C/W)
and ConductionConvection Number N
u
Pe
L
5.3 Evaporation Cooling of Gaseous Streams
5.3.1 Cooling of Gaseous Streams
5.3.2 Cooling of Liquid and Gaseous Streams by Wall Seepage
and Surface Evaporation
5.4 Combustion Heating of Gaseous Streams
5.4.1 Conservation Equations, Adiabatic Flame Temperature,
and FarField Thermal Conditions
5.4.2 Preheat and Reaction Regions
5.4.3 Adiabatic Flame Speed u
f .1
and Thickness
5.4.4 Nonadiabatic Flame Speed: Lateral Heat Losses
MassoudKaviany 2011
Contents
5.4.5 Effect of Thermal Conductivity: Effective Thermal
Conductivity
5.4.6 Volumetric Radiation Heat Transfer: Radiant (Photon)
Conductivity k
r
)
5.4.7 Effect of FreeStream Turbulence: Turbulence Intensity Tu
5.5 Joule Heating of Gaseous Streams
5.5.1 Thermal Plasma Generators
5.5.2 Thermal Plasma Classication
5.5.3 IntegralLength Analysis
5.6 GasStream Volumetric Radiation
5.7 Summary
5.8 References
5.9 Problems
6 Convection: SemiBounded Fluid Streams
6.3.1 Turbulent Convection Heat Flux
6.3.3 Variation of Turbulent Mixing Length Near Solid Surface
6.6.1 Theoretical Maximum Heat Flux for Evaporation and
Condensation q
max
(Kinetic Limit)
6.6.3 Dropwise and Film Condensation
6.6.4 ImpingingDroplets Surface Convection with Evaporation
6.8.4 HotWire Anemometry
6.9 SurfaceConvection Evaporation Cooling
6.11 References
6.12.9 SurfaceConvection Evaporation Cooling
6.12.10 General
7 Convection: Bounded Fluid Streams
7.6.6 Dielectric and Inert HeatTransfer Fluids
7.8 References
7.9.7 General
8 Heat Transfer in Thermal Systems
8.1 Primary Thermal Functions
8.1.1 Primary Functions of Heat Transfer Media
8.1.2 Primary Thermal Functions of Bounding Surfaces
8.1.3 Heat Transfer Material
8.2 Thermal Engineering Analysis
8.2.1 Simplications, Approximations, and Assumptions:
Modeling
8.2.2 Nodal Energy Equation
8.2.3 Simultaneous Inclusion of Various Heat Transfer
Mechanisms
8.2.4 Optimization
MassoudKaviany 2011
Contents
8.3 Examples
8.4 Summary
8.5 References
8.6 Problems
Nomenclature
Glossary
Appendix A: Some Thermodynamic Relations
A.1 Simple, Compressible Substance
A.1.1 Internal Energy and Specic Heat Capacity at Constant
Volume
A.1.2 Specic Enthalpy and Specic Heat Capacity at Constant
Pressure
A.2 Phase Change and Heat of Phase Change
A.3 Chemical Reaction and Heat of Reaction
A.4 References
Appendix B: Derivation of DifferentialVolume Energy Equation
B.1 Total Energy Equation
B.2 Mechanical Energy Equation
B.2.1 Mass Conservation Equation
B.2.2 Species Conservation Equation
B.2.3 Momentum Conservation Equation
B.2.4 Mechanical Energy Equation
B.3 Thermal Energy Equation
B.4 Thermal Energy Equation: Enthalpy Formulation
B.5 Thermal Energy Equation: Temperature Formulation
B.6 Conservation Equations in Cartesian and Cylindrical Coordinates
B.6.1 Cartesian Coordinates, x = (x. y. z) and u = (u. :. n)
B.6.2 Cylindrical Coordinates, x = (r. . z) and u = (u
r
. u
. u
z
)
B.7 BoundingSurface Energy Equation with Phase Change
B.8 References
Appendix C: Tables of Thermochemical and Thermophysical Properties
C.1 Tables
Periodic Table and Phase Transitions
Atmospheric Thermophysical Properties
Mass Transfer and Thermochemical Properties of Gaseous Fuels
LiquidGas Surface Tension
Saturated LiquidVapor Properties
C.2 References
MassoudKaviany 2011
1
Introduction and Preliminaries
1.1 Applications and History
1.1.3 History, Frontiers, and Integration
Heat transfer is part of thermal science and engineering and has been developed by
advances made in thermal science and the integration of these advances into the ther
mal engineering practice. The mechanisms of heat transfer, namely, conduction, con
vection, and radiation, as well as fundamentals of thermodynamics, thermal physics
including thermodynamics, thermometry, thermal chemistry, and uid mechanics
form the basis of heat transfer (as was shown in Chart 1.2). Table 1.1 gives a list of
historical contributions to the fundamentals of heat transfer, starting from Galileos
air thermometer [13]. These include the Newton law of convection cooling, the
Fourier law of conduction heat transfer, and the Planck law of blackbody radiation
(emissive power). Recent advances not listed have been in the areas of multiphase
heat transfer media with reaction and phase change. As microelectronics and nano
technology advance, the heat transfer at small scales is examined using the advances
in the microscale theory of thermal transport (solid and uid states, including phase
transformations). The utility of direct thermalelectrical energy conversion has also
resulted in further advances in thermalelectrical (i.e., thermoelectrics) solidstate
physics.
The undergraduate heat transfer books have evolved from McAdams emphasis
on the surfaceconvection heat transfer and its applications in process heat trans
fer (in particular in heat exchangers, with the use of empirical correlations), to the
inclusion of the results of the boundarylayer theory for surfaceconvection heat
transfer. Also included in the evolution are elaborations on the surface and volu
metric radiation heat transfer, and addressing of the liquidgas and solidliquid phase
change. Applications have extended to energy harvesting, utilization, and conserva
tion, as well as to transportation, electronics, and to the particulars in heat transfer
in biological systems, materials processing, and manufacturing.
Access to high speed computers and software has also inuenced the presenta
tion and utility of heat transfer analysis.
MassoudKaviany 2011
Table 1.1. Historical events in the development of heat transfer and its related disciplines
Year/Event Year/Event
1597 Galileos air thermometer
1609 Sanctorius rst clinical thermometer
1648 Pascals measurement of pressure
variation with altitude pressure variation
with altitude
1660 Boyles law for isothermal gas
1684 Leibnitzs rst article on calculus
1687 Newtons law of viscosity
1690 Papins piston steam engine
1698 Saverys steam engine
1701 Newtons law of surface convection
1703 Amontons use of the boiling point of
water as a xed point of thermometer
1712 Newcomens atmospheric pressure steam
engine
1714 Fahrenheits thermometer
1738 Bernoullis publication of Hydrodynamica
1742 Celsius centigrade scale
1756 Leidenfrosts observation of the boiling
curve
1757 Blacks distinctions of three modes of
heat transfer (conduction, convection,
and radiation)
1765 Watts steam engine
1774 Blacks measurement of latent heat of ice
melting
1788 Lagranges publication of M ecanique
Analytique
1798 Rumfords description of the conversion
of frictional work to heat
1799 Introduction of the Metric System
1800 Continuous electric current
1803 Daltons atomic theory
1811 Fouriers law of heat conduction,
Avogadros denition of mole
1818 DulongPetits law of molecular heat
1821 Seebecks static electronic potential
between two junctions of dissimial metals
at different temperatures
1822 NavierStokes momentum conservation
equation
1824 Carnot cycle
1826 Amp` eres publication of conductivity of
nonmetal solids, Electrodynamics
1827 Ohms law of electrical conduction
1829 P eclets uid conductionconduction heat
transfer ratio
1834 Clapeyrons equation of state for a perfect
gas
1835 Poissons measurement of unsteady
temperature distribution in a bar
1842 Mayers energy conservation law (rst law
of thermodynamics)
1843 Joules measurement of the mechanical
equivalent of heat
1847 Helmholtzs formulation of the energy
conservation law
1848 Kelvins concept of absolute temperature
1859 Kirchhoffs law of radiation, Maxwells
electromagnetic equations
1860 Maxwells thermaluctuation energy
distribution law
1873 Stefans solidliquid phase change
analysis, van der Walls equation of state
1875 Grashofs equation for heat exchange
1876 Gibbs chemical thermodynamics theory
1879 Introduction of Reynolds number
1879 Oberbecks analysis of thermobuoyant
ow and heat transfer
1881 Lorentzs relation between electrical and
thermal conductivity
1883 Graetzs analysis of ow and heat transfer
in a tube
1884 Boltzmanns derivation of blackbody
emissive power
1893 Wiens displacement law of thermal
radiation
1894 Reynolds equation for turbulent ow
1900 Plancks radiation law for photon
emission and quantum theory, Drudes
electrongas theory of conductivity of
metals
1902 Gibbs statistical mechanics published
1904 Prandtls boundarylayer theory
1906 Nernsts heat theorem (third law of
thermodynamics)
1911 Einsteins microscale thermal conductivity
model used for amorphous dielectrics
1914 Debyes theory of phonons
1914 Debyes theory of phonons for thermal
conductivity of nonmetal solids, Millikans
photoelectric experiment verifying
Plancks constant
1915 Nusselts basic law of heat transfer rate
1925 Prandtls turbulent mixing length theory
1928 Blochs theory of quantum statistical
mechanics of electrons in metallic crystal
1932 Jakobs description of heat transfer in
boiling
1933 McAdams heat transfer textbook
1934 Nukiyamas description of boiling curve
1948 Kutateladzes description of transition
boiling
1959 Zubers hydrodynamic theory for critical
heat ux, Callaways model for phonon
thermal conductivity.
MassoudKaviany 2011
EXAMPLE 1.1. FAM
The universe (as a large thermal system with many subsystems) and its dynam
ics offer a large range of temperature, length, time, and mass scales to con
sider. The socalled natural (or Planck) scales of length, time, and mass are
L
P
= h
P
,(2m
P
c
o
) = (h
P
G
N
)
1,2
,(2c
3
o
)
1,2
= 1.61605 10
35
m, t
P
= L
P
c
o
=
[(h
P
G
N
),c
5
o
]
1,2
= 5.39056 10
44
s, and m
P
= [(h
P
c
o
),(2G
N
)]
1,2
= 2.17671
10
8
kg, where G
N
is the Newton constant of gravity, G
N
= 6.67259
10
11
m
3
/kgs
2
, c
o
is the speed of light in vacuum, c
o
= 2.99792458 10
8
m/s,
and h
P
is the Planck constant, h
P
= 6.626075 10
34
kgm
2
/s (or Js).
10
6
1
10
6
10
12
10
18
10
24
10
30
t , s
10
33
10
43
Approaching 0 K
Approaching 0 s
10
23
D
o
m
i
n
a
t
e
d
b
y
Q
u
a
n
t
u
m
G
r
a
v
i
t
y
D
o
m
i
n
a
t
e
d
b
y
G
r
a
n
d
U
n
i
f
i
e
d
F
o
r
c
e
Prediction
Gravitational Force Freezes
Extreme in
High Temperature
Extreme in
Low Temperature
Neutrons and Protons Form
Helium Nuclei Form
Neutral Atoms Form
D
o
m
i
n
a
t
e
d
b
y
E
l
e
c
t
r
o
w
e
a
k
F
o
r
c
e
D
o
m
i
n
a
t
e
d
b
y
E
l
e
c
t
r
i
c
F
o
r
c
e
10
13
10
3
10
7
10
17
(100,000 yr)
(Today,
15 Billion yr)
(1 yr)
T
,
K
Figure Ex. 1.1. Predicted average temperature T) = T)
V
of the universe from the big
bang event to present. Human beings are estimated to have appeared about 400,000 years
(10
13
s) before today.
Our knowledge of the temperature history of the universe is rather incom
plete. Based on existing theories, during the periods marked by very short
elapsed times (measured from the time of the big bang), the universe was very
small [8]. The predicted (with many simplifying assumptions) volumeaverage
temperature of the universe T)
V
(t) is depicted in Figure Ex. 1.1. The four fun
damental forces are shown with periods over which they dominated. The elec
tromagnetic force keeps atoms together and is the basis of all chemical reactions.
The strong nuclear force binds neutrons and protons together in the nucleus and
governs nuclear reactions such as ssion and fusion. The weak nuclear force
determines the radioactive decay leading to emission of alpha and beta particles
from the nucleus. The gravitational force is responsible for largescale structures
such as the universe and inuences our lives on the earth.
Based on the theory used, the initial temperature of the universe (occurring
in a volume no smaller in linear dimension than L
P
and no earlier in time than
MassoudKaviany 2011
t
P
, and having energy larger than E
P
) was extraordinarily large. But, the volume
expansion has also been so intense that the universe has cooled drastically. The
relationship between time and average temperature is approximated (using the
same relation for all eras) as [8]
t
2
t
1
=
1
2
_
3
8G
N
_
1,2
_
1
_
1.22 10
32
T)
4
V.2
_
1,2
1
_
1.22 10
32
T)
4
V.1
_
1,2
_
.
This is based on conserved mass in a sphere as density changes as )
V
= 1.22
10
32
(kg/m
3
K
4
)T)
4
V
(K
4
).
The volumeaveraged (over the entire universe) temperature T)
V
(K)
presently is only about 2.736 K (measured from the background cosmic emis
sion). This temperature is used in radiation heat transfer analysis when extrater
restrial bodies are exposed to deep space. The highest temperature that the
universe has experienced is at the elapsed time of about 10
43
s (Planck time
scale) at the time of the bigbang event. The theory predicts this temperature to
be over 10
32
K. After a year, the average temperature had dropped to 10
6
K.
The expansion of the universe occurs as the total energy remains constant,
while the potential and kinetic energy (and the volume of the universe) change
with time.
EXAMPLE 1.2. FAM
Using the earth as a reference, the natural systems can be classied as terrestrial
or extraterrestrial. As an example of a terrestrial system, consider the temper
ature variations within and around the earth. The surface and timeaveraged
surface temperature for the earth surface T)
A
is about 15
C.
In the solid phase, heat is transferred by conduction, while in the liquid
phase heat is transferred by conduction and convection. The heat is generated
volumetrically by radioactive decay (an atomic or molecularbond energy con
version) mostly in the crust and in the mantle and is designated by s
r.
(W/m
3
).
The existence of the hot geotherms makes for a higher local temperature, com
pared to the regions with the cold geotherms. At the surface of the earth, heat
ows from within the earth by conduction. The amount of heat ow is estimated
as Q
k
= 4 10
13
W [4]. The amount of heat ow per unit surface area is desig
nated by the heat ux q
k
, and is given by q
k
= Q
k
,A
k
, A
k
= 4R
2
. The heat ux
is q
k
= 0.078 W/m
2
, which is much less than that arriving at the surface by the
solar irradiation (which is discussed below and is q
r.i
= 172.4 W/m
2
).
The radiation from the sun and the other parts of the universe impinges
upon (called irradiation) the earth. The solar irradiation (q
r.i
)
s
just outside the
earth atmosphere is (q
r.i
)
s
= 1,353 W/m
2
. The sun has an average radius of
MassoudKaviany 2011
T
A
,
o
C
S
e, S
e,
0 200
100
15
o
C
400 600
50
150
200
T
A
p
10
8
10
6
10
4
10
2
1
Stratosphere
Ozone Layer
Surface Radiation
Emission
A
i
r
p
l
a
n
e
s
F
l
y
(
1
0
k
m
)
S
p
a
c
e
S
h
u
t
t
l
e
s
F
l
y
(
3
8
0
k
m
)
Mesosphere
7% of Irradiation
Reflected
Thermosphere
Radio ( > 1 m)
Infrared (1000 > > 0.77
m)
Ultraviolet (390 > > 10 nm)
XRays (10 > > 0.001 nm)
Cosmic and Gamma Rays ( < 0.001 nm)
Visible Light (0.77 > > 0.39
m)
14% of Irradiation
Absorbed by Entire
Atmosphere (500 km)
Heat Transfer by
Electromagnetic Radiation
(Solar Irradiation)
(b) Irradiation, Temperature, and Pressure Distribution in the Earth Atmosphere
24% of Irradiation
Reflected by Clouds
4% of Irradiation Reflect
ed by Oceans and Land
51% of Irradiation
Absorbed by Oceans
and Land
p, atm
(1 atm = 1.013 x 10
5
Pa)
q
r,
r, km
Log Scale
Linear Scale
q
r,i
0
s
e,
q
r
s
e,
Troposphere
Figure Ex. 1.2. (b) Temperature and pressure distributions within the earths gaseous
atmosphere. The solar radiation impinging on the earth surface (i.e., irradiation) and
radiation emitted from the earth surface, are also shown.
6.75 10
5
km and is at an average distance of 1.5 10
8
km from the
earth. The average (time and spaceaveraged) solar irradiation is about
(q
r.i
)
s
(R
2
,4R
2
) = 338.3 W/m
2
, where R
2
is the earth projected area and
4R
2
is the surface area. This radiation passes in part through the earth atmo
sphere (i.e., is transmitted) and is in part absorbed or reected. About 51% of
this mostly solar radiation coming toward the earth (i.e., owing through the
projected surface area of the earth) arrives on the surface of the earth, as shown
in Figure Ex. 1.2(b). This is an average of (q
r.i
)
s
)
A
= 172.4 W/m
2
arriving at the
earth surface. We will discuss solar irradiation in Section 4.5.2.
This energy is emitted from the earth surface (solid and liquid) as infrared
radiation and results in a steady temperature. The greenhouse effect prevents
some of this infrared radiation from leaving the earth atmosphere. The earth
atmosphere (about 500 km thick) absorbs all or part of the short wavelength
radiation (cosmic and gamma rays, Xrays, and ultraviolet). The longer wave
lengths (visible, infrared, and radio) are mostly transmitted. The temperature
of the earth gaseous atmosphere initially undergoes a few minima, as the dis
tance from the surface of the earth increases, and then begins to rise. The region
with increasing temperature with increase in r, is called the thermosphere. The
pressure decreases monotonically with the distance from the earth surface and
MassoudKaviany 2011
1,200 1,000 1,400
t, yr
Measured
Approximate
Trend Line
C
O
2
M
o
l
e
F
r
a
c
t
i
o
n
,
p
p
m
1,600 1,800
280
320
360
2,000
(c) Measured CO
2
History
Figure Ex. 1.2. (c) Time variation of the atmospheric (the earth) CO
2
mole fraction (in
parts per million) [4].
nearly vanishes in the thermoshpere. Surface radiation emission occurs from the
earth surface and is designated as s
e.c
in Figure Ex. 1.2(b).
The trace species (such as H
2
O and CO
2
), as compared to the abundant
species (N
2
and O
2
), are strong infrared radiation absorbers and inuence the
temperature of the earth surface. These are called the greenhouse gases and
these gases absorb some of the radiant heat emitted from the earth surface
q
r.c
, which results in the earth being maintained warmer. The current average
mole fraction of CO
2
in the atmosphere is 367 parts per million (ppm). This is
the CO
2
mole fraction multiplied by 10
6
, and may be rising due to production
of CO
2
in combustion (about 60% from automobiles) (causing an increase in
the earth temperature). Figure Ex. 1.2(c) shows the variation of atmospheric
0.8
0.4
0.0
 0.4
 0.8
1860 1880 1900 1920 1940 1960 1980 2000
t, yr
<
T
>
A
(
r
=
0
,
t
)
<
T
>
A
(
r
=
0
,
1
9
6
1
<
t
<
1
9
9
0
y
r
)
,
o
C
Measured
}
(d) Measured Surface Temperature History
Figure Ex. 1.2. (d) Change in the average earth surface temperatures <T>
A
for the past
140 years [4]. The change is shown with respect to the average taken between 1961 and
1990.
MassoudKaviany 2011
CO
2
mole fraction with respect to time. Date before 1958 is from air bubbles
trapped inice layers. The atmospheric thermophysical properties, including pres
sure p(Pa), which decreases exponentially with distance r(m), density (kg/m
3
),
gravitational constant g(m/s
2
), and the molecular meanfree path
f
(m), are
listed in Table C.7, Appendix C, as a function of the distance from the earth
surface r.
Figure Ex. 1.2(d) shows the increase in the average surface temperature
(suggested to be related to increasing CO
2
mentioned above), over the past
140 years (reported in year 2000). The results are from three independent mea
surements, but all supporting the same trends [4]. Note the large rate of increase
in recent years.
EXAMPLE 1.3. FAM
As another example of a natural thermal system, consider the human body
temperature regulation (thermophysiology). The average internal temperature
of the human body (core temperature) is maintained to within 0.5
C around
the average temperature of 37
S
lg
=
M
lg
Lh
lg
[6, 16, 20], where
M
lg
(kg/s) is the evaporationrate andLh
lg
(J/kg)
is the heat of evaporation for water. Figure Ex. 1.3(a) renders these mechanisms
and gives the fraction of surface heat loss (Q
r
, Q
k
, and Q
ku
) and surface energy
conversion from an inactive nude body to an ambient at normal ambient condi
tions (at T
s.
= T
f .
= T
= 20
C or rises above 55
C. The shivering
setpoint and the sweating setpoint (onset of sweating as T
b
elevates) are also
shown. The average basal heat production
S
r.c
is about 20 cal/s (or 72 kcal/hr),
or 83.74 J/s (1 cal = 4.1868 J), or 83.74 W. An average human has surface area
of nearly 1.7 m
2
.
Figure Ex. 1.3(d) shows some typical metabolic heat production rates
S
r
per unit surface area A, during typical human activities. During very strenuous
activities, up to 600 W/m
2
of heat is generated [18].
EXAMPLE 1.4. FAM
As a more complex example of energy conversion involving heat transfer, con
sider the heat transfer aspects of the pistoncylinder in a sparkignition, internal
combustion engine. This may be considered a thermal actuation example with
the gas temperature raised using a chemicalband energy conversion
S
r.c
. Figure
Ex. 1.4 shows typical temperatures occurring at various locations in the cylinder,
piston, valves, and the surrounding casing, and in the uids of such an engine
[16].
Also shown are the cycleaveraged temperatures designated as T (overbar
indicates time average). The cycleaveraged energy conversion is by chemical
reaction
S
r.c
(W) and the fraction of this energy converted to thermal energy
MassoudKaviany 2011
W
m
= S
m,p
= (0.25 to 0.40)S
r, c
Shaft Work
Coolant, T = 105
o
C
Cylinder Wall, T = 185
o
C
Intake Manifold, T = 60
o
C
Exhaust Flow, T = 450
o
C
Typical Temperature s
Heat Transfer and
Energy Conservation
Exhaust Valve, T = 650
o
C
Piston Face, T = 300
o
C
Piston Ring, T = 220
o
C
Piston Skirt, T = 190
o
C
Spark Plug, T = 600
o
C
Oil, T = 70
o
C
Intake Valve, T = 250
o
C
S
r, c
, Energy Conversion
by Combustion
Q
oil
= (0.05 to 0.15)S
r, c
Heat Transfer to Oi l
Q
ambient
= (0.02 to 0.10)S
r, c
Heat Transfer to Ambient
Q
exhaust
= (0.20 to 0.45)S
r, c
Heat Transfer through Exhaust
Q
coolant
= (0.10 to 0.30)S
r, c
Heat Transfer to Coolant
Figure Ex. 1.4. Pistoncylinder heat transfer and typical cycleaveraged temperatures in
sparkignition internal combustion engine. The fraction of heat generated by reaction
and converted to various other forms of energy are also shown.
leaves as heat transfer through the exhaust Q
exhaust
, coolant Q
coolant
, ambient
Q
ambient
, and oil Q
oil
, and as shaft work
W
m
=
S
m.p
. Typical efciency
W
m
,
S
r.c
is approximately 0.25 to 0.40 and is dependent on the engine speed (revolutions
per minute, rpm), among other parameters.
We will discuss a detail thermal analysis of an internal combustion engine,
with and added insitu regenerator, in Example 8.5 of Chapter 8.
1.5 Heat Transfer Materials, and Heat Flux Tracking
1.5.1 Three Phases of Matter: Intermolecular and Intramolecular Forces
Matter consists of localized regions of very high density (i.e., atoms and molecules)
separated by regions of nearly zero density. The separation distance is determined by
the attraction and repulsion forces (bonding forces) between atoms and molecules
(i.e., particles). These forces are due to ionic bonds, covalent bonds, and van der
Waals forces (dipoledipole, dipoleinduced, and dispersive forces). All properties
of matter are determined by these forces.
The interaction between constituent particles (i.e., molecules) ranges from no
interaction for ideal gases to very strong interaction for metallic solids. The inter
molecular attraction is due to the shortrange electrical forces, called the dipole
dipole attraction. For polar molecules (such as HCl), in molecules with strong hydro
gen bond (such as H
2
O), and in ionic molecules (such as NaCl), these dipoledipole
forces of attraction are very strong. In nonpolar molecules (such as CH
4
), there is a
weaker attraction (called the London forces), due to the induced dipole caused by
MassoudKaviany 2011
the continuous motion of the electrons. The most detailed account of the interac
tion is given by the quantummechanical interaction among the particles. The next
level in simplicity is the classical mechanics description such as hardspheres mod
els. There is also a phenomenological description based on intermolecularpotential
models. This treats particles (i.e., atoms and molecules) as point masses with inter
acting forces that depend on the separation distance between these particles. The
simplest model is the twobody LennardJones intermolecular potential. It is used to
describe properties (thermodynamic and transport) of gases, liquids, and solids that
are dominated by the van der Waals forces. This potential has a weak, longrange
attractive force that depends on the separation distance r(m) as r
6
, a negative
well of depth LE
LJ
(J), and a steep repulsion force as r
12
, for distances less than
r
o
(m). Typically, r
o
is of the order one tenth of an
A (
A = 10
10
m) for liquids and
solids, and LE
LJ
is of the order of 10
21
J. Each molecule is located at the potential
well minimum and vibrates with a small amplitude around this equilibrium position.
The intermolecular force is d(LE
LJ
),dr.
In gaseous state, atoms and molecules (uid particles) have a rootmeansquare
(rms) speed between collisions (u
2
f
)
1,2
(m/s) which is also called the rms thermal
speed. The time average of this velocity u
f
, over many collisions, is zero. The uid
particles have an average kinetic energy E
f .k
and an average interparticle potential
(or dissociation) energy LE
R
g
M
T
_
1,2
average thermal speed of ideal gas (1.16)
N
A
m = M. R
g
= N
A
k
B
= 8.3145 10
3
J/kmoleK
universal gas
constant,
(1.17)
where k
B
(J/K) is the Boltzmann constant, N
A
(number of molecules per kmole) is
the Avogadro number, M(kg/kmole) is the molecular weight, R
g
(J/kmoleK) is the
(J) is required. Then as T increases, the required kinetic energy will reach
and surpass the dissociation energy and the particles will be separated. At lower
temperature, while the temperature is raised, rst solid, then liquid, and nally gas
phases are observed. Yet at higher temperature ionization and other dissociations
will occur (leading to the plasma state).
Whena particle of kinetic energy 3k
B
T,2(J) collides withanassociated(or bond)
pair, and when this energy is (roughly) larger than LE
,
no dissociation will occur. The clusters become connected and larger conglomerates
will formin which the particles are tightly connected. This is a constrained (i.e., rigid)
structure and is a property of the solid phase.
The isothermal compressibility (1/Pa) is dened as
1
:
:
p
[
T
=
1
p
[
T
1
E
p
isothermal compressibility and bulk modulus E
p
,
E
p
=
E
s
[3(1
P
)]
isothermal solids, (1.18)
where E
s
is the Young modulus and
p
is the Poisson ratio.
The magnitude of decreases as the kinetic energy decreases or as LE
increases,
i.e., as more clusters or ordered structures are formed, decreases. For an ideal gas,
= 1,p, and therefore, is larger at lower pressures. The isothermal compressibility
for gases is several orders of magnitude larger than that of the liquids, which in turn
have an isothermal compressibility a few orders of magnitude larger than the solids.
MassoudKaviany 2011
The compressibility is a manifestation of the average spacing between the particle
centers l
m
.
For gases, the meanfree path
f
(m) is dened as the distance traveled by parti
cles over a time interval Lt(s), u
f
Lt(m), divided by the number of collisions experi
enced over Lt, and is given by
f
=
1
2
1,2
1
d
2
m
n
=
1
2
1,2
k
B
T
d
2
m
p
. p =
R
g
M
T.
n =
N
A
M
=
p
k
B
T
meanfree path
for ideal gas,
(1.19)
where n(1,m
3
) is the number of atoms per unit volume, and d
m
(m) is the collision
diameter.
For example, d
m
is 2.74
A for hydrogen, 3.68
A for nitrogen, 4.60
A for water
vapor, 4.59
A for carbon dioxide, 3.61
A for oxygen, 4.14
A for methane, and 3.50
A
for air. Then at standard atmospheric temperature and pressure, from(1.19) we have
the meanfree path
f
as high as
f
= 117.7 nm for H
2
, as low as
f
= 41.8 nm for
water vapor, and
f
= 66.3 nm for air.
The meanfree path for atmospheric air as a function of altitude is given in
Table C.7 in Appendix C. Note that the average intermolecular spacing l
m
is the
cube root of 1,n, i.e., l
m
= (1,n)
1,3
or l
m
= (k
B
T,p)
1,3
(m). Then for air at atmo
spheric standard temperature and pressure (STP) this spacing l
m
is 3.45 nm (34.5
f
. The ratio of
f
to L is called the Knudsen number Kn
L
. Therefore we make a
continuum treatment (called viscousow regime) of a gas when
Kn
L
=
f
L
 0.1
Knudsen number, for continuum
(viscousow regime) treatment,
(1.20)
where subscript L indicates that the Knudsen number is based on L. For Kn
L
> 10,
we have a molecularow regime and direct interaction of the uid particles with
the bounding surfaces should be included. The regime 0.1 Kn
L
10 is called the
transitional regime and both viscous and molecularow behavious exist.
From the mechanical properties viewpoint, the solids are divided into metals,
ceramics (oxides and nonoxides), polymers, and composites (medium made of two
or more distinct materials). Fromthe electrical properties view, the solids are divided
into metal (or conductors), nonmetals (or insulators or dielectrics), semimetals, and
semiconductors. (However, from the solidstate physics viewpoint the division is
metals, nonmetals, and semiconductors). From the heat transfer view, the solids are
divided into conductors, insulators, and composites. The thermal and electrical prop
erties of solids will be discussed in Sections 3.1 and 3.2. The liquids are generally
divided into metals and nonmetals. Liquids metals have very high electrical and
MassoudKaviany 2011
thermal conductivity, compared to nonmetals. Nonmetal liquids are divided into
organic and nonorganic. The gases are divided as condensables, when in the tem
perature range of application they can coexist with their liquid or solid phase, and
noncondensables (or inert). The periodic table, including the STP phase state of the
elements and some of their physical properties, is listed as Table C.2, in Appendix C.
1.5.2 Microscale Energies of Matter: Discrete and Continuous Energies
Heat transfer is the result of spatial variation of temperature caused by a change
in the local energy states. This change in energy can be due to conversion of bond
(chemical and physical) energy, or electromagnetic energy, or mechanical energy, or
a combination of these. The energy of matter is related to its molecular structure
which allows for discrete energy states and these atomic (or particle) energies are
presented in eV units. The product k
B
T(J) is also given in units of electronvolt,
eV. This is a unit of energy when an electron is moved through a difference in
electric potential of one volt. Since the charge of an electron is e
c
= 1.6022 10
19
C, one eV is 1 eV = 1.6022 10
19
J. At room temperature, T = 298 K, k
B
T =
0.0257 eV. Particles with energies greater than 10 eV (i.e., T = 116,000 K), are
considered to be high energy and can overcome nearly all energy bonds (or potential
gaps), and therefore, become free. As a reference, a photon with a wavelength of
0.5 jm (green light) has energy E
ph
= h
P
f = 2.48 eV. Note that eV is used for
energy per microscale energy carrier, as compared to for example J/kg, which is
used per kg mass. Now to compare these with the energy in matter, for ice (H
2
O)
at T = 273 K, the dominant latticevibration energy (i.e., phonon energy h
P
f ) per
atom is 0.54 10
2
eV, hydrogenbond breakage (ionization) energy per molecule is
Lh
ij
= 0.58 eV, heat of combustionof methane CH
4
per molecule is Lh
r.c
= 10.36 eV,
and typical valenceelectron bond breakage energy per electron Lh
ij
is 10 eV. These
energies can be compared to phasechange energies; for water, the enthalpy of
melting per molecule Lh
sl
is 0.062 eV, the enthalpy of evaporation per molecule
Lh
lg
(at 1 atm) is 0.39 eV, and the enthalpy of sublimation per molecule Lh
sg
is
0.49 eV [12]. Therefore, at room temperature, a much smaller energy is needed to
cause a phase change (e.g., melting), compared to that needed for ionization. When
interacting, there may be one microscale heat carrier (single) or more (multiple)
required to supply the needed energy for conversion. For example, several photons
may be required to provide an electronic gap transition.
As compared to the single, reference latent heat or heat of reaction of molecules,
each molecule may have many electronic and vibration (and rotational and trans
lational) energies. The number of such energy states depends on the molecular
structure. Some of the vibrational energies (including those in crystals, i.e., phonons)
can be very small, but the population density of these states may be large (deter
mined from statistical classical and quantum mechanics). We will give examples of
these distribution functions for uid particles, electrons, phonons and photons.
Figure 1.7 gives typical values of the discrete h
P
f and continuous thermal k
B
T
energies, associated with various atomic (or molecular) energy phenomena. The
MassoudKaviany 2011
thermal energy k
B
T can be used to gauge other energies. For example (covalent)
bond energies are generally between 100 to 300k
B
T. The lowest energies are asso
ciated with rotational and some electronic energy state transitions, while the large
energies are associated with nucleus ssion and fusion. The electromagnetic spec
trum (e.g., microwave, infrared, visible, ultraviolet) will be discussed in Chapter 4. In
Figure 1.7, the smaller transition energies are associated with large quantum num
bers, and there the quantum and classical models yield the same continuous results
(e.g., vibrational, translational and rotational). As the transitional energy increases,
the difference becomes clear, for example in the electron energy gap which is pre
dicted only by the quantum mechanics.
All the energy of matter is divided into sensible heat energy and energy that
can be converted to and from heat. In Chapter 2, we will introduce the various
energy conversion mechanisms
S
i
(chemical and physicalbond, electromagnetic,
and mechanical) of the heat transfer media. Chapter 3 reviews the contributions to
the specic heat capacity (energy storage) c
:
and c
p
.
This molecular view of energy is combined with the molecular or microscale of
thermal energy transport discussed in Section 1.4.2 to give an unied view of physics
of heat transfer. The thermal conductivity k of the various phases are described
by transport of the freemolecule, electron, and phonon microscale heat carriers.
In Chapter 4, we will discuss photons and the radiation properties of heat transfer
media and surfaces. In Chapters 5 to 7, we will examine convection heat carriers in
laminar and turbulent ows.
1.7 Conservation of Mass, Species, and Momentum
In a manner similar to that used for the energy equation, we write belowthe integral
volume conservation equations for the total mass M, species i mass M
i
, and momen
tum u of uid owing through a control volume.
1.7.1 Conservation of Mass
The law of conservation of mass states that the rate of change of mass within the
control volume M(kg), i.e., M,t(kg/s), is equal to the net rate of ow of mass
M[
A
(kg/s) through (i.e., out of) the control surface. This gives the mass conservation
equation as
M
t
A
= 0
1 2
(1.24)
integralvolume mass conservation equation,
where
1 net rate of mass storage in the control volume
2 net rate of mass ow out of the control surface.
MassoudKaviany 2011
0
Microwave
Far
Infrared
Near
Infrared
Visible
Ultraviolet XRays Gamma Rays
10
4
10
2
1 10
2
10
4
10
8
10
6
f = 2.41 x 10
10
Hz 2.41 x 10
12
2.41 x 10
14
2.41 x 10
16
2.41 x 10
18
2.41 x 10
20
2.41 x 10
22
0
.0
0
1
, T
y
p
i
c
a
l
R
o
t
a
t
i
o
n
a
l
E
n
e
r
g
y
0
.0
2
7
, H
i
g
h
P
h
o
n
o
n
E
n
e
r
g
y
o
f
C
u
0
.0
6
2
, M
e
l
t
i
n
g
I
c
e
H2
O
, o
p
t
i
c
a
l
p
h
o
n
o
n
o
f
S
i
0
.1
, T
y
p
i
c
a
l
V
i
b
r
a
t
i
o
n
a
l
E
n
e
r
g
y
0
.1
3
, E
l
e
c
t
r
o
n
E
n
e
r
g
y
G
a
p
f
o
r
p

t
y
p
e
B
i 2
T
e 3
0
.2
5
, P
h
o
t
o
n
E
n
e
r
g
y
f
o
r
=
5
m
0
.4
2
, E
v
a
p
o
r
a
t
i
o
n
o
f
H2
O
1
.1
2
, E
l
e
c
t
r
o
n
E
n
e
r
g
y
G
a
p
o
f
S
i
2
.0
, F
e
r
m
i
E
l
e
c
t
r
o
n
E
n
e
r
g
y
o
f
K
2
.5
, P
h
o
t
o
n
E
n
e
r
g
y
f
o
r
G
r
e
e
n
L
i
g
h
t
=
0
.5
m
7
.1
, F
e
r
m
i
E
l
e
c
t
r
o
n
E
n
e
r
g
y
o
f
C
u
1
3
.6
, E
l
e
c
t
r
o
n
B
o
n
d
E
n
e
r
g
y
o
f
H
1
0
6
,
M
e
g
a
t
r
o
n
E
l
e
c
t
r
o
n
A
c
c
e
l
e
r
a
t
i
o
n
Continuous
Energy
Discrete
Energy
1
0
4
, F
i
n
e
E
l
e
c
t
r
o
n
i
c
E
n
e
r
g
y
T
r
a
n
s
i
t
i
o
n
s
0
.1
9
4
, V
i
b
r
a
t
i
o
n
a
l
E
n
e
r
g
y
o
f
O
2 , a
l
s
o
T
=
2
,2
5
0
K
0
.0
5
, E
n
e
r
g
y
o
f
T
w
o
C
o
l
l
i
d
i
n
g
M
o
l
e
c
u
l
e
s
a
t
T
=
2
9
8
K
T
=
1
1
,6
0
0
K
E
i
n
s
t
e
i
n
T
e
m
p
e
r
a
t
u
r
e
f
o
r
G
r
e
e
n
L
i
g
h
t
(
=
0
.5
m
)
T
=
6
0
,0
0
0
K
, B
o
n
d
E
n
e
r
g
y
o
f
O
2
8
,6
1
3
, T
=
1
0 8
K
h
P
f , eV
k
B
T , eV
Ionization Energy in Radioactive
Decay, Fission, and Fusion Systems
High Energy Low Energy
h
P
f h
P
c
o
k
B
k
B
T
E
= =
T
E
= 2.877 x 10
4
K
10
3
x
1
0
5
,
E
l
e
c
t
r
o
n
i
n
C
a
t
h
o
d
e
R
a
y

T
u
b
e
1
.4
x
1
0
8
,
E
l
e
c
t
r
o
n
E
m
i
s
s
i
o
n
i
n
R
a
d
i
o
a
c
t
iv
e
D
e
c
a
y
o
f
K
4
.2
x
1
0
8
,
A
l
p
h
a
P
a
r
t
i
c
l
e
E
m
i
s
s
i
o
n
i
n
R
a
d
i
o
a
c
t
iv
e
D
e
c
a
y
o
f
U
2
3
8
4
.2
, D
i
s
s
o
c
i
a
t
i
o
n
o
f
N
a
C
l
Translation
E
f
= k
B
T
3
2
Vibration E
f
(T)
I
n
f
r
a
r
e
d
V
i
s
i
b
l
e
X

R
a
y
E
p
h
=
h
P
f p
h
Nucleus
Electron Orbit E
e
e

Electron
Interatomic Bond
E
r, Internuclear Distance
E
e
E
f
Rotation E
f
(T)
1
.
8
x
1
0 4
,
R
o
t
a
t
i
o
n
a
l
E
n
e
r
g
y
o
f
O
2 ,
a
l
s
o
T
=
2
.
0
7
K
0
.
0
0
2
6
S
e
n
s
i
b
l
e
H
e
a
t
o
f
H
2 O
(
G
a
s
)
,
T
=
1
0
K
0
.
0
1
S
e
n
s
i
b
l
e
H
e
a
t
o
f
H
2 O
(
L
i
q
u
i
d
)
,
T
=
1
0
K
0
.
0
2
7
,
D
e
b
y
e
T
e
m
p
e
r
a
t
u
r
e
f
o
r
C
u
0
.
0
2
6
,
T
=
2
9
8
K
h
P
f
D
k
B
T
D
=
T
D
= 315 K
1
.7
x
1
0
7
,
R
e
l
e
a
s
e
d
E
n
e
r
g
y
i
n
F
u
s
i
o
n
o
f
D
e
u
t
e
r
i
u
m
a
n
d
T
r
i
t
i
u
m
Figure 1.7. Typical magnitudes of the discrete h
P
f (top) and continuous thermal k
B
T (bottom) energy in various atomic energy levels and phenomena.
The photon energy regimes are also shown. The proton emission can be due to molecular motion or electron and nucleus dissociation (shown left).
M
a
s
s
o
u
d
K
a
v
i
a
n
y
2
0
1
1
Similar to (1.23), we rearrange this and write
M[
A
in terms of the mass ux
vector as
A
_
A
( m s
n
) dA =
_
A
(u s
n
) dA =
M
t
V
=
t
_
V
dV
1 2
(1.25)
integralvolume mass conservation equation,
where
1 net rate of mass ow out of the control surface
2 negative of rate of mass storage within the control volume.
Here we have used the mass ux vector m(kg/m
2
s) = u, in a similar manner
as the heat ux vector q(J/m
2
s). The mass ux vector is the product of the density
and the velocity vector. The mass conservation equation is also called the continuity
equation.
1.7.2 Conservation of Species
The species conservation principle, similar to the conservation of energy and total
mass, states that in a control volume the time rate of change of the mass of species A
(one of the species in a mixture made of n species) M
A
(kg), is equal to the sum of the
negative of the net ow of species i through the control surface [which is the surface
integral of the species A mass ux vector m
A
(kg/s)] and the rate at which the species
is formed within the control volume
N
A
[
V
(which is the volume integral of volumetric
production rate) n
A
(kg/m
3
s). This gives the species conservation equation as
M
A
_
A
( m
A
s
n
) dA=
M
A
t
N
A
[
V
t
_
V
A
dV
_
n
A
dV
1 2 3
(1.26)
m
A
= m
D,A
m
u,A
integralvolume species A conservation equation,
where
1 net rate of ow of species A out of the control surface
2 negative of rate of storage of species A within the control volume
3 rate of production of species A within the control volume.
The species A mass ux vector m
A
, dened in (1.26), has a diffusion mechanism
(similar to conduction) and a convection mechanism. These are further discussed in
Appendix B, Section B.2.2. The summation of (1.26), over all species A, gives the
total mass conservation equation (1.25).
MassoudKaviany 2011
Net Surface Force Acting
on Control Surface A, F

A
:
Pressure Force:
Viscous Stress Force:
A
(S
.
s
n
) dA
A
p s
n
dA
A
(u)(u
.
s
n
) dA
Net Rate of Momentum Flow
Out of Control Surface:
Gravitational
Force
Electromagnetic
Force
f
e
g
s
n
Control Volume V
Control Surface A
u(x)
Net Volumetric Force Acting
on Control Volume V, F

