sensitive to the
size of the contact area. Thermal conductivity of semi-infinite body
under the sliding contact area, dynamic coefficient of friction and
contact area load are next in the order of sensitivity. However, these
sensitivities have the same order of magnitude as sensitivity to the
size of the contact area. Relative velocity of two bodies and thermal
diffusivity of semi-infinite body under the sliding contact area come
next in the average temperature order of sensitivity. Sliding contact
area average temperature sensitivities to these two independent
variables are significant, and they have half the magnitude as the
sensitivities to previous tier’s independent variables.
Sliding contact area average temperature has its lowest sensitivity
to thermal conductivity of the sliding square contacting area. This
sensitivity is two orders of magnitude lower than the sensitivity to the
size of the contact area.
231
REFERENCES
[1] Carslaw, H. S. and Jaeger, J. C., “Conduction Of Heat In Solids,” Second

Edition, Clarendon Press, Oxford, 1986.
[2] Eckert, E. R. G. and Drake, Jr., R. M., “Heat And Mass Transfer,”
McGrawHill Book Company, Inc., New York, 1959.
[3] Fraden J., “Handbook Of Modern Sensors,” Second Edition, Springer-
Verlag New York, Inc., New York, 1996.
[4] Guenin, B. “Conduction Heat Transfer In A Printed Circuit Board,”
Electronics Cooling Magazine, Volume 4, Number 2, 1998.
[5] Holman, J. P., “Heat Transfer,” Second Edition, McGraw-Hill Book
Company, New York, 1968.
[6] Incropera, F. P. and DeWitt, D. P., “Fundamentals Of Heat And Mass
Transfer,” Fourth Edition, John Wiley & Sons, New York, 1996.
[7] Jaeger, J. C., “Moving Sources Of Heat And The Temperature At Sliding
Contacts,” Journal And Proceedings Of The Royal Society Of New South
Wales, Volume 79, pp. 203–224, 1942.
[8] Jones, J. B. and Hawkins, G. A., “Engineering Thermodynamics,” John
Wiley & Sons, Inc., New York, 1960.
[9] Kline, S. J. and McClintock, F. A., “Describing Uncertainties in
Single-Sample Experiments,” Mechanical Engineering, Volume 75,
pp. 3–8, 1953.
233
[10] Kreith, F., “Principles Of Heat Transfer,” Second Edition, International

Textbook Company, Scranton, Pennsylvania, 1965.
[11] Kutz, M. “Heat Transfer Calculations,” Mc-Graw Hill, New York, 2005.
[12] Lee, W. J., Kim, Y., and Case, E. D., “The Effect Of Quenching Media
On The Heat Transfer Coefficient Of Polycrystalline Alumina,” Journal
Of Materials Science, Volume 28, pp. 2079–2083, 1993.
[13] Lin, S. and Jiang, Z., “ An Improved Quasi-Steady Analysis For Solving
Freezing Problems In A Plate, A Cylinder And A Sphere,” ASME
Journal Of Heat Transfer, Volume 125, pp. 1123–1128, 2003.
[14] London, A. L. and Seban, R. A., “Rate Of Ice Formation,” Transactions
Of The ASME, Volume 65, pp. 771–778, 1943.
[15] Roshenow, W. M. and Choi, H. Y., “Heat, Mass, And Momentum
Transfer,” Prentice-Hall, Inc., Englewood Cliffs, New Jersey, 1961.
[16] Schneider, P. J., “Conduction Heat Transfer,” Addison-Wesley
Publishing Company, Inc., Reading, Massachusetts, 1957.
[17] Timoshenko, S and Goodier, J. N., “Theory Of Elasticity,” Second
Edition, McGraw-Hill Book Company, Inc., New York, 1951.
234

Everyday Heat Transfer

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Everyday Heat Transfer

Hochgeladen von

Copyright:

Verfügbare Formate

EVERYDAY HEAT

INFORMATION CONTAINED IN THIS WORK HAS BEEN OBTAINED BY THE

ASME shall not be responsible for statements or opinions advanced in papers

Library of Congress Cataloging-in-Publication Data

Chapter 1 Heat Loss from Walls in a Typical House ............................. 5

