B48BD

Process Engineering C

Student Guide

Heriot-Watt University
Riccarton Campus, Edinburgh EH14 4AS, United Kingdom
Produced by Heriot-Watt University, 2015
Copyright © 2015 Heriot-Watt University

All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means without express permission from the
publisher.

This material is prepared to support the degree programs in Chemical and Petroleum
Engineering.

Distributed by Heriot-Watt University

B48BD Process Engineering C

Acknowledgements

Thanks are due to the members of Heriot-Watt, School of Engineering and Physical Science
who planned and generated this material.

We would like to acknowledge the assistance and contributions from colleagues across the
University and students in preparing this and support material.
Contents
1 Introduction to Heat Transfer...................................................................................1
1.1 Introduction 2
1.2 Reminders 3
1.3 Temperature 4
1.3.1 Thermal Equilibrium 6
1.3.2 Zeroth Law of Thermodynamics 6
1.3.3 Temperature Measurements 7
1.4 Modes of Heat Transfer 8
1.4.1 Conduction 8
1.4.2 Convection 9
1.4.3 Boiling and Condensation 10
1.4.4 Thermal Radiation 12
1.4.5 Heat Loss by Radiation 13
1.5 Process Heat Transfer 15
1.6 Dimensional Analysis 18
1.7 Laminar and Turbulent Flow 24
1.8 Bibliography 26

2 Thermal Conduction................................................................................................ 27
2.1 Introduction 28
2.2 Conduction Through a Plane Wall 29
2.3 Conduction Through a Composite Wall 32
2.4 Conduction Through a Cylindrical Pipe Wall 40
2.5 Conduction Through a Composite Cylinder 43
2.6 Conduction Through a Boundary Layer 47
2.7 Tutorials 48
2.8 Bibliography 53

3 Convection and Radiation...................................................................................... 55

3.1 Introduction 57
3.2 Convection 59
3.3 Lagging 64
3.3.1 Optimum lagging thickness 64
3.3.2 Critical lagging thickness 64
3.4 Resistances 65
3.4.1 Resistances in Series 65
3.4.2 Individual Resistances 67
3.4.3 Case Study – Jacketed Vessel 68
3.4.4 Fouling Resistances 70
3.5 Film Heat Transfer Coefficient 72
3.5.1 Forced Convection Inside Cylindrical Tubes 73
3.5.2 Forced Convection Outside Cylindrical Tubes 75
3.5.3 Forced Convection Over Flat Plates 78
3.5.4 Single Phase Film Coefficient 79
3.6 Radiation 80
3.6.1 Energy Emitted by a Black Body (Plank’s Distribution Law) 81

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

 3.6.2 Stefan-Boltzmann Law 82
3.6.3 Emissivity and Absorptivity Kirchoff’s Law 84
3.6.4 Net Heat Transferred by Radiation 85
3.7 Tutorials 88
3.8 Bibliography 92

4 Heat Transfer Equipment........................................................................................ 91

4.1 Introduction 93
4.2 Heat Transfer Equipment 94
4.2.1 Double Pipe (Concentric Tube) Heat Exchanger 94
4.2.2 Gasketed Plate Heat Exchanger 95
4.2.3 Welded Plate Exchanger 96
4.2.4 Spiral Heat Exchanger 96
4.2.5 Counter-current Heat Exchanger 97
4.3 Mean Driving Force 100
4.4 Temperature Profiles 101
4.5 Practical Heat Exchangers 102
4.5.1 Shell-and-Tube H/E 102
4.5.2 Plate H/E 104
4.6 Heat Exchanger Basics 105
4.6.1 Log Mean Temperature Difference Method 107
4.6.2 Justification for Using the Log Mean as the Average 109
4.7 Heat Exchanger Design 111
4.7.1 Calculations of H/E Modes of Operation 111
4.7.2 Performance/Simulation Mode H/E Calculations 113
4.7.3 Maximum Possible Heat Transfer Between Two Streams 115
4.8 Number of Heat Transfer Units (NTU) Method 117
4.8.1 Analytic Solutions for the Effectiveness of H/E 118
4.8.2 Comparison of Two Heat Exchanger Approaches 121
4.9 Tutorials 122
4.10 Bibliography 127

5 Heat Exchanger Design.........................................................................................129

5.1 Introduction 131
5.2 Hydraulic Mean Diameter 132
5.3 Double Pipe (Concentric Tube) Heat Exchanger 133
5.4 Mean Temperature Difference in Multi Tube-Side Pass Exchangers 138
5.5 Shell and Tube Heat Exchanger 143
5.5.1 Tubes 145
5.5.2 Tube Side Velocity 146
5.5.3 Shells 146
5.5.4 Baffles 147
5.5.5 Fluid Allocation 148
5.6 Film Heat Transfer Coefficient 149
5.6.1 Tube-Side (Single Phase) 149
5.6.2 Shell-Side (Single Phase) 149
5.7 Heat Transfer in Vessels Using Jackets and Coils 151
5.7.1 Heat Transfer Coefficients for Coils 153
5.7.2 Heat Transfer Coefficient for Jackets 153
5.8 Transient Heat Transfer 154

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

 5.8.1 Lumped Capacitance 154
5.8.1.1 Unsteady State Heating or Cooling 155
5.8.1.2 Time Constant for Heating and Cooling 156
5.8.2 Distributed Capacitance 158
5.9 Tutorials 163
5.10 Bibliography 165

6 Boiling and Condensation.................................................................................... 167

6.1 Introduction 169
6.2 Condenser Theory & Design 171
6.2.1 Desuperheating and Subcooling 174
6.2.2 Mixtures 174
6.2.3 Pressure Drop in Condensers 176
6.3 Reboiler Theory & Design 181
6.3.1 Types of Reboiler 181
6.3.2 Choice of Reboiler Type 183
6.4 Boiling Heat Transfer 184
6.4.1 Pool Boiling 184
6.4.2 Critical Heat Flux 185
6.4.3 Film Boiling 185
6.4.4 Convective Boiling 185
6.4.5 Convective Heat Transfer Coefficients 186
6.5 Design of Reboiler 188
6.5.1 Forced Circulation Reboilers 188
6.5.2 Thermosyphon Reboilers 188
6.5.3 Kettle Reboilers 191
6.6 Bibliography 194

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

 Topic 1

Introduction to Heat Transfer

Contents

1.1 Introduction 3
1.2 Reminders 4
1.3 Temperature 5
1.3.1 Thermal Equilibrium 7
1.3.2 Zeroth Law of Thermodynamics 7
1.3.3 Temperature Measurements 8
1.4 Modes of Heat Transfer 9
1.4.1 Conduction 9
1.4.2 Convection 10
1.4.3 Boiling and Condensation 11
1.4.4 Thermal Radiation 11
1.4.5 Heat Loss by Radiation 12
1.5 Process Heat Transfer 14
1.6 Dimensional Analysis 17
1.7 Laminar and Turbulent Flow 23
1.8 Bibliography 25

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 2

Prerequisite knowledge
 Engineering mathematics.
 Engineering chemistry.
 Mass and energy balances for both closed and open systems.
 Detailed understanding of the ideal gas law.
 Thermodynamic principles and definitions.
 Detailed understanding of both the steady-flow energy equation and the
non-flow energy equation.

Learning Objectives
By the end of this topic, you should be able to:

 Write down and defined thermal equilibrium and the Zeroth Law of
Thermodynamics
 Convert between common empirical and absolute temperature scales.
 Differentiate between conduction, convection and radiative modes of heat
transfer.
 Outline main facts concerning radiative heat transfer.
 Understand the application of dimensional analysis in Heat Transfer.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 3

1.1 Introduction
Heat transfer is one of the core engineering topics that is common to many
branches of engineering but are more important for process and mechanical
engineers. Heat transfer occurs within a system or between systems when a
temperature difference is present. Electrical engineers has its own particular
issues with heat, usually trying to remove heat from integrated circuit, CPU
processors, The basic principles are the same, and perhaps the way we analyse
problems are the same, the end result and application however is different in terms
of scale. In many respects, this section should be called “Energy Transfer” as in
most engineering cases the energy we transfer is in the form of heat.

Energy balances are crucial in Chemical Engineering since they help to define what
utilities are needed, the energy cost and the size of the heating or cooling
equipment that will be needed. Large energy changes are associated with the
following: separation of fine mixtures, as in distillation; phase change, as in
evaporation, condensation or solidification; chemical reactions, both endothermic
and exothermic.

Smaller energy changes are associated with: pumping and transportation of fluids;
separation of coarsely mixed materials (sand/water); heating or cooling without
phase change.

The non-flow and steady-flow energy equations are both expressions of the First
Law of Thermodynamics, which encapsulates the well-known energy conservation
principle.

Heat and work are both forms of energy. They are energy exchange terms and not
energy storage terms. The First Law is about balancing “quantities”. In this regard 1
(kJ) of work is the same “quantity” of energy as 1 (kJ) of heat.

The Second Law of Thermodynamics is about grading energy in terms of “quality”:

 Work may be converted completely into heat, but heat cannot be completely
converted into work; therefore, 1 (kJ) of work is a higher “quality” (higher
grade of energy) than 1 (kJ) of heat.
 A fixed quantity of heat at a higher temperature is a higher grade more
useful form of energy (higher quality) than when it is at a lower temperature.

Although the quantity of energy is always conserved, energy quality for any real
process tends to decline:

 Work naturally tends to become degraded into heat.

 Heat naturally tends to flow from a higher to a lower temperature.
 Therefore, any device or process that works against this natural tendency –
i.e. heat engines, turbines, compressors, refrigerators, heat pumps, etc.
must have their efficiency limited by the Second Law.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 4

1.2 Reminders
Sensible heat is the heat involved when there is a temperature change.

Specific heat capacity is the heat needed to increase the temperature of 1 kg of a

material by 1oC. This can be Cp (constant pressure) or CV (constant volume). For
liquid water Cp = 4.2 kJ/(kgoC) and for air Cp = 1.0 kJ/(kgoC).

Latent heat is the heat of vaporisation. For water at 0oC this is 2500 kJ/kg.

Please note:

"Specific" means per unit mass, e.g. specific volume is m3/kg.

"Molar" means per mole, e.g. molar volume is m3/mol.

Cross-sectional area of a pipe = area of a circle = d 2   r 2
4
Surface area of a pipe = area of a cylinder = perimeter of circle x length
  d L  2 r L
m3 m
Volume flow = velocity x CSA   m2
s s
.
For sensible heat: Q  m c p T
.
For latent heat : Q  m

The T that you use in a heat balance is a change in stream temperature,

.
Q  m c p T  m c p Tout  Tin 

The T that you use in a heat transfer rate equation is a driving force; the
difference in temperature of two streams, Thot – Tcold, or the log mean difference,
ΔTLM.
T  T2
TLM  1
 T 
ln 1 
 T2 

Driving Force . T
Rate of heat transfer  , i.e. Q  .
Resistance R

. 1
Or, more usually Q  UA TLM where UA 
R

“Flux” often means “per unit area” so:

. .
Heat transfer rate, Q , has units kJ/s but heat flux, Q f , has units kJ/(m2s).

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 5

1.3 Temperature
An empirical temperature scale is one where arbitrary numbers are placed against
a pair of reproducible reference states – see empirical Celsius and Fahrenheit
scales in Table 1.1:

Table 1.1, Empirical Celsius and Fahrenheit scales.

Reproducible Celsius Celsius Fahrenheit Fahrenheit
Reference State Scale “Degrees” Scale “Degrees”
Boiling Point of
Pure Water 100oC 212oF

Freezing Point of 100 180

Pure Water o
0C degre o
32 F degre
es es

Between these fixed points, the empirical Celsius scale has 100 “degrees”, while
the empirical Fahrenheit scale has 180 “degrees”:

 Therefore, each oC is equivalent to 180/100 = 1.8 oF.

 There is an offset of 32 degrees between the scales.
 Using TC for the Celsius scale and TF for the Fahrenheit scale the
conversion between the two scales is given by

TF  (1.8 TC )  32 ………………………………..…(1.1)

An absolute scale places the zero point at the absolute zero of temperature. By
international agreement in 1956 the Triple Point of pure water was chosen as the
reproducible reference state for the Kelvin scale

The Triple Point pressure is 0.006112 bar and by international agreement the
triple point temperature was assigned a precise value of 273.16 K.

This value was chosen so that there was approximately 100 K between the “ice
point” and the “steam point” of pure water, which meant that one degree on the
Celsius scale is the same as one degree on the Kelvin scale.

The Celsius scale TC is now defined in terms of the absolute Kelvin scale TK as
follows:

TC  TK  273.15……………………………………(1.2)

Notice that the Celsius scale is only shifted with respect to the Kelvin scale and the
size of a degree is the same in both cases – thus, temperature difference is the
same for both scales and (W/m oC) is the same as (W/m K).

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 6

The Triple Point is now the reproducible reference state and since its value was
fixed at 273.16 K (exactly), it temperature on the Celsius scale is given by

TCTP  273.16  273.15  0.01o C

Thus, the Triple Point temperature of water is either 273.16 K or 0.01oC.

Likewise the temperature of ice/water or the “ice point” is still 0oC, so that its
temperature on the Kelvin scale is given by

TK ICE  0  273.15  273.15 K

The temperature of steam/water or the “steam point” is still 100oC, so that its
temperature on the Kelvin scale is again given by

TKSTEAM  100  273.15  373.15 K

Knowing that absolute zero must correspond to 0 K, absolute zero on the Celsius
scale is also given by:

TC Abs zero  0  273.15  -273.15 o C

On the Fahrenheit scale absolute zero can be readily be found

TF Abs zero  1.8   273.15  32  -459.67 o F

The Fahrenheit absolute temperature scale is the “Rankine” scale, so that absolute
zero also corresponds to zero rankine.

TR  TF  459.67 ………………………………….(1.3)

In summary the various fixed reference states are listed in Table 1.2:

Table 1.2, Fixed reference states.

Celsius Kelvin Fahrenheit Rankine
Reference Scale Scale Scale Scale
State o
TC ( C) TK (K) TF (oF) TR (R)
Boiling Point of
100 373.15 212 671.67
Pure Water
Triple Point
0.01 273.16 32.018 491.688
Pure Water
Freezing Point
0 273.15 32 491.67
of Pure Water
“Absolute Zero” -273.15 0 -459.67 0

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 7

1.3.1 Thermal Equilibrium

A system that is in equilibrium implies that there are no driving forces across the
boundary (between system and surroundings). This in turn implies complete
uniformity of all thermodynamic properties throughout the system.

Thermal equilibrium implies no temperature difference across the boundary,

which means complete uniformity of temperature throughout the system; under
these conditions there can be no heat transfer.

If you bring a hot object and a cold one into a position where they can exchange
heat, the heat will flow from the hot object to the cold one. Eventually there will be
no net heat flow and the two objects are in thermal equilibrium. The thing that is the
same about the objects in thermal equilibrium is called temperature.

Temperature is a difficult concept that we happen to have grown familiar with. We

can sense differences in temperature but we can also be misled. Metal feels colder
than expanded polystyrene (why?). If you put your hand in water, its temperature
depends on whether your hand was previously touching something cold or hot. It is
useful to remember that a thermometer reads its own temperature. For accurate
temperature readings you must be sure that the thermometer is in thermal
equilibrium with the material that you want to measure.

