Académique Documents
Professionnel Documents
Culture Documents
ON
1
11 ME 354: HEAT TRANSFER (4 – 0 – 0 −4)
Course Objectives:
1. Develop the ability to apply energy balance to heat transfer systems to solve engineering problems.
2. Learn the ability to apply the definitions of conduction, convection and radiation heat transfer to
engineering systems.
3. Predict the thermal response of engineering systems to energy transfer mechanisms for transient
and steady state situations.
4. To have working knowledge of variety of engineering correlations to heat transfer analysis and
thermal design of engineering components and systems.
Course Contents
UNIT – I 12 Hours
Chapter 1: Mechanisms of Heat Transfer- Basic laws governing each mechanism; combined
mechanisms; Illustrative examples
Chapter 2: Conduction Basic Equations and Boundary Conditions: Conduction Basic Equations:
one dimensional conduction equation in rectangular, cylindrical and spherical coordinates; thermal
diffusivity;3-dimensional conduction equation in Cartesian coordinates; boundary conditions of first,
second and third kind; radiation boundary condition ;illustrative examples on formulation of
conduction problems.
Chapter 3: One Dimensional Steady State Conduction: Analysis of one dimensional steady state
conduction in slab, radial conduction in cylinders and spheres without and with heat generation; one-
dimensional steady state conduction in composite medium- concept of thermal potential and thermal
resistance; contact resistance; critical thickness of insulation for cylinders and spheres; analysis of
fins of uniform cross section with different tip conditions; one-dimensional steady state conduction in
solids with variable thermal conductivity. Illustrative examples
UNIT – II 10 Hours
Chapter 4 : Transient Conduction: Lumped system analysis with illustrative examples; Criterion
for neglecting internal temperature gradients in transient conduction analysis ;One dimensional
transient conduction in a slab subjected to convective boundary condition- solution of this problem in
the form of Transient-Temperature Chart. Similar charts for radial transient conduction in an infinite
cylinder and in a sphere; illustrative examples; Use of these charts to solve multi-dimensional
transient conduction problems-illustrative examples. Transient temperature charts for semi-infinite
solids; Conduction shape factor: Illustrative examples
Chapter 5: Finite Difference Methods for Conduction: Finite Difference equations for one
dimensional steady state conduction in slabs, cylinders and spheres; Finite difference equations for
two dimensional steady state conduction; Explicit finite difference equations for one dimensional
transient conduction in slabs, cylinders and spheres; Implicit Scheme for one dimensional transient
conduction; Illustrative examples.
Chapter 6: Convection-Basic Concepts: Basic concepts for flow over bodies-Velocity boundary
layer, thermal boundary layer, drag coefficient, general expression for heat transfer coefficient in
2
terms of temperature gradient; illustrative examples. Concepts for flow through duct- hydraulic
diameter; hydrodynamically developing and developed flow; thermally developing and thermally
developed flow; general expression for pressure drop and heat transfer coefficient for flow through
ducts; illustrative examples; Concept of turbulence- Prandtl’s mixing length theory; velocity
distribution in turbulent flow through tubes; Dimensionless Parameters in Forced Convection Flow
and their physical significance.
Chapter 7: Forced convection for flow inside ducts: Analysis of hydro dynamically and thermally
developed laminar flow-COUETTE flow, flow inside a circular tube; expressions for friction factor
and pressure drop for hydro dynamically and thermally developed laminar and turbulent flows;
hydrodynamic and thermal entrance lengths; use of correlations to determine pressure drop and heat
transfer coefficient for hydro dynamically and thermally developed flow through tubes; illustrative
examples.
Chapter 8: Forced convection for flow over bodies: Integral method of analysis for laminar
incompressible boundary layer over a flat plate; Correlations for drag coefficient and heat transfer
coefficient for flow over a flat plate, flow across a cylinder, flow across a sphere and flow across tube
bundles
UNIT – IV 08 Hours
Chapter 9: Free convection :- Dimensionless parameters for free convective heat transfer problems;
Correlations for free convection from vertical plane surfaces, vertical cylinders, horizontal and
inclined plane surfaces, horizontal cylinders, enclosed spaces; combined forced and free convection;
illustrative examples
Chapter 10: Boiling & Condensation:- Nusselt’s theory for laminar film condensation on a vertical
plane surface; correlations for determining heat transfer coefficient for laminar and turbulent film
condensation on a plane vertical surface and horizontal tubes. Illustrative examples..
UNIT – V 10 Hours
Chapter 11: Heat Exchangers: Classification of heat exchangers; overall heat transfer coefficient;
expressions for mean temperature difference for parallel flow, counter flow heat exchangers;
correction factors for other type of heat exchangers; Limitations of LMTD method ;Effectiveness-
NTU method for heat exchanger analysis; illustrative examples
Chapter 12: Radiation Heat Transfer Among Surfaces in a Non-Participating Medium: Basic
concepts and terms used in radiation heat exchange analysis; Planck’s law,Stefan-Boltzman law,
Wein’s displacement Law,Kirchoff’s law and Lambert’s Law; Radiation heat exchanger between two
parallel infinite black and gray surfaces; Radiation exchange between two finite surfaces – Concept of
View Factor; View factor algebra; Hottel’s Cross string formula; Net work method for analysis of
radiation heat exchange in an enclosure.
Text Book :
“Fundamentals of Heat & Mass Transfer”, F.P.Incropera, D.P.Dewitt, T.L.Bergman ,A.S.Lavine,
K.N.Seetharamu, T.R.Seetharam,Wiley IndiaPvt.Ltd,New Delhi, Sixth Edition, 201
Reference Books:
1. “ Heat Transfer – A basic Approach”,M.Necati Ozisik, McGraw Hill International Edition, 1985
2. ”Principles of Heat Transfer”, Frank Kreith & Mark S. Bohn , Sixth Edition, Thomson Learning,
2001
3
CHAPTER 1
1.2. Heat Transfer Mechanisms:- There are three mechanisms by which heat transfer can
take place. All the three modes require the existence of temperature difference. The three
mechanisms are: (i) conduction, (ii) convection and (iii) radiation
1.2.1Conduction:- It is the energy transfer that takes place at molecular levels. Conduction is
the transfer of energy from the more energetic molecules of a substance to the adjacent less
energetic molecules as a result of interaction between the molecules. In the case of liquids
and gases conduction is due to collisions and diffusion of the molecules during their random
motion. In solids, it is due to the vibrations of the molecules in a lattice and motion of free
electrons.
Fourier’s Law of Heat Conduction:- The empirical law of conduction based on
experimental results is named after the French Physicist Joseph Fourier. The law states that
the rate of heat flow by conduction in any medium in any direction is proportional to the area
normal to the direction of heat flow and also proportional to the temperature gradient in that
direction. For example the rate of heat transfer in x-direction can be written according to
Fourier’s law as
4
Qx α – Ax ( ∂T∂x ) …………….
(1.1)
Or Q x = − k x Ax ( ∂T∂x ) …………….. ..
(1.2)
In equation (1.2), Qx is the rate of heat transfer in positive x-direction through area A x (Ax can
vary in the direction of x, for example, radial conduction in cylinders and spheres) of the
medium normal to x-direction, (dT/dx) is the temperature gradient and kx is the constant of
proportionality and is a material property called “thermal conductivity” of the material for
conduction in x-direction. Since heat transfer has to take place in the direction of decreasing
∂T
temperature, ( )
∂x
has to be negative in the direction of heat transfer. Therefore negative
sign has to be introduced in equation (1.2) to make Q x positive in the direction of decreasing
temperature, thereby satisfying the second law of thermodynamics. If equation (1.2) is
divided throughout by Ax we have
5
The mechanism of heat conduction in liquids is more complicated due to the fact
that the molecules are more closely spaced, and they exert a stronger inter-molecular force
field. The values of k for liquids usually lie between those for solids and gases. Unlike gases,
the thermal conductivity for most liquids decreases with increase in temperature except for
water. Like gases the thermal conductivity of liquids decreases with increase in molar mass.
In the case of solids heat conduction is due to two effects: the vibration of lattice
induced by the vibration of molecules positioned at relatively fixed positions , and energy
transported due to the motion of free electrons. The relatively high thermal conductivities of
pure metals are primarily due to the electronic component. The lattice component of thermal
conductivity strongly depends on the way the molecules are arranged. For example, diamond,
which is highly ordered crystalline solid, has the highest thermal conductivity at room
temperature.
1000 Silver
Copper
Solid
metals
100 Sodium
Liquid
metals
Steel Oxides
10
Non-
k [W/(m-K)] Mercury Metallic
solids
1.0 Water
Fibres
Non-
Metallic Insulating He, H2
liquids Materials Non-
0.1 Metallic Evacuated
Plastics gases Insulating
Wood Oils materials
Foams
CO2
0.01
Fig. 1.1: Typical range of thermal conductivities of various materials
Unlike metals, which are good electrical and heat conductors, crystalline solids
such as diamond and semiconductors such as silicon are good heat conductors but poor
electrical conductors. Hence such materials find widespread use in electronic industry.
Despite their high price, diamond heat sinks are used in the cooling of sensitive electronic
6
components because of their excellent thermal conductivity. Silicon oils and gaskets are
commonly used in the packaging of electronic components because they provide both good
thermal contact and good electrical insulation.
One would expect that metal alloys will have high thermal
conductivities, because pure metals have high thermal conductivities. For example one would
expect that the value of the thermal conductivity k of a metal alloy made of two metals with
thermal conductivities k1 and k2 would lie between k1 and k2.But this is not the case. In fact k
of a metal alloy will be less than that of either metal.
Thermal Diffusivity:- This is a property which is very helpful in analyzing transient heat
conduction problem and is normally denoted by the symbol α . It is defined as follows.
Heat conducted k
α = -------------------------------------- = -------- (m2/s) ……(1.4)
Heat Stored per unit volume ρCp
It can be seen from the definition of thermal diffusivity that the numerator represents the
ability of the material to conduct heat across its layers and the denominator represents the
ability of the material to store heat per unit volume. Hence we can conclude that larger the
value of the thermal diffusivity, faster will be the propagation of heat into the medium. A
small value of thermal diffusivity indicates that heat is mostly absorbed by
the material and only a small quantity of heat will be conducted across the material.
1.2.2. Convection :- Convection heat transfer is composed of two mechanisms. Apart from
energy transfer due to random molecular motion, energy is also transferred due to
macroscopic motion of the fluid. Such motion in presence of the temperature gradient
contributes to heat transfer. Thus in convection the total heat transfer is due to random
motion of the fluid molecules together with the bulk motion of the fluid, the major
contribution coming from the latter mechanism. Therefore bulk motion of the fluid is a
necessary condition for convection heat transfer to take place in addition to the temperature
gradient in the fluid. Depending on the force responsible for the bulk motion of the fluid,
convective heat transfer is classified into “forced convection” and “natural or free
convection”. In the case of forced convection, the fluid flow is caused by an external agency
7
like a pump or a blower where as in the case of natural or free convection the force
responsible for the fluid flow (normally referred to as the buoyancy force) is generated within
the fluid itself due to density differences which are caused due to temperature gradient within
the flow field. Regardless of the particular nature of convection, the rate equation for
convective heat transfer is given by
q = h ∆T …………………………………….. (1.5)
where q is the heat flux, ∆T is the temperature difference between the bulk fluid and the
surface which is in contact with the fluid, and ‘h” is called the “convective heat transfer
coefficient” or “surface film coefficient”. Eq.(1.5) is generally referred to as the Newton’s
law of cooling.If Ts is the surface temperature , Tf is the temperature of the bulk fluid and if
Ts > Tf, then Eq. (1.5) in the direction of heat transfer can be written as
The heat transfer coefficient h depends on (i) the type of flow (i.e. whether the flow
is laminar or turbulent), (ii) the geometry of the body and flow passage area, (iii) the thermo-
physical properties of the fluid namely the density ρ, viscosity μ, specific heat at constant
pressure Cp and the thermal conductivity of the fluid k and (iv) whether the mechanism of
convection is forced convection or free convection. The heat transfer coefficient for free
convection will be generally lower than that for forced convection as the fluid velocities in
free convection are much lower than those in forced convection. The heat transfer
coefficients for some typical applications are given in table 1.2.
1.2.3. Thermal Radiation:- Thermal radiation is the energy emitted by matter (solid, liquid
or gas) by virtue of its temperature. This energy is transported by electromagnetic waves (or
alternatively, photons).While the transfer of energy by conduction and convection requires
the presence of a material medium, radiation does not require.Infact radiation transfer occurs
most effectively in vacuum.
8
Consider radiation transfer process for the surface shown in Fig.1.2a.Radiation that
is emitted by the surface originates from the thermal energy of matter bounded by the
surface, and the rate at which this energy is released per unit area is called as the surface
emissive power E.An ideal surface is one which emits maximum emissive power and is
called an ideal radiator or a black body.Stefan-Boltzman’s law of radiation states that the
emissive power of a black body is proportional to the fourth power of the absolute
temperature of the body. Therefore if Eb is the emissive power of a black body at temperature
T 0K, then
G ρG E
qs
qsurr
Fig.1.2: Radiation exchange: (a) at a surface and (b) between a surface and large surroundings
Eb α T 4
Or Eb = σ T 4 ………………………………….(1.7)
−8
σ is the Stefan-Boltzman constant (σ = 5.67 * 10 W / (m2 – K4) ). For a non black surface
the emissive power is given by
E = ε σ T 4…………………………………(1.8)
where ε is called the emissivity of the surface (0 ≤ ε ≤ 1).The emissivity provides a measure
of how efficiently a surface emits radiation relative to a black body. The emissivity strongly
depends on the surface material and finish.
Radiation may also incident on a surface from its surroundings. The rate at which the
radiation is incident on a surface per unit area of the surface is calle the “irradiation” of the
9
surface and is denoted by G. The fraction of this energy absorbed by the surface is called
“absorptivity” of the surface and is denoted by the symbol α. The fraction of the
incident energy is reflected and is called the “reflectivity” of the surface denoted by ρ and
the remaining fraction of the incident energy is transmitted through the surface and
is called the “transmissivity” of the surface denoted by τ. It follows from the definitions of α,
ρ, and τ that
α+ ρ + τ = 1 …………………………………….(1.9)
Therefore the energy absorbed by a surface due to any radiation falling on it is given by
Gabs = αG …………………………………(1.10)
The absorptivity α of a body is generally different from its emissivity. However in many
practical applications, to simplify the analysis α is assumed to be equal to its emissivity ε.
Radiation Exchange:- When two bodies at different temperatures “see” each other, heat is
exchanged between them by radiation. If the intervening medium is filled with a substance
like air which is transparent to radiation, the radiation emitted from one body travels through
the intervening medium without any attenuation and reaches the other body, and vice versa.
Then the hot body experiences a net heat loss, and the cold body a net heat gain due to
radiation heat exchange between the two. The analysis of radiation heat exchange among
surfaces is quite complex which will be discussed in chapter 10. Here we shall consider two
simple examples to illustrate the method of calculating the radiation heat exchange between
surfaces.
As the first example let us consider a small opaque plate (for an opaque surface τ
= 0) of area A, emissivity ε and maintained at a uniform temperature T s. Let this plate is
exposed to a large surroundings of area Asu (Asu >> A) whish is at a uniform temperature Tsur
as shown in Fig. 1.2b.The space between them contains air which is transparent to thermal
radiation.
Qem = A ε σ Ts4
The large surroundings can be approximated as a black body in relation to the small plate.
Then the radiation flux emitted by the surroundings is σ T sur4 which is also the radiaton flux
incident on the plate. Therefore the radiation energy absorbed by the plate due to emission
from the surroundings is given by
Qab = A α σ Tsur4.
The net radiation loss from the plate to the surroundings is therefore given by
10
Assuming α = ε for the plate the above expression for Qnet reduces to
The above expression can be used to calculate the net radiation heat exchange between a
small area and a large surroundings.
As the second example, consider two finite surfaces A1 and A2 as shown in Fig. 1.3.
Surroundings
A2, ε2, T2
A1, ε1, T1
The surfaces are maintained at absolute temperatures T1 and T2 respectively, and have
emissivities ε1 and ε2. Only part of the radiation leaving A1 reaches A2. Similar
considerations apply for the radiation leaving A2.If it is assumed that the radiation from the
surroundings is negligible when compared to the radiation from the surfaces A 1 and A2 then
we can write the expression for the radiation emitted by A1 and reaching A2 as
where F1 – 2 is defined as the fraction of radiation energy emitted by A1 and reaching A2.
Similarly the radiation energy emitted by A2 and reaching A1 is given by
where F2 – 1 is the fraction of radiation energy leaving A 2 and reaching A1. Hence the net
radiation energy transfer from A1 to A2 is given by
Q1 – 2 = Q1→2 − Q2→1
11
= [F1− 2 A1ε1σ T14] − [F2− 1 A2 ε2 σ T24]
F1-2 is called the view factor (or geometric shape factor or configuration factor) of A 2 with
respect to A1 and F2 - 1 is the view factor of A1 with respect to A2.It will be shown in chapter
10 that the view factor is purely a geometric property which depends on the relative
orientations of A1 and A2 satisfying the reciprocity relation, A1 F1 – 2 = A2 F2 – 1.
Qr = hrA ΔT
For the example of radiation exchange between a surface and the surroundings
[Eq. (1. 11)] using the concept of radiation heat transfer coefficient we can write
First Law on rate basis :- The rate at which thermal and mechanical energy enters a control
volume, plus the rate at which thermal energy is generated within the control volume, minus
the rate at which thermal and mechanical energy leaves the control volume must be equal to
the rate of increase of stored energy within the control volume.
Consider a control volume shown in Fig. 1.4 which shows that thermal and
mechanical energy are entering the control volume at a rate denoted by Ein, thermal and
12
.
Eg . .
. Est Eout
Ein
.
mechanical energy are leaving the control volume at a rate denoted by Eout. The rate at
.
which energy is generated within the control volume is denoted by Eg and the rate at
.
which energy is stored within the control volume is denoted by Est. The general form of the
energy balance equation for the control volume can be written as follows:
. . . .
Ein + Eg − Eout = Est ……………………………(1.16)
.
Est is nothing but the rate of increase of energy within the control volume and hence can be
d Es
written as equal to
dt
First Law over a Time Interval Δt :- Over a time interval Δt, the amount of thermal and
mechanical energy that enters a control volume, plus the amount of thermal energy
generated within the control volume minus the amount of thermal energy that leaves the
control volume is equal to the increase in the amount of energy stored within the control
volume.
The above statement can be written symbolically as
The inflow and outflow energy terms are surface phenomena. That is they are associated
exclusively with the processes occurring at the boundary surface and are proportional to the
surface area.
The energy generation term is associated with conversion from some other form
(chemical, electrical, electromagnetic, or nuclear) to thermal energy. It is a volumetric
phenomenon.That is, it occurs within the control volume and is proportional to the magnitude
of this volume. For example, exothermic chemical reaction may be taking place within the
control volume. This reaction converts chemical energy to thermal energy and we say that
energy is generated within the control volume. Conversion of electrical energy to thermal
energy due to resistance heating when electric current is passed through an electrical
conductor is another example of thermal energy generation
13
Energy storage is also a volumetric phenomenon and energy change within the
control volume is due to the changes in kinetic, potential and internal energy of matter within
the control volume.
A. Conduction
Example 1.1:- Heat flux through a wood slab 50 mm thick, whose inner and outer surface
temperatures are 40 0 C and 20 0 C respectively, has been determined to be 40 W/m2. What is
the thermal conductivity of the wood slab?
Known: T1 = 40 0 C; T2 = 20 0 C; L = 0.05 m;
T1 q = 40 W / m2.
T1
q Find: Thermal conductivity,k
T2
T2 Assumptions: (i) steady state conduction
Solution: For steady state conduction across the thickness of the slab and noting that the slab
is not generating any thermal energy, the first law equation for the slab can be written as :
Rate at which thermal energy (conduction) is entering the slab at the surface x = 0 is equal to
the rate at which thermal energy is leaving the slab at the surface x = L.
That is
14
Qx k ( T 1−T 2 )
Heat flux = q = =
A L
qL 40∗0.05
Hence k= = = 0.1 W / (m – K)
T
( 1−T 2 ) ( 40−20 )
Example 1.2:- A concrete wall, which has a surface area of 20 m2 and thickness 30 cm,
separates conditioned room air from ambient air.The temperature of the inner surface of the
wall is 25 0 C and the thermal conductivity of the wall is 1.5 W / (m-K).Determine the heat
loss through the wall for ambient temperature varying from ─ 15 0 C to 38 0 C which
correspond to winter and summer conditions and display your results graphically.
Q = kA(T1 – T2) / L
15
Heat loss Q for different values of T2 ranging from – 15 0 C to + 38 0 C are obtained from Eq.
(1) and the results are plotted as shown: Scale x-axis : 1cm= 5 C ; y-axis : 1cm =1000W
Example 1.3:-
What is the
thickness
required of a
masonry wall
having a
thermal
conductivity of
0.75 W/(m-K),
if the heat
transfer rate is
to be 80 % of
the rate
through
another wall having thermal conductivity of 0.25 W/(m-K) and a thickness of 100 mm? Both
walls are subjected to the same temperature difference.
16
Q1 k1 L2
---- = ----------
Q2 k2 L1
Q2 k1
Or L1 = ----------- L2 = (1 / 0.80) * (0.75/0.25) x 100 = 375 mm
Q1 k2
B. Convection:
Example 1.4:- Air at 40 0 C flows over a long circular cylinder of 25 mm diameter with
an embedded electrical heater. In a series of tests, measurements were made of power
per unit length, P required to maintain the surface temperature of the cylinder at 300 0 C
for different stream velocities V of the air. The results are as follows:
(a) Determine the convective heat transfer coefficient for each velocity and display your
results graphically. (h = P / 20.43)
(b)Assuming the dependence of the heat transfer coefficient on velocity to be of the
form h = CV n , determine the parameters C and n from the results of part (a).
Ts
V,T∞
D
Find: (a) Heat transfer coefficient as a function of velocity of the fluid (b) To find the values
of C and n in the equation h = CVn
17
Assumptions: (i) Heat transfer from the surface of the solid to the flowing fluid is by
convection only.(i.e. radiation heat transfer between the surface of the solid and the fluid is
negligible)
Solution: If h is the surface heat transfer coefficient then the power dissipated by the cylinder
by convection is given by
where As is the area of contact between the fluid and the surface of the cylinder.
Values of h for different flow velocities are obtained and tabulated as follows:
(a) A graph of h versus V can now be plotted as shown in Fig. P 1.4 (a).
18
(b) h = CVn
Therefore ln h = ln C + n ln V …………………………(2)
If ln h is plotted against ln V it will be straight line and the slope of which will give the value
of n. Also the intercept of this line w.r.t the axis on which ln V is plotted will give the value
of ln C from which C can be determined. The log –log plot is as shown in Fig. P 1.4(b).
Scale X axis 1cm=0.25; Y axis 1cm=0.5
ln C = 3.1 or C = 22.2
19
Example 1.5:- A large surface at 50 0 C is exposed to air at 20 0 C. If the heat transfer
coefficient between the surface and the air is 15 W/(m 2-K), determine the heat transferred
from 5 m2 of the surface area in 7 hours.
Ts
h
Find: Qtotal
Assumption (i) Steady state heat transfer; (ii)Heat transfer from the surface of the solid to
the surroundings is by convection only.(i.e. radiation heat transfer between the surface of the
solid and the surroundings is negligible)
h
Ts
D
T∞
Find: Q
20
Assumptions: (i) Heat transfer from the surface of the solid to the surrounding fluid is by
convection only.(i.e. radiation heat transfer between the surface of the solid and the fluid is
negligible)(ii) Steady state heat transfer
Solution: Q = hAs(Ts - T∞) = h 4πR2 (Ts - T∞) = 15 * 4π * (0.25/2)2 * (120 – 20) = 294.52 W
C. Radiation:
d1
= 96.85 W
If the surface is having an emissivity of 0.8 then
Example 1.8:- A vacuum system as used in sputtering conducting thin films on micro
circuits, consists of a base plate maintained at a temperature of 300 K by an
electric heater and a shroud within the enclosure maintained at 77 K by circulating
liquid nitrogen. The base plate insulated on the lower side is 0.3 m in diameter and has
an emissivity of 0.25.
(a) How much electrical power must be provided to the base plate heater?
21
(b) At what rate must liquid nitrogen be supplied to the shroud if its latent heat of
vaporization is 125 kJ/kg?
Sketch:
Find: (a) Qr ; (b) Mass of nitrogen vapourized per unit time, mN2
Assumptions:
Solution: Surface area of the top surface of the base plate = As = (π / 4)d12 = (π / 4) * 0.32
= 0.0707 m2
4 4
(a) Qr = ε1σ As (T1 – T2 )
. 8.08
mN2 = Qr / hfg = ---------------- = 6.464 * 10-5 kg/s or 0.233 kg/s
125 x 1000
Example 1.9:- A flat plate has one surface insulated and the other surface exposed to the
sun. The exposed surface absorbs the solar radiation at a rate of 800 W/m2 and dissipates
heat by both convection and radiation into the ambient at 300 K. If the emissivity of the
surface is 0.9 and the surface heat transfer coefficient is 12 W/(m2-K), determine the
surface temperature of the plate.
22
Qsolar
T∞
Qr
Qconv Ts , ε , h
Insulated
Find: Ts
Assumptions: (i) steady state heat transfer;(ii) the emissivity and the surface heat transfer
coefficient is uniform throughout the surface.
Solution: Energy balance equation for the top surface of the plate is given by
Qsolar = Qr + Qconv
Trial 1:- Assume Ts = 350 K. Then LHS of Eq. (1) = 972.6 which is more than RHS of Eq.
(1). Hence Ts < 350 K.
Trial 2 :- Assume Ts = 340 K. Then LHS of Eq. (1) = 932.6 which is slightly less than
RHS. Therefore Ts should lie between 340 K and 350 K but closer to 340 K.
Trial 3:- Assume Ts = 342.5 K. Then LHS of Eq.(1) = 942.5 = RHS of Eq. (1).
Therefore Ts = 342.5 K
Example 1.10:- The solar radiation incident on the outside surface of an aluminum
shading device is 1000 W/m2. Aluminum absorbs 12 % of the incident solar energy and
dissipates it by convection from the back surface and by combined convection and
23
radiation from the outer surface. The emissivity of aluminum is 0.10 and the convective
heat transfer coefficient for both the surfaces is 15 W/(m2 –K). The ambient temperature
of air may be taken as 20 0 C. Determine the temperature of the shading device.
q solar qr
qc1
qc2
Assumptions: (i) Steady state heat transfer; (ii) the emissivity and the surface heat transfer
coefficient are uniform throughout the surface
Solution: Solar radiation flux absorbed by aluminum = qa = α qsolar = 0.12 * 1000 = 120 W /
m2.
qa = qr + qc1 + qc2
Or 120 = 5.67 * 10 ─ 8* 0.10 (Ts4 – 0.12 * 2934) + (Ts – 293) * (15 + 15)
Trail 1:- Assume Ts = 300 K. Then LHS = 15981 which is > RHS.
24
Trail 2 :- Assume Ts = 295 K. Then LHS = 15710.73 which is < RHS. Hence Ts should lie
between 300K and 295 K.
Trial 3 :- Assume Ts = 297 K . Then LHS = 15819 which is almost equal to RHS (Within
0.34 %)
Therefore Ts = 297 K.
25
CHAPTER 2
A(x) q’’’
Qx Qx + Δx
O
x Δx
of heat transfer in x – direction entering into the volume element at x, A(x) area of heat flow
at the section x ,q′′′ is the thermal energy generation within the element per unit volume and
Qx+Δx is the rate of conduction out of the element at the section x + Δx. The energy balance
equation per unit time for the element can be written as follows:
[ Rate of heat conduction into the element at x + Rate of thermal energy generation within
the element − Rate of heat conduction out of the element at x + Δx ]
26
∂E
i.e., Qx + Qg – Qx+Δx =
∂t
∂2 Q x
or Qx + q′′′ Ax Δx – {Qx + ( ) Δx + ( ) (Δx) / 2! + …….}
∂Q x
∂x ∂ x2
2
∂
= [ρA(x)ΔxCpT]
∂t
Neglecting higher order terms and noting that ρ and Cp are constants the above equation
simplifies to
Or − ( ∂Q∂ x ) + q′′′ A = ρA
x
x x Cp ( ∂T∂t )
Using Fourier’s law of conduction , Qx = − k Ax ( ∂T∂x ) , the above equation simplifies to
∂
−
∂x
{− k Ax ( ∂T∂x ) } + q′′′ Ax = ρAx Cp ( ∂T∂t )
∂
Or {1/Ax }
∂x
{− k Ax ( ∂T∂x ) } + q′′′ = ρ Cp ( ∂T∂t ) ………
……………(2.1)
Eq. (2.1) is the most general form of conduction equation for one-dimensional unsteady state
conduction.
2.2.1.Equation for one-dimensional conduction in plane walls :- For plane walls, the area
of heat flow Ax is a constant. Hence Eq. (2.1) reduces to the form
∂
∂x
{k ( ∂T∂x ) } + q′′′ = ρ Cp ( ∂T∂t )
……………………..(2.2)
(i) If the thermal conductivity of the solid is constant then the above equation reduces to
27
2
( )
∂T
∂x
2 + (q′′′ / k) = (1/α ) ( ∂T∂t )
………………………(2.3)
(ii) For steady state conduction problems in solids of constant thermal conductivity
temperature within the solid will be independent of time (i.e.(∂T/ ∂t) = 0)
and hence Eq. (2.3) reduces to
( )
d T
dx
2 + (q′′′ / k) = 0 …………………………….
(2.4)
(iii) For a solid of constant thermal conductivity for which there is no thermal energy
generation within the solid q’’’ = 0 and the governing equation for steady state conduction is
obtained by putting q’’’ = 0 in Eq. (2.4) as
d2T
( )
dx
2 = 0 ………………………
(2.4)
R
Qr
r
Qr
For radial conduction in cylinders, by convention the radial coordinate is denoted by ‘r’
instead of ‘x’ and the area of heat flow through the cylinder of length L,at any radius r is
given by Ax = A(r) = 2πrL. Hence substituting this expression for Ax and replacing x by r in
Eq. (2.1) we have
28
1 ∂
{ ( 2 πrL ) ∂r
{k 2πrL ( ∂T∂r ) } + q′′′ = ρ Cp ( ∂T∂t )
1 ∂
Or
r ∂r
{ {k r ( ∂T∂r ) } + q′′′ = ρ Cp ( ∂T∂t ) ………..
……….(2.5)
(i) For cylinders of constant thermal conductivity the above equation reduces to
1 ∂
r ∂r
{ {r ( ∂T∂r ) } + q′′′ / k =
1 ∂T
( )
α ∂t
…….
…………….(2.6)
where α = k / ρCp.
(ii) For steady state radial conduction (i.e. (∂T/ ∂t) = 0 ) in cylinders of constant k, the above
equation reduces to
1 d
r dr
{r ( dTdr ) } + q′′′ / k = 0
…………………….(2.7)
(iii) For steady state radial conduction in cylinders of constant k and having no thermal
energy generation (i.e. q’’’ = 0) the above equation reduces to
d
dr
{r ( dTdr ) } =0 ………………………
(2.8)
1 ∂
{
4πr 2 }
∂r
{k 4π r2 ( ∂T∂r ) } + q′′′ = ρCp ( ∂T∂t )
1 ∂
Or
r
2
∂r
{k r2 ( ∂T∂r ) } + q′′′ = ρ Cp ( ∂T∂t ) …….
…………………(2.9)
(i) For spheres of constant thermal conductivity the above equation reduce to
29
1 ∂
r 2
∂r
{ r2 ( ∂T∂r ) } + q′′′ / k =
1
α ( ∂T∂t )
……………………..(2.10)
(ii) For steady state conduction in spheres of constant k the above equation further reduce to
1 d
r 2
dr
{ r2 ( dTdr ) } + q′′′ / k = 0
……………………(2.11)
(iii) For steady state conduction in spheres of constant k and without any thermal energy
generation the above equation further reduces to
1 d
r
2
dr
{ r2 ( dTdr ) } =0
………………………(2.12)
Equation in compact form:- The general form of one – dimensional conduction equations
for plane walls, cylinders and spheres {equations (2..2), (2.5) and (2.9)} can be written in a
compact form as follows:
1 ∂
rn ∂r
{k rn ( ∂T∂r ) } + q′′′ = ρ Cp ( ∂T∂t ) ………..
……….(2.13)
30
z
y
Δz
x
Qx + Δx
Qx
Δy
Δx
Qy Qz
Fig. 2.3: Conduction heat transfer across the six faces of a volume element
= − ∂ / ∂x [− kx Δy Δz(∂T / ∂x)] Δx
Hence the net rate of conduction into the element from all the three directions
Qin = {∂ / ∂x[kx (∂T / ∂x)] + ∂ / ∂y[ky (∂T / ∂y)] + ∂ / ∂z[kz (∂T / ∂z)] } Δx Δy Δz
31
Qin + Qg = ∂E / ∂t
{∂ / ∂x[kx (∂T / ∂x)] + ∂ / ∂y[ky (∂T / ∂y)] + ∂ / ∂z[kz (∂T / ∂z)] } + q′′′ = ρ Cp (∂T / ∂t)
……………………(2.14)
Equation (2.14) is the most general form of conduction equation in Cartesian coordinate
system. This equation reduces to much simpler form for many special cases as indicated
below.
Special cases:- (i) For isotropic solids, thermal conductivity is independent of direction; i.e.,
kx = ky = kz = k. Hence Eq. (2.14) reduces to
{∂ / ∂x[k (∂T / ∂x)] + ∂ / ∂y[k (∂T / ∂y)] + ∂ / ∂z[k (∂T / ∂z)] } + q′′′ = ρ Cp (∂T / ∂t)
……………………..(2.15)
(ii) For isotropic solids with constant thermal conductivity the above equation further reduces
to
Eq.(2.16) is called as the “Fourier – Biot equation” and it reduces to the following forms
under specified conditions as mentioned below:
(v) Steady state conduction without heat generation [i.e., (∂T / ∂t) = 0 and q’’’ = 0]:
32
2.3.2. Three dimensional conduction equation in cylindrical coordinate system:
P(x,y,z)
x
θ
r
1 ∂ ∂T 1 ∂ ∂T ∂ ∂T ∂T
r ∂r (
kr + ) k
∂ r r2 ∂ θ ∂ θ ( ) + k( )
∂z ∂ z
+ q’’’ = ρCp
∂t
…………..(2.20)
33
z
OP’ = r sin φ.Hence
x = r sin φ cos θ ;
P(x,y,z)
y = r sin φ sin θ ;
z = r cos φ
r
φ
O
x
θ
P’
Fig: 2.5: Spherical coordinate system
y
Using the relation between x, y ,z and r, θ and φ, the conduction equation (2.15) can be
transformed into the equation in terms of r, θ and φ as follows.
1 ∂ ∂T 1 ∂ ∂T 1 ∂ ∂T
r ∂r
2 (
kr2 )
+ 2 k
∂ r r ( sin φ ) ∂ θ ∂θ
2 ( ) + 2
r sin φ ∂φ(k sin φ
∂φ ) + q’’’ = ρCp
∂T
∂t
…………….(2.21).
2.4.Boundary and Initial Conditions:
The temperature distribution within any solid is obtained by integrating the above conduction
equation with respect to the space variable and with respect to time.The solution thus
obtained is called the “general solution” involving arbitrary constants of integration. The
solution to a particular conduction problem is arrived by obtaining these constants which
depends on the conditions at the bounding surfaces of the solid as well as the initial
condition. The thermal conditions at the boundary surfaces are called the “boundary
conditions” . Boundary conditions normally encountered in practice are:
(i) Specified temperature (also called as boundary condition of the first kind),
(ii) Specified heat flux (also known as boundary condition of the second kind),
(iii) Convective boundary condition (also known as boundary condition of the third kind) and
(iv) radiation boundary condition. The mathematical representations of these boundary
conditions are illustrated by means of a few examples below.
34
2.4.1. Specified Temperatures at the Boundary:- Consider a plane wall of thickness L
whose outer surfaces are maintained at temperatures T0 and TL as shown in Fig.2.6. For one-
dimensional unsteady state conduction the boundary conditions can be written as
y
T(x,t)
T0 TL T(x,y)
T = φ(x)
L T2
b
Ψ(y)
a
x
T1
x
Fig. 2.6: Boundary condition Fig.2.7: Boundary conditions of
of first kind for a plane wall first kind for a rectangular plate
Consider another example of a rectangular plate as shown in Fig. 2.7. The boundary
conditions for the four surfaces to determine two-dimensional steady state temperature
distribution T(x,y) can be written as follows.
2.4.2. Specified heat flux at the boundary:- Consider a rectangular plate as shown in Fig.
2.8 and whose boundaries are subjected to the prescribed heat flux conditions as shown in the
figure. Then the boundary conditions can be mathematically expressed as follows.
35
qb T(x,y)
y
q0 qa
b
a
x
insulated
2.4.3. Boundary surface subjected to convective heat transfer:- Fig. 2.9 shows a plane
wall whose outer surfaces are subjected to convective boundary conditions.The surface at x =
0 is in contact with a fluid which is at a uniform temperature T i and the surface heat transfer
coefficient is hi. Similarly the other surface at x = L is in contact with another fluid at a
uniform temperature T0 with a surface heat transfer coefficient h0. This type of boundary
condition is encountered in heat exchanger wherein heat is transferred from hot fluid to the
cold fluid with a metallic wall separating the two fluids. This type of boundary condition is
normally referred to as the boundary condition of third kind. The mathematical representation
of the boundary conditions for the two surfaces of the plane wall can be written as follows.
36
T(x)
Surface in contact with fluid at T0 with surface heat transfer coefficient h0
L
Surface in contact with fluid at Ti with surface heat transfer coefficient h i
x
Fig. 2.9: Boundaries subjected to convective heat transfer for a plane wall
T(x,t)
Surface with emissivity ε is radiating heat to the surroundings at Ts 0K
x
Fig. 2.10: Boundary surface at x = L subjected to radiation heat transfer for a plane wall
37
2.4.5. General form of boundary condition (combined conduction, convection and
radiation boundary condition): There are situations where the boundary surface is
subjected to combined conduction, convection and radiation conditions as illustrated in Fig.
2.11.It is a south wall of a house and the outer surface of the wall is exposed to solar
radiation. The interior of the room is at a uniform temperature T i. The outer air is at uniform
temperature T0 . The sky, the ground and the surfaces of the surrounding structures at this
location is modeled as a surface at an effective temperature of Tsky.
x
L
qradiation
qconduction αqsolar
qconvection
38
2.5. Illustrative Examples:
Example2.1.By writing an energy balance for a differential cylindrical volume element in the
‘r’variable (r is any radius), derive the one-dimensional time dependent heat conduction
equation with internal heat generation and variable thermal conductivity in the cylindrical
coordinate system.
Sketch :
r
Qr
dr
Qr
Qg
Qr+dr
Qr+dr
Solution:
A cylindrical element of thickness dr in the radial direction at a radius r is shown in the figure
above. For unsteady state one dimensional radial conduction with heat generation is given by
Qr + Qg – Qr+dr = (∂E / ∂t)
where Qr is the rate of conduction into the element at radius r = ─ k 2πrL (∂T /∂r)
(∂E / ∂t) is the rate of increase of the energy of the element. = 2π rL dr ρCp (∂T / ∂t)
where dV=2πrLdr-------- volume
39
Substituting these expressions in Eq.(1) we get
Example2.2.By writing an energy balance for a differential spherical volume element in the
variable (r is any radius), derive the one-dimensional time dependent heat conduction
equation with internal heat generation and variable thermal conductivity in the spherical
coordinate system.
Sketch:
dr
Qr+dr
Qr
Qg
Solution: Consider a spherical element of thickness dr at any radius r as shown in the figure
above. The energy balance equation for one – dimensional radial unsteady state conduction
with heat generation is given by
Qg = rate of heat generation within the element = (4/3)π [(r + dr)3 – r3 ] q′′′
40
(∂E / ∂t) = rate of increase of energy of the element = ρ (4/3)π [ (r + dr) 3 – r3 ](∂T/∂t)
(r + dr)3 – r3 = 3 r2 dr.
Therefore Qg = 4 π r2 dr q′′′
Substituting the expressions for Qr, Qg and (∂E / ∂t) in Eq. (1) we have
Simplifying the above equation and noting that if k is given to be constant we have
Find: Differential equation for one dimensional time dependent heat conduction with heat
generation
Assumptions: One dimensional time dependent radial heat conduction with heat generation
Solution: (a) The general form of conduction equation for an isotropic solid in rectangular
coordinate system is given by
∂ / ∂x (k∂T / ∂x) + ∂ / ∂y (k∂T / ∂y) + ∂ / ∂z (k∂T / ∂z) + q′′′ = (ρ Cp) (∂T / ∂t)
…………..(1)
For steady state conduction (∂T / ∂t) = 0 ; For one dimensional conduction in x – direction
we have
∂T / ∂y = ∂T / ∂z = 0 . Therefore ∂T / ∂x = dT / dx .
Therefore Eq. (1) reduces to
41
d / dx (k dT / dx) + q′′′ = 0.
d2 T / dx2 + q′′′ / k = 0.
(b) The general form of conduction equation in cylindrical coordinate system is given by
(1 / r) ∂ / ∂r (kr ∂T / ∂r) + (1 / r2) ∂ / ∂θ(k ∂T / ∂θ) + ∂ / ∂z (k∂T / ∂z) + q′′′ = ρCp( ∂T / ∂t)
For steady state conduction, ( ∂T / ∂t) = 0 ; For one-dimensional radial conduction we have
∂T / ∂θ = 0 and ∂T / ∂z = 0. Therefore ∂T / ∂r = dT / dr. With these simplifications the
general form of conduction equation reduces to
(1 / r) d / dr (r dT/dr) + q′′′ / k = 0.
(a) The general form of conduction equation in spherical coordinate system is given by
+ {1/(r2 sin φ)} ∂ /∂φ (k sin φ ∂T/∂φ) + q′′′ = ρCp (∂T/ ∂t) ………..(1)
For steady state conduction (∂T ∂t) = 0 ; For one dimensional radial conduction we have
Sketch :
42
h,T∞
q0
Boundary conditions are :
L (i)at x = 0; ─ k (dT / dx)x = 0 = q0
Example2.5.Consider a solid cylinder of radius R and height Z. The outer curved surface of
thecylinder is subjected to a uniform heating electrically at a rate of q0 W / m2.Both the
circular surfaces of the cylinder are exposed to an environment at a uniform
temeperature T∞ with a surface heat transfer coefficient h.Write the mathematical
formulation of the boundary conditions for the solid cylinder.
Solution:
h,T∞ Boundary conditions are:
z (i) at r = 0; (∂T/∂r) = 0 (axis of symmetry)
R
(ii) at r = R; k(∂T/∂r) = q0
h , T∞
Example2.6. A hollow cylinder of inner radius ri, outer radius r0 and height H is subjected
to thefollowing boundary conditions.
43
(a) The inner curved surface is heated uniformly with an electric heater at a constant rate of
q0 W/m2,
(b) the outer curved surface dissipates heat by convection into an ambient at a uniform
temperature, T∞ with a convective heat transfer coefficient,h
(c) the lower flat surface of the cylinder is insulated, and
(d) the upper flat surface of the cylinder dissipates heat by convection into the ambient at T ∞
with surface heat transfer coefficient h. Write the mathematical formulation of the boundary
conditions for the hollow cylinder.
(iv) at z = H,
Example2.7. A plane wall of thickness L and with constant thermal properties is initially
at a uniform temperature Ti. Suddenly one of the surfaces of the wall is subjected to
heatingby the flow of a hot gas at temperature T∞ and the other surface is kept
insulated. Theheat transfer coefficient between the hot gas and the surface exposed to it
is h. There is no heat generation in the wall. Write the mathematical formulation of the
problem to determine the one-dimensional unsteady state temperature within the wall.
44
T = Ti at t = 0
L
Insulated h∞,T∞
T = T(x,t)
Solution:
Governing differential equation to determine T(x,t) is given by
The boundary conditions are : (i) at x = 0, (∂T / ∂x)x=0 = 0. (Insulated) for all t >0
Example2.8. A copper bar of radius R is initially at a uniform temperature Ti. Suddenly the
heating of the rod begins at time t=0 by the passage of electric current, which generates
heat at a uniform rate of q’’’ W/m3. The outer surface of the dissipates heat into an
ambient at a uniform temperature T∞ with a convective heat transfer coefficient h.
Assuming that thermal conductivity of the bar to be constant, write the mathematical
formulation of the heat conduction problem to determine the one-dimensional radial
unsteady state temperature distribution in the rod.
h,T∞
T = Ti at t ≤ 0
q’’’ for t ≤ 0
R
Solution:
45
The governing differential equation to determine T(r,t) is given by
Example2.9. Consider a solid cylinder of radius R and height H. Heat is generated in the
solid at auniform rate of q’’’ W/m3. One of the circular faces of the cylinder is insulated and
theother circular face dissipates heat by convection into a medium at a uniform
temperature of T∞ with a surface heat transfer coefficient of h. The outer curved surface
of the cylinder is maintained at a uniform temperature of T0. Write the mathematical
formulation to determine the two-dimensional steady state temperature distribution
T(r,z) in the cylinder.
(iv) at z =rH,
─ k (∂T/∂z)z=H = h (T |z=H ─ T∞)
for all x.
Insulated
for all r.
Example2.10. Consider a rectangular plate as shown in Fig. 2.10. The plate is generating
heatat a uniform rate of q’’’ W/m3. Write the mathematical formulation to determine two
dimensional steady state temperature distribution in the plate.
46
h,T∞
T0
bb
a
a
x
Insulated
Fig. 2.10 : Schematic for example 2.10
Example2.11. Consider a medium in which the heat conduction equation is given in its
simple form as
It can be seen from this equation that T depends on one space variable x and the time variable
t. Hence the problem is one dimensional transient conduction problem. No heat generation
term appears in the equation indicating that the medium is not generating any heat.The
47
thermal conductivity of the medium does not appear within the differential symbol indicating
that the conductivity of the medium is constant.
Example2.12. Consider a medium in which the heat conduction equation is given in its
simple form as
(1/r) d / dr(r k dT/dr) + q’’’ = 0.
It can be seen from this equation that the temperature T depends only on one space variable
‘r’ and it does not depend on time t. Also the heat generation term q’’’ appears in the
differential equation.Hence the problem is a one-dimensional steady state conduction
problem with heat generation. Since the thermal conductivity appears within the differential
symbol, it follows that the thermal conductivity of the medium is not a constant but varies
with temperature.
Example2.13. Consider a medium in which heat the heat conduction equation in its simplest
form is given as
(1/r2) ∂/∂r (r2 ∂T /∂r) = (1/α) (∂T/∂t)
It can be seen from the above equation that the temperature T depends on two space variables
r and z and does not depend on time. There is the heat generation term appearing in the
equation and the thermal conductivity k appears within the differential symbol ∂/∂r and
∂/∂z. Hence the problem is two-dimensional steady state conduction with heat generation in
a medium of variable thermal conductivity.
Example2.14. Consider a medium in which the heat conduction equation is given in its
simplest form as
(1/r) ∂/∂r (k r ∂T /∂r) + ∂/∂z (k ∂T /∂z) + q’’’ = 0
48
(b) Is heat transfer one-, two- or three-dimension?
(c)Is there heat generation in the medium?
(d)Is the thermal conductivity constant or variable?
1 ∂ 2T
2 2
(1/r ) ∂/∂r (r ∂T /∂r) + ---------- [ -------] = (1/α) (∂T/∂t)
r2 sin 2 φ ∂θ2
It can be seen from the given equation that the temperature T depends two space variables r
and θ and it also depends on the time variable t. There is no heat generation term appearing in
the given equation . Also the thermal conductivity k do not appear within the differential
symbol. Hence the given equation represents two-dimensional, steady state conduction in a
medium of constant thermal conductivity and the medium is not generating any heat.
Example2.15. Consider the north wall of a house of thickness L. The outer surface of the
wall exchanges heat by both convection and radiation.The interior of the house is
maintained at a uniform temperature of Ti, while the exterior of the house is at a
uniform temperature T0. The sky, the ground, and the surfaces of the surrounding structures
at this location can be modeled as a surface at an effective temperature of T sky for radiation
heat exchange on the outer surface.The radiation heat exchange between the inner surface of
the wall and the surfaces of the other walls, floor and ceiling are negligible.The convective
heat transfer coefficient for the inner and outer surfaces of the wall under consideration are
hi and h0 respectively.The thermal conductivity of the wall material is K and the emissivity of
the outer surface of the wall is ‘ε 0’. Assuming the heat transfer through the wall is steady and
one dimensional, express the mathematical formulation (differential equation and boundary
conditions) of the heat conduction problem
49
Surface in contact with fluid at Ti and surface heat transfer coefficient hi
x
L T (x)
qradiation
surface in contac
at T0 and surface heat
qconduction transfer coefficient h0
qconvection
ε0
Solution: The problem is one-dimensional steady state conduction without any heat
generation and the wall is of constant thermal conductivity. Hence the governing differential
equation is
d2T / dx2 = 0.
50
CHAPTER 3
3.2.1. The Plane Wall (The Infinite Slab):- The statement of the problem is to determine the
temperature distribution and rate of heat transfer for one dimensional steady state conduction
in a plane wall without heat generation subjected to specified boundary conditions.
51
T = T(x)
T1 T2
Qx
R = L /(Ak)
L
Fig. 3.1: One dimensional steady state conduction in a slab
The governing equation for one − dimensional steady state conduction without heat
generation is given by
d2T
=0 ……………..(3.1)
d x2
T = C1x + C2 ………………………………(3.2)
where C1 and C2 are constants which can be evaluated by knowing the boundary conditions.
( T 2−T 1 )
Or C1 =
L
Substituting for C1 and C2 in Eq. (3.2), we get the temperature distribution in the plane wall
as
52
x
T(x) = (T2 – T1) + T1
L
Or T(x) – T1 x
------------ = -------- ……………………………..(3.3)
(T2 – T1) L
For a plane wall A(x) = constant = A. From Eq. (3.3), dT/dx = (T2 – T1) / L.
Eq. (3.5) is analogous to Ohm’s law for flow of electric current. In this equation (T1 – T2) can
be thought of as “thermal potential”, R can be thought of as “thermal resistance”,so that the
plane wall can be represented by an equivalent “thermal circuit” as shown in
Fig.3.1.The units of thermal resistance R are 0 K / W.
53
Surface in contact with a fluid at To with heat transfer coefficient ho
x
L
T1 T2
Rci R Rco
Ti
To
Qx Qx Qx Qx
Fig.3.2: Thermal Circuit for a plane wall with convective boundary conditions
Qx = hi A [Ti – T1]
where R = L /(Ak) is the thermal resistance offered by the wall for conduction and
(T2 – To)
Qx = --------------- ………………………………………..(3.6c)
Rco
Where Rco = 1 / (hoA) is the thermal resistance offered by the fluid at the surface at x = L for
convection. It follows from Equations (3.6a), (3.6b) and (3.6c) that
54
(Ti – T1) (T1 – T2) (T2 – T0)
Qx = --------------- = ------------------ = --------------
Rci R Rco
(Ti – To)
Or Qx = ------------------- ……………………………………(3.7)
[Rci + R + Rco]
The governing differential equation for one-dimensional steady state radial conduction in a
hollow cylinder of constant thermal conductivity and without thermal energy generation is
given by Eq.(2.10b) with n = 1: i.e.,
d
--- [r (dT / dr)] = 0 ………………………….(3.8)
dr
r (dT / dr) = C1
T(r) = C1 ln r + C2 ………………………..(3.9)
where C1 and C2 are constants of integration which can be determined by knowing the
boundary conditions of the problem.
Hollow cylinder with prescribed surface temperatures: Let the inner surface at r = r 1 be
maintained at a uniform temperature T1 and the outer surface at r = r 2 be maintained at
another uniform temperature T2 as shown in Fig. 3.3.
T1 = C1 ln r1 + C2 ………………………….(3.10a)
T2 = C1 ln r2 + C2 ………………………….(3.10b)
55
C1 = ---------------- = -------------------
[ln r1 – ln r2] ln (r1 / r2)
(T1 – T2)
and C2 = T1 − ------------------ ln r1
ln (r1 / r2)
T2
T1
r2
r1
Eq. (3.11) gives the temperature distribution with respect to the radial direction in a hollow
cylinder. The plot of Eq. (3.11) is shown in Fig. 3.4.
Expression for rate of heat transfer:- For radial steady state heat conduction in a hollow
cylinder without heat generation energy balance equation gives
Qr = Qr|r = r1 = Qr|r = r2
56
Hence (dT / dr)|r = r1 = {[ T2 – T1] / ln (r2 / r1) }(1/ r1).
Substituting the expressions for A(r)|r = r1 and (dT / dr)|r = r1 in Eq. (3.12) we get the expression
for rate of heat transfer as
2 π L k (T1 – T2)
Qr = -------------------------- ……………………………….(3.13)
ln (r2 / r1)
(T – T1)
(T2 – T1)
0 r / r1
1.0 r2 / r1
Fig. 3.4:Radial temperature distribution for a hollow cylinder with prescribed surface temperature
ln (r2 / r1) 1
where R = ----------------- = --------- ……………………………..(3.14b)
2πLk k Am
Where Am = (A2 – A1) / ln (A2 / A1), when A2 = 2π r2 L = Area of the outer surface of the
cylinder and A1 = 2π r1 L = Area of the inner surface of the cylinder, and Am is logarithmic
mean area.
Hollow cylinder with convective boundary conditions at the surfaces:- Let for the hollow
cylinder, the surface at r = r1 is in contact with a fluid at temperature Ti with a surface heat
transfer coefficient hi and the surface at r = r2 is in contact with another fluid at a temperature
To as shown in Fig.3.5.By drawing the thermal circuit for this problem and using the concept
57
of thermal resistance it is easy and straight forward to write down the expression for the rate
of heat transfer as shown.
(Ti – T1)
Now Qr = hiAi(Ti – T1) = 2π r1L hi (Ti – T1) = -------------- ……………..(3.15a)
Rci
(T1 – T2)
Also Qr = -------------- …………………………………………………..(3.15c)
R
Surface in contact
With fluid at Ti and
Surface heat transfer
Coefficient hi
r1
r2
Qr
Rco R Rci
(T2 – To)
And Qr = --------------- ………………………………………………….(3.15e)
Rco
1
Where Rco = --------------…………………………………..(3.15f)
(2πr2Lho)
58
(Ti – To)
Or Qr = ---------------------- ……………………….(3.16)
Rci + R + Rco
where Rci, R and Rco are given by Eqs.(3.15b), (3.15d) and (3.15f) respectively.
The governing differential equation for one-dimensional steady state radial conduction in a
hollow sphere without thermal energy generation is given by Eq.(2.10b) with n = 1: i.e.,
d
--- [r2 (dT / dr)] = 0 ………………………….(3.17)
dr
r2 (dT / dr) = C1
T(r) = − C1 / r + C2 ………………………..(3.18)
where C1 and C2 are constants of integration which can be determined by knowing the
boundary conditions of the problem.
59
C1 = ------------------- and C2 = T1 + --------------------------
[1 / r2 – 1 / r1] r1[1 / r2 – 1 / r1]
Surface at temperature T2
Surface at temperature T1
r2
r1
Fig. 3.6: Radial conduction in a hollow sphere with prescribed surface temperatures
Or T(r) – T1 [1 / r2 – 1 / r]
----------------- = ---------------------- ……………………………(3.20)
[T1 – T2] [1 / r2 – 1 / r1]
(ii) Expression for Rate of Heat Transfer:- The rate of heat transfer for the hollow sphere is
given by
Now at any radius for a sphere A(r) = 4π r2 and from Eq. (3.20)
1
dT / dr = [T1 – T2] ------------------ (1 / r2)
[1 / r2 – 1 / r1]
Substituting these expressions in Eq. (3.21) and simplifying we get
4 π k r1 r2 [T1 – T2]
Qr = -------------------------- ……………………………………...(3.22)
60
[r2 – r1]
Where R is the thermal resistance for the hollow sphere and is given by
Hollow sphere with convective conditions at the surfaces:- Fig. 3.7 shows a hollow sphere
whose boundary surfaces at radii r1 and r2 are in contact with fluids at temperatures Ti and T0
with surface heat transfer coefficients hi and h0 respectively.
r1
Fig. 3.7: Radial conduction in a hollow sphere with convective conditions at the two boun
The thermal resistance network for the above problem is shown in Fig.3.8
Where Qci = heat transfer by convection from the fluid at Ti to the inner surface of the hollow
sphere and is given by
[Ti – T1]
Qci = hi Ai [Ti – T1] = --------------- ……………………..(3.25)
Rci
61
Qr Qco
Ti Qci To
Rci R Rco
.
Fig. 3.8: Thermal circuit for a hollow sphere with convective boundary conditions
Rci = 1 / (hiAi) = the thermal resistance for convection for the inside surface
And Qco = Rate of heat transfer by convection from the outer surface of the sphere to the
outer fluid and is given by
[T2 – T0]
Qco = ho Ao [T2 – To] = --------------- ……………(3.26a)
Rco
[Ti – To]
62
Qr = ---------------------- …………………………………………(3.27)
[Rci + R + Rco]
There are many engineering applications in which heat transfer takes place through a
medium composed of several different layers, each having different thermal conductivity.
These layers may be arranged in series or in parallel or they may be arranged with combined
series-parallel arrangements. Such problems can be conveniently solved using electrical
analogy as illustrated in the following sections.