V
:
Rate of Volumetric
Momentum Storage
in Control Volume V:
V
(g + f
e
) dV
V
u dV
s
n
Figure 1.11. Conservation of uid momentum applied to a control volume V having surface
area A. The momentum owing through the control surface and stored within the volume, as
well as various external forces, are shown.
1.7.3 Conservation of Momentum
The law of conservation of uid momentum (the Newton second law) states that the
timerate of change of the linear momentum u(Ns/m
3
) within a control volume is
equal to the sum of the net rate of momentum owing out of its control surface and
all the external forces
i
F
i
acting on the control surface and volume. The external
force vector F(N) is described below. Consider the control volume and surface shown
in Figure 1.11. For this control volume, we rearrange and write the integralvolume
linear momentum u conservation equation [where all the terms have a unit of N
(newton)] as
_
A
(u)(u s
n
) dA =
t
_
V
udV
i
F
i
1 2 3
(1.27)
integralvolume momentum conservation equation,
where
1 net rate of momentum ow out of the control surface (also called convective acceleration
term)
2 negative of rate of momentum storage in the control volume (also called local acceleration
term)
3 sum of all external forces acting on the control surface and volume.
Here we consider four external forces, namely two surface and two volumetric
forces. These are the surface pressure p(N/m
2
) (always acting toward, and perpen
dicular to the control surface), the surface viscous stress tensor S
j
(N/m
2
) acting on
MassoudKaviany 2011
the control surface, the volumetric gravitational force g(N/m
3
) acting on the control
volume, and the volumetric electromagnetic force f
e
(N/m
3
) also acting on the control
volume. The external forces are written as
i
F
i
= F
p
F
j
F
g
F
e
F
=
_
A
ps
n
dA
_
A
(S
j
s
n
)dA
_
V
gdV
_
V
f
e
dV (1.28)
F external forces.
As an additional external force, a prescribed surface force F may be used as a
closure or boundingsurface prescribed condition. The viscous stress tensor S
j
is a
secondorder tensor and is discussed in terms of its components in Section B.1 of
Appendix B. For the Cartesian coordinates, the nine components of the viscous stress
tensor S
j
are given in (B.7). For an incompressible, Newtonian uid (which is the one
considered in our treatment), these components, e.g., the viscous stress
yx
(N/m
2
)
(where the rst index indicates the surface normal direction and the second the
direction of the force), are given by (B.33). For an incompressible Newtonian uid,
the viscous stress is equal to the product of the uid viscosity j(Pas) and the uid
strain rate (i.e., the rate of deformation).
There are other forces, such as the surface tension at the interface of two immis
cible uids (such as liquid and gas), which can be included.
Note that the momentum equation is a vector equation. In practice, after choos
ing a coordinate system (e.g., linear or radial), the momentum equation is decom
posed by writing it for each of the directions in that coordinate system. This should
be viewed in contrast to the energy conservation and continuity (mass conservation)
equations, which are scalar equations.
1.7.4 Other Conserved Quantities
There are other relevant quantities that are conserved and are described (or gov
erned) by the appropriate conservation equations. Among these are the angular
momentumconservation equation, the electric charge density conservation equation
(part of the Maxwell equations for electromagnetic elds and waves), the radiation
intensity I
r
conservation equation (called the equation of radiation transfer), etc.
Heat transfer analysis involves the use of the energy equation along with the
continuity and momentum equations for convection and a form of equation of
radiation transport (the simplest form is that for surface radiation transport).
The mass, species, and linear momentum conservation equations are further
discussed in Sections B.2.1 to B.2.3, in Appendix B, where they are presented in
alternative forms.
The microscale heat carriers are also described by a carrierdistribution conser
vation equation (called the Boltzmann transport equation). In general, this equation
can be used for any particle (e.g., electron, uid particle, phonon, photon) in any
MassoudKaviany 2011
phase (gas, liquid, solid). We will briey refer to this equation in Section 3.2. The
Boltzmann transport equation
can be considered as the most general conservation
equation, such that the energy, mass, and moment conservation equations can be
derived starting from this equation.
EXAMPLE 1.13. FAM
A picnic cooler, shown in Figure Ex. 1.13, contains ice and receives heat from
the ambient in the amount Q[
A
= 5 W. Assume that the temperature within
the cooler is uniform and at the melting temperature of ice.
Determine the rate of melting of the ice
M
sl
. The phase change can be
represented as an energy conversion
S
sl
=
M
sl
Lh
sl
, and the other energy con
versions can be assumed negligible. The heat of melting of ice Lh
sl
is given in
Table C.6.
Ice and Water
Constant, Uniform
Temperature
S
sl
q
k
q
ku
q
ku
q
u
q
u
q
r
q
r
Q 
A
A
dT
dt
V
= 0
Figure Ex. 1.13. Heat leakage into a cooler resulting in melting of ice.
SOLUTION
As long as there is any ice remaining, one assumes that the entire inside volume
of the cooler is at the melting temperature of ice.
In the energy equation (1.23), with all the terms on the righthand side being
negligible except for the phase change which is represented by
S
sl
, we have
Q [
A
=
S
sl
=
M
sl
Lh
sl
.
The general formof the Boltzmanntransport equationis writtenfor a particle probability distribution
function f , and in the phase space (dened by the particle spatial coordinates and momentum). The
Boltzmann transport equation states that the rate of change of f in phase space id balanced by the
f ,t f ,t[
s
s
f
, where f ,t[
s
is the collision scattering of f (which results in a net increase
or decrease in f ) and s
f
stands for other generation/removal rates of f . The mass, momentum and
energy conservation equation can be derived from the Boltzmann transport equation by taking the
product of f and the zeroth, rst and second power of the particle momentum, respectively, and
then integrating over all the possible particle momentum space. This particlebased energy equation
should then be expanded to include all the other energy conversion mechanisms.
The Boltzmann transport equation is
f
t
u f F
p
f =
f
t
s
s
f
.
where F is the applied (external) force and
p
is the gradient in the momentum space.
MassoudKaviany 2011
From Table C.6, we have for water (at one atmosphere pressure)
Lh
sl
= 3.336 10
5
J/kg Table C.6
T
sl
= 0
C Table C.6.
Solving for
M
sl
, we have
M
sl
=
Q[
A
Lh
sl
=
5(W)
3.336 10
5
(J/kg)
= 1.499 10
5
(kg/s) = 53.96 g/hr.
COMMENT
The heat ows from the ambient to the cooler by surface convection and sur
face radiation. This heat then is conducted through the cooler wall and is then
removed by surface convection and radiation to the inside volume (or by conduc
tion to the waterice region). Once all the ice is melted, the inside temperature
begins to increase and will gradually reach the temperature of the ambient air.
Note that we have assumed that the ice and water are in thermal equilibrium
(i.e., the ice is not subcooled or the water is not superheated with respect to the
equilibrium melting temperature).
1.10 References
[1] Barron, R.F., 1985, Cryogenic Systems, Second Edition, Oxford University Press,
Oxford.
[2] Berman, R., 1979, Thermal Conduction in Solids, Clarendon Press, Oxford.
[3] Callen, H. B., 1985, Thermodynamics and Introduction to Thermostatistics, Sec
ond Edition, Wiley, New York.
[4] Caron, J.M., Gauthier, A., Schaaf, A., Ulysse, J., and Wozniak, J., 1989, La
Planete Terre, Ophrys, Paris. Also, Sarmiento, J.L., and Gruber, N., 2002, Sinks
for Anthropogenic Carbon, Physics Today, August 2002, American Institute
of Physics, New York. Also, Watt, K. G., 1991, Is the Greenhous GasClimate
Signal Hiding in the Deep Ocean? Climate Change, Volume 18, pp. iiivi.
[5] Fox, R.W., and McDonald, A.I., 1992, Introduction to Fluid Mechanics, Wiley,
New York.
[6] Guyton, A.C., 1981, Textbook of Medical Physiology, Sixth Edition, W.B. Saun
der Company, Philadelphia.
[7] Hewitt, G.F., Shires, G.L., and Bott, T.R., 1994, Process Heat Transfer, CRC
Press, Boca Raton.
[8] Hubson, A., 1995, Physics: Concepts and Connections, PrenticeHall, Engle
wood Cliffs. Also, Howking, S., 2001, The Universe in a Nutshell, Bantan Books,
New York. Also, Weinberg, S., 1977, The First Three Minutes, Basic Books, New
York.
[9] Incropera, F.P., and DeWitt, D.P., 1996, Introduction to Heat Transfer, Third
Edition, Wiley, New York.
[10] Kattani, A.M., 1970, Direct Energy Conversion, AddisonWesley, Reading.
MassoudKaviany 2011
[11] Kern, D.Q., 1965, Process Heat Transfer, McGrawHill, New York.
[12] Kock, G.S.H., 1994, Latent Heat Transfer: An Introduction to Fundamentals,
Oxford University Press, Oxford.
[13] Layton, E.T., and Lienhard, J.H., Editors, 1988, History of Heat Transfer, Amer
ican Society of Mechanical Engineers, New York.
[14] Meijer, G.C.M., and Herwaarden, A.W., Editors, 1994, Thermal Sensors, Insti
tute of Physics Publishing, Bristol.
[15] Mills, A.F., 1999, Basic Heat and Mass Transfer, Prentice Hall, Upper Saddle
River, Chicago.
[16] Pulkrabek, W.W., 1997, Engineering Fundamentals of the Internal Combustion
Engine, PrenticeHall, Upper Saddle River.
[17] Rubinsky, B., 1997, Microscale Heat Transfer in Biological Systems at Low
Temperatures, Experimental Heat Transfer, Vol. 10, pp. 129.
[18] Shuran, M., and Nelson, R.A., 1991, Quantication of Energy Expenditure
by Infrared Thermography, American Journal of Clinical Nutrition, Vol. 53,
pp. 13611367.
[19] Tabor, D., 1995, Gases, Liquids, and Solids, and Other States of Matter, Cam
bridge University Press, Cambridge.
[20] Todd, R.H., Allen, D.K., and Alting, L., 1994, Fundamental Principles of Man
ufacturing Processes, Industrial Process Inc., New York.
[21] Van Wylen, G., Sonntag, R., and Borgnakke, C., 1994, Fundamentals of Classical
Thermodynamics, Wiley, New York.
[22] von Baeyer, H.C., 1999, Warmth Disperses and Time Passes: The History of Heat,
Modern Library, New York.
[23] Walton, A.J., 1989, Three Phases of Matter, Oxford University Press, Oxford.
[24] Whitaker, S., 1983, Fundamental Principles of Heat Transfer, Krieger Publishing
Company, Malabar.
[25] Webb, P., 1980, The Measurement of Energy Exchange in Man: An Analysis,
American Journal of Clinical Nutrition, Vol. 33, pp. 12991310.
[26] Zemansky, M.W., 1964, Temperatures Very Low and Very High, Dover, New
York.
MassoudKaviany 2011
2
Energy Equation
2.2 Uniform Temperature in One or More Directions: Energy
Equation for Volumes with One or More Finite Lengths
2.2.3 Discrete Temperature Nonuniformity: FiniteSmall Volume
Energy Equation
We started the discussion of the energy equation with the integralvolume form
(1.23). Then in order to allow for the continuous spatial variation of T and q,
we chose an innitesimal volume. This led to the development of the differential
volume form of the energy equation (2.1), by taking the limit as V LV 0 (i.e.,
innitesimal volume).
We now take a step back, i.e., start before taking the limit as LV 0, and
use the form of energy equation for a nitesmall volume (i.e., LV is small but not
approaching zero). Figure 2.5 shows a nitesmall volume within a heat transfer
medium. Within this small, but nite volume LV, we assume that the temperature
is uniform (i.e., a nitevolume averaged T) and the energy conversion is also rep
resented by a uniform, averaged magnitude s
i
[
LV
LV. This allows for a discrete,
spatial variation of T and q. This is most useful in obtaining numerical solutions for
media with nonuniform temperature distributions.
From (2.8), we have the energy equation for small (but nite volume) LV as
q .
_
LA
(q s
n
)dA
LV
=
Q [
LA
LV
=
t
c
p
T
i
s
i
energy equation. (2.12)
We now assume that c
p
T and s
i
are uniform over this small volume and write
_
LA
(q s
n
)dA = Q[
LA
=
d
dt
(c
p
T)
LV
LV
i
s
i
[
LV
LV
nitesmall volume energy equation,
with uniform temperature.
(2.13)
MassoudKaviany 2011
A
FiniteSmall Volume V
Neighboring
FiniteSmall
Volume
y
x
z
q(x)
q(x)
q
r
q
u
q
k
(c
p
T)
V
V + s
i
V
V
d
dt A
(q
.
s
n
) dA = Q
A
=
(c
p
T)
V
V + s
i
V
V
d
dt
A
(q
.
s
n
) dA
Energy storage and conversion, and the influence of neighboring finitesmall control volumes,
cause a change in direction and magnitude of the heat flux vector across the finitesmall volume.
Figure 2.5. Conservation of energy for a nitesmall control volume, taken as part of a heat
transfer medium.
These terms are also shown in Figure 2.5. Note that the small rectangular volume
LV has a uniform temperature T and in general this temperature is time dependent
(so we have used the total time derivative dT,dt). We now expand the surface heat
transfer term and write it in terms of the three mechanisms for the six surface areas.
We begin by addressing each of the six surface areas.
Figure 2.6 renders the energy equation for a nite rectangular volume having
lengths Lx. Ly. Lz with LV = LxLyLz. For simplicity only the x component of
the heat ux vector is shown. We place the location (x. y. z) at the center of the
rectangular control volume, as shown in Figure 2.6. Then the locations of the corners
of the control volume will be those marked in the gure. Next we write the lefthand
side of (2.13), for these Cartesian coordinates and the rectangular volume, as
(q s
n
LA)
xLx,2
(q s
n
LA)
xLx,2
(q s
n
LA)
yLy,2
(q s
n
LA)
yLy,2
(q s
n
LA)
zLz,2
(q s
n
LA)
zLz,2
=
d
dt
(c
p
T)
LV
LV
i
s
i
[
LV
LV. (2.14)
MassoudKaviany 2011
Cartesian Coordinates Presentation of a FiniteSmall Volume and the Energy Equation
(q
.
s
x
dA)
x x/2
y
z
x
(x x/ 2,
y y/ 2,
z z/ 2)
(x x/ 2,
y y/ 2,
z z/ 2)
(x x/ 2,
y y/ 2,
z z/ 2)
(x x/ 2,
y y/ 2,
z z/ 2)
Conduction:
(k )
x x/2
A
x x/2
T
x
Convection:
(c
p
u
x
T)
x x/2
A
x x/2
Radiation:
q
r
x x/2
A
x x/2
V
A
(q
.
s
x
dA)
x x/2
(k )
x x/2
A
x x/2
T
x
(c
p
u
x
T)
x x/2
A
x x/2
q
r
x x/2
A
x x/2
A
x
x x/2
A
x
x x/2
d
dt
(c
p
T )
V
V
Energy Storage
Energy Conversion
s
i
V
V
x
y
z
i
s
n
= s
x
s
n
= s
x
Figure 2.6. Cartesian coordinates presentation of a nitesmall volume, showing the x com
ponents of the heat ux vector at its boundaries and energy conversion and storage terms.
We now use (1.7) to write q in terms of the three mechanisms of heat transfer.
For simplicity we write the extended version only for the x component of the heat
ux vector. This is shown in Figure 2.6. The result is
k
T
x
[
xLx,2
LA
xLx,2
k
T
x
[
xLx,2
LA
xLx,2
(c
p
Tu
x
)
xLx,2
LA
xLx,2
(c
p
Tu
x
)
xLx,2
LA
xLx,2
q
r.x
[
xLx,2
LA
xLx,2
q
r.x
[
xLx,2
LA
xLx,2
(q s
n
LA)
yLy,2
(q s
n
LA)
yLy
(q s
n
LA)
zLz,2
(q s
n
LA)
zLz
=
d
dt
(c
p
T)
LV
LV
i
s
i
[
LV
LV. (2.15)
This is the nitesmall volume energy equation. Further specication can be
made about the conduction terms (by representing conduction heat transfer by a
resistance as will be shown in Chapter 3).
2.3 Thermal Energy Conversion Mechanisms
2.3.1 Chemical or PhysicalBond Energy Conversion
(B) Nuclear Fission Reaction
A nuclear reaction leading to division of an atomic nucleus into two nuclei of com
parable mass is called nuclear ssion. This can occur naturally (spontaneous ssion)
or under bombardment with a neutron or other carriers of energy (induced ssion).
MassoudKaviany 2011
For the neutron bombardment of a radioactive element (e.g., uranium), the thermal
energy generation is given by
S
r.f i
V
= s
r.f i
= n
n
A
nu
u
n
V
Lh
r.f i
volumetric heating by nuclear ssion. (2.22)
where n
n
(1/m
3
) is the number density of the neutrons, A
nu
,V(1/m) is the volume
specic (i.e., per unit volume) collision cross section for neutron nucleus, u
n
(m/s) is
the average velocity of the neutrons, and Lh
r.f i
(J/kg) is the specic energy released
(negative).
EXAMPLE 2.8. FAM
In a lightwatermoderated uranium reactor, the bombardment of uranium rods
by neutrons results in the conversion of nuclearbond energy to thermal energy.
The neutron ux is n
u
u
n
= 10
17
neutron/m
2
s, and the specic effective ssion
cross section is A
nu
,V = 47.90 m
2
/m
3
. The heat released per collision is Lh
r.f i
=
180 MeV/collision or 2.884 10
11
J/collision (exceptionally high).
Determine the volumetric energy conversion rate.
SOLUTION
From (2.22), the volumetric energy conversion rate is
s
r.f i
= n
u
u
n
A
nu
V
Lh
r.f i
= 10
17
(neutron/m
2
s) 47.90(m
2
/m
3
) [2.884 10
11
(J/neutron)]
= 1.381 10
8
W/m
3
.
COMMENT
Note this heat release rate is very large and smaller than that generated by the
combustion of methane in air at the adiabatic ame temperature.
(C) Nuclear Fusion Reaction
This reaction occurs between various nuclei of light elements (e.g., isotopes of hydro
gen) inthe thermal plasma state, i.e., a gas phase at hightemperature (witha complete
ionization of the species). This ionized state allows for the colliding nuclei to react.
The rate of thermal energy generation is
S
r.f u
V
= s
r.f u
= n
1
n
2
(Au)
12
Lh
r.f u
volumetric heating by nuclear fusion. (2.23)
where n
1
(1/m
3
) and n
2
(1/m
3
) are the number densities of the two colliding nuclei,
(Au)
12
(m
3
/s) is the average value of the product of the mutual collision cross section
and the relative velocity, and Lh
r.f u
(J/kg) is the specic energy released per collision.
MassoudKaviany 2011
In the hydrogenhydrogen (protonproton) solar reaction, the energy produced is
Lh
r.f u
= 2.67 MeV per He atom produced.
For a deuteriumdeuterium fusion reaction, an extremely large temperature of
about 4 10
8
K is needed. Since the plasma is charged, it is kept away from any
solid surface by a magnetic connement. Because of the high temperature required,
it is called a thermonuclear reaction.
2.3.2 Electromagnetic Energy Conversion
(E) Volumetric and Surface Radiation Absorption
An electromagnetic wave (or phonon) of a given wavelength (or frequency) can
be absorbed into a molecule by adding to the vibrational energy of the molecule,
and a conversion to thermal energy (as compared to electronic energy) results. This
requires a resonance between the impinging wave and the vibrational energy level
transition (or excitement). This is the volumetric and surface radiation heating of
a radiation absorbing medium. A general treatment of radiation transport in an
absorbing, reecting (or scattering), and emitting medium is beyond our scope here
and is postponed to further studies [18]. Surface radiation is extensively examined
in Chapter 4. For a semitransparent, absorbing medium, a simple treatment can
be made by including the volumetric absorption in the energy conversion term.
For opaque media, surface absorption is also addressed here. Surface emission is
considered next.
Consider an impinging radiation heat ux q
r.i
(called irradiation) arriving at
the bounding surface of a semitransparent medium. This irradiation can be due to
the surface thermal emission [discussed in Section 2.3.2(F)] or due to electron or
particle beams, or due to nonthermal lasers. Then a fraction of this heat ux may be
reected. This fraction is
r
q
r.i
, where
r
is the surface reectivity. This is shown in
Figure 2.9(e). The remainder will penetrate into the medium. At a location x (from
the bounding surface of the medium), the change due to absorption and scattering
[here we represent this simply as a single attenuation) effect in q
r.i
as it penetrates
(i.e., transmits) along x] is given by
q
r
(x)
q
r.i
= (1
r
)e
ex
x
local radiation ux. (2.42)
where
ex
(1/m) is the extinction coefcient or attenuation coefcient. The extinction
coefcient is the inverse of the photon meanfree path
ph
(m), i.e.,
ph
= 1,
ex
. Then
the volumetric rate of absorption s
r.
is found from q
r
.
Fora onedimensional
ex
= 1/
ph
(1/m)
Fused Quartz
(Crystalline), = 4 m
LowIron Glass
(Amorphous), = 0.4 m
Fused Quartz
(Crystalline), = 8 m
LowIron Glass
(Amorphous), = 8 m
Air, STP
10
4
10
2
1 10
2
10
4
10
6
10
8
0
Porosity = 0.9
10 kg/m
3
T = 830 K
p = 1 atm
= 15 m
S
e.
(x)
V
= s
e.
(x)
dq
r
(x)
dx
= q
r.i
(1
r
)
ex
e
ex
x
volumetric radiation
absorption.
(2.43)
The simple treatment in (2.43) allows us to address the volumetric radiation due to
absorption only.
For a layer of thickness L (or at a location L from the surface of a layer with
thickness larger thanL), the product
ex
Lis calledthe dimensionless optical thickness
ex
, i.e., optical thickness of an absorbing layer. This is
ex
=
ex
L optical thickness
ex
. (2.44)
The optically thin and optically thick layers are dened as
ex
 0.1 optically thin (transparent) medium (2.45)
ex
> 10 optically thick (opaque) medium. (2.46)
The range 0.1
ex
10 is called the semitransparent medium. Figure 2.13 gives
typical values of
ex
(1/m) for some gases, liquids, and solids. The materials exhibit
strong dependence of
ex
onthe wavelengthof the irradiation, (m). The wavelengths
of interest in thermal radiation will be discussed in Chapter 4. For example, water
strongly absorbs the larger wavelengths (infrared range), as compared to the shorter
wavelengths (visible range).
Most gases are generally treated as very optically thin (i.e.,
ex
 0.1, or ide
ally transparent or ballistic photon transport), and liquids are optically thin for
MassoudKaviany 2011
some wavelengths (for example water in the visible wavelength range), but other
wise they are optically thick (or diffusive photon transport). Metallic (electrically
conducting) solids are optically very thick, i.e.,
ex
> 10, or opaque. Nonconduct
ing (or dielectric) solids can be optically thin, e.g., glass, which is optically thin in
the visible wavelength range. We will discuss the role of electronic and vibrational
energy states of matter in the emission/absorption of radiation in Section 4.1.2.
Note that
ex
depends on the layer thickness L. Then for materials that absorb a
noticeable portion of the irradiation, as L becomes sufciently large, no radiation is
transmitted.
For opaque surfaces (no signicant penetration of radiation, i.e., no transmis
sion), we have
ex
. Then the amount of radiation absorbed at the surface
S
e.
,A
is
S
e.
A
= (1
r
)q
r.i
=
r
q
r.i
surface radiation absorption on opaque surfaces, (2.47)
where
r
is the surface absorptivity and for opaque surfaces we have
r
= 1
r
. We
will discuss laser surface heating, and some of its applications, in Section 4.5.
EXAMPLE 2.12. FAM
In a materialprocessing application, depicted in Figure Ex. 2.12, volumetric
radiation absorption heating of a solid is made by a shorttime irradiation of
q
r.i
= 10
10
W/m
2
. The surface reectivity is
r
= 0.2 and the medium has a very
large extinction coefcient
ex
= 10
3
1/m.
Determine the optical thickness
ex
, the local radiative heat ux q
r
(x), and
the volumetric radiation absorption rate s
e.
(x) at four different locations: (i)
x = 10
4
m, (ii) 10
3
m, (iii) 10
2
m, and (iv) 10
1
m, from the surface.
Interface
Absorbing Medium with
Extinction Coefficient
ex
(1/m)
Attenuated Radiation
Heat Flux q
r
(x)
Irradiation
q
r,i
Reflected Irradiation
r
q
r,i
x
s
r,
Volumetric
Absorption
q
r
(x)
(1
r
)q
r,i
Figure Ex. 2.12. Laser irradiation of an absorbing medium.
MassoudKaviany 2011
SOLUTION
The local optical thickness
ex
, radiative heat ux q
r
, andthe volumetric radiation
absorption s
e.
are given by (2.42) and (2.43), i.e.,
ex
=
ex
x
q
r
(x) = q
r.i
(1
r
)e
ex
x
s
e.
(x) = q
r.i
(1
r
)
ex
e
ex
x
.
Using the numerical values we have
ex
= 10
3
(1/m) x(m)
q
r
(x) = 10
10
(W/m
2
) (1 0.2) exp[10
3
(1/m) x(m)]
s
e.
(x) = 10
10
(W/m
2
) (1 0.2) 10
3
(1/m) exp[10
3
(1/m) x(m)].
Now for the four positions, we have
(i) At x = 10
4
m,
ex
= 0.1
q
r
= 8 10
9
(W/m
2
) exp[10
3
(1/m) 10
4
(m)] = 7.24 10
9
W/m
2
s
e.
= 8 10
12
(W/m
2
) exp[10
3
(1/m) 10
4
(m)] = 7.24 10
12
W/m
3
.
(ii) At x = 10
3
m,
ex
= 1
q
r
= 8 10
9
(W/m
2
) exp[10
3
(1/m) 10
3
(m)] = 2.94 10
9
W/m
2
s
e.
= 8 10
12
(W/m
2
) exp[10
3
(1/m) 10
3
(m)] = 2.94 10
12
W/m
3
.
(iii) At x = 10
2
m,
ex
= 10
q
r
= 8 10
9
(W/m
2
) exp[10
3
(1/m) 10
2
(m)] = 1.632 10
5
W/m
2
s
e.
= 8 10
12
(W/m
2
) exp[10
3
(1/m) 10
2
(m)] = 3.632 10
8
W/m
3
.
(iv) At x = 10
1
m,
ex
= 100
q
r
= 8 10
9
(W/m
2
) exp[10
3
(1/m) 10
1
(m)] = 2.98 10
34
W/m
2
s
e.
= 8 10
12
(W/m
2
) exp[10
3
(1/m) 10
1
(m)] = 2.98 10
31
W/m
3
.
COMMENT
Note the small optical thickness and attenuation [i.e., q
r
(x) does not decrease
signicantly] and the large volumetric heating rate for x = 10
4
m (i.e., 0.1 mm
MassoudKaviany 2011
or 100 jm from the surface). The radiative heat ux and the volumetric heating
rates vanish completely at x = 10
1
m (i.e., 10 cm from the surface), when the
optical thickness is
ex
= 100. The volumetric generation rates on the order of
10
12
W/m
3
are among the highest heating rates. In laser welding, the beam is
moved or the duration of radiation is kept short (of the order of 10
2
s) to
achieve the proper phase change and to prevent the undesirable effects of the
overheating. We will discuss laser irradiation in Sections 4.5.1 and 4.5.2.
2.3.3 Mechanical Energy Conversion
(A) PressureCompressibility Cooling or Heating
Pressurecompressibility cooling or heating is a reversible volumetric conversion of
mechanical energy tothermal energy, due tothe uidcompressibility. This volumetric
conversion rate is derived in Appendix B and is given by (B.46). It is repeated
here as
S
m.p
V
= s
m.p
=
f
T
_
t
u
_
p
volumetric pressurecompressibility
cooling or heating,
(2.50)
where
f
(1/K) is the volumetric thermal expansion coefcient dened in Appendix
A, by (A.10). The pressurecompressibility work is signicant in gases and upon
expansion (i.e., reduction in pressure) expansion cooling occurs. Upon compres
sion, compressive heating occurs. Here we are not addressing the corresponding
changes that occur in the spatial variation of temperature which inuences the heat
transfer.
The temperature of a gas stream can be reduced by expansion (isenthalpic
expansion) and then this stream can be used for heat exchange. This expan
sion cooling is characterized by the JouleThompson coefcient j
JT
= (T,p)
h
=
[T(,T)
p
:],c
p
which is nonzero for nonideal gases, and should be positive for
expansion cooling. Methane has a large, positive j
JT
(p. T) and is used in staged
cooling to cryogenic temperatures.
EXAMPLE 2.14. FUN
Air is expanded in a diverging nozzle, as shown in Figure Ex. 2.14. Assume an
idealgas behavior, a steady ow, a uniform onedimensional ow throughout
the nozzle, and a linear onedimensional pressure distribution.
(a) Determine the local volumetric expansion cooling rate as a function of the
local velocity u
x
and the assumed constant gradient of pressure (dp,dx)
o
.
(b) Evaluate the magnitude of s
m.p
for a velocity of u
x
= 10
2
m/s and (dp,dx)
o
=
10
5
Pa/m.
MassoudKaviany 2011
Fluid Stream
(Gas)
s
m,p
u
x
= u
x
(x) only : Uniform, OneDimensional Fluid Velocity
p = p
o
+
dp
dx
x : Linear, OneDimensional Pressure Distribution,
Expansion Nozzle
o
dp
dx
< 0, i.e., Pressure Decreasing with x
o
x, u
x
u = u
x
s
x
Figure Ex. 2.14. Gas ow in an expanding, onedimensional channel.
SOLUTION
(a) From (2.50), we have
s
m.p
=
f
T
_
t
u
_
p.
For a steady ow, p,t = 0. For an ideal gas from (1.19), we have
p: =
R
g
M
T ideal gas.
From the denition of
f
given by (A.10), we have
f
1
:
:
T
[
p
.
Then for
f
for an ideal gas, we have
f
=
1
:
T
_
R
g
Mp
T
_
p
=
R
g
M
1
:p
=
1
T
.
i.e., the volumetric thermal expansion coefcient for gas is equal to 1,T, where
T is in absolute scale (K). Then
f
T = 1 for an ideal gas.
For the pressure distribution, we have a linear variationalong the x direction.
This is written as p
o
(p,x)
o
x.
MassoudKaviany 2011
Returning to (u )p, we have for the onedimensional ow
u = u
x
s
x
u = (u
x
s
x
)
_
x
s
x
y
s
y
z
s
z
_
= u
x
x
(u )p = u
x
x
_
p
o
_
p
x
_
o
x
_
= u
x
_
p
x
_
o
.
Finally, we obtain
s
m.p
= u
x
_
p
x
_
o
.
(b) Using the numerical values, we have
s
m.p
= 10
2
(m/s) (10
5
)(Pa/m)
= 10
7
Pa/s = 10
7
W/m
3
.
COMMENT
Note that this air speed is about 1/3 of the sonic speed at STP and this pressure
gradient corresponds to a drop of one atm pressure over one meter of distance.
Note that u
x
is not constant in a variablearea channel and it should decrease
along x (as governed by the conservation of mass). Finally, note that the minus
sign in s
m.p
indicates heat absorption and that 10
7
W/m
3
is a signicant cooling
rate.
(B) Viscous Heating
Viscous heating is an irreversible volumetric conversion of mechanical energy to
thermal energy, due to the uid viscosity. This volumetric conversion rate is derived
in Appendix B (B.39) and is repeated here as
S
m.j
V
= s
m.j
= S
j
: u volumetric viscousdissipation heating, (2.51)
where S
j
(Pa) is the viscous stress tensor, u(1/s) is dyadic product of vectors
and u (discussed in Appendix B), and u is the uid velocity. For the Cartesian
coordinates and Newtonian uids, s
m.j
is given by (B.42). For a twodimensional
MassoudKaviany 2011
(x. y), incompressible uid ow, (B.42) becomes
S
j
: u = j
f
_
u
y
x
u
x
y
_
2
. (2.52)
where u = (u
x
. u
y
) and j
f
(Pas) is the uid viscosity.
In order to calculate s
j
, the velocity eld u must be known. The viscous heating
is signicant in high viscosity liquids (e.g., in lubrication), and under some special
conditions in gases.
EXAMPLE 2.15. FUN
In hydrodynamic lubrication, a moving solid is separated from a stationary
solid by a lubricant (a highviscosity uid), while moving with respect to the
stationary solid. Assume a onedimensional ow of the lubricant, as shown in
Figure Ex. 2.15. Also assume a Newtonian behavior for the uid.
(a) Determine the volumetric viscous heating in terms of the uid viscosity j
f
,
the relative velocity Lu
i
, and the distance between the solid surfaces L (the gap
width).
L
y
y
x, u
x
0
Steel Ball (Moving)
Outer Race (Stationary)
OneDimensional Lubricant Velocity: u
x
= u
i
y
L
0 u
i
u
i
s
m,
Lubricant Stream
u
x
Outer Race (Stationary)
Steel Ball
Lubricant
Inner Race (Moving)
u = u
x
s
x
Figure Ex. 2.15. Velocity distribution in the gap space between a moving and a stationary
solid surface.
MassoudKaviany 2011
(b) Using j
f
= 3.660 10
1
Pa/s (for engine oil at T = 310 K, listed in Table
C.23), Lu
i
= 1 m/s, and L = 0.2 mm, determine the magnitude of s
m.j
.
SOLUTION
(a) From (2.51) we have
s
m.j
= S
j
: u.
The product S
j
: u is given in (2.52) for a Newtonian uid. Here, for the one
dimensional ow, we have
u = u
x
s
x
u
y
= u
z
= 0. u
x
= u
x
(y) only.
This onedimensional velocity is linear in x, as shown in Figure Ex. 2.15, and is
given by u
x
= Lu
i
y,L.
Then for (2.52), we have
s
m.j
= S
j
: u = j
f
_
u
x
y
_
2
= j
f
_
d
dy
_
Lu
i
y
L
_
_
2
= j
f
Lu
2
i
L
2
.
(b) For the numerical values of the parameters given, we have
s
m.j
= 3.66 10
1
(Pas)
_
1(m/s)
2 10
4
(m)
_
2
= 9.151 10
6
W/m
3
.
COMMENT
Note that if this generated heat is not removed, the oil can disintegrate and leave
deposits on the solid surfaces. This oil disintegration deteriorates the lubricant
performance.
2.4.3 Prescribed BoundingSurface Thermal Conditions
The boundingsurface thermal condition can also be prescribed. Here prescription is
a value assigned to a quantity. This value may be unknown and is found as part of the
problem solution). For example if the boundingsurface temperature is maintained
uniform (along the surface) and constant (with respect to time), then we have a
prescribed, constant surface temperature condition. This is shown in Figure 2.16.
By prescribing a quantity at the bounding surface, we exclude consideration of its
variation outside of this bounding surface (this is a closure condition). In some cases
the magnitude of the surfacenormal component of the heat ux vector is maintained
uniform along the surface and constant with time. This is called a prescribed surface
heat ux condition. This is also depicted in Figure 2.16.
The boundingsurface condition is also called the boundary condition, because
the differentialvolume energy equation would require these conditions (or con
straints) in order to have a particular solution. These will be further discussed in the
following chapters.
MassoudKaviany 2011
q
s
on A
ji
(Prescribed)
or
T
s
on A
ji
(Prescribed)
s
n, j
s
n,i
BoundingSurface Heat Flux Vector
from Medium i to Surface
BoundingSurface Heat Flux Vector
from Surface to Medium j
Heat Transfer
Medium j
Control Surface on
Bounding Surfaces
A
ij
= A
ji
Prescribed BoundingSurface Thermal Conditions
A
ij
Heat Transfer
Medium i
q
s
q
i
(q
.
s
n
)
i
A
ji
+ (q
s
.
s
n
)
j
A
ij
= S
S
Figure 2.16. Prescribed thermal conditions at a bounding surface.
Insome heat transfer problems, a knowledge of the temperature onthe bounding
surfaces leads to the determination of the heat ow rate within the medium (e.g.,
conduction, Section 3.3.1) q(x). For the prescribed, uniform surface temperature
condition, (2.57) becomes
T = T
i
= T
j
= T
s
prescribed, uniform surface temperature on A
ij
= A
j i
. (2.64)
where subscript s stands for surface (bounding). This applies to any point along A
ij
and A
j i
.
In other cases, the surface temperature is unknown and some information about
the surface heat ow rate is specied. This is similar to an inverse problem, where
T
s
is to be determined. The simplest of these is when q
s
on the bounding surface is
normal (i.e., uniform bounding surface temperature) and is uniform and prescribed.
Then (2.62), with no surface energy conversion, becomes
[k
i
(T s
n
)
i
(c
p
Tu s
n
)
i
(q
r
s
n
)
i
]A
ij
q
s
A
j i
=
S
energy equation with prescribed q
s
on A
j i
= A
ij
. (2.65)
Note that positive q
s
is owing out of the bounding surface toward the j medium.
In Chapters 3, 4, 5, 6, and 7 we will examine the boundingsurface conditions of
prescribed T
s
and q
s
on A
sf
= A
f s
.
2.4.4 FarField Thermal Conditions
In intramedium heat transfer, the medium may be semibounded (i.e., semiinnite)
or unbounded (innite), or may be approximated to be semibounded or unbounded.
Thenthe thermal conditions inthe fareldlocations needtobe specied. For thermal
MassoudKaviany 2011
equilibrium (i.e., uniform temperature) in the far eld, which is the condition of
uniform fareld temperature, we have
T = T
is the specied fareld temperature. Note that this would result in a zero
fareld temperature gradient, i.e., T = 0, for x .
Similarly, the mechanical equilibrium (i.e., uniform force or motion) may be
specied for the fareld locations. This for a uid ow is given by
u
f
= u
f .
as x fareld mechanical equilibrium, (2.67)
where u
f .
is the specied fareld velocity.
In mass transfer problems, chemical equilibrium species that the chemical
potential is uniform. This can be stated in terms of the uniformity of the species
concentrations.
EXAMPLE 2.18. FUN
At a gassolid interface, shown in Figure Ex. 2.18, with a surface area A,
S
i
= 0,
and there is uniform temperature T
sf
= T
s
= T
f
. Heat ows to the surface on
the solid side by conduction and heat ows away from the surface on the gas
side by conduction and radiation.
(a) Write the appropriate boundingsurface energy equation.
(b) For the values of Q
k.f
= 600 W, Q
r.f
= 200 W, and Q
k.s
= 800 W, show that
the energy conservation is satised.
(c) Show the conduction heat ow in the gas and the solid phase in terms of
the appropriate gradients of the steadystate, onedimensional gas and solid
temperature distributions T
f
= T
f
(x) and T
s
= T
s
(x).
SOLUTION
Since the bounding surface has a uniform temperature, the heat ows only
normal (i.e., perpendicular) to the surface.
u = 0
T
f
=T
f
(x)
T
s
=T
s
(x)
Uniform
Temperature
T
sf
= T
s
= T
f
s
n,s
s
n,f
x, s
x
Solid
Gas Q
k,f
= 600 W Q
r,f
= 200 W
Q
k,s
= 800 W Q
r,s
= 0
A
sf
= A
fs
A
fs
Control Surface
Wrapped around A
Figure Ex. 2.18. Boundingsurface energy conservation for a gassolid interface.
MassoudKaviany 2011
(a) Since there is no surface energy conversion, the appropriate surface energy
equation is (2.62). This equation, for the case of no motion on the surface (u = 0),
becomes
Q[
A
= A
sf
[k
s
(T s
n
)
s
k
g
(T s
n
)
f
(q
r
s
n
)
s
(q
r
s
n
)
f
]
= 0 uniform condition on A
s f
= A
f s
.
Now using the heat transfer mechanisms designation shown in Figure Ex. 2.18 ,
we write the above surface energy equation as
Q
k.f
Q
r.f
Q
k.s
Q
r.s
= 0 surface energy equation.
Note that all Qs are taken as positive when they are owing along the surface
normal s
n
.
(b) Now substituting for the appropriate terms in the above energy equation,
we have
600(W) 200(W) 800(W) 0 = 0.
Note that since Q
k.s
 0, this heat ows toward the bounding surface from the
solid phase side.
(c) We now write the conduction heat transfer terms Q
k.f
and Q
k.s
in terms of
the spatial variations of T
f
= T
f
(x) and T
s
= T
s
(x). We note that for the gas
side we have s
n
= s
x
, and using (1.11) we write
Q
k.f
= A
f s
k
f
(T s
n
)
f
= A
f s
k
f
(T s
x
)
f
on A
f s
.
Upon using (1.14) and noting that T
y
= T
y
(x), we have
Q
k.f
= A
f s
k
f
dT
f
dx
on A
f s
.
Similarly, for the solid side, we have
Q
k.s
= A
sf
k
s
(T s
n
)
s
= A
sf
k
f
[T (s
x
)]
f
on A
sf
= A
f s
Q
k.s
= A
sf
k
g
dT
s
dx
on A
sf
= A
f s
.
COMMENT
Note the difference in the sign appearing after the equal signs for Q
k.f
and Q
k.s
.
The origin of x axis is at the surface and then x increases moving into the gas
phase. If near the bounding surface T decreases in the x direction (for both the
gas and solid phase), the xdirection derivatives of T
f
and T
s
are negative, i.e.,
dT
f
dx
 0 and
dT
s
dx
 0.
Then, since kis always positive (andfor vacuumit is zero), Q
k.f
= A
f s
k
f
dT
f
,dx
will be positive and Q
k.s
= A
sf
k
s
dT
s
,dx will be negative. Also note that these
derivatives are evaluated at the bounding surface, i.e., at A. Since here the
MassoudKaviany 2011
temperature varies only in the x direction, we can replace the specication at A
with the specication x = 0. Then the conduction heat transfer rates become
Q
k.f
= A
f s
k
f
dT
f
dx
at x = 0
or Q
k.f
= A
f s
k
f
dT
f
dx
[
x=0
Q
k.s
= A
sf
k
s
dT
s
dx
at x = 0
or Q
k.s
= A
sf
k
s
dT
s
dx
[
x=0
.
Also note that across the interface we have the continuity of temperature
given by (2.57), i.e.,
T
f
= T
s
at x = 0.
EXAMPLE 2.19. FUN
Consider the interface of a deep pool of water evaporating into a moving gaseous
ambient that is at a higher temperature, as shown in Figure Ex. 2.19. Heat ows
into the interface by surface convection and radiation from the gas side, and by
conduction from the liquid side. Assume that the water surface temperature is
known (prescribed) as T
lg
= T
lg
(p
g
) where p
g
is the watervapor pressure. Also
assume that the fareld temperatures for the gas and liquid are known as T
g.
and T
l.
.
(a) Write the interfacial thermal conditions, i.e., the temperature continuity and
the surface energy equation.
(b) Qualitatively show the temperature distributions within the liquid and the
gas.
SOLUTION
(a) The continuity of temperature across the interface, combined with the
prescribed surface temperature, gives
T
l
= T
g
= T
lg
(p
g
).
The boundingsurface energy equation that includes phase change as the
surface energy conversion is (2.62). The energy conversion term is given by
(2.26). Here we use i = l for the liquid phase and j = g for the gas phase. Then
neglecting radiation on the liquid side and using gasside surface convection
q
ku.g
(this will be explained in Chapter 6), we have
A
lg
[q
r.g
q
ku.g
q
k.l
] =
S
lg
= A
lg
m
lg
Lh
lg
on A
lg
= A
gl
.
Note that as shown in Figure Ex. 2.19, we have taken q
r.q
, q
ku.g
, and q
k.l
to be
pointing away from the surface (i.e., being along s
n
). Since they actually all ow
toward the surface, they all will be negative. The heating use in evaporation
rate per unit area m
lg
Lh
lg
will be positive (because heat is absorbed during
liquidgas phase change, i.e., Lh
lg
> 0.
MassoudKaviany 2011
Gas
Liquid
S
lg
= A
lg
m
lg
h
lg
T
l,
T
g,
T
lg
= T
lg
(p
g
)
q
r,g
q
ku,g
q
k,l
s
n
s
n
x
T
Prescribed
Temperature
Prescribed FarField
Gas Temperature
Control Surface
Wrapped around A
lg
Prescribed FarField
Liquid Temperature
0
A
gl
= A
lg
A
lg
Figure Ex. 2.19. Surface evaporation with the heat supply from the gas and the liquid sides.
(b) The heat for surface (interface) evaporation ows from the gaseous ambi
ent and also from the liquid. This is shown in Figure Ex. 2.19. Here we assume
that the water vapor pressure at the interface is known, and therefore, the sur
face temperature is known. The temperature distribution around the interface,
in the gas and in the liquid, are also shown in Figure Ex. 2.19. The temperature
in the gas increases with the distance from the interface and asymptotically (i.e.,
gradually) reaches T
g.
(prescribed). The temperature in the liquid also reaches
T
l.
far from the interface.
COMMENT
Inpractice, the cooler liquidontopis heavier andsinks, causing a thermobuoyant
motion. We will discuss surface evaporation in Section 6.9. There we will show
that the water surface temperature is determinedfromthe heat andmass transfer
and cannot be prescribed.
2.5 Heat Transfer Analysis
The objective of heat transfer analysis is the design and optimization of thermal
systems by controlling the magnitude, direction, and spatialtemporal variations of
each mechanism of heat transfer.
In the following chapters we will discuss the parameters that inuence q, i.e.,
generate, resist, enhance, or redirect q. For example, an increase in the conductivity
increases q
k
, an increase in velocity increases q
u
, and an increase in surface emissivity
increases q
r
. The geometry also plays a role on q and in q. By changing the
direction of ow (e.g., ow parallel or perpendicular or oblique to a surface) q
u
and
q
u
can be controlled.
The subjects of heat transfer analysis are rendered in Chart 2.5. The heat transfer
problem may be on a subcomponent (e.g., surfaces partitioning uids in a heat
exchanger). The heat transfer may also be on a component (e.g., a heat exchanger
containing these surfaces), or it might be on a system (e.g., a gasturbine cycle
containing this heat exchanger). For each of these levels, central to the heat transfer
MassoudKaviany 2011
Subjects of Heat Transfer Analysis
Component System Subcomponent
Conservation
of Energy
Heat
Transfer
Medium
Overall Energy
Conservation Analysis
Multiple Heat
Transfer Media
Multiple Com
ponent Energy
Conservation
Analysis
Overall Energy
Conservation Analysis
Multiple Heat
Transfer Media
Multiple Subcom
ponent Energy Con
servation Analysis
Heat Flux
Vector
Storage
Conduction Convection Radiation
Conversion
Chart 2.5. Subjects of heat transfer analysis for systems and their components.
analysis are the energy conservation equation, the heat transfer media, and the
bounding surfaces.
Here we discuss aspects of heat transfer analysis that deals with nonuniformity
of temperature and that require the application of the differentialvolume energy
equation, multidimensionality of heat ow, and time variation. Then we introduce
the thermal circuit modeling.
2.5.1 Integration of DifferentialVolume Energy Equation
Heat transfer is a result of thermal nonequilibrium in a medium or between media.
As shown in Chart 2.6, in some applications the integralvolume energy equation
is sufcient and no local or distributed temperature variation is necessary. This
is called the lumpedcapacitance analysis. We will discuss the lumpedcapacitance
analysis in Sections 3.4.6, 4.6.2, and 6.8.3. However, a general treatment should
allow for a pointwise or local treatment of the heat ow. The differentialvolume
energy equation states the local thermal interactions within the medium or the
media. In order to have a closure, we either extend the medium or the media to
where there is a thermal equilibrium (i.e., thermalequilibrium boundary condi
tions), or we specify (i.e., prescribe) the thermal condition at a particular spatial
location (i.e., prescribed boundingsurface thermal conditions). We then need to
Integration of Differential Energy Equation
for Temperature or Heat Flux Vector
Temporal Variation
Steady State Transient
Uniform
(Lumped)
One Dimensional
Spatial Variation
Periodical
Nonperiodical
Three Dimensional
No Motion Motion
Fluid Velocity Field
Non
uniform
Velocity
Field
Uniform
Velocity
Field
Significant Negligible
Radiation Heat Transfer
Energy Conversion
Constant Temperature
Dependent
Two Dimensional
Chart 2.6. Aspects of the thermal problem affecting the solution to the energy equation.
MassoudKaviany 2011
integrate the differentialvolume energy equation up to these thermalequilibriumor
nonequilibrium boundaries (heat ows across a thermalnonequilibrium boundary,
but not across a thermalequilibrium boundary).
In the differentialvolume energy equation (2.1), q
r
can also be expressed in
terms of the local temperature T. As discussed in Section 2.7, s may also depend on
the local temperature T. Then the differentialvolume energy equation expresses or
governs the local spatial and temporal variations of temperature, i.e., T = T(x. t).
Knowing the properties of the heat transfer medium (thermodynamic properties,
. c
p
, etc., transport properties, k. j, etc., and its reactivity and other energy con
versions) and its boundingsurface thermal conditions, we seek to determine the
temperature distribution. The temperature distribution, T = T(x. t), is related to
the heat ux vector distribution q(x. t) through the denition of the three modes of
heat transfer as given in (2.1).
The questions we ask may include the time rate of increase in the local tempera
ture of a piece of dielectric material placed in a microwave eld in order to determine
when a given temperature is reached within or on the surface of the material. We
may ask how the direction of the uid owing over a solid being cooled by this
uid affects the solid surface temperature and the rate of heat owing through this
surface.
To obtain the variations of q(x. t) and T(x. t), we need to solve the differential
equation (2.1). When solving for T, we need to nd the solution to a secondorder
differential equation. When more than one space variable is needed, this becomes a
secondorder partial differential equation. When temporal, as well as spatial, varia
tions are needed, this becomes a partial differential equation with the time derivative
being rstorder and the space derivatives remaining secondorder. Chart 2.6 shows
the various aspects inuencing the solution to the temperature eld (i.e., distribu
tion). There are various analytical methods used in solving this differential equation.
There are also numerical methods used to very closely approximate the solution
(Section 3.7 and Chapter 8).
Throughout the text, we will discuss the techniques used in solving the energy
equation for T and q. Chart 2.7 gives aspects of the energy equation discussed in
each chapter. Conduction is discussed Chapter 3, where because of the simplicity of
conduction heat transfer, and yet its signicant occurrences, we discuss the various
solution methods and solutions. In the following chapters, we progressively include
radiation and convection.
q
k
, c
p
T, s q
r
, q
k
, c
p
T , s
.
q = c
p
T + s
t
q
k
, q
u
, q
r
, q
ku
, c
p
T, s q
u
, q
k
, s
t
SurfaceConvection
and Conduction
(Chapters 6,7)
Conduction,
Convection,
and Radiation
(Chapter 8)
Conduction
(Chapter 3)
q
u
, q
k
, q
ku
, c
p
T
t
Radiation and
Conduction
(Chapter 4)
Convection
and Conduction
(Chapter 5)
Chart 2.7. Various heat transfer mechanisms and volumetric energy storage and conversion
in the energy equation, as discussed in each chapter.
MassoudKaviany 2011
2.5.2 Single and Multidimensional Heat Flow
In any practical nonisothermal system, the temperature varies throughout the sys
tem. However in some systems the temperature variation and heat oware dominant
in a single direction. This assists in the analysis, and as a result, there are available
many closedform solutions to the heat transfer problems with a dominant single
dimension for heat ow. There are many cases where these solutions are justi
ably applicable. If not valid, then two and threedimensional heat ow analyses are
used.
The Cartesiancoordinate system is used for the planar systems (one
dimensional) and the rectangular systems (two and threedimensional). For cylin
drical, spherical, and other curved surfaces, the radialcoordinate systems are used.
There are practical examples in which the region inuenced by the heat trans
fer is a very small volume adjacent to the surface, where the thermal changes are
occurring. In this case, only a thin region near the surface is analyzed (the socalled
skineffect region) and it is assumed that the heat transfer medium is semiinnite
with equilibrium fareld conditions. An example is the penetration of the seasonal
variation of temperature into the earths crust. This penetration distance is only sev
eral meters (compared to the earths radius of 6,371 km). Another example is the
thermal boundary layer formed in a uid moving parallel to a planar, solid surface
(with the thermal nonequilibrium existing between the uid and the solid surface).
This thermal boundarylayer thickness is of the order of one millimeter for air in a
laminar, parallel ow.
In some cases, we choose to average the variations over one, two, or three
spatial coordinate axes. For example, in uid ow and heat transfer in a tube we may
average the heat owacross the tube cross section(intwospace variables). Inanother
example, we may assume local thermal equilibrium and dene a volumeaveraged
local thermal conductivity for a composite such as the berglassair insulation. We
denote the space average (also referred to as spatial average) for a quantity as ),
i.e., with a pair of angular brackets. This is dened as
) =
1
L
_
L
0
dx space averaging. (2.68)
Here we used a length L; however, we may also use an area A, or a volume V, for
the averaging.
EXAMPLE 2.20. FUN
Consider heat ow from a at, semiinnite solid surface to a uniform uid
owing parallel to this surface. The solid surface is at a uniform temperature
T
s
. The uid arrives at the solid surface for the left with a uniform temperature
T
f .
 T
s
. This is shown in Figure Ex. 2.20(a). Assume that the uid velocity is
uniform.
MassoudKaviany 2011
(a) Is the temperature distribution in the uid onedimensional, two
dimensional, or threedimensional?
(b) Qualitatively drawthe temperature distribution at the three locations shown.
Note that heat removal by uid through surface convection is an added heat
to the uid. The uniform uid temperature distribution is shown, for location
x = 0, in Figure Ex. 2.20(a).
Uniform Fluid Temperature
T
f
= T
f,
T
s
= 80 C
Location 1 Location 2 Location 3
T
s
q
k
q
u
q
ku
q
u
x, T
f
T
f,
y
z
Fluid Flow
(Uniform Velocity)
T
f,
= 20 C
Solid
Surface
Solid
Fluid
Figure Ex. 2.20. (a) Uniform uid ow parallel to a solid surface and surface convection.
SOLUTION
As indicated, we have neglected the effect of the presence of the solid surface
on the uid velocity distribution and assumed a onedimensional uid velocity.
(a) As heat ows fromthe surface to the uid by surface convection, it penetrates
the uid by conduction along the y direction and ows along with uid in the
x direction. To a lesser extent, the heat is also conducted along the x direction.
There is no convection along the y direction. Since the solid surface and the
incoming uid have no variation of temperature in the z direction, no heat ows
inthe zdirection. Therefore, the temperature eldinthe uidis twodimensional,
i.e., in x and y directions (in xy plane).
(b) The twodimensional uid temperature distribution, at the three prescribed
locations, is shown in Figure 2.20(b). Note that the fareld uid temperature
remains T
f .
, because the effect of heat owing from the surface is conned to a
small region near the surface. This is stated by (2.66). Note that the uid and solid
temperatures are the same at the surface, due to the continuity of temperature
across the surface. This is stated by (2.64), where T
s
is prescribed (i.e., treated as
a known or a closure condition). As the heat transferred by surface convection
accumulates in the uid, the region affected by this heat transfer, i.e., the uid
boundary layer, grows.
COMMENT
As more and more heat is accumulated in the uid moving along the x direction,
its ability to remove heat from the surface deteriorates. Therefore, parallel ows
are not the most effective owarrangement for surfaceconvection heat transfer.
MassoudKaviany 2011
Thermal Boundary Layer
y
z
x, T
f
u
f,
T
f,
T
f
T
f
T
f
T
f,
T
s
T
f,
T
s
T
f,
T
s
T
s
Parallel (to Surface)
Fluid Flow
Fluid
Solid
Surface
Solid
Figure Ex. 2.20. (b) Fluid temperature distribution at three locations along the surface.
2.5.3 Time Dependence and Steady State
Every system is in a transient behavior, although the temporal changes may be
occurring over elapsed times that are longer than that of practical interest. When
this is the case, then a steadystate behavior is justiably assumed. An example will be
the heat transfer in a desktop computer that is left on for a few hours. Then a steady
state behavior is assumed when the thermal variations caused by the variations in
the computer functions are not substantial compared to the steady thermal behavior.
There are cases in which the transient behavior is of interest as in the temporary
(or sudden) exposure of a person to very cold temperatures. Another example is
the rapid immersion and withdrawal of a piece of rubber in a liquid nitrogen bath in
order to increase its toughness near its surface for drilling and machining purposes.
In some cases, we are interested in the timeaveraged behavior. For example,
consider a time variation that is periodic over a period (s) = 1, = 1,2f , where
f (Hz or 1/s) is the frequency and (rad/s) is the angular frequency. We denote the
time average (also referred to as the temporal average) for a quantity as , i.e.,
with an overbar. This is dened as
=
1
_
0
dt time averaging. (2.69)
This time period can be short (as in reciprocating engines with of the order
of 0.01 s) or very large (as in the global warming in the earths atmosphere). In
convection heat transfer, the uid ow may be turbulent, then the corresponding
timeaverage is taken over the periods of the uctuations in the uid ow. Some of
these periods can be as small as 0.001 s.
Since even steady state, practical uid ows are complex, here we will limit our
treatment only to convective heat transfer with timeaveraged, steadystate ows.
However, when treating with solids, we will treat the transient conduction heat
transfer.
In some transient problems, the temperature variation inside the control volume
is not as signicant as that occurring outside it. Then a uniform temperature is
MassoudKaviany 2011
assumed within the control volume and this is referred to as the lumpedcapacitance
treatment.
EXAMPLE 2.21. FUN
In particle spray coating applications, particles are own and heated in a hot
gaseous stream, as shown in Figure Ex. 2.21(a). Once the gas leaves the nozzle,
its temperature drops due to heat transfer to the particles and entrainment of
the ambient gas. This is shown in Figure Ex. 2.21(b). Using the average particle
velocity u
s
, the time of ight, measured from the location x = 0, is found from
t = x,u
p
. Then the spatial variation of temperatures can be shown as temporal
variations. The solid particle temperature rises due to heating. The surface of the
particle would have a temperature higher than that at the center of the particle.
This is also shown in Figure Ex. 2.21(b). This temperature history for the solid
particle may indicate solid softening, but no wholeparticle melting.
(a) By comparing the given temperature history of (i) the gas, (ii) the particle
surface, and(iii) the particle center, determine the largest temperature difference
u
p
(a) Spray Coating Using Heated Particles
(b) Temperature of Gas and Particle Along the Flight Path
x
r
R
Nozzle
Hot Gas
Particles
Coating
Surface Temperature
of Particle at the
Exit of Nozzle
Gas Temperature
at Exit of Nozzle
Moving
Substrate
Particle Volume
V
s
VolumeAverage
Particle Temperature
T
s
(r = R)
T
s
(r = R, t)
T
f
 T
s
(r = R, t)
T
s
(r = 0)
T
s
(r = 0, t)
T
f
T
s
A
s
R
ku
L
0
T
f
T
f
T
s
L
u
p
t
x
u
p
T
f
T
s
(t)
T
Q
s
Q
sf
T
f
Figure Ex. 2.21. (a) and (b) Temperature of a particle heated in a gaseous stream, where
the gas and particle are discharged through a nozzle.
MassoudKaviany 2011
in this thermal system.
(b) Write the energy equation for the particle.
SOLUTION
(a) As shown in Figure Ex. 2.21(b), the temporal temperature variation of T
f
is
small compared to T
f
T
s
(r = R). This indicates that the heat transfer from the
gas to the surface of the particle Q
f s
can be approximated using a timeaveraged
temperature for the gas, i.e.,
T
f
=
1
L
n
,u
s
_
L
n
,u
s
0
T
f
dt.
where L
n
is the distance between the nozzle and the substrate surface. The other
surface heat transfer is the prescribed heat ow rate Q
s
. Also, the temperature
variation within the particle, i.e., T
s
(r = R. t) T
s
(r = 0. t), is small compared to
the external temperature variation, i.e., T
f
T
s
(r = R. t). This again indicates
the Q
f s
can be approximated using a volumeaveraged temperature for the
solid, i.e.,
T
s
)(t) =
1
V
s
_
V
s
T
s
dV . T
s
(r = R. t).
(b) In Chapter 6, the surfaceconvection heat transfer Q
f s
will be related to
T
f
T
s
)(t) through a surfaceconvection resistance R
ku
(
C/W), i.e.,
Q
f s
=
T
f
T
s
)(t)
R
ku
= Q
sf
.
Then from (2.9), with no energy conversion, we have
Q[
A.s
= Q
sf
Q
s
=
T
f
T
s
)(t)
R
ku
= (c
p
V)
s
dT
s
)(t)
dt
.
COMMENT
When T
f
= T
f
(t) and T
s
= T
s
(x. t) have variations that are of the order of (i.e.,
comparable to) T
f
T
s
), then both the temporal variation of T
f
and the spatial
variationof T
s
shouldbe includedinthe analysis. The ratioof internal conduction
resistance to surfaceconvection resistance is called the Biot number N
ku
= Bi
and will be discussed in Section 6.8. When Bi  0.1, the internal conduction
resistance can be neglected.
2.5.4 Thermal Circuit Models
Although a general analogy between electrical current ow and heat transfer rate
does not exist, the limitedanalogy andmore importantly the ease of use andthe visual
advantage of the thermal circuit presentation allow for an effective problem solving
methodology. Whenthe heat ows by various mechanisms, betweena mediumandits
MassoudKaviany 2011
surroundings, it is compared to electrical current ow. The thermal nonequilibrium
is compared with electrical potential nonequilibrium, the heat ow with current
ow, and the thermal resistance with electrical resistance. Figure 2.17 shows this
thermal circuit model for a thermal node i. This allows for developing thermal
circuit diagrams using thermal potentials for heat ow T
i
T
j
(at thermal nodes
T
i
and T
j
) and thermal resistances R
t.ij
to heat ow (here t stands for the heat
transfer mechanisms, e.g., k for conduction). Here we have written the radiation
heat transfer separately. This is because, as will be shown in Chapter 4, we choose a
different potential when representing Q
r.i
. The summation of currents at the node
(i.e., the Kirchhoff law) is assumed analogous to the energy conservation law. The
node i can be a control surface or a control volume. These analogies are developed
throughout the text. The determination of unknowns in the thermal circuits can be
done by hand calculations for a small number of thermal nodes. Otherwise computer
based solvers (software) can be used. We will use the thermal circuits extensively and
examples are given throughout the text. The thermal resistances are summarized in
Chapter 8 and listed in Table 8.1 and 8.2.
Heat Transfer for Surface i to Surface j
Presented by Resistance
Other Surface
Heat Transfer for Surface i
(Called Prescribed Surface Heat Flow Rate Q
i
)
Surface Radiation
Heat Transfer from Surface i
Volumetric Energy Equation for Thermal Node i
R
t,ij
Q
t,ij
Q
i
Q
r,i
T
j
T
i
Control Surface A
i
and Volume V
i
Volumetric Energy Storage
and Conversion
(c
p
V)
i
+ S
i
dT
i
dt
Q
j
Figure 2.17. Volumetric energy equation for volumetric node i and its thermal circuit model.
2.5.5 Summary of Methodology for Heat Transfer Analysis
The methodology used for heat transfer analysis in the text, follows these general
steps.
(i) Sketch of Thermal Problem
Make a sketch of the physical problem. As much as possible use the appropriate
geometrical scales, i.e., show what is large and what is small.
(ii) Heat Transfer Medium
On the sketch identify the medium or media through which heat ows. Show the
various phases, i.e., gas, liquid, and solids, and show the anticipated hot and cold
regions. Identify the heat transfer objectives.
MassoudKaviany 2011
(iii) Fluid Flow
Identify and show any uid ow and its direction. If thermobuoyancy is signicant,
show the direction of the gravity vector.
(iv) Heat Transfer Mechanisms
Identify and show the signicant mechanisms of heat transfer. The surface area for
each mechanism of heat transfer, i.e., for each heat ux, q
k
, q
u
, q
ku
, and q
r
, may be
different. Identify the areas (crosssectional or surface area) associated with each
mechanism of heat transfer, e.g., A
k
, A
ku
, A
u
, and A
r
.
(v) Energy Conversion Mechanisms
Identify and show the signicant mechanisms of energy conversion (surface and
volume). Determine if a steady state can be assumed or if the volumetric energy
storage/release should be included. Show the locations of various energy storage
and conversions.
(vi) Bounding Surfaces
When there are bounding surfaces, identify and show the direction of heat ow
crossing these surfaces (with respect to the normal unit vector for these bounding
surfaces). When a surface has a uniform temperature, the heat ows perpendicular
to the surface. Identify what is known about the thermal conditions at the surface
(temperature, heat ux, phase change, or energy conversion).
(vii) Heat Flow Direction
Draw (i.e., track) the heat ux vector, starting from the source of energy or the
location of the highest prescribed temperature. For surfaces, also identify Q[
A
. Note
that in the energy equation, the surface net heat ow rate Q[
A
is positive when it
ows out of the control surface. So, if it is known that heat ows into the surface, it
can be represented by using an inward arrow and assigning a negative sign to that
Q[
A
, i.e., using Q[
A
.
(viii) BoundingSurface Energy Equation
When only the heat transfer to and from a bounding surface A
ij
is of interest, the
energy equation is applied to this massless interface. The energy equation is (2.60)
Q[
A
=
_
A
j i
(q s
n
)dA
_
A
ij
(q s
n
)dA
= Q[
A
ij
Q[
A
j i
=
S on A
ij
and A
j i
integralsurface energy equation.
(2.70)
(ix) Volumetric Energy Equation
When there is a volumetric storage or conversion of energy, the integralvolume
energy equation (2.9) is used, i.e.,
Q[
A
=
_
A
(q s
n
)dA= c
p
V
dT
dt
S
integralvolume energy equation
with volumetric terms.
(2.71)
MassoudKaviany 2011
Identify and show the control volume and place a control surface over this volume.
Then use this integralvolume energy equation and write the surface heat trans
fer mechanisms. If there are any energy conversions, write the appropriate energy
conversion relations. If there is storage, state the initial temperature.
(x) Thermal Nodes and Thermal Circuit Analysis
It is convenient to apply the energy equation to a surface or volume using the
concepts of thermal node and thermal resistances and sources. The thermal node
can be a control volume, with its control surface, or it can be a bounding surface (i.e.,
no mass). Write the energy equation for the node. For each node i we can write the
energy equations (2.70) and (2.71) as
Q[
A.i
= Q
i
Q
r.i
j
Q
t.ij
=
S
i
integralsurface energy
equation for thermal node i
(2.72)
Q[
A.i
= Q
i
Q
r.i
j
Q
t.ij
= (c
p
V)
i
dT
i
dt
S
i
integralvolume energy
equation for thermal
node i,
(2.73)
where
Q
t.ij
=
T
i
T
j
R
t.ij
resistive surface heat transfer. (2.74)
For the surface energy equation, A
i
wraps around the bounding surface.
(xi) Numerical Solution: Explicit and Implicit Relations
Determine the number of unknowns and write the sufcient number of relationships
and equations. Then solve the numerical problem.
When the energy equations and the associated relations allow for proper substi
tutions leading to an explicit relation for the unknowns in terms of the knowns, then
the numerical evaluation proceeds in a sequence where each unknown is explicitly
evaluated. However, there are many cases where such explicit relations cannot be
found. Then the unknowns are determined through implicit relations. The implicit
relations are generally solved numerically by iteration with a prescribed accuracy,
which is called the convergence criterion. For a single, implicit equation for a single
unknown, the iterations can be performed by hand. There are many computer algo
rithms, which are called solvers and are used for obtaining simultaneous solutions to
nonlinear algebraic equations or coupled ordinary differential equations (here these
equations are the energy equations and the associated relations). We will discuss
numerical solutions in Section 3.7 and in Chapter 8.
(xii) Advanced Analysis
The differentialvolume (or combineddifferentialintegral length) heat transfer anal
ysis is used here to obtain the thermal resistance R
t.ij
. We begin with the differential
MassoudKaviany 2011
volume energy equation (2.1), i.e.,
_
LA
(q s
n
)dA
LV 0
= q =
t
c
p
T s differentialvolume energy equation, (2.75)
and the stated boundary, i.e., bounding surface, and initial (if transient) conditions,
assign a coordinate axis (or axes), and show the direction of the positive axis (or
axes). The solution to the spatial distributions of the temperature and heat ux vec
tor (and uid velocity vector, if any, from the uid momentum equation) requires
space and time (if transient) integrations. This analysis is generally considered an
advanced analysis and is demonstrated throughout the book with the intention of
developing resistance relations to be used in the circuit analysis. In the transient anal
yses, when needed, we use a combined differentialintegral length energy equation
(Section 2.2.2), examples of which will be given in Chapters 6 and 7.
2.5.6 Solution Format for EndofChapter Problems
The following format is used in the solution manual and is suggested as guide.
(i) Restate Problem Statement (GIVEN)
Reword and restate the problem knowns, assumptions, and simplications.
(ii) Redraw the Physical Problem (SKETCH)
Show the directions of the heat (and when appropriate uid) ow. Identify the
mechanisms of heat transfer and energy conversion.
(iii)Restate Problem Questions (OBJECTIVE)
Write the objectives of the problemand state the questions asked and the unknowns.
(iv) Solve the Problem (SOLUTION)
The solution includes some or all of the following steps.
r
With the appropriate bounding surfaces, dene the control volumes and control
surfaces.
r
Draw the thermal circuit diagram.
r
Write the appropriate relations for energy equation, along with the appropriate
energy conversion and energy storage.
r
Write the appropriate relations for the heat transfer rates and thermal
resistances.
r
Determine the thermochemical and thermophysical properties from available
tables, graphs, and relations.
r
When needed, use a solver (e.g., accompanying MATLAB).
r
Determine the numerical values for the unknowns.
MassoudKaviany 2011
(v) Examine Results and Make Additional Comments (COMMENT)
Examine the numerical results by comparing them to similar available results and to
what is intuitively expected. State what insights you have gained from the exercise.
2.7 References
[1] Angrist, S.W. 1982, Direct Energy Conversion, Fourth Edition, Allyn and Bacon,
Boston.
[2] Ashcroft, N.W., and Mermin, N.D., 1976, Solid State Physics, Saunders Company,
Philadelphia. Also, Kittle, C., 1986, Introduction to Solid State Physics, Wiley,
NewYork. Also, Bube, R.H., 1992, Electrons in Solids, Third Edition, Academic,
San Diego. Also, Kaviany, M., 2008, Heat Transfer Physics, Cambridge, New
York.
[3] Baulch, D.L., et al. 1994, Summary Table of Evaluated Kinetic Data for Com
bustion Modeling: Supplement I, Combustion and Flame, Vol. 98, 11, pp. 5979.
[4] Benedict, R.P., 1977, Fundamentals of Temperature, Pressure, and Flow
Measurements, Second Edition, Wiley, New York.
[5] Decher, R., 1997, Direct Energy Conversion: Fundamentals of Electric Power
Generation, Oxford University Press, New York.
[6] Doebelin, E.O., 1990, Measurement Systems, Application and Design, Fourth
Edition, McGrawHill, New York.
[7] Ensminger, D., 1988, Ultrasonics: Fundamentals, Technology, Applications,
Marcel Dekker, New York. Also, ter Haar, G., 2001, Acoustic Surgery,
Physics Today, December 2001. Also, Cheeke, J.D.N., 2002, Fundamentals and
Applications of Ultrasonic Waves, CRC Press, Boca Raton. Also, Bhatia, A.B.,
1967, Ultrasonic Absorption, Oxford, London.
[8] Fogler, H.S., 1992, Elements of Chemical Reaction Engineering, Second Edition,
PrenticeHall, Englewood Cliffs.
[9] Lehmann, J.F., Editor, 1990, Therapeutic Heat and Cold, Fourth Edition,
Williams & Wilkins, Baltimore.
[10] Metaxas, A.C., and Meredith, R.J., 1983, Industrial Microwave Heating, Institute
of Electrical Engineers, London.
[11] Nussbaum, G., 1982, Physical Aspects of Hyperthermia, American Association
of Physicists in Medicine, American Institute of Physics, New York.
[12] Parker, S.P., Editor, 1991, Encyclopedia of Physics, Second Edition, McGraw
Hill, New York.
[13] Peters, N., and Rogg, B., Editors, 1993, Reduced Kinetic Mechanisms for Appli
cations in Combustion Systems, SpringerVerlag, Berlin.
[14] Pierce, A.D., 1991, Acoustics, An Introduction to Its Physical Principles and
Applications, Acoustical Society of America, Woodbury.
[15] Rowe, D.M., Editor, 1995, CRCHandbook of Thermoelectrics, CRCPress, Boca
Raton. Also, Goldsmid, H.J., 1986, Electronic Refrigeration, Pion, London.
[16] Sahraoui, M., and Kaviany, M., 1994, Direct Simulation vs. VolumeAveraged
Treatment of Adiabatic, Premixed Flame in a Porous Medium, International
Journal of Heat and Mass Transfer, Vol. 17, pp. 28172834.
[17] Seaborg, G.T., 1970, Nuclear Applications and Technology, Vol. 9, pp. 830850.
[18] Siegel, R., and Howell, J.R., 2002, Thermal Radiative Heat Transfer, Fourth
Edition, Taylor & Francis, New York.
[19] Westbrook, C.K., and Dryer, F.L., 1981, Simplied Reaction Mechanisms
for the Oxidation of Hydrocarbon Fuels in Flames, Combustion Science and
Technology, Vol. 27, pp. 8143.
MassoudKaviany 2011
[20] Whitaker, S., 1983, Fundamental Principles of Heat Transfer, Krieger Publishing
Company, Malabar.
2.8 Problems
2.8.4 General
PROBLEM 2.41. FUN
Consider a human body fever resulting in an increase of body temperature from
37
C to 41
S
r.c
)
1
, with a body temperature of T
1
= 37
= 90 W, c
p
= 3,000 J/kgK, T
1
(t = 0) = 37
C, T
= 20
C, Lh
r.ATP
= 7,300
cal/mole ATP, N
c
= 62 10
12
cell/body, D
c
= 7.5 jm, M
ATP
= 0.5032 kg/mole ATP,
ATP
= 1,500 kg/m
3
.
PROBLEM 2.42. FUN
The Seebeck coefcient
S
is dened as L
s
,LT (Figure 2.10) and is generally
a measured quantity. However, there is an approximate prediction based on the
Boltzmann transport equation that gives
S.p
=
k
B
e
c
_
(LE
e.g
)
p
k
B
T
(s
5
2
)
_
.
MassoudKaviany 2011
where k
B
is the Boltzmann constant, e
c
is the charge of an electron, (LE
e.g
)
p
is
the difference between the electron Fermi energy and the valence energy [Figure
Pr. 2.42], and (s 5,2) is interpreted as the electron and hole reduced kinetic energy
transported by the current. The electron Fermi energy represents the energy of the
highest lled electron state in a periodic lattice. Using s = 0.5 indicates that the
electrons and holes are scattered by lattice vibrations (phonons).
Within a ptype semiconductor, there are free electrons and vacant electron sites,
but the number of vacancies is much greater than the number of free electrons. If all
of the free electrons found and lled a vacancy there would still be a net deciency
of electrons within the material. Overall, ptype materials have more positive charge
carriers (holes) than negative charge carriers (electrons). The holes are the dominant
sources of current, whereas in ntype materials, the electrons are.
Conduction
Bands
Fermi Level
Valence Bands
Filled Orbits
in Atom(s)
Electron e
E
l
e
c
t
r
o
n
e
Hole
H
o
l
e
Hot,
T
h
Cold,
T
c
Increasing T,
s
(E
e,g
)
p
T,
s
E
e
,
J
T, K
(i)
(ii)
Acceptor Level
Figure Pr. 2.42. (i) Electron energy states E
e
for ptype semiconductors. (ii) Static electric
potential L
s
caused by a temperature difference LT.
To understand the relationship between temperature and electrical potential,
refer to Figure Pr. 2.42(i). At low temperatures, electrons reside in the valence band.
As the temperature increases, some electrons acquire enough energy to jump the
gap from the valence band into an acceptor level that has been created by adding
impurities to the material. When the electrons leave the valence band, they leave
behind holes, which can be thought of as positive charge carriers. Electrons in the
acceptor level are not free electrons; they are still bond to their atoms. The holes,
however, can be made to move, if electrons are exchanged between adjacent atoms.
If a temperature gradient exists within the semiconductor, more electrons will
be promoted in the warmer region than in the cooler region. The promoted electrons
leave behind energetic holes, which carry heat because of their kinetic energy. Heat,
and therefore holes, will tend to ow toward the cooler region of the semiconductor.
However, the holes also carry charge. The ow of positive charge away from the
warm charge will accumulate at the cold end. This sets up an electric eld within the
semiconductor. This is shown in Figure Pr. 2.42(ii).
For ptype Bi
2
Te
3
, L(E
e.g
)
p
= 0.13 eV. Find the Seebeck coefcient for this
material, and compare the result to the value in Table C.9(a). Comment on any
differences.
MassoudKaviany 2011
s = 0.5. T = 300 K. Use Table C.1 in Appendix C.
PROBLEM 2.43. FUN
The discrete energy a photon carries is related to the frequency of the radiation by
E
ph.
= h
P
f = h
P
c
.
where h
P
is the Planckconstant, f is the frequency of the radiation, is the wavelength
of the radiation, and c is the speed of light. Semiconductors are characterized by an
energy gap between the valence band, LE
e.g
, which holds bond electrons, and the
conduction band, where electrons are free to move around. This energy gap is also
called the forbidden band, as shown in Figure Pr. 2.43.
E
e
,
J
}
}
}
Valence Bands
Filled Orbits
in Atom(s)
Conduction
Bands
Forbidden Band,
E
e,g
Figure Pr. 2.43. Electron potential energy states and conduction and valence electrons.
An incident photon with sufcient energy can promote an electron across the
band gap, thereby freeing it in a current.
(a) Undoped silicon has a band gap energy LE
e.g
= 1.11 eV at T = 300 K. Using the
equation for the energy of a photon, determine the maximum wavelength required
for a single photon to promote an electron across the gap. Using Figure 4.1, classify
the range of radiation.
(b) Crystalline silicon is transparent to farinfrared radiation. Can you explain why?
(c) Amorphous sodalime glass is transparent to visible light. What can you conclude
about its gap energy? Give a quantitative description for this energy, i.e., estimate
LE
e.g
(eV).
(d) If a 100watt light bulb was constructed in such a way that it only emitted blue
light of wavelength
blue
= 400 nm, how many photons would it emit per second?
What if it were emitting only red light at
red
= 700 nm? First determine which
wavelength produces more photons per second, then compare with predictions.
PROBLEM 2.44. FUN
In quantum mechanics, an electron is treated as a wave governed by the Schr odinger
wave equation. In its onedimensional timeindependent form, this is
_
h
2
P
8
2
m
d
2
dx
2
_
= E
e
.
MassoudKaviany 2011
where E
e
is the energy of the electron, is potential energy, is wave potential.
The crystalline lattice of a semiconductor serves as a periodic potential for
electrons. A simple, onedimensional model for this is the KronigPenney Model,
which assumes square potential wells of a nite height,
0
. The wells have width b,
and are spaced a distance a. This is shown in Figure Pr. 2.44.
a b a + b = l
m
(x)
0
x
0
b
l
m
(x)
Figure Pr. 2.44. Periodic potential for electron in a lattice.
Notice that the potential repeats every a bunits, which is the lattice spacing of
the crystal l
m
.
In the limit barrier height
0
increasing without limit, while the width bdecreases
to zero, the wells become delta functions, we have [2]
cos kl
m
= (P,l
m
) sinl
m
cos l
m
. for b 0.
0
.
The term = (2mE
e
, h
2
P
)
1,2
is related to the electron momentum, l
m
is the lattice
constant, and P is an indication of the strength of the electron wave potential. Note
that this equationis only satisedwhen1 (P,l
m
) sinl
m
cos l
m
1. All other
values of E
e
= l
m
are not allowed, and these disallowed regions are what give rise
to the energy gaps in semiconductors.
(a) Plot (P,l
m
) sinl
m
cos l
m
for l
m
> 0, for P = 6. Treat l
m
as a single param
eter x.
(b) Highlight the allowed regions in the plot.
(c) Comment on the width of the allowed regions as l
m
increases.
PROBLEM 2.45. FUN
The acoustic attenuation coefcient
ac
of gases, liquids, and solids varies by several
orders of magnitude. This is related to the rate of conversion of the acoustic waves
(a mechanized energy) into the thermal energy. Consider nitrogen (diatomic) gas
at T = 298 K and p = 1 atm, and ultrasonic waves with f = 1 MHz. The measured
attenuation coefcient for nitrogen under these conditions is given as
ac
= 16.4 1/m
in Table C.11.
(a) Determine the speed of sound in nitrogen gas under these conditions.
(b) Determine the acoustic attenuation coefcient
ac
of nitrogen gas using (2.55),
and compare this with the measured result of Table C.11.
(c) Determine the acoustic heating rate s
m.ac
using an acoustic intensity I
ac
=
300 W/m
2
.
(d) Determine
ac
for argon (monoatomic) gas.
MassoudKaviany 2011
PROBLEM 2.46. FUN
A quartz slab at T = 100 K is used to test the acoustic attenuation coefcient
ac
using ultrasonic waves with f = 1 GHz. The acoustic attenuation coefcient can be
obtained by [Section 2.3.3(D)]
ac
=
3
2
G
(2f )
2
kT
a
5
s
tan
1
(4f
p
)
4f
p
for f
p
> 1.
where T is the absolute temperature, k is the thermal conductivity,
p
is the phonon
relaxation time, and
G
is the Gr uneisen constant dened as
G
=
3
s
E
s
c
:
.
(a) Determine the acoustic attenuation coefcient
ac
for quartz at T = 100 K.
(b) Compare this result with the experimental value of 36 1/m.
Assume isotropic properties (quartz is an anisotropic crystal) and T = 100 K. Use
the thermal conductivity k = 69 W/mK, phonon relaxation time
p
= 1.5 10
11
s,
density = 2,650 kg/m
3
, speed of sound a
s
= 5,000 m/s, specic heat capacity c
:
=
2075 J/kgK, linear thermal expansion coefcient
s
= 7.5 10
6
1/K, and Young
modulus E
s
= 97.2 10
9
Pa.
MassoudKaviany 2011
3
Conduction
3.3 SteadyState Conduction
3.3.4 Conduction Thermometry
We now consider two heat transfer devices, one for the measurement of the surface
heat ux and one for the measurement of the conductivity.
(A) Heat Flux Meter
In order to measure the rate of heat ow through surfaces, a heat ux meter (also
called a heat ux sensor) can be mounted on the surface. The sensor measures the
temperature difference across a reference layer (a dielectric, such as a polymer, is
used, which also has a low thermal conductivity). It is desirable that the thermal
resistance of this sensor to the heat ow be negligible such that it would not notice
ably hinder the heat ow to the surface. High conductivity adhesives are used for
the surface mounting of the sensor. Figure 3.22(a) renders a mounted sensor. The
thickness of the sensor is a fraction of one mm and the resistance A
k
R
k.s
, where
A
k
is the sensor heat ow crosssectional area, is about 0.2
C/(W/m
2
). Therefore,
this sensor will be most useful for mounting on surfaces having resistances R
k
such
that A
k
R
k
A
k
R
k.s
= 0.2
C/(W/m
2
). The temperature difference across the refer
ence layer is measured. The output of the thermopiles [a collection of thermocouples
showninFigure 3.22(c)] across the reference layer is the Seebeckpotential (discussed
in Section 2.3.2) and varies almost linearly. The sensor may use chromelalumel ther
mocouples (designated as Ktype in Figure 2.10) and at room temperature may
have an output of about 10 jV/(W/m
2
), which has the units of L,q
s
, where q
s
is
the surface ux. The output of a typical sensor, i.e., L,q
s
, is also shown in Fig
ure 3.22(c) as a function of the surface temperature T
s
. This variation is due to the
temperature dependence of the Seebeck coefcient
S
and the conductivity of the
polymer k
s
.
MassoudKaviany 2011
q
s
Polymeric Wrapper
ThinFoil Thermopiles
(Collection of Thermocouples)
R
k,s
T
s
A
k
HeatFlux Meter
(a) HeatFlux Meter
(b) Differential Thermopile
Polymer Core
Heat Transfer Surface
R
k
Material B
Comparative Size
Material A
(+)
(+)
Hot Junctions
Polymer
Cold Junctions
Thermopile
Thermocouple
()
() (+) () (+) () (+) ()
, Measured Voltage
q
s
q
k
(c) Typical Electrical Output and Its Temperature Dependence
184
6
8
10
12
14
16
18
/
q
s
,
V
/
(
W
/
m
2
)
T
s
, C
112 73 32 38 93 149
Experiment
204
Figure 3.22. (a) Surface heat ux meter. (b) Meter electrical output. (c) Typical electrical
output, i.e., electrical potential per unit heat ux, as a function of temperature.
(B) ThermalConductivity Meter
The thermal conductivity of various materials is measured by the thermal conduc
tivity meter (also called a device). In general, devices are divided into steadystate
and transient. Here we will discuss a steadystate device, and later we will discuss a
transient device (Section 3.5.3).
MassoudKaviany 2011
Bottom
Heater
Liquid Cooled
Heat Sink
T
A
x
Clamp
Top Guard
Heater
Bottom Guard
Heater
Thermocouple
Thermocouple
(a) Rendering of
Conductivity Meter
(c) Thermal Circuit
Diagram
(b) Physical Model
Reference
Material A
Surface Area A
Reference
Material B
Test Material
Contact Pressure
p
c
T L
L
A
L
B
L
A
k
A
A
k
T
B
T
A
T
T
B
Insulation
R
Coolant
Flow
Coolant
Flow Coolant
Q
k
Q
k
() (+)
()
(+)
L
B
k
B
A
k
L
kA
k
q
k
q
k
q
k
q
k
q
ku
Top Heater
q
u
S
e,J
S
e,J
Figure 3.23. A thermal conductivity meter based on guarded, comparative longitudinal heat
ow. (a) Rendering of a conductivity meter. (b) Physical model. (c) Thermal circuit diagram.
We consider the standard test method of thermal conductivity of solids by means
of the guarded, comparative longitudinal heat ow technique (American Society of
Testing and Materials, ASTM, test number E1225) [28]. This method is suitable
for specimens with intermediate conductivity (0.2 k 200 W/mK) and in the
temperature range between 90 to 1300 K. This covers most solids, i.e., polymers,
ceramics, and metals as well as composites. A schematic of the test apparatus is
given in Figure 3.23(a).
The test specimen is a circular cylinder with the length nearly the same as the
diameter. This test specimen is clamped between two identical reference pieces (or
specimens) of known thermal conductivity, k
A
and k
B
, and length, L
A
and L
B
, and
of identical crosssectional area A
k
. A temperature gradient LT,L and an average
temperature LT are established for the test specimen by controlling the top heater
to provide a heat ow toward the bottom heater and the liquidcooled heat sink.
The bottom heater provides the control for the average temperature. This results in
heat ow through the test specimen, as well as the two reference specimens. This is
shown in Figure 3.23(b). The radial (i.e., lateral) heat ow (heat losses) is minimized
by placing guard heaters around the stack of the test and reference specimens. These
active guard heaters are adjusted to the temperature gradient in the stack.
This series arrangement of the resistance is shown in Figure 3.23(c). Then with
negligible lateral heat losses, the steadystate heat owrate throughthe tworeference
MassoudKaviany 2011
specimens should be equal. Since parasitic heat losses are present, an average of the
heat ow rate Q
k
= A
k
q
k
through specimens A and B is used, i.e.,
Q
k
= q
k
A
k
=
1
2
_
A
k
k
A
LT
A
L
A
A
k
k
B
LT
B
L
B
_
. (3.91)
This heat also ows through the test specimen and we have
Q
k
= A
k
q
k
= A
k
k
LT
L
. (3.92)
Now by solving for k between these two equations and noting that the cross
sectional areas through which the heat ows are equal for all specimens, we have
k =
L
LT
1
2
_
k
A
LT
A
L
A
k
B
LT
B
L
B
_
. (3.93)
The larger the magnitudes of L, L
A
, L
B
, LT, LT
A
, and LT
B
, the lower will be
their uncertainties. However, the larger Ls and LTs, the more pronounced will be
the heat losses. In general, L
A
and L
B
are selected to be of the order 1 cm and each
LT of the order 10
C, T
2
= 40
C.
D
2
>> D
1
D
1
T
1
= T
ls
A Large Piece of Copper,
Ideally Insulated
(Except for Droplet Location),
k, T
2
A Freezing
Hemispherical
Water Droplet
(a) Physical Model (b) Thermal Circuit Model
No Contact
Resistance
R
k,c
= 0
Q
k,12
R
k,12
V T
1
= T
ls
T
2
S
ls Q
1
= 0
Figure Ex. 3.15. (a) A water droplet freezing on a cold copper cylinder. (b) Thermal
circuit model for the heat transfer.
SOLUTION
(a) The thermal circuit diagram is shown in Figure Ex. 3.15(b). There is no
prescribed heat transfer rate, i.e., Q
1
= 0. The conduction heat loss from the
droplets results in its phase exchange and during the solidication the droplet
temperature is assumed to remain the same (i.e., no liquid subcooling).
(b) The energy equation for the droplet is obtained from Figure Ex. 3.15(a) as
Q
k.12
=
S
ls
.
MassoudKaviany 2011
The conduction heat transfer is determined from (3.128), i.e.,
Q
k.12
=
T
1
T
2
R
k.12
.
where R
k.12
is the conduction resistance given in Table 3.3(b) for surface objects.
From Table 3.3(b) for a surface object having a circular cross section in
contact with a semiinnite surface (since D
2
D
1
), we have
R
k.12
=
1
2kD
1
Table 3.3(b).
The thermal conductivity of copper at T = 250 K is found from Table C.14 as
k = 406 W/mK Table C.14.
Then
R
k.12
=
1
2 406(W/mK) 0.014(m)
= 0.08797 K/W.
Q
k.12
=
0(
C) (40)(
C)
0.08797(K/W)
= 454.7 W.
From Tables C.4 and C.23, for water we have at T = 270 K
Lh
sl
= 3.336 10
5
J/kg Table C.4
1
= 1,002 kg/m
3
Table C.23.
From Table 2.1, we have
S
ls
=
M
sl
Lh
ls
.
For a droplet of mass M
1
freezing over an elapsed time Lt, we have
S
ls
=
M
1
Lt
Lh
ls
=
1
V
1
Lt
Lh
ls
=
1
D
3
2 6
Lh
ls
1
Lt
.
Using the energy equation we have
Q
k.12
=
S
ls
=
1
D
3
12
Lh
ls
1
Lt
or
Lt =
1
D
3
12
Lh
ls
Q
k.12
MassoudKaviany 2011
and since Lh
ls
= Lh
sl
, we have
Lt =
1
D
3
12
Lh
ls
Q
k.12
=
1,002(kg/m
3
)
(0.014)
3
(m
3
)
2 6
3.336 10
5
(J/kg)
454.7(W)
= 0.5280 s.
COMMENT
This is a rather short freezing time. If the droplet was initially at a temperature
higher than the freezing temperature, the sensible heat storage would also need
to be included. The results of Section 3.6.2 may be used for this case.
3.5 DistributedCapacitance (Nonuniform Temperature)
Transient: T = T(x. t)
3.5.3 TimeDependent Surface Temperature: SemiInnite Slab
It can be shown that when the surface temperature is a prescribed function of time,
T
s
= T
s
(t), the solution to the onedimensional, transient temperature distribution
for a semiinnite slab is [17]
T(x. t) = T(t = 0)
x
2()
1,2
_
t
=0
T
s
() T(t = 0)
(t )
3,2
exp
_
x
2
4(t )
_
d.
(3.155)
This reduces to (3.142) when T
s
is constant.
The heat ux at the surface is found from (3.143) and by differentiating the
above expression with respect to x and then evaluating the resultant at x = 0. The
result is
q
ck
= q
ck
(t) =
k
2()
1,2
_
t
0
T
s
() T(t = 0)
(t )
3,2
d. (3.156)
Again, this reduces to (3.144), when T
s
is constant.
It can also be shown that when the surface temperature varies in a piecewise
linear manner, as shown in Figure 3.34, the solution for T(x. t) and q
ck
(t) are
T(x. t) T(t = 0) =
n1
i=1
__
(a
i
a
i1
)
_
t
x
2
2
_
(b
i
b
i1
)
_
_
erf
_
x
[4(t t
i
)]
1,2
_
erf
_
x
(4t)
1,2
__
x
2
1,2
(a
i
a
i1
)
MassoudKaviany 2011
T
s,o
t
o
= 0
T
s,1
T
s,2
T
s
= a
n
t + b
n
T
s,n
T
s
t
1
t
2
t
3
t
i
t
n1
t
t
n
x
SemiInfinite Slab
Discrete Variation of Surface Temperature with Respect to Time T
s
= T
s
(t)
q
k
q
ck
T
s
= T
s
(t)
Figure 3.34. Prescribed, piecewise linear variationof the surface temperature of a semiinnite
slab.
_
[4(t t
i
)]
1,2
exp
_
x
2
4(t t
i
)
_
(4t)
1,2
exp
_
x
2
4t
___
[a
n
t b
n
T(t = 0)] erfc
_
x
(4t)
1,2
_
a
n
x
2
1,2
_
(4t)
1,2
exp
_
x
2
4t
_
x
1,2
erfc
_
x
(4t)
1,2
_
_
(3.157)
q
ck
(t) =
T
s
(t) T(t = 0)
A
k
R
ck
(t)
= k
T(x. t)
x
[
x=0
=
n1
i=1
_
[(a
i
a
i1
)t (b
i
b
i1
)]
_
_
kc
p
(t t
i
)
_
1,2
_
kc
p
t
_
1,2
_
c
p
2
1,2
(a
i
a
i1
){[4(t t
i
)]
1,2
(4t)
1,2
]
_
[a
n
t b
n
T(t = 0)]
_
kc
p
t
_
1,2
a
n
_
kc
p
t
_
1,2
. (3.158)
where
a
n
=
T
s.n
T
s.n1
t
n
t
n1
. b
n
= T
s.n
a
n
t
n
.
Equation (3.157) reduces to (3.142) when T
s
is constant.
These relations are useful in the numerical integration of the nodal energy
equation (which will be used in Chapter 8).
MassoudKaviany 2011
3.5.4 Thermal Diffusivity Meter
The thermal diffusivity of materials is measured from transient conduction behavior
of these materials when they are subjected to sudden (or gradual) changes in their
thermal boundary conditions. One of the techniques is the ash method of measure
ment of thermal diffusivity (ASTM test number E1461) [29]. This method is suitable
for opaque (i.e., radiation does not penetrate far into the material), homogeneous
(i.e., uniformproperty throughout the specimen) solids with thermal diffusivity rang
ing from 10
7
to 10
4
m
2
/s and in the temperature range 100 to 2,500 K, in vacuum
or in an inert gaseous environment. Figure 3.35 gives a schematic of the meter and
its principles. In this method, a small, thin (thickness L) disk specimen is subjected
to a high intensity, very short duration Lt irradiation Q
r.i
, i.e., Q
r.i
, on one of its pla
nar surfaces (x = 0). This heat input raises the surface temperature and this change
penetrates into the specimen and reaches the other planar surface (back surface at
x = L). The thermal diffusivity is determined from the examination of the measured
temperature (by an infrared detector or by thermocouples) history of the back sur
face. This temperature history is also shown in Figure 3.35 and has the following
characteristics. If the pulse is very rapid (Lt 0), then the surface temperature
at x = L begins rising monotonically at t = t
o
, where t
o
is the time it takes for the
thermal penetration (or thermal boundary layer) to reach x = L. If the pulse length
is not very short, a maximum is found shortly after t
o
and the temperature drops for
a short time before it begins to rise monotonically thereafter.
After a long elapsed time, the surface temperature at x = L, T(x = L. t) will
reach a maximum T
max
and will remain at this value if there are no heat losses. When
there are heat losses by radiation, conduction, and convection (radially and also
from the front and back surfaces), the temperature begins to drop after reaching the
maximum value. These nonideal effects are also shown in Figure 3.35. The solution
to (3.134) for the transient, onedimensional heat ow, i.e., T(x. t), for a very short
period of heat input is given by [19]
T(L. t) T(t = 0)
T
max
(L. t) T(t = 0)
= 1 2
i=1
(1)
i
e
2
2
i
2
t,L
2 shorttime solution
in terms of T
max
(L).
(3.159)
The results of (3.159) are also plotted in Figure 3.35. The time is normalized with
respect to the elapsed time required for T(L. t) to reach T
i
(T
max
T
i
),2. This
time is called the half time t
1,2
. This half time is found from (3.159) and is given as
[18]
t
1,2
L
2
= Fo
1,2
= 0.1388 relation used for determination of . (3.160)
Then when t
1,2
is measured in the experiment, and L
2
is known, is determined
from this equation.
The measurement uncertainties are due to the measurement of L (generally
selected to be between 1.5 to 4 mm), which appear as L
2
in (3.160), and due to the
measurement of T(x = L. t). The magnitude of T
max
T(t = 0) is kept between 2
to 5
T
(
t
=
0
)
T
m
a
x
(
x
=
L
,
t
)
T
(
t
=
0
)
Prediction
Scaled Temperature
History
0
t
o
+ t
1/2
t
T
max
+ T(t = 0)
2
[T = T(t = 0) + ]
T
max
T(t = 0)
Ideal
(Prediction)
Due to Finite
Pulse Period
Due to
Heat Losses
t
o
T
max
T(t = 0)
2
T(t = 0) +
T(x = L, t)
q
k
k, c
p
q
r
Figure 3.35. (a) A schematic of the ash thermal diffusivity meter. (a) Dimensional tempera
ture response. (c) Dimensionless temperature response.
3.6 LumpedCapacitance (Uniform Temperature) Transient:
InternalExternal Conduction Number N
k.i
 0.1. T = T(t)
3.6.3 Multiple Nodes
When we consider transient heat transfer among multiple media, with each medium
represented by a transient, lumped node, the coupled (i.e., dependent) energy equa
tions (3.161) must be solved. These are coupled ordinary differential equations that
need to be solved simultaneously. There are various numerical methods suggested
for their solutions and these are discussed in [20,23]. Depending on the expected
time variation of the temperatures (e.g., sudden changes due to transient energy
conversions), a suitable method is selected.
The time integration is made by discretizing the time into small increments Lt.
Arelatively high order approximation, which is necessary for dealing with rapid time
variations and for cases where the time derivative can change rapidly, is described
in [20]. This is called the BaderDeuhard method (BDM). In this semiimplicit
MassoudKaviany 2011
method, Lts are selected by monitoring the local truncation errors resulting from
choosing a nite Lt instead of an innitesimal Lt.
Once the initial condition for each node, T
i
= T
i
(t = 0), is given, the time inte
gration can be made using this numerical approximation method. This is one of the
methods used in MATLAB.
EXAMPLE 3.22. FAM.S
A solidied tin solder droplet, in the form of a hemisphere, initially at T
1
(t = 0),
is heated to its melting temperature T
sl
using a carbon steel AISI 1010 hot
needle. This is shown in Figure Ex. 3.22. The needle is initially at T
2
(t = 0).
Assume uniform temperatures for the solder drop and for the needle (N
k.1
 1,
N
k.2
 0.1). The heat is transferred from the needle to the solder drop through a
contact resistance (R
k.c
)
12
, which is assumed to remain constant. All other heat
transfer is assumed negligible.
(a) Draw the thermal circuit diagram.
(b) Determine the elapsed time needed to reach the melting temperature of the
solder droplet T
lg
.
(c) Plot the variation of temperatures T
1
(t) and T
2
(t), with respect to time.
(d) Show that N
k
 0.1 for the solder sphere and the needle.
T
1
(t = 0) = 20
C, T
2
(t = 0) = 500
C, (R
k.c
)
12
= 10
4
K/W
L
Thermal Contact
Resistance
(R
k,c
)
12
(a) Physical Model
(b) Thermal Circuit Model
Ideally Insulating
Surface
A Hemispherical Tin Solder
Droplet with Diameter, D
1
, and
Initial Temperature, T
1
(t = 0)
Printed Circuit
Board
Leads
Hot, Carbon
Steel Needle
T
2
(t = 0)
D
2
(c
p
V)
1
dT
1
dt
(c
p
V)
2
dT
2
dt
T
1
(t)
T
2
(t)
(R
k,c
)
12
Q
k,12
Q
2
Q
1
Figure Ex. 3.22. (a) A hot needle used to mold solder droplet. (b) Thermal circuit model.
SOLUTION
(a) Since we are only interested in T
1
T
lg
, the phase change energy conversion
is not addressed and Figure Ex. 3.22(b) shows the thermal circuit model for
two bodies (nodes) each having a uniform, timedependent temperature and
exchanging heat through a conduction contact resistance.
MassoudKaviany 2011
(b) From (3.161), for the two nodes in Figure Ex. 3.22(b), we have
Q
1
Q
k.12
=
d
dt
(c
p
V)
1
T
1
S
1
Q
2
Q
k.21
=
d
dt
(c
p
V)
2
T
2
S
2
.
where we have noprescribedheat transfer rates, i.e., Q
1
= Q
2
= 0, andnoenergy
conversion, i.e.,
S
1
=
S
2
= 0.
For the conduction heat transfer rate, we have
Q
k.12
=
T
1
(t) T
2
(t)
(R
k.c
)
12
.
Then by assuming constant (c
p
V)
i
, we have
T
1
(t) T
2
(t)
(R
k.c
)
1.2
= (c
p
V)
1
dT
1
dt
T
2
(t) T
1
(t)
(R
k.c
)
1.2
= (c
p
V)
2
dT
2
dt
.
The volumes are
V
1
=
D
3
1
2 6
. V
2
=
D
2
2
4
L
2
V
1
=
(10
3
)
3
(m
3
)
2 6
= 2.617 10
10
m
3
V
2
=
(0.5 10
3
)
2
(m
2
) 0.015(m)
4
= 2.944 10
9
m
3
.
The thermophysical properties of tin and carbon steel AISI 1010 are given in
Table C.16 as
tin :
1
= 7,310 kg/m
3
. c
p.1
= 227 J/kgK Table C.16
carbon steel AISI 1010 :
2
= 7,830 kg/m
3
. c
p.2
= 434 J/kgK. Table C.16
Then
T
1
(t) T
2
(t)
10
4
(K/W)
= 4.343 10
4
(J/K)
dT
1
(t)
dt
T
2
(t) T
1
(t)
10
4
(K/W)
= 1.000 10
2
(J/K)
dT
2
(t)
dt
.
The above two, coupled rstorder ordinary differential equations are solved
using a solver such as MATLAB using the initial temperatures
T
1
(t = 0) = 20
C = 293.15 K
T
2
(t = 0) = 500
C = 773.15 K.
Since the time at which T
1
= T
lg
= 505 K is not known, the integration is
allowed up to the time at which this condition occurs.
MassoudKaviany 2011
The results show that
T
1
(t) = T
lg
when t = 3.399 s.
(c) The variations of T
1
(t) and T
2
(t) with respect to time are shown in Figure
Ex. 3.22(c).
Note the temperature of the needle does not change noticeably (due to
larger c
p
V).
0 0.68 1.36 2.04 2.72 3.40
250
360
470
580
690
800
T
2
(t)
T
1
= T
ls
= 550 K
T
1
(t = 0) = 293.15 K
T
2
(t = 0) = 773.15 K
T
1
,
T
2
,
K
t, s
T
2
(t = 3.399 s) = 761.9 K
T
1
(t)
Figure Ex. 3.22. (c) Variation of the temperature of the solder drop and the needle, with
respect to time.
(d) The internalexternal conduction number is given by (3.161), i.e.,
N
k.1
=
R
k.i
R
k.ij
.
Then for tin hemisphere, we have (choosing the radius as the conduction length)
R
k.1
=
D
1
,2
D
2
1
4
k
s.1
=
2
D
1
k
s.1
hemisphere.
and for the carbon steel cylinder, we have, again choosing the length as the
conduction length,
R
k.2
=
L
D
2
2
4
k
s.2
=
4L
D
2
2
k
s.2
cylinder.
Then from Table C.14 and C.16, we have
tin for T = 400 K : k
s.1
= 62 W/mK Table C.14
carbon steel 1010 : k
s.2
= 64 W/mK Table C.16.
MassoudKaviany 2011
Then
N
k.1
=
2
D
1
k
s
1
(R
k.c
)
12
=
2
10
3
(m) 62(W/mK)
1
10
4
(K/W)
= 1.027 10
3
 0.1 for solder sphere
N
k.2
=
4 1.5 10
2
(m)
(5 10
4
)
2
(m
2
) 64(W/mK)
1
10
4
(K/W)
= 5.965 10
5
 0.1 for needle.
COMMENT
This problem could have been solved by initially comparing (c
p
V)
1
and
(c
p
V)
2
and then noting that (c
p
V)
1
(c
p
V)
2
; we would have assumed
that T
1
remains constant. Then the closedform solution (3.172) could be used
for T
1
(t).
Note that using (3.172), the time constant
1
given by (3.173) is
1
= (c
p
V)
1
R
k.12
= 4.343 10
4
(J/K) 10
4
(K/W)
= 4.343 s.
Then the time constant required to reach T
1
= T
lg
is nearly one time constant.
3.7 Discretization of Medium into FiniteSmall Volumes
In the distributedcapacitance treatment of Section 3.5, we considered the special
boundingsurface conditions of prescribed T
s
or q
s
. These lead to the closedform
solutions given in Table 3.5. The more general boundingsurface conditions, includ
ing those with surface heat transfer represented by resistances, are generally treated
numerically. The transient, distributed energy equation (3.161) can be solved numer
ically. One of the practical, numerical methods is the nitesmall volume method.
In this method, we discretize the heat transfer medium, as was shown in Figure
2.5, and also in Figure 3.39 (for a onedimensional heat ow) and Figure 3.40 (for
a twodimensional heat ow). Because of this spatial discretization, and because it
assumes that the increments are small but nite (as compared to innitesimal), this
is called the nitesmall volume numerical analysis of the energy equation. In Sec
tion 2.2.3 we discussed dividing the medium into small volumes and then wrote the
energy equation assuming a uniform temperature within each nitesmall volume.
This would require that for each of the volumes we have N
k
 0.1, i.e., the internal
conduction resistance is negligible. Then the dimension of this nitesmall volume is
taken such that N
k
 0.1.
Here as examples, we consider a one and a twodimensional, transient conduc
tion in a planar geometry (Cartesian coordinate system). We use the conduction
resistance for the conduction heat transfer among adjacent nitesmall volumes. In
MassoudKaviany 2011
x
y
z
KEY: T
i
xDirection Location Index
L
z
L
y
L
x
Node i1
Node i+1
FiniteSmall Volume
Located at (x): (i)
Node i
(R
k
)
i to i 1
(R
k
)
i to i +1
L
z
with Temperature T
i
Q
t,0
(Q
k
)
01
(Q
t
)
0
Q
o
S
o
Q
o
T
i1
T
i T
i+1
T
N
T
N+1
x
x/2
(Q
t
)
N+1 to
Q
N+1
OneDimensional, Transient Heat Flow: SmallFinite Volume Analysis
L
y V
i
= x L
y
L
z
x
Volumetric Terms
A
k,x
= L
y
L
z
(c
p
V)
i
+ (s V)
i
dT
i
dt
T
0
T
1
x/2
(Q
k
)
i to i+1
(Q
k
)
i to i 1
T
i 1
T
i
T
i+1
Figure 3.39. Finitesmall volume analysis of onedimensional, conductive heat transfer in a
discretized heat transfer medium. The volumetric and boundary nodes are shown.
Problem 3.28 we show that the resistance expressions in Table 3.2 also apply to very
small volumes.
3.7.1 OneDimensional, Transient Conduction
We begin with a onedimensional, transient conduction with a planar geometry, as
shown in Figure 3.39. The heat ow is assumed important and dominant in the x
direction. The heat transfer medium is discretized as shown in Figure 3.39. The rst
nite (but small) volume or node is designated as i = 1 and is to the left. This volume
is at a uniform temperature T
1
and has a volume LV
1
= LxL
y
L
z
. This is because
of the prescribed nite lengths L
y
and L
z
and the xdirection nitesmall length of
Lx. This node is called an interior node (as compared to bounding surface node).
The second node on the left is node 2 and it is also an interior node with again
LV
z
= LxL
y
L
z
and has as its left neighbor T
1
and its right neighbor is T
3
. Using
MassoudKaviany 2011
the index i, the left neighbor is T
i1
and the right neighbor is T
i1
, as shown in
Figure 3.39.
Here we consider the division of the heat transfer medium into equal, small
volumes. In Section 3.7.3 we will discuss nonuniform divisions. The length L
x
is
divided into N increments with Lx dened as
Lx
L
x
N
dividing L
x
into N equalvolume nodes. (3.175)
We now proceed with the transient energy equation for each of the nodes. Then
we discuss the solution to these coupled ordinary differential equations.
(A) Energy Equation for Interior Nodes
We start with the energy equation (3.161) written for a onedimensional heat ow.
We write the surface heat ow rates, for each node i, with conduction as the only
mechanism. This is shown in Figure 3.39. The result is
Q [
A.i
=
j
Q
k.ij
= (Q
k
)
i to i1
(Q
k
)
i to i1
= (c
p
LV)
i
dT
i
dt
s
i
LV
i
energy equation for node T
i
. (3.176)
(Q
k
)
i to i1
=
T
i
T
i1
(R
k
)
i to i1
. (Q
k
)
i to i1
=
T
i
T
i1
(R
k
)
i to i1
. (3.177)
where A
k.x
= L
y
L
z
and (R
k
)
i to i1
= Lx,L
y
L
z
k, etc.
Substituting these resistances into the energy equation (3.176), we have
T
i
T
i1
(R
k
)
i to i1
T
i
T
i1
(R
k
)
i to i1
= A
k
k
T
i
T
i1
Lx
A
k
k
T
i1
T
i
Lx
= (c
p
LV)
i
dT
i
dt
s
i
LV
i
energy equation
for node T
i
.
(3.178)
This is an ordinary differential equation and is referred to as an initialvalue
problem.
(B) BoundingSurface Energy Equation
For the bounding surfaces, nodes i = 0 and i = N 1, we write the surface energy
equation (2.62) as
Q
0
(Q
k
)
01
(Q
t
)
0
=
S
0
Q
N1
(Q
k
)
N1 to N
(Q
t
)
N1 to
=
S
N1
energy equations for
boundingsurface nodes.
(3.179)
The energy balance on the left boundary is also shown in Figure 3.39. Here Q
0
andQ
N1
are the prescribedsurface heat transfer rates. Also(Q
t
)
0
and(Q
t
)
N1 to
are the resistancebased surface heat transfer rates to the external ambients of
surfaces 0 and N 1. These can be by mechanisms of conduction, surface convection,
MassoudKaviany 2011
and/or surface radiation. Also,
S
o
and
S
N1
are surface energy conversions (such as
surface friction heating). The conduction heat transfer to the interior nodes, (Q
k
)
01
and (Q
k
)
N1 to N
, and the external heat transfer rates, are given by
(Q
k
)
01
=
T
0
T
1
(R
k
)
01
= A
k
k
T
0
T
1
Lx,2
. (Q
t
)
0
=
T
0
T
(R
t
)
0
(Q
k
)
N1 to N
=
T
N1
T
N
(R
k
)
N1 to N
= A
k
k
T
N1
T
N
Lx,2
. (Q
t
)
N1 to
=
T
N1
T
(R
t
)
N1 to
.
(3.180)
For the case of a prescribed surface temperature, say T
at i = 0 and/or at
i = N 1 surfaces, we set the resistance (R
t
)
01
and/or (R
t
)
N1
equal to zero.
Then, for example, we will have T
0
= T
at surface i = 0 for (R
t
)
01
= 0.
EXAMPLE 3.23. FAM.S
Human tissue cells die when maintained at over 43
C, T(t = 0) =
37
C,
ac
= 14 m
1
, I
ac.0
= 1.65 10
3
W/m
2
, L
x
= 18 cm, L
y
= 5 cm, L
z
= 5 cm.
SOLUTION
As shown in Figure Ex. 3.23(b), we divide L into N equal nitesmall volumes
to have
Lx =
L
N
.
MassoudKaviany 2011
(a) We rst consider the interior nodes 1 i N. The energy equation for node
1, shown in Figure Ex. 3.23(b), is (3.176), i.e.,
Q
k.10
Q
k.12
= (c
p
LV)
1
dT
1
dt
2
ac
I
ac
(x
i
)LV
1
.
where
Q
k.10
=
T
1
T
0
R
k.10
. R
k.10
=
Lx,2
L
y
L
z
k
Q
k.12
=
T
1
T
2
R
k.12
. R
k.12
=
Lx
L
y
L
z
k
LV
1
= LxL
y
L
z
I
ac
(x
1
) = I
ac.0
e
2
ac
x
1
.
The distance x
i
is measured to the center of the nitesmall volume. This
represents an average s
m.ac
within the volume.
Since T
0
is prescribed, we do not need to write a surface energy equation.
For the interior nodes 2 i N 1, (3.176) becomes
(Q
k
)
i to i1
(Q
k
)
i to i1
= (c
p
LV)
i
dT
i
dt
2
ac
I
ac
(x
i
)LV
i
.
where
(Q
k
)
i to i1
=
T
i
T
i1
(R
k
)
i to i1
. (R
k
)
i to i1
=
Lx
L
y
L
z
k
(Q
k
)
i to i1
=
T
i
T
i1
(R
k
)
i to i1
. (R
k
)
i to i1
=
Lx
L
y
L
z
k
LV
i
= L
y
L
z
Lx
I
ac
(x
i
) = I
ac.o
e
2
ac
x
i
.
For node N, (3.176) becomes
Q
k.N to N1
Q
k.N to N1
= (c
p
LV)
N
dT
N
dt
2
ac
I
ac
(x
N
)LV
N
.
where
Q
k.N to N1
=
T
N
T
N1
R
k.N to N1
. R
k.N to N1
=
Lx
L
y
L
z
k
Q
k.N to N1
=
T
N
T(t = 0)
R
k.N to N1
. R
k.N to N1
=
Lx,2
L
y
L
z
k
LV
i
= L
y
L
z
Lx
I
ac
(x
N
) = I
ac.0
e
2
ac
x
N
.
(b) Next, we consider the surface nodes i = 0, N 1, For these nodes, we use
prescribed temperature, i.e.,
T
0
is prescribed
MassoudKaviany 2011
and
T
N1
= T
= T(t = 0).
35
37
39
41
43
45
47
49
0.000 0.012 0.024 0.036 0.048 0.060 0.072 0.084 0.180
T
,
C
x, m
(c) Predicted Temperature Distribution for Two Elapsed Times
Surface
(x = 0)
Surface
Temperature
T
o
Uniform Termperature within
Each FiniteSmall Volume
Thermal Damage
Threshold
t = 560 s (Elapsed Time
for Reaching Thermal
Damage Threshold)
t = 1,160 s
Linear Interpolation
(Curve Fit)
Width of Damaged Tissue,
Designed to be the Tumor
Figure Ex. 3.23. (c) Predicted temperature distribution for two elapsed times. The loca
tion of damaged tissue is shown with shading.
(c) The initial conditions are
T
i
= T(t = 0) i = 1. N.
(d) Solving the above set of equations using MATLAB, the temperature prole
can be predicted as shown in Figure Ex. 3.23(c) for times t = 560 s (the time
the thermal damage threshold is rst achieved) and t = 1,160 s (ten minutes
after). Since we have suddenly changed the surface temperature at x = 0 to
T
0
= 36
T
i.j
T
i1.j
R
k.Lx
T
i.j
T
i.j 1
R
k.Lz
T
i.j
T
i.j 1
R
k.Lz
= (c
p
LV)
i.j
dT
i.j
dt
s
i.j
LV
ij
energy equation for node T
i.j
. (3.184)
Here we have used R
k.Lx
= Lx,A
k.z
k and R
k.Lz
= Lz,A
k.x
k to show the internodal
conduction resistance over a distances Lx and Lz.
MassoudKaviany 2011
i, j1
i1, j i+1, j
i, j+1
x, i
x
z
y
z, j
KEY: T
i,j
TwoDimensional, Transient Heat Flow: SmallFinite Volume Analysis
SmallFinite Volume
Located at (x,z): (i, j)
zDirection Location Index
xDirection Location Index
T
1,1
T
0,1
T
N,0
T
N,1
T
N,M
T
N+1,M
Node (i, j)
L
y
L
y
with Temperature T
i, j
L
x
L
z
A
z
= x L
y
(Q
k
)
(i, j) to (i, j+1)
A
x
= z L
y
V = x z L
y
Volumetric Terms
(c
p
V)
i,j
+ (s V)
i,j
Q
i, M+1
Q
N+1,j Q
0,j
Q
i,,0
i, j1
i1, j i+1, j
i, j+1
x
z
i, j
dT
i,j
dt
T
1,M
T
1,M+1
(Q
k
)
(i, j) to (i+1, j)
(Q
k
)
(i, j) to (i, j1)
(Q
k
)
(i, j) to (i1, j)
Figure 3.40. Finitesmall volume analysis of twodimensional, conductive heat transfer in a
discretized heat transfer medium. The volumetric nodes and boundary are shown.
(B) Bounding Surface Energy Equation
Similar to the treatment of the onedimensional heat ow, we have the bounding
surface energy equation (2.62), for the four surfaces, as
Q
0.j
(Q
k
)
0.j to 1.j
(Q
t
)
0.j to
=
S
0.j
Q
N1.j
(Q
k
)
N1.j to N.j
(Q
t
)
N1.j to
=
S
N1.j
Q
i.0
(Q
k
)
i.0 to i.1
(Q
t
)
i.0 to
=
S
i.0
Q
i.M1
(Q
k
)
i.M1 to i.M
(Q
t
)
i.M1 to
=
S
i.M1
bounding node
energy equations.
(3.185)
As with the onedimensional heat ow treatment, for a prescribed temperature
at the bounding surface, the external resistance R
t
is set to zero. Then, for example,
at surface node (0. j ), this would result in T
0.j
= T
for (R
t
)
0.j to
= 0.
MassoudKaviany 2011
Again note that
(Q
k
)
i.0 to i.1
=
T
i.0
T
i.1
(R
k
)
i.0 to i.1
= A
k
k
T
i.0
T
i.1
Lx,2
. (3.186)
etc.
EXAMPLE 3.24. FAM.S
Extended surfaces (ns) are attached to a surface that is part of an electronic
component with Joule heating. This is shown in Figure Ex. 3.24(a).
In practice, a uniform temperature is assumed along the y and z directions
while allowing for temperature variation along the x direction. Here assume
that the temperature also varies along the z direction. Use the nitesmall vol
ume treatment and divide the length along the x direction into ten segments,
i.e., L,10, and divide the length along the z direction into three segments, i.e.,
l,3. Then determine the heat ow rate through the base, i.e., x = 0 plane. The
surfaceconvection heat ux is represented by a surfaceconvection resistance.
For example, for a surface node T
i.j
, this is given by
(Q
ku
)
i.j to f .
=
T
i.j
T
f .
(R
ku
)
i.j
(R
ku
)
i.j
=
n
(A
ku
)
i.j
Nu)
n
k
f
.
where (A
ku
)
i.j
is the surfaceconvection area for surface node (i. j ).
(a) Draw the thermal circuit diagram.
(b) For the conditions given below determine the heat ow rate through the
base Q
k
(x = 0).
(c) Plot the variation of temperature along the x axis at the three z locations.
Comment on any difference among these three curves.
q
u
q
u
q
ku
q
ku
S
e,J
w
l
L
q
k
Air Flow
T
f,
, R
ku
, k
f
Prescribed T
s,o
Fin
Solid: High Conductivity k
s
(e.g., Aluminum)
Base
x
y
z
T = T(x,z)
(a) Physical Model of a Rectangular Fin
Figure Ex. 3.24. (a) A rectangular n as part of the extended surface used as a hot sink.
MassoudKaviany 2011
(d) Repeat calculation of (b) for a range of L, i.e., 2 L,l 80. Then plot
Q
k
(x = 0) with respect to L
C. T
f .
= 30
C.
T
2,2
T
2,1
T
1,1
T
1,0
T
s,0
T
s,0
z
y
x
T
s,0
T
2,0
T
1,2
T
2,3
T
1,3
T
1,4
T
f,
T
f,
(R
k
)
1,3 to 2,3
(R
k
)
1,3 to 1,4
(R
ku
)
1,3 to f,
T
2,4
x
L
z
l
w
T
f,
T
f,
T
10,0
T
10,1
T
10,2
T
10,3
T
10,4
T
11,1
T
11,2
T
11,3
T
f,
T
f,
T
f,
T
f,
T
f,
T
f,
(R
k
)
1,3 to s,0
(b) Division of Fin into FiniteSmall Volumes
Figure Ex. 3.24. (b) Division of the n volume into nitesmall volumes, each having a
uniform temperature T
i.j
(t).
SOLUTION
(a) Figure Ex. 3.22(b) shows the division of the n volume into smaller, nite
volumes. There are thirty such nitesmall volumes. Each volume is designated
with a central node temperature T
i.j
. The surface volumes also have surface
nodes showing their surface temperature. These surface nodes are connected to
the uid fareld temperature node T
f .
through surfaceconvection resistances.
The connection between interior nodes is by conduction resistance. Note that
Lx =
L
10
. Lz =
l
3
.
(b) The base heat transfer Q
k
(x = 0) is based on Figure Ex. 3.24(b), i.e.,
Q
k
(x = 0) = (Q
k
)
s.0 to 1.1
(Q
k
)
s.0 to 1.2
(Q
k
)
s.0 to 1.3
.
This steadystate heat ows into the n as conduction at x = 0, is then conducted
along the n, while being removed by surface convection from the n surface.
In terms of the nodal temperature and internodal resistances, Q
k
(x = 0) is
Q
k
(x = 0) =
T
s.0
T
1.1
(R
k
)
s.0 to 1.1
T
s.0
T
1.2
(R
k
)
s.0 to 1.2
T
s.0
T
1.3
(R
k
)
s.0 to 1.3
.
MassoudKaviany 2011
These temperatures T
1.j
are determined from nodal energy equations (3.184),
and here we have a steady state and no nodal energy conversion.
The surface temperatures are determined from (3.185) and (3.186), along
with the surface convection heat transfer given in the problem statement.
For example, for the interior node T
1.2
, we have
T
1.2
T
s.0
(R
k
)
1.2 to s.0
T
1.2
T
2.2
(R
k
)
1.2 to 2.2
T
1.2
T
1.1
(R
k
)
1.2 to 1.1
T
1.2
T
1.3
(R
k
)
1.2 to 1.3
.
Also for surface node T
10.4
, we have
(Q
k
)
10.4 to 10.3
(Q
ku
)
10.4 to f .
= 0
T
10.4
T
10.3
(R
k
)
10.4 to 11.3
T
10.4
T
f .
(R
ku
)
10.4 to f .
= 0.
T
s,0
T
,
C
x, mm
80
79
78
77
76
75
0
0 1 2 3 4 5 6 7 8 9 10
T
1
(1,2,3)
T
2
(1,2,3)
Curve Fit
T
3
(1,2,3)
T
4
(1,2,3)
T
5
(1,2,3)
T
6
(1,2,3)
T
7
(1,2,3)
T
8
(1,2,3)
T
9
(1,2,3)
L
T
10
(1,2,3)
T
11
(1,2,3)
Temperatures at Those z Loca
tions and on the Surface Are
Within Less than One Percent
(c) Discretized Temperature Variation along Fin Length
Figure Ex. 3.24. (c) Variation of the n temperature along the length, at various z locations.
Here the resistances are
(R
k
)
10.4 to 10.2
=
L
10.4 to 10.3
(A
k
)
10.4 to 10.3
k
s
.
(A
k
)
10.4 to 10.3
= Lxn. L
10.4 to 10.3
= Lz,2.
(R
ku
)
10.4 to f .
=
n
LxnNu)
n
k
f
(neglecting the sides)
(R
k
)
1.2 to s.0
=
Lx,2
Lznk
s
(R
k
)
1.2 to 2.2
=
Lx
Lznk
s
.
etc.
The energy equation for the interior nodes, and the surface nodes are solved
simultaneously using a solver (e.g., MATLAB). The result for Q
k
(x = 0) is
Q
k
(x = 0) = 7.370 W.
MassoudKaviany 2011
(c) The variation of temperature along the x axis is shown in Figure Ex. 3.24(c).
Since the results for the three z locations (and also for the surface nodes) are
very close (to less than 1%), the differences cannot be shown. Figure Ex. 3.24(c)
shows how the temperature decays along the n as heat is removed by surface
convection. A continuous curve t to the discrete temperature distribution is
also shown.
Q
k
(
x
=
0
)
,
W
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
2 4 6 8 10 12 14 16 18 20 30 40 50 60 70 80
(d) Variation of Heat Flow with Dimensionless Fin Length
L
*
=
L
l
Figure Ex. 3.24. (d) Variation of heat owing into the n with respect to the dimensionless
n length.
(d) The variation of Q
k
(t = 0) with respect to L,l is shown in Figure Ex. 3.24(d).
Since the tip of the n (i.e., T
11.(0.1.2.3.4)
) approaches T
f .
as L,l , addition
of extra length does not further increase Q
k
(x = 0) for L,l > 70.
COMMENT
When the resistance to heat ow is determined by the resistances outside the
medium (as compared to those inside), then the assumption of uniform temper
ature is justiable. Here along the z direction, the conduction resistance within
the n is negligible compared to the surface convection and conduction along
the x direction. Then by neglecting conduction along the z direction a closed
form solution can be found. We will discuss this resistance comparison further
in Section 6.8.2.
3.7.3 Nonuniform Discretization
In general, the spatial variation of temperature within the heat transfer medium is
not uniform. An example is the transient temperature distribution shown in Figure
3.33. There, for low Fourier numbers (i.e., elapsed times), the temperature variation
MassoudKaviany 2011
x
y
z
L
x
T
N
Assignment of Smaller Volumes Adjacent to Surface 0
T
1
x
1
/2
x
1
x
2
x
3
x
i
x
N
x
N
/2
x
2
x
1
T
2
T
3
T
i
Surface 0
Surface N+1
Figure 3.41. Nonuniform division of the medium into nitesmall volumes.
is concentrated near the bounding surface experiencing the sudden change in the
surface temperature. Under these conditions it is benecial to use nonuniform, small
volumes with the smaller volumes concentrated in a region of large spatial change
in temperature.
For the onedimensional heat transfer of Figure 3.39, the corresponding nonuni
form discretization is shown in Figure 3.41. One method for the nonuniform dis
cretization is to use the spacing ratio r and then the locations of the control surfaces
are given by
x
i
=
r
i
1
r
N
1
L
x
(3.187)
Then the spacing between two adjacent control surfaces is Lx
i
= x
i
x
i1
. For the
spacing ratio, values in the range of 1.05 to 1.15 are commonly used.
3.8 Conduction and SolidLiquid Phase Change:
Stefan Number Ste
l
As an example of a heat transfercontrolled solidliquid phase change, consider
melting of a thick slab of solid. For a pure substance, the solid and liquid phases will
have the same constituents and a distinct interface will appear during melting. This
is depicted in Figure 3.42. We consider melting of a semiinnite slab with the heat
transfer provided by raising the temperature at location x = 0 to a value higher than
the melting temperature T
sl
. At the zero elapsed time t = 0, the melt interface (or
melt front) is at x = 0. As the elapsed time increases, the front penetrates a distance
= (t), with (t = 0) = 0. As with other phase changes, the phase change may not
occur at T
sl
and a metastable solid at a temperature higher than T
sl
may continue
to exist until a threshold excess temperature (i.e., above T
sl
) is reached and then
the phase change occurs (i.e., the solid may be superheated before melting). For
MassoudKaviany 2011
Front Penetration Distance
0
LiquidSolid Interface Confining
Wall
Heat Conducted
through Con
fining Wall
(s
n
)
l
(s
n
)
s
T
x
t
1
> 0 t
2
>t
1
t
3
>t
2
t = 0
x
T
l,0
T
l
=T
l
(x,t)
T
l,0
T
sl
u
(t) u
(t)
Melting Front Speed
u
(t)
T
sl
T
s
=T
sl
(t) (t)
Heat Supplied for
Phase Change, S
sl
/A
sl
Liquid
Melting Front FarField Thermal Condition:
Phase Change Material
Solid
(t)
(a) Melting Front Penetrates, (t)
(b) Heat for Phase Change S
sl
/A
sl
, is Supplied by Conduction, q
k
(c) Temperature Distribution
(d) Significant Dimensionless Variable and Parameter
T
s
= T
sl
Heat Absorbed
by Melting h
sl
No Heat Transfer
(Uniform T)
Heat Supplied
by Conduction
through Liquid
k
l
Heat Stored
in Liquid
l
c
p,l
(t)
, Stefan Number: Ste
l
=
c
p,l
(T
l,0
T
sl
)
h
sl
Similarity Variable: =
x
2(
l
t)
1/2
.
.
q
k
q
k
q
k
q
k
(c
p
)
l
T
l
t
Figure 3.42. Onedimensional, transient solidliquid phase change. (a) Penetration of the
melting front. (b) Heat ux vector tracking. (c) Temperature distribution. (d) Signicant
dimensionless variable and parameter.
simplicity, we assume that the solid is initially at the melting temperature T
sl
and
that the slightest rise in the temperature will result in melting. We also assume that
the melt (i.e., the liquid) will not undergo any motion due to thermobuoyancy (due
to the liquid temperature nonuniformity) or due to phasedensity buoyancy (due to
the difference in the density of the two phases).
For the problem depicted in Figure 3.42, the liquidphase energy equation con
tains the spatial rate of change of q
k
and the volumetric storage term, as given by
(3.129), and is written for constant properties as
k
l
2
T
l
x
2
= (c
p
)
l
T
l
t
(3.188)
or
2
T
l
x
2
1
l
T
l
t
= 0.
l
=
_
k
c
p
_
l
volumetric energy equation for liquid. (3.189)
The solid is at a uniform temperature T
sl
(i.e., volumetric energy conversion or
storage is allowed in the solid). Thus the solid is represented only as a bounding
surface. The boundingsurface conditions are, at x = 0, a prescribed temperature
MassoudKaviany 2011
T
l.0
for t > 0. At the melting front x = = (t) we use the continuity of temperature
(2.57) and the interfacial energy balance, with interfacial phase change, (2.62). These
thermal conditions are, with the energy conversion term given in Table 2.1,
T
s
= T(x = 0. t > 0) = T
l.0
left boundingsurface condition (3.190)
T[x = (t). t > 0] = T
sl
meltingfront temperature (3.191)
q
k
[x =
(t)]A
sl
= k
l
T
l
x
[
x=(t)
A
sl
=
S
sl
= m
sl
Lh
sl
A
sl
=
l
u
F
Lh
sl
A
sl
=
l
Lh
sl
d
(t)
dt
A
sl
meltingfront
energy equation,
(3.192)
where
u
(t)
dt
meltingfront speed. (3.193)
Here we have neglected the radiative heat transfer in (2.62) and have used (s
n
)
s
and
the enthalpy of phase change Lh
sl
= Lh
ls
(heat is absorbed during the melting
and released during solidication, i.e., freezing). We have used a zero temperature
gradient in the solid [as shown in Figure 3.42(c)]. We have also assumed that the solid
and liquid densities are equal. Solving (3.189) with conditions (3.190) to (3.192), the
solution for the temperature distribution in the melt is found similar to that given in
Section 3.5.1 and is left as an end of chapter problem.
T
l
(x. t) T
l.0
T
sl
T
l.0
=
erf
_
x
2(
l
t)
1,2
_
erf(
o
)
=
erf()
erf(
o
)
transient temperature
distribution in liquid
(3.194)
=
x
2(
l
t)
1,2
. (3.195)
where is the similarity variable, identical to (3.136), and the constant
o
is obtained
from the interfacial energy equation (3.192) and is
o
e
2
o
erf(
o
) =
c
p.l
(T
l.0
T
sl
)
1,2
Lh
sl
Ste
l
1,2
implicit relation for
o
. (3.196)
where
Ste
l
c
p.l
(T
l.0
T
sl
)
Lh
sl
Stefan number. (3.197)
Here
o
is the solution to the implicit equation (3.196). This is also called the root
of the transcendental. The Stefan number Ste
l
is the ratio of the liquid sensible heat
to the specic heat of melting. The heat of melting (opposite in sign and equal in
MassoudKaviany 2011
Table 3.7. Roots
o
for the phasechange
transcendental equation
Ste
l
1,2
o
Ste
l
1,2
o
0.0 0.0 1.4 0.8603
0.1 0.2895 1.5 0.8797
0.2 0.3990 1.6 0.8978
0.3 0.4774 1.7 0.9150
0.4 0.5394 1.8 0.9312
0.5 0.5910 1.9 0.9466
0.6 0.6352 2.0 0.9612
0.7 0.6740 2.5 1.0251
0.8 0.7085 3.0 1.0773
0.9 0.7395 3.5 1.1215
1.0 0.7678 4.0 1.1596
1.1 0.7936 5.0 1.2229
1.2 0.8175 6.0 1.2742
1.3 0.8397 10.0 1.4149
magnitude to heat of fusion) for some elements and some pure substances are listed
in the periodic table (Table C.2), and in Tables C.4 to C.6.
The magnitudes of
o
for some values of Ste
l
,
1,2
are given in Table 3.7. Note
that T
l.0
is limited by the boiling temperature of the substance T
lg
.
The front location
(t) = 2
o
(
l
t)
1,2
meltingfront location, measured from x = 0 (3.198)
u
dt
=
1,2
l
o
t
1,2
meltingfront speed. (3.199)
Then the rate of phasechange heat transfer is found from (3.192), i.e.,
q
k
[(t)] = k
l
T
l
x
[
(t)
=
l
Lh
sl
1,2
l
o
t
1,2
rate of heat transfer
to liquidsolid interface.
(3.200)
This shows that similar to the other conduction penetration depths presented in
Section 3.5.2, here the penetration depth
is proportional to t
1,2
. Note that q
k
does
not explicitly depend on T
l.0
T
sl
, but depends on T
l.0
T
sl
through
o
. Similarly,
the front speed u
F
decreases with elapsed time as t
1,2
, as in (3.153).
Note that for Lh
sl
_c
p.l
(T
l.0
T
sl
), i.e., Ste
l
, we have erf(
o
) =
1 from Table 3.5, and we recover the solution for transient conduction without
phase change, i.e., (3.140). Note that in heat storage applications with solidliquid
phase change, the phasechange material can be selected based on the available
temperatures. Table C.5 lists materials with melting temperatures T
sl
between 0 and
MassoudKaviany 2011
260
C. This allows for storage of what is considered relatively low temperature
energy.
EXAMPLE 3.25. FAM
The rate of heat stored in a solid by melting is time dependent, as evident from
(3.200). This rate of heat transfer decreases as the elapsed time increases. This
is shown in Figure Ex. 3.25.
t
T
l
x
(a) PhaseChange Heat Transfer Rate q
k
[
(t)]
(b) Time Variation of q
k
[
(t)]
q
k
[
(t)]
q
k
[
(t)]
q
k
Proportional to t
1/2
(t)
T
s
= T
l,0
T
sl
Solid
Liquid
0
0
0
= m
sl
h
sl
S
sl
A
k
.
.
Figure Ex. 3.25. The rate of heat stored in a solid by melting.
Using some lowmelting temperature pure metals as the phasechange mate
rial, i.e., (i) lead, (ii) mercury, and(iii) sodium, compare the rate of heat storage in
phase change at the elapsed times of 60, 600, and 6,000 s. Use T
l.0
T
lg
= 100
C
and the properties given in Tables C.2, C.23, and C.24.
SOLUTION
From (3.200) we have
q
k
[
(t)] =
l
Lh
sl
1,2
l
o
t
1,2
.
This shows a decrease in q
k
as t increases. In order to determine q
k
, in
addition to the properties, the constant
o
needs to be evaluated. Table 3.7 gives
MassoudKaviany 2011
o
as a function of
Ste
l
1,2
=
c
p.l
(T
l.0
T
sl
)
1,2
Lh
sl
.
The properties are taken as available in the periodic table, Table C.2, and
Tables C.23 and C.24, i.e.,
(i) lead (Pb) c
p.l
= 159 J/kg/K Table C.24
T
sl
= 207.20
C k
l
= 16.1 W/mK
l
= 10,540 kg/m
3
Lh
sl
= 2.3 10
4
J/kg Table C.2
(ii) mercury (Hg) c
p.l
= 139.3 J/kg/K Table C.23
T
sl
= 38.83
C k
l
= 8.54 W/mK
l
= 13,529 kg/m
3
Lh
sl
= 1.14 10
4
J/kg Table C.2
(iii) sodium (Na) c
p.l
= 1,360 J/kg/K Table C.24
T
sl
= 97.72
C k
l
= 72.3 W/mK
l
= 860.2 kg/m
3
Lh
sl
= 1.13 10
5
J/kg Table C.2.
Note the relatively high c
p.l
, k
l
, and Lh
sl
for sodium.
From these properties, we have
(i) lead
Ste
l
1,2
=
159(J/kgK) 100(K)
1,2
2.32 10
4
(J/kg)
= 0.387
l
=
k
l
(c
p
)
l
=
16.1(W/m
2
K)
10,540(kg/m
3
) 159(J/kgK)
= 9.61 10
6
m
2
/s
(ii) mercury
Ste
l
1,2
=
139.3(J/kgK) 100(K)
1,2
1.14 10
4
(J/kg)
= 0.690
l
=
k
l
(c
p
)
l
=
8.54(W/m
2
K)
13,529(kg/m
3
) 139.3(J/kgK)
= 4.53 10
6
m
2
/s
(iii) sodium
Ste
l
1,2
=
1,360(J/kgK) 100(K)
1,2
1.13 10
5
(J/kg)
= 0.679
l
=
k
l
(c
p
)
l
=
72.3(W/m
2
K)
860.2(kg/m
3
) 1,300(J/kgK)
= 6.45 10
5
m
2
/s.
From Table 3.7, we have
(i) lead
o
(Ste
l
,
1,2
= 0.387) = 0.5313 Table 3.7
(ii) mercury
o
(Ste
l
,
1,2
= 0.690) = 0.6701 Table 3.7
(iii) sodium
o
(Ste
l
,
1,2
= 0.679) = 0.6658 Table 3.7.
MassoudKaviany 2011
We now substitute the numerical values into the above expression for q
k
,
i.e.,
(i) lead q
k
(t = 60 s) = 0,5401(kg/m
3
)
2.3 10
4
(J/kg)
(4.53 10
6
)
1,2
(m
2
/s)
1,2
0.5313
(60)
1,2
(s)
1,2
= 3.5390 10
4
W/m
2
q
k
(t = 600 s) = 1.219 10
4
W/m
2
q
k
(t = 6,000 s) = 3.6553 10
3
W/m
2
(ii) mercury q
k
(t = 60 s) = 13,529(kg/m
3
)
1.14 10
4
(J/kg)
(4.53 10
6
)
1,2
(m
2
/s)
1,2
0.6701
(60)
1,2
(s)
1,2
= 2.846 10
4
W/m
2
q
k
(t = 600 s) = 8.999 10
3
W/m
2
q
k
(t = 6,000 s) = 2.846 10
3
W/m
2
(iii) sodium q
k
(t = 60 s) = 860.2(kg/m
3
)
1.13 10
5
(J/kg)
(6.45 10
5
)
1,2
(m
2
/s)
1,2
0.6658
(60)
1,2
(s)
1,2
= 6.713 10
4
W/m
2
q
k
(t = 600 s) = 2.123 10
4
W/m
2
q
k
(t = 6,000 s) = 6.713 10
3
W/m
2
.
COMMENT
This time dependence of the rate of energy storage in phasechange materials
should be considered in the design of such energy storage systems. The heat
withdrawal, i.e., solidication, will have a similar time dependence.
3.9 Thermal Expansion and Thermal Stress
Thermal strain is an elastic strain that results from expansion of solids with increase
in their temperature. When the solid displacement is constrained, this thermal strain
results in thermal stress and the study of such mechanical behavior is called the
thermoelasticity.
Thermal expansion is an increase in the average distance between atoms in
the solid crystalline lattice. At absolute zero temperature (i.e., T = 0 K), the inter
molecular spacing is at its minimum value. As the temperature increases and energy
is stored in the vibrational (and other forms) of energy, the average intermolecular
spacing representing the time averaged position increases (due to asymmetry of the
potential energyintermolecular distance well, called the separation well). For sim
plicity, we will discuss the case of a uniform temperature in Cartesian coordinate
system.
MassoudKaviany 2011
0
1 x 10
5
2 x 10
5
3 x 10
5
4 x 10
5
0 1,000 2,000 3,000 4,000
PEEK
POM
PC
Pb
Metal
Polymer
Ceramic
Glass
Al
Cu
Fe
Ti
Nb
MgO
Ta
W
B
4
C
Al
2
O
3
Fused Silica
SodaLime
T
sl
, C
T = 300 K
Figure 3.43. Linear thermal expansion coefcient
s
for various solids, as a function of their
melting temperature T
sl
.
The linear thermal expansion coefcient
s
(1/K) is dened as
s
1
L
o
L
T
[
p
o
linear thermal expansion coefcient
s
. (3.201)
where L
o
and p
o
are the reference length and pressure.
We now expand (a Taylor series expansion) Laround L
o
(and temperature T
o
),
retain only the rst derivative, and write
L = L
o
L
T
[
p
o
(T T
o
) 0. (3.202)
Now using
s
from (3.201), we have
L = L
o
L
o
s
(T T
o
). (3.203)
Then the unconstrained thermal strain (dimensionless) is given by
LL
t
=
LL
L
o
=
LL
o
L
o
=
s
(T T
o
)
unconstrained dimensionless
thermal strain LL
t
.
(3.204)
Typical values for
s
(1/K), at T = 300 K, are given for various solids in Fig
ure 3.43, as a function of their melting temperature [9]. The solid polymer examples
shown are PC (polycarbonate, a polyester with carbonate ester linked with aro
matics), POM (polyoxymethylene), and PEEK (polyetheretherketone). In general,
s
decreases with increase in melting temperature. This is especially true for pure
metals. The values in Figure 3.43 are for metals, polymers, ceramics, and glasses.
Note that in general the polymers and some pure metals have high linear thermal
expansion coefcients, while some ceramics and glasses have rather low linear ther
mal expansion coefcients. The magnitudes of
s
, for some metallic and nonmetallic
solids, are also listed in Tables C.16 and C.17.
MassoudKaviany 2011
For a unidirectional thermal stress, when the temperature is changed from T
o
to
uniform T while preventing any thermal strains (i.e., an ideally constrained system),
the thermal stress
t
(Pa) is determined from
t
= E
s
s
(T T
o
)
ideally constrained, unidirectional
thermal stress
t
,
(3.205)
where E
s
is the Young modulus of elasticity. For constrained solids, upon heating
the stress is compressive thermal stress (
t
 0) and upon cooling a tensile thermal
stress (
t
> 0) results.
The Newton second law, (1.27), is written for solid particles and this imposes the
condition on the stress eld of the body. The strain conditions state the kinematically
possible deformed state and this is called the compatibility equation. Then, the
stresses and strains are related by the constitutive law for the materials. In the linear
elastic regime, this is the Hooke law.
For a general constraint condition and a threedimensional stressstrain in an
isotropic thermal material, since equal thermal strains occur in all directions, the
Hooke law for elastic behavior is written in the Cartesian coordinates as [33]
LL
x
=
1
E
s
[
xx
P
(
yy
zz
)]
s
(T T
o
) (3.206)
LL
y
=
1
E
s
[
yy
P
(
xx
zz
)]
s
(T T
o
) (3.207)
LL
z
=
1
E
s
[
zz
P
(
xx
yy
)]
s
(T T
o
)
threedimensional Hooke
law for general
constraint conditions,
(3.208)
where
ii
is the normal stress in the i direction and
P
is the Poisson ratio. Note that
each stress component results in a strain in the same direction, and also two Poisson
ratio strains of opposite sign in the other direction. For most materials
P
= 0.25,
and for structural steel
P
= 0.30.
Upon using the unconstrained (
ij
= 0) or ideally constrained (LL
i
= 0) con
dition, from (3.206) to (3.208) we will have (3.204) and (3.205). When LL
i
or
ii
are prescribed, then solutions are found from (3.204) to (3.208) for the unknown
quantities.
The above discussion is for a uniform temperature distribution. For T = T(x. t)
the local stress and strain distributions (including axisymmetric geometries) are
needed [5]. Examples of these are given as the end of chapter problems.
EXAMPLE 3.26. FAM
Some industrial sprinklers [shown in Figure Ex. 3.26(a)] are actuated when
a cylindrical metal rod thermally expands and fractures a glass restriction at
one of its ends. The metal cylinder in this sprinkler activation mechanism is
made of aluminum with a modulus of elasticity E
s
= 68 GPa and has an initial
MassoudKaviany 2011
Glass Restriction
StressInduced Rupture
Actuated Sprinkler
Aluminum
Cylindrical Rod
Water Supply
D = 5 mm
L = 2.5 cm
(c
p
V)
1
dT
1
dt
Figure Ex. 3.26. (a) An industrial sprinkler actuator.
temperature of T
1
(t = 0) = 20
t
= E
s
s
[T
1
T(t = 0)].
where
t
=
max
= 300 10
6
MPa
s
= 23.1 10
6
1/K Table C.16
E
s
= 68 GPa.
Then
300 10
6
(Pa) = 68 10
9
(Pa) 23.1 10
6
(1/K) [T
1
T
1
(t = 0)](K)
[T
1
T
1
(t = 0)] = 185.5
C
T
1
= 205.5
C.
Now using the solution to the singlenode energy equation with constant surface
heat transfer rate, i.e., (3.169), we determine the time needed to reach this
temperature, i.e.,
T
1
(t) = T
1
(t = 0)
Q
S
1
(c
p
V)
1
t.
where
S
1
= 0 and Q
1
= 20 W.
MassoudKaviany 2011
0 2 4 6 8 10 12 14
0
50
100
150
200
208
o
C
T(t =0) = 20
o
C
11.09 s
Actuation Time
250
t, s
T
,
o
C
Figure Ex. 3.26. (b) Variation of actuation temperature with respect to time. The actua
tion time is also marked.
Here
V
1
=
D
2
4
L =
(0.005)
2
(m
2
)
4
(0.025)(m) = 4.9 10
7
m
3
.
From Table C.16, we have
1
= 2,702 kg/m
3
, and c
p.1
= 903 J/kgK. Then
T
1
= 205.5
C = 20
C
20(W)
[2,702(kg/m
3
)] [903(J/kgK)][4.9 10
7
(m
3
)]
t.
or
t = 11.09 s.
COMMENT
We have assumed that the net heat transfer rate to the sprinkler is constant, Q
1
=
20 W. This is not accurate because this net heat transfer rate, which is the sum of
the surfaceconvection, surfaceradiation, and conduction losses varies, with the
sprinkler temperature T
1
(t). This is expected to decrease Q
1
with increase in t.
Therefore, T
1
(t) will rise at a smaller rate with increase in elapsed time [instead
of the linear increase assumed here and shown in Figure Ex. 3.26(b)]. These will
be discussed in Chapters 4 and 6.
3.11 References
[1] Angrist, S.W., 1982, Direct Energy Conversion, FourthEdition, AllynandBacon,
Boston.
[2] Ashcroft, N.W., and Mermin, N.D., 1976, Solid State Physics, Saunders Company,
Philadelphia.
[3] ASHRAE Handbook, 1997, Fundamentals, American Society of Heating,
Refrigeration, and AirConditioning Engineers, Inc., Atlanta.
[4] Avedisian, C.T., Osborne, B., McLeod, F.D., and Curley, C.M., 1999, Measuring
Bubble Nucleation Temperature on the Surface of a Rapidly Heated Thermal
Ink Jet Heater Immersed in a Pool of Water, Proc. Roy. Soc. Lond. A, Volume
455, pp. 38753899.
[5] Barber, J.B., 2000, Intermediate Mechanics of Materials, McGrawHill, New
York.
MassoudKaviany 2011
[6] Berman, R., 1979, Thermal Conduction in Solids, Clarendon Press, Oxford. Also
Grimwall, G., 1999, Thermophysical Properties of Materials, Elsevier, Amster
dam, and Chaikin, P.M. and Lubensky, T.C., 1995 Principles of Condensed Matter
Physics, Cambridge University Press, Cambridge. Also, Srivastava, G.P., 1990,
The Physics of Phonons, Adam Hilger, Bristol.
[7] Carslaw, H.S., and Jaeger, J.C., 1986, Conduction of Heat in Solids, Second
Edition, Clarendon Press, Oxford.
[8] Chen, G., 1996, Nonlocal and Nonequilibrium Heat Conduction in the Vicinity
of Nanoparticles, ASME Journal of Heat Transfer, Volume 118, pp. 539545.
[9] Dowling, N.R., 1993, Mechanical Behavior of Materials, PrenticeHall, Upper
Saddle River.
[10] Johnson, D.A., 1992, The Effect of Heating Pulse on Drop Size and Stability
in Thermal InkJet, Society of PhotoOptical Instrumentation Engineers (SPIE)
Journal, Volume 1670, pp. 224229.
[11] Kaviany, M., 1999, Principles of Heat Transfer in Porous Media, Corrected Sec
ond Edition, Springer, New York.
[12] Kittel, C., and Kroener, H., 1998, Thermal Physics, W.H. Freeman and Company,
New York. Also, Kittle, C., 1986, Introduction to Solid State Physics, Wiley, New
York.
[13] Kraus, A.D., and BarCohen, A., 1983 Thermal Analysis and Control of Elec
tronic Equipment, McGrawHill, New York.
[14] Levine, I. N., 1988, Physical Chemistry, Third Edition, McGrawHill, New York.
[15] Liley, P. E., Makita, T., and Tanaka, Y., 1988, Properties of Inorganic and Organic
Fluids, Ho, C.Y., Editor, Hemisphere Publishing Corporation, New York.
[16] Materials Research Society, 1996 Bulletin, July, p.11.
[17] Meyers, G.E., 1987, Analytical Methods in Conduction Heat Transfer, Genium
Publishing, Schenectady.
[18] Ozisik, M.N., 1993, Heat Conduction, Second Edition, Wiley, New York.
[19] Parker, W.J., Jenkins, R.J., Butler, C.P., and Abbolt, G.L., 1961, Flash Method
of Thermal Diffusivity, Heat Capacity, and Thermal Conductivity, Journal of
Applied Physics, Volume 32, pp. 16791684.
[20] Press, W.H., Teukolsky, S.A., Vetterling, W.T., and Flannery, B.P., 1992, Numer
ical Recipes in Fortran, Chapter 16, Cambridge University Press, Cambridge.
[21] Rezanka, I., Thermal Ink Jet, A Review, Society of Photooptical Instrumen
tation Engineers (SPIE) Journal, Volume 1670, pp. 192200.
[22] Rohsenow, W.M., Hartnett, J.P., andGonic, E.N., 1985, Handbookof Heat Trans
fer, Second Edition, McGrawHill, New York.
[23] Romonelli, M.J., 1960, RungeKutta Method for the Solution of Ordinary Dif
ferential Equations, in Mathematical Methods for Digital Computers, Wiley,
New York.
[24] Rowe, D.M., Editor, 1995, CRCHandbook of Thermoelectrics, CRCPress, Boca
Raton.
[25] Schneider, P.J., 1985, Conduction, in Handbook of Heat Transfer, Second
Edition, Rohsenow, W. M., et al., Editors, McGrawHill, New York. Also,
Yovanovich, M.M., 1998, Conduction and Thermal Contact Resistance (Con
ductance), in Handbook of Heat transfer, Third Edition, Rohsenow, W.M.,
et al., Editors, McGrawHill, New York.
[26] Smith, K., and Jensen, H.H., 1989, Transport Phenomena, Clarendon Press,
Oxford.
[27] Srivastava, G.P., 1990, The Physics of Phonons, Adam Hilger, Bristol. Phelan,
P.E., 1998, Applications of Diffusive Mismatch Theory to Prediction of Thermal
Boundary Resistance in thinFilmHighT
c
Superconductors, ASMEJournal of
Heat Transfer, Volume 170, pp. 3743.
MassoudKaviany 2011
[28] Standard Test Method for Thermal Conductivity of Solids by Means of the
GuardedComparativeLongitudinal Heat Flow Technique, ASTM: 122587,
American Society for Testing and Materials, Philadelphia.
[29] Standard Test Method for Thermal Diffusivity of Solid by the Flush Method,
ASTM: E146192, American Society for Testing and Materials, Philadelphia.
[30] Tabor, D., 1991, Gases, Liquids, and Solids: And Other States of Matter, Third
Edition, Cambridge University Press, Cambridge.
[31] Taylor C.F., 1977, The Internal Combustion Engine in Theory and Practice, MIT
Press, Cambridge.
[32] Tien, C.L., and Lienherd, J.H., 1971, Statistical Thermodynamics, Holt, Rine
hard, and Winston, New York.
[33] Timoshenko, S.P., and Goodier, J.N., 1970, Theory of Elasticity, Third Edition,
McGrawHill, New York.
[34] Touloukian, Y.S., and Ho, C.Y., 1972, Editors, Thermophyisical Properties of
Matter, Volume 1, Thermal Conductivity of Metallic Solids, Volume 2, Thermal
Conductivity of Nonmetallic Solids, Plennum, New York.
[35] Ultramet, Advanced Materials Solutions, Pacoima, California.
[36] Walton, A.J., 1989, Three Phases of Matter, Second Edition, Clarendon Press,
Oxford.
[37] Woods, L.C., 1993, An Introduction to the Kinetic Theory of Gases and Magne
toplasmas, Oxford University Press, Oxford.
[38] Zeng, S.Q., Hunt, A., and Greif, R., 1995, Geometric Structure and Thermal
Conductivity of Porous Medium Silica Gel, ASME Journal of Heat Transfer,
Volume 117, pp. 10551058.
[39] Zircar Product Inc., Technical Data, Florida, New York.
3.12 Problems
3.12.8 Multinode Systems and FiniteSmall Volume Analysis
PROBLEM 3.72. FAM
Due to defects in the brake pad or the rotor geometry, the friction heat generation
S
m.F
may not have a uniform distribution over the brake padrotor contact surface.
This results in a hot spot at the locations of high contact, and due to the thermal
expansion, these hot spots continue to have further increase in contact pressure.
Eventually very high temperatures and a failure occurs. Consider the friction energy
conversion occurring over a rotor surface. The rotor is idealized as a ring of inner
radius R
i
, outer radius R
o
, and thickness l, as shown in Figure Pr. 3.72. Under normal
contact, the energy conversion will be equally distributed over the entire contact
surface and a uniform temperature T
1
(t) can be assumed. Under hotspot contact,
assume that the energy is dissipated over a ring with the inner and outer radii R
i.1
and R
o.1
(with the same thickness l), resulting in a uniform temperature T
1
(t) (i.e.,
N
k.1
 0.1) and that the rest of the rotor is at a constant temperature T
2
with heat
owing by conduction from T
1
(t) to T
2
with a constant resistance R
k.12
. This is only
a very rough approximation.
(a) Draw the thermal circuit diagram for (i) normal contact with no heat transfer,
and (ii) hotspot contact with Q
k.12
as the heat transfer.
MassoudKaviany 2011
Friction Energy
Conversion
An Idealized DiscBrake Rotor
For HotSpotting,
the Rest of Rotor at
Constant Temperature, T
2
HotSpot Contact
Area, T
1
(t)
R
o
l
S
m,F
R
o,1
R
i,1
R
i
Figure Pr. 3.72. A discbrake rotor with (i) normal padrotor contact, and (ii) with hotspot
partial surface contact.
(b) Determine the temperature T
1
(t) for cases (i) and (ii) after an elapsed time of
t = 4 s.
(c) Comment on the difference in T
1
(t = 4 s) for cases (i) and (ii).
S
m.F
= 30 kW, R
o
= 18 cm, R
i
= 13 cm, R
o.1
= 16 cm, R
i.1
= 15 cm, l = 1.5 cm,
(c
p
)
1
= 3.5 10
6
J/m
3
K, R
k.12
= 1
C/W, T
1
(t = 0) = 20
C, T
2
= 20
C.
PROBLEM 3.73. FUN
A thermal barrier coating in the form of spray deposited, stabilized zirconia parti
cles, is used as a thin layer to protect a substrate. Figures Pr. 3.73(i) and (ii) show
a typical thermal barrier coating and a representative twodimensional conduction
network model. The heat conduction through the gas lling the voids is neglected.
The thermal conductivity for the bulk (one hundred percent theoretical density)
stabilized zirconia is k
s
= 1.675 W/mK, and the porosity of the coating is approx
imately c = 0.25. The geometrical properties of the representative network model
are given in Table Pr. 3.73.
Substrate
11
17
8
13 12
16
19
20
22
21
18
14
6
9
3 2
4
5
1
10
15
l
y
x
T
2
k
s
T
1
Adiabatic
Arm
Node
Adiabatic
k
f
= 0
Thin Thermal
Protection Film
Pore
7
L
(i) Physical Model (ii) Network Model
L
x
L
y
Q
k,24
R
k,24
Figure Pr. 3.73. (i) Micrograph of a thermal barrier coating, and (ii) a twodimensional,
representative conduction network model.
MassoudKaviany 2011
Table Pr. 3.73. The geometrical properties of the
representative network model
Arm l, jm L, jm Arm l, jm L, jm
17 6 16.8 1116 12.8 14
24 12 9.8 1213 32 15.4
35 20 12.6 1217 16 21
45 12 11.2 1314 8 8.4
46 3.2 14 1315 8 12.6
59 20 18.2 1419 12 16.8
68 20 14 1518 13 18.2
69 4 16.8 1520 4.8 26.6
78 10 22.4 1621 30 14
711 20 26.6 1718 12 8.4
910 14 15.4 1721 10 25.2
1013 8 18.2 1822 12 5.6
1112 9 28 1920 4 8.4
(a) Determine the estimatedeffective thermal conductivity k)
yy
(along the y axis) for
the lm layer by using the twodimensional thermal circuit diagram given in Figure
Pr. 3.73(ii). (The network model also represents the thermal circuit diagram for the
layer.) Neglect the thermal conductivity of the gas ling the pores (k
f
= 0). Take
the length along the z axis (perpendicular to page) n = 1 m. Use as the temperature
at the lower boundary L
z
, T
1
= 225
(x. y) =
T(x. y) T
1
T
2
T
1
. x
=
x
L
x
. y
=
y
L
y
.
(b) Plot the results for N = 3, 15, and 21.
(c) Compare the results with the exact series solution
T
(x. y) =
2
n=1
1 (1)
n1
n
sin
_
nx
L
x
_
sinh(ny,L
x
)
sinh(nL
y
,L
x
)
by showing the results on the same plot.
L
x
= L
y
= 20 cm, Lx = Ly = L
x
,N.
L
y
TwoDimensional
Temperature Distribution
T = T(x,y)
Finite Small
Volume
L
z
L
x
T
1
, T
*
= 0
T(0,1)
T(1,0)
T(1,1)
T
1
, T
*
= 0
T
1
, T
*
= 0
T
2
, T
*
= 1
x
y
x
y
Figure Pr. 3.74. Twodimensional, steadystate conduction in a rectangular medium. The dis
cretized nitesmall volumes are also shown.
PROBLEM 3.75. FUN
A porcelain workpiece (in form of a circular disk) is ablated by laser irradiation.
The piece is held inside a cooling ring, as shown in Figure Pr. 3.75, to maintain its
outer surface at a temperature T
s
. For the ablation, the temperature of the ceramic
much reach a threshold temperature T
sg
(i.e., a sublimation temperature), over the
area of interest. The radiation is absorbed only over the surface A
r.
, and there is
2R
r
A
r,
2R
r
R
A
ku
= A A
r,
T
s
l
R
z
r
Workpiece
A
ku
R
ku
FarField Air
Temperature
T
f,
Laser Beam
Cooling Ring
Coolant Flow
Node 1
Workpiece
T
s
(Maintained
Temperature)
Laser Generator
Laser Irradiation, q
r,i
Figure Pr. 3.75. A porcelain workpiece is irradiated for ablation. The workpiece is cooled at
its periphery by a coolant carrying ring.
MassoudKaviany 2011
surface convection over the rest of the area A
ku
. Assume a steadystate heat transfer
and a uniform temperature along the z and axes [i.e., T = T(r) only]. Divide the
porcelain piece into N segments, i.e., Lr = R,N, and apply the nitesmall volume
energy equation (3.176) to each segment.
(a) Draw the thermal circuit diagram for the entire disk.
(b) Determine the segment temperature T
i
, for N = 5.
Assume all irradiation is absorbed on the surface. For the central node, node
1, use R
r
,2 as the inner surface location for the determination of the conduction
resistance.
R = 3 cm, R
r
= R,5, l = 3 mm, q
r.i
= 10
6
W/m
2
, T
s
= 90
C, A
ku
R
ku
=
10
3
K/(W/m
2
), T
f .
= 120
C.
PROBLEM 3.76. FUN.S
The effective thermal conductivity of some porous media can be determined using
the random network model. In one of these models, a regular lattice is used,
but the locations of the nodes, within the regular lattice, are generated randomly and
then connected, forming a network. The thicknesses (for a two or threedimensional
geometry) of these connectors (i.e., arms) are then assigned based on the porosity
and any other available information. The network can represent the solid or the
uid part of the medium. A 3 3 square unitcell, twodimensional random network
model is shown in Figure Pr. 3.76. This is determined by randomly selecting the
location of each node within its unit cell space. The coordinates of each node, and
the length and the thickness for each arm are given in Table Pr. 3.76.
Assume that the heat transfer between adjacent nodes is one dimensional and
steady. For the arms, use the thermal conductivity of aluminum(Table C.16). Assume
that the thermal conductivity of the uid is much smaller than that of the solid. The
left and right boundaries of the medium are ideally insulated. The temperature for
Node 11
w
Node 6 Node 1
2
7 12
13
1 mm
1 mm
8
9
4
3
Node 5 Node 10 Node 15
14
Q
k,hc
T
h
= 200 C
x
y
L
23
l
23
T
c
= 100 C
R
k,54
Q
k,54
Arm
Figure Pr. 3.76. A twodimensional regular (periodic) lattice with random location of nodes
within the lattice, making for a random network model.
MassoudKaviany 2011
Table Pr. 3.76. Coordinates of each node (within a regular lattice),
and the lengths and the thicknesses for each arm, forming a
twodimensional random network model
Node (x. y),(mm,mm) L
i.j
, mm l
i.j
, mm
1 (0.76,0.00) L
1.2
= 0.715 l
1.2
= 0.414
2 (0.13,0.34) L
2.3
= 1.581 l
2.3
= 0.213
L
2.7
= 1.059 l
2.7
= 0.055
3 (0.92,1.71) L
3.4
= 0.622 l
3.4
= 0.152
L
3.8
= 1.105 l
3.8
= 0.197
4 (0.47,2.14) L
4.5
= 0.932 l
4.5
= 0.344
L
4.9
= 0.970 l
4.9
= 0.376
5 (0.83,3.00)
6 (1.73,0.00) L
6.7
= 0.788 l
6.7
= 0.151
7 (1.17,0.54) L
7.8
= 1.047 l
7.8
= 0.232
L
7.12
= 1.596 l
7.12
= 0.215
8 (1.93,1.26) L
8.9
= 1.059 l
8.9
= 0.469
L
8.13
= 0.314 l
8.13
= 0.125
9 (1.44,2.15) L
9.10
= 0.935 l
9.10
= 0.071
L
9.14
= 1.592 l
9.14
= 0.316
10 (1.83,3.00)
11 (2.38,0.00) L
11.12
= 0.854 l
11.12
= 0.417
12 (2.75,0.77) L
12.13
= 0.887 l
12.13
= 0.222
13 (2.18,1.45) L
13.14
= 1.602 l
13.14
= 0.118
14 (2.84,2.91) L
14.15
= 0.577 l
14.15
= 0.322
15 (2.27,3.00)
the lower boundary of the medium, [(x. y) = (x. 0)], is maintained at T
c
= 100
C,
(i.e., T
c
= T
1
= T
6
= T
11
= 100
C, (i.e., T
h
= T
5
= T
10
= T
15
= 200
C).
(a) Draw the thermal circuit diagram using the geometrical data. Write the energy
equation for each node, along with the conduction heat transfer relation for each
arm.
(b) Determine the total conduction heat transfer rate, Q
k.hc
.
(c) Using
Q
k.hc
A
k
(T
h
T
c
)k)
L
.
for L = 10
3
m and A
k
= 3 10
6
m
2
, determine the effective thermal conductiv
ity k).
3.12.9 SolidLiquid Phase Change
PROBLEM 3.77. FAM
The friction heating during skating over ice layers causes melting, and the thickness
of this melt at the end of the blade
C, T
sl
= 0
C, L = 0.20 m, u
p
= 2 m/s.
Use properties of water given in Tables C.4 and C.27 (at T = 275 K).
u
s
g
Melt Formation
Under Skate Blade
Length of Blade in
Contact with Ice
Ice
T
s
= T
l,o
x
L
C, and (b) T
l.0
=
15
C.
Assume that the heat transfer through the liquid water is one dimensional and
occurs by conduction only, that the ice is at the melting temperature T
ls
, and neglect
the sensible heat of the liquid water. Use the properties of water at T = 273 K from
Table C.23 and Table C.6.
PROBLEM 3.80. DES
In a thermostat used to control the passage of coolant through secondary piping
leading to the heater core of an automobile, solidliquid phase change is used for dis
placement of a piston, which in turn opens the passage of the coolant. The thermostat
is shown in Figure Pr. 3.80(i). The phasechange material is a wax that undergoes
approximately 15% volume change upon solidication/melting. The response time
of the thermostat is mostly determined by the time required for complete melting
of the wax. This in turn is determined by the speed of penetration of the melting
front into the wax and the time for its complete penetration. Two different designs
are considered, and are shown in Figures Pr. 3.80(ii) and (iii). In the design shown in
Figure Pr. 3.80(iii), the wax reservoirs have a smaller diameter D
2
compared to that
Hot Water
Closed
(ii) Single Reservoir (iii) Three Reservoirs
Wax
Melting Front (t)
(t)
(t)
Open
Piston
T
f,
> T
sl
(p
l
)
Wax
Liquid
Solid
Solid
Liquid
Volume Increased Due
to Melting Wax
Bellow
Q
A
(t)
T
s
T
l,0
T
sl
D
1
D
2
High Conduc
tivity Solid
(i) Automotive Coolant Thermostat
Figure Pr. 3.80. (i) An automotive coolant thermostat. The solidliquid phase change actuates
the piston. (ii) The singlereservoir design. (iii) The threereservoir designs.
MassoudKaviany 2011
of the rst design, shown in Figure Pr. 3.80(ii). Therefore, it is expected that in the
threereservoir design the wax will melt faster.
To solve the melting problem using the analysis of Section 3.8, we need to
assume that the front is planar. Although this will result in an overestimation of the
time required for melting, it will sufce for comparison of the two designs. Due to
unavailability of complete properties for the wax, use the phasechange properties
of Table C.5 for parafn and the properties of engine oil in Table C.23 for the wax
liquid phase. The melting temperature is a function of pressure and is represented
by the ClausiusClapeyron relation (A.14). Here, neglect the pressure variation and
use a pressure of one atm and a constant melting temperature.
Determine the time it takes for the complete melting of the wax in the one and
threereservoir designs.
Use the properties of engine oil at T = 310 K. T
l.0
= 80
C, D
1
= 8 mm, D
2
=
3 mm.
3.12.10 Thermal Expansion and Thermal Stress
PROBLEM 3.81. FAM
In a grinding operation, material is removed from the top surface of a small piece of
pure copper with dimensions shown in Figure Pr. 3.81. The grinding wheel is pressed
against the copper workpiece with a force F
c
= 50 Nand there is an interfacial veloc
ity Lu
i
= 20 m/s. The coefcient of friction between the two surfaces is j
F
= 0.4.
The copper is initially at temperature T
1
(t = 0) = 20
C.
(a) Assuming a uniform copper temperature T
1
(t), i.e., N
t.1
 0.1, and a constant
surface heat loss rate per unit surface area q
1
= 675 W/m
2
, draw the thermal circuit
diagram.
(b) Assuming an unconstrained expansion, determine the elapsed time needed to
cause the copper length L to thermally expand by LL = 0.5 mm. Neglect all non
thermally induced stresses and strains.
L = 5 cm
A
c
w = 2.5 cm
l = 0.5 cm
(S
m,F
)
1
(Due to Grinding)
Workpiece
Uniform
Temperature T
1
(t)
(N
t,1
< 0.1) y x
z
q
1
Figure Pr. 3.81. Friction heating of a copper workpiece by grinding resulting in thermal
expansion.
PROBLEM 3.82. FUN
The thermal stress in an idealized discbrake rotor can be determined using some
simplifying assumptions. The castiron rotor is shown in Figure Pr. 3.82, along with
a prescribed temperature distribution T = T(r).
MassoudKaviany 2011
r
R
S
m,F
T
R
T
0
r
R 0
zz
= 0
(Planar Stress)
(r)
DiscBrake Rotor
rr
(r)
Figure Pr. 3.82. The temperature distributionwithina discbrake rotor andthe induced, planar
axisymmetric thermal stresses.
The temperature distribution is steady and one dimensional and the stress tensor
wouldhave planar, axisymmetric stresses givenby principal components
rr
(r).
(r).
It can be shown that these stresses are expressed as [5]
rr
(r) =
E
s
s
r
2
_
r(T T
o
)dr a
1
a
2
r
2
(r) =
E
s
s
r
2
_
r(T T
o
)dr E
s
s
(T T
o
) a
1
a
2
r
2
zz
(r) = 0.
For cast iron, E
s
= 200 GPa, and
s
is listed in Table C.16. Also R = 17 cm, T
o
=
100
C, T
R
= 400
C.
(a) Determine the integration constants a
1
and a
2
using the mechanical conditions.
Note that at r = R, there is no radial stress (for surface), and that at r = 0 the stresses
should have a nite magnitude.
(b) Plot the distribution of the radial and tangential rotor thermal stresses with
respect to the radial location.
PROBLEM 3.83. FUN
Consider the thermal stress due to a nonuniform temperature T = T(r) in an alu
minum rod encapsulated in a glass shell. This is shown in Figure Pr. 3.83. Since the
thermal expansion coefcient
s
is much smaller for the fused silica glass (Figure
3.43), we assume that the periphery of the aluminum rod is ideally constrained.
For the axisymmetric geometry and temperaturestress conditions we have here the
stress and strain distributions as given by [5].
rr
(r) =
E
s
s
r
2
_
r(T T
o
)dr a
1
a
2
r
2
(r) =
E
s
s
r
2
_
r(T T
o
)dr E
s
s
(T T
o
) a
1
a
2
r
2
LR(r) =
s
(1
P
)
r
_
r(T T
o
)dr
a
1
(1
P
)r
E
s
(1
P
)a
2
E
s
r
.
MassoudKaviany 2011
Assumed Ideally
Constrained Surface
Aluminum Rod
Fused Silica Glass Shell
T
o
T
R
T(r)
R
r
rr
Figure Pr. 3.83. Thermal stress induced in an aluminum rod encapsulated in a fusedsilica
glass shell.
The temperature distribution within the aluminum rod is estimated as
T(r) = T
R
(T
o
T
R
)
_
1
r
2
R
2
_
.
The two constants of integration, a
1
and a
2
, are determined using the mechanical
conditions of a nite stress at r = 0 and an ideal constraint at r = R.
(a) Determine the integration constants a
1
and a
2
and write the expression for
rr
(r = R).
(b) Using the conditions given below, determine the temperature T
R
at which the
ultimate compression stress of the glass
max.g
is reached, i.e.,
rr
(r = R) =
max.g
.
max.g
= 300 MPa, E
s
= 68 GPa,
P
= 0.25, T
o
= 80
C, R = 4 cm.
3.12.11 General
PROBLEM 3.84. DES.S
A microthermoelectric cooling device with legtype p and ntype element to elimi
nate substrate head losses inherent on other designs, is shown in Figure Pr. 3.84. With
a number of np pairs N
te
= 128, a quartz slab is cooled by LT = T
T
q
= 20
C
with a load of Q
c
= 0.005 W, using a current J
e
= 0.0017 A, from a 3 V battery.
Here, 0.005 W is also the predicted steadystate load once the quartz slab is cooled.
The performance of this device is affected by the thermal contact resistance R
k.c
between the layers, which are depends on the fabrication process.
(a) Draw the thermal circuit diagram, including the following temperatures, T
, T
p
,
T
h
, T
c
, and T
q
(assume the copper temperature is T
q
). Include the contact resistances
at the interfaces bounding the p and ntype elements, i.e., between T
q
and T
c
, and
T
h
and T
p
, and a conduction resistance R
k.c
between T
p
and T
.
(b) Write the energy equation, including energy conversion terms for the tempera
ture nodes at T
c
andT
h
, andndexpressions for the temperature drops betweeneach
of the temperature nodes above. This should form a linear system of six equation.
(c) Using these six equations from part (b), solve for the temperatures through the
cooler using MATLAB. Plot the temperature difference T
T
q
, as a function of
MassoudKaviany 2011