1.3.2 Zeroth Law of Thermodynamics

This can be stated quite simply:

“If body A and body B are each equal in temperature to a third body C, then body A
and body B must be equal in temperature to each other”.

This is how thermometers work:

 Time must be allowed the thermometer to reach thermal equilibrium.

 Avoid radiation effects (such as direct sunlight when measuring
atmospheric air temperature). This is why air temperature is usually
measured in the shade. Under these circumstances, direct exchange of
solar radiation cannot heat the thermometer above the prevailing air
temperature to give a false reading.
 Avoid too much heat being conducted away from a metal thermometer
pocket.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 8

1.3.3 Temperature Measurement

There are a wide variety of thermometers on the market. One of the oldest types is
the “expansion” thermometer.

Expansion thermometers use the fact that gases and liquids expand on heating. An
obvious example is a mercury-in-glass thermometer. Some industrial thermometers
pass the fluid into a Bourdon tube, which turns a needle on a scale - the same
principle as in the Bourdon pressure gauge.

However, the expansion is quite small so that the liquid volume in the bulb must be
large in relation to the capillary tube volume. Water is not used since its expansion
is non-linear – having a minimum volume at 4oC.

Thermocouples are based on the “thermo-electric” effect. They use the voltage
difference produced in connections between two different metals to give an
electrical signal that change with temperature.

Resistance thermometers change their electrical resistance with temperature and

this in turn changes their electrical signal. Pyrometers heat up a wire. They do not
require any physical contact with the material whose temperature is being
measured. You can match the colour of the wire to the colour of a very hot (eg. red-
hot or white-hot) material. Apparently humans can measure the temperature within
1oC at 1000oC by this method.

Instrumented pyrometers measure the radiation at selected wavelengths in order to

estimate the temperature. Modern plants use resistance thermometers or
thermocouples and send the signal (a current of 4-20 mA or a voltage of 0-5 V) to
the control room.

Pyrometers are also very useful for measuring the temperature of moving objects
where direct physical contact would be difficult or impossible.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 9

1.4 Modes of Heat Transfer

Heat can be transferred by conduction, convection and radiation.

1.4.1 Conduction

Conduction is a thermal transfer process associated with solids. Consider a solid

material either metal or polymer based. The atoms and molecules are fixed in
space but are vibrating. If one side of the solid is hotter than the other, molecules
will be vibrating faster at that side than at the other side. The vibration will pass
from molecule to molecule through the solid until the vibration is the same
throughout the solid. Energy has passed through the solid. Heat has been
transferred through the solid. This process of passing energy from molecule to
molecule through the material is called conduction.

Heat transfer by conduction may also occur in fluids. If the fluid is flowing under
laminar flow conditions, then heat transfer will occur by conduction.

Also, when a hot vapour condenses onto the outside of a cold pipe, a thin laminar
film of condensate forms on the outside of the pipe. Heat is transferred by
conduction through this laminar film to the cold surface.

One dimensional steady-state heat conduction law is known as Fourier’s Law and
is given by

dT
Q  kA ………………………………………….…(1.4)
dx

Where,

Q  The rate of heat transfer (W)

k  The thermal conductivity of the solid or fluid (W/m K)
A  The heat transfer area, perpendicular to heat flow (m2)
dT / dx  The temperature gradient (K/m)

The minus sign is needed because the temperature gradient is negative in the
direction of heat flow ( T final  Tinital ).

Conduction can be enhanced by passing high energy electrons from molecule to

molecule; this is more likely in good electrical conductors and so copper and
aluminium (metals in general) are good heat conductors. The molecules in liquids
and gases are further apart (especially in gases) so conduction in fluids is much
poorer than in solids. If there are air pockets in the solid, e.g. in expanded
polystyrene and duvets, conduction is poorer.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 10

1.4.2 Convection

Think of holding your hand above a hot radiator. You can feel the hot air rising. This
air will mix with cold air and transfer heat to it. This process of fluids moving and
exchanging heat is called convection. It can happen naturally (natural convection)
or be improved by a fan, pump or mixer (forced convection). If a pipe has a fluid
running through it and is heated on the outside, convection carries heat from the
inside wall of a pipe to the rest of the fluid in the pipe.

Note that this is only true in turbulent flow; in laminar flow the process relies mostly
on conduction.

There are two ways (at least) of picturing the heat transfer process here. Both
imagine a film of fluid next to the wall - hence the idea of "film" heat transfer
coefficients. The thickness of the film is undefined but imagine it to be, say 1 mm.

 First picture: the film is static and the resistance to heat transfer comes from
conduction through the film into the mixed, turbulent fluid.

 Second picture: elements (meaning patches) of fluid about 1 mm thick stick

to the wall, get heated up, and then get mixed into the bulk again by
turbulence.

Whichever picture you use, it is clear that the thickness of the film will be smaller or
the rate of replacement of elements of fluid will be faster if you increase the
turbulence of the fluid - stir it faster or have higher Reynolds numbers.

Steady-state heat convection law is given by the expression below:

Q  hAT ………………………………………….…(1.5)

Where,
Q  The rate of heat transfer (W)
h  The fluid heat transfer coefficient (W/m2 K)
A  The heat transfer area, perpendicular to heat flow direction (m2)
T  The temperature difference in the direction of heat flow (K)

If hot water is circulated by a pump through a cooler tube, then heat will be
transferred by forced convection from the body of the hot water to the cooler metal
(the more turbulence within the fluid, the faster the rate of heat transfer).

If water is heated on a stove, then “free convection cells” will form within the body
of fluid; these are caused by hotter/less dense water rising and colder/more dense
water falling – thus, heat is transferred by free convection.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 11

Consider a car radiator: heat is transferred by forced convection, from hot water
within the radiator; by conduction, through the radiator metal wall; and finally by
forced convection to the cold surroundings.

Heat transfer coefficients are found experimentally and many correlations have
been developed and published – broadly speaking, these correlations depend on
the following:

 Whether heat transfer is by forced convection or natural convection.

 Whether the flow regime is laminar or turbulent.

(Whenever hot vapours condense on to cold tubes, or whenever liquids are
boiling, then special correlations are required.)

1.4.3 Boiling and Condensation

Boiling and condensing are special cases of convection in which there is a phase
change. Boiling and condensing give high heat transfer. A tempting explanation is
that in boiling, a small mass of fluid can take up a lot of heat (latent heat) from the
surface as it changes phase and then carry that heat away into the bulk fluid.
Similarly in condensing, a small mass can deposit a lot of heat at the surface as it
changes phase. However tempting this explanation is, it is the turbulence caused
by the bubbles that is the most important factor in giving high heat transfer
coefficients with boiling (C&R vol 16th ed, p490). In Topic 6 you will become familiar
with the reboilers and condensers theory and design.

1.4.4 Thermal Radiation

Thermal radiation involves electromagnetic transfer of heat from a hot body to a

cold body. The heat does not need to be transferred through any substance,
indeed it may even be transferred through a complete vacuum.

All bodies emit thermal radiation, but this is usually invisible (infra-red region of the
electromagnetic spectrum). However, if a body is hot enough, then it will emit
visible light – the wavelength depends on its temperature.

The rate of thermal radiation from an “ideal blackbody” is given by the Stefan-
Boltzmann Law, as follows:

Q  A T 4 ………………………………………….…(1.6)

Where,

Q  The rate of heat transfer (W)

  The Stefan-Boltzmann constant (5.67x10-8 W/m2 K4)
A  The heat transfer area (m2)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 12

T  The absolute temperature (K)

Corrections are usually required to equation (1.6) to deal with the following
complications:

 “Non-black” behaviour – no body is a “perfect” emitter so that some

adjustment is needed to equation (1.6).

 A body not only emits thermal radiation, but receives thermal radiation from
other bodies – thus, equation (1.6) is modified to allow for a net energy
exchange between two bodies having different temperatures.

 Since thermal radiation is a directed form of energy transfer, the “view” that
different bodies have of each other is important – thus, equation (1.6) is
modified to include these "View Factors".

In the case of a car radiator, not only is heat transferred through the radiator
successively by forced convection, conduction and then forced convection to the
surroundings, but also it is radiated away from the hot metal to the cold
surroundings.

If the area is brought over to the LHS of equation (1.4), (1.5) or (1.6), so that rate of
heat transfer is divided by area, then this is called the “heat flux” Q f .

With heat transfer it is easier to start with heat transfer through flat (plane)
surfaces. Heat transfer through pipes is complicated by the fact that the curved
area, perpendicular to the direction of flow, continually changes.

Heat transfer is made more complex, not only by the different modes of heat
transfer, but also by whether or not these modes occur in series or in parallel.

1.4.5 Heat Loss by Radiation

In most situations, the heat transfer by radiation tends to be forgotten or at least is

thought to be negligible. In the examples below, radiation is significant. The heat
transferred due to radiation is given by:


Q  A Ts4  T f4 
where σ is the Stefan-Boltzmann constant = 5.6703x10-8 W/m2K4
T is the absolute temperature
Q is the heat loss per unit surface area (W/m2)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 13

Example : 1.4.1

Problem:

A 25 mm diameter copper pipe is used to carry hot water. The external surface of
the pipe has a convective heat transfer coefficient of h=6 W/m2 K. The external pipe
wall temperature is 80°C and the ambient air is 5°C. Calculate the heat loss due to
convection.

Solution:

As a basis, choose a 1m length of pipeline.

The heat loss by convection is given by:

Q  hc AT2  T1 

Surface area based on 1m pipeline:

A   d L    0.0251  0.079 m2

Heat loss is then:

Q  6 2  0.079m2 80  5  35.3 W
W
mK

Heat loss by radiation is given by


Q  A Ts4  T f4 

Q  5.6703108
W
2 4
mK
 
80  2734  5  2734  42.6 W

Overall heat loss is then

.
Q overall  35.3 W  42.6 W  78 W

..................................................

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 14

1.5 Process Heat Transfer

Heat is transferred in the direction of decreasing temperature when a temperature
gradient exists between different parts of a system. In mechanical engineering we
are usually concerned with conduction, convection and radiation of heat from a
solid surface into a flowing or stagnant fluid e.g. from a hot bearing into air blown
across the bearing to keep it cool. In process engineering we are concerned with
transferring heat from one hot fluid, into another cold fluid but passing the heat
through a solid barrier between the two fluids. The theory will be the same
regardless of the end application.

In many transfer processes, we model the process by considering some form of

stagnant barrier which in this case is between the solid surface and the fluid
surface. In process engineering heat transfer application, we have two interfaces,
one on either side of the solid between the hot and cold fluids (Figure 1.1).

Figure 1.1, Heat transfer through solid plates.

In Figure 1.1 we have the temperature shown in terms of position through the solid
barrier between the hot and cold fluids. Away from the boundaries, we think of the
fluid to be well mixed and the temperature is therefore uniform. The film barriers
close to the solid surface offer resistance to heat transfer so the temperature rises
or falls depending which side it is on. Note that the diagram shows a linear profile
of temperature change but this may not be the case in practice.

Also remember that the thickness of the two films will depend on the properties of
the fluid, viscosity etc., and if there are any eddy currents from fluid movement.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 15

Another example of heat transfer is shown in a concentric tube heat exchanger

(H/E) along with possible temperature profiles in Figure 1.2.

Hot Fluid In

Cold Fluid Out

Cold Fluid In

Hot Fluid Out

Temp

Distance along the heat exchanger

Figure 1.2, Concentric tube heat H/E and temperature profile.

You can tell that there is no phase change – unless there are more than one
component in which case the bubble point changes as the material evaporates or
condenses. The temperature change of the hot stream is greater than that of the
cold stream; think why this might be?

Thermodynamics tells us:

1st Law:  Enthalpyhot =  Enthalpycold + Heat Losses

Often heat losses are neglected in heat exchanger calculations.

2nd Law: Thot > Tcold at all points along the exchanger

Thermodynamics does not tell us anything about rates and so does not tell us the
size of heat exchanger required to transfer a specified amount of heat.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 16

There are three key questions for chemical engineers about heat exchangers:

 What is the area required for a given duty?

 How much heat will be transferred for a given area?

 What are the surface temperatures?

We will look at the three mechanisms of heat transfer (conduction, convection and
radiation) separately but a combination of these mechanisms is involved in most
practical situations. Now we will look quantitatively at these three modes of heat
transfer.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 17

1.6 Dimensional Analysis

Physical quantities such as thermal conductivity, heat flux, specific heat are
expressed in dimensions in the field of heat transfer. These dimensions include
time, length, mass and temperature.

You will have seen in the last year and a half, in Fluids and Heat Transfer that
dimensionless numbers are common in correlations. Have you been sleepless
with worrying where they came from? If so you should probably get out more.
However, they are found using one of two common methods of dimensional
analysis: Raleigh’s Method and the Buckingham Pi method. We’ll use the latter.

Here are the stages of the analysis:

 Assume that you will use a units system with mass, length, time and
temperature: M, L, T and θ. There are m of these; m = 3 or 4.

 Write down the dependent variable.

 Write down a list of variables that affect the dependent variable. The
number of these plus the dependent variable is n.

 Write down 3 or 4 (depending whether you have temperature) of the

independent variables according to these rules:

– The dimensions M, L,T (and θ if needed) must appear at least once.

– It must not be possible to make up a dimensionless group with these
variables.

 Work out what M, L, T and θ are equivalent to in terms of the variables.

 Divide the independent variables you have not used by their units and
substitute variables for units.

 Rearrange the dimensionless groups, if necessary, to get common

dimensionless groups.

The following two worked examples should be clearer than the above description.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 18

Example : 1.6.1

Problem:

Find a correlation to give the pressure drop in a rough pipe with turbulent flow.

Solution:

The dependent variable is the pressure drop ΔP with the unit N/m2 or kg/(m.s2).

Independent variables (that ΔP depends on) are:

Independent variables Units

Pipe diameter, d m
Fluid density, ρ kg/m3
Fluid viscosity, μ kg/(m.s)
Fluid velocity, v m/s
Pipe length, Lp m
Absolute pipe roughness, ε m

Choose 3 variables since we will use M, L and T. Not d and ε since d/ε is
dimensionless. Let’s choose d, ρ and v - these include units of M, L and T.

Write these 3 variables down in fundamental units:

d L
M
 3
L
L
v
T

Solve for M, L and T:

Ld
M   L3   d 3
L d
T 
v v
Now work out dimensionless ratios for the other variables, ΔP, μ and ε:

P P Pd d 2 P
  
kg M d 3 v 2 v 2
2
ms LT 2
  d d 
   Reynold’s Number (or at least 1/Re)
kg M d 3
v  vd
ms LT

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 19

Lp Lp Lp
  Length to diameter ratio
m L d

  
  Relative roughness
m L d

So we have 4 dimensionless groups and a correlation:

a c
P   vd      L p 
b
 Constant      
v    d   d 
2

You may see that ΔP, μ and ε each occur once while d, ρ and v can appear in more
than one dimensionless number so they are sometimes termed the recurring set.

...........................................

Example : 1.6.2

Problem:

Find a correlation to give the film heat transfer coefficient for flow across a pipe with
turbulent flow.