Composite Plane wall:- (i) Layers in series: Consider a plane wall consisting of three
layers in series with perfect thermal contact as shown in Fig. 3.10.The equivalent thermal
resistance network is also shown. If Q is the rate of heat transfer through an area A of the
composite wall then we can write the expression for Q as follows:
L1 L2 L3
k1 k2 k3
T1 T2 T3 T4
Rci R1 R2 R3 Rco
Ti T1 T2 T3 T4 To
Fig. 3.10: A composite plane wall with three layers in series and the equivalent thermal
resistance network
(Ti – T1) (T1 – T2) (T2 – T3) (T3 – T4) (T4 – Tco)
Q = -------------- = --------------- = ------------- = ------------ = ----------------
Rci R1 R2 R3 Rco
63
Or Q = --------------------------------- = ------------…………………………….(3.28)
Rci + R1 + R2 + R3 + Rco Rtotal
If U is the overall heat transfer coefficient for the composite wall shown in Fig. (3.10) then
Q = U A (Ti – To) ………………………………...(3.29)
Comparing Eq. (3.28) with Eq. (3.29) we have the expression for U as
1
U = --------------- ……………………………………..(3.30)
A Rtotal
1 1
Or U = --------------------------------= ----------------------------------------------------------------
A [ Rci + R1 + R2 + R3 ] A[1/(hiA) + L1/(Ak1) + L2/(Ak2) + L3/(Ak3)+ 1/(hoA)]
1
Or U = --------------------------------------------……………………………………(3.31)
[ 1/hi + L1 / k1 + L2 / k2 + L3 / k3 + 1/ho]
(ii) Layers in Parallel:- Fig.3.11 shows a composite plane wall in which three layers are
arranged in parallel. Let ‘b’ be the dimension of these layers measured normal to the plane of
the paper. Let one surface of the composite wall be in contact with a fluid at temperature T i
and surface heat transfer coefficient hi and the other surface of the wall be in contact with
another fluid at temperature To with surface heat transfer coefficient ho. The equivalent
thermal circuit for the composite wall is also shown in Fig. 3.11. The rate of heat transfer
through the composite wall is given by
Q = Q1 + Q2 + Q3 ………………………….(3.32)
Where R1 = {L / (H1bk1)}
F
64
L H1
k2
H3
k3
b
Q1
R1
Ti
T1 Q2 T2 To
Rci R2 Rco
Q3
R3
FFig. 3.11: Schematic and equivalent thermal circuit for a composite wall with
layers in parallel
(T1 – T2)
Similarly Q2 = -------------- …………………………………………………(3.33b)
R2
Where R2 = {L / (H2bk2)}
and
(T1 – T2)
and Q3 = -------------- = ……………………….. ………………………….(3.33c)
R3
Where R3 = {L / (H3bk3)}
65
Where 1 / Re = 1/R1 + 1/R2 + 1/R3
Composite Coaxial Cylinders :- Fig. 3.12. shows a composite cylinder having two layers in
series. The equivalent thermal circuit is also shown in the figure.The rate of heat
k1
r2 r3
r1
Ti T1 T2 T3 To
Q Rci R1 R2 Rco
Fig. 3.12: Schematic and thermal circuit diagrams for a composite cylinder
(Ti – T1) (T1 – T2) (T2 – T3) (T3 – To) (Ti – T0)
Now Q = ------------- = ------------ = ----------- = ------------- = -----------------------------
Rci R1 R2 Rco [Rci + R1 + R2 + Rco]
……………..(3.36)
66
1 1
Where Rci = 1 / [hiAi] = -------------- ; R1 = ---------- ln (r2 / r1)
2 π r1L hi 2 π L k1
1 1
Rco = 1 / [hoAo] = -------------- ; R2 = ---------- ln (r3 / r2)
2 π r3L ho 2 π L k2
Overall Heat Transfer Coefficient for a Composite Cylinder:- For a cylinder the area of
heat flow in radial direction depends on the radius r we can define the overall heat transfer
coefficient either based on inside surface area or based on outside surface area of the
composite cylinder. Thus if Ui is the overall heat transfer coefficient based on inside surface
area Ai and Uo is the overall heat transfer coefficient based on outside surface area Ao then
(Ti – T0)
Now UiAi (Ti – To) = -----------------------------
[Rci + R1 + R2 + Rco]
Substituting the expressions for Ai, Rci,R1,R2 and Rco in the above equation we have
1
2 π r1L Ui = --------------------------------------------------------------------------------------------
[1 /(2πr1Lhi) + {1/(2πLk1)}ln (r2 / r1) + {1/(2πLk2)}ln (r3 / r2) + 1/(2πr3Lho)]
1
Or Ui = ------------------------------------------------------------------------- ……..(3.38)
[ 1/hi + (r1 / k1) ln (r2/r1) + (r1/k2) ln (r3/r2) + (r1/r3) (1/ho) ]
1
Uo = ------------------------------------------------------------------------------ …..(3.39)
[(r3/ r2) (1/hi ) + (r3 / k1) ln (r2/r1) + (r3/k2) ln (r3/r2) + (1/ho) ]
67
Composite Concentric Spheres:- Fig.3.13 shows a composite sphere having two layers with
the inner surface of the composite sphere in contact with fluid at a uniform temperature Ti
and surface heat transfer coefficient hi and the outer surface in contact with another fluid at a
uniform temperature To and surface heat transfer coefficient ho. The corresponding thermal
circuit diagram is also shown in the figure.
k2
k1
r1 r1
r2
r3
Ti T1 T2 T3
Q Rci R1 R2 Rco
Eq. (3.36) is also applicable for the composite sphere of Fig. 3.13 except that the expression
for individual resistance will be different. Thus
( T i−T 0 )
Q= …………………………….(3.40)
( R ci + R1 + R2 + Rco )
1 1 ( r 2−r 1 ) 1 1
where Rci = = ; R1 = ; Rco = =
( hi A i ) 4 π r 21 h i 4 π k1 r1 r2 ( h0 A 0 ) 4 π r 23 h 0
;
68
( r 3−r 2 )
R1 =
4 π k2 r2 r3
Example 3.1: Consider a plane wall 100 mm thick and of thermal conductivity 100
W/(m-K). Steady state conditions are known to exist with T1 = 400 K and T2 = 600 K.
Determine the heat flux (magnitude and direction) and the temperature gradient dT/dx
for the coordinate system shown in Fig. P3.1.
T(x) T(x)
T(x)
T2 T2
T2
T1 T1
T1
L L L
x
x x
(a) (b) (c)
T2
T1
69
(b) Sketch for example 3.1(b):
Solution:It can be seen fromT(x)
the figure shown that temperature is decreasing with increase in x or in othe
Therefore dT/dx = (T1 – T2)/L
= (400 – 600) / 0.1 = ─ 2000 0 C / m.
qx = ─ k dT/dx = ─ 100 * (─ 2000)
= + 200,000 W / m2.
+ ve sign for qx indicates that heat transfer is taking place in the + ve direction of x.
T2
T1
L
= 2000 0C / m.
Example 3.2.Fig. P3.2 shows a frustum of a cone (k = 3.46 W/m-K). It is of circular cross
section with the diameter at any x is given by D = ax, where a = 0.25. The smaller cross
section is at x1 = 50 mm and the larger cross section is at x 2 = 250 mm. The corresponding
surface temperatures are T1 = 400 K and T2 = 600 K. The lateral surface of the cone is
completely insulated so that conduction can be assumed to take place in x-direction only.
(i) Derive an expression for steady state temperature distribution, T(x) in the solid and
(ii) calculate the rate of heat transfer through the solid.
70
Sketch for example 3.2:
T2 T1
.Known: X1 = 50 mm;X2 = 250 mm; D= ax with a = 0.25;T1 = 400 K; T2 = 600 K;
k = 3.46 W/m-K
Assumptions:
D2 (i) Steady state conduction
D1 (ii) One dimensional conduction through the cone; (iii) therm
D
X1
x
X2
Solution: By Fourier’s law, the rate of heat transfer in x-direction across any plane at a
distance x from the origin ‘o’ is given by
Qx = ─ k Ax (dT/dx).
For steady state conduction without heat generation Qx will be a constant. Also at any x,
D = ax. Therefore
dT = ─ (4/πa2k) Qx (dx/x2)
( 4 Qx)
Or T = T1 ─ --------------- [(1 / x ) – (1 / X1)] …………(1)
71
(π a2 k)
At x = X2, T = T2. Substituting this condition in Eq.(1) and solving for Qx we get
(π a2 k) (T2 – T1)
Qx = ----------------------------- …………………….(2)
4 (1/X2 – 1/X1)
Substituting this expression for Qx in Eq. (1) we get the temperature distribution in the cone
as follows:
Substituting the given numerical values for X1, X2, T1 and T2 in Eq.(3) we get the temperature
distribution as follows:
(600 – 400) [ 1/ x – 1/0.05]
T(x) = 400 + ------------------------------------
[ 1/0.25 – 1/0.05}
Example 3.3.A plane composite wall consists of three different layers in perfect thermal
contact. The first layer is 5 cm thick with k = 20 W/(m-K),the second layer is10 cm thick
with k = 50 W/(m-K) and the third layer is 15 cm thick with k = 100 W/(m-K).The outer
surface of the first layer is in contact with a fluid at 400 0C with a surface heat transfer
coefficient of 25 W/(m2–K),while the outer surface of the third layer is exposed to an
ambient at 30 0C with a surface heat transfer coefficient of 15 W/(m2-K).Draw the
equivalent thermal circuit indicating the numerical values of the thermal resistances and
calculate the heat flux through the composite wall. Also calculate the overall heat transfer
coefficient for the composite wall.
72
Sketch for example 3.3:
L1 L2 L3
hi h0
k1 k2 k3
Ti
To
Q
Rci R1 R2 R3 Rc0
Ti = 400 0 C ; T0 = 30 0 C.
Assumptions: (i) Steady state conduction (ii) One dimensional conduction through the
composite wall; (iii) thermal conducvity of each layer is constant.
Solution:
1
Rci = 1 / (hiA1) = ---------------- = 0.04 m2 – K / W (A1 = A2 = A3 = A4 = 1 m2)
25 * 1
0.05
R1 = L1 /(k1A1) = --------------- = 0.0025 m2 – K / W.
20 * 1
0.10
R2 = L2 / (k2A2) = ---------------- = 0.002 m2 – K / W.
50 * 1
73
0.15
R3 = L3/ (k3A3) = ------------------ = 0.0015 m2 – K / W.
100 * 1
1
Rco = 1 / (h0A4) = ---------------- = 0.067 m2 – K / W.
15 * 1
Or ∑R = 0.113 m2-K/W.
= 3274.34 W / m2.
If ‘U’ is the overall heat transfer coefficient for the given system then
Q 1 1
U = ---------------- = ------------- = -------------- = 8.85 W / (m2 – K).
(Ti – T0) ∑R 0.113
Example 3.4. A composite wall consisting of four different materials is shown in Fig
P3.10. Using the thermal resistance concept determine the heat transfer rate per m2 of
the exposed surface for a temperature difference of 300 0 C between the two outer
surfaces. Also draw the thermal circuit for the composite wall.
74
T1
k1 k2 k2 k4
2m
k3
k31 m
4cm 10 cm 5 cm
Assumptions: (i) Steady state conduction (ii) One dimensional conduction through the
composite wall; (iii) thermal conducvity of each layer is constant.
Solution:
0.04
0
R1 = L1 / (A1k1) = ---------------- = 0.0002 C / W.
2 * 100
0.10
0
R2 = L2 / (A2k2) = ---------------- = 2.50 C / W.
1 * 0.04
0.10
0
R3 = L3 / (A3k3) = --------------- = 0.005 C / W.
1 * 20
0.05
0
R4 = L4 / (A4k4) = ------------------ = 0.00036 C / W.
2 * 70
75
Q R2
R1 R4
R3
Thermal potential = T1 – T4
R2 and R3 are resistances in parallel and they can be replaced by a single equivalent resistance
Re, where
R2 R3 2.5 * 0.005
1 / Re = 1 / R2 + 1 / R3 or Re = --------------- = ----------------------- = 0.00499 0C/W
(R2 + R3) (2.5 + 0.005)
76
ro heater
r1
ri
hi,Ti
Qi kA
Q0 ho,To
kB
Ri Rci Qi
Qtotal
Ro Rco Qo
Find: (b) Expression to determine Temperature of heater Th; (ii) q0 / qi
Assumptions: (i) Steady state conduction (ii) One dimensional radial conduction through the
composite wall; (iii) thermal conductivity of each layer is constant; (iii)No contact resistance
between the surfaces of the two layers
1 1 1
Ri = ------------- ln (r1/ri) ; Rci = ------------------ = ---------------------
(2π LkA) hi A i (2π ri L hi)
1 1 1
Ro = ------------- ln (ro/r1) ; Rco = ------------------ = ---------------------
(2π LkB) ho A o (2π ro L ho)
77
(Th – Ti) (Th – To)
Therefore q h = ------------------------------- + --------------------------------
[ (r1/kA) ln (r1/ri) + r1/(rihi)] [(r1/kB) ln (ro/r1) + r1/(roho)
The temperature Th of the heater can be obtained from the above equation.
Example 3.6. A hollow aluminum sphere with an electrical heater in the centre is used to
determine the thermal conductivity of insulating materials. The inner and outer radii of the
sphere are 15 cm and 18 cm respectively and testing is done under steady state conditions
with the inner surface of the aluminum maintained at 250 0 C. In a particular test, a spherical
shell of insulation is cast on the outer surface of the aluminum sphere to a thikness of 12 cm.
The system is in a room where the air temperature is 20 0 C and the convection coefficient is
30 W/(m2 – K). If 80 W are dissipated by the heater under steady state conditions, what is the
thermal conductivity of the insulating material?
ho,To
r3 Known: r1 = 0.15 m ; r2 =
0.18
r2 m ; r3 = 0.18 + 0.12 =
0.3 mr1
; k1 = 204 W/(m-K) from tables; k2 = 0.30 W/(m-K); ho = 30 W/(m2-K);Q = 60 W; T1 =
Find: k2
T1
k1
k2
Assumptions: (i) Steady state conduction (ii) One dimensional radial conduction through
the composite wall; (iii) thermal conductivity of each layer is constant; (iii)No contact
resistance between the surfaces of the two layers
Solution:
(r2 – r1) (0.18 – 0.15)
R1 = ---------------- = ------------------------------ = 4.335 x 10 ─ 4 0
C / W.
4π k1 r1 r2 4 π * 204 * 0.18 * 0.15
78
(r3 – r2) (0.30 – 0.18)
0
R2 = ---------------- = ------------------------------ = 0.177 / k2 C / W.
4π k2 r2 r3 4 π * 0.30 * 0.30 * 0.18
1 1
0
Rco = 1 / (hoAo) = ------------------- = --------------------- = 0.0295 C / W.
4π r32 ho 4π * (0.3)2 * 30
(T1 – To)
Q = -------------------- or R2 = (T1 – To) / Q – (R1 + Rco)
R1 + R2 + Rco
Example 3.7. In a hollow sphere of inner radius 10 cm and outer radius 20, the inner surface
is subjected to a uniform heat flux of 1.6 x 10 5 W/m2 and the outer surface is maintained at a
uniform temperature of 0 0C.The thermal conductivity of the material of the sphere is 40 W /
(m – K).Assuming one-dimensional radial steady state conduction determine the temperature
of the inner surface of the hollow sphere.
T0
R2
q0
R1
Find: T/r= R1
Assumption: (i)Steady state one dimensional radial conduction without heat generation; (ii)
thermal conductivity of the sphere is constant.
79
Solution: The governing equation for one-dimensional steady-state radial conduction in a
sphere without heat generation is given by
d/dr ( r2 dT / dr ) = 0 …………………………………..(1)
(ii) at r = R2 T(r) = 0.
r2 (dT/dr) = C1
or dT / dr = C1 / r2 …………………………(2)
─ kC1 / R12 = q0
Or C1 = ─ q0 R12 / k
Or C2 = C1 / R2 = ─ (q0R12) / (kR2)
q0 R12 q0 R12
T(r) = -------------- ─ -------------------
kr k R2
80
Or T(r) = (40 / r) ─ 200
3.2.5. Thermal Contact Resistance: In the analysis of heat transfer problems for composite
medium it was assumed that there is “perfect thermal contact” at the interface of two layers.
This assumption is valid only the two surfaces are smooth and they produce a perfect contact
at each point.But in reality, even flat surfaces that appear smooth to the naked eye would be
rather rough when examined under a microscope as shown in Fig. 3.14 with numerous peaks
and valleys.
T2
T1
Rcont
LA LB
Tc2 T2
Fig.3.14: Temperature drop across a contact resistance
The physical significance of thermal contact resistance is that the peaks will form good
thermal contact, but the valleys will form voids filled with air.As a result the air gaps act as
insulation because of poor thermal conductivity of air.Thus the interface offers some
resistance to heat conduction and this resistance is called the “thermal contact
resistance,Rcont”. The value of Rcont is determined experimentally and is taken into account
while analyzing the heat conduction problems involving multi-layer medium.The procedure
is illustrated by means of a few examples below.
Example 3.8. A composite wall consists of two different materials A [k = 0.1 W/(m-k)] of
thickness 2 cm and B[ k = 0.05 W/(m-K)] of the thickness 4 cm. The outer surface of layer A
is in contact with a fluid at 200 0C with a surface heat transfer coefficient of 15 W/(m 2-K) and
the outer surface of layer B is in contact with another fluid at 50 0 C with a surface heat
transfer coefficient of 25 W/(m2-K). The contact resistance between layer A and layer B is
0.33 (m2-K) /W. Determine the heat transfer rate through the composite wall per unit area of
the surface. Also calculate the interfacial temperatures and the inner and outer surface
temperature.
81
Sketch for example 3.8:
Known: Ti = 200 0 C ; T0 = 50 0 C ; h0,T0
hi,Ti
Rcont
hi = 15 W/(m2 – K) ; h0 = 25W/(m2– K)
kA kB
kA = 0.1 W/(m-K) ; kB = 0.05 W/(m-K)
LA LB
Rcont = 0.33 (m2 – K) /W.
Assumptions: (i)T1
Steady state conduction (ii) One dimensional conduction through the composite wall; (i
Tc1
Tc2 T2
(200 – 50)
Heat flux = q = (Ti – To) / ∑R = ------------------ = 104.4 W/m2
1.437
82
Check : To = T2 – q Rco = 54.15 – (104.4 * 0.04) = 49.97 0 C
Example 3.9. A very thin electric heater is wrapped around the outer surface of a long
cylindrical tube whose inner surface is maintained at 5 0C. The tube wall has inner and
outer radii of 25 mm and 75 mm respectively and a thermal conductivity of 10 W/(m-K).
The thermal contact resistance between the heater and the outer surface of the tube
per unit length is 0.01 (m-K) / W. The outer surface of the heater is exposed to a fluid
with a temperature of – 10 0C and a convection coefficient of 100 W/(m2-K). Determine
the heater power required per metre length of the tube to maintain a heater
temperature of 25 0C.
h,T∞
Heater
r2
r1
Qtube T1 T2
Qamb
Qamb
Rco
Qtotal
Rcont Rcond Qtube
Assumptions: (i) Steady state conduction (ii) One dimensional conduction through the
composite wall; (ii) thermal conductivity of each layer is constant ; (iii)No contact resistance
between the surfaces of the two layers
83
Solution:
1
Rco = 1 / (h Ao) = ----------------------- = 0.0212 (m- K) / W.
100 * 2π * 0.075
1 1
Rcond = ------------- ln (r2 / r1) = ------------------ ln (0.075 / 0.025) = 0.0175 (m-K) / W.
2π L k (2* π * 1 * 10)
[25 – (─ 10)]
Qamb = (T2 - T∞) / Rco = ------------------- = 1651 W / m.
0.0212
3.3. One Dimensional Steady State Conduction With Heat Generation: The governing
equation for one – dimensional steady state conduction in solids which are generating is
given as follows.
The following examples illustrate the method of analysis of steady state heat conduction
In solids generating heat.
Example 3.10. A plane wall of thickness L and thermal conductivity k has one of its
surfaces insulated and the other surface is kept at a uniform temperature T1. Heat is
generated in the wall at a rate q’’’(x) where q’’’(x) = q0 cos{(πx) / (2L)} W / m3
where q0 is a constant.
(a) Develop an expression for one-dimensional steady state temperature distribution in the
solid and (b) develop an expression for the temperature of the insulated surface.
84
L
Insulated T1
T = T(x)
q’’’= q0 cos (πx/2L)
Known: (i) Expression for heat generation q’’’ = q0 cos ( 2πxL ) W/m3;(ii) the surface at x
= 0 is insulated i.e. (dT/dx)/x=0 = 0; (iii) T = T1 at x = L
Find: (a) expression for temperature distribution in the solid T(x); (ii) T/x = 0
Assumptions: (i) One dimensional steady state conduction ; (ii) solid is of constant thermal
conductivity
Substituting the given expression for q’’’ the above equation reduces to
(ii) at x = L, T = T1
Substituting the values of C1 and C2 in Eq. (3) we get the temperature distribution as
85
T(x) = -(π/2L)2 (q0/k) cos (πx/2L) + T1 …………………..(4).
T = T(r)
q’’’
h, T∞
T∞ = 500C; h = 60 W/(m2 – K)
Assumptions: (i) one dimensional radial conduction; (ii) steady state conduction; (iii)
thermal conductivity of the rosd is constant
86
Integrating once w.r.t. r we get
r dT/dr + (q′′′ r2) / 2k = C1
q′′′ R q′′′ R2
Or C2 = ----------- + -------------- + T∞
2h 4k
Substituting the expressions for C1 and C2 in Eq.(3) we get the temperature distribution in the
cylinder as
q′′′ R2
T(r) = T∞ + ------------- [ 1 –(r/R)2 ] + (q’’’R) / 2h …………………(4)
4k
3 * 10 5 * (0.05)2
Now q′′′ R2 / (4 k) = ----------------------- = 18.75 0 C.
4 * 10
5
3 x 10 x 0.05
q′′′ R / (2h) = ------------------------ = 125 0 C
2 x 60
87
Example 3.12. In a cylindrical fuel element for a gas-cooled nuclear reactor, the heat
generation rate within the fuel element due to fission can be approximated by the
equation
q’’’ = qo [1 – (r/R)2] W/m3,
where R is the outer radius of the fuel element and q o is a constant. The outer surface of the
cylinder is maintained at a uniform temperature T o. Assuming one-dimensional radial
conduction obtain an expression for the temperature distribution in the element. If R = 2 cm,
k = 10 W/(m-K) and qo = 1.16 x 10 5 W/m3, what would be the temperature difference
between centre temperature and the outer surface temperature.
T0
q’’’ = q0 [ 1 – (r/R)2]
Assumptions: (i) one dimensional steady state conduction; (ii) Thermal conductivity of the
cylinder is constant
q0 r [ 1 – (r/R)2 ]
d/dr(r dT/dr) + ---------------------- = 0 ……………………..(1)
k
88
Boundary conditions are: (i) at r = 0, dT / dr = 0 (axis of symmetry)
(ii) at r = R, T = T0
q0 r2 r4
(r dT/dr) + ---- [----- ─ -----------] = C1
K 2 4R2
Or q0 r r3 C1
dT/dr + ---- [----- ─ -----------] = ----------- ………..(2)
K 2 4R2 r
Integrating once again w.r.t. r we have
q0 r2 r4
T(r) = ─ ---------- [---------- ─ --------- ] + C1 ln r + C2 .....................(3)
k 4 16R2
Condition(i) in Eq.(2) gives
0 + 0 = C1 / 0 or C1 = 0.
q0 R2 R4
T0 = ─ ------ [ -------- ─ ---------- ] + C2
k 4 16 R2
3 q0 R2
Or C2 = T0 + ----------------
16 k
Substituting the expressions for C1 and C2 in Eq. (3) we have
q0 r2 r4 3 q0 R2
T(r) = ─ ---------- [---------- ─ ---------] + T0 + ---------------
k 4 16R2 16 k
Or q0 r2 r4 3 q0 R2
T(r) ─ T0 = ─ ---------- [---------- ─ ---------] + ---------------
k 4 16R2 16 k
89
Example3.13. Develop an expression for one-dimensional radial steady state
temperature distribution in a solid sphere of radius R in which heat is generated at a
rate given by
Solution: The governing differential equation to find the one-dimensional radial steady state
temperature distribution in a sphere with heat generation is given by
Multiplying throughout by r2 and substituting the given expression for q′′′ we have
(ii) at r = R, T = T0.
q0 r3 r4
2
(r dT/dr) + ----- [ ------ ─ ---------- ] = C1
k 3 4R
q0 r r2
Or dT/dr = ─ ----- [ ------ ─ ---------- ] + C1 / r2 ………………….(2)
k 3 4R
Integrating once again w.r.t. r we get
90
q0 r2 r3
Or T(r) = ─ ----- [ ------ ─ ---------- ] ─ C1 / r + C2 …………….(3)
k 6 12R
or C2 = T0 + q0 R2 / 12k
q0 r2 r3 q0 R2
T(r) = ─ ---- [ ------ ─ ---------- ] + T0 + -----------
k 6 12R 12 k
q0 R2
or T(r) = T0 + ---------- [ 1 ─ 2 (r/R)2 + (r/R)3 ]
12k
Example 3.14. A plane wall of thickness 2L is generating heat according to the law
where qo, β, and Tw are constants and T is the temperature at any section x from the mid-
plane of the wall. The two outer surfaces of the wall are maintained at a uniform temperature
Tw. Determine the one-dimensional steady state temperature distribution, T(x) for the wall.
Tw
Tw
2L
91
Known: The two outer surfaces of the wall maintained at Tw; Solid is generating heat at a
rate given by q’’’ = q0 [1 – β(T – Tw)].
Assumptions: (i) One dimensional steady state conduction; (ii) Solid is of constant thermal
conductivity
or d2θ / dx2 − qo βθ /k = − qo / k
where m2 = q0 β /k ……………………………...(1b)
θh(x) = A1 e mx + A2 e − mx…………………………….(4)
The term qo/k makes the governing differential equation non-homogeneous. Since this is a
constant θp(x) is also assumed to be constant. Thus let θp(x) = C, where C is a constant.
92
Substituting this solution in Eq. (5) we get
m 2C = qo / k
Or C = qo /(km2)
θ(x) = A1 e mx + A2 e − mx + 1 / β …………………..(7)
The constants A1 and A2 in Eq.(7) can be determined by using the two boundary conditions,
which are:
(ii) at x = L, T = Tw ; i.e., θ = 0
Or A1 = A2.
(1 / β)
Or A1 = ----------------------------
[e mL + e − mL]
Substituting the expressions for A1 and A2 in Eq. (7) we get the temperature distribution in
the plane wall as
(1 / β )
θ(x) = T(x) – Tw = -------------------- [ e mx + e − mx ] − 1 / β
[e mL + e − mL]
1 e mx + e − mx
Or T(x) – Tw = ---- [ ----------------- − 1]
β e mL + e − mL
93
3.4. Critical Radius of Insulation:- For a plane wall adding more insulation will result in a
decrease in heat transfer as the area of heat flow remains constant .But adding insulation to a
cylindrical pipe or a conducting wire or a spherical shell will result in an increase in thermal
resistance for conduction at the same will result in a decrease in the convection resistance of
the outer surface because of increase in surface area for convection. Therefore the heat
transfer may either increase or decrease depending on the relative magnitude of these two
resistances.
3.4.1.Critical Radius of Insulation for Cylinder:- Let us consider a cylindrical pipe of outer
radius rs maintained at a constant temperature of T s. Let the pipe now be insulated with a
material of thermal conductivity k and outer radius r. Let the outer surface of the insulation
be in contact with a fluid at a uniform temperature T ∞ with a surface heat transfer coefficient
h. Then the thermal circuit for this arrangement will be as shown in Fig.3.15.
Surface
r in contact with a fluid at T∞ and surface hea
Ts
rs
Ts
Rins Rco
Fig.3.15: Schematic of a cylindrical pipe covered with an insulation and exposed to an ambient and
The rate of heat transfer from the pipe to the ambient is given by
It can be seen from Eq. (3.44) that if Ts and h are assumed not to vary with ‘r’ then Q
depends only on r and the nature of variation of Q with r will be as shown in Fig.3.16.
The value of r at which Q reaches a maximum can be determined as follows.
94
(Ts – To)
Eq. (3.44) can be written as Q = -----------
F(r)
ln (r / rs) 1
where F(r) = ---------- + --------------
2πLk 2πrLh
Qmax
Qbare
r
rs rcr = k / h
Or r = k / h.
The above expression for critical radius of insulation has been arrived with the assumption
that the change in outer radius of insulation r will have no effect on the inside surface
temperature of insulation or on the outer surface heat transfer coefficient h. Sparrow
considered a heat transfer coefficient that varies as
95
h – r –m │Ts − T∞│n …………………………(3.46a)
For this case, it is found that the heat transfer is maximized when
(1−m ) k
r = rcr = ……………………….(3.46b)
(1+ n ) h
(1−m )
The quantity is a correction factor (≤ 1) accounting for the r and ΔT
(1+ n )
dependencies of
h. As an example, consider forced convection flow across a cylinder. For Reynolds number
based on diameter between 4000 and 40,000, m = 0.382 and n=0. Correspondingly
Eq.(3.46b) gives the correction factor
(1−m )
= 0.618
(1+ n )
As a second example for free convection about a horizontal cylinder for which m = n = 0.25,
the correction factor will be 0.6.
It can be seen from Fig.(3.16) that if the outer radius of the bare tube or bare wire is
greater than the critical radius then, any addition of insulation on the tube surface
decreases the heat loss to the ambient. But if the outer radius of the tube is less than the
critical radius , the heat loss will increase continuously with the addition of insulation
until the outer radius of insulation equals the critical radius. The heat loss becomes
maximum at the critical radius and begins to decrease with addition of insulation beyond
the critical radius.
The value of critical radius rcr will be the largest when k is large and h is
small. The lowest value of h encountered in practice is about 5 W/(m2 – K) for free
convection in a gaseous medium and the thermal conductivity of common insulating
materials is about 0.05 W/(m – K). Hence the largest value of rcr that we may likely to
encounter is given by
0.05
rcr = --------- = 0.01 m = 1 cm
5
The critical radius would be much less in forced convection (it may be as low as 1mm)
because of large values of h associated with forced convection. Hence we can insulate hot
water or steam pipes freely without worrying about the possibility of increasing the heat
loss to the surroundings by insulating the pipes.
The radius of electric wires may be smaller than the critical radius. Therefore, the
plastic electrical insulation may enhance the heat transfer from electric wires, there by
keeping their steady operating temperatures at lower and safer levels.
3.4.2.Critical Radius Insulation for a Sphere:- The analysis described above for cylindrical
pipes can be repeated for a sphere and it can be shown that for a sphere the critical radius
of insulation is given by
2k
96
rcr = ------- …………………………………..(3.47)
h
h,T∞ h, T∞
(a) Conductor without Insulation. (b) Conductor with critical thickness of
Insulation
Known: D = 0.008 m; Ts = 130 0C ; for insulation k = 1.2 W/(m-K); h = 120 W/(m2-K);
T∞ = 30 0C
Assumptions: (i) one dimensional steady radial conduction (ii) the thermal conductivity of
insulation is constant (iii) by adding the insulation the outer surface temperature of the wire
will not change; (iv) the surface heat transfer coefficient for wire with and without insulation
remain same
Solution:(a) When the conductor is bare the rate of heat loss to the ambient is given by
1 1
R insulation = -------------- ln (Dc / D) = ----------------------- ln (0.02 / 0.008)
2π L k 2 π * 1.0 * 1.2
= 0.1215 (m – 0 C) / W.
1 1
97
Rco = 1 / (h Ac) = ---------------- = ------------------------ = 0.133 (m – 0 C)/W.
π Dc L h π * 0.02 * 1 8 120
Example 3.16. An electrical current of 700 A flows through a stainless steel cable
having a diameter of 5 mm and an electrical resistance of 6x10 ─ 4 ohms per metre
length of the cable. The cable is in an environment at a uniform temperature of 30 0 C
and the surface heat transfer coefficient of 25 W/(m2 – 0C).
(a) What is the surface temperature of the cable when it is bare?
(b)What thickness of insulation of k = 0.5 W/(m – K) will yield the lowest value of the
maximum insulation temperature? What is this temperature when this thickness is used?
Known: Current flow through the cable = I = 700 A; diameter of cable = D = 5 mm;
Electrical resistance of cable / metre length = ρ = 6* 10−4 Ω/m; h= 25 W/(m2-K);
T∞ = 300C; for insulation k =0.5 W/(m-K)
Find: (a) Ts when the wire is bare; (ii) thickness of insulation for lowest value of maximum
insulation temperature; (iii) lowest value of maximum insulation temperature
Assumptions: i) one dimensional steady radial conduction (ii) the thermal conductivity of
insulation is constant (iii) by adding the insulation the outer surface temperature of the wire
will not change; (iv) the surface heat transfer coefficient for wire with and without insulation
remain same
(b) When the cable is covered with insulation: Lowest value of maximum insulatiuon
temperature occurs when the outer radius of insulation is equal to critical radius.
98
1 1
Rinsulation = --------------- ln (rc / ro) = ------------------- ln (0.02 / 0.0025) = 0.662 (m-K)/W
(2π L k) (2π * 1 * 0.5)
1 1
Rco = 1 / (hAo) = ------------------- = ------------------------------ = 0.318 (m-K) / W.
(2π rcL h) (2 * π * 0.02 * 1 * 25)
(Ts - T∞)
Q = --------------------- or Ts = T∞ + Q (Rinsulation + Rco) = 30 + 294 x (0.662 +0.318)
Rinsulation + Rco
Or Ts = 318.12 0 C.
Surface
r in contact with a fluid at T∞ and surface heat transfer coefficient h
Ts
rs
T∞
Ts
Rins Rco Q
Known: Outer radius of the bare wire = r s = 1 mm = 0.001 m ; Length of the wire = L = 10
m ; outer radius of plastic insulation = r = 1 + 1 = 2 mm = 0.002 m ;
Current through the wire = I = 10 A ; Voltage drop in the wire = V = 8 V ; Ambient
temperature = T∞ = 30 0C ; Thermal conductivity of the plastic cover = k = 0.15 W /(m– K) ;
Surface heat transfer coefficient = h = 24 W /(m2 – K).
99
Find: (i) Interface temperature = Ts ; (ii) Whether Ts increases or decreases when the
thickness of insulation is doubled.
Assumptions: i) one dimensional steady radial conduction (ii) the thermal conductivity of
insulation is constant (iii) by adding the insulation the outer surface temperature of the wire
will not change; (iv) the surface heat transfer coefficient for wire with and without insulation
remain same
Solution: (i) Q = VI = 8 x 10 = 80 W.
Since rcr > r, increasing the thickness of plastic insulation will increase the heat transfer rate if
Ts is held constant or for a given heat transfer rate the interface temperature Ts will decrease
till the critical radius is reached. Now when the thickness is doubled then
r = 3 mm = 0.003 m . Therefore
ln ( 0.003 / 0.001)
Rins = ----------------------------- = 0.1166 K / W
2 * π * 10 * 0.15
1
Rco = ------------------------------ = 0.221 K / W
2 x π x 10 x 0.003 x 24
and Ts = 30 + 80 * 0.3376 = 57 0C
100
3.5.1 .Introduction: Convection: Heat transfer between a solid surface and a moving fluid is
governed by the Newton’s cooling law: Q = hA(Ts-T∞), where Ts is the surface temperature
and T∞ is the fluid temperature. Therefore, to increase the convective heat transfer, one can
• Increase the temperature difference (Ts-T∞) between the surface and the fluid.
• Increase the convection coefficient h. This can be accomplished by increasing the fluid flow
over the surface since h is a function of the flow velocity and the higher the velocity, the
higher the h. Example: a cooling fan.
• Increase the contact surface area A. Example: a heat sink with fins.
Many times, when the first option is not in our control and the second option (i.e. increasing
h) is already stretched to its limit, we are left with the only alternative of increasing the
effective surface area by using fins or extended surfaces. Fins are protrusions from the base
surface into the cooling fluid, so that the extra surface of the protrusions is also in contact
with the fluid. Most of you have encountered cooling fins on air-cooled engines
(motorcycles, portable generators, etc.), electronic equipment (CPUs), automobile radiators,
air conditioning equipment (condensers) and elsewhere.
After making the assumption of One-Dimensional Conduction, this equation reduces to the
form:
This is a second order, ordinary differential equation and will require 2 boundary conditions
to evaluate the two constants of integration that will arise.
Consider the cooling fin shown in Fig. 4.1 :
Heat and Mass Transfer Prof. P
The fin is situated on the surface of a hot surface at T s and surrounded by a coolant at
temperature T , which cools with convective coefficient, h. The fin has a uniform cross
∞
sectional area, Ac, (This is the area through with heat is conducted.) and an overall length, L.
Note that as energy is conducted down the length of the fin, some portion is lost, by
convection, from the sides. Thus the heat flow varies along the length of the fin.We further
note that the arrows indicating the direction of heat flow point in both the x and y directions.
This is an indication that this is truly a two- or three-dimensional heat flow, depending on the
geometry of the fin. However, quite often, it is convenient to analyse a fin by examining an
equivalent one–dimensional system. The equivalent system
101
y
Fin attached
x to the primary surface and exposed to surroundings at T∞ with a heat transfer coe
Qfin
Primary surface at T0
Qc = hPΔx (T - T∞)
Δx
will involve the introduction of heat sinks (negative heat sources), which remove an amount
of energy equivalent to what would be lost through the sides by convection.
Consider a differential length of the fin. Across this segment the heat loss will be h(PΔx)(T-
T ), where P is the perimeter around the fin. The equivalent heat sink would be ─ q’’’Ac Δx
∞
which is the equation for a fin with a constant cross sectional area. This is the Second Order
Differential Equation that we will solve for each fin analysis. Prior to solving, a couple of
simplifications should be noted. First, we see that h, P, k and A c are all independent of x in
102
the defined system (They may not be constant if a more general analysis is desired.). We
replace this ratio with a constant.
Under steady state conditions, the energy balance equation for the fin element shown above
can be written as
Qx = Qx+dx + Qconv
For a fin of uniform cross sectional area Ax = Ac = constant. Also if the thermal conductivity
of the fin material is constant then the above equation reduces to
2
d T
– m2 (T - T∞) = 0 ……………………
d x2
(3.50)
The quasi one dimensional heat transfer problems are those where heat is transferred in two
directions, but temperature changes are important only in one direction. We can examine
when temperature variation across the fin can be neglected in comparison to the temperature
variation along the length of the fin.Consider the heat transfer in y direction for the fin of
103
length L, width W and uniform thickness t. At any x ,if the rate of heat transfer in x direction
is of the same order of magnitude as the rate in y direction we have
−k Ax ( ∂T∂x ) ≈ −k Ay ( ∂T∂y )
Using first order linear approximation, the above equation can be written as
( ΔT ) y
( ΔT )x
−k t W ≈ −k L W t
L
2
( ΔT ) y t 2
Or
( ΔT ) x
≈ ()
L
For fins t is very much smaller than L; therefore, for the heat transfers to be of the same order
of magnitude ,it follows that the temperature variation in y direction must be much smaller
∂T ∂T
than the temperature variation in x direction: i.e.
∂y
≪ ( ) ( )
∂x
.
Solution to the fin equation: We notice that equation (3.50) is non-homogeneous (due to
the T∞ term). Recall that non-homogeneous differential equations require both a
homogeneous and a particular solution. We can make this equation homogeneous by
introducing the fin temperature relative to the surroundings:
Let θ ≡ T - T∞ …………………………………….(3.51)
Solution for Equation (3.53): Eq. (3.53) is a second order linear homogeneous differential
equation the general solution for which is given by
104
Generally the exponential solution is used for very long fins, the hyperbolic solutions for
other cases.
Boundary Conditions:
Since the solution results in 2 constants of integration we require 2 boundary conditions. The
first one is obvious, as one end of the fin will be attached to a hot surface and will come into
thermal equilibrium with that surface.
Hence, at the fin base (x = 0),
θ =T0 - T∞ = θ0 ……………………………..(3.56a) e
of Science
The second boundary condition depends on the condition imposed at the other end of the
fin. There are various possibilities, as described below.
Very long fins:
For very long fins, the end located a long distance from the heat source will approach the
temperature of the surroundings. Hence,
θ(∞) = 0 …………………………………(3.56b)
Substitute the second condition into the exponential solution of the fin equation:
θ(∞)= 0 = Ae m∞ + Be ─m∞
The first exponential term is infinite and the second is equal to zero. The only way that this
equation can be valid is if A = 0. Now apply the second boundary condition.
θ (0) = θ 0 =B
θ(x) = θ 0 e ─mx
…………………………. (3.57)
If we wish to find the heat flow through the fin, we may apply Fourier Law:
At any x, the rate of heat conduction is given by
Q = ─ kAc(dT/dx) = ─ kAc(dθ/dx)
All the heat dissipated by the fin to the surroundings is coming from the fin base. Hence the
rate of heat transfer from the fin to the surroundings is given by
Qfin = Q|x=0
= ─ kAc(dθ/dx)x=0 ………………(3.58)
105
The fin with insulated tip:
Assume that the tip is insulated and hence there is no heat transfer at the fin tip.
Hence the boundary condition at the tip is given by
At x = L, dθ/dx = 0.
Substituting this condition and the boundary condition at x = 0 and simplifying we get the
temperature profile for this case as
cosh m(L−x )
θ(x) = θ0 ….
cosh mL
……………(3.60)
We may find the heat flow at any value of x by differentiating the temperature profile and
substituting it into the Fourier Law:
If we compare this result with that for the very long fin, we see that the primary difference in
form is in the hyperbolic tangent term. That term, which always results in a number equal to
or less than one, represents the reduced heat loss due to the shortening of the fin
Indian Institute of Science Bangalore
Other tip conditions:
We have already seen two tip conditions, one being the long fin and the other being the
insulated tip. Two other possibilities are usually considered for fin analysis: (i) a tip
subjected to convective heat transfer, and (ii) a tip with a prescribed temperature. The
expressions for temperature distribution and fin heat transfer for all the four cases are
summarized in the table below.
106
Table 3.1: Expressions for Temperature distribution and rate of heat dissipation from a
fin for different tip conditions
cosh m(L−x )
Tip Insulated θ(x) = θ0 Qfin = kmAcθ0tanh
cosh mL
mL
3.5.2.4.Performance of a fin:
The rate of heat transfer from the fin will be maximum when the entire fin surface is at the
fin base temperature (This means that the fin is having an infinite thermal conductivity): i.e.,
107
(Qfin)max = hPL(To - T∞) = hPLθ0 …………………....(3.63)
Fin Effectiveness: Fins are used to enhance the heat transfer, and the use of fins on a surface
cannot be recommended unless the enhancement in heat transfer justifies the added cost and
complexity associated with the fins. In fact, there is no assurance that by adding fins on a
surface, the heat transfer rate will enhance.The performance of the fin is judged on the basis
of enhancement in heat transfer relative to the no-fin case.To know how effective an
extended surface as compared to a bare primary surface is, another parameter called
effectiveness is used. It is given by
Q fin η h A f θ0 Af
From Eq. (3.64) ε= = = θ
h A b θ 0 h A b θ0 Ab 0
………………(3.65)
Hence, the fin effectiveness can be determined easily when the efficiency of the fin is known,
or vice-versa.
Considerations in the design and selection of fins: The rate of heat transfer from a very
long fin of uniform cross section under steady state conditions is given by Eq.
(3.59).Substituting this expression in Eq.(3.65) we get
(3.66)
ε long =
√ kP
hA
…………………
Several important conclusions can be drawn from Eq.(3.66) which helps in design and
selection of fins for various applications.
108
(i) The therml conductivity k of the fin material should be as high as possible.Hence
fins are made of materials like copper.aluminium are iron. The most widely used
material is aluminium because of its low cost and weight and its resistance to
corrosion.
(ii) The ratio of the perimeter to the cross-sectional area of the fin P/A should be as
high as possible. This criterion is satisfied by employing thin plate fins and
slender pin fins.
(iii) The use of fins is most efficient in applications involving low convection heat
transfer coefficients. Thus the use of fins is easily justified when the medium is
gas instead of a liquid and the heat transfer mechanism is by free convection
(iv) rather than by forced convection.Therfore in a liquid to gas heat exchangers like
automobile radiators, fins are placed on the gas side of the heat exchangers.
Find: (i) Qfin if the rod is long; (ii) Qfin if L = 10 cm and negligible heat loss from the tip;
(iii)Qfin if L = 25 cm and with heat loss from the tip
Assumptions: (i) one dimensional steady state conduction along the length of the fin; (ii)
thermal conductivity of the rod is constant
or m = (√ 50∗0.02
4∗64
) = 16 m−1
109
(ii) Rod of finite length with negligible heat loss from the tip:
For rod of finite length with negligible heat loss from the tip we have
= 23.16 W
(iii) Rod of finite length with heat loss from the tip:
When the heat loss from the rod tip is not negligible with he = h , then we can use the same
formula as in case (ii) with modified length Le given by
Hence mLe = 16 * 0.105 = 1.68 and tanh mLe = tanh 1.68 = 0.9329
Example 3.19. A thin rod of uniform cross section A, length L and thermal conductivity
k is thermally attached from its ends to two walls which are maintained at temperatures
T1 and T2. The rod is dissipating heat from its lateral surface to an ambient at
temperature T∞ with a surface heat transfer coefficient h.
(a) Obtain an expression for the temperature distribution along the length of the rod
(b) Also obtain an expression for the heat dissipation from the rod to the ambient
T1 k h,T∞
x T2
Known: Temperatures at the two ends of the rod, rod cross sectional area A,thermal
conductivity of the rod k, ambient temperature T∞ and surface heat transfer coefficient h.
110
Find: (a) Temperature distribution T(x) along the length of the rod; (b)Rate of heat
dissipation from the rod to ambient
Assumptions: (i) one dimensional steady state conduction along the length of the fin; (ii)
thermal conductivity of the rod is constant
Solution: The general solution for the one-dimensional steady-state temperature distribution
along the length of a rod dissipating heat by convection from its lateral surface is given by
(ii) at x = L, T = T2 or θ = T2 - T∞ = θL (say).
111
Qamb
x
Q|x=0
Q |x=L
L−x
From Eq. (2) we have (dθ / dx) = −m [ L
θ cosh mx+ θL cosh m ( ¿ ) ]
sinh mL
−m [ θL +θ 0 cosh mL ]
Therefore (dθ / dx)|x = 0 =
sinh mL
−m [ θL cosh mL+ θ0 ]
and (dθ / dx)|x = L =
sinh mL
Example 3.20. Heat is generated at a constant rate of q’’’ W/m3 in a thin circular rod of
length L and diameter D by the passage of electric current. The two ends of the rod are
maintained at uniform temperatures with one end at temperature T0 and the other end
at 0 0 C, while heat is being dissipated from the lateral surface of the rod to an ambient
at 0 0C with a surface heat transfer coefficient h.
(a) Derive the one-dimensional steady state energy equation to determine the
temperature distribution along the length of the rod
(b) Solve the above equation and obtain the temperature distribution.
112
00 C h,T∞
q’’’ W/m3
T = T0
x
L
Qamb
Qx Qx+dx
dx Qg
x
Known: Thin rod of diameter D and length L is generating heat at a constant rate of q’’’
W/m3;One end of the rod is maintained at 0 0C and the other end at T0; Lateral surface of the
rod is exposed to ambient at T∞ with a surface heat transfer coefficient h.
Assumptions: (i) one dimensional steady state conduction along the length of the rod; (ii)
thermal conductivity of the rod is constant
Solution: Since the rod is generating heat and dissipating heat to the ambient, the governing
differential equation to determine the one-dimensional steady state temperature distribution
has to be obtained from first principles as illustrated below.
Consider an elemental length ‘dx’of the rod as shown in the figure above. The various
energies crossing the boundaries of the rod as well as the energy generated are also shown in
the figure. For steady state condition the energy balance equation for the rod element can be
written as
Qx + Qg = Qx+dx + Qamb
Or Qx + Qg = Qx + (dQx/dx) dx + Qamb
Or (dQx/dx) dx + Qamb = Qg
113
Or (d2T / dx2) – (hP / kA) (T - T∞) = ─ (q’’’ / k)
Let T - T∞ = θ and (hP / kA) = m2. then the above equation reduces to
where θh(x) satisfies the homogeneous part of the differential equation namely
and θp(x) is the particular integral which satisfies Eq. (1). Solution to Eq.(3) is given by
θ(x) = C1 e mx + C2 e ─ mx -----------------------------(4)
To find θp(x) : Since the RHS of Eq.(1) is a constant let us assume θp(x) = B, where B is a
constant. Substituting this solution in Eq.(1) we have
0 – m2 B = ─ (q’’’ / k)
Or B = (q’’’ / km2)
(ii) at x = L, T = T0
0 = T∞ + C1 + C2 + (q’’’/ km2)
114
Or C1 e mL + C2 e ─ mL = T0 ─ T∞ ─ (q’’’/km2) ----------------------------- (b)
T0 ─ {T∞ + (q’’’/km2)}{1 ─ e ─ mL }
Solving for C1 we get C1 = --------------------------------------------------
{ e mL ─ e ─ mL }
T0 ─ {T∞ + (q’’’/km2)}{1 ─ e ─ mL }
’’’ 2
C2 = ─{ T∞ + (q /km ) }─ --------------------------------------------------
{ e mL ─ e ─ mL }
{ T∞ + (q’’’/km2) }[ ─ e mL + e ─ mL + 1 ─ e ─ mL] ─ T0
C2 = -----------------------------------------------------------------------------------------------
{ e mL ─ e ─ mL }
{ T∞ + (q’’’/km2) }[ 1 ─ e mL] ─ T0
C2 = ----------------------------------------------
{ e mL ─ e ─ mL }
Substituting the expressions for C1 and C2 in Eq. (5) and simplifying we get
T0(e mx - e ─ mx)
’’’ 2
T(x) = T∞ + (q /km ) + --------------------- +
{ e mL ─ e ─ mL }
Example 3.21. Two very long slender rods of the same diameter are given. One rod is
115
of aluminum (k = 200 W/(m-K)). The thermal conductivity of the other
rod is not known. To determine this, one end of each rod is thermally attached to a metal
surface maintained at a uniform temperature T0. Both rods are losing heat to the ambient air
at T∞ by convection with a surface heat transfer coefficient h. The surface temperature of
each rod is measured at various distances from hot base surface. The temperature of the
aluminum rod at 40 cm from the base is same as that of the rod of unknown thermal
conductivity at 20 cm from the base. Determine the unknown thermal conductivity.
Ta
xa
ka
Tb = Ta
xb
kb = ?
Known: Two slender rods A and B are of same diameter;Base temperature T0 is same for
both the rods;Conductivity of rod A = ka = 200 W/(m-K);Ambient temperature T∞ same for
both the rods. Surface heat transfer coefficient h same for both the rods; Ta = Tb with xa = 0.4
m and xb = 0.2m
Assumptions: (i) one dimensional steady state conduction along the length of the rods; (ii)
thermal conductivity of the two rods is constant (iii)rods are assumed to be of infinite length
θ (x) = θ0 e ─ mx ……………………………..(1)
Or mb = 2 ma
116
Or √ [(hPb) / (kbAb)] = 2√[ (hPa) / (kaAa)]
Example 3.22. Show that for a finned surface the total heat transfer rate is given by
Qtotal = [η β + (1 – β)] a h θ0 = ή a h θ0
Where η = fin efficiency ; β= af / a : af = surface area of the fin, a = total heat transfer area
(i.e. finned surface + unfinned surface) ; θ0 = T0 - T∞, with T0 = fin base temperature and T ∞
= ambient temperature, and ή = area – weighted fin efficiency.
Solution:
Qtotal = Qfin + Qbare
Where Qtotal = Total heat transfer rate, Qfin = Heat transfer rate from the finned surface
Example 3.23. The handle of a ladle used for pouring molten lead at 327 0 C is 30 cm long
and is made of 2.5 cm x 1.5 cm mild steel bar stock (k = 43 W/(m-K)). In order to reduce the
grip temperature, it is proposed to make a hollow handle of mild steel plate 1.5 mm thick to
the same rectangular shape. If the surface heat transfer coefficient is 14.5 W/(m 2-K) and the
ambient temperature is 27 0C, estimate the reduction in the temperature of the grip. Neglect
the heat transfer from the inner surface of the hollow shape.
Sketch for example 3.23 (a) : (a) When the handle is made of solid steel bar:
117
2.5 cm
Known: h = 14.5 W/(m2-K) ;
1.5 cm k = 43 W/(m-K)
θ0 = 327 – 27 = 300 0 C
Cross section of the handle
( hP)(1/2) √ [14.5 * 8 * 10 ─ 2]
Therefore m = ------ = -------------------------- = 8.48 (1/m)
(kA)(1/2) √[43 * 3.75 * 10 ─ 4]
When the heat loss from the tip of the handle is neglected the temperature at any point along
the length of the handle is given by
cosh m(L – x)
θ(x) = θ0 ----------------------
cosh mL
Or T(x)|x=L = 47 + 27 = 74 0 C.
= 1.11 cm 2 = 1.11 * 10 − 4 m2
1.5 cm
P = 2 * [ 2.5 + 1.5 ] = 8 cm = 8 * 10 − 2 m
_________ √(14.5 * 8 * 10 − 2)
1.5 mm thick
m = √(hP) / (kA) = ----------------------------
√( 43 * 1.11 * 10 − 4)
118
Or m = 15.59 1/m. Therefore mL = 15.59 * 0.3 = 4.68
θ0 (327 – 27)
θ(x)|x=L = -------------------- = ----------------- = 5.57 0 C.
cosh mL cosh 4.68
Example 3.24. Derive an expression for the overall heat transfer coefficient across a plane
wall of thickness ‘b’ and thermal conductivity ‘k’ having rectangular fins on both sides.
Given that over an overall area A of the wall, the bare area on both sides, not covered by the
fins are Au1 and Au2, the fin efficiencies are η1 and η2, and the heat transfer coefficients h1 and
h2.
Solution:
Let Ti be the temperature of the fluid in contact with the surface 1, T0 be the temperature of
the fluid in contact with surface 2, T1 be the temperature of surface 1 and T2 be the
temperature of surface 2.Let Ti >T0. Then the rate of heat transfer from Ti to T0 is given by
Q = Qbare + Q fin
= hiAu1 (Ti – T1) + hi η1Af1(Ti – T1)
(Ti – T1)
Q = ------------------ ---------------- …………………………(1)
[(1 /h1Au1) + (1/h1η1Af1) ]
(T2 – T0)
Q = ------------------ ---------------- …………………………(2)
[(1 /h2Au2) + (1/h2η2Af2) ]
119
(Ti – T1) + (T1 – T2) + (T2 – T0)
Q = --------------------------------------------------------------------------
[(1 /h1Au1) + (1/h1η1Af1) + (1 /h2Au2) + (1/h2η2Af2) +(b/Ak)]
(Ti – T0)
Q = -------------------------------------------------------------------------- ……(4)
[(1 /h1Au1) + (1/h1η1Af1) + (1 /h2Au2) + (1/h2η2Af2) +(b/Ak)]
Q = UA(Ti – T0)
(Ti – T0)
= ----------------- ………………………………….(5)
(1/UA)
From Eqs. (4) and (5) we have
1
U = ------------------------------------------------------------------------------
A [(1 /h1Au1) + (1/h1η1Af1) + (1 /h2Au2) + (1/h2η2Af2) +(b/Ak)]
Example 3.24. Calculate the effectiveness of the composite pin fin shown in Fig.P3.22.
Assume k1 = 15 W/(m-K), k2 = 50 W/(m-K) and h = 12 W/(m2 – K).
Solution:
k1 3 mm = d1
d2 = 10 mm
x k2
Qc
K1=15 W/m-k , K2=50W/m-k,
h=12 W/(m2 –K)
Qx Qx+dx
(b) Energy transfer across the surfaces of the fin element
dx
120
Energy balance equation for the fin element is given by
Qx = Qx+dx + Qc = Qx + (dQx/dx) dx + Qc
Or dQx / dx + Qc = 0 ………………………………………..(1)
Qx consists of two components namely the heat transfer Qx1 through the material of thermal
conductivity k1 and the rate of heat transfer Qx2 through the material of conductivity k2.
hP2
2 2
(d T / dx ) − ------------------- (T – T∞) = 0
(k1A1 + k2A2)
When the heat loss from the fin tip is negligible , the solution to equation (2) is given by
121
Noting that hP2 / (k1A1 + k2A2) = m2, the above expression for η simplifiers to
tanh mL
η = ------------------------ ……………………(4)
mL
P2 = π x 0.01 0.0314 m.
√ [ 12 x 0.0314]
m = ----------------------------------------------------- = 10.12
√[(15 x 7.1 x 10 − 6 ) + (50 x 7.15 x 10 − 5)]
tanh (1.012)
η = ------------------ = 0.757
1.012
Example 3.25. A plane wall 4 cm thick has one of its surfaces in contact with a
fluid at 130 0C with a surface heat transfer coefficient of 250 W/(m2 – K) and the
other surface is in contact with another fluid at 30 0C with a surface heat transfer
coefficient of 500 W/(m2-K). The thermal conductivity of the wall varies with
temperature according to the law
k = 20 [ 1 + 0.001 T]
where T is the temperature. Determine the rate of heat transfer through the wall and
the surface temperatures of the wall.
Solution:
Rci = Thermal resistance for convection at the surface at Ti = 1/(hiA) = 1 / (250 x 1)
= 0.004 m2 – K /W
122
Or Rco = 0.002 m2-K/W
Now Q = (Ti – T1) / Rci, where T1 = Surface temperature in contact with fluid at Ti.
0.04
Or R = ---------------------------
[21.6 – 2 * 10 − 5Q]
For physically meaningful solution T1 should lie between Ti and To. This is possible only
If
Q = (1.41 * 10 6 − 1.39 * 10 6) / 2 = 10000 W.
Example 3.26. The thermal conductivity of a plane wall varies with temperature
according to the equation
k(T) = k0 [ 1 + β T2 ]
where k0 and β are constants.
123
(a) Develop an expression for the heat transfer through the wall per unit area of the wall
if the two surfaces are maintained at temperatures T1 and T2 and the thickness of the
wall is L.
(b) Develop a relation for the thermal resistance of the wall if the heat transfer area is A.
