AC Course PRINTING 2016-2018
AC Course PRINTING 2016-2018
AC Course PRINTING 2016-2018
CHAPTER
MASS TRANSFER 14
o this point we have restricted our attention to heat transfer problems
T that did not involve any mass transfer. However, many significant heat
transfer problems encountered in practice involve mass transfer. For ex-
ample, about one-third of the heat loss from a resting person is due to evapo-
CONTENTS
14–1 Introduction 718
14–2 Analogy between Heat
ration. It turns out that mass transfer is analogous to heat transfer in many and Mass Transfer 719
respects, and there is close resemblance between heat and mass transfer rela- 14–3 Mass Diffusion 721
tions. In this chapter we discuss the mass transfer mechanisms and develop re- 14–4 Boundary Conditions 727
lations for the mass transfer rate for some situations commonly encountered 14–5 Steady Mass Diffusion
in practice. through a Wall 732
Distinction should be made between mass transfer and the bulk fluid motion 14–6 Water Vapor Migration
(or fluid flow) that occurs on a macroscopic level as a fluid is transported from in Buildings 736
one location to another. Mass transfer requires the presence of two regions at 14–7 Transient Mass Diffusion 740
different chemical compositions, and mass transfer refers to the movement of 14–8 Diffusion in a
a chemical species from a high concentration region toward a lower concen- Moving Medium 743
tration one relative to the other chemical species present in the medium. The 14–9 Mass Convection 754
primary driving force for fluid flow is the pressure difference, whereas for 14–10 Simultaneous Heat
mass transfer it is the concentration difference. Therefore, we do not speak of and Mass Transfer 763
mass transfer in a homogeneous medium.
We begin this chapter by pointing out numerous analogies between heat and
mass transfer and draw several parallels between them. We then discuss
boundary conditions associated with mass transfer and one-dimensional
steady and transient mass diffusion. Following is a discussion of mass trans-
fer in a moving medium. Finally, we consider convection mass transfer and
simultaneous heat and mass transfer.
717
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 718
718
HEAT TRANSFER
14–1 ■
INTRODUCTION
Water Salty It is a common observation that whenever there is an imbalance of a
water
commodity in a medium, nature tends to redistribute it until a “balance” or
Salt “equality” is established. This tendency is often referred to as the driving
force, which is the mechanism behind many naturally occurring transport
(a) Before (b) After phenomena.
FIGURE 14–1 If we define the amount of a commodity per unit volume as the concentra-
Whenever there is concentration tion of that commodity, we can say that the flow of a commodity is always in
difference of a physical quantity in the direction of decreasing concentration; that is, from the region of high con-
a medium, nature tends to equalize centration to the region of low concentration (Fig. 14–1). The commodity sim-
things by forcing a flow from the ply creeps away during redistribution, and thus the flow is a diffusion process.
high to the low concentration region. The rate of flow of the commodity is proportional to the concentration gradi-
ent dC/dx, which is the change in the concentration C per unit length in the
flow direction x, and the area A normal to flow direction and is expressed as
or
· dC
Q kdiff A (14-1)
dx
719
CHAPTER 14
the number moving to the left, resulting in zero net transfer of N2 or O2 across
an imaginary plane.
The molecules in a gas mixture continually collide with each other, and the
diffusion process is strongly influenced by this collision process. The collision
of like molecules is of little consequence since both molecules are identical
and it makes no difference which molecule crosses a certain plane. The colli-
sions of unlike molecules, however, influence the rate of diffusion since un-
like molecules may have different masses and thus different momentums, and
thus the diffusion process will be dominated by the heavier molecules. The
diffusion coefficients and thus diffusion rates of gases depend strongly on
temperature since the temperature is a measure of the average velocity of gas
molecules. Therefore, the diffusion rates will be higher at higher temperatures.
Mass transfer can also occur in liquids and solids as well as in gases. For ex-
ample, a cup of water left in a room will eventually evaporate as a result of
water molecules diffusing into the air (liquid-to-gas mass transfer). A piece of
solid CO2 (dry ice) will also get smaller and smaller in time as the CO2 mole-
cules diffuse into the air (solid-to-gas mass transfer). A spoon of sugar in a
cup of coffee will eventually move up and sweeten the coffee although the
sugar molecules are much heavier than the water molecules, and the mole-
cules of a colored pencil inserted into a glass of water will diffuse into the
water as evidenced by the gradual spread of color in the water (solid-to-liquid
mass transfer). Of course, mass transfer can also occur from a gas to a liquid
or solid if the concentration of the species is higher in the gas phase. For ex-
ample, a small fraction of O2 in the air diffuses into the water and meets the
oxygen needs of marine animals. The diffusion of carbon into iron during
case-hardening, doping of semiconductors for transistors, and the migration of
doped molecules in semiconductors at high temperature are examples of solid-
to-solid diffusion processes (Fig. 14–3). Air Air
Another factor that influences the diffusion process is the molecular Water vapor CO2
spacing. The larger the spacing, in general, the higher the diffusion rate.
Therefore, the diffusion rates are typically much higher in gases than they are
in liquids and much higher in liquids than in solids. Diffusion coefficients in
gas mixtures are a few orders of magnitude larger than these of liquid or solid Liquid Dry
water ice
solutions.
14–2 ■
ANALOGY BETWEEN (a) Liquid to gas (b) Solid to gas
HEAT AND MASS TRANSFER
We have spent a considerable amount of time studying heat transfer, and we
could spend just as much time (perhaps more) studying mass transfer. How- Coffee Iron
ever, the mechanisms of heat and mass transfer are analogous to each other,
and thus we can develop an understanding of mass transfer in a short time Carbon
with little effort by simply drawing parallels between heat and mass transfer. Sugar
Establishing those “bridges” between the two seemingly unrelated areas will (c) Solid to liquid (d) Solid to solid
make it possible to use our heat transfer knowledge to solve mass transfer
FIGURE 14–3
problems. Alternately, gaining a working knowledge of mass transfer will help
Some examples of mass transfer that
us to better understand the heat transfer processes by thinking of heat as involve a liquid and/or a solid.
a massless substance as they did in the nineteenth century. The short-lived
caloric theory of heat is the origin of most heat transfer terminology used
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 720
720
HEAT TRANSFER
70°C
today and served its purpose well until it was replaced by the kinetic theory.
Mass is, in essence, energy since mass and energy can be converted to each
Heat other according to Einstein’s formula E mc2, where c is the speed of light.
Therefore, we can look at mass and heat as two different forms of energy and
10°C exploit this to advantage without going overboard.
Heat concentration
Mass concentration Temperature
70%
The driving force for heat transfer is the temperature difference. In contrast,
CO2 the driving force for mass transfer is the concentration difference. We can
Mass view temperature as a measure of “heat concentration,” and thus a high tem-
perature region as one that has a high heat concentration (Fig. 14–4). There-
10%
CO2
fore, both heat and mass are transferred from the more concentrated regions to
the less concentrated ones. If there is no temperature difference between two
FIGURE 14–4 regions, then there is no heat transfer. Likewise, if there is no difference be-
Analogy between tween the concentrations of a species at different parts of a medium, there will
heat and mass transfer. be no mass transfer.
Conduction
Thermal You will recall that heat is transferred by conduction, convection, and radia-
radiation
tion. Mass, however, is transferred by conduction (called diffusion) and con-
No mass vection only, and there is no such thing as “mass radiation” (unless there is
radiation something Scotty knows that we don’t when he “beams” people to anywhere
Hot
in space at the speed of light) (Fig. 14–5). The rate of heat conduction in a di-
Mass rection x is proportional to the temperature gradient dT/dx in that direction and
body
is expressed by Fourier’s law of heat conduction as
FIGURE 14–5 · dT
Unlike heat radiation, there is Q cond kA (14-2)
dx
no such thing as mass radiation.
where k is the thermal conductivity of the medium and A is the area normal to
the direction of heat transfer. Likewise, the rate of mass diffusion m· diff of a
chemical species A in a stationary medium in the direction x is proportional to
Temperature the concentration gradient dC/dx in that direction and is expressed by Fick’s
profile law of diffusion by (Fig. 14–6)
A
· dT dCA
Qcond = – kA —
dx
m· diff DAB A (14-3)
dx
where DAB is the diffusion coefficient (or mass diffusivity) of the species in
x the mixture and CA is the concentration of the species in the mixture at that
location.
Concentration profile
of species A
It can be shown that the differential equations for both heat conduction and
A mass diffusion are of the same form. Therefore, the solutions of mass dif-
dC fusion equations can be obtained from the solutions of corresponding heat
m· diff = –DAB A ——A conduction equations for the same type of boundary conditions by simply
dx
switching the corresponding coefficients and variables.
721
CHAPTER 14
Heat generation occurs throughout the medium and exhibits itself as a rise in
temperature. Similarly, some mass transfer problems involve chemical reac-
tions that occur within the medium and result in the generation of a species
throughout. Therefore, species generation is a volumetric phenomenon, and
the rate of generation may vary from point to point in the medium. Such reac-
tions that occur within the medium are called homogeneous reactions and are
analogous to internal heat generation. In contrast, some chemical reactions
result in the generation of a species at the surface as a result of chemical re-
actions occurring at the surface due to contact between the medium and the
surroundings. This is a surface phenomenon, and as such it needs to be treated
as a boundary condition. In mass transfer studies, such reactions are called
heterogeneous reactions and are analogous to specified surface heat flux.
Convection
You will recall that heat convection is the heat transfer mechanism that in-
volves both heat conduction (molecular diffusion) and bulk fluid motion.
Fluid motion enhances heat transfer considerably by removing the heated
fluid near the surface and replacing it by the cooler fluid further away. In the
limiting case of no bulk fluid motion, convection reduces to conduction. Like-
wise, mass convection (or convective mass transfer) is the mass transfer
mechanism between a surface and a moving fluid that involves both mass dif-
fusion and bulk fluid motion. Fluid motion also enhances mass transfer con-
siderably by removing the high concentration fluid near the surface and
replacing it by the lower concentration fluid further away. In mass convection,
we define a concentration boundary layer in an analogous manner to the ther-
mal boundary layer and define new dimensionless numbers that are counter-
parts of the Nusselt and Prandtl numbers.
The rate of heat convection for external flow was expressed conveniently by
Newton’s law of cooling as
·
Q conv hconv As(Ts T) (14-4)
where hconv is the heat transfer coefficient, As is the surface area, and Ts T
is the temperature difference across the thermal boundary layer. Likewise, the
rate of mass convection can be expressed as (Fig. 14–7) Mass
transfer Concentration
m· conv hmass As(Cs C) (14-5) coefficient difference
where hmass is the mass transfer coefficient, As is the surface area, and Cs C
→
Mass 678
→
is a suitable concentration difference across the concentration boundary layer. convection: m· conv hmass As(Cs C)
Various aspects of the analogy between heat and mass convection are ex- Heat ·
plored in Section 14–9. The analogy is valid for low mass transfer rate cases convection: Q conv hconv As(Ts T)
123
→
in which the flow rate of species undergoing mass flow is low (under 10 per-
→
cent) relative to the total flow rate of the liquid or gas mixture.
Heat
transfer Temperature
coefficient difference
14–3 ■
MASS DIFFUSION
Fick’s law of diffusion, proposed in 1855, states that the rate of diffusion of a FIGURE 14–7
chemical species at a location in a gas mixture (or liquid or solid solution) is Analogy between convection heat
proportional to the concentration gradient of that species at that location. transfer and convection mass transfer.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 722
722
HEAT TRANSFER
A + B mixture B Although a higher concentration for a species means more molecules of that
species per unit volume, the concentration of a species can be expressed in
several ways. Next we describe two common ways.
A
1 Mass Basis
On a mass basis, concentration is expressed in terms of density (or mass con-
centration), which is mass per unit volume. Considering a small volume V at
V = VA = VB
a location within the mixture, the densities of a species (subscript i) and of the
m = mA + mB mixture (no subscript) at that location are given by (Fig. 14–8)
ρ = ρA + ρB
Partial density of species i: i mi /V (kg/m3)
C = CA + CB
Mass basis:
Total density of mixture: m/V m /V
i i
m m ρ
ρA = —–A , ρ = — , wA = —A– Therefore, the density of a mixture at a location is equal to the sum of the
V V ρ
densities of its constituents at that location. Mass concentration can also be
Mole basis:
expressed in dimensionless form in terms of mass fraction w as
N N C
CA = —A– , C = — , yA = —A–
V V C mi mi /V i
Relation between them:
Mass fraction of species i: wi m (14-6)
m /V
ρ M
CA = —A– , wA = yA —–A
MA M Note that the mass fraction of a species ranges between 0 and 1, and the con-
FIGURE 14–8 servation of mass requires that the sum of the mass fractions of the con-
Different ways of expressing the concentration stituents of a mixture be equal to 1. That is, wi 1. Also note that the density
of species A of a binary mixture A and B. and mass fraction of a constituent in a mixture, in general, vary with location
unless the concentration gradients are zero.
2 Mole Basis
On a mole basis, concentration is expressed in terms of molar concentration
(or molar density), which is the amount of matter in kmol per unit volume.
Again considering a small volume V at a location within the mixture, the mo-
lar concentrations of a species (subscript i) and of the mixture (no subscript)
at that location are given by
Partial molar concentration of species i: Ci Ni /V (kmol/m3)
Total molar concentration of mixture: C N/V N /V C
i i
Again the mole fraction of a species ranges between 0 and 1, and the sum of
the mole fractions of the constituents of a mixture is unity, yi 1.
The mass m and mole number N of a substance are related to each other by
m NM (or, for a unit volume, CM) where M is the molar mass (also
called the molecular weight) of the substance. This is expected since the mass
of 1 kmol of the substance is M kg, and thus the mass of N kmol is NM kg.
Therefore, the mass and molar concentrations are related to each other by
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 723
723
CHAPTER 14
i
Ci (for species i) and C (for the mixture) (14-8)
Mi M
where M is the molar mass of the mixture which can be determined from
N M Ni
N M y M
m i i
M i i i (14-9)
N N
The mass and mole fractions of species i of a mixture are related to each
other by
i Ci Mi Mi
wi yi (14-10)
CM M
Two different approaches are presented above for the description of con-
centration at a location, and you may be wondering which approach is better
to use. Well, the answer depends on the situation on hand. Both approaches
are equivalent, and the better approach for a given problem is the one that
yields the desired solution more easily.
724
HEAT TRANSFER
Higher Lower But the concentration of a species in a gas mixture or liquid or solid solution
concentration concentration can be defined in several ways such as density, mass fraction, molar concen-
of species A of species A
tration, and mole fraction, as already discussed, and thus Fick’s law can be ex-
pressed mathematically in many ways. It turns out that it is best to express the
dC concentration gradient in terms of the mass or mole fraction, and the most ap-
slope = ——A
dx propriate formulation of Fick’s law for the diffusion of a species A in a sta-
tionary binary mixture of species A and B in a specified direction x is given by
Area (Fig. 14–10)
A
m· diff, A d(A /) dwA
CA(x) Mass basis: jdiff, A DAB DAB (kg/s · m2)
A dx dx
Concentration
profile of species A N· diff, A d(CA/C) dyA
Mole basis: j̄diff, A CDAB CDAB (mol/s · m2) (14-12)
A dx dx
x
Here jdiff, A is the (diffusive) mass flux of species A (mass transfer by diffusion
Mass basis: per unit time and per unit area normal to the direction of mass transfer, in
dwA
m· diff = –ρADAB ——
kg/s · m2) and j̄diff, A is the (diffusive) molar flux (in kmol/s · m2). The mass
dx flux of a species at a location is proportional to the density of the mixture at
d(ρA/ρ)
= – ρADAB ———
that location. Note that A B is the density and C CA CB is the
dx molar concentration of the binary mixture, and in general, they may vary
dρA
= –ADAB —— (if ρ = constant)
throughout the mixture. Therefore, d(A /) dA or Cd(CA /C) dCA. But
dx in the special case of constant mixture density or constant molar concentra-
tion C, the relations above simplify to
Mole basis:
dA
· dyA
Ndiff, A = –CADAB —— Mass basis ( constant): jdiff, A DAB (kg/s · m2)
dx dx
d(CA/C) dCA
= –CADAB ——— – Mole basis (C constant): j̄diff, A DAB (kmol/s · m2) (14-13)
dx dx
dCA
= –ADAB —— (if C = constant)
dx The constant density or constant molar concentration assumption is usually
FIGURE 14–10 appropriate for solid and dilute liquid solutions, but often this is not the case
Various expressions of Fick’s law of for gas mixtures or concentrated liquid solutions. Therefore, Eq. 14–12 should
diffusion for a binary mixture. be used in the latter case. In this introductory treatment we will limit our con-
sideration to one-dimensional mass diffusion. For two- or three-dimensional
cases, Fick’s law can conveniently be expressed in vector form by simply re-
placing the derivatives in the above relations by the corresponding gradients
(such as jA DAB wA).
Mass Concentration
Remember that the constant of proportionality in Fourier’s law was defined
diffusivitygradient as the transport property thermal conductivity. Similarly, the constant of pro-
→
Thermal Temperature are usually determined experimentally. The kinetic theory of gases indicates
conductivity gradient that the diffusion coefficient for dilute gases at ordinary pressures is essen-
tially independent of mixture composition and tends to increase with temper-
FIGURE 14–11 ature while decreasing with pressure as
Analogy between Fourier’s law of
DAB, 1 P2 T1 3/2
heat conduction and Fick’s
law of mass diffusion.
DAB
T 3/2
P
or
DAB, 2 P1 T2
(14-14)
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 725
725
CHAPTER 14
This relation is useful in determining the diffusion coefficient for gases at dif- TABLE 14–1
ferent temperatures and pressures from a knowledge of the diffusion coeffi-
cient at a specified temperature and pressure. More general but complicated Binary diffusion coefficients of
some gases in air at 1 atm pressure
relations that account for the effects of molecular collisions are also available.
(from Mills, Ref. 13, Table A.17a,
The diffusion coefficients of some gases in air at 1 atm pressure are given in p. 869)
Table 14–1 at various temperatures.
The diffusion coefficients of solids and liquids also tend to increase with Binary Diffusion Coefficient,*
temperature while exhibiting a strong dependence on the composition. The m2/s 105
diffusion process in solids and liquids is a great deal more complicated than T, K O2 CO2 H2 NO
that in gases, and the diffusion coefficients in this case are almost exclusively 200 0.95 0.74 3.75 0.88
determined experimentally. 300 1.88 1.57 7.77 1.80
The binary diffusion coefficient for several binary gas mixtures and solid 400 5.25 2.63 12.5 3.03
and liquid solutions are given in Tables 14–2 and 14–3. We make these two 500 4.75 3.85 17.1 4.43
observations from these tables: 600 6.46 5.37 24.4 6.03
700 8.38 6.84 31.7 7.82
1. The diffusion coefficients, in general, are highest in gases and lowest 800 10.5 8.57 39.3 9.78
in solids. The diffusion coefficients of gases are several orders of 900 12.6 10.5 47.7 11.8
magnitude greater than those of liquids. 1000 15.2 12.4 56.9 14.1
2. Diffusion coefficients increase with temperature. The diffusion 1200 20.6 16.9 77.7 19.2
coefficient (and thus the mass diffusion rate) of carbon through iron 1400 26.6 21.7 99.0 24.5
during a hardening process, for example, increases by 6000 times as 1600 33.2 27.5 125 30.4
the temperature is raised from 500°C to 1000°C. 1800 40.3 32.8 152 37.0
2000 48.0 39.4 180 44.8
Due to its practical importance, the diffusion of water vapor in air has been
the topic of several studies, and some empirical formulas have been developed *Multiply by 10.76 to convert to ft2/s.
for the diffusion coefficient DH2O–air. Marrero and Mason proposed this popu-
lar formula (Table 14–4):
TABLE 14–2
Binary diffusion coefficients of dilute gas mixtures at 1 atm
(from Barrer, Ref. 2; Geankoplis, Ref. 5; Perry, Ref. 14; and Reid et al., Ref. 15)
Substance Substance T, DAB or DBA, Substance T, DAB or DBA,
A B K m2/s Substance A B K m2/s
Air Acetone 273 1.1 105 Argon, Ar Nitrogen, N2 293 1.9 105
Air Ammonia, NH3 298 2.6 105 Carbon dioxide, CO2 Benzene 318 0.72 105
Air Benzene 298 0.88 105 Carbon dioxide, CO2 Hydrogen, H2 273 5.5 105
Air Carbon dioxide 298 1.6 105 Carbon dioxide, CO2 Nitrogen, N2 293 1.6 105
Air Chlorine 273 1.2 105 Carbon dioxide, CO2 Oxygen, O2 273 1.4 105
Air Ethyl alcohol 298 1.2 105 Carbon dioxide, CO2 Water vapor 298 1.6 105
Air Ethyl ether 298 0.93 105 Hydrogen, H2 Nitrogen, N2 273 6.8 105
Air Helium, He 298 7.2 105 Hydrogen, H2 Oxygen, O2 273 7.0 105
Air Hydrogen, H2 298 7.2 105 Oxygen, O2 Ammonia 293 2.5 105
Air Iodine, I2 298 0.83 105 Oxygen, O2 Benzene 296 0.39 105
Air Methanol 298 1.6 105 Oxygen, O2 Nitrogen, N2 273 1.8 105
Air Mercury 614 4.7 105 Oxygen, O2 Water vapor 298 2.5 105
Air Napthalene 300 0.62 105 Water vapor Argon, Ar 298 2.4 105
Air Oxygen, O2 298 2.1 105 Water vapor Helium, He 298 9.2 105
Air Water vapor 298 2.5 105 Water vapor Nitrogen, N2 298 2.5 105
Note: The effect of pressure and temperature on DAB can be accounted for through DAB ~ T3/2/P. Also, multiply DAB values by 10.76 to convert them to ft2/s.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 726
726
HEAT TRANSFER
TABLE 14–3
Binary diffusion coefficients of dilute liquid solutions and solid solutions at 1 atm
(from Barrer, Ref. 2; Reid et al., Ref. 15; Thomas, Ref. 19; and van Black, Ref. 20)
(a) Diffusion through Liquids (b) Diffusion through Solids
Substance Substance T, Substance Substance B T,
A (Solute) B (Solvent) K DAB, m2/s A (Solute) (Solvent) K DAB, m2/s
Ammonia Water 285 1.6 109 Carbon dioxide Natural rubber 298 1.1 1010
Benzene Water 293 1.0 109 Nitrogen Natural rubber 298 1.5 1010
Carbon dioxide Water 298 2.0 109 Oxygen Natural rubber 298 2.1 1010
Chlorine Water 285 1.4 109 Helium Pyrex 773 2.0 1012
Ethanol Water 283 0.84 109 Helium Pyrex 293 4.5 1015
Ethanol Water 288 1.0 109 Helium Silicon dioxide 298 4.0 1014
Ethanol Water 298 1.2 109 Hydrogen Iron 298 2.6 1013
Glucose Water 298 0.69 109 Hydrogen Nickel 358 1.2 1012
Hydrogen Water 298 6.3 109 Hydrogen Nickel 438 1.0 1011
Methane Water 275 0.85 109 Cadmium Copper 293 2.7 1019
Methane Water 293 1.5 109 Zinc Copper 773 4.0 1018
Methane Water 333 3.6 109 Zinc Copper 1273 5.0 1013
Methanol Water 288 1.3 109 Antimony Silver 293 3.5 1025
Nitrogen Water 298 2.6 109 Bismuth Lead 293 1.1 1020
Oxygen Water 298 2.4 109 Mercury Lead 293 2.5 1019
Water Ethanol 298 1.2 109 Copper Aluminum 773 4.0 1014
Water Ethylene glycol 298 0.18 109 Copper Aluminum 1273 1.0 1010
Water Methanol 298 1.8 109 Carbon Iron (fcc) 773 5.0 1015
Chloroform Methanol 288 2.1 109 Carbon Iron (fcc) 1273 3.0 1011
T 2.072
DH2O–Air 1.87 1010 (m2/s), 280 K T 450 K (14-15)
P
TABLE 14–4
In a binary ideal gas mixture of where P is total pressure in atm and T is the temperature in K.
species A and B, the diffusion The primary driving mechanism of mass diffusion is the concentration
coefficient of A in B is equal to gradient, and mass diffusion due to a concentration gradient is known as the
the diffusion coefficient of B in A,
ordinary diffusion. However, diffusion may also be caused by other effects.
and both increase with temperature
Temperature gradients in a medium can cause thermal diffusion (also called
DH2O–Air or DAir–H2O the soret effect), and pressure gradients may result in pressure diffusion.
at 1 atm, in m2/s Both of these effects are usually negligible, however, unless the gradients are
T, °C (from Eq. 14–15)
very large. In centrifuges, the pressure gradient generated by the centrifugal
0 2.09 105 effect is used to separate liquid solutions and gaseous isotopes. An external
5 2.17 105 force field such as an electric or magnetic field applied on a mixture or solu-
10 2.25 105 tion can be used successfully to separate electrically charged or magnetized
15 2.33 105 molecules (as in an electrolyte or ionized gas) from the mixture. This is called
20 2.42 105 forced diffusion. Also, when the pores of a porous solid such as silica-gel are
25 2.50 105
smaller than the mean free path of the gas molecules, the molecular collisions
30 2.59 105
35 2.68 105
may be negligible and a free molecule flow may be initiated. This is known as
40 2.77 105 Knudsen diffusion. When the size of the gas molecules is comparable to the
50 2.96 105 pore size, adsorbed molecules move along the pore walls. This is known as
100 3.99 105 surface diffusion. Finally, particles whose diameter is under 0.1 m such as
150 5.18 105 mist and soot particles act like large molecules, and the diffusion process of
such particles due to the concentration gradient is called Brownian motion.