Quartz
T
q
T
p
Figure Pr. 3.84. A schematic of the microelectric cooler, showing the various parameters.
A
k
R
k.c
from1.0 10
10
to 3.0 10
6
K/(Wm
2
). Sketch the temperature distribution
in the cooler for A
k
R
k.c
= 10
8
. 10
6
. and 3 10
6
K/(Wm
2
).
(d) Plot the coefcient of performance
cop
, as dened by (3.126), for the device,
for contact resistances A
k
R
k.c
from 1.0 10
10
to 3.0 10
6
K/(Wm
2
). What effect
does an increase in the contact resistance have on the device performance? (Hint:
When T
= 298 K,
R
e
= (
e
L,A)k)
p
(
e
L,A
k
)
n
, 1,R
k.hc
= (A
k
k,L)
p
(A
k
k,L)
n
, silicon substrate
resistance is R
k.s
= 25,000 K/W. Use the thermoelectric properties for p and
ntype Bi
2
Te
3
in Table C.9(a) for Bi
2
Te
3
and Sb
2
Te
3
.
PROBLEM 3.85. FUN.S
Molecular uctuations in solids propagate as either transverse or longitudinal waves.
Atransverse wave is a distortion that displaces the atoms in a direction perpendicular
to the direction of propagation, and a longitudinal wave displaces the atoms in the
direction of propagation. These are shown in Figure Pr. 3.85(a).
(i) Transverse Lattice VibrationWave (ii) Longitudinal Lattice Vibration Wave
lm
yj
Atom
x
(Direction of
Wave Propagation)
xj
Atom j
y
y
x
Figure Pr. 3.85. (a) Transverse and longitudinal lattice vibration waves.
MassoudKaviany 2011
The angular frequency and wavelength give phase velocity u
p
as
u
p
=
2
.
It is convenient to use = 2,, where k is the wave number. Then
u
p
=
.
This allows us to dene another velocity, the group velocity u
g
, as
u
g
=
.
The group velocity is the speed at which the boundaries of a pulse travel, and
the phase velocity is the speed of the wave within the pulse. If the group velocity
and the phase velocity are not equal, then the speed at which the wave propagates
depends on its wavelength and the system is said to be dispersive. Dispersion is
the reason a prism separates white light into a spectrum of colors. Upon entering
the prism, the longer wavelengths of light travel faster than the shorter ones. When
the light exits the prism, the different wavelengths have separated enough that
different colors may be seen.
Consider transverse waves in an innite onedimensional lattice. A one
dimensional lattice is a line of atoms, and assume that each atom is only affected by
the force due to its two adjacent neighbors. We can assume that the wave propagates
along the xaxis, while the atoms themselves are displaced along their individual y
axis. In real solids, the lattices are three dimensional, and each atom may be affected
by a number of neighbors.
Each atomexerts an attractive force upon its neighbors. When a transverse wave
passes through the lattice, the atoms are displaced from their common axis. We will
be concerned with the ycomponent of the force, F
j
, exerted on the j th atom. To
simplify the analysis, we will assume that the magnitude of the displacement from
the xaxis is much smaller than the lattice spacing l
m
. We will also assume that a
constant attractive force, F, is exerted between adjacent atoms. These are shown in
Figure Pr. 3.85(b).
y
j1
 y
j
y
j
 y
j+1
(j1)l
m
jl
m (j+1 )l
m
F
j
x
y
j  1
j
j + 1
j 1
F
j
j +1
Figure Pr. 3.85. (b) Designation of atomic location, force, and displacement.
MassoudKaviany 2011
If the atom located at x = j l
m
is displaced by an amount y
j
, then the force it feels
from the atom at x = (j 1)l
m
is
F
j 1
j
= F cos
= F
y
j 1
y
j
l
m
.
The approximation is justiable, because the displacement y
j
is assumed to be
much smaller than the interatomic distance, l
m
. Similarly, the upward force exerted
by the neighbor at x = (j 1)l
m
is
F
j 1
j
= F sin
= F
y
j
y
j 1
l
m
.
We may rewrite the Newton equation of motion, F = ma, in the form a = F,m,
for the atom located at x = j l
m
:
d
2
y
j
dt
2
=
F
ml
m
y
j 1
2
F
ml
m
y
j
F
ml
m
y
j 1
.
Note that this is a purely classical interpretation, and any quantum effects have
been neglected.
(a) Derive the dispersion relation for a transverse wave by inserting the equation for
a wave
y(x. t) = Ae
i(j l
m
t)
.
into the equation derived above and solve for = f (k). Use the following identities
cos x =
1
2
_
e
ix
e
ix
_
. 1 cos 2x = 2 sin
2
x.
(b) The lattice parameter for Pb is l
m
= 0.495 nm. Use an interatomic forces F =
0.931 nN and plot the dispersion curve for lead Pb, = f (), where f () is the
dispersion relationship found in part (a). m = M,N
A
.
(c) Recall that the group velocity is the speed of propagation of a disturbance. In the
longwavelength limit, this is the speed of sound in a solid. Find the speed of sound
in Pb (lead) for small .
PROBLEM 3.86. FUN.S
Amorphous materials have lower thermal conductivities than crystalline structures
of the same composition and similar densities. One example of this is silica (SiO
2
)
which can be found in both states, namely, amorphous glass or crystalline quartz.
Figure Pr. 3.86 shows the solid structure of these two phases and the variation of
their thermal conductivities with respect to temperature (quartz is anisotropic, and
the results are for the cdirection). The modied Einstein approximation for the
minimum thermal conductivity of amorphous materials k
min
is
k
min
=
_
6
_
1,3
k
B
n
2,3
i
u
p.i
_
T
T
i
_
2
T
i
,T
_
0
x
3
e
x
(e
x
1)
2
dx. T
i
= u
p.i
h
P
2k
B
_
6
2
n
_
1,3
.
MassoudKaviany 2011
Unit Cell
4.91 A
0.1
1
10
100
350 100 150 400 300 200 250 50
T, K
k
,
W
/
m

K
Quartz(c)
Quartz(c) (Experimental)
Amorphous Silica (Experimental)
Quartz(c) (Simulated)
Quartz
Amorphous
Figure Pr. 3.86. Effect of temperature on thermal conductivity of crystalline and amorphous
SiO
2
.
where the sum is taken over the single longitudinal and two transverse modes of
sound, u
p.i
is the phonon speed, T
i
is the temperature corresponding to the cutoff
frequency for each mode, and n is the number density of atoms. This model is
based on the energy transferred between an atom and its neighbors through damped
oscillations about their equilibrium positions, and ignores energy transferred by
longwavelength interactions.
k
B
= 1.3807 10
23
J/K, h
P
= 6.6260 10
34
Js, n = 6.61 1028 1/m
3
,
u
p.1
= u
p.l
= 5.8 10
3
m/s, u
p.2
= u
p.3
= u
p.t
= 3.7 10
3
m/s.
(a) Show that at high temperatures k
min
approaches a constant value
lim
T
k
min
=
_
6
_
1,3
k
B
n
2,3
2
i
u
p.i
.
Note that as T becomes large, x becomes small, so it is reasonable to use the Taylor
series approximation, e
x
. 1 x.
(b) Plot k
min
on the same axes as Figure Pr. 3.12. Note that using MATLAB we
can set the derivative of the variable y equal to the integrand, i.e., y
/
= f (t). Then
y =
_
f (t)dt. This is performed separately for the three integrals in the summation.
(c) Using the ideas of the microscale thermal conduction in Section 3.2, explain the
large difference in the magnitudes of the thermal conductivity of quartz and glass.
PROBLEM 3.87. FUN
The Boltzmann transport equation (BTE) discussed in Section 1.7.4, is expressed as
f
t
u
p
f
F
f
u
p
=
f
s
MassoudKaviany 2011
Using the relaxation time assumption and the assumptions of onedimensional,
steadystate, no external force, BTE is linearized and has simple form as
f
f
o
p
.
Under the local equilibrium condition, we have
df
dx
=
df
dT
dT
dx
=
df
o
dT
dT
dx
.
u
p.x
df
o
dT
dT
dx
=
f
o
p
.
and
f
= f
o
p
u
p.x
df
o
dT
dT
dx
.
This phonon distribution function, f
h
P
2
D
p
()d.
where D
p
() is the phonon density of states function.
Using the Debye model, derive (3.26).
PROBLEM 3.88. FUN.S
The thermal contact conductance depends on the properties of the materials in
contact, including their surface roughness, the contact pressure and temperature.
Figure 3.25 gives the thermal contact conductance for pairs of materials, as a function
of the contact pressure and the average surface roughness of the pair materials. A
general correlation for the contact conductance in the elastoplastic regime is given
in [Sridhar, M.R., and Yovanovich, M.M., 1996, "Thermal Contact Conductance
of Tool Steel and Comparison with Model, Int. J. Heat Mass Transfer, Vol. 39,
pp. 831839]. For material pair 1 and 2 in contact, the contact conductance 1,(A
k
R
k.c
)
is given by
1
A
k
R
k.c
=
1
2(2)
1,2
mk
m
(f
ep
)
1,2
exp(
2
)
2
)
1,2
_
1
_
f
ep
2
erfc()
_
1,2
_
3,2
= erfc
1
_
2
f
ep
p
c
E
h
_
.
2
k
m
=
1
k
1
1
k
2
. m = (m
2
1
m
2
2
)
1,2
.
2
)
1,2
= (
2
1
)
2
2
))
1,2
.
where p
c
is the contact pressure, k
m
is the harmonic mean conductivity of the
materials (k
1
and k
2
) in contact, m is the equivalent surface slope of the surface pair
(dimension of m),
2
)
1,2
is the average surface roughness, and f
ep
is the elastoplastic
function. In general, f
ep
depends on the deformation regime (elastic, plastic, or
transitional) and has a general value between 0.5 to 1.0, and 0.8 to 1.0 for the tool
steels. Also, E
h
is the elastoplastic hardness and depends on the deformation regime.
(a) Use the experimental result (from the above reference), for the workhardened
tool steel, given in Table Pr. 3.88, and plot (in loglog scales) these results for the
MassoudKaviany 2011
contact conductance as a function of the ratio between the contact pressure and the
elastoplastic hardness.
(b) On the same graph, plot the prediction from the above correlation, for p
c
/E
h
between 10
5
to 10
2
. Then comment on any agreement (or lack of it) between the
experiment and predicted results. Use k
1
= k
2
= 40 W/mK, m = 0.0892,
2
)
1,2
=
0.981 jm, f
ep
= 0.92.
Note that erfc refers to the complimentary error function, given by erfc(x) = 1
erf(x), and erfc
1
(x) is the inverse complimentary error function [i.e., erfc
1
(x) = y,
or erfc(y) = x].
Table Pr. 3.88. Experimental results for contact conductance as a function of contact
pressures, for a workhardened tool steel pair
p
c
,E
h
10
4
2 10
4
4 10
4
7 10
4
10
3
1,A
k
R
k.c
, W,(m
2
K) 9.1 10
2
1.8 10
3
3.3 10
3
5.1 10
3
7.6 10
3
PROBLEM 3.89. FAM
Diving wet suit are made of a layer of textured neoprene (synthetic rubber) and
when it is wet, the air bubble trapped in the suit (on the texture surface) reduces
the heat conduction. There is also a thin layer of water trapped in the suit and is
warmed by the body and this also reduces the heat loss. There are three major heat
loss areas of the body, namely, the inner thighs (near the femoral artery), under
the arms (near brachial artery), and the head and neck. It is important for the wet
suit to cover all these areas. The thickness of the suit l is selected (from 3 mm to
about 7 mm) based on the duration of the dive and the water temperature T
n
(far
from the body). The surface convection resistance between the suit and water is R
ku
.
The effective conductivity of the neoprene with trapped air and water lm is given
with an empirical value k). Assume that the human body is at a uniform temper
ature T
s
and the metabolic heat generation is
S
rc
during diving. Model the human
body as a cylinder, and neglect the heat losses at the two ends. These are shown in
Figure Pr. 3.89.
A
ku
L
Water
T
w
Geometric Model
of Human Body
Neglect End
Heat Losses
T
s
l
Neoprene
layer
D
S
r, c
Figure Pr. 3.89. Heat loss from the diving body with a wet suit (neoprene layer) to water.
MassoudKaviany 2011
(a) Draw the thermal circuit diagram for the heat transfer from the diving body with
a wet suit.
(b) Determine the steadystate metabolic heat generation
S
rc
for a diver in a T
n
=
15
o
C, wearing a wet suit with thickness (i) l = 3 mm, and (ii) 6 mm.
Use k) = 0.05 W/mK, D= 0.45 m, L = 1.70 m, T
s
= 34
o
C, and A
ku
R
ku
= 1.8
10
3
K/(W/m
2
).
PROBLEM 3.90. FAM
The automobile airmassmeter measures the mass of air ingested by the engine (to
keep the desired airfuel ratio and to reduce the pollutant emission), using the hot
wire anemometry technique. This is based on the wire temperature with a constant
Joule heating, since the wire temperature is related to the air velocity that ows over
the wire. The response time of the hot wire should be kept short to allow for control
under rapid transient driving.
Consider a hot wire of a diameter D and length L, made of gold. During a
transient air ow, the hot wire temperature changes from an initial temperature
T
1
(t = 0) to a nal temperature T
1
, with a Joule heating
S
e.j
, and a constant air
temperature of T
2
. These are shown in Figure Pr. 3.90. Neglect the heat losses from
the two ends.
L
D
S
e,J
(+)
Hot Wire
()
Air Flow
T
2
T
1
(t)
R
ku,12
Figure Pr. 3.90. Hotwire anemometry used to measure a transient air ow.
(a) Using an internal conduction resistance of R
k.1
= 1,Lk and an external resistance
of R
ku.12
= 60 K/W, determine if the assumption of a uniform temperature can be
made.
(b) Draw the thermal circuit diagram.
(c) How long will it take for the hot wire to cool down to T
1
(given below)?
Use the properties of gold from Table C.16.
D= 3 10
5
m, L = 0.05 m, T
1
(t = 0) = 50
o
C, T
1
= 45
o
C, T
2
= 25
o
C,
S
e.j
=
0.25 W, R
ku.12
= 60 K/W.
PROBLEM 3.91. FAM
In pathology of human skin burns (due to exposure to high temperature), the degree
of a burn refers to the extent of damage of the tissue, and is classied based on the
MassoudKaviany 2011
temperature elevation (above normal, 37
o
C) and the duration (time) of the damaged
tissue. For example,
(i) a rstdegree burn designation means that the tissue has been at a temperature
higher than 58
o
C for more than 12 s,
(ii) a seconddegree burn designation means that the tissue has been at a temperature
higher than 53
o
C for more than 50 s, and
(iii) a thirddegree burn designation means that the tissue has been at a temperature
higher than 45
o
C for more than 35 s.
Generally, during an exposure of the skin surface to a high temperature, different
regions can have different degrees of burn. For example, the region close to the skin
surface is more likely to have a serious burn (rstdegree burn) than the region far
from the skin surface.
In an experiment, the surface of skin is suddenly brought in contact with a copper
slab at a temperature of 80
o
C for a duration of 300 s. The initial temperature of the
skin is 37
o
C. Assume the skin is a semiinnite medium.
(a) In Figure Pr. 3.91, plot T versus x, qualitatively for several different elapsed times
t
i
= t
1
. t
2
. t
3
. . . . . 300 s. (Do not perform any calculations.)
(b) Determine the regions (measured from the skin surface) for the above three
different degrees of burn (i)(iii), at the end of the 300 s. Hint: It is not sufcient for
the tissue to have a certain degree of burn by reaching the designated temperature
at the end of 300 s. But, the tissue must be at or above that designated temperature
for the designated duration.
Use c
p
= 3.7 10
6
J/m
3

o
C, and k = 0.293 W/m
o
C.
T,
o
C
x
80
37
0
Skin surface
Skin surface Copper Tissue
x
Figure Pr. 3.91. Fill in the transient temperature distribution qualitatively in the semiinnite
tissue, for different elapsed times, up to t = 300 s.
MassoudKaviany 2011
PROBLEM 3.92. FAM
In ultrafast laser pulse heating of a metal (as in micromachining), with short pulses
(10 ps or less), due to greater intensities of electromagnetic elds, the energy is
absorbed entirely by the electron gas of the metal. The much greater masses
of the ion cores (lattice) barely have time to respond, before the laser pulse has
passed.
Due to low electron gas heat capacity (product of density and specic heat
capacity) for a brief time, there are two temperatures in the metal, i.e., a very hot
electron gas and a lattice remains at ambient temperature. The laser pumpelectron
probe experiments are intended to probe this brief thermal nonequilibrium. The
laser pump provides the energy, and the electron probe provides a direct measure of
the lattice structure by means of electron diffraction (the relative change in intensity
of diffraction spots yields information on lattice vibration, DebyeWaller effect,
which is a measure of temperature).
Laser Pulse
D
1
L
1
Laser Absorption
S
e,
r, 1
Surroundings
T
5)
o
C. No convection and conduction heat transfer.
Use
1
= 10 ps = 10 10
12
s, a
1
=
S
e.
,V
1
n
c
k
B
,
S
e.
= 1.8 10
8
W, n
c
= 5.9
10
28
1/m
3
, k
B
= 1.381 10
23
J/K, D
1
= 600 jm, L
1
= 100
A = 100 10
10
m, t
l
=
150 fs = 150 10
15
s, T
1
(t = 0) = T
2
= 25
o
C, T
1.f
= 200
o
C, T
= 20
o
C, c
r.1
= 0.9,
(c
p
)
1
= 2.490 10
6
J/m
3
K.
MassoudKaviany 2011
4
Radiation
4.3 Thermal Radiometry
Radiation measurements (i.e., radiometry) can be divided into classes, depending on
the wavelength (or frequency) of interest. An example will be the Xray radiometry.
Here we address thermal radiometry which is mostly inthe visible and infraredrange.
A discussion of thermal radiation thermometry is given in [7]. Here we examine the
pyranometer and the infrared surfacetemperature sensor.
4.3.1 Pyranometer
There are two types of thermal radiation sensors, namely, thermal and photoelec
tronic sensors [3]. The photoelectronic sensors (or detectors) directly convert energy
of irradiating photons into electrical signals. Among the thermal sensors are the
bolometer detectors, where the temperature dependence of the electrical resistivity
of the receiver
e
=
e
(T) (the receiver is heated upon irradiation) is used to deter
mine the irradiation ux q
r.i
. The pyranometer is another thermal sensor used for
measurement of solar radiation (i.e., irradiation by the sun) q
r.i
, where the receiver,
which is a collection of thermocouple junctions (a thermopile), is heated by the irra
diation and produces an electrical potential, [as described in Section 2.3.2(B)], which
is used for quantitative detection. It is assumed that the detector is exposed only to
the solar irradiation. Although this name is generally used for the measurement of
the solar radiative heat ux, it is also used for instruments in indirect temperature
measurements. This is discussed below in Section 4.3.2.
The black and white pyranometer, shown in Figure 4.8(a), has a detector consist
ing of a differential thermopile (i.e., several thermocouple junctions) with the hot
junction receivers blackened and the coldjunction receivers whitened. The receiver
is of radial wirewound (constantan, an alloy, as material) and plated thin lm (cop
per) construction [also shown in Figure 4.8(a)]. This is the type T thermocouple
shown in Figure 2.10.
Builtin temperature compensation is made with a thermistor circuitry that
uses the dependence of the electrical resistivity
e
on temperature. The precision
MassoudKaviany 2011
q
r,i
Fused
Quartz
Glass
Black Paint
Under the White and Black Coating:
(ii) Thermopile Under Black
and White Paints
(i) Solar Pyranometer
Thermocouple
Junction
Copper
(Plated Thin Film)
, Measured Voltage
Constantan (Wire)
D = 0.2 mm
Electrical Insulator
Substrate
T
b
T
w
Figure 4.8. (a) Asolar pyranometer showing the sensor surface and the thermocouples placed
beneath the blackened surface (T
b
) and whitened surface (T
n
).
fused quartz glass hemisphere, also depicted, has a good transmission from 0.2 to
4.0 jm. This transmission range covers the ultraviolet, visible and part of near
infrared, making this particular pyranometer suitable for measurement of the solar
irradiation q
r.i
(W/m
2
). Transmission of quartz is discussed in Section 4.3.2.
The output is about 10 jV/(W/m
2
) and is linear to within 1% for q
r.i
from 0
to 1,400 W/m
2
. This upper limit is slightly larger than the average maximum solar
irradiation q
r.i
s
= 1,353 W/m
2
available just outside the earths atmosphere.
The sun has a diameter of 13.9 10
5
kmand its outermost layer has an equivalent
blackbody temperature of 5,780 K (the center of sun has an estimated temperature
of 2 10
7
K). The sun is assumed to be a blackbody, c
r
= 1, and according to (4.13)
emits Q
r.c
= A
r
E
b
= 3.8 10
26
W, out of which 1.7 10
17
W is intercepted by the
MassoudKaviany 2011
D
q
r,i
r
E
b
( ) T
s
T
s
T
s
D
L
r
L
Emissive Power
Aperture
, Thermocouple
Voltage Output Internal Compensation
Circuit and Signal Processor
Thermopile
(i) Infrared SurfaceTemperature Sensor
(iii) Electrical Circuitry
(iv) Spectral Transmissivity of Lens Materials (L = 1 mm)
(ii) Thermopile
Uncoated
Germanium
Coated (10 m)
Germanium
Sapphire
SodaLime Glass
Visible Near Infrared
Fused Quartz
Irradiation q
r,i
Reflected Irradiation:
Assumed Negligible
Thermopile
SpotWelded
Thermocouple
Lens
()
(+)
0.3
0
1.0
1.0 3.0 10 30
r,
, m
q
r,i
.
Absrobed
Irradiation S
e,
r,
S
e,
.
Figure 4.8. (b) An infrared surfacetemperature sensor showing (i) sensor, (ii) thermopile, (iii)
related electronic circuits, and (iv) spectral transmissivity of different lens materials (including
coated and noncoated germanium) in the visible and near infrared range.
earth located 1.5 10
8
km from the sun. The budget (i.e., breakdown into various
fractions) of the solar energy received by the earth is given in Figure Ex. 1.2(a).
4.3.2 Infrared SurfaceTemperature Sensor
In addition to the measurement of irradiation, infrared pyrometry is used for the
surface temperature measurement. This is done by relating the measured irradiation
ux q
r.i
to the surface temperature T of the emitter (it can also be applied to volu
metric emission). In order to make the relation, generally the total, hemispherical
MassoudKaviany 2011
emissivity of the surface c
r
is also used. The concept is depicted in Figures 4.8(b)(i)
to (iii). The builtin signal converter results in an output similar to that of a thermo
couple (so they are called surface infrared thermocouples).
It is assumed that q
r.i
= c
r
E
b
, i.e., the irradiation to the sensor is equal to the
surface emissive power (i.e., the background irradiation reected by the surface is
assumed negligible).
The lens focuses the radiant energy onto a thermopile (a collection of several
thermocouples). Depending on the surface temperature and the wavelength band
over which a signicant fraction of the energy is emitted, different lens materials
are used. Figure 4.8(b)(iv) also shows some typical lens materials and their spectral
transmissivity
r.
as a function of wavelength. Note that the germanium lens stops
all the visible radiation from reaching the thermopile. This is desirable in eliminating
the visible background radiation. Also note that the germanium lens allows for
the transmission of the long wavelengths (near infrared), while the conventional
sodalime glass (and borosilicate and fused silica glass) does not. The left (or short
wavelength) limit of photon absorption is electronic (corresponding to electron
energy gap E
e.g
) and the right limit (long wavelength) is by multiphonon absorption
E
p
. The principle region of transparency for germanium is between = 2 and 20 jm
and its refractive index has a high value of n
ex.
L
, where
ex.
is the spectral extinction (or absorption) coefcient and is related to the spectral
extinction index through
ex.
= 4
, 1,
ph
.
A reference temperature or junction is needed and this is provided by an appro
priate reference sensor (generally a thermocouple) as part of the compensation
circuitry [3]. For increased accuracy, a cold reference maintained at a low tempera
ture is used. Because of the required signal conditioning (as part of the compensation
and signal conversion circuitry), the temperature range of the applicability is limited.
By bypassing this circuitry, a larger applicable range can be used (but the ease of
readout and linearity will be lost). In some sensors, a value of c
r
= 0.9 is assumed
and some other sensors would allow for the input of value of c
r
. Accuracy within 1
C
is found for those operating over a small temperature range (e.g., 0 to 100
C).
A ratio infrared thermometer alleviates the need for the input of the emissivity.
It samples the irradiation at two different wavelengths and uses the ratio and the
absolute values of these to determine the emissivity and the temperature.
4.5 Prescribed Irradiation and Nongray Surfaces
In Section 4.4 we discussed enclosure radiation exchange, where the irradiation and
radiosity for each surface were determined as part of the enclosure heat transfer for
The expression for the absorption coefcient is derived from the Poynting vector which is the cross
product of electric and magnetic eld.
MassoudKaviany 2011
10
8
10
10
10
12
10
14
10
7
10
9
10
11
10
13
t, s
Laser
Drilling
Laser
Glazing
Laser Transform
ation Hardening
Laser Welding
and Cutting
10
8
10
6
10
4
10
4
10
2
1 10
7
10
5
10
3
10
1
10
10
21
10
20
10
19
10
22
10
15
10
9
10
14
10
16
10
6
10
8
(q
r, i
)
l
t = 10
10
J/m
2
(
q
r
,
i
)
l
,
W
/
m
2
Spot Superheating of Metals
Plasma Forming
Laser Shock
Hardening
Fetmosecond
Nanoscale
Figure 4.15. Application of irradiation heating by laser in manufacturing processes. The laser
irradiationux (q
r.i
)
l
andthe durationof irradiationare shownfor various processes. (Adapted
from Barnas, C.M., reproduced by permission c 1993, ASME.)
gray, diffuse, and opaque surfaces. In some applications, a spectral q
r.i.
or total q
r.i
irradiation is prescribed (as in solar irradiation or irradiation by a laser or another
engineered source). In most prescribed irradiations, the dominant wavelength range
of irradiation is different than that associated with the surface emission. An example
is in solar irradiation on a solarcollector surface where the dominant wavelengths
are in the visible range for the irradiation and infrared for the surface emission.
Therefore, the total absorptivity and the total emissivity of the surface will not
be equal (i.e., the surface behaves as a nongray surface). Here we will consider
surface radiation from a surface without invoking the radiation enclosure and other
assumptions made in Section 4.4. We begin by discussing some irradiation sources
and then we analyze the nongray, nonenclosure surfaceradiation heat transfer.
4.5.1 Laser Irradiation (q
r.i
)
l
Laser irradiation heating is used in manufacturing. Wavelength in the ultraviolet,
visible, and near infrared range are used, depending on the material and objective.
The irradiation ux q
r.i
required for some of these processes is very large. Fig
ure 4.15 gives examples of some manufacturing processes, the associated required
laser irradiation ux (q
r.i
)
l
, and the duration of irradiation Lt. The applications
(objectives) include spot superheating, plasma forming, shock hardening, glazing,
drilling, welding, cutting, and hardening. Some require a large (q
r.i
)
l
and a small
Lt, and some a smaller (q
r.i
)
l
and a larger Lt. With high (q
r.i
)
l
) and Lt  10
13
s,
MassoudKaviany 2011
10
13
10
14
10
15
10
12
10
11
10
10
10
9
10
8
10
7
10
6
10
5
10
4
No Welding Possible
Electron
Beam
Vaporization
Dominated
Vaporization,
Conduction,
and Melting
(Keyholing)
(b) Keyhole Geometry in Laser Welding (a) Surface Heat Flux for Welding
Depth of Weld
Penetration
u
F
, Speed of Travel
Radial
Conduction
Dominated,
with Melting
Negligible
Melting
CO
2
Laser
( = 10.4 m)
Glow
Discharge
Electron
Beam
Plasma
Arc Processes
No Welding Possible
w
Melted Zone
Width Molten Pool
Surface Evaporation
Keyhole
Workpiece
l
Irradiation, (q
r,i
)
l
, W/m
2
q
s
,
(
q
r
,
i
)
l
,
W
/
m
2
Figure 4.16. (a) Application of lasers in welding. (b) The keyhole melt by irradiation heating.
(From Lancaster, J.F., reproduced by permission c 1986, Pergamon.)
several nonequilibrium phenomena occur in metallic solids. These include nonequi
librium between electron (having higher temperature) and phonon (having lower
temperature), nonequilibrium of vibrational (phonon) modes, and heating at con
stant volume. For (q
r.i
)
l
> 10
18
W/m
2
, the irradiated matter (gas, liquid, or solid) will
ionize. Then the formed plasma also interacts with the incoming photons.
Figure 4.16 shows the surface heat ux, q
s
or (q
r.i
)
l
, required in welding [1]. The
highpower density (i.e., (q
r.i
)
l
> 10
10
W/m
2
) welding allows for deep melting (more
specically, the depth of the weld is larger than the width), as depicted in Figure
4.16. The CO
2
lasers (these are gas lasers with a narrow band of emission around
= 10.4 jm, i.e., near infrared, which is invisible) can provide the needed power
density (q
r.i
)
l
. Absorption of these near infrared photons in solids is by interaction
with phonons (lattice vibration). The lasers can provide for very ne focusing and
reliable continuous power. When the beamis allowed to pass through the workpiece,
it is called breakthrough [6]. The gravitational force and surface tension oppose the
breakthrough, while the beam pressure, vapor pressure, and the recoil pressure
(the force exerted by the vapor molecules departing the evaporation surface) assist.
Metals generally have a low absorptivity, and therefore, high irradiation uxes q
r.i
are needed.
Lasers use solidstate or gaseous nonequilibrium molecular excitation for
phonon emission (as compared to equilibrium thermal emission), and therefore,
operate at lower temperatures. The laser beam is also focused to increase (q
r.i
)
l
.
Note that to achieve (q
r.i
)
l
= 10
10
W/m
2
by thermal radiation, a blackbody source of
T = 2 10
4
K would be needed (which is above the melting temperature of solids).
So, the high irradiation ux is due to nonthermal emission and in highly focused
MassoudKaviany 2011
2,400
1,600
800
H
2
O
H
2
O
H
2
O
CO
2
O
2
O
3
0
0.2 0.8
Measured on Earth Surface
Blackbody Emission, T = 5,762 K, q
r,
= 6.250 10
7
W/m
2
Measured Outside Earth's Atmosphere, (q
r,i
)
s,max
= 1,353 W/m
2
O
3
, H
2
O, O
2
, and CO
2
Absorption Bands
Visible
IR
UV
1.4 2.0 2.6
(
q
r
,
i
)
,
s
,
W
/
m
2