Solution:
K = k0 [ 1 + βT2]
T1
T2
L
L
Or Qx = − k0[1 +βT2]A(dT/dx)
(T1 – T2)
Qx = -------------------------------------------------------
1
---------------------------------------------------
(k0A/L) [1 + (β/3)(T12 +T1T2 + T22)]
124
Therefore thermal resistance of the wall is given by
1
R = ---------------------------------------------------
(k0A/L) [1 + (β/3)(T12 +T1T2 + T22)]
Example 3.27. Find the steady-state heat flux through the composite slab made up of two
materials as shown in Fig. P 3.26. Also find the interfacial temperature. The thermal
conductivities of the two materials vary linearly with the temperature in the following
manner:
k1 = 0.05 [1 + 0.008 T] W/m-K and k2 = 0.04 [1 + 0.0075 T] W/m-K
600 0 C 30 0 C
k1 k2
5 cm 10 cm
T3
Data:- T1 = 600 0 C ; T2 = 30 0 C ;
T2
L1 = 0.05 m ; L2 = 0.10 m ;
T1
k1 k2
k1 = 0.05 [ 1 + 0.008T ] ;
k2 = 0.04 [ 1 + 0.0075T ]
To find T3 and q
L1 L2
125
Mean thermal conductivity for the first layer is given by
= 0.17 + 2 * 10 − 4 T3 W/(m-K)
= 496.9 0 C.
126
Therefore km1 = 0.17 + 2 * 10 − 4 * 496.9 = 0.269 W / (m-K)
Example 3.28. Consider a slab of thickness L in which heat is generated at a uniform rate of
q’’’ W/m3. The two boundary surfaces are maintained at temperatures T1 and T2. The thermal
conductivity of the slab varies with temperature according to the equation
k(T) = k0 [ 1 + βT ]
where k0 and β are constants. Develop an expression for the heat flux q(x) in the slab.
Measurements show that steady-state conduction through a plane wall without heat
generation produced a convex temperature distribution such that the mid-point temperature
was ΔT0 higher than expected for a linear temperature distribution. Assuming that the
thermal conductivity has a linear dependence on temperature [k = k 0(1 + βT)], where k0 and
β are constants,develop a relationship to evaluate β in terms of ΔT0, T1 and T2.
Solution:
L
k = k0 [ 1 + βT ]
ΔT0 k = k0
T1
T2
For constant thermal conductivity k0, the temperature distribution in the wall is linear and is
given by
127
T(x) = T1 – (T1 – T2) x / L
When the thermal conductivity varies with temperature the temperature distribution in the
wall is determined as follows.
Integrating the above equation between x = 0 and any x at which the temperature is given by
T*(x) we have
x T*
∫Qxdx = ∫− k0A(1 + βT) dT
T1
1
Or Qx x = − k0A [(T* - T1) + (β/2)(T*2 – T12)] ……………(3)
Therefore
___________________________________________
T* = − (1/β) ± √ (1/β2) – (2km / βk0)(x / L)(T1 – T2) + T12 + (2 / β)T1
_____________________________________
T*|x=L/2 = − (1/β) ± √ (1/β2) – (km / βk0)(T1 – T2) + T12 + (2 / β)T1
__________________________________________________
= − (1/β) ± √ (1/β2) – [k0{1 + β(T1 + T2)/2}/ βk0](T1 – T2) + T12 + (2 / β)T1
128
_______________________________________________
= − (1/β) ± √ (1/β2) – [{1 + β(T1 + T2)/2}/β ](T1 – T2) + T12 + (2 / β)T1
Hence we have
______________________________________________
(T1 + T2)/2 +ΔT0 = − (1/β) ± √ (1/β2) – [{1 + β(T1 + T2)/2}/β ](T1 – T2) + T12 + (2 / β)T1
Example 3.29. A slab of thickness ‘L’ has its two surfaces at x=0 and x = L maintained at
uniform temperatures of T0 and TL respectively. The thermal conductivity of the slab has
spatial variation according to the law k = k 0 [1 + α x], where ko and α are constants. Obtain
expressions for (i) temperature distribution in the slab, and (ii) rate of heat transfer through
the slab assuming one dimensional steady state conduction.
Solution:
k = k0 [1 + α x]
T1 T2
The governing differential equation for one dimensional steady state conduction without
heat generation is given by
d dT
k ( )
dx dx
=0
129
d dT
(k [ 1+αx ]
dx 0 dx ) =0 …………………………..(1)
{ [
d
dy
k0 y
dT
dy ( )
α α ]} =0
Or { [
d
dy
k0 y
dT
dy ( )
α ]} =0
[ ( )]
k0 y
dT
dy
α = C1
C1 dy
Or dT = ( )
α k0 y
C1
T= ( )
α k0
ln y + C2
C1
T= ( )
α k0
ln (1 + α x) + C2 ………………..(2)
Eq.(2) is the general solution of Eq.(1). The values of C1 and C2 can be obtained from the two
boundary conditions at x = 0 and at x = L as follows.
C1
Condition (ii) in Eq. (2) gives T2 = ( )
α k0
ln (1 + α L) + T1
130
α k0
Or C1 = (T2 – T1) .
ln (1+ αL )
Substituing the values of C1and C2 in Eq.(2) we get the temperature distribution as:
ln ( 1+αx )
T(x) = (T2 – T1) + T1
ln (1+ αL )
A (T2 – T1) α 1
= − [k0 (1 + α x)] ------------------- x -----------------
ln(1 + α L) (1 + α x)
k0 α A (T1 – T2)
Or Qx = -----------------------------
ln(1 + α L)
Example 3.30. If in the above problem the thermal conductivity varies with distance
according to the law
K = k0 [1 + α x2]
Obtain expressions for (i) the temperature distribution T(x) and (ii) the rate of heat transfer.
Solution:
131
L
k = k0 [1 + α x 2]
T1 T2
The governing differential equation for one dimensional steady state conduction without heat
generation is given by
√α = sec 2 y (dy/dx)
√α
2
or (dy/dx) = √α / sec y = ------------------- .
[ 1 + α x2 ]
Or d/ dy [k0 √α (dT/dy)] = 0.
132
Or dT = [C1 /( k0 √α)] dy.
[T2 – T1]k0 √α
Or C1 = ------------------
tan −1(√α L)
Qx = − k0 [1 + α x2] A (dT/dx)
√α
2
= − k0 [1 + α x ] A [T2 – T1] ---------------------------
[1 + α x2] tan −1 (√α L)
k0 A √α [T1 – T2]
Or Qx = -------------------------------
tan −1 (√α L)
Example 3.31. A hollow cylinder has its internal surface at radius r 1 maintained at a
uniform temperature T1 and external surface at radius r2 maintained at a uniform
temperature T2. The thermal conductivity of the material of the cylinder varies with radius
according to the law k = k0 [1 + α r], where k0 and α are constants. Derive expressions for (i)
radial temperature distribution in the cylinder and (ii) rate of heat transfer through the
cylinder. Assume one-dimensional radial steady state conduction in the cylinder.
Example 3.32. A hollow cylinder has its internal surface at radius r 1 maintained at a
uniform temperature T1 and external surface at radius r2 maintained at a uniform
temperature T2. The thermal conductivity of the material of the cylinder varies with radius
according to the law k = k0 [1 + α r2], where k0 and α are constants. Derive expressions for
133
(i) radial temperature distribution in the cylinder and (ii) rate of heat transfer through the
cylinder. Assume one-dimensional radial steady state conduction in the cylinder.
Example 3.33. A hollow sphere has its internal surface at radius r1 maintained at a uniform
temperature T1 and external surface at radius r2 maintained at a uniform temperature T2. The
thermal conductivity of the material of the cylinder varies with radius according to the law k
= k0 [1 + α r], where k0 and α are constants. Derive expressions for (i) radial temperature
distribution in the cylinder and (ii) rate of heat transfer through the cylinder. Assume one-
dimensional radial steady state conduction in the cylinder.
Example 3.34. A hollow sphere has its internal surface at radius r1 maintained at a uniform
temperature T1 and external surface at radius r2 maintained at a uniform temperature T2. The
thermal conductivity of the material of the cylinder varies with radius according to the law k
= k0 [1 + α r2], where k0 and α are constants. Derive expressions for (i) radial temperature
distribution in the cylinder and (ii) rate of heat transfer through the cylinder. Assume one-
dimensional radial steady state conduction in the cylinder.
Chapter 4
Transient Conduction
4.1.Introduction:- In general, the temperature of a body varies with time as well as
position.In chapter 3 we have discussed conduction in solids under steady state conditions for
which the temperature at any location in the body do not vary with time. But there are many
practical situations where in the surface temperature of the body is suddenly altered or the
surface may be subjected to a prescribed heat flux all of a sudden. Under such circumstances
the temperature at any location within the body varies with time until steady state conditions
are reached. In this chapter, we take into account the variation of temperature with time as
well as with position.However there are many practical applications where in the temperature
variation with respect to the location in the body at any instant of time is negligible. The
134
analysis of such heat transfer problems is called the “lumped system analysis”. Therefore in
lumped system analysis we assume that the temperature of the body is a function of time
only.
4.2. Lumped system analysis:- Consider a solid of volume V, surface area A, density ρ,
Specific heat Cp and thermal conductivity k be initially at a uniform temperature Ti.Suddenly
let the body be immersed in a fluid which is maintained at a uniform temperature T∞, which
is different from Ti.The problem is illustrated in Fig.4.1.Now if
V = volume
A=surface area
ρ = density
Cp = specific heat
k = conductivity
T(t) is the temperature of the solid at any time t, then the energy balance equation for the
solid at time t can be written as
Rate of increase of energy of the solid = Rate of heat transfer from the fluid to the solid
hA
Or ( dTdt ) =
ρV C p
[T∞ − T(t)]
135
hA
For convenience, a new temperature θ(t) = T(t) - T∞ is defined and denoting m =
ρV C p
the above equation can be written as
Eq.(4.1) is a first order linear differential equation and can be solved by separating the
variables. Thus
dθ
= − m dt
θ
Or θ = C e – mt …………………………………...(4.2)
θ(t) = θi e − mt
θ (t)
or = e – mt ………………………………
θi
(4.3)
Since LHS of Eq.(4.3) is dimensionless, it follows that 1/m has the dimension of time and is
called the time constant.Fig. 4.2 shows the plot of Eq.(4.3) for different values of m. Two
observations can be made from this figure and Eq. (4.3).
1.Eq. (4.3) can be used to determine the temperature T(t) of the solid at any time t or to
determine the time required by the solid to reach a specified temperature.
2.The plot shows that as the value of m increases the solid approaches the surroundings
temperature in a shorter time.That is any increase in m will cause the solid to
respond more quickly to approach the surroundings temperature.
136
1.0
θi
t
Fig.4.2: Dimensionless temperature as a function of time for a
solid with negligible internal temperature gradients
The definition of m reveals that increasing the surface area for a given volume and the heat
transfer coefficient will increase m. Increasing the density, specific heat or volume decreases
m.
Ls = V /A ………………………………(4.4)
For solids like slabs, infinite cylinder, and sphere, it has been found that the error by
neglecting internal temperature gradients is less than 5 %, if
137
Hence a very low value of Biot number indicates that resistance for heat transfer by
conduction within the solid is much less than that for heat transfer by convection and
therefore a small temperature gradient within the body could be neglected.
Example 4.1. A copper cylinder 10 cm diameter and 15 cm long is removed from a liquid
nitrogen bath at ─ 196 0 C and exposed to room temperature at 30 0 C. Neglecting internal
temperature gradients find the time taken by the cylinder to attain a temperature of 0 0C,
with the following assumptions: Surface heat transfer coefficient = 30 W / m2 – K.
Density of the copper cylinder = 8800 kg / m3.Specific heat of the cylinder = 0.38 kJ/(kg-K)
Thermal conductivity of the cylinder = 350 W / (m-K).
D = 10 cm or R = 0.05 m; L = 0.15 m
T∞ D h = 30 W/(m2 – K); k = 350 W / (m-K) ;
h T(t) = 0
Assumptions: (i)Internal temperature gradients are negligible; (ii) solid properties are
constant
Solution: Biot Number = hR / k = 30 x 0.05 / 350 = 0.0043 which is << 0.1. Hence internal
= 4.785 x 10 − 4 1/s
T(t) − T∞
Now ------------------ = e − (hA/ρVcp) t
138
Ti – T∞
0 − 30
Hence ------------------- = exp (− 4.785 x 10 − 4 x t)
− 196 – 30
Example 4.2: A thin copper wire having a diameter D and length L (insulated at
the ends) is initially at a uniform temperature of T0. Suddenly it is exposed to a gas
stream, the temperature of which changes with time according to the equation
T∞ = Tf (1 ─ e─ ct) + T0
where Tf, T0 and c are constants. The surface heat transfer coefficient is h. Obtain an
expression for the temperature of the wire as a function of time t.
T(0) = T0
L
Known: Rod of diameter D and length L, initially at T0, is suddenly exposed to en
environment whose temperature varies with time according to the law T∞ = Tf (1 ─ e─ ct) + T0,
where Tf, T0 and c are constant
Assumptions: (i) Internal temperature gradients for the rod is negligible; (ii) Thermal
properties of the rod are constant
Solution: Let T(t) be the temperature of the cylinder at any time t. Energy balance for the
cylinder for a time interval dt is given by
dT
Or = (hA/ρVCp) [T∞ - T(t)]
dt
139
Putting m = (hA/ρVCp), the above equation reduces to
dT
+ m T(t) = m T∞
dt
dT
+ m T(t) = m [T0 + Tf (1 – e - ct)]
dt
dT
or + m [T(t) – T0] = mTf (1 – e – ct )
dt
dθ
+ m θ (t) = mTf (1 – e – ct ) ………………………
dt
(1).
dy
This equation is of the form + Py = Q, which is solved by multiplying throughout by
dx
an integrating factor and then integrating. For equation (1) the integrating factor is
d
or (emtθ) = mTf [ e mt – e (m – c)t]
dt
m
or θ(t) = Tf − Tf e − ct + C1e − mt …………………(2)
( m−c )
When t = 0 , T(0) = T0 i.e., θ(0) = 0. Substituting this condition in equation (2) we get
m
or 0 = Tf − Tf + C1
( m−c )
140
c
Or C1 = ( ( m−c ))
Tf.
Substituting this expression for C1 in equation (2) we get the temperature of the cylinder as
m c
θ(t) = Tf − Tf e − ct + Tf e − mt
( m−c ) ( m−c )
m c
Or T(t) – T0 = Tf [ 1 − e − ct + e − mt]
( m−c ) ( m−c )
T(0) = T0
h,Tf
Known: Solid sphere of radius R is initiallt at temperature T0; Suddenly at time t = o the
sphere is generating heat at q’’’ W/m3 and simultaneously the sphere is losing heat to the
ambient at Tf with a surface heat transfer coefficient h.
141
Assumptions: (i) Internal temperature gradients for the rod is negligible; (ii) Thermal
properties of the rod are constant
Solution:Energy balance equation for the sphere at any time t can be written as
( dθdt ) + mθ = q0 ………………………………..(1)
d
or (θemt) = q0emt
dt
get C1 = (T0 – Tf) – (q0 / m). Therefore the temperature in the sphere as a function of time is
given by
142
Example 4.4: A solid steel ball (ρ =8000 kg/m3 ; cp = 0.42 kJ/kg-K) 5 cm in diameter is at a
uniform temperature of 450 0 C. It is quenched in a controlled environment which is initially
at 90 0C and whose temperature increases linearly with time at the rate of 10 0C per minute.
If the surface heat transfer coefficient is 58 W/(m2-K), determine the variation of the
temperature of the ball with time neglecting internal temperature gradients. Find the value of
the minimum temperature to which the ball cools and the time taken to reach this minimum
temperature.
T(0) = Ti
h,Tf
dTf
at t = 0, Tf = 90 0C; ( )
dt
= 10 0C / min = (1/6) 0 C / s
dTf
Therefore a = 90 0 C and b = ( )
dt
= (1/6) 0 C / s.
Or Tf = 90 + t / 6 , t in seconds ……………….(1)
Energy balance equation for the sphere at any time t can be written as
hA
Or
dT
( )
dt
= ( ρV C p ) [Tf(t) – T(t)]
143
Letting m = ( ρVhAC )
p
the above equation can be written as
d
or (T e mt) = m [90 + t / 6 ] e mt
dt
(T e mt) = m ∫[90 + t / 6 ] e mt dt + C1
Or T(t) = m e − mt ∫[90 + t / 6 ] e mt dt + C1 e − mt
When t = 0 , T(t) = Ti. Substituting this condition in the above equation and solving for C1 we
get
C1 = [Ti – 90 + 1 / 6m]
144
(360 m + 1/6) e − mt = 1/6
Or e mt = (2160 m + 1) -----------(4)
4 π R2 h
Now m = ( hA
ρV C p
2.07 x 10 − 3. …(5)
) =
( ρG p()4
3
π R3 ) =
3h
ρC pR
=
3∗58
8000∗0.025∗0.42∗10
3 =
T(t)=9.4857 + (230.5152)*(0.9979)t
d2T
Now from the above expression for T(t), > 0.
d t2
mt = 1.7
Or t = 1.7 / m = 1.7 / (2.07 x 10 − 3) = 821 s = 13.7 min
Example 4.5: A house hold electric iron has a steel base [ρ =7840 kg/m3 ; cp = 450 J/(kg-
K) ;k = 70 W/(m-K)] which weighs 1 kg. The base has an ironing surface area of 0.025 m2
and is heated from the other surface with a 250 W heating element. Initially the iron is at a
uniform temperature of 20 0 C with a heat transfer coefficient of 50 W/(m2-K). (a) Determine
the temperature of the plate 300 seconds after the heater switch is on .(b) What would be the
equilibrium temperature of the iron if the control of the iron box did not switch of the
current?
145
Q
Qc h ,T∞
A
Assumptions: (i) Internal temperature gradients for the rod is negligible; (ii) Thermal
properties of the iron are constant
1.275 x 10 − 4
L = V / A = ------------------ = 0.005 m
0.025
50 x 0.005
Bi = (hL / k) = -------------------- = 0.00364
70
Since Bi < 0.1, it can be assumed that temperature gradients within the plate are negligible.
Hence the temperature of the plate depends only on time till steady state condition is reached.
Energy balance at any time t for the plate can be written as
Q − Qc = ρVCp ( dTdt )
Or Q – hA(T - T∞) = ρVCp ( dTdt )
Or ( dTdt ) + m(T - T∞) = (Q / ρVCp) ……………………………..(1)
146
( dθdt ) + m θ = (Q / ρVCp)
Multiplying the above equation by the integrating factor e mt,( e∫mdt=emt) we get
d
Or (θe mt) = (Q / ρVCp) emt
dt
When t = 0, T = Ti or θ = Ti - T∞ = 20 – 20 = 0 0 C.
0 = (Q / ρVCpm) + C1 or C1 = − (Q / ρVCpm)
θ = (Q / ρVCpm) [ 1 − e − mt ]
θ = (Q / hA) [ 1 − e − mt ] …………………………………..(3)
Q 250 50∗0.025
hA
=
50∗0.025
= 200 ; m = 1∗450
= 2.8 * 10 − 3
Or T = 113.7 + 20 = 133.7 0 C.
(b) When the control switch is not switched off and the iron is left in the ambient, steady state
condition will be attained as t tends to ∞ so that the heat transferred to the base plate will be
convected to the ambient. i.e., Q = Qconvection
Or T = 220 0 C.
147
This answer can also be obtained by putting t = ∞ in equation (3) and solving for T.
4.3 One-dimensional Transient Conduction ( Use of Heissler’s Charts): There are many
situations where we cannot neglect internal temperature gradients in a solid while analyzing
transient conduction problems. Then we have to determine the temperature distribution
within the solid as a function of position and time and the analysis becomes more complex.
However the problem of one-dimensional transient conduction in solids without heat
generation can be solved readily using the method of separation of variables.The analysis is
illustrated for solids subjected to convective boundary conditions and the solutions were
presented in the form of transient – temperature charts by Heissler. These charts are now
familiarly known as “Heissler’s charts”.
2L
Surfaces exposed to a fluid at T∞ with heat transfer coefficient h for time t ≥ 0
T = Ti at t =
T = T(x,t)
x
Fig.4.3: Geometry, coordinates and boundary conditions for one dimensional tr
conduction in a slab
148
2
∂T ∂T
Governing differential equation: = (1/α)
∂x
2
∂t
………………………..(4.7a)
It is more convenient to analyze the problem by using the variable θ(x,t), where
θ(x,t) = T(x,t) - T∞. Then equations (4.7a) to (4.7d) reduce to the following forms:
∂2 θ ∂θ
= (1/α)
∂ x2 ∂t
………………………..(4.8a)
∂θ
(i) at x = 0, = 0 for all t > 0 (axis of symmetry) …………………………..
∂x
(4.8c)
d2 X
Y ( )
d x2
= (X/α) ( dYdt )
Or
149
2
( )
1 d X
X d x2
=
1
Yα ( dYdt )
…………………….(4.10)
LHS of Eq. (4.10) is a function of x only and the RHS of Eq. (4.10) is a function of t
only.They can be equal only to a constant say − λ2.(The reason to choose the negative sign is
to get a physically meaningful solution as explained later in this section).Hence we have two
equations namely
with C1, C2 and D are constants of integration. Substituting these solutions in Eq.(4.9) we
have
θ(x,t) = D exp (− αλ2t) [C1 cos (λx) + C2 sin (λx)]
Eq.(4.15) is the general solution involving the constants A1, A2 and λ which can be
determined using the two boundary conditions and the initial condition as illustrated below.
∂θ
Now from Eq. (4.15), = λ exp (− αλ2t) [ −A1 sin (λx) + A2 cos (λx)]
∂x
Substituting boundary condition (i) we have 0 = λ exp (− αλ2t) [0 + A2] for all t.
∂θ
and = λ exp (− αλ2t) [ −A1 sin (λx) ]
∂x
150
k λ A1 exp (− αλ2t) sin (λL) = h A1 exp (− αλ2t) cos (λL)
where Bi = hL / k.
Equation (4.17) is called the “characteristic equation” and has infinite number of roots
namely λ1, λ2, λ3, ..............Corresponding to each value of λ we have one solution and hence
there are infinite number of solutions. Sum of all these solutions will also be a solution as the
differential equation is linear. Therefore the solution θ(x,t) can be written as follows.
To find An:- The constants An in Eq. (4.18) can be found using the orthogonal property of
trigonometric functions as shown below.Substituting the initial condition we have
θi = ∑ An cos (λnx)
Multiplying both sides of Eq.(4.18) by cos λmx and integrating w.r.t ‘x’ between the limits 0
and L we have L L
∫ θi cos (λmx) dx = ∫ ∑ An cos (λmx) cos (λnx) dx
1 1
θ(x,t)
----- = ∑ (An* exp (− λn* 2 Fo) cos (λn* x / L) ………………………………(4.19)
θi
4.3.2.Heissler’s Charts for transient conduction:- For values of Fo > 0.2 the above series
solution converges rapidly and the solution will be accurate within 5 % if only the first term
in the series is used to determine the temperature. In that case the solution reduces to
151
θ(x,t)
----- = A1* exp (− λ1* 2 Fo) cos (λ1* x /L) ………………………………(4.20)
θi
From the above equation the dimensionless temperature at the centre of the slab (x =0) can be
written as
θ(0,t)
----- = A1* exp (− λ1* 2 Fo) ………………………………….(4.21)
θi
The values of A1* and λ1* for different values of Bi are presented in the form of a table (See
Table 4.1). These values are evaluated using one term approximation of the series solution.
Table 4.1: Coefficients used in one term approximate solution of transient one dimensional heat
conduction in plane walls, cylinders, and spheres (Bi = hL/k for a plane wall of thickness 2L, and Bi =
hro/k for a cylinder or sphere of outer radius r0)
152
2.00 1.076 1.178 1.5995 1.3384 2.028 1.479
9 5 8 3
3.00 1.192 1.210 1.7887 1.4191 2.288 1.622
5 2 9 7
4.00 1.264 1.228 1.9081 1.4698 2.455 1.720
6 7 6 2
5.00 1.313 1.240 1.9898 1.5029 2.570 1.787
8 3 4 0
6.00 1.349 1.247 2.0490 1.5253 2.653 1.833
6 9 7 8
7.00 1.376 1.253 2.0937 1.5411 2.716 1.867
6 2 5 3
8.00 1.397 1.257 2.1286 1.5526 2.765 1.892
8 0 4 0
9.00 1.414 1.259 2.1566 1.5611 2.804 1.910
9 8 4 6
10.0 1.428 1.262 2.1795 1.5677 2.836 1.924
9 0 3 9
20.0 1.496 1.269 2.2880 1.5919 2.985 1.978
1 9 7 1
30.0 1.520 1.271 2.3261 1.5973 3.037 1.989
2 7 2 8
40.0 1.532 1.272 2.3455 1.5993 3.063 1.994
5 3 2 2
50.0 1.540 1.272 2.3572 1.6002 3.078 1.996
0 7 8 2
100. 1.555 1.273 2.3809 1.6015 3.1102 1.999
0 2 1 0
∞ 1.570 1.273 2.4048 1.6021 3.141 2.000
8 2 6 0
It can also be concluded from Eq.(4.20) at any time ‘t’ the ratio θ(x,t) / θ(0,t) will be
independent of temperature and is given by
θ(x,t)
------ = cos (λ1* x /L) ……………………………………(4.22)
θ(0,t)
Heissler has represented Eq. (4.21) and (4.22) in the form of charts and these charts are
normally referred to as Heissler’s charts. Eq. (4.21) is plotted as Fourier number Fo versus
dimensionless centre temperature θ(0,t) / θi using [Fig.4.4(b)].
reciprocal of Biot number1 / Bi as the parameter [Fig.4.4(a)], where as Eq. (4.22) is plotted
as θ(x,t) / θ(0,t) versus reciprocal of Biot number using the dimensionless distance x / L as
the parameter.In Fig.[4.4(a)], the curve for 1/Bi = 0 corresponds to the case
h → ∞, or the outer surfaces of the slab are maintained at the ambient temperature T∞. For
large values of 1 / Bi, the Biot number is small, or the internal conductance is large in
comparison with the surface heat transfer coefficient. This in turn, implies that the
temperature distribution within the solid is sufficiently uniform and hence lumped system
analysis becomes applicable.
Fig. (4.5) shows the dimensionless heat transferred Q / Q 0 as a function of
dimensionless time for different values of the Biot number for a slab of thickness 2L. Here Q
153
represents the total amount of thermal energy which is lost by the slab up to any time t during
the transient conduction heat transfer. The quantity Q0, defined as
represents the initial thermal energy of the slab relative to the ambient temperature.
Example 4.6: A brick wall ( α = 0.5 x 10 ─ 6 m2/s, k = 0.69 W/(m-K) and ρ = 2300 kg/m 3 ) of
10 cm thick is initially at a uniform temperature of 230 0 C. The wall is suddenly exposd to a
convective environment at 30 0 C with a surface heat transfer coefficient of 60 W/(m 2-K).
Using the transient-temperature charts, determine
(a) the centre temperature at ½ hour and 2 hours after the exposure to the cooler
ambient,
(c) energy removed from the wall per m2 during ½ hour and during 2 hours.
(d) What would be the time taken for the surface of the wall to reach a
temperature of 40 0C.
2L
154
Assumptions: (i) one dimensional conduction; (ii) thermal conductivity of the solid is
constant
60 x 0.05
Bi = hL / k = --------------- = 4.35.
0.69
Since Bi > 0.1, internal temperature gradients cannot be neglected. i.e. T = T(x,t)
( 0.5 x 10 − 6 x 1800)
1/ Bi = 1 / 4.35 = 0.23 ; Fo = (αt / L2) = -------------------------- = 0.36
0.052
(T0 - T∞)
θ0 = --------------- = 0.8 (T0 = Centre temperature of the slab)
(Ti - T∞)
=190 0 C.
(0.5 x 10 − 6 x 7200)
(ii) when t = 2 h = 7200 s we have Fo = ---------------------------- = 1.44.
(0.052)
(T0 – T∞)
From chart, -------------- = 0.125.
(Ti – T∞)
(T|x=L – T∞)
Hence from the chart ---------------- = 0.275
(T0 – T∞)
= 74 0 C.
155
(ii) t = 7200 s. Hence T|x=L = T∞ + 0.275 (T0 – T∞) = 30 + 0.275 x (55 – 30)
= 36.9 0 C.
(c) (i) Bi2 Fo = 4.35 2 x 0.36 = 6.81; From chart Qremoved / Qmax = 0.50 .
0.69 x 1 x 0.1
Qmax = ρCpV(Ti – T∞) = (k / α)V(Ti – T∞) = ------------------- x (230 – 30)
0.5 x 10 − 6
= 27.6 x 10 6 = 27.6 MJ / m
Now for x / L = 1.0 and 1 / Bi = 0.23 from the chart the ratio of surface temperature
difference to the centre temperature difference can be read as
From the chart corresponding to this value of θ (0,t) / θi and 1 / Bi = 0.23, the Fourier
number can be read as
156
T∞,h
R Ti
Find: (a) T/r=0 and T/r = R after t = 10 min; (b) Qremoved during 10 minute period
Assumptions: (i) one dimensional conduction; (ii) thrmal conductivity of the solid is
constant
(T0 – T∞)
-------------------- = 0.08 ; Hence T0 = T∞ + 0.08 (Ti - T∞)
(Ti – T∞)
= 34.4 0 C.
(T|r=R – T∞)
r / R = 1.0 ; 1 / Bi = 0.1 ; From chart -------------- = 0.13
(T0 – T∞)
157
From energy chart for the infinite cylinder, Q / Qmax = 0.875
Known: R = 0.05 m ; Ti = 30 0 C ; T∞ = 2 0 C ; T0 = 10 0 C ;
Solution:
(0.3 x 0.05 2)
Therefore t = ------------------ = 5357 s = 1.5 h
(1.4 x 10 − 7)
158
Solution using Tables: For the given problem we have
(T0 – T∞)
Therefore -------------- = 0.286 = 1.7362 exp[− λ12Fo]
(Ti – T∞)
0.2925∗R2 0.2925∗0.052
Or t= = = 5223 s = 1 h 27 min.
α 1.45∗10−7
Problem 1:- The solid is initially at a uniform temperature Ti and suddenly at time t>0
the boundary-surface temperature of the solid is changed to and maintained at a uniform
temperature T0 which may be greater or less than the initial temperature Ti.
159
Problem 2:- The solid is initially at a uniform temperature Ti and suddenly at time t>0 the
boundary surface of the solid is subjected to a uniform heat flux of q0 W/m2.
Problem 3:- The solid is initially at a uniform temperature Ti. Suddenly at time t>0 the
boundary surface is exposed to an ambience at a uniform temperature T ∞ with the surface
heat transfer coefficient h. T∞ may be higher or lower than Ti.
Solution to Problem 1:- The schematic for problem 1 is shown in Fig. 4.10. The
mathematical formulation of the problem to determine the unsteady temperature distribution
in an infinite solid T(x,t) is as follows:
The governing differential equation [Eq. 4.7(a)] is
2
∂T ∂T
= (1/α) ……………………
∂x
2
∂t
4.7(a)
It is convenient to solve the above problem in terms of the variable θ(x,t), where θ(x,t) is
defined as
T ( x , t )−T ∞
θ(x,t) =
T i −T ∞
……………………..4.25
The governing differential equation in terms of θ(x,t) will be
0
x
Fig. 4.10: Semi-infinite solid with specified surface temperature T0 for t > 0
160
The initial condition will be at time t = 0, θ(x,0) = Ti − T∞ …………………….4.26(b)
This problem has been solved analytically and the solution θ(x,t) is represented graphically
as θ(x,t) as a function of the dimensionless variable x / [2√(αt)] as shown in Fig. 4.11.
In engineering applications, the heat flux at the boundary surface x = 0 is also
of interest. The analytical expression for heat flux at the surface is given by
k(T0 – Ti)
qs(t) = -------------- ……………………..4.27
√(παt)
Solution to problem 2:- The schematic for this problem is shown in Fig. 4.11.
T(x,t) = Ti at t = 0
q0 W/m2
for t > 0 x
Fig. 4.11: An infinite solid subjected to a constant heat flux at x = 0 for t > 0
Governing differential equation in terms of T(x,t) and the initial condition are same that for
problem 1[i.e. equations 4.26(a) and 4.26(b)].
2q0
T(x, t) = Ti + ------ (αt) ½ [ (1 / √π) exp (− ξ2) + ξ erf (ξ) − ξ ] …………………..(4.28 a)
k
___ 2 ξ
161
where ξ = x / (2√ αt ) and erf (ξ) = ------- ∫ exp (− y2) dy ……………………...(4.28b)
√π 0
Here erf (ξ) is called the “error function” of argument ξ and its values for different values of
ξ are tabulated.
Solution to Problem 3 :- The solid is initially at a uniform temperature T i and suddenly for t
>0 the surface at x = 0 is brought in contact with a fluid at a uniform temperature T ∞ with a
surface heat transfer coefficient h [see Fig. 4.12]. For this problem the solution is represented
in the form of a plot where the dimensionless temperature [1 − θ(x,t)] is plottedagainst
dimensionless distance x / √(αt), using h√(αt) / k as the parameter. It can be noted that the
case h → ∞ is equivalent to the boundary surface ay x = 0 maintained at a constant
temperature T∞.
0
x
Fig. 4.12: Semi-infinite solid with surface at x = 0 is subjected to convective heat transfer coefficien
Example 4.9:-A thick stainless steel slab [α = 1.6 x 10 ─ 5 m2/s and k = 61 W/(m-K)] is
initially at a uniform temperature of 150 0 C. Its surface temperature is suddenly lowered to
20 0 C. By treating this as a one-dimensional transient conduction problem in a semi-infinite
medium, determine the temperature at a depth 2 cm from the surface and the heat flux 1
minute after the surface temperature is lowered
162
Sketch for example 4.9:
0
x
Assumptions: (i) one dimensional conduction; (ii) thermal properties of the solid is constant
Solution: (i)
x 0.02
ξ = -------------- = --------------------------------- = 0.323
2 √ (αt) 2 x √ ( 1.6 x 10 − 5 x 60)
T(x,t) – T0
From chart, --------------------- = 0.35
Ti – T0
(ii)
163
Example 4.10:- A semi-infinite slab of copper (α = 1.1 x 10 ─ 4 m2/s and k = 380 W/(m-K) is
initially at a uniform temperature of 10 0 C. Suddenly the surface at x = 0 is raised to 100 0C.
Calculate the heat flux at the surface 5 minutes after rising of the surface temperature . How
long will it take for the temperature at a depth of 5 cm from the surface to reach 90 0 C?
0
x
Assumptions: (i) one dimensional conduction; (ii) thermal conductivity of the solid is
constant.
Solution:
T(x,t) – T0 90 – 100
θ(x,t) = ------------------ = --------------------- = 0.11 . From chart ξ = 0.1
Ti – T0 10 – 100
x x2 0.05 2
ξ = -------------- or t = ------------------- = --------------------------------- = 586 s = 9.46 min
2√ (αt) 4 α ξ2 4 x 1.1 x 10 − 4 x (0.1) 2
164
Example 4.11:-A thick bronze [α = 0.86* 10 ─ 5 m2/s and k = 26 W/m-K] is initially at 2500C
Suddenly the surface is exposed to a coolant at 25 0 C. If the surface heat transfer coefficient
is 150 W/(m2-K), determine the temperature 5 cm from the surface 10 minutes after the
exposure.
0
x
Assumptions: (i) one dimensional conduction; (ii) thermal conductivity of the solid is
constant.
Solution:
x 0.05
ξ = ---------------- = ------------------------------------ = 0.35
2 √(α t) 2 x √ ( 0.86 x 10 − 5 x 600)
____ __________________
h √(α t) 150 x √ [ 0.86 x 10 − 5 x 600]
-------------- = -------------------------------------- = 0.414
K 26
[T(x,t) – T∞]
165
Therefore from chart 1 − ------------------------ = 0.15
(Ti – T∞)
Example 4.12:- A thick wood [α = 0.82 x 10 ─ 7 m2/s and k = 0.15 W/(m-K)] is initially at 20
0
C. The wood may ignite at 400 0 C. Suddenly the surface of the wood is exposed to gases at
500 0 C. If the surface heat transfer coefficient is 45 W/(m2-K), how long will it take for the
surface of the wood to reach 400 0 C ?
0
x
Solution:
____ _______________
h √ (α t) 45 x √ (0.82 x 10 − 7 x t)
---------- = ------------------------------- = 0.086 √t
k 0.15
____ T(x,t) – T∞ (400 – 500)
x / 2√( α t) = 0. Also 1 − -------------------- = 1 − ------------------- = 0.9799
Ti – T∞ (20 – 5000)
___
Hence from chart h √(α t) / k = 2.75.
166
_
Therefore 0.086 √ t = 2.75 or t = 1023 s = 17 min.
y
h1, T∞
L2
h2, T∞
−L1 o L1
x
h2,T∞
−L2
−L2
h1,T∞
167
∂T
at y = 0, = 0 (axis of symmetry)
∂y
…………….4.29d
∂T
at y= L2, − k = h2 [T −T∞]
∂y
………………..4.29e
The initial condition is : at t = 0, T = Ti …………………4.29f
∂2 θ ∂2 θ 1 ∂θ
+ = α ∂t in – L1 ≤ x ≤ L1 and – L2 ≤ y ≤ L2
∂ x2 ∂ y2
……………4.31a
∂θ
The boundary conditions are: at x = 0, = 0 (axis of symmetry) ...
∂x
………….4.31b
∂θ
at x= L1, − k = h1 θ
∂x
……………..4.31c
∂θ
at y = 0, = 0 (axis of symmetry)
∂y
………….4.31d
∂θ
at y= L2, − k = h2 θ
∂y
…………..4.31e
The initial condition is : at t = 0, θ = 1 ……………4.31f
It can be shown that the solution for the problem given by Eq. 4.31a to 4.31f can be
expressed as a product of solutions of two one − dimensional problems θ 1(x,t) and θ2(y,t) as
illustrated below.
Substituting this in Eq. 4.31a to 4.31f and rearranging we have θ 1 (x,t) to be the solution of
the following one-dimensional problem:
∂2 θ 1 1 ∂ θ1
= in – L1 ≤ x ≤ L1
∂ x2 α ∂t
…………4.33a
168
∂θ 1
The boundary conditions are: at x = 0, = 0 (axis of symmetry) ...
∂x
………….4.33b
∂θ 1
at x= L 1, − k = h1 θ1
∂x
…………..4.33c
The initial condition is : at t = 0, θ1 = 1 …………4.33d
∂2 θ 2 1 ∂ θ2
= in – L2 ≤ y ≤ L2
∂ y2 α ∂t
…………4.34a
∂θ 2
The boundary conditions are: at y = 0, = 0 (axis of symmetry) ...
∂x
………….4.34b
∂θ 2
at x= L 2, − k = h1 θ1
∂x
…………..4.34c
The initial condition is : at t = 0, θ2 = 1 …………4.34d
The above one − dimensional problems for θ1(x,t) and θ2(y,t) are exactly the same as that
whose solution is given by the transient temperature chart for infinite slabs. Thus the solution
for two− dimensional transient conduction problem for a rectangular region – L1 ≤ x ≤ L1
and – L2 ≤ y ≤ L2 can be constructed as the product of two one-dimensional transient
conduction problems for infinite slabs.This concept of product solution is also illustrated in
Fig.4.14(a)
2L2 x
2L1
169
The basic idea developed here can be extended to other configurations. For example the
product solution ,θ(r,x,t)for two dimensional transient conduction in a finite cylinder of
radius R and height 2L will be the product of one dimensional conduction solution for an
infinite slab, θ1 (x,t) and one dimensional solution for an infinite cylinder, θ2(r,t).This is
illustrated in Fig.4.14(b).
x
r
2L
h2,T∞
h1,T∞
170
2L2 x
2L1
Known: 2L1 = 5 cm; 2L2 = 4 cm; α = 1.6 * 10−5m2/s; h1= h2= 500 W/(m2-K); T∞ =250C;
Ti = 2250C; t=2 min; k = 60 W/(m-K)
Assumptions:(i) Surface heat transfer coefficient is constant and is same for all surfaces,i.e.
h1= h2; (ii)Thermal properties of the solid are constant
Solution:
(i) θ(0,0,t) = θ1 (0,t) * θ2(0,t)
T ( x , y ,t )−T ∞ T 1 ( x , t ) −T ∞ T 2 ( y , t )−T ∞
i.e. = *
T i−T ∞ T i−T ∞ T i −T ∞
h1 L1 500∗0.025 αt 1.6∗10−5∗(2∗60)
To find θ1 (0,t): Bi = = = 0.21 ; Fo = =
k 60 L21 0.0252
= 3.072
T ( L1 , L2 ,2 min ) −T ∞
(ii) = θ1 (L1,t) * θ2(L2,t)
T i−T ∞
θ1 (L1,t)
θ1 (L1,t) = ------------ * θ1 (0,t)
171
θ1 (0,t)
x
For the corner
L1
= 1 and Bi = 0,21 as calculated before.
Hence from Heisller’s chart
θ1 (L1,t)
--------------- = 0.9 . Hence θ1 (L1,t) = 0.9 * 0.58 = 0.522
θ1 (0,t)
T ( L1 , L2 ,2 min ) −T ∞
Therefore = 0.522 * 0.46 = 0.24
T i−T ∞
r
2L
172
R
Known: Short cylinder with R = 2.5 cm and 2L = 4 cm; α = 1.6 * 10−5m2/s; k = 60 W/m-K;
Assumption: (i) Two-demensional conduction; (ii) Thermal properties of the cylinder are
constant; (iii) the surface heat transfer coefficient is constant and is same for all boundary
surfaces; (iv) The solid is not generating any heat
hR 500∗0.025
To find θ1 (0,2 min) : Bi = = = 0.208 or 1/Bi = 4.8
k 60
αt −5
1.6∗10 ∗2∗60
Fo = 2 = 2 = 3.1
R 0.025
From Heisller’s chart for infinite cylinder we have θ1 (0,2 min) = 0.31.
hL 500∗0.02
Similarly for infinite slab, Bi = = = 0.167 or 1 / Bi = 6.0
k 60
αt −5
1.6∗10 ∗2∗60
Fo = 2 = 2 = 4.8
L 0.02
(ii) θ(1.5 cm,2 cm,2 min) = θ1 (1.5 cm,2 min) * θ2(2 cm,2 min)
173
1.5
To find θ1 (1.5 cm,2 min): for r = 1.5 cm, r/R = = 0.6 and 1/Bi = 4.8.
2.5
θ 1 ( 1.5 cm ,2 min )
Hence from chart = 0.95 .
θ 1 ( 0,2 min )
2
To find θ2 (2 cm,2 min): For infinite slab we have x/ L = = 0.5 and 1 / Bi = 6.0
4
θ 2 ( 2 cm ,2 min )
Therefore from chart we have = 0.94
θ 2 ( 0,2 min )
Therefore . θ(1.5 cm,2 cm,2 min) = θ1 (1.5 cm,2 min) * θ2(2 cm,2 min)
Example 4.15:An aluminium ingot (k = 200 W/m-K;ρ= 2700 kg/m 3; Cp = 890 J/kg-K) of
dimensions 6 cm X 5 cm X 3 cm is initially at a uniform temperature of 175 0C. Suddenly the
ingot is immersed in a fluid which is maintained at a uniform temperature of 25 0C with the
surface heat transfer coefficient of 250 W/(m 2-K). Determine (i) The temperature at the
centre of the ingot 2 minutes after immersion and (ii)temperature at one of the corners after
the same time interval.
174
2L2
2L3
2L1
Assumptions: (i)Properties of the ingot are constant; (ii) the surface heat transfer
coefficient is constant and is same for all boundary surfaces; (iii) The solid is not generating
any heat
Solution:
k 200
To find θ1 (0,2 min): α =
ρC p
= = 8.323 * 10 −5 m2/s
2700∗890
h L1 250∗0.03 αt −5
8.323∗10 ∗2∗60
Bi = = = 0.0375 ; Fo = = = 11.1
k 200 L21 0.03
2
175
To find θ1 (L1,2 min): x/L1 = 1and 1/Bi = 1/0.0375 = 27
CHAPTER 5
NUMERICAL METHODS IN CONDUCTION
176
5.1. Introduction: Analytical solutions that allow for the determination of the exact
temperature distribution in a solid are only available for limited ideal cases.Graphical
solutions have been used to gain an insight into complex conduction heat transfer problems,
where analytical solutions are not available, but they have limited accuracy and are primarily
used for two-dimensional problems. For solids having complicated geometries, boundary
conditions, and temperature − dependent thermal properties, numerical method of solution
has to be used.Advances in numerical computing now allow for complex heat transfer
problems to be solved rapidly on computers using numerical techniques..The current
numerical techniques include: finite-difference analysis; finite element analysis (FEA); and
finite-volume analysis. In general, these techniques are routinely used to solve problems in
heat transfer, fluid dynamics, stress analysis, electrostatics and magnetics, etc.In this chapter
the use of finite-difference analysis to solve conduction heat transfer problems is illustrated.
In the finite difference method the governing partial differential equation of heat conduction
is approximated by a set of algebraic equations for temperature at a number of nodal points
over the region. Hence the first step in the analysis is the finite difference representation of
the governing partial differential equation into a set of algebraic equations, either by using
Taylor’s series expansion and neglecting higher order terms or by writing the energy balance
equation for each node in the region. Then the set of algebraic equations thus obtained are
solved for unknown temperatures.
5.2.1. Finite difference formulation from differential equation using Taylor’s Series
Expansion:
The governing differential equation for one dimensional steady-state conduction in a plane
wall of thickness L with heat generation is given by
L
Δx =
M
5.2
Δx Δx
L
177
0 1 2 i−1 i i+1 M−1 M
x
iΔx
i−½ i+½
d 2 T (x)
The second derivative of temperature at a node i can be represented in
d x2
dT ( x )
finite differences as shown below.The first derivative of temperature at locations i
dx
+ ½ and
i − ½ can be approximated using Taylor’s series expansion as
dT ( x ) T i+1 −T i
│ i + ½) ≈
dx Δx
5.3(a)
dT ( x ) T i−T i−1
│ (i − ½) ≈
dx Δx
5.3(b)
d 2 T ( x)
Now the second derivative of temperature at node i can be approximated as
d x2
dT ( x) dT ( x )
d 2 T (x) │ ( i+½ )− │(i−½)
│i ≈ dx dx
d x2 Δx
T i−1−2 T i +T i+1
= 2 5.4
Δx
178
T i−1−2 T i +T i+1 ( q'g' ' ) i
+ =0
Δx 2 k
Δx 2
Or Ti-1 − 2Ti + Ti+1 = −( q'g' ' ) i 5.5
k
Where ( q 'g' ' )i represents the heat generation per unit volume at node i.Eq. 5.5 is known as
the finite difference form of the heat conduction equation 5.1 and is valid for all interior
nodes, i.e. i= 1, 2, 3,…………………….M – 1, of the region.In this equation the heat
generation q'i ' ' , thermal conductivity k and mesh size Δx are known quantities. Then Eq.
5.5 gives
M – 1 algebraic equations for the M + 1 unknown node temperatures Ti, i= 0,1,2,3……M+1.
Two additional relations are needed to make the number of equations equal to the number of
unknown node temperatures Ti. These two relations are obtained from the finite difference
representation of the two boundary conditions at nodes i=0 and i=M.
Eq. 5.5 can also be obtained by writing the energy balance equation for the control volume
surrounding the node i (shown by hatched portion in Fig.5.1) as follows. For steady state
conduction the energy balance equation can be written as
Rate of conduction into the control volume from the left + Rate of heat generation within the
control volume – Rate of heat conduction out of the control volume to the right = 0.
Assuming linear variation of temperature between the adjacent nodes and using Fourier’s of
conduction, Eq.5.6 can be written as
179
R(i +1) ,i = ( kAΔx )( i+1) , i
= Thermal Resistance between nodes i+1 and i
5.9b
Eq.5.8 is the finite difference formulation of the one-dimensional steady state conduction
problem with heat generation. This equation is more general than Eq.5.5, as it allows for the
variation of thermal conductivity and cross sectional area with position along the x-axis.If k
and A are constants, Eq. 5.7 reduces to Eq.5.5
where T0 and TM are known temperatures. Then the temperatures at the nodes i = 0 and
i = M in Fig.5.1 are taken as
T│i=0 = T0 5.11a
Eq.5.11a and 5.11b provide the two additional relations needed to make the number of
equations equal the number of unknown node temperatures.
(ii)Boundaries with prescribed heat flux: Let the boundary surface at x = 0 is subjected to
a prescribed heat flux of q0 and the boundary surface at x = L is subjected to a prescribed
heat flux of qM as shown in Fig. 5.2.
q0
0 1 2 M–2 M−1 M
qM
Δx
Δx Δx
2
Δx
2
180
Fig. 5.2 Nomenclature for finite difference formulation for prescribed surface heat flux
boundary conditions
The energy balance equation for a differential volume element surrounding the grid point ‘o’
can be written as follows:
q0 A0 + Qg – Q 0,1 = 0 5.12
( T 0 −T 1 )
q0 A0 + A0(Δx/2)q0’’’ – kA0 =0
( Δx )
Similarly for the volume element surrounding the grid point M, the energy balance equation
after simplification reduces to
For the insulated boundary at x = 0 we have q0 = 0 and for the insulated boundary at x = L we
have qM = 0.Substituting these conditions in Eq. 5.13 and 5.14 we get the finite difference
equations for the boundary nodes as
T∞,h
0 1 2 M–2 M−1 M
T∞, h
Δx Δx
Δx Δx
2 2
Fig. 5.3 Nomenclature for finite difference formulation for convection boundary
conditions
The energy balance equation for the element surrounding node 0 can be written as
181
( T 0 −T 1 )
hA0[T∞ − T0] + A0(Δx/2)q0’’’ – kA0 =0
( Δx )
2 ' ''
2 Δx h Δx q0 2 Δx h
Or 2T1 – ( 2+
k )
T0 +
k
+
k
T∞ = 0 for i = 0
5.17
2 Δx h Δx 2 q'M''
and 2TM −1 – ( 2+
k
TM +) k
2 Δx h
+ k
T∞ = 0 for i = M
5.18
L
0 1 2 3 4 5
T∞, h
Δx
Δx Δx
Δx
2 2
Find: Finite difference equations to determine the unknown temperatures; (ii) Express these
equations in matrix form.
Assumptions: (i) One dimensional steady state conduction: (ii) Thermal conductivity of the
slab is constant
182
Solution: For intermediate nodes 1 to 5 the finite difference equations are given by Eq.5.5:
−q 'i' ' Δx 2
Ti-1 − 2Ti + Ti+1 = i = 1,2,…….,5
k
−q '1' ' Δx 2
T0 – 2T1 + T2 =
k
−7.2∗10 7∗0.0022
Or 50 – 2T1 + T2 =
18
−7.2∗10 7∗0.0022
T1 – 2T2 + T3 =
18
Or T1 – 2T2 + T3 = − 16 (2)
T2 – 2T3 + T4 = − 16 (3)
For the boundary nod 5 the finite difference equation is given by Eq.5.18:
2 ' ''
2 Δx h Δx q5 2 Δx h
2T4 – ( 2+
k )T5 +
k
+
k
T∞ = 0
2 7
2∗0.002∗200 2∗0.002∗200
2T4 − ( 2+
18 )
T5 + 0.002 ∗7.2∗10
18
+
18
* 100 = 0
183
[ ] {} { }
−2 1 0 0 0 T1 −66
1 −2 1 0 0 T2 −16
0 1 −2 1 0 T3 = −16 (6)
0 0 1 −2 1 T4 −16
0 0 0 −2 −2.044 T5 −20.44
The above system of equations are solved using ONLINE solutions using MATLAB and the
unknown temperatures are found to be as shown in the Table 5.1 below. In the same table the
exact solution is also given for comparison. The exact solution for this example is given by
x
T(x) = 50 + 5*
L
+ 200 * (19− Lx ) * ( xL )
(7)
Table 5.1: Comparison betweenfinite difference solution and exact solution for one
dimensional steady state conduction in a slab
x/L 0.2 0.4 0.6 0.8 1.0
0
Temperature in C T1 T2 T3 T4 T5
Finite Difference Solution 119.05 172.09 209.13 230.18 235.22
Exact Solution 119.00 172.00 209.00 230.00 235.00
T∞, h
L he
Δx
Δx Δx
Δx
2 2
184
Fig. 5.4(a) Nomenclature for finite difference formulation for a fin of uniform cross
section
Finite difference equations for intermediate nodes 1 to M -1: For the sub region surrounding
an intermediate node i the difference equation is obtained by writing the energy balance
equation for the sub region shown in Fig.5.4(b)
Qconvection
iii
Node i
Q i-1,i Qi,i+1
Fig5.4(b) Thermal Energy crossing the boundaries of the volume element surrounding
the node i
Rate of conduction into the sub region from the node i-1 to node i
– Rate of conduction out of the region from node i to node i+1 – Rate of convective heat
transfer from the boundary surface of the region into the ambient = 0
or kA [ T i−1−T i
Δx ] – kA [ T i−T i+1
Δx ] – hPΔx [Ti − T∞] = 0
Finite difference equations for the outer boundary node M: The sub region surrounding the
outer boundary node M is shown Fig.5.4(c).
Qc1
Node M - 1 Node M
Qc2
h e, T ∞
185
QM-1,M
Fig5.4(c) Thermal Energy crossing the boundaries of the volume element surrounding
the boundary node M
Energy balance equation for the sub region surrounding the node 5 can be written as
Q4-5 – Qconvection = 0
i.e.
kA [ T M −1−T M
Δx ] − hP 2
Δx
[TM − T∞] − heA [ T M −T ∞ ] =0
If the heat loss from the tip is negligible or if the tip is insulated then he = 0. In that case Eq.
(5.20) reduces to
Example 5.2:An iron rod (k = 50 W/m-K) of length 5 cm and diameter 2 cm protrudes from a
wall and is exposed to an ambient at 200C with a surface heat transfer coefficient of 100 W/
(m2-K).The base of the rod is at 320 0C and the heat transfer from the tip of the rod is
negligible. Determine the temperature distribution along the length of the rod using finite
difference method and compare the solution with the exact solution.
k h,T∞
x Insulated
Find: Temperature distribution along the length using finite difference method and compare
this with the exact solution
Assumptions: (i) One dimensional steady state conduction along the length of the rod
186
(ii)Thermal conductivity of the rod is constant
Solution: The rod is divided into 5 equal subdivisions as shown below. Then Δx = L/5 = 5/5
= 1 cm = 0.01 m
T∞, h
L
Insulated
0 1 2 3 4 5
Δx
Δx Δx
Δx
2 2
For 5 sub regions, the difference equation for intermediate nodes are given by Eq.5.20
Finite difference equations for the boundary node 5: Since the heat loss from the tip is
negligible the difference equation for node 5 is giv en by Eq.(5.21). i.e.
2 θ4 – [2 + (NΔx)2] θ5 = 0 (2)
2
4∗100∗0.01
Now (NΔx) = 2
( 50∗0.02 ) = 0.04
Substituting this numerical value in Eq. (1) and (2) we get the finite difference equations for
temperatures of the nodes as
Therefore we have
187
for i = 4, θ3 – 2.04 θ4 + θ5 = 0 (5d)
The above system of equations is solved using ONLINE Linear Algebra Tool Kit and the
results are compared with exact solution in the table 5.2 given below
Table 5.2: Comparison between finite difference solution and exact solution for one
dimensional steady state conduction in a fin of uniform cross section
5.4. Finite Difference Formulation for One-dimensional Steady state Radial Conduction
in cylinders and spheres:
5.4.1. Cylindrical Geometry: The finite difference formulation for the problem of one
dimensional radial steady state conduction in a solid cylinder is illustrated in this section.Let
the solid cylinder is of radius R and is generating thermal energy at rate of q’’’(r) W/m3.The
R
region 0 ≤ r ≤ R is divided into M cylindrical subregions, each of thickness Δr = as
M
shown in Fig.5.5.
TM
TM-1
Ti+1
Ti
Ti-1
T1
T0
0 1 i−1 i i+1 M-1 M
Δr
2
188
Δr
iᴧr
Fig.5.5 Nomenclature for finite difference formulation for one dimensional radial
conduction in a solid cylinder
5.4.1.1Finite difference equation for an intermediate node i: Energy balance equation for
the sub region surrounding the grid point i can be written per unit length of the cylinder as
k [ (
2 π iΔr−
Δr
2 )
∗1
] [ ] T i−1−T i
Δr
+ ( 2 πiΔr∗Δr∗1 ) q'i ' '
− k[ ( ) ] [2 π iΔr +
Δr
2
∗1
T i−T i+1
Δr ] =0 i = 1,2,3……
M-1
5.4.1.2.Finite difference equation for the innermost node 0: The energy balance equation
for the sub region surrounding the node 0 can be written as
)[ ]
2
Δr Δr T 0 −T 1
q'0' ' π ( )
2
*1 – k ( 2π
2
∗1
Δr
=0
q'0' ' Δr 2
Or − T0 + T1 = −
4k
(5.22b)
189
Qconv
QM-1.M
o M–1 M
Δr
2
Energy balance equation for the volume element surrounding the node M can be written as
T M−1 – T m
Or k * 2π [ ( M −1 ) Δr+
Δr
2
∗1] ¿
Δr
'''
+ q M ∗2
πMΔr∗Δr
2
∗1
[ 1−
1
2M ] TM-1 – [( 1−
1
2M
+
hΔr
)k ] TM = −
q'M'' Δr 2
2k
−
hΔr
K
T∞
(5.22c)
Eqs.(5.22a) to (5.22c) form a set of M+1 simultaneous equations which can be solved for
unknown temperatures T0 to TM+1.
(ii)Boundary surface is maintained at a specified temperature,TM: For this case , since the
temperature at node M is specified, the number of unknown temperatures will be M and Eqs.
(5.22a) and (5.22b) will form a set of M equations which can be solved for unknown
temperatures T0,T1, …………………,TM-1.