Large particles (those whose diameter is greater than 1 m) are not affected
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 727
727
CHAPTER 14
Then the mass fractions of constituent gases are determined from Eq. 14–10 to
be
MN2 28.0
N2: wN2 yN2 (0.781) 0.754
M 29.0
MO2 32.0
O2: wO2 yO2 (0.209) 0.231
M 29.0
MAr 39.9
Ar: wAr yAr (0.01) 0.014
M 29.0
x
Therefore, the mass fractions of N2, O2, and Ar in dry standard atmosphere are
75.4 percent, 23.1 percent, and 1.4 percent, respectively. Air
14–4 ■
BOUNDARY CONDITIONS
We mentioned earlier that the mass diffusion equation is analogous to the heat yH
2O, gas side (0)
diffusion (conduction) equation, and thus we need comparable boundary con- 0
ditions to determine the species concentration distribution in a medium. Two
common types of boundary conditions are the (1) specified species concen- Jump in yH
2O, liquid side
concentration = 1.0
tration, which corresponds to specified temperature, and (2) specified species
flux, which corresponds to specified heat flux. Water
Despite their apparent similarity, an important difference exists between Concentration
temperature and concentration: temperature is necessarily a continuous func- profile
tion, but concentration, in general, is not. The wall and air temperatures at a FIGURE 14–13
wall surface, for example, are always the same. The concentrations of air on Unlike temperature, the concentration
the two sides of a water–air interface, however, are obviously very different of species on the two sides of a
(in fact, the concentration of air in water is close to zero). Likewise, the liquid–gas (or solid–gas or
concentrations of water on the two sides of a water–air interface are also dif- solid–liquid) interface are
ferent even when air is saturated (Fig. 14–13). Therefore, when specifying a usually not the same.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 728
728
HEAT TRANSFER
Using Fick’s law, the constant species flux boundary condition for a diffus-
ing species A at a boundary at x 0 is expressed, in the absence of any blow-
ing or suction, as
dyA dwA
T(x) CDAB j̄A, 0 or DAB jA, 0 (14-17)
Insulated
dx x0 dx x0
dT(0)
surface ——– = 0
dx
·
Q(0) = 0 where j̄A, 0 and jA, 0 are the specified mole and mass fluxes of species A at the
boundary, respectively. The special case of zero mass flux ( j̄A, 0 jA, 0 0)
corresponds to an impermeable surface for which dyA(0)/dx dwA (0)/
x dx 0 (Fig. 14–14).
To apply the specified concentration boundary condition, we must know the
CA(x) concentration of a species at the boundary. This information is usually ob-
Impermeable dCA(0) tained from the requirement that thermodynamic equilibrium must exist at the
surface ——– —=0
dx interface of two phases of a species. In the case of air–water interface, the con-
·
mA(0) = 0 centration values of water vapor in the air are easily determined from satura-
tion data, as shown in Example 14–2.
FIGURE 14–14
An impermeable surface in EXAMPLE 14–2 Mole Fraction of Water Vapor
mass transfer is analogous to an at the Surface of a Lake
insulated surface in heat transfer.
Determine the mole fraction of the water vapor at the surface of a lake whose
temperature is 15°C and compare it to the mole fraction of water in the lake
Air (Fig. 14–15). Take the atmospheric pressure at lake level to be 92 kPa.
92 kPa, 15°C
SOLUTION The mole fraction of the water vapor at the surface of a lake and
Saturated air
the mole fraction of water in the lake are to be determined and compared.
yH O, air side = 0.0185
2 Assumptions 1 Both the air and water vapor are ideal gases. 2 The mole frac-
tion of dissolved air in water is negligible.
yH ≅ 1.0 Properties The saturation pressure of water at 15°C is 1.705 kPa (Table A–9).
2O, liquid side
Lake Analysis The air at the water surface will be saturated. Therefore, the partial
15°C pressure of water vapor in the air at the lake surface will simply be the satura-
tion pressure of water at 15°C,
FIGURE 14–15
Pvapor Psat @ 15°C 1.705 kPa
Schematic for Example 14–2.
Assuming both the air and vapor to be ideal gases, the mole fraction of water
vapor in the air at the surface of the lake is determined from Eq. 14–11 to be
729
CHAPTER 14
Water contains some dissolved air, but the amount is negligible. Therefore, we
can assume the entire lake to be liquid water. Then its mole fraction becomes
Discussion Note that the concentration of water on a molar basis is 100 per-
cent just beneath the air–water interface and 1.85 percent just above it, even
though the air is assumed to be saturated (so this is the highest value at 15°C).
Therefore, huge discontinuities can occur in the concentrations of a species
across phase boundaries.
730
HEAT TRANSFER
TABLE 14–6
Henry’s constant H (in bars) for selected gases in water at low to moderate
pressures (for gas i, H Pi, gas side /yi, water side)
(from Mills, Ref. 13, Table A.21, p. 874)
Solute 290 K 300 K 310 K 320 K 330 K 340 K
H2S 440 560 700 830 980 1140
CO2 1280 1710 2170 2720 3220 —
O2 38,000 45,000 52,000 57,000 61,000 65,000
H2 67,000 72,000 75,000 76,000 77,000 76,000
CO 51,000 60,000 67,000 74,000 80,000 84,000
Air 62,000 74,000 84,000 92,000 99,000 104,000
N2 76,000 89,000 101,000 110,000 118,000 124,000
Gas A
yA, gas side
2. Henry’s constant increases (and thus the fraction of a dissolved gas
yA, liquid side in the liquid decreases) with increasing temperature. Therefore, the
Gas: A dissolved gases in a liquid can be driven off by heating the liquid
Liquid: B
(Fig. 14–16).
3. The concentration of a gas dissolved in a liquid is proportional to the
partial pressure of the gas. Therefore, the amount of gas dissolved in a
yA, gas side yA, liquid side liquid can be increased by increasing the pressure of the gas. This can
or be used to advantage in the carbonation of soft drinks with CO2 gas.
PA, gas side
———— yA, liquid side Strictly speaking, the result obtained from Eq. 14–18 for the mole fraction of
P dissolved gas is valid for the liquid layer just beneath the interface and not
or necessarily the entire liquid. The latter will be the case only when thermo-
PA, gas side = HyA, liquid side dynamic phase equilibrium is established throughout the entire liquid body.
FIGURE 14–16
Dissolved gases in a liquid can EXAMPLE 14–3 Mole Fraction of Dissolved Air in Water
be driven off by heating the liquid.
Determine the mole fraction of air dissolved in water at the surface of a lake
Air whose temperature is 17°C (Fig. 14–17). Take the atmospheric pressure at lake
level to be 92 kPa.
Saturated air Pdry air, gas side
SOLUTION The mole fraction of air dissolved in water at the surface of a lake
is to be determined.
ydry air, liquid side
Lake Assumptions 1 Both the air and water vapor are ideal gases. 2 Air is weakly sol-
17°C uble in water so that Henry’s law is applicable.
Properties The saturation pressure of water at 17°C is 1.92 kPa (Table A–9).
FIGURE 14–17 Henry’s constant for air dissolved in water at 290 K is H 62,000 bar (Table
14–6).
Schematic for Example 14–3.
Analysis This example is similar to the previous example. Again the air at the
water surface will be saturated, and thus the partial pressure of water vapor in
the air at the lake surface will be the saturation pressure of water at 17°C,
Assuming both the air and vapor to be ideal gases, the partial pressure of dry air
is determined to be
731
CHAPTER 14
Note that with little loss in accuracy (an error of about 2 percent), we could
have ignored the vapor pressure since the amount of vapor in air is so small.
Then the mole fraction of air in the water becomes
We mentioned earlier that the use of Henry’s law is limited to dilute gas–
liquid solutions; that is, a liquid with a small amount of gas dissolved in it.
Then the question that arises naturally is, what do we do when the gas is
highly soluble in the liquid (or solid), such as ammonia in water? In this case
the linear relationship of Henry’s law does not apply, and the mole fraction of
a gas dissolved in the liquid (or solid) is usually expressed as a function of the
partial pressure of the gas in the gas phase and the temperature. An approxi-
mate relation in this case for the mole fractions of a species on the liquid and
gas sides of the interface is given by Raoult’s law as
Pi, gas side yi, gas side P yi, liquid side Pi, sat(T) (14-19)
where Pi, sat(T) is the saturation pressure of the species i at the interface tem-
perature and P is the total pressure on the gas phase side. Tabular data are
available in chemical handbooks for common solutions such as the ammonia–
water solution that is widely used in absorption-refrigeration systems.
Gases may also dissolve in solids, but the diffusion process in this case can
be very complicated. The dissolution of a gas may be independent of the
structure of the solid, or it may depend strongly on its porosity. Some dissolu-
tion processes (such as the dissolution of hydrogen in titanium, similar to the
dissolution of CO2 in water) are reversible, and thus maintaining the gas con-
tent in the solid requires constant contact of the solid with a reservoir of that
gas. Some other dissolution processes are irreversible. For example, oxygen
gas dissolving in titanium forms TiO2 on the surface, and the process does not
reverse itself.
TABLE 14–7
The concentration of the gas species i in the solid at the interface Ci, solid side
is proportional to the partial pressure of the species i in the gas Pi, gas side on the Solubility of selected
gas side of the interface and is expressed as gases and solids
(for gas i, Ci, solid side /Pi, gas side)
Ci, solid side Pi, gas side (kmol/m3) (14-20) (from Barrer, Ref. 2)
Gas Solid T, K kmol/m3 · bar
where is the solubility. Expressing the pressure in bars and noting that the
O2 Rubber 298 0.00312
unit of molar concentration is kmol of species i per m3, the unit of solubility is
N2 Rubber 298 0.00156
kmol/m3 · bar. Solubility data for selected gas–solid combinations are given in CO2 Rubber 298 0.04015
Table 14–7. The product of the solubility of a gas and the diffusion coefficient He SiO2 293 0.00045
of the gas in a solid is referred to as the permeability , which is a measure H2 Ni 358 0.00901
of the ability of the gas to penetrate a solid. That is, DAB where DAB is
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 732
732
HEAT TRANSFER
That is, there will be 0.027 kmol (or 0.054 kg) of H2 gas in each m3 volume of
nickel adjacent to the interface.
TABLE 14–8
Analogy between heat conduction
and mass diffusion in a
14–5 ■
STEADY MASS DIFFUSION
stationary medium THROUGH A WALL
Mass Diffusion Many practical mass transfer problems involve the diffusion of a species
Heat Mass Molar through a plane-parallel medium that does not involve any homogeneous
Conduction Basis Basis chemical reactions under one-dimensional steady conditions. Such mass
T wi yi
transfer problems are analogous to the steady one-dimensional heat conduc-
k DAB CDAB tion problems in a plane wall with no heat generation and can be analyzed
q· ji j̄ i
similarly. In fact, many of the relations developed in Chapter 3 can be used
DAB DAB for mass transfer by replacing temperature by mass (or molar) fraction, ther-
L L L mal conductivity by DAB (or CDAB), and heat flux by mass (or molar) flux
(Table 14–8).
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 733
733
CHAPTER 14
wA, 1
DAB dwA (14-21)
where the mass transfer rate m· diff, A and the wall area A are taken out of the
integral sign since both are constants. If the density and the mass diffusion
coefficient DAB vary little along the wall, they can be assumed to be constant.
The integration can be performed in that case to yield · T1 – T2
Q = ———
R
wA, 1 wA, 2 A, 1 A, 2 T1 T2
m· diff, A, wall DAB A DAB A (kg/s) (14-22)
L L R
(a) Heat flow
This relation can be rearranged as
wA, 1 wA, 2 wA, 1 wA, 2 1 – 2
m· diff, A, wall (14-23) I = ———
L/DAB A Rdiff, wall Re
1 2
where Re
L (b) Current flow
Rdiff, wall
DAB A
wA, 1 – wA, 2
is the diffusion resistance of the wall, in s/kg, which is analogous to the elec- m· diff, A = ———— —
Rmass
trical or conduction resistance of a plane wall of thickness L and area A (Fig. wA, 1 wA, 2
14–20). Thus, we conclude that the rate of mass diffusion through a plane Rmass
wall is proportional to the average density, the wall area, and the concentra- (c) Mass flow
tion difference across the wall, but is inversely proportional to the wall thick- FIGURE 14–20
ness. Also, once the rate of mass diffusion is determined, the mass fraction Analogy between
wA(x) at any location x can be determined by replacing wA, 2 in Eq. 14–22 by thermal, electrical, and mass
wA(x) and L by x. diffusion resistance concepts.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 734
734
HEAT TRANSFER
The preceding analysis can be repeated on a molar basis with this result,
· yA, 1 yA, 2 CA, 1 CA, 2 yA, 1 yA, 2
Ndiff, A, wall CDAB A DAB A (14-24)
L L Rdiff, wall
m· diff, A
where Rdiff, wall L/CDAB A is the molar diffusion resistance of the wall in
B
s/kmol. Note that mole fractions are accompanied by molar concentrations
and mass fractions are accompanied by density. Either relation can be used to
r2 determine the diffusion rate of species A across the wall, depending on
r1 whether the mass or molar fractions of species A are known at the boundaries.
Also, the concentration gradients on both sides of an interface are different,
and thus diffusion resistance networks cannot be constructed in an analogous
wA, 1
manner to thermal resistance networks.
In developing these relations, we assumed the density and the diffusion co-
efficient of the wall to be nearly constant. This assumption is reasonable when
wA, 2
a small amount of species A diffuses through the wall and thus the concentra-
tion of A is small. The species A can be a gas, a liquid, or a solid. Also, the
FIGURE 14–21
wall can be a plane layer of a liquid or gas provided that it is stationary.
One-dimensional mass diffusion
The analogy between heat and mass transfer also applies to cylindrical and
through a cylindrical or
spherical shell.
spherical geometries. Repeating the approach outlined in Chapter 3 for heat
conduction, we obtain the following analogous relations for steady one-
dimensional mass transfer through nonreacting cylindrical and spherical
layers (Fig. 14–21)
wA, 1 wA, 2 A, 1 A, 2
m· diff, A, cyl 2LDAB 2LDAB (14-25)
ln(r2/r1) ln(r2/r1)
wA, 1 wA, 2 A, 1 A, 2
m· diff, A, sph 4r1r2DAB r2 r1 4r1r2DAB r2 r1 (14-26)
735
CHAPTER 14
where PA, 1 and PA, 2 are the partial pressures of gas A on the two sides of the
wall. Similar relations can be obtained for cylindrical and spherical walls by
following the same procedure. Also, if the permeability is given on a mass
basis (in kg/m · s · bar), then Eq. 14–29 will give the diffusion mass flow rate.
Noting that 1 kmol of an ideal gas at the standard conditions of 0°C and
1 atm occupies a volume of 22.414 m3, the volume flow rate of the gas
through the wall by diffusion can be determined from
· ·
Vdiff, A 22.414Ndiff, A (standard m3/s, at 0°C and 1 atm)
The volume flow rate at other conditions can be determined from the ideal gas
· ·
relation PAV NA RuT.
· CA, 1 CA, 2
Ndiff 4r1r2DAB r r
2 1
(0.087 0) kmol/m3
4(2.34 m)(2.40 m)(1.2 1012 m2/s)
2.40 2.34
1.228 1010 kmol/s
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 736
736
HEAT TRANSFER
The mass flow rate is determined by multiplying the molar flow rate by the
molar mass of hydrogen, which is M 2 kg/kmol,
·
m· diff MNdiff (2 kg/kmol)(1.228 1010 kmol/s) 2.46 1010 kg/s
Therefore, hydrogen will leak out through the shell of the container by diffusion
at a rate of 2.46 1010 kg/s or 7.8 g/year. Note that the concentration of
hydrogen in the nickel at the inner surface depends on the temperature and
pressure of the hydrogen in the tank and can be determined as explained in
Example 14–4. Also, the assumption of zero hydrogen concentration in nickel
at the outer surface is reasonable since there is only a trace amount of hydro-
gen in the atmosphere (0.5 part per million by mole numbers).
14–6 ■
WATER VAPOR MIGRATION IN BUILDINGS
Moisture greatly influences the performance and durability of building mate-
rials, and thus moisture transmission is an important consideration in the con-
struction and maintenance of buildings.
The dimensions of wood and other hygroscopic substances change with
moisture content. For example, a variation of 4.5 percent in moisture content
causes the volume of white oak wood to change by 2.5 percent. Such cyclic
changes of dimensions weaken the joints and can jeopardize the structural in-
tegrity of building components, causing “squeaking” at the minimum. Excess
moisture can also cause changes in the appearance and physical properties of
materials: corrosion and rusting in metals, rotting in woods, and peeling of
paint on the interior and exterior wall surfaces. Soaked wood with a water
content of 24 to 31 percent is observed to decay rapidly at temperatures 10 to
38°C. Also, molds grow on wood surfaces at relative humidities above 85 per-
cent. The expansion of water during freezing may damage the cell structure of
Dry Wet porous materials.
insulation insulation Moisture content also affects the effective conductivity of porous mediums
such as soils, building materials, and insulations, and thus heat transfer
through them. Several studies have indicated that heat transfer increases al-
most linearly with moisture content, at a rate of 3 to 5 percent for each percent
· · increase in moisture content by volume. Insulation with 5 percent moisture
Q 1.25 Q
content by volume, for example, increases heat transfer by 15 to 25 percent
relative to dry insulation (ASHRAE Handbook of Fundamentals, Ref. 1,
Chap. 20) (Fig. 14–24). Moisture migration may also serve as a transfer mech-
anism for latent heat by alternate evaporation and condensation. During a hot
and humid day, for example, water vapor may migrate through a wall and con-
0% 5% dense on the inner side, releasing the heat of vaporization, with the process re-
moisture moisture versing during a cool night. Moisture content also affects the specific heat and
FIGURE 14–24 thus the heat storage characteristics of building materials.
A 5 percent moisture content can Moisture migration in the walls, floors, or ceilings of buildings and in other
increase heat transfer through wall applications is controlled by either vapor barriers or vapor retarders. Vapor
insulation by 25 percent. barriers are materials that are impermeable to moisture, such as sheet metals,
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 737
737
CHAPTER 14
heavy metal foils, and thick plastic layers, and they effectively bar the vapor
from migrating. Vapor retarders, on the other hand, retard or slow down the
flow of moisture through the structures but do not totally eliminate it. Vapor
retarders are available as solid, flexible, or coating materials, but they usually
consist of a thin sheet or coating. Common forms of vapor retarders are rein-
forced plastics or metals, thin foils, plastic films, treated papers, coated felts,
and polymeric or asphaltic paint coatings. In applications such as the building
of walls where vapor penetration is unavoidable because of numerous open-
ings such as electrical boxes, telephone lines, and plumbing passages, vapor
retarders are used instead of vapor barriers to allow the vapor that somehow
leaks in to exit to the outside instead of trapping it in. Vapor retarders with
a permeance of 57.4 109 kg/s · m2 are commonly used in residential
buildings.
The insulation on chilled water lines and other impermeable surfaces that
are always cold must be wrapped with a vapor barrier jacket, or such cold
surfaces must be insulated with a material that is impermeable to moisture.
This is because moisture that migrates through the insulation to the cold
surface will condense and remain there indefinitely with no possibility of
vaporizing and moving back to the outside. The accumulation of moisture in
such cases may render the insulation useless, resulting in excessive energy
consumption.
TABLE 14–9
Atmospheric air can be viewed as a mixture of dry air and water vapor, and
the atmospheric pressure is the sum of the pressure of dry air and the pressure Saturation pressure of water at
of water vapor, which is called the vapor pressure P. Air can hold a certain various temperatures
amount of moisture only, and the ratio of the actual amount of moisture in the Saturation
air at a given temperature to the maximum amount air can hold at that tem- Temperature, °C Pressure, Pa
perature is called the relative humidity . The relative humidity ranges from 40 13
0 for dry air to 100 percent for saturated air (air that cannot hold any more 36 20
moisture). The partial pressure of water vapor in saturated air is called the 32 31
saturation pressure Psat. Table 14–9 lists the saturation pressure at various 28 47
temperatures. 24 70
The amount of moisture in the air is completely specified by the tempera- 20 104
ture and the relative humidity, and the vapor pressure is related to relative hu- 16 151
12 218
midity by
8 310
P Psat (14-30)
4 438
0 611
5 872
where Psat is the saturation (or boiling) pressure of water at the specified tem- 10 1,228
perature. Then the mass flow rate of moisture through a plain layer of thick- 15 1,705
ness L and normal area A can be expressed as 20 2,339
25 3,169
P, 1 P, 2 1 Psat, 1 2 Psat, 2 30 4,246
m· A A (kg/s) (14-31) 35 5,628
L L
40 7,384
50 12,349
where is the vapor permeability of the material, which is usually expressed 100 101,330
on a mass basis in the unit ng/s · m · Pa, where ng 1012 kg and 1 Pa 105 200 1.55 106
bar. Note that vapor migrates or diffuses from a region of higher vapor pres- 300 8.58 106
sure toward a region of lower vapor pressure.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 738
738
HEAT TRANSFER
739
CHAPTER 14
R-Value, R-Value,
Construction m2 · °C/W s · m2 · Pa/ng
1. Outside surface, 24 km/h wind 0.030 —
2. Painted wood bevel lapped siding 0.14 0.019
3. Wood fiberboard sheeting, 13 mm 0.23 0.0138
4. Glass fiber insulation, 90 mm 2.45 0.0004
5. Painted gypsum wallboard, 13 mm 0.079 0.012
6. Inside surface, still air 0.12 —
TOTAL 3.05 0.0452
The indoor conditions are 20°C and 60 percent relative humidity while the
outside conditions are 16°C and 70 percent relative humidity. Determine if
condensation or freezing of moisture will occur in the insulation.
SOLUTION The thermal and vapor resistances of different layers of a wall are
given. The possibility of condensation or freezing of moisture in the wall is to be
investigated.
Assumptions 1 Steady operating conditions exist. 2 Heat transfer through the
wall is one-dimensional. 3 Thermal and vapor resistances of different layers of
the wall and the heat transfer coefficients are constant.
Properties The thermal and vapor resistances are as given in the problem
statement. The saturation pressures of water at 20°C and 16°C are 2339 Pa
and 151 Pa, respectively (Table 14–9).
Analysis The schematic of the wall as well as the different elements used in its
construction are shown in Figure 14–25. Condensation is most likely to occur
at the coldest part of insulation, which is the part adjacent to the exterior
sheathing. Noting that the total thermal resistance of the wall is 3.05
m2 · °C/W, the rate of heat transfer through a unit area A 1 m2 of the wall is
· Ti To [20 (16)°C]
Q wall A (1 m2) 11.8 W
Rtotal 3.05 m2 · °C/W
The thermal resistance of the exterior part of the wall beyond the insulation is
0.03 0.14 0.23 0.40 m2 · °C/W. Then the temperature of the insula-
tion–outer sheathing interface is 6
5
· 4
TI To Q wall Rext 16°C (11.8 W)(0.40°C/W) 11.3°C 3
2
1
The saturation pressure of water at 11.3°C is 234 Pa, as shown in Table
FIGURE 14–25
14–9, and if there is condensation or freezing, the vapor pressure at the insu-
Schematic for Example 14–6.
lation–outer sheathing interface will have to be this value. The vapor pressure at
the indoors and the outdoors is
Then the rate of moisture flow through the interior and exterior parts of the wall
becomes
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 740
740
HEAT TRANSFER
P, 1 P, l
m· , interior A
P
R
A
interior R, interior
(1404 234) Pa
(1 m2) 94,355 ng/s 94.4 g/s
(0.012 0.0004) Pa · m2 · s/ng
P, 1 P, l
m· , exterior A
P
R exterior
A
R, exterior
(234 106) Pa
(1 m2) 3902 ng/s 3.9 g/s
(0.019 0.0138) Pa · m2 · s/ng
That is, moisture is flowing toward the interface at a rate of 94.4 g/s but flow-
ing from the interface to the outdoors at a rate of only 3.9 g/s. Noting that the
interface pressure cannot exceed 234 Pa, these results indicate that moisture
is freezing in the insulation at a rate of
This corresponds to 7.82 g during a 24-h period, which can be absorbed by the
insulation or sheathing, and then flows out when the conditions improve. How-
ever, excessive condensation (or frosting at temperatures below 0°C) of moisture
in the walls during long cold spells can cause serious problems. This problem
can be avoided or minimized by installing vapor barriers on the interior side of
the wall, which will limit the moisture flow rate to 3.9 g/s. Note that if there
were no condensation or freezing, the flow rate of moisture through a 1 m2 sec-
tion of the wall would be 28.7 g/s (can you verify this?).
14–7 ■
TRANSIENT MASS DIFFUSION
The steady analysis discussed earlier is useful when determining the leakage
rate of a species through a stationary layer. But sometimes we are interested in
the diffusion of a species into a body during a limited time before steady op-
erating conditions are established. Such problems are studied using transient
analysis. For example, the surface of a mild steel component is commonly
hardened by packing the component in a carbonaceous material in a furnace
Carbonaceous at high temperature. During the short time period in the furnace, the carbon
material molecules diffuse through the surface of the steel component, but they pene-
Hardened trate to a depth of only a few millimeters. The carbon concentration decreases
surface
exponentially from the surface to the inner parts, and the result is a steel com-
Steel shaft
Carbon ponent with a very hard surface and a relatively soft core region (Fig. 14–26).
Soft core
The same process is used in the gem industry to color clear stones. For ex-
ample, a clear sapphire is given a brilliant blue color by packing it in titanium
and iron oxide powders and baking it in an oven at about 2000°C for about a
month. The titanium and iron molecules penetrate less than 0.5 mm in the sap-
phire during this process. Diffusion in solids is usually done at high tempera-
tures to take advantage of the high diffusion coefficients at high temperatures
FIGURE 14–26 and thus to keep the diffusion time at a reasonable level. Such diffusion or
The surface hardening of a mild steel “doping” is also commonly practiced in the production of n- or p-type semi-
component by the diffusion of carbon conductor materials used in the manufacture of electronic components. Dry-
molecules is a transient mass ing processes such as the drying of coal, timber, food, and textiles constitute
diffusion process. another major application area of transient mass diffusion.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 741
741
CHAPTER 14
since the total density or total molar concentration of dilute solutions is usu- CA, s
ally constant ( constant or C constant). Therefore, other measures of CA(x, t)
concentration can be used in Eq. 14–36.
A quantity of interest in mass diffusion processes is the depth of diffusion at
CA, i 0
a given time. This is usually characterized by the penetration depth defined x
δdiff Tangent line to
as the location x where the tangent to the concentration profile at the surface
concentration
(x 0) intercepts the CA CA, i line, as shown in Figure 14–27. Obtaining the gradient at x = 0
concentration gradient at x 0 by differentiating Eq. 14–36, the penetration Semi-infinite
depth is determined to be medium
FIGURE 14–27
Therefore, the penetration depth is proportional to the square root of both the The concentration profile of species A
diffusion coefficient and time. The diffusion coefficient of zinc in copper at in a semi-infinite medium during
1000°C, for example, is 5.0 1012 m2/s. Then the penetration depth of zinc transient mass diffusion and the
in copper in 10 h is penetration depth.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 742
742
HEAT TRANSFER
0.01 0.0015
0.012 0.0015
0.81 erfc x
2 DABt
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 743
743
CHAPTER 14
x2 (0.0005 m)2
t 4505 s 1 h 15 min
4DAB(0.17)2
4 (4.8 1010 m2/s)(0.17)2
Discussion The steel component in this case must be held in the furnace for
1 h and 15 min to achieve the desired level of hardening. The diffusion coeffi-
cient of carbon in steel increases exponentially with temperature, and thus this
process is commonly done at high temperatures to keep the diffusion time at a
reasonable level.
