Analytical Heat Transfer

ANALYTICAL HEAT TRANSFER
Mihir Sen
Department of Aerospace and Mechanical Engineering
University of Notre Dame
Notre Dame, IN 46556
March 20, 2015
Preface
These are lecture notes for AME60634: Intermediate Heat Transfer, a second course on heat transfer
for undergraduate seniors and beginning graduate students. At this stage the student can begin to
apply knowledge of mathematics and computational methods to the problems of heat transfer.
Thus, in addition to some undergraduate knowledge of heat transfer, students taking this course are
expected to be familiar with vector algebra, linear algebra, ordinary differential equations, particle
and rigid-body dynamics, thermodynamics, and integral and differential analysis in fluid mechanics.
The use of computers is essential both for the purpose of computation as well as for display and
visualization of results.
At present these notes are in the process of being written; the student is encouraged to make
extensive use of the literature listed in the bibliography. The students are also expected to attempt
the problems at the end of each chapter to reinforce their learning.
I will be glad to receive comments on these notes, and have mistakes brought to my attention.
Mihir Sen
Department of Aerospace and Mechanical Engineering
University of Notre Dame
c by M. Sen, 2015.
Copyright
Contents
Preface
Review
1 Introductory heat transfer

1.1 Fundamentals . . . . . . . . . . . . . . . .
1.1.1 Definitions . . . . . . . . . . . . .
1.1.2 Energy balance . . . . . . . . . . .
1.1.3 States of matter . . . . . . . . . .
1.2 Conduction . . . . . . . . . . . . . . . . .
1.2.1 Governing equation . . . . . . . .
1.2.2 Fins . . . . . . . . . . . . . . . . .
1.2.3 Separation of variables . . . . . . .
1.2.4 Similarity variable . . . . . . . . .
1.2.5 Lumped-parameter approximation
1.3 Convection . . . . . . . . . . . . . . . . .
1.3.1 Governing equations . . . . . . . .
1.3.2 Flat-plate boundary-layer theory .
1.3.3 Heat transfer coefficients . . . . . .
1.4 Radiation . . . . . . . . . . . . . . . . . .
1.4.1 Electromagnetic radiation . . . . .
1.4.2 View factors . . . . . . . . . . . .
1.5 Boiling and condensation . . . . . . . . .
1.5.1 Boiling curve . . . . . . . . . . . .
1.5.2 Critical heat flux . . . . . . . . . .
1.5.3 Film boiling . . . . . . . . . . . . .
1.5.4 Condensation . . . . . . . . . . . .
1.6 Heat exchangers . . . . . . . . . . . . . .
1.6.1 Parallel- and counter-flow . . . . .
1.6.2 HX relations . . . . . . . . . . . .
1.6.3 Design methodology . . . . . . . .
1.6.4 Correlations . . . . . . . . . . . . .
1.6.5 Extended surfaces . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . .
ii
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2
2
2
2
3
3
3
3
5
6
6
7
9
9
9
9
10
11
11
11
11
11
11
11
13
14
15
16
16
16
CONTENTS
II
iii
No spatial dimension
18
2 Steady and unsteady states

2.1 Electrical analog (steady) . . . . . . . . . . . . . . . . .
2.2 Relaxation . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . .
2.4 Electrical analog (unsteady) . . . . . . . . . . . . . . . .
2.5 Variable heat transfer coefficient . . . . . . . . . . . . .
2.5.1 Radiative cooling . . . . . . . . . . . . . . . . . .
2.5.2 Convective with weak radiation . . . . . . . . . .
2.6 Radiation in an enclosure . . . . . . . . . . . . . . . . .
2.7 Long time behavior . . . . . . . . . . . . . . . . . . . . .
2.7.1 Linear analysis . . . . . . . . . . . . . . . . . . .
2.7.2 Nonlinear analysis . . . . . . . . . . . . . . . . .
2.8 Time-dependent T . . . . . . . . . . . . . . . . . . . .
2.8.1 Linear . . . . . . . . . . . . . . . . . . . . . . . .
2.8.2 Oscillatory . . . . . . . . . . . . . . . . . . . . .
2.9 Two-fluid problem . . . . . . . . . . . . . . . . . . . . .
2.10 Two-body problem . . . . . . . . . . . . . . . . . . . . .
2.10.1 Convective . . . . . . . . . . . . . . . . . . . . .
2.10.2 Radiative . . . . . . . . . . . . . . . . . . . . . .
2.11 Linear systems . . . . . . . . . . . . . . . . . . . . . . .
2.11.1 First-order constant coefficient model . . . . . .
2.11.2 First-order variable coefficient model . . . . . . .
2.11.3 Higher-order constant coefficient model . . . . .
2.11.4 Stretched exponential function . . . . . . . . . .
2.12 Fractional derivatives . . . . . . . . . . . . . . . . . . . .
2.12.1 Experiments with shell-and-tube heat exchanger
2.12.2 Fractional time derivative in heat equation . . .
2.12.3 Self-similar trees . . . . . . . . . . . . . . . . . .
2.12.4 Continued fractions . . . . . . . . . . . . . . . .
2.12.5 Mittag-Leffler functions . . . . . . . . . . . . . .
2.13 Nonlinear systems . . . . . . . . . . . . . . . . . . . . .
2.13.1 Separable model . . . . . . . . . . . . . . . . . .
2.13.2 Lyapunov stability . . . . . . . . . . . . . . . . .
2.13.3 Power-law cooling . . . . . . . . . . . . . . . . .
2.14 Models with heating . . . . . . . . . . . . . . . . . . . .
2.14.1 Constant heating . . . . . . . . . . . . . . . . . .
2.14.2 Periodic heating . . . . . . . . . . . . . . . . . .
2.14.3 Delayed effect of heating . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
19
19
20
20
20
21
23
24
24
25
25
26
26
27
27
28
30
30
30
31
31
31
31
34
34
35
38
39
41
43
43
43
43
43
45
47
47
49
49
3 Control
3.1 Introduction . . . . . . . . . . .
3.2 Systems . . . . . . . . . . . . .
3.2.1 Systems without control
3.2.2 Systems with control .
3.3 Linear systems theory . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
50
50
51
51
52
53
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
CONTENTS
iv
3.3.1 Ordinary differential equations

3.3.2 Algebraic-differential equations
3.4 Nonlinear aspects . . . . . . . . . . . .
3.4.1 Models . . . . . . . . . . . . .
3.4.2 Controllability and reachability
3.4.3 Bounded variables . . . . . . .
3.4.4 Relay and hysteresis . . . . . .
3.5 System identification . . . . . . . . . .
3.6 Control strategies . . . . . . . . . . .
3.6.1 Mathematical model . . . . . .
3.6.2 On-off control . . . . . . . . . .
3.6.3 PID control . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
53
54
54
54
55
55
55
55
56
56
57
58
58
4 Physics of conduction
4.1 Heat carriers . . . . . . . . . . . . . .
4.1.1 Free electrons and holes . . . .
4.1.2 Phonons . . . . . . . . . . . . .
4.1.3 Material particles . . . . . . . .
4.1.4 Photons . . . . . . . . . . . . .
4.2 Maxwell-Boltzmann distribution . . .
4.3 Plancks radiation law . . . . . . . . .
4.4 Diffusion by random walk . . . . . . .
4.4.1 Brownian motion . . . . . . . .
4.4.2 One-dimensional . . . . . . . .
4.4.3 Multi-dimensional . . . . . . .
4.5 Phonons . . . . . . . . . . . . . . . . .
4.5.1 Single atom type . . . . . . . .
4.5.2 Two atom types . . . . . . . .
4.5.3 Types of phonons . . . . . . . .
4.5.4 Heat transport . . . . . . . . .
4.6 Molecular dynamics . . . . . . . . . .
4.7 Thin films . . . . . . . . . . . . . . . .
4.8 Boltzmann transport equation . . . . .
4.8.1 Relaxation-time approximation
4.9 Interactions and collisions . . . . . . .
4.10 Moments of the BTE . . . . . . . . . .
4.11 Onsager reciprocity . . . . . . . . . . .
4.12 Rarefied gas dynamics . . . . . . . . .
References . . . . . . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
61
61
61
61
61
61
61
62
63
63
63
64
64
64
65
66
67
67
67
67
67
68
69
69
69
69
69
III
One spatial dimension
5 Conduction
5.1 Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2 Fin theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
70
71
71
71
CONTENTS
5.2.1 Long time solution . . . . . . . . . . .

5.2.2 Shape optimization of convective fin .
5.3 Fin structure . . . . . . . . . . . . . . . . . .
5.4 Fin with convection and radiation . . . . . .
5.4.1 Annular fin . . . . . . . . . . . . . . .
5.5 Perturbations of one-dimensional conduction
5.5.1 Temperature-dependent conductivity .
5.5.2 Eccentric annulus . . . . . . . . . . . .
5.6 Transient conduction . . . . . . . . . . . . . .
5.7 Greens functions . . . . . . . . . . . . . . . .
5.8 Duhamels principle . . . . . . . . . . . . . .
5.9 Linear diffusion . . . . . . . . . . . . . . . . .
5.10 Nonlinear diffusion . . . . . . . . . . . . . . .
5.11 Stability by energy method . . . . . . . . . .
5.11.1 Linear . . . . . . . . . . . . . . . . . .
5.11.2 Nonlinear . . . . . . . . . . . . . . . .
5.12 Self-similar structures . . . . . . . . . . . . .
5.13 Non-Cartesian coordinates . . . . . . . . . . .
5.14 Thermal control . . . . . . . . . . . . . . . .
5.15 Multiple scales . . . . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
71
72
74
74
76
76
76
77
80
80
80
81
82
86
86
87
87
88
88
90
90
6 Forced convection
6.1 Hydrodynamics . . . . . . . . . . . . . . . . . . . . .
6.1.1 Mass conservation . . . . . . . . . . . . . . .
6.1.2 Momentum equation . . . . . . . . . . . . . .
6.1.3 Long time behavior . . . . . . . . . . . . . . .
6.2 Energy equation . . . . . . . . . . . . . . . . . . . .
6.2.1 Known heat rate . . . . . . . . . . . . . . . .
6.2.2 Convection with known outside temperature
6.3 Single duct . . . . . . . . . . . . . . . . . . . . . . .
6.3.1 Steady state . . . . . . . . . . . . . . . . . . .
6.3.2 Unsteady dynamics . . . . . . . . . . . . . . .
6.3.3 Perfectly insulated duct . . . . . . . . . . . .
6.3.4 Constant ambient temperature . . . . . . . .
6.3.5 Periodic inlet and ambient temperature . . .
6.3.6 Effect of wall . . . . . . . . . . . . . . . . . .
6.4 Two-fluid configuration . . . . . . . . . . . . . . . .
6.5 Flow between plates with viscous dissipation . . . .
6.6 Regenerator . . . . . . . . . . . . . . . . . . . . . . .
6.7 Radial flow between disks . . . . . . . . . . . . . . .
6.8 Networks . . . . . . . . . . . . . . . . . . . . . . . .
6.8.1 Hydrodynamics . . . . . . . . . . . . . . . . .
6.8.2 Thermal networks . . . . . . . . . . . . . . .
6.9 Thermal control . . . . . . . . . . . . . . . . . . . .
6.9.1 Control with heat transfer coefficient . . . . .
6.9.2 Multiple room temperatures . . . . . . . . . .
6.9.3 Two rooms . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
92
92
92
92
95
95
96
97
97
98
99
100
100
100
102
103
103
104
105
106
107
109
110
110
110
111
CONTENTS
vi
6.9.4 Temperature in long duct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
7 Natural convection
7.1 Modeling . . . . . . . . . . . . . . . . .
7.1.1 Mass conservation . . . . . . . .
7.1.2 Momentum equation . . . . . . .
7.1.3 Energy equation . . . . . . . . .
7.2 Known heat rate . . . . . . . . . . . . .
7.2.1 Steady state, no axial conduction
7.2.2 Axial conduction effects . . . . .
7.2.3 Toroidal geometry . . . . . . . .
7.2.4 Dynamic analysis . . . . . . . . .
7.2.5 Nonlinear analysis . . . . . . . .
7.3 Known wall temperature . . . . . . . . .
7.4 Mixed condition . . . . . . . . . . . . .
7.4.1 Modeling . . . . . . . . . . . . .
7.4.2 Steady State . . . . . . . . . . .
7.4.3 Dynamic Analysis . . . . . . . .
7.4.4 Nonlinear analysis . . . . . . . .
7.5 Perturbation of one-dimensional flow . .
7.6 Thermal control . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
115
115
115
116
117
118
118
122
126
131
138
147
148
149
150
154
158
160
160
160
8 Moving boundary
8.1 Stefan problems . . . . . . .
8.1.1 Neumanns solution
8.1.2 Goodmans integral
Problems . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
167
167
167
168
168
IV
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Multiple spatial dimensions
9 Conduction
9.1 Steady-state problems . . .
9.2 Transient problems . . . . .
9.2.1 Two-dimensional fin
9.3 Radiating fins . . . . . . . .
9.4 Non-Cartesian coordinates .
Problems . . . . . . . . . . . . .
169
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
170
170
170
170
171
171
171
10 Forced convection
10.1 Low Reynolds numbers . . . .
10.2 Potential flow . . . . . . . . .
10.2.1 Two-dimensional flow
10.3 Leveques solution . . . . . .
10.4 Multiple solutions . . . . . .
10.5 Plate heat exchangers . . . .
10.6 Falkner-Skan boundary flows
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
173
173
173
173
173
173
174
177
CONTENTS
vii
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
11 Natural convection
11.1 Governing equations .
11.2 Cavities . . . . . . . .
11.3 Marangoni convection
Problems . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
178
178
178
178
178
12 Porous media
12.1 Governing equations . . . . . . . . . . . . .
12.1.1 Darcys equation . . . . . . . . . . .
12.1.2 Forchheimers equation . . . . . . .
12.1.3 Brinkmans equation . . . . . . . . .
12.1.4 Energy equation . . . . . . . . . . .
12.2 Forced convection . . . . . . . . . . . . . . .
12.2.1 Plane wall at constant temperature .
12.2.2 Stagnation-point flow . . . . . . . .
12.2.3 Thermal wakes . . . . . . . . . . . .
12.3 Natural convection . . . . . . . . . . . . . .
12.3.1 Linear stability . . . . . . . . . . . .
12.3.2 Steady-state inclined layer solutions
Problems . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
179
179
179
179
180
180
180
180
183
183
185
185
188
193
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
13 Moving boundary
194
13.1 Stefan problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
Complex systems
195
14 Radiation
196
14.1 Monte Carlo methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
14.2 Equation of radiative transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
15 Boiling and condensation
197
15.1 Homogeneous nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
16 Bioheat transfer
198
16.1 Mathematical models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
17 Heat exchangers
17.1 Fin analysis . . . . . . . . . .
17.2 Porous medium analogy . . .
17.3 Heat transfer augmentation .
17.4 Maldistribution effects . . . .
17.5 Microchannel heat exchangers
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
199
199
199
199
199
199
CONTENTS
viii
17.6 Radiation effects . . . . . . . . . . .

17.7 Transient behavior . . . . . . . . . .
17.8 Correlations . . . . . . . . . . . . . .
17.8.1 Least squares method . . . .
17.9 Compressible flow . . . . . . . . . .
17.10Thermal control . . . . . . . . . . .
17.11Control of complex thermal systems
17.11.1 Hydronic networks . . . . . .
17.11.2 Other applications . . . . . .
17.12Conclusions . . . . . . . . . . . . . .
Problems . . . . . . . . . . . . . . . . . .
18 Soft computing
18.1 Genetic algorithms . . . .
18.2 Artificial neural networks
18.2.1 Heat exchangers .
Problems . . . . . . . . . . . .
VI
Appendices
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
199
199
200
200
200
200
202
202
202
205
205
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
206
206
206
206
206
209
A Additional problems
210
Index
258
Part I
Review
Chapter 1
Introductory heat transfer

It is assumed that the reader has had an introductory course in heat transfer of the level of [?, ?, ?,
?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?]. More advanced books are, for example, [?, ?]. A
classic work is that of Jakob [?].
1.1
Fundamentals
1.1.1
Definitions
Temperature is associated with the motion of molecules within a material, being directly related to
the kinetic energy of the molecules, including vibrational and rotational motion. Heat is the energy
transferred between two points at different temperatures. The laws of thermodynamics govern the
transfer of heat. Two bodies are in thermal equilibrium with each other if there is no transfer of
heat between them. The zeroth law states that if each of two bodies are in thermal equilibrium with
a third, then they also are in equilibrium with each other. Both heat transfer and work transfer
increase the internal energy of the body. The change in internal energy can be written in terms of
a coefficient of specific heat1 as M c dT . According to the first law, the increase in internal energy
is equal to the net heat and work transferred in. The third law says that the entropy of an isolated
system cannot decrease over time.
Example 1.1
Show that the above statement of the third law implies that heat is always transferred from a high
temperature to a low.
1.1.2
Energy balance
The first law gives a quantitative relation between the heat and work input to a system. If there is
no work transfer, then
Mc
dT
=Q
dt
1 Since we will usually not consider materials with change of volume under stress, we will not distinguish between
the specific heat at constant pressure and that at constant volume.
1.2. Conduction
where Q is the heat rate over the surface of the body. A surface cannot store energy, so that the
heat flux coming in must be equal to that going out.
1.1.3
States of matter
We will be dealing with solids, liquids and gases as well as the transformation of on to the other.
Again, thermodynamics dictates the rules under which these changes are possible. For the moment,
we will define the enthalpy of transformation2 as the change in enthalpy that occurs when matter is
transformed from one state to another.
1.2
Conduction
[?, ?, ?, ?, ?, ?, ?, ?]
The Fourier law of conduction is
q = kT
where q is the heat flux vector, T (x) is the temperature field, and k(T ) is the coefficient of thermal
conductivity.
1.2.1
Governing equation
T
= (k T ) + g
t
1.2.2
Fins
Fin effectiveness f : This is the ratio of the fin heat transfer rate to the rate that would be if
the fin were not there.
Fin efficiency f : This is the ratio of the fin heat transfer rate to the rate that would be if
the entire fin were at the base temperature.
Longitudinal heat flux
qx = O(ks
Tb T
)
L
Transverse heat flux

qt = O(h(Tb T ))
The transverse heat flux can be neglected compared to the longitudinal if
qx qt
which gives a condition on the Biot number
Bi =
hL
1
k
1.2. Conduction
T
b
x
L
Figure 1.1: Schematic of a fin.
Consider the fin shown shown in Fig. 1.1. The energy flows are indicated in Fig. 1.2. The
conductive heat flow along the fin, the convective heat loss from the side, and the radiative loss from
the side are
dT
dx
qh = hdAs (T T )
qk = ks A
4
qr = dAs (T 4 T
)
where, for a small enough slope, P (x) dAs /dx is the perimeter. Heat balance gives
Ac
T
qk
+
dx + qh + qr = 0
t
x
(1.1)
from which
Ac
T
4
ks (A
) + P h(T T ) + P (T 4 T
)=0
t
x x
where ks is taken to be a constant.

The initial temperature is T (x, 0) = Ti (x). Usually the base temperature Tb is known. The
different types of boundary conditions for the tip are:
Convective: T /x = a at x = L
Adiabatic: T /x = 0 at x = L
Known tip temperature: T = TL at x = L
Long fin: T = T as x
2 Also
called the latent heat.
1.2. Conduction
q (x)+q (x)
r
h
q (x+dx)
k
q (x)
k
Figure 1.2: Energy balance.

Taking
=
=
=
a() =
p() =
T T
Tb T
ks t
Fourier modulus
L2 c
x
L
A
Ab
P
Pb
where the subscript indicates quantities at the base, the fin equation becomes
a
+ m2 p + p ( + )4 4 = 0
a
where
Pb hL2
ks Ab
Pb L2 (Tb T )3
=
ks Ab
T
=
Tb T
m2 =
1.2.3
Separation of variables
Steady-state coduction in a rectangular plate.

2 T = 0
Let T (x, y) = X(x)Y (y).
1.2. Conduction
1.2.4
Similarity variable
T
2T
= 2
t
x
1.2.5
Lumped-parameter approximation
T ,1
Tw,1
Tw,2
T ,2
Figure 1.3: Wall with fluids on either side.

Consider a wall with fluid on both sides as shown in Fig. 1.3. The fluid temperatures are
T,1 and T,2 and the wall temperatures are Tw,1 and Tw,2 . The initial temperature in the wall is
T (x, 0) = f (x).
Steady state
In the steady state, we have
h1 (T,1 Tw,1 ) = ks
Tw,1 Tw,2
= h2 (Tw,2 T,2 )
L
from which
Tw,1 Tw,2
h2 L Tw,2 T,2
h1 L T,1 Tw,1
=
=
ks T,1 T,2
T,1 T,2
ks T,1 T,2
Thus we have
T,1 Tw,1
Tw,1 Tw,2
T,1 T,2
T,1 T,2
Tw,1 Tw,2
Tw,2 T,2
T,1 T,2
T,1 T,2
The Biot number is defined as
Bi =
hL
ks
h1 L
1
ks
h2 L
if
1
ks
if
1.3. Convection
Transient
T
ks 2 T
=
t
c x2
There are two time scales: the short (conductive) tk0 = L2 c/ks and the long (convective) th0 = Lc/h.
In the short time scale conduction within the slab is important, and convection from the sides is
not. In the long scale, the temperature within the slab is uniform, and changes due to convection.
The ratio of the two tk0 /th0 = Bi. In the long time scale it is possible to show that
Ls c
dT
+ h1 (T T,1 ) + h2 (T T,2 ) = 0
dt
where T = Tw,1 = Tw,2 .

Convective cooling
A body at temperature T , such as that shown in Fig. 1.4, is placed in an environment of different
temperature, T , and is being convectively cooled. The governing equation is
Mc
dT
+ hA(T T ) = 0
dt
(1.2)
with T (0) = Ti . We nondimensionalize using

T T
Ti T
hAt
=
Mc
=
(1.3)
(1.4)
The nondimensional form of the governing equation (1.2) is

d
+ =0
d
the solution to which is
= e
(1.5)
This is shown in Fig. 1.5 where the nondimensional temperature goes from = 1 to = 0. The
dimensional time constant is M c/hA.
1.3
Convection
[?, ?, ?, ?, ?, ?, ?, ?, ?, ?].
Newtons law of cooling: The rate of convective heat transfer from a body is proportional the
difference in temperature between the body and the surrounding fluid. Thus, we can write
q = hA(Tb Tf ),
where A is the surface area of the body, Tb is its temperature, Tf is that of the fluid, and h is the
coefficient of thermal convection.
1.3. Convection
T
T
Figure 1.4: Convective cooling.
1.4. Radiation
1.3.1
Governing equations
For incompressible flow

V =0
+ V V = p + 2 V + f
t
T
+ V T = (k T ) +
c
t
1.3.2
Flat-plate boundary-layer theory
Forced convection
Natural convection
1.3.3
Heat transfer coefficients
Overall heat transfer coefficient

Fouling
Bulk temperature
Nondimensional groups
UL
Prandtl number =
hL
Nusselt number N u =
k
Stanton number St = N u/P r Re
Reynolds number Re =
Colburn j-factor j = St P r2/3

2w
Friction factor f =
U 2
1.4
Radiation
[?, ?, ?, ?]
Emission can be from a surface or volumetric. Monochromatic radiation is at a single wavelength. The direction distribution of radiation from a surface may be either specular (i.e. mirror-like
with angles of incidence and reflection equal) or diffuse (i.e. equal in all directions).
The spectral intensity of emission is the radiant energy leaving per unit time, unit area, unit
wavelength, and unit solid angle. The emissive power is the emission of an entire hemisphere.
Irradiations is the radiant energy coming in, while the radiosity is the energy leaving including the
emission plus the reflection.
The absorptivity , the reflectivity , and transmissivity are all functions of the wavelkength
. Also
+ + = 1
1.4. Radiation
10
Integrating over all wavelengths

++ =1
The emissivity is defined as
=
E (, T )
Eb (, T )
where the numerator is the actual energy emitted and the denominator is that that would have been
emitted by a blackbody at the same temperature. For the overall energy, we have a similar definition
=
E(T )
Eb (T )
so that the emission is

E = T 4
For a gray body is independent of .
Kirchhoffs law: = and = .
1.4.1
Electromagnetic radiation
[?]
Electromagnetic radiation travels at the speed of light c = 2.998 108 m/s. Thermal radiation
is the part of the spectrum in the 0.1100 m range. The frequency f and wavelength of a wave
are related by
c = f
The radiation can also be considered a particles called phonons with energy
E = ~f
where ~ is Plancks constant.
Maxwells equations of electromagnetic theory are
D
t
B
E=
t
D=
H=J+
B=0
where H, B, E, D, J, and are the magnetic intensity, magnetic induction, electric field, electric
displacement, current density, and charge density, respectively. For linear materials D = E, J = gE
(Ohms law), and B = H, where is the permittivity, g is the electrical conductivity, and is the
permeability. For free space = 8.8542 1012 C2 N1 m2 , and = 1.2566 106 NC2 s2 ,
For = 0 and constant , g and , it can be shown that
H
2H
g
=0
t2
t
E
2E
2 E 2 g
=0
t
t
The speed of an electromagnetic wave in free space is c = 1/ .

2 H
1.5. Boiling and condensation
11
Blackbody radiation
Planck distribution [?]
E =
C1
5 [exp (C2 /T ) 1]
Wiens law: Putting dE /d = 0, the maximum of is seen to be at = m , where

m T = C3
and C3 = 2897.8 mK.
Stefan-Boltzmanns law: The total radiation emitted is
Z
E d
Eb =
0
= T 4
where = 5.670 108 W/m2 K4 .

d2 T
= 2 T 4
dx2
1.4.2
View factors
1.5
Boiling and condensation
[?, ?, ?]
1.5.1
Boiling curve
1.5.2
Critical heat flux
1.5.3
Film boiling
1.5.4
Condensation
Nusselts solution
1.6
Heat exchangers
[?, ?]
Shell and tube heat exchangers are commonly used for large industrial applications. Compact
heat exchangers are also common in industrial and engineering applications that exchanger heat
between two separated fluids. The term compact is understood to mean a surface to volume ratio
of more than about 700 m2 /m3 . The advantages are savings in cost, weight and volume of the heat
exchanger.
The fin efficiency concept was introduced by Harper and Brown (1922). The effectiveness-N T U
method was introduced by London and Seban in 1941.
A possible classification of HXs is shown in Table 1.1.
1.6. Heat exchangers
According to
Transfer processes
Surface
compactness
Construction
Flow arrangement
Number of fluids
Heat transfer
12
Table 1.1: Classification of HX (due to Shah [?], 1981

Types of HXs
Examples
Direct contact
Indirect contact
(a) direct transfer,
(b) storage, (c)
fluidized bed
Compact
Non-compact
Tubular
(a) double pipe
(b) shell and tube
(c) spiral tube
Plate
(a) gasketed,
(b) spiral,
(c) lamella
Extended surface (a) plate fin,
(b) tube fin
Regenerative
(a) rotary disk
(b) rotary drum
(c) fixed matrix
Single pass
(a) parallel flow
(b) counterflow
(c) crossflow
Multipass
(a) extended surface
cross counter flow,
(b) extended surface
cross parallel flow,
(c) shell and tube
parallel counterflow
shell and tube
mixed, (d) shell
and tube split
flow, (e) shell and
tube divided flow
Plate
Two fluid
Three fluid
Multifluid
Single-phase convection mechanisms on both sides
Single-phase convection on one side, two-phase
convection on other side
Two-phase convection on both sides
Combined convection and radiative heat transfer
13
cold
(a) parallel flow
hot
cold
(a) counter flow
hot
Figure 1.6: Parallel and counter flow.

1.6.1
Parallel- and counter-flow
We define the subscripts h and c to mean hot and cold fluids, i and o for inlet and outlet, 1 the end
where the hot fluids enters, and 2 the other end. Energy balances give
dq = U (Th Tc ) dA
dq = m
c Cc dTc
dq = m
h Ch dTh
(1.6)
(1.7)
(1.8)
where the upper and lower signs are for parallel and counterflow, respectively. From equations (1.7)
and (1.8), we get
1
1
dq
= d(Th Tc )
(1.9)
m
h Ch
m
c Cc
Using (1.6), we find that
U dA
1
1
m
h Ch
m
c Cc
d(Th Tc )
Th Tc
which can be integrated from 1 to 2 to give
1
(Th Tc )1
1
U A
= ln
m
h Ch
m
c Cc
(Th Tc )2
From equation (1.9), we get
qT
1
1
+
m
h Ch
m
c Cc
= (Th Tc )2 (Th Tc )1
where qT is the total heat transfer rate. The last two equations can be combined to give
qT = U ATlmtd
14
where
Tlmtd =
(Th Tc )1 (Th Tc )2
ln[(Th Tc )1 /(Th Tc )2 ]
is the logarithmic mean temperature difference.

For parallel flow, we have
Tlmtd =
(Th,i Tc,i ) (Th,o Tc,o )

ln[(Th,i Tc,i )/(Th,o Tc,o )]
Tlmtd =
(Th,i Tc,o ) (Th,o Tc,i )

ln[(Th,i Tc,o )/(Th,o Tc,i )]
while for counterflow it is
We an write the element of area dA in terms of the perimeter P as dA = P dx, so that

q(x)
m
c Cc
q(x)
Th (x) = Th,1
m
h Ch
Tc (x) = Tc,1
Thus
dq
+ qU P
dx
1
1
m
h Ch
m
c Cc
With the boundary condition q(0) = 0, the solution is
1
Th,1 Tc,1
1
1 exp U P
q(x) = 1
1
m
h Ch
m
c Cc
m
h Ch m
c Cc
1.6.2
HX relations
The HX effectiveness is
=
Qmax
Ch (Th,i Th,o )
=
Cmin (Th,i Tc,i )
Cc (Tc,o Tc,i )
=
Cmin (Th,i Tc,i )
where
Cmin = min(Ch , Cc )
Assuming U to be a constant, the number of transfer units is
NTU =
AU
Cmin
The heat capacity rate ratio is CR = Cmin /Cmax .
15
Effectiveness-N T U relations
In general, the effectiveness is a function of the HX configuration, its N T U and the CR of the fluids.
(a) Counterflow
=
1 exp[N T U (1 CR )]
1 CR exp[N T U (1 CR )]
so that 1 as N T U .
(b) Parallel flow
=
1 exp[N T U (1 CR )]
1 + CR
(c) Crossflow, both fluids unmixed

Series solution (Mason, 1954)
(d) Crossflow, one fluid mixed, the other unmixed
If the unmixed fluid has C = Cmin , then
= 1 exp[CR (1 exp{N T U Cr })]
But if the mixed fluid has C = Cmin
= CR (1 exp{CR (1 eN T U )})
(e) Crossflow, both fluids mixed
(f) Tube with wall temperature constant
= 1 exp(N T U )
Pressure drop
It is important to determine the pressure drop through a heat exchanger. This is given by
G2
1
A1
1
p
=
(Kc + 1 2 ) + 2( 1) + f
(1 2 Ke )
p1
21 p1
2
Ac m
2
where Kc and Ke are the entrance and exit loss coefficients, and is the ratio of free-flow area to
frontal area.
1.6.3
Design methodology
Mean temperature-difference method

Given the inlet temperatures and flow rates, this method enables one to find the outlet temperatures,
the mean temperature difference, and then the heat rate.
Effectiveness-NTU method
The order of calculation is N T U , , qmax and q.
1.6.4
Correlations
1.6.5
Extended surfaces
16
0 = 1
Af
(1 f )
A
where 0 is the total surface temperature effectiveness, f is the fin temperature effectiveness, Af is
the HX total fin area, and A is the HX total heat transfer area.
Problems
1. For a perimeter corresponding to a fin slope that is not small, derive Eq. 1.1.
2. The two sides of a plane wall are at temperatures T1 and T2 . The thermal conductivity varies with temperature
in the form k(T ) = k0 + (T T1 ). Find the temperature distribution within the wall.
3. Consider a cylindrical pin fin of diameter D and length L. The base is at temperature Tb and the tip at T ;
the ambient temperature is also T . Find the steady-state temperature distribution in the fin, its effectiveness,
and its efficiency. Assume that there is only convection but no radiation.
4. Show that the efficiency of the triangular fin shown in Fig. A.36 is
f =
1 I1 (2mL)
,
mL I0 (2mL)
where m = (2h/kt)1/2 , and I0 and I1 are the zeroth- and first-order Bessel functions of the first kind.
t
L
Figure 1.7: Triangular fin.

5. A constant-area fin between surfaces at temperatures T1 and T2 is shown in Fig. A.37. If the external temperature, T (x), is a function of the coordinate x, find the general steady-state solution of the fin temperature
T (x) in terms of a Greens function.
x
T1
T (x)
T2
Figure 1.8: Constant-area fin.

6. Using a lumped parameter approximation for a vertical flat plate undergoing laminar, natural convection, show
that the temperature of the plate, T (t), is governed by
dT
+ (T T )5/4 = 0
dt
Find T (t) if T (0) = T0 .
17
7. Show that the governing equation in Problem 3 with radiation can be written as
d2 T
4
m2 (T T ) (T 4 T
) = 0.
dx2
Find a two-term perturbation solution for T (x) if 1 and L .
8. The fin in Problem 3 has a non-uniform diameter of the form

D = D0 + x.
Determine the equations to be solved for a two-term perturbation solution for T (x) if 1.
Part II
No spatial dimension
18
Chapter 2
Steady and unsteady states

2.1
Electrical analog (steady)
In Fig. 2.1, the resistances are:

Rcond =
=
=
Rconv =
Rrad =
L
kA
ln(r2 /r1 )
2Lk
(1/r1 ) (1/r2 )
2k
1
,
hA
1
,
hr A
conduction (Cartesian) ,
conduction (cylindrical),
conduction (spherical),
convection,
linearized rediation with hr = (T Ts )(T 2 + Ts2 ).
T,1
R1
T1
T2
R2
T,2
R3
Figure 2.1: Electrical analog of steady heat transfer.
19
2.2. Relaxation
2.2
20
Relaxation
This is the return of a perturbed system to a time-independent state, as shown in Fig. 2.2. The
time constants tc of both falling and rising states are defined by
Z
T (t) T ()
tc =
dt.
T (0) T ()
0
0.8
0.6
0.4
0.2
0
0
t
Figure 2.2: Two typical relaxation responses
2.3
Experiment
Temperature change in a room shown in Fig. 2.31 . Possibilities for discrepancy from an exponential
are (a) higher-order relaxation, or (b) nonlinearities.
2.4
Electrical analog (unsteady)
The voltage across each one of the following electrical components in Fig. 2.4 is
VR = iR resistance R,
Z
1
i(t) dt,
capacitance C,
VC =
C
d2 i
VL = L 2 , inductance L.
dt
1 Ch
avez
& Sen (2014).
2.5. Variable heat transfer coefficient
21
Figure 2.3: Temperature measurement compared to exponential.

c
Ti
R
C
Figure 2.4: Electrical analog of unsteady heat transfer.

The total voltage V (t) applied across the circuit must be equal to the voltage drops across each
one of the components. Thus
Z
1
iR +
i(t) dt = V,
C
from which
di
1
+ i=
dt C
1
1
di
+
i=
dt RC
R
R
dV
,
dt
dV
.
dt
Eqs. (2.9) and (2.1) are analogs of each other, with i, RC and
2.5
(2.1)
1 dV
R dt
corresponding to T , tc , and Q.
Variable heat transfer coefficient
If, however, the h is slightly temperature-dependent, then we have

d
+ (1 + ) = 0
d
(2.2)
22
which can be solved by the method of perturbations. We assume that

( ) = 0 ( ) + 1 ( ) + 2 2 ( ) + . . .
(2.3)
To order 0 , we have
d0
+ 0 = 0
d
0 (0) = 1
which has the solution
0 = e
To the next order 1 , we get
d1
+ 1 = 02
d
= e2
1 (0) = 0
the solution to which is

1 = et + e2
Taking the expansion to order 2
d2
+ 2 = 20 1
d
= 2e2t 2e3
2 (0) = 1
with the solution
2 = e 2e2 + e3
And so on. Combining, we get
= e (e e2 ) + 2 (e 2e2 + e3 ) + . . .
Alternatively, we can find an exact solution to equation (2.2). Separating variables, we get
d
= d
(1 + )
Integrating
ln
= + C
+ 1
The condition (0) = 1 gives C = ln(1 + ), so that

ln
(1 + )
=
1 +
23
This can be rearranged to give

=
e
1 + (1 e )
A Taylor-series expansion of the exact solution gives
1
= e 1 + (1 e
= e 1 (1 e ) + 2 (1 e )2 + . . .
= e (e e2 ) + 2 (e 2e2 + e3 ) + . . .
2.5.1
Radiative cooling
If the heat loss is due to radiation, we can write

Mc
dT
4
+ A(T 4 T
)=0
dt
Taking the dimensionless temperature to be defined in Eq. (1.3), and time to be

=
A(Ti T )3 t
Mc
and introducing the parameter

=
T
Ti T
(2.4)
we get
d
+ 4 = 4
d
where
=+
Writing the equation as
d
= d
4 4
the integral is
1
ln
4 3
1
3 tan1
2
= + C
Using the initial condition (0) = 1, we get (?)
1 1 ( + T )( 1)
T 1
1
=
ln
+ tan
2 3 2 ( T )( + 1)
+ (T /)
2.6. Radiation in an enclosure
2.5.2
24
Convective with weak radiation
The governing equationis

Mc
dT
4
+ hA(T T ) + A(T 4 T
)=0
dt
with T (0) = Ti . Using the variables defined by Eqs. (1.3) and (1.4), we get
d
+ + ( + 4 )4 4 = 0
d
(2.5)
where is defined in Eq. (2.4), and
(Ti T )3
h
If radiative effects are small compared to the convective (for Ti T = 100 K and h = 10 W/m2 K
we get = 5.67 103 ), we can take 1. Substituting the perturbation series, Eq. (2.3), in
Eq. (2.5), we get

4
d
0 + 1 + 2 2 + . . . + 0 + 1 + 2 2 + . . . + 0 + 1 + 2 2 + . . .
d
+4 0 + 1 + 2 2 + . . . + 6 2 0 + 1 + 2 2 + . . . 4 3 0 + 1 + 2 2 + . . . = 0
In this case
d
+ ( 0 ) + (4 s4 ) = 0
d
(0) = 1
As a special case, of we take = 0, i.e. T = 0, we get
d
+ + 4 = 0
d
which has an exact solution
=
2.6
1
1 + 3
ln
3 (1 + )3
Radiation in an enclosure
Consider a closed enclosure with N walls radiating to each other and with a central heater H. The
walls have no other heat loss and have different masses and specific heats. The governing equations
are
Mi ci
N
X
dTi
4
Ai Fij (Ti4 Tj4 ) + Ai FiH (Ti4 TH
)=0
+
dt
j=1
where the view factor Fij is the fraction of radiation leaving surface i that falls on j. The steady
state is
T i = TH
(i = 1, . . . , N )
2.7. Long time behavior
25
Linear stability is determined by a small perturbation of the type

Ti = TH + Ti
from which
Mi ci
N
X
dTi
3
3
+ 4TH
Ai Fij (Ti Tj ) + 4TH
Ai FiH Ti = 0
dt
j=1
This can be written as

M
2.7
dT
3
= 4TH
AT
dt
Long time behavior
The general form of the equation for heat loss from a body with internal heat generation is
d
+ f () = a
d
with (0) = 1. Let
f () = a
Then we would like to show that as t . Writing
= +
we have
d
+ f ( + ) = a
d
2.7.1
Linear analysis
If we assume that is small, then a Taylor series gives

f ( + ) = f () + f () + . . .
from which
d
+ b = 0
d
where b = f (). The solution is
= Ceb
so that 0 as t if b > 0.
(2.6)
2.8. Time-dependent T
2.7.2
26
Nonlinear analysis
Multiplying Eq. (2.6) by , we get
1 d 2
( ) = f ( + ) f ()
2 d
Thus
d 2
( ) 0
d
if and [f ( + ) f ()], as shown in Fig. 2.5, are both of the same sign or zero. Thus 0 as
.
3/2
a=b
+
P stable
Both P +and P
unstable
G/2
_
P stable
3/2
a=-b
2.8
Time-dependent T
Let
Mc
dT
+ hA(T T (t)) = 0
dt
with
T (0) = Ti
2.8. Time-dependent T
2.8.1
27
Linear
Let
T = T,0 + at
Defining the nondimensional temperature as
T T,0
Ti T,0
(2.7)
d
+ = A
d
(2.8)
=
and time as in Eq. (1.4), we get
where
A=
aM c
hA(Ti T,0 )
The nondimensional ambient temperature is

=
aM c
hA(Ti T,0 )
Cilm The solution to Eq. (2.8) is

= Ce + A A
The condition (0) = 1 gives C = 1 + A, so that
= (1 + A)e + A( 1)
The time shown in Fig. 2.6 at crossover is
c = ln
1+A
A
and the offset is

= A
as .
2.8.2
Oscillatory
Let
T = T + T sin t
where T (0) = Ti . Defining
=
T T
Ti T
2.9. Two-fluid problem
28
Figure 2.6: Response to linear ambient temperature.

and using Eq. (1.4), the nondimensional equation is
d
+ = sin
d
where
T
Ti T
M c
=
hA
The solution is
= Ce +
sin( )
(1 + 2 ) cos
where
= tan1
From the condition (0) = 1, we get C = 1 + /(1 + 2 ), so that
e +
sin( )
= 1 +
1 + 2
(1 + 2 ) cos
2.9
Two-fluid problem
Suppose there is a body in contact with two fluids at different temperatures T,1 and T,2 , like in
the two examples shown in Fig. 2.7. The governing equation is
Mc
dT
+ h1 A1 (T T,1 ) + h2 A2 (T T,2 ) = 0
dt
2.9. Two-fluid problem
29
T ,2
T
T ,2
T ,1
T ,1
(b)
(a)
Figure 2.7: Two-fluid problems.

where T (0) = Ti . If T,1 and T,2 are constants, we can nondimensionalize the equation using the
parameters for one of them, fluid 1 for instance. Thus we have
T T,1
Ti T,1
h1 A1 t
=
Mc
=
from which
d
+ + ( + ) = 0
d
with (0) = 1, where
h2 A2
h1 A1
T,1 T,2
=
Ti T,1
The equation can be written as

d
+ (1 + ) =
d
with the solution
= Ce(1+)
1+
The condition (0) = 1 gives C = 1 + /(1 + ), from which
e(1+)
= 1+
1+
1+
For = 0, the solution reduces to the single-fluid case, Eq. (1.5). Otherwise the time constant of
the general system is
t0 =
Mc
h1 A1 + h2 A2
2.10. Two-body problem
30
2.10
Two-body problem
2.10.1
Convective
Suppose now that there are two bodies at temperatures T1 and T2 in thermal contact with each
other and exchanging heat with a single fluid at temperature T as shown in Fig. 2.8.
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Figure 2.8: Two bodies in thermal contact.