m
, m
Measured Spectral Solar Irradiation
O
3
CO
2
H
2
O
H
2
O
Figure 4.19. Effective spectral absorptivity (also emissivity) and absorption (also used for
emission) bands for a gaseous CO
2
volume. The numbers shown are band designations in jm.
CO
2
bands are in the near infrared range. The results are for T
f
= 830 K, p
CO
2
= 10 atm,
and a mean beam length of
ph
) = 0.388 m. (From Kaviany, M., reproduced by permission
c 2001, SpringerVerlag.)
The concentrations of the gaseous species CO
2
and H
2
O is given in terms of
their partial pressures, p
CO
2
and p
H
2
O
, and the total gas pressure p. Figure 4.19
shows the spectral absorptivity and absorption (also used for emission) bands of
CO
2
at p
CO
2
= 10 atm and T
f
= 830 K [4]. The various bands are designated with
the wavelength at the center of the band. All of the CO
2
absorption bands are in
the near infrared range. Note that decreases to the right and all these bands are
in the near infrared. Other band absorptions are H
2
O (2.7 and 6.3 jm), CO (2.35,
4.67 jm), CH
4
(3.3 and 7.6 jm), and NO (5.35 jm). Again, all of these bands are
in the infrared range (not observable by eye). Due to the high concentration in the
gaseous combustion products, CO
2
and H
2
O are the most signicant contributors to
the gaseous combustion radiation.
The volume of the gas is also signicant in determining the irradiation from the
gas to bounding surfaces. On the one hand the radiation from these gaseous species
and soot attenuates as it travels toward the irradiation surface, and on the other
hand the larger the volume of radiating gas the larger the irradiation. In general,
an average length for this volumetric radiation is used. This is called the mean
beam length
ph
) (an average path length traveled by a photon within a volume of
absorbing/emitting medium). The mean beam length
ph
) is a geometric parameter
and is equivalent to the radius of a hemispheric gas volume emitting the same heat
ux to the center of its base as the average ux emitted to the area of interest by the
actual volume of the gas.
Then the effective emissivity of the radiating gas c
r.f
) is given in terms of
the product of
ph
) and p
CO
2
, p
H
2
O
, or c
s
(for a given total pressure, fuel, and gas
temperature T
f
). The pressure dependence represents the higher effective emissivity
MassoudKaviany 2011
expected as the number density n(1/m
3
) of the participating molecules (here H
2
O
and CO
2
) increases. The number density is related to pressure through (3.18). Also
the larger
ph
), the larger is the total number of such molecules. The products
p
i
ph
), where i stands for H
2
O or CO
2
, as low as 0.001 atmm can result in a
noticeable c
r.f
). Because c
r.f
) depends on
ph
), it is also called the emittance.
For the case of absorbing/emitting gas (no scattering), the equation of radiation transfer, which
describes the conservation of radiation intensity (and radiation heat ux vector), becomes (in a
onedimensional radiation heat transfer with a gray gas)
d(q
r.i
)
f
dx
= ( s
e.
)
f
( s
e.c
)
f
=
ex
(q
r.i
)
f
ex
SB
T
4
f
.
This is a representation of the equation in the footnote of Section 2.3.2(E). After integration and
for a uniform T
f
and using (q
r.i
)
f
= (q
r.i
)
f
(x = 0) at x = 0, we have
(q
r.i
)
f
= (q
r.i
)
f
(x = 0)e
ex
x
SB
T
4
f
(1 e
ex
x
)
r.f
)(q
r.i
)
f
(x = 0) c
r.f
)
SB
T
4
f
.
Here e
ex
x
is called the transmittance
r.f
), and 1 e
ex
x
is called the absorptance
r.f
) and is
equal to the emittance c
r.f
) for the assumed gray gas.
The emission/absorptionby CO
2
andH
2
Oandby soots are integratedover the volume (the geometry
can be one, two, or three dimensional, requiring integration in one or more spatial directions) of
the gas and averaged over the surface. This is represented as the average gas (i.e., ame) irradiation
heat ux (q
r.i
)
f
and is given by (4.62). Note that
ph
) represents the average photon travel distance,
from locations within the gas volume to the bounding surface. In (4.62),
ex
, which is a measure of
the gas species and soot constituents, is combined with
ph
) and is represented by c
r.f
).
MassoudKaviany 2011
Table 4.4. Mean beam length
ph
) for several contained gas geome
tries. (FromOzisik, M.N., reproduced by permission c 1973 Wiley)
Hemisphere Radiating to Element at
Center of Base
Circular Cylinder of SemiInfinite
Height Radiating to Concave
Bounding Surface
Circular Cylinder of SemiInfinite
Height Radiating to Entire Base
R
R
R
Sphere Radiating to Its Surface
Infinite Slab of Gas Radiating to
Both Bounding Planes
Cube Radiating to a Face
Rectangular Parallelepiped
L x L x 4L Radiating to:
L x 4L Face:
L x L Face:
All Faces:
2L x 6L Face:
L x 6L Face:
L x L Face:
All Faces:
Rectangular Parallelepiped
L x 2L x 6L Radiating to:
q
r, i
q
r, i
L
L
q
r, i
R
q
r, i
q
r, i
q
r, i
2L
6L
L
4L
L
L
highly sooted gases, the CO
2
and H
2
O emissions will be mostly absorbed by the
soot and become insignicant. Subject to these limitations and assuming a negligible
band overlap and interaction (i.e., Lc
r
= 0), we can sum the emissivities as
c
r.f
) = c
r.CO
2
) c
r.H
2
O
) c
r.soot
) Lc
r
(overlap and interaction). (4.63)
Figure 4.20 gives the three components of this gaseous emissivity (also called emit
tance).
MassoudKaviany 2011
r
,
C
O
2
0.004
200
p = 1 atm
p = 1 atm
p
CO
2
ph
, atmm
p
H
2
O
ph
, atmm
500 1,000 1,500 2,000
0.010
0.020
0.050
0.100
0.001
0.001
0.002
0.005
0.01
0.02
0.05
0.1
0.2
0.5
1
2
5
10
0.003
0.008
0.02
0.05
0.1
0.3
0.8
2.0
0.200
0.250
T
f
, K
(a) Emissivity of Gases Containing CO
2
(b) Emissivity of Gases Containing H
2
O
r
,
H
2
O
0.001
0.01
0.1
1.0
0 500 1,000 1,500 2,000 2,500 3,000
T
f
, K
T
f
= 2,000 K
1,500
1,250
1,000
750
(c) Emissivity of Gases Containing Propane Soot
s
ph
, m
r
,
s
o
o
t
1.0
0.1
0.01
10
8
0.3 x 10
8
10
7
10
6
S
e.
= A
r.
r.s
q
r.i
= A
r.
_
0
r.
(q
r.i.
)
i
d
irradiation absorbed
by surface.
Alternatively, when the total irradiation and irradiation surface temperature T
i
are
prescribed,
S
e.
can be written as
S
e.
=
A
r.
q
r.i
_
0
r.
E
b.
(T
i
)d
_
0
E
b.
(T
i
)d
. (4.64)
where A
r.
is the surface receiving the irradiation and we have chosen q
r.i
to be
designated as positive to be consistent with the treatment so far [i.e., (4.22)].
Note that the spectral Kirchhoff law (4.10) is usable here, i.e.,
r.
= c
r.
.
Similarly, when the spectral emissivity c
r.
and the surface temperature T
s
are
known, the amount of energy emitted (treated as a surface energy conversion) is
found from Table 2.1, and is
S
e.c
= A
r.c
c
r.s
E
b
(T) = A
r.c
_
0
c
r.
E
b.
(T
s
)d emitted radiation, (4.65)
where A
r.c
is the surface emitting the radiation.
Table C.19 shows that the total emissivity and total absorptivity are not equal
when considering solar irradiation and surface emission. For example, glass is a good
infrared emitter and a poor solar (i.e., visible) absorber.
Figure 4.21 shows a prescribed irradiation ux q
r.i
as well as the surface heat
transfer Q[
A
from an opaque surface at temperature T
s
, having nongray surface
radiation properties
r.s
and c
r.s
.
Now using the surface energy equation (2.62), we have
Q [
A
=
_
A
(q s
n
)dA=
S
i
surface energy equation
=
S
e.
S
e.c
S = A
r.
r.s
q
r.i
A
r.c
c
r.s
E
b
(T
s
)
S
surface energy
conversion,
(4.66)
where we have included surface emission and
S combines all nonradiation surface
energy conversions.
MassoudKaviany 2011
S
e,
= A
r,
r,s
q
r,i
q
r,i
S
e,
= A
r, r,s
E
b
(T
s
)
S
Q
A
Control
Surface A
Opaque Surface
Other Modes of Heat Tranfer
(Including Phase Change)
Irradiation
Nongray Surface (
r

r
) with Prescribed Irradiation and Surface Emission
s
n
Emission
A
r,
r,s
q
r,i
Reflection
Absorption
T
s
,
r,s
,
r,s
Figure 4.21. Prescribed irradiation of nongray surfaces.
Here we have treated irradiation and surface emission as surface energy conver
sion terms
S
e.
and
S
e.c
. Note that the surfaces for absorption and emission may not
be the same (e.g., irradiation may be upon only one side of a surface, while emission
can be from both sides).
EXAMPLE 4.9. FAM
A cubic workpiece (each side having a dimension of 5 cm) made of pure iron
is irradiated by a laser beam as depicted in Figure Ex. 4.9(a). The beam has a
continuous irradiation A
r.
(q
r.i
)
l
= 2 10
4
W. After an elapsed time, the average
temperature of the workpiece is assumed to be constant and at T
s
= T
sl
=
1,538
S
e.c
= A
r.c
c
r.s
E
b.s
),
S
sl
results in melting of the workpiece. Assume that other than surface absorption,
emission, and phase change, there is no other signicant heat transfer (i.e.,
Q [
A
= 0). The surface absorptivity is
r.s
= 0.3 and the surface emissivity is
c
r.s
= 0.4. The heat of fusion of the elements is given in the periodic table, Table
C.2, in Appendix C.
(a) Draw the thermal circuit diagram.
(b) Using the surface energy equation (4.66), determine the rate of melting,
N
sl
=
S
sl
,Lh
sl
.
SOLUTION
(a) The thermal circuit diagram is shown in Figure Ex. 4.9(b). Here we have
no surface heat transfer, Q[
A
= 0, and we treat the nongray surface irradiation
and emission as energy conversion. In the surface energy equation (4.66),
S
sl
is
the only other surface energy conversion. Then , for the case of no surface heat
transfer, we have
Q [
A
= 0 =
S
i
= A
r.
r.s
(q
r.i
)
l
A
r.c
c
r.s
E
b
(T
s
)
S
sl
.
MassoudKaviany 2011
(a) Physical Model (b) Thermal Circuit Model
L
y
L
x
L
x
= L
y
= L
z
= 5 m
z
y
x
L
z
A
r,
Q
A
r
(q
r,i
)
l
(q
r,i
)
l S
e,
= A
r,
q
r,
A
r,
Laser
Workpiece (Pure Iron)
at Uniform Temperature
T
s
= T
sl
s
n
Q
A
= 0
Surface Radiation
No Other Surface
Heat Transfer
Surface Phase Change
S
sl
S
e,r
= A
r,
r
q
r,i
A
r,
q
r,
S
e,
= A
r,
r
q
r,i
T
s
Figure Ex. 4.9. A workpiece (substrate) irradiated by a laser. (a) Physical model. (b)
Thermal circuit diagram.
(b) The above energy equation is written for
S
sl
as
S
sl
= A
r.
r.s
(q
r.i
)
l
A
r.c
c
r.s
E
b
(T
s
)
N
sl
Lh
sl
A
r.c
= 6L
2
x
for L
x
= L
y
= L
z
.
Solving for
N
sl
we have
N
sl
=
r.s
A
r.
(q
r.i
)
l
A
r.c
c
r.s
E
b
(T
s
)
Lh
sl
.
Using the numerical values, we have
Lh
sl
= 2.473 10
5
J/kg Table C.2
N
sl
=
_
0.3 (2 10
4
)(W) 6(0.05)
2
(m
2
) 0.4 5.670 10
8
(W/m
2
K
4
)
(1,538 273.15)
4
(K
4
)
_
,
_
2.473 10
5
(J/kg)
_
=
6 10
3
(W) 3.660 10
3
(W)
2.473 10
5
(J/kg)
= 3.909 10
2
kg/s = 39.09 g/s.
COMMENT
Note that once the melting temperature is reached, and under the conditions
assumed, a high rate of melting is achievable. This melting is adjacent to the
location of the laser impingement. Considering the small beam diameters (of
the order of 1 mm), very rapid melting rates become possible. In practice the
workpiece is moved or the laser is turned off after a short pulse.
MassoudKaviany 2011
EXAMPLE 4.10. FAM.S
Solar irradiation is used to heat a bounded water stream (coolant) as shown in
Figure Ex. 4.10(a). The location on the earth is 20
S and 20
E. Assume that a
typical summer daytime average irradiation is 2.5 times the annual average given
in Figure 4.18. The solar collector surface is a nickel, Tabor solar absorber (Table
C.19). The surface geometry is square with L = 1 m. The heat is removed from
the underneath of the collector surface by surface convection to the coolant
owing with an average entrance temperature of T
f
)
0
= 25
C. The average
surfaceconvection resistance is R
ku
)
L
= 0.05 K/W. Neglect all other heat losses
from the collector surface.
(a) Draw the thermal circuit diagram.
(b) Determine the collector surface temperature T
s
.
(b) Thermal Circuit Model
(a) Physical Model of a Solar Collector
Q
ku
=
T
s
T
f 0
R
ku L
R
ku L
Ideal Insulation
Q
ku L
=
T
s
T
f 0
R
ku L
S
e,
= A
r, r,s
SB
T
s
4
A
r,
= A
r,
A
ku
S
e,
+ S
e,
= A
r,
r,s
q
r,i
A
r, r,s
SB
T
s
4
S
e,
= A
r,
r,s
q
r,i
A
r,
(1
r,s
)(q
r,i
)
s
Coolant, T
f 0
L
L
(q
r,i
)
s
Control Surface
T
f 0
T
s
T
s
,
r,s
Q
s
= 0
Figure Ex. 4.10. A solar collector irradiated by the sun, with a coolant removing the
collected energy. (a) Physical model. (b) Thermal circuit diagram.
SOLUTION
(a) The thermal circuit diagram is shown in Figure Ex. 4.10(b).
(b) From the surface energy equation for nongray, irradiated surfaces (4.66), we
have
Q [
A
= Q
ku
)
L
=
S
e.
S
e.c
= A
r.
r.s
(q
r.i
)
s
A
r.c
c
r.s
SB
T
4
s
.
MassoudKaviany 2011
From Table C.19, we have for nickel: Tabor solar collector (at T = 308 K)
r.s
= 0.85. c
r.s
= 0.11.
The area is
A
r.c
= A
r.
= A
ku
= LL = 1 (m) 1 (m) = 1 m
2
.
Then we substitute for Q
ku
and we have
Q
ku
)
L
=
T
s
T
f
)
0
R
ku
)
L
= A
r.c
[
r.s
(q
r.i
)
s
c
r.s
SB
T
4
s
].
FromFigure 4.21, we have (q
r.i
)
s
= 250 W/m
2
for the location 20
S and 20
E.
Using the numerical values, we have
T
s
298.12(K)
5 10
2
(K/W)
= 1(m
2
) [0.85 2.5 250(W/m
2
)
0.11 5.670 10
8
(W/mK
4
) T
4
(K
4
)].
The solution for T
s
is found numerically, and we have
T
s
= 321.2 K
= 48.05
C.
COMMENT
Note that this surface temperature results in a thermobuoyant ow, and there
fore, a heat loss to the ambient air. Using a lower average surfaceconvection
resistance, for example, R
ku
)
L
= 0.01 K/W, leads to a lower surface temperature
of T
s
= 302.8 K = 29.65
C.
EXAMPLE 4.11. FAM.S
Apropane (C
3
H
8
)air ame is formed above a nongray stainlesssteel (Type 410)
surface, as shown in Figure Ex. 4.11(a). The gas is at temperature T
f
= 1,800 K,
and it irradiates the surface whose thickness is L = 15 cm, which is smaller
than the boundary surface dimensions, which are L
1
= 50 cm and n = 60 cm
PropaneAir Flame
(b) Thermal Node Model (a) Physical Model
Q
s
S
e,
+ S
e,
T
s
L
L
1
>> L
w>> L
T
f
(q
r,i
)
f
Q
s S
e,
= A
r, r,s
SB
T
s
4
S
e,
= A
r,
r ,s
(q
r,i
)
f
T
s
,
r,s
Figure Ex. 4.11. The inniteslab propaneair ame irradiating to its bounding surface.
(a) Physical model. (b) Thermal circuit diagram.
MassoudKaviany 2011
(so an inniteslab gas geometry can be used). The partial pressure of CO
2
is
p
CO
2
= 0.10 atm. The partial pressure of H
2
Ois p
H
2
O
= 0.13 atm. The soot volume
fraction is c
s
= 10
7
. Assume that Lc
r
= 0 (i.e., no band overlap or interactions).
(a) Draw the thermal circuit diagram.
(b) Determine the irradiation ux from the ame (gas).
(c) Determine the surface temperature T
s
for a backsurface heat removal rate
of Q
s
= 2.1 10
4
W.
SOLUTION
(a) The thermal circuit diagram is shown in Figure Ex. 4.11(b). The irradiation
and surface emissions are treated as energy conversion. The surface is nongray
(c
r
and
r
for the surface are listed in Table C.19).
(b) The ame irradiation ux is found from (4.62), i.e.,
(q
r.i
)
f
= c
r.f
)
SB
T
4
f
.
To determine c
r.f
), we begin with the mean beam length
ph
). From
Table 4.4, we have
ph
) = 1.8L Table 4.4
= 1.8 0.15(m)
= 0.27 m.
Then, from Figure 4.20(a), for CO
2
, we have
CO
2
: p
CO
2
ph
) = 0.10(atm) 0.27(m)
= 0.027 atmm
c
r.CO
2
) . 0.05 for T
f
= 1,800 K Figure 4.20(a).
For H
2
O, from Figure 4.20(b), we have
H
2
O : p
H
2
O
ph
) = 0.13(atm) 0.27(m)
= 0.0351 atmm
c
r.H
2
O
) . 0.06 for T
f
= 1,800 K Figure 4.20(b).
For soot from Figure 4.20(c), we have
Soot : c
s
ph
) = 10
7
0.27(m)
= 2.7 10
8
m
c
r.soot
) . 0.06 for T
f
= 1,800 K Figure 4.20(c).
MassoudKaviany 2011
Now assuming that all band overlaps and interactions are negligible, we
have from (4.63)
c
r.f
) = c
r.CO
2
) c
r.H
2
O
) c
r.soot
)
= 0.05 0.05 0.06 = 0.16.
Then, the irradiation from the ame is
(q
r.i
)
f
= 0.16 5.67 10
8
(W/mK) (1,800)
4
(K
4
)
= 9.52 10
4
W/m
2
= 95.2 kW/m
2
.
This is rather substantial.
(c) The surface temperature is determined from the surface energy equation
(4.66), i.e.,
Q
s
=
S
e.
S
e.c
= A
r.
r.s
(q
r.i
)
f
A
r.c
c
r.s
SB
T
4
s
.
where
A
r.
= A
r.c
= Ln = 0.5(m) 0.5(m)
= 0.3 m
2
.
The solidsurface emissivity and absorptivity are found from Table C.19, i.e.,
stainlesssteel type 410: c
r.s
= 0.13.
r.s
= 0.76 for T
s
= 308 K Table C.19.
Solving the above energy equation for the surface temperature T
s
, we have
T
s
=
_
A
r.
r.s
(q
r.i
)
f
Q
s
A
r.c
c
r.s
SB
_
1,4
=
_
0.3(m
2
) 0.76 9.52 10
4
(W/m
2
) 2.1 10
4
(W)
0.3(m
2
) 0.13 5.67 10
8
(W/m
2
K
4
)
_
1,4
=
_
2.171 10
4
(W) 2.1 10
4
(W)
2.211 10
9
(W/K
4
)
_
1,4
= 751.6 K = 478.5
C.
COMMENT
Note that our treatment of the gas emissivity uses some simplications (negligi
ble band overlaps and interactions). Also, the surface temperature is higher than
the temperature in Table C.19 for the surface radiation properties. Therefore,
all the radiation properties are rather approximated.
MassoudKaviany 2011
4.8 References
[1] Banas, C.M., 1993, Multikilowatt Lasers in Manufacturing, ASME J. Engi
neering for Gas Turbine and Power, Volume 115, pp. 172176.
[2] Brewster, M.Q., 1992, Thermal Radiative Transfer and Properties, Wiley, New
York.
[3] Harrison, T.R., 1960, Radiative Pyrometry, and Its Underlying Principles of Radi
ant Heat Transfer, Wiley, New York.
[4] Kaviany, M., 2001, Principles of Convective Heat Transfer, Second Edition,
SpringerVerlag, New York.
[5] Kreith, F., and Kreider, J.F., 1978, Principles of Solar Engineering, Hemisphere
Publishing Corporation, Washington.
[6] Lancaster, J.F., Editor, 1986, The Physics of Welding, Second Edition, Pergamon
Press, Oxford.
[7] Modest, M.F., 1993, Radiative Heat Transfer, McGrawHill, New York.
[8] Ozisik, M.N., 1973, Radiative Transfer and Interactions with Conduction and
Convection, Wiley, New York.
[9] Ozisik, M.N., 1985, Heat Transfer: A Basic Approach, McGrawHill, New York.
[10] Siegel, R., and Howell, J.R., 2002, Thermal Radiative Heat Transfer, Fourth
Edition, Taylor & Francis, New York.
[11] Kaviany, M., 2008, Heat Transfer Physics, Cambridge, New York.
4.9 Problems
PROBLEM 4.31. FUN
When thermal radiation penetrates a semitransparent medium, e.g., a glass plate,
reection, absorption, and transmission occur at the interface between two adjacent
media along the radiation path, e.g., each of the glass/air interfaces for a glass plate
surrounded by air. These multiple absorptions and reections result in an overall
attenuation, absorption, and reection of the radiation incident in the glass plate.
These effects are modeled as overall transmittance, absorptance, and emittance. The
glass plate is then assumed diffuse and gray.
Consider a glass plate bounded by its two innite, parallel surfaces, as shown
in Figure Pr. 4.31. The glass plate has a transmittance
r.2
= 0.1 and an absorptance
r.2
= 0.7. Its temperature T
2
is assumed uniform across the thickness. The surfaces
are opaque, diffuse, and gray with total emissivity c
r.1
= c
r.3
= 0.5.
(a) Show that the net heat transfer by surface radiation from both sides of the glass
plate is given by
Q
r.a
A
r.2
=
1
r.2
(1
2
)
_
[
r.2
(1
2
) 1](q
r.o
)
a
(q
r.o
)
b
(1 )c
r.2
_
Q
r.b
A
r.2
=
1
r.2
(1
2
)
_
[
r.2
(1
2
) 1](q
r.o
)
b
(q
r.o
)
a
(1 )c
r.2
_
.
where =
r.2
,
r.2
.
(b) Write a system of equations that would allow for the solution of the problem and
identify the set of variables that are being solved for and the number of variables
that need to be known.
MassoudKaviany 2011
(q
r,i
)
3
Q
3
Q
2
L
w
(q
r,i
)
1
(q
r,o
)
a
Surface a
Opaque Surface 3
Opaque Surface 1
Surface b
Semitransparent
Medium, T
2
(q
r,o
)
b
r, 3
(q
r,i
)
3
r, 1
(q
r,i
)
1
r, 3
E
b,3
r, 1
E
b,1
(q
r,o
)
3
(q
r,i
)
a
r, 2
(q
r,i
)
a
r, 2
(q
r,i
)
b
r, 2
E
b,2
r, 2
E
b,2
r, 2
(q
r,i
)
b
r, 2
(q
r,i
)
a
r, 2
(q
r,i
)
a
r, 1
(q
r,i
)
1
r, 2
(q
r,i
)
b
(q
r,i
)
b
(q
r,o
)
1
Q
1
r, 2
,
r, 2
,
r, 2
r, 2
,
r, 2
,
r, 2
r, 3
, T
3
r, 1
, T
1
r, 2
(q
r,i
)
2
Figure Pr. 4.31. Radiative heat transfer across a glass plate. The glass is placed between two
parallel solid surfaces.
(c) If there is no other heat transfer from the glass plate, write the energy equation
to be solved the glass plate temperature T
2
and write the expression for T
2
.
(d) Compare the results of this problem with the results of Problem Pr. 4.31. Is
there an analogy between the transmittance of the glass plate and the porosity of the
screen?
4.9.6 Inclusion of Substrate
PROBLEM 4.35. DES
Consider the efciency of a solar collector with and without a glass cover plate. In a
simple model for a solar collector, all the surfaces around the collector, participating
in the radiation exchange, are included by dening a single average environment
temperature T
a
. A solar collector and the various heat ows considered are shown
in Figure Pr. 4.35.
The irradiation from the sun has a magnitude (q
r.i
)
s
= 800 W/m
2
and the irra
diation from the atmosphere (diffuse irradiation) is given by (q
r.i
)
a
=
SB
T
4
a
, where
T
a
= 290 K is the effective atmospheric temperature. The cover plate (made of
low iron glass) has a total transmittance for the solar irradiation of
r.2
= 0.79. For
the infrared emission, the glass surface can be considered diffuse and gray with an
emissivity c
r.2
= 0.9. The absorber plate (coated with the matte black) is opaque,
diffuse, and gray and has absorptivity
r.1
= 0.95. The glass plate is at a temperature
MassoudKaviany 2011
Insulation
Environment, T
a
Box
Water
Vacuum
w
L
Absorber Plate
T
1
(q
r,i
)
s
(W/m
2
)
(q
r,i
)
a
(W/m
2
)
r, 2
(q
r,i
)
a+s
(W/m
2
)
Q
k,1a
(W)
Coolant Flow
Cover Glass Plate
T
2
.
(S
e,
)
2
(W)
Q
u L
(W)
Q
u 0
(W)
Q
ku 2a
(W)
Figure Pr. 4.35. A solar collector with a cover glass plate and the associated heat transfer
terms.
T
2
= 310 K and heat losses by surfaceconvection heat transfer occur at the rate of
200 W/m
2
. The interior of the solar collector is evacuated and the bottom heat losses
by conduction are at a rate of q
k.1a
= 40 W/m
2
. The collector surface is rectangular
with dimensions n = 1 m and L = 2 m. For the cases (i) with a glass cover, and (ii)
with no glass cover, determine the following. (Use T
1
= 340 K.)
(a) Determine the amount of heat transferred to the coolant Q
u
)
L0
(W).
(b) Determine the thermal efciency , dened as the heat transferred to the coolant
divided by the total irradiation.
PROBLEM 4.36. FUN
Cirrus clouds are the thin, high clouds (usually above 6 km) in the form of trails or
streaks composed of delicate white laments, or tenuous white patches and narrow
bands. Due to the atmospheric conditions at these high altitudes (low temperature
and high relative humidity), these clouds contain large amounts of ice crystals. An
important effect of these clouds in the atmosphere is the absorption and emission of
thermal radiation, which plays an important role in the upper troposphere water and
the heat budget. This may signicantly affect the earths climate and the atmospheric
circulation.
A cirrus uncinus cloud (a hooklike cloud appearing at an altitude between 5
to 15 km, and indicating a slowly approaching storm) has an average thickness L.
The extinction coefcient for the cloud is
ex
= 2.2 10
3
1/m. The intensity of the
irradiation at the top of the cloud is (q
r.i
)
s
= 830 W/m
2
, as shown in Figure Pr. 4.36.
Assume that the surface reectivity of the cloud is zero.
(a) Determine the amount of radiationabsorbedby the cloud
S
e.
,A, for L = 1,000 m.
(b) The lengthaveraged cloud temperature T)
L
is dened as
T)
L
=
1
L
_
L
0
Tdx.
MassoudKaviany 2011
(q
r,i
)
s
(W/m
2
)
e,x
Cloud
Earth's Surface
L
x
s
e,
Figure Pr. 4.36. Irradiation heating of a cloud layer.
Determine the time variation of the lengthaveraged cloud temperature dT)
L
,dt
for the cloud thickness L = 1,000 m. Neglect the heat transfer by surface convection
and the energy conversion due to phase change (i.e., the air is in equilibrium with
the cloud).
Assume that the volumeaveraged density and the specic heat of the cloud are
= 0.8 kg/m
3
and c
p
= 2,000 J/kgK.
PROBLEM 4.37. FUN
Consider a cirrus cloud with an average thickness L = 1 km and at an altitude of
r = 8 km from the earths surface. The top of the cloud is exposed to the deep
space and receives solar radiation at the rate (q
r.i
)
s
= 1,353 W/m
2
. The bottom of
the cloud is exposed to the earths surface. The extinction coefcient for the cloud is
ex.L
= 2.2 10
3
1/m. This is shown in Figure Pr. 4.37.
(q
r,i
)
s
(W/m
2
)
T
2
= 297 K
r, 2
= 0.9
T
1
= 238 K
r, 1
= 1
Deep Sky
T
3
= 3 K
r, 3
= 1
Cloud
Earth's Surface
(ii) Ice Particle on Bottom
(i) Ice Particle on Top
L = 1 km
x
r
s
e,
s
sg
+
Figure Pr. 4.37. Iceparticle heat transfer for a particle (i) at the top, and (ii) at the bottom of
a cloud layer.
(i) A spherical ice particle at the top of the cloud has a temperature of T
1
= 35
C,
and a diameter d = 100 jm. The ice surface is opaque, diffuse, and gray with an emis
sivity c
r.1
= 1. The particle is moving under the effect of the draft air currents and
has no preferred orientation relative to the solar irradiation. The deep sky behaves
MassoudKaviany 2011
as a blackbody with a temperature of T
3
= 3 K.
(ii) Another ice particle at the bottom of the cloud has the same temperature,
dimensions, and surface radiation properties, but it is exposed to the earths surface.
The earths surface is opaque, diffuse, and gray, has a surfaceaveraged temperature
T
2
= 297 K, and a total emissivity c
r.2
= 0.9. For each of these two particles perform
the following analyses.
(a) Track the heat transfer vector and show the energy conversions. Note that the
particles lose heat by surface convection, and that phase change (frosting or subli
mation) also occurs.
(b) Draw the thermal circuit diagram for the particles.
Assuming that the cloud is optically thick (i.e., there is a signicant attenuation of
the radiation across the cloud) the radiation heat transfer between two positions 1
and 2 in the cloud with emissive powers E
b.1
and E
b.2
is given by
Q
r.12
=
(E
b.1
E
b.2
)
3
ex.L
,4A
r
.
This resistance can be added in series with the surfacegrayness and viewfactor
resistances between two surfaces enclosing the cloud.
(c) Determine the net heat transfer for each particle.
(d) Neglecting the surfaceconvection heat transfer and assuming a steadystate
condition, determine the rate of growth by frosting (or by sublimation) of the ice
particles. For the heat of sublimation use Lh
sg
= 2.843 10
6
J/kg and for the density
of ice use
s
= 913 kg/m
3
.
(e) From the above results, are these particles expected to grow or decay in size?
Would the radiation cooling or heating rate differ for particles of a different size?
PROBLEM 4.38. DES
A atplate solar collector is modeled as a surface with equivalent total absorptance
r.1
and emittance c
r.1
, which represent the surface and wavelength average of the
absorptivity and emissivity of all the internal and external surfaces (it also accounts
for the transmissivity of the cover plate). The solar collector and the model are shown
in Figure Pr. 4.38.
r, 1
,
r, 1
Solar
Irradiation
(q
r,i
)
s
Water
Inlet,
Q
u 0
Water
Outlet,
Q
u L
Air Flow
(Wind)
(q
r,i
)
s
(W/m
2
)
Q
k,1
(W)
Q
ku,1
(W)
L = 2 m
w = 1 m
(q
r,i
)
a
(W/m
2
)
(i) (ii)
Q
u L
Q
u 0
Q
1
= 0
Water
Stream
(S
e,
)
1
.
(S
e,
)
1
.
Figure Pr. 4.38. (i) and (ii) A solar collector and the collector plate heat transfer.
MassoudKaviany 2011
The solar collector receives solar irradiation (q
r.i
)
s
= 800 W/m
2
and atmospheric
irradiation (q
r.i
)
a
=
SB
T
4
a
, where T
a
= 290 K is the effective atmospheric temper
ature (it accounts for the emission and scattering of the radiation by the atmo
sphere). This radiation is incident on the collector plate surface, which has area
Ln = 1 2(m
2
). The collector surface temperature is T
1
= 310 K and it loses
heat by surfaceconvection heat transfer at a rate q
ku
)
1
= 400 W/m
2
and also by
surface radiation emitted to the surroundings (
S
e.c
)
1
. The bottom of the collector
loses heat by conduction at a rate q
k.1
= 50 W/m
2
. The total absorptance of the
solar collector is
r.1
= 0.9 and the total emittance is c
r.1
= 0.3.
(a) Determine the amount of heat transferred to the uid Q
u
)
L0
= Q
u
)
L
Q
u
)
0
(W).
(b) Determine the thermal efciency , dened as the ratio of the heat transferred
to the uid to the total irradiation.
PROBLEM 4.39. DES
Table Pr. 4.39 gives a short list of materials used as selective surface coatings, i.e.,
coatings that have different absorption properties for shorter and longer wave
length ranges. Other materials are listed in Table C.19. For each of the two appli
cations below, choose from Table Pr. 4.39 the coating that results in the optimum
performance.
(a) In a solar collector, the absorber plate is exposed directly to the solar and
atmospheric irradiation (q
r.i
)
a
, as shown in Figure Pr. 4.39(i). The intensity of the
solar radiation is (q
r.i
)
s
= 800 W/m
2
. The net radiation emitted by the atmosphere is
given by (q
r.i
)
a
=
SB
T
4
a
, where T
a
= 290 K is the effective atmospheric temperature
(i.e., the sky temperature). The absorber plate has a surface Ln = 1 2 m
2
and it is at a temperature of T
1
= 80
r,s
,
IR
Radiative Cooler
T
1
= 250 K
r,s
,
IR
(q
r,i
)
s
(W/m
2
)
(q
r,i
)
a
(W/m
2
) (q
r,i
)
s
(W/m
2
)
L
w
Q
u L
(W)
Q
u 0
(W)
(q
r,i
)
sky
(W/m
2
)
Q
1
(W)
(S
r,
)
1
(S
r,
)
1
Figure Pr. 4.39. (i) A solar collector. (ii) A radiative cooler. Both use selective radiation
absorption.
Choose a material that wouldgive the maximumheat transfer Q
r.1
(W) anddetermine
this heat ow rate.
PROBLEM 4.40. DES.S
The high temperature of the automobile exhaust can be used to promote catalytic
reactions and conversions (this is called a closecoupled converter) of some gaseous
pollutants (such as unburned hydrocarbons) over catalytic metaloxide surfaces.
When the converter is placed close to, but downstream of an internal combustion
engine, the exhaust pipe leading to the converter is thermally insulated. One scenario
is the addition of a radiation shield around of the outside of the exhaust pipe. This is
shown in Figure Pr. 4.40. The gap between the pipe and the shield contains air, and
there heat transfer occurs by conduction and by surface radiation.
Air
T
1
= 800 K
r, 1
= 0.7
Ambient Surface:
T
3
= 300 K
r, 3
= 1
Exhaust Pipe
Shield
D
2,o
D
1
D
2,i
r, 2
= 0.1
q
u
(W/m
2
)
q
r
(W/m
2
) q
k
(W/m
2
)
q
ku
(W/m
2
) q
r
(W/m
2
)
L
Figure Pr. 4.40. Insulation of an automobile exhaust pipe.
MassoudKaviany 2011
The external surface of the shieldexchanges heat withits surroundings by surface
radiation and surface convection. The surfaceconvection heat loss of the external
surface of the shield is estimated as q
ku
= 4,000 W/m
2
. The surroundings behave as
a blackbody at T
3
= 300 K. The exhaust pipe has an outside diameter D
1
= 5 cm, a
surface temperature of T
1
= 800 K, and its surface is assumed to be diffuse, opaque,
and gray with a surface emissivity of c
r.1
= 0.7. The shield has an inside diameter
D
2.i
= 5.4 cm and an outside diameter D
2.o
= 5.5 cm and is made of chromium
coated carbon steel AISI 1042. Its surface is assumed to be opaque, diffuse, and
gray and has a surface emissivity c
r.2
= 0.1. For the thermal conductivity of air use
k
a
= 0.04 W/mK.
(a) Draw thermal circuit diagram
(b) Determine the heat transfer rate from the exhaust pipe to the ambient, for a
length L = 1 m for the pipe.
(c) Comment on the effect of the pipe wall conduction thermal resistance on the
total heat transfer from the exhaust pipe.
(d) Keeping all the other conditions the same, would the heat transfer from the
exhaust pipe increase or decrease with an increase of the inside diameter of the
shield (while keeping the thickness constant)? (Suggestion: Plot the variation of Q
1
for 5.2 cm D
2.i
7 cm.)
PROBLEM 4.41. DES
During continuous thermal processing of silicon wafers, to heat the wafers to a
desired temperature, the wafers are stacked vertically and moved through an evac
uated, cylindrical radiation oven at speed u
n
. This is shown in Figure Pr. 4.41(i).
Consider a unit cell formed by two adjacent wafers. This is shown in Figure Pr.
4.41(ii). The wafers have a diameter D
1
= 100 mm, the distance separating the
wafers is l = 30 mm, and the wafers are placed coaxially to the oven wall, which is
also cylindrical. The diameter of the ceramic oven is D
2
= 300 mm. The oven surface
is opaque, diffuse, and gray with surface emissivity c
r.2
= 0.9 and surface tempera
ture T
2
= 800 K. The wafers surface is also diffuse, opaque, and gray with emissivity
c
r.1
= 0.01.
(i) (ii)
D
1
Wafers
Wafer
T
1
= 500 K
r, 1
= 0.01
Oven Wall
T
2
= 800 K
r, 2
= 0.9
Axis of Symmetry
Unit Cell
Furnace
u
w
l
D
2
Q
r, 12
Figure Pr. 4.41. (i) Silicon wafers moving and heating in a radiation oven. (ii) The unit cell
formed by two adjacent wagers is also shown.
MassoudKaviany 2011
(a) Assuming that the wafers have a uniform temperature T
1
= 500 K, determine
the net heat transfer by surface radiation between the oven and a wafer Q
r.1
.
(b) During this process, the wafers enter the oven at an initial temperature lower
than 500 K. As they move through the oven, they are heated by surface radiation
from the oven walls until they reach a near steadystate temperature. Assuming that
the axial variation of temperature (along the thickness) is negligible, quantitatively
sketch the radial distribution of the wafer temperature at several elapsed times (as
the wafer moves through the oven).
(c) Apply a combined integraldifferential length energy equation to a wafer. Use the
integral length along the thickness and a differential length along the radius. Express
the surfaceradiation heat transfer in terms of a differential radiation resistance,
which depends on a differential view factor.
4.9.7 General
PROBLEM 4.58. FUN.S
Infrared transmission substrate and thinlmmaterials are often used as interference
lters. High absorption characteristics over the range of undesirable wavelengths is
made, while maintaining reasonably high transmission over the wavelength range
of interest. Consider three infrared transmission optics lters (Section 4.3.2), ger
manium, silicon and cadmium telluride. The spectral extinction index
is found
experimentally as a function of wavelength and for the three materials recorded in
Table Pr. 4.58 and shown graphically in Figure Pr. 4.58.
10
10
5
15 20 25 30 35 40
10
6
10
4
10
3
10
2
Near Infrared Range Far Infrared Range
, m
Si
Ge
CdTe
T = 300 K
Figure Pr. 4.58. Variation of infrared spectral extinction index
. jm
. jm
. jm
. jm
10.00 2.00 10
6
18.18 3.10 10
5
10.00 7.50 10
5
17.24 3.90 10
4
20.00 0
10.26 2.25 10
6
18.75 3.80 10
5
10.26 1.10 10
4
17.54 4.40 10
4
20.41 0
10.53 2.50 10
6
19.35 4.00 10
5
10.53 1.20 10
4
18.18 3.70 10
4
20.83 1.70 10
6
10.81 3.00 10
6
20.00 3.70 10
5
10.81 1.60 10
4
18.52 3.40 10
4
21.28 4.00 10
6
11.11 3.00 10
6
20.69 3.20 10
5
11.11 2.00 10
4
19.23 4.00 10
4
21.74 7.00 10
6
11.43 3.50 10
6
21.43 5.00 10
5
11.24 2.03 10
4
20.00 3.00 10
4
22.22 1.30 10
5
11.76 4.80 10
6
22.22 7.00 10
5
11.36 2.50 10
4
21.05 2.40 10
4
22.73 1.90 10
5
12.00 5.40 10
6
22.86 8.60 10
5
11.49 2.01 10
4
22.22 1.40 10
4
23.26 3.00 10
5
12.24 5.20 10
6
23.53 1.40 10
4
11.63 1.90 10
4
23.53 1.10 10
4
23.81 3.70 10
5
12.50 5.20 10
6
24.24 1.00 10
4
11.76 1.70 10
4
25.00 1.02 10
4
24.39 4.30 10
5
12.77 5.10 10
6
25.00 8.50 10
5
11.90 1.50 10
4
26.67 1.20 10
4
25.00 6.00 10
5
13.04 5.80 10
6
25.81 1.90 10
4
12.05 1.90 10
4
28.57 1.10 10
4
25.64 7.60 10
5
13.33 7.00 10
6
26.67 2.20 10
4
12.20 2.05 10
4
30.77 1.40 10
4
26.32 9.00 10
5
13.64 6.50 10
6
27.59 3.80 10
4
12.35 2.00 10
4
33.33 1.50 10
4
27.03 1.00 10
4
13.95 5.50 10
6
28.58 7.50 10
4
12.50 1.80 10
4
27.78 1.30 10
4
14.29 5.00 10
6
29.63 7.00 10
4
12.90 2.20 10
4
28.57 2.40 10
4
14.49 5.50 10
6
30.77 3.30 10
4
13.33 2.70 10
4
29.41 5.00 10
4
14.71 1.10 10
5
32.00 2.80 10
4
13.51 3.20 10
4
30.30 9.00 10
4
14.93 7.20 10
6
33.33 3.00 10
4
13.79 2.60 10
4
31.25 1.50 10
3
15.15 8.00 10
6
34.78 3.30 10
4
14.29 1.00 10
4
32.26 2.00 10
3
15.38 1.60 10
5
36.36 2.80 10
4
14.49 1.30 10
4
33.33 8.00 10
3
15.79 1.30 10
5
38.10 2.00 10
4
14.81 1.40 10
4
34.48 7.00 10
3
16.22 9.50 10
6
40.00 2.40 10
4
15.38 1.60 10
4
35.71 6.50 10
3
16.67 1.00 10
5
16.00 8.00 10
4
37.04 6.20 10
3
17.14 1.50 10
5
16.13 1.30 10
3
38.46 9.00 10
3
17.65 3.40 10
5
16.67 7.50 10
4
40.00 5.70 10
3
(a) Plot transmissivity (or transmittance)
r.
= e
4
d,
as a function of wavelength,
for each material using a substrate thickness d = 1 mm.
(b) Note that these three materials showdistinctly different transmittance at different
wavelengths. How does the atomic structure of these three materials affect their
absorption characteristics, and can this be used to predict the behavior of other,
similar materials?
PROBLEM 4.59. FUN.S
A carbon dioxide laser generates light by placing a large electric potential across a
glass tube lled with a gas mixture containing carbon dioxide. This electric potential
energizes the CO
2
molecules driving them into a higher energy state which causes
them to emit light at = 10.4 jm (near infrared range). This light will then travel
down the tube to one of two reecting lenses (mirrors); the front and rear reecting
lenses. The rear lens is ideally 100% reective which drives the infrared emissions
back up the tube towards the front lens. This front lens material is chosen to be
MassoudKaviany 2011
Plasma Discharge Tube
Rear
Lens
(Mirror)
Front Lens (Mirror)
(Ge, or Si)
Gas
Inlet
Gas
Outlet
Coolant Inlet
Coolant Outlet
Endplate
Mirror
Cells
(+) High Voltage
() High Voltage
Insulator
Focused
Laser Beam
CO
2
Figure Pr. 4.59. CO
2
laser, showing the front and rear lenses.
highly transparent to the wavelength of light being emitted (here at = 10.4 jm).
This lens is also coated with a thin lm of a highly reective material which reects
most of the light back to the tube again to be focused, only the most focused photons
(those traveling perpendicular to the lens) are allowed through. These photons
not reected by the front lens escape as a focused laser beam. This is shown in
Figure Pr. 4.59
Germanium and silicon are two examples of materials that could potentially be
used as the front lens in a CO
2
laser assembly. To be considered, a lens must be
reasonably transparent to allow the laser to propagate through effectively and to
not heat up signicantly. Consider a lowpowered CO
2
laser tted with lenses made
of germanium and silicon. The laser radiation is emitted at = 10.4 jm and peak
irradiation ux (q
r.i
)
l
= 10
8
W/m
2
through a lens with diameter of D
1
= 2 mm and
thickness of L
1
= 1 mm. Assume that the irradiation ux is uniform over the focal
surface.
(a) Draw the thermal circuit diagram and derive an expression for the volumetric
energy conversion for absorption of irradiation, (
S
e.
)
1
, by integrating (2.43) along
the thickness; and using the expression for
ex.
= 4
r
= 0.8, c
r
= 0.9, T
1
(t = 0) = T
= 298.15 K.
Ge: c
p
= 310 J/kgK, = 5,323 kg/m
3
, k = 60.2 W/mK,
( = 10.4 jm) =
2.5 10
6
.
Si: c
p
= 678 J/kgK, = 2,330 kg/m
3
, k = 148 W/mK,
_
_
Fo
_
1,2
Fo
Fo
2
8
Fo
3
32
15Fo
4
512
_
Fo =
4)t
D
2
 0.6
2k)DR
ck
=
32
3
2
2
3,2
Fo
1,2
_
1
1
3(4Fo)
1
6(4Fo)
2
1
12(4Fo)
3
_
Fo 0.6.
Acc. V Spot WD 20m
20.0 kV 4.0 13.6 (Y86Yb8Er6)3O2 938nm 800mW
q
r,i
Prescribed
Irradiation
Q
r,1
Gray Body Surface
Radiation
Q
ck
Transient Spreading
Conduction
Laser Beam
Laser
Drilled
Hole
Formed
Microtube
Wall
Doped Yttria
Compacted
Powder, d
p
,
T
S
e,
S
sl
Absorption of Radiation,
Phase Change
T
1
(t)
Surface Temperature
L(t)
Doped Yttria
Compact
d
p
, ,
r,1
,
k
s
, k
f
r,1
(q
r,i
)
Reflected Irradiation
T
A
1
D
(ii) Heat Transfer Model
(i) SEM Micrograph
D
Figure Pr. 4.60. (i) A scanning electron microscope (SEM) image of a hole bored in a doped
yttria compact (ii) Heat transfer model of an yttria nonporous compact being irradiated,
resulting in phase change of the compact.
MassoudKaviany 2011
S
sl
=
M
sl
Lh
sl
, and
S
e.
is given in Table 2.1. When calculating radiation losses,
assume the compact is opaque with reectivity
r.1
.
(a) Draw the thermal circuit diagram for the compact during melting.
(b) Calculate the effective thermal diffusivity (3.130) of the compact ), using (3.11),
(3.12), and (3.28) to determine ), c
p
), and k).
(c) Using MATLAB or other numerical simulation, plot the temperature T
1
of the
irradiated surface as a function of time for T
1
 T
sl
using (4.66) for the energy
balance where Q
r.1
can be accounted for as part of Q [
A
(note that for t = 0 s,
R
ck
= 0 W/mK).
(d) Using MATLAB or other numerical simulation, plot the bored length L as a
function of time for 0  t  10 ms (hint: in your MATLAB code, Q
r.1
should be
related to L and the rate of melting,
M
sl
(kg/s), should be related to dL,dt).
T
sl
= 2,703 K, T
= 298 K, D= 40 jm, q
r.i
= 3.2 10
8
W/m
2
, c
r.1
= 0.5, c =
0.35, k
s
= 15 W/mk, k
f
= 0.026 W/mk,
s
= 5,046 kg/m
3
,
f
= 1.177 kg/m
3
, c
p.s
=
502.4 J/kgK, c
p.f
= 1,005 J/kgK, Lh
sl
= 4.65 10
5
J/kg.
PROBLEM 4.61. FAM
A spherical water droplet of diameter D
d
is at a constant temperature T
lg
and
undergoes evaporationwiththe heat suppliedby radiationonly fromits surroundings
at temperature T
= 500 K, c
r.d
= 0.9 and c
r.s
=
0.05.
Use Table C.27 and T
lg
for the heat of evaporation of water.
Radiation Shield
r,s
Surroundings
T
1
D
d
Droplet
T
lg
,
r,d
Surroundings
T
1
S
lg
D
d
Droplet
T
lg
,
r,d
S
lg
D
s
(i) No Radiation Shield (ii) With Radiation Shield
Figure Pr. 4.61. (i) Evaporation with heat supplied by radiation only, without radiation shield;
(ii) Same, but with a radiation shield.
MassoudKaviany 2011
5
Convection: Unbounded Fluid Streams
There is a close correspondence betweenturbulence andstatistical mechan
ics. Statistical mechanics is concerned with the properties of a large number
of particles whose motion conforms to certain motion and collision laws,
whereas turbulence is concerned with continuum whose motion conforms
to the NavierStokes equations. In both cases the interest lies in asymptotic
statistical states which are independent of a wide class of initial conditions.
G.K. Batchelor
We begin discussion of convection heat transfer by considering an unbounded uid
stream undergoing a temperature change along the stream ow direction, where this
change is caused by an energy conversion. Convection heat transfer is the transfer
of heat from one point to another by a net (i.e., macroscopic) uid motion (i.e., by
currents of gas or liquid particles) given by the uid velocity vector u
f
. The uid
motion can be void of random uctuations, which is called a laminar ow, or it can
have these random uctuations, which is called a turbulent ow. Convection heat
transfer examines the magnitude, direction, and spatial and temporal variations of
the convection heat ux vector q
u
in a uid stream. In this chapter, we examine q
u
within a uid stream without directly considering the heat transfer between the uid
and a bounding solid surface (i.e., here we consider only intramedium conduction
and convection of unbounded uid streams). In addition to the occurrence of the
intramedium (i.e., bulk) convection, in many practical applications involving cool
ing and heating of uids, the review of the intramedium convection allows for an
introductory treatment of laminar, steadystate, uniform, onedirectional uid ow
without addressing the effect of the uid viscosity emanating from the bounding
surfaces, which would require inclusion of the momentum conservation equation
(1.27). The interfacial heat transfer between a uid and a solid surface, i.e., the
surfaceconvection heat ux vector q
ku
, is inuenced by the uid viscosity, since the
viscosity results in the nonuniformity of the uid velocity near the solid surface.
Also due to the surfaceconvection heat transfer, the uid temperature changes, and
this change for semibounded uid stream is conned to a region adjacent to the
bounding surface (while the fareld temperature is unchanged); this is treated in
MassoudKaviany 2011
Convection Heat Transfer
Unbounded
Fluid Streams
Fluid Temperature
Variations Along Stream
(Chapter 5)
SemiBounded
Fluid Streams
Bounded
Fluid Streams
Fluid Temperature
Variations Adjacent to
Bounding Surface
(Chapter 6)
Fluid Temperature
Variations Along
and Across Stream
(Chapter 7)
Chart 5.1. Treatment of convection heat transfer, based on presence or lack of surface con
vection and extent of temperature change in the uid temperature.
Chapter 6. When the uid stream is bounded, the surfaceconvection heat transfer
alters the uid temperature across and along the stream, and this is discussed in
Chapter 7. Chart 5.1 shows the division of the treatment of convection heat transfer
among the three chapters. The reason for using the designations, unbounded, semi
bounded, and bounded uid streams, is to maintain the focus on the uid ow and
its thermal behavior (as compared to any bounding solid).
The uid motion is categorized as the viscous ow regime [for small Knudsen
number, (1.20), Kn
L
 1], or the molecular ow regime (for large Knudsen number
Kn
L
> 1). The viscous owregime is in turn divided into the laminar owregime and
the turbulent ow regime. In laminar ow, there are no random velocity uctuations.
This is shown in Figure 5.1(a)(i), where the uid particle trajectory is smooth. In
homogeneous (does not change in space) turbulent ows, there are random velocity
uctuations, as shown in Figure 5.1(a)(ii). Turbulent ows are represented by time
averaged uid velocity u
f
and by the rms of the uctuation velocity components,
e.g., (u
/
f
2
)
1,2
. In the absence of a bounding surface, the turbulence may be generated
downstream of a gridnet (a mesh) placed in the stream.
The grid generated, highReynolds number, homogeneous turbulence has a particular statistical
uctuation property as shown in its kinetic energy distribution function E
f .t
. The wavelength of
the uctuations (or turbulence) is limited on the small size by the Kolmogoroff wavelength
K
(Batchelor, G. K., 1982, The Theory of Homogeneous Turbulent Flow, Cambridge, Cambridge).
Then the energy spectrum function is
E
f .t
= c
K
c
2,3
t
(
f .t
2
)
5,3
K
f .t
.
where c
K
is a universal constant (Kolmogoroff constant, equal to1.44) andc
t
is the average dissipation
rate of the kinetic energy of turbulent ow. As the wavelength decreases its contribution to the total
energy decreases.
MassoudKaviany 2011
S
12
(c) Convection and Internodal Energy Conversion in Unbounded Fluid Streams
(b) ConductionConvection Internodal Resistance in Unbounded Fluid Streams
Q
u,1
Q
u,2
(Q
k,u
)
12
T
f,1
T
f,2
T
f,1 T
f,2
(R
k,u
)
12
OneDimensional
Fluid Flow
u
f u
f
OneDimensional
Fluid Flow
u
f u
f
u
f
(i) Laminar Fluid Stream
(a) Laminar and Tubulent Unbounded Fluid Streams
(ii) Turbulent (Homogeneous) Fluid Stream
t
u
f
u
f
u
f
= u
f
+ u
f
'
'
u
f
2
1/2
'
y,r
Mean Velocity
x,u
z,w
rms of Turbulent
Fluctuation Velocity
Turbulent Velocity Fluctuation
u
f
'
2
u
f
2
Turbulence Intensity, Tu
u
f
u
f
Random
Fluid Velocity
Fluctuations
No Random
Fluid Velocity
Fluctuations
Figure 5.1. (a) Laminar and turbulent uid streams. (b) The conductionconvection resistance.
(c) The internodal energy conversion in uid stream.
twophase system. This allows us to address some of the advantages of porous solids,
under the condition of local thermal equilibrium (i.e., when the surfaceconvection
heat transfer becomes innite, i.e., q
ku
).
Figure 5.1(c) shows the internodal energy conversion addressed in this chap
ter. Far from the energy conversion location, both upstream and downstream, the
temperature is uniform and only convection exits (Q
u.1
and Q
u.2
). The uid under
study may be in liquid or gas (or saturated) state. Upon heat transfer, the uid may
undergo reaction (including chemical reaction and dissociation)
S
12
=
S
r.c
, or phase
change
S
12
=
S
ij
.
A chemical classication of uids is given in Chart 5.2. The division is into
compounds (organic and inorganic) and elements. Inert heat transfer uids are used
for single or twophase ow heat transfer in closed cycles (Tables C.3 to C.4 give a
list of various uids). Organic, reactive uids are used for the conversion of their
MassoudKaviany 2011
Chemical Classification of Fluids
Normal Paraffins
(e.g., methane)
Acetylenes
(e.g., acetylene)
Alcohols
(e.g., methanol)
Aldehydes
(e.g., formaldehyde)
Halogenated Hydrocarbons
[e.g., trichlorofluoromethane (R11)]
Elements
(e.g., oxygen, ni
trogen, mercury)
Others
Olefins
(e.g., propylene)
Aromatics
(e.g., toluene)
Phenols
(e.g., phenol)
Ketones
(e.g., acetone)
Inorganic Substances
(e.g., air, water)
Heat Transfer
(e.g., Dowtherm A)
Organic Substances
Chart 5.2. A chemical classication of uids.
chemicalbond energy into thermal energy and formation of hot gaseous streams
(Table C.21).
Parafns are singlebonded, openchain saturated hydrocarbons (i.e., no more
hydrogen can be added to the molecules). They are represented by C
n
H
2n2
. Olens
are openchain hydrocarbons containing a double bond (i.e., unsaturated). They are
represented by C
n
H
2n
. Acetylenes are openchain unsaturated hydrocarbons con
taining one carboncarbon triple bond. They are represented by C
n
H
2n2
. Aromatics
hydrocarbons have benzene (C
6
H
6
) ring structures as their building blocks. They are
represented by C
n
H
2n6
. Alcohols are organic compounds with one hydroxyl (OH)
group substituting for one hydrogen atom. They are represented by C
n
H
2n1
OH.
Naphthenes (or cycloparafns) are singlebond, unsaturated hydrocarbons. They are
also represented by C
n
H
2n
. Table C.21 list some commonly used parafns, olens,
naphthenes, aromatics, and some alcohols and other fuels.
Evaporating and condensing streams (including droplets) are also used for gen
eration of cold and hot gases (Table C.26). Depending on the temperature and the
thermal energy needed, a selection is made among these uids.
In this chapter we will examine steadystate, uniform, onedimensional ow and
convectionconduction heat transfer with surface or volumetric energy conversion
s. In particular, we examine how the source term s is used to generate hot or cold
streams. Examples are gaseous combustion (chemicalbond energy convection) and
gaseous induction or arc heating (Joule heating). In addition to expansion cooling,
one method of creating a cold gaseous stream is by evaporation cooling. Cooling
towers use the evaporation of droplets, in a mixture of gasdroplets, to cool a hot
exhaust gas or to create cold gases for use in climate control. We discuss aspects of
the surfaceevaporation cooling and will expand on it in Chapter 6.
This chapter covers the primary aspects of simultaneous heat and mass transfer
through examples of evaporation, combustion, and ionization. In Chapter 6, the
MassoudKaviany 2011
surfaceconvection mass transfer will be discussed in connection with evaporation
cooling (Section 6.9).
5.1 OneDimensional ConductionConvection
Energy Equation
The steadystate, uniform, laminar, onedimensional uid ow along the x axis is
designated with velocity u
f
. Therefore, we simply have a uniform uid ow and a
uniform pressure distribution.
The differentialvolume energy equation is the simplied form of (2.1), i.e.,
(q
k
q
u
q
r
) =
_
LA
[(q
k
q
u
q
u
) s
n
]dA
LV
= s
differentialvolume
energy equation for
LV 0.
(5.1)
Then for the onedimensional heat ow in the x direction, this reduces to
d
dx
(s
x
q
k
)
d
dx
(s
x
q
u
)
d
dx
(s
x
q
r
) = s
spatial rate of spatial rate of spatial rate of volumetric
change of q
k
change of q
u
change of q
r
energy
in x direction in x direction in x direction conversion.
(5.2)
For the case of no radiation heat transfer, we have
d
dx
(k
f
dT
f
dx
f
c
p.f
u
f
T
f
) = s
differentialvolume
energy equation for
onedimensional simultaneous
conduction and convection.
spatial rate of spatial rate of volumetric
change of q
k
change of q
u
energy
in x direction in x direction conversion
(5.3)
When we integrate this between points 1 and 2 along the x axis, we have
_
k
f
dT
f
dx
f
c
p.f
u
f
T
f
_
2
1
=
_
2
1
s dx integrated energy equation. (5.4)
This shows that when s = 0, the sum of the conduction and convection heat
transfer terms remains constant, while each term can change causing the other to
change.
When s ,= 0, both the conduction and convection heat transfer terms change
along x to accommodate this volumetric source (or sink) of heat.
We rst introduce the conductionconvection resistance R
k.u
. Then we consider
three examples. One is the liquid ow (from a reservoir) and heat transfer through
a permeable wall (i.e., seepage) with no volumetric energy conversion, but with the
complete surface evaporation of the seepage liquid. This demonstrates evaporative
cooling (because the evaporating surface will be at a lower temperature than the
MassoudKaviany 2011
Convection Heat Transfer
in Unbounded Fluid Streams
Axial Convection
versus Axial Conduc
tion: Peclet Number
Convection
Conduction
Resistance
Evaporation
Cooling of Gaseous
Streams
Combustion
Heating of Gaseous
Streams
Joule Heating
of Gaseous
Streams
Allowable
Seepage
Speed
Conduction
Convection
Region
Flame
Speed
Conduction
Reaction
Region
HeatTransfer
Controlled
Flame Speed
One
Dimensional
Analysis
Generation
of Thermal
Plasmas
Classification
of Thermal
Plasmas
GasStream
Volumetric Radiation
Laminar
Flow
Porous
Medium
Turbulent
Flow
Chart 5.3. Aspects of intramedium convection discussed in Chapter 5.
liquidreservoir temperature). Inthe other twoexamples, we consider the combustion
of a premixed gaseous fuel and oxidant, and the Joule heating of charged gases by
induction heating. The gaseous mass (i.e., species) transfer is also briey discussed.
The radiation loss from hot gaseous streams is also discussed. Chart 5.3 lists the
topics discussed in this chapter.
5.2 Parallel ConductionConvection Resistance R
k.u
(
C/W) and
ConductionConvection Number N
u
Pe
L
The steadystate, laminar, uniform, onedimensional uid ow considered is shown
in Figure 5.2. We begin with the differential volume LV 0 shown in the gure and
then integrate the differentialvolume energy equation between locations x = 0 and
x = L, as shown in the gure. Here we assume that there is no energy conversion,
s = 0. We also assume that k
f
and (c
p
)
f
are constant. Then (5.3) becomes
k
f
d
2
T
f
dx
2
(c
p
)
f
u
f
dT
f
dx
= 0
differentialvolume
energy equation for
permeable wall.
spatial rate of spatial rate of
change of q
k
change of q
u
in x direction in x direction
(5.5)
Here we note that both the conduction and convection heat ow are along the
x direction. This is called here the axialaxial conductionconvection heat ow. In
Chapter 6 we will consider conduction and convection heat ows perpendicular to
each other. We now make the variables T
f
and x dimensionless (i.e., normalized)
using T
f .1
T
f .2
and L as the scales. Asterisks indicate dimensionless quantities.
Then the dimensionless form of (5.5) becomes
d
2
T
f
dx
2
u
f
L
k
f
,(c
p
)
f
dT
f
dx
= 0
(5.6)
MassoudKaviany 2011
T
f,2
T
f,1
u
f
V 0
u
f
k
f
,
f
, c
p, f
L
x, u
f
x
1
x
2
q
k
q
k
OneDimensional
Unbounded Fluid Stream
Uniform
Fluid Flow
ConductionConvection Heat Transfer
in a OneDimensional Fluid Flow
Uniform
u
f
Uniform
u
f
Figure 5.2. A onedimensional, laminar uid ow with parallel conduction and convection
heat transfer.
or
d
2
T
f
dx
2
Pe
L
dT
f
dx
= 0.
spatial rate of spatial rate of
change of q
k
change of q
u
in x direction in x direction
(5.7)
where
Pe
L
=
u
f
L
f
.
f
=
k
f
(c
p
)
f
. T
f
=
T
f
T
f .2
T
f .1
T
f .2
. x
=
x
L
dimensionless
parameters
and variables.
(5.8)
The P eclet number Pe
L
is the ratio of the axialconduction resistance L,k
f
to
the axialconvection resistance 1,(c
p
)
f
u
f
, i.e.,
Pe
L
R
k.f
R
u.f
N
u
u
f
L
f
=
(c
p
)
f
u
f
k
f
,L
P eclet number as ratio of
conduction to convection resistances
Pe
L
 0.1 conduction regime, Pe
L
> 10 convection regime, (5.9)
where conductance is the inverse of the resistance.
For Pe
L
> 10, convection conductance dominates over conduction. For Pe
L

0.1, conduction is the dominant mechanism of heat transfer.
Note that by the normalization, in (5.7) we have 0 T
f
1 and 0 x
1.
The solution to the secondorder, ordinary differential equation (5.7) requires
two thermal boundingsurface conditions. We use prescribed temperatures, i.e.,
MassoudKaviany 2011
0.0
0.2
0.4
0.6
0.8
1.0
T
f
*
=
T
f
T
f
,
2
T
f
,
1
T
f
,
2
0.0 0.2 0.4 0.6 0.8 1.0
Pe
L
= 0
0.5
1.0
Conduction Only
2
5
10
30
x* =
x
L
Figure 5.3. Distribution of the dimensionless temperature with respect to the dimensionless
location along the wall, for various P eclet numbers.
(2.63). These temperatures are already discussed (the two prescribed temperatures
T
f .1
, and T
f .2
), i.e.,
T
f
(x = 0) = T
f .1
or T
f
(x
= 0) = 1 (5.10)
T
f
(x = L) = T
f .2
or T
f
(x
= 1) = 0. (5.11)
By integrating the energy equation(5.7) twice andusing the boundary conditions
(5.10) and (5.11), we have the solution for T
f
= T
f
(x
) as
T
f
(x
) = 1
e
Pe
L
x
1
e
Pe
L
1
or
T
f
(x) T
f .2
T
f .1
T
f .2
= 1
e
u
f
x,
f
1
e
u
f
L,
f
1
temperature
distribution.
(5.12)
Note that in (5.12) both boundary conditions are satised (i.e., at x
= 0 we have
T
f
= 1 and at x
= 1 we have T
f
= 0). The temperature distribution, for var
ious values of Pe
L
, is shown graphically in Figure 5.3. For Pe
L
= 0, only con
duction occurs and a linear temperature distribution is obtained [to obtain the
solution for Pe
L
0 from (5.12), the Taylor series expansion of the exponential
terms is required]. For large Pe
L
, the temperature distribution will have a nearly
uniform region near x
= 1. For yet
larger Pe
L
, the region of nonuniform temperature is conned to a small region
adjacent to x
f
e
Pe
L
for (Q
k.u
)
12
= Q
x=L
. (5.13)
This can be written as
(R
k.u
)
12
=
e
Pe
L
1
A
k.u
k
f
L
Pe
L
e
Pe
L
=
L
A
k.u
k
f
e
Pe
L
1
Pe
L
e
Pe
L
conduction
convection
resistance.
(5.14)
This resistance is depicted in Figure 5.4.
Note that when the ow direction is reversed, the P eclet number becomes nega
tive, but the resistance (R
k.u
)
12
remains positive. Then the temperature distribution
will be the mirror image of the positive Pe
L
results shown in Figure 5.3. Then the
region of temperature drop will be conned to a region adjacent to x
= 0.
In the case of Pe
L
= 0, we have (by taking the derivative of the numerator and
the denominator, with respect to Pe
L
and then setting Pe
L
= 0, left as an end of
chapter problem)
(R
k.u
)
12
=
L
A
k.u
k
= R
k.12
for Pe
L
= 0
(approximated as Pe
L
 0.1)
conduction resistance,
(5.15)
which is the conductiononly resistance for a plane wall, given in Table 3.2.
Note that as u
f
(or Pe
L
) becomes large, (R
k.u
)
12
givenby (5.14) becomes equal to
1,A
k.u
(c
p
u)
f
, i.e., convection dominates the axial heat ow. Also note that we have
assumed a laminar ow. In Chapters 6 and 7 we use this high Pe
L
approximation
of axial heat ow (i.e., highPe
L
limit or asymptotics) to signicantly simplify the
analysis.
MassoudKaviany 2011
Table 5.1. Conductionconvection resistance from bounded uid streams
Heat Flow Rate Potential Resistance
(Q
k.u
)
12
= (q
k.u
)
12
A
k.u
T
f .1
T
f .2
,
C (R
k.u
)
12
=
L
A
k.u
k
f
e
Pe
L
1
Pe
L
e
Pe
L
,
C/W
= (T
f .1
T
f .2
),(R
k.u
)
12
, W Pe
L
=
u
f
L
f
By examining (5.14), we further normalize the conductionconvection resistance
using the conduction resistance, i.e.,
(R
k.u
)
12
=
(R
k.u
)
12
R
k.12
=
e
Pe
L
1
Pe
L
e
Pe
L
dimensionless conduction
convection resistance.
(5.16)
Again, this shows that as Pe
L
becomes large, we will have (R
k.u
)
12
= 1,[A
ku
(c
p
)
f
u
f
] = 1,(
Mc
p
)
f
, where
M
f
= A
k.u
f
u
f
is the stream mass ow rate. Table 5.1
shows the electrical resistance analogy of the conductionconvection heat ow.
EXAMPLE 5.1. FUN
In a radiant porous burner, shown in Figure Ex. 5.1(a), onedimensional, steady
premixed natural gasair ows through a porous ceramic foam for reaction on
the surface of the foam. Upstream of the surface ame, heat ows toward the
premixed gas by conduction and convection. Assume that the solid and uid are
in local thermal equilibrium such that a single local temperature T
f
(x) describes
the temperature for this twophase system.
(a) Draw the thermal circuit diagram for the conductionconvection heat trans
fer through the gas and the gasporous medium.
(b) For the conditions given below, determine the resistances (R
k.u
)
1.2
and
(R
k.u
)
2.3
.
(c) Determine T
f .2
and plot a qualitative variation of T
f
with respect to axial
location.
Determine the gas properties using air at T = 1,000 K. u
f
= 0.90 cm/s, k) =
0.8 W/mK, L
1
= L
2
= 2 cm, n = l = 10 cm, T
f .1
= 300 K, and T
f .3
= 1,700 K.
SOLUTION
(a) Figure Ex. 5.1(b) shows the thermal circuit diagram.
(b) The conductionconvection resistance is given by (5.14), i.e.,
(R
k.u
)
12
=
e
Pe
L.1
1
A
k.u
k
f
L
1
Pe
L.1
e
Pe
L.1
(R
k.u
)
23
=
e
Pe
L.2
1
A
k.u
k
f
)
L
2
Pe
L.2
e
Pe
L.2
.
MassoudKaviany 2011
(a) Physical Model
Premixed Natural
GasAir Flow
L
2
w
l
L
1
x
u
f
(c
p
)
f
, k
f
T
f,1
T
f,2
T
f,3
k
OpenCell
Ceramic Foam
Flame
A
k,u
Q
r
Q
u
S
r,c
(b) Thermal Circuit Model
(c) Temperature Distributions
(R
k,u
)
12
(Q
k,u
)
13
= (Q
k,u
)
12
= (Q
k,u
)
23
(R
k,u
)
23
S
r,c
x
T
f
, K
T
f,3
= 1,700
T
f,1
T
f,3
T
f,2
T
f,2
= 1,000
0 L
1
L
1
+ L
2
T
f,1
= 300
Ceramic Foam
Pe
L,1
= 107.1 Pe
L,2
= 9
Gas
Q
r
Q
u
Figure Ex. 5.1. (a) Radiant porous burner. (b) Thermal circuit model. (c) Temperature
distribution along the gas and ceramic foam.
where
Pe
L.1
=
u
f
L
1
f
. Pe
L.2
=
u
f
L
2
)
.
f
=
k
f
(c
p
)
f
. ) =
k)
(c
p
)
f
A
k.u
= ln
Using Table C.22 for air at T = 1,000 K, we have
k
f
= 0.0672 W/mK
f
= 0.354 kg/m
3
c
p.f
= 1,130 J/kgK
f
=
k
f
(c
p
)
f
=
0.0672(W/mK)
0.354(kg/m
2
) 1,130(J/kgK)
= 1.680 10
4
m
2
/s
) =
k)
(c
p
)
f
=
0.8(W/mK)
0.354(kg/m
2
) 1,130(J/kgK)
= 2.000 10
3
m
2
/s.
MassoudKaviany 2011
Then
Pe
L.1
=
0.9(m/s) 0.02(m)
1.680 10
4
(m
2
/s)
= 1.071 10
2
Pe
L.2
=
0.9(m/s) 0.02(m)
2.000 10
3
(m
2
/s)
= 9.000.
The resistance becomes
(R
k.u
)
12
=
e
1.07110
2
1
0.1(m) 0.1(m) 0.0672(W/mK)
0.02 (m)
1.071 10
2
e
1.07110
2
= 0.2779
C/W
(R
k.u
)
23
=
e
9.000
1
0.1(m) 0.1(m) 0.8(W/mK)
0.02 (m)
9.000 e
9.000
= 0.2777
C/W.
(c) From Figure Pr. 5.6(b), we have
(Q
k.u
)
13
= (Q
k.u
)
12
T
f .1
T
f .3
(R
k.u
)
12
(R
k.u
)
23
=
T
f .1
T
f .2
(R
k.u
)
12
or
T
f .2
= T
f .1
(T
f .1
T
f .3
)
(R
k.u
)
12
(R
k.u
)
12
(R
k.u
)
23
= 300(K) (300 1,700)(K)
0.2779(
C/W)
(0.2779 0.2777)(
C/W)
= 300(K) 700.3(K) = 1,000 K.
Since the tworesistances are equal, the temperature T
f .2
is the arithmetic average
of T
f .1
and T
f .3
.
The quantitative temperature distribution is given in Figure Ex.5.1(c). From
Figure 5.3, we note that for Pe
L.1
= 9.000 the temperature distribution is uniform
over a smaller portion of respective length than it is for Pe
L.2
= 107.1.
COMMENT
Note that although Pe
L
is smaller in the foam region, the resistances are equal.
This is because the numerator of (5.14) becomes e
Pe
L
for large Pe
L
and then
e
Pe
L
cancels from both the numerator and the denominator. Then the effect of k
disappears for this highPe
L
limit. In simultaneous uid ow and heat transfer
in porous media, the nonuniformity of uid velocity leads to an addition to the
effective conductivity k). This addition is called the thermal dispersion and is
similar to the turbulent conductivity (which will be discussed in Section 6.3) and
is dependent on the uid velocity u
f
[6].
MassoudKaviany 2011
Evaporation Cooling
Volumetric Cooling of Gas Streams
Trickle Bed Others Droplet Spray
Surface Cooling of Liquid Re
servoirs and Substrates
Wall Seepage Others Liquid Films
on Substrates
Porous Bed
Gas
Stream
Droplet
Nozzle
Gas
Liquid
Vapor
(Gas)
Vapor (Gas)
Liquid Reservoir
Porous
Wall
u
f,1
T
f,1
Q
u,1
Q
u,1
Q
u,2
Q
loss
Q
loss
Q
u,2
Q
k
Q
u
Q
loss
Q
u
M
l
S
lg
M
l
M
l
T
l
<T
f,1
u
f,2
T
f,2
<T
f,1
T
f,2
<T
f,1
u
f,1
T
f,1
S
lg
Makeup
Liquid
Substrate
Seepage
S
lg
u
g
S
lg
u
f,2
.
.
M
l
.
Q
k,u
Q
loss
Chart 5.4. Characterization of evaporation cooling of gas streams, liquid reservoirs, and sub
strates.
5.3 Evaporation Cooling of Gaseous Streams
The heat absorbed in the liquid evaporation can be supplied by gaseous streams
or by liquid bodies and their substrates thus causing a reduction in their tempera
tures. Chart 5.4 shows these evaporation cooling methods. The heat supplied for the
evaporation can be by conduction, convection, surfaceconvection, radiation, or a
combination of these. The gaseous stream can ow through a bed where the liquid is
supplied by a trickling stream and the heat absorbed from the gas stream (for liquid
evaporation) cools the stream as it exits the bed. The gaseous stream can also be
cooled by spraying into it. The liquid droplets evaporate, absorbing heat from the
gaseous stream.
The liquid reservoir can be cooled by allowing seepage through its container wall
and evaporation of this seepage liquid in the container surface. In other applications,
evaporation of liquid ows draws heat from their substrates.
In this section, we briey discuss volumetric cooling of gas streams and then
examine in detail the seepage cooling of liquids.
MassoudKaviany 2011
5.3.1 Cooling of Gaseous Streams
Consider a gaseous stream of velocity u
f .1
, density
f .1
, specic heat capacity (c
p.f
)
1
,
and temperature T
f .1
, with a crosssectional area A
u
. This stream is cooled by intro
duction and evaporation of a mass ow rate
M
l
(kg/s) of a liquid at temperature
T
l
= T
f .1
into the stream. Then we have
M
lg
=
M
l
. At a sufciently far distance from
the location of evaporation, the temperature is uniformand we have convection only
(i.e., Q
u.1
and Q
u.2
). Assuming that all the liquid evaporates, the exit temperature
T
f .2
is found by applying the integralvolume energy equation (2.9) to the control
volume shown in Chart 5.3, i.e.,
Q[
A.1
Q[
A.2
Q
loss
=
S
lg
Q
u.1
Q
u.2
Q
loss
=
M
l
Lh
lg
M
f .1
c
p.f
T
f .1
M
f .2
c
p.f
T
f .2
Q
loss
=
M
l
Lh
lg
A
u
(
f
c
p.f
u
f
T
f
)
1
A
u
(
f
c
p.f
u
f
T
f
)
2
Q
loss
=
M
l
Lh
lg
integralvolume energy equation with internodal energy conversion,
M
lg
=
M
l
. (5.17)
The heat loss term Q
loss
contains surface convection Q
ku
, surface radiation Q
r
,
and conduction Q
k
. For the case of Q
loss
= 0, we have an adiabatic evaporation
cooling. In general, Q
loss
is negative, i.e., parasitic heat ows into the control volume
from the surroundings (i.e., heat is gained).
For the case of
M
l
_
M
f .1
, we have
M
f .2
.
M
f .1
i.e., we can neglect the mass
added to the gas stream.
When only a portion of the liquid supplied is evaporated, then
M
lg