Chapter 2 Conduction Heat Transfer in a Printed Circuit Board .... 13

Chapter 3 Heat Transfer from Combustion Chamber Walls.............. 25

Chapter 4 Heat Transfer from a Human Body During

Chapter 5 Efficiency of Rectangular Fins .............................................. 41

Chapter 6 Heat Transfer from a Hot Drawn Bar .................................. 51

Chapter 7 Maximum Current in an Open-Air Electrical Wire .......... 65

Chapter 8 Evaporation of Liquid Nitrogen in a

Chapter 9 Thermal Stress in a Pipe ........................................................ 85

Chapter 10 Heat Transfer in a Pipe with Uniform Heat

Chapter 11 Heat Transfer in an Active Infrared Sensor .................. 103

Chapter 12 Cooling of a Chip ................................................................. 113

Chapter 13 Cooling of a Chip Utilizing a Heat Sink

Chapter 14 Heat Transfer Analysis for Cooking in a Pot ................ 131

Chapter 15 Insulating a Water Pipe from Freezing ........................... 139

Chapter 16 Quenching of Steel Balls in Air Flow .............................. 147

Chapter 17 Quenching of Steel Balls in Oil......................................... 155

Chapter 18 Cooking Time for Turkey in an Oven .............................. 161

Chapter 19 Heat Generated in Pipe Flows due to Friction ............. 169

Chapter 20 Sizing an Active Solar Collector for a Pool ................... 179

Chapter 21 Heat Transfer in a Heat Exchanger ................................. 195

Chapter 22 Ice Formation on a Lake .................................................... 203

Chapter 23 Solidification in a Casting Mold ....................................... 213

Chapter 24 Average Temperature Rise in Sliding Surfaces

References .................................................................................................... 233

Index .............................................................................................................. 235

Everyday engineering problems in heat transfer can be very

analyze them in the region of interest. The results of such sensitivity

y = f(x1, x2, x3, … xn)

then the uncertainty U for the dependent variable can be

U = [(∂y/∂x1 u1)2 + (∂y/∂x2 u2)2 + (∂y/∂x3 u3)2 + … + (∂y/∂xn un)2]0.5

where ∂y/∂x1, ∂y/∂x2, ∂y/∂x3, …, ∂y/∂xn are the sensitivities of the

variables. For the derivation of the fundamental heat transfer

H eat loss from the vertical walls of a house is analyzed under

Q/A = hout (Tout – Touter wall outer surface) (1-5)

Q/A = (Tin – Tout)/[(1/hin) + (tinner wall/kinner wall) + (tinsulation/kinsulation)

The denominator in Eq. (1-6) represents all the thermal resistances

Q/A = (Tin – Tout)/[(1/hin) + Rinner wall tinner wall + Rinsulation tinsulation

where the deﬁnitions of the variables with their assumed nominal

Tin = 68ºF (inside temperature)

The heat loss through a wall due to changes in convection heat

Figure 1-3 Wall heat loss versus insulation R-value

Figure 1-4 Wall heat loss versus insulation thickness

∂(Q/A)/∂hout = (Tin – Tout)/{h2out [(1/hin) + Rinner wall tinner wall

Sensitivity of heat loss to the outside convection heat transfer

These two sensitivities are similar, as can be expected, since the

∂ (Q/A) / ∂t insulation, BTU/hr-ft2-in

material R-value and insulation thickness. Sensitivities approach

House Wall House Wall

C onduction heat transfer in printed circuit boards (PCBs) has been

Qin-plane = Q1i-p + Q2i-p + … + Qni-p (2-1)

W Σti kin-plane (TL=0 – TL=L)/L = W t1 k1 (TL=0 – TL=L)/L

In-plane conduction heat transfer in a PCB represents a parallel

(1/Rin-plane) = (1/R1) + (1/R2) + … + (1/Rn) where Ri = L/(kitiW) (2-3)

kin-plane = Σ(kiti)/Σti (2-4)

Through-thickness (t-t) conduction heat transfer in a PCB represents

Qthrough-thickness = Q1t-t = Q2t-t = … = Qnt-t (2-5)

which can be expanded into following equations:

W L kthrough-thickness (Tt=0 – Tt=Σti)/Σti = W L k1 (Tt=0 − Tt=t1)/t1