The mathematical model of the thermal process is
dT1
ks Ac
+
(T1 T2 ) + hA(T1 T ) = Q1
dt
L
ks Ac
dT2
+
(T2 T1 ) + hA(T2 T ) = Q2
M2 c2
dt
L
M1 c1
2.10.2
Radiative
111111111111
000000000000
000000000000
111111111111
Figure 2.9: Bodies with radiation.
dT1
ks Ac
+
(T1 T2 ) + A1 F1s (T14 Ts4 ) + A1 F12 (T14 T24 ) = Q1
dt
L
ks Ac
dT2
+
(T2 T1 ) + +A1 F2s (T24 Ts4 ) + A2 F21 (T24 T14 ) = Q2
M2 c2
dt
L
M1 c1
Without radiation
Q1 = Q2 =
2kA
T1 T2
L
2.11. Linear systems
31
2.11
Linear systems
2.11.1
First-order constant coefficient model

dT
T
+
= Q(t).
dt
tc
(2.9)
where tc is the time constant.

Solution
T = aet/tc
2.11.2
First-order variable coefficient model
Linear equation
dT
+ f ( )T = 0.
d
R
Multiplying by the integrating factor e

R
f (s) ds
f (s) ds
gives
dT
+ f ( )T = 0,
d
R
dT R f (s) ds
e0
+ e 0 f (s) ds f ( ) T = 0,
d
i
R
d h
T e 0 f (s) ds = 0.
d
Integrating
T e
f (s) ds
T =ae
T 0 as if
2.11.3
R
0
R
0
= a,
f (s) ds
f (s) ds > 0 for all .
Higher-order constant coefficient model

(D + 1 )(D + 2 ) . . . (D + n )y = 0,
This is equivalent to the system of equations

(D + n )y = yn1
(D + n1 )yn1 = yn2
(D + n2 )yn2 = yn3
..
.
(D + 2 )y2 = y1
(D + 1 )y1 = 0
32
The solution is
y = a1 e1 t + a2 e2 t + . . . + an en t
Example 2.1
Let
y1 = e1 t ,
1 t
y2 = ae
(2.10a)
2 t
+ +(1 a)e
(2.10b)
where 1 = 1, 2 = 0.1, a = 0.9.
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
10
t
Figure 2.10: Multiple time constants; Eq. (2.10a) is continuous line, Eq. (2.10b) is dashed line.
Example 2.2
Let ai = a/2i1 , i = /2i1 for i = 1, 2, . . . , n and y(0) = 1 , then

y = aet +
n1
a t/2
a
e
+ . . . + n1 et/2
.
2
2
From the initial condition

a
a
1 = a + + . . . + n1 ,
2
2
`
= 2a 1 2n ,
33
since
n
X
ar i1 =
i=1
Thus
a=
Alternative formulation
D + 1
a(1 r n )
.
1r
1
2 (1 2n )
n Dn + n1 Dn1 + . . . + 1 D + 0 y = 0
D + 2
0
..
D + n1
y1
y2
.
.
.
1 yn1
yn
D + n
Example 2.3
See Fig. 2.11.
Figure 2.11: Boxes.

dT1
1
1
(T1 T ) = Q,
+ (T1 T2 ) +
dt
t1
t
1
1
dT2
+ (T2 T1 ) + (T2 T3 ) = 0,
dt
t1
t2
dT3
1
1
+ (T3 T2 ) + (T3 T4 ) = 0,
dt
t2
t3
..
.
1
1
dTn1
+
(Tn1 Tn2 ) +
(Tn1 Tn ) = 0,
dt
tn2
tn1
dTn
1
+
(Tn Tn1 ) = 0.
dt
tn1

0
0

..
= .

0
0
2.12. Fractional derivatives
34
The characteristic equation of the homogeneous equation is

Compare this to the PDE
T
T
1
k(r)r 2
= 2
t
r r
r
Example 2.4
Example: Second-order system
(D + 1 )(D + 2 )y = 0
Solution is
y = a1 e1 t + a2 e2 t
y(t) 0 as t if 1 and 2 are real and positive.
2.11.4
Stretched exponential function
y(t) = et
is a solution of
dy
+ t1 y = 0.
dt
2.12
Fractional derivatives
Cauchys formula for repeated integration is

Z n1
Z x Z 1
f (n ) dn . . . d1 ,
...
In =
a
a
a
Z x
1
=
(x y)n1 f (y) dy.
(n 1)! a
Generalize n to real number, and factorial to gamma function to get the Riemann-Liouville fractional
integral
Z x
1
f (t)
IRL f (x) =
dt.
() a (x t)1
Differentiating n times we have the Riemann-Liouville fractional derivative
Z x
f (t)
dn
1
dt, n 1 < n.
DRL
f (x) =
(n ) dxn a (x t)1+n
35
Figure 2.12: Fractional derivatives DRL

f (x) of x2 ; (a) = 0.2, 0.4, 0.6, 0.8, 1.0.
1, 1.2, 1.4, 1.6, 1.8, 2.0. From Schumer et al. (2001)
(b) =
The Caputo fractional derivative is
DC
f (x)
1
=
(n )
f (n) (t)
dt,
(x t)1+n
n 1 < n.
The Gr
unwald-Letnikov fractional derivative is
1
h0 h
DG
f (x) = lim
(xa)/h
m=0
(1)m
( + 1)
f (x mh).
m! ( m + 1)
Examples of fractional derivative are in Fig. 2.12.
2.12.1
Experiments with shell-and-tube heat exchanger
For this experiment [?], a shell-and-tube heat exchanger, shown in Fig. 2.13, was tested and the
inlet and outlet temperatures on both the cold and hot side were recorded. The inlet flow rates
were held constant in both the cold and hot sides and the inlet temperatures were controlled to
remain constant. After the system reached equilibrium for a given set of conditions, a step change
36
Tc,o
Th,o
Th,i
Tc,i
Figure 2.13: Shell-and-tube heat exchanger.
(on-off) in the flow rate of the cold-side was applied. Experimental results of two tests are shown in
Fig. 2.14, where the non-dimensional temperature is given by
(t) =
and time is non-dimensionalized by
Th,o (t) Th,o (0)

,
Th,o () Th,o (0)
=
t
,
r
where r is the rise time, defined as the time required for Th,o (t) to reach 85% of Th,o ().
The approximations are the following.
first-order
second-order
fractional-order
Data Set #1
Model
First-order
Second-order
Fractional-order
= 1.8196
c1
1.0401
0.097011
0.22381
0.24855
Data Set #2
Model
First-order
Second-order
Fractional-order
= 1.8196
c1
1.2536
0.12359
0.33183
0.2976
c1 D1 y(t) + c2 y(t) = u(t)

2
(2.11a)
c1 D y(t) + c2 D y(t) + c3 y(t) = u(t)

c2
1
t E,+1 ( t )]
c1 D y(t) + c2 y(t) = u(t) [Sol.
c1
c1
c2
0.028554
0.59086
2.1700
2.0221
c2
-0.32904
0.65039
1.8782
2.0476
c3
0.89968
1.8911
c3
0.85564
1.7481
(2.11b)
(2.11c)
error2
3.5284e-03
7.6902e-04
9.5609e-05
1.3355e-04
P
error2
1.4135e-03
1.0973e-04
2.7757e-05
4.9404e-05
P
The three models proposed in Eq. 2.11 were used to find best fit approximations. In both data
sets, the fractional-order model resulted in a better approximation than either the first or second
order models. Data set #1 is shown to be fit best with a derivative of order = 1.8911 and data
set #2 with a derivative of order = 1.7481. The average = 1.8196 also outperforms both the
first and second order models, despite the same number of free parameters.
37
0.8
(t)
0.6
0.4
0.2
0
0
0.2
0.4
0.6
0.8
0.8
(a) Test data 1

1
0.8
(t)
0.6
0.4
0.2
0
0
0.2
0.4
0.6
(b) Test data 2
Figure 2.14: Best fit approximations to shell-and-tube data using first, second, and fractional-order
models. Dotted line is filtered data, dashed line is first-order model, dash-dot is second-order model,
and solid line is fractional-order model.
2.12.2
38
Fractional time derivative in heat equation
Consider a semi-infinite body, < x 0, shown in Fig. 2.15 [?]. The initial temperature is
T (0, x) = T0 , and the surface temperature is T (0, t) = Ts (t).
Ts
Figure 2.15: Semi-infinite body.

Writing (t, x) = T (t, x) T0 ,
2
= 2,
t
x
t 0,
< x 0,
(0, x) = 0,
(t, 0) = s (t),
lim (t, x) < 0.
b x), where2
The Laplace transform in time is (s,
d2 b
s b = 2 .
dx
The solution that is bounded for x is
From this
so that at x = 0
2 The
b 0) exp x s/ .
b = (s,
p
b b
s
= (s, 0)
exp x s/ ,
x
1 b
1 b
(s, 0) = (s,
0).
s x
Laplace transform of f (t) is F (s), and that of 0 Dt f (t) is s F (s)
Pn1
k=0
sk 0 Dt1k f (t)t=0 .
39
i=1
i=2
i=3
Z3,1
Z3,2
Z3,3
Z3,4
Z3,5
Z3,6
Z3,7
Z3,8
Z2,1
Z2,2
Z1,1
Vin
Z1,2
i=N
Z2,3
Z2,4
ZN,1
Vout
ZN,2N
Figure 2.16: n-generational tree.

The inverse transform of this gives
1/2
0 Dt
(t, 0) = (t, 0),

t
from which
1
1/2
(t, 0) = 0 Dt [(t, 0)] .
x
The heat flux at x = 0 is
(t, 0),
x
k
1/2
= 0 Dt
(t, 0) .
t
qs = k
Alternatively, since
1/2
Dt
1/2
Dt
= Dt1 ,
Dx1 Dx1 = Dx2 ,

we have
1/2
Dt
2.12.3
T =
Dx1
1
Dx T
Self-similar trees
See Fig. 2.16.

Potential-driven flow
V = iZ
where
V = potential difference
i = flow rate
Z = impedance
40
Z
(sC)1
Figure 2.17: Equivalent impedance for fractance.

c
Figure 2.18: CAPTION HERE.

The equivalent impedance for a capacitor-resistor tree [?] called fractance, is in Fig. 2.17.
1
1
1
=
+
1
Z
(sC) + Z
R+Z
from which
Z=
R
sC
1/2
Converting back from the Laplace domain,

V (t) =
(R/C)1/2
(1/2)
i( )
d,
(t )1/2
Potential-driven flow networks [?]

L(q) = p
qi,j = qi+1,2j1 + qi+1,2j
qi+1,2j1
qi,j
qi+1,2j
41
Overall behavior of tree

LN qN = pin pout ,
where
1
LN =
1
L1,1 +
1
1
1
+
L2,1 + . . . L2,2 + . . .
+
L1,2 +
1
1
1
+
L2,3 + . . . L2,4 + . . .
Self-symmetric trees
1
LN =
1
L1 +
L2 +
1
1
+
L1 + . . . L2 + . . .
1
1
1
+
L1 + . . . L2 + . . .
For N
L =
1
1
1
+
L1 + L
L2 + L
L =
2.12.4
L1 L2
Continued fractions
1
x=
1+
1+
=
1
1+x
x2 + x 1 = 0
1 + 5
2
x=
1 5
1
1 +
golden mean
42
1.4
1.2
LN (s)
0.8
n=
0.6
n=6
n=4
0.4
n=1
0
n=2
10
s
(a) L1 = s, L2 = 1
35
30
n=8
LN (s)
25
n=
20
n=4
15
n=2
10
n=1
5
0
0
10
s
(b) L1 = s + 1, L2 = s2 + 2s
Figure 2.19: L shown with LN for different values of N . As N increases, LN converges to L in

the Laplace domain while in the time domain LN converges to L .
2.13. Nonlinear systems
2.12.5
43
Mittag-Leffler functions
Define the two-parameter function

E, (z) =
k=0
zk
(k + )
Thus
E1,1 (z) =
k=0
2.13
Nonlinear systems
2.13.1
Separable model
X zk
zk
=
= ez .
(k + 1)
k!
k=0
d
+ f () = 0.
d
(2.12)
from which
d
= d.
f ()
Integrating
Z
2.13.2
d
= + a.
f ()
Lyapunov stability
From Eq. (2.12)

d
+ f () = 0,
d
d 2
= f (),
d
2
2 0+ for if f () > 0 for all .

2.13.3
Power-law cooling
Assume: (a) Constant properties. (b) Power law heat transfer.

Energy balance gives
Mc
dT
+ hA(T T ) = 0,
dt
d
hA
+
= 0,
dt
Mc
T (0) = T0
(2.13)
2.13. Nonlinear systems
44
where
= T T .
From the definitions
hL
,
k
gL3
Ra =
,
Nu =
and using the correlation

N u = C1 Ran ,
we have
k
N u,
L
C1 k n
=
Ra ,
L
n
C1 k gL3
n .
=
L
h=
Thus Eq. (2.13) becomes

d
+ C2 n+1 = 0,
dt
where C2 =
C1 kA
M cL
n1 d = C2 dt.
gL3
For n = 0, N u and h are constant. Upon integration

ln = C2 t + ln a,
= aeC2 t
With T (0) = T0 ,
a = T0 T ,
so that
T T
= eC2 t .
T0 T
= e ,
(2.14)
where
T T
,
T0 T
= C2 t.
(2.15a)
(2.15b)
2.14. Models with heating
45
However, for n 6= 0, integration gives
1 n
= C2 t a,
n
n = n (C5 t + a) ,
1/n
1/n
= n1/n (C2 t + a)
so that
T = T + n1/n (C2 t + a)
From the initial condition
T0 = T + (na)1/n ,
1
a = (T0 T )n ,
n
so that
1/n
1
T = T + n1/n C2 t + (T0 T )n
.
n
1/n
=n
1/n
1
+
n
= (n + n)
1/n
(2.16)
where
T T
,
T0 T
= C2 (T0 T )n t.
(2.17a)
(2.17b)
As n 0, the independent and dependent variables, Eq. (2.17), tend to Eq. (2.15) and, since
lim (n + 1)
1/n
n0
= e ,
the solution Eq. (2.16) tends to Eq. (2.14). Fig. 2.20 shows that, in spite of their different forms,
Eqs. (2.16) and (2.14) give very similar numerical results for small positive and negatives values of
n.
2.14
Models with heating

dT
Mc
+ hA(T T ) = G(t),
{z
} |{z}
| {zdt} |
I
II
III
T = T0 at t = 0,
46
0.8
0.6
0.4
0.2
0
0
0.5
1.5
Figure 2.20: Continuous line is Eq. (2.14); dashed and dotted lines are Eq. (2.16) with n = 0.1 and
0.1, respectively.
where the terms and their units are
I = rate of accumulation of internal energy,
II = rate of heat loss to surroundings,
III = rate of heating,
[(kg) (J/kgK) (K/s) = W],

[(W/m2 ) (m2 ) (K) = W],
[W].
Writing
T T
,
T0 T
Mc
t
,
= , tc =
tc
hA
=
we have
M c(T0 T ) d
+ hA(T0 T ) = G.
M c/hA
d
Dividing by hA(T0 T ), we get the nondimensional equation
d
+ = ,
d
where
( ) =
Mc
G.
hA(T0 T )
47
The solution of the homogeneous equation is

h = a e .
2.14.1
Constant heating
If G = G0 , i.e. = 0 , is constant, then the particular solution is

p = 0 .
The complete solution is
= h + p ,
= a e + 0 .
From the initial condition (0) = 1, we get
a = 1 0 ,
so that
= (1 0 ) e + 0 .
For or t , = 0 and T = T +
2.14.2
M cG0
hA .
Periodic heating
Let
( ) = 0 + 1 cos .
where the nondimensional and dimensional frequencies, and respectively, are related by = t.
Assume the particular solution to be
p = C1 + C2 sin + C3 cos .
Substituting in the equation, we get
(C2 cos t C3 sin ) + (C1 + C2 sin + C3 cos ) = 0 + 1 cos t.
Comparing coefficients
C1 = 0 ,
C2 + C3 = 1 ,
C3 + C2 = 0.
The constants are
C1 = 0 ,
1
,
C2 =
1 + 2
1
,
C3 =
1 + 2
48
so that the complete solution is

( ) = a e + 0 +
1
1
sin +
cos .
1 + 2
1 + 2
For long time, , only the particular solution is left.

We can write
1
1
sin +
cos = A cos( + ),
2
1+
1 + 2
= A cos cos sin sin ,
and compare terms to get
1
,
1 + 2
1
.
A cos =
1 + 2
A sin =
so that
tan = ,
A = 1 .
Alternatively
( ) = 0 + 1 ei .
Assuming
p = C1 + C2 ei .
then
C1 = 0 ,
1
C2 =
,
1 + i
so that
1 i
e ,
1 + i
1 i i
e ,
= 0 + 1
1 + 2
= 0 + 1 ei( +) .
p = 0 +
The complex number a + ib can be represented in polar form as (a2 + b2 )1/2 tan1 (b/a) so
that
= 0 + 1 ( ),
p = 0 + 1 ( + ).
49
Tmax
Tmin
Figure 2.21: Ambient temperature variation.

2.14.3
Delayed effect of heating

d
+ = ( d ).
d
Problems
1. Show that the temperature distribution in a sphere subject to convective cooling tends to become uniform as
Bi 0.
2. Check one of the perturbation solutions against a numerical solution.
3. Plot all real (, ) surfaces for the convection with radiation problem, and comment on the existence of
solutions.
4. Complete the problem of radiation in an enclosure (linear stability, numerical solutions).
5. Lumped system with convective-radiative cooling with nonzero 0 and s .
6. Find the steady-state temperatures for the two-body problem and explore the stability of the system for
constant ambient temperature.
7. Consider the change in temperature of a lumped system with convective heat transfer where the ambient
temperature, T (t), varies with time in the form shown. Find (a) the long-time solution of the system
temperature, T (t), and (b) the amplitude of oscillation of the system temperature, T (t), for a small period t.
Chapter 3
Control
3.1
Introduction
There are many kinds of thermal systems in common industrial, transportation and domestic use
that need to be controlled in some manner, and there are many ways in which that can be done.
One can give the example of heat exchangers [?,?], environmental control in buildings [?,?,?,?,?,?],
satellites [?, ?, ?, ?], thermal packaging of electronic components [?, ?], manufacturing [?], rapid
thermal processing of computer chips [?, ?, ?], and many others. If precise control is not required,
or if the process is very slow, control may simply be manual; otherwise some sort of mechanical or
electrical feedback system has to be put in place for it to be automatic.
Most thermal systems are generally complex involving diverse physical processes. These include
natural and forced convection, radiation, complex geometries, property variation with temperature,
nonlinearities and bifurcations, hydrodynamic instability, turbulence, multi-phase flows, or chemical
reaction. It is common to have large uncertainties in the values of heat transfer coefficients, approximations due to using lumped parameters instead of distributed temperature fields, or material
properties that may not be accurately known. In this context, a complex system can be defined as
one that is made up of sub-systems, each one of which can be analyzed and computed, but when put
together presents such a massive computational problem so as to be practically intractable. For this
reason large, commonly used engineering systems are hard to model exactly from first principles,
and even when this is possible the dynamic responses of the models are impossible to determine
computationally in real time. Most often some degree of approximation has to be made to the
mathematical model. Approximate correlations from empirical data are also heavily used in practice. The two major reasons for which control systems are needed to enable a thermal system to
function as desired are the approximations used during design and the existence of unpredictable
external and internal disturbances which were not taken into account.
There are many aspects of thermal control that will not be treated in this brief review. The most
important of these are hardware considerations; all kinds of sensors and actuators [?, ?] developed
for measurement and actuation are used in the control of thermal systems. Many controllers are
also computer based, and the use of microprocessors [?, ?] and PCs in machines, devices and plants
is commonplace. Flow control, which is closely related to and is often an integral part of thermal
control, has its own extensive literature [?]. Discrete-time (as opposed to continuous-time) systems
will not be described. The present paper will, however, concentrate only on the basic principles
relating to the theory of control as applied to thermal problems, but even then it will be impossible
to go into any depth within the space available. This is only an introduction, and the interested
50
3.2. Systems
51
w(t)
?
u(t)
-
plant
x(t)
y(t) -
w(t)
controller.
Figure 3.1: Schematic of a system
without
?
?
?
yr (t) - Cj e(t)- controller
u(t)
-
plant
x(t)
y(t)
-
Figure 3.2: Schematic of a system with comparator C and controller.

reader should look at the literature that is cited for further details. There are good texts and
monographs available on the basics of control theory [?, ?, ?, ?], process control [?, ?, ?, ?], nonlinear
control [?], infinite-dimensional systems [?,?], and mathematics of control [?,?] that can be consulted.
These are all topics that include and are included within thermal control.
3.2
Systems
Some basic ideas of systems will be defined here even though, because of the generality involved, it
is hard to be specific at this stage.
3.2.1
Systems without control
The dynamic behavior of any thermal system (often called the open-loop system or plant to distinguish it from the system with controller described below), schematically shown in Fig. 3.1, may be
mathematically represented as
Ls (x, u, w, ) = 0,
(3.1)
where Ls is a system operator, t is time, x(t) is the state of the system, u(t) is its input, w(t) is some
external or internal disturbance, and is a parameter that defines the system. Each one of these
quantities belongs to a suitable set or vector space and there are a large number of possibilities. For
example Ls may be an algebraic, integral, ordinary or partial differential operator, while x may be
a finite-dimensional vector or a function. u(t) is usually a low-dimensional vector.
In general, the output of the system y(t) is different from its state x(t). For example, x may be
a spatial temperature distribution, while y are the readings of one or more temperature measurement
devices at a finite number of locations. The relation between the two may be written as
Lo (y, x, u, w, ) = 0,
(3.2)
where Lo is the output operator.

The system is single-input single-output (SISO) if both u and y are scalars. A system is said
to be controllable if it can be taken from one specific state to another within a prescribed time with
the help of a suitable input. It is output controllable if the same can be done to the output. It is
3.2. Systems
52
important to point out that output controllability does not imply system controllability. In fact, in
practice for many thermal systems the former is all that is required; it has been reported that most
industrial plants are controlled quite satisfactorily though they are not system controllable [?]. All
possible values of the output constitute the reachable set. A system is said to be observable if its
state x can be uniquely determined from the input u and output y alone. The stability of a system
is a property that leads to a bounded output if the input is also bounded.
3.2.2
Systems with control
The objective of control is to have a given output y = yr (t), where the reference or set value yr is
prescribed. The problem is called regulation if yr is a constant, and tracking if it is function of time.
In open-loop control the input is selected to give the desired output without using any information from the output side; that is one would have to determine u(t) such that y = yr (t) using
the mathematical model of the system alone. This is a passive method of control that is used in
many thermal systems. It will work if the behavior of the system is exactly predictable, if precise
output control is not required, or if the output of the system is not very sensitive to the input. If
the changes desired in the output are very slow then manual control can be carried out, and that
is also frequently done. A self-controlling approach that is sometimes useful is to design the system
in such a way that any disturbance will bring the output back to the desired value; the output in
effect is then insensitive to changes in input or disturbances.
Open-loop control is not usually effective for many systems. For thermal systems contributing
factors are the uncertainties in the mathematical model of the plant and the presence of unpredictable
disturbances. Internal disturbances may be noise in the measuring or actuating devices or a change
in surface heat transfer characteristics due to fouling, while external ones may be a change in the
environmental temperature. For these cases closed-loop control is an appropriate alternative. This
is done using feedback from the output, as measured by a sensor, to the input side of the system, as
shown in Fig. 3.2; the figure actually shows unit feedback. There is a comparator which determines
the error signal e(t) = e(yr , y), which is usually taken to be
e = yr y.
(3.3)
The key role is played by the controller which puts out a signal that is used to move an actuator in
the plant.
Sensors that are commonly used are temperature-measuring devices such as thermocouples,
resistance thermometers or thermistors. The actuator can be a fan or a pump if the flow rate is to
be changed, or a heater if the heating rate is the appropriate variable. The controller itself is either
entirely mechanical if the system is not very complex, or is a digital processor with appropriate
software. In any case, it receives the error in the output e(t) and puts out an appropriate control
input u(t) that leads to the desired operation of the plant.
The control process can be written as
Lc (u, e, ) = 0,
(3.4)
where Lc is a control operator. The controller designer has to propose a suitable Lc , and then Eqs.
(A.13)(3.4) form a set of equations in the unknowns x(t), y(t) and u(t). Choice of a control strategy
defines Lo and many different methodologies are used in thermal systems. It is common to use on-off
(or bang-bang, relay, etc.) control. This is usually used in heating or cooling systems in which the
heat coming in or going out is reduced to zero when a predetermined temperature is reached and
set at a constant value at another temperature. Another method is Proportional-Integral-Derivative
3.3. Linear systems theory
53
(PID) control [?] in which

u = Kp e(t) + Ki
e(s) ds + Kd
3.3
de
.
dt
(3.5)
Linear systems theory
The term classical control is often used to refer to theory derived on the basis of Laplace transforms.
Since this is exclusively for linear systems, we will be using the so-called modern control or statespace analysis which is based on dynamical systems, mainly because it can be extended to nonlinear
systems. Where they overlap, the issue is only one of preference since the results are identical.
Control theory can be developed for different linear operators, and some of these are outlined below.
3.3.1
Ordinary differential equations
Much is known about a linear differential system in which Eqs. (A.13) and (3.2) take the form
dx
= Ax + Bu,
dt
y = Cx + Du,
(3.6)
(3.7)
where x Rn , u Rm , y Rp , A Rnn , B Rnm , C Rpn , D Rpm . x, u and y are

vectors of different lengths and A, B, C, and D are matrices of suitable sizes. Though A, B, C, and
D can be functions of time in general, here they will be treated as constants.
The solution of Eq. (3.6) is
x(t) = eA(tt0 ) x(t0 ) +
eA(ts) Bu(s) ds.
(3.8)
t0
where the exponential matrix is defined as

eAt = I + At +
1
1 2 2
A t + A3 t3 + . . . ,
2!
3!
with I being the identity matrix. From Eq. (3.7), we get
Z t
eA(ts) Bu(s) ds + Du.
y(t) = C eAt x(t0 ) +
(3.9)
t0
Eqs. (3.8) and (3.9) define the state x(t) and output y(t) if the input u(t) is given.
It can be shown that for the system governed by Eq. (3.6), a u(t) can be found to take x(t)
from x(t0 ) at t = t0 to x(tf ) = 0 at t = tf if and only if the matrix
..
.. 2 ..
.. n1
M = B . AB . A B . . . . . A
B Rnnm
(3.10)
is of rank n. The system is then controllable. For a linear system, controllability from one state to
another implies that the system can be taken from any state to any other. It must be emphasized
that the u(t) that does this is not unique.
3.4. Nonlinear aspects
54
Similarly, it can be shown that the output y(t) is
.
.
.
.
N = D .. CB .. CAB .. CA2 B .. . . .
controllable if and only if
..
. CAn1 B Rp(n+1)m
is of rank p. Also, the state x(t) is observable if and only if the matrix
T
..
..
..
n1
2
Rpnn
P = C . CA . CA . . . . CA
is of rank n.
3.3.2
Algebraic-differential equations
This is a system of equations of the form

E
dx
= A x + B u,
dt
(3.11)
where the matrix E Rnn is singular [?]. This is equivalent to a set of equations, some of which are
ordinary differential and the rest are algebraic. As a result of this, Eq. (3.11) cannot be converted
into (3.6) by substitution. The index of the system is the least number of differentiations of the
algebraic equations that is needed to get the form of Eq. (3.6). The system may not be completely
controllable since some of the components of x are algebraically related, but it may have restricted
or R-controllability [?].
3.4
Nonlinear aspects
The following are a few of the issues that arise in the treatment of nonlinear thermal control problems.
3.4.1
Models
There are no general mathematical models for thermal systems, but one that can be used is a
generalization of Eq. (3.6) such as
dx
= f (x, u).
dt
(3.12)
where f : Rn Rm 7 Rn . If one is interested in local behavior about an equilibrium state x = x0 ,

u = 0, this can be linearized in that neighborhood to give
dx
f f
=
x +
u
dt
x
u
0
= Ax + Bu ,
(3.13)
where x = x0 + x and u = u . The Jacobian matrices (f /x)0 and (f /u)0 , are evaluated at the
equilibrium point. Eq. (3.13) has the same form as Eq. (3.6).
3.5. System identification
3.4.2
55
Controllability and reachability
General theorems for the controllability of nonlinear systems are not available at this point in time.
Results obtained from the linearized equations generally do not hold for the nonlinear equations.
The reason is that in the nonlinear case one can take a path in state space that travels far from the
equilibrium point and then returns close to it. Thus regions of state space that are unreachable with
the linearized equations may actually be reachable. In a thermal convection loop it is possible to go
from one branch of a bifurcation solution to another in this fashion [?].
3.4.3
Bounded variables
In practice, due to hardware constraints it is common to have the physical variables confined to
certain ranges, so that variables such as x and u in Eqs. (A.13) and (3.2), being temperatures, heat
rates, flow rates and the like, are bounded. If this happens, even systems locally governed by Eqs.
(3.6) and (3.7) are now nonlinear since the sum of solutions may fall outside the range in which x
exists and thus may not be a valid solution. On the other hand, for a controllable system in which
only u is bounded in a neighborhood of zero, x can reach any point in Rn if the eigenvalues of A have
zero or positive real parts, and the origin is reachable if the eigenvalues of A have zero or negative
real parts [?].
3.4.4
Relay and hysteresis
A relay is an element of a system that has an input-output relationship that is not smooth; it may be
discontinuous or not possess first or higher-order derivatives. This may be accompanied by hysteresis
where the relationship also depends on whether the input is increasing or decreasing. Valves are
typical elements in flow systems that have this kind of behavior.
3.5
System identification
To be able to design appropriate control systems, one needs to have some idea of the dynamic
behavior of the thermal system that is being controlled. Mathematical models of these systems
can be obtained in two entirely different ways: from first principles using known physical laws, and
empirically from the analysis of experimental information (though combinations of the two paths
are not only possible but common). The latter is the process of system identification, by which
a complex system is reduced to mathematical form using experimental data [?, ?, ?]. There are
many different ways in which this can be done, the most common being the fitting of parameters
to proposed models [?]. In this method, a form of Ls is assumed with unknown parameter values.
Through optimization routines the values of the unknowns are chosen to obtain the best fit of the
results of the model with experimental information. Apart from the linear Eq. (3.6), other models
that are used include the following.
There are many forms based on Eq. (3.12), one of which is the closed-affine model
dx
= F1 (x) + F2 (x)u
dt
The bilinear equation for which F1 (x) = Ax and F2 (x) = N x + b is a special case of this.
3.6. Control strategies
Volterra models, like

y(t) = y0 (t) +
56
Z
X
i=1
...
ki (t; t1 , t2 , t3 , . . . , ti )u(t1 ) . . . u(ti ) dt1 . . . dti
for a SISO system, are also used.

Functional [?], difference [?] or delay [?] equations such as
dx
= A x(t s) + B u
dt
also appear in the modeling of thermal systems.
Fractional-order derivatives, of which there are several different possible definitions [?,?,?,?,?]
can be used in differential models. As an example, the Riemann-Liouville definition of the nth
derivative of f (t) is
Z t
1
dm+1
n
D
f
(t)
=
(t s)mn f (s) ds,
a t
(m n + 1) dtm+1 a
where a and n are real numbers and m is the largest integer smaller than n.
3.6
Control strategies
3.6.1
Mathematical model
Consider a body that is cooled from its surface by convection to the environment with a constant
ambient temperature T . It also has an internal heat source Q(t) to compensate for this heat loss,
and the control objective is to maintain the temperature of the body at a given level by manipulating
the heat source. The Biot number for the body is Bi = hL/k, where h is the convective heat transfer
coefficient, L is a characteristics length dimension of the body, and k is its thermal conductivity.
If Bi < 0.1, the body can be considered to have a uniform temperature T (t). Under this lumped
approximation the energy balance is given by
Mc
dT
+ hAs (T T ) = Q(t),
dt
(3.14)
where M is the mass of the body, c is its specific heat, and As is the surface area for convection.
Using M c/hAs and hAs (Ti T ) as the characteristic time and heat rate, this equation becomes
d
+ = Q(t)
dt
(3.15)
Here = (T T )/(Ti T ) where T (0) = Ti so that (0) = 1. The other variables are now nondimensional. With x = , u = Q, n = m = 1 in Eq. (3.6), we find from Eq. (3.10) that rank(M )=1,
so the system is controllable.
Open-loop operation to maintain a given non-dimensional temperature r is easily calculated.
Choosing Q = r , it can be shown from the solution of Eq. (3.15), that is
(t) = (1 r )et + r ,
that r as t . In practice, to do this the dimensional parameters hAs and T must be
exactly known. Since this is usually not the case some form of feedback control is required.
3.6.2
57
On-off control
In this simple form of control the heat rate in Eq. (3.14) has only two values; it is is either Q = Q0
or Q = 0, depending on whether the heater is on or off, respectively. With the system in its on
mode, T Tmax = T + Q0 /hAs as t , and in its off mode, T Tmin = T . Taking the
non-dimensional temperature to be
T Tmin
Tmax Tmin
=
the governing equation is
d
+ =
dt
1 on
,
0 off
(3.16)
the solution for which is

=
1 + C1 et
C2 et
on
.
off
(3.17)
We will assume that the heat source comes on when temperature falls below a value TL , and goes
off when it is rises above TU . These lower and upper bounds are non-dimensionally
TL Tmin
,
Tmax Tmin
TU Tmin
U =
.
Tmax Tmin
L =
The result of applying this form of control is an oscillatory temperature that looks like that
in Fig. 3.3, the period and amplitude of which can be chosen using suitable parameters. It can be
shown that the on and off time periods are
1 L
,
1 U
U
,
= ln
L
ton = ln
(3.18)
toff
(3.19)
respectively. The total period of the oscillation is then

tp = ln
U (1 L )
.
L (1 U )
If we make a small dead-band assumption, we can write

L = r ,
U = r + ,
where 1. A Taylor-series expansion gives
1
1
tp = 2
+
+ ...
r
1 r
The period is thus proportional to the width of the dead band. The frequency of the oscillation
increases as its amplitude decreases.
58
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
Figure 3.3: Lumped approximation with on-off control.