M
l
and
we determine
M
l
from the mass transfer rate considerations. In Section 6.9, we
will examine surfaceconvection mass transfer resistance R
Du
, where D stands for
the mass diffusivity. There we will derive the relation between
M
lg
and the vapor
densities at the liquidgas interface and in the fareld regions of the stream.
EXAMPLE 5.2. FAM
In a trickle bed evaporator, water ows over synthetic ber and along the gravity
vector while air is own through the bed, as shown in Figure Ex. 5.2(a). The air
stream enters at temperature T
f .1
and provides the heat for the evaporation of
water and exits at a lower temperature T
f .2
. Assume that the evaporation rate
M
l
is known and is much smaller than
M
f .1
.
(a) Draw the thermal circuit including internodal energy conversion.
(b) For the conditions given below, determine the exit temperature T
f .2
.
T
f .1
= 35
C, u
f .1
= 0.3 m/s, l = n = 0.5 m,
M
l
= 10
3
kg/s, Q
loss
= 0.
Evaluate the air and water properties at T = 300 K.
SOLUTION
(a) The thermal circuit diagram is shown in Figure Ex. 5.2(b).
MassoudKaviany 2011
g
u
f,2
T
f,2
< T
f,1
(
f,w
)
2
Air Flow
l
w
Water Distributor
(a) Physical Model: Evaporative Cooling of Gas Stream
Fan
Drain
RunAway
Water Collector
Synthetic Fiber Bed
u
f,1
T
f,1
(
f,w
)
1
M
l
S
lg
= M
lg
h
lg
M
l
(c) Neglected Heat and Mass Transfer Resistances (b) Thermal Circuit Model
(R
ku
)
s
(R
Du
)
s
(
f,w
)
s
(
f,w
)
T
lg
A
ku
T
f,
T
f,2
Q
u,2
Q
loss
Q
u,1 T
f,1
q
ku
, Q
ku
= A
ku
q
ku
m
Du
, M
lg
= A
ku
m
Du
S
lg
=
M
lg
h
lg
M
lg
Single Fiber Liquid Film
Q
loss
Figure Ex. 5.2. Evaporation cooling of air stream in trickle bed evaporator. (a) Physical
model. (b) Thermal circuit model. (c) Heat and mass transfer to a single ber in the
trickle bed.
(b) The exit temperature T
f .2
is determined from the energy equation (5.17),
i.e.,
T
2
= T
1
M
l
Lh
lg
Q
loss
(
Mc
p
)
f
= T
1
M
l
M
f
Lh
lg
c
p.f
= T
1
M
l
A
u
f .1
u
f .1
Lh
lg
c
p.f
.
where we assume that
M
f .2
=
M
l
M
f .1
for
M
l
_
M
f .1
.
M
f .2
=
M
f .1
M
f
.
MassoudKaviany 2011
From the geometry, we have
A
u
= ln
M
f
= A
u
f .1
u
f .1
= ln
f .1
u
f .1
.
From Table C.22 for air at T = 300 K, we have
f .1
=
f .2
= 1.177 kg/m
3
Table C.22
c
p.f
= 1,005 J/kgK Table C.22.
From Table C.27 for water at T = 300 K, we have
Lh
lg
= 2.438 10
6
J/kg Table C.27.
Then
T
2
= 35(
C)
10
3
(kg/s)
0.5(m) 0.5(m) 1.177(kg/m
3
) 0.3(m/s)
2.438 10
6
(J/kg)
1,005(J/kg)
= 35(
C) 27.48(
C) = 7.519
C.
COMMENT
Note that
M
l
= 10
3
kg/s _
M
f .1
= 88.28 10
3
kg/s.
The water ows at a rate of 10
3
kg/s = 1 g/s and since this is a lowliquid ow
rate, the name trickle ow is used. Note that the air is cooled down substantially.
Figure Ex. 5.2(c) shows that in practice surface convection heat and mass transfer
resistances determine
M
lg
(which can be smaller
M
l
resulting in excess and
runaway liquid ow rate). We will discuss these in Section 6.9 in connection with
the surfaceconvection heat transfer.
5.3.2 Cooling of Liquid and Gaseous Streams by Wall Seepage and
Surface Evaporation
Againwe consider steadystate, uniform, onedimensional owof liquidfroma liquid
stream toward a permeable wall (i.e., a porous wall, e.g., a nonglazed clay slab), and
then seepage of this liquid through the permeable wall, and nally its complete
evaporation at the gaseous ambient interface. This is shown in Figure 5.5. The liquid
velocity at surface 1 is u
f
and remains the same until surface 2 is reached, and there
a complete evaporation occurs. The goal here is to derive the relation between u
f
(the allowable seepage velocity for complete evaporation) and the thermodynamic
and transport properties of the liquid and the permeable wall. The heat for the
evaporation is generally from both the liquid reservoir and the gaseous ambient.
This type of surface evaporation is also encountered in sweatcooling of the human
body.
MassoudKaviany 2011
q
k,u
q
1
q
u
q
k
k
Liquid
f
, c
p, f
, u
f
Fluid Uniform
Velocity, u
f
Intramedium
Conduction
Convection
Uniform Temperature
T
f,1
Reservoir
Heat Transfer
Uniform
Temperature
T
f,2
= m
lg
h
lg
= m
f
h
lg
GasSide Heat
Transfer, q
2
= 0
Vapor
Liquid Stream
(or Reservoir)
T
f,1
T
L
T
f,2
Convection
and
Conduction
Gaseous Stream
Surface Evaporation
1 2
Porous Wall
(Clay)
EvaporationCooled
Control Surface
s
n
s
n
x
1
Pe
L
= 0
1 0
0
Uniform Gas Temperature
Uniform Liquid
Temperature
Increasing Pe
L
L 0
x
x
*
T
*
Neglected
Mass Transfer
Resistance
to Vapor Flow
Surface
Evaporation
Water Cooler
Permeable
Wall
S
lg
A
k,u
Figure 5.5. Liquid seepage and complete surface evaporation in a surfaceevaporation water
cooler.
Here we assume that the liquid temperature in the reservoir is uniform and at
T
f .1
and that at the evaporation surface the temperature is T
f .2
(due to evaporative
cooling, we have T
f .2
 T
f .1
). Here we assume that T
f .2
is the same as the uniform
ambient gaseous temperature, T
f .2
= T
f .2
(p
g
), i.e., saturation temperature at vapor
pressure p
g
. This assumption does not allow for any heat transfer from the ambient
gas to the evaporation surface. In a practical systems, where the evaporation is into
a gaseous mixture of an inert (i.e., noncondensible) gas and a vapor, T
f .2
is still
the saturation temperature at the surface, but at the partial pressure of the vapor.
Also, for an ambient gas with a relative humidity less than unity, this surface partial
pressure is larger than that far from the surface (i.e., the fareld partial pressure).
We proceed with prescribed T
f .1
and T
f .2
 T
f .1
. We will address the gaseous mass
transfer at the evaporation surface in Section 6.9.
MassoudKaviany 2011
Liquid Stream
L
Surface Energy
Equation for Liquid
Reservoir Surface
Separate
Mechanisms
Surface Energy Equation
for Evaporation Surface
s
n
m
f
h
lg
q
2
= 0 (at x = L
+
)
q
1
q
1
Gas Stream
Control Surface Control Surface
Surface
Evaporation
s
n
s
n
s
n
x
Porous Wall
(SolidLiquid)
A
k,u
A
k,u
q
k
= k
x = 0
dT
f
dx
q
k
q
u
S
lg
Combined
Presentation of
Mechanisms
q
2
= 0
u
f
u
f
q
k,u
Figure 5.6. Energy equation for the liquidporous wall and the porous wallgas interface.
(A) Allowable Seepage Speed: Jakob Number Ja
l
We nowapply the surface energy equationtothe evaporationsurface. This is depicted
in Figure 5.6. We use (2.62) with no radiation and note that the temperature gradient
on the gas side of the interface is assumed zero. This results in no heat transfer
between the surface and the gas stream, i.e., Q
2
(x = L
) =
S
lg
= A
k.u
m
f
Lh
lg
T
f .1
T
f .2
(R
ku
)
12
[
x=L
= A
k.u
m
f
Lh
lg
evaporationsurface
energy equation,
(5.18)
where A
k.u
is the conductionconvection surface area (note that we have used both
k and u subscripts) and k) is the effective conductivity of the porous wall. We have
used (5.13) for (Q
k.u
)
12
.
Here k) is the effective thermal conductivity of the porous soliduid and
includes an effect due to the motion of the uid. This effect is then called dispersion
and is discussed in [6]. Equation (5.18) shows that the heat conductionconvection
to surface 2 results in a complete evaporation. Substituting for the resistance from
Table 5.1, (5.18) then becomes
k)u
f
f
e
Pe
L
e
Pe
L
1
(T
f .1
T
f .2
) = m
f
Lh
lg
evaporationsurface energy equation
with no gasside heat transfer.
(5.19)
MassoudKaviany 2011
Noting that m
f
= (u)
f
, and with rearranging, the above is written as
1 e
Pe
L
= Ja
l
. Ja
l
c
p.l
(T
f .1
T
f .2
)
Lh
lg
Jacob number as ratio
of sensible heat to
evaporation heat.
(5.20)
The Jakob number Ja
l
is the ratio of the sensible heat available in the liquid
(note that we used c
p.l
to emphasize that the uid phase of interest here is the liquid)
to the heat of evaporation. Note that Ja
l
has a similarity with Ste
l
(Stefan number)
used in Section 3.8 for melting. Table C.6 gives the maximum value for Ja
l
, for some
liquids, at one atm pressure. Equation (5.20) can be further rearranged to give
Pe
L
=
u
f
L
f
= ln(1 Ja
l
) (5.21)
or
u
f
=
f
L
ln(1 Ja
l
)
allowable seepage speed
for complete evaporation.
(5.22)
From (5.22), we observe that when Ja
l
= 1, i.e., when the heat content of the
liquid (between temperatures T
f .1
and T
f .2
) is equal to the heat of evaporation
(evaluated at T
f .2
), then there is no limit on the seepage velocity u
f
(there is a
hydrodynamic limit, but it is not considered here). The maximum temperature
difference, T
f .1
T
f .2
, is limited by the boiling temperature at T
f .1
and freezing
temperature at T
f .2
Table C.6 is constructed using these limits. Since Ja
l
is less than
unity for all substances listed, (5.22) can be used to determine the allowable seepage
speed, corresponding to complete evaporation.
Note that the larger the value of
f
= k),(c
p
)
f
, the larger is the allowable
seepage velocity (or speed) u
f
.
(B) Liquid Stream Cooling Rate
This evaporative cooling can be used to cool the liquid stream from which the liquid
ows. Starting from (2.62), for the surface at x = 0, with no surface radiation, no
surface energy conversion, and noting that the liquid arrives and leaves the surface
at the same temperature, the heat conduction is the only heat transfer mechanism at
x = 0. This is shown in Figure 5.6. The reservoir heat conduction, i.e., the reservoir
heat ux at the liquidporous wall interface, is q
1
= Q
1
,A
k.u
. The conduction heat
transfer rate at x = 0 is found by evaluating the derivative dT
f
,dx at x = 0 and then
multiplying the resultant by k). The derivative is found by differentiating (5.12).
Then (2.62) for the x = 0 surface becomes
q
1
=
Q
1
A
k.u
= k)
dT
f
dx
[
x=0
=
k)(T
f .2
T
f .1
)
L
Pe
L
e
Pe
L
1
rate of reservoir
liquid cooling
(heat withdrawal
rate per unit area).
(5.23)
MassoudKaviany 2011
Equation (5.23) shows that as Pe
L
0, more heat is conducted at x = 0, and as
Pe
L
increases, this heat conduction contribution diminishes (as is evident from the
temperature distribution shown in Figure 5.2). As Pe
L
, from (5.19), we nd
that q
k
(x = 0) = 0. This corresponds to the adiabatic evaporation cooling, where the
entire supply of heat for the evaporation is from the sensible heat of the liquid.
When the gaseous ambient is an airwater vapor mixture, the watervapor pres
sure at the evaporating surface (and the surface temperature T
f .2
) would depend
on the air velocity and the fareld temperature and watervapor pressure (this is
discussed in Section 6.9).
In evaporation cooling of the human body as discussed in Example 1.3, sweating
at the surface cools the supercial tissue (e.g., dead cells and the fat tissue). Then
by conduction, this cooling effect penetrates further into the tissues and reaches the
muscle tissues and the blood vessels, causing cooling of the heat generating muscle
tissues and owing blood (i.e., cooling of perfusing blood).
EXAMPLE 5.3. FAM
A permeable, watercooling cylindrical jar (the radius of the jar is large, so that
a planar wall geometry can be assumed) has a wall thickness L = 3 mm and
an effective thermal conductivity k) = 0.8 W/mK. The water within the jar is
at T
f .1
= 20
Lh
lg
c
p.l
(1 e
Pe
L
).
The P eclet number Pe
L
is dened by (5.9) as
Pe
L
=
u
f
L
f
.
f
=
k)
(c
p
)
f
=
k)
(c
p
)
l
.
From Table C.27, we have for T = 300 K
c
p.f
= 4,179 J/kgK Table C.27
f
= 997 kg/m
3
Table C.27
Lh
lg
= 2.438 10
6
J/kg Table C.27.
Using the numerical values, we have
f
=
0.8(W/mK)
997(kg/m
3
) 4,179(J/kgK)
= 1.920 10
7
m
2
/s
Pe
L
=
10
5
(m/s) 3 10
3
(m)
1.913 10
7
(m
2
/s)
= 0.01562.
Then
T
f .2
= 20(
C)
2.438 10
6
(J/kg)
4,178(J/kg
C)
(1 e
0.01562
) = 10.93
C.
Also,
Ja
l
=
c
p.l
(T
f .1
T
f .2
)
Lh
lg
=
4,179(J/kg
C) (20 10.92)(
C)
2.438 10
6
(J/kg)
= 0.01556.
(b) The heat withdrawal rate per unit area is given by (5.23) as
q
1
=
k)
L
Pe
L
e
Pe
L
1
(T
f .1
T
f .2
).
Using the numerical values, we have
q
1
=
0.8(W/m
C)
3 10
3
(m)
0.01562
e
0.01562
1
[20(
C) 10.92(
C)]
q
1
= 2.338 10
3
W/m
2
.
MassoudKaviany 2011
(c) The conductionconvection resistance is given by (5.14) as
(R
k.u
)
12
=
L
A
k.u
k)
e
Pe
L
1
Pe
L
e
Pe
L
.
Using the numerical values, we have
(R
k.u
)
12
=
3 10
3
(m)
1(m
2
) 0.8(W/m
C)
e
0.01562
1
0.01562 e
0.01562
= 3.721 10
3
C/W.
COMMENT
Note that the conduction resistance is given by (5.15), i.e.,
R
k.12
=
L
A
k.u
k)
=
3 10
3
(m)
1(m
2
) 0.8(W/m
C)
= 3.75 10
3
C/W.
Since we have used A
k.u
= 1 m
2
, we write
A
k.u
(R
k.u
)
12
= 3.721 10
3
C,(W/m
2
).
This shows that since Pe
L
is small, the heat transfer at the reservoir
permeable wall surface is dominated by conduction. This is evident when we
compare R
k.u
)
12
and R
k.12
above.
In comparing these resistances, note that a product of the area and the
resistance less than 10
5
C/(W/m
2
) is considered very small (for example, as in
the discussion of the contact resistance in Section 3.3.5).
The Jakob number used here is Ja
l
= 0.01556 _1. Table C.6 shows that
values as high as Ja
l
= 0.19 are possible for water. For this, a larger T
f .1
T
f .2
is
required.
5.4 Combustion Heating of Gaseous Streams
In fueloxidant combustion, prior to reactions the fuel and the oxidant may be
mixed (as in carburetors) or the fuel may be directly injected into the oxidant (as
in fuelinjected engines). These are depicted in Figure 5.7. As another example of
steadystate, uniform, onedimensional ow and heat transfer with energy conver
sion, we consider the gaseous, laminar premixed combustion. The volumetric energy
conversion rate s
r.c
is given by (2.18). The effect of turbulence will be briey discussed
at the end of this section. The fuel (such as hydrocarbons) and the oxidizer (such as
oxygen), along with some inert gases (such as nitrogen in air), are mixed together
and ow at velocity u
f .1
. Figure 5.7(b) depicts this reacting stream. In practice, the
premixed gas may ow through a short and wide tube, so we neglect the effect of the
conning tube wall.
The ow may also be through a porous medium. The velocity u
f .1
is referred to
as the ame speed. In principle, the ame moves at the speed u
f .1
, toward the ow,
and since the ow speed is also u
f .1
, it exactly opposes the ame propagation; thus
the ame appears stationary. This ame speed u
f .1
, depends on the thermochemical
MassoudKaviany 2011
(c) Premixed Flame in a Porous Medium
(b) Premixed Flame in a Plain Medium
Hot Products
of Reaction
Porous Slab
Flame
Casing
u
f,1
, T
f,1
Cold Premixed Fuel
Oxidant (Reactants) Stream
Cold Premixed Fuel
Oxidant (Reactants) Stream
u
f,1
, T
f,1
u
f,2
, T
f,2
Hot Products
of Reaction
u
f,2
, T
f,2
x
x
Flame Not Influenced
by Tube Wall (One
Dimensional, Adiabatic)
Flame Influenced by Tube
Wall (Multidimensional,
with Heat Loss)
Tube
S
r,c
S
r,c
(a) Injected Fuel Flame
u
f,0
, T
f,0
Cold Oxidant Stream
x
u
f,F
, T
f,F
Cold Fuel
Stream
Hot Products
of Reaction
u
f,2
, T
f,2
Tube
Flame
S
r,c
Figure 5.7. (a) Injectedfuel ame in a plain medium. (b) Premixed, nearly onedimensional
gaseous owandcombustionina plainmedium. (c) Premixed, nearly onedimensional gaseous
ow and combustion in a porous medium.
and thermophysical properties of the gas. Therefore, a given fueloxidant mixture,
and inlet pressure and temperature, has a unique adiabatic (no lateral heat losses)
ame speed u
f .1
. If the gas ow has a speed larger than this unique speed, the ame
blows away (i.e., ame blow out occurs).
If the speed is lower than this unique speed, the ame will move into the tube
(i.e., ashback) occurs. Our goal here is to derive the relation between u
f .1
and the
thermochemical and thermophysical properties (including the chemical reaction).
The stoichiometric chemical reaction is given with the stoichiometric coefcients
i
(moles of species i) as
F
F
O
O
I
I
P
P
I
I stoichiometric reaction. (5.24)
where F stands for the fuel, O for the oxidizer, I for the inert species, and P for the
products.
MassoudKaviany 2011
x
q
u
q
u
q
u
q
k
q
T
f
u
f,1
T
f,2
T
f,1
T
f
0
n
r,F
h
r,F
, ChemicalBond
Energy Conversion
u
f,2
,
f,2
, T
f,2
u
f,1
=u
F
,
f,1
, T
f,1
ConductionConvection
(Preheat) Region
Conduction
Reaction Region
=
p
+
r
Flame
Thickness
Gaseous Premixed
Fuel and Oxidizer
(and Inert Gases)
Gaseous Product
Conduction, Convection,
and ChemicalBond
Energy Conversion
Beginning of
Preheat Region
Uniform Fluid
Velocity
End of Reaction
Region
OneDimensional, Adiabatic Premixed Flame
S
r,c
u
f,2
Uniform Fluid
Velocity,
u
f,2
> u
f,1
(Due to Decrease
in Density)
Figure 5.8. An idealized, steadystate, onedimensional ow and combustion.
As an example, consider the commonly used methaneair combustion given by
the stoichiometric reaction in Example 2.5, which is repeated here as
CH
4
2O
2
7.52N
2
CO
2
2H
2
O7.52N
2
methaneair stoichiometric
reaction.
(5.25)
The air is assumed to be made of oxygen and nitrogen only. Here, according to (5.24),
we have
P
P = CO
2
2H
2
O (i.e., the carbon dioxide and water vapor are treated as
a single product).
In (2.18), i.e., s
r.c
= n
r.F
Lh
r.F
, the volumetric rate of the reaction is designated
by the rate of consumption of the fuel n
r.F
(kg/m
3
s). The heat of reaction is dened
per kg of the fuel (i.e., Lh
r.F
). At a location, the mixture is initially ignited, and
thereafter, given sufcient amount of fuel and oxidant, the heat supplied by the
chemicalbond energy conversion s
r.c
is able to sustain the reaction.
A schematic of the idealized, steadystate, onedimensional ame structure con
sidered is given in Figure 5.8. We consider the case where there are no lateral or
axial heat losses (this is called the adiabatic) ame. We also consider the case of
stoichiometric reactions (i.e., all the fuel and oxidizer are consumed). Then the
temperature after the completion of reaction, i.e., T
f .2
, shown in Figure 5.8, is the
adiabatic ame temperature for this stoichiometric reaction. In the high temperature
MassoudKaviany 2011
Table 5.2. Standard heat of reaction Lh
r.F
(also called caloric value) for some gaseous
fuels with oxygen (at thermodynamic STP),
in order of decreasing heat of reaction
Fuel Lh
r.F
, J/kg
hydrogen, H
2
142.1 10
6
methane, CH
4
55.53 10
6
ethane, C
2
H
6
51.92 10
6
propane, C
3
H
8
50.40 10
6
ethylene, C
2
H
4
50.33 10
6
acetylene, C
2
H
2
48.26 10
6
butane, C
4
H
10
49.59 10
6
benzene, C
6
H
6
42.29 10
6
pentane,C
5
H
12
49.05 10
6
hexane, C
6
H
14
44.92 10
6
heptane, C
7
H
16
44.95 10
6
octane, C
8
H
18
48.37 10
6
methanol, CH
3
OH 21.56 10
6
ethanol, C
2
H
5
OH 28.83 10
6
gasoline fuel, C
7
H
17
44.89 10
6
diesel fuel, C
14.4
H
24.9
43.31 10
6
hydrogen sulde, H
2
S 16.51 10
6
carbon monoxide, CO 10.11 10
6
region, shown in Figure 5.8, the volumetric gaseous radiation becomes noticeable
(but not dominating). For simplicity, we do not include radiation along with conduc
tion and convection. However, in Section 5.6, we address gasstream radiation as a
postreaction heat transfer.
Tables 5.2 andC.21(a) and(b) (inAppendix C) list the heat of reactionLh
r.F
(J/kg
fuel) of some gaseous fuels, and Figure 5.9 shows the adiabatic ame temperature of
some gaseous fuels reacting in air or oxygen [11].
The equivalence ratio + is dened as the ratio of the actual (
F
,
O
)
ac
to the
stoichiometric (
F
,
O
)
st
, i.e.,
+ =
_
O
_
ac
_
O
_
st
equivalence ratio. (5.26)
For a fuelrichmixture we have + > 1, andfor a fuelleanmixture we have +  1.
The stoichiometric mixtures have + = 1. Most of the reactions listed in Figure 5.9
are for + = 1.
Consider reaction of acetylene or methane with pure oxygen. Figure 5.9 gives
the adiabatic ame temperatures for these reactions. Since acetylene C
2
H
2
has a
higher mass fraction
F
,
f
, its reaction with air results in a higher adiabatic ame
temperature, for nearly equal heat of reaction Lh
r.F
(Table 5.2). From Figure 5.9, for
T
f .1
= 16
C and + = 1, T
f .2
for the acetyleneair combustion is 2,632
C, while it is
MassoudKaviany 2011
Acetylene in O
2
Hydrogen in O
2
Natural Gas in O
2
Acetylene in Air
Carbon Monoxide
Gasoline in Air
Natural Gas in Air
Natural Gas in Air
Natural Gas in Air
Coke Oven Gas in Air
Natural Gas in Air
Methane in Air
Natural Gas in Air
Natural Gas in Air
Natural Gas in Air
Blast Furnace Gas in Air
Hydrogen in Air
Natural Gas in Air
Natural Gas in Air
Methanol in Air
Producer Gas in Air
Natural Gas in Air
16
16
16
16
16
760
482
316
16
16
16
16
16
16
16
16
760
593
16
16
16
3,110
2,974
2,827
2,632
2,468
2,243
2,160
2,093
1,988
1,961
1,918
1,871
1,816
1,591
1,454
2,210
2,260
2,310
2,204
1,904
1,654
1,649
1.00
Fuel
Air (or O
2
)
Temperature
T
f,1
, C
Equivalence
Ratio
Predicted Adiabatic
Flame Temperature
T
f,2
, C
1.00
1.00
1.00
1.00
16 1.00
1.11 (Rich)
1.00
1.00
1.00
1.00
1.00
1.18 (Rich)
0.86 (Lean)
0.70 (Lean)
1.00
1.00
1.00
1.00
1.00
1.00
1.49 (Rich)
Figure 5.9. Predicted adiabatic ame temperature for some gaseous fuelair or fueloxygen
combustions. (From Read, R.J., reproduced by permission c 1984 North American Manu
facturing.)
MassoudKaviany 2011
Table 5.3. Chemical kinetic constants for onestep oxidation , given by
n
r.F
= a
r
a
F
F
a
O
O
e
LE
a
,R
g
T
, for some hydrocarbons
Formula Name M, kg/kmole a
F
a
O
a
r
LE
a
, J/kmole
CH
4
methane 16 0 0 1.3 10
8
2.1 10
8
CH
4
methane 16 0.3 1.3 5.3 10
8
2.0 10
8
C
2
H
6
ethane 30 0.1 1.65 4.3 10
8
1.3 10
8
C
3
H
8
propane 44 0.1 1.65 4.8 10
8
1.3 10
8
C
4
H
10
butane 58 0.15 1.6 5.1 10
8
1.3 10
8
C
5
H
12
pentane 72 0.25 1.5 4.9 10
8
1.3 10
8
C
6
H
14
hexane 86 0.25 1.5 5.0 10
8
1.3 10
8
C
7
H
16
heptane 100 0.25 1.5 5.0 10
8
1.3 10
8
C
8
H
18
octane 114 0.25 1.5 5.0 10
8
1.7 10
8
C
9
H
20
nonane 128 0.25 1.5 8.5 10
8
1.3 10
8
C
10
H
22
decane 142 0.25 1.5 5.4 10
8
1.3 10
8
C
6
H
6
benzene 78 0.1 1.85 4.2 10
8
1.3 10
8
C
7
H
8
methyl benzene 92 0.1 1.85 2.7 10
8
1.3 10
8
CH
3
OH methanol 32 0.25 1.5 6.7 10
7
1.3 10
8
C
2
H
5
OH ethanol 46 0.15 1.6 1.8 10
8
1.3 10
8
1,918
C for the methaneair combustion. From Table 5.2, also note that a hydrogen
oxygen reaction has a very high heat of reaction, compared to the hydrocarbons.
Tables C.21(a) and (b) further list the heat of reaction of various pure and
industrial gaseous fuels (and their compositions).
Table 5.3 lists the constants in the chemical kinetic model given by (2.19) used for
the oxidation of some gaseous fuels. The general onestep chemical reaction is that
given by (5.24). For the oxidation of methane we have (5.25). Note that for methane
in the rst model a
F
a
O
= 0.0 and in the second model a
F
a
O
= 1.0. This second
model is a rstorder kinetic model. Here we have n
r.F
(kg/m
3
s), LE
a
(J/kmole),
F
(kg/m
3
), and
O
(kg/m
3
), and the unit of a
r
depends on the constants a
F
and a
O
.
For simplicity, we will use the zerothorder model. The analysis of higherorder
reactions are discussed in [4,10,13].
5.4.1 Conservation Equations, Adiabatic Flame Temperature,
and FarField Thermal Conditions
For the steadystate, uniform, onedimensional ow, the energy, mass (i.e., continu
ity), andspecies (e.g., fuel,
f
) conservationequations, for LV 0, are the simplied
forms of (5.3), (B.23), and (B.30), i.e.,
d
dx
_
k
f
dT
f
dx
f
c
p.f
u
f
T
f
_
= s
r.c
= n
r.F
Lh
r.F
energy equation
spatial rate of change volumetric rates of
of T in x direction energy conversion
(5.27)
MassoudKaviany 2011
d
dx
m
f
=
d
dx
f
u
f
= 0
continuity equation
spatial rate of change
of m in x direction (5.28)
d
dx
_
D
m.F
f
d
dx
f
F
u
f
_
= n
r.F
fuelspecies
conservation
equation,
spatial rate of change volumetric rate of formation
of
F
in x direction of the species F (5.29)
where k
f
,
f
, and c
p.f
are the gas mixture properties. We have used (2.18) for the the
energy conversion terms. From Figure 3.2(b), we note that c
p.f
for CH
4
, O
2
, N
2
, CO,
and H
2
O are temperature dependent and a proper average value should be used
for the mixture c
p
. As discussed in Section 3.1.1, the specic heat capacity for an
idealgas mixture is c
p.f
= Y
f .i
(c
p.f
)
i
,
f
. In (5.27) to (5.29), n
r.F
is the volumetric
rate of formation of fuel species (it is negative here becomes the fuel is consumed
during oxidization), Lh
r.F
is the heat of reaction,
F
,
f
is the mass fraction of the
fuel species, and D
m.F
(m
2
/s) is the species mass diffusion coefcient of the fuel
diffusing into the rest of the mixture species. Table C.20(a) list D
m.i
for some gaseous
species. As is evident, the magnitude of D
m.F
is of the order of 10
5
m
2
/s. This makes
D
m.F
similar in magnitude to
f
for gases.
Note the similarity between the energy and fuelspecies conservation equations, i.e., (5.27) and
(5.29). Here we have chosen a reaction kinetic model (5.30) that is independent of the species
concentrations. Therefore, we do not need to include (5.29) in the analysis. In general, this is not the
case and the species concentrations inuence the reaction rate, e.g., in the reaction kinetic model
given in (2.19). Then (5.29) has to be solved simultaneously with (5.27). When (5.29) has to be solved,
some simplifying assumptions are made. Here we have assumed that k
f
and c
p.f
are constant, and
if the product
f
D
m.F
is also constant, and (k,c
p
)
f
,D
m.F
= Le = 1, then (5.27) and (5.29) would
be mathematically identical. The Lewis number Le signies the ratio of speciesmass diffusivity to
thermal diffusivity D
m.F
,
f
.
MassoudKaviany 2011
x
T
f
T
f,2
T
f,1
Adiabatic (Q
loss
= 0)
Upstream Bounding Surface Downstream Bounding Surface
Flame
S
r,c
= (
F
u
F
A
u
)
1
h
r,F
Q
u,1
= A
u
(
f
c
p, f
u
f
T
f
)
1
Q
u,2
= A
u
(
f
c
p, f
u
f
T
f
)
2
Q
loss
(Lateral
Heat Losses)
Nonadiabatic (Q
loss
> 0)
u
f
Internodal Energy Conversion: Integral Volume
around OneDimensional Premixed Flame
A
V
A
1
A
2
s
n
s
n
T
f,1 T
f,2
Figure 5.10. Internodal energy conversion shown by the integral volume taken around a
onedimensional premixed ame.
where 1 and 2 refer to locations far downstream (nonreacted) and far upstream
(reacted) of the ame. We note that u
f
= u
f
(x) and
f
=
f
(x), such that u
f
(x
) = u
f .1
and
f
(x ) =
f .1
.
Similarly, for T
f .1
designating the temperature of the incoming nonreacted mix
ture, we use T
f .2
for the temperature of the completely reacted mixture.
To determine T
f .2
, we use the far upstream and downstream conditions and
note the lack of any gradient of temperature in these locations. Then we can use
the integralvolume energy equation (2.9) on the volume shown in Figure 5.10, and
using (2.18), i.e.,
Q [
A.1
Q [
A.2
Q
loss
=
S
r.c
(5.32)
Q
u.1
Q
u.2
Q
loss
=
S
r.c
(5.33)
A
u
(
f
c
p.f
u
f
T
f
)
1
A
u
(
f
c
p.f
u
f
T
f
)
2
Q
loss
=
M
F
Lh
r.F
= A
u
(
F
u
f
)
1
Lh
r.F
integralvolume energy equation, i.e., internodal energy conversion,
for complete reaction.
(5.34)
Note that this energy equation describes the internodal energy conversion between
thermal nodes T
f .1
and T
f .2
.
MassoudKaviany 2011
Now using (5.31), we solve for T
f .2
(ame temperature)
T
f .2
= T
f .1
Lh
r.F
c
p.f
_
f
_
1
Q
loss
A
u
f .1
c
p.f
u
f .1
ame temperature. (5.35)
Here we have assumed that the fuel is completely reacted at location 2; then the
temperature of the mixture rises by T
f .2
T
f .1
, and this increase is given by (5.35).
When Q
loss
= 0, (3.35) gives the adiabatic ame temperature. Note that among the
hydrocarbons with smaller Lh
r.F
, the higher (
F
,
f
)
1
results in the higher T
f .2
. This
mass fraction can be increased by increasing the number of C atoms in the molecule
relative to the H atoms. This is the reason acetylene results in a high adiabatic ame
temperature (as shown in Figure 5.7).
Also note that the constant c
p.f
used actually varies with the concentration of
the various species in the gas mixture and with temperature. Table C.30 lists c
p.f
for gaseous methane, ethane, and propane, and for nitrogen and oxygen. The water
vapor c
p.f
is listed in Table C.27. Then a concentrationtemperature averaged c
p.f
should be used in (5.35).
We now rewrite (5.27) as
k
f
d
2
T
f
dx
2
m
f
c
p.f
dT
f
dx
= n
r.F
Lh
r.F
differentialvolume
energy equation.
(5.36)
Note that Lh
r.F
is the heat of reaction per kg of the fuel.
To solve for T
f
, two boundary conditions are required. These are the fareld
(upstream and downstream) thermal equilibrium conditions given by (2.65). Then
using the prescribed T
f .1
and the adiabatic ame temperature T
f .2
found from(5.35),
we have
T
f
= T
f .1
(and
F
=
F.1
) for x upstream thermal condition (5.37)
T
f
= T
f .2
= T
f .1
Lh
r.F
(
F
,
f
)
1
c
p.f
(and also
F.2
= 0)
for x downstream thermal condition. (5.38)
5.4.2 Preheat and Reaction Regions
Our goal is to nd an explicit expression for the ame speed u
f .1
in terms of the ther
mochemical and thermophysical properties. In (5.36), the energy conversion term
is temperature dependent [i.e., (5.30)], and by making a few justiable assumptions
and simplications, we can arrive at a simple expression for u
f .1
.
One of the analysis methods used is to divide the ame into two regions. One
is the socalled preheat region, where the reaction is frozen (i.e., neglected), and
then in (5.32) only conduction and convection are allowed. This region is rendered
in Figure 5.8 and ends where the temperature of the mixture reaches the ignition
temperature T
i
. In general, T
i
 T
f .2
and Table C.21 gives some typical autoignition
MassoudKaviany 2011
temperatures. Here we will assume that T
i
is nearly equal to T
f .2
, and therefore, T
i
need not to be prescribed separately. The other region is the conductivereaction
region, where the convection term in (5.36) is assumed negligible, compared to the
conduction and reaction terms. This region is also shown in Figure 5.8. By choosing
x = 0 to mark the boundary between the conductionconvection (i.e., preheat) and
the conductionreaction (i.e., reaction) regions, we can write (5.36) as
k
f
d
2
T
f
dx
2
m
f
c
p.f
dT
f
dx
= 0  x 0
conductionconvection
(or preheat) region
spatial rate of spatial rate of
change of q
k
change of q
u
in x direction in x direction
(5.39)
k
f
d
2
T
f
dx
2
= n
r.F
Lh
r.F
0 x 
conductionreaction
(or reaction) region.
spatial rate of energy conversion
change of q
k
in x direction
(5.40)
The conductionconvection region is similar to that studied in Section 5.2. Here
the P eclet number based on the ame thickness is not expected to be large. So both
conduction and convection are signicant. The conductionreaction region is similar
to the conductionJoule heating of Section 3.3.6, but here the energy conversion
depends on the local temperature.
At x = 0, we use the thermal boundary condition given in Section 2.4, i.e., the
continuity of temperature (2.57) and the surface energy equation (2.62). Here q
r
= 0
andc
p
T(u s
n
) for bothsides of the surface cancel eachother. These conditions lead
to the matching of the temperatures and the conduction heat ow at the interface
separating the two regions, i.e.,
T
f .0
= T
f .0
= T
i
continuity of temperature (5.41)
A
u
k
f
dT
f
dx
[
0
= A
u
k
f
dT
f
dx
[
0
interregion, surface energy equation. (5.42)
where A
u
is the ow crosssectional area, 0
. (5.43)
Next, we rearrange (5.40), before integration, by rst multiplying it by dT
f
,dx,
i.e.,
dT
f
dx
d
dx
_
dT
f
dx
_
=
n
r.F
Lh
r.F
k
f
dT
f
dx
0
x  (5.44)
MassoudKaviany 2011
or
d
dx
_
dT
f
dx
_
2
=
2 n
r.F
Lh
r.F
k
f
dT
f
dx
0
x  . (5.45)
Upon integration and evaluation at 0
_
2
=
2Lh
r.F
k
f
_
T
f .2
T
f .1
n
r.F
dT
f
. (5.46)
Here in order to integrate over the maximum temperature difference T
f .2
T
f .1
,
we have used n
r.F
= 0 (i.e., frozen reaction) for the preheat region x  0, in which
T
f
 T
i
.
Next we assume that T
i
T
f .1
. T
f .2
T
f .1
, and substitute (5.43) into (5.42),
and evaluate it at 0
and set the resultant equal to the square root of (5.46). The
result is
dT
f
dx
[
0
=
m
f
c
p.f
(T
i
T
f .1
)
k
f
=
u
f .1
f .1
c
p.f
(T
f .2
T
f .1
)
k
f
=
u
f .1
Lh
r.F
F.1
k
f
(5.47)
dT
f
dx
[
0
=
_
2Lh
r.F
k
f
_
T
f .2
T
f .1
n
r.F
dT
f
_
1,2
=
dT
f
dx
[
0
=
u
f .1
Lh
r.F
F.1
k
f
. (5.48)
where in (5.47) we have used (5.35), and in (5.48) we have used (5.47).
This matching of the two temperature gradients given in (5.48) is also shown
in Figure 5.11(a). Note that in the conductionconvection region, the temperature
distribution is the exponential form given by (5.12).
5.4.3 Adiabatic Flame Speed u
f .1
and Thickness
Now we solve (5.48) for u
f .1
, and nd
u
f .1
=
_
2k
f
f .1
c
p.f
(T
f .2
T
f .1
)
_
1,2
_
1
F.1
_
T
f .2
T
f .1
n
r.F
dT
f
_
1,2
. (5.49)
The integration in (5.49) is simplied as follows. We note that the volumetric
reaction rate, i.e., n
r.F
, is signicant only when T
f
approaches T
f .2
. Then for T
f
approaching T
f .2
(i.e., when T
i
. T
f .2
), we can use the Taylor series expansion to
expand 1,T
f
around 1,T
f .2
as
1
T
f
=
1
T
f .2
T
f
T
f .2
T
2
f .2
rstorder approximation of 1/T
f
around 1/T
f .2
. (5.50)
MassoudKaviany 2011
0.0
10
20
30
40
Asymptotics Analysis
Numerical Experiment
50
0
8
6
4
2
0 10 20 30
E
a
= 2.10 x 10
8
J/kg
a
r
= 1.3 x 10
8
kg/m
3
s
1.0 0.6 0.0 0.4
0
0.5
1.0
0.2
(a) Flame Structure and ConductionConvection and ConductionReaction Regions
Slopes Match
Conduction
Convection
Region
Conduction
Reaction
Region
T
f,1
T
f
u
F
, T
f
F,1
F /
F, 1
F
/
F
,
1
p
x
x
T
f
0
u
f
,
1
,
c
m
/
s
T
f
/
T
f
,
1
T
f
/ T
f,1
x, mm Ze
(c) Comparison between Exact Numerical
Results and Results of High Ze
Approximation, for Flame Speed
(b) Predicted Dimensionless Temperature
and Fuel Mass Concentration Distrib
utions from High Ze Approximation
F,1
F,1
Figure 5.11. (a) Structure of premixed ame. (b) Distribution of dimensionless temperature
and fuel mass concentration within the ame. (c) Comparison of measured and predicted
ame speed. The results are for adiabatic, stoichiometric methaneair combustion.
Now substituting this in (5.30), we have
n
r.F
= a
r
e
LE
a
,R
g
T
f
. a
r
e
(LE
a
,R
g
)[1,T
f .2
(T
f .2
T
f
),T
2
f .2
]
= a
r
e
LE
a
,R
g
T
f .2
e
LE
a
(T
f .2
T
f
),R
g
T
2
f .2
volumetric reaction rate. (5.51)
where the rst exponent term is independent of T
f
and will be factored out of the
integrand. For the second term, which contains T
f
, we integrate this term to nd
_
T
f .2
T
f .1
e
LE
a
(T
f .2
T
f
),R
g
T
2
f .2
dT
f
=
R
g
T
2
f .2
LE
a
_
LE
a
(T
f .2
T
f .1
),R
g
T
2
f .2
0
e
LE
a
(T
f .2
T
f
),R
g
T
2
f .2
d
LE
a
(T
f .2
T
f
)
R
g
T
2
f .2
. (5.52)
MassoudKaviany 2011
The dimensionless quantity LE
a
(T
f .2
T
f
),R
g
T
2
f .2
is calledthe Zeldovichnum
ber Ze, i.e.,
Ze =
LE
a
(T
f .2
T
f .1
)
R
g
T
2
f .2
Zeldovich number. (5.53)
where Ze is the ratio of the activation energy for the reaction to the kinetic energy
and is generally much larger than unity (i.e., high activation energy for reaction) for
hydrocarbon oxidation. This allows us to write (5.52) as
_
Ze
0
exp
_
Ze
T
f .2
T
f
T
f .2
T
f .1
_
d
_
Ze
T
f .2
T
f
T
f .2
T
f .1
_
. 1
for Ze > 5 high activation energy approximation. (5.54)
Now using (5.54) and (5.51) with (5.49), we have the nal expression for u
f .1
as
u
f .1
=
_
2k
f
a
r
f .1
c
p.f
F.1
e
LE
a
,R
g
T
f .2
R
g
T
2
f .2
LE
a
(T
f .2
T
f .1
)
_
1,2
=
_
2k
f
a
r
f .1
c
p.f
F.1
Ze
e
LE
a
,R
g
T
f .2
_
1,2
laminar, premixed ame speed. (5.55)
Note that T
f .2
T
f .1
is determined from (5.35) and for the case of Q
loss
= 0, the
ame temperature T
f .2
and the ame speed u
f .1
are called the adiabatic laminar
ame temperature and speed. As Q
loss
increases the ame temperature and the
ame speed increase, up to a limit where the ame temperature will be too low to
sustain the ignition and combustion (this is called the quenching limit).
The laminar ame speed u
f .1
is proportional to k
1,2
f
, i.e., u
f .1
increases with an
increases in k
f
. Since
F.1
is proportional to 1,p
1
(with p
1
= p
2
= p), from (5.55)
u
f .1
is proportional to 1,p
1,2
(or p
0.5
), i.e., the ame speed decreases with increase
in the pressure. This is observed experimentally, and depending on the chemical
kinetic model used, p
0.1
or p
0.2
are also predicted [10]. We have assumed that c
p.f
and k
f
are constant and we need to determine k
f
and c
p.f
at an average temperature.
A simple, and nearly accurate choice would be to use (T
f .2
T
f .1
),2. The density
f .1
is found from the ideal gas relation (3.18), i.e.,
p
1
=
f .1
R
g
M
1
T
f .1
ideal gas. (5.56)
where M
1
is the upstream mixture molecular weight given by
M
1
=
CH
4
M
CH
4
O
2
M
O
2
N
2
M
N
2
CH
4
O
2
N
2
upstream, mixture molecular weight. (5.57)
The fuel density
F.1
is found from the stoichiometry relation
F.1
=
F
M
F
R
M
R
f .1
upstream, fuel density. (5.58)
MassoudKaviany 2011
The ame thickness is the sum of the thicknesses of the preheat and reaction
regions, i.e., =
p
r
(Figures 5.6 and 5.9). The thickness of the conduction
convection (or preheat) region
p
is estimated using the concept of axial conduction
convection discussed in Section 5.1.1. From (5.39), we use a unity preheat P eclet
number Pe
.p
(i.e., nearly equal conduction and convection heat uxes in the preheat
region), and thus we have
Pe
.p
=
u
f .1
f
. 1 (5.59)
or
p
.
f .1
u
f .1
thickness of preheat region, (5.60)
where
f .1
=
k
f
f .1
c
p.f
. (5.61)
Note that for Pe
.p
= 1 from Figure 5.3, we expect only a small deviation from
the conduction temperature distribution (Pe
.p
= 0).
The thickness of the conductionreaction region
r
is much smaller than
p
and
is also estimated using the equal strength of conduction and reaction in (5.40). This
leads to setting the product of the conductionreaction region P eclet number Pe
.r
and the Zeldovich number Ze equal to unity, i.e.,
Pe
.r
Ze =
u
f .1
f .1
LE
a
(T
f .2
T
f .1
)
R
g
T
2
f .2
. 1 or
r
.
f .1
u
f .1
1
Ze
thickness of
reaction region.
(5.62)
If we assume the pressure is uniform across the ame and the mixture molecular
weight remains the same, then the ratio of the two densities
f .1
and
f .2
is given by
the ideal gas law (5.56) as
f .1
f .2
=
T
f .2
T
f .1
. (5.63)
Figure 5.11(a) shows the two regions and the match in the temperature at the
interregion surface. It also shows the preheat and reactionregion thicknesses
p
and
r
. The variation of the normalized temperature T
f
,T
f .1
and the normalized fuel
density
f
,
f .1
within the ame, for the methaneair combustion at one atmpressure,
are shown in Figure 5.11(b). Note that the ame thickness, =
p
r
, is small,
i.e., less than one mm. The results obtained by using the above approximations (i.e.,
division of the ame region into two regions and assuming a large Ze) are compared
in Figure 5.11(c) with those obtained without these approximations by solving the
energy equation numerically. Note that the two predictions are in close agreements.
MassoudKaviany 2011
EXAMPLE 5.4. FAM
Methane is premixed and burned in air at one atm pressure, with a nonreacted
temperature T
f .1
= 16
C listed in
Figure 5.9.
(a) Determine the Zeldovich number Ze and the adiabatic ame speed u
f .1
.
(b) Compare this ame speed with the experimental value given in Table C.21.
(c) Determine the ame thickness .
The average specic heat capacity of the mixture can be determined from the
heat of reaction given in Table 5.2 and by using (5.35). The average mixture
thermal conductivity can be taken as that of air at the average ame temperature
(T
f .1
T
f .2
),2. The density
f .1
is determined from the idealgas law.
SOLUTION
(a) The ame speed is given by (5.55), i.e.,
u
f .1
=
_
2k
f
a
r
f .1
c
p.f
F.1
Ze
e
LE
a
,R
g
T
f .2
_
1,2
.
We need to determine the thermophysical properties k
f
.
f .1
, and c
p.f
. This
will be done later. The chemical kinetic constants a
r
and LE
a
are given for the
zerothorder methaneair reaction in Table 5.3 as
a
r
= 1.3 10
8
kg/m
3
s
LE
a
= 2.10 10
8
J/kmole.
The universal gas constant is listed in Table C.1(b), i.e.,
R
g
= 8.314 10
3
J/kmoleK.
From Figure 5.9, we have
T
f .2
= 1,918
C = 2,191.15 K.
The Zeldovich number Ze is dened by (5.53) as
Ze =
LE
a
(T
f .2
T
f .1
)
R
g
T
2
f .2
.
For the far upstream, we are given
T
f .1
= 16
C = 289.15 K.
Using the numerical value for Ze, we have
Ze =
2.10 10
8
(J/kmole)(2,191.15 289.15)(K)
8.314 10
3
(J/kmoleK)(2,191.15)
2
(K
2
)
= 10.
This is larger than the value of 5 assumed in the approximations given by (5.54).
Therefore, we expect the results of Section 5.4.3 to be valid.
MassoudKaviany 2011
The nonreacted methane mass fraction
F.1
,
f .1
is found from (5.57) using
the stoichiometric reactantsinert mixture given by (5.25) as
F.1
f .1
=
CH
4
M
CH
4
CH
4
M
CH
4
O
2
M
O
2
N
2
M
N
2
=
1(12.011 41.008)(kg/kmole)
[1(12.011 41.008) 2215.999 7.52214.007](kg/kmole)
=
16.043
290.70
= 0.05519.
where the molecular weights are found from the periodic table, Table C.2. Next
we determine the average specic heat capacity c
p.f
of the mixture from (5.35)
by using Lh
r.F
from Table 5.2, i.e.,
c
p.f
=
Lh
r.F
(
F.1
,
f .1
)
T
f .2
T
f .1
=
(5.553 10
7
)(J/kg) 0.05519
(2,191.15 289.15)(K)
= 1,611 J/kgK.
The average nonreacted (upstream) molecular weight M
1
is given by (5.57),
i.e.,
M
1
=
CH
4
M
CH
4
O
2
M
O
2
N
2
M
N
2
CH
4
O
2
N
2
=
[1(12.011 41.008) 2215.999 7.52214.007](kg/kmole)
(1 2 7.52)
=
290.70
10.52
= 27.63 kg/kmole.
The density of the mixture is found from the idealgas relation (3.18) and for a
uniform pressure p
1
, i.e.,
f .1
=
p
1
(R
g
,M
1
)T
f .1
=
(1.013 10
5
)(Pa)
8.3145 10
3
(J/kmoleK)
27.63(kg/kmole)
(289.15)(K)
= 1.164 kg/m
3
.
The average mixture thermal conductivity is approximated as that of air at the
average temperature across the ame T
f
)
, i.e.,
T
f
)
=
T
f .1
T
f .2
2
= 1,240.15 K.
From Table C.22, at T = 1,240 K, we have k
f
= 0.07939 W/mK. Again note
that k
f
changes signicantly with temperature.
MassoudKaviany 2011
Now we substitute all of the above into the expression for u
f .1
and have
u
f .1
=
_
2 0.07939(W/mK) 1.3 10
8
(kg/m
3
s)
1.164(kg/m
3
) 1,611(J/kgK) 0.05519 1.164(kg/m
3
) 10.0
e
[2.1010
8
(J,kg)],[8.31410
3
(J,kgK)2,191.15(K)]
_
1,2
u
f .1
= [1.713 10
4
(m
2
/s
2
) 9.855 10
6
]
1,2
= 0.4109 m/s = 41.09 cm/s.
(b) Table C.21 lists the stoichiometric methaneair amespeed as u
f .1
= 33.8
cm/s. The nearly good agreement here is noticeable considering the approxima
tions and simplications made in the analysis.
(c) The ame thickness is given by (5.60) and (5.62) as
=
p
r
=
f .1
u
f .1
f .1
u
f .1
1
Ze
=
f .1
u
f .1
_
1
1
Ze
_
=
0.07939,(1.164 1,611)(m
2
,s)
0.4109(m/s)
_
1
1
10
_
= 1.133 10
4
m = 0.1133 mm.
COMMENT
Note the rather extensive list of the thermophysical and thermochemical prop
erties that are used. The thermophysical properties are temperature dependent
and here we have had a signicant change in the temperature. The liquid and
gaseous properties of some hydrocarbons, along with oxygen and nitrogen, are
given in Table C.30 as a function of temperature. The chemical kinetic is also sim
plied greatly by the use of the zerothorder reaction. Note that this estimated
is smaller than the numerical result shown in Figure 5.11(b).
5.4.4 Nonadiabatic Flame Speed: Lateral Heat Losses
In many applications, there are lateral heat losses to surfaces containing the ame.
We can determine the ame speed for nonadiabatic ames by including the heat
losses (Q
loss
,= 0). This is done by determining T
f .2
T
f .1
in (5.55) and using (5.35).
Due to lateral heat losses, the reacted gas (i.e., ue gas) temperature T
f .2
decreases
and the ame speed decreases.
In most cases, the ame is multidimensional and even transient, and the energy
equation is then solved numerically. The numerical treatment will also allow for the
inclusion of complex, multistep chemical reactions and complex chemical kinetic
models. Central to the premixed ames is the dependence of the ame speed u
f .1
on the thermophysical, thermochemical, geometrical, and other properties of the
combustion system.
The combustionenergy release rate is highandis comparable toplasmagas Joule
heating. In these problems, the energy conversion rate and its distribution cannot
MassoudKaviany 2011
simply be prescribed; they need to be determined along with the heat transfer (and
uid ow) analysis.
5.4.5 Effect of Thermal Conductivity: Effective Thermal Conductivity
Equation (5.55) shows that u
f .1
is proportional to k
1,2
f
. One option for increasing
the ame speed (i.e., to burn more fuel and have a high m
f
=
f .1
u
f .1
) is to increase
the conductivity. One method of enhancement of k
f
is to burn the premixed gas in
a porous solid. The porosity c [i.e., void fraction dened by (3.27)] is chosen to be
large so the gas can ow without a large pressure drop. The effective conductivity k)
for the gasporous solid (with conductivity of each phase being k
f
and k
s
) composite
is signicantly larger than that for the gas alone (as discussed in Section 3.2.4).
Ceramic foams (e.g., silicon carbide SiC foams) provide such a high porosity (i.e.,
c . 0.90) along with a relatively large k),k
f
, while remaining nearly inert up to very
high temperatures. When the pores are rather large, the assumption of local thermal
equilibriumbetween the gas and the solid phases may not hold. Then allowance must
be made for this local temperature difference between the phases. This is done by
using two separate energy equations for the gas and the solid phases and by allowing
for surfaceconvection heat transfer between them (this is called the twomedium or
the local thermal nonequilibrium treatment).
The porous medium also allows for surfacemediated reactions (i.e., catalytic
reactions) that aim at decreasing the ignition temperature or the destruction of the
gaseous pollutant products.
5.4.6 Volumetric Radiation Heat Transfer: Radiant (Photon) Conductivity k
r
)
In Chapter 4, we addressed the volumetric gas radiation heat transfer for uniform
temperature gases. In Section 2.3.2, the volumetric absorption of radiation was dis
cussed using the extinction coefcient
ex
. The simultaneous treatment of the absorp
tion, emission, and scattering of radiation in a nonisothermal heat transfer mediumis
rather complex. When a porous medium is used, the radiation heat transfer through
the gasporous medium can be approximated using the concept of radiant conduc
tivity. This assumes a large optical thickness
ex
[dened in (2.44)] and treats the
radiation heat transfer as local (as compared to distant and integral for optically
thin media). This approximation applies to moderate and low porosities. Under the
assumptionof a signicant attenuationof the radiation(i.e., a socalledthickmedium
approximation), the radiative heat ux vector q
r
is given as
q
r
=
_
4
3
ex
)
_
E
b
=
_
16
3
ph
SB
T
3
_
T
k
r
)T =
_
16
SB
T
3
3
ex
)
_
T. k
r
) =
16
3
SB
T
3
ex
)
radiant conductivity k
r
) for optically thick (
ex
=
ex
L) porous media, (5.64)
MassoudKaviany 2011
where k
r
) is the effective radiant (or photon) conductivity,
ex
) is the effective
extinction coefcient [dened in (2.42)], and
ph
is the photon meanfree path. Then
we combine the effective and radiant conductivities as
q
k
q
r
= (k) k
r
))T. k = k) k
r
) total conductivity. (5.65)
A few simple correlations exist. They apply to various conditions, e.g., very
small particles, foams, etc. For example, for a packed bed of opaque cubic particles
(each side l) with a porosity c, the radiant conductivity k
r
) is related to the surface
emissivity c
r
, the solid conductivity k
s
, and l, through
k
r
) = 4l
SB
T
3
F
r
. F
r
=
c
r
N
1
r
(1 c)
1,3
c
r
N
1
r
(2 c
r
)
. N
1
r
=
k
s
4
SB
T
3
l
. (5.66)
where N
r
is the dimensionless conductionreaction number dened in (4.74). The
derivation is left as an end of chapter problem.
For a spherical particle, an equivalent linear dimension (having the same volume
for sphere and cubic) can be used by setting l = (,6)
1,3
D in the above equation,
where D is the diameter of the sphere.
For foams, the geometric parameters are the number of pores per meter n
p
(1/m)
and the porosity. Then using a corresponding cubic unit cell, we can make an approx
imation by using cubic particles with linear dimension l = (1 c)
1,3
,n in (5.66).
Note that as c 1, (5.66) results in k
r
) 0. This is because no solid is present
for the surface radiation to occur. Also as k
s
0, (5.66) results in k
r
) 0, because
the absorbed radiation on one side is not able to conduct to the opposite surface.
This shows that k
r
) is proportional to T
3
and to l. Comparing (5.64) and (5.66),
the meanfree path of the photons is proportional to l.
Now returning to the problem of premixed ame in porous media, from (5.65)
we note that k
r
) adds to the effective conductivity k) and the sum of these is
expected to be very large at high temperatures. Then, from (5.55), by replacing k
f
by k) k
r
), we note that this increase in k = k) k)
r
results in a much higher
ame speed (i.e., an increase in the burning rate) and a lower ammability limit. The
lower ammability limit means the fuels with more than stoichiometric air mixtures
or fuels with lower Lh
r.F
(i.e., low caloric fuels) can be used.
5.4.7 Effect of FreeStream Turbulence: Turbulence Intensity Tu
The gaseous forced ow, caused for example by the motion of a fan, has a velocity
that contains some temporal variations. For example, since the blades are separated
from each other by voids, the ow eld produced by the rotation of the blades will
include this timeperiodic effect. In addition, the ow over the blades may undergo
a ow separation and then ow eddies are formed that travel with ow.
These temporal velocity variations are generally rather random and are called
the turbulent uctuations. The reason they are called uctuations is that upon time
MassoudKaviany 2011
averaging over a long period of time, a mean velocity u
f
is found that no longer
changes as the time integration (i.e., integration over period ) is further extended
(i.e., ). Based on this, the instantaneous uid velocity u
f
is decomposed into
a mean velocity component u
f
, and a uctuating (or turbulent) velocity component
u
/
f
, i.e.,
u
f
= u
f
u
/
f
decomposition of uid velocity into
mean and uctuating components
(5.67)
u
f
=
1
_
0
u
f
dt. mean uid velocity. (5.68)
These velocity uctuations are called the freestream turbulence.
For the onedimensional ow considered here, and by assuming that the ow
turbulence is isotropic (i.e., all three components of the velocity uctuations behave
the same), the root mean square (rms) of the turbulent velocity is dened as
(u
/
f
2
)
1,2
= (
1
_
0
u
/
f
2
dt)
1,2
. rms of turbulent uctuations. (5.69)
This is similar to the molecular uctuation of gases and the thermal speed (u
2
m
)
1,2
of
Section 1.5.1. In practice, the freestream turbulence is designated by the turbulence
intensity Tu dened for a onedimensional mean ow with isotropic turbulence as
Tu
u
/
f .1
2
u
2
f .1
turbulence intensity. (5.70)
Through semiempiricism, it is shown that the turbulent ame speed u
f .1
is related to
the laminar ame speed u
f .1
through (u
/
f .1
2
)
1,2
and one of the simplest correlations,
which is valid for low turbulence intensity, is [12]
u
f .1
u
f .1
= 1
u
/
f .1
2
u
2
f .1
= 1 Tu
_
u
f .1
u
f .1
_
2
turbulent ame speed u
f .1
. (5.71)
where u
f .1
is found from (5.55). Equation (5.71) is solved for u
f .1
, once Tu and u
f .1
are known. This shows a signicantly higher ame speed associated with a large
freestream turbulence intensity.
The increase in the ame speed is due to turbulent thermal mixing (which is
represented by the turbulent thermal conductivity k
t
, as discussed in Section 6.3.2),
and also due to the turbulent mixing of the species.
In general, the turbulent ame speed u
f .1
also depends on the Damk ohler num
ber Da (which is the ratio of the reaction rate to the transport or mixing rate). To
determine Da, an empirical estimation of the turbulent ame thickness is needed
[12].
MassoudKaviany 2011
10
6
10
8
10
10
10
12
10
14
10
16
10
18
10
20
10
22
10
24
Typical Temperatures for Natural and Engineered Thermal Plasmas
T
f
, K
n
e
,
1
/
m
3
10 10
3
10
5
10
7
10
9
10
11
Thermonuclear
Fusion
Magnetohydrodynamic
Energy Con
version
Shock
Tubes
Hollow
Cathode
Thermionic
Energy Conversion
Fluorescent
Lamps
Glow
Discharges
Solar
Corona
Flames
Earth
Ionosphere
Interstellar
Space
Interplanetary
Space
Low Pressure
Arcs
10
23
All Solids Melt
Conduction Electron in Solid Conductors
Figure 5.12. Typical temperatures for thermal plasmas and their associated electron densities.
(From Konuma, M., reproduced by permission c 1992 SpringerVerlag.)
5.5 Joule Heating of Gaseous Streams
Electrically charged gases (i.e., plasmas) are heated by Joule heating through impos
ing an electric current. Gases become charged (i.e., ionized) at high temperatures
(T
f
> 2,000 K). When very high temperature gases (i.e., T
f
> 4,000 K) are needed,
Joule heating is used and the process is called the thermal plasma generation. Figure
5.12 shows the typical temperatures in natural and engineered thermal plasmas and
the associated electron density n
e
(1/m
3
). A thermal plasma is made of electrons,
ions, and neutral molecules and atoms. Because electrons are much lighter than the
rest of the species in the plasma, they heat up much faster. Therefore, unless there
are many collisions among all the species, there can be a thermal nonequilibrium
among them. This equilibrium is generally achieved, except for very short elapsed
times and at very low pressures.
Among the common gases used in thermal plasmas are Ar, H
2
, O
2
, N
2
, air, Ar
H
2
, ArHe, and CO
2
. Laboratory units with the capability of
S
e.J
= 30 to 50 kW and
industrial units of up to 1 MW power are used. Because of the high temperatures,
there are some desired heat losses to prevent the solids in contact with the plasma
from melting. There are also some undesired heat losses. Therefore the temperature
in the gas drops over a short distance, but the center of the plasma has temperatures
of about 10,000 K.
MassoudKaviany 2011
0.0
0.0
100% H
2
Experiment Mixture of Argon and Hydrogen Gases
(Volumetric Percentage)
5,000 10,000
T, K
k
,
W
/
m

K
15,000 20,000
5.0
10.0
15.0
90% H
2
50% H
2
30% H
2
100% Ar
Figure 5.13. Thermal conductivity of argonhydrogen gas mixture at elevated temperatures.
(From Boulos, M.I., reproduced by permission c 1991 IEEE.)
Thermal plasmas are used in manufacturing and materials processing, for exam
ple inheating of sprays of paint particles indry painting, inchemical vapor deposition,
in welding, and in surfacematerial removal. Because of the free electrons, plasmas
have a high electrical conductivity
e
, and depending on the gas, they can also have
a high thermal conductivity. Figure 5.13 gives the thermal conductivity of argon
hydrogen (ArH
2
) mixture, as a function of temperature and for various hydrogen
volume fractions [2]. Note that around T = 3,500 K, the conductivity of hydrogen
reaches a value of 15 W/mK (because k
e
> k
m
in ionized gases, this magnitude is
typical of some metallic solids). Because of this high thermal conductivity, such ther
mal plasmas can very effectively heat the entrained particles (heating rates, dT,dt,
of the order of 10
5
to 10
7
C/s are achieved).
The gas is charged by injection of charged particles, by pilot electrodes, by
highfrequency starting circuits, or by an initial increase in the gas temperature, for
example, by an acetyleneoxygen torch. For the Joule heating of charged gases, the
steadystate energy equation is that give by (5.1). In thermal plasma generation, the
volumetric radiation heat transfer becomes signicant and needs to be included. We
use (2.30) as the volumetric energy conversion and (5.1) becomes
(q
k
q
u
q
r
) = s
ij
s
e.J
= s
ij
e
e
e
e
e
= s
ij
e
e
2
e
differentialvolume energy equation, (5.72)
where s
ij
is the dissociationionization energy conversion rate and s
ij
 0, and
e
e
(V/m) is the electriceld intensity vector and can be steady or oscillating (i.e.,
periodic). Then a portion of s
e.J
is used for the dissociationionization.
Figure 5.14(a) shows the equilibrium composition of the argon plasma as a
function of temperature and at one atm pressure. Note that the concentration of
free electrons n
e
rises rapidly as T = 1.5 10
4
is reached and increases at a much
lower rate nearly after that. The number of neutral species n
A
decreases rapidly for
MassoudKaviany 2011
10
19
10
20
10
21
10
22
10
23
10
24
5 x 10
3
1.5 x 10
4
Air
Oxygen
Nitrogen
p = 1 atm
Argon
2.5 x 10
4
3.5 x 10
4
T, K
(a) Equilibrium Composition of Argon Plasma (p = 1 atm)
(b) Electrical Conductivity of Air, Argon, Nitrogen, and Oxygen Plasmas
n
i
,
1
/
m
3
n
Ar
++ (Double Ion)
n
Ar
+ (Single Ion)
n
e
(Electron)
n
t
(Total)
n
Ar
(Neutral)
10
10
2
10
3
10
4
5 x 10
3
1.0 x 10
4
1.5 x 10
4
2.0 x 10
4
2.5 x 10
4
3.0 x 10
4
e
,
1
/
o
h
m