Solution:

A joule = N m = kg m2/s2

The dependent variable is h with the unit J/(s m2 K) or kg/(s3 K). The independent
variables (that h depends on) are

Independent variables Units

Pipe diameter, d m
Fluid density, ρ kg/m3
Fluid viscosity, μ kg/(m s)
Fluid velocity, v m/s
Fluid thermal conductivity, k J/(s m K) or kg m/(s3 K)
Fluid specific heat, Cp J/(kg K) or m2/(s2 K)
Choose 4 variables since we need M, L, T and θ. Let’s choose d, ρ, v and k - these
include M, L, T and θ. Write these down in fundamental units:

d L
M
 3
L
L
v
T

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 20

ML
k
T 3t

Solve for M, L, T and θ:

Ld
L d
T 
v v
M   L3   d 3
ML d 3dv3 dv3
  
kT 3 kd 3 k

Now work out dimensionless ratios for the other variables, h, μ and Cp:

h h hd d 3 dv3 hd
   Nusselt Number
kg M d 3 v3 k k
s 3 K T 3

  d d 
   Reynold’s Number again! Magic!
kg M d v dv
3

ms LT

Cp Cp C p d 2 dv3 C p dv
   Nameless Group
m2 L2 d 2 v2 k k
s2K T 2

There is no guarantee that we will get groups that people liked and named but,
since the groups are dimensionless, we can multiply or divide them as we wish.

Multiply Nameless Group by Reynold’s Number to get:

Cp
Prandtl Number
k

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 21

So we have 3 dimensionless groups and a correlation:

 Cp  
a b
hd   vd 
 Constant    
k     k 

.......................................

You do not need to know this but admit it; it is (almost) interesting. Dimensionless
numbers can often be considered as ratios of two forces or effects. As examples:

InertialForces
Re =
Viscous Forces

Convection Diameter
Nu = or
Conduction Thicknessof ThermalBoundary Layer

MomentumDiffusivity
Pr =
T hermalDiffusivity

Whilst relying on Wikipedia generally is a bad idea, it’s often a great starting point
for more “reliable” references. http://en.wikipedia.org/wiki/Dimensionless_Number

In evaluating dimensionless groups there is often some ambiguity in the choice of

values which must be resolved. The physical dimension for flow over plates is
generally taken as the distance along the plate, for flow in or around ducts it is the
hydraulic or equivalent diameter defined:

4 Cross sectionalarea
de 
WettedPerimeter

As expected, for a circular duct or pipe, diameter d, this is given by:

4 d 2
de  4  d
d

For a square duct, side length x,

4x2
de  x
4x
and for a rectangular duct, width a and depth b:

4ab
de 
4a  b 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 22

For flow through an annulus having inner and outer diameters d1 and d2,
respectively the hydraulic diameter may be calculated:

4  d 22  d12 
  4 d 2  d1 d 2  d1 
de  4  4  d 2  d1 
 d 2  d1   d 2  d1 

which is equal to twice the thickness of the annular gap.

Thermophysical fluid properties (density, viscosity etc.) vary with temperature. It is

important, particularly if the temperature difference between the wall and the fluid is
large, that the appropriate temperature is chosen. Normally this is the fluid bulk
temperature, but some correlations require properties to be evaluated at the mean
film temperature:

T fluid  Twall
T film 
2

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 23

1.7 Laminar and Turbulent Flow

If one imagines a stack of paper or thick card initially arranged as a rectangular

block, to be sheared as shown in Figure 1.3, it can be seen that the individual
cards, or lamina, slide over each other. There is no movement of material
perpendicular to the shear direction.

Figure 1.3, Shear applied to parallel sheets.

Similarly, in laminar fluid flow there is no mixing of the fluid and the fluid can be
regarded as a series of layers sliding past each other. If the flow is laminar a thin
filament of dye inserted in the fluid will remain as a thin filament as it follows the
flow.

Consideration of a simple laminar flow allows us to define viscosity. Figure 1.4

illustrates the velocity profile for a laminar flow of a fluid over a flat plate:

Free stream
velocity

Plate
v

Figure 1.4, Velocity profile in laminar flow over a flat

plate.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 24

The absolute or dynamic viscosity of a fluid, , is defined by:

dv
 
dy

where  is the shear stress. At the wall, the velocity of the fluid must be zero, and
the wall shear stress is given by:

 dv 
 w    
 dy  w

The kinematic viscosity of a fluid is defined:


v


In practice, laminar flow is observed at low speeds, in small tubes or channels, with
highly viscous fluids and very close to solid walls.

If the fluid layers seen in laminar flow break up and fluid mixes between the layers
then the flow is said to be turbulent. The turbulent mixing of fluid perpendicular to
the flow direction leads to a more effective transfer of momentum and internal
energy between the wall and the bulk of the fluid.

Turbulent flow is the more common regime for bulk flow in most heat transfer
equipment, but laminar flow is encountered in highly compact heat exchangers and
those handling very viscous fluids.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 25

1.8 Tutorials

1. Thermal conductivities are usually temperature dependent. Often k varies

linearly with T so that k = a + b T. Hence, starting with Fourier's Law for planar
geometry we could modify the derivation:
T2
. x


Q dx   A
0 T1
 a  bT dT
T2
.

 a  bT dT
A
Q 
xT
1
.
Show that this integrates to give Q 
k av A
T1  T2  where kav is the average of k at
x
T1 and T2.
Hint: Remember that algebraically (x2 – y2) = (x + y)(x – y)

2. Carry out a dimensional analysis of power input to stirred tank containing a

Newtonian fluid to show that the following relationship is reasonable. Assume
that all designs are geometrically similar, ie one length dimension defines the
system. Np = f(Re, Fr)
P
N p  Power Number 
N 3 d 5
Nd 2
Re  Reynold's Number 

2
N d
Fr  Froude Number 
g

(You could take power to be W/m3 and the Power Number would have d2 rather
than d5. You could also take, say, d for a vessel dimension and D for a stirrer
dimension and you would get another dimensionless ration: d/D. There is a
version of this in C&R vol 1, p283ff)

1.9 Bibliography
1. "C&R vol. 1" refers to Chemical Engineering, J.M.Coulson, J.F.Richardson et
al., vol. 1, 6th ed., Butterworth, 1999

2. "C&R vol. 6" refers to Chemical Engineering Design, R.K.Sinnott vol 6, 4th ed.,
Butterworth, 2005

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 1. INTRODUCTION TO HEAT TRANSFER 26

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 27

Topic 2

Thermal Conduction

Contents

2.1 Introduction 28
2.2 Conduction Through a Plane Wall 29
2.3 Conduction Through a Composite Wall 32
2.4 Conduction Through a Cylindrical Pipe Wall 40
2.5 Conduction Through a Composite Cylinder 43
2.6 Conduction Through a Boundary Layer 47
2.7 Tutorials 48
2.8 Bibliography 53

Prerequisite knowledge

 Thermodynamic principles and definitions.

 General knowledge about conduction, convection and radiative modes of
heat transfer.
 Dimensional analysis
 Process heat transfer
 Detailed understanding of both the steady-flow energy equation and the
non-flow energy equation.

Learning Objectives

By the end of this topic, you should be able to:

 Have qualitative ideas behind conduction.

 Know how heat is transferred and why thermal conductivity is higher in
solids than gases.
 Understand the importance of conduction in plane, spherical and cylindrical
pipe walls.
 Draw resistance network diagram for conduction.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 28

2.1 Introduction
An experimental property of materials is their thermal conductivity, k. Table 1
shows the thermal conductivities of some common materials. The carriers of the
energy transfer are molecules, atoms, electrons and photons.

Solids are better conductors than liquids, which, in turn, are better than gases.
Stainless steel is a significantly poorer conductor than ordinary steel. Water is a
better conductor than most liquids. Gas molecules at the same temperature have
1
the same average kinetic energy ( mv2 ) so hydrogen molecules, which are very
2
light, have very high velocities and can pass the energy from one molecule to
another molecule faster. "Lagging" is a layer of insulation that is put on pipes and
vessels to reduce unwanted heat loss or gain.

Conduction depends on the thermal conductivity. Typical values of common

materials are presented in Table 2.1:

Table 2.1: Typical thermal conductivities (C&R vol 1 6th ed. p389)
Material Thermal conductivity, k, W/(m K)
Copper 377
Mild steel 45
Stainless steel 16
Solids 85% magnesia lagging 0.07
Glass 1.09
Fibreglass 0.04
PVC 0.19
Sand 0.15 - 0.25
Benzene 0.16
Water 0.62
Lubricating oil 0.15
Liquids
Ethylene Glycol 0.25
Silicone oil 0.10
Air 0.024
Hydrogen 0.17
CO2 0.0146
o
Gases Steam (125 C) 0.016
Methane 0.03
Propane 0.015

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 29

2.2 Conduction Through a Plane Wall

Figure 2.1 shows the main ideas for conduction through a plane wall.

0 x

Hot
Side

T1 Cold Cold
Side
T2
Wall
Area, A
Heat
Flow, Q

Hot

Direction of Heat Flow

(Rate = Q)

Figure 2.1, Steady-state conduction through a plane wall

Note: The relevant area, A, is the area at right angles to the heat flow. The dotted
line is the temperature profile for steady state. "x" has been used for both the
direction and the thickness.

To analyse this situation we can use Fourier’s Law for heat conduction in one
dimension. For any plane in the wall parallel to the area, A:

. dT
Q  kA
dx
.
The negative sign follows from the convention that heat flow, Q is in the x
.
direction. As x increases, T decreases so that dT/dx is negative. At steady state, Q
has the same value at any x.

For a plane wall and constant k, integrating from x = 0 and T = T1 to x = x and T = T:

. x T


Q dx  kA dT
0

T1

.
 Q x  kA(T1  T )

 . 
kA .
Q
Q  (T1  T ) and T  T1    x
x  kA 
 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 30

Notice that the temperature variation is linear in x. At x = x, T = T2:

.
Q
kA
T1  T2  .............................(2.1)
x

A little test: At steady state, if k increases with increasing T, will the temperature
profile resemble (a), (b) or (c) in this figure?

T1 T1 T1

T2 T2 T2

(a) (b) (c)

Figure 2.2, Temperature profiles for k = f(T)

Fourier's Law, like many other processes, takes the form:

DrivingForce
Flow 
Resistance

Equation (1.7) can be rearranged to give:

. T1  T2
Q
x
kA
. T
or Q ........................(2.2)
R

The driving force for heat transfer, ΔT = T1 - T2 ........................(2.3)

x
The resistance to heat transfer, R  ...................................(2.4)
kA

The resistance concept is an excellent way of visualising heat transfer problems –

heat transfer is now analogous to Ohm’s Law for electricity:

 Voltage drop is analogous to temperature drop.

 Flow of electrical current is analogous to heat flux.
 Electrical resistance is analogous to thermal resistance.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 31

Through equation (2.4), the higher the thermal conductivity k (W/m K) then the
lower the resistance R .

Double glazing works on the principle that two layers of glass, separated by a thin
layer of some low conductivity gas, has a higher resistance to heat transfer than a
single pane of glass on its own.

Some people (including C&R vol. 1) do not include the area in the resistance. It is
possible to use conductance rather than resistance but it is not the conventional
approach. You would have:

Flow  Conductance  DrivingForce

kA
The conductance for heat transfer using Equation (2.1) is
x
To increase heat flow: Increase the driving force and/or Decrease the resistance.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 32

2.3 Conduction Through a Composite Wall

If we consider heat flow through a composite wall, a solid made up from two solids,
each with different thermal conductivities and of different thickness.

Figure 2.3 shows a composite plane wall. Just from looking at the diagram you
should be able to tell which layer has the largest thermal conductivity and which
has the largest resistance.

k1 k2 k3

T1

T2

T3

T4

x1 x2 x3
Figure 2.3, Steady-state conduction through a composite plane wall.

.
At steady state, the rate of heat flow, Q must be the same through each layer, so:

. k1 A
For Layer (1) Q (T1  T2 )
x1
. k A
For Layer (2) Q  2 (T2  T3 )
x2
. k A
For Layer (3) Q  3 (T3  T4 )
x3

If we look at the driving forces for heat transfer and extract the temperature
differences:

.
x1
T1  T2  Q
k1 A
. x
T2  T3  Q 2
k2 A
. x
T3  T4  Q 3
k3 A

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 33

You can see that the driving forces are proportional to the resistances.
Adding up the driving forces gives:

.  x x 
(T1  T2 )  (T2  T3 )  (T3  T4 )  T1 T 4   Q 1  2  3 
x
 k1 A k 2 A k3 A 

Therefore we have

. T1  T4 Totaldriving force
Q  ...........(2.5)
 Resistances Totalresistance

Define a term called the overall heat transfer coefficient:

.
Q  U A Toverall ........................(2.6)

where

Toverall  T1  T4

and U is the overall heat transfer coefficient, given by

1 x x x
 1  2  3
UA k1 A k 2 A k3 A

R
1
 R1  R2  R3  i ............(2.7)
U i

Just as electrical resistances in series are additive, so thermal resistances in series

are also additive; this is what makes the resistance-analogy so valuable.

The result may be extended to any number of thermal resistances in series as

follows:
RT  
Ri ……………………….(2.8)
i

Example : 2.3.1

Problem:

A furnace is constructed with 0.2 m of firebrick, 0.1 m of insulating brick, and 0.2 m
of building brick. The inside surface temperature is 1200K and the outside surface
temperature is 330K.
a) Find the heat loss per unit area of wall.

b) Find the interface temperatures between the layers.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 34

c) What thickness of insulating brick is required to reduce the heat loss to 500
W/m2?

d) Recalculate the interface temperatures if this insulation is applied.

T1 = 1200K T2 (?) T3 (?) T4 = 330K

(1) (2) (3)

Firebrick Insulating Ordinary
Brick Brick

0.20 m 0.10 m 0.20 m

k = 1.4 k = 0.21 k = 0.7 W/(m K)

Solution:

a) Total resistance,
x1 x x
R  2  3  0.905K/W
k1 A k 2 A k3 A

 U A  1.105W/K

.
 Q  U A Toverall  961W

or we could use
. T 1200 330
Q   961 W
R 0.905

b) For each layer we can write

. kA
Q T
x

Firebrick:
. k1 A
Q (T1  T2 )
x1

1.4
 961  (1200 T2 )
0.2

 T2  1063ºC

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 35

Insulating brick:
. k2 A
Q (T2  T3 )
x2

0.21
 961  (1063 T3 )
0.1

 T3  605 ºC

Check using the ordinary brick layer.

.
We could calculate Q for this layer or, as here, get another estimate of T3:

. k3 A
Q (T3  T4 )
x3

0.7
 961 (T3  330)
0.2

 T3  605 ºC

.
(c) and (d) Now increase the thickness of insulating brick. Q  500W

T1 = 1200K T2 (?) T3 (?) T4 = 330K

(1) (2) (3)

Firebrick Insulating Ordinary
Brick Brick

0.20 m ? 0.20 m

k = 1.4 k = 0.21 k = 0.7 W/(m K)

Firebrick:
. k1 A
Q (T1  T2 )
x1
1.4
 500  (1200 T2 )
0.2

 T2  1129 ºC

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 36

Ordinary brick:
. k3 A
Q (T3  T4 )
x3
0.7
 500  (T3  330)
0.2

 T3  473 ºC

Insulating brick:
. k2 A
Q (T2  T3 )
x2

0.21
 500  (1129 473)
x2

 x2  0.275m

Alternative method

Total resistance,
T 1200 330
R .
  1.74 K/W
500
Q

0.2 x2 0.2
R    1.74
1.4 0.21 0.7

 x2  0.275m

Repeat (b) to give interface temperatures:

T2  1129 ºC and T3  473 ºC.