3.6.3
PID control
The error e = r and control input u = Q can be used in Eq. (3.5), so that the derivative of Eq.
(3.15) gives
(Kd + 1)
d2
d
+ (Kp + 1) + Ki = Ki r ,
dt2
dt
with the initial conditions

= 0 at t = 0,
Kp + 1
Kp r
d
=
0 +
dt
Kd + 1
Kd + 1
at
t = 0.
The response of the closed-loop system can be obtained as a solution. The steady-state for Ki 6= 0
is given by = r . It can be appreciated that different choices of the controller constants Kp , Ki
and Kd will give overdamped or underdamped oscillatory or unstable behavior of the system. Fig.
3.4 shows two examples of closed-loop behavior with different parameter values.
Problems
1. Two lumped bodies A and B in thermal contact (contact thermal resistance Rc ) exchange heat between
themselves by conduction and with the surroundings by convection. It is desired to control their temperatures
at TA and TB using separate internal heat inputs QA and QB .
(a) Check that the system is controllable.
(b) Set up a PID controller where its constants are matrices. Determine the condition for linear stability of
the control system. Show that the case of two independent bodies is recovered as Rc .
(c) Calculate and plot TA (t) and TB (t) for chosen values of the controller constants.
2. Apply an on-off controller to the previous problem. Plot TA (t) and TB (t) for selected values of the parameters.
Check for phase synchronization.
59
1.5
(b)
1
(a)
0.5
0.5
1.5
10
20
30
40
50
60
70
Figure 3.4: Lumped approximation with PID control; Ki = Kd = 0.1, r = 0.5, (a) Kp = 0.1,
(b) Kp = 0.9.
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
3. A number of identical rooms are arranged in a circle as shown, with each at a uniform temperature Ti (t). Each
room exchanges heat by convection with the outside which is at T , and with its neighbors with a conductive
thermal resistance R. To maintain temperatures, each room has a heater that is controlled by independent but
identical proportional controllers. (a) Derive the governing equations for the system, and nondimensionalize.
(b) Find the steady state temperatures. (c) Write the dynamical system in the form x = Ax and determine
the condition for stability1 .
1 Eigenvalues
of an N N , circulant, banded matrix of the form

2
b
c
0
... 0
6 a
b
c
... 0
6
6 0
a
b
... 0
6
6 .
.
.
..
..
.
.
6 ..
.
.
.
.
6
4 0 ...
0
a
b
c
0
...
0
a
3
a
0 7
7
0 7
7
7
7
7
c 5
b
are j = b + (a + c) cos{2(j 1)/N } i(a c) sin{2(j 1)/N }, where j = 1, 2, . . . , N .
60
T
i1
i
T
i+1
Chapter 4
Physics of conduction
4.1
Heat carriers
[3]
4.1.1
Free electrons and holes
Free electrons, as opposed to the valence electrons, move with an applied potential field, and contribute to transport of electrical current and heat.
4.1.2
Phonons
These are vibrations of atoms that propagate in a wave-like manner. There is more on this in Section
4.5.
4.1.3
Material particles
This transport is due to the bulk motion of a material which could be a solid, liquid or gas. For a
gas, the motion of the atoms is usually what transports heat. For a continuum, it is usually referred
to as advection, and will be discussed in detail in the Chapters on convection.
4.1.4
Photons
This is a particle that corresponds to an electromagnetic wave and is thus what enables radiation.
There is more on this in Section 4.3.
4.2
Maxwell-Boltzmann distribution
Different distributions apply for different types of particles. The one that applies to classical particles is the Maxwell-Boltzmann distribution: for a single oscillator in a system of oscillators, the
probability p(E) of possessing energy between E and E + dE is proportional to eE/kT dE.
61
4.3. Plancks radiation law
4.3
62
Plancks radiation law
Planck assumed that the energy of each oscillator is a multiple of ~, i.e. E = n~, so that
en~/kT
p(E) = P n~/kT ,
n=0 e
= (1 e~/kT ) en~/kT , since
The average energy of each mode of oscillation is

hEi =
yn =
n=0
1
.
1y
E p(E),
n=0
= (1 e~/kT )
~
e~/kT
n~en~/kT ,
k=0
= (1 e~/kT )~
=
e~/kT
,
(1 e~/kT )2
The number of vibrational modes per unit volume having frequency between and + d is
[To be clarified.]
N () d =
2
d.
2 c3
Thus
N ()hEi d =
2
~
d
2 c3 e~/kT 1
The energy density per unit volume per unit frequency is

S=
~
2
2 c3 e~/kT 1
The density above is for thermal equilibrium, so setting inward=outward gives a factor of 1/2 for
the radiated power outward. Then one must average over all angles, which gives another factor of
1/2 for the angular dependence which is the square of the cosine. Thus the radiated power per unit
area from a surface at this temperature is Sc/4. [To be clarified.]
Spectral radiance is
B (, T ) =
This can be converted to
B (, T ) =
1
2h 3
c2 eh/kT 1
per unit frequency.
2hc2
1
5 ehc/kT 1
per unit wavelength,
where c = and B (, T ) d = B (, T ) d.
1 Quantum
Mechanics with Applications, D.B. Beard and G.B. Beard, Allyn and Bacon, Boston, 1970
4.4. Diffusion by random walk
63
4.4
Diffusion by random walk
4.4.1
Brownian motion
Langevin equation
m
d2 x
dx
=
+ FR (t).
2
dt
dt
for a particle of mass m, position x that is undergoing a viscous drag of (dx/dt) and a random
force FR (t). At the two extremes of time, the solutions are
(a) Short time t m/: ballistic
hx2 i hx20 i =
kT 2
t
m
(b) Long time t m/: viscous fluid

hx2 i hx20 i =
2kT
t
The Fokker-Planck equation for the evolution of a probability density function P (x, t) is
P
1
+ (uP ) = 2 DP
t
2
4.4.2
One-dimensional
Consider a walker along an infinite straight line moving with a step size x taken forward or
backward with equal probability in a time interval t. Let the current position of the walker be
x, and the probability that the walker is in an interval [x dx/2, x + dx/2] be P (x, t). Since
there is equal probability of the walker in the previous time step to have been in the interval
[x x dx/2, x x + dx/2] or [x + x dx/2, x + x + dx/2], we have
1
[P (x x, t t) + P (x + x, t t] .
2
Assuming that x and t are small, Taylor series expansions give
P (x, t) =
(4.1)
P
P
x
t
x
t
2
2
1 P
P
1 2P 2
2
+
x
xt
+
t
2 x2
xt
2 t2
1 3P
1 3P
1 3P 3
1 3P
x3
x2 t
xt2
t
3
2
2
6 x
2 x t
2 xt
6 t3
P (x x, t t) = P
+...,
where the terms on the right side are evaluated at (x, t). Substituting in Eq. (4.1) we get
2P
1 2P
1 3P
P
D 2 =
t
x2 + . . . ,
t
x
2 t2
2 x2 t
where D = x2 /2t. If we let x 0 and t 0 such that D is constant, we get the diffusion
equation
P
2P
=D 2.
t
x
4.5. Phonons
64
Figure 4.1: Lattice of atoms of a single type

4.4.3
Multi-dimensional
Let P = P (r, t), and r be a step in any direction where |r| is a constant. Then
Z
P (r, t) =
P (r + r, t t) dS
S
where S is a sphere of radius |r| centered on r. Also

P (r + r, t t) = P +
P
1 2P
P
ri
t
+ ...
ri
t
2 ri ri
etc.
4.5
Phonons
Phonons are lattice vibrations that can be considered to be quasi-particles. [4] is a review of microscale heat exchangers, and [1] of photonic devices.
[2, 6, 912]
4.5.1
Single atom type
A lattice of atoms of a single type is shown in Fig. 4.1. The mass of each atom is m, the spring
constants are c, and a is the mean distance between the atoms. For a typical atom n, Newtons
second law gives
m
d2 xn
= c(xn+1 xn ) c(xn xn1 )
dt2
= c(xn+1 2xn + xn1 ).
Let
xi = x
bei(nkat) ,
then the dispersion relation is

=
2c
m
1/2
(1 cos ka)
1/2
The phase velocity is

vp =
2c
mk 2
1/2
(1 cos ka)
1/2
4.5. Phonons
65
xi
yi
Figure 4.2: Lattice of atoms of two different type

and the group velocity is
vg =
c 1/2
a sin ka
.
1/2
2m
(1 cos ka)
For ka 0, we have
vg = a
The thermal conductivity
c 1/2
.
m
k = ke + kp
where ke and kp are those due to electron and phonon transports. We can also write
k=
1
cvg l
3
where c is the specific heat, and l is the mean free path.

4.5.2
Two atom types
Newtons second law gives

m1
m2
d2 xi
= c(yi xi ) c(xi yi1 )
dt2
= c(yi 2xi + yi1 ),
d2 y i
= c(xi+1 yi ) c(yi xi )
dt2
= c(xi+1 2yi + xi ).
Let
xi = x
bei(nkat) ,
so that
yi = ybei(nkat) ,
m1 x
b 2 = c yb 2b
x + ybeika ,
ika
m2 yb 2 = c x
be 2b
y+x
b ,
4.5. Phonons
66
k
Figure 4.3: Schematic of dispersion relations; A = acoustic phonons, O = optical phonons.
which can also be written as
2c m1 2
c(1 + eika )
c(1 + eika )
2c m2 2
This means that
x
b
yb
0
0
(2c m1 2 )(2c m2 2 ) c2 (1 + eika )(1 + eika ) = 0,

which simplifies to
m1 m2 4 2c(m1 + m2 ) 2 + 2c2 (1 cos ka) = 0.
The solution is
q
1
2
2
=
2c(m1 + m2 ) 2c m1 + m2 + 2m1 m2 cos ka .
2m1 m2
2
The positive sign corresponds to the optical and the negative to the acoustic mode.
4.5.3
Types of phonons
Frequency
Optical: high frequency
Acoustic: low frequency
Direction of oscillation
Longitudinal: in direction of velocity
Transverse: normal to velocity
4.6. Molecular dynamics
4.5.4
67
Heat transport
Phonons contribute to the internal energy of a material. In the Debye model for a monotomic
crystallind solid
3 Z ~D /kT
kT
kT
x4 ex
cv,p = 9
dx,
n
m ~D
ex 1
0
where m is the mass of an atom, n is the number of atoms per unit volume, is the angular frequency
of the phonon, and D is the Debye frequency, i.e. the maximum frequency of vibration.
Phonons also participate in the transport of heat from one location to the other. The thermal
conductivity k is
k=
1
ncv u
3
where u is the average speed of the carrier, and is its mean free path.
4.6
Molecular dynamics
This is the technique of following individual molecules using Newtonian mechanics and appropriate
intermolecular forces. One such commonly used force field is provided by the Lennard-Jones potential

12 6
VLJ = 4
r
r
4.7
Thin films
4.8
Boltzmann transport equation
The classical distribution function f (r, v, t) is defined as number of particles in the volume dr dv in
the six-dimensional space of coordinates r and velocity v. Following a volume element in this space,
we have the balance equation [5, 7, 8], usually called the BTE:

f
f
f
,
(4.2)
+ v f + a
=
t
v
t scat
where a = dv/dt is the acceleration due to an external force. The term on the right side is due to
collisions and scattering. The heat flux is then
Z
q(r, t) = v(r, t)f (r, , t)D d,
(4.3)
where D() is the density of energy states .
4.8.1
Relaxation-time approximation
Under this approximation
f
t
=
scat
f0 f
,
(4.4)
4.9. Interactions and collisions
68
where = (r, v). Thus

f
f
f0 f
+ v f + a
=
.
t
v
(4.5)
Several further approximations can be made.

(a) Fouriers law
Assume /t = 0, a = 0 and f = f0 in the left side of Eq. (4.5) so that
f = f0 v f0 .
(4.6)
Introducing explicitly the dependence of f on temperature, we can write

f0 =
df0
T.
dT
(4.7)
Using Eqs. (4.6) and (4.7) in (4.3), we get

q = kT,
(4.8)
where
k=
df0
v v
dT
vf0 D d = 0.
D d,
(4.9)
since
(b) Cattaneos equation

Assume = constant and f = (df0 /dT )T . Multiply Eq. (4.5) by vD d and integrate to
get
Z
Z
Z
1
df0
T vD d =
v
vf D d +
vf D d.
t
dT
Using Eq. (4.3), this gives

q+
q
= kT,
t
where k is given by Eq. (4.9). This is Cattaneos equation that can be compared to Fouriers law,
Eq. (4.8).
4.9
Interactions and collisions
The heat carriers can all interact or collide with each other and with themselves, leaading to a
contribution to the (f /t)scat term in the BTE. In addition solids have impurities, grain bvoundaries
and dislocations with which they can interact as well.
4.10. Moments of the BTE
4.10
69
Moments of the BTE
Eq. (4.2) can be multiplied by vn to get the n-th moment of the BTE. These are the conservation
equations: conservation of mass for n = 0, of momentum for n = 1, of mechanical energy for n = 2,
and so on.
4.11
Onsager reciprocity
J =
L f
where L is the Onsager matrix of phenomenological coefficients.
4.12
Rarefied gas dynamics
Knudsen number
Kn =
References
[1] G. Chen. Heat transfer in micro- and nanoscale photonic devices. In C.-L. Tien, editor, Annual Review of Heat
Transfer, volume 7, chapter 1. Begell House, New York, 1996.
[2] H. Ibach and H. L
uth. Solid-State Physics. Springer, New York, 1990.
[3] M. Kaviany. Heat Transfer Physics. Cambridge University Press, New York, 2008.
[4] M.-H. Kim, S.Y. Lee, S.S. Mehendale, and R.L. Webb. Microscale heat exchanger design for evaporator and
condenser applications. In J.P. Hartnett, T.F. Irvine, Y.I. Cho, and G.A. Greene, editors, Advances in Heat
Transfer, volume 37, pages 297429. Academic Press, Amsterdam, 2003.
[5] C. Kittel and H. Kroemer. Thermal Physics. W.H. Freeman & Co., San Francisco, 2nd edition, 1980.
[6] W.W. Liou and Y. Fang. Microfluid Mechanics: Principles and Modeling. McGraw Hill, New York, 2006.
[7] A. Majumdar. Microscale energy transport in solids. In C.-L. Tien, A. Majumdar, and F.M. Gerner, editors,
Microscale Energy Transport, chapter 1. Taylor and Francis, Washington, DC, 1998.
[8] A. Majumdar. Microscale transport phenomena. In W.M. Rohsenow, J.P Hartnett, and Y.I. Cho, editors,
Handbook of Heat Transfer, chapter 8. McGraw-Hill, New York, 1998.
[9] H. Smith and H.H. Jensen. Transport Phenomena. Clarendon Press, Oxford, U.K., 1989.
[10] P. Tabeling. Introduction to Microfluidics. Oxford University Press, Oxford, UK, 2005.
[11] D.Y. Tzou. Macro- to Microscale Heat Transfer: The Lagging Behavior. Taylor and Francis, Washington, DC,
1997.
[12] Z.M. Zhang, C.J. Fu, and Q.Z. Zhu. Optical and thermal radiative properties of semiconductors related to
micro/nanotechnology. In J.P. Hartnett, T.F. Irvine, Y.I. Cho, and G.A. Greene, editors, Advances in Heat
Transfer, volume 37, pages 179296. Academic Press, Amsterdam, 2003.
Problems
Part III
One spatial dimension
70
Chapter 5
Conduction
5.1
Structures
Fig. 5.1 shows a complex shape consisting of conductive bars. At each node
X
qi = 0
i
For each branch

X ki Ai
Li
from which
(Ti T0 ) = 0
P
T0 = P
5.2
Fin theory
5.2.1
Long time solution
k i Ai
Li Ti
k i Ai
i Li
+ f () = 0
Figure 5.1: Complex conductive structures.

71
(5.1)
5.2. Fin theory
72
where f () includes heat transfer from the sides due to convection and radiation. The boundary
conditions are either Dirichlet or Neumannn type at = 0 and = 1. The steady state is determined
from
d
d
(5.2)
a
+ f () = 0
d
d
with the same boundary conditions. Substituting = + in equation (5.1) and subtracting (5.2),
we have
a
+ f ( + ) f () = 0
where is the perturbation from the steady state. The boundary conditions for are homogeneous.
Multiplying by and integrating from = 0 to = 1, we have
dE
= I1 + I2
d
(5.3)
where
Z
1 1
a( )2 d
E=
2 0
Z 1
I1 =
a
d
0
Z 1
f ( + ) f () d
I2 =
(5.4)
(5.5)
(5.6)
Integrating by parts we can show that
1 Z
2
1

d
I1 = a
d
a

d
0
0
Z 1 2
d
d
a
=
d
0
(5.7)
(5.8)
since the first term on the right side of equation (5.7) is zero due to boundary conditions. Thus we
know from the above that I1 is nonpositive and from equation (5.4) that E is nonnegative. If we
also assume that
I2 0
(5.9)
then equation (5.3) tells us that E must decrease with time until reaching zero. Thus the steady
state is globally stable. Condition (5.9) holds if [ and f ( + ) f ()] are of the same sign or both
zero; this is a consequence of the Second Law of Thermodynamics.
5.2.2
Shape optimization of convective fin
Consider a rectangular fin of length L and thickness as shown in Fig. 5.2. The dimensional equation
is
d2 T
m2 (T T ) = 0
dx2
5.2. Fin theory
73
T
b
Figure 5.2: Rectangular fin.

where
m=
2h
ks
1/2
We will take the boundary conditions

T (0) = Tb
dT
(L) = 0
dx
The solution is
T = T (T T ) [tanh mL sinh mx cosh mx]
The heat rate through the base per unit width is
dT
q = ks
dx
x=0
= ks (Tb T )m tanh mL
Writing L = Ap /, we get
q = ks
2h
ks
1/2
"
Ap
(Tb T ) tanh
2h
ks
1/2 #
Keeping Ap constant, i.e. constant fin volume, the heat rate can be maximized if
"
1/2 #
2h
2h
2h 1/2 3 5/2 1 1/2
Ap
1/2
2 Ap
opt sech
Ap
( )opt + opt tanh
=0
ks opt
ks
2
2
ks opt
This is equivalent to
3opt sech2 opt = tanh opt
where
opt =
Ap
opt
2h
ks opt
5.3. Fin structure
74
Numerically, we find that opt = 1.4192. Thus

opt
Lopt
5.3
"
1/2 #2/3
Ap ks Ap
=
opt
2h
"
1/2 #2/3
ks Ap
= opt
2h
Fin structure
Consider now Fig. 5.1 with convection.
5.4
Fin with convection and radiation
Steady state solutions

Equation (5.1) reduces to
d
d
a
d
+ m2 p + p ( + )4 4 = 0
Uniform cross section

For this case a = p = 1, so that
d2
+ m2 + ( + )4 4 = 0
2
d
Convective
With only convective heat transfer, we have
d2
+ m2 = 0
d 2
the solution to whiich is

= C1 sinh m + C2 cosh m
the constants are determined from the boundary conditions. For example, if
(0) = 1
d
(1) = 0
d
we get
= tanh m sinh m + cosh m
5.4. Fin with convection and radiation
75
Example 5.1
If
T (0) = T0
T (L) = T1
then show that
"
T (x) = T1 T0 cosh
!# sinh x hP
kA
hP
q
+ T0 cosh
kA
sinh L hP
kA
hP
kA
Radiative
The fin equation is
Let
d2
+ ( + )4 4 = 0
2
d
=+
so that
d2
+ 4 = 4
d 2
As an example, we will find a perturbation solution with the boundary conditions

(0) = 1 +
d
(1) = 0
d
We write
= 0 + 1 + 2 2 + . . .
The lowest order equation is
d0
= 0,
d 2
0 (0) = 1 + ,
d0
(1) = 0
d
which gives
0 = 1 +
To the next order
d1
= 40 4 ,
d 2
1 (0) = 0,
d1
(1) = 0
d
5.5. Perturbations of one-dimensional conduction
76
with the solution

2
[(1 + )4 4 ]
2
1 = [(1 + )4 4 ]
The complete solution is
2
= (1 + ) + [(1 + ) ] [(1 + )4 4 ]
2
4
+ ...
so that
2
= 1 + [(1 + ) ] [(1 + )4 4 ]
2
4
+ ...
Convective and radiative

?
5.4.1
Annular fin
Example 5.2
Show that under suitable conditions, the temperature distribution in a two-dimensional rectangle tends
to that given by a one-dimensional approximation.
5.5
Perturbations of one-dimensional conduction
5.5.1
Temperature-dependent conductivity
[?]
The governing equation is
d
dx
dT
k(T )
dx
Ph
(T T ) = 0
A
with the boundary conditions

T (0) = Tb
dT
(L) = 0
dx
we use the dimensionless variables
T T
Tb T
x
=
L
77
Consider the special case of a linear variation of conductivity
T T
k(T ) = k0 1 +
Tb T
so that
(1 + )
d2
+
d 2
d
d
m2 = 0
(0) = 1
d
(L) = 0
d
where
m2 =
P hL2
Ak0 (Tb T )
Introduce
() = 0 () + 1 () + 2 () + . . .
Collect terms of O(0 )
d2 0
m2 = 0
d 2
0 (0) = 1
d0
(1) = 0
d
The solution is
() = cosh m tanh m sinh m
To O(1 )
d2 0
d2 1
2
1
0
d 2
d 2
d0
d
= m2 (1 tanh2 m) cosh 2m m2 tanh m sinh 2m

1 (0) = 0
d1
(1) = 0
d
The solution is
5.5.2
Eccentric annulus
Steady-state conduction in a slightly eccentric annular space, as shown in Fig. 5.3 can be solved by
regular perturbation [?]. The radii of the two circles are r1 and r2 .
78
r
2
a
2
1
r
r^
T=T
2
T=T
1
Figure 5.3: Eccentric annulus.
We will use polar coordinates (r, ) with the center of the small circle as origin. The two circles
are at r = r1 and r = rb, where
r22 = a2 + rb2 + 2ab
r cos .
Solving for rb, we have
rb =
r22 a2 (1 cos2 ) a cos .
In the quadratic solution, the positive sign corresponding to the geometry shown in the figure has
been kept.
The governing equation for the temperature is
2
1
1 2
T (r, ) = 0.
+
+
r2
r r r2 2
The boundary conditions are
T (r1 , ) = T1 ,
T (b
r, ) = T2 .
With the variables
T T2
.
T1 T2
r r1
R=
.
r2 r1
r1
.
d=
r2 r1
a
,
=
r2 r1
=
we get
2
1
1
+
+
2
2
R
R + d R (R + d) 2
(R, ) = 0,
79
and
(0, ) = 1,
b ) = 0,
(R,
where
rb r1
b
R()
=
r r1
p2
= (1 + d)2 2 (1 cos2 ) cos d,
The perturbation expansion is
(R, ) = 0 (R, ) + 1 (R, ) + 2 2 (R, ) + . . . .

Substituting in the equations, we get
1
1
2
+
+
0 + 1 + 2 2 + . . . = 0,
2
2
2
R
R + d R (R + d)
2
b
R()
= 1 cos (1 cos2 ) + . . . ,
2
0 (0, ) + 1 (0, ) + 2 2 (0, ) + . . . = 1,
b ) + 1 (R,
b ) + 2 2 (R,
b ) + . . . = 0.
0 (R,
b = 1 to give
Using Eq. (5.11), (5.13) can be further expanded in a Taylor series around R
d0
(1, ) + . . . = 0.
0 (1, ) + 1 (1, ) cos
dR
Collecting terms to order O(0 ), we get
2
1
2
1
0 = 0,
+
+
R2
R + d R (R + d)2 2
0 (0, ) = 1,
0 (1, ) = 0,
which has the solution

0 (R, ) = 1
To order O(1 )
ln(1 + R/h)
.
ln(1 + 1/h)
2
1
1
2
+
+
R2
R + d R (R + d)2 2
1 = 0,
1 (0, ) = 0,
d0
(1, ),
dR
cos
.
=
(1 + h) ln(1 + 1/h)
1 (1, ) = cos
The solution is
1 (R, ) =
R 2h
R cos
.
(1 + 2h) ln(1 + 1/h) R + h
(5.10)
(5.11)
(5.12)
(5.13)
5.6. Transient conduction
5.6
80
Transient conduction
Let us propose a similarity solution of the transient conduction equation

2T
1 T
=0
2
x
t
(5.14)
(5.15)
as
T = erf
2 t
(5.16)
Taking derivatives we find1 [Clarify]
T
1
x2
=
exp
x
4t
t
2
x
x2
T
=
exp
x2
4t
2 3 t3
x
x2
T
=
exp
t
4t
2 t3
so that substitution verifies that equation (5.16) is a solution to equation (5.14).
Alternatively
1
x2
T (x, t) =
exp
4t
4t
In fact all multiples, derivatives and integrals of a solution of Eq. (5.14) are also solutions. They
differ in the boundary conditions that they satisfy.
5.7
Greens functions
For bounded domains

http://people.math.gatech.edu/~xchen/teach/pde/heat/Heat-Green.pdf
5.8
Duhamels principle
For inhomogeneous equations in infinite domains

The solution of the ODE
dy
ay = g(t),
dt
1 The
y(0) = x
error function is defined to be

erf (x) =
et dt.
so that
2
d
2
erf (x) = ex .
dx
5.9. Linear diffusion
81
is
y(t) = S(x, t) +
S(g(s), t s)ds,
where S(x, t) is the solution of

dS
aS = 0,
dt
S(0) = x
Similarly, the solution of [Clarify]

T
2T
2 = g(x, t),
t
x
T (x, 0) = f (x)
(5.17)
is obtained from the solution S(x, t) of

2S
S
2 = 0,
t
x
S(x, 0) = f (x),
(5.18)
which is
S(f, t) =
f (y)G(x y, 2t) dy.
where
x2
1
exp
G(x, t) =
4t
4t
is the Gaussian kernel. The solution of the inhomogeneous equation is then
Z t
S(g(, s), t s) ds,
T (x, t) = S(f, t)(x) +
0
Z t
Z
g(y, s)G(x y, t s) dy ds.
f (y)G(x y, 2t) dy +
=
0
5.9
Linear diffusion
Let T = T (x, t) and

T
2T
= 2
t
x
in 0 x L, with the boundary and initial conditions T (0, t) = T1 , T (L, t) = T2 , and T (x, 0) =
f (x). The steady state solution is
T (x) = T1 +
T2 T1
x.
L
With
T (x, t) = T T
5.10. Nonlinear diffusion
82
we have the same equation

T
2T
= 2
t
x
(5.19)
but with the conditions: T (0, t) = 0, T (L, t) = 0, and T (x, 0) = f (x) T .

Following the methodology outlined in Section ??, we consider the eigenvalue problem
d2
=
dx2
with (0) = (L) = 0. The operator is self-adjoint. Its eigenvalues are
i =
i2 2
,
L2
and its orthonormal eigenfunctions are

i (x) =
2
ix
sin
.
L
L
Thus we let
(x, t) =
ai (t)i (x),
i=1
so that
daj
=
dt
j
L
aj ,
with the solution

(
aj = Cj exp
j
L
2 )
t .
Thus
T (x, t) =
X
i=1
Cj exp
j
L
2 ) r
2
ix
t
sin
.
L
L
(5.20)
The solution shows that T 0, as t . Thus = 0 is a stable solution of the problem. It

must be noted that there has been no need to linearize, since Eq. (5.19) was already linear.
5.10
Nonlinear diffusion
The following diffusion problem with heat generation is considered in [?]

2T
T
= 2 + f (T ),
t
x
83
with < x < , t 0, and 1. The initial condition is taken to be

T (x, 0) = g(x)
1
=
.
1 + ex
(5.21)
(5.22)
Consider two time scales, a fast, short one t1 = t, and a slow, long scale t2 = t. Thus
=
+
.
t
t1
t2
Assuming an asymptotic expansion of the type
T = T0 (x, t1 , t2 ) + T1 (x, t1 , t2 ) + . . .
we have a Taylor series expansion
f (T ) = f (T0 ) + f (T0 ) + . . .
Substituting and collecting terms of O(0 ), we get
T0
= f (T0 ),
t1
(5.23)
with the solution

Z
T0
1/2
dr
= t1 + (x, t2 ).
f (r)
(5.24)
The lower limit of the integral is simply a convenient value at which g(x) = 0.5. Applying the initial
condition gives
=
g(x)
1/2
dr
.
f (r)
The terms of O() are

2 T0
T0
T1
= f (T0 )T1 +
2
t1
x
t2
Differentiating Eq. 5.24 with respect to x gives
T0
= f (T0 ) ,
x
x
so that
2 T0
T0
2
=
f
(T
)
+
f
(T
)
0
0
x2
x x
x2
2
@2
+ f (T0 )
= f (T0 )f (T0 )
x
x2
(5.25)
84
Also, differentiating with respect to t2 gives

T0
= f (T0 )
t2
t2
Substituting in Eq. 5.25,
"
2 #
T1
.
= f (T0 )T1 + f (T0 )
+ f (T0 )
t1
x2
t2
x
Since
f (T0 ) T0
ln f (T0 ) =
t1
f (T0 ) t1
= f (T0 )
we have

T1
= f (T0 )T1 + f (T0 )
+
ln
f
(T
)
.
0
t1
x2
t2
t1
The solution is
"
T1 = A(x, t2 ) + t1
2
+
x2
t2
!#
ln f (T0 )
f (T0 )
which can be checked by differentiation since

"
2 #
T1
2
f (T0 )
=
+ f (T0 )
2
t1
x
t2
x
#
"
2
2
T0

ln f (T0 ) f (T0 )
+
+ A + t1
x2
t2
x
t1
"
#
2
2
=
f (T0 ) + T1 f (T0 ).
+
f
(T
)
0
x2
t2
x
where Eq. 5.23 has been used.
To suppress the secular term in Eq. 5.26, we take
2
2
= 0.
+
(x,
t
)
1
2
x
t2
x
where = f (T0 ). Let
w(x, t2 ) = e ,
so that its derivatives are
w
= e
x
x
2
2
2w

2
+
e
=
e
x2
x
x2
w
= e
t2
t2
(5.26)
85
We find that
"
2 #
2
2w
w
= e
+
2
2
x
t2
x
t2
x
= 0.
The solution is
Z
1
w=
2
R(x + 2r t2 )er dr,
(5.27)
where R(x) = w(x, 0). This can be confirmed by finding the derivatives
Z
2
1
2w
=
R (x + 2r t2 )er dr
2
x

Z
2
r
1
w
R (x + 2r t2 ) er dr
=
t2

t2
Z
#
"
r2

1
R (x + 2r t2 )2 t2 e
= R (x + 2r t2 )
dr
2
1
=
R (x + 2r t2 )er dr

and substituting. Also,
R(x) = exp [(x, 0)]
#
" Z
g(x)
dr
= exp
f (r)
1/2
so that the final (implicit) solution is
Z
T0
1/2
dr
1
1
r 2
R(x + 2r t2 )e
dr
= t1 + ln
f (r)
Fishers equation: As an example, we take f (T ) = T (1 T ), so that the integral in Eq. 5.24 is

Z
T0
1/2
T0
dr
= ln
r(1 r)
T0 1

1
1 + e(t1 +)
w
=
w + et1
T0 =
Thus
w=
T0 et1
1 T0
5.11. Stability by energy method
86
and from Eq. 5.22,

R(x) = w(x, 0)
= ex
Substituting in Eq. 5.27 and integrating,
w = exp x + 2 t2
so that
1
1 + exp [x t(1 + 2 )/]
T0 =
This is a wave that travels with a phase speed of (1 + 2 )/.
5.11
Stability by energy method
5.11.1
Linear
As an example consider the same problem as in Section 5.9. The deviation from the steady state is
governed by
2T
T
= 2
t
x
with T (0, t) = 0, T (L, t) = 0.
Define
E(t) =
so that E 0. Also
dE
=
dt
1
2
T 2 dx
T
dx
t
2T
dx
x2
0
L Z
Z L
L
2
T
T
dx
T
=

x
x
0
0
T
x
dx
so that
dE
0.
dt
Thus E 0 as t whatever the initial conditions.
5.12. Self-similar structures
5.11.2
87
Nonlinear
Let us now re-do the problem for a bar with temperature-dependent conductivity. Thus
T
T
k(T )
,
=
t
x
x
with T (0, t) = T1 and T (L, t) = T2 . The steady state, T (x), is governed by
d
dT
k(T )
= 0.
dx
dx
Let the deviation from the steady state be
(x, t) = T (x, t) T (x).
Thus
=
t
x
k()
x
where = (x, t), with (0, t) = (L, t) = 0. The steady state is = 0. Let
Z L
1
2 dx.
E(t) =
2 0
so that E 0. Then
dx
t
0
Z L
k()
dx
=
x
x
0
L Z
2
L

= k()
dx.
k()
x
x
0
dE
=
dt
Due to boundary conditions the first term on the right is zero, so that dE/dt 0. Thus E 0 as
t .
5.12
Self-similar structures
Consider the large-scale structure shown in Fig. 5.4 in which each line i (indicated by i = 0, 1, . . .)
is a conductive bar. The length of each bar is Li = L/ i and its diameter is Di = D/ i . The
beginning is at temperature T0 and the ambient is T .
The total length of the structure is
LT = L0 + 2L1 + 4L2 + 8L3 + . . .
= ...
The total volume of the material is
2
VT =
D0 L0 + 2D12 L1 + 4D22 L2 + 8D32 L3 + . . .
4
= ...
Both of these are finite if < c = . . ..
5.13. Non-Cartesian coordinates
88
$T_\infty$
3
3
1
T
0
3
3
Figure 5.4: Large-scale self-similar structure.
5.13
Non-Cartesian coordinates
Toroidal, bipolar.
5.14
Thermal control
Partial differential equations (PDEs) are an example of infinite-dimensional systems that are very
common in thermal applications [?, ?]. Exact controllability exists if the function representing the
state can be taken from an initial to a final target state, and is approximate if it can be taken to a
neighborhood of the target [?]. Determination of approximate controllability is usually sufficient for
practical purposes.
Consider a system governed by
X
= AX + Bu,
t
(5.28)
with homogeneous boundary and suitable initial conditions, where A is a bounded semi-group operator [?], and B is a another linear operator. The state X(, t) is a function of spatial coordinates
and time t. If A is self-adjoint, then it has real eigenvalues and a complete orthonormal set of
eigenfunctions m (), with m = 0, 1, 2 . . ., which forms a complete spatial basis for X. It is known [?]
that the system is approximately state controllable if and only if all the inner products
hB, m i =
6 0.
(5.29)
The lumped approximation in this chapter, valid for Bi 1, is frequently not good enough
for thermal systems, and the spatial variation of the temperature must be taken into account. The
system is then described by PDEs that represent a formidable challenge for control analysis. The
simplest examples occur when only one spatial dimension is present.
5.14. Thermal control
89
=0
=L
TL
Figure 5.5: One-dimensional fin with convection.

Fig. 5.5 shows a fin of length L with convection to the surroundings [?]. It is thin and long
enough such that the transverse temperature distribution may be neglected. The temperature field
is governed by
2T
T
= 2 (T T ),
t
(5.30)
where T (, t) is the temperature distribution that represents the state of the system, T is the
temperature of the surroundings, t is time, and is the longitudinal coordinate measured from
one end. The thermal diffusivity is , and = hP/cAc where h is the convective heat transfer
coefficient, Ac is the constant cross-sectional area of the bar, P is the perimeter of the cross section,
is the density, and c is the specific heat. For simplicity it will be assumed that is independent of
. The end = 0 will be assumed to be adiabatic so that (T /)(0, t) = 0.
Since a linear system that is controllable can be taken from any state to any other, we can
arbitrarily assume the fin to be initially at a uniform temperature. There are two ways in which
the temperature distribution on the bar can be controlled: in distributed control2 the surrounding
temperature T is the control input and in boundary control it is the temperature of the other end
T (L, t) of the fin.
(a) Distributed control: The boundary temperature T (L, t) = TL is fixed. Using it as a reference
temperature and defining = T TL , Eq. (5.30) becomes,
2
= 2 + (t)
t
(5.31)
with the homogeneous boundary and initial conditions (/)(0, t) = 0, (L, t) = 0, and (, 0) = 0.
The operators in Eq. (5.28) are A = 2 / 2 , B = , and u = . A is a self-adjoint
operator with the eigenvalues and eigenfunctions
(2m + 1)2 2
m =
,
4L2
r
2
(2m + 1)
m =
cos
,
L
2L
respectively. Inequality (5.29) is satisfied for all m, so the system is indeed state controllable. It can
be shown that the same problem can also be analyzed using a finite-difference approximation [?].
(b) Boundary control: Using the constant outside temperature T as reference and defining =
T T , Eq. (5.30) becomes
2
= 2 ,
t
2 This
term is also used in other senses in control theory.
5.15. Multiple scales
90
with the initial and boundary conditions (/)(0, t) = 0, (L, t) = TL (t) T , and (, 0) = 0.
To enable a finite-difference approximation [?], the domain [0, L] is divided into n equal parts
of size , so that Eq. (1) becomes
di
= i1 (2 + )i + i1 ,
dt
where = / 2 . The nodes are i = 1, 2, . . . , n + 1, where i = 1 is at the left and i = n + 1 at the
right end of the fin in Fig. 5.5. With this Eq. (1) can be discretized to take the form of Eq. (3.6),
where x is the vector of unknown i . Thus we find
(2 + )
2
0
0
..
(2 + )
.
..
..
..
A=
Rnn ,
.
.
.
0
..
B = [0, , ] R .
(2 + )
The boundary conditions have been applied to make A non-singular:

adiabatic, and at the right end n+1 is the control input u.
The controllability matrix M is
0
n1
0
..
..
..
..
.
.
.
M = .
3
0
0
0
2
2 2 (2 + )
(2 + ) 3 + (2 + )2
at the left end the fin is
..
.
The rank of M is n, indicating that the state of the system is also boundary controllable.
5.15
Multiple scales
Solve
2 (T1 T2 )
T1
=
t
x2
T2
2 (T2 T1 )
= R
t
x2
where 1, and with a step change in temperature at one end. Let
t = t0 + t1
Problems
1. From the governing equation for one-dimensional conduction
d
dT
k(x, T )
= 0,
dx
dx
5.15. Multiple scales
91
T
b
x
L
Figure 5.6: Longitudinal fin of concave parabolic profile.
with boundary conditions
T
,
2
T
T (L) = T
.
2
T (0) = T +
show that the magnitude of the heat rate is independent of the sign of T if we can write k(x, T ) = A(x) (T ).
2. Consider a rectangular fin with convection, radiation and Dirichlet boundary conditions. Calculate numerically
the evolution of an initial temperature distribution at different instants of time. Graph the results for several
values of the parameters.
3. Consider a longitudinal fin of concave parabolic profile as shown in the figure, where = [1 (x/L)]2 b . b
is the thickness of the fin at the base. Assume that the base temperature is known. Neglect convection from
the thin sides. Find (a) the temperature distribution in the fin, and (b) the heat flow at the base of the fin.
Optimize the fin assuming the fin volume to be constant and maximizing the heat rate at the base. Find (c)
the optimum base thickness b , and (d) the optimum fin height L.
Chapter 6
Forced convection
In this chapter we will considering the heat transfer in pipe flows. We will take a one-dimensional approach and neglect transverse variations in the velocity and temperature. In addition, for simplicity,
we will assume that fluid properties are constant and that the area of the pipe is also constant.
6.1
Hydrodynamics
6.1.1
Mass conservation
For a duct of constant cross-sectional area and a fluid of constant density, the mean velocity of the
fluid, V , is also constant.
6.1.2
Momentum equation
The forces on an element of length ds, shown in Fig. 6.1, in the positive s direction are: fv , the
viscous force and fp , the pressure force. We can write
fv = w P ds
p
fp = A ds
s
where w is the magnitude of the wall shear stress, and p is the pressure in the fluid. Since the mass
of the element is A ds, we can write the momentum equation as
A ds
dV
= fv + fp
dt
from which we get

w P
1 p
dV
+
=
dt
A
s
Integrating over the length L of a pipe, we have
4w
p1 p2
dV
+
=
dt
D
L
92
6.1. Hydrodynamics
93
p+dp
fp
s
f
ds
Figure 6.1: Forces on an element of fluid.

where p1 and p2 are the pressures at the inlet and outlet respectively, and the hydraulic diameter is
defined by D = 4A/P .
For fully developed flow we can assume that w is a function of V that depends on the mean
velocity profile, and acts in a direction opposite to V , so that we can write
dV
+ T (|V |)V = p
dt
(6.1)
where T (V ) = |4w /DV | is always positive, = 1/L, and p = p1 p2 is the pressure difference
that is driving the flow. The wall shear stress is estimated below for laminar and turbulent flows.
Laminar
The fully developed laminar velocity profile in a circular duct is given by the Poiseuille flow result
4r2
ux (r) = um 1 2
D
where u is the local velocity, r is the radial coordinate, um is the maximum velocity at the centerline,
and D is the diameter of the duct. The mean velocity is given by
V =
4
D2
D/2
ux (r) 2r dr
Substituting the velocity profile, we get

V =
um
2
The shear stress at the wall w is given by
ux
w =
r
4
= um
D
8V
=
D
The wall shear stress is linear relationship

w = V
r=D/2
6.1. Hydrodynamics
94
where
=
8
D
so that
T (|V |)V =
32
V
D2
Turbulent
For turbulent flow the expression for shear stress at the wall of a duct that is usually used is
f 1 2
w =
V
4 2
so that
T (V ) =
f
|V |V.
2D
Here f is the Darcy-Weisbach friction factor1 . The friction factor is also a fucntion of |V |, and may
be calculated from the Blasius equation for smooth pipes
f=
0.3164
Re1/4
where the Reynolds number is Re = |V |D/, or the Colebrook equation for rough pipes
1
2.51
e/Dh
+
= 2.0 log
3.7
f 1/2
Re f 1/2
where e is the roughness at the wall, or similar expressions.
In the flow in a length of duct, L, without acceleration, the pressure drop is given by
p A = w P L
where A is the cross-sectional area, and P is the inner perimeter. Thus

1 2
L
V
p = f
4A/P
2

1
L
=f
|V |V
D
2
Example 6.1
Consider a long, thin pipe with pressures p1 and p2 ate either end. For t 0, p1 p2 = 0 and there is no
flow. For t > 0, p1 p2 is a nonzero constant. Find the resulting time-dependent flow. Make the assumption
that the axial velocity is only a function of radial position and time.
1 Sometimes, confusingly, the Fanning friction factor, which is one-fourth the Darcy-Weisbach value, is used in the
literature.
6.2. Energy equation
6.1.3
95
Long time behavior
Consider the flow in a single duct of finite length with a constant driving pressure drop. The
governing equation for the flow velocity is equation (6.1). The flow velocity in the steady state is a
solution of
T (V )V = p
(6.2)
where p and V are both of the same sign, say nonnegative. We can show that under certain
conditions the steady state is globally stable. Writing V = V + V , equation (6.1) becomes
dV
+ T (V + V )(V + V ) = p
dt
Subtracting equation (6.2), we get
dV
= T (V + V )(V + V ) + T (V )(V )
dt
Defining
E=
1 2
V
2
so that E 0, we find that

dV
dE
=V
dt
dt
= V T (V + V )(V + V ) T (V )V
= V V T (V + V ) T (V ) V 2 T (V + V )
If we assume that T (V ) is a non-decreasing function of V , we see that
V V T (V + V ) T (V ) 0
(6.3)
regardless of the sign of either V or V , so that
dE
0
dt
Thus, E(V ) is a Lyapunov function, and V = V is globally stable to all perturbations.
6.2
Energy equation
Consider a section of a duct shown in Fig. 6.2, where an elemental control volume is shown. The
heat rate going in is given by
Q = AV cT kA
T
s
where the first term on the right is due to the advective and second the conductive transports. c is
the specific heat at constant pressure and k is the coefficient of thermal conductivity. The heat rate
going out is
Q+ = Q +
Q
ds
s
6.2. Energy equation
96
Q
Q
_
Q
s
ds

The difference between the two is
Q
ds
s
T
2T
= AV c
kA 2 ds
s
s
Q+ Q =
(6.4)
Furthermore, heat is gained from the side at a rate dQ, which can be written as
dQ = q ds
(6.5)
where q is the rate of gain of heat per unit length of the duct.
An energy balance for the elemental control volume gives
Q + dQ = Q+ + A ds c
T
t
where the last term is the rate of accumulation of energy within the control volume.
Substituting equations (6.4) and (6.5) in (6.6) we get the energy equation
T
q
k 2T
T
+V
=
+
t
s
Ac c s2
The two different types of heating conditions to consider are:
6.2.1
Known heat rate
The heat rate per unit length, q(s), is known all along the duct. Defining
x
L
(T Ti )V AC
=
Lq
tV
=
L
=
(6.6)
6.3. Single duct
97
gives
d2
+
2 =1
d
where
=
k
LV c
Boundary conditions may be = 0 at = 0, = 1 at = 1.