5.4.2. Spherical Geometry: Following the procedure illustrated in section 5.4.1for radial
steady state conduction in a solid cylinder, the finite difference formulation for radial
conduction in a solid sphere can be written as follows:
(5.23)
190
2 2
Making the approximation that ( 1−
1
2i ) ≈ [ 1−
1
2i ] and ( 1+
1
2i ) ≈ [ ]
1+
1
2i
Eq. (5.24a) is similar to Eq.(5.23a) for a solid cylinder except that 1/2i is replaced by 1/i.
Δr 2 q'0' '
6T1 – 6T0 = −
k
(5.24b)
[( ) ]
' '' 2
1 2 hΔr
2
1 hΔr q M Δr
( 1−
2M ) TM-1 − 1−
2i
+
K
TM = −
K ∞
T −
2k
2 2
Making the approximation that ( 1−
1
2i ) ≈ [ 1−
1
2i ] and ( 1+
1
2i ) ≈ [ ]
1+
1
2i
We get
[1−
M
1
] TM-1 − [( ) ] T
1−
1
M
+
hΔr
k M =−
hΔr
K ∞
T −
q'M'' Δr 2
2k
(5.24c)
Eqs. (5.24a) to (5.24c) form a set of M+1 simultaneous equations which can be solved for
unknown temperaturatures T0 to TM.
Example 5.3: A long cylindrical fuel element of radius 1 cm and thermal conductivity of 25
W/(m-K) generates energy at a constant rate of 5* 10 8 W/m3. The boundary surface is
maintained at a uniform temperature of 1000C.Assuming one-dimensional radial conduction,
determine the radial temperature distribution in the fuel element using finite difference
method and compare the values with exact solution for the problem.
Known: R = 0.01m; k = 25 W/(m-K); q'i ' ' = 5* 10 8 W/m3 for all the nodes;TM = 1000C
191
Find: Radial steady state temperature distribution in the cylinder by finite difference method
and compare with exact solution
0.01
Δr = = 0.002 m
5
The finite difference equation for the innermost node ‘0’ is obtained from Eq.(5.22b) as
q'0'' Δr 2 8 2
T0 – T1 = − = − 5∗10 ∗0.002
4k 4∗25
2 '''
Hence for i = 1 we have [ 1−
1
2i ] Ti-1 – 2Ti + [ ] 1+
1
2i
Ti+1 = −
Δr qi
k
=−¿ 80
Eqs. (1) to (5) are solved using Online Linear Algebra Tool Kit and tabulated as shown in the
Table below. This table also gives the exact solution at the 5 nodes for comparison. The exact
solution is given by
192
Table 5.3:: Comparison between finite difference solution and exact solution for one
dimensional steady state radial conduction in a cylinder
The numerical solution agrees very well with the exact solution (the maximum deviation
being 0.143 %) and hence there is no need to divide the region 0 ≤ r ≤ R into more number of
sub regions to improve the accuracy. It can also be seen from the above table that by dividing
the given region into 10 sub regions the numerical predictions exactly matches with the exact
solution
Example 5.4. A spherical fuel element of radius 1 cm and thermal conductivity of 25 W/(m-
K) generates energy at a constant rate of 7.5* 10 8 W/m3. The boundary surface is
maintained at a uniform temperature of 1000C.Assuming one-dimensional radial conduction,
determine the radial temperature distribution in the fuel element using finite difference
method and compare the values with exact solution for the problem.
Known: R = 0.01m; k = 25 W/(m-K); q'i ' ' = 7.5* 10 8 W/m3 for all the nodes; TM = 1000C
Find: Radial steady state temperature distribution in the sphere by finite difference method
and compare with exact solution
0.01
Δr = = 0.002 m
5
Finite difference equation for the inner most node ‘0’ is obtained from Eq.(24b) as
193
For intermediate nodes form i = 1 to 4 Eq.(5.24a) is applicable:
Eqs. (1) to (5) are solved using online Linear Algebra Tool Kit and the solution is shown in
Table below.The table also gives the exact solution for comparison. The exact solution is
given by
[ ( )]
2
r
T(r) = 100 + 500 1+
R
Table 5.3:: Comparison between finite difference solution and exact solution for one
dimensional steady state radial conduction in a sphere
Example 5.5: A long triangular fin attached to a surface is made of Aluminium (k = 180 W/
(m2-K).The fin is 5 cm long having a base thickness of 1 cm.The base is maintained at a
uniform temperature of 200 0C. The fin is exposed to an ambient at 25 0C and the convective
heat transfer coefficient of 15 W/(m2-K). Using finite difference method with six equally
spaced nodes determine(a) the temperature at the nodes, (b) the rate of heat dissipation from
the fin to the ambient and (c) the fin efficiency
T0 θ
194
0 5
b
Known: L = 0.05 cm; b = 0.01 cm; k = 180 W/(m-K); T0= 200 0C; T∞ = 250C; h = 15
W/(m2-K) ; Δx = L/5 = 0.05/5 = 0.01m ;
Find: (a) Temperatures at nodes 1 to 5; (b) Qfin; (c) fin efficiency η
Assumptions: (i) one dimensional steady state conduction; (ii) thermal conductivity of the fin
is constant; (iii) heat dissipation to the ambient is by convection only
Solution: Volume element surround ing an intermediate node i is shown below with adjacent
nodes.
Δx
A i- 1/2 Length =
cos θ
Fin Base Node i
Ai+1/2
Node (i – 1) Node (i + 1)
Δx
iΔx
Energy balance equation for the volume element surrounding node i can be written as
k A i- ½ ( T Δx−T )
i−1 i
− k A i+ ½ ( T Δx−T )
i−1 i
−h
2 wΔx
( T −T ∞ ) = 0
cos θ i , j
195
2 wΔx
−h ( T −T ∞ ) = 0
cos θ i , j
Or [ 1 Δx
1 – (i− )
2 L ] Ti – 1 − [2 – 2i
Δx
L
−
h Δx 2
kLsin θ
] Ti + [ 1 Δx
1 – (i+ )
2 L ] Ti + 1
2
h Δx
= − T∞ i = 1,2,3,and 4
kLsin θ
(2)
0.01
Now tan θ = b/2L = = 0.1 Hence sin θ = 0.0995
2∗0.05
The finite difference equation for node i = 5 is obtained by writing the energy balance
equation for the volume element having a length of Δx / 2 surrounding node 5 as shown in
Fig. below
Δx/(2cos θ)
4 5
Δx/2 Δx/2
k *2w
Δx
2
tan θ ( T Δx−T )
4 5
– h *2w *
Δx/2
cos θ
(T5 - T∞) = 0
hΔx hΔx
or T4 – [1 + ]T5 = − T∞
k sinθ k sinθ
Hence for node 5 we have T4 – 1.008 T5 = − 0.2 (e)
Solving Eqs. (a) to (e) using ONLINE Linear Algebra Tool Kit we get the unknown
temperatures as follows:
196
Node 1 2 3 4 5
Temp (0 C) 199.76 198.99 197.31 195.70 194.30
(b) Qfin = Sum of heat dissipated from all the volume elements to the surroundings
Δx/2 Δx
= h* 2w (T0 - T∞) + h*2W [(T1 − T∞) + (T2 − T∞) + (T3 − T∞)
cos θ cos θ
Δx/ 2
+ (T4 − T∞)] + h* 2w (T5 − T∞)
cos θ
wΔx
=h [T0 + 2(T1 + T2 + T3 + T4) + T5 −10 T∞]
cos θ
15∗1∗0.01
= * [200 + 2* (199.76 + 198.99 + 197.31 + 195.7) + 194.3 – 10 *25]
cos 5.71
= 260.46 W
2 wL 15∗2∗1∗0.05
(c) Qmax = h Afin,total (T0 − T∞) = h (T0 − T∞) = * (200 – 25) = 263.8
cos θ cos 5.71
W
260.46
Fin efficiency = η = = 0.987 = 98.7%
263.8
Example 5.6: A pin fin of variable circular cross section has a diameter of 2 cm at its base
and 1 cm at its tip and is 12 cm long. It is made of a material of thermal conductivity of 1 W/
(m-K). The fin base temperature is 125 0C and the fin is exposed to an ambient at 25 0C with a
surface heat transfer coefficient of 10 W/(m 2-K).Determine the steady state temperature
distribution along the length of the fin using finite difference method
Δx/2
Δx/2
θ D2
0 1 2 3 4 5 6
D1
197
Δx
L
T0 = 1250C; T∞ = 250C
Find: (a) Steady state temperature distribution along the length using finite difference
method; (b)Qfin
Assumptions: (i) one dimensional steady state conduction; (ii) thermal conductivity of the fin
is constant; (iii) heat dissipation to the ambient is by convection only
Solution: The fin is divided equally into 6 subregions each of length Δx so that
( D1−D2 ) ( 0.02−0.01 )
Δx = L/6 = 0.12/6 = 0.02 m. Also tan θ = = = 0.0417 or θ =
2L 2∗0.12
2.388 0
Difference equations for intermediate nodes from 1 to 5: The volume element surrounding
an intermediate node ‘i’ along with corresponding nomenclature is shown in Fig. E5.6(a).
Δx / cos θ
X
i–1 i i +1
i Δx
Δx
Ai–½ Ai +1/2
Fig. E5.6(a): Nomenclature for the volume element surrounding an intermediate
node ‘i’
Energy balance equation for the volume element shown can be written as
kAi – ½ ( T Δx−T )
i−1 i
− kAi +½ ( T −TΔx )
i i+ 1
– h Asi [Ti −T∞] = 0 i = 1 to 5
(1)
hΔx
Let φ = T − T∞. and Hi = Asi
k
198
Ai – ½ φi-1 – [Ai – ½ + Ai +½ + H i ] φi + Ai +½ φi+1 = 0 i = 1 to 5 (2)
Δx Δx
where Ai – ½ = π [D2/2 + (L – iΔx + ) tan θ]2 ; Ai +½ = π [D2/2 + (L – iΔx − ) tan
2 2
θ]2
Asi =
0.02
cos 2.388
∗2∗π∗
0.01
2 [
+ ( 0.12−0.02 i )∗0.0417
]
= 1.258* 10 – 3 – 1.049 * 10 – 6 i
Numerical values of Ai – ½ , Ai +½ and Hi are found for different values of i from 1 to 5 are
tabulated as shown below.
Substituting the calculated values of A i – ½ , Ai +½ and Hi for different nodes the final form of
Eqs.(3) for nodes 1 to 5 c an be written as
199
i = 3: 1.960 φ2 – 7.302 φ3 + 1.577 φ 4 = 0
(c)
Difference equations for node 6: The volume element surrounding node 6 is shown in Fig.
E5.6(b).
A5+1/2
Δx
2 cos θ
5 6
D2
Δx/2 Δx/2
Energy balance equation for the volume element shown can be written as follows:
Q5-6 – Qconvection = 0
[ π D 22
)]
2
D2 Δx D Δx
Where A5+1/2 = π ( 2
+
2
tan θ ) and As6 =
4 (
+
Δx
2 cos θ )
∗2 π∗ 2 +
2 (2
tanθ
hΔx
A5+1/2 φ5 – [A5+1/2 + A s6 ¿ φ6 = 0
k
2
0.01 0.02
Now A5+1/2 = π ( 2
+
2
∗0.0417 ) = 9.219 * 10 – 5
200
2
π∗0.01 0.02 0.01 0.02
and As6 =
4
+ (
2cos 2.388
∗2 π∗ ) 2
+
2 (
∗0.0417 ) = 4.192 * 10 −4
15∗0.02
Or 9.219 * 10 – 5 φ5 – [9.219 * 10 – 5 + * 4.192 * 10 −4] φ6 = 0
1
Or φ5 – 2.364 φ6 = 0 (f)
Eqs. (a) to (f) can be solved using any standard technique of solving a system of linear
equations.
5.5. Finite Difference Formulation for Two-dimensional Steady State conduction with
Energy Generation
∂2 T ∂2 T ' ''
+ + q (x , y) = 0 in region R
∂ x2 ∂ x2 k
(5.25)
y
(i – 1, j) (i, j+1) (i, j) (i+1, j)
Δx
(i, j-1)
(j +1)Δy
jΔy
Δy
(j – 1)Δy
201
x
(i −1)Δx iΔx (i +1)Δx
Fig.5.7: Rectangular mesh of size Δx, Δy and intermediate node (i,j) surrounded by 4
neighboring nodes
Difference equation for all intermediate nodes: Control volume surrounding the nodal point
i,j is shown in Fig.5.7.The energy balance equation for this control volume under steady state
conditions can be written as follows:
T i , j−1−T i , j
k (Δx *1) + (Δx *Δy*1) q'i ', 'j = 0
Δy
q'i 'j'
Ti – 1,j + Ti ,j+1 + Ti + 1,j + Ti ,j – 1 − 4 Ti ,j + Δx 2
( )
k
=0
(5.27)
Eq.(5.27) is the finite difference form of the heat conduction equation for any interior nodal
point (i,j).
(i , j+1)
(i +1, j)
q0 ( i , j)
202
Energy ba;ance equation for the control volume surrounding the nodal point (i,j) can be
written as:
q0*(Δy*1) + Q(i , j+1) , (i,j) + Q(i , j+1) , (i,j) + Q(i , j+1) , (i,j) + Qg =0 (5.28)
Using Fourier’s law for heat conduction the above equation reduces to
( T i , j−1−T i , j )
+ k (Δx/2 *1)
Δy
+ ( Δy∗Δx
2
∗1 ) q ' ''
i, j =0
If the boundary containing the node (i,j) then q0 = 0. In that case Eq. (5.29) reduces to
(i , j+1)
(i +1, j)
( i , j)
Energy balance equation for the control volume surrounding the nodal point (i,j) can be
written as:
Qconvection + Q(i , j+1) , (i,j) + Q(i , j+1) , (i,j) + Q(i , j+1) , (i,j) + Qg =0
203
( T i , j+1 −T i , j ) ( T i +1, j−T i , j )
h(Δy*1)[T∞ − Ti j] + k (Δx/2 *1) + + k (Δy *1)
Δy Δx
( T i , j−1−T i , j )
+ k (Δx/2 *1)
Δy
+ ( Δy∗Δx
2
∗1 ) q ' ''
i, j =0
Substituting Δx = Δy and simplifying, the above equation reduces to
If the boundary containing the node (i,j) is insulated then h = 0 and there fore Eq.(5.30)
reduces to Eq. (29a).
(i) Difference equation for exterior corner nodal point subjected to convection with ambient
temperature at T∞ and surface heat transfer coefficient h.
Q2-convection
(i , j) (i +1, j)
Q1-convection
Δy Fig.5.10 Nomenclature for an exterior nodal point
(i, j – 1) Δx
T i+1, j−T i , j )
h ( Δy2 ∗1) [T −T ∞ i, j ] +h ( Δx2 ∗1)[ T −T ∞ i, j ] +k ( Δy2 ∗1) ( Δx
[ ]
Δx
Δx ( T i , j −1 −T i , j ) ∗Δy
+k ( 2
∗1 ) Δy
+ 2
2
∗1 q'i ', 'j
=0
204
2 ' ''
2Ti+1, j + 2Ti, j−1 – [ 4+
4 hΔx
k ]
T i, j = −
4 hΔx
k
T∞ −
Δx qi , j
k
(5.31)
(ii) Difference equation for nodal point at the intersection of two convection boundaries
with ambient temperature at T∞ and surface heat transfer coefficient h.
The nomenclature for the nodal point (i,j) at the intersection of two convection boundaries is
shown in Fig. 5.11.
(i, j+1)
Energy balance equation for the control volume surrounding the node (i,j) can be written as
follows:
Q(i-1,j) – (i,j) + Q(i,j+1) – (i,j) + Q(i+1,j) – (i,j) + Q(i ,j-1) – (i,j) + Qconvection + Qg = 0
[ ]
Δx
T i , j−1−T i , j
+ k (Δx/2 *1) ( Δx ) + h [ Δx Δy
2
+
2 ] [T∞ − T,j] + 3 2
∗Δy
2
∗1 q'i ', 'j =
205
Consider a node (i,j) located next to a curved surface as shown in Fig.5.12.Let the two stirngs
from the node (i,j) intersect the curved boundary at say points A and B as shown in Fig.5.12.
Node (i-1,J)
A
Node (i,j)
ξΔy
B
Δy Fig. 5.12 Nomenclature for a node adjacent to a
Δx ηΔx curved boundary
Node (i,j-1)
Assuming that the temperatures at A and Bi.e. TA and TB are known the energy balance
equation for the control volume surrounding the node (i,j) can be written as follows
( T i−1, j−T i , j ) ( T A −T i , j )
k [( Δy ξΔy
2
+
2
∗1 ) ]
Δx
+k [( Δx ηΔx
2
+
2
∗1 ) ]
ξΔy
+k
( T B−T i , j )
[( Δy ξΔy
2
+
2
∗1) ]
ηΔx
T i , j−1−T i , j )
+k ([ Δx2 + ηΔx2 )∗1] ( Δy
+ ( Δx2 + ηΔx2 ) ( Δy2 + ξΔy2 ) q ' ''
i, j =0
Example 5.7: A solid body of square cross section of side 2 cm. is generating heat at a
uniform rate of 2*10 6W/m3.The thermal conductivity of the body is 15 W/(m-K).The left
surface of the body is insulated and the bottom surface is maintained at a uniform
temperature of 1000C.The right surface is subjected to a uniform heat flux of 5000 W/m 2 and
the top surface is in contact with a fluid at a temperature of 25 0C with a surface heat
transfer coefficient of 75 W/(m2-K).Assuming Δx = Δy = 1 cm obtain the finite difference
equations for the unknown temperatures assuming steady state conduction at the six nodes
and solve for these temperatures.
206
Sketch for example 5.7
Node 2
Δx h = 75 W/(m2-K), T∞ = 25 0C
1 2 3
ΔΔΔ
7 8 9
T = 100 0C
Known: Δx = Δy = 0.01 m; k = 15 W/(m-K); q'i ' ' =2*10 6W/m3 for all ‘i’;qs = 5000 W/m2;
Find: (i) Finite difference equations to find unknown temperatures T1 to T6 (ii) Numerical
values of T1 to T6 by solving the finite difference equations
Assumptions: (i) steady state two dimensional conduction heat transfer; (ii) thermal
conductivity and heat generation in the solid are constant
h, T∞
1 2
Insulated Δy/2
Energy balance equation for the volume element surrounding node 1 per unit width measured
normal to the paper can be written as:
207
T 4 −T 1
k
Δx
2
∗1 ( Δy ) +k
Δy
2 ( T Δx−T )
2 1
+h
Δx
2
*1 ( T ∞−T 1 ) +
Δx
( 2
∗Δy
2 )
' ''
∗1 q1
=0
[ ]
2
hΔx hΔx
T4 + T2 – 2+
k
T1 = −
k
− ( )
Δx ' ' '
2k
q1
[ ]
2
75∗0.01 75∗0.01
T4 + T2 – 2+
15
T1 = −
15
– ( )
0.01
2∗15
∗2∗10 6
Finite difference equation for node 2: Volume element surrounding node 2 with the
nomenclature is shown in Fig. E5.7(2).
Δx h, T∞
1 2 3
Δy/2
Δx/2 Δy/2
5
Energy balance equation for the volume element surrounding node 2 per unit width measured
normal to the paper can be written as:
T 1−T 2 T 3 −T 2
k ( Δy2 ∗1)( Δx ) +k ( Δy2 ∗1)( Δx ) + k ( Δx∗1 ) ( T Δy−T )
5 2
+ h (Δx *1) (T∞ -
T2)
208
+ ( Δx∗Δy
2
∗1 ) q ' ''
2 =0
Substituting Δy = Δx and simplifying we get
Finite difference equation for node 3: Volume element surrounding node 3 with the
nomenclature is shown in Fig. E5.7(3).
h, T∞
2 3
Δy/2 qs
Δy Δx/2
Δx 6
Energy balance equation for the volume element surrounding node 3 per unit width measured
normal to the paper can be written as:
T 2−T 3 T 6 −T 3
Or k ( Δy
2
∗1)( Δx ) +k ( Δx
2
∗1)( Δy ) +h ( Δx2 ∗1) (T∞ − T3)
Δx
+ ( Δy
2
∗1 ) qs +
( 2
∗Δy
2 )
∗1 q'3' ' =0
209
Substituting the given data in the above equation we have
Finite difference equation for node 4: Volume element surrounding node 4 with the
nomenclature is shown in Fig. E5.7(4).
h, T∞
1
Insulated
4 5
7
Δx
Energy balance equation for the volume element surrounding node 4 per unit width measured
normal to the paper can be written as:
T 1−T 4
Or k ( Δx
2
∗1)( Δy ) + k (Δy*1) ( T Δx−T )
5 4
+k ( Δx
2
∗1
T 7 −T 4
)( Δy )
+ ( Δy∗Δx
2
∗1 ) q ' ''
4 =0
210
2
0.01
T1 – 4T4 + 2T5 = − 100 − * 2 * 10 6
15
Finite difference equation for node 5: Volume element surrounding node 5 with the
nomenclature is shown in Fig. E5.7(5).
2
5
4 5 6
Δx Δy
8
Energy balance equation for the volume element surrounding node 3 per unit width measured
normal to the paper can be written as:
Or k Δy ( T Δx−T )
4 5
+ k Δx ( T Δy−T )
2 5
+ k Δy ( T Δx−T )
6 5
+ k Δx ( T Δy−T )
8 5
+ (Δx*
Or T2 + T4 + T6 – 4 T5 = − 113.33 (e)
Finite difference equation for node 6: Volume element surrounding node 6 with the
nomenclature is shown in Fig. E5.7(6).
211
3
Δx
Δy
5 6
qs
Energy balance equation for the volume element surrounding node 6 per unit width measured
normal to the paper can be written as:
Or k Δy ( T Δx−T )
5 6
+k ( Δx
2
∗1
T 3−T 6
)( Δy ) +k ( Δx
2
∗1
T 9 −T 6
Δy )( ) + (Δy*1)qs
+ ( Δy∗Δx
2
∗1 ) q ' ''
6 =0
Substituting Δy = Δx and simplifying we get
2
2 Δx Δx '''
T3 + 2T5 – 4T6 = − T9 – q – q6
k s k
2∗0.01∗5000 0.012
T3 + 2T5 – 4T6 = − – * 2* 10 6
15 15
Eqs. (a) to (f) are solved using Online Linear Algebra Tool Kit and the nodal temperatures
are found to be as follows:
T6 = 119.030C
212
Example 5.8:Hot combustion gases of a furnace are flowing through a concrete chimney
[(k = 1.4 W/(m-K)] of rectangular cross section as shown in Fig. E5.8.The flow section of
the chimney is 20 cm X 40 cm, and the thickness of the wall of the chimney is 10 cm. The
average temperature of the hot gases in the chimney is Ti = 280 0C, and the average
convective heat transfer coefficient inside the chimney is h i = 75 W/(m2-K). The chimney is
losing heat from its outer surface to the ambient air at 15 0C by convection with a surface
heat transfer coefficient of ho = 18 W/(m2-K). Using the finite difference method with Δx = Δy
= 10 cm and taking full advantage of the symmetry, determine (a) the finite difference
formulation of the problem to find steady two-dimensional temperatures at the nodal points,
(b) the temperatures at these nodal points, and (c) the rate of heat loss for a 1-m-long section
of the chimney.
h0, T0
k
10 cm
20 cm
10 cm
H
10 cm 40 cm 10 cm
h i , Ti
213
Sketch for example 5.8 with nodes:Because of symmetry only one fourth of the cross section
of the chimney is considered for analysis as shown below.
Axis of symmetry
(Insulated boundary)
1 2
hi,Ti
Δx
3 4 5 6
4
Δy/2
Δy/2
7 8 9 10
Δx/2 h0,T0
Known: k = 1.4 W/(m-K); Δx = Δy = 0.1 m ; hi =75 W/(m2-K); Ti = 280 0C;
h0 = 18 W/(m2-K); T0 = 15 0C
Find: (i) finite difference equations to determine steady two dimensional temperatures at
different nodes in the chimney T1 to T10 (ii) temperatures at the nodes and (iii) rate of heat
transfer for 1 m height of the chimney
Assumptions: (i) Two dimensional steady state conduction; (ii) Thermal conductivity of the
chimney is constant; (iii) No radiation losses from the outer surface of the chimney to the
surroundings
Solution: Finite difference equation for node 1: The volume element surrounding node 1 is
shown in Fig. E5.8(1). Energy balance equation for the volume element can be written as
Δy/2
hi,Ti
Δy
Δx/2
214
Q5-1 + Q2-1 + Qconvection = 0
hi Δx hi Δx
−[2 + ¿ T1 + T2 + T5 = − [ ¿ Ti
k k
75∗0.1 75∗0.1
−[2+ ]* T1 + T2 + T5 = − [ ]*280
1.4 1.4
Finite difference equation for node 1: The volume element surrounding node 2 is shown in
Fig. E5.8(2). Energy balance equation for the volume element can be written as
Δy/2
ho,To
Δy
Δx/2
215
Substituting for the given quantities we have
75∗0.1 75∗0.1
T1 − [ 2 + 1.4
]* T2 + T6 = − [
1.4
]*15
Finite difference equation for node 3: The volume element surrounding node 3 is shown in
Fig. E5.8(3). Energy balance equation for the volume element can be written as follows.
hi,Ti
3 4
Δy/2
Δy
Δx/2
7
Axis of Symmetry (Insulated)
T 7−T 3 T 4 −T 3
Or k ( Δx
2
∗1 )(Δy ) +k ( Δy
2
∗1 )( Δx ) + hi ( Δx2 ∗1) ( T Δ−Ty )
i 3
=0
+ hi Δx hi Δx
T7 – [2 ¿ T3 + T4 = − Ti
k k
75∗0.1 75∗0.1
T7 − [ 2 + 1.4
]* T3 + T4 = − [
1.4
]*15
Finite difference equation for node 4: The volume element surrounding node 4 is shown in
Fig. E5.8(4). Energy balance equation for the volume element can be written as follows.
Δx h i , Ti
216
4
3 5
Δy/2
Δx/2 Δy/2
8
T 3−T 4
Or k ( Δ2y ∗1)( Δx ) + k (Δx *1) ( T Δ−Ty )
8 4
+k ( Δ2y ∗1)(
T 5−T 4
Δx )
+ hi(Δx*1)(Ti – T4) = 0
2∗75∗0.1 2∗75∗0.1
T3 − [ 4 + ]* T4 + 2T8 = − * 280
1.4 1.4
Finite difference equation for node 5: The volume element surrounding node 5 is shown in
Fig. E5.8(5). Energy balance equation for the volume element can be written as follows.
hi,Ti
Δy/2
4 5 6
Δy/2
Δx/2 Δx/2
217
Q4-5 + Q1-5 + Q6-5 + Q9-5 + Qconvection = 0
T 4−T 5 T 1−T 5 T 6 −T 5
Or k ( Δy
2
∗1)( Δx ) +k ( Δx
2
∗1 )( Δy ) + k ( Δy∗1 ) ( Δx )
T 9 −T 5
+ k ( Δx∗1 )
Δy ( ) + hi ([ Δx2 + Δy2 )∗1]( T −T )
i 5 =0
Finite difference equation for node 6: Volume element surrounding node 6 with the
nomenclature is shown in Fig. E5.8(6). Energy balance equation for the volume element can
be written as follows:
Δx
Δy ho, To
5 6
10
218
Q5-6 + Q3-6 + Q10-6 + Qconvection = 0
Δx Δx
∗1 ∗1
2 2
Or k (Δy*1) ( T Δx−T )
5 6
+k T 3−T 6 +k T 10−T 6
(¿ )
( Δy ) (¿ )
(
Δy )
+ ho(Δy *1) (To – T6) = 0
Finite difference equation for node 7: Volume element surrounding node 7 with the
nomenclature is shown in Fig. E5.8(7). Energy balance equation for the volume element can
be written as follows:
T 3−T 7 T 8 −T 7
Or k ( Δx2 ∗1)( Δy ) +k ( Δy2 ∗1)( Δx ) + ho ( Δx2 ∗1) (To – T7) = 0
Axis of symmetry
3
Δx
Δy
7 8
ho, To
219
Fig. E5.8(7) Volume element surrounding node 7
T3 – [2 + hoΔx / k] T7 + T8 = − ho( Δx / k) To
Finite difference equation for node 8: Volume element surrounding node 8 with the
nomenclature is shown in Fig. E5.8(8). Energy balance equation for the volume element can
be written as follows:
4
Δx
Δy
7 8 9
ho,To
220
Finite difference equation for node 9: Volume element surrounding node 9 with the
nomenclature is shown in Fig. E5.8(9). Energy balance equation for the volume element can
be written as follows:
5
Δx
Δy
8 9 10
ho, To
T 8 −T 9
Or k ( Δy2 ∗1)( Δx ) + k (Δx *1) ( T Δ−Ty )
5 9
+k ( Δy2 ∗1)(
T 10−T 9
Δx )
+ ho (Δx *1)(To – T9) = 0
Finite difference equation for node 10: Volume element surrounding node 10 with the
nomenclature is shown in Fig. E5.8(10). Energy balance equation for the volume element can
be written as follows:
221
Δx
Δy
9 10
ho, To
T 9 −T 10 T 6 −T 10
Or k ( Δy2 ∗1)( Δx ) +k ( Δx2 ∗1)( Δy ) + ho ( Δx2 + Δy2 )∗1 [To –T10] =
0
Equations (a) to (j) form a system of 10 linear equations and are listed below:
222
[ai,j]{Ti} = {ci}
5.6.1.Finite difference equations for intermediate nodes: The cylinder is divided into small
sub regions as shown in Fig. 5.13 (a).The volume element surrounding the node i,j is shown
Δz iΔr
Δr
jΔz
223
z
i,j+1
i+1,j
i,j
i-1,j
i, j-1
separately in Fig.5.13(b). The volume element is in the form of a circular ring of thickness Δz
in the z-direction and of radial thickness Δr in the radial direction as shown in the
figure.Energy balance equation for this volume element can be written as follows:
Or k*2π iΔr*Δr (T −T i , j
i , j−1
Δz ) + k*2π (iΔr + Δr/2)Δz (T −T i , j
i +1, j
Δr )
+ k*2π iΔr*Δr
'''
(T −T i , j
i , j+1
Δz ) + k*2π (iΔr − Δr/2)Δz (T −T i , j
i−1, j
Δr ) + 2πiΔr*Δr*Δz
qi , j =0
1 1 Δr 2
Ti,j-1 + Ti,j+1 + ( 1−
2i ) Ti-1,j + ( 1+
2i ) Ti+1,j − 4Ti,j = −
k
'''
qi , j
(5.34)
(j+1)Δz
224
jΔz
(j-1)Δz (i,j)
(i+1,j)
(i,j-1) (i,j+1) Δr
(i+1)Δr Δz
(i-1,j)
iΔr
(1,j)
Δz
Δr
Axis of cylinder
T 0, j−1 – T 0, j T 1, j – T 0, j
k π (Δr/2)2 ( Δz ) + k* 2π (Δr/2) Δz ( Δr ) + k π (Δr/2)2
(T – T 0, j
0, j+1
Δz )
225
+ π (Δr/2)2Δz q'0,' 'j = 0
2
T 0, j−1 + T 0, j +1 + 4 T 1, j − 6 T 0, j = − ( )
Δr '' '
k
q 0, j j = 1,2,3……N-1
(5.35)
5.6.3.Finite difference equations for intermediate nodes on the outer curved surface of the
cylinder subjected to convection boundary condition: Fig.5.13(d) shows a volume element
surrounding the node (M,j) on the outer surface of the cylinder subjected to the convection
boundary condition.Energy balance equation for this volume element can be written as
follows:
hM, T∞
jΔz
Δr/2
Δz Δr
M-1,j
[ ) ](
T M , j−1−T M , j
[ ) ](
T M , j +1−T M , j
2 2
kπ
2
(
( MΔr ) – MΔr−
Δr
2 Δz ) + kπ
2
(
( MΔr ) – MΔr−
Δr
2 Δz )
226
[ )]
2
T M−1, j −T M , j
+ k*2π ( MΔr−
Δr
2 ) Δz ( Δr ) +
2
(
( MΔr ) – MΔr−
Δr
2
Δz q'M' ' , j
+ 2π M Δr Δz hM (T∞ − TM,j) = 0
3 M −1 2 h M Δr Δr 2 2 h M Δr T ∞
TM,j-1 + TM,j+1 – [4 ( 4 M −1
+
k) ¿ TM,j = −
k
q'M' ' , j −
k
j = 1,2,3,…N-1 (5.36)
5.6.4.Finite difference equations for intermediate nodes on the outer curved surface of the
cylinder subjected to prescribed heat flux condition: Fig.5.13(e) shows a volume element
surrounding the node (M,j) on the outer surface of the cylinder subjected to the convection
boundary condition.Energy balance equation for this volume element can be written as
follows:
(qs)M,j
jΔz
Δr/2
Δz Δr
M-1,j
227
[ ) ](
T M , j−1−T M , j
[ ) ](
T M , j +1−T M , j
2 2
kπ
2
(
( MΔr ) – MΔr−
Δr
2 Δz ) + kπ
2
(
( MΔr ) – MΔr−
Δr
2 Δz )
) [ )]
2
T M−1, j −T M , j
+ k*2π ( MΔr−
Δr
2 ) Δz ( Δr
+
2
(
( MΔr ) – MΔr−
Δr
2
Δz q'M' ' , j
+ 2π M Δr Δz (qs)M,j = 0
3 M −1 2
2 ( q s ) M , j Δr
TM,j-1 + TM,j+1 – 4 ( 4 M −1 ) TM,j = −
Δr
k
'''
qM , j −
k
j = 1,2,3,…N-1 (5.37)
5.6.5.Finite difference equations for intermediate nodes on the outer flat surface of the
cylinder subjected to convection boundary condition: Fig.5.13(f) shows a volume element
surrounding the node (i,0) on the flat surface at z = 0 of the cylinder subjected to the
convection boundary condition. Energy balance equation for this volume element can be
written as follows
i+1,0
h0,T∞ Δr
i,0 i,1
i-1,0 iΔr
Δz
228
i, 0
[( Δr 2
)]
2
Δr Δz Δr
k*2π (iΔr +
2
) ( )
2
T i +1,0−T ¿ + k π
( ¿¿ Δr )
iΔr +
2 ) (− iΔr−
2
(T i,1 −T i , 0
Δz )
i,0
[( Δr 2
)]
2
Δr Δz Δr
+ k*2π (iΔr −
2
) ( ) 2
T i−1,0 −T ¿
( ¿ ¿ Δr )
+ h0 π iΔr +
2) ( − iΔr−
2
(T∞ − Ti,0)
[( Δr 2
) ] ( Δz2 )
2
Δr
+π iΔr +
2 ) (− iΔr−
2
'''
qi , 0 = 0
1 1 2 h0 Δr 2 h0 Δr
( 1+
2i ) Ti+1,0 + ( 1−
2i ) Ti−1,0 + 2 Ti,1 – [2 +
k
¿ Ti,0 = −
k
T∞ −
Δr 2 '''
qi , 0
2k
i = 1,2,3,……………M-1 (5.38)
Similarly the energy balance equation for the volume element surrounding the node i,N [See
Fig.5.13(g)] can be written as follows:
1 1 2 h0 Δr 2 h N Δr
( 1+
2i ) Ti+1,N + ( 1−
2i ) Ti−1,N + 2 Ti,N-1 – [2 +
k
¿ Ti,N = −
k
T∞ −
Δr 2 '''
qi , N
2k
i = 1,2,3,……………M-1 (5.39)
i+1,N
Δr
hN,T∞
i,N-1 i,N
229
i-1,N iΔr
Δz
5.6.6.Finite difference equations for volume elements surrounding the nodes (0,0) and
(0,N): The nomenclature for the volume elements surrounding the nodes (0,0) and (0,N) are
shown in Fig. 5.13(h). The energy balance equation for the volume element surrounding the
node (0,0) can be written as:
Δz h0,T∞ hN,T∞ Δz
(1,0) (1,N)
Δr Δr
Axis of symmetry
Fig.5.13(k): Nomenclature for volume elements surrounding nodes (0,0) and (0,N)
+ π(Δr/2)2(Δz/2) q'0,0
''
=0
(5.40)
230
Similarly it can be shown that the energy balance equation for the volume element
surrounding the node (0,N) is given by
(5.41)
5.6.7.Finite difference equations for volume elements surrounding the nodes (M,0) and
(M,N):
The nomenclature for the volume elements surrounding the nodes (M,0) and (M,N) are
shown in Fig. 5.13(k). The energy balance equation for the volume element surrounding the
node (M,0) can be written as:
hM,T∞
Δr Δr
Δz Δz
Fig.5.13(k): Nomenclature for volume elements surrounding nodes (M,0) and (M,N)
231
Assuming Δz = Δr and simplifying we get
(2− 21M ) M(Δr/2) [TM,1 – TM,0] + (1− 21M ) MΔr [TM-1,0 – TM,0]
2
1 1
+ hMMΔr (Δr/k) [2 −
4M
] [T∞ − TM,0] + ( 2−
2M ) MΔr ( )
Δr
4k
q'M' ' ,0 =0
Or ( 1−
1
4M ) TM,1 + ( 1−
1
2M ) TM-1,0 – {( 1−
1
2M
h Δr
+ M
k
2−) 1
4M ( )} TM,0
= {( 1−
1
2M
h Δr
)
+ M
k
2−
1
4M( )} T∞ − ( 2−
1 Δr 2
2M 4k ) '''
q M ,0 (5.42)
Similarly for the volume element surrounding the node (M,N) the energy balance equation
reduces to
= {( 1−
1
2M
+)
h M Δr
k
2−( 1
4M )} T∞ − ( 2−
1 Δr 2
2M 4k ) '''
q M ,0 (5.43)
5.7.1.Finite difference equations for intermediate nodes: The cylinder is divided into small
sub regions and the volume element surrounding the node i,j will be as shown in Fig.5.14.
(i,j+1)
(i+1,j)
(i,j)
(i-1,j)
232
(i, j-1)
Δθ
Δθ
iΔr Δr
Fig.5.14: Nomenclature for the volume element surrounding an intermediate node (i,j)
Energy balance equation for the volume element surrounding the node (i,j) can be written as
follows:
Q(i-1,j) – (i,j) + Q(i+1,j) – (i,j) + Q(i,j+1) – (i,j) + Q(i,j-1) – (i,j) + (Q'g' ' )(i , j) =0
T i , j−1−T i , j
[Δr*1] ( iΔr Δθ )
T i , j+1 −T i , j
+ k [Δr*1] ( iΔr Δθ ) ' ''
+ (iΔrΔθ*Δr*1) ( q g )(i, j ) = 0
1 1 1 1
Or [1 − ] Ti−1,j + [1 + ] Ti+1,j + Ti,j-1 + Ti,j+1
2i 2i (iΔθ )2 (iΔθ )2
2 Δr 2
− [2 +
(iΔθ )2
] Ti,j = −
k
( q 'g' ' )(i, j )
5.7.2 Finite difference equations for innermost node (0,0): The volume element
surrounding the innermost node (0,0) will be as shown in Fig.5.15.Energy balance equation
233
Δθ
(1,j)
(0,0)
Δr
for the volume element surrounding the node (0,0) can be written as follows:
Δr
Δθ∗1
2
Or k T 1, j −T 0,0 + ( Δr2 Δθ∗Δr ∗1 ) ( q '' '
)
g (0,0) = 0 for j = 1,2,3, …..N-1
(¿ )
( Δr ) 2
2
Δr
Or T1,j – T0,0 = −
2k
(q 'g'' )(0,0) for j = 1,2,3, …..N-1
(5.45)
5.7.3.Finite difference equations for nodes on the outer surface with convective boundary
condition: The volume element surrounding a node (M,j) on the outer surface of the cylinder
will be as shown in Fig.5.16.Energy balance equation for this element can be written as
follows:
(M,j+1)
h,T∞
(M,j)
(M-1,j)
234
Δθ
Δθ
(M,j-1)
MΔr
Q(M-1,j) – (M,j) + Q(M,j+1) – (M,j) + Q(M,j−1) – (M,j) + Qconvection + (Qg)M,j = 0 for j = 1,2,3,….N
T M−1, j −T M , j T M , j−1−T M , j
Or k { (MΔr −
Δr
2
)Δθ*1} ( Δr ) + k(
Δr
2
* 1) ( MΔrΔθ ) + k(
Δr
2
* 1) ( T MΔrΔθ
−T
M , j+1
) M, j
Δr Δr
+ h (MΔrΔθ*1)(T∞ − TM,j) + (MΔr −
4
)Δθ*
2
*1 (q 'g'' )(M , j) = 0 for j
=1,2,3,….N
Simplifying we get
2
hΔr
=− ( )
k
T∞ − ( 1−
1
4M )( ) (q
Δr
2
' ''
)
g (M , j) for j=1,2,3,..N
(5.46)
Example 5.9.A a heating element in the form of a short solid cylinder of radius 2 cm and
length 6 cm is generating heat at a uniform rate of 2*106 W/m3. The thermal conductivity of
the heating element may be taken as 50 W/(m-K). The element is exposed to the surroundings
at a uniform temperature of 300C with a surface heat transfer coefficient of 200 W/(m2-
k).Assuming two dimensional steady state conduction in r and z direction (i) obtain the finite
difference equations to determine the steady state temperature in the cylinder assuming Δr =
Δz = 1 cm (ii) solve the system of equations thus obtained for temperature distribution in the
cylinder.Assume that the circular faces of the cylinder are insulated.
4 5 6 7 R Insulated
235
Δr
0 1 2 3 z
Δz
Find: (i) Finite difference equations to determine the temperatures at discrete points in the
cylinder;(ii Temperatures from the finite difference equations thus formed
Assumptions: (i) two dimensional steady state conduction in r and z directions only
(ii)Thermal conductivity and the heat generation are constants.(iii) Heat transfer coefficients
from all the outer surfaces (the curved surface and the flat surface)are same.
Solution: Since there are two symmetry lines as shown in the figure only one fourth of the
cylinder need be considered for finite difference formulation .Since Δr = Δz = 1 cm there will
be 12 nodes as shown in the figure above.
Δr
0 1
Δz
The nomenclature for the volume element surrounding node 0 is shown in the figure
above.Energy balance equation for this volume element can be written as
Q4-0 + Q1-0 + Qg = 0
Or k 2π (Δr/2)(Δz/2) ( T Δr−T )
4 0
+ k π (Δr/2)2 ( T Δr−T )
1 0 '''
+ π (Δr/2)2)(Δz/2) ( q g )0
= 0
Δr 2 '''
4(T4 – T0) + 2(T1 – T0) = − ( q g )0
k
236
'''
Substituting for Δr,k and ( q g )0 we get
0.012
− 6T0 + 2T1 + 4T4 = − *2*10 6
50
Finite difference Equations for node ‘3’: The volume element surrounding the node 3 is
shown in figure below.Energy balance equation for the volume element surrounding node 3
can be written as
7 Insulated
Axis of symmetry
Δr
2 3
Δz
Q7-3 + Q2-3 + Qg = 0
T 7 −T 3 T 2−T 3
Or k 2π (Δr/2)(Δz/2) ( Δr ) + k π (Δr/2)2 ( Δr ) '''
+ π (Δr/2)2)(Δz/2) ( q g )3 =
0
Finite Difference Equations for nodes 1 and 2: The volume elements surrounding nodes 1
and 3 are shown below.
5 6
Δz
Δr
0 2 1 1 2 3
Axis of symmetry
Energy balance equation surrounding the node 1 can be written as
237
k π(Δr/2)2 ( T Δz−T )
'''
0 1
+ k 2π (Δr/2)Δz ( T Δr−T )
5 1
+ k π(Δr/2)2 ( T Δz−T )
2 1
+π
(Δr/2)2Δz ( q g )1 =0
0.012
T0 – 6T1 + T2 + 4T5 = − * 2* 106
50
k π(Δr/2)2 ( T Δz−T )
1 2
+ + k 2π (Δr/2)Δz ( T Δr−T )
6 2
+ k π(Δr/2)2 ( T Δz−T )
3 2
'''
+ π (Δr/2)2Δz ( q g )2 =0
0.012
T1 + 4T6 + T3 – 6T2 = − * 2* 106
50
Finite Difference Equations for node 4: The volume element surrounding the node 4 is
shown in the figure below.Energy balance equation for the element is
238
Or k 2π (Δr/2) (Δz/2) ( T Δr−T )
0 4
+ k 2π (Δr +Δr/2) (Δz/2) ( T Δr−T )
8 4
+
[( )] (
2 2
T 5−T 4
kπ Δr +
Δr
2 ) (
− Δr−
Δr
2 Δz )
[( Δr 2
)]
2
Δr
+π Δr +
2 ) (
− Δr−
2
(Δz/2) ( q g ) 4
'''
=0
4 5
Δr
Δz
0
2 Δr 2 '''
T0 – 8T4 + 3T8 = − (qg )4
k
'''
Substituting the values for Δr,k, and ( q g )0 we get
Finite Difference Equations for nodes 5 and 6: Both these nodes are internal nodes and
hence the heat transfer across the boundary surfaces of these nodes are only by
conduction.The volume elements surrounding node 5 is shown below.
9
5
4 6
Δr
Δz
239
1
Fig.: Volume element surrounding node 5
Energy balance equation for the volume element surrounding node 5 can be written as
[( ) ( ) ] ( T Δz−T )
Δr 2 Δr 2
Or k π Δr +
2
−
2
4 5
+ k 2π ( Δr + Δr2 ) Δz ( T Δr−T )
9 5
[( ) ( ) ] ( T Δz−T )
2 2
+kπ Δr +
Δr
2
−
Δr
2
6 5
+ k 2π ( Δr2 ) Δz ( T Δr−T )
1 5
[( ) ( ) ] Δz (q )
Δr 2 Δr 2
' ''
+π Δr + − g 5 =0
2 2
2 Δr 2 '''
T1 + 2T4 − 8T5 + 2T6 + 3T9 = − ( q g )5
k
'''
Substituting the values for Δr,k, and ( q g )0 and simplifying we get
Finite Difference Equations for node 7: The volume element surrounding node 7 is shown
below.The energy balance equation is
11
Δr Insulated
6 7
240
Δz
3
[( ) ( ) ] ( T Δz−T )
Δr 2 Δr 2
Or kπ Δr +
2
−
2
6 7
+ k 2π (Δr +Δr/2) (Δz/2) ( T Δr−T )
11 7
+
[( ) ( )]
Δr 2 Δr 2
'''
+π Δr + − (Δz/2) ( q g )7 =0
2 2
2 Δr 2 '''
T3 + 4T6 – [ 8 ] T7 + 3T11 = − ( q g )7
k
'''
Substituting the values for Δr,k,h,T∞ and ( q g )7 and simplifying we get
Finite Difference Equations for node 8: The volume element surrounding the node 8 is
shown in the figure below. Energy balance equation for the element is given by
8 h,T∞ 9
Δr
Δz
Symmetry line
Δr
241
Or
2
T 4 −T 8
k 2π (Δr+ Δr/2)Δz/2) ( Δr ) +kπ ( 2
( 2 Δr ) −
2{ } ) ( T Δz−T )
3 Δr 9 8
+
2
h{2π 2Δr (Δz/2)}(T∞ −T8) + π ( ( 2 Δr ) −
2
{ } )( Δz/2)
3 Δr
2
'''
( q g )8 =0
7 Δr 2 '''
6 T4 – [13 + 8 (hΔr/k)]T8 + 7T9 = − 8 (hΔr/k)T∞ − ( q g )8
k
'''
Substituting the values for Δr,k,h,T∞ and ( q g )8 and simplifying we get
Finite Difference Equations for nodes 9 and 10 : The volume element surrounding node 9
is shown in figure below.Energy balance equation for the volume element shown is
h,T∞
8 9 10
Δr
Δz
5
Axis of cylinder
Δr
Or
k π{(2Δr)2 – (Δr + Δr/2)2} ( T Δz−T )
8 9
+ k 2π(Δr + Δr/2)Δz ( T Δr−T )
5 9
+
T 10−T 9
k π{(2Δr)2 – (Δr + Δr/2)2} ( Δz ) + h2π(Δr + Δr/2)Δz (T∞ − T9)
'''
+ π{(2Δr)2 – (Δr + Δr/2)2}Δz ( q g )9 =0
242
Substituting Δz = Δr and simplifying we get
'''
12 T5 + 7T8 – [26 + (12hΔr/k)]T9 + 7T10 = − (12hΔr/k)T∞ − 7 (Δr2/k) ( q g )9
'''
Substituting the values for Δr,k,h,T∞ and ( q g )8 and simplifying we get
h,T∞
9 10 11
Δr
Δz
6
Axis of cylinder
Δr
It can be shown that the energy balance equation for this volume element will be
Finite Difference Equations for node 11: The volume element surrounding node 11 is
shown in the figure below. The energy balance equation for this element is as follows
h,T∞
10 11 Insulated
Δr
Δz
7
Fig. Volume element surrounding node 11
Δr
243
Q10-11 + Q7-11 + Qconvection + (Qg)11 = 0
Or
T 10−T 11 T 7 −T 11
k π[(Δr + Δr/2)2 – (Δr/2)2] ( Δz ) + k 2π (Δr + Δr/2)Δz ( Δr )
+ h 2π(2Δr)(Δz/2)(T∞ − T11)
'''
+ π[(Δr + Δr/2)2 – (Δr/2)2](Δz/2) ( q g )11 =0
'''
Substituting the values for Δr,k,h,T∞ and ( q g )8 and simplifying we get
Equations (1) to (12) form a system of linear equations which can be solved by using any of
the standard techniques.
In steady state conduction problems the solution obtained is valid for any time since under
steady state conditions, the temperatures in the solid do not change with time. But in transient
conduction problems the temperatures change with time as well as with position. Therefore
the finite difference solution of transient problems requires discretization in time in addition
to discretization in space as shown in Fig.5.17 for one dimensional transient conduction
problem.The discretization in time domain is done by selecting a suitable time step Δt and
solving for unknown nodal temperatures for each time step Δt until the solution at the desired
time is obtained.In a transient conduction problem a superscript (say ‘k’) is used to indicate
the index or counter of time steps with k = 0 corresponding to the specified initial condition.
Thus Tik represents the temperature at the node ‘i’ after time steps kΔt have elapsed.The
energy balance equation for transient conduction involves one additional term representing
the change in the energy content of the solid with time, Hence this equation for a volume
element for time duration of Δt can be written as follows:
{Net amount of heat transfer into the volume element across all its surfaces during time
interval Δt} + {Heat generated within the element during time interval Δt}
244
= {The amount of increase of the energy content of the element during the time
interval Δt}.
k +1 k+1 k+1
T Ti-1 Ti Ti+1
k +1
Δt
k
k–1 k
Ti+1
k
Ti k
Ti-1
Δt Δx Δx
x
0 1 2 i–1 i i +1
Where the rate of heat transfer Q normally consists of conduction terms for interior nodes,
but may involve convection, surface heat flux and radiation for boundary nodes.
Noting that ΔEelement = ρ ΔVCp ΔT, where ρ is the density,ΔV is the volume and Cp is the
specific heat of the element and dividing Eq. (5.44) by Δt we have
ΔT
∑Q + [Qg]element = ρΔV Cp
Δt
(5.45)
All sides
245
Eq.(5.45) is applicable to volume element surrounding any node ‘i’ in the domain.ΔT
represents the change in the temperature of the volume element surrounding the node i for a
time step of Δt and this can be written as [Tik+1 – Tik]. Hence Eq. (5.45) reduces to
∑Q + [Qg]element = ρΔV Cp
( T ik+1−T ki )
Δt
(5.45)
All sides
The term ∑Q involves temperatures at adjacent nodes i – 1 and i +1 and question arises as
All sides
whether to use these temperatures at previous time step ‘k’ or the new time step ‘k+1’. Both
these approaches are used in practice. The approach using the previous time step is called the
explicit method where as the method using the new time step is called the implicit method
Therefore Eq.(5.45) takes the following forms for the two methods:
The explicit and implicit formulations given in Eqs.(5.46) and (5.47) are quite general and
can be used in any coordinate system and is applicable to multi dimensional transient
conduction problems also except that the volume element in multi dimensional problems will
have more surfaces and thus involve more terms in the summation.
The explicit method is easy to implement but imposes a limit on the allowable time step
in order to avoid instabilities in the solution and the implicit method requires the nodal
temperatures to be solved simultaneously for each time step but imposes no limit on the
magnitude of the .time step
246
(Q'g' ' )i
Qi-1,I Qi+1,i
x
0 1 2 3 i-1 i i+1 M-1 M
ΔE
Qi-1,i+ Qi+1,i + (Qg) i =
Δt
T ki−1−T ki T ki +1−T ik
Or k AΔx ( Δx ) + k AΔx ( Δx ) + AΔx q'g' ' = ρAΔxCp
T ki +1−T ki
( Δt )
T ki−1 +T ki+1−2 T ik
( )
k +1 k
( T i −T i
Δt ) =α
( Δx )2
αΔt
Where τ=
Δx 2
(5.49)
τ is the Fourier number for the element under consideration with characteristic length Δx.
247
Eq. (5.48) is called the explicit scheme because the temperature at any node i at the present
time step (k+1) is explicitly expressed in terms of the temperatures at the previous time step
k.
Restriction on τ (Stability Criterion): The explicit method, though easy to use has an
undesirable feature which restricts its utility. The method is not unconditionally stable and
there is a restriction on the value of the parameter τ if the scheme has to be stable.The
restriction is that in Eq.(5.48) ( 1−2 τ ) ≥ 0 i.e.
0<τ≤ ½ (5.50)
This restriction implies that for given values of α and Δx, the time step Δt cannot exceed the
limit imposed on it by Eq. (5.50). This is called the stability criterion. It can be shown
mathematically or by invoking second law of thermodynamics that the stability criterion is
satisfied if the coefficients of all T ik in the expressions for T ik+1 are greater than or
equal to zero for all nodes.If this stability criterion is violated then the solution becomes
unstable as shown in Fig.5.19.It is instructive to determine this stability criterion by the
following physical argument.
k
Let at any time step k the temperatures T i−1 and T ik+1 at nodes i-1 and i+1 are
equal but less than T ik at node i between them. Then if the value of τ exceeds ½, the
coefficient ( 1−2 τ ) becomes negative. Then, according to Eq.(5.48),for negative value of
( 1−2 τ ) , the temperature T ik+1 at node i at the next time step will be less than that at the
neighbouring two nodes. This is not possible according to the second law of
thermodynamics, since we had assumed that T ik higher than that at neighboring nodes.
Hence to obtain meaningful solution it is necessary that ( 1−2 τ ) should not be negative.
Exact solution
Finite Difference Equations for Nodes on Boundaries: Eq. (5.48) can be used for all
internal nodes (i = 1,2,3,……………M-1).If the boundary temperatures are specified then we
have M-1 equations to solve for the temperatures at M-1 nodes.If the boundaries are
subjected to convection or prescribed surface heat flux then we have to develop the finite
difference equations for the boundary nodes also to determine the boundary temperatures.The
248
time dependent energy balance equation for the volume element surrounding the node 0 (see
Fig.5.20) subjected to convection can be written as follows:
k
(Q'g' ' ) 0
h 0 , T∞
0 1
Δx/2
Δx
k ΔE
Q1-0 + Qconvection + (Q'g' ' )0 =
Δt
T k1 −T 0K
Or kA ( Δx ) + h0 A (T ∞−T k0 ) + A(Δx/2)
k
(q 'g'' )0 = ρ A(Δx/2) Cp
T k0 +1−T k0
( Δt )
Solving for T k+1
0 we get
Similarly for the other boundary node M subjected to convection boundary condition the
finite difference equation can be written as
h M Δx k hM Δx 2
T k+1
M = ( 1−2 τ −2 τ
k )TM + 2τ T kM−1 + 2τ
k
T∞ + τ
Δx ' ' ' k
k
( qg )0
(5.52)
k k
For stability of the scheme the coefficients of T 0 and T M should be ≥ 0. i.e.
249
(1−2 τ −2 τ h kΔx ) ≥ 0
0
and (1−2 τ −2 τ MΔx
k )
≥0
1
Or τ≤
2( 1+ h0 Δx /k )
(5.53a)
1
and τ≤
2(1+ hM Δx /k)
(5.53b)
Example 5.10.A marble slab[k = 2W/(m – 0C) and α = 1*10 −6 m2/s],2 cm thick is initially at
a uniform temperature of 200 0C. Suddenly one of its surfaces is lowered to 00C and is
maintained at that temperature, while the other surface is kept insulated. Develop an explicit
finite difference scheme to determine the temperature distribution in the slab as a function of
position after 80 seconds have elapsed by dividing the slab into 5 sub-regions and also find
the heat flux at the boundary surface which is maintained at 00C after 80 seconds.
Insulated T5
0
0 1 2 3 4 5 x
Find: (i) Finite difference equations to determine temperatures at nodes 0 to 4at different
time steps; (ii) q5k
Solution: The finite difference equations for internal nodes from 1 to 4 are given by Eq.
(5.48) and (5.49) with q'g' ' = 0 and M = 5. With these conditions we have
Ti
k+1
=τ (T ki−1+ T ki+ 1) + ( 1−2 τ ) T ki i = 1,2,3,and 4.
250
For i = 2 T 2k+1 = τ (T k1 +T k3 ) + ( 1−2 τ ) T k2 (b)
For i = 3 T3
k+1
=τ (T k2 +T k4 ) + ( 1−2 τ ) T k3 (c)
For i = 4 T k+1
4 =τ (T k3 +T k5 ) + ( 1−2 τ ) T k4 (d)
k k
( T 0 −T 1
Δx ) = 0 for all k.
k+1 k+1
Therfore we have T 0 =T 1 (e)
αΔt
Choosing 1 − 2τ = 0 we have τ = ½ =
Δx 2
Δx 2 0.004 2
Or Δt = = = 8 s.