m
T, K
Figure 5.14. (a) Equilibrium composition of argon plasma as a function of temperature, at
one atm pressure. (b) Electrical conductivity of air, argon, nitrogen, and oxygen plasmas as a
function of temperature, at one atmpressure. (FromLancaster, J.F., reproduced by permission
c 1986 Pergamon.)
T > 10
4
K and becomes relatively rare at T = 3.5 10
4
K. In Section 3.2.2(C)
we mentioned that solid semimetals have electron concentration of 10
23
to 10
28
electrons/m
3
and this places these thermal plasmas in the electronic conduction
range of semimetals. The electrical conductivity of several gases is given in Figure
5.14(b), as a function of temperature. Again note that for T > 10
4
K a plateau is
MassoudKaviany 2011
Thermal Plasma Generation
Arc (Electrode)
Radiofrequency
Range ( f between
240 kHz and 40 MHz)
Direct
Current
(DC)
Arc Nontransferred
Arc Transferred
Coolant
Alternating
Current
(AC)
Hybrid Other
Induction Coupling
(Electrodeless)
Induction and
Combustion
Heating
Others
Microwave
Range
( f 1 GHz)
Thermal
Plasma Jet
Cathode ()
Inlet
Gas
Flows
Inlet Gas Flows
Inlet Gas
Flows
Anode (+)
Thermal
Plasma Jet
Shroud Gas Slot
Carrier Gas Port
Cathode Gas Slot
Cathode ()
Workpiece,
Anode (+)
Plasma
Jet
9,600 K
6,000 K
Workpiece
(Cooled)
Induction
Coil
Induction Coupling
u
f,1
u
f,1
u
f,2
u
f,3
u
f,2
u
f,1
u
f,2
u
f,3
q
u
q
u
q
ku
S
e,J
S
e,J S
e,J
Workpiece
(cooled)
Thermal
Plasma Jet
Cooled
Surface
Chart 5.5. A classication of thermal plasma generators.
reached. These electrical conductivities are very large for the gas phase, but smaller
than the conductivities given in Table C.8 for metals (note that
e
= 1,
e
).
Determination of s
ij
involves determining the most signicant dissociation and
ionization reactions, which in turn requires knowledge of the associated equilibrium
compositions and the energy needed for breaking each of these bonds. This is not
addressed here and details can be found in [7].
5.5.1 Thermal Plasma Generators
The different types of thermal plasma generators are classied in Chart 5.5. The
rst level of the classication is based on the presence or absence of electrodes. The
electroded generators (i.e., arc generators) can be based on the direct or alternat
ing currents, while the induction coupled generators use an alternating current. In
electroded devices, current ows between the cathode and the anode. The nontrans
ferred arcs have a short distance between the electrodes, while the transferred arcs
MassoudKaviany 2011
have a large distance (of the order of one cm). The hybrid generator may combine
electroded and electrodeless generators or may combine Joule heating with combus
tion heating to generate thermal plasmas with a desired temperature and chemical
composition.
5.5.2 Thermal Plasma Classication
Figure 5.12 shows a classication of plasmas based on the temperature and the elec
tron density [1,2]. Again, note that as the electron density increases,
e
increases,
thus increasing the heat generation. For the calculation of the temperature distribu
tion, in addition to the energy equation (5.72), the momentum equation containing
the electromagnetic volumetric force f
e
should be included. The electron density
(and other charged species) is also accounted for through the electron conservation
equation [similar to the species equation (5.29)].
In Chapter 6, we will examine the heating of surfaces and particles by ow of hot
gases (i.e., surfaceconvection heat transfer). As we have noted, these very hot gases
can be generated by combustion, by Joule heating, or by a combination of both.
5.5.3 IntegralLength Analysis
Consider a onedimensional ow (an idealization), as shown in Figure 5.15. For
this onedimensional ow, we can use the integralvolume energy equation (2.9).
The fareld thermal conditions are the upstream, cold gas at temperature T
f .1
, and
the downstream hot gas at temperature T
f .2
. We allow for lateral heat losses (by
surface convection) Q
loss
= A
ku
q
ku
, where A
ku
is the surface area of the conning
walls, which are constantly cooled to prevent a meltdown. In Section 5.6, we discuss
radiation losses from gas streams. There we treat the radiant emission from the
gas as an energy conversion. For the areas designated in Figure 5.15, and assuming
convection heat transfer only at the two (upstream and downstream) ends, we have
the integralvolume energy equation (2.9) as
Q [
A
= Q [
A.1
Q [
A.2
Q
loss
= Q
u.1
Q
u.2
Q
loss
=
S
12
= [(
f
c
p.f
u
f
T
f
)
1
(
f
c
p.f
u
f
T
f
)
2
]A
u
A
ku
q
ku
=
_
V
s
ij
dV
_
V
s
e.J
dV =
S
ij
S
e.J
integralvolume energy equation, internodal energy conversion, (5.73)
where again
S
ij
is the energy used for the ionizationdissociation (
S
ij
 0), and in
terms of the plasma generator (or torch) diameter D
t
and length L
t
(shown in
Figure 5.14), we have
A
u
= D
2
t
,4. A
ku
= D
t
L
t
. V = D
2
t
L,4. (5.74)
MassoudKaviany 2011
x
L
t
q
u
q
ku
q
r
q
r
q
k
T
T
f,2
T
f,1
T
f
u
f,2
T
f,2
u
f,1
T
f,1
End of
Joule Heating
Adiabatic
With Lateral
Heat Losses
Beginning of
Joule Heating
Induction Coil
OneDimensional, Nonadiabatic Joule Heating of a Charged Gas
Cold Gas Hot Plasma
A
u
A
ku
e j
e
D
t
A
u
s
e,J
T
f,1
T
f,2
T
f,2
Q
u,2
= A
u
(
f
c
p, f
u
f
T
f
)
2
Q
u,1
= A
u
(
f
c
p, f
u
f
T
f
)
1
q
u
q
u
q
ku
q
r
u
f,1
Uniform Fluid
Velocity
u
f,2
Uniform Fluid
Velocity
Q
loss
Q
loss
(Lateral Heat Loss)
Figure 5.15. A onedimensional inductioncoupled, thermal plasma generation. The lateral
surfaceconvection heat losses are also shown.
Note that similar to (5.32), this energy equation describes internodal energy
conversion between thermal nodes T
f .1
and T
f .2
. In (5.72),
S
e.J
is the Joule heating
rate (or power) and is a fraction of the electrical power used in the induction coil.
The mass ow rate is conserved (for no mass crossing the lateral boundaries)
and as in (5.31), we have
m
f
= (
f
u
f
)
1
= (
f
u
f
)
2
. (5.75)
EXAMPLE 5.5. FAM
In inductioncoupled Joule heating of plasmas, relatively large gas velocities
(as high as 1,000 m/s) are used. Consider a plasma torch heating a owing
nitrogen gas with the following specications and with a heat loss of Q
loss
=
0.15(
S
ij
S
e.J
).
Determine the exit temperature T
f .2
.
D
t
= 5 cm, m
f
= 50 kg/m
2
s,
S
ij
S
e.J
= 5 10
4
W, T
f .1
= 25
C, c
p.f
=
1,400 J/kgK.
MassoudKaviany 2011
SOLUTION
The integralvolume energy equation (5.73) gives the desired relation. We will
also use (5.74), so that we can use the mass ow rate, i.e.,
m
f
c
p.f
A
u
(T
f .2
T
f .1
) =
S
ij
S
e.J
Q
loss
integralvolume energy equation.
We solve for T
f .2
from the above, and the result is
T
f .2
= T
f .1
S
ij
S
e.J
Q
loss
m
f
c
p.f
A
u
.
We now use Q
loss
= 0.15
S
e.J
and have the nal relation as
T
f .2
= T
f .1
0.85(
S
ij
S
e.J
)
m
f
c
p.f
A
u
. A
u
= D
2
t
,4.
Note that we have been using a constant c
p.f
and an average value must be
used.
Using the numerical values, we have
T
f .2
= 25(
C)
0.85(5 10
4
)(W)
200(kg/m
2
s) 1,400(J/kgK)(0.05)
2
(m
2
),4
= 25(
C) 3,864(
C) = 3,889
C.
COMMENT
In practice, the owand temperature across the tube (i.e., torch) are not uniform.
The center of the plasma torch (i.e., at the centerline of tube) is called the eye of
the plasma and has its highest temperature. In addition, axial heat losses occur,
causing an axial temperature nonuniformity.
5.6 GasStream Volumetric Radiation
Radiation heat transfer from an isothermal hot gas stream to the containing solid
walls can be calculated using a simple, approximate method. This uses the gas emis
sion and irradiation to the surface, as discussed in Section 4.5.3, along with a gas
absorption treatment.
Figure 5.16(a) shows the heat transfer between a hot stream of gas and its
bounding surface. The radiation exchange between a nongray gas and a nongray
surface is shown as energy conversions, s
e.
(absorption) and s
e.c
(emission).
The uid temperature decrease due to the radiation heat loss, i.e., T
f
= T
f
(x).
Using the differential volume along x, (5.3), and assuming a uniform temperature
T
f
across the tube (and no surfaceconvection heat transfer), we have
dq
u
dx
= ( mc
p
)
f
dT
f
dx
= s
e.
s
e.c
onedimensional gas energy equation (5.76)
or
A
u
( mc
p
)
f
dT
f
dx
= P
r
(
r.f
c
r.s
SB
T
4
s
r.s
c
r.f
SB
T
4
f
). (5.77)
where P
r
is the perimeter for radiation surface.
MassoudKaviany 2011
L
R
(a) Radiation Exchange Between a Nongray
Gas Stream and its Bounding Surface
(b) Gas Temperature Distribution
(c) Internodal Energy Conversion Representation
M
f
, T
f,1
A
u
A
r
= P
r
L
r,s
,
r,s
s
e,
+ s
e,
M
f
, T
f,2
P
r
x 0
q
u
q
u
0 L
T
T
f,1
T
f
(x)
T
f,2
T
s
x
Q
u,2
Q
u,1
Radiation Cooling
S
12
= S
e,
+ S
e,
Uniform Surface
Temperature, T
s
Well Mixed Gas Stream T
f
= T
f
(x)
Figure 5.16. Volumetric radiation heat transfer between a hot emitting gas stream and its
cold bounding surface. (a) Physical model. (b) Axial temperature distribution. (c) Internodal
energy conversion representation.
Note that we could have represented s
e.
s
e.c
as q
r
, as indicated in the
footnote of Sections 2.3.2(E) and 4.5.3.
In (5.77), we have used the product of the emissive power of the surface c
r.s
SB
T
4
s
and the absorptivity
r.f
, for the absorbed radiation by gas. The product of gaseous
irradiation to the surface c
r.f
SB
T
4
f
and the surface absorptivity
r.s
gives the heat
loss from the gas.
We can write (5.77), by separating the variables, as
P
r
c
r.f
r.s
SB
A
u
( mc
p
)
f
_
L
o
dx =
_
T
f .2
T
f .1
dT
f
a
1
T
4
s
T
4
f
(x)
. (5.78)
where
a
1
=
r.f
c
r.s
r.s
c
r.f
. (5.79)
The axial distribution of the gas stream cooled by volumetric radiations is shown
in Figure 5.16(b). Equation (5.78) is similar to (4.81) and the solution to (5.78) is
similar to (4.82). Here we simplify the solution by assuming that the bounding surface
MassoudKaviany 2011
temperature is smaller than the gas temperature to the extent that T
4
s
_T
4
f
. Then
(5.78) can be readily integrated to give
LP
r
c
r.f
r.s
SB
A
u
( mc
p
)
f
A
r
c
r.f
r.s
SB
(
Mc
p
)
f
=
1
T
3
f .2
1
T
3
f .1
. (
Mc
p
)
f
= A
u
( mc
p
)
f
(5.80)
or
1
T
3
f .2
=
1
T
3
f .1
A
r
c
r.f
r.s
SB
(
Mc
p
)
f
T
4
s
T
4
f .1
_1
uid exit temperature T
f .2
for
negligible surface emission.
(5.81)
This shows that as the gas mass owrate
M
f
decreases, or as the gas emissionsurface
absorption A
r
c
r.f
r.s
increases, the uid exit temperature T
f .2
drops further below
T
f .1
. The radiation heat loss results in a change in the uid convection heat ow
rate, i.e.,
Q
u
)
L0
Q
u.1
Q
u.2
Q
loss
= A
u
( mc
p
)
f
(T
f .2
T
f .1
)
S
e.c
)
L
.
S
e.c
)
L
=
_
L
0
P
r
r.s
c
r.f
SB
T
4
f
dx.
radiation loss
from gas stream.
(5.82)
where T
f .2
is found from (5.81). This internodal energy conversion is shown in
Figure 5.16(c).
Inclusion of the surface emission requires numerical integration. The surface
convection heat transfer, which is also important, is neglected here for simplicity.
In Chapter 7, we will discuss surfaceconvection heat transfer of bounded streams.
Simultaneous inclusionof volumetric radiationandsurface convectionwouldrequire
numerical solution.
EXAMPLE 5.6. FAM
The products of combustion (i.e., ue gas) of methaneair, with a low stoichiom
etry ratio (i.e., excess air), pass through a circular channel of radius R = 15 cm
and length L = 3 m, before arriving at a processing site where it is used for an
evaporationseparation process. The channel is made of stainless steel type 410.
This is shown in Figure Ex. 5.6(a). The walls of this channel are at T
s
= 550
C,
while the ue gas enters the channel at T
g
= 1,300
S
e.c
)
L
.
Use the properties of air at T = 1,500 K (Table C.22).
MassoudKaviany 2011
Flue Gas
u
f,1
, T
f,1
T
f,2
L >> R
r, s
S
e,
+ S
e,
Q
u,2
Q
u,1
Q
loss
= 0
Q
loss
= 0
Channel Wall
(a) Physical Model
(b) Thermal Circuit Model
Gas Radiation
Absorption and Emission
Internodal Energy Conversion
S
12
= S
e,
Mean Beam Length, <
ph
>
R
Uniform Surface
Temperature, T
s
, S
e,
<< S
e,
A
r
c
r.f
r.s
SB
A
u
( mc
p
)
f
.
where
A
r
= 2RL = 2 0.15(m) 3(m) = 2.826 m
2
A
u
= R
2
= 0.15
2
(m
2
) = 0.07665 m
2
m =
f .1
u
f .1
.
MassoudKaviany 2011
From Table C.22, we have for air at T = 1,500 K
f .1
= 0.235 kg/m
3
c
p.f
= 1,244 J/kgK.
From Table C.19, we have
stainlesssteel type 410 :
r.s
= 0.76 Table C.19.
The gas emissivity is determined as outlined in Section 4.5.3. The mean beam
length
ph
) is found from Table 4.4 for long cylinders, and is
ph
) = 1.9R = 1.9 0.15(m) = 0.114 m Table 4.4.
From Figures 4.23(a) to (c), we have
p
CO
2
ph
) = 0.06(atm) 0.114(m) = 6.840 10
3
atmm
c
r.CO
2
) . 0.035 at T
f
= 1,500 K Figure 4.23(a)
p
H
2
O
ph
) = 0.12(atm) 0.114(m) = 0.01368 atmm
c
r.H
2
O
) . 0.03 at T
f
= 1,500 K Figure 4.23(b)
c
s
ph
) = 10
7
0.114(m) = 1.140 10
8
m
c
r.soot
) . 0.02 at T
f
= 1,500 K Figure 4.23(c).
Then assuming no band overlapping or other interactions, we have from (4.63)
c
r.f
) = c
r.CO
2
) c
r.H
2
O
) c
r.soot
)
= 0.035 0.03 0.02 = 0.085.
The exit ue gas temperature is
1
T
3
f .2
=
1
(1,300 273.15)
3
(m
3
)
2.826(m
2
) 0.085 0.76 5.670 10
8
(W/m
2
K
4
)
0.07065(m
2
) 0.235(kg/m
3
) 1,244(J/kgK)
= 2.569 10
10
(1/K
3
) 5.011 10
10
(1/K
3
)
or
T
f .2
= 1,096 K.
MassoudKaviany 2011
Then from (5.82), with Q
loss
= 0, we have
Q
u
)
L0
= Q
u.2
Q
u.1
= A
u
(u)
f .1
c
p.f
(T
f .2
T
f .1
)
S
e.c
)
L
= 0.07065(m) 0.235(kg/m
3
) 0.1(m/s) 1,244(J/kgK)
(1,096 1,573.5)(K)
= 985.6 W radiation heat loss.
COMMENT
The gas emissivity correlations can be made more accurate by adding the
band overlapping and other interactions. The air properties were evaluated at
T = 1,500 K, and the average gas temperature is (T
f .2
T
f .1
),2 = (1,573.15
1,096)(K),2 = 1,335 K. More accurate calculations can be made by evaluating
the properties at this average temperature.
5.7 Summary
For the case of a onedimensional, uniform ow (i.e., uid ow in absence of a
solid surface), we have examined the combined resistance due to conduction and
convection R
k.u
. We have shown that for large Pe clet numbers this resistance is not
signicant and that for zero P eclet number this reduces to the conduction resistance.
We have given some examples on how a uid (gas or liquid) can be heated or
cooled (by energy conversion) to create hot or cold streams for subsequent heat
transfer usages. These energy conversions are represented with internodal energy
conversion. In the premixed gaseous combustion, we have shown how the adiabatic
ame speed u
f .1
depends on the thermophysical and thermochemical properties.
We have discussed Joule heating of gases (to create thermal plasmas). The heat
losses from the hot streams to the bounding surfaces, by surfaceconvection and gas
radiation emission, have also been briey discussed.
The hot or cold streams may be used in surfaceconvection heat transfer, to
heat or cool a solid or indirectly heat or cool another uid stream (as in direct and
indirect heat exchangers). We will discuss these surfaceconvection heat transfers in
Chapters 6 and 7.
5.8 References
[1] Berocci, R., Bonizzoni, G., and Sindoni, E., Editors, 1996, Thermal Plasmas for
Hazardous Waste Treatment, World Scientic, Singapore.
[2] Boulos, M. I., 1991, Thermal Plasma Processing, IEEETransactions onPlasma
Science, Volume 19, pp. 10781089.
[3] Etemadi, K., and Mostaghimi, J., Editors, 1991, Heat Transfer in Thermal Plasma
Processing, ASME HTDVol. 161, American Society of Mechanical Engineers,
New York.
MassoudKaviany 2011
[4] Glasman, I., 1977, Combustion, Academic Press, New York.
[5] Kaviany, M., 1986, Heat Withdrawal by Evaporation from Container Surface,
in Heat Transfer in Air Conditioning and Refrigeration Equipment, Kohler, J.A.
and Lee, J.W.B., Editors, ASME HTDVol. 65, 6165, ASME, New York.
[6] Kaviany, M., 1999, Principles of Heat Transfer in Porous Media, Corrected Sec
ond Edition, SpringerVerlag, New York.
[7] Konuma, M., 1992, Film Deposition by Plasma Techniques, SpringerVerlag,
Berlin.
[8] Kuo, K. K.Y., 1986, Principles of Combustion, Wiley, New York.
[9] Lancaster, J.F., 1986, The Physics of Welding. Pergamon Press, Oxford.
[10] Peters, N., and Rogg, B., Editors, 1993, Reduced Kinetic Mechanisms for Appli
cations in Combustion Systems, SpringerVerlag, Berlin.
[11] Read, R.J., 1984, North America Combustion Handbook, Third Edition, North
America Manufacturing Company, Cleveland.
[12] Turns, S.R., 1996, An Introduction to Combustion, McGrawHill, New York.
Also, Borman, G.L., and Ragland, K.W., 1998, Combustion Engineering,
McGrawHill, New York. Also, Shepherd, I.G., and Cheng, R.K., 2001, The
Burning Rate of Premixed Flames in Moderate and Intense Turbulence, Com
bustion and Flame, Volume 127, pp. 20662075.
[13] Westbrook, C.K., and Dryer, F.L., 1981, Simplied Reaction Mechanisms for
the Oxidation of Hydrocarbon Fuels in Flames, Combustion Science and Tech
nology, Volume 27, pp. 8143.
5.9 Problems
5.9.1 ConductionConvection Resistance
PROBLEM 5.1. FAM
In order to protect exhaust line walls from exposure to hightemperature exhaust
gases, these walls are covered by a sacricial layer. This is shown in Figure Pr. 5.1.
Upon exposure to high temperature exhaust gas and a rise in temperature, this sacri
cial layer undergoes a pyrolytic thermal degradation, produces pyrolytic gases, and
x
Q
1
Q
s
Q
s
Q
u
Q(x = 0)
T
f,1
T
f,2
Gas Evolved
by Pyrolysis of
Sacrificing Layer
Exhaust Gas
FlueGas
Stream
S
r,c
Sacrificing Layer
(a FiberFiller
Composite)
u
f
D
L
l
k
Control Surface A
1
Combustion
Chamber
Figure Pr. 5.1. Exhaust line walls covered by sacricial layers.
MassoudKaviany 2011
becomes porous. The pyrolytic gas ows toward the heated surface, thus providing
for transpiration cooling and prevention of the large heat load Q
s
from reaching the
wall. Treat the pyrolytic gas as air at T = 600 K and assume a steadystate gas ow
that is uniform through the layer. Assume that the area for conductionconvection
is A
k.u
= Dl and use the planar presentation of the resistance as given by (5.14).
(a) Draw the thermal circuit diagram and show the energy equation for surface node
T
f .1
.
(b) For the conditions given below, determine the rate of heat owing into the
wall Q
1
.
T
f .1
= 300 K, T
f .2
= 900 K, k) = 0.5 W/mK, u
f
= 50 cm/s, D= 80 cm, l = 1 m,
L = 1.5 cm.
PROBLEM 5.2. FAM
The axial conductionconvection resistance may be large, when compared to other
heat transfer resistances. In ow through a tube, as shown in Figure Pr. 5.2, the
axial conductionconvection resistance R
k.u
is compared to the average convection
resistance R
u
)
L
(this will be discussed in Chapter 7). The average velocity of the
uid owing in the tube is u
f
and the average uid inlet and outlet temperatures are
T
f .1
and T
f .2
.
(a) For the conditions given below determine the ratio of R
k.u
,R
ku
when the uid
is air.
(b) Determine the ratio of R
k.u
,R
ku
when the uid is engine oil.
The average convection resistance is given (for the case of small NTU, to be
discussed in Chapter 7) by
R
u
)
L
=
1
3.66Lk
f
.
u
f
= 0.2 m/s, D= 5 cm, and L = 30 cm.
Evaluate the properties for air at T = 350 K from Table C.22, and for engine oil
at T = 350 K from Table C.23.
L
D
R
k,u
R
u L
T
s
T
f,2
T
f,1
Average
Convection Resistance
Axial Conduction
Convection Resistance
u
f
A
k,u
= D
2
/4
Fluid
Stream
Figure Pr. 5.2. Comparison of lateral (surfaceconvective) and axial (conductionconvective)
resistances.
MassoudKaviany 2011
PROBLEM 5.3. FAM
Transpiration surface cooling refers to owing a uid through a permeable solid
toward the surface to intercept and remove a large amount of heat owing to the
surface. This imposed heat input Q
s
is called the heat load. The owing uid opposes
the axial conduction heat transfer and results in a lower surface temperature,
compared to that of conduction heat transfer only (i.e., no permeation). This is
shown in Figure Pr. 5.3. Air is made to ow through a porous ceramic slab to protect
a medium (a substrate) beneath the ceramic.
(a) Draw the thermal circuit diagram.
(b) For the conditions given below, determine the surface temperature T
f .2
.
(c) For comparison, determine T
f .2
using u
f
= 0, i.e., (5.15).
Q
s
= 10
3
W, u
f
= 10 cm/s, T
f .1
= 20
C, T
f .2
= 100
C.
For the gas use the properties of air at T = 900
C, k) = 20 W/mK.
Evaluate the water properties at T = 350 K.
5.9.2 Evaporation Cooling of Gaseous Streams
PROBLEM 5.8. FUN
Capillary pumping (or wicking) refers to ow of liquid through and toward the
porous solids by the force of surface tension (an intermolecular force imbalance at
the liquidgas interface). In capillary pumped evaporators, heat is also provided to
the porous solid surface such that the liquid is completely evaporated on the surface.
Figure Pr. 5.8(a) shows three capillarypumped evaporators, distinguished by the
relative direction of the heat and liquid ows. These are used in heat pipes (to be
discussed in Example 8.1) and in enhanced, surface evaporations.
Figure Pr. 5.8(b) renders a counter heatwater capillary evaporator, where the
liquid ow rate
M
l
owing through the wick(distributed as attached, permeable
cylinders) is evaporated at location L
1
. Assume that the liquid is at T
f .1
= T
f .2
= T
lg
at x = L
1
, such that (Q
k.u
)
12
= 0. The temperature at T(x = L
1
) is the saturation
temperature, so the heat for evaporation is provided by conduction in the region
adjacent to the surface, i.e., L
1
x L
1
L
2
. Then Q
k.23
is determined from
Table 3.2. This simple thermal circuit model is also shown in Figure Pr. 5.8(b).
Liquid Wick (Wet
Porous Solid)
Vapor
Vapor
Liquid
Liquid
Substrate
g
g
S
lg
q
k
(i) Counter Heat
Liquid Flow
(ii) Cross Heat
Liquid Flow
(iii) CounterCross
HeatLiquid Flow
q
k
q
k
CapillaryPumped Evaporators
S
lg
Vapor
S
lg
Substrate
Substrate
Figure Pr. 5.8. (a) Various capillary pumped evaporators, based on the relative direction of
the heat and liquid ows.
MassoudKaviany 2011
L
1
x
g
L
2
T
3
S
lg
, T
f,2
= T
lg
T
f,2
= T
lg
Vapor
Water Flow
M
l
, T
f,1
= T
f,2
Wick (Wet,
Porous Solid)
(ii) (i)
Surface Evaporation
S
e,J
M
l
q
k,23
Counter HeatLiquid Flow
Capillary Evaporator
Thermal Circuit
Diagram
Q
k,23
R
k,23
T
3
S
lg
Substrate
Figure Pr. 5.8. (b) A counter heatliquid ow capillary evaporator with a simple conduction
heat transfer model for the region adjacent to surface L
1
x L
1
L
2
.
(a) Determine the liquid mass ow rates
M
l
and m
l
=
M
l
,A
k
, for the following
conditions.
(b) Comment on the practical limit for the reduction of L
2
.
R = 0.5 mm, L
2
= 150 jm, T
lg
= 100
C, T
3
= 105
C, k) = 10 W/mK.
Determine saturated water properties from table C.27, at T = 373.15 K.
PROBLEM 5.9. FAM
In order to protect a substrate from high temperatures resulting from intense
irradiation, evaporation transpiration cooling is used. This is shown in Figure Pr. 5.9.
Liquid water is supplied under a porous layer and this liquid is evaporated by
the heat reaching the liquid surface, which is at temperature T
lg
= T
f .1
. The heat
ow to the liquid surface is only a fraction of the prescribed irradiation, because
the watervapor owintercepts and carries away a fraction of this heat by convection.
(a) For the condition given in Figure Pr. 5.9, draw the thermal circuit diagram.
(b) Determine
M
lg
and T
f .2
.
k
Makeup
Water
Stream
Porous Layer
Physical Model: Evaporation Transpiration Cooling
Water
Vapor
Stream
Heat Returned
by Convection
Heat Used for
Evaporation
Permeable, Sintered
Metal Particles
Substrate
Q
1
= 0
x
S
lg
= M
lg
h
lg
(1
r,2
) q
r,i A
k,u
A
k,u
r,2
q
r,i
q
r,i
l
L
w
T
f,2
T
lg
= T
f,1
Figure Pr. 5.9. Evaporation transpiration cooling.
MassoudKaviany 2011
f
= 0.596 kg/m
3
, c
p.f
= 2,029 J/kgK, Lh
lg
= 2.257 10
6
J/kg (water at 100
C),
k) = 15 W/mK, L = 40 cm, n = 15 cm, L = 1.5 cm, T
lg
= T
f .1
= 100
C,
r.2
= 0.9,
q
r.i
= 10
5
W/m
2
,
l
= 958 kg/m
3
.
Note that from conservation of mass across the liquid surface,
f
u
f
=
l
u
l
.
PROBLEM 5.10. FAM
Anesthetic drugs are supplied as liquid and are evaporated, mixed with gases (such
as oxygen), and heated in portable vaporizer units for delivery to patients. This
is shown in Figure Pr. 5.10. The drugs, such as enuorane, isouorane, etc., have
thermophysical properties similar to that of refrigerant R134a (Table C.28). The
drug is sprayed into the vaporizer. The heating of the gas mixture (drug and oxygen)
is by surface convection and here it is prescribed by Q
ku
)
L
. This heat in turn is
provided by Joule heating from a heater wrapped around the tube. Assume that
droplets evaporate completely.
(a) Draw the thermal circuit diagram for the control volume V.
(b) Determine the exit uid stream temperature T
f .2
, for the conditions given below.
T
f .1
= 15
C,
M
l
= 2 10
5
kg/s,
M
O
2
= 2 10
4
kg/s, Q
ku
)
L
= 6 W.
Use the specic heat capacity of oxygen (at T = 300 K, Table C.22) for the
mixture, and use Lh
lg
from Table C.28, at p = 1 atm.
L
S
lg
, Evaporation
Warm Gas Mixture
Stream, T
f,2
Surface Convection
Droplets
Liquid Nozzle
(Injector)
Oxygen Stream
M
O
2
, T
O
2
= T
f,1
Liquid Anesthetic
Drug Stream
M
l
, T
l
= T
f,1
Control Volume, V
Control Surface, A
To Patient
Q
ku L
Figure Pr. 5.10. A liquid anesthetic drug is evaporated, mixed with oxygen, and heated in a
vaporizer. The drug is sprayed into the vaporizer.
PROBLEM 5.11. FAM
In surface evaporation from permeable membranes, the heat for evaporation is
partly provided by the ambient gas (by surface convection) and partly by the liquid
reservoir (through the conductionconvection heat transfer through the membrane).
This is shown in Figure Pr. 5.11. The ambient gas may contain species other than
the vapor produced by the evaporation. These other species are called the inert or
noncondensables and provide a resistance to the vapor mass transfer. This will be
discussed in Section 6.9 and here we do not address the mass transfer resistance and
assume that the gas is made of the vapor only. Consider using superheated steam to
evaporate water froma permeable membrane. We assume that the gas is moving and
has a fareld temperature T
f .
and that there is a surfaceconvection heat transfer
MassoudKaviany 2011
M
l
M
l
M
l
Gas Stream
T
f,
, u
f,
Liquid
(Water)
Permeable
Membrane
L
k
T
f,
R
ku,2
(R
ku
)
12
(Q
k,u
)
12
Q
ku,2
T
lg
= T
f,2
A
k,u
= A
ku
S
lg
T
f,1
Figure Pr. 5.11. Water supplied through a permeable membrane is evaporated on the gasside
interface. The thermal circuit diagram is also shown.
resistance R
ku.2
between the surface and the gas stream. This is also shown in
Figure Pr. 5.11.
The surface temperature T
f .2
is equal to the saturation temperature T
lg
(p
g
).
Also, since (R
k.u
)
12
depends on
M
l
, the liquid mass ow is determined such that it
simultaneously satises (Q
k.u
)
12
and
S
lg
.
For the following conditions, determine
M
l
.
T
f .1
= 100
C, T
f .
= 110
C, T
lg
= 95
C, A
k.a
= A
ku
= 1 m
2
, R
ku.2
= 0.25
K/W, k) = 1 W/mK, L = 1 cm.
Determine the water properties at T = 373.15 K, from Table C.27.
5.9.3 Combustion Heating of Gaseous Streams
PROBLEM 5.12. FAM
Consider an adiabatic methaneair ame in a packedbed of spherical alumina par
ticles of diameter D= 1 mm and a bed porosity c = 0.4. Assume the average tem
perature of the bed to be T = 1,300 K. The inlet conditions are T
f .1
= 289 K and
p
1
= 1 atm and the mixture is stoichiometric.
(a) Determine the effective thermal conductivity for the bed. Use the radiant (pho
ton) conductivity correlation for spheres given by [6]
k
r
) = 4D
SB
T
3
F
r
= 4D
SB
T
3
_
a
1
c
r
tan
1
_
a
2
c
r
_
k
s
4D
SB
T
3
_
a
3
_
a
4
_
correlation for radiant conductivity for packed bed of particle with 0.4 c 0.6,
where F
r
is the radiant exchange factor, a
1
= 0.5756, a
2
= 1.5353, a
3
= 0.8011, and
a
4
= 0.1843.
(b) Determine the effective radiant conductivity.
(c) Using the sum of these conductivities, at the above average temperature, deter
mine the adiabatic ame speed.
Use the results of Example 5.4, as needed.
MassoudKaviany 2011
PROBLEM 5.13. FAM
To achieve the same ame speed that was obtained using the porous medium in
Problem Pr.5.12, turbulent ow may be used. The turbulent intensity affects the
ame speed. Using the same adiabatic, stoichiometric methaneair ame, determine
the needed turbulent intensity Tu to achieve the same ame speed.
The laminar ame speed is u
f .1
= 0.4109 m/s and the packedbed ame speed is
1.30 m/s.
PROBLEM 5.14. FUN
In a premixed fueloxidant stream, as the fuel is oxidized, its density
F
decreases and
this decrease inuences the combustion and chemical kinetics. Consider a rstorder
chemicalkinetic model for reaction of methane and oxygen given by
n
r.F
= a
r
F
e
LE
a
,R
g
T
f
.
Start with the fuelspecies conservation equation (B.51). Assume onedimensional,
incompressible ow, and negligible mass diffusion.
Derive the relation for the fuel density as it undergoes reaction over a length L
with a velocity u
f
, inlet density
F.0
, and temperature T
f .0
.
Use a constant, average temperature T
f
= (T
f .0
T
f .L
),2 to represent the aver
age temperature over length L. The nal expression is
F.L
=
F.0
e
(a
r
L,u
f
)e
LEa,Rg T
f
.
PROBLEM 5.15. FAM
To estimate the ame speed of the atmospheric (p
1
= 1 atm) gasoline (assume it
is octane) reaction with air (premixed), assume that the chemical kinetic model for
this reaction can be approximated as being zeroth order and represented by a pre
exponential factor a
r
andanactivationenergy LE
a
giveninTable 5.3 for zerothorder
reaction of methane and air. The reaction is represented by
C
8
H
18
12.5O
2
47.0N
2
8CO
2
9H
2
O47.0N
2
.
Determine (a) the Zeldovichnumber Ze, and(b) the ame speedu
f .1
, for a premixed
gasolineair ame.
Use Figure 5.9 for the adiabatic ame temperature, Table 5.2 for the heat of
reaction, and evaluate the properties of air at the average ame temperature T
f
)
=
(T
f .1
T
f .2
),2. Determine the density
f .1
from the idealgas law using T
f .1
= 16
C.
PROBLEM 5.16. FUN
In order to achieve higher ame temperatures, pure oxygen is used instead of
air in burning hydrocarbons. Consider stoichiometric methaneoxygen premixed
combustion.
(a) Determine the adiabatic ame temperature T
f .2
.
(b) Determine the laminar, adiabatic ame speed u
f .1
.
MassoudKaviany 2011
(c) Compare this ame speed with the laminar, adiabatic ame speed of methaneair
in Example 5.4 (i.e., u
f .1
= 0.4109 m/s)
T
f .1
= 16
C, c
p.f
= 3,800 J/kgK.
Use the chemical kinetic constants for the zerothorder reaction given in
Table 5.3. Use the thermal conductivity of air at the average gas temperature
T
f
)
= (T
f .1
T
f .2
),2 for the mixture.
PROBLEM 5.17. FAM.S
Surfaceradiation drying of wet pulp in paper production uses permeable ceramic
foams for both combustion and surface emission. This is shown Figure Pr. 5.17.
The premixed, gaseous fuelair ows into the foam, and after an initial ignition, it
undergoes steady combustion. The ue gas heats the foam and leaves while the foam
radiates to the load (wet paper). The steady combustion requires a mixture ow
rate that in turn is determined by the heat transfer. This ow rate, or specically the
velocity u
f .1
, can be several times the laminar, adiabatic ame speed given by (5.55).
Consider stoichiometric methaneair combustion with the mixture arriving at T
f .1
and at 1 atm pressure.
(a) Draw the thermal circuit diagram.
(b) Assuming local thermal equilibrium between the gas and the permeable solid
(i.e., the ceramic foam), determine the radiantsurface temperature T
s
= T
f .2
and
the radiation heat transfer rate Q
r.2p
.
(c) What is the radiant heat transfer efciency of this heater?
Assume that F
2p
= 1 and both the radiantburner surface and the wet paper
are blackbody surfaces.
u
f .1
= 0.2 m/s, T
f .1
= 16
C, a = 25 cm, n = 60 cm, T
p
= 60
C, c
p.f
= 1,611 J/kg
K,
f .1
= 1.164 kg/m
3
, (
F
,
f
)
1
= 0.05519.
T
f,2
= T
s
Pulp, T
p
S
r,c Occurring at
Top of Heater
Q
u,2
Q
r, 2p
Q
u,1
a
w
Paper Drying System
Premixed
MethaneAir
Stream, T
f,1
, u
f,1
Figure Pr. 5.17. A surfaceradiation burner used for drying a wetpulp sheet.
MassoudKaviany 2011
PROBLEM 5.18. FUN
In burning gaseous fuel in a tube or in a porous medium, it is possible to locally
create gas temperatures above the adiabatic ame temperature. This is done by
conduction of heat through the bounding solid (tube or solid matrix) from the high
temperature region to the lower temperature region. This is called heat recirculation
and the process of combustion with this local increase in the gas temperature is called
the superadiabatic combustion. This is rendered for a premixed fuel (methane) and
oxidant (air) in Figure Pr. 5.18. In this idealized rendering, three different regions
are identied, namely the gaspreheat, combustion, and solidheating regions. The
heat recirculation begins as surface convection from the ue gas to the solid, then
it is conducted along the solid (owing opposite to the gas ow), and is nally
returned to the gas (premixed fueloxidant) by surface convection. The idealized gas
temperature distribution (for the three regions) is also shown in the gure and can be
represented by the internodal energy conservation equation and the temperatures
labeled T
f .1
to T
f .4
. The surface convection out of the solidheating region is given
per unit gas ow crosssectional area, i.e., Q
ku
,A
u
.
(a) Draw the thermal circuit diagram.
Determine (b) T
f .2
, (c) T
f .3
, (d) T
f .4
, and (e) the excess temperature LT
excess
=
T
f .4
T
f .3
for stoichiometric, premixed methaneair combustion.
(f) Comment on the effect of using temperaturedependent specic heat capacity on
the predicted excess temperature.
Q
ku
,A
u
= 5 10
4
W/m
2
, T
f .1
= 20
C.
S
r,c
T
f,1
T
f,2
T
f,3
T
f,4
T
f,3
T
f,2
T
f,1
x
T
f
T
excess
T
f,4
Tube
Wall
(i) Local Superadiabatic Combustion Heat Circulation
Through Axial Tube Conduction
(ii) Ideal Axial Temperature Distribution
Premixed
MethaneAir
Stream
Gas
Preheat
Region
Heat Recirculation
Solid
Heating
Region
Com
bustion
Region
A
u
Q
ku
= Q
k,s
Q
ku
= Q
k,s
(Mc
p
)
f
T
combustion
Figure Pr. 5.18. Heat recirculation and local superadiabatic temperature in combustion in a
tube.
MassoudKaviany 2011
Assume negligible heat loss in the combustion region (i.e., an adiabatic combus
tion region). Use the heat of combustion of methane Lh
r.F
and the stoichiometric
mass fraction of the fuel from Table C.21(a), and a constant specic heat capacity
corresponding to air at T = 1,500 K.
PROBLEM 5.19. FAM
Consider nonadiabatic (i.e., lateral heat losses Q
loss
not being negligible), one
dimensional ow and reaction of premixed methaneair gaseous mixture. There
is a heat loss, per unit ow crosssectional area, Q
loss
,A
u
= q
loss
= 10
5
W/m
2
. The
mixture is stoichiometric and the initial temperature and pressure are T
f .1
= 298 K
and p
1
= 1 atm.
Determine the nal temperature T
f .2
.
Use the Table C.21(a) data for adiabatic, stoichiometric ame (
F
,
f
, Lh
r.F
, and
u
f .1
), and for the average specic heat use c
p.f
= 1,611 J/kgK.
PROBLEM 5.20. FAM.S
The adiabatic ame temperature and speed are referred to the condition of no lat
eral heat losses (Q
loss
= 0) in combustion of a unidirectional premixed fueloxidant
stream. The presence of such losses or gains decreases the ame temperature T
f .2
,
given by (5.35), and also the ame speed given by (5.55). This can continue until
the ame temperature decreases below a threshold temperature required to sustain
ignition and combustion.
(a) Consider a laminar, stoichiometric premixed methane air combustion. For the
conditions given below, plot the ame temperature T
f .2
and the ame u
f .1
speed for
the given range of Q
loss
.
(b) Comment on the quenching of the ame as Q
loss
increases and T
f .2
decreases.
T
f .1
= 16
C,
f .1
= 1.164 kg/m
3
, (
F
,
f
)
1
= 0.05519, c
p.f
= 1,611 J/kgK, k
f
=
0.07939 W/mK, A
u
= 10
2
m
2
, 0 Q
loss
800 W.
Use Tables 5.2 and 5.3 for Lh
r.F
, a
r
, and LE
a
.
PROBLEM 5.21. FAM
A surfaceradiation burner, which uses distributed, direct fuel supply, is shown in
Figure Pr. 5.21. The radiation is from the impermeable surface having a uniform
temperature T
s
and facing the radiationload surface at temperature T
L
. The oxidant
stream (air) is at mass ow rate
M
O
and temperature T
f .1
.
The fuel stream (methane) is divided into three smaller streams with each ow
rate designated as
M
F.i
. i = 1. 2. 3. The product stream
M
P
leaves at the exit port
at temperature T
f .4
. Here we assume that T
f .4
= T
f .3
= T
f .2
= T
s
. This is a design
requirement that in practice is obtained by using more than three fuel stream ports
and by taking the pressure drop in the fuel membrane and the combustion chamber
into the account.
(a) Draw the thermal circuit diagram.
(b) Determine the surface temperature T
s
.
MassoudKaviany 2011
Distributed, Direct Fuel Stream, T
f,1
Radiation Load, Q
r,sL
w
L
M
P
= M
F,4
T
f,3
= T
s
T
f,4
= T
s
T
L
,
r,L
T
s
,
r,s
S
r,c
S
r,c
S
r,c
M
F,3
L/3
M
F,1
Oxidant Stream
M
O
, T
f,1
PowderFilled
Combustion Chamber
Product Stream
Permeable Fuel
Supply Membrane
Impermeable
Radiation Surface
RadiationLoad Surface
M
F,2
T
f,2
= T
s
Figure Pr. 5.21. A distributed, direct fuel supply surfaceradiation burner. The radiation sur
face is impermeable.
(c) Determine the burner efciency = Q
r.sL
,
S
r.c
,
S
r.c
= Lh
r.F
i
M
F.i
.
c
r.s
= 0.9, c
r.L
= 1, F
sL
= 1, T
L
= 700 K, T
f .1
= 300 K,
M
O
= 0.013 kg/s,
M
F.1
=
3 10
4
kg/s,
M
F.2
=
M
F.3
= 2 10
4
kg/s.
Use an integralvolume energy equation for each of the three segments. Assume
a constant specic heat capacity c
p.f
= 1,600 J/kgK.
5.9.4 Joule Heating of Gaseous Streams
PROBLEM 5.22. FAM
A watercooled thermal plasma generator, used for spray coating and shown in
Figure Pr. 5.22, is to produce an argon gas plasma stream with an average exit
temperature of T
f .2
= 5,000 K. The Joule heating is by direct current (dc) and uses
J
e
= 200 A, and L = 150 V provided by a power supply. The heat transfer to the
Thermal Plasma
(Argon Gas)
Particle Coating
Cathode ()
Argon Gas
Spray Coating Using Thermal Plasma
to Melt Entrainned Particles
Coolant
(Water)
Particle
Stream
Anode (+)
Substrate
M
f
, T
f,1
S
e,J
u
s
Q
loss
Q
ku
Plasma Stream
Figure Pr. 5.22. Generation of an argon plasma stream using the Joule heating.
MassoudKaviany 2011
water coolant is Q
ku
and other heat losses are given by Q
loss
. Assume
S
ij
= 0.
(a) Draw the thermal circuit diagram.
(b) Determine the argon gas ow rate
M
f
.
Q
ku
= 5 kW, Q
loss
= 3 kW, T
f .1
= 300 K.
Use a constant specic heat capacity for the ionized argon and use a value equal
to ve times that for argon at T = 1,500 K.
PROBLEM 5.23. FAM
An acetyleneoxygen torch has 70% by weight of its stoichiometric oxygen provided
by a pressurized tank and this is called the primary oxygen. Due to the fast chemical
reactionof C
2
H
2
andO
2
andthe fast diffusionof O
2
, the remaining 30%of the oxygen
is provided by entraining the ambient air and this is called the secondary oxygen.
These are shown in Figure Pr. 5.23. Along with the entrained oxygen, nitrogen is also
entrained and this inert gas tends to lower the nal temperature T
f .2
.
The products of combustion (i.e., the ue gas) ows over a surface to be welded.
Thena fractionof the sensible heat is transferredtothe surface by surfaceconvection
heat transfer Q
ku
, and the remainder ows with the gas as Q
u.2
.
T
f,1
T
f,2
M
f
Q
u,2
Q
u
Q
u
M
O
2
, M
N
2
Entrained
(Secondary)
Oxygen
Flame
Acetylene Fuel
Primary Oxygen
Torch
Surface Convection, Q
ku
Intramedium
Convection
S
r,c
M
O
2
M
C
2
H
2
Plasma
Stream
Figure Pr. 5.23. An acetyleneoxygen torch used for welding of a surface. The oxygen is
provided by a tank and by entraining air.
(a) Draw the thermal circuit diagram for this combustion and heat transfer.
(b) For the conditions given below, determine the ue gas temperature T
f .2
.
T
f .1
= 20
C, p
1
= 1 atm, Q
ku
= 0.
Use a constant and uniform specic heat of c
p.f
= 3,800 J/kgK.
PROBLEM 5.24. FAM
The acetyleneoxygen, amecutting torch is used with low carbon and low alloy
irons. A simple mixer that requires pressurized oxygen and acetylene is shown in
MassoudKaviany 2011
T
f,1
T
f,2
Q
u,1
Q
u,2
Q
3
T
3
O
2
, For Acetylene and Iron
AcetyleneOxygen Cutting Torch
with Extra Heat Generated by
Oxidation of Iron Workpiece
Q
ku
Q
ku
(S
r,c
)
12
C
2
H
2
, Acetylene
(S
r,c
)
3
3Fe + 2O
2
Fe
3
O
4
C
2
H
2
+ O
2
2CO
2
+ H
2
O
2
Q
32
Figure Pr. 5.24. A cutting torch with reacting workpiece.
Figure Pr. 5.24. In addition to the heat provided by the reaction
C
2
H
2
2.5O
2
2CO
2
H
2
O.
the excess oxygen provided by the torch reacts with the iron and releases heat. One
of the reactions is
3Fe 2O
2
Fe
3
O
4
.
This reaction is highly exothermic, i.e., Lh
r.F
= 6.692 10
6
J/kg of Fe. Fig
ure Pr. 5.24 shows this iron oxidation. The rate of this reaction is controlled by
the speed of the torch moving on the cutting surface, and with other variables.
Assume that this reaction will add an extra energy conversion such that we can
approximate the contribution for this iron oxidation by adding 30% to the heat of
reaction of C
2
H
2
. We also model the excess oxygen by adding 40%, by weight, to
the stoichiometric oxygen needed to burn the acetylene.
(a) Draw the thermal circuit diagram.
(b) Determine the combustion product gas temperature T
f .2
.
T
f .1
= 20
C, c
p.f
= 3,800 J/kgK.
5.9.5 GasStream Volumetric Radiation
PROBLEM 5.25. FAM
The combustion product gas from a replace chamber enters its chimney at a ow
rate
M
f
and a temperature T
f .1
. This is shown in Figure Pr. 5.25. Assume that the
surfaceconvection heat transfer to the chimney wall is negligible.
(a) Show that the surface radiation emitted by the chimney wall is negligible
compared to that emitted by the combustion product gas, under the following
conditions.
MassoudKaviany 2011
Chimney
Burning Wood Logs
Air Supply
M
f
, Flue
Gas Stream
Fireplace
Air Stream
S
e,
+ S
e,
T
s
,
r,s
q
u
T
f,1
T
f,2
L
R
Figure Pr. 5.25. A chimney bounding the combustion product gas stream exiting a replace
chamber.
(b) Determine the combustion product gas stream exit temperature T
f .2
, for the
these conditions.
T
s
= 120
C, T
f .1
= 600
C,
M
f
= 2.5 10
3
kg/s, R = 15 cm, L = 5 m, c
r.f
= 0.1
(for large soot concentration).
Use c
p.f
for air at T = 600 K. Use c
r.s
=
r.s
for reclay brick.
MassoudKaviany 2011
6
Convection: SemiBounded Fluid Streams
6.3 Parallel, Turbulent Flow: Transition Reynolds
Number Re
L.t
6.3.1 Turbulent Convection Heat Flux
As with (6.54) and (6.55), the convection heat ux vector is similarly averaged, and
we have
q
u
= q
u.m
q
u.t
= (c
p
)
f
u
f
T
f
(c
p
)
f
u
/
f
T
/
f
mean and turbulent
convection heat uxes.
mean convection turbulent convection
heat ux vector q
u
heat ux vector q
t
(6.1)
The turbulent uctuations contribute to the heat transfer in a way similar to
the molecular uctuations discussed in Section 3.2.1. There we used the molecular
velocity uctuations (u
2
m
)
1,2
and through (3.14) showed how the heat was transfered
by these uctuations. Here, the combined effects of the turbulent velocity uctuation
(u
/
f
u
/
f
)
1,2
, the ability of the uid to store/release heat (c
p
)
f
, and the ability of the
uctuations to carry the uid content a short distance before reaching thermal equi
librium also result in heat transfer in the presence of a temperature nonuniformity.
This is shown in Figure 6.11(b).
From (6.56), we write the y component (i.e., perpendicular to the surface) of the
turbulent convection heat ux as
(q
u.t
)
y
= (c
p
)
f
:
/
f
T
/
f
turbulent convection heat ux along y direction. (6.2)
We now proceed to relate :
/
f
T
/
f
to mean quantities that are more readily mea
sured or predicted.
MassoudKaviany 2011
6.3.2 Microscale Turbulent Convection: Turbulent Conductivity k
t
and Turbulent Mixing Length
t
Using an analogy with the conduction in gases (due to molecular uctuations), as
given by (3.14), the heat transfer due to turbulent uctuations is expressed as
(q
u.t
)
y
= (c
p
)
f
:
/
f
T
/
f
Pr
1
t
(c
p
)
f
(:
/2
f
)
1,2
t
dT
f
dy
turbulent convection heat
ux along y direction,
(6.3)
where
t
is the turbulent meanfree path and is also called the turbulent mixing
length [shown in Figure 6.11(b)]. This is a representation of the distance traveled by
a uid particle before reaching equilibrium (either thermal or mechanical). It varies
with the distance from the bounding surface. The constant Pr
t
(called the turbulent
Prandtl number) is determined empirically and is near unity. Here (:
/2
f
)
1,2
is the root
mean square of the turbulent velocity uctuation in the y direction.
Since in (6.58) the turbulent heat transfer is related to the gradient of the mean
temperature, a turbulent thermal conductivity k
t
is dened by
q
t.y
k
t
dT
f
dy
turbulent convection heat ux representeed
with turbulent conductivity.
(6.4)
Then from (6.59) and (6.58), k
t
is found as
k
t
Pr
1
t
(c
p
)
f
(:
/2
f
)
1,2
t
denition of turbulent thermal conductivity. (6.5)
This is similar to (3.16).
Next (:
f
/2
)
1,2
is related to the gradient of the mean velocity along the surface,
i.e., [u
f
,y[. This is because, from the examination of the momentum equation,
we nd that the magnitude of the uctuations increases as the velocity gradient
becomes larger. Next we use the turbulent mixing length
t
again and have (:
f
/2
)
1,2
t
[u
f
,y[. Then (6.60) becomes
k
t
= Pr
1
t
(c
p
)
f
2
t
u
f
y
.
t
=
k
t
(c
p
)
f
turbulent conductivity
and diffusivity.
(6.6)
where
t
is the turbulent thermal diffusivity (or eddy diffusivity). Note that Pr
t
=
t
,
t
, where
t
is the turbulent kinematic viscosity.
We can extend the results to the threedimensional heat ow. The turbulent
convection heat ux vector (c
p
)
f
u
/
f
T
/
f
is proportional to the gradient of T
f
, i.e.,
T
f
. Therefore, as T
f
becomes larger, (c
p
)
f
u
/
f
T
/
f
increases. Then, assuming that
k
t
is isotropic (i.e., the same in all directions), from (6.58) the vectorial form of q
t
is
q
u.t
= (c
p
)
f
u
/
f
T
/
f
k
t
T
f
.
turbulent convection heat ux vector
is related to gradient of mean temperature.
(6.7)
We note that in general k
t
is anisotropic (and should be represented by a tensor).
MassoudKaviany 2011
6.3.3 Variation of Turbulent Mixing Length Near Solid Surface
For fullyturbulent ows at high Reynolds numbers, k
t
can be larger than k (the
molecular thermal conductivity) at some locations near the solid surface. Also, k
t
vanishes in the laminar sublayer. Therefore, k
t
(or u
/
f
and
t
) varies with the location
from the surface and enhances the convection heat transfer. For parallel ows, this
enhancement is noticeable in the direction perpendicular to the ow (y direction),
where the mean velocity :
f
is small and the heat transfer is otherwise dominated by
molecular conduction.
Empirical results show that for the ow over a semiinnite at plate, the vari
ation of the mixing length in the turbulent boundary layer, adjacent to surface, is
approximated as [9]
t
=
_
_
0. 0 y
l.i
y.
l.i
y 0.21
0.086
. 0.21
y 
, above which
t
is proportional to the
boundary layer thickness. Outside the boundary layer (in the freestream) there are
no turbulent uctuations and
t
= 0. As will be discussed in the following section,
.
The distribution of the mean, xdirection velocity u
f
is generally determined
from empirical relations that include the dependence on y and Re
L
[9].
Figures 6.11(c) and (d) give a rendering of this enhanced heat transfer. The
uctuations in the velocity and in the temperature are characterized by the turbulent
eddies (i.e., small vortices with a large range of sizes and frequencies). The turbulent
mixing length
t
is also shown.
6.6 LiquidGas Phase Change
6.6.1 Theoretical Maximum Heat Flux for Evaporation and
Condensation q
max
(Kinetic Limit)
The continuous vapor (or condensate) production rate is controlled by the rate of
heat delivered (or removed) to the liquidvapor interface and by the rate of vapor (or
condensate) removal. In evaporation in a pool of liquid (i.e., no net liquid motion), a
theoretical kineticlimit maximumin q
ku
is observed at a particular surface superheat
T
s
T
lg
. Pool boiling will be discussed in Section 6.6.2. This maximum (called the
critical heat ux) q
ku.CHF
is determined by the heat delivery (heat transfer limit)
and the liquid supply and vapor removal (uid dynamics limit) limitations. When
MassoudKaviany 2011
x
Thermobuoyant
Liquid Motion
BubbleInduced
Liquid Motion
T
s,1
T
s
Heated Substrate
k
s
Pool of
Liquid
T
l
, u
l
Apex
Condensation
Evaporation
Microlayer
Evaporation
Vapor Heating
Peripheral
Evaporation
Latent Heat
Transport
Vapor
Scape
Vapor Scape
Departing
Bubble
Vapor
Bubble
Nucleation
Surface
Convection
Due to
Makeup
Liquid
Vapor
Surface
Evaporation
Fractions of SurfaceConvection Heat Flux Used in Surface Evaporation
and in Sensible Surface Convection
q
ku
T
T
o
T
s
T
lg
T
l
R
g
q
k
l
Figure 6.17. Nucleate poolboiling surface convection and a rendering of the division of the
surfaceconvectionheat ux among the various routes tothe liquidphase andinthe nucleating,
growing, and departing vapor bubbles.
the liquid is not a pure uid (i. e., is multicomponent), then there is also a mass
diffusion limit. We will discuss the mass diffusion limit in relation to the liquid
surface evaporation in Section 6.9.
Since the surfaceconvection heat transfer involves both sensible heating of
liquid and gas (vapor) and evaporation, an approximate division of q
ku
among these
is shown in Figure 6.17 for a pool of liquid under steadystate heating frombelow. The
ow of heat experiences resistances as it ows through the heated solid surface and
towardthe vapor bubble nucleatedonthe surface andalsotowardthe liquid. Because
of the periodic bubble formation and departure, the temperature distribution in the
solid and the adjacent liquid changes rapidly. The convection through the liquid
occurs with the motion inuenced by the thermobuoyancy (unstable liquiddensity
distribution) and by the phasedensity (or bubble) buoyant motion. Among the two
buoyant motions, the contribution of the bubbleinduced motion is expected to be
more signicant. The heat ow toward the vapor is through the solidvapor interface
(expected to be a small fraction of q
ku
) and through the liquidvapor interface. The
heat ow through a thin liquid layer underneath the attached bubble (called the
microlayer) is expected to be large compared to that through the remainder of the
bubble periphery. The heat transfer to the bubble is rendered as a physicalbond
energy conversion, and for the example given in Figure 6.17, is only a fraction of q
ku
.
A generalization of evaporationcondensation under phasedensity ow is given
in Figure 6.18. For evaporation, the solid surface wetting by the liquid is required
for a continuous bubble formation. For an ideally nonwetting liquid, only a vapor
MassoudKaviany 2011
Evaporation
SurfaceConvection Evaporation and Condensation
Melting Limit
(Bounding Surface)
Solidification Limit
(Fluid)
N
onw
etting (Vapor
Film
Evaporation)
N
o
n
w
e
t
t
i
n
g
(
D
r
o
p
l
e
t
C
o
n
d
e
s
a
t
i
o
n
)
W
e
t
t
i
n
g
(
B
u
b
b
l
e
E
v
a
p
o
r
a
t
i
o
n
)
(+)
0
0
()
()
Condensation
(+)
T
s
T
lg
, K
q
k
u
,
W
/
m
2
(q
ku, max
)
c
W
e
t
t
i
n
g
(
L
iq
u
id

F
i
l
m
C
o
n
d
e
n
s
a
tio
n
)
q
ku,CHF
= (q
ku, max
)
e
Figure 6.18. Variation of the solid surface temperature with respect to the surface heat ux
for heat addition to a liquid (T
s
> T
lg
), i.e., evaporation, and heat removal from a vapor
(T
s
 T
lg
), i.e.,condensation. The maximum heat ux added for the nucleate boiling and the
maximum heat ux withdrawn for the dropwise condensation are also shown.
lm is formed (after a threshold surface superheat T
s
T
lg
, which can be small).
For wetting (small contact angle
c
) liquids, a maximum in q
ku
indicates the smallest
combined resistance to the heat delivery (to the liquidvapor interface) and vapor
removal. Boiling will be further discussed in Section 6.6.2. Condensation will be
discussed in Section 6.6.3.
A similar behavior is expected for condensation (shown in Figure 6.18 with neg
ative q
ku
and T
s
T
lg
) and for a nonwetting (large contact angle) liquid a dropwise
condensation (similar to the surface bubble nucleation) occurs.
As the combined resistance to the heat delivery for evaporation (or removal
for condensation) and vapor (or condensate) removal is decreased, q
ku.max.e
(or
[q
ku
[
max.c
) can be increased up to the limit determined by the kinetic theory.
The molecular ux evaporating from a liquid surface is that given in terms of the
molecular properties. Now assuming that the liquid evaporation takes place at an
innite rate and that the vapor molecules are immediately removed (no molecules
returning to the liquidvapor interface,
g.
= 0), the maximum mass ux of the gas
molecule for the liquidgas phase change m
lg
(kg/m
2
s) is given by using the same
molecular ux we used in Section 3.2.1, i.e.,
m
lg.max
=
1
4
mn
g
u
f
) . u
f
) =
_
8
R
g
T
lg
M
_
1,2
. mn
g
=
g
(6.9)
maximum mass ux at the liquid vapor interface.
MassoudKaviany 2011
q
m
a
x
,
W
/
m
2
q
k
u
,
m
a
x
,
i
*
10
10
0.1
1
0.01
0.001
10
3
10
2
10
1
1
Experiment
Condensation
Dense
Gas
LiquidGas
Coexistence
Evaporation
10
9
10
8
0 0.2 0.4 0.6 0.8
R12
(Refrigerant)
Ethanol
Water
Predicted
g,
= 0
1.0
(a)
Theoretical Maximum SurfaceConvection Heat Flux with LiquidGas Phase Change
and Fractions of It Experimentally Achieved
(b)
p
p
c
Not Achieved Yet
p
p
c
Figure 6.19. (a) Variation of the predicted, kineticlimit heat ux for evaporation of water,
ethanol, and Freon12, with respect to the reduced pressure. (b) Variation of the ratio of
the measured maximum heat ux for evaporation and condensation to the predicted limit,
with respect to the reduced pressure. The results are for
g.
= 0. (From Gambill, W.R., and
Lienhard, J.H., reproduced by permission c 1989, ASME.)
The mass ux is the same as that used in (3.13) and the average molecular speed
is given by (1.16) for the determination of the gas thermal conductivity. Then the
heat ux associated with this mass ux is the maximum liquidvapor, phasechange
heat transfer q
max
(W/m
2
) and is
q
max
( s
lg
,A)
max
= m
lg.max
Lh
lg
=
_
R
g
T
lg
2M
_
1,2
g
Lh
lg
maximum liquidvapor
phase change heat
ux, for
g.
= 0.
(6.10)
This assumes a zero fareld vapor density (i.e., vacuum). When there is a
nonzero fareld vapor density, then we have
q
max
= ( s
lg
,A)
max
= m
lg.max
Lh
lg
=
_
R
g
2M
_
1,2
Lh
lg
(
g
T
1,2
lg
g.
T
1,2
lg.
). (6.11)
As the temperature T
lg
increases,
g
Lh
lg
increases and q
max
increases until the
critical point is approached, and then Lh
lg
decreases and is zero at p
c
. The results
for q
max
for water, ethanol, and R12 (also called Freon12, Freon being a trademark
for nonammable uorinated hydrocarbons used in refrigeration cycles) are shown
in Figure 6.19(a) (we will discuss the heat transfer uids in Section 7.6.6). The
MassoudKaviany 2011
Surface Convection with
LiquidGas Phase Change
Continuous LiquidGas
Interface: Layered Phases
Liquid in Contact
with Solid
Gas in Contact
with Solid
Liquid and Gas in Contact
with Solid (Contact Line):
Nonlayered Phases
Surface
Bubbles
Surface
Droplets
Stable
Liquid
Films
Unstable
Liquid
Films
Phase
Density
Buoyant
Flow
Forced
Buoyant
Flow
Phase
Density
Buoyant
Flow
Forced
Buoyant
Flow
Stable
Vapor
Films
Unstable
Vapor
Films
Formation
and Growth
Departure Surface
Formed
Droplets
Impinging
Droplets
Formation
and Growth
Removal
Phase
Density
Buoyant
Flow
Forced
Buoyant
Flow
Phase
Density
Buoyant
Flow
Forced
Buoyant
Flow
Phase
Density
Buoyant
Flow
Forced
Buoyant
Flow
Surface
Adherence
Levitation
Chart 6.6. A classication of surfaceconvection heat transfer with liquidgas phase change,
based on the uid phase or phases in contact with the solid.
results show the variation of q
max
as a function of the reduced pressure p,p
c
. Note
the sudden drop near p,p
c
= 0.9. As is shown, very large heat ux is possible by
delivering the heat directly to the liquidvapor interface and by removing the vapor
once it is formed. This ideal heat transfer (i.e., no thermal resistances) would require
that T
s
= T
lg
= T
g
.
Due to various hydrodynamic and thermal resistances that will be discussed
below, in practice only a fraction of q
max
is obtained. This is designated as q
ku.max
.
For evaporation this is shown as q
ku.max.e
and for condensation q
ku.max.c
. Then the
dimensionless measured maximum solid boundingsurface heat ux, dened as,
q
ku.max.i
=
[q
ku
[
max.i
q
max
=
[q
ku
[
max.i
(
S
lg
,A)
max
. i = e. c
dimensionless, maximum
surfaceconvection heat ux
for evaporation and
condensation
(6.12)
is less than unity. A compilation [5] of the available experimental results for q
ku.max.i
,
is given in Figure 6.19(b). The results suggest that since
g
Lh
lg
increases with p,p
c
,
then less vapor (or liquid) is produced for a given q
ku
, as p,p
c
increases. Then a
large fraction of the heat ows through the liquid (or vapor). Numerous techniques
introducing high velocities, swirls, jets, and subcooling have been used in the results
compiled in Figure 6.19(b). As shown in Figure 6.19(b) only 10% of the theoretical
maximum has yet been achieved.
Consider an experiment using water with q
max
= 2 10
9
W/m
2
supplied through
a thin solid (shown in Figure 6.17) slab of silver (high k
s
, Table C.14), for minimum
conduction resistance while having some mechanical strength, with a L = 1 mm
thickness. Then this would require a T
sl
T
s
= LT = 4,700 K across this thin slab.
This is not physically possible without melting the slab. Therefore, we have to devise
MassoudKaviany 2011
special transient experiments for achieving the higher surface heat uxes shown in
Figure 6.19(b).
With this introduction into the surfaceconvection heat transfer with liquidgas
phase change, we nowconsider some specic solid geometries, orientations, and uid
ows. Chart 6.6 lists a division of the surface evaporationcondensation based on the
uid phase or phases in contact with the solid. The case of only a thin lm of the
liquid phase in contact with a solid is called lm evaporation or lm condensation.
These are less complex compared with the cases where both uid phases contact the
surface (i.e., bubbles and droplets).
Here we will rst consider only at surfaces with no forced ow(i.e., evaporation
in pool boiling and buoyant dropwise and lmcondensation). Then among the
forced ows, we consider the forced, impingingdroplets evaporation.
EXAMPLE 6.6. FAM
(a) Using Table C.4, in Appendix C, which lists the phasetransition properties
for p = 1.013 10
5
Pa, evaluate kinetic limit of evaporation heat transfer rate
q
max
(W/m
2
) for (i) water, (ii) ethanol, and (iii) R12.
(b) Compare the values with those shown in Figure 6.19(a). Use (6.95), i.e.,
g.
= 0.
SOLUTION
(a) From (6.95), we have
q
max
= (
S
lg
,A) = m
lg.max
Lh
lg
=
_
R
g
T
lg
2M
_
1,2
g
Lh
lg
for
g.
= 0.
From the idealgas relation (3.18), we write
g
=
p
g
M
R
g
T
lg
.
Now substituting into (6.95), we have
q
max
=
_
R
g
T
lg
2M
_
1,2
g
Lh
lg
=
_
R
g
T
lg
2M
_
1,2
p
g
M
R
g
T
lg
Lh
lg
=
_
M
2R
g
T
lg
_
1,2
p
g
Lh
lg
.
From Table C.4, we have
(i) water: p = 1.013 10
5
Pa
Lh
lg
= 2.256 10
6
J/kg Table C.4
M = 18.015 kg/kmole Table C.4
T
lg
= 373.2 K Table C.4
p
c
= 2.212 10
7
Pa Table C.4,
MassoudKaviany 2011
(ii) ethanol: p = 1.013 10
5
Pa
Lh
lg
= 8.409 10
5
J/kg Table C.4
M = 46.069 kg/kmole Table C.4
T
lg
= 351.5 K Table C.4
p
c
= 6.38 10
6
Pa Table C.4,
(iii) R12: p = 1.013 10
5
Pa
Lh
lg
= 1.651 10
5
J/kg Table C.4
M = 120.91 kg/kmole Table C.4
T
lg
= 243.4 K Table C.4
p
c
= 4.12 10
6
Pa Table C.4.
We also note that from (1.17), or from Table C.1(b), we have
R
g
= 8.3145 10
3
J/kmoleK.
Here p
g
= 1.013 10
5
Pa, and we have
(i) water: q
max
=
_
18.015(kg/kmole)
2 8.3145 10
3
(J/kmole K) 373.2(K)
_
1,2
1.013 10
5
(kg/ms
2
) 2.256 10
6
(J/kg)
= 2.195 10
8
W/m
2
p
g
p
c
=
1.013 10
5
(Pa)
2.212 10
7
(Pa)
= 4.580 10
3
.
(ii) ethanol: q
max
=
_
46.069
2 8.3145 10
3
351.5
_
1,2
1.013 10
5
8.409 10
5
= 1.349 10
8
W/m
2
p
g
p
c
=
1.013 10
5
6.38 10
6
= 1.588 10
2
.
(iii) R12: q
max
=
_
120.91
2 8.3145 10
3
243.4
_
1,2
1.013 10
5
1.651 10
5
= 5.157 10
7
W/m
2
p
g
p
c
=
1.013 10
5
4.12 10
6
= 2.459 10
2
.
(b) The results are in agreement with the results of Figure 6.19(a).
COMMENT
Note that water and R12 are generally used in pressurized systems (i.e., p
g
>
1.013 10
5
Pa), and this aids in increasing q
max
[as shown in Figure 6.19(a)].
MassoudKaviany 2011
c
Thin Absorbed Film
Droplet
Nucleation
and Growth
(a) Dropwise Condensation Adjacent
to Vertical Plate
(b) Filmwise Condensation Adjacent
to Vertical Plate
g
Departure
Coalescence
q
ku
l
(x)
y
Solid
Liquid
T
s
T
lg
q
k
q
k
q
ku
q
ku
x, u
y
Solid Vapor
g
Vapor
x, u
k
s
k
s
L L
T
o
T
s
T
lg
T
o
q
u
m
gl
h
gl
Figure 6.23. (a) Dropwise condensation showing the droplet nucleation, growth, departure,
and coalescence on a cooled, vertical surface bounding a saturated vapor. (b) Filmwise con
densation.
6.6.3 Dropwise and Film Condensation
We nowexamine the surfacedroplet nucleation, growth, and departure on a vertical,
cooled surface bounding a saturated vapor, as depicted in Figure 6.23. Most existing
analyses of droplet formationanddynamics, oncooledsurfaces bounding a vapor, are
for an otherwise quiescent vapor. The interest in the dropwise condensation, versus
the lmwise condensation, is that the surface droplets with a spherical cap shape
offer a small conductive resistance to the heat owaround their peripheries. This can
increase the heat transfer rate, and therefore, the condensation rate. Various surface
coatings (e.g., Teon) that increase the contact angle of the liquid, i.e., hinder wetting,
have been examined. Figure 6.24 shows the variation of the surfaceconvection heat
transfer by condensation, with respect to surface subcooling T
lg
T
s
[21]. Note the
higher heat transfer rates associated with the dropwise condensation. The transition
to lmcondensation is also shown. We have seen such q
ku
versus surface temperature
behavior in Figure 6.18 (in third quadrant).
(A) Condensation Regimes for Vertical Plate
The condensation of vapor on a cooled solid surface occurs when the surface tem
perature T
s
becomes lower than a threshold (or incipient) temperature. This is
designated by the threshold supercooling (or supersaturation), i.e., T
lg
T
s
, where
T
lg
is the saturation temperature of the vapor. As with the incipient surface bubble
nucleation, this supercooling would depend on the uid and the surface properties
and conditions. As with the vapor bubbles, the general practice is to modify the
homogeneous nucleation theory to accommodate the surface energy, e.g., contact
angle
c
.
MassoudKaviany 2011
7 x 10
5
6 x 10
5
5 x 10
5
4 x 10
5
3 x 10
5
2 x 10
5
10
5
0 10 20 30 40 50 60 70 80 90
0
Dropwise Condensation
Theory
Transition
Transition From Dropwise to Filmwise Condensation on a Vertical Plate
Flimwise Condensation Theory
Dropwise Condensation Filmwise Condensation
Experiment
Fluid: Ethanediol, u
f,
= 0, T
lg
= 386 K, p
g
= 3.8 x 10
3
Pa,
TeflonFilm Coated Vertical Copper Surface
q
k
u
L
,
W
/
m
2
T
lg
T
s
, K
Figure 6.24. Measured and predicted variation of the surface heat ux with respect to the
surface subcooling for dropwise and lmwise (and transition) condensation on a coated,
vertical surface. (From Stylianou, S.A., reproduced by permission c 1983, Elsevier Science.)
In discussions of vapor condensation, instead of the surface subcooling T
lg
T
s
,
the vapor supersaturation ratio S is used. This is the ratio of the vapor pressure
needed for the incipient surfacedroplet nucleation p
g
, to the saturation pressure at
the same vapor temperature T
lg
(i.e., isothermal compression) p
g
(T
lg
), i.e.,
S =
p
g
p
g
(T
lg
)
supersaturation ratio. (6.108)
Beyond the threshold subcooling, the number of nucleations per unit area and
the growth rate of the formed droplet increase as the subcooling increases. Then,
for small values of liquid subcooling, an isolateddroplets regime exists followed
by the interactingdroplets regime, which results in the droplet coalescence and
eventually in the liquidlm ow. The growth of a single, surfacenucleated droplet
is by condensation on the droplet surface and is controlled by the heat transfer rate
through the droplet and across the dropletsolid interface. Due to the gravity (i.e.,
negative phasedensity buoyancy) or the drag force at the liquidvapor interface
A
lg
(for the case of a owing vapor), the droplet moves when a departure size
MassoudKaviany 2011
(or an equivalent diameter) is reached. This growth and departure are depicted in
Figure 6.23(a).
(B) Correlation for Film Condensation on Vertical Plate
The boundarylayer solution gives the local condensate lmthickness
l
(m), the local
rate of phase change
M
gl
(kg/ms), the local liquid velocity at the liquidgas interface
u
l
(m/s), and Nu
L
and Nu
L
)
L
as [16]
l
(L) =
_
4k
l
l
(T
eq
T
s
)L
gL
lg
Lh
lg
_
1,4
(6.109)
M
gl
(L) = L
lg
g
2
l
L,(3
l
) (6.110)
u
l
(at
l
) = L
lg
g
2
l
,j (6.111)
Nu
L
=
L
l
=
_
L
lg
gLh
lg
L
3
4k
l
l
(T
lg
T
s
)
_
1,4
. (6.112)
and using (6.50), we have
Nu)
L
=
q
ku
)
L
L
(T
lg
T
s
)k
l
= 0.9428
_
gL
lg
Lh
lg
L
3
k
l
l
(T
lg
T
s
)
_
1,4
laminar, nonwavy
lm condensation
on a vertical,
semiinnite plate.
(6.113)
Note that L
lg
=
l
g
, and
M
gl
is the mass rate of gas to liquid phase change,
over length L for a unit depth (n = 1 m). The condensation lm becomes unstable
(wavy) and then turbulent when the accumulated condensate becomes larger than a
threshold value. The onset of lmcondensation turbulence is empirically given by
4[q
ku
)
L
[L
j
l
Lh
lg
= 1,800
onset of turbulence in
verticalplate lm condensation.
(6.114)
A value as low as 33 is used for this quantity for the onset of waviness [16].
So far, we have addressed liquidgas phase change with the uid motion and
liquid or vapor supplied to the surface by phasedensity buoyant motion. In practice,
the liquid or vapor may be forced over the heat transfer surface and the effect of
this forced ow or the heat transfer rate and the rate of phasechange can be sub
stantial. The uid being forced to ow can be semibounded or completely bounded.
Among the semibounded forced ows are ow boiling and condensation over at
plates, cylinder, and spheres. The existing correlations are compiled and presented
in [17].
6.6.4 ImpingingDroplets Surface Convection with Evaporation
In order to continuously remove large heat uxes from a surface (of the order of 10
6
W/m
2
=1 MW/m
2
) while using a small amount of liquid, forced droplet spray cooling
MassoudKaviany 2011
q
ku
q
k
q
ku
q
k
q
ku
q
k
Flooded Surface
Liquid Film
Gas Flow
Nozzle
(b) Multidroplet
Impingement
(a) Isolated Droplet
Impingement
SurfaceBubble
Nucleation or
Vapor Film
Recoil
Gas
Flow
Dense Droplet
Concentration
DropletImpingement Cooling
Gas Flow
T
s
T
o
k
s
L
r,s
u
d
, m
d
T
l,
< T
lg
g
We
D
> 80 We
D
< 80
t
1
t
2
>t
1
t
4
>t
3
t
2
>t
1
t
4
>t
3
t
3
>t
2
t
1
t
3
>t
2
Breakup
u
d
D
Figure 6.25. A rendering of the impinging and evaporating droplets showing the droplet
impingement dynamics. (a) For the isolated droplet (for two different droplet Weber num
bers). (b) For the multidroplet impingement (with interdroplet interactions upon surface
impingement).
is used. Figure 6.25 gives a rendering of the surface dropletimpingement cooling.
Chart 6.7 gives a description of the heat transfer mechanisms [4]. During droplet
contact with the surface, contact conduction occurs and this results in evaporation
and vapor convection. Droplets also entrain (cause to move) the local gas (e.g., air
andvapor) andthus owing gas exchanges heat withthe surface andthis is designated
as bulk convection. Radiation heat transfer between the droplets and the surface
can also become signicant.
Impinging droplets can reduce the resistances to evaporation vapor removal,
thus allowing for approachtoward, the upper limit (kinetic theory) of the evaporation
rate. The resistance to the heat transfer to the liquidvapor interface can be reduced
by sweeping the heated solid surface by rapidly moving liquid in close contact (due
to droplet inertia). The resistance to vapor removal can be reduced by entraining the
vapor in the rapidly moving gas surrounding the droplets. In order to understand and
reduce these resistances, the droplet impingement dynamics (for multiple droplets),
which are inuenced by the surface heat transfer, need to be modeled. However,
because of the complexity of the droplet impingement dynamics (i.e., the droplet
surface interaction), the interdroplet interactions (i.e., interaction before, during,
and after impingement), and the existence of various evaporation regimes (e.g., the
surfacebubble nucleation and the vaporlm regimes), no comprehensive model is
yet available [4].
MassoudKaviany 2011
SurfaceImpact
Conduction
BulkGas Surface
Convection
ImpingingDroplet Surface Heat Transfer
DropletRelated
Conduction and Convection
Vapor
Gas
Surface Radiation
to Droplets
Local Vapor
Surface Convection
Local EntrainedGas
Surface Convection
q
ku
q
ku
q
ku
q
k
q
r
Solid
Surface
Droplet
Nozzle
Droplet
q
u
q
u
q
ku
q
r
Droplet
Evaporation
q
k
s
lg
.
Chart 6.7. Heat transfer mechanisms in impingingdroplet surface convection.
(A) IsolatedDroplet Dynamics and Evaporation: Weber Number
The breakup of a droplet upon impingement on a surface is described by the droplet
Weber number We
D
, based on the droplet diameter D, i.e.,
We
D
=
l
u
2
d
D
l.
d
_
1
m
d
),
l.
( m
d
),
l
)
o
_
1
T
s
T
l.
1,720(T
s
T
l.
)
0.088
D)
1.004
u
d
)
0.764
( m
d
),
l.
)
2
( m
d
),
l
)
o
for 180
C  T
s
T
l.
 380
C
vaporlm regime in
impingingdroplet cooling,
(6.116)
MassoudKaviany 2011
0
10
6
2 x 10
6
0 100 200 300
q
k
u
L
,
W
/
m
2
T
s
T
lg
, K
m
d
, kg/m
2
s
0.347
DropletImpingement Cooling: High Mass Fluxes
0.729
1.43
1.64
D = 480 m, We
D
= 67 84,
u
d
, m/s
3.23
3.20
3.21
3.58
Water, T
l,
= 20
o
C, T
lg
= 100
o
C
VaporFilm
Regime
Transition Surface
Nucleation Regime
Surface
Nucleation
Regime
q
u
Figure 6.26. Dropletspray cooling of a heated surface showing the variation of the mea
sured surface heat ux with respect to the surfaceconvection superheat, for four different
droplet mass uxes. (FromDeb, S., and Yao, S.C., reproduced by permission c 1989, Elsevier
Science.)
where
d
3.68 10
4
l.
Lh
lg.
(T
s
T
l.
)
1.691
D)
0.062
(6.117)
Lh
lg.
c
p.l
(T
lg
T
l.
) Lh
lg
(6.118)
_
m
d
)
l
_
o
5 10
3
m/s. (6.119)
A Nusselt number based on the gasphase conductivity is dened as
Nu)
L
L
A
ku
R
ku
)
L
k
g
. (6.120)
The correlation is based on experimental results for water droplet spray on a
nickelplated copper surface, in an otherwise stagnant air. The surface radiation heat
transfer also becomes signicant, and its effect is indirectly included in the above
correlation.
Note that the vaporlmregime has a rather small q
ku
and for signicant surface
convection heat transfer, the nucleate boiling regime should be used.
MassoudKaviany 2011
EXAMPLE 6.8. FAM
Consider water dropletspray cooling of the plastic surface of Example 6.7, and
in order to assure lm boiling, consider T
s
= 200
C. The average
droplet diameter is D) = 500 jm = 0.5 mm, the average droplet velocity is
u
d
) = 2 m/s, the average droplet mass velocity is m
d
),
l.
= 2 10
3
m/s.
(a) Draw the thermal circuit diagram.
(b) Determine the rate of surfaceconvection cooling.
(c) Determine the average surfaceconvection resistance R
ku
)
L
.
Evaluate the water properties at T = 373 K.
L
m
lg
h
lg
T
s
> T
lg
Water
Nozzle
T
l,
< T
lg
q
u
q
ku
u
d
D
SurfaceConvection Heat Transfer: Droplet Impingement and Evaporation
A
ku
(a) Physical Model (b) Thermal Circuit Model
T
l,
T
s
Q
s
Q
ku L R
ku L
Figure Ex. 6.8. Surfaceconvection heat transfer by droplet impingement lm boiling.
(a) Physical model. (b) Thermal circuit model.
SOLUTION
(a) The thermal circuit diagram is shown in Figure 6.8(b).
(b) The rate of surfaceconvection heat transfer is given by (6.116) as
Q
ku
)
L
=
T
s
T
l.
R
ku
)
L
=
A
ku
(T
s
T
l.
)
A
ku
R
ku
)
L
.
where A
ku
= L
2
and A
ku
R
ku
)
L
is given for impingingdroplet lm evaporation
by the correlation (6.116). Then for Q
ku
)
L
we have
Q
ku
)
L
= A
ku
l.
Lh
lg.
_
m
d
l.
_
d
_
1
m
d
),
l.
( m
d
),
l
)
o
_
A
ku
1. 720 (T
s
T
l.
)
0.912
D)
1.004
u
d
)
0.764
( m
d
),
l.
)
2
( m
d
),
l
)
o
.
MassoudKaviany 2011
where ( m
d
),
l
)
o
= 5 10
3
m/s and
d
are given by (6.119) and (6.117). Here,
from Tables C.23 and C.4, we have for water at T = 373 K
l.
= 960.5 kg/m
3
Table C.23
(c
p.l
)
(T
lg
T
l.
) Lh
lg
= 4,218(J/kgK) (100 10)(K) 2.256 10
= 2.636 10
6
.
Also, from (6.117) we have
d
=
3.68 10
4
l.
Lh
lg.
(T
s
T
l.
)
1.691
D)
0.062
d
=
3.68 10
4
960.5(kg/m
3
) 2.636 10
6
10
6
(J/kg)
(200 10)
1.691
(K)
1.691
(5 10
4
)
0.062
(m)
0.062
= 0.1661.
Now we evaluate Q
ku
)
L
as
Q
ku
)
L
= (0.2)
2
(m)
2
960.5(kg/m
3
) 2.636 10
6
(J/kg) 2 10
3
(m/s)
0.1661
_
1
2 10
3
(m/s)
5 10
3
(m/s)
_
(0.2)
2
(m
2
) 1,720
(200 10)
0.912
(K)
0.912
(5 10
4
)
1.004
(m)
1.004
(2)
0.746
(m/s)
0.746
(2 10
3
)
2
(m/s)
2
5 10
3
(m/s)
= 2.019 10
4
(W) 8.100 10
3
(W) = 2.819 10
4
W.
(c) The average surfaceconvection resistance is given by (6.115), i.e.,
R
ku
)
L
=
T
s
T
l.
Q
ku
)
L
= 6.740 10
3
C/W
or
A
ku
R
ku
)
L
= (0.2)
2
(m
2
) 6.740 10
3
(
C/W) = 2.696 10
4
C/(W/m
2
).
COMMENT
Note that, although it is close, q
ku
)
L
= 5.568 10
6
W/m
2
is slightly larger than
that shown in Figure 6.26. This is because the correlation uses other data in
addition to those given in Figure 6.26. Compared to the fullydeveloped nucleate
pool boiling cooling of Example 6.7, (T
s
T
lg
= 25
T
l.
= 190
C
singlephase, laminar, forced parallel ow 7.007 10
4
impingingdroplets, lmvapor regime 2.696 10
4
T
s
T
l.
= 190
C
singlephase, turbulent, forced parallel ow 6.615 10
5
fullydeveloped nucleate pool boiling, T
s
T
lg
= 15
C 3.044 10
5
singlephase, turbulent, highspeed single impinging jet 1.590 10
6
heat transfer rate Q
ku
)
L
is higher, but the surfaceconvection resistance R
ku
)
L
is also larger than the nucleate pool boiling.
Table Ex. 6.8 summarizes the surfaceconvection resistances in Examples
6.3 to 6.5, 6.7, and 6.8. Here the product A
ku
R
ku
)
L
is used and the results are
arranged in order of decrease in the surfaceconvection resistance A
ku
R
ku
)
L
.
Note that for the conditions used, a single impinging jet (single phase) has
resulted in the smallest resistance.
6.8 Inclusion of Substrate
6.8.4 HotWire Anemometry
An anemometer is an instrument used for the measurement of the velocity of moving
uids. A hotwire anemometer uses the surfaceconvection heat transfer from a
heated wire (and its relationship to the fareld uid velocity u
f .
) to indirectly
measure the fareld velocity u
f .
. The hotwire anemometer probe is made of
a probe body, which has two supporting needles, a wire sensor at the one end,
and electrodes at the other end. These are shown in Figure 6.33(a). The wire is
usually made of tungsten or platinum, with L = 1 mm and D= 5 jm [14]. This
aspect ratio L,D of 200 ensures a minute end effect. The wire may have a thin
metallic coating and it is welded to the support needles, as shown in the gure. It
is heated electrically. The most common anemometer operates under a constant
temperature. The circuit for this electronic unit is also shown in Figure 6.33(b). The
circuit supplies the current needed to maintain a constant resistance
e
(i.e., constant
temperature), under varying uid velocity u
f .
. The circuit contains a Wheatstone
bridge, which has the sensor as one of its arms. In addition, there are two xed
resistors and one adjustable resistor in the bridge. These are shown in the gure.
A differential feedback amplier senses the bridge conductance and adds/reduces
current to maintain the sensor temperature constant. The resistance of the wire
R
e
(T
s
), is related to the uid fareld velocity u
f .
.
MassoudKaviany 2011
q
ku
q
r
T
s
(x)
u
f,
T
f,
u
f,
T
f,
D
L
x
x
A
k
= D
2
/4
A
ku
= P
ku
x = D x
End Thermal
Condition T T
f,
Sensor (Probe) Body (Metallic Cover
with Ceramics or Epoxy Filler)
Electrodes
(a) HotWire Anemometer Sensor
(b) Supporting Electrical Circuitry
Support Needles
Wire Sensor
Fluid Flow
Support
Needle
Wire Sensor
Bridge Unbalance
Sensed Here
r,s
q
k
t
(c
p
T)
s
, s
e,J
j
e
J
e
( )
(+)
, Measured Anemometer
Output Voltage
Feedback
Amplifier
Variable
Resistance R
e
Fixed Resistances
Sensor
Fluid
Flow
( (+)
Figure 6.33. (a) The hotwire anemometer probe. (b) Its supporting electrical circuitry. The
volumetric heat generation, intramedium conduction, and surface convection are also shown.
The resistance R
e
(T
s
) is related to the resistivity
e
(T
s
). The relation between
e
and T
s
is linearized as
e
(T
s
) =
e.o
[1
e
(T
s
T
o
)] linear
e
T
s
relation, (6.159)
where
e.o
and T
o
are the reference resistivity and temperature. For platinum, the
thermal coefcient of resistivity
e
(1/K) is 3.510
3
1,
4
D
q
r
x
q
k
=
t
(c
p
)
s
T
s
s
e.J
transient
energy
equation
(6.160)
or
P
ku
(q
ku
q
r
) A
k
x
q
k
= A
k
t
(c
p
)
s
T
s
J
2
e
e
A
k
(6.161)
or
P
ku
(q
ku
q
r
) A
k
x
k
s
T
s
x
= A
k
t
(c
p
)
s
T
s
J
2
e
e
A
k
. (6.162)
where A
k
= D
2
,4 and P
ku
= Dare dened in Figure 6.33. Under negligible radia
tion (i.e., c
r.s
0), constant k
s
and c
p
, and by using q
ku
= Nu)
D
(k
f
,D)(T
s
T
f .
),
we write (6.162) as
P
ku
Nu)
D
k
f
D
(T
s
T
f .
) A
k
k
s
2
T
s
x
2
= (c
p
)
s
A
k
T
s
t
J
2
e
e
A
k
transient
energy
equation.
(6.163)
(A) Temperature Distribution along Sensor
Here we are interested in determining T
s
= T
s
(t). Nowusing the electrical resistance
R
e
as denedby (2.36), i.e., R
e
(T
s
) =
e
(T
s
)L,A
k
, or
e
(T
s
) = R
e
(T
s
)A
k
,L, andusing
e
as given by (6.159), and for steady state (i.e., ,t = 0), (6.163) becomes
d
2
T
s
dx
2
e
J
2
e
R
e
(T
o
) P
ku
LNu)
D
k
f
D
A
k
k
s
L
(T
s
T
f .
)
=
J
2
e
R
e
(T
f .
)
A
k
k
s
L
steadystate energy equation. (6.164)
Equation (6.164) is integrated using the thermal conditions at the tip surfaces.
These tip surfaces are assumed to be at the prescribed uid fareld temperatures
T
f .
(this assumes that the support needles are in thermal equilibrium with the
surrounding uid). These tip thermal conditions are
T
s
(x = 0) = T
s
(x = L) = T
f .
wireend thermal conditions. (6.165)
This equality of thermal conditions at both ends is assumed to have a symmetry
(i.e., dT
s
,dx = 0) with respect to the plane passing through x = L,2. We use this
condition and move the origin of the coordinate system to this central plane and
dene the new coordinate axis x
1
as x
1
= x L,2. Using this, the integration of
MassoudKaviany 2011
(6.164) gives the temperature distribution along x
1
as
T
s
(x
1
) = T
f .
a
1
a
2
_
cosh a
1,2
2
x
1
cosh a
1,2
2
L,2
1
_
. x
1
= x L,2
temperature
distribution
along sensor
a
1
=
J
2
e
R
e
(T
f .
)
A
k
k
s
L
. a
2
=
e
J
2
e
R
e
(T
o
) P
ku
LNu)
D
k
f
D
A
k
k
s
L
. (6.166)
The aspect ratio L,Ddetermines the extent of the uniformity of T
s
in the central
portion of the hot wire.
The goal of the analysis of the temperature distribution is to choose an L,Dsuch
that T
s
is nearly uniform over most of the wire. This would allow for using a single
T
s
(or the average T
s
) over the wire, without a signicant loss of accuracy.
The average sensor temperature T
s
)
L
is found from
T
s
)
L
1
L
_
L,2
L,2
T
s
(x
1
)dx
1
denition of average sensor temperature. (6.167)
Using T
s
= T
s
(x
1
) from (6.166), we have
T
s
)
L
= T
f .
a
1
a
2
_
tanh(a
1,2
2
L,2)
a
1,2
2
L,2
1
_
average sensor temperature. (6.168)
This average temperature is used in the determination of the uid velocity u
f .
and in the timeresponse analysis of the sensor.
(B) Fluid VelocityElectrical Resistance Relation
Now using the average sensor temperature T
s
)
L
, we drop the axial conduction term
from the energy equation (6.163) and rearrange the resulting equation as
P
ku
Nu)
D
k
f
D
(T
s
)
L
T
f .
) =
J
2
e
e
(T
s
)
L
)
A
k
.
A
ku
Nu
D
)
k
f
D
(T
s
)
L
T
f .
) = J
2
e
R
e
(T
s
)
L
) (6.169)
or
Nu)
D
=
J
2
e
R
e
(T
s
)
L
)
(A
ku
k
f
,D)(T
s
)
L
T
f .
)
energy equation.
(6.170)
Now using (6.159), we can remove T
s
)
L
from the denominator and have a
relationship between Nu)
D
and R
e
(T
s
)
L
) as
Nu)
D
= f [R
e
(T
s
)
L
). R
e
(T
f .
). T
f .
. T
o
]
=
J
2
e
R
e
(T
s
)
L
)R
e
(T
f .
)
e
(A
ku
k
f
,D)[R
e
(T
s
)
L
) R
e
(T
f .
)][1
e
(T
f .
T
o
)]
. (6.171)
MassoudKaviany 2011
Next Nu)
D
is related to the uid velocity u
f .
through empirical correlations.
Among these correlations for forcedow surface convection for circular cylinders
(in cross ow) is that given in Table 6.4, i.e.,
Nu)
D
= Nu)
D
(Re
D
. Pr)
= a
1
Re
a2
D
Pr
1,3
Nusselt number correlation for cross
ow over long cylinders (Table 6.4),
(6.172)
where
Pr =
_
jc
p
k
_
f .
. Re
D
=
u
f .
D
f .
. (6.173)
and a
1
and a
2
are found in Table 6.3, and depend on the specic ranges of Re
D
.
Note that (6.171) is the relation
Nu)
D
(u
f .
) = f [R
e
(T
s
)
L
). R
e
(T
f .
). T
f .
. T
o
]
relation between velocity and temperature. (6.174)
and on the righthand side all variables are measured [R
e
(T
s
)
L
). R
e
(T
f .
)
L
), T
f .
]
or are specied (T
o
). This leads to the determination of u
f .
using (6.172), once
Nu)
D
is determined from (6.171). This is the method of determination of the uid
velocity using the hot wire anemometry.
(C) Sensor Response Time: Lumped Capacitance
The time variation of u
f .
can also be measured, if the wire can respond rapidly
enough to the changes. The speed of this response is hindered by the thermal storage
of the sensor. Returning to the transient energy equation (6.162), with negligible
radiation and no axial conduction (using the average sensor temperature), we have
P
ku
Nu)
D
k
f
D
(T
s
)
L
T
f .
) = (c
p
)
s
A
k
dT
s
)
L
dt
J
2
e
e
A
k
energy
equation
(6.175)
or using (6.159) we have
J
2
e
R
e
(T
s
)
L
) =
(c
p
)
s
A
k
L
e
R
e
(T
o
)
dR
e
dt
P
ku
LNu)
D
k
f
D
(T
s
)
L
T
f .
) (6.176)
or
J
2
e
R
e
(T
s
)
L
) =
(c
p
)
s
A
k
L
e
R
e
(T
o
)
dR
e
(T
s
)
L
)
dt
P
ku
LNu)
D
k
f
D
e
R
e
(T
o
)
[R
e
(T
s
)
L
) R
e
(T
f .
)]. (6.177)
Now using the decomposition of (u
f .
and R
e
), similar to the discussion of
Section6.8.3, the time constant for the above equationis found. This leads tothe time
averaged velocity u
f .
and the uctuating velocity u
/
f .
(T
f .
is assumed constant)
MassoudKaviany 2011
and the corresponding R
e
and R
/
e
. The result for the time constant
s
(s) is [14]
s
=
(c
p
)
s
V[R
e
(T
s
)
L
) R
e
(T
f .
)]
J
2
e
e
R
e
(T
f .
)R
e
(T
o
)
time constant. (6.178)
where J
e
is assumed constant and V = A
k
L.
Atypical value for
1
is of the order of 0.01 s (i.e., uctuation up to 100 Hz can be
measured). Smaller time constants are possible using the frequency compensation
method [14].
EXAMPLE 6.17. FAM
In an air mass ow meter used in measurement of air intake into internal com
bustion engines, three or more hotwire anemometers are placed across the air
ow channel, as shown in Figure Ex. 6.17(a). Multiple hot wires are used to nd
an average velocity across the ow channel. The velocity is measured from the
heat loss from the hot wire and the density is determined from the temperature
measurement.
L
D
S
e,J
A
ku
T
f,
(a) Air Mass Flow Meter
Each Hot Wire
(b) Thermal Circuit Model
Electronic Signal
Processor
Honeycomb Structure for
OneDimensionality of Flow
Thermocouple
Air Flow
A
ku
e
( T
s L
)
(+)
Hot Wire
( )
J
e
T
s L
T
f,
u
f,
T
s L
R
ku D
Q
ku D
Figure Ex. 6.17. (a) A multisensor hotwire anemometer. (b) Thermal circuit model.
(a) Draw the thermal circuit model.
(b) Using (6.170) and (6.172), and the conditions and values given below, deter
mine the air velocity u
f .
.
T
s
)
L
= 100
C, T
f .
= 15
f
.
The air and tantalum properties are found from Table C.22 and Table C.8, at
T = 300 K, as
k
f
= 0.0267 W/mK Table C.22
f
= 1.566 10
5
m
2
/s Table C.22
Pr = 0.69 Table C.22
e
= 1.629 10
8
ohmm Table C.8.
Also,
A
ku
= DL = 10
4
(m) 8 10
3
(m) = 2.512 10
6
m
2
R
e
= 1.629 10
8
(ohmm) 8 10
3
(m),7.85 10
9
(m
2
) = 0.01660 ohm.
Substituting for the numerical values, we have
a
1
_
u
f .
10
4
(m)
1.566 10
5
(m
2
/s)
_
a
2
(0.69)
1,3
=
(5)
2
(A
2
) 0.016660(ohm) 10
4
(m)
2.512 10
6
(m
2
) 0.0267(W/mK)(100 15)(K)
.
We begin by guessing that 0  Re
D
 4 10
3
, and from Table 6.4, we have
a
1
= 0.683. a
2
= 0.466.
Solving for u
f .
, using a solver, we have
u
f .
= 32.77 m/s.
The Reynolds number is
Re
D
=
32.77(m/s) 10
4
(m)
1.566 10
5
(m
2
/s)
= 209.3  4 10
3
.
COMMENT
By choosing a smaller diameter D, a smaller current can be used.
MassoudKaviany 2011
6.9 SurfaceConvection Evaporation Cooling
In Section 5.3 and so far in all examples and problems, we have not addressed the
resistance to vapor owing through the inert (noncondensible) gaseous species and
away from evaporation surface. This resistance is zero when the vapor molecules
can be readily removed while maintaining a uniform vapor pressure adjacent to
the evaporation surface. In practice the vapor pressure is not uniform, because the
gas contains other species (called inert or noncondensable, because of their higher
boiling temperatures). A common example is water vapor in air (mostly nitrogen
and oxygen). Then this mass diffusion limit has to be at rest. Here we consider the
combined heat and mass transfer limits to surfaceconvection evaporation.
Similar tothe surfaceconvectionheat transfer resistance, the surfaceconvection
mass transfer rate
M
A
[
A
for species A (e.g., water vapor), appearing in (1.26), is
represented by a surfaceconvection mass transfer resistance R
Du
)
L
, and the species
A vaporpressure potential p
A.s
p
A.
. Here D stands for D
m.A
(m
2
/s), which is
the mass diffusion coefcient and is a uid property similar to k for heat transfer
[Table C.20(a)]. Instead of the vaporpressure potential, it is also customary to use
the vapordensity potential (
f .A
)
s
(
f .A
)
, where
f
is the gasmixture density. Using this we
write the species A surfaceconvection mass transfer rate as
M
A
[
A
=
M
Du
)
L
=
(
f .A
,
f
)
s
(
f .A
,
f
)
R
Du
)
L
surfaceconvection species A
mass transfer rate for
(
f .A
,
f
)
s
_1.
(6.179)
Figure 6.34 shows the surfaceconvection mass transfer. Here, we have assumed
that (
f .A
,
f
)
s
is small enough such that we can neglect the convection of species A
across the surface (as compared to that due to surface convection).
We can allow for the convection of species A across the liquidgas interface (in addition to surface
convection), i.e.,
M(
f .A
,
f
)
s
, where
M is the mass ow rate crossing the evaporation interface.
Note for the single component liquid considered here, only species Acrosses the liquidgas interface,
therefore,
M[
A
=
M
A
[
A
.
The convection across the liquid interface is added to the surfaceconvection mass transfer, where
the surfaceconvection term is similar to (6.23) for heat transfer. By adding these, we have
M
A
[
A
=
M[
A
_
f .A
f
_
s
f
D
m.A
(
f .A
,
f
)
y
[
y=0
=
M
A
[
A
_
f .A
f
_
s
f
D
m.A
(
f .A
,
f
)
y
[
y=0
or
M
A
[
A
=
f
D
m.A
(
f .A
,
f
)
y
[
y=0
1 (
f .A
,
f
)
s
.
Nowrelating the surfaceconvection termto the surfaceconvection mass transfer resistance R
Du
)
L
,
we have
M
A
[
A
=
(
f .A
,
f
)
s
(
f .A
,
f
)
R
Du
)
L
[1 (
f .A
,
f
)
s
]
.
MassoudKaviany 2011
R
Du L
M
A
V
(
f,A
)
M
Du L
(
f,A
)
s
Liquid
(Mixture of Volatile
Species A and Other
Inert Species)
Surface Conditions
(
f,A
)
s
,
f,s
, T
s
Gas
SurfaceConvection Mass Transfer
Rate (Evaporation / Condensation)
Gas Flow and
FarField Conditions
SurfaceConvection Mass Transfer Resistance
Rate of
Storage and Generation
of Species A
Control Volume V
Mass of Species A
in Volume V
+ N
A
V
u
f,
, (
f,A
) ,
f,
, T
f,
Surface Convection Area A
ku
s
n
m
Du,A
m
u,A
SurfaceConvection
Mass Flux Vector
d
dt
M
A
V
Figure 6.34. Surface evaporation showing simultaneous heat and mass transfer.
Then the surfaceconvection mass transfer resistance R
Du
)
L
(s/kg) is similar to
the heat transfer resistance R
ku
)
L
. For the case of an evaporation surface, where the
heat and mass transfer surface are the same, the two resistances are related as
R
Du
)
L
= R
ku
)
L
c
p.f
Le
2,3
. Le =
D
m.A
f
surfaceconvection mass transfer
resistance related to heat
transfer and Lewis number Le,
(6.180)
where Le is the Lewis number (also discussed in Section 5.4.1) and is similar to the
Prandtl number. Lewis number signies the ratio of species A mass boundarylayer
thickness to thermal boundarylayer thickness. When Le 0 (i.e., D
m
0), the
mass transfer resistance tends to innity, i.e., no surfaceconvection mass transfer
occurs. In (6.180), the surfaceconvection resistance R
ku
)
L
is evaluated as discussed
in Section 6.7.
Tables C.20(a) and (b) list the value of D
m
for some gaseous species in mixtures.
As temperature increases, D
m
increases. As D
m.A
increases R
Du
)
L
decreases, result
ing in a larger mass transfer rate. Now we apply (6.179) to a integralvolume liquid
body containing species A. Then the species A mass conservation equation (1.26)
Comparing to (6.179), this shows a larger species Amass transfer rate, when this convection transport
is added. We note that when convection across the interface is signicant, its effect on R
Du
)
L
should
be included.
For the case of
(
f .A
,
f
)
s
_1.
we will have (6.179).
MassoudKaviany 2011
becomes, as shown in Figure 6.34,
M
A
[
A
=
(
f .A
,
f
)
s
(
f .A
,
f
)
R
Du
)
L
=
d
dt
M
A
[
V
N
A
[
V
integralvolume species i mass
conservation equation,
(6.181)
where M
A
[
V
is the instantaneous mass of species i within the volume, and
N
A
[
V
is
the generation rate of species i within the volume.
When we examine only the evaporation surface, then the above mass conserva
tion equation will be applied to the control surface with the volumetric terms set to
zero.
If there is no generation of species A, i.e.,
N
A
[
V
= 0, and when we use the
notation lg to signify the interfacial liquidgas mass transfer, and if species A is the
only constituent of the liquid phase, then (6.181) becomes,
M
lg
[
A
=
(
f .A
,
f
)
s
(
f .A
,
f
)
R
Du
)
L
=
d
dt
M
A
[
V
integralvolume species A mass
conservation with surface evaporation
and nite, mass transfer resistance.
(6.182)
EXAMPLE 6.18. FAM.S
A water layer is heated by the moving ambient air and the heat transfer results
in the surface evaporation of the water. This is shown in Figure Ex. 6.18(a). The
parallel air ow contains some water vapor and the fareld value is (
f .n
)
.
(a) Write the energy equation for the surface with the surface convection given
in terms of R
ku
)
L
. Draw the thermal and mass circuit diagrams.
(b) Write the evaporation rate
M
lg
and the energy equation in terms of the
resistance R
ku
)
L
.
(c) Write the ClausiusClapeyron relation in terms of T
s
and (
f .n
)
s
.
(d) Explain how T
s
and (
f .
)
s
are determined.
(e) Evaluate R
ku
)
L
.
(f) Determine T
s
and (
f .
)
s
.
(g) Determine
M
lg
and Q
ku
)
L
.
Determine all properties [except T
s
and (
f .
)
s
] at T
f .
. Assume that
f .s
=
f .
.
SOLUTION
Heat ows from the gaseous ambient to the evaporating surface and the amount
of this heat ow depends on the fareld conditions T
f .
,
f .
, and u
f .
. These
conditions also determine the surface temperature T
s
.
MassoudKaviany 2011
x
T
s
T
l
T
T
f,
Gas (Water
Vapor and Air)
Liquid (Water)
Q
s
= 30 W
Q
ku L
S
lg
= M
lg
h
lg
0
x
f,w
(
f,w
)
(
f,w
)
s
0
Evaporation Surface
A
ku
= 1 x 1 (m) T
s
= T
lg
[(
f,w
)
s
]
M
Du L
= M
lg
Air, Parallel Flow
u
f,
= 1 m/s
(
f,w
) = 0.001 kg/m
3
T
f,
= 350 K
WaterVapor
Density Distribution
Temperature
Distribution
(a) Physical Model
Figure Ex. 6.18. (a) Evaporation cooling of a water surface.
(a) The thermal circuit model for the surface is shown Figure Ex. 6.18(b). The
surface energy equation (2.72) becomes
Q[
A
= Q
s
Q
ku
)
L
=
T
s
T
f .
R
ku
)
L
=
S
lg
.
The energy conversion term
S
lg
is found from Table 2.1, and is
S
lg
=
M
lg
Lh
lg
.
(b) The surfaceconvection mass transfer rate
M
lg
is written using (6.179), i.e.,
M
lg
=
(
f .n
,
f
)
s
(
f .n
,
f
)
R
Du
)
L
.
Using this in the energy equation, we have
Q
s
T
s
T
f .
R
ku
)
L
=
(
f .n
,
f
)
s
(
f .n
,
f
)
R
Du
)
L
Lh
lg
.
From (6.180), we have
R
Du
)
L
= R
ku
)
L
c
p.f
Le
2,3
.
Using this, the energy equation becomes
Q
s
T
s
T
f .
R
ku
)
L
=
(
f .n
,
f
)
s
(
f .n
,
f
)
R
ku
)
L
c
p.f
Le
2,3
Lh
lg
.
(c) The ClausiusClapeyronrelationis (A.14), andis repeatedhere (for T
s
= T
ls
)
p
f .n
= (p
f .n
)
o
exp
_
M
n
Lh
lg
R
g
_
1
T
s
1
T
lg.o
__
ClausiusClapeyron relation.
Assuming an idealgas behavior for water vapor, we have
p
f .n
= (p
f .n
)
o
R
g
M
n
f .n
T
lg
ideal gas.
MassoudKaviany 2011
Surface Mass
Transfer
Superheated Vapor
at Far Field
Saturated Vapor
at Surface
Surface Heat
Transfer
Q
ku L
S
lg
= M
lg
h
lg
T
f,
T
s
= T
lg
[(
f,w
)
s
]
A
ku
= 1 x 1 (m
2
)
M
lg
Q
s
R
Du L
(
f,w
/
f
)
(
f,w
/
f
)
s
R
ku L
M
Du L
= M
lg
b) Thermal Circuit Model
Figure Ex. 6.18. (b) Thermal circuit model.
Then using T
s
= T
lg
, we have
(
f .n
)
s
=
(
f .n
)
o
T
lg.o
T
s
exp
_
M
n
Lh
lg
R
g
_
1
T
s
1
T
lg.o
__
.
Here we use the saturation data for water given in Table C.27.
(d) The above ClausiusClapeyron relation and the energy equation are solved
simultaneously for T
s
and (
f .n
)
s
.
(e) We now determine R
ku
)
L
starting from (6.49), i.e.,
R
ku
)
L
=
L
A
ku
Nu
L
k
f
)
.
In order to determine Nu)
L
, we have to evaluate the Reynolds number given
by (6.45), i.e.,
Re
L
=
u
f .
L
f
.
From Table C.22 for air (neglecting the effect of the water vapor at T = 350 K),
and also from Table C.20(a), we have
f
= 2.030 10
5
m
2
/s Table C.22
c
p.f
= 1,007 J/kgs Table C.22
k
f
= 0.0300 W/mK Table C.22
f
= 2.944 10
5
m
2
/s Table C.22
Pr = 0.69 Table C.22
f
= 1.012 kg/m
3
Table C.22
D
m.n
= 2.20 10
5
m
2
/s Table C.20(a)
Le = D
m.n
,
f
= 0.7473.
Then
Re
L
=
1(m/s) 1(m)
2.030 10
5
(m
2
/s)
= 4.926 10
4
 5 10
5
.
MassoudKaviany 2011
The Nu)
L
correlation is that given in Table 6.3, i.e.,
Nu)
L
= 0.664Re
1,2
L
Pr
1,3
Nu)
L
= 0.664(4.926 10
4
)
1,2
(0.69)
1,3
= 130.2.
Then
R
ku
)
L
=
1(m)
1(m) 1(m) 130.2 0.0300(W/mK)
= 0.2560 K/W.
(f) The two equations to be solved for T
s
and (
f .n
)
s
are the energy equation
[with the mass transfer rate represented by R
ku
)
L
] and the ClausiusClapeyron
relation. Both are given above.
FromTable C.27, at T = 300 K(the reference temperature we chose above),
Lh
lg
= 2.438 10
6
J/kg
(
f .n
)
o
= 0.0256 kg/m
3
.
The two equations are
30(W)
(T
s
350)(K)
0.2560(K/W)
=
[(
f .n
)
s
,1.012] (0.001,1.012) 2.438 10
6
(J/kg)
0.2560(K/W) 1,007(J/kgK) (0.7473)
2,3
(
f .n
)
s
=
0.0256(kg/m
3
) 300(K)
T
s
(K)
exp
_
18(kg/kmole)
8,315(J/kmoleK)
2.43810
6
(J/kg)
_
1
T
s
(K)
1
300(K)
__
.
These two equations are solved simultaneously using a solver. The results are
(
f .n
)
s
= 0.02909 kg/m
3
T
s
= 302.3 K.
(g) Using these, we have
M
lg
=
(0.02909,1.012) (0.001,1.012)
0.2560(K/W) 1,007(J/kgK) (0.7473)
2,3
= 8.884 10
5
kg/s = 0.08884 g/s
Q
ku
)
L
=
(302.3 350)
0.2560(K/W)
= 186.3 W.
COMMENT
The water surface temperature is determined from the surfaceconvection
conditions.
Note that if Q
s
= 0, then R
ku
)
L
would drop out of the energy equation.
Then, because the surfaceconvection heat and mass transfer are related, the
energy equation becomes the statement of relation only between these two
terms, the speed of the air stream will not inuence T
s
.
MassoudKaviany 2011
6.11 References
[1] ASHRAE, Handbook, 1997, Fundamentals, American Society of Heating,
Refrigeration, and Air Conditioning Engineers, Atlanta.
[2] Bernardin, J.D., and Mudawar, I., 1997, Film Boiling Heat Transfer of Droplet
Streams and Sprays, International Journal of Heat and Mass Transfer, Volume
40, pp. 25792593.
[3] Carey, V.P., 1992, LiquidVapor PhaseChange Phenomena, Hemisphere Pub
lishing Company, Washington.
[4] Deb, S., and Yao, S.C., 1989, Analysis of Film Boiling Heat Transfer of
Impacting Sprays, International Journal of Heat and Mass Transfer, Volume 32,
pp. 20992112.
[5] Gambill, W.R., and Lienhard, J.H., 1989, An UpperBound for Critical Boiling
Heat Fluxes, ASME Journal of Heat Transfer, Volume 111, pp. 815818.
[6] Gebhart, B., Jaluria, Y., Mahajan, R.L., and Sammakia, B., 1988, Buoyancy
Induced Flow and Transport, Hemisphere Publishing Corporation, Washington.
[7] Hinze, J., 1975, Turbulence, Second Edition, McGrawHill, New York.
[8] Kaviany, M., 2001, Principles of Convective Heat Transfer, Second Edition,
SpringerVerlag, New York.
[9] Kays, W.M., and Crawford, M.T., 1993, Convective Heat and Mass Transfer,
Third Edition, McGrawHill, New York. Also, Mills, A.F., 1999, Heat Transfer,
Second Edition, Prentice Hall Upper Saddle River.
[10] Kraus, A.D., and BarCohen, A., 1995, Design and Analysis of Heat Sinks, Wiley,
New York.
[11] Lahey, R.T., Jr., Editor, 1992, Boiling Heat Transfer, Modern Developments and
Advances, Elsevier, Amsterdam.
[12] Lienhard, J.H., 1987, A Heat Transfer Textbook, Second Edition, PrenticeHall,
Englewood Cliff.
[13] Liter, S.G., and Kaviany, M., 2001, PoolBoiling LiquidChoking Limits within
and above Modulated PorousLayer Coatings, International Journal of Heat
and Mass Transfer, Volume 44, pp. 42874311.
[14] Lomas, C.G., 1986, Fundamentals of Hot Wire Anemometry, Cambridge Univer
sity Press, Cambridge.
[15] Martin, H., 1977, Heat and Mass Transfer Between Impinging Jets and Solid
Surfaces, in Advances in Heat Transfer, Volume 13, Hartnett, J.P., and Irvine,
T.F., Jr., Editors, Academic Press, New York.
[16] Raithby, G.D., and Hollands, K.G., 1985, Natural Convection, in Handbook of
Heat Transfer, Fundamentals, Second Edition, Rohsenow, W.M., et al., Editors,
McGrawHill, New York.
[17] Rohsenow, W.M., and Grifth, P., 1985, Condensation, in Handbook of
Heat Transfer, Fundamentals, Second Edition, Rohsenow, W.M., et al., Editors,
McGrawHill, New York.
[18] Rohsenow, W.M., 1985, Boiling, in Handbook of Heat Transfer, Fundamentals,
Second Edition, Rohsenow, W.M., et al., Editors, McGrawHill, New York.
[19] Schneider, P.J., 1985, Conduction, in Handbook of Heat Transfer, Fundamen
tals, Second Edition, Rohsenow, W.M., et al., Editors, McGrawHill, New York.
[20] Stylianou, S.A., and Rose, J.W., 1983, DroptoFilmwise Condensation Transi
tion: Heat Transfer Measurement for Ethanediol, International Journal of Heat
and Mass Transfer, Volume 26, pp.747760.
[21] Walton, A.J., 1989, Three Phases of Matter, Second Edition, Oxford University
Press, Oxford.
[22] Whitaker, S., 1983, Fundamental Principles of Heat Transfer, Krieger Publishing
Company, Malabar.
MassoudKaviany 2011
6.12 Problems
6.12.9 SurfaceConvection Evaporation Cooling
PROBLEM 6.50. FUN.S
In surfaceconvection evaporation cooling by water seepage of the evaporating
liquid through a permeable wall, the heat required for evaporation is provided by
the ambient gas and the liquid reservoir. This is shown in Figure Pr. 6.50, which is
similar to Figure 5.5, except here the gas surface convection is included.
(a) Sketch the qualitative temperature and water vapor density distributions, and
draw the thermal and mass circuit diagrams.
(b) Write the energy equation for the evaporation surface, along with the relations
for Q
k.u
(from Section 5.3), Q
ku
,
S
lg
, and (
f .n
)
s
= (
f .n
)
s
(T
s
).
(c) Write the watervapor species mass conservation equation for the evaporation
surface.
(d) Determine the surface temperature T
s
for the conditions given below. Note that
u
l
= m
lg
,
l
is unknown.
(e) Determine the rate of heat owing to the evaporating surface from the gas
stream Q
ku
and from the liquid reservoir Q
k.u
.
A
ku
R
ku
)
L
= 5 10
4
C/(W/m
2
), (
f .n
)
= 0.0005 kg/m
3
, T
f .
= 300 K, L = 20
cm, n = 20 cm, L
n
= 6 mm, k) = 0.8 W/mK, T
1
= 293 K.
Evaluate all properties at T = 300 K [except for T
s
and (
f .n
)
s
], and assume
f .s
=
f .
.
A
k,u
l
u
l
= M
l
S
lg
= M
lg
h
lg
Water Reservoir Air
Water
Vapor
FarField
Conditions
Porous Wall
(SolidLiquid)
Control Surface
w
x
L
L
w
M
g
u
f,
, T
f,
WaterVapor
Density (
f,w
)
A
ku
A
ku
M
lg
k
q
u
Q
ku L
Q
k,u
Figure Pr. 6.50. Evaporation cooling of a surface. The heat is provided by the liquid reservoir
and the ambient air.
MassoudKaviany 2011
PROBLEM 6.51. FUN.S
Awater droplet of initial diameter D(t = 0) is in an air streamwith a fareld velocity
u
f .
, while the droplet is moving in the same direction with a velocity u
d
. The air
stream has a fareld temperature T
f .
and water vapor density (
f .n
)
. This is
shown in Figure Pr. 6.51. The droplet has a uniform temperature T
s
(t) = T
lg
(t) at
any time, which is determined from the spontaneous heat transfer and evaporation
rates. Neglect the sensible heat storage/release in the droplet as T
s
varies with time.
u
f,
T
f,
(
f,w
)
T
f,s
(
f,w
)
s
M
Du
(t)
Evaporating Droplet
D(t) u
d
Figure Pr. 6.51. An evaporating droplet with surface convection heat and mass transfer.
(a) Draw the thermal and mass circuit diagrams.
(b) For the conditions given below, plot the variation of the droplet diameter and
volume as a function of time, up to the time the droplet vanishes.
(c) Using the droplet velocity, determine the length of ight before the droplet
vanishes.
D(t = 0) = 4 mm, T
f .
= 350
C, u
f .
= 0.5 m/s, u
d
= 0.1 m/s, (
f .n
)
= 0,
f .
=
f .s
.
Evaluate the air properties at T = 500 K and water properties at T = T
lg
=
300 K.
PROBLEM 6.52. FAM.S
Electric hand dryers provide hot air ow for the evaporation of thin water layers
over human skin. The rate of evaporation, and hence the elapsed time for drying,
depends on the air temperature and velocity and the waterlayer thickness. This is
shown in Figure Pr. 6.52(i). The hands can be modeled as a cylinder with the air
owing across it, as shown in Figure Pr. 6.52(ii).
(a) Draw the thermal and mass circuit diagrams.
(b) Determine the evaporation rate
M
lg
and the watersurface temperature T
s
= T
lg
,
for the conditions given below.
(c) Assuming that the mass transfer rate is constant, determine the elapsed time for
the evaporation of a water layer with thickness l.
u
f .
= 0.8 m/s, T
f .
= 35
C, (
f .n
)
C, T
f .
=
20
C, (
f .n
)
= 0.001 kg/m
2
.
Determine the properties of air at T = 325 K. Determine the water vapor prop
erties at T = 80
y
_
(k
f
k
t
)
T
f
y
_
.
Assume that (u
/2
f
),x _(u
/
f
:
/
f
),y and that (u
/
f
T
/
f
),x _(:
/
f
T
/
f
),y. The
Reynolds stress, u
/
f
:
/
f
is represented as
t
u
f
,y where
t
is the turbulent vis
cosity. The vertical (ydirection) turbulent convection heat ux q
t.y
= (c
p
)
f
:
/
f
T
/
f
is
represented by (6.59) as k
t
T
f
,y, where k
t
is the turbulent thermal conductivity.
(b) Near the solid surface, it can be assumed that the mean velocity and temperature
elds vary only in the vertical direction. That is, u
f
= u
f
(y) and T
f
= T
f
(y). With
this assumption, and a no slip boundary condition at the uidsolid interface, use the
continuity equation to show that the above momentum and energy equations can be
written as
y
_
(
f
t
)
u
f
y
_
= 0
1
(c
p
)
f
y
_
(k
f
k
t
)
T
f
y
_
= 0.
MassoudKaviany 2011
To proceed further, the forms of
t
and k
t
must be specied. One model that has been
developed introduces the concept of the turbulent mixing length
t
, dened in (6.61)
with respect to the turbulent thermal conductivity as k
t
= Pr
1
t
(c
p
)
f
2
t
u
f
,y. The
variation of
t
across the boundary layer is given by (6.63).
(c) As discussed in Section 6.3, experimental results show that there is a laminar
viscous sublayer close to the wall, where there are no turbulent uctuations (i.e.,
t
= 0). Starting from the momentum equation found in part (b), and by assuming
that
t
f
outside the sublayer and using the mixing length model to nd an
expression for k
t
, show that
u
f
y
=
_
_
_
1 laminar viscous sublayer
1
y
= y(
s
,
f
)
1,2
,
f
and the dimension
less horizontal velocity u
f
= u
f
,(
s
,
f
)
1,2
. For the analysis, assume that
t
= y is
valid from the edge of the sublayer to the top of the boundary layer. The square root
of the ratio of
s
, the wall shear stress, and the uid density
f
, is called the friction
velocity, (
s
,
f
)
1,2
. Note that at the solid surface (y = 0), j
f
u
f
,y =
s
.
(d) Empirically, an effective laminar viscous sublayer thickness is found to be
y
.crit
= 10.8. Using the sublayer velocity prole as a boundary condition and
= 0.41, integrate the results of part (c) to show that the velocity prole in the
turbulent portion of the boundary layer is
u
f
= 2.44 ln y
5.00. y
10.8.
Inthe next problem, the temperature prole is determinedandusedwiththis velocity
prole to nd Nu
L
.
PROBLEM 6.55. FUN
This problem is a continuation of the previous one. The next step is to nd the
temperature prole in the boundary layer.
(a) Noting that the convection heat ux in the vertical direction, q
u.y
, is given by
(k
f
k
t
) T
f
,y, use the energy equation from part (b) of the previous problem
to show that q
u.y
is a constant, which will be denoted as q
s
(the surface heat ux).
Then, by using the dimensionless vertical coordinate y
0
dy
1,Pr
t
,
f
.
where
t
= k
t
,(c
p
)
f
and T
f
= [(T
s
T
f
)(
s
,
f
)
1,2
],[q
s
,(c
p
)
f
].
(b) Analogous to the laminar viscous sublayer, there is a laminar thermal sublayer
close to the wall, where there are no turbulent uctuations (i.e.,
t
= 0). Empirically,
MassoudKaviany 2011
for Pr = 0.7, an effective laminar thermal sublayer thickness is found to be y
.crit
=
13.2. Furthermore, if in the turbulent region (y
t
f
, separate the integrand found in part (a) into two parts and show that the
expression for T
f
becomes
T
f
= 13.2Pr
_
y
13.2
f
dy
t
. y
13.2.
(c) By using the mixing length model to nd an expression for
t
, the value of from
the previous problemand Pr
t
=
t
,
t
= 0.9 (valid for Pr . 1), showthat the equation
for the temperature prole in the turbulent portion of the boundary layer is
T
f
= 13.2Pr 2.20 ln y
5.66. y
13.2.
As in part (c) of the previous problem, assume that
t
= y is valid from the edge of
the sublayer to the top of the boundary layer.
(d) By denoting the dimensionless velocity and temperature at the edge of the
boundary layer as u
f .
and T
f .
, and eliminating lny
f .
= 0.9u
f .
13.2Pr 10.2.
(e) The Stanton number St, and the skin friction coefcient c
f
, are dened as
St =
Nu
L
Re
L
Pr
=
q
s
(T
s
T
f .
)(c
p
)
f
u
f .
c
f
=
s
0.5
f
u
2
f .
.
Using the denitions of u
f .
and T
f .
from the nondimensionalization procedure,
and the result of part (d), show that
St =
(c
f
,2)
1,2
T
f .
=
c
f
,2
0.9 (c
f
,2)
1,2
(13.2Pr 10.2)
.
(f) Using an integral method, the skin friction coefcient can be approximated as
c
f
,2 = 0.0287Re
0.2
L
[9]. With this form of the skin friction coefcient, the denom
inator of the Stanton number expression derived in part (e) can be approximated
as Pr
0.4
for 0.5 Pr 1.0 and 5 10
5
Re
L
5 10
6
. With these approximations,
show that
Nu
L
= Nu)
L
= 0.0287Re
4,5
L
Pr
3,5
.
which has a similar form to (6.65), although the prefactor and power of the Prandtl
number are slightly different.
PROBLEM 6.56. FUN
In the boundarylayer ow and heat transfer over a smooth, semiinnite at plate,
the ow is initially laminar, but will transition to turbulence when the Reynolds
MassoudKaviany 2011
number based on the location from the leading edge reaches a critical value, Re
L.t
.
In order to use (6.50) to evaluate the average Nusselt number, Nu)
L
, over the plate,
the laminar and turbulent regions must be considered separately.
(a) Using the expressions for the local Nusselt number, Nu
L
, in the laminar and
turbulent regions, given by (6.44) and (6.65) respectively, develop an expression for
the average Nusselt number for a plate with length L > L
t
.
(b) For a transition Reynolds number Re
L.t
= 5 10
5
, show that this expression
reduces to (6.67).
PROBLEM 6.57. FUN
Surfaceconvection Nusselt number correlations are often evaluated by comparison
with limited exact solutions and by comparison with experimental data. Table
Pr.6.57 lists the experimental result for thermobuoyant ow adjacent to a vertical
plate. The Nusselt numbers are for air, Pr = 0.71, and Rayleigh numbers are in the
laminarow regime (Ra
L
10
9
).
Table Pr. 6.57. Experimental data for Nusselt number at given Raleigh numbers, Pr = 0.71
Ra
L
7 24 8.510
2
1.910
4
2.910
5
2.010
6
8.310
6
4.010
7
1.010
8
3.510
8
Nu)
L
1.5 2.5 3.4 6.4 11.8 19 28 40 50.5 75
(a) Plot the experimental data results given in Table Pr.6.57, using loglog scales.
(b) On the graph of part (a), also plot the correlation given by (6.89), using Pr = 0.71
and Ra
L
= 1 to 10
9
.
(c) On graph of part (b), also plot the laminar portion Nu
L.l
), of (6.92), for the same
range of Rayleigh numbers, and the same Prandtl number.
(d) Using the Taylor series expansion on the laminar portion of (6.92) and then
taking the natural log, show that this expression and (6.89) are identical for large
values of Ra
L
.
PROBLEM 6.58. FAM
To heat a solid sphere of diameter D
s
and emissivity c
r.s
, it is placed inside a long
circular tube and a hot stream of air, with velocity u
f .
and temperature T
f .
, is
owing across (cross ow) it. The tube has a diameter D
t
, length L
t
, emissivity
c
r.t
, and temperature T
t
. These are shown in Figure Pr. 6.58. The system is in
steadystate. Since the tube is long, the radiation heat transfer through the two ends
can be negligible (i.e., the view factor from the sphere to tube ends is zero). So, a
twosurface enclosure can be assumed for the spheretube.
(a) Draw the thermal circuit diagram.
(b) Determine the temperature of the sphere T
s
.
Use D
s
= 0.01 m, c
r.s
= 0.9, u
f .
= 5 m/s, T
f .
= 500 K, D
t
= 0.1 m, L
t
= 1 m,
c
r.t
= 0.05, T
t
= 300 K.
MassoudKaviany 2011
D
t
Solid Sphere Tube
T
t
,
r, t
L
t
D
s
T
s
,
r, s
f
N
f
A
ku.f
= 400 mm
2
, T
f .
= 20
o
C.
Neglect contact resistances. Evaluate all thermopysical properties at T = 300 K.
MassoudKaviany 2011
7
Convection: Bounded Fluid Streams
7.6 Heat Exchange Between Two Bounded Streams
7.6.6 Dielectric and Inert HeatTransfer Fluids
When a heat transfer uid circulates in a heat exchanger (i.e., undergoes heating and
cooling) ina closedcycle, thenthis uidcanbe selectedbasedonits melting andfreez
ing temperatures, its chemical inertness (i.e., reactiveness with the heat exchanger
surfaces), as well as its thermodynamic and transport properties. Heattransfer u
ids are synthetic organic uids especially designed for their noncorrosiveness when
used in metallic heat exchangers. Among these uids are those with low and high
boiling temperatures (at one atm pressure). Figure 7.17 lists some uids that have
boiling temperatures below, near, or above the room temperature. In Appendix C,
Table C.26 lists two heat transfer uids (Dowtherm J and Dowtherm A) that are
synthesized to perform as liquid phase or as liquidvapor phase uids (i.e., undergo
phase change).
Refrigerants are the working uid in the refrigeration, air conditioning, and heat
pumping cycles. The chemical inertness of the refrigerants, in addition to the ther
modynamic phase equilibrium and the thermodynamic and transport properties, are
important in their selection. Chlorouorocarbon (CFC) uids, such as Refrigerant
11 (or R11), are destructive to atmospheric ozone. Table C.3 lists the designations of
the refrigerants, along with their physical properties. The hydrouorocarbon (HFC)
uids, such as Refrigerant 134a (or R134a), and the dielectric ourocarbon (FC)
uids are not expected to be environmentally harmful.
The dielectric uorocarbons can ow directly over electrical carrying conduc
tors for removal of the Joule heating. In Table C.3, the list is arranged in order of
increasing boiling temperature T
lg
(p
g
= 1.013 10
5
Pa). Table C.26 gives the ther
mophysical properties of these uids at one atm. Table C.28 gives the thermophysical
properties of R134a at various pressures.
Auorocarbon can be selected based on the desired boiling temperature. Figure
7.18 shows the variation of the saturation temperature of several ourocarbon uids,
MassoudKaviany 2011
R11
CFC12
FC72
Some Commonly Used Heat Exchanger Fluids with Boiling Temperatures
(at one atm pressure) Not Far from Room Temperature
T
lg
,
o
C
40 0 40 80 120 160 200
FC104
FC77
FC43 FC70 FC71
HFC134a R113 FC84 Water FC75 FC40
240
273
o
C
(0 K)
Figure 7.17. Some uids with boiling temperatures (p
g
= 1 atm) near, below, and above the
room temperature.
Typical Vapor Pressures of Dielectric Fluorocarbon (FC) Liquids
20
T
l
g
,
o
C
p
g
, Pa
10
6
10
0
10
20
30
40
50
60
80
100
120
140
160
210
180
240
10
5
(one atm)
10
4
10
3
F
C