6.2.2
Convection with known outside temperature
The heating is now convective with a heat transfer coefficient U , and an external temperature of
T (s). Thus,
q = P U (T T )
Defining
x
L
T Ti
=
T Ti
tV
=
L
=
gives
d2
+
2 + H = H
d
where
k
LV c
U L
H=
V Ac
=
6.3
Single duct
Consider the duct that is schematically shown in Fig. 6.3. The inlet temperature is Tin (t), and the
outlet temperature is Tout (t), and the fluid velocity is V . The duct is subject to heat loss through
its surface of the form U P (T T ) per unit length, where the local fluid temperature is T (s, t) and
the ambient temperature is T (t). U is the overall heat transfer coefficient and P the cross-sectional
perimeter of the duct.
We assume that the flow is one-dimensional, and neglect axial conduction through the fluid
and the duct. Using the same variables to represent non-dimensional quantities, the governing
non-dimensional equation is
+
+ H = 0
6.3. Single duct
98
T (t)
Tin (t)
T (s, t)
Tout (t)
Figure 6.3: Fluid duct with heat loss.

where the nondimensional variables are
s
L
tV
=
L
T T
=
T
=
The characteristic time is the time taken to traverse the length of the duct, i.e. the residence time.
The ambient temperature is
T (t) = T + Te (t)
where the time-averaged and fluctuating parts have been separated. Notice that the nondimensional
mean ambient temperature is, by definition, zero. The characteristic temperature difference T will
be chosen later. The parameter = U P L/AV c represents the heat loss to the ambient.
6.3.1
Steady state
No axial conduction
The solution of the equation
d
+ H( 1) = 0
d
with boundary condition (0) = 0 is
() = 1 eH
With small axial conduction
We have
d2 d
H( 1) = 0
d 2
d
where 1, and with the boundary conditions (0) = 0 and (1) = 1 .

We can use a boundary layer analysis for this singular perturbation problem. The outer solution
is
out = 1 eH
The boundary layer is near = 1, where we make the transformation
X=
6.3. Single duct
99
This gives the equation

d2 in
din
H(in 1) = 0
dX 2
dX
To lowest order, we have
d2 in
din
=0
dX 2
dX
with the solution

in = A + BeX
The boundary condition in (X = 0) = 1 gives 1 = A + B, so that
in = A + (1 A)eX
The matching conditions is
outer ( = 1) = in (X )
so that
A = 1 eH
The composite solution is then
= 1 eH + (1 1 + eH )e(1)/ + . . .
6.3.2
Unsteady dynamics
The general solution of this equation is
Z t
Ht e
e T (t ) dt eHt
T (s, t) = f (s t) +
(6.7)
The boundary conditions T (0, t) = Tin (t) and T (s, 0) = T0 (s) are shown in Fig. 6.4. The solution
becomes
(
Rt
Tin (t s)eHs + HeHt ts eHt Te (t ) dt for t s

R
T (s, t) =
(6.8)
t
T0 (s t)eHt + HeHt 0 eHt Te (t ) dt
for t < s
The t < s part of the solution is applicable to the brief, transient period of time in which the fluid
at time t = 0 has still not left the duct. The later t > s part depends on the temperature of the
fluid entering at s = 0. The temperature, Tout (t), at the outlet section, s = 1, is given by
(
Rt
Tin (t 1)eH + HeHt t1 eHt Te (t ) dt for t 1

R
(6.9)
Tout (t) =
t
for t < 1
T0 (1 t)eHt + HeHt 0 eHt Te (t ) dt
It can be observed that, after an initial transient, the inlet and outlet temperatures are related
by a unit delay. The outlet temperature is also affected by the heat loss parameter, , and the
ambient temperature fluctuation, Te . The following are some special cases of equation (6.9).
6.3. Single duct
100
t>s
T (0, t) = Tin (t)
@
R
t=s
t<s
T (s, 0) = T0 (s)
s
Figure 6.4: Solution in s-t space.

6.3.3
Perfectly insulated duct
If H = 0 the outlet temperature simplifies to
Tin (t 1)
Tout (t) =
T0 (1 t)
for t 1
for t < 1
The outlet temperature is the same as the inlet temperature, but at a previous instant in time.
6.3.4
Constant ambient temperature
For this Te = 0, and equation (6.9) becomes
Tin (t 1)eH
Tout (t) =
T0 (1 t)eHt
for t 1
for t < 1
This is similar to the above, but with an exponential drop due to heat transfer.
6.3.5
Periodic inlet and ambient temperature
We take
Tin (t) = T in + Tbin sin t
Te (t) = Tb sin t
so that equation (6.9) becomes

i
h
bin sin (t 1) eH
T
+
T
in
q
1+e2H
Tout (t) =
b H 2eH cos
sin(t + )
+
T
H 2 +2
H
Ht
2
T0 (1 t)e
+ 2 2 Tb H + 2 sin(t + )
H +
where
H(1 eH cos 1) + eH sin 1

(1 eH cos 1) HeH sin 1
tan =
H
tan =
(6.10)
(6.11)
for t 1
for t < 1
6.3. Single duct
101
T1
T2
Figure 6.5: Effect of wall.

The outlet temperature has frequencies which come from oscillations in the inlet as well as the ambient temperatures. A properly-designed control system that senses the outlet temperature must take
the frequency dependence of its amplitude and phase into account. There are several complexities
that must be considered in practical applications to heating or cooling networks, some of which are
analyzed below.
6.3. Single duct
6.3.6
102
Effect of wall
The governing equations are

T
T
2T
+ V Ac
kA 2 + hi Pi (T T ) = 0
t
x
x
Tw
2 Tw
+ hi Pi (Tw T ) + ho Po (Tw T ) = 0
w Aw cw
kw Aw
t
x2
Ac
Nondimensionalize, using
x
L
tV
=
L
T T
=
Ti T
Tw T
w =
Ti T
=
we get
+
2 + Hin ( w ) = 0
2
w
w
w
+ Hin (w ) + Hout w = 0
2
where
kw
w V Aw cw L
hin Pin L
=
w Aw cw V
hout Pout L
=
w Aw cw V
w =
Hin
Hout
In the steady state and with no axial conduction in the fluid

d
+ Hin ( w ) = 0
d
w
d2 w
+ Hin (w ) + Hout w = 0
d 2
If we assume w = 0 also, we get

w =
Hin
Hin + Hout

dT
+ Hw = 0
d
6.4. Two-fluid configuration
103
Figure 6.6: Two-fluids with wall.

where
Hw =
6.4
w
w
Hin
Hout
w
w
Hin + Hout
Two-fluid configuration
Consider the heat balance in Fig. 6.6. Neglecting axial conduction, we have
Tw
+ h1 (Tw T1 ) + h2 (Tw T2 ) = 0
t
T1
T1
+ 1 V1 c1
+ h1 (T1 Tw ) = 0
1 A1 c1
t
x
T2
T2
+ 2 V2 c2
+ h2 (T2 Tw ) = 0
2 A2 c2
t
x
w Aw cw
6.5
Flow between plates with viscous dissipation
Consider the steady, laminar flow of an incompressible, Newtonian fluid between fixed, flat plates at
y = h and y = h. The flow velocity u(y) is in the x-direction due to a constant pressure gradient
P < 0. The plane walls are kept isothermal at temperature T = T0 , and the viscosity is assumed to
decrease exponentially with temperature according to
= exp(1/T ).)
The momentum equation is then
d
dy
(T )
du
dy
=P
with boundary conditions u = 0 at y = h. Integrating, we get

(T )
du
= Py + C
dy
(6.12)
6.6. Regenerator
104
hightemperature solution
1
lowtemperature solution
0
1
0.8
0.6
0.4
0.2
0.2
0.4
0.6
0.8
Figure 6.7: Two solutions of Eq. (6.13) with boundary conditions (6.14) for a = 1.
Due to symmetry du/dy = 0 at y = 0 so that C = 0. There is also other evidence for this.
The energy equation can be written as
2
du
d2 T
=0
k 2 + (T )
dy
dy
with T = T0 at y = h, where k has been taken to be a constant. The second term corresponds
to viscous heating or dissipation, and the viscosity is assumed to be given by Eq. (6.12). We nondimensionalize using
= (T T0 )
y
=
h
The energy equation becomes
d2
+ a 2 e = 0
d 2
(6.13)
where
a=
P 2 h4
k0

= 0 for = 1
(6.14)
There are two solutions that can be obtained numerically (by the shooting method, for instance)
for the boundary-value problem represented by Eqs. (6.13) and (6.14) for a < ac and above which
there are none. There are other solutions also but they do not satisfy the boundary conditions on
the velocity. As examples, two numerically obtained solutions for a = 1 are shown in Fig. 6.7.
The bifurcation diagram corresponding to this problem is shown in Fig. 6.8 where S is the
slope of the temperature gradient on one wall.
6.6
Regenerator
A regenerator is schematically shown in Fig. 6.9.
6.7. Radial flow between disks
105
25
20
15
10
3
a
Figure 6.8: Bifurcation diagram.
Figure 6.9: Schematic of regenerator.
Mc
6.7
dT
+ mc(T
in Tout ) = 0
dt
Radial flow between disks
This is shown in Fig. 6.10.
ur =
C
r
qr = ur 2rHT k2rH
dT
dr
where H is the distance between the disks. With dqr /dr = 0, we get
d
d dT
(rur T ) = k (r
)
dr
dr dr
r
1
r
2
Figure 6.10: Flow between disks.
6.8. Networks
106
For the boundary conditions T (r1 ) = T1 and T (r2 ) = T2 , the temperature field is
T (r) =
T1 ln(r2 /r) T2 ln(r1 /r)

ln(r2 /r1 )
Example 6.2
Redo the previous problem with a slightly eccentric flow.
6.8
Networks
A network consists of a number of ducts that are united at certain points. At each junction, we
must have
X
Ai Vi = 0
(6.15)
i
where Ai are the areas and Vi the fluid velocities in the ducts coming in, the sum being over all the
ducts entering the junction. Furthermore, for each duct, the momentum equation is
dVi
out
+ T (Vi )Vi = pin
+ p
i pi
dt
where p is the pressure developed by a pump, if there happens to be one on that line. We must
distinguish between two possible geometries.
(a) Two-dimensional networks: A planar or two-dimensional network is one that is topologically
equivalent to one on a plane in which every intersection of pipes indicates fluid mixing. For such a
graph, we know that
E =V +F 1
where E, V and F are the number of edges, vertices and faces, respectively. In the present context,
these are better referred to as branches, junctions and circuits, respectively.
The unknowns are the E velocities in the ducts and the V pressures at the junctions, except
for one pressure that must be known. The number of unknowns thus are E + V 1. The momentum
equation in the branches produce E independent differential equations, while mass conservation at
the juntions give V 1 independent algebraic relations. Thus the number of
(b) Three-dimensional networks: For a three-dimensioanl network, we have
E =V +F 2
If there are n junctions, they can have a maximum of n(n 1)/2 lines connecting them. The number
of circuits is then (n2 3n + 4)/2. The number of equations to be solved is thus quite large if n is
large.
6.8. Networks
6.8.1
107
Hydrodynamics
The global stability of flow in a network can be demonstrated in a manner similar to that in a
finite-length duct. In a general network, assume that there are n junctions, and each is connected
to all the rest. Also, pi is the pressure at junction i, and Vij is the flow velocity from junction i to j
defined to be positive in that direction. The flow velocity matrix Vij is anti-symmetric, so that Vii
which has no physical meaning is considered zero.
The momentum equation for Vij is
dVij
+ Tij (Vij )Vij = ij (pi pj )
dt
(6.16)
The network properties are represented by the symmetric matrix ij . The resistance Tij may or
may not be symmetric. To simplify the analysis the network is considered fully connected, but
Tij is infinite for those junctions that are not physically connected so that the flow velocity in the
corresponding branch is zero. We take the diagonal terms in Tij to be also infinite, so as to have
Vii = 0.
The mass conservation equation at junction j for all flows arriving there is
n
X
Aij Vij = 0
for j = 1, . . . , n
i=1
where Aij is a symmetric matrix. The symmetry of Aij and antisymmetry of Vij gives the equivalent
form
n
X
Aji Vji = 0
for j = 1, . . . , n
(6.17)
i=1
which is simply the mass conservation considering all the flows leaving junction j.
The steady states are solutions of
Tij (V ij )V ij = ij (pi pj )
n
X
(6.18)
n
X
Aij V ij = 0or
Aji V ji = 0
(6.19)
i=1
i=1
We write
Vij = V ij + Vij
pi = p + pi
Substituting in equations (6.16)(6.17), and subtracting equations (6.18) and (6.19) we get
n
X
dVij
= Tij (V ij + Vij )(V ij + Vij ) Tij (V ij )V ij + ij (pi pj )
dt
Aij Vij = 0or
n
X
i=1
i=1
Defining
n
E=
1 X X Aij 2
V ,
2 j=1 i=1 ij ij
ij > 0
Aji Vji = 0
6.8. Networks
108
we get
n
X X Aij
dVij
dE
=
Vij
dt
dt
j=1 i=1 ij
=
n
n X
X
Aij
j=1 i=1
ij
(6.20)
Vij Tij (V ij + Vij )(V ij + Vij ) Tij (V ij )V ij

+
n
n X
X
j=1 i=1
Aij Vij (pi pj )

(6.21)
The pressure terms vanish since

n
n X
X
Aij Vij pi =
n
n X
X
Aij Vij pi
i=1 j=1
j=1 i=1
n
X
i=1
=0
pi
n
X
j=1
Aij Vij
n
X
Aij Vij
and
n
n X
X
Aij Vij pj
n
X
j=1
j=1 i=1
pj
=0
i=1
The terms that are left in equation (6.21) are similar to those in equation (6.3) and satisfy the same
inequality. Since E 0 and dE/dt 0, the steady state is globally stable. For this reason the
steady state is also unique.
Example 6.3
Show that the flow in the the star network shown in Fig. 6.11 is globally stable. The pressures p1 , p2 and
p3 are known while the pressure p0 and velocities V10 , V20 and V30 are the unknowns.
For branches i = 1, 2, 3, equation (6.1) is
dVi0
+ Ti0 (Vi0 )Vi0 = i0 (pi p0 )
dt
(6.22)
Equation (6.15) at the junction gives

3
X
Ai0 Vi0 = 0
(6.23)
i=1
In the steady state

Ti0 (V i0 )V i0 = i0 (pi p0 )
3
X
i=1
Ai0 V i0 = 0
(6.24)
(6.25)
6.8. Networks
109
20
10
30
Figure 6.11: Star network.

and p = p + p in equations (6.22) and (6.23) and subtracting equations (6.24)
Substituting Vi0 = V i0 + Vi0
0
0
0
and (6.25), we find that
3
X
dVi0
)(V i0 + Vi0
) Ti0 (V i0 )V i0 i0 p0
= Ti0 (V i0 + Vi0
dt
Ai0 Vi0
=0
(6.27)
i=1
If we define
E=
3
1 X Ai0 2
V i
2 i=1 i0
we find that
3
X Ai0
dE
dVi0
=
Vi0
dt
dt
i=1 i0
3
3
X
X
Ai0
Ai0 Vi0
)(V i0 + Vi0
) Ti0 (V i0 )V i0 p0
Vi0 Ti0 (V i0 + Vi0
i=1
i=1 i0
The last term vanishes because of equation (6.27). Thus

3
3
X
X
dE
Ai0
Ai0 2
Vi0 V i0 Ti0 (V i0 + Vi0

V i0 Ti0 (V i0 + Vi0
=
) Ti0 (V i0 )
)
dt
i=1 i0
i=1 i0
Since E 0 and dE/dt 0, E is a Lyapunov function and the steady state is globally stable.
6.8.2
[?]
Thermal networks
(6.26)
110
6.9
Thermal control
6.9.1
Control with heat transfer coefficient
6.9.2
Multiple room temperatures
Let there be n interconnected rooms. The wall temperature of room i is Tiw and the air temperature
is Tia . The heat balance equation for this room is
dTiw
= hi Ai (Tia Tiw ) + Ui Aei (T e Tiw )
dt
dT a
1 X a
Mia ca i = hi Ai (Tiw Tia ) + ca
(mji + |maji |)Tja
dt
2
j
Mia cw
(6.28)
1 X a
(mij + |maij |)Tia + qi
ca
2
j
(6.29)
where T e is the exterior temperature, mij is the mass flow rate of air from room i to room j. By
definition mij = mji . Since mii has no meaning and can be arbitrarily taken to be zero, mij is an
anti-symmetric matrix. Also, from mass conservation for a single room, we know that
X
maji = 0
j
Analysis
The unknowns in equations (6.28) and (6.29) are the 2n temperatures Tiw and Tia .
(i) Steady state with U = 0
(a) The equality
XX
XX
(maji + |maji |)Tja
(maij + |maij |)Tia = 0
i
can be shown by interchanging i and j in the second term. Using this result, the sum of equations
(6.28) and (6.29) for all rooms gives
X
qi = 0
i
which is a necessary condition for a steady state.

(b) Because the sum of equations (6.28) and (6.29) for all rooms gives an identity, the set of equations is not linearly independent. Thus the steady solution is not unique unless one of the room
temperatures is known.
Control
The various proportional control schemes possible are:
Control of individual room heating
qi = Ki (Tia Tiset )
111
Control of mass flow rates

maji = fij (Tja , Tia , Tiset )
Similar on-off control schemes can also be proposed.
6.9.3
Two rooms
Consider two interconnected rooms 1 and 2 with mass flow m from 1 to 2. Also there is leakage of
air into room 1 from the exterior at rate m, and leakage out of room 2 to the exterior at the same
rate. The energy balances for the two rooms give
M1 ca
M2 ca
1
dT1
= U1 A1 (T e T1 ) + (m + |m|)(T e T1 )
dt
2
1
(m |m|)(T2 T1 ) + q1
2
(6.30)
1
+ (m + |m|)(T1 T2 ) + q2
2
(6.31)
1
dT2
= U2 A2 (T e T2 ) (m |m|)(T e T2 )
dt
2
The overall mass balance can be given by the sum of the two equations to give
M1 ca
dT2
dT1
+ M2 ca
= U1 A1 (T e T1 ) + U2 A2 (T e T2 ) + |m|T e
dt
dt
1
1
+ (m |m|)T1 (m + |m|)T2
2
2
+q1 + q2
One example of a control problem would be to change m to keep the temperatures of the two rooms
equal. Delay can be introduced by writing T2 = T2 (t ) and T1 = T1 (t ) in the second to last
terms of equations (6.30) and (6.31), respectively, where is the time taken for the fluid to get from
one room to the other.
6.9.4
Temperature in long duct
The diffusion problem of the previous section does not have advection. Transport of fluids in ducts
introduces a delay between the instant the particles of fluid go into the duct and when they come
out, which creates a difficulty for outlet temperature control. The literature includes applications
to hot-water systems [?, ?] and buildings [?, ?]; transport [?] and heater [?, ?] delay and the effect of
the length of a duct on delay [?] have also been looked at.
A long duct of constant cross section, schematically shown in Fig. 6.12 where the flow is driven
by a variable-speed pump, illustrates the basic issues [?, ?]. The fluid inlet temperature Tin is kept
constant, and there is heat loss to the constant ambient temperature T through the surface of the
duct.
With a one-dimensional approximation, energy conservation gives
T
T
4h
+v
+
(T T ) = 0,
t
cD
(6.32)
112
=0
Tin
flow =
=L
velocity = v(t)
temperature = T (, t)
Tout (t)
Figure 6.12: Schematic of duct.

with the boundary condition T (0, t) = Tin , where T (, t) is the fluid temperature, t is time, is the
distance along the duct measured from the entrance, v(t) is the flow velocity, h is the coefficient
of heat transfer to the exterior, is the fluid density, c is its specific heat, and D is the hydraulic
diameter of the duct. The flow velocity is taken to be always positive, so that the = 0 end is
always the inlet and = L the outlet, where L is the length of the duct. The temperature of the
fluid coming out of the duct is Tout (t).
Using the characteristic quantities of L for length, cD/4h for time, and hL/cD for velocity,
the non-dimensional version of Eq. (6.32) is
+v
+ = 0,
t
(6.33)
where = (T T )/(Tin T ), with (0, t) = 1. The other variables are now non-dimensional.
Knowing v(t), this can be solved to give
Z t
v(s) ds ,
(6.34)
(, t) = et f
0
where the initial startup interval in which the fluid within the duct is flushed out has been ignored;
f is an arbitrary function. Applying the boundary condition at = 0 gives
Z t
t
v(s) ds .
(6.35)
1=e f
0
The temperature at the outlet of the duct, i.e. at = 1, is
Z t
t
v(s). ds
out (t) = e f 1
0
Eqs. (6.35) and (7.1) must be simultaneously solved to get the outlet temperature out (t) in terms
of the flow velocity v.
The problem is non-linear if the outlet temperature Tout (t) is used to control the flow velocity
v(t). The delay between the velocity change and its effect on the outlet temperature can often
lead instability, as it does in other applications [?, ?, ?, ?]. Fig. 6.13 shows a typical result using
PID control in which the system is unstable. Shown are the outlet temperature, flow velocity and
residence time of the fluid in the duct, all of which ultimately achieve constant amplitude oscillations.
Problems
4
1. Determine the single duct solutions for heat loss by radiation q = P (Tsur
T 4 ).
2. A sphere, initially at temperature Ti is being cooled by natural convection to fluid at T . Churchills correlation
for natural convection from a sphere is
1/4
0.589 RaD
Nu = 2 + h
i4/9 ,
1 + (0.469/Pr )9/16
113
0.5
Tout
0.4
0.3
0.2
50
100
150
50
100
150
50
100
150
1.5
0.5
1.4
1.2
1
0.8
0.6
Figure 6.13: Outlet temperature, velocity and residence time for Ki = 5 and Kp = 2.5 [?].
where
g(Ts T )D3
.
Assume that the temperature within the sphere T (t) is uniform, and that the material properties are all
constant. Derive the governing equation, and find a two-term perturbation solution.
RaD =
3. The velocity field, u(r), for forced convection in a cylindrical porous medium is given by
u + r 1 u s2 u + s2 Da = 0,
where s and the Darcy number Da are parameters. A WKB solution for small Da has been reported as2
"
#
es(1r)
u = Da 1
.
r
Re-do to check the analysis.
4. Consider one-dimensional steady-state flow along a pipe with advection and conduction in the fluid and lateral
convection from the side. The fluid inlet and outlet temperatures given. Use the nondimensional version of
the governing equation to find the inner and outer matched temperature distributions if the fluid thermal
conductivity is small.
5. Plot the exact analytical and the approximate boundary layer solutions for Problem 4 for a small value of the
conduction parameter.
6. Show that no solution is possible in Problem 4 if the boundary layer is assumed to be on the wrong side.
7. Consider one-dimensional unsteady flow in a tube with a non-negligible wall thickness, as shown in Fig. A.38.
There is conduction along the fluid as well as along the wall of the tube. There is also convection from the
outer surface of the tube to the environment as well as from its inner surface to the fluid. Find the governing
equations and their boundary conditions. Nondimensionalize.
2 K. Hooman and A.A. Ranjbar-Kani, Forced convection in a fluid-saturated porous-medium tube with isoflux wall,
International Communcications in Heat and Mass Transfer, Vol. 30, No. 7, pp. 10151026, 2003.
114
11111111111111111111
00000000000000000000
00000000000000000000
11111111111111111111
11111111111111111111
00000000000000000000
00000000000000000000
11111111111111111111
00000000000000000000
11111111111111111111
Figure 6.14: Flow in tube with non-negligible wall thickness.
Chapter 7
Natural convection
7.1
Modeling
Let us consider a closed loop, shown in Fig. 7.1, of length L and constant cross-sectional area A filled
with a fluid. The loop is heated in some parts and cooled in others. The temperature differences
within the fluid leads to a change in density and hence a buoyancy force that creates a natural
circulation. The spatial coordinate is s, measured from some arbitrary origin and going around the
loop in the counterclockwise direction.
We will make the Boussinesq approximation by which the fluid density is constant except in
the buoyancy term. We will also approximate the behavior of the fluid using one spatial dimensions.
Thus, we will assume that the velocity u and temperature T are constant across a section of the
loop. In general both u and T are functions of space s and time t, though we will find that u = u(t).
7.1.1
Mass conservation
Consider an elemental control volume as shown in Fig. 7.2. The mass fluxes in and out are
m = 0 uA
m + = m +
m
ds
s
For a fluid of constant density, there is no accumulation of mass within an elemental control volume,
so that the mass flow rate into and out of the control volume must be the same, i.e. m = m+ . For
s
g
Figure 7.1: A general natural convective loop.

115
7.1. Modeling
116
+
m
_
m
s
ds
Figure 7.2: Mass flows in an elemental control volume.

a loop of constant cross-sectional area, this implies that u is the same into and out of the control
volume. Thus u is independent of s, and must be a function of t alone.
7.1.2
Momentum equation
The forces on an element of length ds, shown in Fig. 7.3, in the positive s direction are: fv , the
viscous force, fp , the pressure force, and fg , the component of the gravity force. We can write
fv = w P ds
p
fp = A ds
s
fg = A ds g
where w is the wall shear stress, and p is the pressure in the fluid. It is impossible to determine
the viscous force fv through a one-dimensional model, since it is a velocity profile in the tube that
is responsible for the shear streass at the wall. For simplicity, however, we will assume a linear
relationship between the wall shear stress and the mean fluid velocity, i.e. w = u. For Poiseuille
flow in a duct, which is strictly not the case here but gives an order of magnitude value for the
coefficient, this would be
=
8
D
The local component of the acceleration due to gravity has been written in terms of
g(s) = g cos
dz
=g
ds
(7.1)
(7.2)
where g is the usual acceleration in the vertical direction, g is its component in the negative s
direction, and dz is the difference in height at the two ends of the element, with z being measured
upwards. The integral around a closed loop should vanish, so that
Z
g(s) ds = 0
(7.3)
The density in the gravity force term will be taken to decrease linearly with temperature, so that
= 0 [1 (T T0 )]
7.1. Modeling
117
Since the mass of the element is 0 A ds, we can write the momentum equation as
0 A ds
du
= fv + fp + fg
dt
from which we get

P
1 p
du
+
u=
[1 (T T0 )] g
dt
0 A
0 s
Integrating around the loop, we find that the pressure term disappears, and
Z
P
L
du
+
u=
T g(s) ds
dt
0 A
L 0
(7.4)
(7.5)
(7.6)
where u = u(t) and T = T (s, t).

7.1.3
Energy equation
Fig. 7.4 shows the heat rates going into and out of an elemental control volume. The heat rate going
in is given by
Q = 0 Aucp T kA
T
s
where the first term on the right is due to the advective and second the conductive transports. cp
is the specific heat at constant pressure and k is the coefficient of thermal conductivity. The heat
rate going out is
Q+ = Q +
Q
ds
s
The difference between the two is

Q+ Q =
Q
ds
s
2T
T
kA 2
= 0 Aucp
s
s
ds
(7.7)
Furthermore, heat is gained from the side at a rate Q, which can be written as
Q = q ds
(7.8)
where q is the rate of gain of heat per unit length of the duct.
An energy balance for the elemental control volume gives
Q + Q = Q+ + 0 A ds cp
T
t
(7.9)
where the last term is the rate of accumulation of energy within the control volume.
Substituting equations (7.7) and (7.8) in (7.9) we get the energy equation
T
q
k 2T
T
+u
=
+
t
s
0 Acp
0 cp s2
(7.10)
7.2. Known heat rate
118
fv
fg
fp
p+dp
p
ds
gravity
Q
Q
_
Q
s
ds
Figure 7.4: Heat rates on an elemental control volume.
7.2
Known heat rate
[?, ?]
The simplest heating condition is when the heat rate per unit length, q(s), is known all along
the loop. For zero mean heating, we have
Z L
q(s) ds = 0
(7.11)
0
q(s) > 0 indicates heating, and q(s) < 0 cooling.

7.2.1
Steady state, no axial conduction
Neglecting axial conduction, the steady-state governing equations are

Z
L
P
u=
T (s)
g (s) ds
0 A
L 0
u
dT
q(s)
=
ds
0 Acp
(7.12)
(7.13)
The solution of equation (7.13) gives us the temperature field

Z s
1
T (s) =
q(s ) ds + T0
0 Acp u 0
where T (0) = T0 . Using equation (7.3) it can be checked that T (L) = T0 also. Substituting in
equation (7.12), we get
Z L Z s
P
(7.14)
q(s ) ds g(s) ds
u=
0 A
0 Acp Lu 0
0
119
Figure 7.5: Bifurcation with respect to parameter H.

from which
s
u=
P Lcp
q(s ) ds
g(s) ds
Two real solutions exist for

Z
q(s ) ds
g(s) ds 0
and none otherwise. Thus there is a bifurcation from no solution to two as the parameter H passes
through zero, where
Z L Z s
q(s ) ds g(s) ds
H=
0
The pressure distribution can be found from equation (7.4)
dp
P u
=
0 1 (T T0 ) g
ds
A
Z s
P u
q(s ) ds g
0 g +
=
A
Acp u 0
(7.15)
(7.16)
from which
P u
s 0
p(s) = p0
A
g(s ) ds +
Acp u
"Z
q(s ) ds
g(s ) ds
where p(0) = p0 . Using equations (7.3) and (7.14), it can be shown that p(L) = p0 also.
Example 7.1
Find the temperature distributions and velocities in the three heating and cooling distributions corresponding to Fig. 7.6. (a) Constant heating between points c and d, and constant cooling between h and a.
(b) Constant heating between points c and d, and constant cooling between g and h. (c) Constant heating
between points d and e, and constant cooling between h and a. (d) Constant heating between points a and c,
and constant cooling between e and g. The constant value is q, and the total length of the loop is L.
120
Figure 7.6: Geometry of a square loop.

Let us write
F (s) =
q(s ) ds
G(s) = F (s)g(s)
Z L
H=
G(s) ds
0
The functions F (s) and G(s) are shown in Fig. 7.7. The origin is at point a, and the coordinate s runs
counterclockwise. The integral H in the four cases is: (a) H = 0, (b) H = qL/8, (c) H =
q L/8, (d) H = qL/4.
The fluid velocity is
s
H
u=
P Lcp
No real solution exists for case (c); the velocity is zero for (a); the other two cases have two solutions each, one
positive and the other negative. The temperature distribution is given by
T T0 =
F (s)
0 Acp u
The function F (s) is shown in Fig. 7.7. There is no real; solution for case (c); for (a), the temperature is
unbounded since the fluid is not moving; for the other two cases there are two temperature fields, one the
negative of the other.
The pressure distribution can be found from equation (7.16).
Example 7.2
What is the physical interpretation of condition (7.15)?
Let us write
H=
L
0
Z L Z
q(s ) ds
g(s) ds
q(s ) ds
0
s
0
q(s ) ds
L Z
0
g(s ) ds
g(s ) ds
L
0
q(s)
g(s ) ds
0
ds
121
s
(a)
(b)
s
(c)
s
(d)
Figure 7.7: Functions F (s) and G(s) for the four cases.
122
The first term on the right vanishes due to equations (7.11) and (7.3). Using equation (7.2), we find that
Z L
H = g
q(s)z(s) ds
(7.17)
0
The function z(s) is another way of describing the geometry of the loop. We introduce the notation
q(s) = q + (s) q (s)
where
q(s) for q(s) > 0

q+ =
0
for q(s) 0
and
0
for q(s) 0
q =
q(s) for q(s) < 0
Equations (7.11) and (7.17) thus becomes
Z L
Z L
q + (s) ds =
q (s) ds
0
(7.18)
(7.19)
H = g
q + (s)z(s) ds
L
q (s)z(s) ds
(7.20)
From these, condition (7.15) which is H 0 can be found to be equivalent to

RL +
RL
q (s)z(s) ds
0 q (s)z(s) ds
< 0R L
RL
+ (s) ds
q
0
0 q (s) ds
This implies that the height of the centroid of the heating rate distribution should be above that of the cooling.
7.2.2
Axial conduction effects
To nondimensionalize and normalize equations (7.6) and (7.10), we take

t
s
s =
L
t =
u
V G1/2
T T0
T =
T G1/2
g
g =
g
q
q =
qm
u =
where
P L
0 A
P 2 2 L
T =
g20 A2
0 A
=
P
qm g20 A2
G= 3 3
P Lcp
V =
123
Substituting, we get
Z 1
du
T g ds
+
u
=
dt
0
T
2T
1/2 T
1/2
+
G
u
=
G
q
+
K
t
s
s2
(7.21)
(7.22)
where
K=
kA
P L2 cp
The two nondimensional parameters which govern the problem are G and K.
Under steady-state conditions, and neglecting axial conduction, the temperature and velocity
are
Z s
1
T (s) =
q (s1 ) ds1
u 0
s
Z 1 Z s
q (s1 ) ds1 g (s ) ds
u =
0
All variables are of unit order indicating that the variables have been appropriately normalized.
For = 8/D, A = D2 /4, and P = D, we get
4
1
Gr
D
G=
8192 P r
L
2
1
D
K=
32 P r
L
where the Prandtl and Grashof numbers are
cp
k
qm gL3
Gr =
2k
Pr =
respectively. Often the Rayleigh number defined by

Ra = Gr P r
is used instead of the Grashof number.
Since u is of O(1), the dimensional velocity is of order (8L/D2 )Gr1/2 . The ratio of axial
conduction to the advective transport term is
K
G1/2
1/2
8
=
Ra
124
Taking typical numerical values for a loop with water to be: = 998 kg/m3 , = 1.003 103 kg/m
s, k = 0.6 W/m K, qm = 100 W/m, g = 9.91 m/s2 , = 0.207 103 K1 , D = 0.01 m, L = 1 m,
cp = 4.18 103 J/kgK, we get the velocity and temperature scales to be
V G1/2 =
T G1/2 =
and the nondimensional numbers as
G = 1.86 102
(7.23)
(7.24)
11
(7.25)
12
(7.26)
(7.27)
K = 4.47 10
Gr = 3.35 10
Ra = 2.34 10
= 3.28 10
Axial conduction is clearly negligible in this context.

For a steady state, equations (7.21) and (7.22) are
Z 1
u =
T g ds
0
d2 T
dT
2 u = q (s )
ds
ds
Integrating over the loop from s = 0 to s = 1, we find that continuity of T and equation (7.11)
imply continuity of dT /ds also.

Conduction-dominated flow
If = G1/2 / 1, axial conduction dominates. We can write
u = u0 + u1 + 2 u2 + . . .
T (s) = T 0 (s) + T 1 (s) + 2 T 2 (s) + . . .
where, for convenience, the asterisks have been dropped. Substituting into the governing equations,
and collecting terms of O(0 ), we have
Z 1
u0 =
T 0 g ds
0
d2 T 0
=0
ds2
The second equation, along with conditions that T 0 and dT 0 /ds have the same value at s = 0 and
s = 1, gives T 0 = an arbitrary constant. The first equation gives u0 = 0.
The terms of O() give
Z 1
u1 =
T 1 g ds
(7.28)
0
d2 T 1
dT 0
= q(s) + u0
ds2
ds
(7.29)
125
The second equation can be integrated once to give

Z s
dT 1
q(s ) ds + A
=
ds
0
and again
T1 =
"Z
q(s ) ds
ds + As + B
Continuity of T 1 (s) and dT 1 /ds at s = 0 and s = 1 give

#
Z "Z
s
B=
q(s ) ds
ds + A + B
A=A
respectively, from which
A=
and that B can be arbitrary. Thus

Z "Z
s
q(s ) ds
T1 =
"Z
q(s ) ds
ds + s
1
0
"Z
ds
q(s ) ds
ds + T1 (0)
where T (0) is an arbitrary constant. Substituting in equation (7.28), gives

#
#
)
Z
Z 1 "Z s
Z 1 (Z s "Z s
q(s ) ds ds
q(s ) ds ds g ds +
u1 =
0
s
g ds
The temperature distribution is determined by axial conduction, rather than by the advective velocity, so that the resulting solution is unique.
Advection-dominated flow
u=
T g ds
d2 T
dT
=q+ 2
ds
ds
where 1. Expanding in terms of , we have

u = u0 + u1 + 2 u2 + . . .
T = T 0 + T 1 + 2 T 2 + . . .
126
To O(0 ), we get
u0 =
T 0 g ds
u0
dT 0
=q
ds
from which
Z s
1
q(s ) ds
u0 0
Z 1 Z s
q(s ) ds g ds
u0 =
T0 =
Axial conduction. therefore, slightly modifies the two solutions obtained without it.
7.2.3
Toroidal geometry
The dimensional gravity function can be expanded in a Fourier series in s, to give

X
2ns
2ns
gnc cos
g(s) =
+ g1s sin
L
L
n=1
The simplest loop geometry is one for which we have just the terms
g(s) = g1c cos
2s
2s
+ g1s sin
L
L
(7.30)
corresponds to a toroidal geometry. Using

g 2 = (g1c )2 + (g1s )2
gc
0 = tan1 1s
g1
equation (7.30) becomes
g(s) = g cos
2s
0
L
Without loss of generality, we can measure the angle from the horizontal, i.e. from three oclock
point, and take 0 = 0 so that
g(s) = g cos (2s/L)
The nondimensional gravity component is
g = cos(2s)
where the * has been dropped.
Assuming also a sinusoidal distribution of heating
q(s) = sin(2s )
127
the momentum and energy equations are

u=
T (s) cos(2s) ds
(7.31)
dT
d2 T
= sin(2s ) + 2
ds
ds
(7.32)
The homogeneous solution is

T h = Beus/ + A
The particular integral satisfies
u dT p
d2 T p
1
= sin(2s )
2
ds
ds
Integrating, we have
1
u
dTp
Tp =
cos(2s )
ds
2
1
[cos(2s) cos + sin(2s) sin ]
=
2
Take
T p = a cos(2s) + b sin(2s)
from which
dT p
= 2a sin(2s) + 2b cos(2s)
ds
Substituting and collecting the coefficients of cos(2s) and sin(2s), we get
u
cos
a + 2b =
2
u
sin
2a b =
2
The constants are
(u/22 ) cos + (1/) sin
4 2 + u2 /2
(1/) cos + (u/22 ) sin
b=
4 2 + u2 /2
a=
The temperature field is given by

T = Th + Tp
Since T (0) = T (1), we must have B = 0. Taking the other arbitrary constant A to be zero, we have
1
1
1
u
u
cos
+
sin
cos(2s)
+
cos
+
sin
sin(2s)
T =
22
22
4 2 + u2 /2
128
The momentum equation gives

u=
(u/22 ) cos + (1/) sin

2(4 2 + u2 /2 )
which can be written as
u + u 4
cos sin = 0
4
2
3
2 2
(7.33)
Special cases are:

=0
Equations (7.31) and (7.32) can be solved to give

1
[cos(2s) cos + sin(2s) sin ]
2u
r
cos
u=
4
T =
On the other hand substituting = 0 in equation (7.33) gives an additional spurious solution
u = 0.