2α 2∗1∗10−6
k+1
T2 = 0.5* (T k1 +T k3 ) (ii)
T4
k+1
= 0.5* (T k3 +T k5 ) (iv)
T k+1
0 = T 1k+1 (v)
Eqs. (i) to (v) are used to calculate the temperatures at different nodes for each time step as
shown in the table below :
k t (s) i=0 1 2 3 4 5
x(m) = 0 00.004 0.008 0.012 0.016 0.020
251
0 0 200 200 200 200 200 200
1 8 200 200 200 200 200 0.00
2 16 200 200 200 200 100 0.00
3 24 200 200 200 150 100 0.00
4 32 200 200 175 150 75 0.00
5 40 187.5 187.5 175 125 75 0.00
6 48 181.2 181.2 156.25 125 62.5 0.00
7 56 168.7 168.7 153.1 109.4 62.5 0.00
8 64 160.9 160.9 139.1 107.8 54.7 0.00
9 72 150 150 134.4 96.9 53.9 0.00
10 80 142.2 142.2 123.5 94.2 48.5 0.00
Heat flux at the boundary surface at x = L after t = 80 s i.e after 10 time steps is given by
10 10
q10
M =k ( T 4 −T 5
Δx ) =2* ( 48.5−0.0
0.004 )
= 24,250W/m 2
Example 5.11. A large uranium plate of thickness 4 cm, thermal conductivity 28 W/(m-K)
and thermal diffusivity of 12.5* 10 −6 m2/s is initially at a uniform temperature of 200
0
C.Heat is generated at a constant rate of 5*10 6 W/m3.Suddenly at time t = 0, one side of the
plate is brought into contact with iced water and is maintained at 0 0C at all times, while the
other side is subjected to convection to an environment at 30 0C with a heat transfer
coefficient of 45 W/(m2-K). Considering a total of three equally spaced nodes, two at the
boundaries and one at the middle of the plate, estimate the temperature distribution in the
plate 150 s after the start of the cooling process using explicit method.
Node 1
L
T0 h,T∞
0 2 x
Δx/2
Δx
Known: L = 4 cm =0.04 m; k = 28 W/(m-K); α = 12.5 * 10 −6 m 2/s; T0 = 00C; M = 3;
T∞ = 300C;h2 = 45 W/(m 2-K); q'g' ' = 5 * 10 6 W/m3;
Find: (i) Finite difference equations to determine temperatures at nodes 1 and 2 after time
150 s
252
Assumptions: (i)the plate is having constant thermal properties and constant rate of heat
generation (ii) one dimensional conduction across the thickness of the plate
Solution: The finite difference equations for internal node 1 is given by Eq. (5.48) and
(5.49) with i = 1 and M = 2. With these conditions we have
' '' 2
q g Δx
T1
k+1
=τ (T k0 +T k2 ) + ( 1−2 τ ) T k1 + τ
k
(a)
αΔt
Where τ=
Δx 2
The finite difference equation for boundary node 2 is given by Eq.(5.52) with M = 2.
h2 Δx 45∗0.02
Δx = L / M = 0.04/2 = 0.02 m ; = = 0.032
k 28
From Eq.(a) for stability of the difference scheme 1−2 τ ≥ 0 or τ ≤ 0.5 (c)
1
Or (1 - 2τ – 2*0.032τ ) ≥ 0 i.e. τ ≤ i.e. ≤ 0.484
2.064
(d)
Therefore we have to choose a value of τ which satisfies both Eqs. (c) and (d)
0.484∗( 0.02 )2
Let τ = 0.484. Then Δt = = 15.5 s.
12.5∗10−6
Any time step ≤ 15.5 s can be used. Since we want the temperature after 150 s it will be
convenient to choose Δt = 15 s so that we get the required temperatures at nodes 1 and 2 after
10 time steps.Therefore
−6
12.5∗10 ∗15
τ= = 0.46875
( 0.02 )2
253
1−2∗0.46875 5∗106 (0.02)2
k+1
T1 = 0.46875* (T k0 +T k2 ) +( ¿) T
k + 0.46875*
1 28
k+1 k k
Or T1 = 0.0625 T 1 + 0.46875 T 2 + 33,482
(i)
The temperatures at nodes 1 and 2 at each time step are calculated using the above two
equations and substituting the temperatures at the previous time step. The results are shown
in Table E5.11
k t (s) i=0 1 2
x(m) = 0 0.02 0.04
0 0 0 200 200
1 15 0 139.7 228.4
2 30 0 149.3 172.8
3 45 0 123.8 179.9
4 60 0 125.6 156.3
5 75 0 114.6 157.1
6 90 0 114.3 146.9
7 105 0 109.5 146.3
8 120 0 108.9 141.8
9 135 0 106.7 141.1
10 150 0 106.3 139.0
254
5.8.1(b). Explicit Method for One-dimensional Transient Radial conduction in a cylinder
The energy balance equation for the volume element surrounding node i (see Fig. 5.5) for a
time step Δt can be written as follows.
ΔE
Qi-1,i+ Qi+1,i + (Qg) i =
Δt
Or
k [ ( ) ] [ ]
2 π iΔr−
Δr
2
∗1
T ki−1−T ik
Δr
+ ( 2 πiΔr∗Δr∗1 ) q'i ' '
[ ( ) ] [ ]
k k k +1 k
+ k 2 π iΔr +
Δr
2
∗1
T i+1−T i
Δr
= ρ *2πiΔr2*1*Cp ( T i −T i
Δt )
i = 1,2,3……M-1
T ik+1 = τ [( ) ( )
1−
1 k
2i
T i−1+ 1+
1 k
2i
T i+1 +q'i ' '
Δr2
k ] + [ 1−2 τ ] T ki
i = 1,2,3……M-1
αΔt
where τ = and α = k/(ρCp )
Δr 2
(5.54)
Referring to Fig. 5.5, the energy balance equation for node 0 can be written as follows:
Δr 2
k ( Δr
2 π ∗1
2 )[
T 0 −T 1
Δr ] + q'0' ' π
2
( Δr2 ) *1 = ρCpπ ( )
2
∗1∗( T 0k+1−T k0 )
Δt
255
After simplification the above equation reduces to
T k+1
0 = 4τ
[ k
T +
1
q'0' ' ( Δr )2
k ] + [ 1−4 τ ] T k0
(5.56)
(iii) Finite difference Equation for outermost node M when outer surface is subjected to
prescribed heat flux: The nomenclature for the volume element surrounding the outer most node M
is shown in Fig. 5.21.Energy balance equation for the volume element can be written as follows:
M qs
M-1
MΔr
256
[ ]
1 1
T k+1
=4 τ
( 1−
2M)
T
k
+ 1−4 τ
( 1−
2M )
T
k
+
4 τ q s Δr
1 +
M
(2− 21M )
M −1
(2− 21M )
M
(
k 2−
2M )
τ q'M' ' ( Δr )2
(5.58)
k
[ ]
1
(iii) Finite difference Equation for outermost node M when outer surface is in contact with a
fluid at a uniform temperature T∞ with a surface heat transfer coefficient h (convective
boundary condition):
The nomenclature for the volume element surrounding the outer most node M is shown in Fig.
5.22.Energy balance equation for the volume element can be written as follows:
M h,T∞
M-1
MΔr
257
Fig.5.21: Nomenclature for volume element surrounding the node M subjected to
convective condition
[ ]
1 1
TM
k+1
=4 τ
( 1−
2M) k
T M −1 + 1−4 τ
( 1−
2M )
−
4 τ hΔr k
TM +
(2− 21M ) (2− 21M ) k ( 2− 21M )
4 τ hΔr T ∞
τ q'M' ' ( Δr )2
1 +
(
k 2−
2M ) k
(5.60)
[ ]
1
Schematic : The cylinder is divided equally as shown in the figure with Δr = R/2, where
Δr 2 h,T∞
Δr/2 1
0 R
.
R is the radius of the cylinder.
258
Find: Temperatures at the nodes after 9 seconds.
Assumptions: (i) Constant physical properties for aluminium; (ii) Radial unsteady state
conduction
The finite difference equation for the internal node 1 is given by Eq.(5.54) with i = 1.
αΔt
where τ = 2 .
Δr
The finite difference equation for the internal node 0 is given by Eq.(5.54) with q'0' ' = 0.
k+1
i.e. T0 = 4τ T 1k + [ 1−4 τ ] T k0
…………………………………(3)
The finite difference equation for the outermost node 2 is given by Eq.(5.60) with q'2' ' = 0.
[ ]
1 1
i.e., T
k+1
=4 τ
( 1− )
4
T
k
+ 1−4 τ
( 1− )
4
−
4 τ hΔr
T
k
+
4 τ hΔr T ∞
2
(2− 14 )
1
(2− 14 ) k (2− 14 )
M
( 14 )
k 2−
k+1 k k
Or T 2 =¿ 1.714τ T 1 + [ 1 – 1.943τ] T 2 + …………………….
(5)
259
1
For stability 1 – 1.943τ ≥ 0 Or τ ≤
1.943
………………………….(6)
The numerical value of τ has to be selected such that all the three stability conditions viz.,
Eqs.(2),(4) and (6) are satisfied. The condition τ ≤ ¼ will satisfy all the three stability
condtions.Hence choose τ ≤ ¼ : i.e. τ ≤ 0.25.
Or Δt ≤ 2.27 s
t 9
Selecting Δt = 2.25 s, τ = 0.247 no of time steps required = = =4
Δt 2.25
With the value of τ = 0.247 the nodal equations (1), (3) and (5) reduce to
T k+1
0 = 0.012 T k0 + 0.988 T 1k (a)
k+1
T1 = 0.123 T k0 + 0.506 T 1k + ¿ 0.37 T 2k
(b)
k+1
And T 2 =¿ 0.423 T 1k + 0.52 T 2k + 2.82 (c)
The temperatures at the three nodes at different time steps are calculated using Eqs. (a), (b)
and (c) and are shown in the table below.
Time step (k) Δt (s) t = kΔt (s) Temperature at different nodes (0C)
T0 T1 T2
0 2.25 0 100 100 100
1 2.25 2.25 100 99.9 97.10
2 2.25 4.50 99.9 98.78 95.57
3 2.25 6.75 98.79 97.63 94.30
4 2.25 9.00 97.64 96.44 93.15
260
5.8.2.Implicit Method for Transient conduction
In the explicit finite difference scheme, the temperature at any node at time (t + Δt)
is calculated from knowledge of temperatures at the same and neighboring nodes for the
preceding time t. Hence determination of temperature at a node at some time is independent
of temperatures at other nodes for the same time. This method, though offers computational
convenience, it suffers from the limitations on the time step Δt. For a given Δx, the time step
is limited by the stability requirements as illustrated in examples 5.10 and 5.11. This dictates
the use of smaller Δt resulting in large number of time steps to determine the transient
tenmpeartures of the nodes after certain time.By employing an implicit scheme the restriction
on Δt can be completely eliminated.In the implicit scheme the space derivatives in finite
difference form are expressed using temperatures of the all the nodes at time t+∆t, while the
backward finite difference form of the derivative of temeperature iwthrespect to time is
used.
Refering to Fig 5.18, for a plane wall the energy balance equation in implicit form for any
internal node can be written as
T ki−1−T k+ 1
T ki +1−T ik+1 T ki +1−T ki
k AΔx ( Δx
i
) + k AΔx ( Δx ) + AΔx q'''
g = ρAΔxCp ( Δt )
i = 1,2,3….. M – 1
Similar equations for the boundary nodes 0 and M can be obtained depending on the
boundary conditions specified.
261
Chapter 6
6.2. Heat Transfer Coefficient:- In engineering application, to simplify the heat transfer
calculations between a hot surface say at temperature Tw and a cold fluid flowing over it at a
bulk temperature T∞ as shown in Fig. 6.1 a term called “heat transfer coefficient,h” is defined
by the equation
q = h(Tw – T∞)………………………………………………..6.1(a)
where q is the heat flux (expressed in W / m2) from the surface to the flowing fluid.
Alternatively if the surface temperature is lower than the flowing fluid then the heat transfer
takes place from the hot fluid to the cold surface and the heat flux is given by
q = h(T∞ – Tw)………………………………………………..6.1(b)
The heat flux in this case takes place from the fluid to the cold surface.If in equations 6.1(a)
and 6.1(b) the heat flux is expressed in W / m2, then the units of heat transfer coefficient will
be W /(m 2 – K) or W / (m 2 – 0 C).
The heat transfer coefficient is found to vary with (i) the geometry of
the body, (ii) the type of flow (laminar or turbulent), (iii) the transport properties of the fluid
262
(density, viscosity,specific heat and thermal conductivity),(iv) the difference in temperature
between the average temperature of the fluid and the surface with which the fluid is in
contact , (v) the position along the surface of the body, and (vi) whether the heat transfer is
by forced convection or free convection. For convection problems involving simple
geometries like flow over a flat plate or flow inside a circular tube, the heat transfer
coefficient can be
u∞, T∞ T∞
x Tw
determined analytically. But for flow over complex configurations, experimental / numerical
approach is used to determine h.There is a wide difference in the range of values of h for
various applications.Typical values of heat transfer coefficients encountered in some
applications are given in Table 6.1.
263
Condensing steam 5500 – 120000
6.3. Basic concepts for flow over a body:- When a fluid flows over a body, the velocity and
temperature distribution at the vicinity of the surface of the body strongly influence the heat
transfer by convection.By introducing the concept of boundary layers (velocity boundary
layer and thermal boundary layer) the analysis of convective heat transfer can be simplified.
6.3.1.Velocity Boundary Layer:- Consider the flow of a fluid over a flat plate as shown in
Fig. 6.2 The fluid just before it approaches the leading edge of the plate has a velocity u ∞
which is parallel to the plate surface.As the fluid moves in x-direction along the plate,
u∞
y
u(x, y)
δ(x)
u(x, y)
x
xcr Turbulent Region
Laminar Region Transition
Fig. 6.2: Velocity boundary layer for flow over a flat plate
those fluid particles that makes contact with the plate surface will have the same velocity as
that of the plate. Therefore if the plate is stationary, then the fluid layer sticking to the plate
surface will have zero velocity (This is called no-slip condition).But far away from the plate
(y = ∞) the fluid will have the velocity u ∞.Therefore starting from the plate surface (y = 0)
there will be retardation of the fluid in x-direction component of velocity u(x,y).This
retardation effect is reduced as we move away from the plate surface.At distances sufficiently
long from the plate(y = ∞) the retardation effect is completely reduced: i.e. u → u ∞ as y → ∞.
This means that there is a region surrounding the plate surface where the fluid velocity
changes from zero at the surface to the velocity u∞ at the outer edge of the region. This region
is called the velocity boundary layer. The variation of the x-component of velocity u(x,y)
with respect to y at a particular location along the plate is shown in Fig. 6.2.The distance
measured normal to the surface from the plate surface to the point at which the fluid attains
99% of u∞ is called “velocity boundary layer thickness” and denoted by δ(x)
Thus for flow over a flat plate, the flow field can be divided into two distinct regions,
namely, (i) the boundary layer region in which the axial component of velocity u(x,y) varies
264
rapidly with y with the result the velocity gradient (∂u /∂y) and hence the shear stress are
very large and (ii) the potential flow region which is outside the boundary layer region,
where the velocity gradients and shear stresses are negligible.
The flow in the boundary layer, starting from the leading edge of the
plate will be initially laminar in which the fluid particles move along a stream line in an
orderly manner. In the laminar region the retardation effect is due to the viscosity of the fluid
and therefore the shear stress can be evaluated using Newton’s law of viscocity.The laminar
flow continues along the plate until a critical distance ‘x cr’ is reached. After this the small
disturbances in the flow begin to grow and fluid fluctuations begin to develop. This
characterizes the end of the laminar flow region and the beginning of transition from laminar
to turbulent boundary layer. A dimensionless parameter called Reynolds number is used to
characterize the flow as laminar or turbulent. For flow over a flat plate the Reynolds number
is defined as
u∞ x
Rex =
ν
……………………….6.2
where u∞ = free-stream velocity of the fluid, x = distance from the leading edge of the plate
and ν = kinematic viscosity of the fluid.
For flow over a flat plate it has been found that the transition from
laminar flow to turbulent flow takes place when the Reynolds number is ≈ 5 x 10 5.This
number is called as the critical Reynolds number Recr for flow over a flat plate.Therefore
u ∞ x cr 5
Re cr = = 5 x 10
ν
………….6.3
The critical Reynolds number is strongly dependent on the surface roughness and the
turbulence level of the free stream fluid. For example, with very large disturbances in the free
stream, the transition from laminar flow to turbulent flow may begin at Re x as low as 1 x 10 5
and for flows which are free from disturbances and if the plate surface is smooth transition
may not take place until a Reynolds number of 1 x 10 6 is reached.But it has been found that
for flow over a flat plate the boundary layer is always turbulent for Re x ≥ 4 x 10 6.In the
turbulent boundary layer next to the wall there is a very thin layer called “the viscous sub-
layer”, where the flow retains its viscous flow character. Next to the viscous sub-layer is a
region called “buffer layer” in which the effect of fluid viscosity is of the same order of
magnitude as that of turbulence and the mean velocity rapidly increases with the distance
from the plate surface.Next to the buffer layer is “the turbulent layer” in which there is large
scale turbulence and the velocity changes relatively little with distance.
Boundary Layer Concept for flow over a Curved Body: Fig. 6.3 illustrates the boundary
layer growth for flow over a curved body.Here ‘x’ coordinate is measured along the curved
surface of the body. Starting from the stagnation point, and at each location along x, the y
265
coordinate is measured normal to the surface of the body.The free stream velocity U ∞, in this
case is not a constant as in the case of flow over a flat plate but is a function of x. The
Fig. 6.3: Boundary layer growth for flow over a curved body
boundary layer thickness δ(x) increases with the distance x along the surface. But, because of
the curvature of the body, after some distance, the velocity profile u(x,y) exhibits a point of
inflexion; that is at this location (ðu / ðy) becomes zero at the wall surface. Beyond the point
of inflexion, reversal of flow takes place, and the boundary layer is said to be detached from
the surface.Beyond the point of flow reversal the flow become very complicated and the boundary
layer analysis is no longer valid.
6.3.2. Drag coefficient and Drag force:- If the velocity distribution u(x,y) in the boundary
layer at any ‘x’ is known then the viscous shear stress at the wall can be determined using
Newton’s law of viscosity. Thus the wall-shear stress, τ w(x) at any location x can be written
as :
where μ is the absolute viscosity of the fluid.The drag coefficient is dimensionless wall shear
stress. Therefore the local drag coefficient, Cx at any ‘x’ is defined as
τ w ( x)
Cx = 1 2
ρu
2 ∞
……………………….6.5
Substituting for τw(x) in the above equation from Eq. 6.4 and simplifying we get
2ν (∂u / ∂y)y = 0
Cx = ---------------------- ……………….6.6
266
u∞2
Therefore if the velocity profile u(x,y) at any x is known then the local drag coefficient C x at
that location can be determined from Eq. 6.6.The average value of C x for a total length L of
the plate can be determined from the equation
L
Cav = (1/L) ∫Cx dx ……………………6.7
0
_
τw
Or Cav = -------------------- ……………….6.8
(1/2) ρu∞2
_
Where τw is the average wall-shear stress for total length L of the plate.
The total drag force experienced by the fluid due to the presence of the plate can be written
as
_
FD = As τw …………………………….6.9
Where As is the total area of contact between the fluid and the plate. If ‘W’ is the width of the
plate then As = LW if the flow is taking place on one side of the plate and A s = 2LW if the
flow is on both sides of the plate.
6.3.3.Thermal boundary layer:- Similar to the velocity boundary layer one can visualize the
development of a thermal boundary layer when a fluid flows over a flat plate with the
temperature of the plate being different from that of the free stream fluid.Consider that a fluid
at a uniform temperature T∞ flows over a flat plate which is maintained at a uniform
temperature Tw.Let T(x,y) is the temperature of the fluid at any location in the flow field.Let
the dimensionless temperature of the fluid θ(x,y) be defined as
T(x,y) – Tw
θ(x,y) = ------------------- …………………………….6.10
T∞ − Tw
The fluid layer sticking to the plate surface will have the same temperature as the plate
surface [T(x,y)y = 0 = Tw] and therefore θ(x,y) = 0 at y = 0.Far away from the plate the fluid
temperature is T∞ and hence θ(x,y) → 1 as y → ∞. Therefore at each location x along the
plate one can visualize a location y = δ t(x) in the flow field at which θ(x,y) = 0.99. δ t(x) is
called “the thermal boundary layer thickness” as shown in Fig. 6.3. The locus of such
267
points at which θ(x,y) = 0.99 is called the edge of the thermal boundary layer. The relative
thickness of the thermal boundary layer δt(x) and the velocity
boundary layer δ(x) depends on a dimensionless number called “Prandtl number” of the
fluid.It is denoted by Pr and is defined as
μCp (μ/ρ) ν
Pr = --------- = ---------- = -------- ………..6.11
k (k/ρCp) α
Where μ is the absolute viscosity of the fluid, Cp is the specific heat at constant pressure
k is the thermal conductivity,υ is the kinematic viscosity and α is the thermal diffusivity of
the fluid.The Prandtl number for fluids range from 0.01 for liquid metals to more than
100,000 for heavy oils. For fluids with Pr = 1 such as gases δ t(x) = δ(x), for fluids with Pr
<< 1,such as liquid metals, δt(x) >> δ(x) and for fluids with Pr >> 1, like oils δt(x) << δ(x).
6.3.4. General expression for heat transfer coefficient:- Let us assume that Tw > T∞.Then
heat is transferred from the plate to the fluid flowing over the plate.Therefore at any ‘x’ the
heat flux is given by
q = − k (∂T /∂y)y=0 ………………………..6.12(a)
In terms of the local heat transfer coefficient hx, the heat flux can also be written as
268
From equation 6.10 we have (∂T /∂y)y=0 = [T∞ − Tw] (∂θ /∂y)y=0. Substituting this expression
in Eq.6.13 and simplifying we get the general expression for hx as
The same expression for hx could be obtained even when Tw < T∞. Equation 6.14 can be used
to determine the local heat transfer coefficient for flow over a flat plate if the dimensionless
temperature profile θ(x,y) is known.
Average heat transfer coefficient:- For a total length L of the plate the average heat transfer
coefficient is given by
L
hav = (1 /L) ∫hxdx ……………………………..6.15a
0
Since (∂θ /∂y)y=0 at any x depends on whether the flow at that section is laminar or turbulent
the expression for hav can be written as
xcr L
hav = (1 /L) { ∫ k [(∂θ /∂y)y=0]laminar dx + ∫ k [(∂θ /∂y)y=0]turbulent dx }……5.16
0
xcr
Relation Between Fluid Friction and Heat Transfer (Reynolds – Colburn Analogy):
For laminar flow over a flat plate an analogy between momentum transfer and heat transfer
exists; that is, it is possible to establish a relation between the local drag coefficient, C x and
the local Nusselt number Nux as shown below.
The excat expression for the local drag coefficient is given by
The exact expression for the local Nusselt number, Nux is given by
269
Stx = -------------------------- = ---------------
(ρ x u∞ /μ) (μ Cp/k) Rex Pr
Eq. 6.18 is also applicable to turbulent flow over a flat plate, but it does not apply to flow
through tubes.This analogy is referred to as “Reynolds – Colburn Analogy”.
Example 6.1:- Assuming the transition from laminar to turbulent flow takes place at a
Reynolds number of 5 x 10 5, determine the distance from the leading edge of a flat plate at
which transition occurs for the flow of each of the following fluids with a velocity of 2 cm/s
at 40 0 C.(i) Air at atmospheric pressure;(ii)Hydrogen at atmospheric pressure;(iii) water;
(iv) Engine oil;(v) mercury.Comment on the type of flow for the 5 fluids if the total length of
the plate is 1 m.
u∞ xcr Recr ν 5 x 10 5 x 17 x 10 − 6
Recr = ----------- or xcr = -------------------- = ---------------------------- = 4.25 m.
ν u∞ 2
5 x 10 5 x 117.9 x 10 − 6
Therefore xcr = ------------------------------- = 29.5 m
2
5 x 10 5 x 0.658 x 10 − 6
Therefore xcr = ------------------------------- = 0.1645 m
2
5 x 10 5 x 0.24 x 10 − 3
Therefore xcr = ------------------------------- = 60 m
270
2
5 x 10 5 x 0.107 x 10 − 6
Therefore xcr = ------------------------------- = 0.027 m
2
3 Water 0.1645 xcr < L Flow is partly Laminar & Partly Turbulent
5 Mercury 0.027 xcr << L Flow is turbulent for almost entire length
Example 6.2:- An approximate expression for the velocity profile u(x,y) for laminar
boundary layer flow along a flat plate is given by
where δ(x) is the velocity boundary layer thickness given by the expression
271
Therefore (∂u / ∂y)y=0 = 2u∞ / δ(x)
0.686 0.686
= ---------------------- = -----------------
{(u∞x) / ν}1/2 (Rex) ½
{ 0.686 (u∞/ν) − ½} L
= ----------------------- ∫ x − ½ dx
0
L
Or 2 x 0.686 1.372
Cav = --------------- = ------------------
(u∞L / ν) ½ (ReL) ½
u∞L 4x2
ReL = ----------------- = ------------------- = 3.793 x 10 5
ν 21.09 x 10 − 6
1.372 1.372
Average drag coefficient = Cav = ------------------- = ------------------------ = 2.228 x 10 − 3
ReL0.5 (3.793 x 10 5)0.5
Drag force assuming that the flow takes place on one side of the plate is given by
_
FD = τw LW = (1/2)ρ u∞2 Cav LW for flow over one side of the plate
272
= (1/2) x 1.00 x 42 x 2.2228 x 10 − 3 x 2 x 2
= 0.071 N
Example 6.3:- An approximate expression for temperature profile θ(x,y) in the thermal
boundary layer region for flow over a flat plate is given by
hx = k (∂θ / ∂y)|y = 0.
Now θ(x,y) = 2(y / δt )− (y / δt)2
2 Rex0.5Pr 1/3
Hence (∂θ / ∂y)|y = 0. = 2 / δt = ---------------------
5.5.x
0.5 1/3
2 k Rex Pr
Or hx = ------------------------ = 0.364 (k / x) Rex0.5 Pr 1/3
5.5.x
hx x
----- = 0.364 Rex0.5 Pr 1/3
K
hx x /k is a dimensionless number involving local heat transfer coefficient and is called
(ii) The average heat transfer coefficient for a total length L of the plate is given by
L
hav = (1 / L) ∫hx dx = (1 / L) ∫ 0.364 (k / x) Rex0.5 Pr1/3 dx
0
= (1 / L) (L 0.5 /0.5) 0.364 Pr1/3 k (U∞ / ν)0.5 = 0.728 (k / L) (U∞L / ν) 0.5 Pr 1/3
273
hav L / k is a dimensionless number involving the average heat transfer coefficient and is
called the “average Nusselt number”.
Example 6.4:- The heat transfer rate per unit width from a longitudinal section x 2 ─ x1 of a
flat plate can be expressed as q12 = h12 (x2 – x1)(Ts - T∞), where h12 is the average heat
transfer coefficient for the section length of (x 2 – x1). Consider laminar flow over a flat plate
with a uniform temperature Ts. The spatial variation of the local heat transfer coefficient is of
the form hx = C x ─ 0.5, where C is a constant.
(a) Derive an expression for h12 in terms of C,x1 and x2.
(b) Derive an expression for h12 in terms of x1, x2, and the average coefficients h1 and h2
corresponding to lengths x1 and x2 respectively.
Solution:
u∞, T∞ q12
Ts
x1
x2
(a) hx = C x − 0.5
_ 1 x2
Therefore h12 = -------------- ∫hx dx
(x2 – x1) x1
1 x2
= ------------------- ∫ C x − 0.5 dx
0
(x2 – x1)
2C
= ------------------ [ x2 0.5 – x1 0.5]
(x2 – x1)
_ x1
(b) h1 = (1/x1) ∫C x − 0.5 dx
0
= 2C / √ x1
_ ___
274
Similarly h2 = 2C / √ x2
x1 __ _ 1 x2 x1
Since ∫ hxdx = x1h1, h12 = -------------- [ ∫ hxdx - ∫ hxdx ]
0 0 0
( x2 – x1)
_ _
_ h2x2 - h1x1 {2C / √ x2 } ─ {2C / √ x1 }
h12 = ----------------- = ----------------------------------- = 2C [√ x2 ─ √ x1]
x2 – x1 (x2 – x1)
6.4. Basic Concepts For Flow Through Ducts :- The basic concepts developed on the
development of velocity and thermal boundary layers for flow over surfaces are also
applicable to flows at the entrance region of the ducts.
6.4.1.Velocity Boundary Layer:- Consider the flow inside a circular tube as shown in
Fig.6.5. Lat uo be the uniform velocity with which the fluid approaches the tube. As the fluid
enters the tube, a “velocity boundary layer” starts to develop along the wall-surface. The
velocity of the fluid layer sticking to the tube-surface will have zero velocity and the fluid
layer slightly away from the wall is retarded. As a result the velocity in the central portion of
the tube increases to satisfy the continuity equation (law of conservation of mass).The
thickness of the velocity boundary layer δ(z) continuously grows along the tube-surface until
it fills the entire tube. The region from the tube inlet up to little beyond the hypothetical
location where the boundary layer reaches the tube centre is called “hydrodynamic entrance
region or hydrodynamically developing region” and the
corresponding length is called “hydrodynamic entrance length Lh”. In the hydrodynamically
developing region the shape of the velocity profile changes both in axial and radial direction,
i.e., u = u(r,z). The region beyond the hydrodynamic entry length is called
“Hydrodynamically developed region”, because in this region the velocity profile is invariant
with distance along the tube,i.e., u = u(r).
If the boundary layer remains laminar until it fills the tube, then laminar flow will prevail
in the developed region. However if the boundary layer changes to turbulent before its
thickness reaches the tube centre, fully developed turbulent flow will prevail in the
hydrodynamically developed region. The velocity profile in the turbulent region is flatter
than the parabolic profile of laminar flow. The Reynolds number, defined as
is used as a criterion for change from laminar flow to turbulent flow. In this definition, u m is
the average velocity of the fluid in the tube, D h is the hydraulic diameter of the tube and ν is
the kinematic viscosity of the fluid. The hydraulic diameter is defined as
275
Hydrodynamic Entrance Hydrodynamically developed
uo Region Region
δ(z) r
R
Lh
z Fully developed profile
u = u(r)
Fig. 6.5: development of velocity boundary layer at entrance region of a tube
4 x Area of flow
Dh = ------------------------------ ……………………(5.20)
Wetted Perimeter
For flows through ducts it has been observed that turbulent flow prevails for
But this critical value is strongly dependent on the surface roughness, the inlet conditions and
the fluctuations in the flow. In general, transition may occur in the range 2000 < Re d < 4000.
It is a common practice to assume a value of 2300 fro transition from laminar flow to
turbulent flow.
6.4.2. Friction Factor and Pressure Drop Relations For Hydrodynamically Developed
Laminar Flow
In engineering applications, the pressure gradient (dp / dz) associated with the flow is a
quantity of interest, because this decides the pumping power required to overcome thr
frictional losses in the pipe of a given length.
Consider a differential length dz of the tube at a distance z from the entrance and
let this length be in the fully developed region. The various forces acting on the fluid
element in the direction of flow are shown in Fig.6.6.
276
(PA)z (pA)z + d/dz(pA) dz
τw Sdz
Fig. 6.6: Forces acting on a fluid element in fully developed floe region
Rate of change of momentum in the direction of flow = 0 because the velocity u does not
vary with respect to z in the fully developed region.
For duct of uniform cross section A is constant. Therefore the above equation reduces to
dp/dz = − τw S /A ……………………………….(6.22)
dp μS
------ = -------- (du/dr)|wall ……………………….(6.23)
dz A
Eq.(6.23) is not practical for the determination of (dp/dz), because it requires the evaluation
of the velocity gradient at the wall. Hence for engineering applications a parameter called
“friction factor, f ” is defined as follows:
(dp/dz) Dh
f = − ------------------- ………………………….(6.24a)
½ (ρum2)
277
Substituting for (dp/dz) from Eq. (6.23) we have
(μS/A) (du/dr)|wall Dh
f = − ---------------------------- …………………..(6.24b)
½ (ρum2)
8μ
f = − ------- (du/dr)|wall …………………………(6.24c)
(ρum2)
Integrating the above equation over a total length L of the tube we have
p2 (½ ) (ρum2) f L
∫dp = − ------------------- ∫dz
p1 Dh 0
6.4.3. Thermal Boundary Layer for Flow Through Tubes: In the case of temperature
distribution in flow inside a tube, it is more difficult to visualize the development of thermal
boundary layer and the existence of thermally developed region. However under certain
heating or cooling conditions such as constant wall-heat flux or constant wall-temperature it
is possible to have thermally developed region.
Consider a laminar flow inside a circular tube subjected to uniform heat flux
at the wall. Let ‘r’ and ‘z’ be the radial and axial coordinates respectively and T(r,z) be the
local fluid temperature. A dimensionless temperature θ(r,z) is defined as
T(r,z) – Tw(z)
θ(r,z) = ------------------- ………………………………..(6.27a)
Tm(z) – Tw(z)
where Tw(r,z) = Tube wall-temperature and Tm(z) = Bulk mean temperature of the fluid.
The bulk mean temperature at any cross section ‘z’ is defined as follows:
278
∫ ρ(2πrdr) u(r,z) Cp T(r,z) ∫ rdr u(r,z)T(r,z)
Tm(z) = --------------------------------- = ---------------------- ………………..(6.27b)
∫ ρ(2πrdr) u(r,z) Cp ∫ rdr) u(r,z)
At the tube wall it is clear that θ(r,z) = 0 and attains some finite value at the centre of the
tube. Thus we can visualize the development of thermal boundary layer along the tube
surface as shown in Fig. 6.7.The thickness of the thermal boundary layer δ t continuously
grows along the tube surface until it fills the entire tube. The region from the tube inlet to the
hypothetical location where the thermal boundary layer thickness reaches the tube centre is
called the “thermal entry section”. In this region the shape of the dimensionless temperature
profile θ(r,z) changes both in axial and in radial directions. The region beyond the thermal
entry section is called as the “thermally developed region”, because in this region the
dimensionless temperature profile θ remains invariant with respect to z. That is in this region
θ = θ(r).It is difficult to explain qualitatively why θ should be independent of z even though
the temperature of the fluid T depends both on r and z. However it can be shown
mathematically that ,for both constant wall-heat flux and constant wall-temperature
conditions, θ depends only on r for large values of z.For constant wall-heat flux condition
the wall-temperature Tw(z) increases with z.
The variation of wall-temperature and the bulk fluid temperature as we proceed along the
length of the tube for constant wall-heat flux conditions is shown in Fig. 6.8. It can be shown
that for constant wall-heat flux condition the temperature difference ΔT between the tube
wall and the bulk fluid remains constant along the length of the tube in the thermally
developed region
The growth of the thermal boundary layer for constant wall-temperature conditions
is similar to that for constant wall-heat flux condition except that the wall temperature does
279
not vary with respect to z. Therefore the temperature profile T(r,z) becomes flatter and flatter
as shown in Fig. 6.9 as we proceed along the length of the tube and eventually the fluid
temperature becomes equal to the wall temperature. Since the wall-temperature remains
constant and the bulk fluid temperature varies along the length the temperature difference
between the tube wall and the bulk fluid varies along the length of the tube as shown in Fig.
5.10.
Tw(z)|z=L
0 L
Fig. 6.8: Variation of tube wall-temperature and bulk fluid temperature along the length of the tube
280
Tw Tw Tw
Tw
Tfi
z
Thermally developing region Thermally developed
Thermal entrance length Lth region
Fig.6.9: Growth of thermal boundary layer for flow through a tube with
constant wall-temperature
Tw
ΔTo
Tfo
ΔTi
Tfi
L
0
z
Fig. 6.10: Variation of bulk fluid temperature along the length of the tube
for tube with constant wall-temperature
6.4.4. Mean Temperature Difference, ΔTm: If Q is the total heat transfer rate between the
fluid and the tube surface , As is the area of contact between the fluid and the surface, h m is
the average heat transfer coefficient for the total length of the tube then we can write
Q = hm As ΔTm …………………………….(6.28)
281
Where ΔTm = mean temperature difference between the tube wall and the bulk fluid.An
expression for ΔTm can be obtained in terms of ΔT i and ΔTo where ΔTi is the temperature
difference between the tube-wall and the bulk fluid at inlet and ΔT o is the temperature
difference between the tube-wall and the bulk fluid at exit as shown below.
dQ
mm
dz
Consider a control volume consisting an elemental length dz of the tube at a distance z from
the inlet as shown in Fig.5.11. Let T m be the bulk temperature of the fluid entering the control
volume and Tm + dTm be the bulk temperature of the fluid leaving the control volume. Let dQ
be the heat transfer rate from the tube wall to the fluid over the length dz.
Energy balance for the control volume neglecting the changes in kinetic and potential
energies in comparison with change in enthalpy gives
where S is the perimeter of the tube.From the above two equations wer have
dTm hSdz
-------------------- = -------------- (6.29c)
[Tw(z) – Tm(z)] m Cp
For a tube with constant wall-heat flux condition, since the temperature difference between
the fluid and the tube surface remains constant along the length of the tube it follows that
ΔTm = [Tw(z)|z=0 − Tfi] = [Tw(z)|z=L − Tfo] …………………(6.29d)
For a tube with constant wall-temperature condition the mean temperature difference is given
by
282
6.4.5. General expression for heat transfer coefficient : Let the fluid be heated as it flows
through the tube. Then at any z the heat flux from the tube surface to the fluid is given by
Fourier’s law as
qw(z) = k (∂T / ∂r)|wall ……………………(6.30)
[Note that when the fluid is heated Tw> Tm so that (∂T / ∂r)|wall will be positive).
If hz is the heat transfer coefficient then
k (∂T / ∂r)|wall
hz = ------------------- ……………………(6.32)
[Tw(z) – Tm(z)]
[T(r,z) – Tw(z)]
Now θ(r,z) = --------------------
[Tm(z) – Tw(z)]
Since (dθ / dr)|wall is independent of z it follows that the heat transfer coefficient h z is
independent of z. This is true both for constant wall-temperature and constant wall-heat flux
conditions.
Example 6.5:- The velocity profile for hydrodynamically developed laminar flow inside a
circular tube of radius R is given by
283
where um is the average velocity of the fluid in the tube.Develop an expression for the friction
factor f and express it in terms of the Reynolds number Re d where Red is defined as Red =
(umD) / ν.
Solution:
u(r) = 2um[ 1 – (r/R)2]
Or f = 64 / Red
Example 6.6:- The velocity profile u(y) for hydro dynamically developed laminar
flow between two parallel plates a distance 2L apart is given by u(y) / um = (3 / 2)
[1– (y /L) 2]where um is the mean flow velocity and the coordinate axis y is as shown
in Fig. P6.6.
2L
o x
H
Fig. P 6.6 : Schematic for problem 5.6
284
y
dz
z τw1
p
z p+ (dp/dz)dz
2L
τw2
H
Solution:
Consider a fluid element of length dz at a distance z from the origin as shown in the
figure. Resultant force acting on the fluid element is given by
For fully developed flow there is no change in the momentum of the fluid in z-direction.
Hence by Newton’s second law F = 0.
It is given that,
u = (3/2) um [ 1 – (y/L)2 ]
= (3μum)/L
= − [(3μum)/ L2]
285
(a) The friction factor f is given by
− (dp/dz) dh [(3μum)/ L2] dh
f = --------------------- = ------------------------
(1/2) ρum2 (1/2) ρum2
4 x 2L
as ( dp/dz) = (1/2) ρ um2 f (1/dh), dh = hyd. Diameter = -------------- = 4L
2
12 x 2 24
= ------------------- = ----------------
(ρumL)/μ ReL
p2 H
Δp = p1 − p2 = − ∫dp = − ∫ −[(3μum) / L2]dz = 3 (H/L)(μum / L)
p1 0
3 (H/L) (ρum2)
= --------------------- = 3 (H/L) (ρum2) / ReL
(ρumL/μ)
Example 6.7:-The friction factor for hydro dynamically developed laminar flow
through a circular tube is given by
Water at a mean temperature of 60 0 C and a mean velocity of 10 cm/s flows inside a tube of
1 cm ID. Calculate the pressure drop for a length of 10 m of the tube and also the
corresponding pumping power required.
Solution:
Properties of water at 60 0 C are : ρ = 985 kg/m3 ; μ = 0.78 x 10 − 3 kg / (m – s);
Mean velocity of water = um = 0.1 m/s ; Di = 0.01 m ; L = 10 m.
286
= 152.68 N / m2.
. π x (0.01)2 x 0.1
Volume flow rate = V = (πDi2/4) um = ------------------------ = 7.85 x 10 − 6 m3/s
4
.
Pumping power = Δp V = 152.68 x 7.85 x 10 − 6 = 1198.5 x 10 − 6 J / s.
where 2L is the distance between the plates. Calculate the temperature difference between
the plate surface and the mean fluid temperature.
Solution:
(hm4L) / k = 8.235
287
vortices varies from a significant fraction of the tube radius at the central region of the tube
to microscopic sizes near the wall where the velocity approaches zero.
In turbulent flow, the fluid properties such as velocity, pressure, and temperature are
subject to fluctuations both with the location in the flow field and with time. Hence the
instantaneous values of these properties can be represented as a sum of a time averaged
mean part and a fluctuating part as follows:
_ _ _ _
u = u + u’; v = v + v’ ; T = T + T’ ; p = p + p’ …………………………………..6.35
_ _ _ _
where u,v,T and p are the instantaneous values ; u,v,T and p are the time-averaged values and
u’,v’,T’ and p’ are the fluctuations. If the temperature of a flowing fluid is recorded at a
location in the flow field and if the flow is turbulent at that location, then the plot of
temperature versus time will be as illustrated in Fig.6.11. In turbulent flow, the fluctuations
are considered superimposed on local average flow.The time averaged value of temperature,
for example, can be defined as
__ 1 t + Δt
T = ------ ∫ T dt …………………………6.36
Δt t
T
T’ = Temperature Fluctuation
__
T = Time –Averaged Temperature
Time,t
Fig: 6.11. Temperature fluctuations with time in turbulent flow
where Δt is a very small time interval which is large enough to record the turbulent
fluctuations but small for the temperature to be unaffected by the external disturbances on the
system.Similar averaging process can be defined for velocity and pressure (according to
Eq.5.36, the time average of fluctuating quantities are all equal to zero).The effects of these
fluctuations occurring in velocity and temperature in turbulent flow are to increase heat
transfer and resistance to fluid flow.
288
Consider turbulent flow along a flat plate as shown in Fig. 6.12.The y coordinate
is measured normal to the plate surface. Let u, and v be the instantaneous
u∞
y
u(x, y)
u(x, y)
x
xcr Turbulent Region
Turbulent Layer
Laminar Region Transition
Buffer Layer
Viscous sublayer
Fig. 6.12: Turbulent boundary layer for flow over a flat plate
values of the components of of velocities along x and y directions and T be the instantaneous
value of the temperature. Let u’,v’ and T’ be the corresponding fluctuations respectively.Let τ
be the shear stress in x direction at a location y from the plate surface.This shear stress can be
considered as to be composed of two parts:
_
viscous shear stress resulting from the mean flow velocity u and turbulent shear stress
resulting from the velocity fluctuations u’ and v’. Then we can write
Here the viscous shear stress term given by Eq. 6.37b is the laminar flow contribution.
The physical significance of the turbulent shear stress term given by Eq. 6.37c is not quite
apparent,but it can be derived by introducing the instantaneous velocities u and v in the x
momentum equation (See Ref.1) and by applying the rules of averaging to terms involving
the cross products of fluctuations.
Now if q is the heat flux in the y direction, then q can be considered as composed of
two parts: the diffusive or conductive heat flux due to gradient of mean temperature and the
turbulent heat flux resulting from the temperature fluctuations T’ and the velocity
fluctuations v’ in the y direction. With this consideration, we can write
289
q = qdiffusive + qturbulent ………………………….6.38a
_
Where qdiffusive = ─ k (ðT / ðy)………………………...6.38b
____
qturbulent = ρcp v’T’ ……………………………...6.38c.
Here the diffusive heat flux given by Eq. 6.38b is similar to that encountered in conduction
heat transfer, but the significance of turbulent heat flux given by Eq. 6.38c is not quite
apparent.It is derived by introducing the instantaneous velovities u and v and the
instantaneous temperature T in the energy equation (See Ref.1) for forced convection and
time averaging these equations.The relations for turbulent shear stress and turbulent heat flux
given by equations 6.37c and 6.38c are not useful for computational purposes unless they are
related to the mean quantities ū and T. A number of models heve been proposed to develop
such relations by semi-empirical hypothesis. For example the ‘mixing length model”
proposed by Prandtl and von Karman have been used extensively and successively in relating
turbulent shear stress and turbulent heat flux to the gradients of average velocity and average
temperature.
6.5.2. Prandtl’s Mixing Length Model: The basic idea behind this model is analogus to the
concept of “mean free path” for the motion of molecules in kinetic theory of gases. The
main difference is that in turbulent flow macroscopic lumps are envisioned.That is, in
turbulent flow in the x-direction along the surface, Prandtl postulated that fluid particles at a
distance y from the wall surface coalesce into macroscopic lumps and then travel, on the
average a distance l in the direction normal to the main flow while retaining their x-direction
momentum before they are dispersed.Thus , if the slow moving lumps enter the fast-moving
layer, they act as drag on the fast moving layer, and the momentum is transferred between
layers as a result of transeverse mixing. It is to be noted that l is an unknown quantity, and in
reality there is no such clear defined distance. Though the concept lacks generality, it has
been found useful in most engineering applications.According to Prandtl’s theory, the
velocity fluctuations can be related to
(ðū / ðy) by the following equations.
where u’ and v’ are of opposite sign and l 1 and l2 are the “mixing lengths for momentum
transport”.Similarly the temperature fluctuation is related to the gradient of mean
tempaerature as _
T’ = l3 (ðT / ðy) ……………………………..6.39c
Where l3 is the “mixing length for energy transport”. Now the turbulent shear stress can be
written as:
___
τturbulent = ─ ρ u’v’ = ρl1l2 (ðū / ðy)2 = ρlm2 (ðū / ðy)2
290
or τturbulent = ρεm (ðū / ðy) ……………………………………..6.40a
εm and εh are called as “eddy diffusivity for momentum transfer” and “eddy difussivity for
heat transfer” respectively. Therefore for turbulent flow the total shear stress and the total
heat heat flux can be written as follows:
Where υ is the kinematic viscosity and α is the thermal diffusivity of the fluid. Equations
6.42a and 6.42b clearly demonstrates the effects of turbulent flow in enhancing the drag as
well as the heat transfer. Depending on the level of turbulence in the flow field, the turbulent
transport properties εm and εh may be an order of magnitude larger than the diffusive
properties υ and α. In general ε m and εh are not the same. Their ratio is defined as “turbulent
Prandtl number, Prt”. Thus
Prt = εm / εh ………………………….6.43.
Pr = υ / α ……………………………..6.44.
The physical significance of Pr and Prt should be distinguished. The Prandtl number is a
physical property of the fluid. It varies from a value much less than 1 for liquid metals to
values much greater than 1 for liquds and ois. It is about 1 for gases. The turbulent Prandtl
number is a property of the flow field and various models have been proposed for its
determination. The simplest model is due to Reynolds who assumed Pr t = 1, which implies
that heat and momentum transfer in turbulent flow takes place exactly by the same process.
The numerical value of Prt may vary between 1 and 2.
291
One of the velocity distribution which has been used widely is based on the concept
of dividing the turbulent flow field into three distinct layers as illustrated in Fig.5.10.A very
thin layer adjacent ot the wall where laminar or viscous shear stress dominate is called the
“viscous sublayer”. Adjacent to this layer is the “buffer layer”, where the viscous and
turbulent shear stresses are of the same order of magnitude and the third layer next to the
buffer layer is called “turbulent layer”, where the turbulent shear stress dominates.For the
steady, turbulent flow of an incompressible, constant property fluid over a smooth surface the
velocity distribution is expressed using the following dimensionless parameters:
ū
+
u = ------------ = Dimensionless velocity
√(τw / ρ)
y
+
y = ------ √(τw / ρ) = Dimensionless distance
υ
where ρ is the density, τw is the shear stress at the wall, υ is the kinematic viscosity and ū
component of mean velocity parallelto the wall surface.
Experiments have shown that for region where y + < 5, turbulent shear stress
contribution is negligible and the region is viscous sublayer.Therefore the shear stress takes
the form τw = μ(ðū / ðy). Integration of this expression for constant τ w with ū = 0 at y = 0
gives the following distribution for viscous sublayer:
The buffer layer is considered to be in the region between y + = 5 and y + = 30. In this region
the velocity distribution is assumed to follow the logarithmic law of the form
u + = A ln y + + B. The constants A and B are determined from requirement that at
y + = 5 ,velocity u + be equal to that of viscous sublayer and be equal to that of turbulent layer
at y + = 30.Hence for the buffer region the velocity distribution will be of the form
The region y + > 30 is considered to be the “turbulent layer” where the laminar
shear stress is negligible in comparison to the turbulent shear stress. By using the mixing
length concept and assuming that the mixing length varies linearly with the distance from the
wall in the form l = κy, it can be shown that the velocity distribution in the turbulent layer has
a logarithmic profile in the form
u + = (1/κ ) ln y + + C …………………………….6.47
where κ is called the universal constant. Experiments have shown that κ = 0.4 and the
constant C has been determined by the correlation of Eq. 6.47 with the measured velocity
profile. For turbulent flow inside a smooth pipe, C = 5.5. Then the velocity distribution in the
turbulent layer is given as
292
u + = 2.5 ln y + + 5.5 for turbulent layer y + > 30 ……..6.48
Though the above velocity distribution [Eq.6.45, Eq. 6.46 and Eq. 6.48] by dividing the
turbulent flow field into three distinct regions appears to be in reasonably good agreement
with the experimental data, the transition from a viscous to turbulent flow regime takes place
gradually.Therefore , he representation of of velocity distribution by three different curves
having discontinuous slopes at locations where they join is not realistis.A more serious
inconsistency of the logarithmic velocity distribution given by Eq.5.48 is that it does not give
zero velocity gradient at the tube centre. For this reason the average velocity for flow inside a
pipe as determined using the above equations overestimates the velocity.But these equations
have been used extensively in literature to study the relation between momentum and heat
transfer in turbulent flow.
λ
λ + = ---- √(τw / ρ) …………………………….6.49
υ
+
If λ < 5, the roughness has no effect on the friction factor.With this consideration three
distinct situations are envisioned for the effects of roughness:
For the hydrodynamically smooth case, the heights of roughness are so small that all
protrusions are covered by the viscous sublayer and hence roughness has no effect on friction
factor. For the transitional case, the protrusions are partly outside the viscous sublayer and
cause some additional rsistance to flow. For the fully rough case, the heights of protrusions
are so large that all protrusions penetrate the viscous sublayer; hence the viscous sublayer no
longer exists and protrusions influence the turbulent mixing.
For the fully turbulent region (y + > 30),the logarithmic velocity profile guven by Eq.
6.47 is applicable if y + is replaced by y / λ. In that acse the velocity distribution is given by
293
u + = (1/κ ) ln (y / λ )+ C ……………………….6.51
where κ = 0.4 as in the case of smooth wall but the constant C is different. A correlation of
this relation with Nikuradse’s experiments has shown that C = 8.5. Then Eq. 5.51 takes the
form
Eq. 6.52. is called logarithmic velocity distribution for turbulent flow in rough pipes in the
fully rough region.
A significant difference between laminar and turbulent flow lies in the fact that, in
turbulent flow the velocity profile is affected by the surface roughness, where as in laminar
flow, the velocity profile is not affected by the surface roughness. Since the velocity
distribution affects the shear stress at the wall and hence the friction factor, it is expected that
in turbulent flow the friction factor will depend not only on the Reynplds number but also on
the relative roughness, where as in laminar flow the friction factor is a function of Reynolds
number only.
derive the logarithmic velocity distribution, u + = (1/κ ) ln y + + C for the fully turbulent
region.
ū y
Where u + = ------------ and y + = ------ √(τw / ρ)
√(τw / ρ) ν
Example 6.12. Show that the velocity distribution in the viscous sublayer is given by
u+ = y +
with u+ and y+ are as defined in example 5.11.
294
Chapter 7
Heat transfer and pressure drop calculations in flow through ducts are required in many
industrial applications. Laminar flow through pipe is encountered generally in compact heat
exchangers, cryogenic systems, heating or cooling of heavy fluids like oils.For such flows
analytic expressions are available for the prediction of friction factor and heat transfer
coefficient.But most of the available correlations to determine heat transfer and friction factor
for turbulent flows are either empirical or semi-empirical in nature.
7.2.1. Couette Flow: Consider a special case of parallel flow of an incompressible fluid
between two parallel plates where one plate is stationary and the other plate is
moving with an uniform velocity U.A distance L separates the two plates (refer Fig.
7.1). The stationary plate is maintained at temperature To and the moving plate at
temperature TL. This type of flow is referred to as COUETTE flow and occurs, for
example, in journal bearing. The continuity, momentum and energy equations for
such a flow are given as follows:
For parallel flow along the plate the continuity equation (6.54b) reduces to
du / dx = 0 (7.1)
The momentum equation (6.59a) reduces to
d2u / dy 2 = 0 (7.2)
295
y
TL
y
u(y) = U
To, u(y) = 0
Fig. 7.1: Schematic for Couette Flow
at y = 0, u(y) = 0 (7.4)
Substituting these values of C1 and C2 in equation (2) we get the velocity distribution as
From equation (7.7) (∂u / ∂y) = U / L substituting this in the energy equation (7.3)
296
at y = 0 T = T0 (7.9)
Substituting the expressions for C1 and C2 in Equation (7.8 ) we get the temperature as
Or T(y) – T0 =
y
L [ ( T L −T 0 ) +
2k ( )]
μ U2
1−
y
L
(7.11)
T ( y )−T 0
T L −T 0
=
y
L[1+
1 μ U2
2 k ( T L−T 0 )
1−
y
L ( )]
(7.12)
μ U2
Now the term can be written as follows:
k ( T L −T 0 )
μ U2 μCp μU 2
= = Pr * E where Pr is the Prndtl
k ( T L −T 0 ) k C p ( T L −T 0 )
number and E is
T ( y )−T 0
T L −T 0
=
y
L [1
1+ Pr E 1−
2
y
L ( )]
(7.13)
T ( y )−T 0 y
Fig. 7.2 shows the plot of as a function of for different values of the
T L −T 0 L
parameter
Pr*E.The case when Pr*E = 0 corresponds to no flow condition. Hence heat transfer from the
hot plate to cold plate is by conduction in y direction. Therefore the temperature distribution
is linear.The physical significance for other values of Pr*E can be understood better if heat
transfer at the wall (either at the upper wall or at the lower wall) is considered.The heat flux
at the upper wall is given by
297
q(y)| y = L = − k (dT /dy)| y = L
T L−T 0
From Eq.(7.12) we have (dT /dy)| y = L =
L (1− 12 Pr∗E )
T L−T 0
Therefore q(y)| y = L = − k
L (1− 12 Pr∗E )
(7.14)
The above result shows that the direction of heat flow at the upper wall for the case T L >
T0 depends on the value of the parameter Pr*E.
1.0
0.8
0.6
0.4
0.2
For Pr * E > 2, the right side of Eq.(7.14) is positive. Therfore the heat flows in the
positive y direction, i.e., from the fluid into the wall though the upper wall is at higher
temperature than the lower wall.
298
For Pr * E < 2, the right side of Eq. (7.14) is negative. Hence the heat flows in the
negative y direction,i.e., from the upper wall to the fluid.
For Pr * E = 2, the right hand side of Eq.(7.14) reduces to zero. Hence there is no heat
flow at the upper wall.
When TL = T0, both plates are at the same temperature, Therfore from Eq. (7.11) the
temperature distribution reduces to
(7.15)
It can be seen from the aboe equation that the maximum temperature occurs at y = L/2
and the maximum temperature will be
μ U2
T max – T0 =
8k
(7.16)
Combining Eq. (7.15) and Eq. (7.16) the the temperature distribution can be expressed in
terms Tmax as
T ( y )−T 0 y y
T L −T 0
= 4
L
1−( )
L
(7.17)
Assumptions: (i)The oil properties are constant; (ii) the flow of oil in the space between the
journal and its bearing is considered as flow between two parallel plates; i.e. the flow is
coquette flow.
Solution: For Couette flow when T0 = TL, the maximum temperature rise is given by Eq.
(7.16): i.e.
2 2
μU 0.25∗6
ΔTmax = Tmax – T0 = = = 9 0C
8k 8∗0.125
299
Example 7.2.A lubricating oil of viscosity μ and thermal conductivity k fills the clearance L
between two rotating cylinders, which can be regarded as two parallel plates in motion for
the purpose of analysis. The velocity of the inner cylinder is U 1 and that of the outer cylinder
is U2.Find the relation for the velocity distribution in the oil layer for (a) the inner and outer
cylinders are rotating in the same direction but U 2 > U1 and (b) the inner and outer cylinders
are rotatibg in the opposite directions. Also find an expression for the shear stress in the fluid
resulting from the rotation.
Known: Viscosity of the oil = μ; Thermal conductivity of oil = k ; Clearnace between the two
rotating cylinders = L;velocity of lower plate=U1; Velocity of upper plate = U2
Find: (a) u(y) when both U1 and U2 are in same direction with U 2 > U1; (b) u(y) and shear
stress when U2 is in opposite direction with respect to U2.
Assumptions: (i) Fluid properties are constant; (ii) The flow between the two rotating
cylinders is considered as flow between two parallel plates (Couette flow)
d2u / dy 2 = 0 (1)
y
y
u(y) = U2
L
u(y) = U1
C2 = U 1 (3)
300
U2 = C1L + U1
( U 2−U 1 )
Or C 1 =
L
(3)
Substituting the expressions for C1 and C2 in eq. (2) we get the velocity distribution as
y
u(y) = ( U 2−U 1) + U1
L
u ( y ) −U 1 y
or =
U 2−U 1 L
(4)
( U 2−U 1 )
Shear stress is given by τ=μ ( dudy ) =μ
L
Schematic for Case (b) :- U2 = − U1 :See Fig.E7.2b
y
y
u(y) = U2
L
u(y) = U1
The general solution Eq.(7.2) holds good for this case also.