Note: Insulating brick now has a much larger proportion of the total
resistance, so takes a larger proportion of the total temperature difference.
(T2 has increased; T3 has decreased).

..............................................

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 37

Example: 2.3.2

Problem:

The inside of a metal furnace wall, thermal conductivity 45 W/m K and thickness 5
cm, is 150oC. If there is 2.5 cm thickness of lagging, with thermal conductivity 0.05
W/m K and, if the outside lagging temperature is 30oC, then calculate the individual
resistances of each element, the overall resistance of both elements, the heat flux
through the overall system and the temperature of the outside furnace metal wall
(the same as the inner lagging temperature).

Assume steady-state operation throughout.

Solution:

The individual resistances are given

First for element “1”

x1 0.05
R1    0.001 (K m2/W)
k1 45

Then for element “2” – the lagging

x2 0.025
R2    0.5 (K m2/W)
k2 0.05

The total resistance for the combined system– the metal and the lagging

RT  R1  R2

 RT  0.5  0.001 0.501 (K m2/W)

.
The heat flux Q f (W/m2) through the combined system is given by

T  T 
Q f  1 3
 x1 x2 
  
 k1 k 2 

150  30
 Q f   239.5 (W/m2)
0.501

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 38

Notice that this calculation could be represented by the following “resistance

network”, where temperature is analogous to voltage, thermal resistance to
electrical resistance and heat flux to current flow.
See below:

This greatly helps to visualise the problem. The intermediate temperature T2 is

found in exactly the same way as if it were a missing intermediate voltage. Just as
the voltage drop across R1 would be

V1  i R1

So, the temperature drop across this first element is simply

T1  Q f R1  239.5  0.001

 T1  0.24 oC

 T1  T2  0.24 oC

 T2  149.76 oC

Notice how small the temperature drop is across the metal (very small resistance)
and how large it is across the lagging (much larger resistance).

………………………………………

Example: 2.3.3

Problem:

Using the same information as the previous problem, calculate the U -value of the
system of two resistances in series and find the rate of heat loss, to the
surroundings, through 10 m2 of lagging.

Solution:

The total resistance for the combined system – the metal and the lagging was
found to be

RT  R1  R2  RT  0.5  0.001 0.501 (K m2/W)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 39

This may be converted to a U -value if we use equation (2.7):

R
1
 i
U i

1
U   2 W/m2 K
0.501

Then the total rate of heat loss through 10 m2 of lagging area is given by equation
(2.6):

Q  UATT

 Q  2  10  (150  30)

 Q  2.4 kW

Each square metre of lagging loses about 0.24 kW or approximately 240 W, which
is the same as 240 J every second. This represents a large heat loss and cost per
annum, which is why the performance of lagging is so important.

When solving heat transfer problems the resistance network approach is used to
find the total thermal resistance. Once this is known it is converted into a U -value,
using equation (2.7), then equation (2.6) may be solved.

In many heat exchanger design problems it is the heat transfer area A (m2) that is
needed. This area determines the size of the heat exchanger.

.............................................

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 40

2.4 Conduction Through a Cylindrical Pipe Wall

We saw that in a plane wall the surface area for the heat transfer rate was
constant. In a cylindrical pipe wall, the surface area is a function of the radius r. It is
very important to define the surface area in which heat transfer occurs and take it
into account when calculating the overall heat transfer coefficient. Figure 2.4 shows
the temperature profile for conduction through a tube wall. The radius, r, is
measured from zero at the centre to r1 and r2 at the inner and outer walls.

Cold Hot

T1

T2

r1 r2

Figure 2.4, Conduction through thick-walled cylinder or sphere.

Fourier’s Law can be written for any radius:

. dT
Q  kA
dr

In this case A is the area of a cylinder

A  2 r L

. dT
Q  k 2 rL
dr

r2 2 T
. dr
Q 
r1
r
 k 2 L dT
T

1

. 2 kLT1  T2 
Hence after integration: Q ..........................(2.9)
 r1 
ln 
 r2 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 41

Figure 2.4 can also represent conduction through a thick-walled spherical shell.
Starting again from Fourier’s Law:

. dT
Q  kA
dr

and in this case: A = 4πr2

This leads to:

. 4 k T1  T2 
Q .....................(2.10)
1 1

r1 r2

Note that the temperature profile is linear for planar conduction but curved for tubes
dT .
and spheres since, as A increases, must decrease to keep Q constant. How
dr
could you use equation (2.9) or (2.10) to plot the temperature profile?

Table 2.2 summarises resistance and heat flow equations for different geometries.

Table 2.2, Resistances for different geometries.

Resistance Heat flow
Plane x . kA
Q T
kA x
r  . 2kL
Tube ln 2  Q T
 r1   r2 
ln 
2kL  r1 
1 1 . 4k
Shell  Q T
r1 r2 1 1

4k r1 r2

Now have a look at Example 2.3.4 on heat loss by conduction from a pipe.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 42

Example : 2.3.4

Problem:

Find heat loss/unit length of pipe for the following situation:

Copper pipe k = 400 W/(m K), 10 cm id, 12 cm od

Insulation k = 0.20 W/(m K), 5 cm thick
Inside Temperature of pipe 110oC
Outside Temperature of insulation 30oC

Solution:

For 1 m length of pipe the total resistance is:

 0.06   0.11
ln  ln 
Re   0.05 
  0.06 
 0.4824K/W
2π 400 x 1 2π 0.2 x 1

. TotalDriving Force 110  30

Q   165.8 W
TotalResistance 0.4824

Alternatively,

1
UA   2.073 W/K
R

.
Q  UA T  2.073 110  30  165.8 W

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 43

2.5 Conduction Through a Composite Cylinder

In industrial applications tubes and pipes often consist of multiple layers. Figure 2.5
shows a composite cylinder with a wall consisting of 4 layers and each layer has a
different thickness and thermal conductivity. Consider this composite cylinder with
the temperatures on the boundaries as T1, T2, T3 and T4.

Figure 2.5, Heat transfer through a composite cylinder.

We have the equation for the heat transfer rate as:

. 2 k1 LT1  T2  2 k 2 LT2  T3  2 k3 LT3  T4 

Q  
 r1   r2  r 
ln  ln  ln 3 
 r2   r3   r4 

By analogy with the composite plane wall we have the combined resistances to
heat transfer as:

. 2 LT1  T4 
Q
1  r1  1  r2  1  r3 
ln   ln   ln 
k1  r2  k 2  r3  k3  r4 

If the inner and outer layers were fluid, e.g., hot water flowing through a pipeline
and cold oil flowing over the outside of the pipe, we can replace two of the terms by
film coefficients:

.
Q
T1  T4 
1 1 r  1
 ln 2  
A1h1 2Lk 2  r3  A2 h2

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 44

The two area terms are included here as the surface areas at the two radiuses are
not the same. Unless there is need to analyse heat transfer in fine detail, as in
modelling a heat transfer device, then the area term can be resolved. If a rough
sizing is needed, then a mean area term can be used.

Example : 2.5.1

Problem:

A pipe carries saturated steam at 110°C. The external air temperature is 30°C.
Pipe dimensions: id = 10cm, od = 12cm
Pipe wall thermal conductivity = 400 W/mK
Inside film coefficient = 1000 W/m2K
Outside film coefficient = 15 W/m2K
Thermal conductivity of the lagging = 0.2 W/mK
a) Investigate the effect of the lagging thickness on the heat loss from the steam
to the surrounding air.
b) If the steam flowrate is 0.15 kg/s, how much lagging is required to prevent no
more than 5% of the steam condensing in a 100m run of pipe?
Solution:

From:
.
Q
T1  T4 
1 1 r  1
 ln 2  
A1h1 2Lk 2  r3  A2 h2
We can add in the additional solid layer to deal with the lagging as:

.
Q
T1  T4 
1 1 r  1  rlaggingod  1
 ln od   ln  
A1h1 2Lk pipe  rid  2Lk lagging  rod  A2 h2

a) With no lagging, based on a 1m length of pipe we have:

1 1
  0.00032
A1h1 2  0.05  1  1000

1 1
  0.1768
A2 h2 2  0.06 115
1 6
ln   7.254  105
2  1  400  5 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 45

Summing these terms gives

1 1 1 r  1
  ln od  
UA A1h1 2Lk pipe  rid  A2 h2

1
= 0.00032 + 0.1768 + 7.254 ´10-5
UA

1
 0.17723
UA

Therefore, for a 1m length of pipe

.
Q  U A T1  T4    110  30  451W
1
0.17723

In other words, the outside film is controlling the process and reducing the
heat transfer.

b) The impact of lagging is best shown by adding in the lagging and plotting the
heat loss from the pipe

.
Q
110  30
1  rlagging od 
0.00032 7.254 105  ln   0.1768
2  1 0.2  0.06 

Lagging Lagging Outside Heat % of total

Thickness Resistance Film Transferred resistance in
Resistance the lagging
(cm) (W/m)
0 0.0000 0.1768 451.4 0
1 0.1227 0.1516 291.3 45
2 0.2289 0.1326 221.0 63
3 0.3227 0.1179 181.4 73
4 0.4065 0.1061 155.9 79
5 0.4823 0.0965 138.1 83
6 0.5516 0.0884 124.9 86
7 0.6153 0.0816 114.7 88
8 0.6743 0.0758 106.6 90
9 0.7292 0.0707 100.0 91
10 0.7805 0.0663 94.4 92

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 46

Figure 2.6 shows that the percentage of heat resistance in the lagging goes
through a maximum deflection point around 2 cm. This introduces the concept of
an optimal thickness of lagging, where adding more does not appreciably change
the heat loss.

100
90
80
70
60
% Resistance in Lagging

50
40
30
20
10
0
0 2 4 6 8 10 12
Lagging Thickness (cm)

Figure 2.6, Graphical results from worked example

Figure 2.7 shows a typical optimization profile looking at the cost of energy lost,
the cost of lagging. The overall cost has a minimum point but one that is difficult
to locate – somewhere between 2.5 and 4 cm.

Figure 2.7, Optional lagging thickness.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 47

2.6 Conduction through boundary layer

For conduction through the thermal boundary layer, if we consider it to be a slab:

. kA
Q (TS  TF )


But since the boundary layer thickness, , is not readily estimated or measured, it is
generally combined with k:

.
Q  h A (TS  TF )
where:
k
h  Film heat transfercoefficient ( W/(m2 K))

 1
Film resistance   (K/W)
kA hA

In a simpler picture illustrated in Figure 2.8, all resistance to heat transfer can be
considered to be concentrated in a "thermal boundary layer" in which the sole
heat transfer mechanism is conduction.

Bulk
Fluid

Thermal Boundary
Layer 

Solid Surface
TS TF
Temperature
Profile

Figure 2.8, Conduction through boundary layer

Notice that thermal conductivity, k, is a physical property but film heat transfer
coefficients, h, depend both on the fluid and the flow geometry. If a fluid is flowing
over the outside of a pipe, the local value of h will be different at every point on the
pipe; high facing the flow and low on the side sheltered from the flow. The value of
h given in correlations is an average value. Film heat transfer coefficients are found
from empirical correlations, which vary depending on the geometry.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 48

2.7 Tutorials
1. Find the heat flux through a brick wall 0.5 m thick when the inner surface
is at 400 K and the outside at 310 K. Thermal conductivity of the brick =
0.7 W/(m K)

Ans: 126 W/m2

2. Calculate the heat loss through a brick wall, 2 m high, 3 m wide and 0.15
m thick, when the inner surface is kept at 400oC and the outer surface at
50oC. Also calculate the temperature half way through the brick.
Assume that the thermal conductivity of brick is constant at 0.70 W/(m
K).

Ans: 9800 W, 225ºC

3. Glass for windows is 5 mm thick and can be fitted single-glazed or

double-glazed with a 20 mm air gap between the panes. The thermal
conductivities of the glass and air are 0.9 W/(m K) and 0.03 W/(m K)
respectively. Find the overall heat transfer coefficients for single-glazed
and double-glazed windows neglecting the convection on either side.
Find also the heat loss per square metre of window for both cases when
the inside temperature is 20oC and the outside temperature is 0oC.
.
Ans: Single-glazed: U = 180 W/(m2 K) Q = 3.6 kW
.
Double-glazed: U = 1.475 W/(m2K) Q = 29.5 W

4. a) A furnace is constructed with a wall of 15 cm thick firebrick, 10 cm of

insulating material and 15 cm of building brick. The inside wall
temperature is maintained at 1000oC and the outside wall temperature at
50oC. If the heat loss through the wall is 1.0 kW/m2, calculate the thermal
conductivity and the inner and outer temperatures of the insulating
material. k for firebrick = 1.4 W/(m K), k for building brick = 0.7 W/(m K)

b) A new furnace is to be designed with the same materials and with the
same inside and outside wall temperatures but with the heat loss
reduced to 0.5 kW/m2. If the thickness of building brick is still to be 15
cm, find the thicknesses of the firebrick and insulation layers such that
the maximum temperature of the insulation is 900oC.

Ans: a) 0.159 W/(m K), 893oC, 264oC

b) firebrick 28.0 cm, insulation 23.6 cm
5. Insulation must be selected for a large process vessel, operating at
230oC. Its external surface area is 60m2 and, at an air temperature of
15oC, the heat loss should be restricted to 6.5 kW.
The vessel wall resistance and inside and outside film resistances may
be neglected. The vessel is sufficiently large to be considered as a flat
surface.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 49

Two insulating materials, some properties of which are shown in the

table, may be used either singly or in combination.

Material 1 Material 2
Thermal Conductivity, W/(m K) 0.15 0.22
Maximum Temperature 270ºC 180ºC
Cost, £/m3 340 220

a) What is wrong with using only material 2?

b) What is the cost of insulation if only material 1 is used?
c) What is the cost of insulation if the minimum permissible thickness of
material 1 is used with a further covering of material 2?

Ans: b) £6079 c) £5835

6. A process vessel, of external surface area 15 m2, has a 5 mm thick steel

wall. Inside the steel wall is a 2 mm thick, corrosion resistant glass lining.
Outside the steel wall is a 2.5 cm thick layer of lagging material of
unknown thermal conductivity. Under normal process conditions, the
inside temperature of the glass lining is 95oC and the outside
temperature of the lagging is 17oC. A heat balance on the vessel shows
the heat loss through the wall to be 6.5 kW. The thermal conductivities of
steel and glass are 45 and 1.2 W/(m K) respectively.

a) Determine the thermal conductivity of the lagging and the highest

temperature to which it is subjected.

b) It is proposed to halve the heat loss by adding a further layer of a

different lagging with a thermal conductivity of 0.10 W/(m K). What
thickness of this lagging is required if the outside temperature of the
lagging is still to be 17ºC?

c) If the new lagging costs £450/m3 and energy savings are worth
£0.014/MJ, how many hours of operation are required to recover its
cost?