We get that u 0.
=0
We get
u=
1
2 2
4 4
q
1 4 2 2
4
The last two solutions exist only when < (16 3 )1/2 .
= /2
The velocity is a solution of

u3 + u4 2 2
=0
2
Figure 7.8 shows u- curves for three different values of . Figure 7.9 and 7.10 show u- curves
for different values of . It is also instructive to see the curve u-Ra, shown in Figure 7.11, since the
Rayleigh number is directly proportional to the strength of the heating.
The bifurcation set is the line dividing the regions with only one real solution and that with
three real solutions. A cubic equation
x3 + px + q = 0
129
= 0.001
0.5
0
-0.5
-1
-2
-3
-1
1
u
= 0.01
0.5
0
-0.5
-1
-2
-3
-1
1
u
0.5
= 0.1
0
-0.5
-1
-2
-3
-1
Figure 7.8: u- curves.
0.2
u
=0
0.1
0.02 0.04 0.06 0.08 0.1
-0.1
-0.2
0.2
u
=0.01
0.1
-0.1
0.02 0.04 0.06 0.08 0.1
-0.2
0.2
u
=0.01
0.1
-0.1
0.02 0.04 0.06 0.08 0.1
-0.2
Figure 7.9: u- curves.
130
=/4
0.2
0.1
-0.1
0.02
0.04
0.06
0.08
-0.2
0.2
u
0.15
=/4
0.1
0.05
0.02
u
0.04
0.06
0.08
0.1
0.1
0.08
0.06
=/4
0.04
0.02
0.02
0.04
0.06
0.08
0.1
Figure 7.10: More u- curves.
0.2
0.1
10000
20000
30000
40000
Ra
-0.1
-0.2
Figure 7.11: u-Ra for = 0.01 radians.
131
0.045
0.04
0.035
0.03
0.025
0.02
0.015
0.01
0.005
0
100
80
60
40
20
20
40
60
80
100
(degrees)
Figure 7.12: Region with three solutions.

has a discriminant
D=
p3
q2
+
27
4
For D < 0, there are three real solutions, and for D > 0, there is only one. The discriminant for the
cubic equation (7.33) is
3
1
1
1
2
4 2 2
cos + ( sin )
D=
27
4
4
The result is shown in Fig. 7.12.
7.2.4
Dynamic analysis
We rescale the nondimensional governing equations (7.21) and (7.22) by

1
u
b
2G1/2
1
Tb
T =
2G1/2
u =
to get
du
+u=
dt
T g ds
1 T
G1/2 K 2 T
T
+
u
= Gq +
t
2 s
2 s2
where the hats and stars have been dropped.
We take g = cos(2s) and q = sin(2s ). Expanding the temperature in a Fourier series,
we get
T (s, t) = T0 (t) +
[Tnc (t) cos(2ns) + Tns (t) sin(2ns)]
n=1
Substituting, we have
du
1
+ u = T1c
dt
2
132
and

dT0 X dTnc
dTnc
+
cos(2ns) +
sin(2ns)
dt
dt
dt
n=1
+u
[nTnc sin(2ns) + nTns cos(2ns)]
n=1
= G [sin(2s) cos cos(2s) sin ]
X
[Tnc cos(2ns) + Tns sin(2ns)]
2n2 G1/2 K
n=1
Integrating, we get
dT0
=0
dt
Multiplying by cos(2ms) and integrating
c
m s
1
1 dTm
c
+ uTm
= G sin m2 G1/2 KTm
2 dt
2
2
Now multiplying by sin(2ms) and integrating

s
1 dTm
m c
1
s
uTm
= G cos m2 G1/2 KTm
2 dt
2
2
Choosing the variables

x=u
1
y = T1c
2
1
z = T1s
2
and the parameters
G
sin
2
G
b = cos
2
c = 2G1/2 K
a=
(7.34)
(7.35)
(7.36)
we get the dynamical system

dx
=yx
dt
dy
= a xz cy
dt
dz
= b + xy cz
dt
(7.37)
(7.38)
(7.39)
133
c2
Figure 7.13: Bifurcation diagram for x.

The physical significance of the variables are: x is the fluid velocity, y is the horizontal temperature
difference, and z is the vertical temperature difference. The parameter c is positive, while a and b
can have any sign.
The critical points are found by equating the vector field to zero, so that
yx=0
a xz cy = 0
b + xy cz = 0
(7.40)
(7.41)
(7.42)
From equation (7.40), we have y = x, and from equation (7.42), we get z = (b+x2 )/c. Substituting
these in equation (7.41), we get
x3 + x(c2 b) ac = 0
(7.43)
This corresponds to equation (7.33), except in different variables.

To analyze the stability of a critical point (x, y, z) we add perturbations of the form
x = x + x
y = y + y
(7.44)
(7.45)
z = z + z
(7.46)
Substituting in equation (7.37)-(7.39), we get the local form

1 1
0
x
0
x
d
y
= z c x y + x z
dt
z
z
y
x c
x y
The linearized version is

x
1
d
y
= z
dt
y
z

x
1
0
c x y
z
x c
No tilt, with axial conduction (a = 0, c 6= 0)

From equation (7.43), for a = 0 we get
x3 + x(c2 b) = 0
(7.47)
(7.48)
134
from which
The z coordinate is
0
b c2
x=y=

b c2
b/c
c
z=
The bifurcation diagram is shown in Figure 7.13.
Stability of conductive solution

The critical point is (0, 0, b/c). To examine its linear stability, we look at the linearized
equation (7.48) to get

1 1
0
x
x
d
y
(7.49)
= b/c c 0 y
dt
z
0
0 c
z
The eigenvalues of the matrix are obtained from the equation
(1 + )
1
0
b/c
(c
+
)
0
0
0
(c + )
which simplifies to
=0
b
(c + ) (1 + )(c + )
=0
c
One eigenvalue is
1 = c
Since c 0 this eigenvalue indicates stability. The other two are solutions of
b
2 + (c + 1) + (c ) = 0
c
which are
"
r
1
(c + 1) (c + 1)2 4(c
2 =
2
"
r
1
(c + 1) + (c + 1)2 4(c
3 =
2
#
b
)
c
#
b
)
c
2 is also negative and hence stable. 3 is negative as long as

r
b
(c + 1) + (c + 1)2 4(c ) < 0
c
135
which gives
b < c2
This is the condition for stability.
In fact, one can also prove global stability of the conductive solution. Restoring the nonlinear
terms in equation (7.47) to equation (7.49), we have
dx
= y x
dt
b
dy
= x cy x z
dt
c
dz
= cz + x y
dt
Let
b
E(x, y, z) = x2 + y 2 + z 2
c
Thus
1 dE
b dx
dy
dz
= x
+ y
+ z
2 dt
c dt
dt
dt
b 2 2b
2
= x + x y cy cz 2
c
c
b
b
= (x y )2 (c )y 2 cz 2
c
c
Since
E
dE
dt
0
0
for 0 b c2 , E is a Liapunov function, and the critical point is stable to all perturbations in this
region. The bifurcation at b = c2 is thus supercritical.
Stability of convective solution
For b > c2 , only one critical point ( b c2 , b c2 , c) will be considered, the other being
similar. We use the linearized equations (7.48). Its eigenvalues are solutions of
(1 + )
0
2 =0
c
c
(c
+
)
b c2
b c2 (c + )
This can be expanded to give
3 + 2 (1 + 2c) + (b + c) + 2(b c2 ) = 0
136
The Hurwitz criteria for stability require that all coefficients be positive, which they are. Also the
determinants
D1 = 1 + 2c
1 + 2c
D2 =
1
1 + 2c
1
D3 =
2(b c2 )
b+c
2(b c2 )
b+c
1 + 2c
0
0
2(b c2 )
should be positive. This requires that
c(1 + 4c)
if
1 2c
c(1 + 4c)
b>
if
1 2c
b<
c < 1/2
c > 1/2
With tilt, no axial conduction (a 6= 0, c = 0)

The dynamical system (7.37)-(7.39) simplifies to
dx
=yx
dt
dy
= a xz
dt
dz
= b + xy
dt

The critical points are ( b, b, a/ b). The linear stability of the point P + given by ( b, b, a/ b)
will be analyzed. From equation (7.48), the solutions of
(1 + ) 1
0
a/ b b = 0
b
b
are the eigenvalues. This simplifies to
a
3 + 2 + (b + ) + 2b = 0
b
(7.50)
For stability the Hurwitz criteria require all coefficients to be positive, which they are. The determinants
D1 = 1
D2 =
D3 =
1
1
1
1
0
2b
b + a/ b
2b
b + a/ b
1
0
0
2b
137
3/2
a=b
+
P stable
Both P +and P
unstable
G/2
_
P stable
3/2
a=-b
Figure 7.14: Stability of critical points P + and P in (b, a) space.
should also be positive. This gives the condition (b + a/ b) 2b > 0, from which, we have
a > b3/2
(7.51)
for stability. The stable and unstable region for P +

isshownin Figure 7.14. Also shown is the
stability of the critical point P with coordinates ( b, b, a/ b). The dashed circles are of radius
G/2, and the angleof tile is also indicated. Using equations (7.34) and (7.35), the stability condition
(7.51) can be written as
1/2
sin
G
>
2
cos3/2
As a numerical example, for the value of G in equation (7.23), P + is stable for the tilt angle range
> 7.7 , and P is stable for < 7.7 . In fact, for G 1, the stability condition for P + can be
approximated as
1/2
G
>
2
The same information can be shown in slightly different coordinates. Using x = b for P +
and equation (7.35), we get
G
x2
=
2
cos
The stability condition (7.51) thus becomes
tan < x
The stability regions for both P + and P are shown in Figure 7.15.
3/2
The loss of stability is through imaginary eigenvalues. In fact, for P + , substituting
in
a=b
equation (7.50), the equation can be factorized to give the three eigenvalues 1, i 2b. Thus
the
nondimensional radian frequency of the oscillations in the unstable range is approximately 2b.
The effect of a small nonzero axial conduction parameter c is to alter the Figure 7.51 in the
zone 0 < b < c.
138
P +exists
but unstable
P+stable
_
P stable
_
P exists
but unstable
Figure 7.15: Stability of critical points P + and P in (, x) space.

7.2.5
Nonlinear analysis
Numerical
Let us choose b = 1, and reduce a. Figures 7.16 and 7.17 show the x-t and phase space representation
for a = 0.9, Figures 7.18 and 7.19 for a = 0.55, and Figures 7.20 and 7.21 for a = 0.53.
The strange attractor is shown in Figures 7.22 and 7.23.
Comparison of the three figures in Figures 7.24 shows that vestiges of the shape of the closed
curves for a = 0.9 and a = 0.9 can be seen in the trajectories in a = 0.
Analytical
The following analysis is by W. Franco.
We start with the dynamical system which models a toroidal thermosyphon loop with known
heat flux
dx
=yx
dt
dy
= a zx
dt
dz
= xy b
dt
For b > 0 two critical points P + and P appear
a
b, b,
(
x, y, z) =
b
139
1.8
1.6
1.4
1.2
0.8
0.6
0.4
0.2
10
15
t
20
25
30
Figure 7.16: x-t for a = 0.9, b = 1.
2.5
2
1.5
1
0.5
0
0.5
2.5
2
1.5
1.5
1
1
0.5
0.5
0
y
0.5
Figure 7.17: Phase-space trajectory for a = 0.9, b = 1.
140
2.5
1.5
0.5
0.5
10
15
t
20
25
30
Figure 7.18: x-t for a = 0.55, b = 1.
4
3
2
1
0
1
2
4
3
2.5
2
1.5
0.5
0
1
y
0.5
1
141
2.5
1.5
0.5
0.5
10
15
t
20
25
30
Figure 7.20: x-t for a = 0.53, b = 1.
4
3
2
1
0
1
2
3
4
3
2.5
2
1.5
0.5
0
1
y
0.5
1
142
50
100
150
t
200
250
300
Figure 7.22: x-t for a = 0, b = 1.
6
4
2
0
2
4
6
5
10
Figure 7.23: Phase-space trajectory for a = 0, b = 1.
143
a=0
6
4
2
4
2
0
0
2
2
4
4
x
a = 0.9
6
4
2
4
2
0
0
2
2
4
4
x
a = - 0.9
6
4
2
4
2
0
0
2
y
2
4
4
x
Figure 7.24: Phase-space trajectories for b = 1.
144
The local form respect to P + is

dx
= y x
dt
dy
a
= x bz
dt
b

dz
= bx + by
dt
3
3
For stability a > b 2 . At a = b 2 the eigenvalues are 1, 2bi, thus a nonlinear analysis through
3
the center manifold projection is possible. Lets introduce a perturbation of the form a = b 2 + and
the following change of variables
a
=
b
= b
rewriting the local form, dropping the primes and regarding the perturbation the system becomes
dx
=yx
dt
dy
2
x z
= +
dt
dz
= x y
dt
for stability > 2 .
Lets apply the following transformation:
x = w1 +
2
2 2
w
+
2
2 2 + 1
2 2 + 1
y = 2w2
2 2 2 + 1
2
w2 +
w3
z = w1 2
2 + 1
2 2 + 1
(7.52)
in the new variables
1
w
= 0
0
0
0
2
0
+ l(w)
2 w + Pw
0
(7.53)
The center manifold projection is convenient to use if the large-time dynamic behavior is of
interest. In many dimensional systems, the system often settles into the same large-time dynamics
irrespective of the initial condition; this is usually less complex than the initial dynamics and can
be described by far simple evolution equations.
We first state the definition of an invariant manifold for the equation
x = N (x)
(7.54)
145
where x Rn . A set S Rn is a local invariant manifold for (7.54) if for x0 S, the solution x(t)
of (7.54) is in S for | t |< T where T > 0. If we can always choose T = , then S is an invariant
manifold. Consider the system
x = Ax + f (x, y)
y = By + g(x, y)
(7.55)
where x Rn , y Rm and A and B are constant matrices such that all the eigenvalues of A have
zero real parts while all the eigenvalues of B have negative real parts. If y = h(x) is an invariant
manifold for (??) and h is smooth, then it is called a center manifold if h(0) = 0,h (0) = 0. The flow
on the center manifold is governed by the n-dimensional system
x = Ax + f (x, h(x))
The last equation contains all the necessary information needed to determine the asymptotic behavior
of small solutions of (??).
Now we calculate, or at least approximate the center manifold h(w). Substituting w1 =
h(w2 , w3 ) in the first component of (7.53) and using the chain rule, we obtain
w
h h 2
= h + l1 (w2 , w3 , h)
(7.56)
w 1 =
,
w2 w3
w3
We seek a center manifold
h = aw22 + bw2 w3 + cw32 + O(3)

substituting in (7.56)
2w3
=
(2aw2 + bw3 , 2cw3 + bw2 )
2w2
aw22 + bw2 w3 + cw32 +
Equating powers of x2 ,xy and y 2 , we find that
k1 w22 + k2 w2 w3 + k3 w32 + O(3)
a = k1 b 2
c = k3 + b 2
k2 + 2 2 (k1 k3 )
b=
8 2 + 1
The reduced system is therefore given by
w 2 = 2w3 + s2 (w2 , w3 )
w 3 = 2w2 + s3 (w2 , w3 )
Normal form: Now we carry out a smooth nonlinear coordinate transform of the type
w = v + (v)
(7.57)
146
to simplify (7.57) by transforming away many nonlinear terms. The system in the new coordinates
is

(v1 v2 ) v12 + v22

2
0
v =
(7.58)
+
2
0
(v2 + v1 ) v12 + v22
where and depend on the nonlinear part of (7.57). This is the unfolding of the Hopf bifurcation.
Although the normal form theory presented in class pertains to a Jacobian whose eigenvalues
all lie on the imaginary axis, one can also present a perturbed version. The eigenvalues are then
close to the imaginary axis but not quite on it. Consider the system
+ f (v)
v = Av + Av
(7.59)
where the Jacobian A has been evaluated at a point in the parameter space where all its eigenvalues
represents a linear expansion of order in the parameters above that
are on the imaginary axis, A
point; a perturbed Jacobian. The perturbation parameter represents the size of the neighborhood in
the parameter space. We stipulate the order of such that the real part of the eigenvalues of A + A
is such that, to leading order, A does not change the coefficients of the leading order nonlinear terms
of perturbed Hopf can always be transformed
of the transformed equation. The linear part A + A
to

The required transformation is a near identity linear transformation
v = u + Bu
such that the linear part of (7.59) is transformed to
z
u = A + AB BA + A
For the Hopf bifurcation if
=
A
a1
a3
a2
a4
then
=
a1 + a4
Therefore for small we can write (7.58) as

v =
0
2
p22
2
v+
p
0
32
p23
p33
v+
f1 (v)
f2 (v)
where the perturbation matrix comes from (7.53). Applying a near identity transformation of the
form v = u + Bu the system becomes

(u1 u2 ) u21 + u22
u =
u+
2
2
2
(u2 + u1 ) u1 + u2
7.3. Known wall temperature
147
which is the unfolding for the perturbed Hopf bifurcation. In polar coordinates we have
r = r + r3
= 2
where
2
= 2
2 (2 2 + 1)
=
40 6 + 40 4 + 12 3 + 10 2 + 12 + 3
4 (8 2 + 1) (2 2 + 1)
Appendix
k1 =
k2 =
k3 =
8 2 + 1
3
(2 2 + 1)

2 + 1 4 2 8 2 2
(2 2 + 1)
2
8 + 1
(2 2 + 1)
(2 2 + 1)
2
= 2
2 + 1
p22 =
p23
From the literature
1
(fxxx + fxyy + gxxy + gyyy )
=
16
1
(fxy (fxx + fyy ) gxy (gxx gyy ) fxx gxx fyy gyy )
16
in our problem f = f1 , g = f2 , x = v1 , y = v2 and = 2.

+
7.3
Known wall temperature
The heating is now convective with a heat transfer coefficient U , and an external temperature of
Tw (s). Thus,
q = P U (T Tw )
Neglecting axial conduction
Z
P
L
u=
T (s)
g (s) ds
0 A
L 0
dT
u
= T Tw (s)
ds
(7.60)
(7.61)
7.4. Mixed condition
148
where = U P/0 Acp 1 . Multiplying the second equation by es/u /u, we get
d s/u
T = es/u Tw
e
ds
u
Integrating, we get
Z
s s /u
s/u
T =e
e
Tw (s ) ds + T0
u 0
Since T (L) = T (0), we get
eL/u
T0 =
u
RL
0
es /u Tw (s ) ds
eL/u 1
The velocity is obtained from
Z
Z
s s /u
P
L s/u
Tw (s ) ds + T0 g(s) ds
u=
e
e
0 A
L 0
u 0
This is a transcendental equation that may have more than one real solution.
Example 7.3
Show that there is no motion if the wall temperature is uniform.
Take Tw to be a constant. Then equation (7.61) can be written as
d(T Tw )
T Tw
The solution to this is
ds
u
T = Tw + K es/u
where K is a constant. Continuity of T at s = 0 and s = L gives K = 0. Hence T = Tw , and, from equation
(7.60), u = 0.
Assume the wall temperature to be

Tw (s) = sin(2s )
The temperature field is
b
T =
r2 r1
7.4
cos(2s )
2(1 + r22 /4 2 )
cos(2s )
2(1 + r1 62/4 2 )
Mixed condition
The following has been written by A. Pacheco-Vega.

It is common, especially in experiments, to have one part of the loop heated with a known heat
rate and the rest with known wall temperature. Thus for part of the loop the wall temperature is
known so that q = P U (T Tw (s)), while q(s) is known for the rest. As an example, consider
s + 2
P U (T T0 ) for 2
q=
q0
for + 2 < s < 2 + 2
where T0 and q0 are constants.
1 The
sign of appears to be wrong.
149
Constant wall
temperature T w
=0, 2
Cooling
water out
g
~
d=2r
R
=
Cooling
water in
Uniform
heat flux q
Figure 7.25: Schematic of a convection loop heated with constant heat flux in one half and cooled
at constant temperature in the other half.
7.4.1
Modeling
[?]
If we consider a one-dimensional incompressible flow, the equation of continuity indicates that
the velocity v is a function of time alone. Thus,
v = v(t).
Taking an infinitesimal cylindrical control volume of fluid in the loop r2 d, see Figure (7.25), the
momentum equation in the -direction can be written as
r2 Rd
dp
dv
= r2 d
gr2 Rd cos( + ) w 2rRd
dt
d
(7.62)
Integrating Eq. (7.62) around the loop using the Boussinesq approximation = w [1 (T
Tw )], with the shear stress at the wall being approximated by that corresponding to Poiseuille flow
in a straight pipe w = 8v/w r2 , the expression of the balance in Eq. (7.62) modifies to
8
g
dv
+
v=
dt
w r2
2
(T Tw ) cos( + ) d
(7.63)
Neglecting axial heat conduction, the temperature of the fluid satisfies the following energy balance
equation
2h (T Tw ), 0
r
v T
T
=
+
w cp
(7.64)
t
R
2
q,
<
<
2
r
150
Following the notation used by Greif et al. (1979), the nondimensional time, velocity and temperature are defined as
=
t
,
2R/V
w=
v
,
V
T Tw
q/h
respectively, where
V =
gRrq
2cp
1/2
Accordingly, Eqs. (7.63) and (7.64) become

dw
+ w =
d
4D
cos( + ) d
(7.65)
and
+ 2w
=
2D,
2D,
0
< < 2
(7.66)
where the parameters D and are defined by

D=
7.4.2
2Rh
w cp rV
16R
w r2 V
Steady State
The steady-state governing equations without axial conduction are

Z 2
w=
cos( + ) d
4D 0
(7.67)
and
D
, 0
w
d
=
D
w ,
(7.68)
< < 2
where w and are the steady-state values of velocity and temperature respectively. Eq. (7.68) can
be integrated to give
A e(D/w) , 0
() =
D
< < 2
w + B,
Applying the condition of continuity in the temperature, such that (0) = (2) and ( ) = ( + )
the constants A and B can be determined. These are
D
1
D 2 e(D/w) 1
A=
B
=
w 1 e(D/w)
w 1 e(D/w)
151
The resulting temperature filed is
() =
D e(D/w)
w 1e(D/w) ,
D
w
2e(D/w) 1
1e(D/w)
0
i
(7.69)
, < < 2
Substituion of Eq. (7.68) in Eq. (7.67), followed by an expansion of cos( + ), leads to

Z
D e(D/w)
cos d
cos
w=
(D/w)
4D
0 w 1e
Z 2
D
2e(D/w) 1
cos
d
+
+
w
1 e(D/w)
D e(D/w)
sin d
sin
(D/w)
4D
0 w 1e
Z 2
D
2e(D/w) 1
sin d
+
+
w
1 e(D/w)
and integration around the loop, gives the steady-state velocity as
cos (D/w) cos + (D/w)2 sin 1 + e(D/w)

i
h
.
+
w2 =
D 2
2
1 e(D/w)
4 1 + w
As a final step, multiplying the numerator and denominator by e(D/2w) and rearranging terms leads
to the expresion for the function of the steady-state velocity
G(w, , D) = w2
cos (D/w) cos + (D/w)2 sin

h
coth(D/2 w) = 0
D 2 i
2
4 1 + w
(7.70)
For = 0, symmetric steady-state solutions for the fluid velocity are possible since G(w, 0, D) is an
even function of w. In this case Eq.(7.70) reduces to
(D/w) 1 + e(D/w)
1
2
i
.
(7.71)
w = + h
2 4 1 + D 2 1 e(D/w)
w
The steady-state solutions of the velocity field and temperature are shown next. Figure 7.26
shows the w curves for different values of the parameter D. Regions of zero, one, two and
three solutions can be identified. The regions of no possible steady-state velocity are: 180 < <
147.5 and 147.5 < < 180 . There is only one velocity for the ranges 147.5 < < 0
and 0 < < 147.5 where 0 varies from 90 at a value of D = 0.001 to 0 = 32.5 when
D = 100. Three velocities are obtained for 0 < < 32.5 and 32.5 < < 0 , except for the
zero-inclination case which has two possible steady-state velocities. The temperature distribution
in the loop, for three values of the parameter D and = 0 is presented in Figure 7.27. From
the curves it can be seen the dependence of the temperature with D. As D increases the
variation in temperature between two opposit points also increases. When has a value D = 0.1 the
heating and cooling curves are almost straight lines, while at a value of D = 1.0 the temperature
152
1
D=10
0.8
D=2.5
0.6
0.4
D=1
0.2
0
-147.5
D=0.1
32.5
-32.5
147.5
0.2
0.4
0.6
0.8
1
150
100
50
50
100
150
Figure 7.26: Steady-State velocity field.

3
D=2.5
2.5
(,D, =0)
2
D=1.0
1.5
D=0.1
0.5
0
0
Figure 7.27: Nondimensional temperature distribution as a function of the parameter D for = 0.

decays exponentially and rises linearly. Similar but more drastic change in temperature is seen when
D = 2.5. Figure 7.28 shows the curves for three different inclination angles with D = 2.5. It
can be seen the increase in the temperature as takes values of = 0 , = 90 and = 135 . This
behaviour is somewhat expected since the steady-state velocity is decreasing in value such that the
fluid stays longer in both parts of the loop. Figure 7.29 shows the steady-state velocity as a function
of D for different angles of inclination . For = 0 we have two branches of the velocity-curve
which are symmetric. The positive and negative values of the velocity are equal in magnitude for
any value of D. For = 45 , the two branches are not symmetric while for = 90 and = 135 ,
only the positive branch exist.
153
= 135
(,D=2.5, )
4
=90
3
=0
0
0
Figure 7.28: Nondimensional temperature distribution as a function of thermosyphon inclination

for D = 2.5.
=0
=45
=90
0.8
0.6
=135
0.4
0.2
0
0.2
0.4
=45
0.6
0.8
=0
1
1
10
10
10
10
Figure 7.29: Velocity w as a function of D for different thermosyphon inclinations .
7.4.3
154
Dynamic Analysis
The temperature can be expanded in Fourier series, such that

= 0 +
[cn (t) cos(n) + sn (t) sin(n)]
(7.72)
n=1
Substituiting into Eqs. (7.65) and (7.66), we have

dw
2
2
+ w =
cos c1
sin s1
d
4D
4D
(7.73)
and

dsn
dc0 X dcn
+
cos (n) +
sin (n)
d
d
d
n=1
+2w
[ncn sin (n) + nsn cos (n)]
n=1
P
2D {c0 + n=1 [cn (t) cos(n) + sn (t) sin (n)]} ,
2D,
(7.74)
< < 2
Integrating Eq. (7.74) from = 0 to = 2 we get

"
#
dc0
1 X sn
c
n
= D 0
[(1) 1] 1
d
n=1 n
(7.75)
Multiplying by cos (m) and integrating from = 0 to = 2
D X s
dcm
2n
+ 2m w sm = Dcm +
(1)m+n 1 2
d
n=1 n
n m2
(7.76)
n6=m
Now multiplying by sin (m) and integrating from = 0 to = 2
2m
D X c
dsm
2m w cm = Dsm +
n (1)m+n 1
d
n=0
m2 n2
n6=m
for m 1.
Choosing the variables
w=w
C0 = c0
Cm = cm
Sm = sm
2D
[1 (1)m ]
m
(7.77)
155
we get an infinte-dimensional dynamical system

dw
2
2
= w +
cos C1
sin S1
d
4D
4D
dC0
D X Sn
= D C0 +
[(1)n 1] + D
d
n=1 n
(7.78)
(7.79)
D X
dCm
2n
= 2m w Sm D Cm +
Sn (1)m+n 1 2
d
n=1
n m2
(7.80)
n6=m
2m
D X
dSm
= 2m w Cm D Sm +
Cn (1)m+n 1
d
n=0
m2 n2
n6=m
2D
[(1)m 1]
(7.81)
m
for m 1. The physical significance of the variables are: w is the fluid velocity, C is the horizontal
temperature difference, and S is the vertical temperature difference. The parameters of the system
are D, and . D and are positive, while can have any sign.
The critical points are found by equating the vector filed to zero, so that
+
2
2
cos C 1 +
sin S 1 = 0
4D
4D
1 X Sn
[(1)n 1] = 0
(C 0 1)
n=1 n
w
2m w S m + D C m
D X
2n
=0
S n (1)m+n 1 2
n=1
n m2
n6=m
2m
D X
2m w C m D S m +
C n (1)m+n 1
n=0
m2 n2
n6=m
2D
[(1)m 1] = 0
m
However, a convenient alternative way to determine the critical points is by using a Fourier
series expansion of the steady-state temperature field solution given in Eq. (7.69). The Fourier
series expansion is
+
n=0
C n cos(n) + S n sin(n)
Performing the inner product between Eq. (7.69) and cos(m) we have
Z
D
1
e(D/w) cos(m) d
w 1 e(D/w) 0
Z
2e(D/w) 1
D 2
cos(m) d
+
+
w
1 e(D/w)
Z 2
Z 2
X
X
=
Cn
Sn
cos(n) cos(m) d +
sin(n) cos(m)d
n=0
n=0
(7.82)
156
Now the inner product between Eq. (7.69) and sin(m) gives
Z
1
D
e(D/w) sin(m) d
w 1 e(D/w) 0
Z
2e(D/w) 1
D 2
sin(m) d
+
+
w
1 e(D/w)
Z 2
Z 2
X
X
=
cos(n) sin(m) d +
sin(n) sin(m)d
Cn
Sn
n=0
n=0
from which we get
1
D 3 2e(D/w) 1
1+
+
2
w 2
1 e(D/w)
D 2
D
1 e(D/w) cos(m)
w
=h
+ 2 2 [1 cos(m)]
D 2 i
(D/w)
m w
1e
m2 + w
D 3
(D/w)
cos(m)
1e
h w 2 i
=
(D/w)
m m2 + D
1e
C0 =
Cm
Sm
(7.83)
(7.84)
(7.85)
To analyze the stability of a critical point (w, C 0 , C 1 , , C m , S 1 , , S m ) we add perturbations of the form
w = w + w
C0 = C 0 +
C1 = C 1 +
..
.
Cm =
S1 =
..
.
C0
C1
(7.87)
(7.88)
(7.89)
C m + Cm
S 1 + S1
Sm = S m +
(7.86)
(7.90)
(7.91)
(7.92)
Sm
(7.93)
157
Substituting in Eqs. (7.78) to (7.81), we obtain the local form

dw
2
2
= w +
cos C1
sin S1
d
4D
4D
D X (1)n 1
dC0
= D C0 +
Sn
d
n=1
n
(7.94)
(7.95)
dCm
= 2m w Sm
2m S m w D Cm
d
2n
D X
(1)m+n 1 Sn 2m w Sm
2
2
n=1 n m
m1
n6=m
(7.96)
dSm
= 2m w Cm
+ 2m C m w D Sm
D X 2m
(1)m+n 1 Cn + 2m w Cm
n=0 m2 n2
n6=m
(7.97)
The linearized version is
dw
2
2
= w +
cos C1
sin S1
d
4D
4D
dC0
D X (1)n 1
= D C0 +
Sn
d
n=1
n
(7.98)
(7.99)
dCm
2m S m w D Cm
= 2m w Sm
d
2n
D X
(1)m+n 1 Sn
n=1 n2 m2
m1
D X 2m
(1)m+n 1 Cn
n=0 m2 n2
m1
n6=m
dSm
(7.100)
+ 2m C m w D Sm
= 2m w Cm
n6=m
(7.101)
In general, the system given by Eqs. (7.78) to (7.81) can be written as

dx
= f (x)
dt
(7.102)
The eigenvalues of the linearized system given by Eqs. (7.98) to (7.101), and in general form
as
dx
= Ax
dt
(7.103)
m1
158
25
20
15
Unstable
10
Stable
5
0
0
0.5
1.5
2.5
Figure 7.30: Stability curve D and for a thermosyphon inclination = 0.

are obtained numerically, such that
|A I| = 0
where A is the Jacobian matrix corresponding to the vector field of the linearized system, I is the
identity matrix, and are the eigenvalues. The neutral stability curve is obtained numerically from
the condition that () = 0. A schematic of the neutral curve is presented in Figure 7.30 for = 0.
In this figure, the stable and unstable regions can be identified. Along the line of neutral stability,
a Hopf-type of bifurcation occurs. Figure 7.31 shows the plot of w curve for a value of the
parameters D = 0.1 and = 0.20029. When = 0, a Hopf bifurcation for both the positive and
negative branches of the curve can be observed, where stable and unstable regions can be identified.
It is clear that the natural branches which correspond to the first and third quadrants are stable,
whereas the antinatural branches, second and fourth quadrants are unstable. The symmetry between
the first and third quadrants, and, between the second and fourth quadrants can be notice as well.
The corresponding eigenvalues of the bifurcation point are shown in Figure 7.32. The number of
eigenvalues in this figure is 42 which are obtained from a dynamical system of dimension 42. This
system results from truncating the infinite dimensional system at a number for which the value of
the leading eigenvalues does not change when increasing its dimension. When we increase the size of
the system, new eigenvalues appear in such a way that they are placed symmetrically farther from
the real axis and aligned to the previous set of slave complex eigenmodes. This behaviour seems to
be a characteristic of the dynamical system itself. Figure 7.33 illustrates a view of several stability
curves, each for a different value of the tilt angle in a D plane at = 0. In this plot, the
neutral curves appear to unfold when decreasing the tilt angle from 75.5 to 32.5 increasing the
region of instability. On the other hand, Figure 7.34 illustrates the linear stability characteristics of
the dynamical system in a w plot for a fixed and three values of the parameter D. The stable
and unstable regions can be observed. Hopf bifurcations occur for each branch of ecah particular
curve. However, it is to be notice that the bifurcation occurs at a higher value of the tilt angle when
D is smaller.
7.4.4
Nonlinear analysis
Let us select D = 1.5, and increase . Figures 7.35 and 7.36 show the w time series and phasespace curves for = 0.95cr , whereas Figures 7.37 and 7.38 show the results for = 1.01cr . The
plots suggest the appearance of a subcritical Hopf bifurcation. Two attractors coexist for cr ,
159
1
0.8
Stable
Unstable
0.6
Hopf Bifurcation
0.4
0.2
0
0.2
0.4
Hopf Bifurcation
0.6
Stable
0.8
Unstable
1
150
100
50
0
[ ]
50
100
150
Figure 7.31: Stability curve w vs. for D = 0.1, and = 0.20029.
150
100
()
50
50
100
150
0.5
0.4
0.3
0.2
0.1
0.1
0.2
()
Figure 7.32: Eigenvalues at the neutral curve for D = 0.1, = 0.20029 and = 0.
25
Unstable
20
15
10
=3.5
=21.5
=-14.5
=39.5
=57.5
=-32.5
=75.5
0
0
Stable
0.5
1.5
2.5
Figure 7.33: Neutral stability curve for different values of the tilt angle .
7.5. Perturbation of one-dimensional flow
160
0.8
D=0.1
0.6
0.4
D=2.5
D=1.0
0.2
w
D=1.0
0
0.2
D=0.1
0.4
0.6
D=2.5
0.8
=4.0
1
150
100
50
50
100
150
Figure 7.34: Curve w vs. = 4.0 and D = 0.1,D = 1.0,D = 2.5.

these being a critical point and a strange attractor of fractional dimension. For > cr , the only
presence is of a strange attractor.
Now we choose D = 0.1, and increase . Figure 7.39 presents a plot of the w curve for
= 0.99cr . Figures 7.40 and 7.41 show the time series plots and phase space representation for
= 1.01cr , and Figures 7.42 and 7.43 for = 20cr . In this case, for < cr we have stable
solutions. For = 1.01cr the figures show a possible limit cycle undergoes a period doubling. This
implies a supercritical Hopf bifurcation. The strange attractor is shown in Figures 7.44 and 7.45.
7.5
Perturbation of one-dimensional flow
7.6
Thermal control
Consider the control of temperature at a given point in the loop by modification of the heating. Both
known heat flux and known wall temperatures may be looked at. In terms of control algorithms,
one may use PID or on-off control.
Problems
1. Find the pressure distributions for the different cases of the square loop problem.
2. Consider the same square loop but tilted through an angle where 0 < 2. There is constant heating
between points a and c, and constant cooling between e and g. For the steady-state problem, determine the
temperature distribution and the velocity as a function of . Plot (a) typical temperature distributions for
different tilt angles, and (b) the velocity as a function of tilt angle.
3. Find the steady-state temperature field and velocity for known heating if the loop has a variable cross-sectional
area A(s).
4. Find the temperature field and velocity for known heating if the total heating is not zero.
5. Find the velocity and temperature fields for known heating if the heating and cooling takes place at two different
points. What the condition for the existence of a solution?
6. What is the effect on the known heat rate solution of taking a power-law relationship between the frictional
force and the fluid velocity?
7. For known wall temperature heating, show that if the wall temperature is constant, the temperature field is
uniform and the velocity is zero.
161
2.5
2
1.5
1
0.5
0
0.5
1
1.5
2
2.5
450
460
470
480
490
500
510
Figure 7.35: Curve w vs. for D = 1.5, = 0.95cr .
2
1
0
1
2
3
4
2
4
2
2
c
1
2
4
Figure 7.36: Phase-space trajectory for D = 1.5, = 0.95cr .
162
2.5
2
1.5
1
0.5
0
0.5
1
1.5
2
2.5
1900
1920
1940
1960
1980
2000
2
1
0
1
2
3
4
2
4
2
2
c
2
4
163
1.5
1.4
1.3
1.2
1.1
1
0.9
0.8
0
500
1000
1500
2000
2500
2
1.5
1
0.5
0
0.5
1
1.5
2
1980
1985
1990
1995
2000
164
1
0.8
0.6
0.4
c
1
0.2
0
0.2
0.4
0.6
0.8
1
1980
1985
1990
1995
2000
1
0.8
0.6
0.4
s
1
0.2
0
0.2
0.4
0.6
0.8
1
1980
1985
1990
1995
2000
165
0.5
0
0.5
1
1
0.5
2
1
0.5
c
1
1
0.6
0.4
0.2
c
1
0
0.2
0.4
0.6
0.8
1088
1090
1092
1094
1096
1098
1100
1102
Figure 7.44: Phase-space trajectory for D = 0.1, = 20cr .
166
0.2
0.1
0
0.1
0.2
0.6
0.4
4
2
0.2
0
c
2
0.2
Figure 7.45: Phase-space trajectory for D = 0.1, = 20cr .