− U2 = C1L + U1
( U 1+ U 2 )
Or C1 = −
L
301
Hence the velocity distribution is given by
y
u(y) = − ( U 1 +U 2 ) + U1
L
U 1−u( y ) y
=
U 1 +U 2 L
(5)
( U 1+ U 2 )
Shear stress is given by τ = μ ( dudy ) = − μ
L
(6)
Example 7.3.Consider Couette flow with heat transfer for which the lower plate moves with
a velocity of U = 15 m/s and is perfectly insulated (see Fig. E7.3). The upper plate is
stationary and is made of material with k up = 1.5 W/(m-K) and thickness L up = 3 mm. Its
outer surface is maintained at Tup = 40 0 C. The plates are separated by a distance of L 0 = 5
mm which is filled with an engine oil of viscosity μ = 0.8N-s/m 2 and thermal conductivity k0
=0.145W/(m- K).
(a) On T(y) ─ y coordinates, sketch the temperature distribution in the oil film and in the
moving plate.
(b) Obtain an expression for the temperature at the lower surface of the film T 0 in terms of
the plate speed U, the stationary plate parameters T up,kup,Lup and the oil parameters μ, k0, L0.
Calculate this temperature for the prescribed conditions.
302
y Tup
Stationary Plate
Lup
oil L0
Moving plate
Fig. E7.3: Schematic for example 7.3
Solution:
d2u /dy2 = 0.
Condition (i) in equation (1) gives C2 = U and condition (ii) in equation (1) gives
C1 = − U / L0.
Substituting the expressions for C1 and C2 in equation (1) we get the velocity distribution as
303
Therefore equation (3) reduces to
k0(d2T / dy2) = −μ (U / L0)2
At y = 0 the surface is insulated i.e. (dT/dy) = 0. Substituting this condition in equation (4)
we have
C1= 0.
At y = L0, the condition is
At lower surface y = 0
304
7.1. Determine the friction factor, the pressure drop and pumping power for fully
developed laminar flow of water at 21 0 C [μ = 9.8 kg/(m-s) ; ρ = 997.4
kg/m3]through a 2.5 cm diameter, 100 m long tube for a mass flow rate of 0.015
kg/s. What are the mean and maximum velocities of flow?
7.2. Determine the friction factor, the pressure drop and pumping power required for
the flow of water at 0.5 kg/s and 40 0 C through a tube of square cross section of
2 cm x 2 cm and 12 m long. What would be the corresponding values if the pipe
is of equilateral-triangular cross section of side 2 cm and length 5 m ?
7.3. Water at 30 0 C with a mass flow rate of 2 kg/s enters a 2.5 cm-ID tube whose
wall is maintained at a uniform temperature of 90 0 C. Calculate the length of the
tube required to heat the water to 70 0 C.
7.4. Water at 20 0 C with a mass flow rate of 5 kg/s enters a circular tube of 5 cm-ID
and 10 m long. If the tube surface is maintained at 80 0 C, determine the exit
temperature of water.
7.5. Air at 27 0 C with a flow rate of 0.01 kg/s enters a rectangular tube 0.6 cm x 1.0
cm in cross section and 2 m long. The duct wall is subjected to a uniform heat
flux of 5 kW/m2. Determine the outlet temperature of air and the duct surface
temperature at the exit assuming that the flow is hydro dynamically and
thermally developed.
7.6. Three kg/min of liquid sodium is heated from a bulk mean temperature of 400 0 C
to 500 0 C, as it flows through a stainless steel tube of 5 cm-ID and 2 mm thick.
The sodium is heated by a constant wall-heat flux, which maintains the tube-wall
temperature at 30 0 C above the bulk mean temperature of sodium all along the
length of the tube. Calculate the length of the tube required. Assume the
following properties for liquid sodium.
ρ = 846.7 kg/m3 ; k = 68.34 W/(m-K) ; Cp = 1.274 kJ/(kg-K) ; Pr = 0.00468 ;
ν =0.2937 x 10 ─ 6 m2 /s.
7.7. Consider hydro dynamically and thermally developed turbulent flow of water with
a mass flow rate of M kg/s inside a circular tube of inside diameter ‘D’. The
Dittus-Boelter equation can be used to determine the heat transfer coefficient. If
the tube’s inside diameter is changed from D to D/2 while the mass flow remains
same, determine the resulting change in the heat transfer coefficient.
7.8. Mercury at a temperature of 100 0 C and with a velocity of 1 m/s enters a 1.25
cm ID tube, which is maintained at a uniform temperature of 250 0 C. Determine
the length of the tube required to heat mercury to a temperature of 200 0C.
305
B. Hydrodynamic and thermal entry lengths:
7.9. Determine the hydro dynamic entry lengths for flow at 60 0 C and at a rate of
0.015 kg/s of water, ethylene glycol and engine oil through a circular tube of 2.5
cm ID.
7.10. Determine the hydro dynamic entry length, thermal entry length and the heat
transfer coefficient for fully developed flow for engine oil at 60 0 C flowing at a
rate of 0.01 kg/s through a square duct 1 cm x 1cm cross section and subjected
to a uniform wall-temperature. Assume the following physical properties for the
engine oil:
ρ = 864 kg/m3 ; Cp = 2047 J/(kg-K) ; k = 0.14 W/(m-K) ; μ = 0.0725 kg/(m-s) ;
Pr = 1050
7.11. Atmospheric air at 25 0 C flows over both the surfaces of a flat plate 1 m long
with a velocity of 5 m/s. The plate is maintained at a uniform temperature of
750C.
(a) Determine the velocity boundary layer thickness, the surface shear stress and
the heat flux at the trailing edge of the plate.
(b) Determine the drag force on the plate and the total heat transfer from the plate
to air.
7.12. Air at 30 0 C flows with a velocity of 10 m/s along a flat plate 4 m long. The plate
is maintained at a uniform temperature of 130 0 C. Assuming a critical Reynolds
number of 2 x 10 5 and width of plate to be 1 m determine (a) the heat flux at
the trailing edge of the plate, (b) the heat transfer from the laminar portion of
the plate, (c) the total heat transfer from the plate and (d) the heat transfer from
the turbulent portion of the plate.
7.13. A highly conducting thin wall 2 m long separates the hot and cold air streams
flowing on both sides parallel to the plate surface. The hot stream at 250 0 C is
flowing with a velocity of 50 m/s while the cold stream at 50 0 C is flowing with a
velocity of 15 m/s. Calculate (a) the average heat transfer coefficients for both
the air streams and the heat transfer between the two streams per metre width
of the plate and (b) the local heat flux at the mid point of the plate. Assume that
the wall is at the arithmetic mean of the temperature of the two streams for the
purposes of calculating the physical properties of the two streams and the critical
Reynolds number to be 2 x 10 5.
306
150 0 C
T∞ = 25 0 C
u∞ = 30 m/s
10 mm
700 mm
50 mm
(a) Find the required power generation in W/m 2 in a module positioned at a distance
of 700 mm from the leading edge of the plate.
(b) Find the maximum temperature in the heat generating module.
7.16. A long cylindrical heating element [k = 240 W/(m-K), ρ = 2700 kg/m 3 and Cp =
900 kJ/(kg-K)] of diameter 10 mm is installed in a duct in which air moves in
cross flow over the heating element at a temperature of 27 0 C with a velocity of
10 m/s.
(a) Estimate the steady state surface temperature of the heater when electrical
energy is being generated at a rate of 1000 W per metre length of the cylinder.
(b) If the heater is activated from an initial temperature of 27 0 C, estimate the time
required for the surface temperature to come to within 10 0 C of its steady state
value.
V (m/s) : 1 2 4 8 12
307
(a) Determine the convection coefficient for each of the above test conditions.
Display your results graphically.
(b) For the corresponding Reynolds number range, determine the suitable constants
C and m for use with an empirical correlation of the form Num = C Redm Pr1/3.
7.18. A thermocouple is inserted into a hot air duct to measure the air temperature.
The thermocouple (T1) is soldered to the tip of a steel thermocouple well of
length 15 cm and inner and outer diameters of 5 mm and 10 mm respectively. A
second thermocouple (T2) is used to measure the duct wall temperature (see Fig.
P 6.18).
T2
Air
at T∞, u∞ L Steel well
Di
Do
Consider the conditions for which the air velocity in the duct u ∞ = 3 m/s and the
two thermocouples register temperatures of T 1 = 450 K and T2 = 375 K.
Neglecting radiation determine the air temperature T ∞. Assume that for steel k =
35 W/(m-K), and for air ρ = 0.774 kg / m3, μ = 251 x 10 ─ 7 N-s / m2, k = 0.0373
W/(m-K), and Pr = 0.686
7.19 Air at atmospheric pressure and 30 0 C flows over a bank of tubes consisting 1
cm OD tubes, 10 rows deep. The velocity of air before it enters the bundle is 1
m/s.
(a) Determine the friction factor and the pressure drop and (b) the average heat
transfer coefficient for the following cases:
(i) Tubes are in equilateral-triangular arrangement with ST / D = SD / D = 1.25
(ii) Tubes are in square arrangement with ST / D = SL / D = 1.25
7.20. Hot flue gases at 375 0 C flow across a tube bank consisting of 1.25 cm OD tubes,
which are maintained at a uniform surface temperature of 30 0 C by flowing water
through the tubes. The tube bundle is 10 rows deep in the direction of flow and
contains 40 tubes in each row. The tubes are 1 m long and have an in-line
arrangement with SL / D = ST / D = 2. he velocity of the flow gases entering the
tube matrix is 7 m/s. Determine the average heat transfer coefficient and the
total heat transfer rate. Assume that thermo-physical properties of the flow gases
to be same as that of air at any temperature.
308
A.Hydro-dynamically and Thermally developed flow through ducts
7.1. Solution:-
Water di
At 200C
Reynolds Number = Red = uav dh / ν, where uav = average velocity of the fluid in the pipe
and dh = hydraulic diameter for the pipe.
Now m = ρ(πdi2/4)uav.
4 x 0.015
Or uav = (4m) / ρ(πdi2) = --------------------------- = 0.0305 m / s
1000 x π x 0.0252
309
The velocity of the fluid at any radius for fully developed laminar flow through a circular
tube is given by
7.2. Solution
0.5
uav = m / (ρab) = -------------------- = 1.26 m / s.
994.59 x 0.022
1.26 x 0.02
Reynolds number Red = uav dh / ν = ------------------- = 38299
0.658 x 10 − 6
For fully developed turbulent flow through a pipe of square cross section the friction factor f
is given by (Moody chart, smooth pipe)
f = 0.02175.
Pressure drop = Δp = (1/2)f (L/dh) ρuav2 = 0.5 x 0.02175 x (12 / 0.02) x 994.59 x 1.262
= 10303 N / m2
310
For a tube of equilateral triangular cross section, dh = 4 {√3 x a2/4} / 3a, where a is the side of
the triangle.
0.5
Average velocity = uav = ----------------------------- = 2.9 m/s
994.59 x (√3/4) x 0.02 2
2.9 x 0.0115
Reynolds number = Red = --------------------- = 50684
0.658 x 10 − 6
Since Red > 2300, flow is turbulent. Hence from Moody chart we have
f = 0.02
7.3. Solution:
Tw
Tfi di Tfo
L
Data :- Tfi = 30 0C ; Tfo = 70 0C ; Tw = 90 0C ; m = 2 kg / s ; di = 2.5 cm = 0.025 m.
311
Since nothing has been specified in the problem regarding the type of flow, it is assumed that
the flow is hydro dynamically and thermally developed.
4x2
2
Average velocity = uav = m / (ρx πdi /4) = ---------------------------- = 4.11 m/s.
990 x π x 0.025 2
Since Red >2300, flow is turbulent. For fully developed turbulent flow the Nusselt number is
given by
= 628
To find the length of the tube L, we write the energy balance equation for the entire length of
the tube as
Heat supplied to fluid from the tube wall = Increase of energy of the fluid
Where ΔTm = logarithmic mean temperature difference = [ΔTi – ΔTo] / ln (ΔTi / ΔTo)
7.4. Solution:
312
Tw
To find Tfo.
Since Tfois not known we cannot determine the bulk fluid mean temperature to know the
properties of the fluid. Hence this problem has to be solved by trial and error method as
shown below
Trial No. 1:- Assume suitable value for Tfo noting that Tfo < Tw.
4x5
2
Average velocity of water = uav = 4m /( πd ρ) = --------------------------
i
π x (0.05)2 x 994.59
= 2.56 m/s.
313
= 704.5
704.5 x 0.628
Heat transfer coefficient = h = Nud k / dh = ---------------------- = 8848.5 W /(m2 – K).
0.05
Heat balance equation for the total length of the tube can be written as
[80 – 20]
= ----------------------------------------------------
exp{(8848.5 x π x 0.05 x 10) / (5 x 4178.4)}
= 30.85 0 C.
Therefore Tfo = 80 – 30.85 = 49.15 0 C. This temperature is quite away from the
assumed value and hence one more iteration is required.
[80 – 20]
Therefore Tfo = 80 − ---------------------------------------------------------
exp {(8328.6 x π x 0.05 x 10) / (5 x 4179.3)}
314
7.5. Solution:
qw
a
Tfi Tfo b
L
Data :- Fluid is air ; Tfi = 27 0 C ; m = 0.01 kg/s; a = 0.01 m; b = 0.006 m; L = 2 m;
qw = 5000 W / m2.
Energy balance equation for total length of the tube can be written as
Properties of air at 42.9 0 C are : ρ = 1.12 kg / m3; 1006.8 J/kg-K; ν = 17.30 x 10 − 6 m2/s;
0.01
Average velocity = uav = m / [ρab] = ------------------------- = 149 m /s.
1.12 x 0.01 x 0.006
149 x 0.0075
315
Reynolds number = Red = uavdh / ν = --------------------- = 64595
17.3 x 10 − 6
Since Red >2300 flow is turbulent. Assuming that the flow is hydrodynamically and
thermally developed we have
Nud = 0.023 Red 0.8Pr n, with n = 0.4 as air is being heated.
140.9 x 0.02745
Heat transfer coefficient = h = Nud k / dh = ---------------------- = 515.7 W/(m2 – K).
0.0075
7.6.Solution:
Data: Fluid is liquid sodium; m = 3/60 =0.05 kg/s; Tfi = 400 0 C; Tfo = 500 0C;di = 0.05 m
4 x 0.05
Average velocity = uav = 4m / ( ρπdi2) = ------------------------ = 0.03 m/s.
846.7 x π x 0.05 2
0.03 x 0.05
Reynolds number = Red = uavdh / ν = ----------------------- = 5107
0.2937 x 10 − 6
Since Red > 2300, flow is turbulent.Assuming the flow to be hydrodynamically and thermally
developed and since Pr << 1 (Liquid metal), the Nusselt number for constant wall heat flux
condition is given by
5.075 x 68.34
Heat transfer coefficient = h = Nud k / dh = ------------------------ = 6936.5 W/(m2 – K).
316
0.05
Energy balance equation for the total length of the tube can be written as
The Dittus-Boetler correlation for hydrodynamically and thermally developed flow is given
by
Nud = hdh/k = 0.023 Red0.8 Pr n ………………..(1)
4M
For a circular tube of diameter D, Red = uavD / ν = --------------
ρπDν
Similarly when the diameter of the tube is reduced to D/2, for the same mass flow rate the
heat transfer coefficient is given by
317
Tw
di
Tfi Tfo
To find L
Bulk mean temperature of mercury = Tm = ½(Tfi + Tfo) = 0.5 x (100 + 200) = 150 0C.
1.0 x 0.0125
Reynolds number = Red = uavdh / ν = ----------------------- = 1.445 x 10 5
0.0865 x 10 − 6
17.4 x 11.425
Heat transfer coefficient, h = Nudk/dh = -------------------- = 15904.5 W/(m2-K)
0.0125
318
= [150 – 50] / ln(150 / 50) = 91 0 C
= 1.952 kg/s
Energy balance equation for the total length of the pipe is given by
= 0.4715 m
6.9. Solution:
Data:- Tfi = 60 0C ; m = 0.015 kg / s; di = dh = 0.025 m.
4 x 0.015
Average velocity = uav = 4m / ρπdi2 = -------------------------- = 0.032 m/s.
985.46 x π x 0.025 2
0.032 x 0.025
Reynolds number = Red = uavdh / ν = --------------------- = 1728
0.478 x 10 − 6
Since Red < 2300, flow is laminar. Hence the hydrodynamic entrance length Lh for a circular
pipe is given by
Lh / dh
-------- = 0.056
Red
= 2.419 m
(ii)Fluid is ethylene glycol: ρ = 1087.66 kg/m3 ; ν = 4.75 x 10 − 6 m2/s.
4 x 0.015
Average velocity = uav = 4m / ρπdi2 = -------------------------- = 0.0281 m/s.
1087.66 x π x 0.025 2
319
0.0281 x 0.025
Reynolds number = Red = ---------------------------- = 147.9
4.75 x 10 − 6
= 0.21 m.
(iii) Fluid is engine oil; ρ = 864.04 kg/m3 ; ν = 0.839 x 10 − 4 m2/s.
4 x 0.015
Average velocity = uav = 4m / ρπdi2 = -------------------------- = 0.0354 m/s.
864.04 x π x 0.025 2
0.0354 x 0.025
Reynolds number = Red = ---------------------------- = 10.55
0.839 x 10 − 4
0.01
2
Average velocity = uav = m / (ρa ) = --------------------- = 0.1157 m/s.
864 x (0.01) 2
320
(Lt / dh)
---------- = 0.041
Pe
(iii)For fully developed laminar flow through a square tube Nusselt number is given by
Nu T = hdh / k = 2.976
Consider the flow of water at a rate of 0.01 kg/s through an equilateral triangular duct of
sides 2 cm and whose walls are kept at a uniform temperature of 100 0C. Assuming the flow
to be hydrodynamically and thermally developed, determine the length of the tube required to
heat the water from 200C to 70 0C.
Solution:
Tw = 100 0C
Water
Tfo=700C
Tfi = 20 0C
m = 0.01 kg/s
L= ?
a = 0.02 m
Bulk mean temperature of water = ½(Tw + T∞) = 0.5x (20 + 70) = 45 0C.
For an equilateral triangular tube, area of flow = A = (√3 / 4)a 2 = (√3 / 4) x 0.02 2
= 1.732 x 10 − 4 m 2
321
4 [(√3 / 4)a 2] a 0.02
Hydraulic diameter = dh = --------------------- = ------------- = ------------ = 0.01155 m
3a √3 √3
m 0.01
Average velocity of water = uav = ----------------- = -------------------------------
ρA 992.3075 x 1.732 x 10 − 4
=0.0582 m/s
0.0582 x 0.01155
Reynolds number = Red = uavdh / ν = -------------------------- = 1124
0.598 x 10 − 6
Since Red <2300 flow is laminar.For thermally developed laminar flow with constant wall-
temperature the Nusselt number is given by
2.47 x 0.63375
Therefore hav = ------------------ = 135.53 W/(m2-K)
0.01155
Mean temperature difference between the surface and the bulk fluid is given by
Rate of heat transfer to water = Q = mcp(Tfo – Tfi) = 0.01 x 4179.9 x (70 – 20)
= 2090 W
Q 2090
Or L = -------------------- = -------------------------------- = 5.13 m
hav 3a ΔTm 135.53 x 3 x 0.02 x 50.1
6.11. Solution:
322
u∞, T∞
Tw
x
L
k = 0.028 W/(m-K).
= 2.775 x 10 5
Assuming the critical Reynolds number to be 5 x 10 5, the flow is laminar at the trailing edge.
Therefore from heat transfer data hand book we have
= 0.0088 m.
= 1.26 x 10 − 3
323
Nux|x=L = 0.332 ReL 0.5 Pr 0.333
= 155.5
= 4.354 W/(m2-K)
6.12. Solution:-
u∞ = 10 m/s (air)
T∞ = 30 0 C Tw = 130 0 C
x
W=1 m
L= 4 m
To find :- (a) qw(x)|x=L; (b) Qlaminar; (c) Qtotal assuming Recr = 2 x 10 5 (d) Qturbulent
324
10 x 4
ReL = u∞L / ν = -------------------------- = 1.86 x 10 6
21.48 x 10 − 6
For turbulent flow of air over a flat plate the local Nusselt number is given by
2714 x 0.03047
Therefore [hx|x=L] L / k = 2803 or hx|x=L = ---------------------- = 20.67 W /(m2-K)
4
Recr ν 2 x 10 5 x 21.48 x 10 − 6
Or xcr = --------- = ------------------------------- = 0.4296 m
u∞ 10
Heat transfer rate from laminar portion = Qlaminar = [hav]laminar [x cr W] (Tw – T∞)
= 790.5.0 W
325
(c)Average Nusselt number for the entire length of the plate is given by
3085 x 0.03047
Hence hav = Nuav k / L = ----------------------- = 23.5 W/(m2 – K)
4
6.13. Solution:
Hot air at T∞1 =250 0 C, u∞1 = 50 m/s
Tw = ½( T∞1+ T∞2)
= 150 0 C
L= 2 m
To find (i) hav for the hot surface ; (ii) hav for the cold surface ; (iii) qw(x)|x=L/2
(i) Mean film temperature for the hot fluid = ½(Tw + T∞1) = 0.5 x (150 + 250) = 200 0C.
50 x 2
Or ReL = ------------------- = 2.869 x 10 6
34.85 x 10 − 6
Since ReL > Recr, flow is partly laminar and partly turbulent. Therefore the average Nusselt
number is given by
326
Where A = 0.037 Recr0.8 – 0.664 Recr 0.5.
4465 x 0.0391
Hence for the hot surface [hav] hot = Nuav k / L = ---------------------- = 87.29 W/(m2-K)
2
(ii) Mean film temperature for the cold surface = ½(150 + 50) = 100 0C.
Properties of air at the mean film temperature are: ν = 23.33 x 10 − 6 m2/s; Pr = 0.693
k = 0.03184 W/(m-K)
15 x 2
ReL = u∞2L / ν = -------------------------- = 1.286 x 10 6
23.33 x 10 − 6
2219 x 0.03184
Hence for the cold surface [hav]cold = -------------------- = 35.33 W/(m2-K).
2
(iii) The rate of heat transfer from the hot air stream to cold air stream is given by
1
Now Rc1 = 1 / [hav]hot A = ------------------- = 0.00573 m2 – K / W.
87.29 x (2 x 1)
1
Similarly Rc2 = 1 / [hav]cold A = -------------------- = 0.01415 m2 – K / W.
35.33 x (2 x 1)
327
(iv) At mid point of the plate x = L / 2.
(2.33 x 250) + 50
Hence Tw = -------------------------- = 189.94 0C
1 + 2.33
Also check for qw(x)|x=L/2 ; qw(x)|x=L/2 = 46.2 x (189.94 – 50) = 6465.2 W/m2
6.14. Solution:
328
Air with u∞ = 30 m/s,T∞ = 25 0C
Tw = 150 0 C
10 mm
X1 = 700 mm Insulated
X2 = 750 mm
Mean film temperature = ½(Tw + T∞) = 0.5 x(25 + 150) = 87.5 0C.
Properties of air at 87.5 0C are : ν = 21.79 x 10 − 6 m2/s; k = 0.03075 W/(m-K);
Pr = 0.695. Assume Recr = 5 x 10 5
Since Rex1 > Recr, the flow is turbulent at x1 and the flow will be turbulent over the module
under consideration. Therefore the average heat transfer coefficient for the module can be
written as x2
[hav]m = {1 / (x2 – x1)} ∫ hxdx
x1
x2
= {1 / (x2 – x1)} ∫{k[ 0.037 (Rex)0.8 – 871]Pr 1/3 /x}dx
x1
k Pr 1/3 x2
k Pr 1/3
= ---------------- {0.037 (u∞/ν)0.8 ∫ x − 0.2 dx - 871∫ (dx /x) }
(x2 – x1)
1/3
k Pr
= ---------------- { 0.04625[Rex20.8 – Rex10.8] – 871 ln (x2 / x1)}
(x2 – x1)
329
0.03075 x (0.695)1/3{0.04625[(10.325 x105)0.8 –(9.325 X 105)0.8] – 871 ln (0.75/0.7)}
= -------------------------------------------------------------------------------------------------------
(0.75 – 0.7)
= 95.16 W / (m2-K)
For the module, power generation = qw = [hav]m {Tw - T∞} = 95.16 x (150 – 25)
(b) Since the bottom surface of the module is insulated, all the heat generated in the module
is transferred to air from the top surface of the module. Hence if q’’’ is the heat generated per
unit volume then
q’’’ (x2 – x1)δW = qw(x2 – x1)W,
where y is the coordinate measured in the direction of the thickness of the module.
The boundary conditions are (i) at y = 0, the surface is insulated i.e. dT/dy = 0
and at y = δ, T = Tw. The solution of Eq.(a) subject to the boundary conditions is given by
Since the bottom surface is insulated, the maximum temperature of the module will be at the
bottom surface (y = 0) and is therefore given by
Tmax = Tw + q’’’δ2 / 2k
6.15. Solution:
330
u∞ = 15 m/s; T∞ = 25 0 C
Tw = 1150C
L d = 0.025 m
= 18703
Therefore drag force = FD = ½(ρu∞2)LD CD = 0.5 x 1.0231 x 152 x 1 x 0.025 x 1.2 = 3.453 N
(ii) From Heat transfer data hand book for gases the average Nusselt number is given by
(μw = μ∞ can be considered for air )
84.04 x 0.0295
Therefore hav = Nuav k/d = -------------------- = 99.17 W/(m2-K)
0.025
6.16 Solution:
331
u∞ = 10 m/s; T∞ = 27 0C
Qw
D = 0.01 m
Given:- Qw = 1000 W/m; ks = 240 W/(m-K); ρs = 2700 kg/m3; Cps = 900 kJ/kg-K
(ii) time ‘t’ required for the surface to reach a temperature of (Tw – 10) 0C
Since Tw is not known it is not possible to read the properties at the mean film temperature.
Hence the problem has to be solved by trail and error procedure.
Trial 1:- Calculations are started using the properties of air at T∞.
Properties of air at 270C are:
Pr = 0.708
10 x 0.01
Red = u∞D / ν = ---------------------- = 6377.5
15.68 x 10 − 6
45.79 x 0.02624
Therefore hav = Nuav k/D = ------------------------- = 120.15 W/(m2- K)
0.01
332
Trial 2:- Assume Tw = 291.93 0C .Mean film temperature = ½(291.93 +27) = 159.5 0C.
10 x 0.01
Red = u∞D / ν = ---------------------- = 3323.4
30.09 x 10 − 6
Nuav = (0.4 x 3323.4 0.5 + 0.06 x 3323.4 2/3) x (0.682) 0.4 = 31.25
Trial 3:- Assume Tw = 303 0 C. Mean film temperature = 0.5 (303 + 27) = 165 0 C.
Pr = 0.682
10 x 0.01
Red = u∞D / ν = ---------------------- = 3238.34
30.88x 10 − 6
Nuav = (0.4 x 3238.34 0.5 + 0.06 x 3238.34 2/3) x (0.682) 0.4 = 30.8
(ii) To determine the time ‘t’ required for the surface of the cylinder to reach a temperature of
(307 – 10 = 2970C).
If dT/dt represents the rate of change of temperature of the surface w.r.t. time during unsteady
state, then energy balance equation for the cylinder can be written as follows:
Rate of increase of energy of the cylinder = Heat transfer rate by convection from
the surroundings to the cylinder +
333
Row 1 Row 2 Row 3
u∞,T∞∞
D
ST
SL
Since the surface temperature of the tubes is not known properties of air are evaluated at T∞.
Hence properties of air at 30 0C are:
ST / D 1.25
Umax = u∞ ------------------ = 1 x ---------------- = 5 m/s
[ST / D – 1] [ 1.25 – 1]
0.01 x 5.887
Reynolds number = Re = DGmax / μ = ------------------------ = 2969
1.983 x 10 − 5
From the graph friction factor f = 0.55 and Z = 1 for square arrangement as ST = SL.
N (Gmax) 2 10 x (5.887) 2
Now Δp = f ---------------- Z = 0.55 x ------------------------- x 1.0 = 161.9 N/ m2
ρ 1.1774
334
Here c2 = 0.27 ; m = 0.63 ; n = 0.
Since N <20, the above value of Nuav has to be multiplied by a correction factor.
u∞,T∞
ST
SD
Given: ST / D = SD / D = 1.25
ST / D
Maximum velocity Umax = u∞ ------------------ = 5 m/s as calculated above
[ST / D – 1]
ST / D 1.25
Or Umax = (½ )u∞ ------------------ = 0.5 x 1 x ------------- = 2.5 m/s
[SD / D – 1] (1.25 – 1)
We have to choose the higher of the two maximum velocities to calculate the Reynolds
number.Hence Umax and Reynolds number will be same as the above case.
335
From chart for staggered tube arrangement, f = 0.7and Z = 1
10 x 5.887 2
Therefore Δp = 0.7 x ----------------- x 1 = 206.1 N/m2
1.1774
= (√3/2)(ST/D)
Or ST/SL = 2 / √3 = 1.155
= 38.54
From the graph c3 = 0.98. Hence [ Nuav] N=10 = 0.98 x 38.54 = 37.8
6.20. Solution:
SL
336
D = 0.0125 m; number of tubes in each row = m = 40; L = 1 m ;
0.0125 x 10.394
Reynolds number = Re = DGmax / μ = ---------------------- = 4985.6
2.606 x 10 − 5
For in-line arrangement from data hand book, c2 = 0.27 and m = 0.63
48.18 x 0.03948
Average heat transfer coefficient = hav = [ Nuav] N=10 k / D = ----------------------
0.0125
= 152.175 W/(m2-K).
Energy balance between the tubes surfaces and hot gases can be written as
Heat transfer from hot gases to the tube surfaces = Q = (πDLNm)hav(T∞ - Tw)
337
Chapter 7 : Free Convective Heat Transfer
A. Free convection from/to plane surfaces:
338
A.Free Convection to or from plane surfaces
7.1. Solution:
Tw = 120 0 C
T∞ = 30 0C
L = 0.3m
Mean film temperature of air = 0.5 x (120 +30) = 750C
x
Properties of air at 750C are:
First we have to establish whether the flow become turbulent within the given length of the
plate by evaluating the Rayleigh number at x = L.
= 1.62 x 10 8
Since RaL < 10 9 flow is laminar for the entire height of the plate. Hence the average Nusselt
number is given by (from data hand book)
60.6 x 0.03
Therefore hav = Nuav k / L = ---------------- = 6.069 W /(m2 – K).
0.3
Total heat transfer fro both sides of the plate per unit width of the plate is given by
339
Qtotal = hav(2LW) ΔT = 6.06 x (2 x 0.3 x 1) X (120 – 30) = 327.726 W/m.
7.2. Solution:
Insulated
qw = 600 W/m2
T∞ = 30 0C
L = 0.25 m
Since Tw is not known, it is not possible to determine the mean film temperature at which
fluid properties have to be evaluated. Hence this problem requires a trial and error solution
either by assuming Tw and then calculate Tw by using the heat balance equation and check for
the assumed value or assume a value for h av ,calculate Tw and then calculate hav and check for
the assumed value of hav.Since it is difficult to guess a reasonable value for Tw to reduce the
number of iterations, it is preferable to guess a reasonable value for h av for air as we know
that for air hav varies anywhere between 5 and 15 W/(m2-K).
Or RaL* = 4.61 x 10 9.
Since RaL* >10 9 flow is turbulent for the entire length of the plate
340
Hence Nuav = 1.25 Nux|x=L = 1.25 x 0.17 x (4.61 x 10 9) 0.2 = 55.37
Therefore hav = 55.37 x 0.02896 / 0.25 = 6.41 W/(m2 – K)
Since the calculated value of hav deviates from the assumed value by about 34 %, one more
iteration is required.
Since the calculated value of hav is very close to the assumed value (error < , the iteration is
stopped. The surface temperature of the plate is therefore given by
Data:- Characteristic length = L = height of the plate = 0.3 m; Tw = 100 0C;T∞= 20 0C;
341
(18.97 x 10 − 6 ) 2
= 86.256 W
T
Heated surface facing downwards with
Heat transfer coefficient hbottom
Or RaL = 1.923 x 10 6.
From data hand book for heated surface facing upwards with constant surface temperature
the average Nusselt number is given by
342
Nutop = htopL/k = 0.54 x (RaL) 0.25 = 0.54 x ( 1.923 x 10 6) 0.25 = 20.11
Hence htop = 20.11 x 0.02896 / 0.075 = 7.76 W/(m2-K)
(b)To find hbottom:- From data hand book for heated surface facing downwards with constant
surface temperature, the average Nusselt number is given by
7.4. Solution: Data:- Horizontal circular plate with D = 0.25 m;Tw = 100 0C ; T∞ = 20 0C
This problem is similar to the previous problem except for the characteristic length. For a
horizontal circular plate of diameter D the characteristic length is given by
From data hand book for heated surface facing upwards with constant surface temperature
the average Nusselt number is given by
(b)To find hbottom:- From data hand book for heated surface facing downwards with constant
surface temperature, the average Nusselt number is given by
343
hbottom = 8.76 x 0.02896 / 0.0625 = 4.059 W/(m2-K)
From data hand book, for inclined plate with heated surface facing upwards the Nusselt number is giv
From data hand book for θ = − 50 0, Grc = 4 x 10 8 which is more than GrL.Hence the above
correlation cannot be used. Instead the following correlation has to be used.
Nuav = 0.59 (GrLcos θ Pr) ¼ = 0.59 x (2.23 x 107 x cos 500) ¼ = 36.3
344
The correlation for Nusselt number when the heated surface is facing downwards is given by
Nuav = 0.56 (GrLcos θ Pr) ¼ = 0.56 x (2.23 x 107 x cos 500) ¼
= 34.45
hav = 34.45 x 0.02791 / 0.25 = 3.84 W/(m2-K).
7.6. Solution:
Tw = 225
W
T∞=200C
L=0.2 m
Mean film temperature = 0.5 x (225 + 20) = 122.5 0C. Properties of air at 122.50C are: β = 1/(122.5 + 273
k = 0.03365 W/(m-K).
1000
Now Q = havLWΔT or W = Q / (havLΔT) = -----------------------------
7.94 x 0.2 x (225 – 20)
= 3.0718 m
345
7.7. Solution:
θ = − 60 0 T∞ = 25 0 C
Insulated
Since Tw is not known, mean film temperature to evaluate the fluid properties cannot be
determined. Hence the problem requires a trial- and – error method by suitably assuming a
value for hav and then check for this assumpotion.
For θ = − 60 0, Grc = 10 8. Since GrL > Grc the average Nusselt number is given by
= 164.45
346
Since the calculated value of hav is quite different from the assumed value one more iteration
is required.
Tw = 121.25 0C ;Tm = 97.18 0C C;Tβ = 49.06 0C; GrL = 2.311 x 10 9;Nuav = 159.96
Hence hav = 6.88 W/(m2-K). This value agrees with the assumed value within 8 %. Hence the
iteration is stopped and the equilibrium temperature of the plate surface is calculated as
= 60 0C.
Properties of air at 60 0C are:Pr = 0.696
347
(L/d)
Since ------- < 0.025, the vertical tube/cylinder can be treated as a vertical flat surface
GrL1/4
For calculating the Nusselt number.
Now RaL = GrLPr = 2.21 x 10 10 x 0.696 = 1.538 x 10 10
Rate of heat transfer = Q = πdLhavΔT = π x 0.05 x 1.5 x (100 – 20) x 4.81 = 90.67 W
(ii) When the pipe is horizontal:- When the pipe is horizontal, the characteristic length is the
diameter. Hence
Rad
½
[Nuav] = 0.60 + 0.387 -----------------------------------
[1 + (0.559/Pr)9/16] 8/27
1/6
5
(5.697x 10 )
½
[Nuav] = 0.60 + 0.387 ----------------------------------- = 4.01 or Nuav = 16.08
[1 + (0.559/0.696)9/16] 8/27
7.9. Solution:
T∞ = 27 0C Tw = ?
qw = 30 W/m d =.025 m
348
Since Tw is not known, it is not possible to evaluate the fluid properties at the mean film
temperature. Hence the problem has to be solved by trial and error solution by assuming a
suitable value for hav and check for the assumed value.
Hence Nuav = 7.84 or hav = 7.84 x 0.028 / 0.025 = 8.78 W/(m2 – K).
Since the calculated value of hav deviates very much from the assumed value one more
iteration is required.
The calculated value of hav agrees with the assumed value within 5 % iteration is stopped.
349
The equilibrium temperature of the surface = Tw = 27 + 30 /(π x 0.025 x 9.5691) = 67 0C.
7.10. Solution:
Q
d
L
550
Wall heat flux = qw = Qw /(πdL) = ---------------------- = 58357 W/m2
(π x 0.01 x 0.3)
Since Tw is not known, fluid properties cannot be evaluated at the mean temperature and hence the proble
For free convection in liquids the order of heat transfer coefficient is around 10 to 1000 W/
(m2-K).Let us assume hav = 1000 W/(m2-K).
Nud = C Rad n with C = 0.125 and n = 1/3 for this value of Rad.
350
Hence hav = 33.908 x 0.639 / 0.01 = 2166.21 W/(m2 –K)
Tw = 20 + 58357 / 1763 = 53.1 0C. Mean film temperature = 0.5 x (20 + 53.1) = 36.55 0C.
Since the calculated value of hav agrees with the assumed value within 4%, iteration is
stopped and the equilibrium temperature of the heater is calculated as
When a heated surface is exposed to air the order of heat transfer coefficient varies between 5
and 20 W/(m2-K).
Tw = 20 + 58357 / 20 = 2938 0C. Mean film temperature = 0.5 x (20 + 2938) = 1479 0C
351
k = 0.108 W/(m-K); ν = 294.3 x 10 − 6 m2/s.
Hence hav = 2.13 x 0.108 / 0.01 = 23 W/(m2-K) This is 13% away from the assumed value
and hence one more iteration is required.
k = 0.0978 W/(m-K)
Nuav = 0.850 x (188) 0.188 = 2.275.Hence hav = 2.275 x 0.0978 / 0.01 = 22.25 W/(m2 – K).
This value of hav agrees with the assumed value within 4% and hence the iteration is stopped.
The equilibrium temperature of the heater is therefore given by
352
Chapter 8 : Condensation & Boiling
8.1. Introduction: Knowledge of heat transfer occurring during change of phase i.e. during
condensation and boiling is very useful in a number of ways. For example in all power and
refrigeration cycles, it is necessary to convert a liquid into a vapour and vice-versa. This is
accomplished in boilers or evaporators and condensers.
The rate of heat transfer during the two types of condensation processes
is quite different. For the same temperature difference between the vapour and the surface,
the heat transfer rates in drop-wise condensation are significantly higher than those in film-
wise condensation. Therefore it is preferable to have drop-wise condensation from the
designer’s point of view if the thermal resistance on the condensing side is a significant part
of the total thermal resistance. However it is generally observed that, although drop-wise
condensation may be obtained on new surfaces, it is difficult to maintain drop-wise
condensation continuously and prolonged condensation results in a change to film-wise
condensation. Therefore it is still the practice to design condensers under the conservative
assumption that the condensation is of film type.
8.3. Nusselt’s theory for laminar film-wise condensation on a plane vertical surface:-
The problem of laminar film-wise condensation on a plane vertical surface was first
analytically solved by Nusselt in 1916.He made the following simplifying assumptions in his
analysis.
(i) The fluid properties are constant.
(ii) The plane surface is maintained at a uniform temperature, Tw which is less than the
saturation temperature Tv of the vapour.
(iii) The vapour is stationary or has a very low velocity and so it does not exert any drag on
the motion of the condensate: i.e., the shear stress at the liquid-vapour interface is zero.
353
(iv) The flow velocity of the condensate layer is so low that the acceleration of the
condensate is negligible.
(v) The downward flow of the condensate under the action of gravity is laminar.
(vi) Heat transfer across the condensate layer is purely by conduction; hence the liquid
temperature distribution is linear.
[τ + (∂τ/∂y)dy]dx
x pdy τdx
pdy
δ
dx
(p+dp)dy (p+dp)dy
ρLdxdyg ρvdxdyg
Expression for (dp/dx) in terms of vapour density ρv can be obtained by making a force
balance for a vapour element as shown in Fig. 8.1(b). The force balance gives
354
or (dp/dx) = ρvg
(∂τ/∂y) = (ρv−ρL)g
(ρv−ρL)g y C1
Or (∂u/∂y) = --------------- + ------- ……(8.2)
μL μL
(ii) at y = δ, (∂u/∂y) = 0.
Condition (i) in Eq. (8.3) gives C2 = 0 and condition (ii) in Eq. (8.2) gives
(ρv−ρL)g δ C1
0 = ------------- + ---------
2 μL μL
(ρv−ρL)g δ
Therefore C1 = − --------------
2
Substituting for C1 and C2 in Eq.(8.3) we get the velocity distribution in the condensate layer
as
g(ρL − ρv)
u(y) = --------------- [ δy – (y2/2)] …………………(8.4)
μL
355
δ
m = ∫ ρL{ g(ρL − ρv) / μL}[ δy – (y2/2)]dy
0
g ρL (ρL − ρv) δ 3
= ------------------ ………………………………..(8.5)
3 μL
g ρL (ρL − ρv) δ 2 dδ
Hence dm = ----------------------
μL
Energy balance for the condensate element shown in the figure can be written as
dq = kL(Tv – Tw)dx / δ
kL μL (Tv – Tw)dx
3
or δ dδ = ----------------------
g ρL (ρL − ρv) hfg
Integrating we get
δ4 kL μL (Tv – Tw)x
----- = --------------------- + C 3
4 g ρL (ρL − ρv) hfg
At x = 0, δ = 0. Hence C3 = 0.
4 kL μL (Tv – Tw)x
or δ = [-------------------------- ] 1/ 4……………………(8.7)
g ρL (ρL − ρv) hfg
356
Now kL (Tv – Tw)dx
----------------- = hx dx [Tv – Tw]
δ
kL g ρL (ρL − ρv) hfg kL3
Therefore hx = --------- = [ --------------------------] 1 / 4
δ 4 μL (Tv – Tw)x
The average heat transfer coefficient for a length L of the plate is given by
L
hav = (1/L) ∫ hxdx ………………………………………(8.9)
0
L
Hence hav = (1/L) C ∫ x − ¼ dx = (C / L) (4/3) L− ¼ = (4/3)C L− ¼
0
8.4. Condensation on Inclined Surfaces : Nusselt,s analysis given above can readily be
extended to inclined plane surfaces making an angle θ with the horizontal plane as shown in
Fig. 8.2.
357
y
θ
g
The component of the gravitational force along the length of the pate is g sin θ.The expressions for local a
…………………………(8.13)
8.5. Condensation on a horizontal tube: The analysis of heat transfer for condensation on
the outside surface of a horizontal tube is more complicated than that for a vertical
surface. Nusselt,s analysis for laminar film-wise condensation on the surface of a horizontal
tube gives the average heat transfer coefficient as
where D is the outside diameter of the tube. A comparison of equations (8.11) and (8.14) for
condensation on a vertical tube of length L and a horizontal tube of diameter D gives
[hav]vertical 0.943
--------------- = ------------(D/L) ¼ = 1.3 (D/L) 1/4 .....................................(8.15)
[hav]horizontal 0.725
This result implies that for a given value of (Tv – Tw), the average heat transfer coefficient for
a vertical tube of length L and a horizontal tube of diameter D becomes equal when L =
358
2.856 D.For example when L = 100 D, theoretically [hav]horizontal would be 2.44 times [hav]vertical.
Therefore horizontal tube arrangements are generally preferred to vertical tube arrangements
in condenser design.
condensate from one tube drains on to tube just below. If it is assumed that the drainage from
one tube flows smoothly on to the tube below, then for a vertical tier of N tubes each of
diameter D, the average heat transfer coefficient for N tubes is given by
This relation generally gives a conservative value for the heat transfer coefficient. Since
some turbulence and some disturbance of condensate are unavoidable during drainage, the
heat transfer coefficient would be more than that given by the above equation.
8.7. Reynolds number for condensate flow: Although the flow hardly changes to turbulent
flow during condensation on a single horizontal tube, turbulence may start at the lower
portions of a vertical tube. When the turbulence occurs in the condensate film, the average
heat transfer coefficient begins to increase with the length of the tube in contrast to its
decrease with the length for laminar film condensation. To establish a criterion for transition
from laminar to turbulent flow, a “Reynolds number for condensate flow” is defined as
follows.
ρL uav Dh
Re = ----------------- …………………..(8.17)
μL
359
where uav is the average velocity of the condensate film and Dh is the hydraulic diameter for
the condensate flow given by
4 x (Cross sectional area for condensate flow) 4A
Dh = --------------------------------------------- --------- = -------
Wetted Perimeter P
4A ρL uav 4M
Therefore Re = -------------------- = ---------------……………..(8.18)
P μL P μL
where M is mass flow rate of condensate at the lowest part of the condensing surface in kg/s.
The wetted perimeter depends on the geometry of the condensing surface and is given as
follows.
πD …..For vertical tube of outside diameter D ………….(8.19 a)
P = 2L …...For horizontal tube of length L …………………(8.19 b)
W ….. For vertical or inclined plate of width W………...(8.19 c)
Experiments have shown that the transition from laminar to turbulent condensation takes
place at a Reynolds number of 1800. The expression for average heat transfer coefficient
for a vertical surface [Eq.(8.11)] can be expressed as follows.
ReL
(hav / kL)( νL2 / g) 1 / 3 = ------------------------ ………………(8.22)
[1.08 ReL1.22 – 5.2]
360
8.8. Turbulent film condensation: For turbulent condensation on a vertical surface,
Kirkbride has proposed the following empirical correlation based on experimental data.
In the above correlation the physical properties of the condensate should be evaluated at the
arithmetic mean temperature of Tv and Tw.
8.9. Film condensation inside horizontal tubes: In all the correlations mentioned above, it
is assumed that the vapour is either stationary or has a negligible velocity. In practical
applications such as condensers in refrigeration and air conditioning systems, vapour
condenses on the inside surface of the tubes and so has a significant velocity. In such
situations the condensation phenomenon is very complicated and a simple analytical
treatment is not possible. Consider, for example, the film condensation on the inside surface
of a long vertical tube. The upward flow of vapour retards the condensate flow and causes
thickening of the condensate layer, which in turn decreases the condensation heat transfer
coefficient. Conversely the down ward flow of vapour decreases the thickness of the
condensate film and hence increases the heat transfer coefficient.
Chato recommends the following correlation for condensation at low vapour
velocities inside horizontal tubes:
This result has been developed for the condensation of refrigerants at low Reynolds number
[Rev = (ρvuvD) / μv < 35,000 ; Rev should be evaluated at the inlet conditions.]
For higher flow rates, Akers, Deans and Crosser propose the following
correlation for the average condensation heat transfer coefficient on the inside surface of a
horizontal tube of diameter D:
hav D
------ = 0.026 Pr 1 / 3 [ReL + Rev(ρL / ρv) ½ ] 0.8 ………..(8.25)
k
The above equation correlates the experimental data within 50 % for ReL > 5000 and
361
Example 8.1: Saturated steam at 1.43 bar condenses on a 1.9 cm OD vertical tube which is
20 cm long. The tube wall is at a uniform temperature of 109 0C . Calculate the average heat
transfer coefficient and the thickness of the condensate film at the bottom of the tube.
Solution: Data:- Tv = Saturation temperature at 1.43 bar = 110 0 C (from steam tables)
Mean film temperature of the condensate (water) = 0.5 x (110 + 109) = 109.5 0C.
k = 0.685 W/(m-K); ν = 0.2714 x 10 − 6 m2/s; hfg = 2230 kJ/kg. Also ρL >>> ρv.
Let us assume that the condensate flow is laminar and later check for this assumption.
= 17,653 W / (m2-K)
Check for Laminar flow assumption:- The relation between hav and Reynolds number at the
bottom of the tube is given by
= 24.72
Since ReL < 1800, our assumption that condensate flow is laminar is correct.
362
Example 8.2:- Saturated steam at 80 0C condenses as a film on a vertical plate 1 m high. The
plate is maintained at a uniform temperature of 70 0C. Calculate the average heat transfer
coefficient and the rate of condensation. What would be the corresponding values if the effect
of ripples is taken into consideration.
Solution:Data:- Tv = 80 0C; Tw = 70 0C; Mean film temperature =0.5 x (80 + 70) = 75 0C.
Assuming laminar film condensation the average heat transfer coefficient is given by
4M
Check for laminar flow assumption :- ReL = --------------, where P = width of the plate for
μL P
4 x 0.0263
vertical flat plate. Hence ReL = ------------------------- = 276
381 x 10 − 6
Since ReL < 1800, the condensate flow is laminar.
Since ReL > 30, it is clear that the effects of ripples have to be considered.
4M 4 hav L (Tv – Tw)
Now ReL = ------------ = -----------------------
μL P μL P hfg
ReL μL P hfg
Hence hav = ------------------ …………………………………………….(1)
4L(Tv – Tw)
363
When the effects of ripples are considered the relation between ReL and hav is given by Eq.
(8.22) as follows:
ReL
1.08 ReL1.22 – 5.2 = ----------------------- Substituting for hav from Eq.(1) we have
(hav/kL)(νL2 /g)1 /3
[ It can be seen that the ripples on the surface increase the heat transfer coefficient by about
15 %].
Example 8.3:- Air free saturated steam at 65 0C condenses on the surface of a vertical tube
of OD 2.5 cm. The tube surface is maintained at a uniform temperature of 35 0C. Calculate
the length of the tube required to have a condensate flow rate of 6 x 10 −3kg/s.
4M 4 x 6 x 10 − 3
Reynolds number = Re = --------------- = ------------------------------- = 544
μL πDo 0.562 x 10 − 3 x π x 0.025
364
Since Re < 1800 flow is laminar. It is more convenient to use Eq.(8.21)
= 3599 W/(m2 – K)
M hfg 6 x 10 − 3 x 2346 x 10 3
Therefore L = ---------------------- = ------------------------------------
hav πDo [Tv – Tw] 3599 x π x 0.025 x (65 – 35)
= 1.66 m
Example 8.4:- Air free saturated steam at 85 0C condenses on the outer surfaces of 225
horizontal tubes of 1.27 cm OD, arranged in a 15 x 15 array. Tube surfaces are maintained
at a uniform temperature of 75 0C. Calculate the total condensate rate per one metre length
of the tube.
Mean film temperature = 0.5 x (85 + 75) = 80 0C. Properties of the condensate (liquid
365
= 7142 x 225 x π x 0.0127 x 1 x (85 – 75) = 641.14 x 10 3 W
Mass flow rate of condensate = M = Q / hfg = 641.14 x 10 3 / 2296 x 10 3 = 0.28 kg/ (s-m)
Example 8.5:- Superheated steam at 1.43 bar and 200 0C condenses on a 1.9 cm OD vertical
tube which is 20 cm long. The tube wall is maintained at a uniform temperature of 109 0C.
Calculate the average heat transfer coefficient and the thickness of the condensate at the
bottom of the tube. Assume cp for super heated steam as 2.01 kJ/(kg-K).
Solution: With a superheated vapour, condensation occurs only when the surface
temperature is less than the saturation temperature corresponding to the vapour pressure.
Therefore for a superheated vapour, the amount of heat to be removed per unit mass to
condense it is given by
Where cp is the specific heat of superheated steam and Tsat is the saturation temperature
corresponding to the vapour pressure. If it is assumed that the liquid – vapour interphase is at
the saturation temperature, then Eq.(8.20 ) still holds good with hfg replaced by
hfg + cpv(Tv – Tsat).
At 1.43 bar, Tsat = 110 0C.Mean film temperature = 0.5 x (110 + 109) = 109.5 0C.
Example 8.6:- Air free saturated steam at 70 0C condenses on the outer surface of a 2.5 cm
OD vertical tube whose outer surface is maintained at a uniform temperature of
50 0C. What length of the tube would produce turbulent film condensation?
366
Solution: Data:- Tv = 70 0C; Tw = 50 0C; Do = 0.025 m; Vertical tube.
Mean film temperature = 0.5 x (70 + 50) = 60 0C. Properties of the condensate (liquid
= 0.0166 kg / s.
=7m
Example 8.7:- Saturated steam at 100 0C condenses on the outer surface of a 2 m long
vertical plate. What is the temperature of the plate below which the condensing film at the
bottom of the plate will become turbulent?
Solution: Data:- Tv = 100 0C; L = 2 m. Since Tw is not known, properties of the condensate at
the mean film temperature cannot be determined and therefore the problem has to be solved
by trial and error procedure as follows:
Trial 1:- The properties of the condensate are read at Tv = 100 0C. The properties are
367
Since the flow has to become turbulent at the bottom of the plate we have
9.81 x 0.683 3
Hence hav = 0.0077 x (1800) x [ ------------------------------ ] 1 / 3
0.4
(0.2824 x 10 − 3 / 958.4) 2
= 5098 W / (m2 – K)
= 72 0C
Trial 2:- Assume Tw = 72 0C. Mean film temperature = 0.5 x (100 + 72) = 86 0C.
9.81 x 0.677 3
Hence hav = 0.0077 x (1800) x [ ------------------------------ ] 1 / 3
0.4
(0.3349 x 10 − 3 / 968.5) 2
Since the calculated value of Tw is quite different from the assumed value, one more iteration
is required.
Trial 3:- Assuming Tw = 60 0C and proceeding on the same lines as shown in trial 2 we
get hav = 4365 W /(m2 – K) and hence Tw = 590C. This value is very close to the value
assumed (difference is within 2 % ). The iteration is stopped. Hence Tw = 59 0C.
368
Example 8.8:- Air free saturated steam at 90 0C condenses on the outer surface of a 2.5 cm
OD, 6 m long vertical tube, whose outer surface is maintained at a uniform temperature of
60 0C. Calculate the total rate of condensation of steam at the tube surface.
Mean film temperature = 0.5 x (90 + 60) = 75 0C. Properties of the condensate at 75 0C
are: kL = 0.671 W/(m-K); μL = 0.3805 x 10 − 3 N-s/m2; ρL= 974.8 kg/m3; At 90 0C, hfg =
2283 x 10 3 J/(kg-K).
We do not know whether the condensate flow is laminar or turbulent. We start the
calculations assuming laminar flow and then check for laminar flow condition.
For laminar flow
g ρL2 kL3 hfg
hav = 0.943 [ ----------------------------- ] 1 / 4
μL(Tv – Tw) L
= 2443
ReL = 4M / (μLπDo). But Mhfg = havπDoL (Tv – Tw) or M / (πDo) = havL (Tv – Tw) / hfg
369
Substituting this expression for ReL in equation (1) we have
4 L (Tv – Tw)
(hav )0.6 = 0.0077 [------------------------- ] 0.4 [νL2 / (gkL3) ] − 1 / 3
hfg μL
370
Where ∆Te = [Tw - Tsat] and is termed as the “excess temperature”. The boiling process is
characterized by the formation of vapour bubbles which grow and subsequently detach from
the surface. Vapour bubble growth and dynamics depend, in a complicated manner, on the
excess temperature ∆Te, the nature of the surface, and the thermo-physical properties of the
fluid, such as its surface tension. In turn the dynamics of vapour bubble growth affect fluid
motion near the surface and therefore strongly influence the heat transfer coefficient.
Boiling may occur under varying conditions. For example if the liquid is
quiescent and if its motion near the surface is due to free convection and due to mixing
induced by bubble growth and detachment, then such a boiling process is called “pool
boiling”. In contrast in “forced convection boiling”, the fluid motion is induced by an
external means as well as by free convection and bubble induced mixing. Boiling may also
be classified as “sub-cooled boiling” and “saturated boiling”. In sub-cooled boiling, the
temperature of the liquid is below the saturation temperature and the bubbles formed at the
surface may condense in the liquid. In contrast, in saturated boiling, the temperature of the
liquid slightly exceeds the saturation temeperature, Bubbles formed at the surface are then
propelled through the liquid by buoyancy forces, eventually escaping from a free surface.
8.13. Pool Boiling Regimes: The first investigator who established experimentally the
different regimes of pool boiling was Nukiyama. He immersed an electric resistance wire
into a body of saturated water and initiated boiling on the surface of the wire by passing
electric current through it. He determined the heat flux as well as the temperature from the
measurements of current and voltage. Since the work of Nukiyama, a number of
investigations on pool boiling have been reported. Fig. 8.4 illustrates the characteristics of
pool boiling for water at atmospheric pressure. This boiling curve illustrates the variation of
heat flux or the heat transfer coefficient as a function of excess temperature ∆Te. This curve
pertains to water at 1 atm pressure.From Eq. (8.27) it can seen that qw depends on the heat
transfer coefficient h and the excess temperature ∆Te.
Free Convection Regime(up to point A):- Free convection is said to exist if ∆Te ≤ 5 0 C. In
this regime there is insufficient vapour in contact with the liquid phase to cause boiling at the
saturation temperature. As the excess temperature is increased, the bubble inception will
eventually occur, but below point A (referred to as onset of nucleate boiling,ONB), fluid
motion is primarily due to free convection effects.Therefore,according to whether the flow is
laminar or turbulent, the heat transfer coefficient h varies as ∆Te1/4 or as ∆Te1/3 respectively so
that qw varies as ∆Te5/4 or as ∆Te4/3.