Ans: a) 0.140 W/(m K), 94.2oC b) 0.018 m c) 742 h

7. a) A spherical storage tank of diameter 3 m contains liquid at 120oC. It is
to be lagged with material of thermal conductivity 0.15 W/(m K).
Assuming that all of the heat transfer resistance lies in the lagging,
calculate the thickness required to restrict heat loss from the tank to 2.0
kW, when the outside temperature of the lagging is 15oC.

b) What limiting heat loss is approached as very large insulation

thicknesses are applied? (Hint: what happens to the equation if r2
becomes very large?)

Ans: a) 0.26 m b) 0.3 kW

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 50

8. Repeat tutorial exercise 2. This time, assume that the thermal

conductivity of brick varies with temperature, T ºC: k = 0.60 + 0.0005 T
W/(m K). Use the equation below:
.
Q  av T1  T2 
k A
x

Ans: 9975 W, 236oC

9. Heat is transferred from one fluid stream to another across a flat metal
heat transfer surface that is 6 mm thick. Calculate the overall heat
transfer coefficient.

Film heat transfer coefficient for the hot fluid = 1.0 kW/(m2K)
Film heat transfer coefficient for the cold fluid = 1.5 kW/(m2K)
Metal thermal conductivity = 20 W/(m K)
One side is clean but the other has a dirt coefficient of 850 W/(m2K).
Ans. U = 318.2 W/(m2K)

10. The graph below shows some temperatures in a two-layer wall of area
15 m2 and its adjoining boundary layers. The heat transfer rate through
the wall is 6000 W and the thermal conductivity of layer A is 0.25 W/(m
K).

a) Calculate the thermal conductivity of layer B.

b) Calculate the unknown temperature on the inside of layer, A.
c) Calculate the film heat transfer coefficient for each of the boundary
layers.
d) Assuming the inside temperature to remain at 150oC, determine the
effect on heat transfer rate and on each of the three inter-layer
temperatures of a halving of the outside film coefficient with a
simultaneous increase in the outside temperature to 25oC.
? 110oC 40oC

Inside A B Outside
150oC 15oC

2 cm 5 cm

Ans: a) 0.286 W/(m K)

b) 142ºC
c) inside 50, outside16 W/(m2 K)
d) 4690 W, 143.7, 118.7 and 64ºC

11. A pipe carrying steam has an outside diameter of 60.3 mm and is insulated
with a 10 cm thick layer of glass wool. The outside surface of the pipe has a

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 51

temperature of 177oC, while the outside surface of the insulation is 51oC.

Determine the heat loss per metre of pipe length.
Thermal conductivity of glass fibre = 0.074 W/(m K)
.
Ans: UA = 0.318 W/K Q = 40.1 W

12. A 2 in. schedule 40 steel pipe (6.033 cm OD) is covered with a 1.5 cm layer
of asbestos insulation followed by a 2.0 cm layer of fibre-glass insulation.
The temperature of the pipe wall is 270oC and the outside surface
temperature of the insulation is 25oC. Estimate the UA value, the heat loss
per metre of pipe, and the temperature at the interface between the
asbestos and fibre-glass.
Thermal conductivity of fibre-glass = 0.074 W/(m K)
Thermal conductivity of asbestos = 0.17 W/(m K)
.
Ans: UA = 0.858 W/K Q = 210 W T = 191oC

13. A steam pipe has a diameter of 50 mm ID and 100 mm OD and is covered

with a 15 mm layer of lagging. Estimate the UA value and the heat loss per
metre length of pipe. Neglect convection coefficients.
klagging = 0.04 W/(m K)
kmetal = 50 W/(m K)
Steam temperature = 150oC
Surface temperature of lagging = 20oC
.
Ans: UA = 0.956 W/K Q = 124.3 W

14. A 5 in. schedule 80 steel pipe (14.130 cm OD, wall thickness 0.953 cm) is
covered by a 3.0 cm layer of high-temperature insulation, followed by a 3.0-
cm layer of 85% magnesia. The temperature of the inside surface of the
steel pipe is 340oC and the temperature of the outside surface of the
magnesia is 22oC.
Thermal conductivity of steel = 43 W/(m K)
Thermal conductivity of high temperature insulation = 0.087 W/(m K)
Thermal conductivity of 85% magnesia = 0.064 W/(m K)
a) What is the rate of heat loss per metre length from the insulated pipe?
.
Ans: UA = 0.771 W/K Q = 245.2 W
b) If the two insulating materials were interchanged what would be the heat
.
loss per metre of pipe? Ans: UA = 0.736 W/K Q = 234.2 W

15. A 24 m high vertical stack has an inside diameter of 0.5 m. It is constructed

with an inside layer of heat-resistant brick that is 0.3 m thick, a 0.3 m thick
layer of building brick, and a 0.01 m thick outer steel shell. Air at 1 atm
pressure and 540oC enters the base of the stack at a velocity of 3 m/s. The
average temperature of the inside surface of the stack is 370oC and that of
the outside surface of the steel shell is 27oC.

Thermal conductivity of heat resistant brick = 2.6 W/(m K)

Thermal conductivity of building brick = 0.87 W/(m K)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 52

Thermal conductivity of steel shell = 62 W/(m K)

Specific heat capacity of air = 1.0 kJ/(kgoC)

Find:

a) The heat loss per metre of stack.

b) The total heat loss from the stack.
c) The mass flow rate of the air.
d) The temperature at which the air leaves the stack?

Ans: a) 2681 W (UA = 7.817 W/K), b) 64.35 kW, c) 0.256 kg/s, d) 288.7oC

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 53

2.8 Bibliography

1. "C&R vol. 1" refers to Chemical Engineering, J.M.Coulson, J.F.Richardson et

al., vol. 1, 6th ed., Butterworth, 1999

2. "C&R vol. 6" refers to Chemical Engineering Design, R.K.Sinnott vol 6, 4th ed.,
Butterworth, 2005

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 2. THERMAL CONDUCTION 54

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 55

Topic 3

Convection and Radiation

Contents
3.1 Introduction 57
3.2 Convection 59
3.3 Lagging 64
3.3.1 Optimum lagging thickness 64
3.3.2 Critical lagging thickness 64
3.4 Resistances 65
3.4.1 Resistances in Series 65
3.4.2 Individual Resistances 67
3.4.3 Case Study – Jacketed Vessel 68
3.4.4 Fouling Resistances 70
3.5 Film Heat Transfer Coefficient 72
3.5.1 Forced Convection Inside Cylindrical Tubes 73
3.5.2 Forced Convection Outside Cylindrical Tubes 75
3.5.3 Forced Convection Over Flat Plates 78
3.5.4 Single Phase Film Coefficient 79
3.6 Radiation 80
3.6.1 Energy Emitted by a Black Body (Plank’s Distribution Law) 81
3.6.2 Stefan-Boltzmann Law 82
3.6.3 Emissivity and Absorptivity Kirchoff’s Law 84
3.6.4 Net Heat Transferred by Radiation 85
3.7 Tutorials 88
3.8 Bibliography 92

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 56

Prerequisite knowledge

 Detailed understanding of the heat transfer processes by conduction.

 General knowledge about convection and radiation.
 Understand fluid mechanics of a boundary layer.
 Turbulent and laminar layers in fully developed flow.

Learning Objectives

By the end of this topic, you should be able to:

 Differentiate between convection and radiative modes of heat transfer.

 Outline main facts concerning radiative heat transfer.
 Understand the application of forced convection inside and outside pipes.
 Be familiar with overall heat transfer coefficient and film heat transfer
coefficient.
 Be familiar with Plank’s Distribution Law and Stephen-Boltzmann Law.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 57

3.1 Introduction

In this Topic we review the heat transfer processes by convection and radiation.
In convection, heat transfer is due to bulk movement of a fluid (eddies or circulating
currents). See also C&R vol 1 p 414. The idea of laminar and turbulent flow lends
itself to a distinction between natural convection and forced convention. However, a
natural convection process can be turbulent but only by the nature of the flow.

Natural Convection - is where the temperature change as a result of heat transfer

causes parts of the fluid to change density and become more buoyant than denser
colder fluid. For instance, the visible shimmer that one sees from the top of a
radiator that is hotter than its surroundings is a sign of lighter hotter gas rising in
colder gas. This movement causes eddies which exchange energy between the hot
and cold regions of the fluid by a natural process.

Forced Convection - is where there is a deliberate flow of fluid over a heat

transfer surface. The transfer of energy is enhanced by the mixing processes that
are generated by the forced movement. It should come as no surprise that the heat
transfer characteristics of laminar and turbulent flows are very different. In forced
convection the magnitude of the Reynolds number provides an indication of
whether the flow is likely to be laminar or turbulent.

Thermal radiation is another mode of heat transfer. All bodies emit heat by thermal
radiation. Thermal radiation is part of the continuous electromagnetic spectrum and
is defined as that electromagnetic radiation between wavelengths 0.1 μm and 100
μm.

Radiant energy emissions are propagated electromagnetically with a maximum

velocity of 3 x 108 m/s in a vacuum. Conduction and convection require some
physical medium – radiation can occur in a complete vacuum.

Usually thermal radiation is not visible (except when using thermal imaging
techniques), but if the temperature of a body is hot enough it will emit part of its
radiant energy as visible light.

When sufficiently hot the object will first glow cherry red in colour but, if its
temperature is high enough, it will appear white – so-called “white hot”.

If an object absorbs thermal radiation it will heat up. In the Middle East cars
become very hot in direct sunlight, which is why most drivers in hot countries try to
park in the shade.

An object may absorb, reflect or transmit thermal radiation. For instance, many
gases tend to both absorb and reflect very little thermal radiation – mostly they just
transmit thermal radiation (the exceptions are “greenhouse” gases).

Solids, on the other hand (except for glass), tend to reflect or absorb thermal
radiation, since they transmit little or no radiation they are “opaque”. The way a
solid behaves is determined by its surface characteristics.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 58

A perfectly black object, called a “perfect blackbody”, will absorb 100% of all
incident radiation. Matt black objects behave very like “perfect blackbodies”;
however, no body is really perfectly black.

Black objects are good absorbers of thermal radiation. They are also good emitters
of thermal radiation for the following reasons:

 Whenever a photon of thermal radiation is incident to a surface, an electron

on the surface will absorb that photon.

 This causes the electron to rise briefly to a higher energy level.

 When the electron falls back to its original energy level thermal energy is
once again released.

 Since one process is the reverse of the other, it follows that good absorbers
make good emitters.

 Black objects make better absorbers (and emitters) than non-black objects,
simply because fewer photons are reflected.

 The emitted radiant energy has a different spectrum of wavelengths when

compared to the incident radiation – the emitted spectrum depends on the
body’s temperature and it surface characteristics.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 59

3.2 Convection
For flow over a flat plate, as shown in Figure 3.1 we may determine whether the
flow in the boundary layer is likely to be laminar or turbulent by applying the
following conditions:

 v d 
Re x     105 Laminar flow
  

 v d 
Re x      106 Turbulent flow
  

where x is the distance from the leading edge of the plate

V

Laminar Transition Laminar

Turbulent
sublayer

Figure 3.1, Development of a boundary layer over a flat plate.

For values of Reynolds number between 105 and 106 the situation is complicated by
two factors.

Firstly, the transition is not sharp, it occurs over a finite length of plate. In the
transition region the flow may intermittently take on turbulent and laminar
characteristics. Secondly, the position of the transition zone depends not only upon
the Reynolds number, it is also influenced by the nature of the flow in the free
stream and the nature of the surface. Surface roughness or protuberances on the
surface tend to trip the boundary layer from laminar to turbulent.

Once the flow is fully developed, as would be some distance into a pipeline for
example, then the boundaries between the laminar sub layer and the turbulent core
will be uniform.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 60

Figure 3.2 illustrates what we imagine the temperature and velocity may be doing
in these regions.

Turbulent Core

Turbulent Boundary
Layer

Laminar Sublayer

Solid Surface 0 V TS TF

Velocity Temperature
Profile Profile

Figure 3.2, Turbulent and laminar layers in fully developed flow.

For flow in pipes, channels or ducts the situation is similar to that for a flat plate in
the entry region, but in long channels the boundary layers from all walls meet and
fully developed temperature and velocity profiles are established.

For fully developed flow in pipes or channels the transition from laminar to turbulent
flow occurs at a Reynolds number, based on the channel hydraulic diameter, of
approximately 2000. As with the boundary layer on a flat plate, the transition may
occur at higher or lower values of Red.

If the flow at entry to the channel contains no turbulence and the channel is very
smooth, laminar flow may be sustained up to Reynolds numbers of 5-10000.
Turbulence may occur at values of Red as low as 1000, but at low Reynolds
numbers will decay if induced by, for example, sharp corners.

As we shall see, heat transfer coefficients are generally higher in turbulent flow
than in laminar flow, and higher in the entry region than in the fully developed
region. Heat exchanger designers may therefore incorporate features, which either
promote turbulence or lead to a geometry which approximates to many short
channels.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 61

Consider two fluids separated by a plane solid wall. Figure 3.3 shows the
temperature profile. The temperature is constant in the well-mixed bulk fluid
regions.

Boundary Wall Boundary

Layer 1 Layer 2

h1 kW h2

TF1

Bulk TW1 Bulk

Fluid 1 Fluid 2
TW2

TF2

X
Figure 3.3, Convection on both sides of a plane solid wall

For heat transfer rate we have:

.
Q
T otalDriving Force

TF1  TF 2 
T otalResistance 1 x 1
 
h1 A kW A h2 A

For a similar situation but in a pipe of length L and with fluid 1 inside and fluid 2
outside:

Fluid 2 T2 Bulk fluid

Boundary layer

Pipe wall

Boundary layer

Fluid 1 r1 r2

Bulk fluid
T1
PIPE

Figure 3.4, Convection on both sides of a cylindrical pipe wall

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 62

For heat transfer rate we have:

. T1  T2
Q
 r  
 ln 2  
  1
1  1 r 1 
2L  h1r1 kW h2 r2 
 
 

Example : 3.2.1

Problem:

A pipe carries saturated steam at 110ºC. The external air temperature is 30ºC.

Pipe ID = 10 cm, OD = 12 cm, kw = 400 W/(m K)

Inside Film h1 = 10000 W/(m2 K)
Outside Film h2 = 15 W/(m2 K)
Lagging kL = 0.20 W/(m K)

a) Investigate the effect of lagging thickness on heat loss.

b) If the steam flow is 0.15 kg/s, how much lagging is required to prevent more
than 5% of the steam condensing in a 100 m run of pipe?

Solution:

a) There are four resistances:

. T1  T2
Q
 r  r  
 ln 2  ln 3  
1  1
  r1 
  r2 

1 
2L  h1r1 kW kL h2 r2 
 
 

With no lagging (r3 = r2):

The resistances (for 1 m of pipe)= 0.00032 + 0.00007 + 0.17684 = 0.17723 K/W

Q 
. 110  30  451W per meterlength of pipe
0.17723

With 5 cm of lagging:

Lagging resistance = 0.482 K/W

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 63

Outside film resistance = 0.096 K/W

(Because of the lagging, this is now at radius 0.11 m rather than 0.06 m.)
.
Hence Q  138 W per metre length of pipe.

You can see that, in this case, the resistances of the inside film and pipe
are so small they may be neglected.

From more calculations you can get Figure 3.5.