8. Study the steady states of the toroidal loop with known wall temperature including nondimensionalization of
the governing equations, axial conduction and tilting effects, multiplicity of solutions and bifurcation diagrams.
Illustrate typical cases with appropriate graphs.
9. Consider a long, thin, vertical tube that is open at both ends. The air in the tube is heated with an electrical
resistance running down the center of the tube. Find the flow rate of the air due to natural convection. Make
any assumptions you need to.
10. For a thin, vertical pipe compare the wall shear stress to mean flow velocity relation obtained from a twodimensional analysis to that from Poiseuille flow.
11. Find the combination of fluid parameters that determines the rate of heat transfer from a closed loop with
known temperature distribution. Compare the cooling rate achieved by an ionic liquid to water in the same
loop and operating under the same temperature difference.
12. Consider a tall natural circulation loop shown in Fig. 7.46 consisting of two vertical pipes of circular cross
sections. The heating pipe has a diameter D, and that of the cooling side is 2D. The heat rate per unit length
coming in and going out are both q. Find the steady state velocity in the loop. Neglect the small horizontal
sections and state your other assumptions.
Figure 7.46: Tall natural circulation loop.

13. Set up a controller for PID control of the velocity x to a given value, xs , in the toroidal natural convection
loop equations
dx
=yx
(7.104)
dt
dy
= a sin xz
(7.105)
dt
dz
= b cos + xy
(7.106)
dt
where a and b are held constant. Use the tilt angle as control input, and show numerical results.
Chapter 8
Moving boundary
8.1
Stefan problems
[?]
The two phases, indicated by subscripts 1 and 2, are separated by an interface at x = X(t). In
each phase, the conduction equations is
2 T1
1 T1
=0
2
x
1 t
1 T2
2 T2
=0
x2
2 t
(8.1)
(8.2)
At the interface the temperature should be continuous, so that

T1 (X, t) = T2 (X, t)
(8.3)
Furthermore the difference in heat rate into the interface provides the energy required for phase
change. Thus
k1
8.1.1
T1
T2
dX
k2
= L
x
x
dt
(8.4)
Neumanns solution
The material is initially liquid at T = T0 . The temperature at the x = 0 end is reduced to zero for
t > 0. Thus
T1 = 0
at
x=0
(8.5)
T2
T0
Assume T1 (x, t) to be
x
T1 = A erf
2 1 t
167
as
(8.6)
8.1. Stefan problems
168
so that it satisfies equations (8.1) and (8.5). Similarly

x
T1 = A erf
2 1 t
x
T2 = T0 B erf
2 2 t
satisfies equation (8.2) and (8.6). The, condition (8.3) requires that
x
x
A erf
= T0 B erf
= T1
2 1 t
2 2 t
This shows that
X = 2 1 t
where is a constant. Using the remaining condition (8.4), we get
r
1 1 2 /2
2
k1 Ae
k2 B
= L1
e
2
The temperatures are
2
2
k2 21 (T0 T1 )e2 /2
L
e
p
=
erf
c1 T1 s
k1 2 T1 erfc( 1 /2 )
x
T1
)
erf (
erf
2 1 t
T0 T1
x
p
T2 = T0
)
erfc(
2 2 t
erfc( 1 /2 )
T1 =
8.1.2
Goodmans integral
Problems
Part IV
Multiple spatial dimensions
169
Chapter 9
Conduction
9.1
Steady-state problems
See [?].
9.2
Transient problems
9.2.1
Two-dimensional fin

cdx dyL
T
= Lk
t
2T
2T
+
2
x
y 2
hdx dy(T T )
which simplifies to
2T
2T
1 T
=
+
m2 (T T )
t
x2
y 2
where m2 = h/kL, the Biot number.
In the steady state
2T
2T
+
m2 (T T )
x2
y 2
Consider a square of unit side with = T T being zero all around, except for one edge
where it is unity.
Let
(x, y) = X(x)Y (y)
so that
1 d2 Y
1 d2 X
=
+ m2 = 2
X dx2
X dy 2
170
9.3. Radiating fins
171
x
y
Figure 9.1: Two-dimensional fin.

This leads to one equation
d2 X
+ 2 X = 0
dx2
with X(0) = X(1) = 1. Thus
X(x) = A sin x + B cos x
where due to the boundary conditions, B = 0 and = n, n = 1, 2, . . .. Another equation is
d2 Y
m2 + 2 Y = 0
2
dy
with
Y (y) = A sinh
m2 + n2 2 y + B cosh
m2 + n2 2 y
The condition Y (0) = 0 gives B = 0. Thus

p
X
(x, y) =
An sin nx sinh m2 + n2 2 y
9.3
Radiating fins
9.4
Non-Cartesian coordinates
Toroidal, bipolar.
Shape factor.
Moving boundary problem at a corner.
Problems
1. Show that the separation of variables solution for 2 T = 0 for a rectangle can also be obtained through an
eigenfunction expansion procedure.
2. Consider steady-state conduction in bipolar coordinates shown in
http://mathworld.wolfram.com/BipolarCylindricalCoordinates.html
with a = 1. The two cylindrical surfaces shown as v = 1 and v = 2 are kept at temperatures T1 and T2 ,
respectively. Sketch the geometry of the annular material between v = 1 and v = 2 and find the temperature
distribution in it by solving the Laplaces equation 2 T = 0.
9.4. Non-Cartesian coordinates
172
3. Set up and solve a conduction problem similar to Problem 2, but in parabolic cylindrical coordinates. Use
Morse and Feshbachs notation as shown in
http://www.math.sdu.edu.cn/mathency/math/p/p059.htm
4. Consider an unsteady one-dimensional fin of constant area with base temperature known and tip adiabatic.
Use the eigenfunction expansion method to reduce the governing equation to an infinite set of ODEs and solve.
5. Consider conduction in a square plate with Dirichlet boundary conditions. Find the appropriate eigenfunctions
for the Laplacian operator for this problem.
Chapter 10
Forced convection
See [?].
10.1
Low Reynolds numbers
10.2
Potential flow
[?]
If the heat flux is written as
q = cuT kT
the energy equation is
q=0
Heat flows along heatlines given by
dx
dy
dz
=
=
cT ux k(T /x)
cT uy k(T /y)
cT uz k(T /z)
The tangent to heatlines at every point is the direction of the heat flux vector.
10.2.1
Two-dimensional flow
10.3
Leveques solution
Leveque (1928)
10.4
Multiple solutions
See [?].
173
10.5. Plate heat exchangers
174
x flow
y flow
Figure 10.1: Schematic of crossflow plate HX.
10.5
Plate heat exchangers
There is flow on the two sides of a plate, 1 and 2, with an overall heat transfer coefficient of U .
Consider a rectangular plate of size Lx Ly in the x- and y-directions, respectively, as shown in Fig.
10.1. The flow on one side of the plate is in the x-direction with a temperature field Tx (x, y). The
mass flow rate of the flow is mx per unit transverse length. The flow in the other side of the plate
is in the y-direction with the corresponding quantities Ty (x, y) and my . The overall heat transfer
coefficient between the two fluids is U , which we will take to be a constant.
For the flow in the x-direction, the steady heat balance on an elemental rectangle of size dxdy
gives
Tx
dx = U dx dy (Ty Tx )
x
cx mx dy
where cx is the specific heat of that fluid. Simplifying, we get

2Cx R
Tx
= Ty Tx
x
(10.1)
where R = 1/2U is proportional to the thermal resistance between the two fluids, and Cx = cx mx .
For the other fluid
2Cy R
Ty
= Tx Ty
y
(10.2)
These equations have to be solved with suitable boundary conditions to obtain the temperature
fields Tx (x, y) and Ty (x, y).
From equation (10.2), we get
Tx = Ty + Cy R
Ty
y
Substituting in equation (10.1), we have

2
1 Ty
Ty
1 Ty
+
+ 2R
=0
Cy x
Cx y
xy
(10.3)
175
Nusselt (Jakob, 1957) gives an interesting solution in the following manner. Let the plate be
of dimensions L and W in the x- and y-directions. Nondimensional variables are
x
L
y
W
Tx Ty,i
Tx,i Ty,i
Ty Ty,i
Tx,i Ty,i
UWL
Cx
UWL
Cy
=
=
x =
y =
a=
b=
The governing equations are then
y
b(x y ) =
a(x y ) =
(10.4)
(10.5)

x = 1
y = 0
at = 0
at = 0
(10.6)
(10.7)
Equation (10.5) can be written as

y
+ by = bx
Solving for y we get
Z
y = eb C() + b
x (, )eb d
From the boundary condition (10.7), we get

C=0
so that
y (, ) = beb
x (, )eb d
Using the same procedure, from equation (10.4) we get

x (, ) = ea + aea
y ( , )ea d
176
Substituting for y , we find the Volterra integral equation

Z Z
a
(a+b)
x ( , )ea +b d d
x (, ) = e
+ abe
0
for the unknown x .

We will first solve the Volterra equation for an arbitrary , where
Z Z
a
(a+b)
x ( , )ea +b d d
x (, ) = e
+ abe
0
Let us express the solution in terms of a finite power series

x (, ) = 0 (, ) + 1 (, ) + 2 2 (, ) + . . . + n n (, )
(10.8)
This can be substituted in the integral equation. Since is arbitrary, the coefficient of each order
of must vanish. Thus
0 (, ) = ea
1 (, ) = abe(a+b)
2 (, ) = abe(a+b)
(a+b)
n (, ) = abe
0 ( , )ea +b d d
1 ( , )ea +b d d
..
.
n1 ( , )ea +b d d
The solutions are

0 = ea
1 = aea (1 eb )
1
2 = a2 2 ea (1 eb beb )
2
1
1
3 =
a3 3 ea (1 eb beb b2 2 eb )
23
2
..
.
1
1
bn1 n1 eb )
n = an n ea (1 eb beb . . .
n!
(n 1)!
Substituting into the expansion, equation (10.8), and taking = 1, we get
x (, ) = 0 (, ) + 1 (, ) + 2 (, ) + . . . + n (, )
where the s are given above.
Example 10.1
Find a solution of the same problem by separation of variables.
10.6. Falkner-Skan boundary flows
177
Taking
Ty (x, y) = X(x)Y (y)
Substituting and dividing by XY , we get
1 Tx
1 dY
+
+ 2R = 0
Cx x
Cy dy
Since the first term is a function only of x, and the second only of y, each must be a constant. Thus we can
write
dX
1
+
X=0
dx
cx mx (a + R)
dY
1
+
Y =0
dy
cy my (a R)
where a is a constant. Solving the two equations and taking their product, we have
x
y
c
exp
+
Ty =
a+R
cx mx (a + R)
cy my (a R)
where c is a constant. Substituting in equation (10.3), we get
c
x
y
Tx =
exp
+
aR
cx mx (a + R)
cy my (a R)
The rate of heat transfer over the entire plate, Q, is given by
Z Ly
Q = Cx
[Tx (Lx , y) Tx (0, y)] dy
0
Ly
lx
2
= cCx Cy exp
Cx (a + R)
C2 (a R)
The heat rate can be maximized by varying either of the variables Cx or Cy .
10.6
Falkner-Skan boundary flows
Problems
1. This is a problem
Chapter 11
Natural convection
11.1
Governing equations
11.2
Cavities
11.3
Marangoni convection
See [?].
Problems
1. This is a problem.
178
Chapter 12
Porous media
[?, ?, ?]
12.1
Governing equations
The continuity equation for incompressible flow in a porous medium is

V =0
12.1.1
(12.1)
Darcys equation
For the momentum equation, the simplest model is that due to Darcy
p =
V + f
K
(12.2)
where f is the body force per unit mass. Here K is called the permeability of the medium and
has units of inverse area. It is similar to the incompressible Navier-Stokes equation with constant
properties where the inertia terms are dropped and the viscous force per unit volume is represented
by (/K)V. Sometimes a term c0 V/t is added to the left side for transient problems, but it
is normally left out because it is very small. The condition on the velocity is that of zero normal
velocity at a boundary, allowing for slip in the tangential direction.
From equations (12.1) and (12.2), for f = 0 we get
2 p = 0
from which the pressure distribution can be determined.
12.1.2
Forchheimers equation
Forchheimers equation which is often used instead of Darcys equation is

p =
V cf K 1/2 |V|V + f
K
where cf is a dimensionless constant. There is still slip at a boundary.
179
12.2. Forced convection
12.1.3
180
Brinkmans equation
Another alternative is Brinkmans equation

p =
V+
2 V + f
K
where
is another viscous coefficient. In this model there is no slip at a solid boundary.
12.1.4
Energy equation
The energy equation is

(c)m
T
+ cp V T = km 2 T
t
where km is the effective thermal conductivity, and

(c)m = cp + (1 )(c)m
is the average heat capacity. The subscripts m refers to the solid matrix, and is the porosity of
the material. An equivalent form is
T
+ V T = m 2 T
t
where
km
cp
(c)m
=
cp
m =
See [1].
12.2
Forced convection
12.2.1
Plane wall at constant temperature
The solution to
u v
+
=0
x y
K
K
v=
u=
is
u=U
v=0
p
x
p
y
181
For
Ux
= P ex 1
m
the energy equation is
u
T
2T
T
+v
= m 2
x
y
y
or
U
T
2T
= m 2
x
y

T (0) = Tw
T () = T
Writing
r
U
m x
T Tw
() =
T Tw
=y
we get
d
T
= (T Tw )
x
d x
r
!
U 1 3/2
d
= (T Tw )
y
x
d
m 2
d
T
= (T Tw )
y
d y
r
d
U
= (T Tw )
d m x
d2 U
2T
= (T Tw ) 2
2
y
d m x
so that the equation becomes
1
+ = 0
2
with
(0) = 0
() = 1
182
We multiply by the integrating factor e
The first integral is
/4
to get
d 2 /4
e
=0
d
= C1 e
/4
Integrating again we have

= C1
/4
d + C2
With the change in variables x = /2, the solution becomes

Z /2
2
= 2C1
ex dx
0
Applying the boundary conditions, we find that C1 = 1/ and C2 = 0. Thus

Z /2
2
2
ex dx
=
0
= erf
2
The heat transfer coefficient is defined as

q
Tw T
km
T
=
Tw T y
= km
y
h=
The local Nusselt number is given by

N ux =
hx
km

=x
y
y=0
r
Ux
1
=
m
1
= P e1/2
x
Example 12.1
Find the temperature distribution for flow in a porous medium parallel to a flat plate with uniform heat
flux.
12.2.2
183
Stagnation-point flow
For flow in a porous medium normal to an infinite flat plate, the velocity field is
u = Cx
v = Cy
Cx
12.2.3
T
T
2T
Cy
= m 2
x
y
y
Thermal wakes
Line source
For P ex 1, the governing equation is
U
2T
T
= m 2
x
y
where the boundary conditions are

T
= 0 at y = 0
y
Z
(T T ) dy
q = (cp )U
(12.3)
(12.4)
Writing
r
U
m x
r
T T U x
() =
q /km
m
=y
we find that
r
r
r
T
m
U
m
1 3/2
d
q
1 3/2 q
x
+
x
=
y
x
km
U
2
d
m
2
km U x
r
r
T
m d
U
q
=
y
km U x d m x
r
m d U
q
2T
=
2
y
km U x d m x
1
= ( + )
2
(12.5)
12.3. Natural convection
184
The conditions (12.3)-(12.4) become
= 0 at = 0
dy = 1
The equation (12.5) can be written as

=
1 d
()
2 d
which integrates to
1
= + C1
2
Since = 0 at = 0, we find that C1 = 0. Integrating again, we get
= C2 e
Substituting in the other boundary condition
Z
Z
d = C2
1=
/4
/4
d = 2 C2
from which
1
C2 =
2
Thus the solution is
2
1
= e /4
2
or
U y 2
1 q p m
) exp(
)
T T =
(
Ux
4m x
2 km
Example 12.2
Show that for a point source
T T =
q
U r 2
exp(
)
4kx
4m x
185
B
y
gravity
x
Figure 12.1: Stability of horizontal porous layer.
12.3
Natural convection
12.3.1
Linear stability
This is often called the Horton-Rogers-Lapwood problem, and consists of finding the stability of a
horizontal layer of fluid in a porous medium heated from below. The geometry is shown in Fig. 12.1.
V =0
p V + g = 0
K
T
+ V T = m 2 T
t
= 0 [1 (T T0 )]
where g = gj. The basic steady solution is
V=0
y
T = T0 + T 1
H
1
y
p = p0 0 g y + T
2y
2
H
For constant heat flux T = qs /km . We apply a perturbation to each variable as

V = V + V
T = T + T
p = p + p
Substituting and linearizing
V = 0

V 0 T g = 0
K
T
T
w = m 2 T
t
H
186
Using the nondimensional variables

x
H
m t
H 2
HV
m
T
T
Kp
m
x =
t =
V =
T =
p =
the equations become, on dropping *s
V =0
p V + Ra T k = 0
T
w = 2 T
t
where
Ra =
0 gKHT
m
From these equations we get

2 w = Ra2H T
where
H =
2
x2
Using separation of variables

w(x, y, z, t) = W (y) exp (st + ikx x)
T (x, y, z, t) = (y) exp (st + ikx x)
Substituting into the equations we get
2
d
2
k s = W
dy 2
2
d
2
k W = k 2 Ra
dy 2
where
k 2 = kx2 + ky2
187
Isothermal boundary conditions

The boundary conditions are W = = 0 at either wall. For the solutions to remain bounded as
x, y , the wavenumbers kx and ky must be real. Furthermore, since the eigenvalue problem is
self-adjoint, as shown below, it can be shown that s is also real.
For a self-adjoint operator L, we must have
(u, Lv) = (Lu, v)
If
vL(u) uL(v) =
then
[vL(u) uL(v)] dV =
=
=
P
Q
+
x
y
ZV
P
Q
+
x
y
dV
(P i + Qj) dV
n (P i + Qj)
If n(P i+Qj) = 0 at the boundaries (i.e. impermeable), which is the case here, then L is self-adjoint.
Thus, marginal stability occurs when s = 0, for which
2
d
2
k = W
dy 2
2
d
2
W = k 2 Ra
k
dy 2
from which
d2
k2
dy 2
W = k 2 Ra W
The eigenfunctions are

W = sin ny
where n = 1, 2, 3, . . ., as long as
Ra =
n2 2
+k
k
For each n there is a minimum value of the critical Rayleigh number determined by
2 2
2 2
n
n
dRa
=2
+k 2 +1
dk
k
k
The lowest critical Ra is with k = and n = 1, which gives
Rac = 4 2
for the onset of instability.
188
y
x
Figure 12.2: Inclined porous layer.

Constant heat flux conditions
Here W = d/dy = 0 at the walls. We write
W = W0 + 2 W1 + . . .
= 0 + 2 1 + . . .
Ra = Ra0 + 2 Ra1 + . . .
For the zeroth order system
d 2 W0
=0
dy 2
with W0 = d0 /dy = 0 at the walls. The solutions is W0 = 0, 0 = 1. To the next order
d2 W1
= W0 Ra0 0
dy 2
d2 1
+ W1 = 0
dy 2
with W1 = d1 /dy = 0 at the walls. Finally, we get Rac = 12.
12.3.2
Steady-state inclined layer solutions
Consider an inclined porous layer of thickness H at angle with respect to the horizontal shown in
Fig. 12.2
Introducing the streamfunction (x, y), where
v=
x
u=
189
Darcys equation becomes

p

= 0 g [1 (T T0 )] sin
x K y

p
+
= 0 g [1 (T T0 )] cos
y K x
Taking /y of the first and /x of the second and subtracting, we have
2 2
0 gK T
T
+
=
cos
sin
x2
y 2
x
y
T
T
y x
x y
= m
2T
2T
+
x2
y 2
Side-wall heating
The non-dimensional equations are
2 2
T
T
+
=
Ra
cos
sin
x2
y 2
x
y
2
2
T
T
T
T
=
+
y x
x y
x2
y 2
where the Rayleigh number is
Ra =?
T
= 0at
x
T
= 1at
= 0,
y
= 0,
A
2
1
y=
2
x=
With a parallel-flow approximation, we assume [?]

= (y)
T = Cx + (y)
The governing equations become
2
d
Ra sin
+ RC cos = 0
2
y
dy
d
d2
C
=0
2
dy
dy
190
An additional constraint is the heat transported across a transversal section should be zero. Thus
Z 1/2
T
uT
dy = 0
(12.6)
x
1/2
Let us look at three cases.
(a) Horizontal layer
For = 0 the temperature and streamfunction are
RaC 2 2
T = Cx y 1 +
4y 3
24
RaC 2
=
4y 1
8
Substituting in condition (12.6), we get
C 10R Ra2 C 2 120 = 0

the solutions of which are
C=0
1 p
10(Ra 12)
Ra
1 p
C=
10(Ra 12)
Ra
C=
The only real solution that exists for Ra 12 is the conductive solution C = 0. For C > 12, there
are two nonzero values of C which lead to convective solutions, for which
RaC
8
12
Nu =
12 RaC 2
c =
For = 180 , the only real value of C is zero, so that only the conductive solution exists.
(b) Natural circulation
Let us take C sin > 0, for which we get
B
1 cosh
c =
C
2
Nu =
2B sinh 2 + C cot
where
2 = RC sin
1 + C cot
B=
cosh 2
and the constant C is determined from
B 2 sinh
=0
1 B cot cosh sinh
C
2C
191
(c) Antinatural circulation

For C sin < 0, for which we get
1 cosh
2
Nu =
2B sinh 2 + C cot
B
c =
C
where
2 = RC sin
1 + C cot
B=
cosh 2
B 2 sin
=0
C
1 B cot cosh sinh
2C
2
End-wall heating
Darcys law is
2
=R
T
T
sin +
cos
x
y

T
= 1at
x
T
= 0at
= 0,
y
A
2
1
x=
2
= 0,
x=
With a parallel-flow approximation, we assume

= (y)
T = Cx + (y)
The governing equations become
d
d
C
=0
2
dy
dy
2
d
R cos
RC sin = 0
2
y
dy
An additional constraint is the heat transported across a transversal section. Thus
Z 1/2
T
dy = 1
uT
x
1/2
Let us look at three cases.
(12.7)
192
(a) Vertical layer

For = 0 the temperature and streamfunction are
B1
B2
sin(y)
cos(y)
B1
B2
=
cos(y) +
sin(y) + B3
C
C
T = Cx +
where
2 = RC
Substituting in condition (12.6), we get
C 10R R2 C 2 120 = 0
the solutions of which are
C=0
1p
C=
10(R 12)
R
1p
C=
10(R 12)
R
The only real solution that exists for R 12 is the conductive solution C = 0. For C > 12, there
are two nonzero values of C which lead to convective solutions, for which
RC
8
12
Nu =
12 RC
c =
For = 180 , the only real value of C is zero, so that only the conductive solution exists.
(b) Natural circulation
Let us take C sin > 0, for which we get
B
1 cosh
C
2
Nu =
2B sinh 2 + C cot
c =
where
2 = RC sin
1 + C cot
B=
cosh 2
B 2 sinh
=0
1 B cot cosh sinh
C
2C
193
(c) Antinatural circulation

For C sin > 0, for which we get
1 cosh
2
Nu =
2B sinh 2 + C cot
B
c =
C
where
2 = RC sin
1 + C cot
B=
cosh 2
B 2 sin
=0
1 B cot cosh sinh
C
2C
Problems
1. This is a problem.
Chapter 13
Moving boundary
13.1
Stefan problems
Problems
194
Part V
Complex systems
195
Chapter 14
Radiation
14.1
Monte Carlo methods
[?]
[?] is a review of the radiative properties of semiconductors.
14.2
Equation of radiative transfer

1
I + (k,s + k,a )I = j + 1 k,s
I +
c t
4c
I d
Problems
1. Consider an unsteady n-body radiative problem. The temperature of the ith body is given by
Mj cj
n
X
dTj
Ai Fij (Ti4 Tj4 ) + Qj
=
dt
i=1
= Aj
n
X
i=1
Fji (Ti4 Tj4 ) + Qj
What kind of dynamic solutions are possible?

2. The steady-state temperature distribution in a one-dimensional radiative fin is given by
dT
+ hT 4 = 0
dx
Is the solution unique and always possible?
3. Show that between one small body 1 and its large surroundings 2, the dynamics of the small-body temperature
is governed by
dT1
M1 c1
= A1 F12 (T14 T24 ) + Q1 .
dt
196
Chapter 15
Boiling and condensation

15.1
Homogeneous nucleation
Problems
197
Chapter 16
Bioheat transfer
16.1
Mathematical models
Good references are [?, ?].
Problems
198
Chapter 17
Heat exchangers
17.1
Fin analysis
Analysis of Kraus (1990) for variable heat transfer coefficients.
17.2
Porous medium analogy
See Nield and Bejan, p. 87 [?].
17.3
Heat transfer augmentation
17.4
Maldistribution effects
Rohsenow (1981)
17.5
Microchannel heat exchangers
Phillips (1990).
17.6
Radiation effects
See Ozisik (1981).

Shah (1981)
17.7
Transient behavior
Ontko and Harris (1990)

For both fluids mixed
dT
+m
1 c1 (T1in T1out ) + mc
2 (T2in T2out ) = 0
dt
For one fluid mixed and the other unmixed, we have
Mc
Ac2
T2
T2
+ V2 Ac2
+ hP (T T1 ) = 0
t
x
199
17.8. Correlations
200
17.8
Correlations
17.8.1
Least squares method
Possible correlations are

y(x) =
n
X
ak xk
k=0
y(x) = a0 + a1/2 x1/2 + a1 x + a3/2 x3/2 + a2 x2 + . . .

y(x) = axm + bxn + . . .
Z n
a(k)xk dk
y(x) =
k=0
where 1 < x < 1
A power-law correlation of the form
y = cxn
satisfies the invariance condition given by equation (??).
17.9
Compressible flow
17.10
Thermal control
A cross-flow heat exchanger model, schematically shown in Fig. A.38, has been studied using finite
differences [?, ?, ?]. Water is the in-tube and air the over-tube fluid in the heat exchanger. This
example includes all the conductive, advective and convective effects discussed before. The governing
equations on the outside of the tube, in the water, and in the wall of the tube are
m
a
a
a
ca (Tin
Tout
) = ho 2ro (Ta Tt ),
(17.1)
L
Tw
Tw
w cw ri2
+m
w cw
= hi 2ri (Tt Tw ),
(17.2)
t
2 Tt
Tt
= kt (ro2 ri2 ) 2 + 2ro ho (Ta Tt ) 2ri hi (Tt Tw ),
(17.3)
t ct (ro2 ri2 )
t
respectively. L is the length of the tube; m

a (t) and m
w (t) are the mass flow rates of air and water;
a
a
Tin
and Tout
(t) are the inlet and outlet air temperatures; Ta (t) is the air temperature surrounding
the tube; Tt (, t) and Tw (, t) are the tube-wall and water temperatures; hi and ho the heat transfer
coefficients in the inner and outer surfaces of the tube; ri and ro are the inner and outer radii of the
tube; ca , cw and ct are the specific heats of the air, water and tube material; w and t are the water
and tube material densities; and kt is the thermal conductivity of the tube material. In addition,
a
a
the air temperature is assumed to be Ta = (Tin
+ Tout
)/2. The boundary and initial conditions are
w
Tt (0, t) = Tw (0, t) = Tin , Tt (L, t) = Tw (L.t), and Tt (, 0) = Tw (, 0). Suitable numerical values were
assumed for the computations.
a
w
The inlet temperatures Tin
and Tin
, and the flow rates m
a and m
w can all be used as control
w
a
. The flow rates present a special difficulty;
or Tout
inputs to obtain a desired outlet temperature, Tout
they appear in nonlinear form in Eqs. (17.1) and (17.2), and the outlet temperature is bounded. Fig.
w
that can be achieved on varying m
w ; temperatures
17.2 shows the steady-state range of values of Tout
outside this range cannot be obtained. It is also seen that the outlet water temperature is hard to
control for large water flow rates. As an example of control dynamics, Fig. 17.3 shows the results of
w
applying PI control on m
w to obtain a given reference temperature Tout
= 23 C.
201
111111111111111111111
000000000000000000000
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
111111111111111111111
000000000000000000000
000000000000000000000
111111111111111111111
Figure 17.1: Schematic of single-tube cross-flow heat exchanger.

23.6
23.5
a [ C]
Tout
23.4
23.3
.
ma =
0.23 kg/s
23.2
23.1
0.17 kg/s
23
22.9
0.1
0.2
0.3
0.4
0.5
0.6
0.7
.
ma
0.8
0.9
[kg/s]
a
Figure 17.2: The relation between Tout
and m
w for different m
a [?].
25
a [ C]
Tout
24.5
24
Kp= 100,000
23.5
10,000
23
1,000
22.5
100
22
21.5
200
400
600
800
1000
1200
1400
1600
1800
2000
t [s]
a
Figure 17.3: Behavior of Tout
as a function of time for Ki = 50 and different Kp [?].
17.11. Control of complex thermal systems
17.11
202
Control of complex thermal systems
The previous sections examined systems that were fairly simple in the sense that mathematical
models could be written down and their behaviors studied. For most practical thermal systems,
this is difficult to do with any degree of precision. In the following, we will look first at thermal
components and then combine them in networks.
17.11.1
Hydronic networks
The science of networks of all kinds has been put forward as a new emerging science [?]. In the
present context this means that a complete understanding of the behavior of components does not
necessarily mean that large networks formed out of these components can be modeled and computed
in real time for control purposes. Controllability issues of heat exchanger networks are reported in [?].
Mathematical models of the dynamics of a piping network lead to differential-algebraic systems [?].
The momentum equation governing the flow in each pipe is differential, while the conservation of
mass condition at each of the junctions is algebraic. Thus, it turns out that only certain flow rates
may be controllable, the others being dependent on these.
There are at present many different strategies for the thermal control of networks, and comparative studies based on mathematical models can be carried out. Fig. 17.4 shows a network in
which three specific control strategies can be compared [?, ?]; each control method works differently
and are labeled VF, MCF and BT in the following figures (details are in [?]). The network has a
primary loop, a secondary loop and a bypass that has the three strategies as special cases. The
primary loop includes a chiller, while the secondary has a water-air cooling coil which serves as a
thermal load. Integral controllers are used to operate the valves V , V , and V to control the air
temperature leaving the cooling coil, TaL (t). Figs. 17.5 and 17.6 show the dynamic response of TaL (t),
the leaving water temperatures TwL (t), and the bypass pressure difference pbp to step changes in
the air velocity over the coil. (t), (t), and (t) are the respective closing fractions of the valves
which change dynamically in response to the error signal. There are some oscillations in all the
variables before they settle down to stable, steady values.
Laboratory experiments with a network of water-to-water heat exchangers have been reported
in [?, ?, ?]; the configuration is shown in Fig. 17.7. The hot water flow is diminished by changing its
controller set point. Figure 17.8 shows the secondary hot water flow rates q8 , q7 and q4 to the heat
exchangers for the three different control strategies. Each curve represents one independent run;
that is, water flow to HXBT and HXCF is zero when testing VF, and so on. The system is taken
to the nominal operating conditions, and then the hot water flow is decreased by a constant value
every 1800 s. The controls drive the system to different operational points while coping with the
changes. The input voltages v7 , v4 and v1 that control flow and the hot water temperature at the
heat exchanger inlet T21 are also shown. It is seen that for certain control parameters, the system
is becomes unstable and the variables oscillate in time.
17.11.2
Other applications
There are a large number of other thermal problems in which control theory has been applied. Agentbased controls have been proposed by complex thermal systems such as in buildings [?], microwave
heating [?], thermal radiation [?], and materials processing and manufacturing [?, ?]. Control of
convection is an important and active topic; this includes the study of convection loops [?, ?, ?, ?, ?],
stabilization and control of convection in horizontal fluid layers [?,?,?,?,?,?,?,?,?,?], and in porous
media [?].
203
1B
B
c
P2
P3
a
P1
T aE
ch
f
T aL
CC
T wE
T wL
Figure 17.4: Network used to study control strategies [?].
T aL ( oC)
27
26
25
75
100
125
150
175
200
225
250
275
300
200
225
250
275
300
t (s )
1
0.5
0
75
100
125
150
175
t (s )
Figure 17.5: Dynamic response of control system to drop in air velocity, method VF, and
L
method MCF; Ta,set
= 26 C, TaE = 30 C [?].
204
T wL ( oC)
7
6
5
100
125
150
175
200
225
250
275
300
200
225
250
275
300
200
225
250
275
300
200
225
250
275
300
200
225
250
275
300
0.5
p bp(N/m 2 )
0
3
x 10
100
125
150
175
t
2
1
0
100
125
150
175
0.5
T aL ( oC)
100
125
150
175
27
26
25
100
125
150
175
t (s )
L
=
Figure 17.6: Dynamic response of control system to drop in air velocity with method BT; Tw,set
2
E
5.5 C, pbp,set = 20000 (N/m ); Ta = 30 C [?].
PID
P1
T21
p
V10
10
T14
T8
HX
T7
q8
q7
q4
HX
BT
T 12
T5
T6
HX
CF
T12
T11
q3
P2
Control Valve
T16
VF
T15
q6
M4
M3
T 17
Pump
T18
HT 2
HT 1
Manual Valve
P3
M6
2
pV
M5
T1
13
q1
P4
m
T4
q5
q9
T3
T9
T 19
T10
T20
q1
12
M1
T2
M2
Figure 17.7: Layout of hydronic network [?].
11
HX
pP
q2
pV
17.12. Conclusions
205
q8 ,q7 ,q 4
(m3/s)
x 10
3
2
1
0
1000
2000
3000
4000
5000
6000
1000
2000
3000
4000
5000
6000
1000
2000
3000
4000
5000
6000
v1 ,v4 ,v 7
(Volts)
3
2
1
0
T21( oC)
40
35
30
t (s)
Figure 17.8: Secondary hot water flows and T21 : BT , CF , V F . The dashed vertical
lines are instants at which the thermal load is changed [?].
17.12
Conclusions
This has been a very brief introduction to the theory of thermal control. The fundamental ideas in
this subject are firmly grounded on the mathematics of systems and control theory which should be
the starting point. There are, however, a few aspects that are particularly characteristic of thermal
systems. Phenomena such as diffusion, convection and advection are common and the systems
are usually complex, nonlinear and poorly predictable dynamically. The governing equations cover
a wide range of possibilities, from ordinary and partial differential equations to functional and
differential-algebraic systems. Furthermore, control theory itself is a vast subject, with specialized
branches like optimal [?], robust [?], and stochastic control [?] that are well developed. Many of the
tools in these areas find applications in thermal systems.
The study of thermal control will continue to grow from the point of view of fundamentals as
well as engineering applications. There are many outstanding problems and issues that need to be
addressed. To cite one specific example, networking between a large number of coupled components
will become increasingly important; it is known that unexpected synchronization may result even
when multiple dynamical systems are coupled weakly [?]. It is hoped that the reader will use
this brief overview as a starting point for further study and apply control theory in other thermal
applications.
Problems
Chapter 18
Soft computing
18.1
Genetic algorithms
[?]
18.2
Artificial neural networks
18.2.1
Heat exchangers
The most important of the components are heat exchangers, which are generally very complex in that
they cannot be realistically computed in real time for control purposes [?, ?, ?]. An approach that
is becoming popular in these cases is that of artificial neural networks (ANN) [?] for prediction of
system behavior both for time-independent [?,?,?,?] and time-dependent operation. It is particularly
suitable for systems for which experimental information that can be used for training is available.
Reviews of artificial neural network applications to thermal engineering [?, ?] and other soft control
methodologies [?, ?, ?] and applications [?] are available.
A stabilized neurocontrol technique for heat exchangers has been described in [?, ?, ?, ?, ?].
Fig. 18.1 shows the test facility in which the experiments were conducted. The objective is to
a
control the outlet air temperature Tout
. Figs. 18.2 shows the results of using neurocontrol compared
with PID; both are effective. Fig. 18.3 shows the result of an disturbance rejection experiment.
a
The heat exchanger is stabilized at Tout
= 36 C, and then the water flow is shut down between
t = 40s and t = 70s; after that the neurocontroller brings the system back to normal operation.
A neural network-based controller is able to adapt easily to changing circumstances; in thermal
systems this may come from effects such as the presence of fouling over time or from changes in
system configuration as could happen in building heating and cooling systems.
[?, ?]
[?]
Problems
206
18.2. Artificial neural networks
207
a
Tout
Air flow
Water flow
(a)
(b)
Figure 18.1: Experimental setup: (a) heat exchanger test facility with wind tunnel and in-draft fan,
a
is the air outlet temperature [?].
(b) heat exchanger with water and air flows indicated; Tout
36.4
ANN
PID
a
Tout [ C]
36.2
36
35.8
35.6
PI
35.4
100 120 140 160 180 200 220 240 260 280 300
t [s]
36.5
36
35.5
35
a
Tout [ C]
34.5
ANN
PID
PI
34
33.5
33
32.5
32
31.5
50
100
150
t [s]
200
250
300
350
Figure 18.2: Time-dependent behavior of heat exchanger using ANN, PID and PI control methodologies [?].
18.2. Artificial neural networks
208
mw [kg/s]
300
200
100
0
50
100
150
34.5
0
50
100
150
200
250
300
350
400
200
250
300
350
400
250
300
350
400
t [s]
a
Tout [ C]
36.5
36
35.5
35
t [s]
m a [kg/s]
0.5
0.4
0.3
0.2
0.1
50
100
150
200
t [s]
Figure 18.3: Time-dependent behavior of heat exchanger with neurocontrol for disturbance rejection
experiment showing flow rates and air outlet temperature [?].
Part VI
Appendices
209
Appendix A
Additional problems
1. Plot all real (, ) surfaces for the convection with radiation problem, and comment on the
existence of solutions.
2. Complete the problem of radiation in an enclosure (linear stability, numerical solutions).
3. Two bodies at temperatures T1 (t) and T2 (t), respectively, are in thermal contact with each
other and with the environment. The temperatures are governed by
dT1
+ hAc (T1 T2 ) + hA(T1 T ) = 0
dt
dT2
+ hAc (T2 T1 ) + hA(T2 T ) = 0
M2 c2
dt
M1 c1
Derive the equations above and explain the parameters. Find the steady-state temperatures
and explore the stability of the system for constant T .
4. Consider the change in temperature of a lumped system with convective heat transfer where
the ambient temperature, T (t), varies with time in the form shown. Find (a) the long-time
solution of the system temperature, T (t), and (b) the amplitude of oscillation of the system
temperature, T (t), for a small period t.
T
Tmax
Tmin
Figure A.1: Ambient temperature variation.