14–8 ■
DIFFUSION IN A MOVING MEDIUM
To this point we have limited our consideration to mass diffusion in a station-
ary medium, and thus the only motion involved was the creeping motion of
molecules in the direction of decreasing concentration, and there was no mo-
tion of the mixture as a whole. Many practical problems, such as the evapora-
tion of water from a lake under the influence of the wind or the mixing of two
fluids as they flow in a pipe, involve diffusion in a moving medium where the
bulk motion is caused by an external force. Mass diffusion in such cases is
complicated by the fact that chemical species are transported both by diffusion
and by the bulk motion of the medium (i.e., convection). The velocities and
mass flow rates of species in a moving medium consist of two components:
one due to molecular diffusion and one due to convection (Fig. 14–29). Air
Diffusion in a moving medium, in general, is difficult to analyze since var-
ious species can move at different velocities in different directions. Turbu- Convection
lence will complicate the things even more. To gain a firm understanding of
the physical mechanism while keeping the mathematical complexities to a Diffusion
minimum, we limit our consideration to systems that involve only two com-
ponents (species A and B) in one-dimensional flow (velocity and other prop- Lake
erties change in one direction only, say the x-direction). We also assume the
total density (or molar concentration) of the medium remains constant. That FIGURE 14–29
is, A B constant (or C CA CB constant) but the densities of In a moving medium, mass transfer is
species A and B may vary in the x-direction. due to both diffusion and convection.
Several possibilities are summarized in Figure 14–30. In the trivial case
(case a) of a stationary homogeneous mixture, there will be no mass transfer
by molecular diffusion or convection since there is no concentration gradient
or bulk motion. The next case (case b) corresponds to the flow of a well-mixed
fluid mixture through a pipe. Note that there is no concentration gradients and
thus molecular diffusion in this case, and all species move at the bulk flow ve-
locity of . The mixture in the third case (case c) is stationary ( 0) and
thus it corresponds to ordinary molecular diffusion in stationary mediums,
which we discussed before. Note that the velocity of a species at a location in
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 744
744
HEAT TRANSFER
FIGURE 14–30
Various quantities associated with a mixture of two species A and B at a location x under one-dimensional flow or no-flow
conditions. (The density of the mixture A B is assumed to remain constant.)
this case is simply the diffusion velocity, which is the average velocity of a
group of molecules at that location moving under the influence of concentra-
tion gradient. Finally, the last case (case d) involves both molecular diffusion
and convection, and the velocity of a species in this case is equal to the sum of
the bulk flow velocity and the diffusion velocity. Note that the flow and the
diffusion velocities can be in the same or opposite directions, depending on
the direction of the concentration gradient. The diffusion velocity of a species
is negative when the bulk flow is in the positive x-direction and the concen-
tration gradient is positive (i.e., the concentration of the species increases in
the x-direction).
Noting that the mass flow rate at any flow section is expressed as m· A
where is the density, is the velocity, and A is the cross sectional area, the
conservation of mass relation for the flow of a mixture that involves two
species A and B can be expressed as
m· m· A m· B
or
A AAA BB A
745
CHAPTER 14
AA BB A B
A B wAA wBB (14-39)
where is called the mass-average velocity of the flow, which is the veloc-
ity that would be measured by a velocity sensor such as a pitot tube, a turbine
device, or a hot wire anemometer inserted into the flow.
The special case 0 corresponds to a stationary medium, which can
now be defined more precisely as a medium whose mass-average velocity is
zero. Therefore, mass transport in a stationary medium is by diffusion only,
and zero mass-average velocity indicates that there is no bulk fluid motion.
When there is no concentration gradient (and thus no molecular mass
diffusion) in the fluid, the velocity of all species will be equal to the mass-
average velocity of the flow. That is, A B. But when there is a con-
centration gradient, there will also be a simultaneous flow of species in the
direction of decreasing concentration at a diffusion velocity of diff. Then the
average velocity of the species A and B can be determined by superimposing
the average flow velocity and the diffusion velocity as (Fig. 14–31) diff, A = 0
A =
A diff, A
B diff, B (14-40)
A
Similarly, we apply the superposition principle to the species mass flow rates
to get Flow
velocity
m· A AA A A( diff, A)A AA Adiff, A A m· conv, A m· diff, A
m· B BB A B( diff, B)A BA Bdiff, B A m· conv, B m· diff, B (14-41) (a) No concentration gradient
Using Fick’s law of diffusion, the total mass fluxes j m· /A can be ex-
pressed as diff, A ≠ 0
A = + diff, A
A dwA dwA
jA A Adiff, A DAB wA(jA jB) DAB
dx dx
B dwB dwB A
jB B Bdiff, B DBA wB(jA jB) DBA (14-42) diff, A
dx dx
Note that the diffusion velocity of a species is negative when the molecular
diffusion occurs in the negative x-direction (opposite to flow direction). The
mass diffusion rates of the species A and B at a specified location x can be ex- (b) Mass concentration gradient and thus
mass diffusion
pressed as
FIGURE 14–31
m· diff, A Adiff, A A A(A )A The velocity of a species at a point is
m·diff, B A ( )A
B diff, B B B (14-43) equal to the sum of the bulk
flow velocity and the diffusion
By substituting the relation from Eq. 14–39 into Eq. 14–43, it can be shown velocity of that species at that point.
that at any cross section
dwA dwB
m· diff, A m· diff, B 0 → m· diff, A m· diff, B → ADAB ADBA (14-44)
dx dx
which indicates that the rates of diffusion of species A and B must be equal
in magnitude but opposite in sign. This is a consequence of the assumption
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 746
746
HEAT TRANSFER
w
That is, in the case of constant total concentration, the diffusion coefficient of
species A into B is equal to the diffusion coefficient of species B into A.
0 We now repeat the analysis presented above with molar concentration C and
x ·
the molar flow rate N. The conservation of matter in this case is expressed as
· · ·
m· diff, A
N NA NB
m· diff, B or
A AA A BB A (14-47)
A B
Canceling A and solving for gives
wA = –wB
dwA dwB
—— = – —— CAA CBB CA CB
dx dx yAA yBB (14-48)
C C A C B
m·
diff, A= –m· diff, B
DAB = DBA where is called the molar-average velocity of the flow. Note that
FIGURE 14–32 unless the mass and molar fractions are the same. The molar flow rates of
In a binary mixture of species A and B species are determined similarly to be
with A B constant, the rates · · ·
NA CAA A CA( diff, A)A CAA CAdiff, A A Nconv, A Ndiff, A
of mass diffusion of species A and B · · ·
are equal magnitude and opposite NB CBB A CB( diff, B)A CBA CBdiff, B A Nconv, B Ndiff, B (14-49)
in direction. ·
Using Fick’s law of diffusion, the total molar fluxes j̄ N /A and diffusion
·
molar flow rates Ndiff can be expressed as
CA dyA dyA
j̄A CA CAdiff, A C CDAB yA( j̄A j̄B) CDAB
C dx dx
CB dyB dyB
j̄B CB CBdiff, B C CDBA yB( j̄A j̄B) CDBA (14-50)
C dx dx
and
·
Ndiff, A CAdiff, A A CA(A )A
·
Ndiff, B CBdiff, B A CB(B )A (14-51)
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 747
747
CHAPTER 14
By substituting the relation from Eq. 14–48 into these two equations, it can
be shown that
· · · ·
Ndiff, A Ndiff, B 0 → Ndiff, A Ndiff, B (14-52)
which again indicates that the rates of diffusion of species A and B must be
equal in magnitude but opposite in sign.
It is important to note that when working with molar units, a medium is said
to be stationary when the molar-average velocity is zero. The average veloc-
ity of the molecules will be zero in this case, but the apparent velocity of the
mixture as measured by a velocimeter placed in the flow will not necessarily
be zero because of the different masses of different molecules. In a mass-
based stationary medium, for each unit mass of species A moving in one
direction, a unit mass of species B moves in the opposite direction. In a mole-
based stationary medium, however, for each mole of species A moving in one
direction, one mole of species B moves in the opposite direction. But this may
result in a net mass flow rate in one direction that can be measured by a
velocimeter since the masses of different molecules are different.
You may be wondering whether to use the mass analysis or molar analysis
in a problem. The two approaches are equivalent, and either approach can be
used in mass transfer analysis. But sometimes it may be easier to use one of
the approaches, depending on what is given. When mass-average velocity is
known or can easily be obtained, obviously it is more convenient to use the
mass-based formulation. When the total pressure and temperature of a mix-
ture are constant, however, it is more convenient to use the molar formulation,
as explained next.
where Ru 8.314 kJ/kmol · K is the universal gas constant. Therefore, the as-
sumption of constant mixture density ( constant) in such cases will not be
accurate unless the gas or gases with variable concentrations constitute a very
small fraction of the mixture. However, the molar density C of a mixture re-
mains constant when the mixture pressure P and temperature T are constant
since
Ru
P RT T CRuT (14-53)
M
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 748
748
HEAT TRANSFER
Diffusion of B
·
j̄A NA/A constant (or jA m· A/A constant)
Bulk flow
The pressure and temperature of the gas–vapor mixture are said to be con-
stant, and thus the molar density of the mixture must be constant throughout
the mixture, as shown earlier. That is, C CA CB constant, and it is more
0 convenient to work with mole fractions or molar concentrations in this case
Liquid A
instead of mass fractions or densities since constant.
FIGURE 14–34 Noting that yA yB 1 and that yA, 0 yA, L, we must have yB, 0 yB, L. That
Diffusion of a vapor A is, the mole fraction of the gas must be decreasing downward by the same
through a stagnant gas B. amount that the mole fraction of the vapor is increasing. Therefore, gas must
be diffusing from the top of the column toward the liquid interface. However,
the gas is said to be insoluble in the liquid, and thus there can be no net mass
flow of the gas downward. Then under steady conditions, there must be an
upward bulk fluid motion with an average velocity that is just large enough
to balance the diffusion of air downward so that the net molar (or mass) flow
rate of the gas at any point is zero. In other words, the upward bulk motion
offsets the downward diffusion, and for each air molecule that moves down-
ward, there is another air molecule that moves upward. As a result, the air ap-
pears to be stagnant (it does not move). That is,
·
j̄B NB /A 0 (or jB m· B /A 0)
The diffusion medium is no longer stationary because of the bulk motion. The
implication of the bulk motion of the gas is that it transports vapor as well as
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 749
749
CHAPTER 14
the gas upward with a velocity of , which results in additional mass flow of
vapor upward. Therefore, the molar flux of the vapor can be expressed as
· dyA
j̄A NA/A j̄A, conv j̄A, diff yA( j̄A j̄B) CDAB (14-54)
dx
since j̄A constant, C constant, and DAB constant. Separating the vari-
ables and integrating from x 0, where yA(0) yA, 0, to x L, where
yA(L) yA, L gives
yA, L
A, 0
dyA
1 yA
0
L j¯A
CDAB
dx (14-57)
Then the molar flux of vapor A, which is the evaporation rate of species A per
unit interface area, becomes
· CDAB 1 yA, L
j̄A NA/A ln (kmol/s · m2) (14-59)
L 1 yA, 0
This relation is known as Stefan’s law, and the induced convective flow de-
scribed that enhances mass diffusion is called the Stefan flow.
An expression for the variation of the mole fraction of A with x can be
determined by performing the integration in Eq. 14–57 to the upper limit of x
where yA(x) yA (instead of to L where yA(L) yA, L). It yields
1 yA j¯A
ln x (14-60)
1 yA, 0 CDAB
Substituting the j̄A expression from Eq. 14–59 into this relation and rearrang-
ing gives
1 yA 1 yA, L x/L x/L
yB yB, L
and yB, 0 yB, 0 (14-61)
1 yA, 0 1 yA, 0
The second relation for the variation of the mole fraction of the stationary
gas B is obtained from the first one by substituting 1 yA yB since
yA yB 1.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 750
750
HEAT TRANSFER
where As is the surface area of the liquid–vapor interface, hfg, A is the latent
heat of vaporization, and MA is the molar mass of species A.
Equimolar Counterdiffusion
Consider two large reservoirs connected by a channel of length L, as shown in
Gas
A
Gas Figure 14–35. The entire system contains a binary mixture of gases A and B at
mixture mixture a uniform temperature T and pressure P throughout. The concentrations of
A+B B A+B
· · species are maintained constant in each of the reservoirs such that yA, 0 yA, L
yA > yB NA NB yA < yB
and yB, 0 yB, L. The resulting concentration gradients will cause the species A
to diffuse in the positive x-direction and the species B in the opposite direc-
tion. Assuming the gases to behave as ideal gases and thus P CRuT, the total
0 x L molar concentration of the mixture C will remain constant throughout the
T, P T, P mixture since P and T are constant. That is,
C CA CB constant (kmol/m3)
FIGURE 14–35
Equimolar isothermal counterdiffusion
This requires that for each molecule of A that moves to the right, a molecule
of two gases A and B.
of B moves to the left, and thus the molar flow rates of species A and B must
be equal in magnitude and opposite in sign. That is,
· · · ·
NA NB or NA NB 0 (kmol/s)
Therefore, the mixture is stationary on a molar basis and thus mass transfer is
by diffusion only (there is no mass transfer by convection) so that
· dyA · dyB
j̄A NA/A CDAB and j̄B NB /A CDBA (14-63)
dx dx
Under steady conditions, the molar flow rates of species A and B can be
determined directly from Eq. 14–24 developed earlier for one-dimensional
steady diffusion in a stationary medium, noting that P CRuT and thus
C P/RuT for each constituent gas and the mixture. For one-dimensional
flow through a channel of uniform cross sectional area A with no homoge-
neous chemical reactions, they are expressed as
751
CHAPTER 14
These relations imply that the mole fraction, molar concentration, and the par-
tial pressure of either gas vary linearly during equimolar counterdiffusion.
It is interesting to note that the mixture is stationary on a molar basis, but
it is not stationary on a mass basis unless the molar masses of A and B are
equal. Although the net molar flow rate through the channel is zero, the net
mass flow rate of the mixture through the channel is not zero and can be
determined from
· · ·
m· m· A m· B NAMA NB MB NA(MA MB) (14-65)
· ·
since N B N A. Note that the direction of net mass flow rate is the flow
direction of the gas with the larger molar mass. A velocity measurement
device such as an anemometer placed in the channel will indicate a velocity
of m· /A where is the total density of the mixture at the site of
measurement.
Noting that the pressure of helium is 1 atm at the bottom of the tube (x 0)
and 0 at the top (x L), its molar flow rate is determined from Eq. 14–64 to be
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 752
752
HEAT TRANSFER
2
(8.314 kPa · m3/kmol · K)(298 K) 15 m 1 atm
3.85 10 kmol/s
12
Therefore,
·
m· helium (NM)helium (3.85 1012 kmol/s)(4 kg/kmol) 1.54 1011 kg/s
m· net m· helium m· air 1.54 1011 112 1012 9.66 1011 kg/s
The mass fraction of air at the bottom of the tube is very small, as shown above,
and thus the density of the mixture at x 0 can simply be taken to be the den-
sity of helium, which is
P 101.325 kPa
helium 0.1637 kg/m3
RT (2.0769 kPa · m3/kg · K)(298 K)
Then the average flow velocity at the bottom part of the tube becomes
753
CHAPTER 14
pressure is 83.5 kPa. The tube is partially filled with water, and the distance
from the water surface to the open end of the tube is 40 cm (Fig. 14–37). Dry Air, B
air is blown over the open end of the tube so that water vapor rising to the top yA, L
Diffusion of vapor
is removed immediately and the concentration of vapor at the top of the tube is L
Diffusion of air
zero. In 15 days of continuous operation at constant pressure and temperature,
Dry air is blown on top of the tube and, thus, yvapor, L yA, L 0. Also, the total
molar density throughout the tube remains constant because of the constant
temperature and pressure conditions and is determined to be
P 83.5 kPa
C 0.0343 kmol/m3
RuT (8.314 kPa · m3/kmol · K)(293 K)
The evaporation rate is given to be 1.23 g per 15 days. Then the molar flow rate
of vapor is determined to be
which gives
for the binary diffusion coefficient of water vapor in air at 20°C and 83.5 kPa.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 754
754
HEAT TRANSFER
14–9 ■
MASS CONVECTION
So far we have considered mass diffusion, which is the transfer of mass due to
a concentration gradient. Now we consider mass convection (or convective
mass transfer), which is the transfer of mass between a surface and a moving
fluid due to both mass diffusion and bulk fluid motion. We mentioned ear-
lier that fluid motion enhances heat transfer considerably by removing the
heated fluid near the surface and replacing it by the cooler fluid further away.
Likewise, fluid motion enhances mass transfer considerably by removing
the high-concentration fluid near the surface and replacing it by the lower-
concentration fluid further away. In the limiting case of no bulk fluid motion,
mass convection reduces to mass diffusion, just as convection reduces to con-
duction. The analogy between heat and mass convection holds for both forced
and natural convection, laminar and turbulent flow, and internal and ex-
ternal flow.
Like heat convection, mass convection is also complicated because of the
complications associated with fluid flow such as the surface geometry, flow
regime, flow velocity, and the variation of the fluid properties and composi-
tion. Therefore, we will have to rely on experimental relations to determine
mass transfer. Also, mass convection is usually analyzed on a mass basis
rather than on a molar basis. Therefore, we will present formulations in terms
of mass concentration (density or mass fraction w) instead of molar concen-
tration (molar density C or mole fraction y). But the formulations on a molar
basis can be obtained using the relation C /M where M is the molar mass.
ρA,
Concentration Also, for simplicity, we will restrict our attention to convection in fluids that
Concentration profile are (or can be treated as) binary mixtures.
boundary ρA, Consider the flow of air over the free surface of a water body such as a lake
y layer
under isothermal conditions. If the air is not saturated, the concentration of
water vapor will vary from a maximum at the water surface where the air is al-
0
x ρA, s
Species A
ways saturated to the free steam value far from the surface. In heat convection,
FIGURE 14–38
we defined the region in which temperature gradients exist as the thermal
The development of a concentration
boundary layer. Similarly, in mass convection, we define the region of the
boundary layer for species A during fluid in which concentration gradients exist as the concentration boundary
external flow on a flat surface. layer, as shown in Figure 14–38. In external flow, the thickness of the con-
centration boundary layer c for a species A at a specified location on the sur-
face is defined as the normal distance y from the surface at which
Concentration Fully developed
entry length region A, s A
A, s A, 0.99
where A, s and A, are the densities of species A at the surface (on the fluid
side) and the free stream, respectively.
Species A
In internal flow, we have a concentration entrance region where the
Concentration boundary layer concentration profile develops, in addition to the hydrodynamic and thermal
Thermal boundary layer entry regions (Fig. 14–39). The concentration boundary layer continues to
Velocity boundary layer develop in the flow direction until its thickness reaches the tube center and
FIGURE 14–39 the boundary layers merge. The distance from the tube inlet to the location
The development of the velocity, where this merging occurs is called the concentration entry length Lc, and
thermal, and concentration boundary the region beyond that point is called the fully developed region, which is
layers in internal flow. characterized by
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 755
755
CHAPTER 14
A, s A
x A, s A, b
0 (14-66)
A, b
1
Acave dA
Ac
A c (14-67)
Momentum diffusivity
Prandtl number: Pr (14-68) Mass transfer: Sc
DAB
Thermal diffusivity
The corresponding quantity in mass convection is the dimensionless Schmidt FIGURE 14–40
number, defined as In mass transfer, the Schmidt number
plays the role of the Prandtl
Momentum diffusivity number in heat transfer.
Schmidt number: Sc (14-69)
DAB Mass diffusivity
756
HEAT TRANSFER
Concentration Note that species transfer at the surface (y 0) is by diffusion only because
profile
wA, wA, of the no-slip boundary condition, and mass flux of species A at the surface
dwA
—– can be expressed by Fick’s law as (Fig. 14–42)
dy y=0
wA
y
Mass diffusion
jA m· A/A DAB
y y0
(kg/s · m2) (14-72)
0 wA, s
Species A This is analogous to heat transfer at the surface being by conduction only and
expressing it by Fourier’s law.
∂C The rate of heat convection for external flow was expressed conveniently by
–DAB —–A = hmass(wA, s – wA, ) Newton’s law of cooling as
∂y y=0
·
FIGURE 14–42 Q conv hconv A(Ts T)
Mass transfer at a surface occurs by
diffusion because of the no-slip where hconv is the average heat transfer coefficient, A is the surface area, and
boundary condition, just like heat Ts T is the temperature difference across the thermal boundary layer.
transfer occurring by conduction. Likewise, the rate of mass convection can be expressed as
m· conv hmass A(A, s A, ) hmassA(wA, s wA, ) (kg/s) (14-73)
where hmass is the average mass transfer coefficient, in m/s; A is the surface
area; A, s A, is the mass concentration difference of species A across the
concentration boundary layer; and is the average density of the fluid in the
boundary layer. The product hmass, whose unit is kg/m2 · s, is called the mass
transfer conductance. If the local mass transfer coefficient varies in the flow
direction, the average mass transfer coefficient can be determined from
hmass, ave
1
A h A
massdA (14-74)
FIGURE 14–43 where hmass is the mass transfer coefficient and DAB is the mass diffusivity. The
In mass transfer, the Sherwood Nusselt and Sherwood numbers represent the effectiveness of heat and mass
number plays the role the Nusselt convection at the surface, respectively.
number plays in heat transfer. Sometimes it is more convenient to express the heat and mass transfer co-
efficients in terms of the dimensionless Stanton number as
hconv 1
Heat transfer Stanton number: St Nu (14-77)
Cp Re Pr
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 757
757
CHAPTER 14
and
hmass 1
Mass transfer Stanton number: Stmass Sh (14-78)
Re Sc
where is the free steam velocity in external flow and the bulk mean fluid
velocity in internal flow.
For a given geometry, the average Nusselt number in forced convection
depends on the Reynolds and Prandtl numbers, whereas the average Sher-
wood number depends on the Reynolds and Schmidt numbers. That is,
Nusselt number: Nu f(Re, Pr) TABLE 14–12
Sherwood number: Sh f(Re, Sc)
Analogy between the quantities that appear
in the formulation and solution of heat
where the functional form of f is the same for both the Nusselt and Sherwood convection and mass convection
numbers in a given geometry, provided that the thermal and concentration
boundary conditions are of the same type. Therefore, the Sherwood number Heat Mass
can be obtained from the Nusselt number expression by simply replacing the Convection Convection
Prandtl number by the Schmidt number. This shows what a powerful tool
T C, y, , or w
analogy can be in the study of natural phenomena (Table 14–12).
In natural convection mass transfer, the analogy between the Nusselt and hconv hmass
Sherwood numbers still holds, and thus Sh f(Gr, Sc). But the Grashof num- thermal concentration
ber in this case should be determined directly from Lc Lc
Re Re
g( s) L3c g(/) L3c
Gr (14-79) g (Ts T) L3c g ( s) L3c
2 2 Gr , Gr
2
2
which is applicable to both temperature- and/or concentration-driven natural Pr Sc
D AB
convection flows. Note that in homogeneous fluids (i.e., fluids with no con-
centration gradients), density differences are due to temperature differences h conv h mass
St Stmass
only, and thus we can replace / by T for convenience, as we did in nat- C p
ural convection heat transfer. However, in nonhomogeneous fluids, density hconv Lc hmass Lc
Nu Sh
differences are due to the combined effects of temperature and concentration k DAB
differences, and / cannot be replaced by T in such cases even when all Nu f (Re, Pr) Sh f (Re, Sc)
we care about is heat transfer and we have no interest in mass transfer. For ex- Nu f (Gr, Pr) Sh f (Gr, Sc)
ample, hot water at the bottom of a pond rises to the top. But when salt is
placed at the bottom, as it is done in solar ponds, the salty water (brine) at the
bottom will not rise because it is now heavier than the fresh water at the top
(Fig. 14–44). 20°C Fresh water
Concentration-driven natural convection flows are based on the densities of
No convection SOLAR
different species in a mixture being different. Therefore, at isothermal condi- currents POND
tions, there will be no natural convection in a gas mixture that is composed of
gases with identical molar masses. Also, the case of a hot surface facing up 70°C ρbrine > ρwater Brine
corresponds to diffusing fluid having a lower density than the mixture (and Salt
thus rising under the influence of buoyancy), and the case of a hot surface fac-
ing down corresponds to the diffusing fluid having a higher density. For ex-
FIGURE 14–44
ample, the evaporation of water into air corresponds to a hot surface facing up
A hot fluid at the bottom will rise and
since water vapor is lighter than the air and it will tend to rise. But this will not
initiate natural convection currents
be the case for gasoline unless the temperature of the gasoline–air mixture at only if its density is lower.
the gasoline surface is so high that thermal expansion overwhelms the density
differential due to higher gasoline concentration near the surface.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 758
758
HEAT TRANSFER
y
Velocity,
temperature, or
Analogy between Friction, Heat Transfer,
CA,
T
concentration
profile
and Mass Transfer Coefficients
Consider the flow of a fluid over a flat plate of length L with free steam con-
ditions of T, , and wA, (Fig. 14–45). Noting that convection at the surface
Tangent line (y 0) is equal to diffusion because of the no-slip condition, the friction, heat
at y = 0 transfer, and mass transfer conditions at the surface can be expressed as
0
f
FIGURE 14–45 Wall friction: s 2 (14-80)
y y0 2
The friction, heat, and mass transfer
q· k
coefficients for flow over a surface are T
Heat transfer: hheat(Ts T) (14-81)
proportional to the slope of the tangent s
y y0
line of the velocity, temperature, and w
y
A
concentration profiles, respectively, Mass transfer: jA, s D AB hmass(wA, s wA, ) (14-82)
y0
at the surface.
These relations can be rewritten for internal flow by using bulk mean proper-
ties instead of free stream properties. After some simple mathematical manip-
ulations, the three relations above can be rearranged as
d(/ ) f L f
Wall friction: d(y/L) y0
2
Re
2
(14-83)
d[(T T )/(T T )]
s s hheat L
Heat transfer: Nu (14-84)
d(y/L) y0 k
w )/(w w )]
d[(wA A, s A, A, s hmass L
Mass transfer: Sh (14-85)
d(y/L) y0 DAB
The left sides of these three relations are the slopes of the normalized veloc-
ity, temperature, and concentration profiles at the surface, and the right sides
are the dimensionless numbers discussed earlier.