Nucleate Boiling Regime(Between points A and C):- Nucleate boiling exists in the range 5 0
C ≤ ∆Te ≤ 30 0 C. In this range, two different flow regimes may be distinguished. In the
region A – B, isolated bubbles form at nucleation sites and separate from the surface,
substantially increasing h and qw. In this regime most of the heat exchange is through
direct transfer from the surface to liquid in motion at the surface, and not through vapour
bubbles rising from the surface. As ∆Te is increased beyond 10 0C (Region B-C), the
nucleation sites will be numerous and the bubble generation rate is so high that continuous
columns of vapour appear. As a result very high heat fluxes are obtainable in this region. In
practical applications, the nucleate boiling regime is most desirable, because large heat fluxes
371
are obtainable with small temperature differences. In the nucleate boiling regime, the heat
increases rapidly with increasing excess temperature
120
107
106
qw, W/m2
105
104
103
∆Te until the peak heat flux is reached. The location of this peak heat flux is called the
burnout point, or departure from nucleate boiling (DNB), or the critical heat flux (CHF). The
reason for calling the critical heat flux the burnout point is apparent from the Fig. 8.4. Such
high values of ∆Te may cause the burning up or melting away of the heating element.
Film Boiling Regime:- It can be seen from Fig. 8.4 that after the peak heat flux is reached,
any further increase in ∆Te results in a reduction in heat flux. The reason for this curious
phenomenon is the blanketing of the heating surface with a vapour film which restricts
liquid flow to the surface and has a low thermal conductivity. This regime is called the film
boiling regime. The film boiling regime can be separated into three distinct regions namely
(i) the unstable film boiling region, (ii) the stable film boiling region and (iii) radiation
dominating region. In the unstable film boiling region, the vapour film is
unstable, collapsing and reforming under the influence of convective currents and the
surface tension. Here the heat flux decreases as the surface temperature increases, because
the average wetted area of the heater surface decreases. In the stable film boiling region, the
heat flux drops to a minimum, because a continuous vapour film covers the heater surface.In
the radiation dominating region, the heat flux begins to increase as the excess temperature
increases, because the temperature at the heater surface is sufficiently high for thermal
radiation effects to augment heat transfer through the vapour film.
372
8.14. Pool Boiling Correlations:
Correlation for The Nucleate Boiling Regime:- The heat transfer in the nucleate boiling
regime is affected by the nucleation process, the distribution of active nucleation sites on the
surface, and the growth and departure of bubbles.Numerous experimental investigations have
been reported and a number of attempts have been made to correlate the experimental data
corresponding to nucleate boiling regime.The most successful and widely used correlation
was developed by Rohsenow. By analyzing the significance of various parameters in relation
to forced - convection effects. He proposed the following empirical relation to correlate the
heat flux in the entire nucleate boiling regime:
In Eq. (8.28) the exponent n and the coefficient Csf are the two provisions to adjust the
correlation for the liquid – surface combination. Table 8-1 gives the experimentally
determined values Csf for a variety of liquid – surface combinations. The value of n should be
taken as 1 for water and 1.7 for all other liquids shown in Table 8 – 1.
373
Table 8 – 1: Values of Csf of Eq. (8.28) for various liquid – surface combinations
Correlations for Peak Heat Flux:- The correlation given by Eq. (8.28) provides information
for the heat flux in nucleate boiling, but it cannot predict the peak heat flux.Based on stability
considerations, Kutateladze and Zuber veloped the following correlation to calculate the peak
heat flux in pool boiling from an infinite horizontal plate facing up.
Π σ* g (ρl – ρv)
qmax = ----- ρv hfg [ ---------------------- ] ¼ [ 1 + ρv / ρl ] ½ ………….(8.29)
24 ρv2
It is apparent from this equation that large values of hfg, ρv, g and σ * are desirable for a large
value of the peak heat flux. For example, water has a large value of hfg; hence the peak heat
flux obtainable with boiling water is high. This equation also shows that a reduced
gravitational field decreases the peak heat flux. For most situations, the quantity
[ 1 + ρv / ρl ] ½ is approximately equal to unity. Hence Eq. (8.29) can be written as
π
qmax = ----- ρv1/2 hfg [σ* g (ρl – ρv) ] ¼ ………………………..(8.30)
24
Lienhard and Dhir improved the analysis and considered the effect of the size of the
horizontal plate. They showed that
374
qmax = 0.149 hfg ρv1/2 [σ* g (ρl – ρv) ] ¼ ……………………….(8.31)
The above expression was shown to be valid as long as the plate is large and the
dimensionless quantity L [g (ρl – ρv) / σ* ] ½ ≥ 2.7, where L is the characteristic dimension of
the plate. For circular plate L is taken as the diameter, while for a square plate it is taken as
the side of the plate.
For the case of horizontal cylinders of radius R, Lienhard & Sun recommended the
following modified form of Eq. (8.30).
π
qmax = ----- ρv1/2 hfg [σ* g (ρl – ρv) ] ¼ f(L’) ……………………..(8.32)
24
where f(L’) = 0.89 + 2.27 exp { − 3.44 √L’ }, and L’ = R [g (ρl – ρv) / σ* ] ½.
Eq. (8.32) is valid for situations in which L’ ≥ 0.15. This equation cal also be used for large
spheres in which case f(L’) = 0.84 and L’ = 4.26. For small spheres
f(L’) = 1.734 / √L’, where 0.15 ≤ L’ ≤ 4.26.
Correlations for Film boiling Regime:- No satisfactory correlation exists for the unstable
film boiling region. For the stable film boiling region, Bromley has derived the following
correlation for horizontal cylinders:
Where ho = average boiling heat transfer coefficient in the absence of radiation W/(m2 –K),
D = outside diameter of the tube,
hfg* = Difference between the enthalpy of the vapour at the mean film temperature,Tm
(Tm = [Tw + Tsat] / 2) and the enthalpy of the liquid at the saturation temperature
where hr is the radiation heat transfer coefficient which is calculated from the formula for the
radiation heat exchange between parallel planes;
1 σ [ Tw4 – Tsat4 ]
hr = ----------------------- ------------------------- ……(8.35)
375
[ 1 / ε + 1 / α – 1] [Tw – Tsat
Eq.(8.34) is difficult to use because a trial and error approach is needed to determine h. When
hr is smaller than ho, Eq. (8.34) can be replaced by
h = ho + ¾ hr ………………………………………(8.36)
Example 8.9:- Saturated water at 100 0C is boiled with a copper heating element having a
heating surface area of 0.04 m2which is maintained at a uniform temperature of 115 0C.
Calculate the surface heat flux and the rate of evaporation of water. Also calculate the
critical flux.
Solution: Given:- Tsat = 100 0C; Tw = 115 0C; Surface area = A = 0.04 m2;
Properties of liquid water at 100 0C are: Cpl = 4216 J/(kg – K); hfg = 2257 x 10 3 J/kg;
Since ∆Te lies between 5 0C and 30 0C, the boiling is in the nucleate regime.
Csf = 0.0130 for water – copper combination and for water n = 1.For nucleate boiling
Hence total heat transfer = Q = Aqw = 0.04 x 4.84 x 10 5 = 19.36 x 10 3 W = 19.36 kW.
19.36 x 10 3
Rate of evaporation = M = Q / hfg = ----------------- = 8.58 x 10 − 3 kg / s.
376
2257 x 10 3
Example 8.10:- A metal clad heating element of 6 mm diameter and emissivity equal to unity
is horizontally immersed in a water bath. The surface temperature of the metal is 255 0C
under steady state boiling conditions. If the water is at atmospheric pressure estimate the
power dissipation per unit length of the heater.
Solution: Given:- Tw = 255 0C ; Tsat = saturation temperature of water at 1 atm = 100 0C;
∆Te = 255 – 100 = 155 0C. Since ∆Te > 120 0C, film boiling conditions will prevail. The heat
transfer in this regime is given by Eq.(8.33) namely
μv = 14.85 x 10 − 6 kg / (m-s).
= 460 W/(m2 – K)
1 σ {Tw4 – Tsat4}
hr = ---------------------- x ------------------------------
[1/ε + 1/α − 1 ] {Tw – Tsat}
377
[1/1+1/1−1] { 528 – 373}
= 21.3 W / (m2-K).
Example 8.11:- A vessel with a flat bottom and 0.1 m2 in area is used for boiling water at
atmospheric pressure. Find the temperature at which the vessel must be maintained if a
boiling rate of 80 kg/h is desired. Assume that the vessel is made of copper and the boiling is
nucleate boiling. Take ρv = 0.60 kg/m3.
Solution: Given:- A = 0.1 m2; Tsat = 100 0C; M = 80 kg/h = 0.022 kg/s; Prl = 1.75
hfg = 2257 x10 3 J /kg; Cpl = 4216 J/(kg-K); ρl = 960.6 kg/m3; σ* = 58.8 x 10 − 3 N/m;
For nucleate boiling Eq.(8.28) is used to calculate the excess temperature .∆Te
4216 x ∆Te
----------------------- = 0.013 x {4.965 x 10 5/(282.4 x 10− 6x 2257 x 103)
2257 x 10 3 x 1.75 _____________________________
x √ 58.8 x 10 − 3 / [9.81 x (960.6 – 0.6)] } 0.33
Or ∆Te = 15.2 0C
Example 8.12:- Calculate the heat transfer coefficient during stable film boiling of water
from a 0.9 cm diameter horizontal carbon tube. The water is saturated and at 100 0C and the
tube surface is at 1000 0C. Take the emissivity of the carbon surface to be 0.8 and assume
that at the average film temperature, the steam has the following properties.
kv = 0.0616 W/(m-K); ρv = 0.266 kg/m3; μv = 28.7 x 10 −6 kg/(m-s); Cpv = 2168 J/(kg-K);
ρl = 958.4 kg/m3
378
Solution: Given:- D = 0.009 m; ∆Te = Tw – Tsat = 1000 – 100 = 900 0C; ε = 0.8; α = 1.0
hfg* = hfg + 0.8 Cpv ∆Te = 2257 x 10 3 + 0.8 x 2168 x 900 = 3818 x 10 3 J/kg.
= 194 W/(m2 – K)
CHAPTER 9
HEAT EXCHANGERS
9.1. Introduction: Heat exchangers are devices in which heat transfer takes place between
two or more fluids.Over the period of time many different types of heat exchangers have
379
been developed to suit particular applications. We find the use of heat exchangers in steam
power plants, chemical processing plants, air conditioning plants, household refrigerators,
automobiles etc.
9.2. Classification of Heat Exchangers: There are many classifications of heat exchangers.
Here we consider classifications based on (i) the heat transfer process, (ii) compactness, (iii)
construction type, (iv) flow arrangement, and (v) heat transfer mechanism.
(i) Classification based on heat transfer process : Based on the heat transfer process, heat
exchangers are classified as
(a) Direct transfer type (or recuperative),
(b) Storage type (or regenerative type)
(c) Direct mixing type.
A “direct transfer type” of heat exchanger is one in which the cold and hot fluids flow
simultaneously through the exchanger and heat is transferred through a wall separating the
two fluids. This type of heat exchanger is the one which is mostly used.
Cold fluid in
Hot
Fluid in Hot
Fluid out
simple example of this type is shown in Fig. 9.1. The heat exchanger has a concentric tube
arrangement. One fluid flows through the inner tube, while the other fluid flows through the
annular space between the two tubes. The heat transfer takes place across the wall of the
inner tube. Although both the fluids flow through the exchanger simultaneously, there is no
mixing of the two fluid There are no moving parts.The most serious defect of this type of
heat exchanger is the fact that , with passage of time, scale and dirt tends to accumulate on
the heat transfer surface. This accumulation (called fouling) increases the thermal resistance
to heat flow so that the performance of the heat exchanger deteriorates.
A “storage type” heat exchanger or regenerator is one in which the heat transfer from
the hot fluid to the cold fluid occurs through a coupling medium in the form of a porous solid
matrix. The hot and cold fluids flow alternately through the matrix, the hot fluid giving heat
380
to the matrix and the cold fluid extracting heat from the matrix. The arrangement is
illustrated in Fig. 9.2.In many applications, a continuous flow has to be
Cold fluid in
maintained on both the hot and cold sides.In such cases, it is a common practice to use a
rotating disc type matrix, every element of which passes from the hot stream to the cold and
back again during each revolution.. Storage type heat exchangers are commonly used in open
–hearth steel melting furnaces and in blast furnaces.It is also used at very low temperatures in
certain cryogenic cycles where a very high degree of performance is required.
Because of alternate flow of hot and cold fluid through the matrix, the storage type
of heat exchanger is self cleaning. This type also results in a much more compact
arrangement than the direct transfer type. The major disadvantages of the storage type are
that some mixing of hot and cold fluids is inevitable and that sealing the hot side from the
cold side in the rotary design presents considerable difficulty.
A “direct contact type” of heat exchanger is one in which the two fluids come in direct
contact with one another in the heat exchanger. Therefore this type of heat exchanger can be
used only for two immiscible fluids.Cooling towers of steam power plants and scrubbers
used in chemical plants are examples of this type. If heat has to be transferred between a gas
and a liquid, the gas is either bubbled through the liquid or the liquid is sprayed in the form
of droplets into the gas. A schematic diagram of a direct contact type heat exchanger is
shown in Fig. 9.3. Very often in this type of heat exchanger in addition to heat transfer there
will be mass transfer also.
381
Fig. 9.3: Direct mixing type of Heat Exchanger
(ii) Classification based on compactness: The definition of compactness with respect to heat
exchangers is quite arbitrary. The ratio of the maximum heat transfer surface area of the heat
exchanger to its volume called as ‘surface area density”, is normally used as the parameter. A
heat exchanger having a surface area density greater than 700 m 2 / m3 is arbitrarily referred to
as a compact heat exchanger. Automobile radiators having an area density of the order of
1100 m2 / m3 and the glass ceramic heat exchangers used in vehicular gas-turbine engines
having a surface area density of 6600 m 2 / m3 are compact heat exchangers. The human
lungs, with an area density of about 20,000 m 2 / m3 , are the most compact heat and mass
exchanger. The very fine matrix regenerator used in sterling engine has an area density
approaching that of the human lung.
The compactness of a heat exchanger becomes very important when heat
exchangers are to be employed in automobiles, marine applications, aircraft, space vehicles,
cryogenic systems, and refrigeration and air conditioning units.
(iv) Classification based on flow arrangement: Heat exchangers are also classified as (a)
parallel flow, (b) counter flow, (c) cross flow and (d) multi-pass flow.
Parallel flow arrangement: In parallel flow arrangement the hot and cold fluids enter the
same end of the heat exchanger, flow through in the same direction and leave together at the
382
other end as illustrated in Fig. 9.4(a). The temperature distribution along the length of the
exchanger is shown in Fig. 9.4(b) In this figure Thi and Tho are the inlet and exit temperatures
Tho
Hot fluid out at Tho
Thi
Fig 9.4 (b) Temperature
Tho distribution in a parallel flow
T exchanger
Tco
of the hot fluid and Tci and Tco are the inlet and exit temperatures of the cold fluid. It is
obvious from the temperature distribution shown that for parallel flow arrangement T co
cannot exceed Tho in order to satisfy the second law of thermodynamics and in practice it is
necessary that there shall be a finite temperature difference between the two fluids in order to
have a finite amount of heat transfer.Hence Tco is always less than Tho for a parallel flow heat
exchanger.The temperature difference between the two fluids is quite large at one end of the
exchanger and decreases considerably as we proceed towards the other end.
Counter flow arrangement:The schematic and temperature distribution for a counter flow
arrangement are shown in Fig.9.5(a) and Fig. 9.5 (b) respectively. In this type of exchanger
both the fluids flow parallel to, but opposite to each other. It can be seen from Fig. 9.5(b) that
383
Tco has to less than Thi to satisfy the second law. However it is possible to have T co > Tho in a
counter flow arrangement.
Tho
Hot fluid out at Tho
Thi
Fig 9.5 (b) Temperature
Tho distribution in a counter flow
Tco
T exchanger
Tci
Cross flow heat exchanger: In this type the two fluids flow through the exchanger at right
angles to each other. The temperature distribution of the two fluids depend on whether each
fluid is mixed or unmixed as it flows through the exchanger. Figure 9.6 illustrates the
direction of flow of the two fluids as well as the temperature distribution for an
384
Cold fluid
in
y
Hot
Fluid
in x
Th = Th (x,y)
Tc = Tc(x,y)
aa
Hot fluid
Outlet
Temperature
distribution
Unmixed-unmixed cross flow heat exchanger. Thermodynamically, the effectiveness for the
cross flow heat exchanger falls in between that for the counter flow and parallel flow
arrangements. The largest temperature difference exists at the corner of the entering hot and
cold fluids such as point a in Fig. 9.6.In a cross flow heat exchanger, a fluid stream is
considered unmixed when it passes through individual flow channels or tubes with no fluid-
mixing between adjacent flow channels.
385
Multipass flow: The multi pass flow arrangement is frequently used in heat exchanger
design as multipassing increases the overall effectiveness individual effectiveness. Fig.9.7
shows some typical arrangements of multipass flow heat exchangers. Fig.9.7(a) is a one shell
pass two tube pass heat exchanger, where as Fig.9.7(b) is a two shell pass four tube pass heat
exchanger.
(a) (b)
Fig.9.7: Multipass flow arrangement:- (a) one shell pass two tube pass arrangement;
(b) Two shell pass four tube pass arrangement.
9.3. Overall Heat Transfer Coefficient and Fouling Factor: In the thermal analysis of heat
exchangers,various thermal resistances in the path of the heat flow from the hot fluid to the
cold fluid are combined in the form of overall heat transfer coefficient, U.For example, for a
double pipe heat exchanger, the total thermal resistance is the sum of the individual components;
i.e., resistance of the inside flow, the conduction resistance in the tube material, and the outside
convective resistance, given by
(A o − Ai)
A ln = ----------------- …………………………………… (9.2d)
ln (A o / Ai)
and hi and ho are inside and outside heat transfer coefficients respectively.
The total heat transfer resistance can be defined in terms of overall heat transfer
coefficient based on either outer or inner areas, as long as the basis is clearly spelled out.
386
For example, based on outer area, we have
1 1 1
Rtotal = +
Aihi kA
+ Aoho
.............................
(9.3)
which after simplifying yields the overall heat transfer coefficient based on inner and
outer areas, respectively as
1
Ui = --------------------------------------------------------------- ……………….(9.4a)
(1/hi)+ [di/(2k)] ln (do/di) + (di/do)( (1/hi)
1
Uo = --------------------------------------------------------------- ……………….(9.4b)
(do/di)(1/hi)+ [d0/(2k)] ln (do/di) + (1/ho)
We note from the above equations that if the wall thickness is negligible Di ≈ Do .For example, in
thin tube heat exchangers or the thermal conductivity of the tube material is very high, the
conduction resistance through the tube may be neglected in Equations(9.4a) and (9.4b) to give
1 1 1 1
≈ = + ………………………………..(9.5)
Ui Uo hi ho
The convection coefficients for the inlet and outlet side of the heat exchanger tube can be
estimated using empirical correlations appropriate for the flow geometry and conditions. During
normal heat exchanger operation, surfaces are subjected to fouling by fluid impurities, rust
formation, and scale depositions, which can markedly increase the resistance to heat transfer
between the fluids. For such situations, one would add the fouling resistance (inside and/or
outside-side) to Equation (9.4a and 9.4 b) to give
1
Ui = -------------------------------------------------------------------------- ………………….(9.6a)
(1/hi) + Fi + [di/(2k)] ln (do/di) +(di/do)Fo + (di/do) (1/ho)
and
1
Uo = ------------------------------------------------------------------------ ………………….(9.6b)
(do/di)(1/hi) + (do/di)Fi + [d0/(2k)] ln (do/di) + Fo + 1/ho
Where di = inside diameter of the inner tube, do = outside diameter of the inner tube,
Fi = Fouling factor for the inside surface of the inner tube,and F o = Fouling factor for the
outside surface of the inner tube.Values of the fouling factor have been determined
experimentally for a number of situations by conducting tests with clean and fouled surfaces.
Some representative values obatained after running the heat exchangers for long periods of
time are given in Table 9.1.
387
Table 9.1 : Representative values of the fouling factor
dQ = U ΔT dA ………………………………..(9.7)
where ΔT = (Th – Tc) is the temperature difference between the bulk mean temperatures of
the two fluids on the two sides of the area dA.
Hot Fluid at Th dA
dQ
Cold fluid at Tc
The total heat transfer rate for the heat exchanger is obtained by integrating Eq. (9.7).
A
Hence Q = ∫ U ΔT dA
0
The mean temperature difference ΔTm for a heat exchanger is defined such that
Q = UA ΔTm ………………………………………...(9.9)
388
ΔTm = (1 / A) ∫ ΔT dA ……………………………(9.10)
0
For a circular tube the surface area of the heat exchanger can either be inside surface area A i
or the outside surface area A0. In that case Eq. (9.10) and (9.11) has to be modified as
follows:
Q = U0 A0 ΔTm ………………………………………...(9.11a)
Or Q = Ui Ai ΔTm ………………………………………...(9.11b)
Where Ui and U0 are the overall heat transfer coefficients based on the outside surface area
and inside surface area respectively.
9.4.1.Mean temperature difference for a parallel flow heat exchanger: The form of the mean
temperature difference, ΔTm , may be determined by applying an energy balance to differential
control volumes (elements) in the hot and cold fluids. As shown in Fig. 9.9, for the case of
parallel flow arrangement, each element is of length dx and the heat transfer surface area is
dA.Energy balance equation for a surface area dA can be written as follows:
Thi
Tho
Tco
Tci
0 A
Area
The heat transfer across the surface area dA can be expressed in terms of the overall heat transfer
coefficient as:
dQ = UdA dT ……………………………….(9.12)
where dT = Th – Tc is the local temperature difference between the hot and cold fluids.
389
From Eq.(9.11) we have dTh = − dQ / Ch and dTc = dQ / Cc.
1 1
Hence dT h – dTc = − U{ + } dA (Th – Tc)
Ch Cc
d (Th−Tc) 1 1
Or =−U { + }dA ………………………………….
( Th−Tc ) Ch Cc
(9.13)
( Tho−Tco ) A
d (Th−Tc ) 1 1
∫ =∫ −U { + }dA
( Th−Tc ) 0 Ch Cc
…………………………. (9.14)
(Thi−Tci)
¿ ln
(Tho−Tco)
( Thi−Tci)
=−U A
1
Ch
+ {
1
Cc } . ……………………………… (9.15)
1 (Tco−Tci)
Now Q = C c (Tco – Tci) or =
Cc Q
------------------------------- (9.16a)
1 (Thi−Tho)
Similarly =
Ch Q
------------------------------- (9.16b)
(Tho−Tco) −UA
ln = { ( Thi−Tho )−(Tco−Tci)}
(Thi−Tci) Q
Or Q = UA ∆T m ………………………(9.17)
∆
∆
Where ∆T m = (¿−∆ Ti) …………………………
ln
(¿ ¿ ∆ Ti)
(9.18a)
390
It can be seen from Eq.(9.18a) that ∆T m is logarithmic mean of ∆ ¿∧∆ Ti and therefore it is
normally referred to as “Logarithmic Mean Temperature Difference (LMTD in short”.
For a counterflow heat exchanger, the energy balance equations for an elemental area dA is given
by
The heat transfer across the surface area dA can be expressed in terms of the overall heat transfer
coefficient as:
dQ = UdA dT ……………………………….(9.12)
where dT = Th – Tc is the local temperature difference between the hot and cold fluids.
391
Or d(T h – Tc) / (Th – Tc) = ─ U{1/Ch ─ 1/Cc} dA
Substituting for 1/Ch and 1/Cc from Eq.(9.16a) and (9.16b) and rearranging we get
(Tho−Tco) −UA
ln = { ( Thi−Tho )−(Tco−Tci)}
(Thi−Tci) Q
Or Q = UA {( Tho ─ Tci) ─ {(Thi ─ Tco)} / ln{(Tho ─ Tci) / (Thi ─ Tco)} ………………...(9.22)
Or Q = UA ∆T m
∆
∆
Where ∆T m = (¿−∆ Ti) …………………………
ln
(¿ ¿ ∆ Ti)
(9.23a)
The expression for ∆Tm for counter flow [Eq.(9.23)] is identical with that for parallel flow except
that the the symbols ∆Ti and ∆To represent different quantities in each case.
Case (i): When one of the fluids is undergoing a change of phase in the heat exchanger:
This situation may arise in the case of a boiler or a condenser. In the boiler the cold fluid will
be undergoing the change of phase at constant pressure so that T co = Tci as shown in Fig 9.10(a).
In this case the mean temperature difference is same both for parallel flow and counter flow
arrangements.Similarly in a condenser the hot fluid will be undergoing a change of phase so that
Tho = Thi as shown in Fig 9.10(b) and hence in this case also the mean temperature difference is
same both for parallel flow and counter flow arrangements.
392
Fig.9.10(a): Mean temperature difference when Fig.9.10(b) Meantemperature difference when
cold fluid is undergoing a change of phase hot fluid is undergoing a change of phase
Case (ii ): When ∆Ti = ∆To: This situation arises only for counterflow arrangement. When ∆Ti =
∆To , Eq. (9.23a) gives
∆Tm = 0 / 0 (Indeterminate).
Thus the mean temperature difference is equal to the difference in temperature between the hot
and cold side at either end of the heat exchanger. It can be shown that in this case both T h and Tc
vary linearly with the same slope.Consequently the value of ∆T is the same all along the length of
the heat exchanger.
9.4.4.Mean temperature difference for a cross flow heat exchanger: In both parallel flow
and counter flow arrangements, the temperature on both sides of the wall separating the two
fluids vary along the length of the heat exchanger,x and therefore are functions of a single
variable. That is both Th and Tc are functions of x only. This is not the case for cross flow
heat exchanger.
393
Case(i): consider a single pass cross flow situation in which both fluids are unmixed (see
Fig.9.6). Here the area A across which the heat transfer takes place is in the plane of the
paper with two fluids flowing on either side, one in the x-direction and the other in the y
direction. It is obvious that both Th and Tc are functions of x and y and the exit profile
temperatures are not uniform.
Consider an elemental surface area dA (dA = dx dy). The rate of heat transfer across this area
is given by
dQ = U(Th ─ Tc)dA = U(Th ─ Tc) dx dy ………………..(9.24)
The determination of ΔTm therefore involves double integration and is rather complicated.
Case(ii): If mixing occurs on one side or on both sides the situation will be different from
that in case(i).By mixing we mean a secondary flow transverse to the flow direction takes
place and this permits a fluid to mix with itself and consequently evens out the temperature
of the fluid in the transverse direction. Therefore if the cold fluid is mixed and the hot fluid is
unmixed and if we assume the cold fluid is flowing in x direction and hot fluid is flowing in
y direction then Tc = f(x) and Th = f(x,y).
Case(iii): If both the fluids are mixed then Tc = f(x) and Th = f(y).
The computation for the mean temperature difference for the three cases of cross flow
have ben done by Bowman, Mueller and Nagle.They have represented their results in the
form of a correction factor,F which is defined as follows:
ΔTm
i.e. F = --------------------------- ………………………………………………(9.27)
(ΔTm)counter flow
Thus the rate of heat transfer for a cross flow heat exchanger can be written as :
F is expressed as a function of two parameters R and P where R and P are given by the
following expressions:
394
R = --------------- and P = ----------------
(t2 ─ t1) (T1 ─ t1)
The parameter R is the ratio of the change in temperatures of the two fluids which is also
equal to the ratio of the heat capacities of the two fluids. P is the ratio of the change in
temperature of one of the fluids to the difference of inlet temperatures of the two fluids.By
definition it follows that R ≥ 0 and 0 ≤ P ≤ 1. Foa cases (i) and (iii) it is immaterial whether T
stands for temperature for hot fluid or cold fluid , but for case (ii) where one fluid is mixed
and the other unmixed,T stands for the temperature of the mixed fluid whether it is hot fluid
or cold fluid and t stands for the temperature of the unmixed fluid
Example 9.1: Water at 25 0 C and a velocity of 1.5 m/s enters a brass condenser tube 6 m long, 1.34
cm ID, 1.58 cm OD and k = 110 W/(m-K). Steam is condensing on the outer surface of the tube with
a heat transfer coefficient of 12,000 W/(m2 – K). The fouling factors for the inner and outer surfaces
are both equal to 0.00018 (m2 – K) / W. Calculate the overall heat transfer coefficient based on (i) the
inside surface area and (ii) the outside surface area.
Solution:
395
ho= 12000 W/m2k
Fo= 0.00018 m2-k/W
k=110 W/m-k
di = 0.0134m
do= 0.0158m
(i) Overall heat transfer coefficient based on inside surface area is given by
1
Ui = -----------------------------------------------------------------------------------
(1/hi )+Fi +[di/2k]loge(do/di)+(di/do)Fo + (di/do)(1/ho)
Since Red > 2300, flow is turbulent.Assuming that the flow is hydrodynamically and
thermally developed we have
1
Ui = ------------------------------------------------------------------------------------------------------
396
(1/6289.5)+0.00018+[0.0134/2x110]loge(0.0158/0.0134)+(0.0134/0.0158)x0.00018
Ui = 1747 W/(m2-K)
(ii) Ui Ai = Uo Ao
Hence Ui π di L = Uo π do L
Or Uo = Ui (di/do)
= 1747 x (0.0134/0.0158)
= 1481.6 W/(m2-K)
Example 9.2: In a heat exchanger hot fluid enters at 60 0 C and leaves at 48 0 C, where as the cold
fluid enters at 35 0 C and leaves at 44 0 C. Calculate the mean temperature difference for (a) parallel
flow, (b) counter flow, (c) single pass cross flow (both fluids unmixed), (d) single pass cross flow
(hot side fluid mixed, cold side fluid unmixed), and (e) single pass cross flow (hot side fluid unmixed,
cold side fluid mixed).
To find: (a) MTD for parallel flow ; (b) MTD for counter flow; (c) MTD for single pass cross flow
with both fluids mixed; (d) single path cross flow with hot fluid mixed and cold fluid unmixed.
Tco
397
For parallel flow, MTD = ------------------- = ---------------- = 11.5 0C
ln(ΔTi / ΔTo) ln (25/4)
(c) MTD for cross flow with both fluids mixed: MTD = F (MTD)counter flow.
From chart , F = 0.90. Heat flux = qx = ─ k dT/dx = ─ 100 * 2000 = −200,000 W/m2
L − ve sign for qx indicates that heat transfer is taking place in
Hence MTD = 0.90 x 14.45 =the − ve0Cdirection of x.
13.01
(d) MTD for cross flow with cold fluid mixed and hot fluid unmixed: Suface in contact
with fluid at To
andforsurface
Since cold fluid is mixed, T stands for temperature of cold fluid and t stands heat of hot
temperature
0 0 0 transfer
fluid. Therefore T1 = Tci = 35 C , T2 = Tco = 44 C, t1 = Thi = 60 C and t2 = Tho = 48 0C.
(a) Force balance on a condensate (b) Force balance oncoefficient
a vapour element
ho
element
(35 ─ 44) (48 ─ 35)at the same distance x from top
Therfore R = ---------------- = 0.75 ; P = --------------- = 0.52
(48 ─60) (35 – 60)
Example 9.3: A simple heat exchanger consisting of two concentric flow passages is used for heating
1110 kg/h of oil (Cp = 2.1 kJ/kg-K) from a temperature of 27 0 C to 49 0C.The oil flows through the
398
inner pipe made of copper (k = 350 W/m-K) with2.54 cm ID and 2.86 cm OD. The surface heat
transfer coefficient on the oil side is 635 W/m 2 –K. The oil is heated by water supplied at a rate of
390
kg/h and at an inlet temperature of 93 0 C. The water side heat transfer coefficient is 1270 W/m2-K.
the fouling factors o the oil side and water side are 0.0001 and 0.0004 m 2-K/W respectively. What is
the length of the heat exchanger required for (i) parallel flow, (ii) counter flow arrangement.
Cold fluid : oil mc = (1110 /3600) = 0.3055 kg/s; Cpc = 2100 J/(kg-K) ; Tci = 27 0C ;
Hot fluid : water ; mh = 390/3600 = 0.1083 kg/s; Cph = 4200 J/kg-K ;Thi = 93oC ;
ho, Fo
water
oil hi , Fi do
di
Solution: Overall heat transfer coefficient based on outside area of the inner tube is given by
1
Uo = ----------------------------------------------------------------------------------------------------
(do/di) (1/hi) + (do/di) Fi + (do/2k) loge(do/di) + Fo + (1/ho)
1
= ----------------------------------------------------------------------------------------------------------------------------
(0.0286/0.0254)(1/635)+(0.0286/0.0254)0.0001+(0.0286/700)loge(0.0286/0.0254)+0.0004+ (1/1270)
= 325 W/(m2-K)
399
Heat balance equation can be written as
Q = Uo π do L π ∆Tm
(ii) Counter flow arrangement : In this case ∆Ti = Thi – Tco = 93 – 49 = 44 0C.
9.5.1. Limitations of Mean Temperature Difference method:- There are generally two types of
problems encountered with heat exchangers. The first type (known as the design type) is the design of
heat exchanger for a given heat load i.e.,for the two fluids, the flow rates, the inlet and exit
temperatures are specified and the heat transfer area required for the heat exchanger is to be
calculated. This type of problem has been solved in Example 9.3 using the mean temperature
difference method. In the second type of problem(known as the rating problem of heat exchanger), for
a given heat exchanger and the two given fluids , only the flow rates and the inlet tempetratures of the
two fluids are specified and the exit temperatures of the two fluids have to be determined.In the mean
temperature difference method,for this type of problem, the exit temperatures of the two fluids have
to be determined by solving two simulataneous equations for both parallel flow and counter flow
arrangements, where as for cross flow and multi pass arrangements the exit temperatures have to be
determined using a trial and error method as illustrated below:
Case (i):Solution for Parallel Flow arrangement: From Eq. (9.15) we have
400
From heat balance equation we have
The RHS of both the equations 9.29(a) and 9.29(b) can bedetermined from the given information and
hence these two rquations can be solved for T ho and Tco
(Tho ─ Tci)
-------------- = exp {─ UA [(1 / C h) ─ (1 / Cc)]
(Thi ─ Tco)
or
Tho + exp {─ UA [(1 / Ch) ─ (1 / Cc)] Tco = exp {─ UA [(1 / Ch) ─ (1 / Cc)] Thi + Tci ………9.29(c)
Eq.9.29(b) holds good for counter flow arrangement also. Hence Eq.9.29(b) and 9.29(c) can be solved
simultaneously to get Tho and Tco.
Case (iii): Solution for crossflow and multi pass arrangements: Unfortunately the solution for these
type of heat exchangers are not as straight forward as the above two types ,as the mean temperature
difference involves the correction factor, F which requires the knowledge of the exit temperatures.
Hence the problem has to be solved by trial and error method[ ] as shown below.
Calculate R ,assume a value for F and read the value of P from the chart for the particular
arrangement.Knowing P, calculate T ho, Tco and Q . Then calculate ΔT m for counterflow arrangement
and check for theassumed value of F by calculating F from the equation
Q
F = -------------------------
UA (ΔTm)counter flow
The procedure is repeated until the solution converges. The method is illustrated for all the three cases
mentioned above with the following example.
Example 9.4. In an oil to water heat exchanger, the oil[C p = 2000 J/(kg-k)] enters the exchanger at
1000C with aheat capacity rate of 3700 W/K. Water(Cp= 4190 J/kg-K) enters the exchanger at 150C
with a mass flow rate of 0.6 kg/s.Determine the exit temperatures of the two fluids for (i) parallel flow
arrangement, (ii) counterflow arrangement and (iii)on shell pass two tube pass arrangement.Assume
the total surface area of the heat exchanger is 10 m 2 and the overall heat transfer coefficient is 500
W/(m2-K)
Cold Fluid is water with : Tci = 150C ; mc = 0.6 kg/s ; Cpc = 4190 J/(kg-K)
U = 500 W/(m2 – K) ; A = 10 m2
Solution: (a) Parallel flow arrangement: Cc = mc Cpc = 0.6 x 4190 = 2514 W/K
401
From Eq.9.29(a) we have Tho ─ Tco = (100 – 15) exp {─ 500 x 10 (1 / 3700 + 1 / 2514)}
Solving for Tho and Tco from Eq.(i) and Eq.(ii) we get
Tho + exp {─ 500x10 [(1 / 3700) ─ (1 / 2514)} Tco = exp {─ 500x10 [(1 / 3700) ─ (1 / 2514)}x 100
+ 15
Solving for Tho and Tco from (ii) and (iii) we get Tho = 57.6 0C ,Tco = 77.470 C and
Trial 1: Assume F = 0.8 (This generally is a mimum value for good design)
Or t 2 = 66.0 0C = Tho
402
ln {(100 – 66.0) / (66 – 15)}
Q 128214
Hence F = ------------------------ = --------------------- = 0.612.
U A (ΔTm)counter flow 500 x 10 x 41.93
The calculated value of F is about 31 % lower than the assumed value. Hence further iteration has to
be carried out. The Table 2 shows the difference between the assumed value of F and calculated
value in % for first 16 iterations.
It can be seen from the above table that the convergence of the iterations is slow and F continues to
fluctuate, but it will converge. This rating problem can be guaranteed convergent using the Newton-
Raphson method (i.e. the solution of f(x) = 0 where function f is non linear in the present case).
The above example illustrates the limitations of the mean temperature difference method for solving
the rating problem of a given heat exchanger. Though the method is straight forward for parallel flow
and counterflow arrangement, but it requires solving two simultaneous equations to find T ho and
Tco..For crossflow and multi pass type heat exchangers the method involves a trial and error solution
wherein the solution converges very slowly iterating on F.This problem can be solved relatively
straight forward by effectiveness-NTU method.
(i) Capacity ratio,C: It is defined as the ratio of minimum heat capacity to the maximum capacity. i.e.
403
Cmin
C = ---------- …………………………………………………………….(9.30)
Cmax
Cc (Thi – Tho)
If Cc < Ch then C = ------------- = -------------------- ……………………………………..(9.31a)
Ch (Tco – Tci)
Ch (Tco – Tci)
If Cc > Ch then C = ------------- = -------------------- ……………………………………..(9.31a)
Cc (Thi – Tho)
(ii) Number of transfer units, NTU:
UA
NTU = --------------- …………………………………..(9.32)
C min
UA
If Cc < Ch, then NTU = ----------------- …………………………………(9.33a)
Cc
UA
If Cc > Ch, then NTU = ----------------- …………………………………(9.33b)
Ch
(iii) Effectiveness, ε: This defined as the ratio of the actual rate of heat transfer in the exchanger to
the maximumpossible rate of heat transfer.
Exercise Problems:
404
Water at
75 0C & Di = 22 mm D0 = 27 mm
0.5 m/s
Determine the overall heat transfer coefficient based on the outside surface area
of the tube. Compare the thermal resistances due to convection, tube-wall
conduction and fouling and make comments.
9.2. A one-shell pass, two-tube pass exchanger is to be designed to heat 0.5 kg/s of
water entering the shell side at 10 0 C. The hot fluid oil enters the tube at 80 0 C
with a mass flow rate of 0.3 kg/s and leaves the exchanger at 30 0 C. The overall
heat transfer coefficient is 250 W/m 2-K. Assuming the specific heat of oil to be 2
kJ/kg-K, calculate the surface area of the heat exchanger required.
9.3. A single pass cross flow heat exchanger uses hot gases (mixed) to heat water
(unmixed) from 30 0 C to 80 0 C at a rate of 3 kg/s.The exhaust gases, having
thermo-physical properties similar to air enter and leave the exchanger at 225
and 100 0 C respectively. If the overall heat transfer coefficient is 200 W/m 2-K,
determine the required surface area of the exchanger.
9.4. A two-shell pass, four-tube pass heat exchanger is used to heat water with oil.
Water enters the tubes at a flow rate of 2 kg/s and at 20 0 C and leaves at 80 0 C.
Oil enters the shell side at 140 0 C and leaves at 90 0 C. If the overall heat
transfer coefficient is 300 W/m2-K, calculate the heat transfer area required.
9.5. A shell and tube heat exchanger is to be designed for heating water from 25 0 C
to 50 0 C with the help of steam condensing at atmospheric pressure. The water
flows through the tubes (2.5 cm ID, 2.9 cm OD and 2 m long) and the steam
condenses on the outside of the tubes. Calculate the number of tubes required if
the water flow rate is 500 kg/min and the individual heat transfer coefficients on
405
the steam and water side are 8000 and 3000 W/ m 2-K respectively. Neglect all
other resistances.
9.6. Show that for counter flow heat exchanger with capacity ratio C = 1, the
effectiveness is given by
ε = NTU / (1 + NTU)
What is the maximum possible effectiveness if the arrangement is (i) parallel flow
and (ii) counter flow?
9.8. Calculate the exit temperature of the hot fluid and inlet temperature of the cold
fluid for a counter flow heat exchanger having the following specifications.
Mass flow rate of hot fluid = 3 kg/s.
Mass flow rate of cold fluid = 0.75 kg/s.
Cp for hot fluid = 1.05 kJ/kg-K.
Cp for cold fluid = 4.2 kJ/kg-K.
Inlet temperature of hot fluid = 500 0 C.
Exit temperature of cold fluid = 85 0 C.
Overall heat transfer coefficient = 450 W/m2-K.
Total surface area of the exchanger = 1 m2.
9.9. In a gas turbine power plant, heat is transferred in an exchanger from the hot
gases leaving the turbine to the air leaving the compressor. The air flow rate is
5000 kg/h and the fuel-air ratio is 0.015 kg/kg. The inlet temperatures on the air
side and the gas side are 170 0 C and 450 0 C respectively. The overall heat
transfer coefficient for the exchanger is 52 W/m 2-K and the surface area of the
exchanger is 50 m 2. If the arrangement is cross flow with both fluids unmixed
determine the exit temperatures of both the fluids and the rate of heat transfer.
Take the specific heats of both the fluids as 1.05 kJ/kg-K.
9.10. A concentric-tube heat exchanger operates on the counter flow mode. The fluid
flowing in the annular space enters the exchanger at 20 0 and leaves at 70 0 C.
The fluid flowing through the inner tube enters at 110 0 C and leaves at 65 0 C.
The length of the exchanger is 30 m. It is desired to increase the outlet
temperature of the cold fluid to 80 0 C by increasing only the length while
maintaining the same mass flow rates, inlet temperatures and tube diameters.
Make any justifiable assumption and calculate the new length.
406
9.11. It is proposed to cool 1000 kg/h of oil from150 0 C to 50 0 C in a heat exchanger
using 1667 kg/h of water at an inlet temperature of 30 0 C. Calculate the surface
required assuming a single pass cross flow arrangement in which the oil is mixed
and the water unmixed. Assume Cp for oil to be 2.087 kJ/kg-K and the overall
heat transfer coefficient to be 550 W/m 2-K. Solve the problem by the mean
temperature difference method as well as by the ε – NTU method.
9.12. Two identical counter flow heat exchangers are available. Water at the rate of
5000 kg/h and at 30 0 C is to be heated by cooling an oil (C p = 2.1 kJ/kg-K) at 90
0
C. The oil flow rate is 2000 kg/h. The heat transfer area is 3 m 2. From the point
of view of maximizing the heat transfer rate, which of the following is the best
arrangement?
(a) Both the fluids flow in series.
(b)The oil flow is split up equally between the two exchangers, while the water flows
in series.
(c) Both oil and water flows are split up equally in both the exchangers.
9.13. A counter flow double pipe heat exchanger is used to heat 1.25 kg/s of water
from 35 0 C to 80 0 C by cooling an oil (Cp = 2.0 kJ/kg-K) from 150 0 C to 85 0 C.
The overall heat transfer coefficient is 850 W/m 2-K. A similar arrangement is to
be built at another location, but it is desired to compare the performance of the
single counter flow heat exchanger with two smaller counter flow heat
exchangers connected in series on the water side and in parallel on the oil side as
shown in Fig. P 9.16. The oil flow is split equally between the two exchangers and
it may be assumed that the overall heat transfer coefficient for the smaller
exchangers is the same as for the large exchanger. If the smaller exchanger
costs 20 % more per unit surface area, which would be the most economical
arrangement – the one large exchanger or the two equal-sized small exchangers?
9.1 Solution:
407
ho
Fo= 0.0002 m2-k/W
Di = 0.022m
Do= 0.027m
k=45 W/m-k
Air at T∞ = 15oC
Uav = 20m/s
To find ho:
408
Red = U∞do/ ν = 20 x 0.027/14.61 x 10 − 6 = 36961
Nud = hodo/k = [ 0.4 Red 0.5 + 0.06 Red 0.6667] x Pr0.4 x ( µ ∞/µ w )0.25
= 0.096 k/W
Total thermal resistance = ∑R = 4.01 x 10-3 + 0.0579 + 7.24 x 10-3 + 2.36 x 10-3
+ 0.096
= 0.1675 k/W
Uo Ao = 1/ ∑R
Uo = 1 / Ao ∑R
= 1 / ( π x 0.027 x 0.1675)
= 70.4 W/(m2-K)
409
Rfi Resistance due to fouling at 0.0579
the inside surface
R Resistance of the tube wall 7.24 x 10-3
for conduction
Rfo Resistance due to fouling at 2.36 x 10-3
the outside surface
Rco Thermal resistance for 96 x 10-3
convection at the outside
surface
The comparison shows that the thermal resistance for convection heat transfer from the outer
surface of the tube due to air is very large compared to the other resistances i.e,
∑R ≈ Rco
9.2 Solution:
U = 250 W/(m2-K)
= 30000 W
Also Q = mc x Cpc (Tco – Tci)
= 24.3oC
(∆Tm)c.f = (80 – 24.3) – (30 – 10) / loge [(80 – 24.3) / (30 – 10)]
= 34.35oC
410
T1 = 10 oC ; T2 = 24.3 oC ; t1 = 80 oC ; t2 = 30 oC
R = T1 – T2 / t2 – t1 = 10 – 24.3 / 30 – 80 = 0.29
P = t2 – t1 / T1 – t1 = 30 – 80 / 10 – 80 = 0.714
Q = U A (∆Tm)
9.3 Solution:
Cross flow HE : Cold water (Unmixed) : mc = 3 kg/s ; Cpc = 4200 J/kg-K ;Tci = 30oC ;
Tco = 80oC ;
Hot gases (Mixed) : Cph = 1005 J/kg-K ;Thi = 225oC ; Tho = 100oC ; U = 200 W/(m2/K)
Also
P = t2 – t1 / T1 – t1 = 80 – 30 / 225 – 30 = 0.256
(∆Tm)c.f = (225 – 80) – (100 – 30) / loge [(225 – 80) / (100 – 30)] = 54.62oC
9.4 Solution:
411
Q = mc x Cpc (Tco – Tci) = 2 x 4200 (80 – 20 ) = 504000 W
P = t2 – t1 / T1 – t1 = 80 – 20 / 140 – 20 = 0.5
R = T1 – T2 / t2 – t1 = 140 – 90 / 80 – 20 = 0.83
(∆Tm)c.f = (140 – 80) – (90 – 20) / loge [(140 – 80) / (90 – 20)] = 64.87oC
9.5 Solution:
Shell and Tube HE : Cold fluid : Water : mc = 8.33 kg/s ;Tci = 25oC ; Tco = 50oC ;
Hence Thi = Tho = Tsat at atmospheric pressure = 99.6o C (from steam tables)
412
1
= ------------------------------------------------------------------ = 1954 W/(m2-K)
(0.029 / 0.025) (1/3000) + 0 + ( 1 / 8000 )
Since Thi = Tho , both parallel flow and counter flow arrangement will give the same
value of (∆Tm).
Hence (∆Tm) = (99.6 – 25) – (99.6 – 50 ) / loge [(99.6 – 25) / (99.6 – 50 ) ] = 61.2 oC
9.6 Solution:
1 - ℮[ - (1 – c ) NTU ]
ε= ------------------------------
1 - c℮[ - (1 – c ) NTU]
1 - ℮0
ε= ------------------ = 0 / 0 = indeterminate.
1 – 1 x ℮0
d/dc [1 - e[ - (1 – c ) NTU ] ]
ε= Ltc--> 1 -------------------------------------
d/dc [1 - ce[ - (1 – c ) NTU] ]
0 – NTU ℮ - (1 – c ) NTU
= Ltc--> 1 ------------------------------------------------------------------
0 – {℮ - (1 – c ) NTU x 1 + c (NTU)℮ - (1 – c ) NTU }
413
0 – NTU NTU
ε = --------------------------- = --------------------
0 – { 1 + NTU } 1 + NTU
9.7 Solution:
mh Cph 281.4
Since mh Cph < mc Cpc , hence c = -------------- = ------------ = 0.84
mc Cpc 335.16
1 - e[ - (1 + c ) NTU ]
ε= ------------------------------ ………………………………………………….. (1)
1+c
From (1),
1 x [ 0 + (1+c)e –(1+c)NTU]
[d ε/d(NTU)] = ------------------------------------- = 0
1+c
1 - ℮[ - (1 + c ) ∞ ] 1–0 1 1
εmax = ---------------------- = --------------- = ----------- = --------------- = 0.5435
1+c 1+c 1+c 1.84
414
For a counter flow HE,
1 - ℮[ - (1 – c ) NTU ]
ε = ------------------------------ ……………………………………… (2)
1 - c℮[ - (1 – c ) NTU]
ε = 1 [ from (2) ]
9.8 Solution:
Counter flow HE :
U = 450 W / (m2- K )
UA UA
Also NTU = ------------------ or ---------------------
mc Cpc mh Cph
450 x 1
= -------------- = 0.1428
3150
415
1 - ℮[ - (1 – c ) NTU ]
ε = ------------------------------
1 - c℮[ - (1 – c ) NTU]
NTU 0.1428
Since c = 1, ε = ---------------- = ------------------- = 0.1250
1 + NTU 1.1428
85 – 62.5
Or Tci = ---------------------- = 25.7oC
(1 – 0.125)
Tho = 440.7oC
9.9 Solution:
Thi = 450oC ;
Cold fluid : Air : mc = 1.39 kg/s; Cpc = Cph = 1050 J/kg-K ; Tci = 170oC ;
U = 52 W / (m2- K ) ; A = 50 m2
416
mc Cpc 1459.5
Since mh Cph > mc Cpc , hence c = -------------- = ------------ = 0.986
mh Cph 1480.5
UA 52 x 50
NTU = ------------------ = -------------- = 1.78
mc Cpc 1459.5
ε = 0.6
= 338 oC
mc Cpc
Hence Tho = Thi – ------------- x [ Tco – Tci ]
mh Cph
9.10 Solution:
Counter flow HE :
case (i) Thi = 110o C ; Tho = 65o C ; Tci = 20o C ; Tco = 77oC ; L = 30 m ; U1
417
Q1 mc x Cpc ( 70 – 20 ) 5
Hence, ------ = -------------------------------- = ------
Q2 mc x Cpc ( 80 – 20 ) 6
110 – 65 70 – 20
Hence --------------------- = --------------
110 - (Tho)2 80 – 20
(110 – 70 ) – (65 – 20 )
Hence for case (i) (∆Tm)1 = ----------------------------------------- = 42.45oC
loge [(110 – 70 ) / (65 – 20 )]
(110 – 80 ) – (56 – 20 )
Hence for case (ii) (∆Tm)2 = ----------------------------------------- = 32.9 oC
loge [(110 – 80 ) / (56 – 20 )]
Q1 U1 π d L1 (∆Tm)1 5
hence , --------- = ---------------------- = -------
Q2 U2 π d L2 (∆Tm)2 6
L1 (∆Tm)1 5
Hence --------------------------- = ----------
L2 (∆Tm)2 6
9.11 Solution:
418
Cross flow HE with oil mixed and water unmixed;
Hot fluid : oil : mh = 0.278 kg/s ; Cph = 2087 J/kg-K ;Thi = 150oC ; Tho = 50oC
Cold fluid : water : mc = 0.463 kg/s ; Cpc = 4200 J/kg-K ;Tci = 30oC;
U = 550 W / (m2 – K )
(∆Tm)c.f = (150 – 59.8) – (50 – 30) / loge [(150 – 59.8) / (50 – 30) ]
= 46.6 oC
Q 58019
A = ------------- = --------------- = 2.83 m2
U (∆Tm) 550 x 37.3
mh Cph 580.2
Since mh Cph < mc Cpc , hence c = -------------- = ------------ = 0.29
mc Cpc 1944.6
when mh Cph < mc Cpc ,
419
ε = ------------------ = ---------------- = 0.83
Thi – Tci 150 – 30
9.12. Solution:
Cold: water: mc= 5000/3600 = 1.39 kg/s ; Tci= 30 oC ;Cpc= 4200 J/(kg-K) assumed
Hot : oil : mh = 2000/3600 = 0.555 kg/s ;Cph = 2100 J/(kg-K) ;Thi= 90oC
Tc*
Th*
Tci , mhcph
EXCHANGER 1
Tho
As mcCpc = (5000/3600) x 4200 = 5833.3 J/(s-K)
420
UA 465*3
NTU = = = 1.196
mhC ph 1166.7
1 - exp[ -(1 - C ) NTU ]
e=
1 - C exp[-(1 - C ) NTU ]
1 - exp[ -(1 - 0.2)1.196] 1 - 0.3841
= =
1 - 0.2 exp[ -(1 - 0.2)1.196] 1 - 0.2*0.3841
= 0.667
Case (2) When oil is split up equally between the two heat exchangers:
UA 465*3
NTU = = = 2.4
mh C ph 583.8
1 - exp[-(1 - C ) NTU ]
e=
1 - C exp[ -(1 - C ) NTU ]
1 - exp[ -(1 - 0.1)2.4] 1 - 0.1153
= =
1 - 0.1exp[-(1 - 0.1)2.4] 1 - 0.1*0.1153
= 0.895
Case (3) Both the oil and water flows are split up equally:
Since ε is the highest in case (2), this arrangement is the best for ” maximizing heat transfer”
421
9.13 Solution:
mh/2 mh/2
Tci = 35oC
mc =1.25 kg/s
Tco=80oC
Tc*
Thi Th1* Th2*
EXCHANGER 2
Tho=85oC
U1 = U2 = U = 850 W / (m2 – K )
Q 236250
Hence mh = ------------------------ = ---------------------------- = 1.817 kg/s
Cp [Thi – Tho] 2000 x [150 – 85 ]
mh Cph 3634.6
Since mh Cph < mc Cpc , hence c = -------------- = ------------ = 0.7
mc Cpc 5250
422
Case (ii) When two smaller heat exchangers are used:
To find ‘ε’ , we should know the exit temperatures of hot and cold fluids for at least one
HE. Since UA and (mCp)min is the same for both the exchangers NTU should be same
Since the oil flow is the same in each exchanger and the average exit oil temperature must be
85oC …..it follows that
Th1* + Th 2*
= 85o C........................................(2)
2
Energy balance on the second heat exchanger gives
Equations 1,2 ,3 may be solved for the three unknowns Th1* , Th2* , Tc* . The solutions are as
follows:
Eqn 1can be rearranged after cross multiplying as:
423
Th 2* + 82
Tc =
*
= 0.345Th 2* + 28.3
2.9
From eqn 5………………. Th1 = 170 - Th 2
* *
Or Th 2 - 706Th 2 + 57055 = 0
*2 *
Therefore………. Th 2 = 93 C
* o
150 - 77
e1 = e 2 = = 0.635
Therefore……………. 150 - 33
1817 *1.16
A1 = = 2.48m 2 = A2
Therefore…………….. 850
Therefore total area required to meet the heat load = 2.482*2= 4.92m2
This is more than the 4.276 m2 required in the one larger heat exchanger. In addition the cost
per unit area is greater so that the most economical choice from the heat transfer point of
view would be the single large heat exchanger.
424
CHAPTER: 10
When a body is placed in an enclosure whose walls are at temperatures below that of
the body, the temperature of the body will decrease even if the enclosure is evacuated. This
process by which heat is transferred from a body by virtue of its temperature, without the aid
of any intersecting medium is called “THERMAL RADIATION”. The actual mechanism of
radiation is not yet completely understood. There are at present two theories by which
radiation phenomenon is explained. According to Maxwell’s electromagnetic theory,
Radiation is treated as electromagnetic waves, while Max Planck’s theory treats radiation as
“photons” or “quanta of energy”. Neither theory completely describes all observed
phenomena. It is however known that radiation travels with the speed of light, c (c = 3x10 8
m/s) in a vacuum. This speed is equal to the product of the frequency of the radiation and the
wavelength of this radiation,
OR c = λν ………………………………….(10.1)
From the viewpoint of electromagnetic wave theory, the waves travel at the speed of
light, while from the quantum theory point of view, energy is transported by photons which
travel at the speed of light. Although all the photons have the same velocity, there is always a
distribution of energy among them. The energy associated with a photon, ep = hν where h is
the Planck’s constant equal to 6.6256 x 10-34 Js. The entire energy spectrum can also be
described in terms of the wavelength of radiation.
425
visible to the naked eye, and this is called light rays. The wavelength spectrum of radiation is
illustrated in Fig 10.1
Thermal Radiation
( 0.1 μm to 100 μm )
λ , μm
In the study of radiation transfer, a distinction should be made between bodies “semi-
transparent” to radiation and those which are “opaque”. If the material is semitransparent to
radiation, such as Glass, salt crystals, and gases at elevated temperatures, then the radiation
leaving the body from its outer surfaces results from emissions at all depths within the
material. The emission of radiation for such cases is a BULK or a VOLUMETRIC
PHENOMENON. If the material is opaque to thermal radiation, such as metals, wood, rock
etc. then the radiation emitted by the interior regions of the body cannot reach the surface. In
such cases, the radiation emitted by the body originates from the material at the immediate
vicinity of the surface (i.e. within about 1mm) and the emission is regarded as a SURFACE
PHENOMENON. It should also be noted that a material may behave as a semi transparent
medium for certain temperature ranges, and as opaque for other temperatures. Glass is a
typical example for such behaviour. It is semi transparent to thermal radiation at elevated
temperatures and opaque at intermediate and low temperatures.
Emissive Power (E): The emissive power of a surface is the energy emitted by the surface at
a given temperature per unit time per unit area for the entire wavelength range, from λ=0 to
∞
426
∞
E = ∫Eλdλ ……………………………………………(10.2)
0
Qt Qρ
Qα
Qτ
When a radiant energy strikes a material surface, part of the radiation is reflected, part is
absorbed, and part is transmitted, as shown in Fig. 10.2. Reflectivity (ρ) is defined as the
fraction of energy which is reflected, absorptivity (α) as the fraction absorbed, and
transmissivity (τ) as the fraction transmitted. Thus ρ + α + τ = 1.