500
450
Heat Flow, W/m

400
350
300
250
200
150
100
50
0
0 2 4 6 8 10

Lagging Thickness, cm

Figure 3.5, Effect of Lagging Thickness on Heat Flow

b) Latent heat of saturated steam at 110oC = 2230 kJ/kg (from steam

tables)

Permitted heat loss for 100 m = 0.15 x 0.05 x 2230 = 16.725 kW

Permitted heat loss for 1 m = 167 W/m

Therefore, from Figure 14 or by calculation using something like goal seek

or solver, 3.5 cm of lagging will be sufficient.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 64

3.3 Lagging
3.3.1 Optimum Lagging Thickness

As more laggings are added the cost of the lagging increases. At first the energy
saving is large but you get a diminishing return as you add more lagging. Figure 3.6
shows how you might find the optimum lagging thickness – about 4 cm in this case.
In this graph the cost values are not important, it is the trend that is relevant.

5000
4500
4000
3500
Annual cost

3000 Cost of lagging

2500 Cost of heat loss
2000 Total annual cost
1500
1000
500
0
0 2 4 6 8 10
Lagging thickness, cm

Figure 3.6, Optimal lagging thickness.

3.3.2 Critical Lagging Thickness

As shown in Figure 3.7, in some rather special cases (very small diameter pipes
and/or relatively poor insulating material), heat loss can increase at first with
increasing lagging thickness before eventually passing through a maximum and
decreasing in the usual way. The thickness at which heat loss is maximised is
known as the critical thickness. If the only significant resistances are those due to
the lagging and the outside convective film, the corresponding outside radius of the
lagging corresponding to the critical thickness can be shown to be k/h. If, as is
usually the case, k/h < outside pipe radius, you do not see this effect – the heat
loss decreases continuously with increasing lagging thickness.

tcrit Lagging Thickness

Figure 3.7, Critical lagging thickness.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 65

3.4 Resistances
There are many systems in which a driving force causes a flow. There will also be
a resistance to this flow. The inverse of resistance is conductance.

3.4.1 Resistances in series

In many heat transfer systems there can be up to five resistances in series, each
with the same heat flux passing through them, as follows:

Table 3.8 shows several examples, which are simplified here; some are better
presented as derivatives.

Table 3.1, Applications of the idea of flux, resistance and driving force.
Flux Resistance Driving Force
DC Electricity V R V
I
R
Planar Heat conduction . k x T
Qf  T
x k
Heat convection . 1 T
Q f  hT
h
Mass diffusion D x C
F C
x D
Momentum  x v
R v
x 

Usually the area, A, will be included in the thermal resistance but it is not included it
in the above table just to be consistent. In any case, since the table gives heat flux,
the area would equal 1 m2.

In many cases the resistances are typically:

 Inner convective resistance – “inner side” of tube/vessel wall.

 Inner conductive dirt/scale resistance – “inner side” of tube/vessel.
 Conductive resistance through tube/vessel wall.
 Outer conductive dirt/scale resistance – “outer side” of tube/vessel.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 66

 Outer convective resistance – “outer side” of tube/vessel wall.

It is usual that only T1 and T6 are known, which is why the overall resistance RT
and the U -value are needed – using these overall values eliminates all the
unknown intermediate temperatures.

The greater part of solving heat exchanger problems lies in calculating all these
resistances and then combining them into an overall U -value.

Allowances must also be made for the fact that dirt and scale accumulates on
surfaces over time – thus, their resistances change with time.

Now consider the double-pipe heat exchanger in Figure 3.9. Steam in the outer
pipe heats water in the inner pipe.

steam

water

Figure 3.9, Double-pipe heat exchanger.

Heat is transferred from the steam through a layer of condensate, through a layer
of dirt and rust, through the pipe wall, through another layer of dirt, and finally
through a film of water into the bulk water. There is a series of resistances to heat
transfer:

 Condensing film
 Dirt
 Metal
 Dirt
 Water film

In many cases the dirt (also called "scale") resistances will be negligible. We need
a way of dealing with resistances in series and it may be easier to look, at first, at a
case where only conduction is involved, as in Figure 3.9.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 67

Notice that in Figure 3.10 we assume that materials in contact are at the same
temperature - in this case both metal and lagging are at temperature T i at the
interface between them.

We do the same in the case of convection – the solid and the fluid touching it are
assumed to be at the same temperature. km and kl are the thermal conductivities of
the metal and lagging.

metal Ti lagging

Heat Flux, Qf km = 45 W/(m K)

kl = 0.05 W/(m K)

Th = 150oC Tc = 30oC

Xm Xl
5 cm 2.5 cm

Figure 3.10, Heat transfer through a lagged planar wall

3.4.2 Individual Resistances

At steady state,
Heat flux through metal = Heat flux through lagging

. (Th  Ti ) (T  Tc )
Q f  km  kl i
xm xl
.
Q f  Heat flux (W/m2 )

For values in Figure 3.8:

(150  Ti ) (T  30)
45  0.05 i
0.05 0.025

Giving Ti = 149.733oC

. (Th  Ti ) (150  149.733)

And now Q f  km  45  240 W/m2
xm 0.05

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 68

3.4.3 Case Study – Jacketed Vessel

Consider a more realistic heat transfer situation:

Agitator

Steam
IN

Condensate
OUT

Heat is transferred from the steam jacket (located on the outside of the vessel) into
the agitated liquid.

There are now five resistances in series:

 Steam-side film resistance (convective resistance R1 ).

 Steam-side dirt factor resistance (conductive resistance R2 ).
 Vessel wall resistance (conductive resistance R3 ).
 Liquid-side dirt factor resistance (conductive resistance R4 ).
 Liquid-side film resistance (convective resistance R5 ).

All these resistances exist between the bulk of the steam in the outer jacket,
through the jacket wall to the bulk of the liquid within the agitated vessel. The only
known temperatures are T1 & T6 - see network below:

In heat transfer problems the “jacket-side” is often labelled “outer” and the
“process-side” is often labelled “inner”. Notice that the heat flux is the same
through each thermal resistance – the heat must pass through each in turn.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 69

Just as in electrical circuits, the total thermal resistance is found by adding the
individual resistances together (this applies to resistances in series):

RT  R
i
i

 RT  R1  R2  R3  R4  R5

1 xdS xV xdl 1
 RT     
hS k dS kV k dl hl

The thickness of the dirt factors are never specified so they are just left as Rdo
(outer dirt resistance) and Rdi (inner dirt resistance), leading to

1 x 1
RT   Rdo  V  Rdi 
hS kV hl
Tidy up the notation to outer film coefficient ho (the service fluid in the jacket need
not always be steam) and inner film coefficient hi (within the vessel) and then drop
the “vessel” subscripts to get
1 x 1
RT   Rdo   Rdi 
ho k hi
Changing from overall resistance RT to overall U -value, leads to

1 1 x 1
  Rdo   Rdi 
U ho k hi
Where,
ho  The outer film coefficient for steam (W/m2 K)
hi  The inner film coefficient for liquid (W/m2 K)
Rdo  The outer dirt factor for steam-side (K m2/W)
Rdi  The inner dirt factor for liquid-side (K m2/W)
x  The thickness of the vessel wall (m)
k  The thermal conductivity of the vessel wall (W/m K)

Once the U -value is calculated, then: Q  UATT

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 70

3.4.4 Fouling Resistances

So far we have considered conduction and convection resistances. Another term is

often used to account for the resistance of solid deposits on the heat transfer
surfaces.

Typical values for these "fouling resistances" (also called scale resistances or dirt
factors) per unit area are quoted in C&R vol. 1 pp518 and 640. They can be
expressed as resistances, R values, or coefficients, hd values. For example, for
liquid organics R = 0.2 (m2 K)/kW or hd = 5 kW/(m2 K).

You should be able to tell whether a quoted fouling value is an R or an hd by

checking the units. In this course we will use the hd form;
1
If you find R values in the literature, use hd  .
R
Fouling can come from several sources but amongst these are:

 Sedimentation Fouling – mainly river streams and cooling water pipelines that
contain suspended solids that settle on the heat transfer surface. These
deposits do not adhere to the surface and can be removed by cleaning.
However, if one of the surfaces is hot then solids can “bake on” causing a more
difficult problem to remedy.

 Inverse Solubility Fouling – In aqueous systems that contain salts, some salts
can come out of solution due to a chemical or thermal change in the system.
Usually solubility of salts will be higher was the water temperature increases. If
there is an attempt to cool a solution to below the solubility point, then salts will
deposit usually on metal surfaces leaving to fouling. Scale formed by this
mechanism is usually very strong and difficult to remove. In hydrocarbons, a
temperature change may cause wax deposits to be formed as longer chain
hydrocarbons come out of solution. Chemical inhibitors can sometimes be
employed to halt the nucleation of solids.

 Chemical Reaction Fouling – where a change in the solution composition due

to a chemical reaction causes the solubility of some components to change and
hence come out of solution.

 Corrosion Fouling – is associated with the corrosion of the heat transfer

metal. In some cases this corrosion layer is protective and reduces further
metal loss however it may result in a layer of material that has poor heat
conductivity characteristics.

 Biological Fouling – many cooling water systems, especially those which are
re-circulated may contain organisms that attach and grow on internal pipework
surfaces. The “slime” on the heat transfer surface can reduce heat transfer
considerably.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 71

Other typical fouling resistances are listed in Table 3.2.

Table 3.2, Fouling resistance of different materials.

Boiler feedwater 0.00009 m2K/W Natural Gas 1.75-3.5 x10-4
Sea Water 1.75-5.3 x10-4 m2K/W Crude Oil Streams (120°C) 3.5-7 x10-4

River Water 3.5-5.3x10-4 m2K/W Crude Oil Streams (160°C) 5.25-7 x10-4

Refrigerants 1.75x10-4 m2K/W Crude Oil Streams (230°C) 9-10.5 x10-4

Compressed Air 2.6-3.5x10-4 m2K/W LPG 3.5 x10-4

Kerosene 3.5-5.3x10-4 m2K/W MEA/DEA Solutions 3.5 x10-4

Source: InTech Europe

Table 3.3 , Thermal conductivity of fouling components.

Foulant Thermal Conductivity (W/mK)
Alumina 0.42
Biofilm 0.6
Calcium Sulphate 0.74
Calcium Carbonate 2.19
Magnesium Carbonate 0.43
Wax 0.24

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 72

3.5 Film Heat Transfer Coefficient

The overall heat transfer coefficient, as indicated earlier, is a function of the thermal
conductivity of the boundary between the hot and cold fluids in the presence of
deposits, which foul the surface and the film coefficients.

Apart from fouling values, the principle factor, which governs the value of the
overall heat transfer coefficient, is the value of the individual film heat transfer
coefficients. Since calculation of the film coefficients dictates the performance of
heat transfer equipment, methods and factors, which govern film coefficients, are
therefore important and should be found before any detailed design of the heat
exchanger is considered.

In principle, factors which affect the film coefficient for heat transfer are:
Fluid properties:

 fluid density kg/m3

 fluid viscosity N s/m2
k fluid thermal conductivity W/m K
CP fluid specific heat capacity J/kg K

Flow and geometric parameters:

v average fluid velocity m/s

d pipe diameter m

Therefore h is a function of , , k, CP, v and d. A relationship between all the

governing variables, in a heat transfer problem, can be manipulated via analysing
the dimensions of the terms to give empirical correlations useful to predict film heat
transfer coefficients.

From an analysis of the units and dimensions listed above, there are 7 variables
with the following 4 fundamental quantities:

[M] mass [L] length

[T] time [] temperature

Any relationship connecting these 7 variables can be expressed in terms of 7-4=3

dimensionless groups. As mentioned before, dimensionless groups are commonly
used to relate forces together e.g. Reynolds number relates momentum to viscous
forces.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 73

Likewise, in heat transfer, the three dimensionless groups are:

hd
Nu = Nusselt number =
k
Cp 
Pr = Prandtl number =
k
 vd
Re = Reynolds number =


Usually, these groups are related together in the form of

Nu  Constant Re n  Prm

Where, the constant and indices n and m depend on the exact situation.

Using empirical equations of this type allows the film heat transfer coefficient to be
related to the fluid properties and flow regime.

3.5.1 Forced Convection inside Cylindrical Tubes

For fully developed turbulent flow inside pipes or ducts where an equivalent
diameter can be used, the correlations for Nusselt number have a very similar
format. The subtle variations presented below highlight the importance of physical
properties particularly those at the hot or cold surface and those in the bulk phase
of the fluid. Correlations presented here have been divided into sections where the
level of detail increases:

a) Dittus-Boelter equation:
Nu  0.023 Re 0.8  Prn

n = 0.4 for heating

n = 0.3 for cooling

This expression is valid for gases and liquids where 0.6< Pr <160, Re >104 and tube
length L to diameter D ratio of L/D>10.

b) Colburn equation:
Nu  0.023 Re 0.8  Pr0.33

The physical properties are evaluated at bulk fluid temperature except for viscosity
in Re which is taken at the average film temperature.
An alternative form of the Colburn equation is to use Stanton number.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 74

Stanton number:
Nu h
St  
Re Pr C P  v

St Pr0.67  jh  0.023Re 0.2

There are plots of jh against Re rather like pipe friction factor charts – see for
example Coulson and Richardson vol 1, p522.

c) For very viscous liquids (Sieder and Tate):

0.14
  
Nu  0.023 Re 0.8  Pr0.33  
 s 
 = viscosity at bulk temperature
s = viscosity at the heat transfer surface temperature

Table 3.4 shows that the viscosity correction term is not highly significant unless
the temperature difference across the boundary layer is large.

Table 3.4: Variation of (/s)0.14 for a bulk temperature of 20ºC.

Surface   /s ( /s)0.14
Temperature, Ts, ºC kg/m.s
20 0.001002 1.000 1.000
30 0.000797 1.257 1.033
50 0.000544 1.842 1.089
80 0.000350 2.863 1.159

d) For liquids with high Prandtl numbers:

Nu  0.0225 Re 0.795 Pr0.495 e 0.0225(ln Pr)

2

Valid for 4×104 < Re < 106, and 0.3 < Pr < 300

In general, the temperature variation of physical properties, particularly viscosity,

can be substantial and through Re and Pr can have a significant effect on film heat
transfer coefficients.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 75

3.5.2 Forced Convection Outside Cylindrical Tubes

The general correlations for forced convection outside cylindrical tubes are the
same as for the inside of tubes but the details depend on the geometric situation.
The following are some typical situations.

 Single tube

 Concentric tubes (outside of inner tube)

 Tube bundles

i) square pitch

ii) triangular pitch

1. Correlations for flow across single tubes and tube banks are summarised
below:

Air flow across a cylinder depends on the Reynolds number(4) and is often
referred to as Hilpert’s correlations:

Nu  0.615 Re 0.466 40  Re  4,000

Nu  0.174 Re 0.618 4,000  Re  40,000
Nu  0.0239 Re 0.805
40,000  Re  400,000

Churchill and Bernstien suggested the following for many fluids:

45
0.62 Re1 2 Pr1 3   Re  
58
Nu  0.3  1    
  0.4  2 3    282000 
14

1    
  Pr  

This equation is only valid provided:

– Wall temperature is constant.
– A wide range of Prandtl numbers.
– Reynolds numbers up to 107, provided Re Pr>0.2
2. Correlations for flow across banks of tubes depend on the bank orientation
– square layout or staggered.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 76

Usually, the average heat transfer coefficient for air flow over the entire
bundle is usually given by

Nu  C Re m

This equation is only valid provided:

– Number of rows >10 as measured in the direction of flow

– 2000 < Re < 40,000

For fluids other than air, (which has Pr = 0.7), we have:

Nu  1.13 Re m Pr1 3

Constants C and m depend on the arrangements of the tube bank Aligned

type of tube bank is represented as:
L

Direction of Flow

A staggered bank is typically:

L

T D

Direction of Flow

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 77

In general however, the Nusselt correlation is written as:

Nu  a Re m Pr1 3 F1 F2

where F1 is a correction factor normally introduced to account for fluid properties

variation between the heat transfer surface and the bulk fluid:

0.26
 Pr 
F1   b  , b=bulk fluid and w=wall.
 Prw 

In effect, F1 is equivalent to a ratio of viscosities at the bulk and wall temperatures,

provided the heat capacity and thermal conductivity are constant. F2 is a factor
which depends on the number of tube banks and is usually F2 =1 for 10 rows of
tubes, but falls to 0.9 for 4 tubes.