759
CHAPTER 14
2/3 2/3
D
hheat Sc
Cp Cp Cp Le2/3
2/3
or (14-89) 1 DAB
hmass Pr AB f
2
For air–water vapor mixtures at 298 K, the mass and thermal diffusivities are Special case: DAB
DAB 2.5 105 m2/s and 2.18 105 m2/s and thus the Lewis number
is Le /DAB 0.872. (We simply use the value of dry air instead of the hmass
hheat 1
f
moist air since the fraction of vapor in the air at atmospheric conditions is Cp 2
low.) Then ( /DAB)2/3 0.8722/3 0.913, which is close to unity. Also, the
Lewis number is relatively insensitive to variations in temperature. Therefore, FIGURE 14–47
for air–water vapor mixtures, the relation between heat and mass transfer co- When the friction or heat transfer
efficients can be expressed with a good accuracy as coefficient is known, the mass transfer
coefficient can be determined directly
hheat Cp hmass (air–water vapor mixtures) (14-90)
from the Chilton–Colburn analogy.
where and Cp are the density and specific heat of air at mean conditions (or
Cp is the specific heat of air per unit volume). Equation 14–90 is known as
the Lewis relation and is commonly used in air-conditioning applications.
Another important consequence of Le 1 is that the wet-bulb and adiabatic
saturation temperatures of moist air are nearly identical. In turbulent flow, the
Lewis relation can be used even when the Lewis number is not 1 since eddy
mixing in turbulent flow overwhelms any molecular diffusion, and heat and
mass are transported at the same rate.
The Chilton–Colburn analogy has been observed to hold quite well in lam-
inar or turbulent flow over plane surfaces. But this is not always the case for
internal flow and flow over irregular geometries, and in such cases specific re-
lations developed should be used. When dealing with flow over blunt bodies,
it is important to note that f in these relations is the skin friction coefficient, not
the total drag coefficient, which also includes the pressure drag.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 760
760
HEAT TRANSFER
761
CHAPTER 14
TABLE 14–13
Sherwood number relations in mass convection for specified concentration at the surface corresponding to the Nusselt
number relations in heat convection for specified surface temperature
Convective Heat Transfer Convective Mass Transfer
1. Forced Convection over a Flat Plate
(a) Laminar flow (Re 5 105)
Nu 0.664 Re0.5L Pr
1/3
, Pr 0.6 Sh 0.664 Re0.5 1/3
L Sc , Sc 0.5
(b) Turbulent flow (5 105 Re 107)
Nu 0.037 Re0.8 L Pr
1/3
, Pr 0.6 Sh 0.037 Re0.8 1/3
L Sc , Sc 0.5
2. Fully Developed Flow in Smooth Circular Pipes
(a) Laminar flow (Re 2300)
Nu 3.66 Sh 3.66
(b) Turbulent flow (Re 10,000)
Nu 0.023 Re0.8 Pr0.4, 0.7 Pr 160 Sh 0.023 Re0.8 Sc0.4, 0.7 Sc 160
3. Natural Convection over Surfaces
(a) Vertical plate
Nu 0.59(Gr Pr)1/4, 105 Gr Pr 109 Sh 0.59(Gr Sc)1/4, 105 Gr Sc 109
Nu 0.1(Gr Pr) , 1/3
109 Gr Pr 1013 Sh 0.1(Gr Sc)1/3, 109 Gr Sc 1013
(b) Upper surface of a horizontal plate
Surface is hot (Ts T) Fluid near the surface is light (s )
Nu 0.54(Gr Pr)1/4, 104 Gr Pr 107 Sh 0.54(Gr Sc)1/4, 104 Gr Sc 107
Nu 0.15(Gr Pr)1/3, 107 Gr Pr 1011 Sh 0.15(Gr Sc)1/3, 107 Gr Sc 1011
(c) Lower surface of a horizontal plate
Surface is hot (Ts T) Fluid near the surface is light (s )
Nu 0.27(Gr Pr)1/4, 105 Gr Pr 1011 Sh 0.27(Gr Sc)1/4, 105 Gr Sc 1011
T 2.072 3002.072
DAB DH2O-air 1.87 1010 1.87 1010 2.54 105 m2/s
P 1
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 762
762
HEAT TRANSFER
D (1.2 m/s)(0.015 m)
Re 1139
1.58 105 m2/s
which is less than 2300 and thus the flow is laminar. Therefore, based on the
analogy between heat and mass transfer, the Nusselt and the Sherwood num-
bers in this case are Nu Sh 3.66. Using the definition of Sherwood num-
ber, the mass transfer coefficient is determined to be
The mass transfer rate (or the evaporation rate) in this case can be determined
by defining the logarithmic mean concentration difference in an analogous
manner to the logarithmic mean temperature difference.
763
CHAPTER 14
PA, s MA
11 Pa 128.2 kg/kmol
wA, s 4.8 104
P Mair 101,325 Pa 29 kg/kmol
which confirms that the low mass flux approximation is valid. The rate of evap-
oration of naphthalene in this case is
m 0.012 kg
m· evap 1.33 105 kg/s
t (15 60 s)
Using the analogy between heat and mass transfer, the average heat transfer
coefficient is determined from Eq. 14–89 to be
2/3
hheat Cp hmass D
AB
2/3
2.141 10
5
m2/s
(1.184 kg/m3)(1007 J/kg · °C)(0.0776 m/s)
0.61 10
5
m2/s
215 W/m2 · °C
Heat
14–10 ■
SIMULTANEOUS HEAT AND Space vehicle
during reentry
Heat
MASS TRANSFER
Many mass transfer processes encountered in practice occur isothermally, and
thus they do not involve any heat transfer. But some engineering applications (a) Ablation (b) Evaporation of
involve the vaporization of a liquid and the diffusion of this vapor into the sur- rain droplet
rounding gas. Such processes require the transfer of the latent heat of vapor-
Heat
ization hfg to the liquid in order to vaporize it, and thus such problems involve rejection
simultaneous heat and mass transfer. To generalize, any mass transfer problem Evaporation
involving phase change (evaporation, sublimation, condensation, melting, Condensation
etc.) must also involve heat transfer, and the solution of such problems needs Vapor
to be analyzed by considering simultaneous heat and mass transfer. Some ex-
amples of simultaneous heat and mass problems are drying, evaporative cool- Liquid
ing, transpiration (or sweat) cooling, cooling by dry ice, combustion of fuel Evaporation
droplets, and ablation cooling of space vehicles during reentry, and even ordi-
Heat
nary events like rain, snow, and hail. In warmer locations, for example, the Heat
snow melts and the rain evaporates before reaching the ground (Fig. 14–51). absorption
To understand the mechanism of simultaneous heat and mass transfer, con- (c) Drying of clothes (d) Heat pipes
sider the evaporation of water from a swimming pool into air. Let us assume FIGURE 14–51
that the water and the air are initially at the same temperature. If the air is Many problems
saturated (a relative humidity of 100 percent), there will be no heat or encountered in practice involve
mass transfer as long as the isothermal conditions remain. But if the air is not simultaneous heat and mass transfer.
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 764
764
HEAT TRANSFER
FIGURE 14–52 where m· is the rate of evaporation and hfg is the latent heat of vaporization of
Various mechanisms of heat transfer water at the surface temperature. Various expressions for m· under various ap-
involved during the evaporation of proximations are given in Table 14–14. The mixture properties such as the
water from the surface of a lake. specific heat Cp and molar mass M should normally be evaluated at the mean
film composition and mean film temperature. However, when dealing with
air–water vapor mixtures at atmospheric conditions or other low mass flux sit-
uations, we can simply use the properties of the gas with reasonable accuracy.
TABLE 14–14
Various expressions for evaporation rate of a liquid into a gas through an
interface area As under various approximations (subscript stands for vapor,
s for liquid–gas interface, and away from surface)
Assumption Evaporation Rate
General m· hmass As(, s , )
hmass As P, s P,
Assuming vapor to be an ideal gas,
P RT
m·
R Ts
T
hmass As P, s P,
Using Chilton–Colburn analogy,
hheat CphmassLe2/3
m·
CpLe2/3R Ts
T
1 1 1 Ts T hmass As M P, s P,
Using , where T m·
Ts T T 2 CpLe2/3 M P
and P RT (Ru /M)T
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 765
765
CHAPTER 14
·
The Q in Eq. 14–91 represents all forms of heat from all sources transferred
to the surface, including convection and radiation from the surroundings and
conduction from the deeper parts of the water due to the sensible energy of the
water itself or due to heating the water body by a resistance heater, heating
coil, or even chemical reactions in the water. If heat transfer from the water
body to the surface as well as radiation from the surroundings is negligible,
which is often the case, then the heat loss by evaporation must equal heat gain
by convection. That is,
which is a relation for the temperature difference between the liquid and the
surrounding gas under steady conditions.
Water: hfg 2454 kJ/kg, P 2.34 kPa; also, P 4.25 kPa at 30°C
Dry air: CP 1.007 kJ/kg · °C, 2.141 105 m2/s
The molar masses of water and air are 18 and 29 kg/kmol, respectively (Table
A–1). Also, the mass diffusivity of water vapor in air at 25°C is DH2O-air 2.50
105 m2/s (Table 14–4).
Analysis Utilizing the Chilton–Colburn analogy, the surface temperature of the
drink can be determined from Eq. 14–92,
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 766
766
HEAT TRANSFER
Note that we could take the Lewis number to be 1 for simplicity, but we chose
to incorporate it for better accuracy.
The air at the surface is saturated, and thus the vapor pressure at the sur-
face is simply the saturation pressure of water at the surface temperature
(2.34 kPa). The vapor pressure of air far from the surface is determined from
Noting that the atmospheric pressure is 1 atm 101.3 kPa, substituting gives
767
CHAPTER 14
Properties Relevant properties for each mode of heat transfer are determined
below in respective sections.
663 W
(b) The air–water vapor mixture is dilute and thus we can use dry air properties
for the mixture at the average temperature of (T Ts)/2 (25 50)/2
37.5°C. Noting that the total atmospheric pressure is 92/101.3 0.9080 atm,
the properties of dry air at 37.5°C and 0.908 atm are
The air at the surface is saturated, and thus the vapor pressure at the surface is
simply the saturation pressure of water at the surface temperature (12.35 kPa).
The vapor pressure of air far from the water surface is determined from
Treating the water vapor and the air as ideal gases and noting that the total at-
mospheric pressure is the sum of the vapor and dry air pressures, the densities
of the water vapor, dry air, and their mixture at the water–air interface and far
from the surface are determined to be
and
The area of the top surface of the water bath is As (3.5 m)(1 m) 3.5 m2 and
its perimeter is p 2(3.5 1) 9 m. Therefore, the characteristic length is
cen58933_ch14.qxd 9/9/2002 10:05 AM Page 768
768
HEAT TRANSFER
As 3.5 m2
Lc p 0.3889 m
9m
Then using densities (instead of temperatures) since the mixture is not homo-
geneous, the Grashof number is determined to be
g( S)L3c
Gr
2
(9.81 m/s2)(1.0684 0.9420 kg/m3)(0.3889 m)3
[(0.9420 1.0684)/2 kg/m3](1.848 105 m2/s)2
2.125 108
Recognizing that this is a natural convection problem with hot horizontal sur-
face facing up, the Nusselt number and the convection heat transfer coeffi-
cients are determined to be
and
T 2.072 310.52.072
DAB DH2O–air 1.87 1010 1.87 1010
P 0.908
3.00 105 m2/s
The Sherwood number and the mass transfer coefficients are determined to be
and
769
CHAPTER 14
Then the evaporation rate and the rate of heat transfer by evaporation become
and
·
Q evap m· hfg (0.00146 kg/s)(2383 kJ/kg) 3.479 kW 3479 W
which is more than seven times the rate of heat transfer by natural convection.
Finally, noting that the direction of heat transfer is always from high to low
temperature, all forms of heat transfer determined above are in the same direc-
tion, and the total rate of heat loss from the water to the surrounding air and
surfaces is
· · · ·
Q total Q rad Q conv Q evap 663 479 3479 4621 W
Discussion Note that if the water bath is heated electrically, a 4.6 kW resis-
tance heater will be needed just to make up for the heat loss from the top sur-
face. The total heater size will have to be larger to account for the heat losses
from the side and bottom surfaces of the bath as well as the heat absorbed by
the spray paint cans as they are heated to 50°C. Also note that water needs to
be supplied to the bath at a rate of 5.25 kg/h to make up for the water loss by
evaporation. Also, in reality, the surface temperature will probably be a little
lower than the bulk water temperature, and thus the heat transfer rates will be
somewhat lower than indicated here.
SUMMARY
770
HEAT TRANSFER
where DAB is the diffusion coefficient (or mass diffusivity) of where PA, 1 and PA, 2 are the partial pressures of gas A on the
the species in the mixture, jdiff, A is the diffusive mass flux of two sides of the wall.
species A, and j̄diff, A is the molar flux. During mass transfer in a moving medium, chemical species
The mole fractions of a species i in the gas and liquid phases are transported both by molecular diffusion and by the bulk
at the interface of a dilute mixture are proportional to each fluid motion, and the velocities of the species are expressed as
other and are expressed by Henry’s law as
A diff, A
Pi, gas side B diff, B
yi, liquid side
H
where is the mass-average velocity of the flow. It is the
where H is Henry’s constant. When the mixture is not dilute, an velocity that would be measured by a velocity sensor and is
approximate relation for the mole fractions of a species on the expressed as
liquid and gas sides of the interface are expressed approxi-
mately by Raoult’s law as wAA wBB
Pi, gas side yi, gas side P yi, liquid side Pi, sat(T)
The special case 0 corresponds to a stationary medium.
Using Fick’s law of diffusion, the total mass fluxes j m· /A in
where Pi, sat(T) is the saturation pressure of the species i at the a moving medium are expressed as
interface temperature and P is the total pressure on the gas
phase side. dwA
The concentration of the gas species i in the solid at the in- jA A Adiff, A wA( jA jB) DAB
dx
terface Ci, solid side is proportional to the partial pressure of the
species i in the gas Pi, gas side on the gas side of the interface and dwB
jB B Bdiff, B wB( jA jB) DBA
is expressed as dx
Ci, solid side Pi, gas side The rate of mass convection of species A in a binary mixture
is expressed in an analogous manner to Newton’s law of cool-
where is the solubility. The product of the solubility of a gas ing as
and the diffusion coefficient of the gas in a solid is referred to
as the permeability , which is a measure of the ability of the m· conv hmass As(A, s A, ) hmass As(wA, s wA, )
gas to penetrate a solid.
In the absence of any chemical reactions, the mass transfer where hmass is the average mass transfer coefficient, in m/s.
rates m· diff, A through a plane wall of area A and thickness L and The counterparts of the Prandtl and Nusselt numbers in mass
cylindrical and spherical shells of inner and outer radii r1 and r2 convection are the Schmidt number Sc and the Sherwood num-
under one-dimensional steady conditions are expressed as ber Sh, defined as
wA, 1 wA, 2 A, 1 A, 2
m· diff, A, wall DAB A DAB A Momentum diffusivity hmass L
L L Sc and Sh
DAB Mass diffusivity DAB
wA, 1 wA, 2 A, 1 A, 2
m· diff, A, cyl 2LDAB 2LDAB
ln(r2/r1) ln(r2/r1) The relative magnitudes of heat and mass diffusion in the ther-
wA, 1 wA, 2 mal and concentration boundary layers are represented by the
m· diff, A, sph 4r1r2DAB r2 r1 Lewis number, defined as
A, 1 A, 2 Sc Thermal diffusivity
4r1r2DAB r2 r1 Le
Pr DAB Mass diffusivity
The mass flow rate of a gas through a solid plane wall under Heat and mass transfer coefficients are sometimes expressed in
steady one-dimensional conditions can also be expressed in terms of the dimensionless Stanton number, defined as
terms of the partial pressures of the adjacent gas on the two
sides of the solid as hconv 1 hconv 1
Stheat Nu and Stmass Sh
PA, 1 PA, 2 Cp Re Pr Re Sc
m· diff, A, wall DAB AB A
L
where is the free-stream velocity in external flow and the
PA, 1 PA, 2 bulk mean fluid velocity in internal flow. For a given geometry
AB A
L and boundary conditions, the Sherwood number in natural or
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 771
771
CHAPTER 14
forced convection can be determined from the corresponding which is known as the Chilton–Colburn analogy. The analogy
Nusselt number expression by simply replacing the Prandtl between heat and mass transfer is expressed more conve-
number by the Schmidt number. But in natural convection, the niently as
Grashof number should be expressed in terms of density dif-
ference instead of temperature difference. hheat CpLe2/3 hmass Cp( /DAB)2/3hmass
When the molecular diffusivities of momentum, heat,
and mass are identical, we have DAB, and thus Pr For air–water vapor mixtures, Le 1, and thus this relation
Sc Le 1. The similarity between momentum, heat, and simplifies further. The heat–mass convection analogy is limited
mass transfer in this case is given by the Reynolds analogy, to low mass flux cases in which the flow rate of species under-
expressed as going mass flow is low relative to the total flow rate of the
liquid or gas mixture. The mass transfer problems that involve
f phase change (evaporation, sublimation, condensation, melt-
Re Nu Sh or
2 ing, etc.) also involve heat transfer, and such problems are an-
f L hheat L hmass L f alyzed by considering heat and mass transfer simultaneously.
or St Stmass
2 k DAB 2
f
St Pr2/3 StmassSc2/3
2
1. American Society of Heating, Refrigeration, and Air 11. W. M. Kays and M. E. Crawford. Convective Heat and
Conditioning Engineers. Handbook of Fundamentals. Mass Transfer. 2nd ed. New York: McGraw-Hill, 1980.
Atlanta: ASHRAE, 1993. 12. T. R. Marrero and E. A. Mason. “Gaseous Diffusion
2. R. M. Barrer. Diffusion in and through Solids. New York: Coefficients.” Journal of Phys. Chem. Ref. Data 1 (1972),
Macmillan, 1941. pp. 3–118.
3. R. B. Bird. “Theory of Diffusion.” Advances in Chemical 13. A. F. Mills. Basic Heat and Mass Transfer. Burr Ridge,
Engineering 1 (1956), p. 170. IL: Richard D. Irwin, 1995.
4. R. B. Bird, W. E. Stewart, and E. N. Lightfoot. Transport 14. J. H. Perry, ed. Chemical Engineer’s Handbook. 4th ed.
Phenomena. New York: John Wiley & Sons, 1960. New York: McGraw-Hill, 1963.
5. C. J. Geankoplis. Mass Transport Phenomena. New York: 15. R. D. Reid, J. M. Prausnitz, and T. K. Sherwood. The
Holt, Rinehart, and Winston, 1972. Properties of Gases and Liquids. 3rd ed. New York:
6. Handbook of Chemistry and Physics 56th ed. Cleveland, McGraw-Hill, 1977.
OH: Chemical Rubber Publishing Co., 1976. 16. A. H. P. Skelland. Diffusional Mass Transfer. New York:
7. J. O. Hirshfelder, F. Curtis, and R. B. Bird. Molecular John Wiley & Sons, 1974.
Theory of Gases and Liquids. New York: John Wiley & 17. D. B. Spalding. Convective Mass Transfer. New York:
Sons, 1954. McGraw-Hill, 1963.
8. J. P. Holman. Heat Transfer. 7th ed. New York: 18. W. F. Stoecker and J. W. Jones. Refrigeration and Air
McGraw-Hill, 1990. Conditioning. New York: McGraw-Hill, 1982.
9. F. P. Incropera and D. P. De Witt. Fundamentals of Heat 19. L. C. Thomas. Mass Transfer Supplement—Heat Transfer.
and Mass Transfer. 2nd ed. New York: John Wiley & Englewood Cliffs, NJ: Prentice Hall, 1991.
Sons, 1985. 20. L. Van Black. Elements of Material Science and
10. International Critical Tables. Vol. 3. New York: Engineering. Reading, MA: Addison-Wesley, 1980.
McGraw-Hill, 1928.
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 772
772
HEAT TRANSFER
PROBLEMS*
Analogy between Heat and Mass Transfer 14–9C Mark these statements as being True or False for a
14–1C How does mass transfer differ from bulk fluid flow? binary mixture of substances A and B.
Can mass transfer occur in a homogeneous medium? (a) The molar concentration of a mixture is always
14–2C How is the concentration of a commodity defined? equal to the sum of the molar concentrations of its
How is the concentration gradient defined? How is the diffu- constituents.
sion rate of a commodity related to the concentration gradient? (b) The ratio of the molar concentration of A to the molar
concentration of B is equal to the mole fraction of
14–3C Give examples for (a) liquid-to-gas, (b) solid-to- component A.
liquid, (c) solid-to-gas, and (d) gas-to-liquid mass transfer. (c) If the mole fraction of component A is greater than
14–4C Someone suggests that thermal (or heat) radiation can 0.5, then at least half of the mass of the mixture is
also be viewed as mass radiation since, according to Einstein’s component A.
formula, an energy transfer in the amount of E corresponds to (d) If both A and B are ideal gases, then the pressure frac-
a mass transfer in the amount of m E/c2. What do you think? tion of A is equal to its mole fraction.
(e) If the mole fractions of A and B are both 0.5, then the
14–5C What is the driving force for (a) heat transfer,
molar mass of the mixture is simply the arithmetic
(b) electric current flow, (c) fluid flow, and (d) mass transfer?
average of the molar masses of A and B.
14–6C What do (a) homogeneous reactions and (b) hetero-
geneous reactions represent in mass transfer? To what do they 14–10C Fick’s law of diffusion is expressed on the mass
·
correspond in heat transfer? and mole basis as m· diff, A ADAB(dwA/dx) and Ndiff, A
CADAB(dyA/dx), respectively. Are the diffusion coefficients
Mass Diffusion DAB in the two relations the same or different?
14–7C Both Fourier’s law of heat conduction and Fick’s law 14–11C How does the mass diffusivity of a gas mixture
·
of mass diffusion can be expressed as Q kA(dT/dx). What change with (a) temperature and (b) pressure?
·
do the quantities Q , k, A, and T represent in (a) heat conduction
and (b) mass diffusion? 14–12C At a given temperature and pressure, do you think
14–8C Mark these statements as being True or False for a the mass diffusivity of air in water vapor will be equal to the
binary mixture of substances A and B. mass diffusivity of water vapor in air? Explain.
(a) The density of a mixture is always equal to the sum 14–13C At a given temperature and pressure, do you think
of the densities of its constituents. the mass diffusivity of copper in aluminum will be equal to the
(b) The ratio of the density of component A to the den- mass diffusivity of aluminum in copper? Explain.
sity of component B is equal to the mass fraction of
14–14C In a mass production facility, steel components are
component A.
to be hardened by carbon diffusion. Would you carry out the
(c) If the mass fraction of component A is greater than
hardening process at room temperature or in a furnace at a high
0.5, then at least half of the moles of the mixture are
temperature, say 900°C? Why?
component A.
(d) If the molar masses of A and B are equal to each 14–15C Someone claims that the mass and the mole frac-
other, then the mass fraction of A will be equal to the tions for a mixture of CO2 and N2O gases are identical. Do you
mole fraction of A. agree? Explain.
(e) If the mass fractions of A and B are both 0.5, then the
molar mass of the mixture is simply the arithmetic 14–16 The composition of moist air is given on a molar basis
average of the molar masses of A and B. to be 78 percent N2, 20 percent O2, and 2 percent water vapor.
Determine the mass fractions of the constituents of air.
Answers: 76.4 percent N2, 22.4 percent O2, 1.2 percent H2O
*Problems designated by a “C” are concept questions, and
students are encouraged to answer them all. Problems designated 14–17E A gas mixture consists of 5 lbm of O2, 8 lbm of N2,
by an “E” are in English units, and the SI users can ignore them. and 10 lbm of CO2. Determine (a) the mass fraction of each
Problems with an EES-CD icon are solved using EES, and component, (b) the mole fraction of each component, and
complete solutions together with parametric studies are included (c) the average molar mass of the mixture.
on the enclosed CD. Problems with a computer-EES icon are
comprehensive in nature, and are intended to be solved with a 14–18 A gas mixture consists of 8 kmol of H2 and 2 kmol of
computer, preferably using the EES software that accompanies N2. Determine the mass of each gas and the apparent gas con-
this text. stant of the mixture.
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 773
773
CHAPTER 14
14–19 The molar analysis of a gas mixture at 290 K and the same as the mole fraction of water in the lake (which is
250 kPa is 65 percent N2, 20 percent O2, and 15 percent CO2. nearly 1)?
Determine the mass fraction and partial pressure of each gas. 14–28C When prescribing a boundary condition for mass
14–20 Determine the binary diffusion coefficient of CO2 transfer at a solid–gas interface, why do we need to specify the
in air at (a) 200 K and 1 atm, (b) 400 K and 0.8 atm, and side of the surface (whether the solid or the gas side)? Why did
(c) 600 K and 3 atm. we not do it in heat transfer?
14–21 Repeat Problem 14–20 for O2 in N2. 14–29C Using properties of saturated water, explain how
you would determine the mole fraction of water vapor at the
14–22E The relative humidity of air at 80°F and 14.7 psia is
surface of a lake when the temperature of the lake surface and
increased from 30 percent to 90 percent during a humidifica-
the atmospheric pressure are specified.
tion process at constant temperature and pressure. Determine
the percent error involved in assuming the density of air to 14–30C Using solubility data of a solid in a specified liquid,
have remained constant. Answer: 2.1 percent explain how you would determine the mass fraction of the
solid in the liquid at the interface at a specified temperature.
14–31C Using solubility data of a gas in a solid, explain how
you would determine the molar concentration of the gas in the
solid at the solid–gas interface at a specified temperature.
14–32C Using Henry’s constant data for a gas dissolved in a
80°F liquid, explain how you would determine the mole fraction of
14.7 psia the gas dissolved in the liquid at the interface at a specified
30% RH temperature.
14–33C What is permeability? How is the permeability of a
Humidifier gas in a solid related to the solubility of the gas in that solid?
14–34E Determine the mole fraction of the water vapor at
the surface of a lake whose temperature at the surface is 60°F,
and compare it to the mole fraction of water in the lake. Take
the atmospheric pressure at lake level to be 13.8 psia.