Most solid bodies do not transmit thermal radiation, so that for many applied problems,
the transmissivity may be taken as zero. Then
ρ + α = 1 ……………………………………(10.3)
427
Specular Radiation and Diffuse Radiation:
Φ1 Ф2
Ф2 = Φ1
When radiation strikes a surface, two types of reflection phenomena may be observed. If
the angle of incidence is equal to the angle of reflection, the radiation is called Specular.
On the other hand, when an incident beam is distributed uniformly in all directions after
reflection, the radiation is called Diffuse Radiation. The two types of radiation are
depicted in Fig. 10.3. Ordinarily, no real surface is either specular or diffuse. An ordinary
mirror is specular for visible light, but would not necessarily be specular over the entire
wavelength range.A rough surface exhibits diffuse behaviour better than a highly
polished surface. Similarly, a highly polished surface is more specular than a rough
surface.
Black Body:
A body which absorbs all incident radiation falling on it is called a blackbody. For a
blackbody, α = 1, ρ = τ = 0. For a given temperature and wavelength, no other body at the
same temperature and wavelength, can emit more radiation than a blackbody. Blackbody
radiation at any temperature T is the maximum possible emission at that temperature. A
blackbody or ideal radiator is a theoretical concept which sets an upper limit to the
emission f radiation. It is a standard with which the radiation characteristics of other
media are compared.
Emissivity of a Surface (ε): The emissivity of a surface is the ratio of the emissive
power of the surface to the emissive power of a black surface at the same
428
temperature. It is denoted by the symbol ε.
i.e. ε = [E/Eb]T.
This is defined as the total energy leaving the surface per unit time per unit area of the
surface. This definition includes the energy reflected by the surface due to some radiation
falling on it.
Irradiation of a surface(G):
This is defined as the radiant energy falling on a surface per unit time, per unit area of the
surface.
Therefore if E is the emissive power, J is the radiosity, e is the irradiation and r the
reflectivity of a surface, then,
J = E + ρG
For an opaque surface ρ + α = 1 or ρ = (1 – α)
This law states that the emissive power of a blackbody is directly proportional to the
fourth power of the absolute temperature of the body. i.e.
Eb α T 4
Or Eb = σT4 ---------------------------------- (10.5)
where σ is called the Stefan – Boltzmann constant.
In SI units σ = 5.669x10-8 W/(m2-K4).
429
10.2.2 PLANCK’S LAW:
C1
Ebλ = ------------------------------ ………………………..(10.6)
λ 5 [ exp (C2 / λT) – 1]
where C1 and C2 are constants whose values are found from experimental data; C 1 = 3.745 x
10-16 Wm2 and C2 = 1.4388 x 10-2 m-K. λ is the wavelength and T is the absolute temperature
in K.
It can be seen from Eq. 10.6 that at a given temperature, E bλ depends only on λ.
Therefore this vale of λ which gives max value for Ebλ can be obtained by differentiating
Eq(10.6) w.r.t λ and equating it to zero.
C1
Ebλ = -----------------------------
[C2 / (yT)] 5 [ e y – 1]
Or ey(5 – y) = 5
C2/λmT = 4.965
Equation (10.7) is called the Wein’s displacement law. From this equation it can be seen that
the wavelength at which the monochromatic emissive power is a maximum decreases with
increasing temperature. This is also illustrated in Fig 10.4(a). Fig 10.4(b) gives a
430
comparison of monochromatic emissive powers for different surfaces at a particular
temperature for different wavelengths.
Ebλ
Fig. 10.4 (a) Black body emissive power as a function of wave length
and Temperature
431
10.2.4 KIRCHOFF’S LAW:
This law states that the emissivity of a surface is equal to its absorptivity when the
surface is in thermal equilibrium with the surroundings.
Proof: Consider a perfect black enclosure i.e. the one which absorbs all the incident radiation
falling on it (see Fig 10.5). Now let the radiant flux from this enclosure per unit area arriving
at some area be qi W/m2.
Black Enclosure
EA
Sample
qi Aα
Now suppose that a body is placed inside the enclosure and allowed to come to thermal
equilibrium with it. At equilibrium, the energy absorbed by the body must be equal to the
energy emitted; otherwise there would be an energy flow into or out of the body, which
would raise or lower its temperature. At thermal equilibrium we may write
EA = qiA α ………………………………….(10.8)
If we now replace the body in the enclosure with a black body of the same size and shape and
allow it to come to thermal equilibrium with the enclosure,
EbA = qiA……………………………………(10.9)
E/Eb = α
432
But by definition E/Eb = e, the emissivity of the body, so that
ε = α ……………………………………….(10.10)
Equation 10.10 is called Kirchoff’s law and is valid only when the body is in thermal
equilibrium with the surroundings. However, while analyzing radiation problems in practice
we assume that Kirchoff’s law holds good even if the body is not in thermal equilibrium with
the surroundings, as the error involved is not very significant.
Example 10.1: The emission of radiation from a surface can be approximated as blackbody
radiation at 1000K.
Solution: The radiation flux emitted by the blackbody over the wavelength interval 0 – l is
given by
λ
[Eb]0 – l = ∫Ebλdλ
0
The integration required in the above equation has been done numerically and the results are
presented in the form of a table. The table gives the value of D 0-λ where
λ
∫Ebλdλ 1 λ
0
D0-λ = --- -∞----------- = ------- ∫ Ebλ dλ
∫Ebλdλ σ T4 0
0
(a) From Table of Radiation properties, for λT = 5 x 1000 = 5000, D0-λ = 0.6337. This
means that 63.37 % of the total emission occurs below λ = 5 μm.
Example 10.2: The monochromatic emissivity of a surface varies with the wavelength in the
following manner:
433
ελ = 0 for λ < 0.3μm
= 0.9 for 0.3μm < λ < 1μm
= 0 for λ > 1μm
Calculate the heat flux emitted by the surface if it is at a temperature of 1500 K
Solution:
Eλ = ε λ Ebλ
∞ 0.3 μm 1 μm ∞
Therefore E = ∫ ε λ Ebλ = ∫ 0.0 Ebλ dλ + ∫ 0.9 Ebλ dλ + ∫ 0.0 Ebλ dλ
0 0 0.3 μm 1 μm
1 μm 1 μm 0.3 μm
= 0.9 ∫Ebλ dλ = 0.9 [ ∫ Ebλ dλ − ∫ Ebλ dλ ]
0.3 μm 0 0
Example 10.3: Calculate the heat flux emitted due to thermal radiation from a black surface
at 60000 C. At what wavelength is the monochromatic emissive power maximum and what is
the maximum value?
λm = 0.002898/6273 = 4.62x10-7 m
2 π C1
(Ebλ)max = ------------------------------------
λmax [ exp {C2 / (λmaxT)} – 1]
2 x π x 0.596 x 10 − 16
= ---------------------------------------------------------------
(4.62 x 10 − 7) 5 x [ exp{ 0.014387 / 0.002898} – 1]
434
= 1.251 x 10 14 W / m2
Solution:
∞
E ∫ελ Ebλ dλ 1 λ1 λ2 λ3
ε = ------ = -0------------- = ------ [ ε1 ∫ Ebλ dλ + ε2 ∫ Ebλ dλ + ε3 ∫ Ebλ dλ ]
Eb σ T4 σ T4 0 λ1 λ2
Example 10.5: the filament of a light bulb is assumed to emit radiation as a black body at
2400K. if the bulb glass has a transmissivity of 0.90 for radiation in the visible range,
calculate the percentage of the total energy emitted by the filament that reaches the ambient
as visible light.Assume the wave length rabge corresponding to the visible range as0.38 μm
to 0.76 μm.
435
Now λ1T = 0.38 x 2400 = 912. Hence D0-λ1 = 0.0002
Only 3.9 % of the total energy enters the ambient as light. The remaining energy produces
heating.
Assumptions:
(i) The medium does not participate in radiation heat exchange between the two
surfaces.
(ii) The surfaces are flat and are at specified uniform temperatures.
J1 = Eb1
T1 G1 = J2
G2 = J1 J2 = Eb2
T2
Fig: 10.6 Radiation heat exchange between two parallel black surfaces.
Since both surfaces are parallel, flat and infinite, radiosity of surface 1 = irradiation of
surface 2 and vice versa. i.e. J1 = G2 and J2 = G1. Since both the surfaces are black, J1 = Eb1 =
σT14 and J2 = Eb2 = σT24
436
Net radiation leaving A1 = Qr1 = A1(J1 – G1) All this energy will reach A2.
J1
T1, α1, ε1, A1 G1 = J2
G2 = J1 J2
T2, α2 ,ε2, A2
Fig: 10.7 Radiation Heat Exchange Between 2 Parallel Infinite Gray Surfaces.
J1 = E1 + (1-α1)G1 (10.12b)
J2 = E2 + (1-α2)G2 (10.12c)
J1 = G2 (10.12d)
J2 = G1 (10.12e)
437
J1 – (1 – α1)G1 = E1 ……………………………..(4.12f)
Equation (4.12c) with the help of Eqns. (10.12d) and Eqns. (10.12e) can be rewritten as
– (1 – α2)J1 + G1 = E2 (10.12g)
E1 + (1 – α1) E2
J1 = ----------------------------
1 – (1 – α1) (1 – α2)
A1
Q1-2 = -------------------------- [ε1E b1 + (1 – α1) ε2E b2 − ε2E b2 − (1 – α2) ε1E b1]
[1 – (1 – α1) (1 – α2)]
438
[1 – (1 – α1) (1 – α2)]
σ A1 (T14 – T24)
Or Q1-2 = --------------------------- …………………………(10.15)
[ 1 / ε1 + 1 / ε2 − 1 ]
10.4.3 PLANE RADIATION SHIELDS: It is possible to reduce the net radiation heat
exchange between two infinite parallel gray surfaces by introducing a third surface in
between them. If the third surface, known as the radiation shield is assumed to be very thin,
then both sides of this surface can be assumed to be at the same temperature.
Fig.10.7 shows a scheme for radiation heat exchange between two parallel infinite gray
surfaces at two different temperatures T 1 and T2 in presence of a radiation shield at a uniform
temperature, T3.
439
T1, α1= ε1, A1
T3, ε13, A3 = A1=A2
Fig: 10.7 Radiation Heat Exchange Between Two Parallel Infinite Gray
surfaces in presence of a radiation shield
Then
(T14 – T34) (T34 – T24)
--------------- = ------------------
X Y
440
Q1-3 / A1 = Q3-2 / A1 = (Q1-2 / A)1 Rad.Shield = σ { T24 – [{T14 +(X/Y)T24}/(1 + X/Y)] } / X
…………………….(10.17a)
σ [ T14 − T24 ]
= ------------------------ ………………………………(10.18b)
2 [2 / ε − 1]
It can be seen from the above equation that when the emissivities of all surfaces are equal,
the net radiation heat exchange between the surfaces in the presence of single radiation shield
is 50% of the radiation heat exchange between the same two surfaces without the presence of
a radiation shield. This statement can be generalised for N radiation shields as follows:
1
[Q1-2 / A]N shields = --------- [Q1-2 / A] without shield ……………(10.18c)
(N + 1)
Example 10.6: Two parallel infinite grey surfaces of emissivities 0.5 are at temperatures of
400K and 300K. Determine the net radiation heat flux between the two surfaces. Also
determine the reduction in radiation flux when a plane radiation shield having emissivity of
0.5 on both its surfaces is placed between the two grey surfaces. Also determine the steady
state temperature of the shield.
Solution:The radiation flux between two grey surfaces is given by
σ (T1 4 – T2 4)
q = Q1-2 / A = -------------------------- .
[ 1 / ε1 + 1 / ε2 – 1]
5.67 x 10 − 8 x (400 4 – 300 4)
Since ε1 = ε2 = ε = 0.5, we have q = -------------------------------------
(2 / 0.5 – 1)
441
Or q = 330.75 W / m2.
When a radiation shield of same emissivity is placed between two grey surfaces, the
temperature of the shield T3 is given by
T3 = [(T14 + T24) / 2 ] ¼
Hence T3 = [(400 4 + 300 4) / 2] ¼
= 360.3 K
Also, since the emissivities of the plates and shield are equal we have
Example 10.7: Two parallel plates are at temperatures T1 and T2 and have emissivities
ε1 = 0.8 and ε2 = 0.5. A radiation shield having the same emissivity ε 3 on both sides is placed
between the plates. Calculate the emissivity ε3 of the shield in order to reduce the radiation
heat loss from the system to one tenth of that without shield.
Solution: Radiation flux between the two plates without the presence of a radiation shield is
given by
σ [ T14 − T24 ]
or q1-2 = -------------------- …………………………(1)
2.25
When a shield is placed between the plates, the radiation flux is given by
442
T14 + ( x / y )T24 �
�
s�
T14 - �
� 1+ x / y �
( q ) 1radshield =
x
1 1 1 1 1
x= + -1 = + - 1 = 0.25 +
e1 e 3 0.8 e 3 e3
1 1 1 1 1
y= + -1 = + -1 = 1 +
e3 e2 e3 0.5 e3
s �
(1 + x / y )T14 - T14 - ( x / y )T24 �
� �
\ ( q ) 1radshield =
x (1 + x / y )
s ( T14 - T24 )
( q ) 1radshield =
x+ y
Solution:
When a radiation field is placed between thick plates the radiation flux is given by
σ (T14 – T34)
(q)1 Rad. shield = ------------------------- with x = [ 1 / ε1 + 1 / ε13 – 1]
X
443
Therefore T3 = [------------------------------- ] 1/4
1 + (11 / 20.25)
= 746.8 K
Example 10.9: Find an expression for the net radiant flux between two infinite parallel
diffuse grey surfaces at temperatures T1 and T2 degrees Kelvin when an infinite opaque plate
of thickness b and thermal conductivity K is placed between them. Assume that all surfaces
have the same emissivity.
Solution: The schematic for the problem is shown in Fig. E10.9. For steady state heat
transfer we have
σ (T14 – T34)
Now Q1-3 / A = ------------------------ …………………………………………….(1)
(2 / ε – 1)
444
T1, ε
T3, ε
T4, ε
Plate of thickness b &
Conductivity k T2, ε
σ (T44 – T24)
Q4-2 / A = ------------------------ …………………………………………….(3)
(2 / ε – 1)
Example 10.10: Calculate the steady heat flow through the composite slab of Fig P10.10
consisting of two large plane walls with an evacuated space in between. The thicknesses
of the walls are 20 and 30cm, they have thermal conductivities of 1.0 and 0.5 W/m-k and the
emissivities of the surfaces facing each other are 0.5 and 0.4 respectively.
Solution :
445
L1 L2
vacuum
Fig. E10.10: Schematic for example 10.10
From Eq. (a) T1 = 473 – 0.25 q, and from Eq. (c) T2 = 313 + 0.7q.
Solving the above equatin by trial and error method we get q = 139 W / m 2.
446
10.6 VIEW FACTOR OR CONFIGURATION FACTOR:
Solid Angle: The solid angle dw subtended by an elemental area dA surrounding point P with
respect to any other point O in space is defined as the component of the area dA in the
direction OP divided by the square of the distance between O and P. This is illustrated in Fig.
10.9. Solid angle is measured in Steradian (Sr).
n
dA
P
n
φ
dω r
r
dω = dA / r2 dω = dA cos φ / r2
I = dE/dw (10.19a)
E = ∫Idw (10.19b)
447
Where the integration is carried out over all directions encompassed by a hemisphere.
Consider an elemental area dA1 whose total emissive power is E 1. This total radiant
energy emitted by dA1 can be intercepted by a hemisphere as shown in Fig 10.10.
dAn
P
θ
OP = r
φ
O
dφ
dA1
Fig 10.10: Radiation from a differential area dA1 into surrounding hemisphere
centered at dA1.
If I is the intensity of radiation at any point P on the surface of the hemisphere due to
emission by an elemental area dA1 at O, then
� dAn �
E1 = I cos q dw = �
� 1 I cos q �
1 2 �
�r �
p
2p � I cos q �r sin q dq �rdf �
E1 = ��
2
�
1
2 �
q =0 f =0
� r �
Assuming that I1 is same in all directions (Lambert’s Law)
p
2p
E1 = I1 ��
2
cos q sin q dq df
q =0 f =0
E1 = p I1 (10.20)
Eb = πIb (10.21)
448
10.6.2: VIEW FACTOR BETWEEN TWO ELEMENTAL SURFACES.
Consider 2 elemental surfaces of area dA1 and dA2 as shown in Fig. 10.11. Let their
normals n1 and n2 make angles φ1 and φ2 with the line joining the centroid of the two
elemental areas. Let dw2 be the solid angle subtended by dA 2 at dA1 and dw1 be the solid
angle subtended by dA1 at dA2.
n1 n2
φ2
dA2
φ1
dA1
Let I1 be the intensity of radiation from dA1 striking dA2 and I2 be the intensity of radiation
from dA2 striking dA1. If dq1->2 is the radiant energy leaving dA1 and striking dA2 then
dq1�2 = I1dA1 cos f1dw2
dA2 cos f2
Or dq1�2 = I1dA1 cos f1
r2
Radiation energy leaving dA1 = dqr1 = E1dA1
Fraction of energy leaving dA1 and striking dA2 is defined as the view factor of dA 2 with
respect to dA1 and is denoted by dF1-2
449
π r2
Similarly, the view factor of dA1 with respect to dA2 is denoted by dF2-1 and given by
n1
n2
φ2 dA2
r A2
A1 φ1
dA1
450
Hence Q1→2 = ∫ ∫[{ I1 dA1 dA2 cos φ2} / r2].
A1 A2
1
Or F1-2 = ------ [∫ ∫{ I1 dA1 dA2 cos φ2}/ ( πr2)] …………………(10.25a)
A1
1
Similarly F2-1 = ------ [∫ ∫{ I1 dA1 dA2 cos φ2}/ ( πr2)] …………………(10.25b)
A2
Properties of view factor: Consider an enclosure consisting of N zones, each of surface area
Ai (i = 1,2,3 … N). The surface of each zone may be plane, convex or concave. For the
enclosure, the following relations hold good.
4. When there are two surfaces, one surface say A1 is completely enclosed by A2 and
if A1 cannot see itself then, F1-2 = 1 and F2-1 = A1/A2
5. The view factor F1-2 between surfaces A1 and A2 (Fig. 10.13) is equal to the sum of
the view factors F1-3 and F1-4 if the two areas A3 and A4 together make up the area
A2.
451
A2 = A3 + A4
A3
A4
A1
View factors for standard configurations: The determination of view factors has been the
object of considerable research. In cases where the integrals in Eq. 10.25 and Eq. 10.26
cannot be solved analytically, numerical methods have been used. Some of these results are
represented graphically for certain standard configurations like
(i) Shape factors between parallel rectangles of equal size.
(ii) Shape factors between rectangles perpendicular to each other and having a
common edge
(iii) Shape factor from an elemental area dA1 to a rectangular area A2
(iv) Shape factor between two coaxial parallel discs
(v) Shape factors for concentric cylinders of finite length etc.
With the help of those charts and View Factor algebra, shape factors between surfaces not
covered above can be determined.
452
10.7 ILLUSTRATIVE EXAMPLES ON VIEW FACTORS:
Example 10.11: Determine the view factor between an elemental area A1 and a circular disc
A2 of radius R. The two areas are parallel to each other and positioned at a distance L form
each other such that the perpendicular to A1 passes through the centre of A2.
R
ρ
Φ2
L Φ1
r
α
2pr d r L2 2r d r
R R
\ F1- 2 = �
0 ( L + r ) p ( L + r )
2 2 2 2
= L2 �
0 ( L + r )
2 2 2
R2
\ F1- 2 = 2 = sin 2 a
L +R 2
453
Example 10.12: Obtain an expression for the shape factor for a conical cavity with respect to
itself. The height of the cavity is H and the semi vertex angle of the cavity is α
(See Fig. E10.12a)
Solution:
A2
H dz
α
z
r
ds
Fig. E10.12(a) : Figure for Example 10.12
= 2 π z dz tan α sec α
H
Therefore A1 = ∫2 π z dz tan α sec α = πH2 tan α sec α
0
πH 2 tan 2 α
454
Hence F1-1 = 1 – F1-2 = 1 − ------------------------- = 1 – sin α
πH2 tan α sec α
Example 10.13: Consider 3 small surfaces each of area dA1 = dA2 = dA3 = 2 cm2 as shown in
fig E10.13. (a) Calculate the solid angle subtended by dA 2 with respect to a point on dA 1 (b)
The solid angle subtended by dA3 with respect to a point on dA 1 and (c) The elemental
Diffuse factors dFdA1-dA2 and dFdA1-dA3.
Φ1 = 30 0; Φ2 = 45 0 ;
r12
n3
n2
n1 r13
Φ1
Fig. E10.13: Figure for example 10.13
dA1
If dw2 -1 is the solid angle subtended by dA2 w.r.t a point on dA1 then
dA2 cos f2 2 cos 450
dw2 -1 = = = 1.57 �10-5 sr
r122 3002
dA3 cos f3
Similarly dw3-1 = 2
(but f3 = 0 0 )
r13
dA3 2
\ dw3-1 = 2
= 2
= 5�10 -5 sr
r13 200
dA2 cos f12 cos f2 2 �cos 300 �cos 450
dFdA1 - dA2 = = = 4.33 �10-6
p ( r12 )
2
p �300 2
Example 10.14: Determine the view factor F1-2 between an elemental surface dA 1 and the
finite rectangular surface A2 for the geometric arrangements shown in Fig E10.14
455
8m 8m
3m A2 3m A2 A2
6m
4m 4m
3m
3m
3m
Soluiton : (a) The configuration in Fig. E10.14(a) is a standard configuration for which the
analytical expression for F1-2 is given by
1 x y y x
F1-2 = ---- [ ------------ tan − 1 {------------- } + ------------- tan − 1 { ----------------} ]
2π √(1 + x2) √(1 + x2) √(1 + y2) √(1 + y2)
1 1 1.33 1.33 1
F1-2 = ---- [ ------------ tan − 1 {------------- } + ------------- tan − 1 { ----------------} ]
2π √(1 + 12) √(1 + 12) √(1 + 1.332) √(1 + 1.332)
456
Or F1-2 = 0.154
(a) The area A2 is divided into two equal areas A3 and A4 both having the common edge
of width L1 = 3 m.Then
= 2 x 0.15 v4 = 0.308
(c) In this case A2 is divided into three equal areas A3,A4, and A5 as shown below
A4 A5 = 3 F1-3
A3 = 3 x 0.154
= 0.462
Example 10.15: Determine the Shape factor F1-2 for the configuration shown in Fig E10.15
457
3m
A1
1m
1m
A2
1m 1m
Solution:
A6 F6-1 A4 F4-1
F1-2 = F1-6 − F1-4 = ---------- − --------------
A1 A1
A6 A4
= ------ [ F6-5 − F6-3 ] − ---------- [ F4-5 − F4-3 ]
A1 A1
Values of F6-5, F6-3, F4-5 and F4-3 can be obtained from chart as follows.
To find F6-5:
458
A5 = A6 = 3 x 2 = 6 m 2 ;
2m
L1 / W = L2/W = 2/3 = 0.667;
2m
A6
6 3x1
Hence F1-2 = ------- [ 0.22 − 0.16 ] − ------- [ 0.32 − 0.27 ] = 0.07
1x3 3x1
Example 10.16: Find F1-2 for the configuration shown in Fig. E10.16
459
2m
2m
2m
A2
A1
2m
A6 = A2 + A4 = 4 x 2 = 6 m2 and
A4 A5 = A1 + A3 4 x 2 = 6 m2.
A1 = A2 = A3 = A4 = 2 x 2 = 4 m2.
A2
A1
F1-2 = F1-6 − F1-4
A3 = (A6F6-1) / A1 − F1-4
460
To find F6-5:- L2 = L1 = 2 m; W = 4 m.
L1 / W = L2 / W = 2 / 4 = 0.5.
From chart F5-6 = 0.25 = F6-5 since A5 = A6. Similarly F1-4 = 0.2.
Example 10.17: Find the Shape Factor F2-1 for the configuration shown in fig E10.17
A2
2m
2m
A1
0.5
2m
461
A1 = A3 + A5; A6 = A2 + A4
A4
A2
A2 = A3 = 0.5 x 2 = 1 m 2
A4 = A5 = 1.5 x 2 = 3 m 2
A1 = A6 = 2 x 2 = 4 m2
A3 A5
Fig. E10.17(a)
Example 10.18: Find the Shape factor F1-2 for the configuration shown in Fig E10.18
= F1-8 + 2 (A4F4-1) / A1
462
a/2
a
A1
a
A2
A7 = A1 + A3 ;
A3
A10 = A4 + A8
A1 A4
A5 A8
A6
a2
Hence F1-2 = 0.15 + 2 x -------------- x [0.23 – 0.20] = 0.27
[(a/2) x a]
Example 10.19 A1 and A2 are two rectangular flat surfaces having a common edge and
inclined at an arbitrary angle α to each other. They are very long along the common edge and
have lengths of ab and ac respectively in the other direction. Show that
463
F1-2 = (ab +ac) – bc
2ab
Or (A1F1-3) / A3 = 1 – (A2F2-3)/A3
(ab + ac – bc)
= -------------------
2 ab
Example 10.20: (Hottel’s cross string formula) Obtain an expression for the view factor
between two flat surfaces, which extend to infinity in one direction.
464
A2 = L2 x 1
A4 = L4 x 1
A6 = L6 x 1
A3 = L3 x 1 A1 = L1 x 1
A5 = L5 x 1
Consider unit width perpendicular to the plane of the paper. A 1, A2, A3 and A4 form an
enclosure. Hence we have
F1-1 + F1-2 + F1-3 + F1-4 = 1 and F1-1 = 0.
= 1 – (A5F5-1) / A1
= 1 – (A5 / A1) [ 1 – F5-3]
465
Hence F1-3 = --------------------------------- = ---------------------------
2 2A1
A1 – A6 + A4
Similarly F1-4 = ------------------------
2A1
A1 – A5 + A3 A1 – A6 + A4
F1-2 = 1 – ------------------- − -------------------
2A1 2A1
Example 10.21: A truncated cone has top and bottom diameters of 10cm and 20cm and a
height of 10cm. Calculate the shape factor between the top surface and the side and the side
and itself.
Solution: To find (i) F2-3 and (ii) F3-3 . Refer to Fig. E10.21.
(i) F2-1 = (A1F1-2) / A2. F1-2 can be directly obtained fom chart as follows:
L1 / r = 10 / 10 = 1 ; and L2 / r = 5 / 10 = 0.5.
π x (10)2
Therefore F2-1 = -------------- x 0.12 = 0.48; F2-1 + F2-2 + F2-3 = 1 and F2-2 = 0
π x (5)2
466
A2
r2
A3
r1 = 10 cm
r2 = 5 cm
L
L = 10 cm
r1
A1
π x (10 2)
Hence F3-1 = (A1F1-3) / A3 = ------------- x 0.88 = 0.525
526.9
Example 10.22: Determine the shape factors for the geometries shown in Fig. E10.23(a) to
E10.23(i)
467
Long groove, A1 Surroundings, A2
A1
Long duct
A2
A1 1.0 Surroundings, A3
Find F1-1 and F1-3
1.0
468
A1
A2
2.0 A2
A1
1.0
A1
A2
A1 A3
1.0
A2
1.5
1.5
1.0
3.0
A1
A2
Fig. E10.23 (i) : F1-2, F2-1, F1-3; hemispherical shell, A1 enclosing a small sphere,A2 surrrounded
1.0
3.0
469
Solution: Refer Fig. E10.23(a):
All the radiation from (1) which goes to the surroundings can be intercepted by an imaginary
surface 2 and 4 as shown in Fig. Now, 1, 2 and 4 form an enclosure.
470
\ F1-1 = 1 - 2 F1- 2 - - - - - - - - - (a )
also F2-1 + F2- 2 + F2- 4 = 1 and F2- 2 = 0
\ F2-1 + F2- 4 = 1 � F2-1 = 1 - F2-4
from chart for two parallel coaxial discs ( 2 ) and ( 4 )
F2- 4 = 0.383
F2-1 = 1 - 0.383 = 0.617
A2 F2 -1 p R 2 1.0
\ F1- 2 = = �0.617 = �0.617 = 0.309
A1 2p RL 2 �1.0
\ F1-1 = 1 - 2 �0.309 = 0.383
F1-3 = 1 - F1-1 = 1 - 0.383 = 0.617
A2 F2-1 = A1F1- 2
A1 p ( 0.52 )
\ F2-1 = F1-2 = �1 = 0.055
A2 2p ( 1.52 )
From eq (b), F2-2 = 1 - ( F2-1 + F2 -3 )
= 1 - ( 0.055 + 0.444 )
\ F2-2 =0.5
471
Refer to Fig. E10.23(e):
Surroundings (3) can be replaced by imaginary surfaces (5) and (4) as shown in
Fig.below.Now, 1, 2, 4 and 5 form an enclosure.Hence
To find F5-( 1+ 4 ) :
d 3
R= = = 0.75
2 L 2 �2
2 R 2 + 1 2 �0.752 + 1
X= = = 3.78
R2 0.752
X- �
X 2 - 4�
� � 3.78 - �
3.782 - 4 �
� �
\ F( 1+ 4 ) -5 = F5-( 1+ 4) = =
2 2
or F5-( 1+ 4 ) = 0.286
� From Eqn (d) F5- 2 = 1 - 0.286 = 0.714
472
A5 F5- 2 p ( 1.52 )
\ F2-5 = = �0.714 = 0.268
A2 2p �1.5 �2
\ From (b) F2- 2 = 1 - 2 �0.268 = 0.465
A1 F1-5
Now F5- 4 = F5-( 1+ 4) - F5-1 = F5-( 1+ 4) -
A5
p ( 0.52 )
= 0.286 - �0.35 = 0.247
p ( 1.52 )
A5 F5- 4 p ( 1.52 )
\ F4-5 = = �0.247 = 0.278
A4 p ( 1.52 - 0.52 )
From (c) F4- 2 = 1 - 0.278 = 0.722
A4 F4- 2 p ( 1.5 - 0.5 )
2 2
Consider two black surfaces of area A1 and A2 and at temperatures T1 and T2 as shown in Fig
10.14. Then radiation leaving A1 and reaching A2 can be written as
473
A2,T2, Eb2
A1, T1,Eb1
Q1->2 = A1Eb1F1-2
The calculation of radiation heat transfer between black surfaces is relatively easy because all
the radiant energy which strikes a surface is absorbed by it. When non black bodies are
involved, the situation is much more complex because all the energy striking a surface will
not be absorbed: part will be reflected back to another heat transfer surface, and part may be
474
reflected out of the system entirely. The problem can become complicated because the
radiant energy can be reflected back and forth between heat transfer surfaces several times.
While deriving the expression for radiation exchange between any two finite grey
surfaces the following assumptions are made
i. All the surfaces are diffuse and uniform in temperature.
ii. The reflection and emissive properties are constant over all the surface
iii. Radiosity and irradiation are uniform over each surface. This assumption is not
strictly correct even for ideal grey diffuse surfaces, but the problems become
exceedingly complex when this restriction is not imposed.
iv. The surfaces are opaque (i.e. Transmissivity is zero)
Qr = A(J – G).
But J = E + (1 – α)G.
Assuming α = ε, J = ε Eb + (1 – ε)G. Or G = (J - ε Eb) / (1 – ε)
Qr = A [J – (J - ε Eb) / (1 – ε)]
(Eb – J) (Eb – J)
Qr = ---------------- = ---------------- ………………..(10.28)
(1 – ε)/(Aε) R
Eq. 10.28 can be interpreted as follows… (Eb-J) can be thought of as thermal potential, R =
(1-ε)/εA can be thought of as thermal resistance offered by the surface for radiation, as Q r is
the radiation heat flow rate. Therefore a radiating surface can be replaced by an element as
shown in Fig 10.15
Qr
Eb R = (1 – ε)/(Aε) J
Fig 10.15: Element representing surface resistance in the Radiation network method
Now let us consider the radiation heat exchange between two surfaces A 1 and A2. Radiation
which leaves A1 and strikes A2 is given by
Q1→2 = A1J1F1-2
475
Q2→1 = A2J2F2-1
J1 J2
R12 = 1 / (A1F12)
The two network elements shown in Fig 10.15 and 10.16 represent the essentials of radiation
network method. To construct a network for a particular radiation heat transfer problem we
need only to connect a “surface resistance” (1-ε)/εA to each surface and a “space resistance”
1/(AiFi-j) between the radiosity potential potentials. This is illustrated below.
The radiation network for the above problem will be as shown in Fig 10.17
J1
T1, α1= ε1, A1 G1 = J2
G2 = J1 J2
T2, α2 = ε2, A2
Eb1 J1 J2 Eb2
Fig 10.17 Radiation network for two parallel infinite grey surfaces
476
(Eb1 – Eb2)
From Fig. 10.17 we can write Qr1 = Q12 = ------------------------
(R1 + R12 + R2)
σ(T14 – T24)
Or Q12 = ---------------------------------------------------------------------
{(1 – ε1) / (A1ε1) + 1 / (A1F12) + (1 – ε2) / (A2ε2)}
For two infinite parallel surfaces A1 = A2 and F12 = 1.0. Hence the above expression
simplifies to
Fig: 10.18 Radiation Net work forHeat Exchange Between Two Parallel Infinite Gray
surfaces in presence of a radiation shield
From Fig 10.18 the net radiation heat transfer from A1 to A2 is given by
(Eb1 – Eb2)
Q12 = ----------------------------------------------
477
[R1 + R13 + R3 + R3’ + R32 + R2]
(Eb1 – Eb2)
= ---------------------------------------------------------------------------------------------------
[(1 – ε1) / (A1ε1) + 1 / (A1F13) + (1 – ε13) / (A3 ε13) + (1 – ε32) / (A3 ε32) + 1 / (A3F32)
+ (1 – ε2) / (A2ε2)]
σA1(T14 – T24)
Q12 = ------------------------------------------------------------------------------------------
[{(1 – ε1) / ε1} + 1 + {(1 – ε13) / ε13} + {(1 – ε32) / ε32}+ 1 + {(1 – ε2) / ε2}]
σA1(T14 – T24)
= -------------------------------------------------------------------- ………………..(10.30)
[ { (1 / ε1) + (1 / ε2) – 1 } + { (1 / ε13) + (1 / ε32) – 1 }]
A2, T2, ε2
A1, T1, ε1
478
(Eb1 – Eb2) σ(T14 – T24)
Qr1 = Q1-2 = -------------------- = -----------------------------------------------------------
(R1 + R12 + R2) [(1 – ε1)/(A1ε1) + 1/(A1F1-2) + [(1 – ε2)/(A2ε2)]
σ(T14 – T24)
= ------------------------------------- ………………………………………(10.31)
[(1/ε1) + (A1/A2){(1/ε2) – 1}]
Eb1 J1 J2 Eb2
Fig. 10.19 (b) : Radiation network for the configuration shown in Fig. 10.19(a)
A1, ε1, T1
A2, ε2 T2
A3, ε3 T3
479
Fig. 10.20 (b) : Radiation Network for a three zone enclosure
Reradiating Surface: In many practical situations one of the zones may be thermally
insulated. In such a case, the net radiation heat flux in that particular zone is zero, because
that surface emits as much energy as it receives by radiation from the surrounding zones.
Such a zone is called a “RERADIATION ZONE” or an “ADIABATIC ZONE”. Fig 10.21(a)
represents a three zone enclosure with surface (3) being the reradiating surface and Fig
10.21(b) the corresponding radiation network.
A1, ε1, T1
A2, ε2 T2
Fig. 10.21 (a) : Radiation in a three – zone enclosure with A3 being a reradiating surface
480
Fig. 10.21 (b) : Radiation network for the configuration shown in Fig. 10.21(a)
For a three zone enclosure under steady state conditions, by I law of thermodynamics.
(1 – ε1) (1 – ε2)
─1 ─1
i.e Req = ----------- + [ A1F12 + {1/(A1F1-3) + 1/(A2F2-3)} ] + ------------
A1ε1 A2ε2
………………. (10.33b)
Example 10.24: Two square plates 1m x 1m are parallel to and directly opposite to each
other at a distance of 1m. The hot plate is at 800K and has an emissivity of 0.8. The clod
plate is at 600K and also has an emissivity of 0.8. The radiation heat exchange takes place
between the plates as well as the ambient at 300K through the opening between the plates.
Calculate the net radiation at each plate and the ambient.
Solution:
481
1m
A1, ε1, T1 1m
1m 3
A2, ε2, T2
Solution: (1 – 0.8)
R1 = (1 – ε1) / (A1ε1) = ----------------- = 0.25
(1 x 1) x 0.8
(1 – 0.8)
Similarly R2 = (1 – ε2) / (A2ε2) = ----------------- = 0.25
(1 x 1) x 0.8
R3 = (1 – ε3) / (A3ε3)
A3 is the area of the surroundings which is very large. Hence R3 = 0.The radiation net work
for this problem is shown in Fig.E 10.24.
From chart F1-2 = F2-1 = 0.20. But F1-1 + F1-2 + F1-3 = 1 and F1-1 = 0.
Hence F1-3 = 1 – F1-2 = 1 – 0.2 = 0.8 = F2-3.
R12 = 1 / (A1F1-2) = 1 / {(1x1) x 0.2} = 5.0. R13 = 1 / (A1F1-3) = 1 / {(1x1) x 0.8} = 1.25
482
\
Eb1 - J1
-
( J 1 - J 2 ) - J1 - J 3 = 0
R1 R12 R13
23.224 - J1 ( J1 - J 2 ) ( J1 - 0.459 )
or - - = 0 _______ ( a )
0.25 5 1.25
similarly Qr2 = Q21 + Q23 � Qr2 - Q21 - Q23 = 0
Eb2 - J 2
-
( J 2 - J1 ) - J 2 - Eb3
=0
R2 R12 R23
7.348 - J 2 ( J 2 - J1 ) ( J 2 - 0.459 )
- - = 0 _______ ( b )
0.25 5 1.25
Solving Eqn (a) and (b) simultaneously we get
J1 = 18.921KW / m2 ; J 2 = 6.709 KW / m 2
Eb1 - J1 23.224 - 18.921
\ Qr1 = = = 17.212KW
R1 0.25
Eb2 - J 2 7.348 - 6.709
Qr2 = = = 2.557 KW
R2 0.25
But Qr1 + Qr2 + Qr3 = 0 � Qr3 = - � �= - [ 17.212 + 2.557 ]
Qr1 + Qr2 �
�
Qr3 = -19.769 KW
Solution:
A1 = A2 = A3 = 1m2
T1 =900K, Є1 = 0.8
T2 = 400K, Є2 = 0.8
The radiation network for the above problem will be as shown for example 10.24
483
1 - e1 1 - 0.8
R1 = = = 0.25
A1e1 1�0.8
1 - e 2 1 - 0.8
R2 = = = 0.25
A2e 2 1�0.8
Using Hottel's cross string formula, we have
F1- 2 =
( A1 + A2 ) - A3 = ( 1 + 1) - 1 = 0.5 = F = F2 -3
1-3
2 A1 2 �1
1 1
\ R12 = = = 2 = R23 = R13
A1 F1-2 1�0.5
-1
�1 1 �
Req = R1 + � + � + R2
R
�12 R13 + R23 �
-1
1
� 1 �
= 0.25 + � + + 0.25 = 1.833
2 2 + 2�
� �
Eb1 - Eb2 s ( T14 - T24 ) 5.67 �10-8 ��
900 4 - 400 4 �
� �
Qr1 = = =
Req Req 1.833
= 19503 W / m 2
Qr1 + Qr2 + Qr3 = 0 and Qr3 = 0 � Qr2 = -Qr1 = -19503 W / m 2
The radiation network for the above problem will be as shown for example 10.24
484
1 - e1 1 - 0.8
R1 = = = 0.318
A1e1 p ( 0.52 ) 0.8
1 - e 2 1 -1
R2 = = =0
A2e 2 A2 �1
1 1
R12 = = = 7.3
A1 F1- 2 p ( 0.5 ) 0.175
2
1 1
R13 = = = 1.54
A1 F1- 2 p ( 0.5 ) 0.825
2
1 1
R23 = = = 1.54
A2 F2-3 p ( 0.5 ) 0.825
2
-1
�1 1 �
Req = R1 + � + � + R2
�R12 R13 + R23 �
-1
�1 1 �
= 0.318 + � + + 0 = 2.484
�7.3 1.54 + 1.54 �
�
Eb1 - Eb 2 s ( T1 - T2 ) 5.67 �10 ��
-8
4 4
�20004 - 1000 4 �
�
\ Qr1 = = =
Req Req 2.484
= 329.14 �103 W = 329.14 KW
Qr1 + Qr 2 + Qr 3 = 0 and Qr 3 = 0
\ Qr 2 = -Qr1 = -329.14 KW
Example 10.27 A spherical tank with diameter 40cm fixed with a cryogenic fluid at 100K is
placed inside a spherical container of diameter 60cm and is maintained at 300K. The
emissivities of the inner and outer tanks are 0.15 and 0.2 respectively. A spherical radiation
shield of diameter 50cm and having an emissivity of 0.05 on both sides is placed between the
spheres. Calculate the rate of heat loss from the system by radiation and find also the rate of
evaporation of the cryogenic liquid if the latent heat of vaporization of the fluid is 2.1x10 5 W-
s/Kg
Solution: The schematic and the corresponding network for the problem will be as shown in
Fig P.4.27
485
Eb1 - Eb 2
Qr1 = Q12 =
R1 + R13 + 2R3 + R32 + R2
s ( T14 - T24 )
=
1 - e1 1 1- e3 1 1- e2
+ +2 + +
A1e1 A1 F1-3 A3e 3 A3 F3- 2 A2e 2
s A1 ( T14 - T24 )
=
1 A1 �2 � A �1 �
+ � - 1�+ 1 � - 1�
e1 A3 �e 3 � A2 �e 2 �
5.67 �10 -8 �4p �( 0.2 2 ) ��
1004 - 3004 �
� �
= 2 2
= -6.83 W
1 �40 �� 2 � �40 ��1 �
+ � �� - 1�+ � �� - 1�
0.15 �50 ��0.05 � �60 �� 0.2 �
Q 6.83
Evaporation Ratio = r1 = 5
= 3.25 �10 -5 Kg / s
hfg 2.1�10
486
4.10 RADIOSITY- MATRIX METHOD FOR RADIATION IN ENCLOSURES
The network method is quite easy to apply for determining the radiation exchange in
enclosures having not more than 3 zones. As the number of zones forming the enclosure
increases, the manipulations involved in the network method becomes enormous and the
method is not so practical. Whereas the radiosity-matrix method is very straight forward and
the method transforms the problem to the solution of an algebraic matrix equation for the
unknown radiosities Ji [i=1,2,3,…….N] once these equations are solved for Ji then the net
radiation flux or the zone temperature at any zone (i) can be immediately determined. This is
illustrated below.
Consider an enclosure made up of N zones.Let Gi be the irradiation at zone i
Let Ji be the Radiosity at zone i .Net radiation heat flux from zone i is given by
qi = Qri / Ai …………………………….(10.34)
Qj – i = AjFj-i Jj = AiFi-j Jj
N
Therefore irradiation for zone i = Gi = Qj – I / Ai = ∑Fi-j Jj ………………………….(10.35)
j=1
(Ebi – Ji)
Also qi = ----------------- ………………………………(10.36a)
{(1-ε i) / εi}
487
Equations (4.36) and (4.37) provide the fundamental relations for obtaining a system of N
algebraic equations to determine the N unknown radiosities. Once the radiosities are known
the net radiation heat flux qi at any zone (i) can be computed using either Eq. 4.36 or 4.36a
The solution depends on the prescribed conditions for each of the zones. Two
situations are of practical interest
1. Temperatures are prescribed for each of the N zones
2. Temperatures are prescribed for some of the zones and the net radiation flux are
prescribed for the remaining zones
T1
T2
-
{T} = - (10.39d)
-
TN
488
The elements mij of matrix [M] can be determined from the following:
δij ─ (1 – εi) Fi-j
mij = ------------------------- , i = 1 to N and j = 1 toN.(10.39e)
εi
Where δij = 1 for i = j and = 0 for i ≠ j. …………………………(10.39f)
ii. Temperature prescribed for some zones and net heat flux prescribed for
others
In many practical situations, temperatures are prescribed for some of the zones and
net heat fluxes for the remaining zones of an enclosure. In such problems we have to
determine the net heat fluxes for the zones for which temperatures are specified and
temperatures for the zones for which the net heat fluxes are prescribed. This can be done
using the same equations (10.36) and (1037) and illustrated below.
Let us assume that temperatures Ti are prescribed for zones i=1, 2, 3 ….k and the net heat
fluxes qi are prescribed for the remaining zones i=k+1, k+2….n
For zones 1 to k since temperatures are prescribed we can use eq. 10.38 that is
N
( Ji / εi) – {(1 – εi) / εi}∑Fi-jJj = σTi4; i= 1 to k……………… (10.40)
j
=1
For zones i = k+1, k+2, … N, with prescribed heat fluxes we can use Eq (4.36) namely
N
qi = Ji - ∑Fi-j Jj ; 1 = k+1, k+2, …… N (10.40a))
j=1
It is more convenient to express Eq. (10.40) and (10.40a) in matrix form as follows.
489
qN
Ones these equations are solved for unknown radiosities Ji, then the unknown radiation fluxes
can be determined from the equation
εi
qi = --------- [Ebi – Ji] ; i = 1,2,3, ……….k
(1 – εi)
…………….(10.41e)
10.10: ILLUSTRATIVE EXAMPLES ON RADIOSITY MATRIX METHOD
Since J3 is already known, we have to solve only for J 1 and J2. Thus the matrix form of
equation for radiosities J1 and J2 can be written as
490
s T14 �
�
m11 m12 ��J1 � �
� �
� �� �= � 4 �
m21 m22 J 2
��� s T2 �
�
d ij - ( 1 - e i ) Fij
mij =
ei
where d ij = 1 for i = j
= 0 for i �j
1 - ( 1 - 0.8 ) �0
\ m11 = = 1.25; s T14 = 5.67 �10 -8 �800 4 = 23224W / m 2
0.8
0 - ( 1 - 0.8 ) 0.2
m12 = = -0.05; s T24 = 5.67 �10-8 �600 4 = 7348W / m 2
0.8
0 - ( 1 - 0.8 ) 0.2
m21 = = -0.05
0.8
1 - ( 1 - 0.8 ) �0
m22 = = 1.25
0.8
�1.25 -0.05��J1 � � 23224 �
\� � � �= � �
-0.05 1.25 �J 2
��� 7348
i.e,
1.25 J1 - 0.05 J 2 = 23224
-0.05 J1 + 1.25 J 2 = 7348
Solving for J1 and J 2 we get J 2 = 6632W / m 2 and J1 = 18845W / m 2
Eb1 - J1 23224 - 18845
\ qr1 = = = 17516W / m 2
�1 - e1 � �1 - 0.8 �
� � � �
� e1 � � 0.8 �
E - J 2 7348 - 6632
qr 2 = b 2 = = 2864W / m 2
�1- e2 � � 1 - 0.8 �
� � � �
� e 2 � � 0.8 �
qr1 + qr 2 + qr 3 = 0 � qr 3 = - [ 17516 + 2864 ] = -20380W / m 2
Example 10.29:- Solve example 10.25 by radiosity-matrix method.
ε2 = 0.8;
This is a case wherein temperature is specified for 2 zones and radiation flux is specified for
the remaining zone.
δij − (1 – εi) Fij
mij = ------------------------ for i = 1,2 and mij = δij − Fij for i = 3
εi
0 – (1 – 0.8) x 0.5
Therefore m11 = (1 – 0) / 0.8 = 1.25 ; m12 = ------------------------- = - 0.125 = m21
0.8
491
(1 – 0) 0 – (1 – 0.8) x 0.5
m22 = ------------ = 1.25 ; m13 = ------------------------ = − 0.125 = m23
0.8 0.8
σ T14
[mij]{Ji} = {S} where {S} = σ T24
0
In expanded form the above equation can be written as:
Substituting the numerical values fo mij , T1, T2 and σ the above three equations can be solved
for J1, J2 and J3.
492
CHAPTER 11: MASS TRANSFER
11.1 INTRODUCTION: Mass transfer can result from several different phenomena. There
is a mass transfer associated with convection, where in, mass is transported from one place to
another in the flow system. This type of mass transfer occurs on a macroscopic level and is
normally treated in the subject of fluid mechanics. When a mixture of gases or liquids is
contained such that, there exists a concentration gradient of one or more of the constituents of
the system, there will be a mass transfer on a microscopic level as the result of diffusion from
regions of high concentration to regions of low concentration.
This chapter deals with some of the simple relations which may be used to calculate mass
diffusion and their relation to heat transfer. However it should be remembered that the
general subject of mass transfer encompasses both mass diffusion on a molecular scale and
the bulk mass transport which may result from a convection process.
Mass diffusion not only occurs on a molecular scale, but also in turbulent systems
where accelerated diffusion rates will occur as a result of the rapid-eddy mixing process.
Mass diffusion may also result from a temperature gradient in a system; this is called
thermal diffusion. Similarly, a concentration gradient can give rise to a temperature gradient
and a consequent heat transfer.
11.2. FICK’S LAW OF DIFFUSION: Consider the system shown in Fig. 11.1. A thin
A thin partition separates two gases A and B. When the partition is removed, the two gases diffuse through
A B
Therefore mA
------ = − D (∂CA / ∂x) ………………
A
493
An expression similar to Eq.(11.1) could also be written for the diffusion of the constituent
A in either y or z direction.
τ = μ (∂u / ∂y)
can be noticed. The heat conduction equation describes the transport of energy, the viscous-
shear equation describes the transport of momentum across the fluid layers, and the Fick’s
diffusion equation describes the transport of mass.
CA
on the left hand side of this plane than on the right hand side. A higher concentration means
that there are more molecules per unit volume. If the system is a gas or a liquid, the
molecules are moving in a random fashion, and the higher the concentration, more
molecules will cross a given plane per unit time. Thus, on the average, more molecules are
moving from left to right across the plane than in the opposite direction. This results in a net
mass transfer from the region of high concentration to the region of low concentration.
In gases the diffusion rates are clearly dependent on the molecular
speed, and consequently we should expect a dependence of the diffusion coefficient on
temperature since temperature indicates the average molecular speed.
494
11.3. DIFFUSION IN GASES: GILLILAND has proposed a semi-empirical equation for
the diffusion coefficient in gases:
T3/2 __________________
D = 435.7 ---------------------------- √[ (1 / MA) + (1 / MB) ]
p [ VA1/3 + VB1/3 ]
……………………….(11.2)
In the above equation D is in cm 2/s, T is in degrees Kelvin, p is the total system pressure in
pascals, and VA and VB are the molecular volumes of constituents A and B and MA and MB
are the molecular weights of constituents A and B.
p = ρRT ………………………………(11.3)
The density ρ represents the mass concentration to be used in Fick’s law. The gas _
constant R for a particular gas may expressed in terms of the universal gas constant R and the
molecular weight of the gas: _
R
RA = ------- …………………………….(11.4)
MA
_
Where R = 8314 J/(kg-mol-K)
_
or R = 0.08205 m3.atm/(kg-mol – K)
495
Consequently Fick’s law of diffusion for component A into component B can be written as
follows:
_
CA = ρA = (pAMA) / (R T)
_
Hence ( ∂CA / ∂x) = [MA / (R T)] (dpA/dx)
mA _
------ = − DAB [MA / (R T)] (dpA/dx) ………………..(11.5)
A
Similarly we can also write for the diffusion of component B into component A as
mB _
------ = − DBA [MB / (R T)] (dpB/dx) ………………..(11.6)
A
Equation (11.6) also assumes isothermal diffusion process.
Equimolal counter diffusion in gases:- Consider a physical situation as shown in Fig. 11.3:
NA and NB represent the steady-state molal diffusion rates of components A and B,
Reservoir of A Reservoir of B
NA NB
Δx
1 2
mA _
NA = ------ = − DAB [A / (R T)] (dpA/dx)
496
MA
mB _
NB = ------ = − DBA [A / (R T)] (dpB/dx)
MB
The total pressure of the system p remains constant at steady state, so that
p = pA + pB = constant
Since each molecule of A is replacing a molecule of B, the molal diffusion rates are equal:
− NA = NB
Illustrative example 11.1:- Consider two large vessels, each containing uniform mixtures of
nitrogen and carbon dioxide at 1 atm, and 288.9 K, but at different concentrations. Vessel 1
contains 90 mole percent N2 and 10 mole percent CO2, where as vessel 2 contains 20 mole
percent N2 and 80 mole percent CO2. The two vessels are connected by a duct of 0.1524 m
inside diameter and length of 1.22 m. Determine the rate of transfer of nitrogen between the
two vessels by assuming that steady-state transfer takes place. The mass diffusivity for the N2
– CO2 mixture at 1 atm and 288.9 K is 0.16 x 10 − 4 m 2/s.
Solution: In this mass transfer process, N2 is transferred from vessel 1 containing higher
concentration of N2 to vessel 2 having lower concentration. In the early stages of mass
transfer, the partial pressure of N2 in both vessels is considered to remain constant so that
steady-state transfer can be assumed. Then the mass transfer process can be characterized as
an equimolal counter diffusion as described above. Hence the mass flux of N2 can be
determined by using Eq. (11.9)
497
Data:- d = 0.1524 m ; length = Δx = 1.22 m; D = 0.16 x 10 − 4 m2/s;
_
R = 0.08205 m3.atm/(kg-mol – K) ; (pN2)1 = 0.9 x 1 = 0.9 atm; (pN2)2 = 0.2 x 1 = 0.2 atm
pw pA
1
Water
p
498
air movement may be necessary to accomplish this, it is assumed that it does alter the
concentration profiles in the air in the tank.We further assume that air and water vapour
behave as ideal gases.
As the water evaporates, it will diffuse upward through the air, and at
steady state this upward movement must be balanced by a downward diffusion of air so that
the concentration at any x will remain constant. But at the surface of the water there can be
no net mass movement of air downward. Consequently there must be a bulk mass movement
upward with a velocity just large enough to balance the diffusion of air
downward. This bulk mass movement then produces an additional mass flux of water vapour
upward. The diffusion of air downward is given by (see Eq. 11.5)
_
mA = − DAB A [MA / (R T)] (dpA/dx) ………………..(11.10)
where A denotes the cross sectional area of the tank. This must be balanced by the bulk-mass
transfer upward. If u is the bulk-mass velocity then
pAMA
mA = − ρAAu = − -------- Au
RT
DAW
or u = -------- (dPA/dx) ……………………….(11.11)
pA
_
(mw)diffusion = − DWA A [MW / (R T)] (dpW/dx) ………(11.12)
pWMW
(mw)bulk = ρWAu = ---_-------- Au
RT
_ pWMWDAW
= − DWA A [MW / (R T)] (dpW/dx) + --------_-------------(dpA/dx) A
499
R T pA
Also p = pW + pA, p = constant. Hence (dpA/dx) = − (dpW/dx) ; DWA = DAW = D. Hence the
expression for (mW)total reduces to
D A MW
(mW)total = − --_-------- (dpW/dx) = [pA + pW] / pA
RT
D A MW
or (mW)total = − --_-------- (dpW/dx) p/ (p −pW) ………………………………..(11.13)
RT
D p MW A D p MW A
(mW)total = --_---------------- ln [(p – pw2) / (p – pw1) = --_-------------- ln [pA2 / pA1]
R T (x2 – x1) R T (x2 – x1)
………………………………….(11.14)
Illustrative example 11.1:- Estimate the diffusion rate of water from the bottom of a test tube
10 mm in diameter and 15 cm long into dry atmospheric air at 250C.
_
Solution:Data:- d = 0.01 m; (x2 – x1) = 0.15 m; T = 25 + 273 = 298 K;R = 8314.3N-m/kMol-
K ;p = 1atm = 101.32 kPa ; pw1 = saturation pressure at 250C = 3.166 kPa
(0.256 x 10 − 4) x 18 x (π x 0.005 2)
(mw)total = ------------------------------------------- ln [(101.32 – 0) / (101.32 – 3.166)]
8314.3 x298 x 0.15
= 3.131 x 10 − 10 kg/s
Steady state Equimolal Counter Diffusion in Liquids:- The analysis for steady state counter
diffusion in liquids follows the same approach as that for gases.
In terms of number of mole flux of species A ,Fick’s law can be written as
500
NA = (mA / MA) = − D (∂CA / ∂x) ……………………..(11.15)
(CA1 − CA2 )
NA = D ---------------- ………………………………….(11.16)
(x2 – x1)
The above equation can also be expressed in terms of mole fractions of species
at locations x1 and x2 as follows.
(χA1 − χA2)
NA = DC -------------- kg-mol /(m2-s) …………………(11.17)
(x2 – x1)
Steady State Unidirectional Diffusion in liquids:- The analysis for the determination of the
concentration distribution during steady state unidirectional diffusion in liquids is exactly
same as that for gases except that in the case of liquids the equation is derived in terms of
molal concentrations CA and CB instead of partial pressures.
( CB2 – CB1)
CB ln = -------------------- ……………………..(11.18 b)
ln (CB2 / CB1)
Also C = CA + CB and CB2 – CB1 = ( CA1 – CA2). Equations (11.18) can also be written in
terms ox mole fractions:
501
( x2 – x1 ) χB ln ( x2 – x1 ) χB ln
( χB2 – χB1)
χB ln = -------------------- ……………………..(11.19 b)
ln (χB2 / χB1)
X = distance,m
Now pA1 / p = (CA)1. Hence pA1 = 0.8 x 1 = 0.8 atm. Similarly pA2 = 0.3 x 1 = 0.3 atm
_
R = 0.08205 m3 – atm / (kg mol – K); MA = 28.86; D = 0.122
= 3.43 x 10 − 8 kg / s.
502
D p MB A
mB = _-------------- ln [pA2 / pA1]
R T (x2 – x1)
= 1.03 x 10 − 7 kg / (s – m2)
503
504