Constants a and m depend on the Reynolds number and the bank arrangement; an
example is given in Table 3.5.

Table 3.5, Constants a and m at different bank arrangements.

Reynolds Number In-line Banks Staggered Banks
a m a m
2
10 to 3x10 0.742 0.431 1.309 0.36
3x102 to 2x105 0.211 0.651 0.273 0.635
2x105 to 2x106 0.116 0.7 0.124 0.7

More effective heat transfer can often be achieved by using tubes which have fins
extending outward increasing the surface area. Unfortunately, these finned tubes
will suffer is highly viscous fluids are used on the outside of tube banks. Exact
descriptions of finned tubes and their properties are beyond the scope of this
section, however the equation for Nusselt number below provides some insight to
the complexity these more efficient surfaces can bring:

0.06 0.11
s 0.7 
0.26
 p1   h 
Nu  0.183Re       Pr0.36 F1 F2 F3
 
h  Df   Df 
   

Briefly, the factors are:

F1 accounts for variations in fluid properties

F2 accounts for the number of tube banks
F3 is a factor for the tube arrangement
Terms s, h etc are dimensions of the fins, their spacing, distance from the
surface etc.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 78

For natural convection, fluid velocity, and hence Re, is no longer significant, but
buoyancy effects become important. Correlations have the form:

Nu = const × Grn × Prm

Grashof number,

 g T d 3  2
Gr 
2

where: β = coefficient of cubic expansion (also called Volumetric Thermal-

Expansion Coefficient or Volume Coefficient of Expansion)
g = gravitational acceleration
∆T = temperature difference across the film

The product βg determines the extent of the natural circulation, or buoyancy, effect
produced by the temperature difference ∆T.
β is defined by:

1  dV 
  
V  dT  P

For an ideal gas, β = 1/T. For real gases you need to work with an equation of state. For
liquids, β can be found from tables of specific volume or density. For example, for water at
50ºC and using specific volumes from steam tables:

1  0.001013 0.001012 4
    4.94 x 10 1/K
0.001012 52  50 

3.5.3 Forced Convection Over Flat Plates

Flow over flat plates has little practical use but it is often quoted in many
engineering texts as an introduction to relating heat transfer to fluid drag;
momentum transport, heat transfer and mass transfer are all related by similar
descriptions of forces which results in a similar set of correlations. For laminar flow
over a flat plate, the Nusselt correlation is:

Nu  0.664 Re1 2 Pr1 3

where the characteristic size for Nu and Re is the length of the plate. For turbulent
flow, the usual form is:

 
Nu  0.0366 Re 0.8  C Pr1 3

where the constant C depends on the transition Reynolds number (4,5) but is
reported as 871 for Re <108, 0.6 < Pr < 60.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 79

Both film heat transfer coefficients and fluid friction factors are related due to the
physical processes active in the flow of fluid over the flat surfaces.

It is not uncommon for texts to highlight two empirical correlations for both Nusselt
number and friction factor. A summary from Hewitt et.al, (1994) for Nusselt
numbers where the surface temperature across the plate is constant, are listed in
Table 3.6.

Table 3.6, Nusselt numbers and different fluid friction factors.

Laminar Transition Turbulent
Nu  0.664 Re1 2 Pr1 3 Nu  0.037 Re 4 5 Pr1 3
for Pr >0.6
 
Nu  0.037 Re 4 5  871 Pr1 3
for 0.6< Pr <60

for 0.6< Pr <60

Nu  0.332 Re1 2 Pr1 3 Transition at Re = 5×105
Nu  0.0296 Re 4 5 Pr1 3
for Pr <0.6

For a flat surface where the heat flux is constant (heat per unit area), the Nusselt
correlation is:

 Laminar flow: Nu  0.453 Re1 2 Pr1 3

 Turbulent flow: Nu  0.0308 Re 4 5 Pr1 3

3.5.4 Single Phase Film Coefficients

Film heat transfer coefficients for gases and liquids depend on the flowrate or more
correctly the velocity of the fluid close to the surface where heat is transferred. A
dimensional analysis of heat transfer processes shows that there are three
important groups which describe the relationship between viscous drag forces,
thermal properties, fluid momentum.

From these groupings, correlations have been developed for particular heat
transfer situations. One of the other difficulties in heat transfer is where liquids and
gases change phase due to changes in temperature. Boiling and condensation
processes are therefore more difficult to analyse. In this section, the theory and
equations will be confined to single phase processes where gases or liquids are
heated and cooled without a change of phase.

Note that some physical properties change rapidly close to the change of phase
points and some equations must be used with caution round the phase change
point.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 80

3.6 Radiation

All materials emit electromagnetic radiation to some degree or other, but this
becomes increasingly important as they get hotter - think of electrical fires, light
filaments etc. Other materials can absorb this radiation - your hand absorbs the
heat if you put it near the electric fire.

Some materials, such as glass, water and air, transmit radiation as well as
absorbing or reflecting it. Of all the radiation landing on your clothes from the Sun,
some is reflected and some is absorbed. Absolutely black clothes would absorb all
of the radiation. Black objects are also better radiators. Any material, solid, liquid or
gas, can radiate and absorb heat - the Sun is not a solid but still radiates heat.

Atoms can switch quantums (or quanta) of their rotational or vibrational energy to
or from one of their electrons. As a simplistic, “non-physics” explanation,
absorption of heat radiation involves an electron absorbing a photon of
electromagnetic radiation, moving to a higher orbital and then switching the energy
to thermal energy as it drops back to its usual orbital.

Radiation involves an electron taking energy from thermal energy, moving to a

higher orbital and emitting a photon of electromagnetic radiation as it drops back to
its usual orbital. Because the same mechanism is involved in absorption and
radiation, materials that are good radiators are good absorbers.

Some atoms/molecules will be better at this process than others so some will
appear black while others are white. Materials may also absorb and radiate better
at some frequencies than others depending on their electron energy levels.

Although we think of radiation as a surface phenomenon, atoms/molecules inside a

solid must also radiate if they are hot enough. What happens to their radiation?
Radiation from molecule to molecule is one of the mechanisms of conduction.

Thermal energy is transmitted in the wavelength range:

0.1 μ (ultra-violet) 100 μ (infrared),

Compared with the visible spectrum: 0.38 μ  0.78 μ.

3.6.1 Energy Emitted by a Back Body (Plank’s Distribution Law)

A heated body emits electromagnetic waves at a rate, E, that depends on T and its
surface characteristics.

E = total emissive power (W/m2)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 81

Of the entire radiation incident on a body, some is reflected, some is absorbed, and
some is transmitted as shown in Figure 3.11.
incident
reflected

absorbed

transmitted

Figure 3.11, Radiation is reflected, absorbed or transmitted

Reflectivity is the fraction of incident radiation reflected.

Absorptivity is the fraction absorbed.
Transmissivity is the fraction transmitted.
Opaque bodies have zero transmissivity. In general, we have:

Reflectivity  Absorptivity  Transmissivity  1

Reflection can be regular or diffuse. Diffuse is more common in heat transfer

situations.

Regular Diffuse

Figure 3.12, Regular and diffuse reflection.

The emissive power at a particular wavelength – the monochromatic emissive

power – has units of W per m2 area per metre wavelength (W/m3). Note that
Energy emitted by a “black body” depends on T only.

Planck’s Distribution Law gives the monochromatic emissive power of a “black

body” at different wavelengths:

3.742 x 1016
E 
  1.439 x 10 2  
5  exp   1
 
  T  

Eλ = monochromatic emissive power, W/m3

λ = wavelength, m
T = absolute temperature, K

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 82

3.6.2 Stefan-Boltzmann Law

White and silver surfaces are very good reflectors of thermal radiation, thus they
are both poor absorbers and poor emitters; in order to retain body heat when a
person is hypothermic, they are often wrapped in insulated silver foil.

Atoms inside a sold surface also emit thermal radiation and this is part of the heat
conduction mechanism. The rate of thermal radiation from an “ideal black body” is
given by the Stefan-Boltzmann Law, as follows:

Eb  T 4

Where,
Eb = total emissive power W/m2
T = ABSOLUTE TEMPERATURE K
 = Stefan-Boltzmann constant = 5.67 x 10-8 W/(m2 K4)

The emissive power of an object is a flux of radiant energy given out by a body and
for a perfect blackbody it depends only on the absolute temperature T (K). The flux
is the rate of heat emission Q (W) per unit area A (m2).

No surface is truly an “ideal blackbody”, so the Stefan-Boltzmann equation must be

modified to allow for real non-black behaviour as follows:

E   T 4

Where, ε is the dimensionless quantity called the “emissivity” and is defined as the
ratio of the emissive power of a non-black surface E to the emissive power of a
perfect black body Eb at the same temperature T .
E

Eb

(It is necessary to emphasise the fact that both E & Eb , in the above expression,
must be at the same temperature T). If ε is constant for all  then the body is
termed an ideal “grey body”.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 83

Figure 3.13Error! Reference source not found. shows the variation of

monochromatic emissive power Eλ with wavelength. Notice that the wavelength of
maximum emission moves towards shorter values as T increases.

Figure 3.13, Black-body spectral emissive power (Engineering Heat Transfer,

W.S.Janna, PWS Engineering, p701, 1986)

You may be puzzled by snow and wonder why light-coloured clothes are cooler
than dark-coloured ones. The confusion comes because emmisvities are averages
which depend on wavelength of EM radiation for “real” bodies. Snow has a low
emissivity for short wavelengths and a high emissivity for long wavelengths. Table
3.7 shows some typical emissivities.
Table 3.7, Some emissivity values mostly from Process Heat Transfer, D.Q.Kern,
McGraw-Hill, 1965
Surface Emissivity
Aluminium foil 0.087
Oil paint 0.92 - 0.96
Water 0.95 - 0.963
Snow 0.985
Rough steel 0.95

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 84

3.6.3 Emissivity and Absorptivity Kirchoff’s Law

Figure 3.14 shows a chamber containing two bodies, a "grey body" with area A and
a "black body" with area Ab. The system is at thermal equilibrium, i.e. for each
body, energy absorbed = energy emitted.

Completely Insulated
area A Vacuum Chamber
area Ab
System in Thermal
Equilibrium
2 small
bodies

Figure 3.14, Black and grey bodies at thermal equilibrium.

Let I = the rate of incident radiation from the walls falling on unit area of either body.
Let a = absorptivity.

“Grey body”: Energy absorbed = Energy emitted

aAI = EA

“Black body”: abAbI = EbAb

E Eb
So: I 
a ab

But a "black body", by definition, is one that absorbs all radiation incident upon it
(i.e. ab = 1).
E
Hence a  .
Eb
But emissivity, ε, of a body is defined by:

E

Eb

So for any body, ε = a. This is Kirchoff’s Law.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 85

3.6.4 Net Heat Transferred by Radiation

Consider a body at T1 radiating to "black" surroundings at T2. The energy emitted

by the body per m2 of area is E   T14 . The black surroundings reflect back none of
this but radiate T2 4 on to the body, of which the body absorbs a proportion "a".

aT2   T2
4 4
Energy absorbed by body per m2 =

Net radiation from the body to its black surroundings:

.
Q  A1 (T1  T2 )
4 4

or flux:
.
Q f   (T1  T2 ) .
4 4

Example : 3.6.1

Problem:

Calculate, for a range of pipe temperatures, the total heat loss per square metre by
radiation and convection from an un-lagged pipe, 50 mm od to air at 283 K.
The emissivity, ε of the pipe = 0.9.

Solution:

 Natural convection from pipe to air:

A suitable correlation for the film coefficient is

0.25
 T 
hc  1.18 

 d0 
0.25
 T 
Hence qc  1.18 
 T
 d0 

Where, T  Tpipe  Tair

 Radiation from pipe to air:


qr   Tpipe
4
 Tair
4
 
 0.9  5.67  108 Tpipe
4
 2834 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 86

Table 3.8 shows the calculations of the heat loss by convection and radiation.
Notice that heat loss by radiation rises more quickly with temperature than
convective losses.

Table 3.8, Heat loss by convection, qc, and radiation, qr.

Tpipe (K) qc (W/m2) qr (W/m2) qtotal (W/m2) qr/qtotal

350 478 438 917 0.48

400 960 979 1939 0.51
450 1498 1765 3263 0.54
500 2078 2862 4940 0.58
550 2693 4342 7036 0.62
600 3338 6286 9624 0.65

……………………………

Example : 3.6.2

Problem:

Calculate the radiant heat flux from a heating panel, emissivity 0.85, when its
temperature is 50oC and again when its temperature is increased to 150oC.

Solution:

Use equation (4.19) at 50oC:

E  0.85 5.67 108  3234

 E  524.6 W/m2

Then use equation (4.19) again at 150oC:

E  0.85 5.67 108  4234

 E  1543 W/m2

……………………………

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 87

Example: 3.6.3

Problem:

Calculate the net radiant heat flux from a heating panel, emissivity 0.85, when its
temperature is 50oC and it is completely surrounded by a room at 20oC.

Assume that the surfaces of the room behave as a perfect black body.

Solution:

In this case body “1” (hot object) is emitting thermal radiation, but it is also receiving
thermal radiation from the room “2” (cold object), thus the net heat flux leaving
surface “1” q1 (W/m2) is amount leaving less the amount arriving:


q1   T14  T24 

 q1  0.85  5.67  108  3234  2934 
 q1  169 W/m2

……………………………

A more detailed treatment of thermal radiation must take into account the number
of surfaces, the “view” that each surface has of each other and the emissivity of
each surface – this is where the resistance approach is helpful.

Thermal radiation is not usually considered in heat exchanger problems unless the
surfaces are at a high temperature, or the overall rate of heat transfer is slow (such
as natural convection), or both.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 88

3.7 Tutorials
1. A copper hot-water line (1.270 cm OD) is to be insulated with a 1.30 cm
layer of material that has a thermal conductivity of 0.04 W/(moC). The
temperature of the outside of the copper pipe may be taken as that of the
water, which is 90oC. The air surrounding the line has a temperature of
21oC, and the air convection coefficient is estimated to be 5.7 W/(m2oC).
Estimate the heat loss per metre length of pipe with and without the layer of
insulation?