14–35 Determine the mole fraction of dry air at the surface of
FIGURE P14–22E a lake whose temperature is 15°C. Take the atmospheric pres-
sure at lake level to be 100 kPa. Answer: 98.3 percent
14–23 The diffusion coefficient of hydrogen in steel is given
as a function of temperature as 14–36 Reconsider Problem 14–35. Using EES (or
other) software, plot the mole fraction of dry air
DAB 1.65 106 exp(–4630/T) (m2/s) at the surface of the lake as a function of the lake temperature
as the temperatue varies from 5°C to 25°C, and discuss the
where T is in K. Determine the diffusion coefficients from results.
200 K to 1200 K in 200 K increments and plot the results. 14–37 Consider a rubber plate that is in contact with nitrogen
14–24 Reconsider Problem 14–23. Using EES (or gas at 298 K and 250 kPa. Determine the molar and mass den-
other) software, plot the diffusion coefficient as sities of nitrogen in the rubber at the interface.
a function of the temperature in the range of 200 K to 1200 K. Answers: 0.0039 kmol/m3, 0.1092 kg/m3
Rubber
plate
Boundary Conditions
14–25C Write three boundary conditions for mass transfer
(on a mass basis) for species A at x 0 that correspond to
specified temperature, specified heat flux, and convection N2
boundary conditions in heat transfer. 298 K
250 kPa
14–26C What is an impermeable surface in mass transfer?
How is it expressed mathematically (on a mass basis)? To what
does it correspond in heat transfer?
ρN2 = ?
14–27C Consider the free surface of a lake exposed to the
atmosphere. If the air at the lake surface is saturated, will the
mole fraction of water vapor in air at the lake surface be FIGURE P14–37
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 774
774
HEAT TRANSFER
14–38 A wall made of natural rubber separates O2 and N2 (d) Other things being equal, doubling the mass fraction
gases at 25°C and 500 kPa. Determine the molar concentra- of the diffusing species at the high concentration side
tions of O2 and N2 in the wall. will double the rate of mass transfer.
14–39 Consider a glass of water in a room at 20°C and 14–44C Consider one-dimensional mass diffusion of species
97 kPa. If the relative humidity in the room is 100 percent and A through a plane wall of thickness L. Under what conditions
the water and the air are in thermal and phase equilibrium, de- will the concentration profile of species A in the wall be a
termine (a) the mole fraction of the water vapor in the air and straight line?
(b) the mole fraction of air in the water. 14–45C Consider one-dimensional mass diffusion of species
14–40E Water is sprayed into air at 80°F and 14.3 psia, and A through a plane wall. Does the species A content of the wall
the falling water droplets are collected in a container on the change during steady mass diffusion? How about during tran-
floor. Determine the mass and mole fractions of air dissolved sient mass diffusion?
in the water. 14–46 Helium gas is stored at 293 K in a 3-m-outer-diameter
14–41 Consider a carbonated drink in a bottle at 27°C and spherical container made of 5-cm-thick Pyrex. The molar con-
130 kPa. Assuming the gas space above the liquid consists of a centration of helium in the Pyrex is 0.00073 kmol/m3 at the
saturated mixture of CO2 and water vapor and treating the inner surface and negligible at the outer surface. Determine
drink as water, determine (a) the mole fraction of the water the mass flow rate of helium by diffusion through the Pyrex
vapor in the CO2 gas and (b) the mass of dissolved CO2 in a container. Answer: 7.2 1015 kg/s
200-ml drink. Answers: (a) 2.77 percent, (b) 0.36 g
5 cm
CO2 Pyrex
H2O
He gas Air
293 K
He
diffusion
FIGURE P14–46
27°C
130 kPa
14–47 A thin plastic membrane separates hydrogen from air.
The molar concentrations of hydrogen in the membrane at the
FIGURE P14–41 inner and outer surfaces are determined to be 0.065 and 0.003
kmol/m3, respectively. The binary diffusion coefficient of hy-
drogen in plastic at the operation temperature is 5.3 1010
Steady Mass Diffusion through a Wall m2/s. Determine the mass flow rate of hydrogen by diffusion
through the membrane under steady conditions if the thickness
14–42C Write down the relations for steady one-dimensional of the membrane is (a) 2 mm and (b) 0.5 mm.
heat conduction and mass diffusion through a plane wall, and
identify the quantities in the two equations that correspond to 14–48 The solubility of hydrogen gas in steel in terms of its
each other. mass fraction is given as wH2 2.09 104 exp(–3950/T)P0,5 H2
where PH2 is the partial pressure of hydrogen in bars and T is
14–43C Consider steady one-dimensional mass diffusion the temperature in K. If natural gas is transported in a 1-cm-
through a wall. Mark these statements as being True or False. thick, 3-m-internal-diameter steel pipe at 500 kPa pressure and
(a) Other things being equal, the higher the density of the the mole fraction of hydrogen in the natural gas is 8 percent,
wall, the higher the rate of mass transfer. determine the highest rate of hydrogen loss through a 100-m-
(b) Other things being equal, doubling the thickness of long section of the pipe at steady conditions at a temperature
the wall will double the rate of mass transfer. of 293 K if the pipe is exposed to air. Take the diffusivity of hy-
(c) Other things being equal, the higher the temperature, drogen in steel to be 2.9 1013 m2/s.
the higher the rate of mass transfer. Answer: 3.98 1014 kg/s
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 775
775
CHAPTER 14
14–49 Reconsider Problem 14–48. Using EES (or determine how long it will take for the pressure inside the
other) software, plot the highest rate of hydro- balloon to drop to 100 kPa.
gen loss as a function of the mole fraction of hydrogen in nat- 14–53 Pure N2 gas at 1 atm and 25°C is flowing through a
ural gas as the mole fraction varies from 5 to 15 percent, and 10-m-long, 3-cm-inner diameter pipe made of 1-mm-thick
discuss the results. rubber. Determine the rate at which N2 leaks out of the pipe if
14–50 Helium gas is stored at 293 K and 500 kPa in a 1-cm- the medium surrounding the pipe is (a) a vacuum and (b) at-
thick, 2-m-inner-diameter spherical tank made of fused silica mospheric air at 1 atm and 25°C with 21 percent O2 and
(SiO2). The area where the container is located is well venti- 79 percent N2.
lated. Determine (a) the mass flow rate of helium by diffusion Answers: (a) 4.48 1010 kmol/s, (b) 9.4 1011 kmol/s
through the tank and (b) the pressure drop in the tank in one
week as a result of the loss of helium gas. Vacuum
14–51 You probably have noticed that balloons inflated with N2
helium gas rise in the air the first day during a party but they
fall down the next day and act like ordinary balloons filled with 1 atm
air. This is because the helium in the balloon slowly leaks out N2 gas
25°C
through the wall while air leaks in by diffusion.
Consider a balloon that is made of 0.1-mm-thick soft rubber
and has a diameter of 15 cm when inflated. The pressure and Rubber pipe
temperature inside the balloon are initially 110 kPa and 25°C. FIGURE P14–53
The permeability of rubber to helium, oxygen, and nitrogen
at 25°C are 9.4 1013, 7.05 1013, and 2.6 1013
kmol/m · s · bar, respectively. Determine the initial rates of dif- Water Vapor Migration in Buildings
fusion of helium, oxygen, and nitrogen through the balloon 14–54C Consider a tank that contains moist air at 3 atm and
wall and the mass fraction of helium that escapes the balloon whose walls are permeable to water vapor. The surrounding air
during the first 5 h assuming the helium pressure inside the bal- at 1 atm pressure also contains some moisture. Is it possible for
loon remains nearly constant. Assume air to be 21 percent oxy- the water vapor to flow into the tank from surroundings?
gen and 79 percent nitrogen by mole numbers and take the Explain.
room conditions to be 100 kPa and 25°C.
14–55C Express the mass flow rate of water vapor through a
wall of thickness L in terms of the partial pressure of water
vapor on both sides of the wall and the permeability of the wall
to the water vapor.
110 kPa
25°C 14–56C How does the condensation or freezing of water
He vapor in the wall affect the effectiveness of the insulation in the
Air
wall? How does the moisture content affect the effective ther-
mal conductivity of soil?
14–57C Moisture migration in the walls, floors, and ceilings
of buildings is controlled by vapor barriers or vapor retarders.
Explain the difference between the two, and discuss which is
more suitable for use in the walls of residential buildings.
14–58C What are the adverse effects of excess moisture
on the wood and metal components of a house and the paint on
the walls?
14–59C Why are the insulations on the chilled water lines
always wrapped with vapor barrier jackets?
FIGURE P14–51 14–60C Explain how vapor pressure of the ambient air is
determined when the temperature, total pressure, and relative
14–52 Reconsider the balloon discussed in Problem 14–51. humidity of the air are given.
Assuming the volume to remain constant and disregarding 14–61 The diffusion of water vapor through plaster boards
the diffusion of air into the balloon, obtain a relation for the and its condensation in the wall insulation in cold weather are
variation of pressure in the balloon with time. Using the results of concern since they reduce the effectiveness of insulation.
obtained and the numerical values given in the problem, Consider a house that is maintained at 20°C and 60 percent
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 776
776
HEAT TRANSFER
Plaster 25°C
board 88 kPa
50% RH Moisture
migration
9.5 mm
Room
Outdoors Aluminum
foil
20°C
97 kPa Vapor
diffusion Milk
60% RH
25°C
FIGURE P14–61
FIGURE P14–66
relative humidity at a location where the atmospheric pressure
is 97 kPa. The inside of the walls is finished with 9.5-mm-thick permeance is 2.9 1012 kg/s · m2 · Pa. The inner diameter of
gypsum wallboard. Taking the vapor pressure at the outer side the glass is 12 cm. Assuming the air in the glass to be saturated
of the wallboard to be zero, determine the maximum amount of at all times, determine how much the level of the milk in the
water vapor that will diffuse through a 3-m 8-m section of a glass will recede in 12 h. Answer: 0.00079 mm
wall during a 24-h period. The permeance of the 9.5-mm-thick
gypsum wallboard to water vapor is 2.86 109 kg/s · m2 · Pa. Transient Mass Diffusion
14–62 Reconsider Problem 14–61. In order to reduce the 14–67C In transient mass diffusion analysis, can we treat the
migration of water vapor through the wall, it is proposed to use diffusion of a solid into another solid of finite thickness (such
a 0.2-mm-thick polyethylene film with a permeance of 2.3 as the diffusion of carbon into an ordinary steel component) as
1012 kg/s · m2 · Pa. Determine the amount of water vapor that a diffusion process in a semi-infinite medium? Explain.
will diffuse through the wall in this case during a 24-h period. 14–68C Define the penetration depth for mass transfer, and
Answer: 6.7 g explain how it can be determined at a specified time when the
14–63 The roof of a house is 15 m 8 m and is made of a diffusion coefficient is known.
20-cm-thick concrete layer. The interior of the house is main- 14–69C When the density of a species A in a semi-infinite
tained at 25°C and 50 percent relative humidity and the local medium is known at the beginning and at the surface, explain
atmospheric pressure is 100 kPa. Determine the amount of how you would determine the concentration of the species A at
water vapor that will migrate through the roof in 24 h if the av- a specified location and time.
erage outside conditions during that period are 3°C and 30 per-
14–70 A steel part whose initial carbon content is 0.12 per-
cent relative humidity. The permeability of concrete to water
cent by mass is to be case-hardened in a furnace at 1150 K by
vapor is 24.7 1012 kg/s · m · Pa.
exposing it to a carburizing gas. The diffusion coefficient of
14–64 Reconsider Problem 14–63. Using EES (or carbon in steel is strongly temperature dependent, and at the
other) software, investigate the effects of tem- furnace temperature it is given to be DAB 7.2 1012 m2/s.
perature and relative humidity of air inside the house on the
amount of water vapor that will migrate through the roof. Let
the temperature vary from 15°C to 30°C and the relative hu-
midity from 30 to 70 percent. Plot the amount of water vapor 1150 K
that will migrate as functions of the temperature and the rela-
tive humidity, and discuss the results.
Carbon
14–65 Reconsider Problem 14–63. In order to reduce the mi-
gration of water vapor, the inner surface of the wall is painted
with vapor retarder latex paint whose permeance is 26 1012 Steel part
kg/s · m2 · Pa. Determine the amount of water vapor that will
diffuse through the roof in this case during a 24-h period.
14–66 A glass of milk left on top of a counter in the kitchen
at 25°C, 88 kPa, and 50 percent relative humidity is tightly
sealed by a sheet of 0.009-mm-thick aluminum foil whose FIGURE P14–70
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 777
777
CHAPTER 14
Also, the mass fraction of carbon at the exposed surface of the (a) The rates of mass diffusion of species A and B are
steel part is maintained at 0.011 by the carbon-rich environ- equal in magnitude and opposite in direction.
ment in the furnace. If the hardening process is to continue (b) DAB DBA.
until the mass fraction of carbon at a depth of 0.7 mm is raised (c) During equimolar counterdiffusion through a tube,
to 0.32 percent, determine how long the part should be held in equal numbers of moles of A and B move in opposite
the furnace. Answer: 9 h directions, and thus a velocity measurement device
14–71 Repeat Problem 14–70 for a furnace temperature of placed in the tube will read zero.
500 K at which the diffusion coefficient of carbon in steel is (d) The lid of a tank containing propane gas (which is
DAB 2.1 1020 m2/s. heavier than air) is left open. If the surrounding air
and the propane in the tank are at the same tempera-
14–72 A pond with an initial oxygen content of zero is to be ture and pressure, no propane will escape the tank
oxygenated by forming a tent over the water surface and filling and no air will enter.
the tent with oxygen gas at 25°C and 130 kPa. Determine the
mole fraction of oxygen at a depth of 2 cm from the surface 14–78C What is Stefan flow? Write the expression for
after 12 h. Stefan’s law and indicate what each variable represents.
14–79E The pressure in a pipeline that transports helium gas
Tent at a rate of 5 lbm/s is maintained at 14.5 psia by venting helium
O2 gas to the atmosphere through a 14-in. internal diameter tube that
25°C extends 30 ft into the air. Assuming both the helium and the
130 kPa atmospheric air to be at 80°F, determine (a) the mass flow rate
of helium lost to the atmosphere through an individual tube,
O2 diffusion
(b) the mass flow rate of air that infiltrates into the pipeline,
and (c) the flow velocity at the bottom of the tube where it is
Pond
attached to the pipeline that will be measured by an anemome-
ter in steady operation.
778
HEAT TRANSFER
14–82 Reconsider Problem 14–81. Using EES (or 14–87C What is a concentration boundary layer? How is it
other) software, plot the mass flow rate of defined for flow over a plate?
hydrogen lost as a function of the diameter of the charging 14–88C What is the physical significance of the Schmidt
valve as the diameter varies from 1 cm to 5 cm, and discuss the number? How is it defined? To what dimensionless number
results. does it correspond in heat transfer? What does a Schmidt num-
14–83E A 1-in.-diameter Stefan tube is used to measure the ber of 1 indicate?
binary diffusion coefficient of water vapor in air at 70°F and
14–89C What is the physical significance of the Sherwood
13.8 psia. The tube is partially filled with water with a distance
number? How is it defined? To what dimensionless number
from the water surface to the open end of the tube of 10 in. Dry
does it correspond in heat transfer? What does a Sherwood
air is blown over the open end of the tube so that water vapor
number of 1 indicate for a plain fluid layer?
rising to the top is removed immediately and the concentration
of vapor at the top of the tube is zero. During 10 days of con- 14–90C What is the physical significance of the Lewis
tinuous operation at constant pressure and temperature, the number? How is it defined? What does a Lewis number of 1
amount of water that has evaporated is measured to be 0.0015 indicate?
lbm. Determine the diffusion coefficient of water vapor in air 14–91C In natural convection mass transfer, the Grashof
at 70°F and 13.8 psia. number is evaluated using density difference instead of tem-
14–84 An 8-cm-internal-diameter, 30-cm-high pitcher half perature difference. Can the Grashof number evaluated this
filled with water is left in a dry room at 15°C and 87 kPa with way be used in heat transfer calculations also?
its top open. If the water is maintained at 15°C at all times also, 14–92C Using the analogy between heat and mass transfer,
determine how long it will take for the water to evaporate explain how the mass transfer coefficient can be determined
completely. Answer: 1125 days
from the relations for the heat transfer coefficient.
14–93C It is well known that warm air in a cooler environ-
Room
ment rises. Now consider a warm mixture of air and gasoline
15°C
(C8H18) on top of an open gasoline can. Do you think this gas
87 kPa Water mixture will rise in a cooler environment?
vapor
14–94C Consider two identical cups of coffee, one with no
sugar and the other with plenty of sugar at the bottom. Initially,
both cups are at the same temperature. If left unattended, which
cup of coffee will cool faster?
14–95C Under what conditions will the normalized velocity,
thermal, and concentration boundary layers coincide during
flow over a flat plate?
14–96C What is the relation ( f/2) Re Nu Sh known as?
Water Under what conditions is it valid? What is the practical im-
15°C portance of it?
14–97C What is the name of the relation f/2 St Pr2/3
StmassSc2/3 and what are the names of the variables in it? Under
what conditions is it valid? What is the importance of it in
FIGURE P14–84 engineering?
14–98C What is the relation hheat Cp hmass known as? For
14–85 A large tank containing ammonia at 1 atm and 25°C is what kind of mixtures is it valid? What is the practical im-
vented to the atmosphere through a 3-m-long tube whose inter- portance of it?
nal diameter is 1 cm. Determine the rate of loss of ammonia 14–99C What is the low mass flux approximation in mass
and the rate of infiltration of air into the tank. transfer analysis? Can the evaporation of water from a lake be
treated as a low mass flux process?
Mass Convection 14–100E Consider a circular pipe of inner diameter D
14–86C Heat convection is expressed by Newton’s law of 0.5 in. whose inner surface is covered with a thin layer of
·
cooling as Q hA(Ts T). Express mass convection in an liquid water as a result of condensation. In order to dry the
analogous manner on a mass basis, and identify all the quanti- pipe, air at 540 R and 1 atm is forced to flow through it with an
ties in the expression and state their units. average velocity of 4 ft/s. Using the analogy between heat and
cen58933_ch14.qxd 9/9/2002 10:06 AM Page 779
779
CHAPTER 14
mass transfer, determine the mass transfer coefficient inside the 14–105 Consider a 5-m 5-m wet concrete patio with an
pipe for fully developed flow. Answer: 0.024 ft/s average water film thickness of 0.3 mm. Now wind at 50 km/h
14–101 The average heat transfer coefficient for air flow is blowing over the surface. If the air is at 1 atm, 15°C, and
over an odd-shaped body is to be determined by mass transfer 35 percent relative humidity, determine how long it will take
measurements and using the Chilton–Colburn analogy between for the patio to dry completely. Answer: 18.6 min
heat and mass transfer. The experiment is conducted by blow- 14–106E A 2-in.-diameter spherical naphthalene ball is sus-
ing dry air at 1 atm at a free stream velocity of 2 m/s over a pended in a room at 1 atm and 80°F. Determine the average
body covered with a layer of naphthalene. The surface area of mass transfer coefficient between the naphthalene and the air
the body is 0.75 m2, and it is observed that 100 g of naph- if air is forced to flow over naphthalene with a free stream ve-
thalene has sublimated in 45 min. During the experiment, both locity of 15 ft/s. The Schmidt number of naphthalene in air at
the body and the air were kept at 25°C, at which the vapor room temperature is 2.35. Answer: 0.0525 ft/s
pressure and mass diffusivity of naphthalene are 11 Pa and
DAB 0.61 105 m2/s, respectively. Determine the heat 14–107 Consider a 3-mm-diameter raindrop that is falling
transfer coefficient under the same flow conditions over the freely in atmospheric air at 25°C. Taking the temperature of the
same geometry. raindrop to be 9°C, determine the terminal velocity of the rain-
drop at which the drag force equals the weight of the drop and
0.75 m2 the average mass transfer coefficient at that time.
Air
14–108 In a manufacturing facility, wet brass plates coming
1 atm out of a water bath are to be dried by passing them through a
Body
2 m /s section where dry air at 1 atm and 25°C is blown parallel to
25°C
25°C their surfaces. If the plates are at 20°C and there are no dry
spots, determine the rate of evaporation from both sides of
Naphthalene a plate.
vapor
FIGURE P14–101
780
HEAT TRANSFER
14–113C During evaporation from a water body to air, under the emissivities of sheet metal and water to be 0.61 and 0.95,
what conditions will the latent heat of vaporization be equal to respectively.
convection heat transfer from the air? Answers: (a) 61,337 W, (b) 1480 W, (c) 3773 W, (d ) 79,960 W,
44.9 kg/h
14–114 Jugs made of porous clay were commonly used to
cool water in the past. A small amount of water that leaks out 14–118 Repeat Problem 14–117 for a water bath temperature
keeps the outer surface of the jug wet at all times, and hot and of 50°C.
relatively dry air flowing over the jug causes this water to 14–119 One way of increasing heat transfer from the head on
evaporate. Part of the latent heat of evaporation comes from the a hot summer day is to wet it. This is especially effective in
water in the jug, and the water is cooled as a result. If the envi- windy weather, as you may have noticed. Approximating the
ronment conditions are 1 atm, 30°C, and 35 percent relative head as a 30-cm-diameter sphere at 30°C with an emissivity of
humidity, determine the temperature of the water when steady 0.95, determine the total rate of heat loss from the head at am-
conditions are reached. bient air conditions of 1 atm, 25°C, 40 percent relative humid-
ity, and 25 km/h winds if the head is (a) dry and (b) wet. Take
the surrounding temperature to be 25°C.
Answers: (a) 40.6 W, (b) 352 W
Water that
leaks out Wet
Evaporation 1 atm
Hot, dry air 30°C
25°C
30°C 40% RH
35% RH
25 km/h
FIGURE P14–114
781
CHAPTER 14
FIGURE P14–127
14–128 The diffusion coefficient of carbon in steel is
given as
782
HEAT TRANSFER
783
CHAPTER 14
and wind is blowing at an average velocity of 40 mph, deter- cern in cold climates as it undermines the effectiveness of the
mine the rate of heat loss from the top surface of the pond by insulation. Investigate how the builders in your area are coping
(a) forced convection, (b) radiation, and (c) evaporation. Take with this problem, whether they are using vapor retarders or
the average temperature of the surrounding surfaces to be 60°F. vapor barriers in the walls, and where they are located in the
14–141E Repeat Problem 14–140E for a solar pond surface walls. Prepare a report on your findings and explain the rea-
temperature of 90°F. soning for the current practice.
Answers: (a) 299,400 Btu/h, (b) 1,057,000 Btu/h, (c) 3,396,000 14–146 You are asked to design a heating system for a swim-
Btu/h ming pool that is 2 m deep, 25 m long, and 25 m wide. Your
client desires that the heating system be large enough to raise
Computer, Design, and Essay Problems the water temperature from 20°C to 30°C in 3 h. The heater
14–142 Write an essay on diffusion caused by effects other must also be able to maintain the pool at 30°C at the outdoor
than the concentration gradient such as thermal diffusion, pres- design conditions of 15°C, 1 atm, 35 percent relative humidity,
sure diffusion, forced diffusion, knodsen diffusion, and surface 40 mph winds, and effective sky temperature of 10°C. Heat
diffusion. losses to the ground are expected to be small and can be disre-
14–143 Write a computer program that will convert the mole garded. The heater considered is a natural gas furnace whose
fractions of a gas mixture to mass fractions when the molar efficiency is 80 percent. What heater size (in Btu/h input)
masses of the components of the mixture are specified. would you recommend that your client buy?
14–144 One way of generating electricity from solar energy 15°C
involves the collection and storage of solar energy in large ar- 1 atm Heat
tificial lakes of a few meters deep, called solar ponds. Solar en- Evaporation 35% RH loss
ergy is stored at the bottom part of the pond at temperatures
close to boiling, and the rise of hot water to the top is prevented
by planting salt to the bottom of the pond. Write an essay on 30°C
the operation of solar pond power plants, and find out how Heating
much salt is used per year per m2. If the cost is not a factor, can Pool
fluid
sugar be used instead of salt to maintain the concentration gra-
dient? Explain.
14–145 The condensation and even freezing of moisture in
building walls without effective vapor retarders is a real con- FIGURE P14–146
Chapter 1
Psychrometry
Chapter 1
Psychrometry
1.1. Introduction:
Atmospheric air makes up the environment in almost every type of air
conditioning system. Hence a thorough understanding of the properties of
atmospheric air and the ability to analyze various processes involving air is
fundamental to air conditioning design.
Atmospheric air is a mixture of many gases plus water vapour and a number
of pollutants (Fig.1.1). The amount of water vapour and pollutants vary from place
to place. The concentration of water vapour and pollutants decrease with altitude,
and above an altitude of about 10 km, atmospheric air consists of only dry air. The
pollutants have to be filtered out before processing the air. Hence, what we process
is essentially a mixture of various gases that constitute air and water vapour. This
mixture is known as moist air.
The moist air can be thought of as a mixture of dry air and moisture. For all
practical purposes, the composition of dry air can be considered as constant. In
1949, a standard composition of dry air was fixed by the International Joint
Committee on Psychrometric data. It is given in Table 1.1.
Based on the above composition the molecular weight of dry air is found to be
28.966 and the gas constant R is 287.035 J/kg.K.
At a given temperature and pressure the dry air can only hold a certain
maximum amount of moisture. When the moisture content is maximum, then the air
is known as saturated air, which is established by a neutral equilibrium between the
moist air and the liquid or solid phases of water.
For calculation purposes, the molecular weight of water vapour is taken as 18.015
and its gas constant is 461.52 J/kg.K.
Atmospheric air
Dust particles,
fumes etc
After filtration
Mixture of permanent
gases (N2,O2,Ar,H2,…)
Water vapour
other pressures also. However, since in most cases the pressures involved are low,
one can apply the perfect gas model to estimate psychrometric properties.
According to the Gibbs-Dalton law for a mixture of perfect gases, the total
pressure exerted by the mixture is equal to the sum of partial pressures of the
constituent gases. According to this law, for a homogeneous perfect gas mixture
occupying a volume V and at temperature T, each constituent gas behaves as
though the other gases are not present (i.e., there is no interaction between the
gases). Each gas obeys perfect gas equation. Hence, the partial pressures exerted
by each gas, p1,p2,p3 … and the total pressure pt are given by:
nR T n R T n R T
p 1 = 1 u ; p 2 = 2 u ; p 3 = 3 u ......
V V V (1.1)
p t = p 1 + p 2 + p 3 + ......
p = pt = pa + pv (1.2)
Dry bulb temperature (DBT) is the temperature of the moist air as measured by a
standard thermometer or other temperature measuring instruments.