1
0
4
F
C

7
7
F
C

8
4
F
C

7
2
F
C

4
0
F
C

4
3
F
C

7
0
F
C

7
1
F
C

7
5
Figure 7.18. Variation of the saturation temperature of some dielectric uorocarbon (FC)
heat transfer uids, with respect to vapor pressure.
with respect to pressure. Depending on the ranges of pressure and temperature of
interest, a uid is selected.
EXAMPLE 7.7. FAM
In a coaxial heat exchanger, shown in Figure Ex. 7.7(a), liquid water is the cold
stream and ows in the inner tube at a mass ow rate
M
c
= 0.5 kg/s, through
a tube diameter D
c
= 1.5 cm. The thickness of this pure copper inner tube is
l = 1.5 mm. Liquid Dowtherm J is the hot stream and ows in the annulus with
MassoudKaviany 2011
a mass ow rate
M
h
= 0.25 kg/s and the inner diameter of the outer tube is
D
h
= 2.5 cm.
(a) Draw the thermal circuit diagram.
(b) Determine the surfaceconvection resistance for the cold uid R
ku.c
for a
length L = 1 m.
(c) Determine the wallconduction resistance R
k.ch
for L = 1 m.
(d) Determine the surfaceconvection resistance for the hot uid R
ku.h
for a
length L = 1m.
(e) Determine the length Lneeded for an overall resistance, given by (7.88), for
R
Y
= 0.001
C/W.
Evaluate the thermophysical properties for the uids at their saturatedstates
and p = 1.013 10
5
Pa, given in Table C.26. For copper, use Table C.4 and
T = 400 K. Also neglect the entrance effects.
D
c
D
h
l
L
M
h
Hot Stream
(a) Physical Model
Overall Resistance in a Coaxial Tube Heat Exchanger
(b) Thermal Circuit Model
R
ku D,c
R
k,ch
R
ku D,h
T
f,h
Cold Stream
M
c
T
f,c
A
ku,h
A
ku,c
Figure Ex. 7.7. (a) Geometry of the coaxial (parallel axes) heat exchanger. (b) Thermal
circuit diagram for the heat exchanger.
SOLUTION
(a) The thermal circuit diagram is shown in Figure Ex. 7.7(b).
(b) The surfaceconvection resistance for the inner tube is determined from the
available correlations for Nu)
D
for circular tubes and annuli given in Tables 7.2
and 7.3. The appropriate correlation is found using the Reynolds number Re
D
dened by (7.36) with the velocity found from the mass ow rate
M
f
dened by
(7.3). These are
Re
D.c
=
_
f
u
f
)D
j
f
_
c
.
f
u
f
)A
u
=
M
f
. A
u.c
=
D
2
c
4
.
MassoudKaviany 2011
or
Re
D.c
=
4
M
f
D
c
j
f
.
From Table C.26, for liquid water, we have
j
f
= 2.775 10
4
Pas Table C.26
k
f
= 0.6790 W/mK Table C.26
Pr = 1.73 Table C.26.
Then
Re
D.c
=
4 0.5(kg/s)
(1.5 10
2
)(m) 2.775 10
4
(Pas)
= 1.592 10
5
> 10
4
turbulentow regime.
For turbulent ow, from Table 7.3, we have
Nu)
D.c
= 0.023Re
4,5
D
Pr
0.4
Table 7.3.
where we have used the exponent for Pr corresponding to the case of the uid
heating.
Then
Nu)
D.c
= 0.023(1.592 10
5
)
4,5
(1.73)
0.4
= 402.3.
From (7.88), for no extended surfaces present on A
ku.c
[i.e., (
f .o
)
c
= 1], we have
R
ku.c
=
_
A
ku.c
Nu)
D.c
k
f .c
D
c
_
1
A
ku.c
= D
c
L
R
1
ku.c
= LNu)
D.c
k
f .c
= 1(m) 402.3 0.6790(W/m
C) = 858.2 10
2
W/
C
R
ku.c
= 0.001165
C/W.
(c) The wallconduction resistance for the cylindrical shell is given in Table 3.1,
i.e.,
R
k.ch
=
ln
D
c
2l
D
c
2Lk
s
Table 3.1.
For pure copper at T = 400 K, from Table C.14, we have
k
s
= 393 W/m
C Table C.14.
Then
R
k.ch
=
ln
(1.5 10
2
2 1.5 10
3
)(m)
1.5 10
2
(m)
2 1(m) 393(W/m
C)
= 7.383 10
5
C/W.
MassoudKaviany 2011
(d) The surfaceconvection resistance for the hot stream owing in the annulus
is found from the correlations in Table 7.2 and Table 7.3. As discussed in Section
7.4.2, the Reynolds number is based on the hydraulic diameter D
h.h
, i.e.,
D
h.h
=
4A
u.h
P
ku.h
=
4[D
2
h
(D
c
2l)
2
],4
[D
h
(D
c
2l)]
= D
h
(D
c
2l).
From Table C.26, for liquid Dowtherm J, we have
j
f
= 1.72 10
4
Pas Table C.26
k
f
= 0.1185 W/mK Table C.26
Pr = 3.48 Table C.26.
Then
Re
D.h
=
f
u
f
)D
h.h
j
f
.
f
u
f
)A
u
=
M
f
. A
u
= [D
2
h
(D
c
2l)
2
],4
Re
D.h
=
M
f
D
h.h
A
u
j
f
=
M
f
[D
h
(D
c
2l)]j
f
=
0.25(kg/s)
[2.5 10
2
(1.5 10
2
2 1.5 10
3
)](m) 1.72 10
4
(Pas)
= 3.380 10
4
> 10
4
turbulent ow regime.
From Table 7.3, we note that the correlation for Nu)
D
is the same as that used
for the turbulent ow in circular tubes, except D
h.h
is used in the correlation, i.e.,
Nu)
D.h
= 0.023(3.380 10
4
)
4,5
(3.48)
0.3
= 140.4.
where we have used the exponent for Pr corresponding to the case of the uid
being cooled.
Then the surfaceconvection resistance, with no extended surface [i.e.,
(
f .o
)
c
= 1], is given by (7.88), i.e.,
R
1
ku.h
= A
ku.h
Nu)
D.h
k
f .h
D
h.h
= (D
c
2l)LNu)
D.h
k
f
D
h.h
= (1.5 10
2
2 1.5 10
3
)(m) 1(m) 140.4
0.1185(W/mK)
[2.5 10
2
(1.5 10
2
2 1.5 10
3
)](m)
= 134.3 W/
C
or
R
ku.h
= 0.007444
C/W.
MassoudKaviany 2011
(e) The overall resistance R
Y
is given by (7.88), i.e.,
R
Y
= R
ku.c
R
k.ch
R
ku.h
= (0.001165 7.383 10
5
0.007444)(
C/W) = 0.008683
C/W.
Since all R
t
s are proportional to 1,L, we have
L =
R
Y
(L = 1 m)
R
Y
(L)
1(m) =
0.008683(
C/W)
0.001(
C/W)
1(m) = 8.68 m.
COMMENT
Note that here the wall conduction resistance is negligible compared to the two
surfaceconvection resistances, which are nearly equal.
EXAMPLE 7.8. FUN
The exhaust gas stream owing through the automobile exhaust pipe (carbon
steel AISI 4130) is cooled by the surfaceconvection (to the air stream ow
ing underneath the automobile) and surfaceradiation heat transfer. Consider
surfaceconvection heat transfer only. The air stream owing over the pipe can
be treated as a semibounded uid stream with a fareld velocity (u
f .c
)
and a
temperature (T
f .c
)
M
h
=
1
2
rpm
60
f
V
d
V
.
where V
d
is the engine total (all cylinders) displacement volume, rpm is the
engine rpm, and
f
is the cylinder inlet density (approximately air at the
ambient temperature and at one atm pressure), and
V
= 1, is the volumetric
efciency. The hot exhaust gas enters at T
f .h
)
0
and exits at T
f .h
)
L
.
(a) Draw the thermal circuit diagram for a location x along the exhaust pipe,
with a differential length Lx along x.
(b) Draw the anticipated temperature distributions along the pipe, for the
exhaust gas, the pipe wall (use a radially averaged temperature), and the
ambient air.
(c) For this problem, would the prediction of T
f .h
)
L
using the twostream
treatment of Section 7.6.4 give the same results as that of Section 7.2.6?
(d) Determine the exit temperature T
f .h
)
L
for the following conditions.
rpm = 2,500,
f
= 1.1 kg/m
3
, V
d
= 2.2 10
3
m
3
, T
f .h
)
0
= 600
C,
(T
f .c
)
= 35
C, (u
f .c
)
= 20 m/s, D
1
= 4.1 cm, D
2
= 4.4 cm, L = 1.5 m.
Evaluate the exhaust gas properties using air at T = 1,000 K, and the ambi
ent air properties at T = 300 K. Assume no heat conduction along the pipe.
MassoudKaviany 2011
(u
f,c
)
(T
f,c
)
(T
f,c
)
Tube Wall
T
s,2
T
s,1
T
s
(x)
M
h
D
2
x
SemiBounded
Air Stream
Automobile
Exhaust Pipe
(a) Physical Model
(b) Thermal Circuit Models
(c) Temperature Distribution Along x Axis
Bounded
Exhaust
Gas Stream
0
L
T
f,h L
T
f,h 0
D
1
/2
D
1
/2
R
ku,h
R
k,hc
R
ku,c
T
f,h
(x)
T
x
x
L 0
x
x
T
f,h 0
T
f,h L
T
s 0
T
s L
Q
u x
T
f,c
T
f,h
(x)
T
s
(x)
A
ku,h
A
ku,c
A
ku,c
A
ku,h
(i) Local
(Mc
p
)
h
R
u L
Q
u L0
(ii) Integrated
Q
u,h 0
Q
u,h L
T
f,h 0
T
f,h L
T
f,c
(Mc
p
)
h
D
1
Figure Ex. 7.8. (a) Heat exchanger between a bounded exhaust gas stream and a semi
bounded cross ow. (b) Thermal circuit diagrams for the local and integral (integrated
along L) models. (c) Anticipated temperature distribution for the uid and tube wall
along the ow. Here T
s
) is the radially averaged tube wall temperature.
SOLUTION
(a) Figure Ex. 7.8(b)(i) shows the local thermal circuit diagram for a location x
along the exhaust pipe, with a different length Lx. Figure Ex. 7.8(b)(ii) shows
the integrated thermal circuit model.
(b) Figure Ex. 7.8(c) shows the anticipated temperature distribution along the x
axis. The highest temperature is the exhaust gas inlet temperature and the low
est temperature is the ambient air fareld temperature (which remains constant
because the uid is assumed to be semibounded, thus with its fareld tempera
ture unaffected). The local wall temperature shown here with a radially average
MassoudKaviany 2011
temperature T
s
)(x) is between the hot and the coldstream temperatures. Note
that no axial conduction is allowed along the pipe.
(c) Since the fareld temperature of the semibound uid (T
f .c
)
does not
change, this represents a stream with an innite heat capacity (i.e., C
c
or C
r
= 0). Then using the twostream heat exchange analysis, we have
from (7.85) and (7.86)
Q
u
)
L0
=
LT
max
R
u
)
L
=
(T
f .c
)
T
f .h
)
0
R
u
)
L
R
u
)
L
=
1
(
Mc
p
)
h
c
he
.
From Table 7.7, for all heat exchangers with C
r
= 0, we have
c
he
=
T
f .h
)
0
T
f .c
)
L
T
f .h
)
0
T
f .c
)
= (1 e
NTU
) Table 7.7.
These are the same as (7.25) and (7.22), with T
f .c
)
replacing T
s
). Then we
can predict T
f .h
)
L
from either of these analyses.
(d) The exit temperature is determined from (7.22) and (7.82), i.e.,
T
f .c
)
L
= T
f .h
)
0
[T
f .h
)
0
T
f .c
)
](1 e
NTU
).
where from (7.74) we have
NTU =
1
R
Y
(
Mc
p
)
min
=
1
R
Y
(
Mc
p
)
h
.
The overall resistance R
Y
is given by (7.87) and shown in Figure Ex. 7.8(b), i.e.,
R
Y
= R
ku.c
R
k.ch
R
ku.h
.
where from (6.124), Table 3.2, and (7.9), we have
R
ku.c
=
D
2
A
ku.c
Nu)
D.2
k
f .c
R
k.ch
=
ln(D
2
,D
1
)
2k
s
L
R
ku.h
=
D
1
A
ku.h
Nu)
D
1
k
f .h
.
For the coldstream resistance, we use the semibounded, crossow Nusselt
number from Table 6.3. The Reynolds number Re
D.2
is
Re
D.2
=
(u
f .c
)
D
2
f
.
MassoudKaviany 2011
From Table C.22, for air at T = 300 K, we have
f
= 1.566 10
5
m
2
/s Table C.22
k
f
= 0.0267 W/mK Table C.22
Pr = 0.69 Table C.22.
Then
Re
D.2
=
20(m/s) 0.644(m)
1.566 10
5
(m
2
/s)
= 5.619 10
4
> 10
4
turbulent ow regime.
From Table 6.3, we have
Nu)
D.2
= 0.027Re
0.805
D.2
Pr
1,3
= 158.9.
R
ku.c
=
0.044(m)
0.022(m) 1.5(m) 158.9 0.0267(W/mK)
= 0.1001
C/W.
For the conduction resistance, from Table C.16, we have for carbon steel AISI
4130
carbon steel : k
s
= 43 W/mK.
Then
R
k.ch
=
ln(0.044,0.041)
2 43(W/mK) 1.5(m)
= 0.0001743
C/W.
For the hotstream resistance, we rst determine the Reynolds number Re
D.1
Re
D.1
=
u
f .h
)D
1
f
.
From Table C.22, for air at T = 1,000 K, we have
f
= 1.173 10
5
m
2
/s Table C.22
f
= 0.354 kg/m
3
Table C.22
k
f
= 0.0672 W/mK Table C.22
Pr = 0.70 Table C.22
c
p.f
= 4,130 J/kg Table C.22.
The average velocity u
f .h
) is found for
u
f .h
) =
M
h
A
u.h
f .h
.
MassoudKaviany 2011
Then
Re
D.1
=
M
f .h
D
1
A
u.h
f .h
f .L
M
h
=
1
2
rpm
60
f
V
d
=
1
2
_
2,500
60
__
1
s
_
1.1(kg/m
3
) 2.2 10
3
(m
3
)
= 5.042 10
2
kg/s
Re
D.1
=
5.042 10
2
(kg/s) 0.041(m)
(0.0205)
2
(m
2
) 1.173 10
5
(m
2
/s) 0.354(kg/m
3
)
= 3.772 10
5
> 10
4
turbulent ow regime.
Note that this corresponds to u
f .h
) = 107.9 m/s, which is very high.
From Table 7.3, we have
Nu)
D.2
= 0.023Re
0.8
D.1
Pr
0.30
. T
s
 T
f .h
)
0
= 0.023 (3.772 10
5
)
0.8
(0.70)
0.30
= 597.8.
Then
R
ku.h
=
0.041(m)
(0.0205)(m) 1.5(m) 597.8 0.0672(W/mK)
= 0.01057
C/W.
Then
R
Y
= (0.1001 0.0001743 0.01057)(
C/W) = 0.1108
C/W.
Next
NTU =
1
0.1108(
C/W)(5.042 10
2
)(kg/s) 1,130(J/kg)
= 0.1583
T
f .h
)
L
= T
f .h
)
0
[T
f .h
)
0
(T
f .c
)
](1 e
NTU
)
= 600(
C) (600 35)(
C) (1 e
0.1584
)
= 600(
C) 82.77(
C) = 517.2
C.
COMMENT
Heat loss by surface radiation, and by conduction to various fasteners, reduces
T
f .h
)
L
below the above value. Note that the local pipe wall temperature is that
which allows for the local surfaceconvection heat ux from the hot stream q
ku.h
to be equal to the surfaceconvection heat ux to the cold streamq
ku.c
. Also note
that since (T
f .c
)
is constant, the analysis could have also been done using the
MassoudKaviany 2011
procedure outlined in Section 7.5 on the inclusion of the bounding surface in a
single uid stream.
7.7 References
[1] ASHRAE Handbook, 1997, Fundamentals, American Society of Heating,
Refrigeration and Air Conditioning Engineers, Atlanta.
[2] Carey, V.P., 1992, LiquidVapor PhaseChange Phenomena, Hemisphere Pub
lishing Company, Washington.
[3] Kakac, S., Bergles, A.E., andMayinger, F., 1981, Heat Exchangers, McGrawHill,
New York.
[4] Kaviany, M., 2001, Principles of Convective Heat Transfer, Second Edition,
SpringerVerlag, New York.
[5] Kays, W.M., and Crawford, M.E., 1993, Convection Heat and Mass Transfer,
Third Edition, McGrawHill, New York.
[6] Kays, W.M., and Perkins, H.C., 1985, Forced Convection, Internal Flow in Ducts,
in Handbook of Heat Transfer, Fundamentals, Second Edition, Rohsenow, W.M.,
et al., Editors, McGrawHill, New York. Also, Ebadian, M.A., and Dong, Z.F.,
1998, Forced Convection Internal Flow in Ducts, in Handbook of Heat Trans
fer, Third Edition, Rohsenow, W.M., et al., Editors, McGrawHill, New York.
[7] Lahey, R.T., Jr., Editor, 1992, Boiling Heat Transfer, Modern Developments and
Advances, Elsevier, Amsterdam.
[8] Mills, A.F., 1995, Basic Heat and Mass Transfer, Irwin, Chicago.
[9] Raithby, G.D., and Hollands, K.G., 1985, Natural Convection, in Handbook of
Heat Transfer, Fundamentals, Second Edition, Rohsenow, W.M., et al., Editors,
McGrawHill, New York.
[10] Rohsenow, W.M., 1985, Boiling, in Handbook of Heat Transfer, Fundamentals,
Second Edition, Rohsenow, W.M., et al., Editors, McGrawHill, New York.
[11] Rohsenow, W.M., and Grifth, P., 1985, Condensation, in Handbook of
Heat Transfer, Fundamentals, Second Edition, Rohsenow, W.M., et al., Editors,
McGrawHill, New York.
[12] Shah, R.K., and Mueller, A.C., 1985, Heat Exchangers, in Handbook of
Heat Transfer, Fundamentals, Second Edition, Rohsenow, W.M., et al., Editors,
McGrawHill, New York.
[13] Whitaker, S., 1983, Fundamental Principles of Heat Transfer, Krieger Publishing
Company, Malabar.
[14] White, F.M., 1994, Fluid Mechanics, Third Edition, McGrawHill, New York.
[15] Woschni, G., 1967, A Universally Applicable Equation for the Instantaneous
Heat Transfer Coefcient in the Internal Combustion Engine, SAE Paper
Number 670931.
7.8 Problems
7.9.7 General
PROBLEM 7.35. FUN
In this and the next two problems, we consider the expression for Nu
D
for a turbulent
uid stream owing in a smooth tube with a uniform wall temperature T
s
. The uid
stream is rendered in Figure Pr. 7.35. We begin by evaluating the fullydeveloped
uid temperature distribution T
f
(r. x).
MassoudKaviany 2011
D=2R
T
s
SurfaceConvection Area:
dA
ku
= Ddx
Fluid Flow CrossSectional
Area: A
u
= D
2
/4
AreaVelocity Averaged Fluid
Temperature: T
f
(x)
r
u
f
u
f
dx
x
A
ku
y
1 2
Flow
Crosssection
Averaged
Velocity:
Local Mean Fluid
Temperature: T
f
(r,x)
q
ku
Surface
Convection
q
u
q
u
Figure Pr. 7.35. Surfaceconvection q
ku
and convection q
u
heat transfer in a smooth tube with
turbulent uid ow. The wall is at uniform temperature T
s
.
The hydrodynamic conditions for the fullydeveloped turbulent regime are that
both the radial meanvelocity component :
f
, and the gradient of the axial mean
velocity component (u
f
,x), are zero everywhere, i.e.,
:
f
= 0.
u
f
x
= 0.
In contrast to the hydrodynamics, the heat transfer conditions lead to nonzero
x derivatives for the average (areavelocity) mean temperature T
f
),x, and the
local mean temperature T
f
,x. The local temperature distribution T
f
(r. x) changes
with x, but (T
s
T
f
),(T
s
T
f
)) does not, and under the thermallyfullydeveloped
condition, we have
x
_
T
s
T
f
T
s
T
f
)
_
= 0.
Noting that T
s
,x = 0, then we have
T
f
x
=
T
s
T
f
T
s
T
f
)
dT
f
)
dx
.
For the turbulent uid stream in a circular tube, under the assumptions of neg
ligible axial conduction
2
T
f
,x
2
= 0 and an axisymmetric temperature eld, the
energy equation (B.62) is written as
u
f
T
f
x
:
f
T
f
r
=
1
r
r
_
r
f
T
f
r
r:
/
f
T
/
f
_
.
where (c
p
)
f
:
/
f
T
/
f
is the turbulent convection heat ux in the r direction. As dened
in (6.58), :
/
f
T
/
f
can be expressed as
:
/
f
T
/
f
Pr
1
t
(:
/
f
2
)
1,2
t
T
f
r
=
t
T
f
r
t
Pr
1
t
(:
/
f
2
)
1,2
t
.
MassoudKaviany 2011
where
t
is the turbulent thermal diffusivity. Note that :
/
f
T
/
f
vanishes in the region
where there is no mean temperature gradient. Here (:
/
f
2
)
1,2
is related to the gradient
of the axial mean velocity through (:
/
f
2
)
1,2
t
[u
f
,r[. Then the turbulent thermal
diffusivity is dened as
t
Pr
1
t
t
2
[
u
f
r
[.
Note that
t
is not uniform across the tube cross section, since the turbulent
mixing length
t
, which is related to the size of the turbulent eddies, changes with
the radial location. The mixing length is smaller than the tube radius, is larger at the
central core of the tube (the turbulent core region), and decreases as the distance
to the wall decreases. Adjacent to the tube wall the ow is laminar and there are
no turbulent eddies (
t
= 0). This region is called the laminar thermal sublayer. The
transition region is called the buffer turbulent region.
Substituting for :
/
f
T
/
f
and using :
f
= 0, the energy equation becomes
u
f
T
f
x
=
1
r
r
_
r(
f
t
)
T
f
r
_
.
The temperature distribution satisfying the above equation for a constant wall
temperature will be used to nd the Nusselt number.
(a) Show that for the thermally fullydeveloped turbulent ow with a uniform wall
temperature we have
T
f
x
=
T
s
T
f
T
s
T
f
)
dT
f
)
dx
.
(b) Show that the mean uid temperature distribution can be expressed as
1 T
=
T
f
T
f
(r = 0)
T
s
T
f
(r = 0)
=
R
2
u
f
)(dT
f
),dx)
T
s
T
f
)
_
r
0
(r
)
r
(
f
t
)
dr
.
where
(r
) =
_
r
0
u
f
u
f
)
T
dr
. T
T
s
T
f
T
s
T
f
(r = 0)
. r
r
R
.
Hint: For part (b), integrate the energy equation over r = 0 to r = r, using the
result from part (a) to eliminate T
f
,x.
PROBLEM 7.36. FUN
In order to evaluate the integral in the relation for the radial distribution of the mean
uid temperature, derived in the previous problem [part (b)], the radial variation of
the turbulent thermal diffusivity
t
(r) is required. Assume that the turbulent Prandtl
number Pr
t
=
t
,
t
is equal to unity, and obtain the radial variation of the turbulent
thermal diffusivity, using
t
(r) found from the measured velocity distributions u
f
(r).
MassoudKaviany 2011
The momentum equation (B.61), written for the turbulent ow is
f
u
f
u
f
x
f
:
f
u
f
r
=
dp
dx
1
r
r
_
rj
f
u
f
r
ru
/
f
:
/
f
_
.
where u
/
f
:
/
f
is the turbulent shear stress (called the Reynolds stress).
The turbulent kinematic viscosity
t
is dened [similar to (6.58)] as
u
/
f
:
/
f
=
t
u
f
r
. where
t
=
j
t
f
.
Based on this relation, and because :
f
= 0 everywhere, the above momentum equa
tion becomes
1
f
dp
dx
=
1
r
r
_
r(
f
t
)
u
f
r
_
.
From this, the total shear stress (molecular plus turbulent) is
f
= (
f
t
)
u
f
r
.
Derive expressions for the radial distribution of turbulent kinematic viscosity
t
for (a) the region adjacent to the wall (laminar sublayer), (b) the region far away
from the wall (turbulent buffer zone), and (c) the turbulent core region, using the
measured mean axial velocity distributions
u
= y
0 y
5, laminar sublayer
u
= 3.05 5.0 ln y
5 y
= 5.5 2.5 ln y
=
y(
s
,
f
)
1,2
f
. u
=
u
f
(
s
,
f
)
1,2
.
Note that the turbulent shear stress varies linearly with the radius as =
s
r,R
(
s
is the wall shear stress). Here r Ry, where y is the distance from the wall.
(d) Plot the velocity distribution in the forms (u
versus y
) and (u
versus r
), and
the turbulent kinematic viscosity distribution (
t
,
f
versus r
), for Re
D
=10,000 and
Pr = 0.01. Consider the tube diameter D= 0.02 m. Comment on the results.
PROBLEM 7.37. FUN
Using the results from the previous two problems, the distribution of the mean uid
temperature can be determined using the turbulent thermal diffusivity distribution.
This uid temperature distribution is given as
1 T
=
T
f
T
f
(r = 0)
T
s
T
f
(r = 0)
=
R
2
u
f
)(dT
f
),dx)
T
s
T
f
)
(r
f
.
where
(r
) =
_
r
0
(r
)
r
(
f
t
),
f
dr
.
MassoudKaviany 2011
Iterations begin with the use of the constant q
s
solution for the temperature
distribution. After computing (r
=
T
s
T
f
T
s
T
f
(r = 0)
= 1
T
f
T
f
(r = 0)
T
s
T
f
(r = 0)
= 1
(r
)
(1)
.
This dimensionless temperature distribution is used in the expression for (r
),
as the next approximation, in an iterative procedure in which all the above compu
tations are repeated. A new dimensionless radial temperature distribution will then
be obtained from the above relation. When a sufcient convergence of the results is
achieved, the Nusselt number Nu
D
is calculated from (7.9) as
Nu
D
=
q
ku
D
(T
s
T
f
))k
f
=
f
f
1
(1)
.
(a) Show that, after the rst iteration, the dimensionless temperature distribution is
given by
T
= 1
(r
)
(1)
.
(b) Show that the Nusselt number is given by
Nu
D
=
f
f
1
(1)
.
Begin by writing the combined integraldifferential length energy equation for
the uid control volume.
(c) The Nusselt numbers predicted by the expression derived in (b) is curve tted,
with an accuracy of 6%, as,
Nu
D
= Nu
D
=
q
ku
D
(T
s
T
f
))k
f
= 5.0 0.025Re
0.8
D
Pr
0.8
.
for Prandtl numbers less than 0.1 (liquid metals).
Compare this correlation with the Nusselt number correlation obtained experi
mentally by Dittus and Boelter (correlation presented in Table 7.3 for uniformT
s
and
0.7  Pr  160), and the correlation suggested by Sleicher and Rouse (correlation
presented in Table 7.3 for uniform T
s
and Pr  0.1). Comment on the results.
PROBLEM 7.38. FUN.S
Using Example 3.23 and Example 7.4, the temperature distribution within a tumor
and the surrounding normal tissue, with perfusion cooling, can be determined. Use
the same acoustic intensity and conditions given in Example 3.23, and include per
fusion cooling using convection heat transfer with n
f
= 6 kg/m
3
s, and T
f
)
0
= T
N
.
(a) Write the energy equation for the interior nodes.
(b) Plot the temperature distribution to 0 x L, and compare these results to that
predicted distribution without perfusion cooling[Figure Ex.3.23(c)].
MassoudKaviany 2011
(c) Plot the temperature distribution to 0 x L, and increasing the acoustic inten
sity by a factor of 5. What is the maximum temperature?
PROBLEM 7.39
Adenosine Triphosphate (ATP) is the primary source of energy at the cellular
level for humans (see Problem 2.41) and is used for muscle contraction, there
is 14 kcal/mole of ATP released during the muscle contraction. The muscle cell
experiences an increase in temperature, so the blood ow (perfusion) can remove
the extra heat generated due to inefciency of muscle energy to work conversion.
A depiction of a muscle and its bers is shown in Figure Pr. 7.39(i) and Figure
Pr. 7.39(ii) renders a muscle ber (cell) with blood perfusion. The blood perfusion
is discussed in Example 7.4 and the convection heat transfer is expressed as
Q
u
)
L0
= ( nc
p
)
f
V
1
(T
1
T
f
)
0
), where n
f
is the volumetric blood ow rate and
c
f .p
is the blood heat capacity.Assume that the average mass of muscle in the body
is approximately 29 kg (about 40% of body mass, there are 640 muscles in the body,
and there is an average of about 500,000 bers per muscle).
Bone
Tendon
Muscle
Muscle Fiber (cell)
(i)
Perfusion (Blood Flow)
Muscle Fiber
(cell)
V
1
(Volume
of Fiber)
(ii)
Metabolic Energy
Conversion (S
r,c
)
1
.
Mechanical (Contraction)
Energy Conversion (S
m,c
)
1
.
T
1
T
f 0
Q
u L0
Figure Pr. 7.39. (i) Schematic of humanmuscle. (ii) Single muscle ber (cell) showing perfusion
effects.
(a) Draw the thermal circuit diagram for a muscle ber with blood perfusion con
vection cooling.
(b) Consider a relatively rigorous physical activity resulting in metabolic energy con
version of (
S
r.c
)
1
= 250 W [Figure Ex.1.13(d)]. Assume a steadystate heat transfer
with a portion of (
S
r.c
)
1
, resulting in muscle work (
S
m.c
)
1
contraction, and the rest in
Q
u
)
L0
. Considering an efciency of 25%, determine the energy Q
u
)
L0
and the
corresponding difference in temperature T
1
T
f
)
0
.
(c) Compare the temperature difference found in (b) with the measured value from
(T
1
T
f
)
0
) = 3 10
3
C and comment. What is the ATP consumption rate?
1
= 1,000 kg/m
3
, c
p.f
= 3,000 J/kgK, Lh
r.ATP
= 14 kcal/mole, n
f
= 6 kg/m
3
s,
M
1
= 29 kg, M
ATP
= 0.5032 kg/mole.
MassoudKaviany 2011
PROBLEM 7.40. FAM
During hockey games, the thickness of the ice surface decreases. This is remedied
by regularly ooding the surface using a machine called a zamboni, which deposits
a layer of liquid water on the ice surface, which then freezes. The surface is cooled
by a refrigerant as shown in Figure Pr. 7.40. The surface is a rectangle with sides
of length l and a. The liquid layer has a thickness of L
1
and uniform temperature
T
1
= T
sl
= 273.15 K. The ice has a thickness of L
2
and its lower surface has a
temperature T
2
. At a distance L below the ice, there are a N tubes of diameter D
buried in concrete that carry liquid ammonia with an inlet temperature T
f
)
0
. Each
tube carries a mass ow rate
M
1
and the tubes are separated by a distance n (center
to center). The tube surface temperature is assumed uniform and equal to T
s
. Heat
transfer between the liquid layer and the ambient is neglected.
(a) Draw the local and integrated (along l) thermal circuits. Neglect conduction
resistance though the tube wall.
(b) Determine the local resistance R
Y
. Find the concrete conduction resistance
assuming that the relations in Table 3.3(a) are applicable. Divide the pertube resis
tance by the number of tubes to obtain the total concrete resistance.
(c) Determine the rate of heat transfer from the liquid layer to the ammonia ow,
i.e.,
Q
u
)
l0
=
T
1
T
f
)
0
R
u
)
l
.
(d) Determine the rate of liquid to solid phase change
M
ls
.
(e) Determine how long it will take for the liquid layer to freeze. Comment on the
magnitude of your result. Is it a reasonable value?
(f) Calculate the outlet ammonia temperature. Comment on the use of a constant
wall temperature for the Nusselt number relationship.
a = 60 m, l = 26 m, L
1
= 0.5 mm, L
2
= 0.032 m, L = 3.8 cm, n = 0.1 m, D= 2.5 cm,
T
f
)
0
= 230 K,
M
1
= 0.1 kg/s, N = 600, k
c
=0.93 W/mK.
D
l
w
L
a
Ammonia
M
1
(Per Tube), T
f 0
Freezing Water Layer,
Temperature T
1
, Thickness L
1
Ice, Thickness L
2
S
ls
Tube Surface,
Tempeature T
s
IceConcrete Interface,
Temperature T
2
Concrete, k
c
Figure Pr. 7.40. A liquid water layer is solidifying on an ice surface.
MassoudKaviany 2011
Evaluate the ammonia properties from Table C.23 at T
f
)
0
. Take the density of
liquid water to be = 1,000 kg/m
3
. The heat of solidication of water is listed in
Table C.4, and the thermal conductivity of ice in Table C.15.
PROBLEM 7.41. FUN.S
A micro gas chromatograph (GC) is heated through electrical resistance (Joule
heating) to achieve a steadystate temperature T
gc
. Heat is transferred from it by
conduction through its substrate (printed circuit board) and by surface convection
(through the top surface), as shown in Figure Pr. 7.41(i). To reduce the heat transfer
from the top surface, a very thin glass plate is placed at a distance L
gap
from this sur
face. The schematic of the heat transfer problem is shown in Figure Pr. 7.41(ii). The
fareld ambient air temperature T
f .
and the substrate temperature T
s
are 24
o
C.
Use a square geometry of a a =3.25 cm3.25 cmfor the chromatograph and glass.
Glass
Sheet
GC
Chip
Glass
Sheet,
T
g
GC Chip,
T
gc
(Uniform
Temperature)
Substrate
Q
k
L
gap
Air, T
f,
g
Cellular Thermobuoyant Motion
Thermobuoyant Motion
(i) (ii)
T
s
a
a
S
e,J
.
Q
ku,g
Q
ku,gc
Figure Pr. 7.41. (i) Heated gas chromatograph with a glass sheet placed a distance L
gap
from
its top surface. There is heat conduction through the substrate (printed circuit board). (ii)
Rendering of the heat transfer from the heated gas chromatograph.
(a) To determine the substrate conduction heat transfer rate Q
k
, the glass sheet is
removed and it is found that T
gc
=92.5
o
Cat a Joule heating rate of 4.22 W. Neglecting
radiation, draw the thermal circuit diagram and determine the conduction heat
transfer resistance.
(b) Using the above conduction resistance (assuming it remains constant), explore
the optimal value for the air gap distance L
gap
(for highest T
gc
), using three different
values for L
gap
, namely 3.3, 7.8, and 27 mm. Also, draw the thermal circuit diagram.
(c) Compare the results of Part (b) above with the experimental results given in the
Table Pr. 7.41 below.
Evaluate the air properties at T = 330 K. Neglect radiation heat transfer.
Table Pr. 7.41. Measured chromatograph temperature
T
gc
for a few values of the air gap distance
L
gap
3.3 mm 7.8 mm 27 mm
T
gc
98
o
C 104
o
C 99
o
C
MassoudKaviany 2011
PROBLEM 7.42. FUN
The JouleThomson process can be used to cool a nonideal gas to extremely low
temperatures, by isenthalpically expanding it from a high to a low pressure. This
process is described by the JouleThomson coefcient,
j
JT
_
T
p
_
h
=
_
T
:
T
p
:
_
1
c
:
.
where j
JT
= j
JT
(T. p). This value is typically found from thermodynamic proper
ties, using h(T. p) or c
p
(T. p). However, over limited regions, this value can be
expressed as a polynomial in absolute temperature T and p.
For methane, a j
JT.o
can be found as
j
JT.o
= 216.48 3.3271T 0.0205T
2
5.734 10
5
T
3
6.04 10
8
T
4
.
where j
JT.o
is the value of j
JT
at p
o
= 0.1 MPa in the reference temperature (T is
in K, p is in MPa). For pressures above p
o
, we can use
j
JT
= j
JT.o
0.1707(p p
o
) 0.0405(p p
o
)
2
.
where p is the pressure in MPa. Note that this is only valid for 115  T  300 K,
0.1  p  10 MPa (excluding the region above the inversion or saturation curves).
Figure Pr. 7.42 shows the inversion curve (where j
JT
= 0) and the saturation curve
for methane in a pT diagram.
0
2
4
6
8
10
100 125 150 175 200 225 250 275 300
T, K
p
,
M
P
a
Inversion
Curve
Saturation
Line
Methane, CH
4
Figure Pr. 7.42. pT diagram for methane, showing the inversion and saturation curves.
(a) Using MATLAB, integrate the differential equation for dT,dp = j
JT
(T. p),
by combining the expressions for j
JT.o
and j
JT
given above. The equation will be
parameterized with p. [Hint: to model an expansion from high to low pressure, use
a third parameter x, and set p = (p
max
p
o
) x.] Solve this equation starting from
initial conditions of p
max
= 10 MPa (i.e., p = 9.9 x; 0  x  9.9) and T(0) = 300
K. What is the nal temperature reached? Plot the resulting isenthalpic curve in a
pT diagram.
MassoudKaviany 2011
(b) Next, consider a staged JT cooler, with a second high pressure methane stream
at T = 300 Kprecooled by the rst (the rst having a larger mass owrate). Assume
a 95% effectiveness for a parallelow heat exchanger. Find the inlet temperature
for the second stream as it exits the heat exchanger. What is the nal temperature
of the second stream as it expands to 0.1 MPa? Plot the resulting isenthalpic curves
for both expansions in the pT diagram.
PROBLEM 7.43. FAM
In an internalcombustion engine with heat regeneration, a moving ceramic foam
is used to recycle the combustion ue gas from the last cycle to heat the air
intaked into the cylinder before combustion. The regenerator is shown in Figure
Pr. 7.43. This is called the air heating period. In the 720
o
crank angle cycle, this is
followed by heating of the foam with the combustion ow gas (this is called the
regeneration period). Assuming that the foam has a uniform and time dependent
temperature, we can estimate the rise T
s
(t = t
f
) T
s
(t = 0) in the foam temper
ature during the passage of hot gas through it (i.e., during the regeneration period t
f
).
(a) Draw the thermal circuit diagram.
(b) Determine the rise in temperature of the foam.
Use c = 0.7, A
ku
,V = 10. 000 m
2
/m
3
, L = 8 mm, D= 12.5 cm,
M
f
= 1 kg/s, t
f
=
10 ms, T
f
)
0
= 1. 500 K, T
s
(t = 0) = 1. 000 K,
s
= 2. 000 kg/m
3
, c
p.s
= 800 J/kgK.
Take the gas properties to be of air at 1,500 K.
T
f
0
(Mc
p
)
f
T
f
L
(Mc
p
)
f
Initial Temperature, T
s
(t = 0)
Uniform Foam
Temperature, T
s
(t)
Heat Storage Ceramic Foam
D
L
Porosity
Specific Surface Area A
ku
/V
Diameter
Figure Pr. 7.43. A foam regenerator used inside the internal combustion cylinder. The foam is
porous and its geometry parameters are shown. The hot gas ows in during the foam heating
(regenerator) period. The foamis assumed to be at uniform, but time dependent temperature.
MassoudKaviany 2011
8
Heat Transfer in Thermal Systems
It is by logic that we prove, but by intuition we discover.
H. Poincar e
As the births of living creatures at rst is illshapen, so are the innovations,
which are the births of time.
F. Bacon
Heat transfer analysis addresses control of heat transfer, in conjunction with
energy storage and conversion, for innovative applications and optimized perfor
mance in thermal systems. This heat transfer can occur at various, cascading length
scales within the system. The analysis is done by modeling the transport, storage,
and conversion of thermal energy as thermal circuits. The elements of these circuits,
i.e., mechanisms and models of resistances, storage, and energy conversion, have
been discussed in Chapters 2 to 7. We now consider their combined usage in some
innovative, signicant, and practical thermal systems. Figure 8.1 renders thermal
engineering analysis, which aims at determining the feasibility of thermallybased
innovative ideas. Innovation starts with a need for signicant ideas in any of the
applications areas we discussed in Charts 1.3(a) to (e), and it seeks to answer a par
ticular current need by using insightful (and perhaps not readily apparent) solutions.
The insight is achieved by knowledge of fundamentals, analysis, and experience.
These innovative ideas are then analyzed for feasibility, and if proven practica
ble, they are designed, using the thermal system analysis covered in the earlier
chapters.
We begin by addressing the primary thermal functions of heat transfer media
and bounding surfaces. Then we summarize the elements of thermal engineering
analysis. These include the need for assumptions and approximations for reduction
of the physical models and conceptual thermal processes to thermal circuit models.
Next we give several examples, in detail, to demonstrate the application of the
fundamentals and relations developed in the text, along with the use of software.
The examples are selected for their innovative potentials.
MassoudKaviany 2011
Control
Surface A
Control
Volume V
q
r
q
u
, q
ku
q
q
q
k
(a) Physical Model
(b) Thermal Circuit Model
= c
p
V + S
dT
dt
Control of q = q
k
+ q
u
+ q
r
(i.e., Magnitude, Direction, and SpatialTemporal
Variation of Each Mechanism) by Selection of Thermal Resistances, in
Conjunction with Energy Storage and Conversion and for Optimal Performance.
Objective:
Next Larger Control
Volume and Surface
Thermal System
A
(q
.
s
n
) dA
Q
A
=
A
(q
.
s
n
) dA
c
p
V + S
dT
dt
Next Larger Control
Volume and Surface
c
p
V + S
dT
dt
Q
A
=
A
(q
.
s
n
) dA
Thermal Engineering Analysis
Analysis and Design
Innovative
Thermally Based
Ideas in An
Application Field
Figure 8.1. Control of heat transfer at various, cascading length scales in thermal systems, and
modeling of systems as thermal circuits. (a) Physical model. (b) Thermal circuit model.
8.1 Primary Thermal Functions
8.1.1 Primary Functions of Heat Transfer Media
We now consider heat transfer media selected (i.e., designed and manufactured)
for particular thermal functions. The primary thermal functions can be categorized
as thermal capacitors, thermal conductors, thermal convectors, thermal converters,
thermal exchangers, thermal insulators, and any combination of these functions. This
classication is listed in Chart 8.1.
MassoudKaviany 2011
Primary Functions of Heat Transfer Media
Thermal
Capacitor
Electromagnetic
Thermo
electrical
Media
Thermo
optical
Media
Frictional,
Displacement,
Pressure
Reaction Others Phase
Change
Fluid
Solid
Fluid
Fluid
Direct Indirect
Thermo
magnetic
Media
Thermo
chemical
Media
Thermo
nuclear
Media
Thermo
mechanical
Media
Thermal
Conductor
Thermal
Exchanger
R
0
Thermal
Convector
R
k,u
0
A
ku
V
, (c
p
k)
R
k
Thermal
Converter
Combined
Thermal
Functions
Thermal
Insulator
R
0
R
0
1/ 2
Chart 8.1. A classication of the primary functions of heat transfer media.
Thermal capacitors store and release sensible heat and generally have a large
specic surface area A
ku
,V (such as porous solids and discontinuous solid surfaces).
In rapid heatingcooling, the storage media should allow for complete penetration
of the heat front. This is achieved by having thin solids and large thermal effusivity
(c
p
k)
1,2
.
Good thermal conductors are generally metals (or quartz or carbon), and are
nonporous. Media that are good thermal conductors, but poor electrical conductors,
include nonoxide ceramics like aluminum nitride.
Good thermal convectors use high speed u
f
and large (c
p
)
f
. Liquids have high
(c
p
)
f
, but achieving high speed for them is more difcult than for gases. Liquids
also generally need to be contained.
Among the thermal converter media are the thermoelectrical, thermomagnetic,
thermooptical, thermomechanical, thermochemical, and thermonuclear media. For
example, thermal sensors have detection media, where these thermal energy conver
sions are used for sensing. An example of a thermal sensor is the thermoelectrical
sensor, e.g., the infrared surfacetemperature sensor discussed in Section 4.3, or the
hotwire anemometer discussed in Section 6.8.4. Another example of application of
thermal converter media is the thermoelectric cooler.
Thermal exchangers (uidsolid and uiduid heat exchangers) allow for heat
exchange with the least overall resistance. The overall resistance includes various
conduction and convection resistances.
Ideal insulators allowfor innite conduction resistance, and near ideal insulation
is achieved by gases and then by porous solids.
A medium may also have combined thermal functions, i.e., may involve more
than one of these functions.
MassoudKaviany 2011
Primary Thermal Functions of Bounding Surfaces
Heat Transfer Mech
anism Converter
(Both Sides of Sur
face Have q = 0)
Thermal Barrier
(One Side of Sur
face Has q 0)
Thermal
Converter
of Energy
Combined
Thermal
Functions
One or Both Sides
Dominated by
Conduction Heat
Transfer
One or Both Sides
Dominated by Sur
faceConvection
Heat Transfer
One or Both Sides
Dominated by Sur
faceRadiation
Heat Transfer
SurfaceConvection
to SurfaceRadia
tion Converter
SurfaceConduction
to SurfaceRadiation
Converter
Reradiators
Others
Mechanical
Phase
Change
Electro
magnetic
Reaction Others
Others
R
k
0
r
= 1, R
r,
0,
A
r
R
r,F
= 1
R
ku
0
Chart 8.2. A classication of the primary thermal functions of bounding surfaces.
8.1.2 Primary Thermal Functions of Bounding Surfaces
The bounding surfaces are also selected for their thermal functions. Chart 8.2 gives
a classication of the primary thermal functions of bounding surfaces. These are
divided into heat transfer mechanism converters, heat barriers, energy converters,
and combined functions.
For example, in a converter of the surface heat transfer mechanism, surface
convection and radiation on the one side of the surface (e.g., a gaseous ambient) may
change (or convert) to surface conduction on the other side (e.g., a solid substrate).
In heat barriers (or heat returners), the heat arriving from one side is mostly
returned at the same side, while a negligible amount of heat ows to the other side.
An example of a heat barrier is a multilayer radiation shield that would reduce, to a
negligible amount, the radiation heat transfer transmitted through the shields.
In energy converters, surface energy conversion occurs (e.g., surface friction
heating).
The surface may also have a combined thermal function involving more than
one of these functions.
8.1.3 Heat Transfer Material
In Chapters 2 to 7, we have dealt with various heat transfer materials. They have
been in uid (gas or liquid) form, in solid form, or in multiphase form. We have also
addressed many of the thermochemical and thermophysical properties.
Chart 8.3 gives a classication of the signicant thermodynamic and transport
properties, and other properties considered in the selection of heat transfer material.
The properties listed in Chart 8.3 include some of the energy conversion prop
erties. Among the properties listed are the melting temperature T
sl
for solids and
MassoudKaviany 2011
Thermochemical and Thermophysical
Properties of Heat Transfer Medium
Gas Liquid Solid
Multiphase
k
g
Nonionized
Thermal
Plasmas
Cold
Plasmas
Ionized
Condensation
T
gl
Flow
Evaporation
T
lg
Solidification
T
ls
Flow
k
l
k
s
(c
p
)
g
g
(c
p
)
l
(c
p
)
s
l
r
,
r
,
ex
g
Energy Conver
sion to and from
Thermal Energy
Reactivity
Melting
T
sl
ac
Organic
Metals Nonmetals
Inorganic
Electrons
and Holes
Phonons
Gas Solid Liquid
e
,
ec
GasLiquid
A
lg
V
e
,
ec
,
ac
ex
Nuclear
Reaction
Electronic
Bond Reaction
Nuclear
Reaction
Electronic
Bond Reaction
Nuclear
Reaction
Electronic
Bond Reaction
Chemical and
Physical Bond
h
r,A
, h
ij
Electromagnetic
Electrical
Magnetic Others
Others
Mechanical
Optical
, , ,
A
sf
V
, ,
SolidFluid
LiquidLiquid
(Immiscible Liquids)
, , k
k k
Chart 8.3. A classication of thermochemical and thermophysical properties of heat transfer
medium.
the boiling temperature T
lg
for liquids. Reactivity (or lack of it, i.e., inertness) is also
also important in selecting solids, liquids, and gases.
In multiphase systems, for the liquidgas systems the surface tension and the
void fraction (which is the same as the porosity used for the soliduid system), and
for the solid uid systems the porosity, are important.
For solid surfaces, the surface emissivity can be important and is altered by
various surface treatments (oxidation or polishing).
8.2 Thermal Engineering Analysis
8.2.1 Simplications, Approximations, and Assumptions: Modeling
In order to transform an actual thermal system into a solvable problem, thermal
modeling is needed. A thermal model is a simplied description or conception of
a thermal system and is used to study its characteristics and to make predictions
about its behavior. Simplications, approximations, and assumptions made in ther
mal modeling and analysis are discussed below.
MassoudKaviany 2011
(A) Simplications
Identify the dominant thermal nonequilibria in the system by marking the media
and/or bounding surfaces having the largest heat transfer rates. This exclusion of
the less important thermal nonequilibria is the most important step in the thermal
modeling.
(B) Approximations
Approximate the geometry to a rectangular or radial systems with as much symmetry
as possible. For surface heat transfer, much of the existing relations are for planar
surfaces, cylinders, spheres, and tubes. Then determine the dominant heat transfer
direction. In solids, the onedimensional heat ow approximation allows the use of
the onedimensional conduction resistances.
(C) Assumptions
Whenever the dominant heat transfer is between a medium and its surroundings,
assume a uniform temperature within the heat transfer medium. Whenever the
dominant heat transfer is from a bounding surface to its surroundings, assume a
uniform boundingsurface temperature. Also, if the time variations in temperature
are less signicant than the spatial variations, assume a steady state. If the variation
of the energy conversion rate within the medium is less than the rate of surface heat
transfer, then assume a uniform heat generation rate.
Once all signicant volumetric and surface thermal nodes are identied, connect
them using the appropriate resistances.
8.2.2 Nodal Energy Equation
Each thermal node T
i
refers to a uniform temperature volume or surface. The ther
mal circuit model is a collection of such nodes connected (i.e., interacting) together.
An energy equation is written for each thermal node T
i
and accordingly this will be
a volumetric or a surface energy equation.
Once the thermal circuit model is ready, we use the appropriate relations, from
Chapters 2 to 7, for the energy storage, energy conversion, and surface heat transfer
rates.
Table 8.1 (graphical) and Table 8.2 (textual) summarize the thermal resistances
identied in Chapter 3 to 7 and show the relation between heat ows, potentials,
and resistances, and the appropriate thermophysical, geometric, ow, and related
parameters. For gray, diffuse opaque surfaces, in addition to the temperature nodes
T
i
, the emissivepower node E
b.i
, and the radiosity node (q
r.o
)
i
, are also used. For
nongray surfaces, surface radiation is included in
S
i
. Table 8.1 gives a graphical
representation of these resistances and the heat ow associated with them.
MassoudKaviany 2011
Table 8.1. Summary of internodal thermal resistances and internodal sources
Transient Conduction Resistance R
ck
(t)
Radiation, SurfaceGrayness Resistance (R
r,
)
i
Radiation, ViewFactor Resistance (R
r,F
)
i
ConductionConvection (Unbounded Fluid Stream)
Resistance R
k,u
SurfaceConvection (SemiBounded Fluid Stream)
Resistance
Average Convection (Bounded Fluid Stream)
Resistance
R
ck
(t)
Q
ck
(t)
T
i
(t = 0) T
i
(R
r,
)
i
Q
r,i
(q
r,o
)
i
E
b,i
R
r,
Q
r,ij
E
b, j
E
b,i
(R
k,F
)
ij
Q
r,ij
(q
r,o
)
j
(q
r,o
)
i
R
k,u
(Q
k,u
)
ij
T
j
T
i
T
j,
=T
j
T
i
T
i
R
ku L
Q
ku L
=Q
ku,ij
R
ku L
R
u L
T
j L
=T
j,L
T
j 0
=T
j,0
Q
u L0
=Q
u,ij
R
u L
Q
A,i
u
f
Internodal Energy Conversion S
ij
T
i
T
j
u
f,
u
f
u
f
NoResistance Connection Between Surface Temperature T
i
and Emissive Power E
b,i
No Resistance
T
i
E
b,i
=
SB
T
i
4
Q
r,i
SteadyState Conduction Resistance R
k,ij
R
k,ij
Q
k,ij
T
j
T
i
S
ij
Q
A, j
All Blackbody Enclosures and Graybody TwoSurface
Enclosures (and ThreeSurface Enclosures with Third
Surface Having Q S = 0), Overall Radiation Resistance is
R
r,
MassoudKaviany 2011
Table 8.2. Summary of heat transfer resistance relations for node T
i
Mechanism Potential,
ij
Heat Flow Rate, Q
t,ij Resistance, R
t,ij
SteadyState
Conduction
Conduction
Radiation
Opaque
Enclosures
Bounded
FluidStream
Convection
SemiBounded
FluidStream
Convection
Unbounded
FluidStream
Convection
Q
k,ij
= , W
T
i
T
j
R
k,ij
Q
k,ij
= , W
T
i
T
j
(t = 0)
R
ck
(t)
ij
T
i
T
j
(t = 0) , C
T
i
T
j
, C
T
i
to E
b,i
=
SB
T
i
4
E
b,i
(q
r,o
)
i
, W/m
2
(q
r,o
)
i
(q
r,o
)
j
, W/m
2
E
b,i
E
b,j
, W/m
2
R
k,ij
= , C/W
L
A
k
k
R
k,ij
= , C/ W
ln(R
j
/ R
i
)
2Lk
R
k,ij
= , C/W
(1/R
i
1/ R
j
)
4k
finite slab:
cylindrical shell:
spherical shell:
R
rck
(t) = , C/W
(t)
1/2
A
k
(c
p
k)
1/2
Q
r,i
, W
Q
r,i
= , W
E
b,i
(q
r,o
)
i
(R
r,
)
i
Q
r,i
= , W
(q
r,o
)
i
(q
r,o
)
j
(R
r,F
)
ij
Q
r,ij
= , W
E
b,i
E
b,j
R
r
for diffuse blackbody n
surface having S Q = 0)
no resistance
(R
r,
)
i
= , 1/m
2
(1
r
)
i
(A
r
r
)
i
(R
r,F
)
ij
= , 1/m
2
1
A
r,i
F
ij
T
i
T
j
, C
T
i
T
j
, C
T
i
T
j,0
, C
(Q
k,u
)
ij
= , W
T
i
T
j
(R
k,u
)
ij
= , W
T
i
T
j
R
ku L (or D)
Q
ku,ij
= Q
ku L (or D)
Q
u,ij
= Q
u L0 = , W
T
i
T
j,0
R
u L
(R
k,u
)
ij
= , C/W
L
A
k
k
e
Pe
L
1
Pe
L
e
Pe
L
Pe
L
=
u
f
L
f
R
ku,ij
= R
ku L (or D)
= , C/W
L (or D)
A
ku
Nu
L (or D)
k
f
Nu
L (or D)
from Tables 6.3 to 6.6
Nu
D
Two Streams: Table 7.7
R
u,ij
= R
u L
=
1
(Mc
p
)
f he
.
= , C/ W
1
(Mc
p
)
f
(1 e
NTU
)
.
NTU = ,
A
ku
Nu
D
k
f
(Mc
p
)
f
.
MassoudKaviany 2011
T
i
Control Surface A
i
and Volume V
i
Thermal Node T
i
and its Associated Surface Heat Transfer
Resistances, and Energy Storage and Conversion
Volumetric Energy Storage
and Conversion
R
k,ij
R
ck
(t)
R
ku L
R
k,u
(R
r,
)
i
(R
r,F
)
ij
(R
r,
)
j
E
b,j
T
j
T
i
(t = 0)
(q
r, o
)
j
Q
r,ij
(q
r, o
)
i
E
b,i
SB
T
i
4
Q
k,ij
Q
ck
(t)
Q
r,i
Q
r,i
(Q
k,u
)
ij
Q
u L0
=Q
u,ij
T
j
u
f
u
f,
u
f
Q
i
(c
p
V)
i
+ S
i
T
j
Conduction
Bounding
Surface Node
Transient
Conduction
Other Surface
Heat Flow Rates
dT
i
dt
Internodal Energy
Conversion Node
S
ij
Energy Conversion
Between Nodes i and j
EmissivePower Node
ConductionConvection
BoundingSurface Node
FarField Fluid Node
SemiBounded
Fluid Motion
BoundedFluid Exit Node
BoundedFluid Motion
BoundedFluid Inlet Node
A Composite
Resistance R
Node
Radiosity Node
Radiosity
Node
u
f
R
u L
R
Q
ku L
=Q
ku,ij
T
j L
=T
j,L
T
j 0
=T
j,0
Q
ij
T
j
T
j
T
j
Q
A,i
Q
A, j
No Resistance
No Resistance
Figure 8.2. A thermal node T
i
and its associated energy storage and conversion and thermal
interactions with neighboring thermal nodes.
Figure 8.2 shows the circuit diagram for a volumetric thermal node T
i
, and
the associated control surface A
i
, control volume V
i
, volumetric energy storage
conversion terms (c
p
V)
i
dT
i
,dt
S
i
, surface heat transfer not represented by
resistances Q
i
, surface radiation heat transfer Q
r.i
, and the heat transfer represented
by thermal resistance Q
t.ij
=
(T
i
T
j
),R
t.ij
. Alsoshowninthe internodal energy
conversion
S
ij
.
MassoudKaviany 2011
For each thermal node T
i
, and from(2.71), (2.73), (3.50), (4.41), (4.42), and (5.73)
(with Q
loss
= 0), we have
Q [
A.i
= Q
i
Q
r.i
j =1
Q
t.ij
= (c
p
V)
i
dT
i
dt
S
i
volumetric
node T
i
(8.1)
Q [
A.i
= Q
i
Q
r.i
j =1
Q
t.ij
=
S
i
surface
node T
i
(8.2)
Q
t.ij
=
T
i
T
j
R
t.ij
thermal
resistance R
t.ij
(8.3)
Q
r.i
=
E
b.i
(T
i
) (q
r.o
)
i
_
1 c
r
A
r
c
r
_
i
. E
b.i
=
SB
T
4
i
node E
b.i
in gray,
diffuse opaque
enclosures
(8.4)
E
b.i
(T
i
) (q
r.o
)
i
_
1 c
r
A
r
c
r
_
i
=
j =1
(q
r.o
)
i
(q
r.o
)
j
1
A
r.i
F
ij
node (q
r.o
)
i
in gray,
diffuse opaque
enclosures
(8.5)
Q [
A.i
Q [
A.j
=
S
ij
internodal energy
conversion.
(8.6)
Then for each node we have one energy equation. In addition for each termR
t.ij
. Q
i
,
and
S
i
, we write a relation. Note that the subscript t indicates the various mechanisms
of heat transfer and that R
t.ij
s are listed in Table 8.2.
The various terms in the volumetric energy equation are summarized in
Chart 8.4.
Figures 8.3 and8.4 showthe anticipatedmagnitudes of the resistances andenergy
conversion rates. The range of magnitude for eight thermal resistances (product of
surface and resistance A
t
R
t
) is shown in Figure 8.3. These are conduction resis
tance across a layer, conductioncontact resistance, transientconduction resistance,
conductionconvection resistance, surfaceconvection resistance, average convection
Other Surface
Heat Flow
(Prescribed)
Surface Heat Flow
by Conduction,
Convection, and Radiation
Energy Equation for Volumetric Thermal Node T
i
Thermal Energy
Storage / Release
Energy Conversion
to or from Thermal Energy
R
k,ij
, R
ck
R
k,u
, R
ku
, R
u
R
r,
, R
r, F
Overall
Resistance
R
Chemical and
Physical Bond
Electro
magnetic
Mechanical
Others
dT
i
dt
T
i
T
j
R
t,ij
Q
A,i
= Q
i
+ Q
r,i
+ = (c
p
V)
i
+ S
i
j
Chart 8.4. Various terms in the volumetric thermal nodal energy equation for node T
i
.
MassoudKaviany 2011
A
t
R
t
,
o
C/(W/m
2
)
A
r
R
r
0 10
8
10
6
Very Thin, High
Conductivity
Conductor
Physical
Bonding
ConductionContact Resistance, A
k
R
k,c
ConductionConvection Resistance, A
k,u
R
k,u
ViewFactor
Resistances, A
r
R
r,F
Unity View
Factor
Zero
View
Factor
SurfaceGrayness Resistance, A
r
R
r,
Black
body
Air, with
Sonic Speed
Water, with
Sonic Speed
SurfaceConvection Resistance, A
ku
R
ku
SteadyState Conduction Resistance, A
k
R
k,12
Ideal
Vacuum
Non
emitter
10
2
Ideal
Insulator
Ideal
Conductor
1 10
4
Average Convection Resistance, A
u
R
u L
Air, with
Sonic Speed
Water, with
Sonic Speed
TransientConduction Resistance, A
k
R
ck
(t)
Phase
Change
Substantial
Vapor Removal
Enhancement
Boundary
Scattering
Range of Magnitude of Product of Surface Area and
Thermal Resistance, for Various Resistance Mechanisms
Figure 8.3. Summary of the magnitudes of the product of the surface area and the thermal
resistance for various resistance mechanisms.
resistance, surfacegrayness resistance, and viewfactor resistance. The lower limit on
all these resistances, except for the viewfactor resistance, is zero, although for most
of them it may take a signicant effort to achieve a zero resistance. The upper limit
on the resistances is innity and is achieved by vacuum or by having zero emissivity
and view factor. In general, for most practical problems, resistances smaller than
10
5
C/(W/m
2
) are considered negligible.
The range of magnitude for fteen volumetric and surface energy conversion
rates is shown in Figure 8.4. These are marked as resulting in heat absorption or heat
release and being either an electromagnetic, chemicalphysical bond, or mechanical
energy conversion. Very high heat release rates are possible with surface absorption
(laser irradiation, for a short period), chemical reaction, or Joule heating. Very high
heat absorption rates are possible with evaporation or by surface emission from very
high temperature surfaces.
8.2.3 Simultaneous Inclusion of Various Heat Transfer Mechanisms
Some of the resistances representing various heat transfer mechanisms have been
obtained while allowing for only a limited number of mechanisms. For example,
in determining the uid conductionconvection resistance R
k.u
, we excluded the
MassoudKaviany 2011
S
e,m
/ V, Microwave Heating
S
m,
/ V , Viscous Heating
S
ij
/ V, Evaporation or
Melting Cooling
S
m,ac
/ V, Ultrasound (Acoustic) Heating
S
m,F
/A, Friction Heating
S
e,P
/A, Peltier Cooling
Heat Release
Range of Magnitude of Surface and Volumetric Thermal Energy
Conversion Rate, for Various Conversion Mechanisms
Heat Absorption
E
l
e
c
t
r
o
m
a
g
n
e
t
i
c
t
o
/
f
r
o
m
T
h
e
r
m
a
l
E
n
e
r
g
y
C
h
e
m
i
c
a
l
a
n
d
P
h
y
s
i
c
a
l
B
o
n
d
t
o
/
f
r
o
m
T
h
e
r
m
a
l
E
n
e
r
g
y
M
e
c
h
a
n
i
c
a
l
t
o
/
f
r
o
m
T
h
e
r
m
a
l
E
n
e
r
g
y
S
r,c
/ V, Exothermic
Chemical Reaction
S
r,c
/ V, Endothermic
Chemical Reaction
S
e,J
/ V, Joule Heating (Solid or Fluid)
S
e,
/ V, VolumetricAbsorption Heating
S
r,n
/ V , Nuclear Reaction
S
e,
/A, SurfaceAbsorption Heating
S
m,p
/ V, GasExpansion Cooling
S
i
/A, W/m
2
S
i
/V, W/m
3
10
6
10
3
10
3
10
6
0 10
9
S
e,
/A, Surface Emission Cooling
S
ij
/ V, Condensation and
Freezing Heating
Figure 8.4. Summary of the magnitude and the sign of various surface and volumetric energy
conversion mechanisms (to and from thermal energy).
surfaceconvection heat transfer, and vice versa in the evaluation of the average
convection resistance R
u
)
L
.
For nitesmall length, these resistances can be used simultaneously without
a loss of accuracy (posed as end of chapter problems in Chapters 3, 5, and 7).
Therefore, we suggest the use of resistances and their relationships given in Table 8.2
evenwhenconduction, convection, surface convection, andsurface radiationall need
to be included (because they are all signicant relative to each other), along with
energy storage and conversion. We just need to discretize the media into nitesmall
volumes.
8.2.4 Optimization
At times we need to minimize or maximize the heat ow rate or temperature while
varying some of the variables affecting them. This problemis called optimization and
the quantity to be optimized is posed in a form designated as f (called the objective
MassoudKaviany 2011
function) such that the minimization of f is sought. This means that if Q
ku
)
L
is to
be maximized, then the objective function will be set as f = 1,Q
ku
)
L
.
In addition to the objective function, which is minimized, there may be stated
constraints. For example, the temperature, given in absolute scale (K), should always
be equal to or greater than zero. The constraints may also arise from some practical
limitation; for example, the sonic speed limit becomes a constraint when creation of
supersonic ow is either not possible with the hardware available, or is undesirable.
There are various mathematical schemes for the search of the minimum value
for the objective function f . These are reviewed in [13]. This reference also provides
the algorithms and the computer program (as a subroutine).
8.3 Examples
The following examples illustrate the use of the materials covered in the text in
solving some thermal engineering problems. The equations are numerically solved
with solvers such as MATLAB.
EXAMPLE 8.1. DES.S
HEAT PIPE WITH WICK: A LIGHTWEIGHT CONDUCTOR
The heat pipe is a closed, lightweight device that allows for a large heat transfer
rate within it by offering a very small thermal resistance R
Y
. The light weight is
accomplished by using a relatively large void space within it. Almost all of the
R
Y
is due to delivery and removal of heat to and from the heat pipe (generally
surfaceconvection resistance), while within the pipe, the resistance is almost
negligible.
Figure Ex. 8.1.1(a) depicts the physical and thermal circuit models of a heat
pipe with a wick. The wick allows for the return of the liquid against gravity
(i.e., the tilt angle is , and for ow against gravity, we have 0   180
).
(a) For the physical and geometrical properties listed in Table Ex. 8.1.1, showthe
thermal circuit diagramand determine the resistances due to surfaceconvection
R
Y.e
and R
Y.c
, tube conduction R
k.e
and R
k.c
, and wick conduction R
k).e
and
R
k).c
.
(b) Formulate the variation of the vapor pressure p
g
and the vapor temperature
T
g
as a function of the heat input Q(W).
Assume that the evaporation and condensation resistances are negligible.
Also assume no pressure drop between the evaporator and condenser (i.e.,
uniform vapor temperature T
g
= T
lg
).
SOLUTION
(a) The thermal circuit diagram for this heat pipe is shown in Figure Ex. 8.1.1(b).
By examining Figure Ex. 8.1.1(b), we note that all heat transfer resistances can be
combined into parallel and series arrangements. Then from Figure Ex. 8.1.1(b),
MassoudKaviany 2011
R
k ,c
R
k,c
(R
ku, f
)
c
L
f
l
(R
ku, b
)
c
u
f,c
T
f,c
k
f,c
u
f,e
T
f,e
k
f,e
(T
f,
)
c
Vapor Flow
(in Core)
Adiabatic
Region
Hot Stream Cold Stream
S
ep
arato
r
Condensation
Region
Evaporation
Region
Heat Transfer in a Heat Pipe with Wick
(a) Physical Model
(b) Thermal Circuit Model
Liquid Flow
(in Wick)
Closed Loop
Motion of Vapor
and Liquid
Surface
Evaporation
Core (Void, for
Vapor Transfer)
Wick
(for Liquid
Transfer)
Fin
g
L
e
L
a
D
2
/2
D
1
/2
L
c
Surface
Condensation
(in Pores)
q
u
q
ku
q
u
q
ku
q
k
q
k
R
k,e
(R
ku, f
)
e
(R
ku, b
)
e
(T
f,
)
e
Q
ku
= S
gl
M
lg
h
lg
Q
u
T
lg
(p
c
) = T
lg
(p
e
)
(Negligible Pressure Drop)
T
lg
(p
e
)
Q
ku
= S
lg
M
lg
h
lg
Pipe
Wall
Fin Base
Fin
Vapor Flow
in Core
Q
c
= Q
Q
e
= Q
Wick
Liquid Flow in Wick
q
k
R
k ,e
q
k
Figure Ex. 8.1.1. (a) A schematic of heat pipe with wick. (b) Its thermal circuit model.
for the heat ow rate Q, we have
Q =
T
f .e
T
f .c
R
Y
=
T
f .e
T
f .c
R
Y.e
R
k.e
R
k).e
R
k).c
R
k.c
R
Y.c
.
The surfaceconvection resistances R
Y.e
and R
Y.c
for semibounded uids
owing over the evaporator and condenser are the extendedsurface resistances
of Section 6.8.2. These are made of parallel arrangements of surface convection
from the bare surface area A
b
and n surface area A
f
. Then from (6.151) and
MassoudKaviany 2011
Table Ex. 8.1.1. Physical and geometrical properties of heat pipe
considered
Parameter Magnitude
tube: copper L
a
0.3 m
L
e
0.1 m
L
c
0.1 m
D
1
,2 6.5 mm
D
2
,2(D
c
= D
e
= D
2
) 7.5 mm
k
s
(copper)(T = 300 K) 401 W/mK
n: copper L
f
20 mm
l(l
c
= l
e
= l) 1 mm
N
f
(evaporator) 20
N
f
(condenser) 20
k
s
(copper)(T = 300 K) 401 W/mK
f .e
=
f .c
wick: copper c 0.5
k
f
(water T
f
= T
lg
) 0.65 W/mK
k) 11 W/mK
l
n
0.5 mm
owing air u
f .e
= u
f .c
10 m/s
properities evaluated
using T
f .e
= 340 K
and T
f .c
= 283 K
(Table C.22)
working uid: water M 18 kg/kmole
(Table C.27) c
p
,c
:
1.31
Lh
lg
2.4 10
6
J/kg
(6.152), we have
R
1
Y.e
= A
b.e
Nu)
D.e
k
f .e
D
e
A
f .e
Nu)
n.e
k
f .e
n
f .e
f .e
R
1
Y.c
= A
b.c
Nu)
D.c
k
f .c
D
c
A
f .c
Nu)
n.c
k
f .c
n
f .c
f .c
.
Table Ex. 8.1.1 lists the physical and geometrical properties of the heat pipe
considered. The pipe ns and the wick are all made of copper. The working
uid (inside the pipe) is water and the evaporator and condenser are heated and
cooled by semibounded air streams with prescribed T
f
and u
f
.
(i) SurfaceConvection Resistance
For the ns, we assume an efciency of unity, i.e.,
f .e
=
f .c
= 1 (it can be
shown that this is justiable for copper ns of moderate length L
f
). For the semi
bounded uid ow Nusselt numbers, we use the surfaceconvection correlation
of Table 6.3. As shown in Table Ex. 8.1.1 we use the same geometry and ow for
the condenser and evaporator regions.
MassoudKaviany 2011
For the base section of the tubes, we use the tube Reynolds number for
semibounded uids, i.e.,
Re
D.e
=
u
f .e
D
2
f
.
The relation for the Nusselt number is found from Table 6.4 as
Nu)
D.e
= a
1
Re
a
2
D.e
Pr
1,3
.
where a
1
and a
2
are constant and their magnitudes are listed in Table 6.3 for
various ranges of Re
D.e
.
For the ns, we use a characteristic length n
f
(n
f .c
= n
f .e
= n
f
) for the
surfaceconvection from
n
2
f
= [(D
2
2L
f
)
2
D
2
2
],4. The narea Reynolds number for the n surface
then is
Re
n.e
=
u
f .e
n
f
f
.
Then the relation for the Nusselt number is found from Table 6.3, and if
Re
n.e
 5 10
5
, we use
Nu)
n.e
= 0.664Re
1,2
n.e
Pr
1,3
.
The nsurface and basesurface areas for the evaporator and condenser
regions are equal. However, the respective air ows are at different temperatures
and the thermophysical properties are evaluated at T
f .e
= 340 K and T
f .c
= 283
K. The surface areas are
A
b.e
= D
2
L
e
D
2
l
e
N
f
A
f .e
= N
f
2
_
(D
2
2L
f
)
2
D
2
2
4
_
.
(ii) TubeWall Resistance
For the cylindrical shell, from Table 3.2, we have
R
k.e
=
ln(D
2
,D
1
)
2kL
e
.
where k is the copper conductivity.
(iii) Wick Resistance
The effective conductivity of the wick k) is found from (3.28), with k
s
being
that of copper and k
f
being that of liquid water. The resistance for the wick is
given similar to that of the tube wall as
R
k).e
=
ln[(D
1
2l
n
),D
1
]
2k)L
e
.
The results for the condenser side are identical to the evaporator side, for
the above two conduction resistances.
MassoudKaviany 2011
Table Ex. 8.1.2. Calculated values
of the resistances for the heat pipe
considered
Variable Magnitude
R
Y.e
1.91 10
1
C/W
R
Y.c
1.88 10
1
C/W
R
k.e
= R
k.c
5.68 10
4
C/W
R
k).e
= R
k).c
1.16 10
2
C/W
(b) Within the pipe we assume that water is the only species and the watervapor
pressure and temperature are related through the ClausiusClapeyron relation
(A.14),
p
g
= p
g.o
exp
_
MLh
lg
R
g
_
1
T
g
1
T
g.o
__
phaseequilibrium condition.
where for water M= 18 kg/kgmole and Lh
lg
is listed in Table C.27. In addition,
the density, pressure, and temperature are related through the ideal gas law
(3.18), i.e.,
g
=
Mp
g
R
g
T
.
The vapor ow rate and speed are given by
M
g
=
M
lg
= A
u
g
u
g
=
Q
Lh
lg
.
Here we consider subsonic vapor speeds, where the sonic speed is given by
(3.20), i.e.,
a
s
=
_
_
c
p
c
:
_
g
R
g
M
T
g
_
1,2
.
For water vapor c
p
,c
:
= 1.31.
Numerical results are obtained by varying 300 K T
f .e
450 K, while
keeping T
f .c
= 283 K.
The computed resistances are listed in Table Ex. 8.1.2 and in Figure Ex. 8.1.2.
Note the surfaceconvection resistances are the largest, followed by the wick
conduction resistance.
The vapor temperature is shown in Figure Ex. 8.1.3(a), along with the con
stant T
f .c
and the variable T
f .e
, as a function of the heat ow rate Q. Note that
as T
f .e
increases, T
g
= T
lg
remains between T
f .c
and T
f .e
. The amount of heat
transferred is between 20 to 450 W. The vapor pressure and the vapor velocity
are shown in Figure Ex. 8.1.3(b). Note that the vapor pressure increases with Q
(or increase in T
f .e
) and surpasses the ambient pressure. On the other hand, the
vapor velocity rst increases and after reaching a peak, decreases, as Qincreases.
MassoudKaviany 2011
R
k,c
(Tube) = 5.68 x 10
4
R
,c
= 1.88 x 10
1 O
C/W
Magnitude of Various Thermal Resistances
R
k ,c
(Wick) = 1.16 x 10
2
R
k,e
(Tube) = 5.68 x 10
4
R
,e
= 1.91 x 10
1
R
t
,
O
C/W
0 10
8
10
7
10
6
10
5
10
4
10
3
10
2
10
1
1
R
k ,e
(Wick) = 1.16 x 10
2
Figure Ex. 8.1.2. Magnitude of various thermal resistances for the heat pipe considered.
This is due to the increase of the vapor density, which increases in a nonlinear
manner with vapor temperature. The sonic speed, which is a function of T
g
, is
also shown for Q = 100 and 200 W. The vapor velocity is much lower than the
sonic speed.
COMMENT
We could have included the vapor pressure drop between the evaporator and
condenser. This would require a friction factor for the viscous vapor ow
[7,9,22,24].
EXAMPLE 8.2. DES.S
THERMAL INKJET: A RAPID, LOCALIZED HEATING DEVICE
The disposable printhead consists of a glass substrate sandwiched between a
nickel manifold containing several nozzles and a plastic body that contains the
ink. This is shown in Figures Ex. 8.2.1(a) and (b). Each nozzle can supply ink on
demand from the printer as the printhead scans across the paper. The droplets
of ink are ejected by instantaneously vaporizing a tiny volume of ink. The vapor
bubble grows rapidly, leading to the inertial of the adjacent ink and propelling
the ink through the nozzle onto the paper. After the electrical pulse to the heater
is turned off, the vapor bubble collapses. This is shown in Figure Ex. 8.2.1(c).
The thinlmJoule heater (electrical resistor) provides the heat to the ink for
the bubble nucleation and growth. Before depositing this thin lm, a dielectric
material, such as silicon dioxide, is deposited on the glass substrate by sputtering.
This barrier lm(shown in Figure Ex. 8.2.1) prevents leaching of impurities from
the glass substrate into the heater lm. The heater lm is tantalumaluminum
alloy and is also sputter deposited. A passivation layer (silicon nitride or silicon
carbide) protects the heater lm from the chemical and physical damages from
the ink and from the erosive damage of the bubble nucleation and collapse.
(a) Draw the transient thermal circuit diagram.
(b) Plot the variation of the heater temperature T
3
, and the other nodal
temperatures as a function of time. Allow for the passive cooling of the system.
MassoudKaviany 2011
0 100 200 300 400 500
250
300
350
400
450
500
T
,
K
Q, W
(a) Gas, Evaporator Air, and Condenser Air Temperature
(b) Vapor Velocity and Pressure
T
f,c
(Prescribed)
Tlg
=
Tg
(P
re
d
ic
te
d
)
T
f,e
(
P
r
e
s
c
r
i
b
e
d
)
0 100 200 300 400 500
1.6
1.4
9
8
7
6
5
4
3
2
1
0
1.2
1.0
0.8
0.6
0.4
0.2
0
p
g
,
a
t
m
(
1
.
0
1
3
x
1
0
5
P
a
)
u
g
,
m
/
s
Q, W
a
s
= 435 m/s a
s
= 460 m/s
u
g
p
g
Figure Ex. 8.1.2. (a) Variation of the vapor temperature within the pipe as a function of
the heat ow rate. (b) Variation of the vapor pressure and speed as a function of the heat
ow rate.
(c) Show the fraction of heat owing into the ink and into the substrate, as
functions of time.
(d) Showthe effect of current (or Joule heating s
e.J
L
3
) on the heater temperature
history.
(e) Show the temperature distribution in the system to elapsed times corre
sponding to the time just before heating stops (t = 4.8 js), just after heating
stops (t = 5.2 js), and some time later (t = 10 js).
Since both the glass substrate and the ink are much larger than the three
thin lm layers, use semiinnite medium approximations for them. Determine
the transient heat transfer in these two semiinnite layers using the results for
the timedependent surface temperature given in Section 3.5.3. Alternatively,
MassoudKaviany 2011
i. Bubble Nucleation
(c) Drop Generation Regimes
(b) PrintHead Orifice Plate (a) Inkjet Print Head
ii. Drop Ejection
Substrate
Heater and
Associated Layers
iii. Bubble Collapse
Orifices
Slot
Wall
Vent
Vent Wall
Standoffs
iv. Channel Refill
Drop
Ink
Bubble
ResistorArray Substrate
Bladder
(Ink Supply)
Bladder Wall
Orifice
Ink
Figure Ex. 8.2.1. A schematic of the thermal inkjet. (a) Inkjet printhead. (b) Printhead
orice plate. (c) Drop generation regimes.
use several nodes for the glass and the ink layers as discussed in the discretized
medium treatment of Section 3.7.
The magnitudes of the various parameters, for the heater layer and the
surrounding layer, are given in Table Ex. 8.2.1. It has been determined exper
imentally [1,46,14,16,19,21] that when the temperature T
4
is raised to about
280
C
k
1
1.10 W/m
C
barrier layer: SiO
2
L
2
3 jm
2
2,190 kg/m
3
c
p.2
740 J/kg
C
k
2
1.5 W/m
C
heater layer: L
3
0.1 jm
tantalumaluminum
e.3
4 10
6
ohmm
3
13,300 kg/m
3
c
p.3
140 J/kg
C
k
3
144 W/m
C
passivation layer: SiC L
4
4 jm
4
3160 kg/m
3
c
p.4
675 J/kgK
k
4
40 W/m
C
ink
5
994.5 kg/m
3
c
p.5
4,256 J/kg
C
k
5
0.628 W/m
C
Joule heating: j
e
2.505 10
10
A/m
2
T
5
= T
1
= T
3
(t = 0) 298 K
duration of heating 5 js
SOLUTION
(a) A schematic of the idealized, onedimensional model and its corresponding
thermal circuit model are shown in Figures Ex. 8.2.2(a) and (b). Since L
3
is small
compared to L
2
and L
4
, the heater is treated as a uniform temperature solid.
(b) The energy equation for node 3 is written from (8.1) as
Q
31
Q
35
= (c
p
V)
3
dT
3
dt
S
e.J
S
e.J
= s
e.J
L
3
A
k
=
e
j
2
e
L
3
A
k
Q
31
=
T
3
(t) T
1
R
k.23
R
ck.21
(t)
Q
35
=
T
3
(t) T
5
R
k.34
R
ck.45
(t)
.
where A
k
is the area for conduction heat transfer. Here we perform our analysis
per unit area and use q
31
, etc.
Since T
4
and T
2
both change with time, the transient conduction resistances
R
cu.45
(t) and R
cu.21
(t) are used and are evaluated using (3.158). Alternatively,
MassoudKaviany 2011
L
2
= 3 m
L
4
= 4 m
T
4
(PassivationInk
Interface)
T
2
(BarrierSubstrate
Interface)
T
3
(Uniform)
T
5
(FarField), Ink
T
1
(FarField), Substrate
(b) Thermal Circuit Model
(a) Physical, OneDimensional Model of Heater and Associated Layers
T
5
R
ck,45
(t)
Q
35
Q
31
R
ck,21
(t)
R
k,34
(Constant PassivationLayer Resistance)
(c
p
V)
3
+ S
e,J
dT
3
dt
R
k,32
(Constant BarrierLayer Resistance)
T
4
(t)
T
2
(t)
T
3
(Uniform Heater Temperature)
T
1
Ink (WaterBased)
Glass (Substrate)
Passivation Layer
(SiC or Si
3
N
4
)
Barrier Layer (SiO
2
)
JouleHeating Layer (Ta
2
Al)
w
L
3
x
Q
31
Q
35
Figure Ex. 8.2.2. (a) A schematic of onedimensional model for thermal inkjet. (b) Its
thermal circuit model.
in the glass and the ink, we can divide the near regions (penetration depths) into
nitesmall volumes (about ten small volumes in each layer). For the glass, the
penetration depth, estimated using (3.148), is about 20 jm and for the ink it is
about 10 jm, after an elapsed time of 10 js.
We have included the storage, i.e., c
p
V of the barrier and passivation layer,
with that of the heater.
Using a solver (e.g., MATLAB), the set of energy equations are solved
numerically. The results for the temperatures T
2
, T
3
, and T
4
are shown in Figure
Ex. 8.2.3(a). Note the rather slow decay of the temperatures (compared to the
fast rise in the temperatures).
MassoudKaviany 2011
2.5 x 10
8
2.0 x 10
8
1.5 x 10
8
1.0 x 10
8
5.0 x 10
7
0.0
5 x 10
7
250
300
350
400
450
500
T
4
(Passivation Layer Ink Interface)
T
3
(Heater)
Passive
Cooling
Active
Heating
Passive
Cooling
Active
Heating
T
3
(t = 5 s) = 581.7 K
T
2
(Barrier Layer Glass Interface)
550
600
650
T
2
,
T
3
,
T
4
,
K
q
3