210
211
(x)
w
b
Figure A.2: Longitudinal fin of concave parabolic profile.

dT1
+ hc Ac (T1 T2 ) + hA(T1 T (t)) = Q(t)
dt
dT2
+ hc Ac (T2 T1 ) + hA(T2 T (t)) = 0,
M2 c2
dt
M1 c1
where Q( t) is internal heat generation that can be controlled. Take T = constant. Show
results for (a) PID and (b) on-off control.
dT1
+ hc Ac (T1 T2 ) + hA(T1 T (t)) = Q1 (t)
dt
dT2
M2 c2
+ hc Ac (T2 T1 ) + hA(T2 T (t)) = Q2 (t)
dt
M1 c1
where Q1 and Q2 are internal heat generation sources, each one of which can be independently
controlled. Using on-off control, show analytical or numerical results for the temperature
responses of the two bodies. If you do the problem analytically, take the ambient temperature,
T , to be constant, but if you do it numerically, then you can take it to be (a) constant, and
(b) oscillatory. For Q1 = Q2 = 0, find the steady state (T 1 , T 2 ) and determine its stability.
7. Consider a longitudinal fin of concave parabolic profile as shown in the figure, where =
[1 (x/L)]2 b . b is the thickness of the fin at the base. Assume that the base temperature is
known. Neglect convection from the thin sides. Find (a) the temperature distribution in the
fin, and (b) the heat flow at the base of the fin. Optimize the fin assuming the fin volume to
be constant and maximizing the heat rate at the base. Find (c) the optimum base thickness
b , and (d) the optimum fin height L.
8. Analyze an annular fin with a prescribed base temperature and adiabatic tip. Determine its
fin efficiency and plot.
212
9. Consider a square plate of side 1 m. The temperatures on each side are (a) 10 C, (b) 10 C,
(c) 10 C, and (d) 10 + sin(x) C, where x is the coordinate along the edge. Find the steadystate temperature distribution analytically. Write a computer program to do the problem
numerically using finite differences and compare with the analytical results. Choose different
grid sizes and show convergence.
10. A plane wall initially at a uniform temperature is suddenly immersed in a fluid at a different
temperature. Find the temperature profile as a function of time. Assume all parameter values
to be unity. Write a computer program to do the problem numerically using finite differences
and compare with the analytical results.
11. Consider the hydrodynamic and thermal boundary layers in a flow over a flat plate at constant
temperature. Starting from the boundary layer equations
u v
+
=0
x y
u
2u
u
+v
= 2
u
x
y
y
T
T
2T
u
+v
= 2
x
y
y
(A.1)
(A.2)
(A.3)
change to variables f () and () and derive the boundary layer equations

2f + f f = 0
Pr
f = 0
+
2
(A.4)
(A.5)
and the boundary conditions. Solve equations (A.4) and (A.5) numerically by the shooting
method for different Pr and compare with results in the literature.
12. For Problem 1, derive the momentum and energy integral equations. Using cubic expansions
for u/u and , derive expressions for the boundary layer thicknesses. [Explain Momentum
integral method?]
13. For natural convection near a vertical plate, show that the governing boundary layer equations
u v
+
=0
x y
u
u
2u
u
+v
= g(T T ) + 2
x
y
y
2
T
T
T
+v
= 2
u
x
y
y
can be reduced to
f + 3f f 2(f )2 + = 0
+ 3 P r f = 0
(A.6)
(A.7)
with appropriate boundary conditions. Solve equations (A.6) and (A.7) numerically by the
shooting method for different P r and compare with results in the literature.
213
C
L
H
y
gravity
B
Figure A.3: Inclined rectangular cavity

14. Write the governing equations for natural convection flow in an inclined rectangular cavity, and
nondimensionalize them. The thermal conditions at the walls of the cavity are: (a) AB heating
with heat flux qs , (b) BC adiabatic, (c) CD cooling with heat flux qs , (d) DA adiabatic.
15. For parallel and counterflow heat exchangers, I obtained the temperature distributions

1
1
Th,1 Tc,1
1 exp{U P
x} ,
(A.8)
Th (x) = Th,1
1 (m
h ch /m
c cc )
m
h ch
m
c cc

Th,1 Tc,1
1
1
1 exp{U P
Tc (x) = Tc,1
x}
(A.9)
(m
c cc /m
h ch ) 1
m
h ch
m
c cc
for the hot and cold fluids, respectively. As usual the upper sign is for parallel and the lower
for counterflow; 1 is the end where the hot fluid enters (from where x is measured) and 2 is
where it leaves. Please check.
16. For a counterflow heat exchanger, derive the expression for the effectiveness as a function of
the NTU, and also the NTU as function of the effectiveness.
17. (From Incropera and DeWitt) A single-pass, cross-flow heat exchanger uses hot exhaust gases
(mixed) to heat water (unmixed) from 30 to 80 C at a rate of 3 kg/s. The exhaust gases,
having thermophysical properties similar to air, enter and exit the exchanger at 225 and 100 C,
respectively. If the overall heat transfer coefficient is 200 W/m2 K, estimate the required area.
18. (From Incropera and DeWitt) A cross-flow heat exchanger used in cardiopulmonary bypass
procedure cools blood flowing at 5 liters/min from a body temperature of 37 C to 25 C in
order to induce body hypothermia, which reduces metabolic and oxygen requirements. The
coolant is ice water at 0 C and its flow rate is adjusted to provide an outlet temperature of
15 C. The heat exchanger operates with both fluids unmixed, and the overall heat transfer
coefficient is 750 W/m2 K. The density and specific heat of the blood are 1050 kg/m2 and 3740
J/kg K, respectively. (a) Determine the heat transfer rate for the exchanger. (b) Calculate
the water flow rate. (c) What is the surface area of the heat exchanger?
19. Determine computationally the view factor between two rectangular surfaces (each of size
L 2L) at 90 with a common edge of length 2L; see Fig. A.4. Compare with the analytical
result.
214
e
i
j
c
l
h
Figure A.4: Two rectangular surfaces.
Figure A.5: Two rectangular surfaces with sphere. CAPTION HERE

20. Calculate the view factor again but with a sphere (diameter L/2, center at a distance of L/2
from each rectangle, and centered along the length of the rectangles) as an obstacle between
the two rectangles; see Fig. A.5.
21. (From Incropera and DeWitt) Consider a diffuse, gray, four-surface enclosure shaped in the
form of a tetrahedron (made of four equilateral triangles). The temperatures and emissivities
of three sides are
T1 = 700K,
1 = 0.7
T2 = 500K,
T3 = 300K,
2 = 0.5
3 = 0.3
The fourth side is well insulated and can be treated as a reradiating surface. Determine its
temperature.
22. An Aoki curve is defined as shown in Fig. A.6. Show that when n , the dimension of
the curve is D = 1 and the length L .
n=0
n=1
n=2
Figure A.6: Aoki curve.
215
T4
T3
T2
3
1
T0
T1
n=3
Figure A.7: Fractal tree.

23. Consider conductive rods of thermal conductivity k joined together in the form of a fractal
tree (generation n = 3 is shown in Fig. A.7; the fractal is obtained in the limit n ). The
base and tip temperatures are T0 and T , respectively. The length and cross-sectional area of
bar 0 is L and A, respectively, and those of bar 1 are 2L/ and A/ 2 , where 1 < 2, and
so on. Show that the conductive heat transfer through rod 0 is
kA
(T0 T ) 1
q=
L
2
24. The dependence of the rate of chemical reaction on the temperature T is often represented by
the Arrhenius function f (T ) = eE/T , where E is the activation energy. Writing T = T /E,
show that f (T ) has a point of inflexion at T = 1/2. Plot f (T ) in the range T = 1/16 to
T = 4 as well as its Taylor series approximation to various orders around T = 1/2. Plot also
the L2 -error in the same range for different orders of the approximation.
other and with the environment. Show that the temperatures are governed by
dT1
+ C(T1 T2 ) + C1, (T1 T ) = Q1
dt
dT2
+ C(T2 T1 ) + C2, (T2 T ) = Q2
M2 c2
dt
where Mi is the mass, ci is the specific heat, the Cs are thermal conductances, and Qi is
internal heat generation. Find the steady state (T 1 , T 2 ) and determine its stability.
M1 c1
26. Using a complete basis, expand the solution of the one-dimensional heat equation
T
2T
= 2
t
x
k
as an infinite set of ODEs.
T
= q0 at x = 0,
x
T = T1 at x = L
(A.10)
(A.11)
216
Tmax
Tmin
Figure A.8: Ambient temperature variation.

27. Show that the governing equation of the unsteady, variable-area, convective fin can be written
in the form
T
a(x)
+ b(x)T = 0
t
x
x
Show that the steady-state temperature distribution with fixed temperatures at the two ends
x = 0 and x = L is globally stable.
28. Nondimensionalize and solve the radiative cooling problem
Mc
dT
4
=0
+ A T 4 T
dt
with T (0) = Ti . [Is this done elsewhere?]
29. For heat transfer from a heated body with convection and weak radiation, i.e. for
o
n
d
4
+ + ( + ) 4 = q
d
with (0) = 1, using symbolic algebra determine the regular perturbation solution up to and
including terms of order 4 . Assuming = 0.1, = 1, q = 1, plot the five solutions (with one
term, with two terms, with three terms, etc.) in the range 0 1.
30. Consider a body in thermal contact with the environment
Mc
dT
+ hA(T T ) = 0
dt
where the ambient temperature, T (t), varies with time in the form shown below. Find (a)
the long-time solution of the system temperature, T (t), and (b) the amplitude of oscillation of
the system temperature, T (t), for a small period t.
31. For a heated body in thermal contact with a constant temperature environment
Mc
dT
+ hA(T T ) = Q
dt
analyze the conditions for linear stability of PID control.
217
32. Show numerical results for the behavior of two heated bodies in thermal contact with each
other and with a constant temperature environment for on-off control with (a) one thermostat,
and (b) two thermostats.
33. Analyze the system controllability of two heated bodies in thermal contact with each other and
with a constant temperature environment for (a) Q1 (t) and Q2 (t) being the two manipulated
variables, and (b) with Q1 (t) as the only manipulated variable and Q2 constant.
34. Run the neural network FORTRAN code in
http://www.nd.edu/~msen/Teaching/IntSyst/Programs/ANN/
for 2 hidden layers with 5 nodes each and 20,000 epochs. Plot the results in the form of exact
z vs. predicted z.
35. Consider the heat equation
T
2T
=
t
x2
with one boundary condition T (0) = 0. At the other end the temperature, T (1) = u(t) is
used as the manipulated variable. Divide the domain into 5 parts and use finite differences
to write the equation as a matrix ODE. Find the controllability matrix and check for system
controllability.
36. Determine the semi-derivative and semi-integral of (a) C (a constant), (b) x, and (c) x where
> 1.
37. Find the time-dependent temperature field for flow in a duct wih constant T and Tin , but with
variable flow rate V (t) = V0 + V sin(t) such that V is always positive. Write a computer
program to solve the PDE, and compare numerical and analytical results.
38. Find the steady-state temperature distribution and velocity in a square-loop thermosyphon.
The total length of the loop is L and the distribution of the heat rate per unit length is
for L/8 x L/4,

Q
Q for 5L/8 x 3L/4,
q(x) =
0
otherwise.
39. Show that the dynamical system governing the toroidal thermosyphon with known wall temperature can be reduced to the Lorenz equations.
40. Draw the steady-state velocity vs. inclination angle diagram for the inclined toroidal thermosyphon with mixed heating. Do two cases: (a) without axial condution1 , and (b) with axial
conduction.
41. Model natural convection in a long, vertical pipe that is being heated from the side at a
constant rate. What is the steady-state fluid velocity in the pipe? Assume one-dimensionality
and that the viscous force is proportional to the velocity.
1 M. Sen, E. Ramos and C. Trevi
no, On the steady-state velocity of the inclined toroidal thermosyphon, ASME J.
of Heat Transfer, Vol. 107, pp. 974977, 1985.
218
g
Q
Q
x
L/4
Figure A.9: Square naturcal circulation loop

42. The Brinkman model for the axial flow velocity, u (r ), in a porous cylinder of radius R is
1 du
d u
u + G = 0,
+
eff
2
dr
r dr
K
where u = 0 at r = R (no-slip at the wall), and u /r = 0 at r = 0 (symmetry at the
centerline). eff is the effective viscosity, is the fluid viscosity, K is the permeability, and G
is the applied pressure gradient. Show that this can be reduced to the nondimensional form
d2 u 1 du
1
+
s2 u +
= 0,
dr2
r dr
M
where M = eff /, Da = K/R2 , s2 = (M Da)1 .
43. Using a regular perturbation expansion, show that for s 1, the velocity profile from equation
(A.12) is
s2
1 r2
1
3 r2 + . . .
u=
4M
16
[What is this?]
44. Using the WKB method, show that the solution of equation (1) for s 1 is
exp {s(1 r)}
+ ...
u = Da 1
r
[What is this?]
45. Consider steady state natural convection in a tilted cavity as shown. DA and BC are adiabatic
while AB and CD have a constant heat flux per unit length. It can be shown that the governing
equations in terms of the vorticity and the streamfunction are
2 2
+
+ =0
x2
x2
T
T

2
Ra
P
r
Pr
+
cos
sin
=0
y x
x y
x2
y 2
x
y
T
2T
2T
T
=0
2
y x
x y
x
y 2
219
C
L
H
y
gravity
B
Figure A.10: Inclined natural cavity

where P r and Ra are the Prandtl and Rayleigh numbers, respectively. The boundary conditions are:
A
at x = ,
2
1
at y = ,
2
= 0,
x
=
= 0,
y
T
= 0,
x
T
= 1.
y
where A = L/H is the aspect ratio. Find a parallel flow solution for using
= (y)
T (x, y) = Cx + (y)
46. Obtain the response to on-off control of a lumped, convectively-cooled body with sinusoidal
variation in the ambient temperature.
47. Determine the steady-state temperature field in a slab of constant thermal conductivity in
which the heat generated is proportional to the exponential of the temperature such that
d2 T
= exp(T ),
dx2
where 0 x 1, with the boundary conditions T (0) = T (0) = 0.
48. In the previous problem, assume that is small. Find a perturbation solution and compare
with the analytical. Do up to O() by hand and write a Maple (or Mathematica) code to do
up to O(10 ).
49. The temperature equation for a fin of constant area and convection to the surroundings at a
constant heat transfer coefficient is
2
d
2
m = 0,
dx2
where = T T . Determine the eigenfunctions of the differential operator for each combination of Dirichlet and Neumann boundary conditions at the two ends x = 0 and x = L.
220
Figure A.11: Slightly tapered 2D fin.

50. Add radiation to a convective fin with constant area and solve for small radiative effects with
boundary conditions corresponding to a known base temperature and adiabatic tip.
51. Find the temperature distribution in a slightly tapered 2-dimensional convective fin with known
base temperature and adiabatic tip.
52. Prove Hottels crossed string method to find the view factor FAB between two-dimensional
surfaces A and B with some obstacles between them as shown. The dotted lines are tightly
stretched strings. The steps are:
(a) Assuming the strings to be imaginary surfaces, apply the summation rule to each one of
the sides of figure abc.
(b) Manipulating these equations and applying reciprocity, show that
Fabac =
Aab + Aac Abc

.
2Aab
(c) For abd find Fabbd in a similar way.

(d) Use the summation rule to show that
Fabcd =
Abc + Aad Aac Abd

2Aab
(e) Show that Fabcd = AA FAB /Aab .

(f) Show the final result
FAB =
Abc + Aad Aac Abd

2AA
53. Complete the details to derive the Nusselt result for laminar film condensation on a vertical
flat plate. Find from the literature if there is any experimental confirmation of the result.
54. Consider the hydrodynamic and thermal boundary layers in a flow over a flat plate at constant
temperature. Use a similarity transformation on the boundary layer equations to get
2f + f f = 0,
Pr
+
f = 0.
2
(A.12)
(A.13)
Using the shooting method and the appropriate boundary conditions, solve the equations for
different Pr and compare with the results in the literature.
221
11111111
00000000
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
11111111
00000000
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
d
cc
Figure A.12: Hottels string method.

55. Solve the steady state conduction equation 2 T = 0 in the area in the figure between the
square and the circle using the MATLAB Toolbox. Edges DA and BC have temperatures of
100 and 0 units, respectively, AB and CD are adiabatic and the circle is at a temperature of
200 units. Draw the isotherms.
56. (From Brauner and Shacham, 1995) Using Eq. 11, write a program to redraw Fig. 2 on a sunny
day (Cl = 0) and a cloudy day (Cl = 1). Assume Ta = 37 C. Use Eq. 8 to calculate hc . Note:
since the physical properties are to be taken at the mean temperature between Ts and Ta , Eq.
11 must be solved numerically.
57. The steady-state temperature distribution in a plane wall of thermal conductivity k and thickness L is given by T (x) = 4x3 + 5x2 + 2x + 3, where T is in K, x in m, and the constants in
appropriate units. (a) What is the heat generation rate per unit volume, q(x), in the wall?
(b) Determine the heat fluxes, qx , at the two faces x = 0 and x = L.
58. (From Incropera and DeWitt, 5th edition) Consider a square plate of side 1 m. Going around,
the temperatures on the sides are (a) 50 C, (b) 100 C, (c) 200 C, and (d) 300 C. Find the
steady-state temperature distribution analytically. Write a computer program to do the previous problem numerically using finite differences and compare with the analytical result. Choose
different grid sizes and show convergence of the heat flux at any wall. Plot the 75, 150, and
250 C isotherms.
59. A plane wall of thickness 1 m is initially at a uniform temperature of 85 C. Suddenly one side
of the wall is lowered to a temperature of 20 C, while the other side is perfectly insulated. Find
the time-dependent temperature profile T (x, t). Assume the thermal diffusivity to be 1 m2 /s.
Write a computer program to do the previous problem numerically using finite differences and
compare with the analytical result.
60. At a corner of a square where the temperature is discontinuous, show how the finite difference solution of the steady-state temperature behaves ompared to the separation-of-variables
solution.
61. Find the view factor of a semi-circular arc with respect to itself.
62. Derive the unsteady governing equation for a two-dimensional fin with convection and radiation.
222
T
i1
i
T
i+1
63. Determine the steady temperature distribution in a two-dimensional convecting fin.

64. A number of identical rooms are arranged in a circle as shown, with each at a uniform temperature Ti (t). Each room exchanges heat by convection with the outside which is at T , and
with its neighbors with a conductive thermal resistance R. To maintain temperatures, each
room has a heater that is controlled by independent but identical proportional controllers. (a)
Derive the governing equations for the system, and nondimensionalize. (b) Find the steady
state temperatures. (c) Write the dynamical system in the form x = Ax and determine the
condition for stability2 .
65. A sphere, initially at temperature Ti is being cooled by natural convection to fluid at T .
Churchills correlation for natural convection from a sphere is
1/4
where
0.589 RaD
Nu = 2 + h
i4/9 ,
9/16
1 + (0.469/Pr )
g(Ts T )D3
.
Assume that the temperature within the sphere T (t) is uniform, and that the material properties are all constant. Derive the governing equation, and find a two-term perturbation solution.
RaD =
66. (a) Show that the transient governing equation for a constant area fin with constant properties
that is losing heat convectively with the surroundings can be written as
2
1
=
m2 .
t
x2
2 Eigenvalues
of an N N , circulant, banded matrix of the form [?]

3
2
b
c
0
... 0 a
6 a
b
c
... 0 0 7
7
6
6 0
a
b
... 0 0 7
7
6
7
6 .
.
.
.
.
..
..
..
..
7
6 ..
7
6
4 0 ...
0
a
b c 5
c
0
...
0
a b
are j = b + (a + c) cos{2(j 1)/N } i(a c) sin{2(j 1)/N }, where j = 1, 2, . . . , N .
223
Figure A.13: Mengers Sponge
(a) Initator
(b) Generator
Figure A.14: Initiator and generator for a Peano (space filling) curve. The generator is recursively
applied to generate the Peano curve.
(b) With prescribed base and tip temperatures, use an eigenfunction expansion to reduce to
an infinite set of ordinary differential equations. (c) Show that the steady state is attracting
for all initial conditions.
67. Construct a fractal that is similar to the Cantor set, but instead remove the middle 1/2 from
each line. Show that the fractal dimension is 1/2.
68. Shown below is Mengers Sponge. Calculate its Hausdorff dimension using each of the following
methods: (a) Dh = log P/ log S, (b) Box Counting (analytical), (c) Box Counting (graphical).
69. Shown below is a Peano curve, a single line that completely fills a unit square. Calculate its
Hausdorff dimension and state if the Peano curve is indeed a fractal.
70. A duct carrying fluid has the cross-section of Kochs curve. Show that the perimeter of the
cross-section is infinite while the flow area is finite.
71. Verify Cauchys formula for repeated integration by (a) integrating f (t) five times, (b) applying
Cauchys formula once with n = 5, (c) applying Cauchys formula twice, once to f (t) with
n = 2 and then to the result with n = 3, showing that
Z Z Z Z Z t
f ( )d = J 5 f (t) = J 3 J 2 f (t)
0
224
for f (t) = 16t3 .

72. Take f (t) = 2t5 and using the Caputo Right Hand Derivative, (a) calculate D3 f (t) (take m = 4,
= 3) and verify that you get the same result as traditional differentation by comparing to
d3 f /dt3 . (b) Calculate D2.5 f (t) and plot the function.
73. Consider the heat flux in a blast furnace taht was calculated to be
p
1/2
q (t) = cp 0 Dt g(t),
where
g(t) = Tsurf (t) T0 ,

which is simply the derivative of order = 1/2 of the temperature difference at the surface.
Assume that the function g(t) is given as g(t) = 14 sin(t/60) where t is in minutes and
the thermocouples sample once per minute, giving the discrete data set gi = 14 sin(i/60).
Calculate the fractional derivative numerically using the first 2 hours of data and plot both
the heat flux at the surface and g(t).
Hint: It is easiest to calculate the binomial coefficients recursively, according to the recursion
formula:

= 1,
(A.14)
0
k+1
.
(A.15)
=
k
k+1
k
Note: In large time intervals (t very large), which would be of interest in this problem, the
calculation we used would not be suitable because of the enormous number of summands
in the calculation of the derivative and because of the accumulation of round off errors. In
these situations, the principle of short memory is often applied in which the derivative only
depends on the previous N points within the last L time units. The derivative with this short
memory assumption is typically written as (tL) Dt .
74. Consider the periodic heating and cooling of the surface of a smooth lake by radiation. The
surface is subject to diurnal heating and nocturnal cooling such that the surface temperature
can be described by Ts (t) = To + Ta sin t. Assume the heat diffusion to be one-dimensional
and find the heat flux at the surface of the lake. The following steps might be useful:
(a) Assume transient one-dimensional heat conduction:
2 T (x, t)
T (x, t)
= 0,
t
x2
with initial and boundary conditions T (x, 0) = To and T (0, t) = Ts (t). Also assume the lake
to be a semi-infinite planer medium (a lake of infinite depth) with T (, t) = 0.
(b) Non-dimensionalize the problem with a change of variables = 1/2 x and (x, t) =
T (x, t) To .
(c) Use the following Laplace transform properties to transform the problem into the Laplace
domain:
f (x, t)
(with 0 initial conditions)
= s F (x, s)
L
t
225
T1
T2
Figure A.15: Plane wall in steady state.

and
f (x, t)
L
F (x, s)
=
x
x
(d) Solve the resulting second order differential equation for (, s) by applying the transformed boundary conditions.
(e) Find / and then substitute (, s) into the result. Now take the inverse Laplace
transform of / and convert back to the original variables. Be careful! The derivative of
order 1/2 of a constant is not zero! (see simplifications in (g) to simplify)
(f) You should now have an expression for T (x, t)/x. Substitute this into Fouriers Law to
calculate the heat flux, q (x, t) = k T (x, t)/x.
(g) Evaluate this expression at the surface (x = 0) to find the heat flux at the surface of the
lake. The following simplifications might be helpful:
C
1/2 C
= 1/2
1/2
t
t
[Cg(t)]
g(t)
=
C
t
t
(h) The solution should look now look like
qs (t) =
k d1/2 (Ta sin t)

1/2
dt1/2
75. Consider radiation between two long concentric cylinders of diameters D1 (inner) and D2
(outer). (a) What is the view factor F12 . (b) Find F22 and F21 in terms of the cylinder
diameters.
76. Temperatures at the two sides of a plane wall shown in Fig. A.42 are TL and TR , respectively.
For small , find a perturbation steady-state temperature distribution T (x) if the dependence
of thermal conductivity on the temperature has the form
T TL
.
k(T ) = k0 1 +
TR TL
77. One side of a plane wall shown in Fig. A.43 has a fixed temperature and the other is adiabatic.
With an initial condition T (x, 0) = f (x), determine the temperature distribution in the wall
T (x, t) at any other time. Assume constant properties.
226
T1
T2
Figure A.16: Plane wall in unsteady state.
Figure A.17: Turbine blade.

78. Using an eigenfunction expansion, reduce the governing PDE in Problem 77 to an infinite set
of ODEs.
79. A room that loses heat to the outside by convection is heated by an electric heater controlled
by a proportional controller. With a lumped capacitance approximation for the temperature,
set up a mathematical model of the system. Determine the constraint on the controller gain
for the system response to be stable. What is the temperature of the room after a long time?
80. A turbine blade internally cooled by natural convection is approximated by a rotating natural
circulation loop of constant cross-section. The heat rate in and out at the top and bottom,
respectively, is Q while the rest of the loop is insulated. Find the steady-state velocity in the
loop. Consider rotational forces but not gravity. State your other assumptions.
81. An infinite number of conductive rods are set up between two blocks at temperatures Ta and Tb .
The first rod has a cross-sectional area A1 = A, the second A2 = A/, the third A3 = A/ 2 ,
and so on, where > 1. What is the total steady-state heat transfer rate between the two
blocks? Assume that the thermal conductivity k is a constant, and that there is no convection.
82. Show that the functions 1 (x) = 2 sin x and 2 (x) = 2 sin 2x are orthonormal in the
interval [0, 1] with respect to the L2 norm. Using these as test functions, use the Galerkin
L
000
111
111
000
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
a
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
..
.
000
111
111
000
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
b
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
Figure A.18: Infinite number of rods.
227
d
R
Figure A.19: Lamp and disk
C
Figure A.20: Radiating triangle.

method to find an approximate solution of the steady-state fin equation
T T = 0,
with T (0) = 0, T (1) = 1.
83. A lamp of q W is radiating equally in all directions. Set up the governing equation for the
temperature T (r) in the disk.
84. Determine the steady-state temperature distribution in the triangle shown, if the hypotenuse
is adiabatic, one of the sides is at one temperature and the other is at another.
85. A ball with coefficient of restitution r falls from height H and undergoes repeated bouncing.
Determine the temperature of the ball as a function of time T (t) if heat loss is by convection
to the atmosphere. Assume that the energy loss at every bounce goes to heat the ball.
86. Heat at the rate of q per unit volume is generated in a spherical shell that lies between R
and R + . If heat loss is by convection on the external surface only, find the steady-state
temperature distribution.
87. Two identical fins of square cross section exchange heat by radiation between them and with
the surroundings. Set up the governing equations for the temperatures T (x) in the fins and
solve for weak radiation.
88. Determine the time-dependent temperature T (t) of a block of weight M g that is being slided
at a constant velocity V across a horizontal surface. Assume that the frictional heat generated
at the interface (with coefficient of friction ) goes entirely to the block, that its temperature
is uniform, and that heat loss from it is only by convection to the surroundings.
89. (From Incropera and DeWitt, 5th edition) The figure shows a conical section fabricated from
pure aluminum. It is of circular cross section having diameter D = ax1/2 , where a = 0.5m1/2 .
The small end is located at x1 = 25 mm and the large end at x2 = 125 mm. The end
228
Figure A.21: Conical fin

D
Figure A.22: Area between square and circle.

temperatures are T1 = 600K and T2 = 400K, while the lateral surface is well insulated. (a)
Derive an expression for the temperature distribution T (x) in symbolic form, assuming onedimensional conditions. (b) Sketch the temperature distribution. (c) Calculate the heat rate
qx .
90. (From Incropera and DeWitt, 5th edition) Copper spheres of 20 mm diameter are quenched
by dropping into a tank of water that is maintained at 280 K. The spheres may be assumed
to reach the terminal velocity on impact and to drop freely though the water. Estimate the
terminal velocity by equating the drag and gravitational forces acting on the sphere. What is
the approximate height of the water tank needed to cool the spheres from an initial temperature
of 360 K to a center temperature of 320 K?
91. (From Incropera and DeWitt, 5th edition) Neglecting the longitudinal temperature gradient
229
in laminar flow in a circular tube, the energy equation is
T
T
T

r
.
u
+v
=
x
r
r r
r
The heat flux at the wall is qs = h(Ts Tm ), where Tm is the mean (or bulk) temperature
of the fluid at a given cross section and Ts is the wall temperature. Slug flow is an idealized
condition for which the velocity is assumed to be uniform over the entire tube cross section
with v = 0, u = constant, and T /x = dTm /dx = constant with respect to r. For uniform
wall heat flux, determine T (r) and the Nusselt number N uD .
92. (From Incropera and DeWitt, 5th edition) Consider natural convection within two long vertical
plates maintained at uniform temperatures of Ts,1 and Ts,2 , where Ts,1 > Ts,2 . The plates are
open at their ends and are separated by the distance 2L. (a) Sketch the velocity distribution
in the space between the plates. (b)Write appropriate forms of the continuity, momentum,
and energy equations for laminar flow between the plates. (c) Evaluate the temperature
distribution, and express your result in terms of the mean temperature Tm = (Ts,1 + Ts,2 )/2.
(d) Estimate the vertical pressure gradient by assuming the density to be a constant m
corresponding to Tm . Substituting from the Boussinesq approximation, obtain the resulting
form of the momentum equation. (e) Determine the velocity distribution. (Hints: The net
flow across any horizontal section should be zero. Assume that the flow is hydrodynamically
and thermally fully developed. Here the Boussinesq approximation means use of the simplified
density variation expression m m (T Tm ).)
93. (From Incropera and DeWitt, 5th edition) For forced convection boiling in smooth tubes, the
heat flux can be estimated by combining the separate effects of boiling and forced convection.
The Rohsenow and Dittus-Boelter correlations may be used to predict nucleate boiling and
forced convection effects, with 0.019 replacing 0.023 in the latter expression. Consider water
at 1 atm with a mean velocity of 1.5 m/s and a mean temperature of 95 C flowing through a
15-mm diameter brass tube whose surface is maintained at 110 C. Estimate the heat transfer
rate per unit length of the tube.
94. (From Incropera and DeWitt, 5th edition) A single-pass, cross-flow heat exchanger uses hot
exhaust gases (mixed) to heat water (unmixed) from 30 to 80 C at a rate of 3 kg/s. The
exhaust gases, having thermophysical properties similar to air, enter and exit the exchanger
at 225 and 100 C, respectively. If the overall heat transfer coefficient is 200 W/m2 K, estimate
the required surface area using LMTD and -NTU method respectively.
95. (From Incropera and DeWitt, 5th edition) Saturated water vapor leaves a steam turbine at
a flow rate of 1.5 kg/s and a pressure of 0.51 bar. The vapor is to be completely condensed
to saturated liquid in a shell-and-tube heat exchanger that uses city water as the cold fluid.
The water enters the thin-walled tubes at 17 C and is to leave at 57 C. Assuming an overall
heat transfer coefficient of 2000 W/m2 K, determine the required heat exchanger surface area
and the water flow rate. After extended operation, fouling causes the overall heat transfer
coefficient to decrease to 1000 W/m2 K, and to completely condense the vapor flow rate. For
the same water inlet temperature and flow rate, what is the new vapor flow rate required for
complete condensation?
96. (From Incropera and DeWitt, 5th edition) Advances in very large scale integration (VLSI) of
electronic devices on a chip are often restricted by the ability to cool the chip. For mainframe
230
computers, an array of several hundred chips, each of area 25 mm2 , may be mounted on a
certain ceramic substrate (backside of substrate insulated). A method of cooling the array
is by immersion in a low boiling point fluid such as refrigerant R-113. At 1 atm and 321 K,
properties of the saturated liquid are = 5.147 104 N s/m2 , cp = 983.8 J/kgK, and
P r = 7.183. Assume values of Cs,f = 0.004 and n = 1.7. (a) Estimate the power dissipated by
a single chip if it is operating at 50% of the critical heat flux. (b) What is the corresponding
value of the chip temperature? (c) Compute and plot the chip temperature as a function of
0.90.
surface heat flux for 0.25 qs /qmax
97. (From Incropera and DeWitt, 5th edition) A thermosyphon consists of a closed container that
absorbs heat along its boiling section and rejects heat along its condensation section. Consider
a thermosyphon made from a thin-walled mechanically polished stainless steel cylinder of
diameter D. Heat supplied to the thermosyphon boils saturated water at atmospheric pressure
on the surfaces of the lower boiling section of length Lb and is then rejected by condensing
vapor into a thin film, which falls by gravity along the wall of the condensation section of
length Lc back into the boiling section. The two sections are separated by an insulated section
of length Li .The top surface of the condensation section may be treated as being insulated.
The thermosyphon dimensions are D = 20 mm, Lb = 20 mm, Lb = 40 mm, and Li = 40
mm. (See figure in book) (a) Find the mean surface temperature, Ts,b , of the boiling surface
if the nucleate boiling heat flux is to be maintained at 30% of the critical heat flux. (b)
Find the mean surface temperature, Ts,c , of the condensation section assuming laminar film
condensation. (c) Find the total condensation flow rate, m,
within the thermosyphon.
98. (From Incropera and DeWitt, 5th edition) A steam generator consists of a bank of stainless
steel (k = 15W/mK) tubes having the core configuration of Fig. A.26 and an inner diameter
of 13.8 mm. The tubes are installed in a plenum whose square cross section is 0.6 m on a
side, thereby providing a frontal area of 0.36 m2 . Combustion gases, whose properties may be
approximated as those of atmospheric air, enter the plenum at 900 K and pass in cross flow
over the tubes at 3 kg/s. If saturated water enters the tubes at a pressure of 2.455 bars and a
flow rate of 0.5 kg/s, how many tubes rows are required to provide saturated steam at the tube
outlet? A convection coefficient of 10,000 W/m2 K is associated with boiling in the tubes.
99. (From Incropera and DeWitt, 5th edition) The spectral, directional emissivity of a diffuse
material at 2000 K has the distribution as in Fig. A.24. Determine the total, hemispherical
emissivity at 2000 K. Determine the emissive power over the spectral range 0.8 to 2.5 m and
for the directions 0 30 .
100. (From Incropera and DeWitt, 5th edition) Solar irradiation of 1100 W/m2 is incident on a large,
flat horizontal metal roof on a day when the wind blowing over the roof causes a convection
heat transfer coefficient of 25 W/m2 K. The outside air temperature is 27 , the metal surface
absortivity for incident solar radiation is 0.60, the metal surface emissivity is 0.20, and the roof
is well insulated from below. Estimate the roof temperature under steady-state conditions.
231
Figure A.23: Steam generator.
Figure A.24: Emissivity of material.
232
2m
1m
2m
Figure A.25: Parallel plates.