Saturated vapour pressure (psat) is the saturated partial pressure of water vapour at
the dry bulb temperature. This is readily available in thermodynamic tables and
charts. ASHRAE suggests the following regression equation for saturated vapour
pressure of water, which is valid for 0 to 100oC.
c
ln(p sat ) = 1 + c 2 + c 3 T + c 4 T 2 + c 5 T 3 + c 6 ln(T ) (1.3)
T
where psat = saturated vapor pressure of water in kilo Pascals
T = temperature in K
The regression coefficients c1 to c6 are given by:
Relative humidity (Φ) is defined as the ratio of the mole fraction of water vapour in
moist air to mole fraction of water vapour in saturated air at the same temperature
and pressure. Using perfect gas equation we can show that:
6
Humidity ratio (W): The humidity ratio (or specific humidity) W is the mass of water
associated with each kilogram of dry air 1 . Assuming both water vapour and dry air to
be perfect gases 2 , the humidity ratio is given by:
kg of water vapour p v V / R v T pv / R v
W= = = (1.5)
kg of dry air p a V / R a T (p t − p v ) / R a
Substituting the values of gas constants of water vapour and air Rv and Ra in
the above equation; the humidity ratio is given by:
pv
W = 0.622
pt − pv (1.6)
For a given barometric pressure pt, given the DBT, we can find the saturated vapour
pressure psat from the thermodynamic property tables on steam. Then using the
above equation, we can find the humidity ratio at saturated conditions, Wsat.
4030(DBT + 235 )
DPT = − 235 (1.7)
4030 − (DBT + 235 ) ln φ
where Φ is the relative humidity (in fraction). DBT & DPT are in oC. Of course, since
from its definition, the dew point temperature is the saturation temperature
corresponding to the vapour pressure of water vapour, it can be obtained from steam
tables or using Eqn.(1.3).
1 Properties such as humidity ratio, enthalpy and specific volume are based on 1 kg of dry air. This is useful as the total
mass of moist air in a process varies by the addition/removal of water vapour, but the mass of dry air remains constant.
2 Dry air is assumed to be a perfect gas as its temperature is high relative to its saturation
temperature, and water vapour is assumed to be a perfect gas because its pressure is low relative to its saturation
pressure. These assumptions result in accuracies, that are, sufficient for engineering calculations (less than 0.7 percent as
shown by Threlkeld). However, more accurate results can be obtained by using the data developed by Goff and Gratch in
1945.
7
Degree of saturation μ: The degree of saturation is the ratio of the humidity ratio W to
the humidity ratio of a saturated mixture Ws at the same temperature and pressure,
i.e.,
W (1.8)
μ=
Ws t ,P
Enthalpy: The enthalpy of moist air is the sum of the enthalpy of the dry air and the
enthalpy of the water vapour. Enthalpy values are always based on some reference
value. For moist air, the enthalpy of dry air is given a zero value at 0oC, and for water
vapour the enthalpy of saturated water is taken as zero at 0oC.
The enthalpy of moist air is given by:
The unit of h is kJ/kg of dry air. Substituting the approximate values of cp and hg, we
obtain:
h = 1.005 t + W (2501 + 1.88 t ) (1.10)
Humid specific heat: From the equation for enthalpy of moist air, the humid specific
heat of moist air can be written as:
c pm = c p + W.c pw (1.11)
Since the second term in the above equation (w.cpw) is very small compared
to the first term, for all practical purposes, the humid specific heat of moist air, cpm
can be taken as 1.0216 kJ/kg dry air.K
Specific volume: The specific volume is defined as the number of cubic meters of
moist air per kilogram of dry air. From perfect gas equation since the volumes
occupied by the individual substances are the same, the specific volume is also
equal to the number of cubic meters of dry air per kilogram of dry air, i.e.,
3 Though the water vapor in moist air is likely to be superheated, no appreciable error results if we assume it to be saturated.
This is because of the fact that the constant temperature lines in the superheated region on a Mollier chart (h vs s) are
almost horizontal.
8
R T R aT
v= a = m3 / kg dry air (1.12)
pa pt − pv
Lines of constant
Lineshumidity
relative of
constant RH Lines of
constant
sp.volume
Saturation curve
(RH = 100%)
Lines of
constant
W
enthalpy
(kgw/kgda)
DBT, oC
Fig.1.2: Schematic of a psychrometric chart for a given barometric pressure
1.3.1. Combined heat and mass transfer; the straight line law
The straight line law states that “when air is transferring heat and mass
(water) to or from a wetted surface, the condition of air shown on a psychrometric
chart drives towards the saturation line at the temperature of the wetted surface”.
For example, as shown in Fig.1.3, when warm air passes over a wetted
surface its temperature drops from 1 to 2. Also, since the vapor pressure of air at 1 is
greater than the saturated vapor pressure at tw, there will be moisture transfer from
air to water, i.e., the warm air in contact with cold wetted surface cools and
dehumidifies. According to the straight line law, the final condition of air (i.e., 2) lies
on a straight line joining 1 with tw on the saturation line. This is due to the value of
unity of the Lewis number, that was discussed in an earlier chapter on analogy
between heat and mass transfer.
water in the duct, there will be heat and mass transfer between water and air. If the
duct is infinitely long, then at the exit, there would exist perfect equilibrium between
air and water at steady state. Air at the exit would be fully saturated and its
temperature is equal to that of water temperature. The device is adiabatic as the
walls of the chamber are thermally insulated. In order to continue the process, make-
up water has to be provided to compensate for the amount of water evaporated into
the air. The temperature of the make-up water is controlled so that it is the same as
that in the duct.
h1 = h 2 − ( W2 − W1 )h f (1.13)
hfg,2 (1.14)
t 2 = t1 − ( w 2 − w 1)
cpm
where hfg,2 is the latent heat of vaporization at the saturated condition 2. Thus
measuring the dry bulb (t1) and wet bulb temperature (t2) one can find the inlet
humidity ratio (W1) from the above expression as the outlet saturated humidity ratio
(W2) and latent heat heat of vaporizations are functions of t2 alone (at fixed
barometric pressure).
On the psychrometric chart as shown in Fig.1.5, point 1 lies below the line of
constant enthalpy that passes through the saturation point 2. t2 = f(t1,W1) is not a
unique function, in the sense that there can be several combinations of t1 and W1
which can result in the same sump temperature in the adiabatic saturator. A line
passing through all these points is a constant wet bulb temperature line. Thus all
inlet conditions that result in the same sump temperature, for example point 1’ have
the same wet bulb temperature. The line is a straight line according to the straight-
line law. The straight-line joining 1 and 2 represents the path of the air as it passes
through the adiabatic saturator.
11
Perfect insulation
Make-up water
(W2-W1) per kgda
W2
W1
t2 t1
Wet wick
DBT
4 By performing energy balance across the wet-bulb, it can be shown that, the temperature measured
by the wet-bulb thermometer is:
for air-water mixtures, the ratio (hc/kwcpm) = Lewis number is ≈1, hence, the wick temperature is approximately
equal to the thermodynamic wet-bulb temperature.
It should be noted that, unlike thermodynamic WBT, the WBT of wet bulb thermometer is not a
thermodynamic property as it depends upon the rates of heat and mass transfer between the wick and air.
13
The units of all the pressures in the above equations should be consistent.
Once the vapor pressure is calculated, then all other properties such as
relative humidity, humidity ratio, enthalpy, humid volume etc. can be calculated from
the psychrometric equations presented earlier.
1.5. Psychrometer:
Any instrument capable of measuring the psychrometric state of air is called a
psychrometer. As mentioned before, in order to measure the psychrometric state of
air, it is required to measure three independent parameters. Generally two of these
are the barometric pressure and air dry-bulb temperature as they can be measured
easily and with good accuracy.
The sling psychrometer is widely used for measurements involving room air
or other applications where the air velocity inside the room is small. The sling
psychrometer consists of two thermometers mounted side by side and fitted in a
frame with a handle for whirling the device through air. The required air circulation (≈
3 to 5 m/s) over the sensing bulbs is obtained by whirling the psychrometer (≈ 300
RPM). Readings are taken when both the thermometers show steady-state readings.
The function of the wick on the wet-bulb thermometer is to provide a thin film
of water on the sensing bulb. To prevent errors, there should be a continuous film of
water on the wick. The wicks made of cotton or cloth should be replaced frequently,
and only distilled water should be used for wetting it. The wick should extend beyond
the bulb by 1 or 2 cms to minimize the heat conduction effects along the stem.
Chapter 2
Psychrometric Processes
2.1. Introduction:
In the design and analysis of air conditioning plants, the fundamental
requirement is to identify the various processes being performed on air. Once identified,
the processes can be analyzed by applying the laws of conservation of mass and
energy. All these processes can be plotted easily on a psychrometric chart. This is very
useful for quick visualization and also for identifying the changes taking place in
important properties such as temperature, humidity ratio, enthalpy etc. The important
processes that air undergoes in a typical air conditioning plant are discussed below.
During this process, the moisture content of air remains constant but its
temperature decreases as it flows over a cooling coil. For moisture content to remain
constant, the surface of the cooling coil should be dry and its surface temperature
should be greater than the dew point temperature of air. If the cooling coil is 100%
effective, then the exit temperature of air will be equal to the coil temperature. However,
in practice, the exit air temperature will be higher than the cooling coil temperature.
Figure 28.1 shows the sensible cooling process O-A on a psychrometric chart. The heat
transfer rate during this process is given by:
ho
hA
A O
W
DBT
Fig.2.1: Sensible cooling process O-A on psychrometric chart
Version 1 ME, IIT Kharagpur
b) Sensible heating (Process O-B):
During this process, the moisture content of air remains constant and its
temperature increases as it flows over a heating coil. The heat transfer rate during this
process is given by:
where cpm is the humid specific heat (≈1.0216 kJ/kg dry air) and ma is the mass flow
rate of dry air (kg/s). Figure 28.2 shows the sensible heating process on a
psychrometric chart.
hB
O B
DBT
Fig.2.2: Sensible heating process on psychrometric chart
When moist air is cooled below its dew-point by bringing it in contact with a cold
surface as shown in Fig.28.3, some of the water vapor in the air condenses and leaves
the air stream as liquid, as a result both the temperature and humidity ratio of air
decreases as shown. This is the process air undergoes in a typical air conditioning
system. Although the actual process path will vary depending upon the type of cold
surface, the surface temperature, and flow conditions, for simplicity the process line is
assumed to be a straight line. The heat and mass transfer rates can be expressed in
terms of the initial and final conditions by applying the conservation of mass and
conservation of energy equations as given below:
ma .w O = ma .w C + mw (2.3)
hO
Cooling coil hw
hC O
Wo
ma ma Wc
C
ho hC
Wo WC
Qt mw
Ts TC To
Fig.2.3: Cooling and dehumidification process (O-C)
m a .h O = Q t + m w .h w + m a .h C (2.4)
from the above two equations, the load on the cooling coil, Qt is given by:
Q t = m a (h O − h C ) − m a ( w O − w C )h w (2.5)
the 2nd term on the RHS of the above equation is normally small compared to the other
terms, so it can be neglected. Hence,
Q t = m a (h O − h C ) (2.6)
It can be observed that the cooling and de-humidification process involves both latent
and sensible heat transfer processes, hence, the total, latent and sensible heat transfer
rates (Qt, Ql and Qs) can be written as:
Q t = Ql + Q s
where Q l = m a (h O − h w ) = m a .h fg ( w O − w C ) (2.7)
Q s = m a (h w − h C ) = m a .c pm (TO − TC )
By separating the total heat transfer rate from the cooling coil into sensible and
latent heat transfer rates, a useful parameter called Sensible Heat Factor (SHF) is
defined. SHF is defined as the ratio of sensible to total heat transfer rate, i.e.,
Δw
tan c = (2.9)
ΔT
1 − SHF Q l m a h fg Δ w 2501Δ w Δw
= = = = 2451 (2.10)
SHF Q s m a c pm Δ T 1.0216 Δ T ΔT
1 ⎛ 1 − SHF ⎞
tan c = ⎜ ⎟ (2.11)
2451 ⎝ SHF ⎠
Thus we can see that the slope of the cooling and de-humidification line is purely
a function of the sensible heat factor, SHF. Hence, we can draw the cooling and de-
humidification line on psychrometric chart if the initial state and the SHF are known. In
some standard psychrometric charts, a protractor with different values of SHF is
provided. The process line is drawn through the initial state point and in parallel to the
given SHF line from the protractor as shown in Fig.2.4.
c
Wo-Wc
-h
ho
c
-h
hw
SH F
c
T − TS (2.12)
BPF = C
TO − TS
It can be easily seen that, higher the by-pass factor larger will be the difference
between air outlet temperature and the cooling coil temperature. When BPF is 1.0, all
the air by-passes the coil and there will not be any cooling or de-humidification. In
practice, the by-pass factor can be decreased by increasing the number of rows in
a cooling coil or by decreasing the air velocity or by reducing the fin pitch.
CF = 1 − BPF (2.13)
d) Heating and Humidification (Process O-D):
During winter it is essential to heat and humidify the room air for comfort. As
shown in Fig.2.5., this is normally done by first sensibly heating the air and then
adding water vapour to the air stream through steam nozzles as shown in the figure.
hD
Heating coil Steam nozzles
hO D
wD
ma ma
TO TD
wO wD wO
hO hD O
Qh mw
TO TD
Mass balance of water vapor for the control volume yields the rate at which steam has
to be added, i.e., mw:
mw = ma (wD − wO ) (2.14)
Q h = m a (hD − h O ) − m w h w (2.15)
where Qh is the heat supplied through the heating coil and hw is the enthalpy of steam.
Since this process also involves simultaneous heat and mass transfer, we can
define a sensible heat factor for the process in a way similar to that of a coolind and
dehumidification process.
As the name implies, during this process, the air temperature drops and its
humidity increases. This process is shown in Fig.2.6. As shown in the figure, this can
be achieved by spraying cool water in the air stream. The temperature of water should
be lower than the dry-bulb temperature of air but higher than its dew-point temperature
to avoid condensation (TDPT < Tw < TO).
ma ma wE
TO TE
wO wE wO
hO hE
Tw TDPT TE TO
It can be seen that during this process there is sensible heat transfer from air to
water and latent heat transfer from water to air. Hence, the total heat transfer depends
upon the water temperature. If the temperature of the water sprayed is equal to the wet-
bulb temperature of air, then the net transfer rate will be zero as the sensible heat
transfer from air to water will be equal to latent heat transfer from water to air. If the
water temperature is greater than WBT, then there will be a net heat transfer from water
to air. If the water temperature is less than WBT, then the net heat transfer will be from
air to water. Under a special case when the spray water is entirely recirculated and is
neither heated nor cooled, the system is perfectly insulated and the make-up water is
supplied at WBT, then at steady-state, the air undergoes an adiabatic saturation
process, during which its WBT remains constant. This is the process of adiabatic
saturation discussed in Chapter 1. The process of cooling and humidification is
encountered in a wide variety of devices such as evaporative coolers, cooling towers
etc.
Hygroscopic
material
O WO
O F
WF
F
TO TF
i) Without condensation: Figure 2.8 shows an adiabatic mixing of two moist air
streams during which no condensation of moisture takes place. As shown in the figure,
when two air streams at state points 1 and 2 mix, the resulting mixture condition 3 can
be obtained from mass and energy balance.
From the above equations, it can be observed that the final enthalpy and
humidity ratio of mixture are weighted averages of inlet enthalpies and humidity ratios.
A generally valid approximation is that the final temperature of the mixture is the
ma,1
ma,1+ma,2 = ma,3
ma,2
As shown in Fig.2.9, when very cold and dry air mixes with warm air at high
relative humidity, the resulting mixture condition may lie in the two-phase region, as a
result there will be condensation of water vapor and some amount of water will leave
the system as liquid water. Due to this, the humidity ratio of the resulting mixture (point
3) will be less than that at point 4. Corresponding to this will be an increase in
temperature of air due to the release of latent heat of condensation. This process rarely
occurs in an air conditioning system, but this is the phenomenon which results in the
formation of fog or frost (if the mixture temperature is below 0oC). This happens in
winter when the cold air near the earth mixes with the humid and warm air, which
develops towards the evening or after rains.
Eliminator
Plates
Make-up
water
Pump
Cooler/heater
Fig.2.10: Air washer
In the air washer, the mean temperature of water droplets in contact with air
decides the direction of heat and mass transfer. As a consequence of the 2nd law, the
heat transfer between air and water droplets will be in the direction of decreasing
temperature gradient. Similarly, the mass transfer will be in the direction of decreasing
vapor pressure gradient. For example,
a) Cooling and dehumidification: tw < tDPT. Since the exit enthalpy of air is less than its
inlet value, from energy balance it can be shown that there is a transfer of total energy
from air to water. Hence to continue the process, water has to be externally cooled.
Here both latent and sensible heat transfers are from air to water. This is shown
by Process O-A in Fig.2.10.
b) Adiabatic saturation: tw = tWBT. Here the sensible heat transfer from air to water is
exactly equal to latent heat transfer from water to air. Hence, no external cooling
or heating of water is required. That is this is a case of pure water recirculation.
This is shown by process O-B in Fig. 2.11. This the process that takes place in a
perfectly insulated evaporative cooler.
d) Cooling and humidification: tWBT < tw < tDBT. Here the sensible heat transfer is from air
to water and latent heat transfer is from water to air, but the total heat transfer is from
water to air, hence, water has to be heated externally. This is shown by Process O-D in
Fig.211. This is the process that takes place in a cooling tower. The air stream
extracts heat from the hot water coming from the condenser, and the cooled water is
sent back to the condenser.
e) Heating and humidification: tw > tDBT. Here both sensible and latent heat transfers are
from water to air, hence, water has to be heated externally. This is shown by Process
O-E in Fig.2.11.
Thus, it can be seen that an air washer works as a year-round air conditioning
system. Though air washer is a and extremely useful simple device, it is not commonly
used for comfort air conditioning applications due to concerns about health resulting
from bacterial or fungal growth on the wetted surfaces. However, it can be used in
industrial applications.
D E
B
W
C
O
DBT
Fig.2.11: Various psychrometric processes that can take place in an air washer
The sensible (QS) and latent (QL) heat transfer rates are given by:
Q S = h C A S (t i − t a )
. (2.18)
Q L = m w .h fg = hD .A S ( w i − w a ).h fg
Q T = Q S + Q L = hC A S (t i − t a ) + hD .A S ( w i − w a ).h fg (2.19)
Since the transport mechanism that controls the convective heat transfer
between air and water also controls the moisture transfer between air and water, there
exists a relation between heat and mass transfer coefficients, hc and hD as discussed in
an earlier chapter. It has been shown that for air-water vapor mixtures,
hC hc
hD ≈ or = Lewis number ≈ 1.0 (2.20)
c pm hD .c pm
h A
[
Q T = Q S + Q L = C S ( t i − t a ) + ( w i − w a ).h fg
c pm
] (2.21)
h A
Q T = Q S + Q L = C S [(hi − h a )] (2.22)
c pm
thus the total heat transfer and its direction depends upon the enthalpy difference (or
potential) between water and air (hi-ha).
if hi > ha; then the total heat transfer is from water to air and water gets cooled
if hi < ha; then the total heat transfer is from air to water and water gets heated
if hi = ha; then the net heat transfer is zero, i.e., the sensible heat transfer rate is equal
to but in the opposite direction of latent heat transfer. Temperature of water remains at
its wet bulb temperature value
mrc
Cooling coil
Supply fan ms=mo+ mrc
mo
By-pass
Slight discomfort
Lethal Lethal
Neutral condition
Studies show that the metabolic rate can be correlated to the rate of
respiratory oxygen consumption and carbon dioxide production. Based on this
empirical equations have been developed which relate metabolic rate to O2
consumption and CO2 production.
Since the metabolic rate is specified per unit area of the human body
(naked body), it is essential to estimate this area to calculate the total metabolic
rate. Even though the metabolic rate and heat dissipation are not uniform
throughout the body, for calculation purposes they are assumed to be uniform.
The total heat transfer rate from the skin Qsk is given by:
Q c = 14.8 V 0.5 (t b − t )
Q r = 11.603 (t b − t s ) (3.5)
0.4
Q e = 181.76 V (p s,b − p v )
In the above equation all the heat transfer rates are in watts, temperatures
are in oC and velocity is in m/s; ps,b and pv are the saturated pressure of water
vapour at surface temperature of the body and partial pressure of water vapour in
air, respectively, in kPa. From the above equations it is clear that the convective
heat transfer from the skin can be increased either by increasing the surrounding
air velocity (V) and/or by reducing the surrounding air DBT (t). The radiative heat
transfer rate can be increased by reducing the temperature of the surrounding
surfaces with which the body exchanges radiation. The evaporative heat transfer
rate can be increased by increasing the surrounding air velocity and/or by
reducing the moisture content of surrounding air.
The heat transfer rate due to respiration Qres is given by:
where Cres = Dry heat loss from respiration (sensible, positive or negative)
Eres = Evaporative heat loss from respiration (latent, always positive)
For comfort, the rate of heat stored in the body Qst should be zero, i.e.,
Since the body temperature depends on the heat balance, which in turn
depends on the conditions in the surroundings, it is important that the
surrounding conditions should be such that the body is able to maintain the
thermal equilibrium with minimum regulatory effort. All living beings have in-built
body regulatory processes against cold and heat, which to some extent
maintains the body temperatures when the external conditions are not favorable.
For example, human beings consist of a thermoregulatory system, which tries to
maintain the body temperature by initiating certain body regulatory processes
against cold and heat.
When the environment is colder than the neutral zone, then body loses
more heat than is generated. Then the regulatory processes occur in the
following order.
When the environment is hotter than the neutral zone, then body loses
less heat than is generated. Then the regulatory processes occur in the following
order.
1. Physiological factors such as age, activity, sex and health. These factors
influence the metabolic rate. It is observed that of these factors, the most
important is activity. Other factors are found to have negligible effect on thermal
comfort.
2. Insulating factor due to clothing. The type of clothing has strong influence on
the rate of heat transfer from the human body. The unit for measuring the
resistance offered by clothes is called as “clo”. 1 clo is equal to a resistance of
about 0.155 m2.K/W. Typical clo values for different types of clothing have been
estimated and are available in the form of tables. For example, a typical business
suit has a clo value of 1.0, while a pair of shorts has a clo value of about 0.05.
3. Environmental factors. Important factors are the dry bulb temperature, relative
humidity, air motion and surrounding surface temperature. Of these the dry bulb
temperature affects heat transfer by convection and evaporation, the relative
humidity affects heat loss by evaporation, air velocity influences both convective
and evaporative heat transfer and the surrounding surface temperature affects
the radiative heat transfer.
Apart from the above, other factors such as drafts, asymmetrical cooling
or heating, cold or hot floors etc. also affect the thermal comfort. The objective of
a comfort air conditioning system is to control the environmental factors so that
comfort conditions prevail in the occupied space. It has no control on the
physiological and insulating factors. However, wearing suitable clothing may help
in reducing the cost of the air conditioning system.
1
Tmrt 4 = T 4 + C V 2 (T − T ) (3.8)
g g a
where:
The derived indices combine two or more direct indices into a single
factor. Important derived indices are the effective temperature, operative
temperature, heat stress index, Predicted Mean Vote (PMV), Percent of People
Dissatisfied (PPD) etc.
Effective temperature (ET): This factor combines the effects of dry bulb
temperature and air humidity into a single factor. It is defined as the temperature
of the environment at 50% RH which results in the same total loss from the skin
as in the actual environment. Since this value depends on other factors such as
activity, clothing, air velocity and Tmrt, a Standard Effective Temperature (SET) is
defined for the following conditions:
Operative temperature (Top): This factor is a weighted average of air DBT and
Tmrt into a single factor. It is given by:
where hr and hc are the radiative and convective heat transfer coefficients and
Tamb is the DBT of air.
ASHRAE has defined a comfort chart based on the effective and operative
temperatures. Figure 3.3 shows the ASHRAE comfort chart with comfort zones
for summer and winter conditions. It can be seen from the chart that the comfort
zones are bounded by effective temperature lines, a constant RH line of 60% and
dew point temperature of 2oC. The upper and lower limits of humidity (i.e. 60 %
RH and 2oC DPT, respectively) are based on the moisture content related
considerations of dry skin, eye irritation, respiratory health and microbial growth.
The comfort chart is based on statistical sampling of a large number of occupants
with activity levels less than 1.2 met. On the chart, the region where summer and
winter comfort zones overlap, people in winter clothing feel slightly warm and
people in summer clothing feel slightly cool. Based on the chart ASHARE makes
the following recommendations:
Inside design conditions for Winter:
Summer zone
W
ET=22.5oC
ET=20oC
ET=26oC
DPT=2oC ET=23.5oC
Top
Fig.3.3: ASHRAE comfort chart for a sedentary person (activity ≈ 1.2 met)
where M is the metabolic rate and L is the thermal load on the body that is the
difference between the internal heat generation and heat loss to the actual
environment of a person experiencing thermal comfort. The thermal load has to
be obtained by solving the heat balance equation for the human body.
[
PPD=100 − 95 exp − (0.03353 PMV4 + 0.2179 PMV2 ) ] (3.11)
where dissatisfied refers to anybody not voting for –1, 0 or +1. It can be seen
from the above equation that even when the PMV is zero (i.e., no thermal load on
body) 5 % of the people are dissatisfied! When PMV is within ± 0.5, then PPD is
less than 10 %.
Where To and Tc are the outdoor and indoor comfort temperature in oC,
respectively. According to the above correlation, higher the outdoor temperature,
higher can be the indoor temperature. This is very important from energy
conservation point-of-view as air conditioning systems are very energy intensive,
and the load on an air conditioning plant can be reduced by maintaining the
indoor temperatures at as high a value as is allowed from thermal comfort point-
of-view.
1. Supply air conditions (air flow rate, DBT, humidity ratio & enthalpy)
2. Coil specifications (Latent and sensible loads on coil, coil ADP & BPF)
In this chapter fixing of supply air conditions and coil specifications for
summer air conditioning systems are discussed. Since the procedure is similar for
winter air conditioning system, the winter air conditioning systems are not discussed
here.