5
,
q
3

1
,
W
/
m
2
(a) Evolution of Heater Temperature
(b) Evolution of Interfacial Heat Fluxes
0 5 10 15 20 25 30
t, s
0 5 10 15 20 25 30
t, s
T(t = 0), Initial Temperature
q
35
+ q
31
q
31
(Heat Flow into Ink)
q
35
(Heat Flow into Substrate)
Figure Ex. 8.2.3. (a) Evolution of the heater and interface temperatures. (b) Evolution
of the rate of heat ow into the ink and into the substrate.
MassoudKaviany 2011
Passivation
Layer
Penetration
Depth
(t)
Penetration
Depth
(t)
Barrier Layer
Heater Layer
(0.1 m)
Glass Ink
(b) Temperature Distribution within Various Layers at Several Elapsed Times
(a) Effect of Joule Heating Rate s
e,J
L
3
on Heater Temperature
250
300
350
400
450
500
550
600
650
T
,
K
250
300
350
400
450
500
550
600
T
3
,
K
8 6 4 2 0 2 4 6 8
x, m
T(t = 0)
t = 10 s (Some
Time After Heat
ing Stops)
t = 5.2 s
(Shortly After
Heating Stops)
t = 4.8 s
(Shortly Before
Heating Stops)
T(t = 0)
0 5 10 15 20 25 30
t, s
s
e,J
L
3
= 2.5 x 10
8
W/m
2
s
e,J
L
3
= 2.0 x 10
8
s
e,J
L
3
= 1.5 x 10
8
s
e,J
L
3
= 1.0 x 10
8
s
e,J
L
3
= 0.5 x 10
8
Figure Ex. 8.2.4. (a) Evolution of the heater temperature for several Joule heating rates.
(b) The temperature distribution in the various layers and substrates, for three different
elapsed times.
MassoudKaviany 2011
Table Ex. 8.2.2. Calculated
values of the variables for the
thermal inkjet considered
Parameter Magnitude
T
3.0
(t = 5 js) 581.7 K
s
e.J
L
3
2.51 10
8
W/m
2
(c) The temperature of the heater returns to the initial temperature in about 100
ms. The amounts of heat owing into the ink q
35
and into the substrate q
31
are
shown in Figure Ex. 8.2.3(b), during the heating and passive cooling periods.
(d) The dependence of the heater temperature on the Joule heating rate is
shown in Figure Ex. 8.2.4(a). For the desired T
3
temperature at the prescribed
time (here after elapsed time of 5 js), the appropriate current should be used.
Note that T
3
increases near linearly with s
e.J
L
3
.
(e) The temperature distribution in the various layers, and within the ink and
substrate, are shown in Figure Ex. 8.2.4(b), for three different elapsed times:
one for just before the desired temperature is reached (near the end of the
heating period), one just after the heating has stopped, and one some time after
the heating has stopped.
Table Ex. 8.2.2 lists the magnitude of the variables determined.
COMMENT:
Active cooling of the heater (by thermoelectricity) can signicantly reduce the
cooling period. Equation (3.158) would require the use of the surface temper
ature at prior nite intervals. This requires the use of an accessible array that
contains all prior surface temperatures.
When a solver is used that does not allow for an easy retrieval of the tem
perature array T
3
(t), then each ink and substrate region adjacent to the heater
is divided into ten or so subregions. The numerical results (e.g., MATLAB)
results are for 14 subdivisions in the ink and 14 in the substrate. The results
are nearly identical to those using solution for the semiinnite region with a
variable surface temperature.
EXAMPLE 8.3. DES.S
PHASECHANGE PNEUMATIC MICROACTUATOR: A
MICROELECTROEMECHANICAL SYSTEM
Displacements are needed in some microelectromechanical systems (MEMS).
The system may be a small motor, pump, etc. Such systems are referred to
as actuators. In integratedcircuit devices, in order to move or regulate the
ow of uids, relatively large displacements (of the order of ten micrometers)
are needed. One of the methods of providing such a forcedisplacement is by
MassoudKaviany 2011
heating an enclosed uid, resulting in a rise in the pressure of this enclosure (i.e.,
cavity) [22]. This uid can be a liquid, a gas, or a liquidgas mixture. In order
to minimize the electric power consumption, a mixture of liquidvapor (with
a relatively high vapor quality) is used. Then the heat added to this enclosed
liquidvapor mixture causes further evaporation and an increase in the cavity
pressure. For quick reversal, active (thermoelectric) heating is used, and once the
desired pressure is reached during heating, the current is reversed and cooling
of the liquid begins.
Figures Ex. 8.3.1(a) to (c) shows a pneumatic microactuator design in which
the thin membrane on top of the cavity is displaced upon increase in the cavity
pressure [3].
Thin, Displacing
Membrane
Orifice
Vapor Space
Gas
Flow
Silicon Heater Evaporation
Silicon Cap
Sealed
Separator
Gas Flow
Substrate
Coolant
Coolant Flow, T
1
, Nu
D,h
Vapor Space
p
g D
h
D
h
/2
p
o
p
g
> p
o
p
g
< p
o
S
e,J
+ S
e,P
T
lg
Thermoelectric
Element
SiO
2
Plate
(Electrical
Insulator)
Liquid Layer
T
3
(Uniform)
T
2
L
TE
L
3
L
l
a
c
a
c
Electrical
Conductor
p
o
p
g
a
c a
w L
c
(+)
( )
D
h
/2
a
TE
L
s
T
f,
, Nu
D,h
S
e,P
S
e,J
.
.
q
k
q
u
q
ku
S
e,P
.
q
k
q
k
Orifice
(a) PhaseChange Based Microactuator (Valve)
(b) ElasticDeform
able Membrane
of Heater and Associated Layers
PhaseChange
Actuated Valve
Figure Ex. 8.3.1. (a) A schematic of the phasechange based pneumatic microactuator
considered. (b) Elasticdeformable membrane. (c) Physical, onedimensional model of
the heater and associated layers.
MassoudKaviany 2011
Table Ex. 8.3.1. Physical and geometrical properties of the
phasechange actuated valve considered
Parameter Magnitude
gas ow in orice region T
f .
298 K
D
h
40 jm
Nu)
D.h
7.54
k
g
0.0263 W/mK
cavity L
c
300 jm
a
c
600 jm
working uid: R11 M 137.37 kg/kgmole
Lh
lg
1.802 10
5
J/kg
k
l
0.0889 W/mK
l
1488.0 kg/m
3
L
l
5 jm
T(t = 0) 298.15 K (25 K)
silicon L
si
10 jm
s
2330 kg/m
3
c
p.s
678 J/kgK
k
s
150 W/mK
thermoelectric layer J
e
15 mA
l
TE
40 jm
N
TE
8 16 junctions
a
TE
20 jm
e
10
5
ohmm
S.p
2.30 10
4
V/K
S.n
2.10 10
4
V/K
TE
9320 kg/m
3
c
p.TE
122 J/kgK
air
1 kg/m
3
c
p.air
1000 J/kgK
substrate: Cu L
s
30 jm
s
8933 kg/m
3
c
p.s
385 J/kgK
k
s
401 W/mK
coolant surfaceconvection : T
1
298 K
Nu)
D.h
7.54
D
h
13.22 jm
k
f
0.0263 W/mK
(a) Draw the thermal circuit diagram.
(b) For the conditions shown in Table Ex. 8.3.1, show the liquidvapor temper
ature and the vapor pressure as a function of time. Note that once the desired
pressure is reached, the direction of the current ow is reversed to begin imme
diate cooling.
(c) Show the thickness of the liquid layer and the rate of heat owing through
the liquid layer and the substrate as a function of time.
MassoudKaviany 2011
T
lg
(t)
T
3
(t)
T
2
(t)
R
k,32
, Constant Thermoelectric Conduction Resistance
R
k,s
0, Substrate Conduction Resistance
R
ku,s
, Substrate SurfaceConvection Resistance
R
k,3lg
(t), LiquidLayer Conduction Resistance
T
f,
R
ku,m
, Membrane SurfaceConvection Resistance
Thermal Circuit Model for PhaseChange Based Actuator
T
1
(c
p
V)
2
dT
2
dt
+ (S
e,J
)
2
+ (S
e,P
)
2
(c
p
V)
3
dT
3
dt
+ (S
e,J
)
3
+ (S
e,P
)
3
(c
p
V)
w
dT
lg
dt
S
lg
Q
3lg
Q
21
Q
32
(Q
lg
)
m
Q
lg,w
Wall Heat
Storage / Release
T
lg
Figure Ex. 8.3.2. The thermal circuit model for the phasechange based pneumatic micro
actuator.
SOLUTION
(a) Figure Ex. 8.3.2 shows the thermal circuit diagram.
(b) For the vaporliquid volume we have the integralvolume energy equation
(8.1), neglecting the storage term,
Q
lg.
Q
3lg
=
S
lg
(c
p
V)
n
dT
lg
dt
.
where
Q
lg.
=
T
lg
(t) T
R
ku
R
1
ku
= A
ku
Nu)
D.h
k
f
D
h
.
Here we have used a channel ow with laminar, fullydeveloped temperature
and velocity elds. Then Nu)
D.h
is found from Table 7.2 and is Nu)
D.h
= 7.54.
The hydraulic diameter D
h
= 40 jm.
For node 3 (upper portion of the thermoelectric module and the silicon layer)
we have
Q
3lg
Q
k.32
= (c
p
V)
3
dT
3
dt
(
S
e.J
)
3
(
S
e.P
)
3
.
MassoudKaviany 2011
where
Q
3lg
=
T
3
(t) T
lg
(t)
R
k.3lg
(t)
R
k.3lg
(t) =
L
l
(t)
A
k
k
l
(c
p
V)
3
.
1
2
(c
p
V)
TE
(c
p
V)
SiO
2
V
TE
= N
TE
a
2
TE
L
TE
.
The contribution of the SiO
2
lm is negligible (about 50 nm thick lm). During
the heating of the liquid, the current J
e
(A) ows such that both
S
e.J
and
S
e.P
are
positive. During the cooling of the liquid, the current is reversed. The Joule and
Peltier heatingcooling is
(
S
e.J
)
3
=
1
2
R
e.32
J
2
e
(
S
e.P
)
3
=
S
T
3
J
e
.
(J
e
> 0 for heating and
J
e
 0 for cooling).
For node 2 (lower portion of the thermoelectric module) we have
Q
k.32
Q
21
= (c
p
V)
2
dT
2
dt
(
S
e.J
)
2
(
S
e.P
)
2
.
where
Q
21
=
T
2
(t) T
1
R
ku
. R
1
ku
= A
ku
Nu)
D.h
k
f
D
h
(c
p
V)
2
.
1
2
(c
p
V)
TE
(c
p
V)
Cu
.
(
S
e.P
)
2
=
2
T
2
J
e
.
The Peltier heating for node 2 is reversed such that for J
e
> 0, (
S
e.P
)
2
 0, etc.
Now examining some of the thermodynamic relations, the amount of
methanol in the cavity is
M= V
l
l
V
g
g
.
We represent the cavity volume as having a crosssectional area A and lengths
L
g
L
l
.
Then we have
M= A[
l
L
l
(t)
g
L
g
(t)]
= A{
l
L
l
(t)
g
[L
c
L
l
(t)]]
p
g
=
R
g
M
g
T
lg
A= a
2
c
.
where L
c
is the total length of the cavity.
MassoudKaviany 2011
The vapor pressure is related to the temperature through the Clausius
Clapeyron relation (A.14). This is written as
p
g
= p
g.o
exp
_
MLh
lg
R
g
_
1
T
lg
1
T
lg.o
__
.
Here we have used Refrigerant R11 with p
g.o
= 1 atm, and T
lg.o
= 296.9 K.
The density of the vapor is related to pressuretemperature through ideal gas
law. The electrical conductor, connecting p and n junctions, is separated from
the copper substrate by a thin SiO
2
layer for electrical insulation.
Table Ex. 8.3.1 lists the physical and geometrical properties of the phase
change based pneumatic microactuator considered.
Heat is added by passing current through the thermoelectric module and
when the desired gas pressure is reached, the current is reversed to allow for
an active cooling, thus resulting in the return of the membrane to its initial
location.
Figure Ex. 8.3.3(a) shows the rise and fall of the various temperatures. Note
that after the heating stops, T
3
remains above T
lg
and heat continues to ow
to the liquidgas interface. The temperature at the thermoelectricsubstrate
interface drops (due to Peltier cooling) during the heating of the liquid layer.
The reverse occurs during the cooling of the liquid layer. The gas pressure
history is shown in Figure Ex. 8.3.3(b). In the pressure history, note the
symmetry in each of the rise and fall branches.
(c) As was mentioned, onlyasmall amount of liquidis neededfor theevaporation
and the rise to the desired pressure. This is shown in Figure Ex. 8.3.4(a), where
the initial thickness of the liquid L
l
= 5 jm and only reduces to 4 jm at the end
of evaporation pressure. The transient heat ow rates, q
21
and q
3lg
, are shown
in Figure Ex. 8.3.4(b). Again note that q
3lg
remains positive, because for this
example the surfaceconvection heat loss is signicant and controls the cooling
period. The substrate initially loses heat to the thermoelectric module and then
during the liquid cooing period heat is added to the substrate.
Table Ex. 8.3.2 lists the magnitude of the variables calculated.
Table Ex. 8.3.2. Computed
values of the variables for
the phasechange actuated
valve considered
Variable Magnitude
p
g
(t = 0) 1.043 atm
p
g.max
2.043 atm
T
lg.max
319.7 K
Lt
heating
37.0 ms
Lt
cooling
9.3 ms
MassoudKaviany 2011
T
3
(ThermoelectricSiO
2
Layer Interface)
T
2
(Thermoelectric
Substrate Interface)
T
f,
T
2
,
T
3
,
T
l
g
,
K
280
270
290
300
310
320
330
t, ms
(a) Evolution of Various Temperatures
Heating Cooling
p
g
/
p
g
(
t
=
0
)
1.0
2.2
2.0
1.8
1.6
1.4
1.2
t, ms
(b) Evolution of Gas Pressure
Desired p
g,max
Cooling Heating
p
g
T
lg
(Liquid
Vapor Interface)
50 40 30 20 10 0
50 40 30 20 10 0
Figure Ex. 8.3.3. (a) Evolution of the liquidgas interfacial temperature (also Si layer and
thermoelectricsubstrate interfacial temperatures). (b) Evolution of the vapor pressure
within the cavity.
COMMENT
Typical membrane displacement for the microactuator considered is about
50 jm [20].
For the performance evaluation, we dene a gure of merit as
Lx
=
LFLx
S
e.J
Lt
t
.
MassoudKaviany 2011
q
3lg
q
21
L
l
(t = 0)
L
l
t
cooling
t
heating
L
l
,
m
4.0
5.0
4.8
4.6
4.4
4.2
50 40 30 20 10
t, ms
(a) Evolution of LiquidFilm Thickness
Cooling Heating
Cooling Heating
q
2

1
,
q
3

l
g
,
W
/
m
2
0
50 40 30 20 10 0
t, ms
(b) Evolution of Interfacial Heat Fluxes
1.5 x 10
6
5.0 x 10
5
1.0 x 10
6
0.0
1.5 x 10
6
1.0 x 10
6
5.0 x 10
5
Figure Ex. 8.3.4. (a) The evolution of the liquidlm thickness. (b) Evolution of the heat
ux owing through the liquidlm layer and through the substrate.
where LF = Lpa
2
c
is the force, Lx is the displacement,
S
e.J
is the Joule heating,
and Lt
t
is the total response time. Substituting the numerical values gives
Lx
=
1.013 10
5
(Pa) (600 10
6
)
2
(m
2
) 50 10
6
(m)
0.015(A)
2
2(ohm) 46.3 10
3
(s)
= 0.0875.
This magnitude is reasonable compared to that of other actuators [23].
In this analysis, the deection of the membrane is neglected. However, its
inclusiondoes not change the nal volume by more than8%. Therefore, inclusion
of the change in cavity volume due to membrane defection is not necessary.
MassoudKaviany 2011
A wide variety of working uids can be used. Fluids with near room tem
perature boiling temperature at one atm will be the proper candidates. These
are listed in Figure 7.18 and Tables C.3 and C.4.
The twodimensional effect, i.e., the heat loss to the side walls, can be
neglected due to the gas lled gaps at the side walls acting as insulators.
EXAMPLE 8.4. DES.S
COMBUSTIONTHERMOELECTRIC TUBE: A DIRECT
ENERGY CONVERSION DEVICE
Exothermic, gaseous reaction in tubes (and porous media in general) allows for
preheating of the reactants by conduction through the tube wall. If in addition
the gas ow is allowed to reciprocate in direction, heat can be stored within the
tube. The combination of the preheating and storage allows for creation of high
temperatures using a very small amount of heat of reaction per kilogram of the
reactants. These high temperatures can in turn be used for direct thermoelectric
energy conversion [8,15,17].
Consider the electric power generation in a tube using reciprocating, direct
fuel injection gaseous ow [6]. The fuel (methane) is injected near the hot
junctions. The ow and heat transfer are shown in Figure Ex. 8.4.1. There is
a central combustion region and two end (adjacent to the combustion region)
thermoelectric regions. The combustion region attains a high temperature, while
the thermoelectric regions have aninner hightemperature endandanouter low
temperature end. The temperature difference across the thermoelectric region
is used for the electric power generation.
(a) Draw the thermal circuit diagram, by discretizing each of the thermoelectric
regions into ve equal nitesmall volumes and dividing the combustion region
into three equal nitesmall volumes.
Allow for energy storage in the solid, but not in the uid (because of the
much smaller gas density).
Allow for conduction in the solid by using a uniform temperature across
the wall thickness and neglect conduction in the gas (here because of the much
smaller gas conductivity).
Allow for surfaceconvection heat transfer between the bounded gas and
the tube wall, by using the average convection resistance.
Allow for a temperaturedependent reaction using the niteorder reaction
kinetic model (2.19). Use an average gas temperature in determining the intern
odal energy conversion term
S
r.c
.
(b) Using the magnitude of the various physical and geometric parameters given
in Table Ex 8.4.1, determine the electrical power generated and the efciency of
the thermoelectric energy conversion device, as a function of the elapsed time
[during each period (s)].
MassoudKaviany 2011
Adiabatic
Surface
Joined with
Small Contact
Resistance
Electrical
Conductor
Electrical Insulator
nType
Material
pType Material
Reciprocating
Gaseous Oxidant
Flow
Reciprocating
Fuel Flow
(+)
( )
e
,
S
, k
P
, (c
p
)
P
k
s
, (c
p
)
s
(a) CombustionThermoelectric Energy Conversion Tube
J
e
J
e
L
TE
Thermoelectric
Storage Region
L
A
Combustion
Region
Reaction
Region
L
TE
D
1
/2
D
2
/2
Thermoelectric
Storage Region
0
T
h
s
L
TE
L
TE
+ L
A
L = 2L
TE
+ L
A
T, s
r
T
s
T
c
s
T
f L
u
f
(t)
T
f 0
s
r,c
x
f
f
(b) Heat Transfer in CombustionThermoelectric Energy Conversion Tube
q
k
q
r
q
u
q
k
q
u
q
k
(+)
( )
q
ku
S
e,P
q
k
q
k
S
e,J
S
e,P
s
r, c
s
r, c
Electrical Conductor, Thermal Capacitor, Thermal Insulator,
Energy Converter (Chemical to Thermal and Thermal to Electrical)
Combustion
Region
Thermal Conductor,
Thermal Capacitor
Thermoelectric
Storage Region
Thermoelectric
Storage Region
u
f
(t)
u
f
(t) u
f
(t)
u
f
(t), k
f
, (c
p
)
f
Reciprocating
Premixed Gaseous
FuelOxidant Flow
Figure Ex. 8.4.1. A schematic of the combustionthermoelectric energy conversion tube.
(a) The geometry and various parameters. (b) The heat ux vector tracking.
SOLUTION
(a) The thermal circuit diagram for this problem is shown in Figure Ex. 8.4.2.
Each thermoelectric region is divided into ve equal nite volumes. The middle
combustion region is divided into three equal nite volumes.
(b) For the solid node, from (8.1) we write the energy equation
Q
r.i
T
s.i
T
f .i
R
u
)
L.i
T
s.i
T
s.j
R
k.ij
= (c
p
V)
i
dT
s.i
dt
(
S
e.J
)
i
(
S
e.P
)
i
nodal, solidphase energy equation.
MassoudKaviany 2011
The surface radiation from any solid node T
s.i
is given by (8.4) and (8.5). Here,
(8.5) is summed over all the solid surfaces and the two end surfaces. The view
factors are determined from Table 4.2. There, the view factor between rings is
given. The case of the view factor from a ring to the end of the cylinder is found
from the same expression by allowing for the ring placed at the end to have a
very large length (i.e., l
2
).
The heat storage in the uid is neglected because of its small density com
pared to the solid. The axial conduction in the uid is also neglected, because it
is small compared to the solid axial conduction. Similarly, for the uid nodes we
have, from (8.1),
T
f .i
T
s.i
R
u
)
L.i
(
Mc
p
)
f
A
u
(T
f .i1
T
f .i
) =
S
r.c.i
internodal, uidphase
energy equation.
Here we have an inlet air A
u
(
f
u
f
))
0
=
M
f .0
and near the hot junction (in the
combustion region), fuel is injected. The amount of fuel injected is
M
F
. The fuel
injection helps to avoid the premature ignition of the fuel and to locate the ame
near the hot junctions for effective heat transfer to the junctions.
We have included the change in the uid temperature resulting from the
surfaceconvection heat exchange with the tube wall, (7.25), for the uid node
entering as i and exiting as i 1. The average convection resistance is given by
(7.27) as
R
u
)
L.i
=
1
(
Mc
p
)
f
(1 e
NTU
i
)
.
where from (7.20)
NTU
i
=
A
ku.i
Nu)
D
k
f
D
1
A
u
u
f
)(c
p
)
f
=
A
ku
Nu)
D
k
f
D
1
( mc
p
)
f
.
Table Ex. 8.4.1 lists the magnitudes of the various parameters considered. The
energy conversion and conduction resistance terms are (Tables 2.1 and 3.2)
S
r.c.i
= A
u
u
f
)(
F.i
F.i1
)Lh
r.F
(
S
e.J
)
i
= R
e.i
J
e
(
S
e.P
)
i
=
S
J
e
T
i
for pn junctions
S
=
S.p
S.n
R
e.i
=
e
L
i
A
k
.
Here when i = h (i.e., hot junction),
S
e.P
is negative (heat absorption).
The fuel density is determined from the fuelspecies conservation equation
(B.31). The mass diffusion and time dependence are neglected. Here we use the
rstorder reaction kinetic model that has been developed for porous media [8],
i.e., n
r.F
= a
r
F
e
LE
a
,R
g
T
f
.
MassoudKaviany 2011
Table Ex. 8.4.1. Physical, chemical, and geometrical properties of the
thermoelectric energy conversion considered
Parameter Magnitude
fuel (methane):
F
0.02746 kg/m
3
Lh
r.F
55.53 10
3
kJ/kgfuel
T
298 K
a
r
2.6 10
8
1/s
LE
a
130.0 10
3
J/mole
M
O
16 kg/kmole
M
F
32 kg/kmole
O
44.6
F
1.0
gas k
f
0.0573 W/mK
f
1.164 kg/m
3
c
p.f
1099 J/kgK
j
f
208.2 10
7
N s/m
2
u
f
) 0.75 m
10 s
tube assembly: number of tubes, N
t
144(= 0.03
2
,D
2
2
)
[0.03 (m) 0.03 (m)] external resistance,
R
e.o
R
e.o
= R
e.hc
solid:
combustion region k
s
30 W/mK
(ceramic foam)
s
3,160 kg/m
3
c
p.s
1,310 J/kgK
D
1
1.3 mm
D
2
2.5 mm
L
A
7.5 mm
c
r
1.0
solid:
thermoelectric region
S.p
.
S.n
195.3, 207.3 jV/K
(SiGe)
e
19.1 johmm
k
s
4.3 W/mK
s
2990 kg/m
3
c
p.s
816.7 J/kgK
D
1.TE
1.3 mm
D
2.TE
2.5 mm
L
TE
60 mm
c
r
1.0
c
r.
1.0
Also, an average temperature T
f
= (T
f .i
T
f .i1
),2 is used. The result, after
integration over the internodal length L
i
(this is shown in Problem 5.14), is
F.i1
=
F.i
exp
_
a
r
L
i
u
f
)
e
(LE
a
,R
g
T
f
)
_
.
where a
r
and LE
a
are the preexponential factor and the activation energy.
MassoudKaviany 2011
Thermal Circuit Model for CombustionThermoelectric Energy Conversion Tube
L
A
3
T
f,1
x
T
f,2
T
f,3
T
f,4
T
f,5
T
f,6
T
f,7
T
f,8
T
f,9
T
f,10
T
f,11
T
f,12
T
f,13
T
f,14
T
s,2
T
s,1
T
s,3
T
s,4
T
s,5
T
s,6
T
s,7
(q
r,o
)
7
T
s,8
E
b,1
T
s,9 T
= (c
p
V)
s,i
dT
s,i
dt
+ (S
e,J
)
i
u
f
(t) u
f
(t)
R
u L,i
Tube Wall
Combustion
Region
Thermoelectric
Region
Thermoelectric
Region
T
s,10
T
s,11
T
s,12
T
s,13
Line of Symmetry
S
r,c
R
,
T
R
,
(R
k
)
i to i+1
(R
r,
)
i
L
TE
5
L
TE
5
A B C A A B B C
A B C
B C B B C
= (c
p
V)
s,i
dT
s,i
dt
+ (S
e,J
+ S
e,P
)
i
= (c
p
V)
s,i
dT
s,i
dt
Figure Ex. 8.4.2. The thermal circuit diagram for the combustionthermoelectric energy
conversion tube.
From (2.40), the power generated is
P
e.o
= R
e.o
J
2
e
= J
e
L =
_
N
t
(
S.p
S.n
)(T
h
T
c
)
R
e.o
R
e.hc
_
2
R
e.o
.
where J
e
is the current, L is the potential difference, N
t
is the number of the
tubes and R
e.o
is the external resistance. We use a bundle of tubes with the
bundle crosssectional area 0.03 (m) 0.03 (m) and the tubes are arranged in
an squarearray arrangement. Figure Ex. 8.4.3 shows the tubebundle model of
the combustionthermoelectric energy converter where the tubes are connected
electrically in series. Each of p and ntype thermoelectric materials uses a half of
the tube crosssectional area. We have N
t
tubes and connect themas an electrical
circuit in series. The electrical resistance R
e.hc
is the total resistance for a current
owing through the inner and outer cylindrical shells and is expressed as
R
e.hc
= 4N
t
e
L
TE
A
k
.
The other relations are
A
u
=
D
2
1
4
A
ku.i
= D
1
L
i
A
k
=
D
2
2
D
2
1
4
MassoudKaviany 2011
u
f
(
x
,
t
)
TubeBundle Model of Combustion
Thermoelectric Energy Converter
Reaction
Region
0
L
TE
L
T
E
L
T
E
+
L
A
L
=
2
L
T
E
+
L
A
x
L
A
s
r, c
L
TE
L
T
c
s
T
h
s
T
f
L
f
T
o
f
u
f
Reciprocating,
Gaseous
Oxident
Flow
Alternating,
Gaseous Fuel Flow
Fuel Flow Chamber
( )
(+)
(+)
J
e
L
TE
L
A
L
TE
u
f
(
x
,
t
)
T
h
e
r
m
o

e
l
e
c
t
r
i
c

S
t
o
r
a
g
e
R
e
g
i
o
n
T
h
e
r
m
o

e
l
e
c
t
r
i
c

S
t
o
r
a
g
e
R
e
g
i
o
n
C
o
m
b
u
s
t
i
o
n
R
e
g
i
o
n
E
l
e
c
t
r
i
c
a
l
C
o
n
d
u
c
t
o
r
Figure Ex. 8.4.3. The tubebundle model of the combustionthermoelectric energy con
verter.
A
r.i
= D
1
L
i
Nu)
D
= 3.66 laminar ow, Re
D.1
 2,300
R
k.ij
=
L
i
A
k
k
s
(R
r.c
)
i
=
_
1 c
r
A
r
c
r
_
i
(R
r.F
)
ij
=
1
A
r.i
F
ij
F
ij
=
_
inner ring to inner ring, Table 4.2
inner ring to end disk, Figure 4.11(a).
Figure Ex. 8.4.4 shows the temperature distribution at t = 5 s (end of a half
cycle with uid owing from left to right) for 13 solid nodes. The nodal uid and
solid temperatures correspond to nodes shown in Figure Ex. 8.4.2. The ame
is near node T
f .5
as evident from the high uid temperature there. Note that
the combustion region has a nearly uniform temperature. Table 8.4.2 lists the
computed Reynolds number for the conditions used in the above gures.
For an asymptotic solution of this problem, we need to have a much larger
number of divisions. Figure Ex. 8.4.5 shows the temperature distribution with
52 solid nodes.
MassoudKaviany 2011
Table Ex. 8.4.2. Calculated
values of variables for the
thermoelectric energy
conversion considered for 13
solid nodes computations
Variable Magnitude
L
TE.i
60 mm/5
L
A.i
7.5 mm/3
Re
D.i
54.5
The exit temperature T
f .L
is found by applying the integralvolume energy
equation (2.9) to the tubegas system and over the cycle .
The instantaneous energy equation, for no conduction and radiation a short
distance from the two ends of the tube, gives
Q
u
Q
loss
=
_
V
c
p
dT
dt
dV
S
r.c
(
S
e.P
)
h
(
S
e.P
)
c
S
e.J
.
T
ad
T
0
0
1
2
3
4
5
6
7
8
9
10
11
n
r
,
F
(
k
g
/
m
3

s
)
,
,
,
Combustion
Region
Predicted Nodal Temperatures at the End of Half Cycle for 14 Fluid Nodes
T
f
T
f,0
u
f
= 0.0538
T
f
T
f
,
0
T
s
T
f,0
T
s
T
f
,
0
f,P
f
n
r,F
f
,
P
f
0
0
0.06 0.0675 0.13
x, m
2L
TE
+ L
A
L
TE
+ L
A
L
TE
Figure Ex. 8.4.4. Temperature distribution along the combustionthermoelectric tube at
various uid and solid nodes for 14 uid nodes and 13 solid nodes.
MassoudKaviany 2011
0
1
2
3
4
5
6
7
8
9
10
11
12
0
0
0.06 0.0675 0.13
x, m
n
r
,
F
(
k
g
/
m
3

s
)
,
,
,
Combustion
Region
Predicted Nodal Temperatures at the End of Half Cycle for 52 Fluid Nodes
2L
TE
+ L
A
L
TE
+ L
A
L
TE
T
f
T
f, 0
u
f
= 0.1101
T
f
T
f
,
0
T
s
T
f, 0
T
ad
T
0
T
s
T
f
,
0
f,P
f
n
r,F
f
,
P
f
Figure Ex. 8.4.5. Temperature distribution along the combustionthermoelectric tube at
various uid and solid nodes for 52 uid nodes and 51 solid nodes.
Upon time integration over a complete period , where at the end of the period
the temperature everywhere is the same as it was at the beginning of the period
(i.e., the spatial temperature prole is periodic in time), we have
_
0
Mc
p.f
(T
f .L
T
f .0
)dt =
_
0
_
S
r.c
(
S
e.P
)
h
(
S
e.P
)
c
S
e.J
Q
loss
_
dt.
Solving for the timeaverage exit temperature T
f .L
, and assuming complete
combustion, we have
T
f .L
= T
f .0
_
0
A
u
f
Lh
r.F
u
f
) (
S
e.P
)
h
(
S
e.P
)
c
S
e.J
Q
loss
Mc
p.f
dt.
Note that (
S
e.P
)
h
 0, while (
S
e.P
)
c
> 0 and the Peltier heating takes place only
at pn junctions. The Joule heating is always positive. Here we have heat loss of
the form
Q
loss
=
T
s.13
T
R
Y.
T
s.1
T
R
Y.
.
MassoudKaviany 2011
x 100
0 10
/2
9 8 7 6 5 4 3 2 1
0
300
200
100
25
20
15
10
5
t, s
J
e
(
m
A
)
,
(
V
)
,
J
e
(
W
)
,
u
f
u
f
For 52 Solid Nodes
J
e
J
e
Figure Ex. 8.4.6. The variation of current, voltage, and power generated, and the ther
moelectric efciency, with respect to the elapsed time, for the conditions considered and
with 52 solid nodes.
where R
Y.
is the sum of the surfaceconvection and surfaceradiation resis
tances between the end surfaces of the tube to the ambient.
The efciency is dened as
(
S
e.P
)
h
(
S
e.P
)
c
S
e.J
A
u
f
Lh
r.F
u
f
)
.
Figure Ex. 8.4.6 shows the power generated J
e
L, current J
e
, potential L, and
the thermoelectric efciency during the cycle. Note that the power generated
is summed for two thermoelectric tubes at both ends.
COMMENT
It was suggested that the thermoelectric regions be each divided into ve nite
small volumes, and the combustion region into three nitesmall volumes. This
is rather arbitrary. In practice, the division into nitesmall volumes is improved
by progressively making further divisions, until the results do not change further
with any further increase in the number of divisions.
Improvements in the efciency can be made by injecting the fuel directly
into the combustion region near the hot junction and by intensive cooling at cold
junction. This results in a nonpremixed or diffusion ame.
MassoudKaviany 2011
EXAMPLE 8.5. DES.S
INCYLINDERTHERMAL REGENERATION: POROUSFOAM
ENGINE REGENERATOR
Althoughnewenergy forms are being developedfor automotive applications, for
example, fuel cells, electric batteries, solar cells, hybrid engines, etc., the internal
combustion engines are still the primary engines used in the automotive, mar
itime, railway, and small stationary electrical energy generation applications. The
diesel engines present a higher efciency than the gasoline engines, especially
for the highpower applications. The study and development of new engines
have been focused on improving the efciency and reducing the pollutant emis
sion. A relatively new and innovative concept [10,11] is the use of an incylinder
porous foam for thermal regeneration and catalytic destruction of gaseous and
particulate pollutants.
The operation of the regenerative engine is shown in Figure Ex. 8.5.1(a).
The porous regenerator is attached to a rod and moves up and down inside
the cylinder, synchronized to the movement of the piston. Figure Ex. 8.5.1(b)
shows the position of the top of the piston and the top of the regenerator, as a
function of the crank angle . For most of the period (720
Regenerator/Converter
Compression
Expansion
Exhaust
Connecting Rod
Crank Shaft
(a) CompressionIgnition Engine
Intake
180
o
360
o
540
o
720
o
Piston
Regenerator/
Converter
(b) Motion of the Regenerator/Converter
0
o
Bottom of Regenerator
Bottom
Dead
Center
(BDC)
Crank Angle,
Distance
from BDC
Expansion
Top of Piston
Exhaust Compression
x
Regenerative
Cooling Stroke
Regenerative
Heating Stroke
Top Dead
Center
(TDC)
Figure Ex. 8.5.1. (a) A compressionignition engine with direct fuel injection. Movement
of the porous regenerator through a complete cycle (720
)
c
b

i
,
b
(R
r,
)
cbp,b
R
ku,tcb DB
R
ku,ti DB
(R
r,
)
cbi,t
(Q
u,ins
)
b
S
e,
+ S
e,
Q
u,exh
Q
u,int
T
f,t
Piston (Aluminum)
Cylinder Head (Ceramic)
(R
k,ch
)
01
(R
k,
)
ch w
T
ch ,1
T
ch ,N
T
ch ,0
(
R
r
,
)
c
h

i
,
t
R
ku,tch DB
S
r,F,t
+S
m,p,t
+S
e,
+ S
e,
S
e,
+ S
e,
Q
u,inj
.
(R
k,
)
cbw
(R
k,
)
pw
R
ku,bp DB
S
e,
+ S
e,
(R
r,
)
chcb,t
. . .
Q
u,if,b
Q
u,if
(Q
u,ins
)
t
(Q
u,ins
)
t
(Q
u,ins
)
b
Figure Ex. 8.5.3. Thermal circuit diagram for the regenerative engine.
(b) Energy equations, mass conservation equations, the equations of state, and
other equations for the submodels (ow, combustion, etc.) are needed to model
the engine. These are given in the complete solution.
The relevant performance indicators are the thermal efciency
T
, the vol
umetric efciency
:
, the net indicated meaneffective pressure p
mep.i.n
, the
indicated specic fuel consumption Lh
1
isf c
, and the fuel conversion efciency
F
.
The net indicated work per cycle W
c.i.n
is given by
W
c.i.n
=
_
p
f
dV
f
.
The mass of fuel burned M
r.F
is
M
r.F
=
_
t
F.s
Lt
r.F
t
F.s
M
r.F
dt.
MassoudKaviany 2011
The thermal efciency
T
is then
T
=
W
c.i.n
M
r.F
Lh
r.F
.
The net indicated meaneffective pressure p
mep.i.n
is
p
mep.i.n
=
W
c.i.n
V
d
.
The volumetric efciency
V
is
V
=
M
a
(
:.e.i
)
M
a.o
.
where M
a
(
:.e.i
) is the amount of mass of air in the cylinder when the intake
valve closes.
The indicated power per cycle P
c.i
is
P
c.i
=
W
c.i.n
N
2(rot,cycle) 60(s,min)
.
The indicated specic fuel consumption Lh
1
isf c
is
Lh
1
isf c
=
M
F.o
W
c.i.n
.
The fuel conversion efciency
F
is given by
F
=
1,Lh
1
isf c
Lh
r.F
.
Figure Ex. 8.5.4(a) shows the cycle in the p
f
V
f
diagram from the beginning
= 0
270
o
1.0
0.0
1.0
2.0
3.0
4.0
0.0
0.2
0.4
0.6
0.8
1.0
360
o
Regenerative
Heating Stroke
Regenerative
Cooling Stroke
450
o
540
o
M
i
n
s
,
k
g
/
s
x
p
/
L
t
,
x
i
/
L
t
x
p
L
t
x
i
L
t
M
ins
(c)
500
1,000
1,500
2,000
2,500
3,000
T
,
K
1.0
0.8
0.6
0.2
0.0
0
o
90
o
180
o
270
o
360
o
450
o
540
o
630
o
720
o
T
s,o
to T
s,L
x
p
/
L
t
,
x
i
/
L
t
x
i
L
t
x
p
L
t
T
f,mix
T
f,t
T
f,b
0
T
ch,o
T
p,o
T
cb,o
Figure Ex. 8.5.4. (a) Predicted p
f
V
f
diagram from the beginning = 0
to the end
= 720
of the cycle for the nonregenerative and regenerative engines. The net indicated
work W
c.i.n
is calculated from the linear scale plot. (b) Predicted mass ow rate through
the insert for the compression and expansion periods. Also shown are the positions of
the top of the piston and bottom of the insert, identifying the regenerative cooling stroke
and the regenerative heating stroke. (c) Predicted crankangle variations of the gas and
solid temperatures.
MassoudKaviany 2011
Table Ex. 8.5.1. Comparison of results for nonregenerative and
regenerative engines
Parameter or Variable Nonregenerative Regenerative
V
d
cm
3
1,718 1,718
V
c
cm
3
110.8 245.4
r
c
16.5 10
M
a.o
g 6.494 6.494
M
F.o
g 0.2165 0.2165
(A,F)
a
30 30
p
f .max
MPa 23.9 top: 30.3
MPa bottom: 30.3
T
f .max
K 1,619 top: 2,305
K bottom: 964
T
exh
K 882 780
W
c.i.n
(720
) kJ 3.97 4.86
P
c.i
(720
) kW 53.0 64.8
p
mep.i.n
MPa 2.31 2.82
F
0.43 0.53
V
0.94 0.92
Lh
1
isf c
g/kWhr 196 160
chamber. During this period, there is a transfer of sensible heat from the gas to
the insert.
Figure Ex. 8.5.4(c) shows the crankangle variations of the gas and solid
temperatures. The regenerative cooling stroke is characterized by a decrease in
the insert temperature and the regenerative heating stroke is characterized by
an increase in the insert temperature. The high peak in top uid temperature
is due to combustion occurring in the hotter gas, when compared to the nonre
generative engine. Also shown in the gure is the mass averaged temperature of
the gas (T
f .mix
). The peak mass averaged temperature (c
p
is assumed constant)
is substantially smaller than the peak topuid temperature.
The comparison of the results for the nonregenerative and the regenerative
engines is made in Table Ex. 8.5.1. The result shows that the thermal efciency
increases from 43% to 53%. The volumetric efciency decrease slightly, from
94% to 92%. For the same power generation, the regenerative engine consumes
a lesser amount of fuel than the nonregenerative engine.
COMMENT
The high peak temperature experienced by the gas in the top chamber during
the premixedame regime of the combustion is a result of the high initial
temperature of the air (before injection starts) and the relatively smaller amount
of air available for combustion. A constant volume adiabatic ame temperature
for a stoichiometric mixture of methane and air at a initial temperature and
pressure of 1,600 K and 75 MPa is T
ad
= 3,845 K, in contrast to the constant
pressure adiabatic ame temperature of T
ad
= 3,418 K.
MassoudKaviany 2011
The insert can also be used as a catalytic converter for the particulates and
some of the exhaust gases. The particulates (soot) can be trapped during the
regenerative heating stroke and then burned during the regenerative cooling
stroke, when the supply of oxygen in the fresh air and the relatively high tem
perature of the solid matrix allow for the soot oxidation. When coated with a
catalytic material, like platinum or palladium, the insert can also oxidize the
unburned hydrocarbons and carbon monoxide during the regenerative heating
stroke, therefore, before the exhaust gas leaves the cylinder.
The local higher temperature of the gas in the top chamber may increase the
production of thermal NO
x
. The insert is not effective in the catalytic reduction
of NO
x
and that should be done at a tail catalytic converter. The less fuel
consumption in the regenerative engine lowers the peak ame temperature and
therefore, reduces the thermal NO
x
formation, while the power generated is
similar to that of the nonregenerative engine.
8.4 Summary
In this nal chapter, we have culminated our heat transfer analysis with a general
discussion of selection (i.e., design) of the heat transfer media and their bounding sur
faces. The chapter began with a review of the various heat transfer functions of these
media and surfaces, along with the geometrical and thermophysicalthermochemical
properties inuencing these functions. Charts 8.2 and 8.3 summarized these heat
transfer functions and properties. Then we summarized the volume and surface
energy equations as given by (8.1) and (8.2). The various thermal resistances devel
oped in the text are summarized in Tables 8.1 and 8.2. A general thermal circuit
diagram for a thermal node i is shown in Figure 8.2. The encountered ranges for
the magnitudes of the thermal resistances and the energy conversions [listed in
Table C.1(d)], and their asymptotic values, are given in Figures 8.3 and 8.4. These
should help the readers to connect and access all the useful relations and data in this
text, once the reader has become familiar with the early chapters.
In this chapter, we have also given some relatively extensive thermal design
examples, Examples 8.1 to 8.5, involving multiple heat transfer and energy conver
sion mechanisms. These examples start with simple thermal systems and move to
more elaborate systems. These examples are solved and results are discussed in the
text.
8.5 References
[1] Asai, A., 1991, Bubble Dynamic in Boiling Under High Heat Flux Pulse Heat
ing, ASME Journal of Heat Transfer, Volume 113, pp. 973979.
[2] Assanis, D.N. and Heywood, J.B., Development and Use of Computer Sim
ulation of the Turbocompound Diesel System for Engine Performance and
Components Heat Transfer Studies, SAE paper 860329, 1986.
[3] Bergstrom, P.L., Ji, J., Liu, Y.N., Kaviany, M., and Wise, K., 1995, Thermally
Driven PhaseChange Microactuation, Journal of Microelectromechanical
MassoudKaviany 2011
Systems, Volume 4, pp. 1017. Also, Rich, C. A. and Wise, K. D., 2003, A
High Flow Thermopneumatic Microvalve with Improved Efciency and Inte
grated State Sensing, Journal of Microelectromechanical Systems, Volume 12,
pp. 201208.
[4] Bhaskar, E.V. and Aden, J.S., 1985, Development of the ThinFilm
Structure for the Inkjet Printhead, HewlettPackard Journal, May 1985,
pp. 2733.
[5] Buskirk, W.A., Heckelman, D.E., Hall, S.T., Hanarek, P.H., Low, R.N., Trueba,
E., and van de Poll, R.R., 1988, Development of a HighResolution Thermal
Inkjet Printhead, HewlettPackard Journal, October 1988, pp. 5561.
[6] Chen, P.H., Chen, W.C., and Chang, S.H., 1997, International Journal of
Mechanical Science, Volume 39, pp. 683695.
[7] Dunn, P.D. and Reay, D.A., 1976, Heat Pipes, Pergamon Press, Oxford.
[8] Echigo, R., Tawata, K., Yoshida, H., and Tada, S., 1995, Effective Heat
ing/Cooling Method for Porous Thermoelectric Device in Reciprocation Flow
Combustion System, ASME/JSME Joint Thermal Engineering Conference,
Lahaina, Hawaii, Volume 4, pp. 389.
[9] Faghri, A., 1995, Heat Pipe Science and Technology, Taylor and Francis,
Washington.
[10] Ferrenberg, A.J., Regenerative Internal Combustion Engine, US patent
4,790,284, Dec. 13, 1988.
[11] Ferrenberg, A.J., Williams, B.E., andMcNeal, S.R., LowHeat RejectionRegen
erated Diesel Engines for Shipboard Mechanical and Electrical Power, Final
Report, Caterpillar Inc., Contact N0016792C0020, Feb. 1995.
[12] Heywood, J.B., 1988, Internal Combustion Engine Fundamentals, McGrawHill,
New York. Also, Borman, G.L., and Ragland, K.W., 1998, Combustion Engi
neering, McGrawHill, New York.
[13] IMSL, 1987, Users Manual, Math/Library Fortran Subroutines for Mathematical
Applications, IMSL, Inc. Houston.
[14] Johnson, D.A., The Effect of Heating Pulse Shape on Drop Size and Stability
in Thermal Inkjet, Society of PhotoOptical Instrumentation Engineers (SPIE)
Journal, Volume 1670, pp. 224228.
[15] Min, D.K., and Shin, H.D., 1991, Laminar Premixed Flame Stabilized Inside a
Honeycomb Ceramic, International Journal of Heat and Mass Transfer, Volume
34, pp. 341356.
[16] Nielsen, N.J., 1985, History of Thinkjet Printhead Development, Hewlett
Packard Journal, May 1985, pp. 410.
[17] Park, C.W. and Kaviany, M., 2000, CombustionThermoelectric Tube, ASME
Journal Heat Transfer, Volume 122, pp. 721729.
[18] Park, C.W. and Kaviany, M., 2002, EvaporationCombustion Affected by In
Cylinder, Reciprocating Porous Regenerator, ASME Journal Heat Transfer,
Volume 124, pp. 184194.
[19] Rezanka, I., 1992, Thermal Inkjet: A Review, Society of PhotoOptical Instru
mentation Engineers (SPIE) Journal, Volume 1670, pp. 192200.
[20] Silverstein, C.C., 1992, Design of Heat Pipe for Cooling and Heat Exchange,
Taylor and Francis, Washington.
[21] Silwell, G.L., Boucher, W.R., and McClelland, P.H., 1985, The Thinkjet Orice
Plate: A Part with Many Functions, HewlettPackard Journal, May 1985,
pp. 3337.
[22] Zdeblick, M.J. and Angell, J.B., 1987, A Microminiature ElectrictoFluidic
Valve, Digest of International Conference on SolidState Sensors and Actuators,
Tokyo, Japan, pp. 827829.
MassoudKaviany 2011
[23] Zhang, Y. and Wise, K.D., 1994, Performance of NonPlanar Silicon Diaphrous
under Large Deections, Journal of Microelectromechanical systems, Volume 3,
pp. 5968.
[24] Zuo, Z. J. and Faghri, A., 1998, A Network Thermodynamic Analysis of
the Heat Pipe, International Journal of Heat and Mass Transfer, Volume 41,
pp. 14731484.
8.6 Problems
8.6.1 Thermal System Design
PROBLEM 8.1. DES.S
To harvest the automobile exhaustgas heat, a shell of bismuthtelluride thermoelec
tric elements is placed around the tailpipe. This is shown in Figure Pr. 8.1. In principle,
placement of the elements closer to the exhaust manifold, where higher gas temper
atures are available, is more benecial. However, high temperature thermoelectric
materials are needed.
The local tailpipe wall temperature is T
f .h
) = T
h
and the elements are placed
between the pipe and an outerceramic shell. The contact resistance between the
thermoelectric elements and the pipe is (R
k.c
)
h
and that with the ceramic is (R
k.c
)
c
.
The ceramic is cooled with surface convection using a crossow air stream with a
fareld temperature T
f .
. The ceramic surface is also cooled by surface radiation
to the surrounding solid surfaces (all surfaces are diffuse, gray opaque) with the
L
TE
a
TE
a
TE
T
c
T
h
Physical Model
Outside Ceramic Lining
Electric Power
Generation
Tailpipe
D
T
s,
T
c
T
h
R
e,o
J
e
2
L
r,
r,c
Hot
Exhaust
Gas
Cold Air Stream
u
f,
, T
f,
Thermoelectric
Modules, N
TE
Pairs
Surrounding Solid Surface
nType
pType
Contact Resistance, (R
ku
)
c
Outside
Ceramic
Lining
Contact Resistance, (R
ku
)
h
J
e
R
e,o
Figure Pr. 8.1. A thermoelectric power generator using the automobile exhaustgas heat
content.
MassoudKaviany 2011
surrounding solid surfaces at T
s.
. The emissivity of the ceramic surface is c
r.c
and
that of surrounding surfaces is c
r.
.
The electrical power generated is maximized with respect to the external
electrical resistance R
e.o
, by choosing R
e.o
= R
e
.
(a) Draw the thermal circuit diagram.
(b) For the conditions below, determine the power output J
2
e
R
e.o
.
(c) Plot the variation of the electrical power generation with respect to the surface
convection resistance R
ku
. Comment on how this resistance can be reduced.
N
TE
= 1,000, a
TE
= 2 mm, L
TE
= 3 mm, L = 30 cm, D= 4 cm, T
h
= 300
C,
T
f .
= 20
C, T
s.
= 20
C, u
f .
= 50 km/hr, c
r.c
= 0.90, c
r.
= 0.85, (R
k.c
)
c
=
(R
k.c
)
h
= 0.
Determine the thermophysical properties of air at T = 350 K.
PROBLEM 8.2. DES
Anesthetic drugs are available in liquid form and are evaporated, heated, and mixed
with other gases (e.g., oxygen) in a vaporizer tube. This is shown in Figure Pr. 8.2.
The heat is supplied through the Joule heating and the tube is ideally insulated on the
outside. The primary and secondary air mix with the evaporated drug, and initially
the temperature of the gas mixture drops from T
i
(inlet condition) to T
f
)
0
(after
the assumed complete evaporation).
(a) Draw the thermal circuit diagram.
(b) Determine the gas mixture temperature after the assumed complete evaporation
and before the surface convection begins.
(c) Determine the required Joule heating rate
S
e.J
for an exit temperature T
f
)
L
=
30
C.
(d) Determine the tube surface temperature T
s
needed to have an exit temperature
T
f
)
L
= 30
C.
Physical Model
( ) (+)
Primary and Secondary
Gas (Oxygen) Stream
Droplets
Joule Heating
Ideal
Insulation
Gas Mixture Out
Liquid Drug and
Gaseous Oxygen
L
D
R
e
J
e
2
M
O
2
, T
i
Liquid (e.g., Enfluorane,
Isofluorane, Halothane, or
Sevofluorane; Properties
Similar to Refrigerant 134a) Stream
M
l
, T
i
T
f L
Q
ku L
T
s
S
lg
T
f 0
Figure Pr. 8.2. An anesthetic liquid evaporatorvapoheater.
MassoudKaviany 2011
D= 2.3 cm, L = 18 cm, T
i
= 20
C,
M
O
2
=
M
p
M
s
= 2.17 10
4
kg/s,
M
l
=
1.66 10
5
kg/s.
Evaluate the thermophysical properties of the drug (assume that they are the
same as those for Refrigerant 134a) at T = 253.2 K, and the properties of oxygen at
T = 300 K. Use the properties of oxygen for the evaluation of the Nusselt number
and (
Mc
p
)
f
.
PROBLEM 8.3. DES
A condenserchemical analyzer uses a stream of cold water, where the temperature
of the stream entering the condenser is controlled to within a small deviation.
An offtheshelf thermoelectric coolerheat exchange unit is used to provide
this cold water stream. This is shown in Figure Pr. 8.3. The heat exchange surface
is assumed to be at a uniform temperature T
s
with the water entering at T
f
)
0
and
exiting at T
f
)
L
. The heat exchange surface is the cold side of a thermoelectric
module (i.e., T
s
= T
c
), where there are N
TE
pairs of bismuthtelluride thermoelectric
elements with the specications given below. The hot side of the thermoelectric
module is connected to an extended surface (ns or heat sink) cooled by air blown
over its surfaces with fareld conditions T
f .
and u
f .
. The ns have an efciency
of unity.
Air Flow
Air Flow
( )
(+)
l
l
Water Flow Channel
with Length, L
Water
w
u
f,
, T
f,
T
f 0
T
s
a
TE
a
TE
L
TE
T
f L
M
f
Physical Model
Channel Length, L
A
ku
Figure Pr. 8.3. A thermoelectric coolerheat exchange unit used to cool a bounded water
stream.
MassoudKaviany 2011
The dominant resistances between T
f .
and T
f
)
0
are due to the surface
convection, conduction in thermoelectric materials, and the average convection
resistance.
(a) Draw the thermal circuit diagram.
(b) Plot the heat removal rate from the water stream Q
u
)
L0
, as function of the
current for 0.2 J
e
1.4 A. (c) Comment on the optimum current.
T
f
)
0
= 20
C,
M
f .n
= 0.01 kg/s, l = 5 mm, L = 35 cm, a
TE
= 1.5 mm, L
TE
= 3.5
mm, N
TE
= 400, A
ku
= 0.05 m
2
, u
f .
= 1 m/s, T
f .
= 25
C, n = 9 cm.
Use properties of air at T = 300 K and water at T = 290 K, and determine
Nu)
n
using n, the n width.
PROBLEM 8.4. DES
The automobile exhaust pollutants (NO, CO, and unburned hydrocarbons) escape
conversion i