101. Maxwells equations of electromagnetic theory are
D
t
B
E=
t
D=
B=0
H=J+
(A.16)
(A.17)
(A.18)
(A.19)
where H, B, E, D, J, and are the magnetic intensity, magnetic induction, electric field,
electric displacement, current density, and charge density, respectively. For linear materials
D = E, J = gE (Ohms law), and B = H, where is the permittivity, g is the electrical
conductivity, and is the permeability. For = 0 and constant , g and , show that
H
2H
g
=0
2
t
t
2E
E
2 E 2 g
=0
t
t
2 H
(A.20)
(A.21)
With = 8.8542 1012 C2 N1 m2 , and = 1.2566 106 NC2 s2 , find the speed of an
electromagnetic wave is free space.
102. (From Incropera and DeWitt, 5th edition) Consider the parallel plates of infinite extent normal
to the page having opposite edges aligned as shown in the sketch. (a) Using appropriate view
factor relations and the results for opposing parallel planes, develop an expression for the
view factor between the plates. (b) Write a numerical code based on the integral definition to
compute the view factor and compare.
103. (From Incropera and DeWitt, 5th edition) A room is 6 m wide, 10 m long and 4 m high. The
ceiling, floor and one long wall have emissivities of 0.8, 0.9 and 0.7 and are at temperatures
of 40 C, 50 C and 15 C, respectively. The three other walls are insulated. Assuming that the
surfaces are diffuse and gray, find the net radiation heat transfer from each surface.
104. (From Incropera and DeWitt, 5th edition) A radiant heater consists of a long cylindrical
heating element of diameter D1 = 0.005 m and emissivity 1 = 0.80. The heater is partially
enveloped by a long, thin parabolic reflector whose inner and outer surfaces emissivities are
2i = 0.10 and 2o = 0.80, respectively. Inner and outer surface areas per unit length of the
reflector are each A2i = A2o = 0.20 m, and the average convection coefficient for the combined
233
Surroundings
Heater
135
Air
Reflector
Figure A.26: Figure for problem 3.

inner and outer surfaces is h2(i,o) = 2 W/m2 K. The system may be assumed to be in an infinite,
quiescent medium of atmospheric air at T = 300 K and to be exposed to large surroundings
at Tsur = 300 K. (a) Sketch the appropriate radiation circuit, and write expressions for each
of the network resistances. (b) If, under steady state conditions, electrical power is dissipated
in the heater at P1 = 1500 W/m and the heater surface temperature is T1 = 1200 K, what is
the net rate at which radiant energy is transferred from the heater? (c) What is the net rate
at which radiant energy is transferred from the heater to the surroundings? (d) What is the
temperature T2 of the reflector?
105. Consider the convective fin equation
d2 T
T = 0,
dx2
where 0 x 1, with x(0) = 1, x(1) = 0. Solve using the following methods. You may have
to transform the dependent or independent variable differently for each method. In each case
show convergence.
(a) Finite differences: Divide into N parts, write derivatives in terms of finite differences,
reduce to algebraic equations, apply boundary conditions, and solve.
(b) Trigonometric Galerkin: Expand in terms of N trigonometric functions, substitute in
equation, take inner products, reduce to algebraic equations, and solve.
(c) Chebyshev Galerkin: Expand in terms of N Chebyshev polynomials, substitute in equation, take inner products, reduce to algebraic equations, and solve.
(d) Trigonometric collocation: Expand in terms of N trigonometric functions, substitute in
equation, take inner products, apply collocation, and solve.
(e) Galerkin finite elements: Divide into N elements, assume linear functions, integrate by
parts, assemble all equations, apply boundary conditions, and solve.
(f) Polynomial moments: Assume dependent variable to be a N th-order polynomial that
satisfies boundary conditions, obtain algebraic equations by taking moments, and solve.
234
106. If eE/T is proportional to the heat generated within a tank by chemical reaction, and there is
heat loss by convection from the tank, show that the temperature of the tank T is determined
by
dT
Mc
= aeE/T hA(T T )
dt
where M is the mass, c is the specific heat, a is a proportionality constant, h is the convective
heat transfer coefficient, A is the surface area of the tank, and T is the ambient temperature.
Nondimensionalize the equation to
dT
)
= e1/T H(T T
dt
(a) For T
= 0.1, draw the bifurcation diagram with H as the bifurcation parameter, and
determine the bifurcation points.
(b) Determine the stability of the different branches of the bifurcation diagram.
107. The governing equations for the dynamic behavior of a toroidal convective loop with known
heat flux has been shown to be
dx
=yx
(A.22)
dt
dy
= a zx
(A.23)
dt
dz
= xy b
(A.24)
dt
(a) Using the same approach and notation as far as possible, modify the dynamical system
to include the effects of axial conduction.
(b) Determine the critical points of the conductive system and analyze their linear stability.
(c) Show that there is a conductive solution for heating that is symmetric about a vertical
diameter, and analyze its nonlinear stability.
108. Consider a toroidal convective loop with known wall temperature. Show that the dynamical
system for this problem can be reduced to the Lorenz equations. Determine the dimension of
the strange attractor for these equations.
109. Graphically compare the numerical and approximate boundary layer solutions in the problem
of one-dimensional flow in a pipe with advection, convection and axial conduction.
110. Consider the energy equation in the one-dimensional flow in a tube of finite wall thickness.
Include convection between the fluid and the tube, between the tube and the environment,
and axial conduction in the tube wall, but neglect axial conduction in the fluid. The inlet
temperature of the in-tube fluid is known, and the tube wall is adiabatic at the two ends.
(a) Show that the governing equations can be reduced to a form
dy
= Ay.
d
is the nondimensional axial coordinate,
i
y = t ,
qt
235
11111111111111111111
00000000000000000000
00000000000000000000
11111111111111111111
00000000000000000000
11111111111111111111
00000000000000000000
11111111111111111111
00000000000000000000
11111111111111111111
Figure A.27: CAPTION HERE.
and
a
a
0
0
1 .
A= 0
b b + c 0
i is the nondimensional temperature of the in-tube fluid, t that of the tube, and qt the
heat flux.
(b) Show that this can also be written as a third-order differential equation.
(c) Solve as much as you can analytically, and then assume numerical values for the parameters (e.g. a = 3, b = 10/3, c = 8/3) and complete the solution.
(d) Which of the cases of the parameters a, b and c going to either zero or infinity are regular
perturbations and which are singular?
111. Derive an expression for heat transfer in a fractal tree-like microchannel net3 .
112. A body is being cooled by natural convection and black-body radiation. Assume lumped
parameters and derive an approximation for its temperature T (t) based on neglecting one
mode of heat transfer for small times and the other for long times. Compare with a numerical
result.
113. Consider the two-dimensional steady-state temperature distribution T (x, y) on the plate ABCDEFG
where AB = BC = DE = FG = GA and CD = BE = EF. The line ABC is adiabatic, AG is at
temperature T1 and the others are at T2 . (a) Assume that the temperature distribution in the
dashed (interior) line BE is f (y) and solve the temperature distributions in the two rectangles
separately. (b) Equate the normal heat flux on either side of BE to get an equation for f (y).
(c) Suggest a numerical solution.
114. Explore a perturbation solution of the previous problem for a geometry in which CD/AG is
very small (instead of being 0.5 above).
115. Fluids A and B exchange heat in the concentric-tube counterflow heat exchanger shown.
Analyze the dynamics of the temperature field for a step change in the inlet temperature of
one of the fluids.
116. Derive the dynamical system for a convection loop with known heat flux that loops around in
a circle twice in a vertical plane before merging with itself. Investigate its static and dynamic
behavior.
3 Y. Chen and P. Cheng, Heat transfer and pressure drop in fractal tree-like microchannel nets, International
Journal of Heat and Mass Transfer, Vol. 45, pp. 26432648, 2002
236
y
A
Figure A.28: CAPTION HERE

111111111111111111111
000000000000000000000
000000000000000000000
111111111111111111111
B
A
B
111111111111111111111
000000000000000000000
000000000000000000000
111111111111111111111

117. Show that a convectively cooled lumped mass that is heated with a time-varying source is
governed by an equation of the type
dT
+ T = Q(t).
dt
If Q = 1 + sin t, find the amplitude of the temperature response as a function of .
118. Show that a radiative fin with heat generation is governed by an equation of the type
d2 T
T 4 + Q = 0.
dx2
for 0 x 1. With Q = 1 and boundary conditions T (0) = 0, T (1) = 2, solve T (x)

numerically for different . From this see if and where boundary layers develop as 0.
119. Consider steady-state conduction in a L 2L rectangular plate, with 1. If temperatures
are prescribed on the short sides and the long sides are insulated, show that the governing
equation in normalized spatial variables can be written as
2
2T
2T
+
=0
x2
y 2
in 0 x 1, 1 y 1 with boundary conditions T (0, y) = f (y), T (1, y) = g(y), (T /y)y=1 =

(T /y)y=1 = 0.
237
(a) Assume an outer expansion of the form

T (x, y) = T0 (x, y) + 2 T1 (x, y) + . . .
and show that
T0 = a0 x + b0 (1 x)
where a and b are undetermined constants.
(b) With boundary layers near x = 0 and x = 1, show that matching with the outer solution
gives
Z 1
1
a0 =
g(y) dy
(A.25)
2 1
Z 1
1
b0 =
f (y) dy
(A.26)
2 1
120. Using multiple scales, solve the Fisher equation
2T
T
= 2 + T (1 T )
t
x
with < x < , t > 0, T (x, 0) = 1/{1 + exp(x)}, and 1. Plot the analytical and
numerical results for = 0.01, = 1 and t = 0, t = 5, and t = 10.
121. Find an asymptotic approximation for
T
T
+
= f (T )
t
x
with < x < , t > 0, T (x, 0) = g(x) and 1 that is valid for large t. Then reduce to
the special case for f (T ) = T (1 T ) and g(x) = 1/{1 + exp(x)}.
122. Show that the temperature T (t) of a lumped body that is subject to convection and radiation
to a constant ambient temperature T is governed by an equation of the form
dT
4
+ a(T T ) + b(T 4 T
) = 0,
dt
and that two steady states are conditionally possible. Analyze the stability of each.
123. The temperature T (t) of an object is controlled to a desired value Ts by blackbody radiation
from a lamp at a temperature T (t). T can be measured and on the basis of that T is
varied by a PID controller. Model the system with control, and determine the condition for
its stability.
124. A long tube held vertically is kept at a uniform temperature Tw that is higher than that of the
surrounding fluid T . There is an upward flow through the tube due to natural convection.
What is the temperature of the fluid leaving the tube? Make suitable assumptions.
125. Consider transient thermal conduction in a convective fin with its two ends kept at fixed
temperatures. Show that the steady-state temperature distribution is globally stable.
238
T
w
m
1
m
1
m
1
239
y
D

126. An infinite one-dimensional lattice consists of repetitions of a unit consisting of two identical
masses followed by a third mass that is different. The springs and distances between the masses
are all identical. Find the small wave number phonon speed of the acoustic mode.
127. The surface of a sphere is divided into n unequal parts numbered 1, 2, . . . , n with areas
A1 , A2 , . . . , An respectively. Show that the view factor
Fij =
Aj
.
A1 + A2 + . . . + An
128. There are at least two ways to obtain a regular perturbation solution for the steady-state
temperature T (x, y) due to conduction without heat generation for the plate shown in the
figure, where 1, For each, set up two terms of the problem (i.e. provide the equations to
be solved and the corresponding boundary conditions), but do not solve.
129. Find the maximum temperature on a thin rectangular plate with uniform temperature at the
edges and constant heat generation per unit area by using a one-term method of moments
approximation.
130. Determine the temperature distribution along a fin of constant area which has lateral convection and also a small amount of radiative cooling. The base temperature is known and the tip
is adiabatic.
240
131. The governing equations for the dynamic behavior of a toroidal natural convection loop with
axial conduction that is symmetrically heated with known heat flux is
dx
= y x,
dt
dy
= zx cy,
dt
dz
= xy b cz.
dt
(A.27)
(A.28)
(A.29)
Draw the bifurcation diagram and determine the linear and nonlinear stability of the conductive
solution.
132. Model and solve the dynamics of a counterflow heat exchanger.
133. Describe how you would solve this problem using more typical methods and what other information would be required.
134. (Chap. 1) A square silicon chip (k = 150 W/m K) is of width 5 mm on a side and of thickness
1 mm. The chip is mounted on a substrate such that its side and back surfaces are insulated,
while the front surface is exposed to a coolant. If 4 W are being dissipated in circuits mounted
to the back surface of the chip, what is the steady-state temperature difference between the
back and front surfaces.
135. (Chap. 1) A square isothermal chip is of width 5 mm on a side and is mounted on a substrate
such that its side and back surfaces are well insulated, while the front surface is exposed to the
flow of a coolant at temperature 85 C. If the coolant is air and the corresponding convection
coefficient is h = 200 W/m2 K, what is the maximum allowable chip power? If the coolant is
a dielectric liquid for which h = 3000 W/m2 K, what is the maximum allowable power?
136. (Chap. 1) An overhead 25 m long, uninsulated industrial steam pipe of 100 mm diameter is
routed through a building whose walls and air are at 25 C. Pressurized steam maintains a
pipe surface temperature of 150 C, and the coefficient associated with natural convection is
h = 10W/m2 K. The surface emissivity is 0.8. (a) What is the rate of heat loss from the steam
line? (b) If the steam is generated in a gas-fired boiler operating at an efficiency of 0.9 and
natural gas is priced at $0.01 per MJ, what is the annual cost of heat loss from the line?
137. (Chap. 1) Three electric resistance heaters of length 250 mm and diameter 25 mm are submerged in a 10 gallon tank of water, which is initially at 295 K. (a) If the heaters are activated,
each dissipating 500 W, estimate the time required to bring the water to a temperature of 335
K. (b) If the natural convection coefficient is given by h = 370(Ts T )1/3 , where h is in W/m2 K
and Ts and T are temperatures of the heater surface and water in K, respectively, what is the
temperature of each heater shortly after activation and just before deactivation? (c) If the
heaters are inadvertently activated when the tank is empty, the natural convection coefficient
associated with heat transfer to the ambient air at T = 300 K may be approximated as
h = 0.70(Ts T )1/3 . If the temperature of the tank walls is also 300 K and the emissivity of
the heater surface is 0.85, what is the surface temperature of each heater under steady-state
conditions?
138. (Chap. 2) Beginning with a differential control volume, derive the heat diffusion equation in
cylindrical coordinates.
241

139. (Chap. 3) A circular copper rod of 1 mm diameter and 25 mm length is used to enhance heat
transfer from a surface that is maintained at 100 C. One end of the rod is attached to this
surface, while the other end is attached to a second surface at 0 C. Air flowing between the
surfaces and over the rods is at 0 C. The convective heat transfer coefficient at the surface of
the rod is 100 W/m2 K. What is the rate of heat transfer by convection from the rod to the
air?
140. (Chap. 4) Using the method of separation of variables, determine the steady-state temperature
field T (x, y) in the rectangle shown below. Make a three-dimensional plot. Also draw the
isotherms on a two-dimensional plot.
Lengths AB = 3, and BC = 2; temperatures at the boundaries at AB, CD and DA are zero,
but at BC is
y,
0y1
T (3, y) =
2 y, 1 < y 2
141. (Chap. 5) In a material processing experiment conducted aboard the space shuttle, a coated
niobium sphere of 10 mm diameter is removed from a furnace at 900 C and cooled to a
temperature of 300 C. Take = 8600 kg/m3 , c = 290 J/kg K, and k = 63 W/m K. (a) If
cooling is implemented in a large evacuated chamber whose walls are at 25 C, determine the
time required to reach the final temperature if the emissivity is 0.1. How long would it take
if the emissivity were 0.6? (b) To reduce the time required for cooling, consideration is given
to immersion of the sphere in an inert gas stream at 25 C with h = 200 W/m2 K. Neglecting
radiation, what is the time required for cooling?
142. (Chap. 5) The density and specific heat of a plastic material are known ( = 950 kg/m3 ,
cp = 1100 J/kg K), but its thermal conductivity is unknown. An experiment is performed in
which a thick sample is heated to a uniform temperature of 100 and then cooled by passing
air at 25 C over one surface. A thermocouple embedded a distance of 10 mm below the
surface records the thermal response of the plastic during cooling. If the convection coefficient
associated with air flow is 200 W/m , and a temperature of 60 is recorded 5 min after the
onset of cooling, what is the thermal conductivity of the material?
143. (Chap. 6) A shaft with a diameter of 100 nm rotates at 9000 rpm in a journal bearing that
is 70 mm long. A uniform lubricant gap of 1 mm separates the shaft and the bearing. The
lubricant properties are = 0.03 Ns/m2 , k = 0.15 W/m K, while the bearing material has
a thermal conductivity of kb = 45 W/m K. (a) Determine the viscous dissipation in W/m3
in the lubricant. (b) Determine the rate of heat transfer in W from the lubricant, assuming
that no heat is lost through the shaft. (c) If the bearing housing is water-cooled and the outer
242
bearing
lubricant
111111111
000000000
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
000000000
111111111
shaft

surface of the housing is 30 C, determine the temperatures at the interface between lubricant
and shaft and lubricant and bearing.
144. (Chap. 7) Consider the wing of an aircraft as a flat plate of length 2.5 m in the flow direction.
The plane is moving at 100 m/s in air that is at a pressure of 0.7 bar and temperature 10 C.
The top surface of the wing absorbs solar radiation at a rate of 800 W/m2 . Estimate the
steady-state temperature of the wing, assuming it to be uniform.
145. (Chap. 8) The evaporator section of a heat pump is installed in a large tank of water, which
is used as a heat source during the winter. As energy is extracted from the water, it begins
to freeze, creating an ice/water bath at 0 , which may be used for air-conditioning during
the summer. Consider summer cooling conditions for which air is passed through an array of
copper tubes, each of inside diameter D = 50 mm, submerged in the bath. (a) If air enters
each tube at a mean temperature of Tm,i = 24 C and a flow rate of m
= 0.01 kg/s, what tube
length L is needed to provide an exit temperature of Tm,o = 14 C? With 10 tubes passing
through a tank of total volume V = 10 m3 , which initially contains 80% ice by volume, how
long would it take to completely melt the ice? The density and latent heat of fusion of ice are
920 kg/m3 and 3.34 105 J/kg, respectively. (b) The air outlet temperature may be regulated
by adjusting the tube mass flow rate. For the tube length determined in part (a), compute
and plot Tm,o as a function of m
for 0.005 m
0.05 kg/s. If the dwelling cooled by this
system requires approximately 0.05 kg/s of air at 16 C, what design and operating conditions
should be prescribed for the system?
146. (Chap. 9) Beverage in cans 150 mm long and 60 mm in diameter is initially at 27 C and is to
be cooled by placement in a refrigerator compartment at 4 C. In the interest of maximizing
the cooling rate, should the cans be laid horizontally or vertically in the compartment? As a
first approximation, neglect heat transfer from the ends.
147. (Chap. 9) A natural convection air heater consists of an array of N parallel, equally spaced
vertical plates, which may be maintained at a fixed temperature Ts by embedded electric
heaters. The plates are of length and width L = W = 300 mm and are in quiescent, atmospheric air at T = 20 C. The total width of the array cannot exceed a value of War = 150
mm. For Ts = 75 C, what is the plate spacing (distance between plates) S that maximizes
heat transfer from the array? For this spacing, how many plates comprise the array and what
is the corresponding rate of heat transfer from the array?
243
148. (Chap. 10) Saturated steam at 1 atm condenses on the outer surface of a vertical 100 mm
diameter pipe 1 m long, having a uniform surface temperature of 94 C. Estimate the total
condensation rate and the heat transfer rate to the pipe.
149. (Chap. 11) A shell-and-tube heat exchanger consists of 135 thin-walled tubes in a double-pass
arrangement, each of 12.5 mm diameter with a total surface area of 47.5 m2 . Water (the
tube-side fluid) enters the heat exchanger at 15 C and 6.5 kg/s and is heated by exhaust gas
entering at 200 C and 5 kg/s. The gas may be assumed to have the properties of atmospheric
air, and the overall heat transfer coefficient is approximately 200 W/m2 K. (a) What are the
gas and water outlet temperatures? (b) Assuming fully developed flow, what is the tube-side
convection coefficient? (c) With all other conditions remaining the same, plot the effectiveness
and fluid outlet temperature as a function of the water flow rate over the range from 6 to 12
kg/s. (d) What gas inlet temperature is required for the exchanger to supply 10 kg/s of hot
water at an outlet temperature of 42 C, all other conditions remaining the same? What is the
effectiveness for this operating condition?
150. (Chap. 11) A boiler used to generate saturated steam is in the form of an unfinned, cross-flow
heat exchanger, with water flowing through the tubes and a high temperature gas in cross
flow over the tubes. The gas, which has a specific heat of 1120 J/kg K and a mass flow rate
of 10 kg/s, enters the heat exchanger at 1400 K. The water, which has a flow rate of 3 kg/s,
enters as saturated liquid at 450 K and leaves as saturated vapor at the same temperature.
If the overall heat transfer coefficient is 50 W/m2 K and there are 500 tubes, each of 0.025 m
diameter, what is the required tube length?
151. (Chap. 12) The energy flux associated with solar radiation incident on the outer surface of
the earths atmosphere has been accurately measured and is known to be 1353 W/m2 . The
diameters of the sun and earth are 1.39 109 and 1.29 107 m respectively, and the distance
between the sun and the earth is 1.5 1011 m. (a) What is the emissive power of the sun? (b)
Approximating the suns surface as black, what is its temperature? (c) At what wavelength
is the spectral emissive power of the sun a maximum? (d) Assuming the earths surface to
be black and the sun to be the only source of energy for the earth, estimate the earths
temperature.
152. (Chap. 12) The spectral reflectivity distribution for white paint can be approximated by the
following stair-step function: = 0.75 for < 0.4 m, = 0.15 for 0.4 < < 3 m,
and = 0.96 for > 3 m. A small flat plate coated with this paint is suspended inside a
large enclosure, and its temperature is maintained at 400 K. The surface of the enclosure is
maintained at 3000 K, and the spectral distribution of its emissivity is = 0.2 for < 2.0
m, and = 0.9 for > 2.0 m. (a) Determine the total emissivity, , of the enclosure. (b)
Determine the total emissivity, , and absorptivity, , of the plate.
153. (Chap. 12) A thermocouple whose surface is diffuse and gray with an emissivity of 0.6 indicates
a temperature of 180 C when used to measure the temperature of a gas flowing through a large
duct whose walls have an emissivity of 0.85 and a uniform temperature of 450 C. (a) If the
= 125
convection heat transfer coefficient between the thermocouple and the gas stream is h
W/m2 K and there are negligible conduction losses from the thermocouple, determine the
temperature of the gas. (b) Consider a gas temperature of 125 C. Compute and plot the
1000
thermocouple measurement error as a function of the convection coefficient for 10 h
W/m2 K.
244
t
L
Figure A.36: Triangular fin [Same as Fig. 1.7 on p. 16].

x
T1
T (x)
T2
Figure A.37: Constant-area fin [Same as Fig. 1.8].

154. (Chap. 13) Consider two diffuse surfaces A1 and A2 on the inside of a spherical enclosure of
radius R. Using the following methods, derive an expression for the viewfactor F12 in terms
of A2 and R. (a) Find F12 by beginning with the expression Fij = qij /Ai Ji . (b) Find F12
using the view factor integral
Z Z
cos i cos j
1
dAi dAj
Fij =
Ai Ai Aj
R2
155. (Chap. 13) Two parallel, aligned disks, 0.4 m in diameter and separated by 0.1 m, are located
in a large room whose walls are maintained at 300 K. One of the disks is maintained at a
uniform temperature of 500 K with an emissivity of 0.6, while the backside of he second disk
is well insulated. If the disks are diffuse, gray surfaces, determine the temperature of the
insulated disk.
156. (Chap. 13) A solar collector consists of a long duct through which air is blown; its cross
section forms an equilateral triangle of side 1 m on a side. One side consists of a glass cover of
emissivity 1 = 0.9, while the other two sides are absorber plates with 2 = 3 = 1.0. During
operation the surface temperatures are known to be T1 = 25 C, T2 = 60 C, and T3 = 70 C.
What is the net rate at which radiation is transferred to the cover due to exchange with the
absorber plates?
157. (Chap. 13) Consider a circular furnace that is 0.3 m long and 0.3 m in diameter. The two
ends have diffuse, gray surfaces that are maintained at 400 and 500 K with emissivities of 0.4
and 0.5, respectively. The lateral surface is also diffuse and gray with an emissivity of 0.8 and
a temperature of 800 K. Determine the net radiative heat transfer from each of the surfaces.
158. Show that no solution is possible in Problem 4 if the boundary layer is assumed to be on the
wrong side.
159. Find the steady-state temperature distribution and velocity in a square-loop natural convection
loop. The total length of the loop is L and the distribution of the heat rate per unit length is:
Q between c and d, and Q between g and h. The rest of the loop is adiabatic.
245
11111111111111111111
00000000000000000000
00000000000000000000
11111111111111111111
11111111111111111111
00000000000000000000
00000000000000000000
11111111111111111111
Figure A.38: Flow in tube with non-negligible wall thickness.
1
0
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
246

L
111
000
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
a
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
..
.
111
000
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
b
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
000
111
160. Consider a long, thin, vertical tube that is open at both ends. The air in the tube is heated
with an electrical resistance running down the center of the tube. Find the flow rate of the air
due to natural convection. Make any assumptions you need to.
161. Obtain the table below by solving the fin conduction equation for the different boundary
conditions.
162. Derive the fin efficiency given below for a circular fin.
163. Read the section on the bioheat equation and solve the following.
164. Use separation of variables to find the steady-state temperature distribution T (x, y) in the
rectangle shown for the following boundary conditions. AB: T = T1 ; BC: adiabatic; CD:
T = T2 ; DA: T = T2 . The dimensions are AB = 2 units, BC = 1 unit. Also write a finitedifference program to find a numerical solution to the problem, and compare with the analytical
solution.
165. Use the finite-element based pde toolbox4 in Matlab to solve the numerical example in the
previous homework. Compare with the analytical result.
166. The exact solution of transient conduction in a plane wall with convection at the surfaces is
given in the book. Derive it.
4 Just
type pdetool in Matlab.
T1
T2
Figure A.42: Plane wall in steady state.
247
T1
T2
Figure A.43: Plane wall in unsteady state.
248
167. Show, following the steps in the book, that the transient heat equation in a semi-infinite solid
can be written as
d2 T
dT
= 2
.
d 2
d
Derive also the constant heat flux and surface convection solutions given.
168. The following problem is an example in the book. Write computer programs and find what is
asked using more nodes and smaller time steps. Also integrate until steady state is reached,
and plot the temperature at a few specific points.
169. Find the non-dimensional form of the energy equation for an incompressible fluid with constant
properties
T
+ u T = 2 T + /cp ,
t
where the dissipation function is
"
2
2
2
2 #
2
2
u
v
w
v
w v
u w
u
+2
+2
+
+
+
= 2
+
+
+
x
y
z
x y
y
z
z
x
in terms of the components of u = (u, v, w). Show that the Peclet and Eckert numbers arise
naturally.
170. For the flat plate boundary layer in forced convection, solve the hydrodynamic and thermal
equations
d3 f
d2 f
+
f
= 0,
d 3
d 2
Pr dT
d2 T
+
f
= 0,
d 2
2
d
2
numerically, with the boundary conditions f (0) = f (0) = T (0) = 0, and f () = T () = 1.

2
Compare f () and ddf2 (0) with values given in the book. Plot dT
d (0) as a function of Pr, and
compare with the approximation for Pr larger than 0.6 given in the book.
171. For natural convection over a vertical flat plate, solve the boundary layer equations
2
d3 f
d2 f
df
+ 3f 2 2
+ T = 0,
d 3
d
d
dT
d2 T
+
3Pr
f
= 0,
d 2
d
numerically, with the boundary conditions f (0) = f (0) = 0, T (0) = 1, and f () = T () =
0. Plot graphs choosing values of Pr to compare with those in the book.
172. For a counter-flow heat exchanger, determine from first principles (a) the heat rate expression,
(b) the log mean temperature difference, and (c) the -NTU relation.
249

173. There is flow on either side of a L L metal plate as shown. Assume that the flow is laminar
on both sides and that the local heat transfer coefficient is given by the boundary layer theory.
Determine the local heat transfer coefficient U (x, y) as a function of position. Neglect the
conductive thermal resistance of the plate. If the incoming temperatures of the two fluids are
Th,i and Tc,i , what are the temperature distributions at the outlets as functions of position.
174. Again consider a square plate as shown above, but with conduction important and the heat
transfer coefficient independent of position. Neglecting the variation of temperature across the
thickness of the plate, derive a steady state equation for the local temperature in the plate
T (x, y). With suitable boundary conditions, solve T (x, y) analytically and compare with a
numerical solution.
175. Show that Maxwell equations (Eqs. 1.531.56, p. 10 in the Notes) reduce to the wave equation
in electric and magnetic fields for zero electrical conductivity.
176. For small electrical conductivity, show that the one-dimensional equation for the electric field
2E
E
2E
2 g
=0
2
x
t
t
has an approximate solution of the form
E = ex f (x at)
where a = 1/ , and is small (so that 2 can be neglected). What is is terms of the
parameters g, and ?
177. Show, by integration, the expression in the book for the view factor between two aligned,
parallel rectangles.
178. Find the steady-state temperature distribution T (x) in a plate x [0, L], with T (0) = T1 , and
T (L) = T2 , and where the thermal conductivity varies linearly with temperature, i.e.
T T1
,
k = k0 1 +
T2 T1
with 1.
250
radiation from heater
r
179. Consider a fin equation
d2
=0
dx2
with (0) = 1, (1) = 1. (a) Solve analytically. (b) Solve using perturbations for 1 with
boundary layers on both sides. (c) Compare the two solutions graphically.
180. Determine the time-dependent temperature in the cooling of a lumped body with radiation
and weak convection. For simplicity, take T = 0.
181. Two bodies are in thermal contact with each other (as in Section 2.6.1, p. 25 but without heating Q). Solve for initial conditions T1 (0) = T2 (0) = T0 . Solve the problem using perturbations
if the contact thermal resistance ks is small. Clearly specify the small parameter . Compare
graphically with the analytical solution of the previous problem.
182. A circular disc of radius R lies on an adiabatic surface and is being heated by a radiative
heater. The heat input per unit area from the radiator is q. The disc is thin enough for the
transverse temperature variation to be neglected. If the temperature at the edge of the disc is
T (R) = T0 , determine the steady-state temperature distribution, T (r), in the disc, where r is
the radial coordinate.
183. Consider a thin plate of thickness , and constant thermal conductivity k. Both sides of the
plate are subjected to convective heat transfer with a constant heat transfer coefficient h. Show
that the temperature field, T (x, y), is governed by
2T
2T
+
m2 (T T ) = 0,
2
x
y 2
where m2 = 2h/k, and T is the temperature of the fluid surrounding the plate.
184. The governing equation for the temperature distribution in a thermal boundary due to flow
over a flat plate at constant temperature is
d2 T
Pr
dT
+
f
= 0,
d 2
2
d
where f () is the known solution of the Blasius equation. The boundary conditions are T (0) =
0 and T () = 1. Show that the solution can be written as
Z
Z
Pr r
f (s) ds dr
exp
2 0
Z
.
T () = Z 0
Pr r
f (s) ds dr
exp
2 0
0
251

185. The inside wall of a long cylinder of radius R and length L has a temperature distribution of
Tw (). Assuming that there is no participating medium inside the cylinder and that the inside
surface acts a black-body, determine the heat flux distribution on the inside wall q() due to
internal radiation.
186. Show that the set of all continuous temperature distributions T (x) in the interval x [a, b]
along a bar with T (a) = 0 and T (b) = 1 does not form a vector space.
187. In the above, show that T (x) would be a vector space if the second condition were changed to
T (b) = 0 (and T could assume any value). What is the dimension of that space?
188. S is a space of sufficiently smooth functions f (x) with x [0, 1] and f (0) = f (1) = 0 that is
endowed with the L2 inner product.
(a) Show that
d2
dx2
that operates on members of S is self-adjoint. Show that
L1 =
L2 =
d2
m2
dx2
where m is a constant, is also self-adjoint.

(b) Find the eigenvalues, n , and eigenfunctions of L1 . Normalize each eigenfunction by
dividing by its norm to obtain an orthonormal set, n (x).
(c) Solve the one-dimensional steady-state fin equation
L2 (T ) = 0
for the temperature distribution T (x), with T (0) = 0, T (1) = 1, where L2 is defined as
above.
(d) Let (x) = T x. Write the differential equation that (x) satisfies in the form L() =
g(x), and show that S. Expand (x) in terms of the eigenfunctions obtained above
as
X
ai i (x).
(x) =
i=0
Substitute into L() = g(x), and take the inner product of the equation with i to
determine the coefficients ai .
(e) Graphically compare the exact and the eigenfunction expansion (using perhaps five terms)
solutions of T (x).
252
convection
x
convection
111111111111111111111
000000000000000000000
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
convection
111111111111111111111
000000000000000000000
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
x
189. Consider one-dimensional flow in a pipe with lateral convection (with constant heat transfer
coefficient h) as shown. The flow velocity is V and the temperature T (x). Neglect the pipe
wall.
(a) Derive the governing energy balance differential equation and non-dimensionalize.
(b) Solve with boundary conditions T (0) = Tin and T (L) = Tout .
(c) Using the boundary layer approximation for small axial conduction, find the inner and
outer solutions as well as the composite solution.
(d) Show that the adjoint differential equation represents a pipe with flow in the opposite
direction.
190. Consider now the pipe with a thick wall which exchanges heat by convection with the fluid and
with the surroundings (with constant but different heat transfer coefficients). Assume axial
conduction in both wall and fluid.
(a) Derive the two governing equations for the wall and fluid temperatures, Tw (x) and Tf (x),
respectively.
(b) Nondimensionalize using = x/L, w = (Tw T )/T , f = (Tf T )/T , T =
Tfin T , where L is the length of the pipe, T is the ambient temperature, and Tfin is
the inlet temperature of the fluid.
T
T
, show that the nondimensional equations can
(c) Writing X = w f
and Y = f w
be written as
d
= A,
d
T
where = [X Y] .
(A.30)
253
(d) Assume boundary conditions Tf = Tin , dTw /dx = 0 at x = 0, and Tf = Tout , dTw /dx = 0
at x = L. Show that Eq. (A.30) can be written as
X(1) = D11 X(0) + D12 Y(0),
(A.31)
Y(1) = D21 X(0) + D22 Y(0),
(A.32)
where Y(0) and Y(1) are known.

(e) Determine X(0) from Eq. (A.32), and use this to find (0).
(f) Solve Eq. (A.30).
(g) Find the total heat lost to the surroundings by convection in terms of w (x).
191. Use the following numerical methods to solve the fin problem: T T = 0 with T (0) = 1 and
T (1) = 0.
(a) Finite differences with n divisions, where n is as large as necessary.
(b) Finite elements with n elements, where n is as large as necessary.
(c) Collocation method with Chebycheff polynomials. Use n polynomials, where n is as
large as necessary. Remember to transform the equation first into one with homogeneous
boundary conditions.
192. Consider the problem of pipe flow with a thick wall given in HW16, Prob. 2. Assume a
vanishingly small axial conduction in both wall and fluid, and find the outer solution for the
boundary layer approximation, and the corresponding heat loss to the surroundings.
193. Using a lumped approximation write the differential equation that governs the time dependence
of the temperature of a body that is convecting and radiating to surroundings at T . Find
the steady state temperature and show that it is linearly stable.
194. Find the correlation dimension of the strange attractor for the Lorenz equations (that model
flow in a thermosyphon under some conditions) [Probs 38, 53, 136]
dx
= (y x),
dt
dy
= x y xz,
dt
dz
= bz + xy.
dt
(A.33)
(A.34)
(A.35)
from data obtained from numerical integration of the equation, where = 10, = 28 and
b = 8/3. Use the following definition of the dimension D: vary r and count the number of
points N (r) within a sphere of radius r; the dimension is given by the slope D = ln N/ ln r.
195. Find the linear stability of the critical points of the dynamical system (that also models flow
in a thermosyphon under different conditions)
dx
= y x,
dt
dy
= a zx,
dt
dz
= xy b.
dt
(A.36)
(A.37)
(A.38)
254

Plot (x, y, z) in three dimensional space for the initial condition (1,1,1) and (i) a = 2, b = 1,
(ii) a = 1, b = 1, and (iii) a = 0, b = 1.
196. The dynamical system for an untilted, toroidal thermosyphon with axial conduction and known
heat flux is
dx
= y x,
dt
dy
= zx cy,
dt
dz
= b + xy cz.
dt
(A.39)
(A.40)
(A.41)
Show that (0, ?, ?) is a globally stable critical point for b < c2 .

197. Consider two thermosyphons that can exchange heat through a common wall5 , as shown in
Fig. A.51. There is constant heating through wall ab and cooling through cd. Derive the
time-dependent governing equations for this conjugate problem, and find the steady state
solution(s).
198. A U-shaped open loop with constant heating from the side is rotated about its axis6 as shown
in Fig. A.52. Determine the steady-state flow rate.
199. Consider a PCR (polymerase chain reaction) thermocycler7 as shown in Fig. A.53, that acts
as a single-phase thermosyphon. The triangle is equilateral with a 2 cm base and 0.4 mm
internal diameter. If water is the working fluid, estimate its cycle time (as a number). Assume
a suitable heat transfer correlation.
5 O. Salazar, M. Sen and E. Ramos, Flow in conjugate natural circulation loops, AIAA Journal of Thermophysics
and Heat Transfer, Vol. 2, No. 2, pp. 180183, 1988.
6 M.A. Stremler, D.R. Sawyers, M. Sen, Analysis of natural convection in a rotating open loop, AIAA Journal of
Thermophysics and Heat Transfer, Vol. 8, No. 1, pp. 100106, 1994.
7 N. Agrawal, Y.A. Hassan, V.M. Ugaz, A Pocket-Sized Convective PCR Thermocycler, Angewandte Chemie
International Edition, Vol. 46, No. 23, pp. 43164319, 2007.
255
11
00
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
00
11
55 C
72 C
95 C
256
convection
heating
convection

200. Consider a rotating torus (such as an automobile tire) that is being cooled by natural convection
through hollow spokes as in Fig. A.54 (a single loop is also shown at the side). Estimate the
heat rate that can be extracted by convection through the walls of the spoke.
201. A lumped mass with internal heat generation is radiating heat to its surroundings. Assume
that the instantaneous temperature of the mass can be measured by a sensor, and the heat
generation can be changed by an actuator. PI control is used to maintain the temperature of
the mass at a constant value. Find the conditions on the PI constants for linear stability.
202. Consider a thermostatically-controlled heating model of a ring of n rooms
dTi
+ Ti + k (Ti Ti1 ) + k (Ti Ti+1 ) = Qi ,
dt
i = 1, . . . , n
where Ti is the temperature of room i (with respect to the surroundings), and t is time. The
third and fourth terms on the left are the heat transfer between room i and its neighbors i 1
and i + 1, respectively. The parameter k represents the strength of this interaction. The heater
Qi goes on if the temperature Ti falls below Tmin , and goes off if it rises above Tmax , so that
(
0 if heater is off,
Qi =
1 if heater is on.
Using n = 3, Tmin = 0.25, Tmax = 0.75, T1 (0) = 0.10, T2 (0) = 0.50, T3 (0) = 0.85, write
a compute program to plot T1 (t), T2 (t), T3 (t) on the same graph for 40 t 50 with (a)
k = 0.28, and (b) k = 0.29.
203. The following questions refer to the book S. Chandrasekhar, Hydrodynamic and Hydromagnetic
Stability, Dover, 1961. Consider only the case of fluid between two rigid plates held at different
temperatures. Show all the steps that are not given in the book.
(a) From the governing equations (2), (19) and (39), make suitable assumptions and get the
steady state solutions for the fluid velocity, temperature, and pressure given in Section 9.
(b) From there, carry out the steps in the book to get to Eqs. (99) and (100). Set the
eigenvalue = 0 to get Eqs. (126) and (127).
(c) Complete the solution in Section 15 (solve Eq. (216) numerically) to get the minimum
Rayleigh number for an even mode to be 1707.762 as shown in Eq. (217).
204. The dispersion relation of gravity waves on the surface of water is
p
= gk tanh(kh) ,
257

where h is the depth of water, and g is the acceleration due to gravity. (a) Find the two
approximations for kh 1 (shallow water), and kh 1 (deep water) to leading order. (b)
Determine the phase and group velocities in these two limits. (c) Calculate the speed of a
tsunami (shallow-water wave) in km/hr for an average ocean depth of 3 km. (d) Show that
the group velocity of a deep-water wave is one-half its phase velocity.
205. Analyze the wave motion in an infinite chain of masses connected by identical springs. The
pattern of masses is m1 , m1 , m2 repeating itself.
206. Consider the longitudinal motion of a string of ten identical masses connected to each other
and to walls on either side by identical springs. Before starting, all masses are stationary and
at equilibrium; then only the first mass is pulled a certain distance to one side and let go.
Calculate numerically the resulting longitudinal motion of each mass as a function of time and
plot the result.
207. Derive the Power Flow, Wave Energy and Power Flow, and Momentum and Mass Flow equations outlined in
http://www.silcom.com/~aludwig/Physics/QM/LatticeWaves.htm
208. Using the one-dimensional formula as a guide, derive a formula for solving the steady-state
diffusion equation in two dimensions using the random-walk method.
209. Assuming that the distribution function f depends on time alone, use the relaxation time
approximation to solve the Boltzmann Transport Equation for f (t).
Index
artificial neural networks, 203
microscale heat transfer, 64
boiling, 10
boiling, curve, 11
Neumann solution, 164

nondimensional groups, 8
nucleation
homogeneous, 194
cavity, 175
compressible flow, 197
computational methods
Monte Carlo, 193
condensation, 10
conduction, 3
convection, 7
cooling, 7
correlations, 15, 197
equation
Brinkmans, 177
Darcys, 176
Forchheimers, 176
extended surfaces, 15
Fin, 68
fin, 196
annular, 73
fins
radiation, 168
fouling, 8
phonons, 64
porous media, 176
forced convection, 177
natural convection, 182
stagnation-point flow, 180
thermal wakes, 180
potential flow, 170
radiation, 9, 193, 196
cooling, 22
enclosure, 23
fin, 72
Reynolds number
low, 170
temperatute
bulk, 8
thin films, 66
two-body, 29
two-fluid, 27
genetic algorithms, 203

Goodmans integral, 165
heat exchanger, 11, 14
counterflow, 14
microchannel, 196
plate, 171
least squares, 197
Leveques solution, 170
maldistribution, 196
Maragnoni convection, 175
258

Analytical Heat Transfer

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Analytical Heat Transfer

Uploaded by

Copyright:

Available Formats

ANALYTICAL HEAT TRANSFER

1 Introductory heat transfer

2 Steady and unsteady states

3.3.1 Ordinary differential equations

One spatial dimension

5.2.1 Long time solution . . . . . . . . . . .

6.9.4 Temperature in long duct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

Multiple spatial dimensions

17.6 Radiation effects . . . . . . . . . . .

Introductory heat transfer

Transverse heat flux

where ks is taken to be a constant.

called the latent heat.

Figure 1.2: Energy balance.

Steady-state coduction in a rectangular plate.

Figure 1.3: Wall with fluids on either side.

where T = Tw,1 = Tw,2 .

with T (0) = Ti . We nondimensionalize using

The nondimensional form of the governing equation (1.2) is

Figure 1.5: Convective cooling.

For incompressible flow

Flat-plate boundary-layer theory

Heat transfer coefficients

Overall heat transfer coefficient

Colburn j-factor j = St P r2/3

Integrating over all wavelengths

so that the emission is

The speed of an electromagnetic wave in free space is c = 1/ .

1.5. Boiling and condensation

Wiens law: Putting dE /d = 0, the maximum of is seen to be at = m , where

where = 5.670 108 W/m2 K4 .

Boiling and condensation

Critical heat flux

1.6. Heat exchangers

Table 1.1: Classification of HX (due to Shah [?], 1981

1.6. Heat exchangers

Figure 1.6: Parallel and counter flow.

Parallel- and counter-flow

which can be integrated from 1 to 2 to give

1.6. Heat exchangers

is the logarithmic mean temperature difference.

(Th,i Tc,i ) (Th,o Tc,o )

(Th,i Tc,o ) (Th,o Tc,i )

while for counterflow it is

We an write the element of area dA in terms of the perimeter P as dA = P dx, so that

With the boundary condition q(0) = 0, the solution is

The heat capacity rate ratio is CR = Cmin /Cmax .

1.6. Heat exchangers

(c) Crossflow, both fluids unmixed

Mean temperature-difference method

1.6. Heat exchangers

Figure 1.7: Triangular fin.

Figure 1.8: Constant-area fin.

1.6. Heat exchangers

8. The fin in Problem 3 has a non-uniform diameter of the form

Steady and unsteady states

Electrical analog (steady)

In Fig. 2.1, the resistances are:

Figure 2.1: Electrical analog of steady heat transfer.

Electrical analog (unsteady)

& Sen (2014).

2.5. Variable heat transfer coefficient

Figure 2.3: Temperature measurement compared to exponential.