In this simple system, there is no outside air and the same air is recirculated
as shown in Fig.4.1. Figure 4.2 also shows the process on a psychrometric chart. It
can be seen that cold and dry air is supplied to the room and the air that leaves the
condition space is assumed to be at the same conditions as that of the conditioned
space. The supply air condition should be such that as it flows through the
conditioned space it can counteract the sensible and latent heat transfers taking
place from the outside to the conditioned space, so that the space can be maintained
at required low temperature and humidity. Assuming no heat gains in the supply and
return ducts and no energy addition due to fans, and applying energy balance across
the room; the Room Sensible Cooling load (Qs,r), Room Latent Cooling Load (Ql,r)
and Room Total Cooling load (Qt,r) are given by:
From cooling load calculations, the sensible, latent and total cooling loads on
the room are obtained. Hence one can find the Room Sensible Heat Factor (RSHF)
from the equation:
ti,Wi,hi
RSHF line
Return
ti,Wi,hi
Fan
i
ms
s
ms
ts,Ws,hs
ADP ts ti
Supply Fan
Qs,c, Ql,c Cooling coil
From the RSHF value one can calculate the slope of the process undergone
by the air as it flows through the conditioned space (process s-i) as:
1 ⎛ 1 − RSHF ⎞
slope of process line s − i, tan θ = ⎜ ⎟ (4.5)
2451 ⎝ RSHF ⎠
Since the condition i is known say, from thermal comfort criteria, knowing the
slope, one can draw the process line s-i through i. The intersection of this line with
the saturation curve gives the ADP of the cooling coil as shown in Fig.4.1. It should
be noted that for the given room sensible and latent cooling loads, the supply
condition must always lie on this line so that the it can extract the sensible and
latent loads on the conditioned space in the required proportions.
Since the case being considered is one of 100 % re-circulation, the process
that the air undergoes as it flows through the cooling coil (i.e. process i-s) will be
exactly opposite to the process undergone by air as it flows through the room
(process s-i). Thus, the temperature and humidity ratio of air decrease as it flows
through the cooling coil and temperature and humidity ratio increase as air flows
through the conditioned space. Assuming no heat transfer due to the ducts and fans,
the sensible and latent heat transfer rates at the cooling coil are exactly equal to the
sensible and latent heat transfer rates to the conditioned space; i.e.,
If the by-pass factor (X) of the cooling coil is known, then, from room
conditions, coil ADP and by-pass factor, the supply air temperature ts is obtained
using the definition of by-pass factor as:
⎛ t − t ADP ⎞
X = ⎜⎜ s ⎟⎟ ⇒ t s = t ADP + X(t i − t ADP ) (4.7)
⎝ t i − t ADP ⎠
Once the supply temperature ts is known, then the mass flow rate of supply air
is obtained from Eqn.(4.1) as:
Q s,r Q s,r
ms = = (4.8)
Cpm (t i − t s ) Cpm (t i − t ADP )(1 − X)
From the mass flow rate of air and condition i, the supply air humidity ratio and
enthalpy are obtained using Eqns.(4.2) and (4.3) as:
Q l,r (4.9)
Ws = Wi −
m sh fg
Q t ,r
h s = hi − (4.10)
ms
From Eqn.(4.8), it is clear that the required mass flow rate of supply air
decreases as the by-pass factor X decreases. In the limiting case when the by-pass
factor is zero, the minimum amount of supply air flow rate required is:
Q s,r
m s,min =
Cpm (t i − t ADP )
Thus with 100 % re-circulated air, the room ADP is equal to coil ADP and the
load on the coil is equal to the load on the room.
Figure 4.2 shows the schematic of the summer air conditioning system with
outdoor air and the corresponding process on psychrometric chart, when the by-pass
factor X is zero. Since the sensible and latent cooling loads on the conditioned space
are assumed to be known from cooling load calculations, similar to the earlier case,
one can draw the process line s-i, from the RSHF and state i. The intersection of this
line with the saturation curve gives the room ADP. As shown on the psychrometric
chart, when the by-pass factor is zero, the room ADP is equal to coil ADP, which in
turn is equal to the temperature of the supply air. Hence from the supply temperature
one can calculate the required supply air mass flow rate (which is the minimum
required as X is zero) using the equation:
Q s, r Q s,r
ms = = (4.12)
Cpm (t i − t s ) Cpm (t i − t ADP )
From the supply mass flow rate, one can find the supply air humidity ratio
and enthalpy using Eqns.(4.9) and (4.10).
Ql r Qs,r
o
m me ti,Wi,hi
ti,Wi,hi
i mrc
s
ms=mo+mrc
ts,Ws,hs
mo
Supply Fan
ADP = ts ti tm to
Cooling coil
Qs,c, Ql,c
Fig.4.2: A summer air conditioning system with outdoor air for ventilation and a
zero by-pass factor
From mass balance of air;
m s = mrc + m o (4.13)
Where mrc is the re-circulated air flow rate and mo is the outdoor air flow rate.
Since either mo or the ratio mo : mrc are specified, one can calculate the amount of
re-circulated air from Eqn.(4.13).
Calculation of coil loads:
From energy balance across the cooling coil; the sensible, latent and total
heat transfer rates, Qs,c, Ql,c and Qt,c at the cooling coil are given by:
Q s, c = m s Cpm (t m − t s )
Q l, c = m sh fg ( Wm − Ws ) 4.14)
Q t , c = Q s, c + Q l, c = m s (hm − h s )
Where ‘m’ refers to the mixing condition which is a result of mixing of the recirculated
air with outdoor air. Applying mass and energy balance to the mixing process one
can obtain the state of the mixed air from the equation:
mo Wm − Wi hm − hi t m − t i
= = ≈ (4.15)
m s Wo − Wi ho − hi t o − t i
Since (mo/ms) > 0, from the above equation it is clear that Wm > Wi, hm > hi
and tm > ti. This implies that ms(hm - hs) > ms(hi - hs), or the load on the cooling coil is
greater than the load on the conditioned space. This is of course due to the fact that
during mixing, some amount of hot and humid air is added and the same amount of
relative cool and dry air is exhausted (mo = me).
From Eqn.(4.1) to (4.3) and (4.14), the difference between the cooling
load on the coil and cooling load on the conditioned space can be shown to be equal
to:
Q s, c − Q s,r = m o Cpm (t o − t i )
Q l, c − Q l,r = m o h fg ( Wo − Wi ) (4.16)
Q t , c − Q t ,r = m o (h o − hi )
From the above equation it is clear that the difference between cooling coil
and conditioned space increases as the amount of outdoor air (mo) increases and/or
the outdoor air becomes hotter and more humid.
The line joining the mixed condition ‘m’ with the coil ADP is the process line
undergone by the air as it flows through the cooling coil. The slope of this line
depends on the Coil Sensible Heat Factor (CSHF) given by:
Q s, c Q s, c
CSHF = = (4.17)
Q s, c + Q l,c Q t ,c
For actual cooling coils, the by-pass factor will be greater than zero, as a
result the air temperature at the exit of the cooling coil will be higher than the coil
ADP. This is shown in Fig.4.3 along with the process on psychrometric chart. It can
be seen from the figure that when X > 0, the room ADP will be different from the coil
ADP. The system shown in Fig.4.3 is adequate when the RSHF is high ( > 0.75).
A/C Room
me= mo i ti, Wi, hi
mrc
Cooling coil
s
Supply fan ms=mo+ mrc
m
mo
o
Supply duct losses
Qt,c=Qs,c+Ql,c By-pass
i
s
ADP ts ti tm to
Fig.4.3: A summer air conditioning system with outdoor air for ventilation and a non-
zero by-pass factor
Normally in actual systems, either the supply temperature (ts) or the
temperature rise of air as it flows through the conditioned space (ti-ts) will be
specified. Then the step-wise procedure for finding the supply air conditions and the
coil loads are as follows:
i. Since the supply temperature is specified one can calculate the required supply air
flow rate and supply conditions using Eqns. (4.8) to (4.10).
ii. Since conditions ‘i’, supply air temperature ts and RSHF are known, one can draw
the line i-s. The intersection of this line with the saturation curve gives the room ADP.
iii. Condition of air after mixing (point ‘m’) is obtained from known values of ms and
mo using Eqn.(4.15).
iv. Now joining points ‘m’ and ‘s’ gives the process line of air as it flows through the
cooling coil. The intersection of this line with the saturation curve gives the coil ADP.
It can be seen that the coil ADP is lower than the room ADP.
vi. From points ‘m’, ‘s’ and coil ADP, the by-pass factor of the cooling coil can be
calculated.
If the coil ADP and coil by-pass factor are given instead of the supply air
temperature, then a trial-and-error method has to be employed to obtain the supply
air condition.
When the latent load on the building is high due either to high outside
humidity or due to large ventilation requirements (e.g. hospitals) or due to high
internal latent loads (e.g. presence of kitchen or laundry), then the simple system
discussed above leads to very low coil ADP. A low coil ADP indicates operation of
the refrigeration system at low evaporator temperatures. Operating the system at low
evaporator temperatures decreases the COP of the refrigeration system leading to
higher costs. Hence a reheat coil is sometimes used so that the cooling coil can be
operated at relatively high ADP, and at the same time the high latent load can also
be taken care of. Figure 4.4 shows an air conditioning system with reheat coil along
with the psychrometric representation of the process. As shown in the figure, in a
system with reheat coil, air is first cooled and dehumidified from point ‘m’ to point ’c’
in the cooling coil and is then reheated sensibly to the required supply temperature ts
using the reheat coil. If the supply temperature is specified, then the mass flow rate
and state of the supply air and condition of the air after mixing can be obtained using
equations given above. Since the heating process in the reheat coil is sensible, the
process line c-s will be horizontal. Thus if the coil ADP is known, then one can draw
the coil condition line and the intersection of this line with the horizontal line drawn
from supply state ‘s’ gives the condition of the air at the exit of the cooling coil. From
this condition, one can calculate the load on the cooling coil using the supply mass
flow rate and state of air after mixing. The capacity of the reheat coil is then obtained
from energy balance across it.
i
i
m c s
o
Reheat coil
Cooling coil
o
m
i
c s
Fig.4.4: A summer air conditioning system with reheat coil for high latent cooling
load applications
Advantages and disadvantages of reheat coil:
Thus the actual benefit of reheat coil depends may vary from system.
ii. The cooling coil should have 2 to 6 rows for moderate climate and 6 to 8 rows in
hot and humid climate. The by-pass factor of the coil varies from 0.05 to 0.2. The by-
pass factor decreases as the number of rows increases and vice versa. The fin pitch
and air velocity should be suitable.
iii. If chilled water is used for cooling and dehumidification, then the coil ADP will be
higher than about 4oC.
Chapter 5
Evaporative, Winter and All
Year Air Conditioning
Systems
5.1. Introduction to evaporative air conditioning systems:
Summer air conditioning systems capable of maintaining exactly the required
conditions in the conditioned space are expensive to own and operate. Sometimes,
partially effective systems may yield the best results in terms of comfort and cost.
Evaporative air conditioning systems are inexpensive and offer an attractive
alternative to the conventional summer air conditioning systems in places, which are
hot and dry. Evaporative air conditioning systems also find applications in hot
industrial environments where the use of conventional air conditioning systems becomes
prohibitively expensive.
Evaporative cooling has been in use for many centuries in countries such as
India for cooling water and for providing thermal comfort in hot and dry regions. This
system is based on the principle that when moist but unsaturated air comes in
contact with a wetted surface whose temperature is higher than the dew point
temperature of air, some water from the wetted surface evaporates into air. The latent
heat of evaporation is taken from water, air or both of them. In this process, the air
loses sensible heat but gains latent heat due to transfer of water vapour. Thus the
air gets cooled and humidified. The cooled and humidified air can be used for
providing thermal comfort.
(t o − t s )
ε= (5.1)
( t o − t o, wbt )
Exhaust air(i)
Conditioned Qt = Qs+Ql
space
Water spray or
wetted surface
Air washer
Supply air (s)
OD air(o)
Blower
Filter
Water pump
i
s w
to,wbt t ts to
.
The amount of supply air required ms can be obtained by writing
energy balance equation for the conditioned space, i.e.,
. Qt
ms = (5.2)
(hi − h s )
where Qt is the total heat transfer rate (sensible + latent) to the building, hi and
hs are the specific enthalpies of return air and supply air, respectively.
(o’)
Primary
air (o) Blower Air-to-air heat
exchanger
Secondary
Water Exhaust air (o)
pump air(e)
e
o’
o
s
t
Fig.5.2: An indirect, evaporative cooling system
Secondary
air out (2)
e
o’
2
1 o
t2 to’ t
Fig.5.3: A two-stage evaporative cooling system
2 s
w
i
m
1
o
t
Fig.5.4: A winter air conditioning system with a pre-heater
The humidification of air can be achieved in several ways, e.g. by bringing the
air in contact with a wetted surface, or with droplets of water as in an air
washer, by adding aerosol sized water droplets directly to air or by direct
addition of dry saturated or superheated steam. Humidification by direct
contact with a wetted surface or by using an air washer are not recommended
for comfort applications or for other applications where people are present in
the conditioned space due to potential health hazards by the presence of
micro-organisms in water. The most common method of humidifying air for
these applications is by direct addition of dry steam to air. When air is
humidified by contact with wetted surface as in an air washer, then
temperature of air decreases as its humidity increases due to simultaneous
transfer of sensible and latent heat. If the air washer functions as an adiabatic
saturator, then humidification proceeds along the constant wet bulb
temperature line. However, when air is humidified by directly adding dry,
saturated steam, then the humidification proceeds close to the constant dry
bulb temperature line. The final state of air is always obtained by applying
conservation of mass (water) and conservation of energy equations to the
humidification process.
.
Q s = m s c pm (t s − t i ) (5.3)
.
Q l = m s h fg ( w s − w i ) (5.4)
.
where ms is the mass flow rate of supply air, cpm is the specific heat of air, hfg
is the latent heat of vapourization of water, ws and wi are the supply and return
air humidity ratios and ts, ti are the supply and return temperatures of air. By
applying mass and/or energy balance equations across individual
components, the amount of sensible heat transfer rate to the pre-heater and
re-heater and the amount of moisture to be added in the humdifier can easily
be calculated.
Figure 5.5 shows another scheme that can also be used for heating
and humidification of air as required in a winter air conditioning system. As
shown in the figure, this system does not consist of a pre-heater. The mixed
air is directly humidified using an air washer (m-1) and is then reheated (1-s)
before supplying it to the conditioned space. Though this system is simpler
compared to the previous one, it suffers from disadvantages such as
possibility of water freezing in the air washer when large amount of cold
outdoor air is used and also from health hazards to the occupants if the water
used in the air washer is not clean. Hence this system is not recommended
for comfort conditioning but can be used in applications where the air
temperatures at the inlet to the air washer are above 0oC and the conditioned
space is used for products or processes, but not for providing personnel
comfort.
Qs,Ql
Return air
Exhaust air(i)
Conditioned
space
Recirculated
air(i)
1
s
w
i
m
t
Fig.5.5: A winter air conditioning system without a pre-heater
Exhaust
Return air
Cooling/
Conditioned Heating
Recirculated
air space load
Supply air
OD air
B
F RH
H CC Hu
1. A duct system that includes a supply air duct, return air duct, cooling and/or
heating coils, humidifiers/dehumidifiers, air filters and dampers
2. An air distribution system comprising various types of outlets for supply air and
inlets for return air
3. Supply and return air fans which provide the necessary energy to move the air
throughout the system
p 1 V12 p V2 p
+ + z 1 = 2 + 2 + z 2 = T = total head (7.1)
ρg 2g ρg 2g ρg
p V2
where is the pressure head, is the velocity head and Z is the static head,
ρg 2g
respectively. Each of the heads has units of length as explained before. The above
equation can be written in terms of static, velocity, datum and total pressures as:
ρV12 ρV22
p1 + + ρgz 1 = p 2 + + ρgz 2 = p T = total pressure (7.2)
2 2
The above equation implies that for frictionless flow through a duct, the total
pressure remains constant along the duct. Since all real fluids have finite
viscosity, i.e. in all actual fluid flows, some energy will be lost in overcoming friction.
This is referred to as head loss, i.e. if the fluid were to rise in a vertical pipe it will
rise to a lower height than predicted by Bernoulli’s equation. The head loss will
cause the total pressure to decrease in the flow direction. If the head loss is denoted
by Hl, then Bernoulli’s equation can be modified to:
p 1 V12 p V2
+ + z 1 = 2 + 2 + z 2 + Hl (7.3)
ρg 2g ρg 2g
To overcome the fluid friction and the resulting head, a fan is required in air
conditioning systems. When a fan is introduced into the duct through which air is
flowing, then the static and total pressures at the section where the fan is located
rise. This rise is called as Fan Total Pressure (FTP). Then the required power input
to the fan is given by:
.
Q .FTP
Wfan = air (7.4)
η fan
The FTP should be such that it overcomes the pressure drop of air as it flows
through the duct and the air finally enters the conditioned space with sufficient
momentum so that a good air distribution can be obtained in the conditioned space.
Evaluation of FTP is important in the selection of a suitable fan for a given
application. It can be easily shown that when applied between any two sections 1
and 2 of the duct, in which the fan is located, the FTP is given by:
ρ(V2 2 − V12 )
FTP = (p 2 − p 1 ) + + ρg(z 2 − z 1 ) + ρgHl (7.5)
2g
Thus to evaluate FTP, one needs to know the static pressures at sections 1
and 2 (p1, p2), air velocities at 1 and 2 (V1, V2), datum at 1 and 2 (Z1, Z2) and the
head loss Hl. Normally, compared to the other terms, the pressure change due to
datum ρg(z 2 − z1) is negligible. If the static pressures at the inlet and exit are equal,
say, to atmospheric pressure (p1= p2 = patm) and the duct has a uniform cross section
(v1=v2), then FTP is equal to the pressure loss due to friction. Thus to find FTP, one
has to estimate the total pressure loss as air flows through the duct from one section
to other.
7.4. Estimation of pressure loss in ducts:
As air flows through a duct its total pressure drops in the direction of flow. The
pressure drop is due to:
1. Fluid friction
Δ p t = Δ p f + Δp d (7.6)
L ⎛ ρV 2 ⎞
Δp f = f ⎜ ⎟
D ⎜⎝ 2 ⎟ (7.7)
⎠
where f is the dimensionless friction factor, L is the length of the duct and D is the
diameter in case of a circular duct and hydraulic diameter in case of a non-circular
⎛ ρVD ⎞
duct. The friction factor is a function of Reynolds number, Re D = ⎜⎜ ⎟⎟ and the
⎝ μ ⎠
relative surface roughness of the pipe or duct surface in contact with the fluid.
For turbulent flow, the friction factor can be evaluated using the empirical
correlation suggested by Colebrook and White is used, the correlation is given by:
1 ⎡ k 2.51 ⎤
= − 2log10 ⎢ s + ⎥ (7.8)
f ⎣ 3.7D (ReD ) f⎦
where ks is the average surface roughness of inner duct expressed in same units as
the diameter D. Evaluation of f from the above equation requires iteration since f
occurs on both the sides of it.
In general in air conditioning ducts, the fluid flow is turbulent. It is seen from
the above equation that when the flow is turbulent, the friction factor is a function of
Reynolds number, hydraulic diameter and inner surface roughness of the
duct material. Table 7.1 shows absolute roughness values of some of the
materials commonly used in air conditioning:
Material Absolute roughness , ε (m)
Galvanized Iron (GI) sheet 0.00015
Concrete 0.0003 to 0.003
Riveted steel 0.0009 to 0.009
Cast Iron (CI) 0.00026
Commercial steel 0.00046
Of the different materials, the GI sheet material is very widely used for air
conditioning ducts. Taking GI as the reference material and properties of air at
20oC and 1 atm. pressure, the frictional pressure drop in a circular duct is given by:
.
0.022243 Q air 1.852 L
Δp f = in N / m 2 (7.9)
4.973
D
.
where Q air is the volumetric flow rate of air in m3/s, L is the length and D is the inner
diameter of the duct in meters, respectively.
Using the above equation, friction charts have been created for estimation of
frictional pressure drop of standard air through circular ducts made of GI sheets.
Figure 7.1 shows the standard chart for estimating frictional pressure drop in
circular ducts made of GI sheets at standard air conditions.
⎛ Δp f ,1 ⎞ ⎛ ρ 1 ⎞ ⎛ Δp f ,1 ⎞ ⎛ T2 ⎞ 0.857
⎜ ⎟ =⎜ ⎟ and ⎜ ⎟ =⎜ ⎟ (7.10)
⎜ Δp f ,2 ⎟ ⎜⎝ ρ 2 ⎟⎠ ⎜ Δp f ,2 ⎟ ⎜⎝ T1 ⎟⎠
⎝ ⎠ ⎝ ⎠
The chart shown above is valid only for circular ducts. For other shapes, an
equivalent diameter has to be used to estimate the frictional pressure drop.
Even though circular ducts require the least material for a given flow rate and
allowable pressure drop, rectangular ducts are generally preferred in practice as
they fit easily into the building construction thus occupying less space, and they are
also easy to fabricate. The ratio of the two sides ‘a’ and ‘b’ of the rectangle (a/b) is
called as aspect ratio of the duct. Since square ducts with aspect ratio 1.0 come
close in performance to a circular duct, it is preferable to use an aspect ratio as close
to unity as possible for best performance.
One can use equation (7.9) and friction chart for circular ducts for estimating
pressure drop through a rectangular duct by using an equivalent diameter. A
rectangular duct is said to be equivalent to a circular duct, if the volumetric
.
flow rate Q air and frictional pressure drop per unit length (ΔPf/L) are same for
both. Equating these two parameters for a rectangular duct and an equivalent
circular duct, it can be shown that the equivalent diameter is given by:
(ab) 0.625
D eq = 1.3 (7.11)
(a + b) 0.25
The above equation is found to be valid for aspect ratio less than or equal
to 1:8. Thus from the known values of the two sides of the duct ‘a’ and ‘b’, one can
find the equivalent diameter Deq. From the equivalent diameter and the air flow rate,
one can estimate the frictional pressure drop per unit length by using either Eq.(7.9)
or the friction chart Fig. 7.1. However, when using equivalent diameter and flow rate
to find the frictional pressure drop from the chart, the velocity values shown on the
chart are not the actual velocities. The actual velocities have to be obtained from
the flow rate and the actual cross-sectional area of the rectangular duct. If a
rectangular duct has to be designed for a given flow rate and a given frictional
pressure drop, then one can first find the equivalent diameter from the friction chart
or from Eq.(7.9) and then find the required dimensions of the duct either by fixing
the aspect ratio or one of the sides.
7.5. Dynamic losses in ducts:
Dynamic pressure loss takes place whenever there is a change in either the
velocity or direction of airflow due to the use of a variety of bends and fittings in air
conditioning ducts. Some of the commonly used fittings are: enlargements,
contractions, elbows, branches, dampers etc. Since in general these fittings and
bends are rather short in length (< 1 m), the major pressure drop as air flows through
these fittings is not because of viscous drag (friction) but due to momentum change.
Pressure drop in bends and fittings could be considerable, and hence should be
evaluated properly. However, exact analytical evaluation of dynamic pressure drop
through actual bends and fittings is quite complex. Hence for almost all the cases,
the dynamic losses are determined from experimental data. In turbulent flows, the
dynamic loss is proportional to square of velocity. Hence these are expressed as:
ρV 2 (7.12)
Δp d = K
2
⎛ ρV 2 ⎞ ⎛ f.L eq ⎞⎛ ρV 2 ⎞
Δp d = K ⎜ ⎟=⎜ ⎟⎜ ⎟ (7.13)
⎜ 2 ⎟ ⎜ D eq ⎟⎜ 2 ⎟
⎝ ⎠ ⎝ ⎠⎝ ⎠
a) Turns, bends or elbows: The most common type of bends used in air
conditioning ducts are 90o turns shown in Fig. 7.2(a).
1’ 2
R2 W/H
1/4
Cb 1/2
R1 1
2
1
R1/R2
(a) Fig.7.2: Airflow through a 90o bend (elbow)
(b)
The cross-section of the elbow could be circular or rectangular. Weisbach
proposed that the dynamic pressure loss in an elbow is due to the sudden expansion
from the vena contracta region (1’) to full cross-section 2 as shown in Fig.7.2(a). The
dynamic pressure drop due to the elbow or 90o turn is found to be a function of the
aspect ratio (W/H), inner and outer radii of the turn (R1 and R2) and the velocity
pressure ρV2/2, i.e.,
⎛ ρV 2 ⎞ ⎛ ρV 2 ⎞
Δp d,b = Cb ⎜ ⎟ = f (( W / H), R 1 , R 2 )⎜ ⎟ (7.14)
⎜ 2 ⎟ ⎜ 2 ⎟
⎝ ⎠ ⎝ ⎠
Turning vanes
The equivalent lengths are available as function of geometry for other types of
turns and bends.
b ) Branch take-offs: Branch take-offs (Fig. 7.3) are commonly used in air
conditioning ducts for splitting the airflow into a branch and a downstream duct. The
dynamic pressure drop from the upstream (u) to downstream (d), Δpu-d is given by:
⎛ ρV 2 ⎞ ⎛ 2
⎜ d ⎟⎜ Vd ⎞
Δp u -d = 0.4 1− ⎟⎟ (7.15)
⎜ 2 ⎟ ⎜⎝ V ⎠
⎝ ⎠ u
where Vd and Vu are the air velocities in the downstream and upstream ducts,
respectively.
The dynamic pressure drop from the upstream (u) to branch (b), Δpu-b is given by:
⎛ ρV 2 ⎞
Δp u -b = C u -b ⎜ d ⎟ (7.16)
⎜ 2 ⎟
⎝ ⎠
d
u
β
b
The value of dynamic loss coefficient Cu-b is available in the form of tables and
graphs as a function of the angle β and the ratio of branch-to-upstream velocity,
Vb/Vu. Cu-b is found to increase as β and Vb/Vu increase.
c) Branch entries: Branch entries (Fig. 7.4) are commonly used in return air ducts.
Similar to branch take-offs, the values of dynamic pressure loss coefficients from
upstream-to-downstream (Cu-d) and from branch-to-downstream (Cu-d) are available
in the form of tables and graphs as functions of upstream, branch and downstream
velocities and the angle β.
d
β
b
Fig.7.4: A branch entry
⎛ ρV 2 ⎞ ⎛ 2
⎜ 1 ⎟⎜ A1 ⎞
Δp d, enl = 1− ⎟⎟ (7.17)
⎜ 2 ⎟⎜⎝ A ⎠
⎝ ⎠ 2
where V1 is the velocity before enlargement, and A1 and A2 are the areas before and
after enlargement, respectively. The above expression, which is obtained analytically
using modified Bernouille’s equation and momentum balance equation is found to
over-predict the pressure loss when the air flow rates are high and under-predict
when the flow rate is low. Correction factors are available in the form of tables for
different enlargements.
Vena contracta
2
1 2 1 1’
⎛ ρV 2 ⎞⎛ A 2 ⎛ ρV 2 ⎞⎛ 1
⎞ ⎞
Δp d, con = ⎜ 2 ⎟⎜ 2
− 1⎟⎟ = ⎜ 2 ⎟⎜⎜ − 1⎟⎟ (7.18)
⎜ 2 ⎟⎜⎝ A 1' ⎠ ⎜ 2 ⎟⎝ C c ⎠
⎝ ⎠ ⎝ ⎠
where V2 is the velocity in the downstream, and A1’ and A2 are the areas at vena
contracta and after contraction, respectively. The coefficient Cc is known as
contraction coefficient and is seen to be equal to area ratio A1’/A2. The contraction
coefficient Cc is found to be a function of the area ratio A2/A1, and the values of Cc as
obtained by Weisbach are shown in Table 7.3.
A2/A1 Cc
0.1 0.624
0.5 0.681
0.8 0.813
1.0 1.000
Filters, cooling and heating coils, dampers etc.: The pressure drop across air
handling unit equipment, such as, air filters, dampers, cooling and heating coils
depend on several factors. Hence, normally these values have to be obtained from
the manufacturer’s data.
Δp s (p s,2 − p s,1 )
R= = (7.22)
Δp v (p v,1 − p v,2 )
Thus for ideal enlargement the Static Regain Factor R is equal to 1.0, whereas it is
less than 1.0 for non-ideal enlargement.
1 2
pt
ps
p
pv
1 2
pt
ps
p
pv
L
High pressure systems: Velocity > 10 m/s, static pressure 15<ps ≤ 25 cm H2O (g)
1. Smaller ducts and hence, lower initial cost and lower space requirement
Recommended air velocities depend mainly on the application and the noise
criteria. Typical recommended velocities are:
Due to the several issues involved, the design of an air conditioning duct
system in large buildings could be a sophisticated operation requiring the use of
Computer Aided Design (CAD) software. However, the following methods are
most commonly used for simpler lay-outs such as the one shown in Fig.8.1.
1. Velocity method
2. Equal Friction Method
3. Static Regain method
4
H
5
G I
A C D
B
FAN
1 E
F
2
3
Fig.8.1: Typical air conditioning duct lay-out
ii. Find the diameters of main and branch ducts from airflow rates and velocities
for circular ducts. For rectangular ducts, find the cross-sectional area from flow
rate and velocity, and then by fixing the aspect ratio, find the two sides of the
rectangular duct
iii. From the velocities and duct dimensions obtained in the previous step, find the
frictional pressure drop for main and branch ducts using friction chart or equation.
iv. From the duct layout, dimensions and airflow rates, find the dynamic pressure
losses for all the bends and fittings
v. Select a fan that can provide sufficient FTP for the index run
vi. Balancing dampers have to be installed in each run. The damper in the index
run is left completely open, while the other dampers are throttled to reduce the
flow rate to the required design values.
The velocity method is one of the simplest ways of designing the duct
system for both supply and return air. However, the application of this method
requires selection of suitable velocities in different duct runs, which requires
experience. Wrong selection of velocities can lead to very large ducts, which,
occupy large building space and increases the cost, or very small ducts which
lead to large pressure drop and hence necessitates the selection of a large fan
leading to higher fan cost and running cost. In addition, the method is not very
efficient as it requires partial closing of all the dampers except the one in the
index run, so that the total pressure drop in each run will be same.
For example, let the duct run A-C-G-H be the index run and the total
pressure drop in the index run is 100 Pa. If the pressure drop in the shortest duct
run (say A-B) is 10 Pa, then the damper in this run has to be closed to provide an
additional pressure drop of 90 Pa, so that the required airflow rate to the
conditioned zone 1 can be maintained. Similarly the dampers in the other duct
runs also have to be closed partially, so that the total pressure drop with damper
partially closed in each run will be equal to the pressure drop in the index run
with its damper left open fully.
In this method the frictional pressure drop per unit length in the main and
branch ducts (Δpf/L) are kept same, i.e.,
(8.3)
Then the stepwise procedure for designing the duct system is as follows:
i. Select a suitable frictional pressure drop per unit length (Δpf/L) so that the
combined initial and running costs are minimized.
ii. Then the equivalent diameter of the main duct (A) is obtained from the
selected value of (Δpf/L) and the airflow rate. As shown in Fig.8.1, airflow rate in
.
the main duct Q A is equal to the sum total of airflow rates to all the conditioned
zones, i.e.,
. . . . . . N . (8.2)
Q A = Q1 + Q 2 + Q 3 + Q 4 + Q 5 = ∑ Qi
i=1
From the airflow rate and (Δpf/L) the equivalent diameter of the main duct (Deq,A)
can be obtained either from the friction chart or using the frictional pressure drop
equation, i.e.,
⎛ ⎞
(14.973 )
⎜ . ⎟
⎜ 0.022243 Q A 1.852 ⎟
D eq, A = ⎜ ⎟
⎛ Δp f ⎞ (8.3)
⎜ ⎜ ⎟ ⎟
⎜ ⎝ L ⎠A ⎟
⎝ ⎠
iii. Since the frictional pressure drop per unit length is same for all the duct runs,
the equivalent diameters of the other duct runs, B to I are obtained from the
equation:
iv. If the ducts are rectangular, then the two sides of the rectangular duct of each
run are obtained from the equivalent diameter of that run and by fixing aspect
ratio as explained earlier. Thus the dimensions of the all the duct runs can be
obtained. The velocity of air through each duct is obtained from the volumetric
flow rate and the cross-sectional area.
v. Next from the dimensions of the ducts in each run, the total frictional pressure
drop of that run is obtained by multiplying the frictional pressure drop per unit
length and the length, i.e.,
⎛ Δp f ⎞ ⎛ Δp f ⎞
ΔPf , A = ⎜ ⎟ .L A ; ΔPf ,B = ⎜ ⎟ .L B ... (38.5)
⎝ L ⎠A ⎝ L ⎠B
vi. Next the dynamic pressure losses in each duct run are obtained based on the
type of bends or fittings used in that run.
vii. Next the total pressure drop in each duct run is obtained by summing up the
frictional and dynamic losses of that run, i.e.,
(8.6)
ΔPA = Δp f,A + Δp d,A ; ΔPB = Δp f,B + Δp d,B ...
viii. Next the fan is selected to suit the index run with the highest pressure loss.
Dampers are installed in all the duct runs to balance the total pressure loss.
This method is commonly used for high velocity systems with long duct
runs, especially in large systems. In this method the static pressure is maintained
same before each terminal or branch. The procedure followed is as given below:
ii. Velocities in each successive runs are reduced such that the gain in static
pressure due to reduction in velocity pressure equals the frictional pressure drop
in the next duct section. Thus the static pressure before each terminal or branch
is maintained constant. For example, Fig.8.2 shows a part of the duct run with
two sections 1 and 2 before two branch take-offs. The velocity at 1 is greater
than that at 2, such that the static pressure is same at 1 and 2. Then using the
static regain factor, one can write:
(
Δp f ,2 + Δp d,2 = R p v,1 − p v,2 ) (8.7)
where Δpf,2 and Δpd,2 are the frictional and dynamic losses between 1 and 2, and
pv,1 and pv,2 are the velocity pressures at 1 and 2 respectively.
ps,1=ps,2
iv. The procedure is followed in the direction of airflow, and the dimensions of the
downstream ducts are obtained.
v. As before, the total pressure drop is obtained from the pressure drop in the
longest run and a fan is accordingly selected.
Static Regain method yields a more balanced system and does not call for
unnecessary dampering. However, as velocity reduces in the direction of airflow,
the duct size may increase in the airflow direction. Also the velocity at the exit of
the longer duct runs may become too small for proper air distribution in the
conditioned space.
.
total pressure drop, ΔPt ∝ (Q) 2 (8.8)
.
or, total pressure drop, ΔPt = C(Q) 2 (8.9)
where C is the resistance offered by the duct system. Once the duct system is
designed and installed, the value of C is supposed to remain constant. However,
if the air filters installed in the duct become dirty and/or if the damper position is
altered, then the value of C changes. Thus variation of total pressure drop with
airflow rate is parabolic in nature as shown in Fig. 8.3. In this figure, the curve A
refers to the performance of the duct at design conditions, while curve B refers to
the performance under the conditions of a dirty filter and/or a higher damper
closure and curve C refers to the performance when the damper is opened more.
From the duct characteristic curve for constant resistance, one can write
.
Δp t ,1 (Q 1 ) 2
=
Δp t ,2 . (8.10)
(Q 2 ) 2
Thus knowing the total pressure drop and airflow rate at design condition
(say 1), one can obtain the total pressure drop at an off-design condition 2, using
the above equation.
B
Δpt A
.
Q
Fig.8.3: Variation of total pressure drop with flow rate for a given duct system
The fan laws are a group of relations that are used to predict the effect of
change of operating parameters of the fan on its performance. The fan laws are
valid for fans, which are geometrically and dynamically similar. The fan laws
have great practical use, as it is not economically feasible to test fans of all sizes
under all possible conditions.
Here the fan laws related to the density of air and the rotative speed of the
fan are considered. The effect of the size of the fan is important at the time of
designing the fan. For a given air conditioning system with fixed dimensions,
fittings etc. it can be easily shown that:
.
airflow rate, Q ∝ ω (8.11)
ρV 2
static pressure rise, Δp s ∝ (8.12)
2
. . . ⎛ ρV 2 ⎞
fan power input, W ∝ Q(Δp s ) + Q⎜ ⎟ (8.13)
⎜ 2 ⎟
⎝ ⎠
From the expression for fan power input (Eqn.(8.13)), it can be seen that
the 1st term on the RHS accounts for power input required for increasing the
static pressure of air and the 2nd term on RHS accounts for the power input
required to impart kinetic energy to air as it flows through the fan. Using the
above relations, the following fan laws can be obtained.
Law 1: Density of air ρ remains constant and the speed ω varies:
. .
Q ∝ ω; Δp s ∝ ω 2 and W ∝ ω 3 (8.14)
.
Law 2: Airflow rate Q remains constant and the density ρ varies:
. .
Q = cons tan t; Δp s ∝ ρ and W ∝ ρ (8.15)
Law 3: Static pressure rise Δps remains constant and density ρ varies:
. 1 1 . 1
Q∝ ; Δp s = cons tan t, ω ∝ and W ∝ (8.16)
ρ ρ ρ
Figure 8.4 shows the variation of FTP of a centrifugal fan (fan performance
curve) and variation of total pressure loss of a duct system (duct performance
curve) as functions of the airflow rate. As shown in the figure, the point of
intersection of the fan performance curve and the duct performance curve yield
the balance point for the combined performance of fan and duct system. Point 1
gives a balance point between the fan and duct system when the rotative speed
of fan is ω1. At this condition the airflow rate is Q1 and the total
Fan performance
Curve at ω1 Duct performance
curve
Δpt,
FTP
Fan performance
Curve at ω2
1
Δpt,1 ω1 > ω2
2
Δpt,2
Q2 Q1
.
Q
b) To avoid draft in the occupied zone. Draft is defined as the localized feeling of
cooling or warmth. Draft is measured above or below the controlled room condition
of 24.4 οC and an air velocity of 0.15 m/s at the center of the room. The effective
draft temperature (EDT) for comfort is given by:
where DBT is the local dry bulb temperature (in oC) and V is the local velocity (m/s).
For comfort, the EDT should be within –1.7oC to +1.1oC and the air velocity should
be less than 0.36 m/s.
⎛N ⎞
ADPI = ⎜ θ ⎟ x100 (9.2)
⎝ N ⎠
where N is the total number of locations at which observations have been made, and
Nθ is the number of locations at which the effective draft temperature is within –1.7oC
to +1.1oC.
The objective of air distribution system design is to select and place the
supply air diffusers in such a way that the ADPI approaches 100 percent. The ADPI
provides a rational way of selecting air diffusers. Studies show that the value of ADPI
depends very much on space cooling load per unit area. A large value of space
cooling load per unit area tends to reduce the value of ADPI.
9.1.2. Space Diffusion Effectiveness Factor (SDEF)
T − Ts
SDEF = ex (9.3)
Tr − Ts
where Tex is the temperature of the exhaust air, Ts is the supply air temperature and
Tr is the temperature of the room air (at the measuring point). A SDEF value of ≤ 1
implies that some amount of cold supply air has not mixed with the room air and is
leaving the conditioned space as exhaust. The space air distribution is considered to
be effective if SDEF ≥ 1.0.
Table 39.1 shows the recommended supply air velocities for diffusers. Since
the air velocity at the supply air outlet is normally much higher than 0.36 m/s and its
temperature is much lower than 24.4oC, it has to mix properly with the room air
before it reaches the occupancy level. This depends on the effective design of the air
distribution system.
a) Velocity of air at the inlet to the supply diffuser: Noise criteria to be observed
e) Room Geometry
f) Room surface temperature: Lower the surface temperature (e.g. with glass)
stronger are the natural convection currents.
h) Room turbulence
The exact prediction of velocity and temperature profiles inside the
conditioned space requires simultaneous solution of mass, momentum and energy
equations for the conditioned space. However in general this task is extremely
complicated due to the several factors that affect airflow and heat transfer inside the
conditioned space. However, a basic understanding of room air distribution requires
the understanding of, buoyancy effects, deflection of air streams and behaviour of
free-stream jets. Normally the location and type of return air grilles do not affect the
air distribution significantly.
Due to the buoyancy effects, a supply air stream that is cooler than the room
air will drop and supply air that is warmer than room air rises. However, from thermal
comfort point-of-view, it is important that the supply air stream does not strike at
occupancy level. Figure 9.1(a) shows the drop of a supply air jet that is cooler than
the room air.
Room
Beam
Room
Fig.9.1(a): Drop of a cool air jet Fig.9.1(b): Deflection of a cool air jet
⎛ ΔT ⎞
Vt 2 = gh⎜⎜ ⎟⎟ (9.4)
⎝ Tr ⎠
where ΔT is the difference between the local temperature of the fluid (Tf) element
and the room air (Tr), Tr is the room air temperature in K, g is the acceleration due to
gravity and h is the height. For equilibrium at a height H, the velocity of the fluid
element should be equal to the entrance velocity of supply air (Vo), i.e.,
Vt = Vo at equilibrium (9.5)
Then from Eqn.(9.1):
⎛ ΔT ⎞
Vo 2 = gH ⎜⎜ d ⎟⎟ (9.6)
⎝ Tr ⎠
where ΔTd is the difference between the temperature of air at supply outlet and the
room air.
The Archimedes number, Ar is then defined as:
gH ⎛ ΔTd ⎞
Ar = ⎜⎜ ⎟⎟ (9.7)
Vo 2 ⎝ T r ⎠
In the above expression for Archimedes number, H may be the height of the room or
the hydraulic diameter, Dh of the room given by:
4 WH
Dh = (9.8)
2( W + H)
where W and H are the width and height of the room, respectively. Archimedes
number conveniently combines the supply air velocity at the outlet, supply to room
temperature difference and the principle dimensions of the room- important factors
that define the air distribution in a room. Several studies show that the airflow pattern
in a room is largely dependent on the Archimedes number. The Archimedes number
can also be viewed as a ratio of Grashof number to the square of Reynolds number
(Ar ≈ Gr/Re2), thus combining the effects of natural convection due to buoyancy and
forced convection due to supply air jet. Archimedes number also affects the heat
transfer between the air inside the conditioned space and the surrounding surfaces.
To avoid cold drafts in the occupied zone, the Archimedes number should not
exceed a maximum value, which depends on the room dimensions. Table 9.2 shows
the maximum Archimedes number values as a function of W/H ratio.
9.2.2. Deflection:
When an air stream strikes a solid surface such as a concrete beam or a wall,
it deflects. Again from comfort criteria, it is essential to ensure that due to deflection,
the supply air does not strike the occupants before it is diffused. Figure 9.1(b) shows
the deflection of a supply air jet as it strikes a solid beam.
Blow or throw:
It is the vertical distance the air moves after leaving the supply outlet and
reaches the end of blow.
Figure 9.2 shows the meaning of drop and throw of free-stream jets.
Room
Entrainment ratio:
As the high velocity jet (called as primary air) leaves the supply air outlet, it
entrains some amount of room air (called as secondary air). Entrainment gives rise
to motion of room air. The entrainment ratio at a distance x from the supply outlet is
defined as the ratio of volumetric flow rate of air at x to the volumetric flow rate of air
at the supply air outlet (x=0), i.e.,
Qx
Entrainment ratio at x,R x = (9.9)
Qx = 0
Spread:
It is the angle of divergence of the air stream after it leaves the supply air
outlet as shown in Fig.9.3. The spread can be both horizontal as well as vertical.
Vanes are normally used in the supply air outlets. These vanes can be straight,
converging or diverging. Figure 9.3 shows the outlet with diverging vanes, for which
the horizontal spread is 60o as shown in the figure. For straight vanes and
converging the spread is equal to 19o both in horizontal and vertical directions.
Converging vanes yield a blow that is about 15% longer than that of straight vanes,
whereas for diverging vanes it is about 50% less than that of horizontal vanes.
Outlet vanes
Room
Supply
air
9.4.Circular jets:
An understanding of the principle of the simple circular jet can be used to
understand the characteristics of most of the commercial supply air diffusers and
grilles. Figure 9.4 shows the airflow pattern in a circular jet. As shown in the figure,
supply air leaves the outlet at a velocity Vo. The velocity decays as the jet enters the
room and entrains the room air. Figure 9.4 also shows the velocity profile. It can be
seen that the velocity of air varies as a function of distance, horizontal x from the
opening along the centerline and the radial distance from the centerline.
Using the mass and momentum balance equations to the circular jet, it has
been shown by Schlichting that the velocity profile for the circular jet is given by:
7.41 Vo A o
V (x, r ) =
[ (
x 1 + 57.5 r 2 / x 2 )] 2 (39.10)
where Vo is the velocity at the outlet, m/s; V(x,r) is the velocity of air in the jet at x
and r, and Ao is the cross-sectional area of the outlet. From the above equation it is
easy to predict that the air velocity in the circular jet decreases as x and r increase,
and as Ao and Vo decrease. Thus a jet sustains its velocity better as the velocity at
the supply outlet increases and/or the area of opening increases. One can also
deduce that since the velocity decreases with x and r, the jet spreads as it flows, so
that the mass of air is always conserved. And from momentum conservation, it can
be deduced that entrainment of room air takes place as the jet moves away from the
supply air outlet.
V
Vo centreline
From Eqn.(9.10), the velocity of air in the circular jet along the centerline (r= 0) is
found to be:
7.41 Vo A o (9.11)
V ( x , r = 0) =
x
From the above expression, the entrainment ratio Rx for the circular jet can be
written as:
∞
∫ V(x, r ).2πr.dr
Qx r =0 0.405 x
Rx = = = (9.12)
Q x =0 A o Vo Ao
Large circular openings are rarely used in actual air distribution systems as
they travel long distances before mixing with room air. As this can cause discomfort
to the occupants, normally diffusers are used in circular jets. These diffusers provide
rapid velocity decay and large entrainment.
2.40 Vo b ⎡ 2⎛ y ⎞⎤
V ( x , y) = ⎢1 − tanh ⎜ 7.67 x ⎟ ⎥ (9.13)
x ⎣ ⎝ ⎠⎦
where b is the width of the opening and y is the normal distance from the central
plane. A comparison between circular and rectangular jets shows that the centerline
velocity decreases more rapidly for a circular jet compared to a rectangular jet. The
rectangular jet entrains less air than a circular jet, as a result it decelerates more
slowly.
Fig. 9.5. Front view of a supply air grille with horizontal and vertical vanes
Ceiling diffusers: A ceiling diffuser consists of concentric rings or inner cones made
up of vanes arranged in fixed directions. Ceiling diffusers can be round, square or
rectangular in shape. Figure 9.6 (a) shows square and rectangular ceiling diffuser,
and Fig. 9.6(b) shows a perforated diffuser. A square diffuser is widely used for
supply air. In the diffusers the supply air is discharged through the concentric air
passages in all directions. The air distribution pattern can be changed by adjusting
the adjustable inner cones or the deflecting vanes. Ceiling diffusers are normally
mounted at the center of the conditioned space. Ceiling diffusers provide large
entrainment ratio and shorter throw, hence are suitable for higher supply air
temperatures and for conditioned spaces with low head space. Ceiling diffusers can
deliver more air compared to grilles and slot diffusers.
Slot
diffuser
Slot
diffuser
b) An integrated layout of light troffer, diffuser and return slots can be formed on
suspended ceilings
c) Improved aesthetics
d) A combination of light troffer and return slot reduces the space cooling load as
the return air absorbs a part of the heat emitted by the lights. However, they
should be designed such that the return air does not come in direct contact
with the tube so that deposition of dust on the fluorescent tube is prevented
Figure 39.8 shows a light troffer-diffuser slot that combines the light troffer, supply air
diffuser and return air slot.
In addition to the above air distribution devices, the floor mounted grilles
and diffusers, low-side wall diffusers, nozzle diffusers etc. are also used for
room air distribution.
9.7. Return air inlets:
Different types of return air inlets are used to return the space air to the air
handling unit. Requirements of return air inlets are:
Similar to supply air outlets, return air inlets can be classified as grilles, registers,
diffusers etc. In many commercial buildings the ceiling plenum is used as return air
plenum. In this case, return slots are used to draw the return air through the ceiling.
In return air inlets the air velocity decreases sharply as the distance from the inlet
increases. Based on noise criteria, the air velocity should be within 3 m/s if the return
air inlet is inside the occupied space and it should be less than 4 m/s if it is above the
occupied space.
a) Entrainment of room air to reduce the air temperature and velocity in the
occupied zone to acceptable levels
b) Reverse air stream in the occupied zone for an even velocity and temperature
distribution
The airflow pattern in the conditioned space is influenced mainly by the type and
location of supply air outlets. The high side outlets, ceiling diffusers and slot diffusers
are most commonly used in air conditioned buildings.
Figure 9.9 shows the airflow pattern using high side outlets installed on a high
sidewall for cooling and heating applications. As the air is discharged from the high
side outlet, due to surface effect (Coanda effect) the air jet tends to stick to the
ceiling as shown in the figure. For cooling applications, the cold supply air entrains
the room air and deflects downwards when it strikes the opposite wall. The reverse
air stream formed due to entrainment fills the occupied space as shown. If the throw
is longer than the length of the room and height of the opposite wall, then the air jet
is deflected by the opposite wall and the floor and enters the occupied zone with high
velocity. On the other hand if the throw is too small, then the air jet drops directly
into the occupied zone before it strikes the opposite wall. Thus both these i.e, a very
long or very short throw can cause draft. For heating, a stagnant zone may form as
shown due to buoyancy effect. However, if the throw is long, the reverse flow can
minimize the stagnant area during heating. For high sidewall outlet, the most suitable
location for return air inlet is on the ceiling outside the air jet as shown in the figure.
Cooling Cooling
Side view End view
Return air
inlet
Heating Cooling
Plan view
Fig.9.9: Airflow pattern using high side outlets for cooling and heating applications
Figure 9.10 shows the airflow pattern using ceiling diffusers for both cooling
and heating applications. It is seen that ceiling diffusers produce a shorter throw, a
lower and more even distribution of air velocity and a more even temperature in the
occupied zone when used for cooling. However, when used for heating it is seen that
a larger stagnant area is formed due to buoyancy effect. Ceiling diffusers are widely
used for conditioned spaces with limited ceiling height and are designed to have a
large entrainment ratio and are widely used in variable air volume systems.
Cooling
Side view Cooling
End view
Heating Cooling
Plan view
Fig.9.10: Airflow pattern using ceiling diffusers for cooling and heating applications
Figure 9.11 shows the airflow patters obtained using slot diffusers installed in
the ceiling in the perimeter and interior zones. The slot diffusers installed in the
perimeter zone discharge air vertically downwards and also in the horizontal
direction. Due to its better surface effect, the air jet remains in contact with the ceiling
for a longer period and the reverse air stream ensures uniformity of temperature and
velocity in the occupied zone. Due to their superior characteristics and better
aesthetics, slot diffusers are widely used in large office spaces with normal ceiling
heights and with VAV systems.
Fig.9.11: Airflow pattern for cooling using slot diffusers
Stratified mixing flow for summer cooling offers the following advantages:
a) Convective heat transfer from the hot roof is effectively blocked by the higher
temperature air in the stagnant upper zone thus reducing the building cooling
load
b) Location of the return air inlets affects the cooling load only when they are
located in the upper zone
However, due to considerably reduced airflow rates, the air distribution and IAQ
may get affected, especially when using with VAV systems. Better insulation and
sealing of the ducts may be required to reduce losses and prevent surface
condensation.
2. Shape, size and ceiling height of the building: Ceiling and slot diffusers are
ideal for buildings with limited ceiling height. For large buildings with large
ceiling heights, high side wall mounted outlets are recommended.
3. Volume flow rate per unit floor area: Sidewall outlets are limited to low
specific volume flow rates as they give rise to higher air velocities in the
occupied zone. Compared to slot diffusers, the ceiling diffusers can handle
efficiently a larger volumetric flow rates. Table 9.3 shows the specific volume
flow rate of different outlets
4. Volume flow rate per outlet: The volume flow rate per supply outlet depends
on the throw required to provide a satisfactory room air distribution. For linear
slot diffusers, the volume flow rate per unit length is important. Its value
normally lies between 23 to 62 L/s.m for linear slot diffusers. In a closed office
with a floor area of about 14 m2 and only one external wall, one ceiling diffuser
is normally sufficient
5. Throw: High side wall outlets have a longer throw than ceiling diffusers.
Square ceiling diffusers and circular ceiling diffusers have similar throw
6. Noise level
7. Total pressure drop: The total pressure loss of supply air as it flows through
a slot diffuser of 19 mm width is normally between 12 to 50 Pascals, whereas
it is between 5 to 50 Pascals for ceiling diffuser. Normally the pressure loss
across the supply outlet should not exceed 50 Pascals
8. Cost and Appearance: Finally the cost and appearance of the supply air
outlets also have to be considered depending upon the specific application
Type of outlet Specific volume flow rate Max. ACH for 3-m
L/s/m2 of floor area ceiling
Grilles 3.0 to 6.0 7
Slot diffuser 4.0 to 20.0 12
Perforated Panel 4.5 to 15.0 18
Ceiling diffuser 4.5 to 25.0 30