Ans. 11.74 W (UA = 0.17 W/K) and 15.7 W (UA = 0.227 W/K)

2. Calculate the following:

a) Calculate the critical outside radius for lagging of thermal conductivity

0.17 W/(m K) surrounding a cylindrical pipe and exposed to air at 20oC.
The air film coefficient is 3.0 W/(m2 K).
b) Calculate the heat loss from a 200oC, 5.0 cm OD pipe i) with no lagging;
ii) with the critical thickness of lagging; iii) with 1.5 x the critical thickness
of lagging.

Ans: a) 56.7 mm (i.e. critical thickness = 31.7 mm for a 5 cm OD pipe)

b) i) 84.8 W/m ii) 105.7 W/m iii) 104.1 W/m

3. A pipe of 40 mm ID, 50 mm OD, which is maintained at 1100 K on its inside

surface, is lagged with 50 mm of insulation. The temperature of the
surroundings is 280 K.

a) Estimate the heat loss per unit length.

b) Estimate the temperature on the outside of the insulation.

Thermal conductivity of insulation = 0.35 W/(m K)

Thermal conductivity of pipe = 25 W/(m K)
Outside film coefficient = 10 W/(m2 K)

Ans: a) 1150 W b) 524 K

4. Use the Dittus-Boelter correlation on page Error! Bookmark not defined.

to find the film heat transfer coefficient for water at 50oC flowing in a 5 cm ID
pipe with a velocity of 2 m/s. Confirm that the flow is turbulent. The following
values are taken from steam tables.

Thermal conductivity = 0.643 W/(m K)

Density = 1000 kg/m3
Viscosity = 0.544 x 10-3 kg/(m s)
Specific heat capacity = 4.18 kJ/(kg K)

Ans. Re = 1.84 x 105 Pr = 3.53 h = 7978 W/(m2K)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 89

5. Determine the inside film heat transfer coefficient for water flowing at
10000 kg/h in a 2.5 cm id pipe if the water temperature is i) 20oC, ii) 80oC.
Use: Nu = 0.023 Re0.8 Pr0.33 and properties given in Table below.
(The main point of this question is to show the strong effect of temperature
on physical properties, particularly viscosity and thermal conductivity, and
how these in turn can influence heat transfer coefficients.)

Ans: i) 13860 W/(m2 K) ii) 24410 W/(m2 K)

Physical Property 20ºC 80ºC

ρ kg/m3 1000 971
cp J/(kg K) 4182 4197
μ kg/(m s) 1.002 x 10-3 0.35 x 10-3
k W/(m K) 0.603 0.67

6. Water from a storage tank is supplied to a process by means of a 650 m

pipeline. Under the most severe weather conditions ambient temperature
and storage tank temperature are expected to have minimum values of 263
K and 278 K respectively. Estimate the minimum flow rate between the
storage tank and the process necessary to prevent ice formation in the
pipeline. The pipeline consists of a 100 mm OD pipe covered with a layer of
magnesia insulation (k = 0.1 W/(m K)) 25 mm thick. A heat transfer
coefficient of 1.2 W/(m2K) at the surface of the insulation is anticipated.
(Hint: The inlet temperature to the pipe is the tank temperature; take the
outlet temperature to be 0oC. Find the UA value and the heat loss using the
.
log mean driving force. Now use Q  m CP T to find m.)
.
Ans: Q = 3.321 kW (UA = 269.3 W/K), m = 0.158 kg/s

7. A pipe 20 mm ID and 30 mm OD carries a fluid at 448 K. The inner and

outer film heat transfer coefficients are 22 and 5 W/(m2K). The ambient
temperature is 293 K. Ignore the resistance of the metal pipe wall.
a) Calculate the heat loss per metre of pipe.
b) What will be the heat loss if a layer of insulation 50 mm thick with a
thermal conductivity of 0.8 W/(m K) is applied to the pipe? Assume that
the outer film heat transfer coefficient stays the same.

Ans: a) 54.5 W (UA = 0.351 W/K) b) 103 W (UA = 0.665 W/K)

(Can you explain why the lagged pipe loses more heat?)

8. Saturated steam flowing at 360 kg/h enters an unlagged 100 mm OD pipe.

Calculate the length of pipeline that can be used before the steam quality
drops below 97% dry i.e. 3% condenses.
If the pipe is lagged with a 2 cm thick layer of insulation so that the
temperature of the exposed surface is reduced to 50oC, what will be the
steam quality at the end of 13 m?

Latent heat of steam = 2210 kJ/kg

Outside surface temperature of pipe, Tw =115ºC

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 90

Emissivity of pipe, lagged and unlagged = 0.9

Surrounding air temperature, Ts = 20ºC

Outside convective film heat transfer coefficient,

ho = 1.6 (T/do)0.25 W/(m2 K)
T = (Tw - Ts) K
do = outside pipe diameter m

Ans: a) 13 m b) 99.1%

9. The Sun has a radius of 6.98 x 108 m and radiates as though it were a black
body at a temperature of 5800 K. The distance from the Sun to the Earth is
1.5 x 1011 m. Estimate the energy impinging on the Earth per square metre
when the Sun is overhead. Area of a sphere = 4r 2
(Just out of interest, about 30% of the radiation is absorbed by the
atmosphere.)

Ans: 1.389 kW

10. A stainless steel pipeline between two gas/oil separators is un-insulated.

The temperature of the oil inside the pipeline is found to fall from 80°C to
60°C. The ambient air flowing over the line is at 3°C and the typical wind
velocity can be 3m/s. If the pipeline internal diameter is 3 inches, and the
wall thickness is 0.25 inches calculate the following:

a. The heat loss from the oil, given that the average velocity of the oil is
1m/s
b. Calculate the outside film coefficient of air over the pipeline
c. Calculate the overall heat transfer coefficient, assuming that the
internal one is 1500 W/m2K.
d. Comment on where the resistance to heat transfer is and what may
help to limit the heat lost from the oil.
Additional Data
Assume that Nusselt number for the air side is calculated with this correlation
Nu = 0.615 Re 0.466

Oil properties: Density 900 kg/m3, viscosity 3 cP, specific heat capacity 2.5
kJ/kgK
Air properties: density 1.3 kg/m3, viscosity 13x10-6 Pas , specific heat
capacity 1.0 kJ/kgK, thermal conductivity 2x10-5 kW/mK
.
Ans: a) Q = 205 kW, b) ha = 15.978 W/m2 K, c) U =16 W/m2 K,
d) The overall HTC is the same numerical value as the film heat transfer
coefficient into the air. It is the heat from the pipe wall into the air that controls
the heat transfer process so changing the conditions inside the pipeline will
not do much for the heat loss. In order to prevent further heat losses,
insulation should be applied to the pipeline.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 91

11. Calculate the total heat loss per metre length by radiation and convection
from an unlagged steam pipe of 50 mm OD at 415 K to air at 290 K.
Air film heat transfer coefficient = 7.0 W/(m2K).
Emissivity of steel pipe = 0.7.

Ans: Radiation 140.8 W, Convection 137.4 W

12. A storage tank 5m diameter and 4m tall containing oil at 40°C, is exposed to
strong winds travelling at 4m/s. Average air temperature is 5°C. Calculate the
rate of heat loss for radiation and convection, and hence the total heat loss.
Assume that the wall temperature of all sides of the tank are the same as the
oil.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 3. CONVECTION AND RADIATION 92

3.8 Bibliography
1. "C&R vol. 1" refers to Chemical Engineering, J.M.Coulson, J.F.Richardson et
al., vol. 1, 6th ed., Butterworth, 1999

2. "C&R vol. 6" refers to Chemical Engineering Design, R.K.Sinnott vol 6, 4th ed.,
Butterworth, 2005

3. A reasonable site for doing unit conversions, including heat transfer coefficients
(called thermal conductance) is: http://www.cleavebooks.co.uk/scol/index.htm

4. Introduction to Heat Transfer, F.P. Incropera et al, 5th ed., Wiley, 2007
Good book. I used it for the section on fins. The library has earlier editions –
536.1INC.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 91

Topic 4

Heat Transfer Equipment

Contents

4.1 Introduction 93
4.2 Heat Transfer Equipment 94
4.2.1 Double Pipe (Concentric Tube) Heat Exchanger 94
4.2.2 Gasketed Plate Heat Exchanger 95
4.2.3 Welded Plate Exchanger 96
4.2.4 Spiral Heat Exchanger 96
4.2.5 Counter-current Heat Exchanger 97
4.3 Mean Driving Force 100
4.4 Temperature Profiles 101
4.5 Practical Heat Exchangers 102
4.5.1 Shell-and-Tube H/E 102
4.5.2 Plate H/E 104
4.6 Heat Exchanger Basics 105
4.6.1 Log Mean Temperature Difference Method 107
4.6.2 Justification for Using the Log Mean as the Average 109
4.7 Heat Exchanger Design 111
4.7.1 Calculations of H/E Modes of Operation 111
4.7.2 Performance/Simulation Mode H/E Calculations 113
4.7.3 Maximum Possible Heat Transfer Between Two Streams 115
4.8 Number of Heat Transfer Units (NTU) Method 117
4.8.1 Analytic Solutions for the Effectiveness of H/E 118
4.8.2 Comparison of Two Heat Exchanger Approaches 121
4.9 Tutorials 122
4.10 Bibliography 127

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 92

Prerequisite knowledge

 Detailed understanding of different modes of heat transfer.

 General knowledge about fluid mechanics.
 General knowledge about conduction and convection in different streams.
 Turbulent and laminar layers in fully developed flow.

Learning Objectives

By the end of this topic, you should be able to:

 Be familiar with heat transfer equipment.

 Differentiate between different types of heat exchangers and their
applications.
 Outline main facts concerning heat exchanger design.
 Understand the mean driving force.
 Carry out performance/simulation calculations.
 Use analytic solutions for the effectiveness of heat exchangers.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 93

4.1 Introduction

Heat exchangers are used to achieve heating of one fluid stream at the same time
cooling another fluid stream but keeping the two streams physically separate. In
most cases, the objective is to bring about the heating or cooling of one stream
only. In this case, the second stream, known as a Utility, serves only to supply or
remove heat for the first stream.

There are many different patterns of heat exchanger in use but the shell and tube
type is much more widely encountered than any other. Other types may be
preferred in special cases, e.g. plate and frame or double tube exchanger when
only low temperature difference is available or spiral exchangers for compactness
and low pressure drops.

In addition to the standard heat exchanger with tubes or channels through which
one fluid is passed, the tube or channel can be enhanced by putting a series of fins
on one or both sides. This increases the contact area and hence the rate of heat
transfer.

Designing heat exchangers is a complex process with computer based software

taking a lot of the load in terms of decision making. Normally specific vendors will
specialize in heat transfer and supply a unit to the required specifications. The
process engineer need therefore only be concerned with generating the design
specification. The operations engineer needs to know how the unit behaves and
what may cause deviations from the intended specification.

In general for heating and cooling vessels three methods can be used:

 Jacket - The vessel is heated or cooled through its wall by a jacket, or

panels, or a coil. Heat transfer is not particularly good and the area
available is small.

 Internal Coils - A coil or bundle of tubes inside the vessel carries steam or
cooling water etc. This gives a larger area than a jacket but the coils may
get fouled and are difficult to clean.

 External heat exchangers - Fluid is pumped out of the vessel, through a

heat exchanger and returned to the vessel. This is expensive but now
allows the heat transfer equipment to be designed independently of the
vessel.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 94

4.2 Heat Transfer Equipment

There are different types of heat exchangers, including:

 Double pipe (concentric tube) heat exchanger

 Shell and tube heat exchanger
 Gasketed plate heat exchanger
 Welded plate heat exchangers
 Spiral heat exchanger

For more details see also C&R vol. 1 and C&R vol. 6.

4.2.1 Double pipe (concentric tube) heat exchanger

Double pipe heat exchangers are commonly constructed as "U-Tube" units,

available in standard sizes, which can be combined to give the area required (see
Figures 4.1 and 4.2). Double pipe heat exchangers are relatively simple and cheap
to construct. The other advantages of using them are:

 Cleaning and maintenance is relatively simple.

 Close to pure counter-current heat exchange.
 Suitable for use of externally finned tubes.

Figure 4.1, Double pipe U-tubes combined

Fluid 1 In

Fluid 2 Out
Fluid 2 In

Fluid 1 Out

Figure 4.2, Double pipe heat exchanger

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 95

The range of their operation includes:

 Tube pressures up to 1400 atm

 Shell pressures up to 300 atm
 Temperatures up to 600oC
 Heat loads up to 1000 kW

4.2.2 Gasketed Plate Heat Exchanger (Plate-and-Frame)

A stack of corrugated metal plates is pressed in a frame and sealed at their edges
by compressible gaskets. This forms a series of interconnected narrow passages
with the hot and cold fluid flowing in alternate passages. Heat is transferred through
the relatively thin, low heat transfer resistance, plates.

Plates can be pressed from a wide range of materials such as mild steel, stainless
steel, titanium, and nickel.

Corrugations on the plates increase the area, promote turbulence and improve
rigidity.

The gaskets can be made of synthetic rubbers, plastics or asbestos-based

materials.

Conditions, in which the heat exchanger can operate, are limited by the gasket
material: temperature < 260oC, pressure < 25 bar.

See C&R vol. 6 p756 (3rd ed. p753) and Foundation Notes for diagrams, and C&R
vol. 6 p761 (3rd ed. p758) for a design example.

The benefits of using Gasketed Plate Heat Exchangers are:

 Flexibility:
– Addition, removal, replacement of plates.
– Possibility of multiple duties on 1 frame.

 Compactness:
– High heat transfer coefficient.
– High area per unit volume.

 Low fabrication costs:

– Pressed rather than welded construction – this may allow
use of a more corrosion-resistant material.

 Accessibility:
– Easily dismantled for cleaning and maintenance.

 Fouling:
– Relatively low fouling rates in spite of small passages.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 96

Gasketed Plate Heat Exchangers are traditionally used in food and beverage
processing and the production of pharmaceuticals. Currently, they have also been
used for the production of various chemicals, oil and gas industry, and power
generation. Figure 4.3 shows this type of heat exchanger in details.

Figure 4.3, Gasketed Plate Heat Exchangers.

4.2.3 Welded Plate Exchangers

Similar concept to the gasketed plate exchanger, but the plate edges are welded
together. They are more expensive to manufacture but the cost differential reduces
with large plate sizes.

Welding increases the upper temperature and pressure limits for plate-type
construction. However, they lack the "flexibility" and "accessibility" benefits of
gasketed plate exchangers. They are also not suitable for the application in which
fouling is a problem.

4.2.4 Spiral Heat Exchanger

Two long metal sheets, welded together along one long edge, wound into a spiral
and sealed between two end plates, forming two concentric spiral channels.

The cold fluid generally enters at the periphery and spirals to the centre while the
hot fluid spirals outwards.

The gap between the plates and the tightness of the spiral can be set to suit the
flows and properties of the two fluids.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 97

The benefits of using spiral heat exchangers are:

 High heat transfer coefficients, low pressure drop

 Swirling path helps prevent fouling (very good for liquids with suspended
solids)
 Easy access when required for cleaning.

Figure 4.4 shows the diagrams of spiral heat exchangers.

Figure 4.4, Spiral heat exchanger – C&R vol 1, p551

4.2.5 Counter-current Heat Exchanger

Consider a simple heat exchanger as shown in Figure 4.5.

Th2
Hot stream
mh kg/s at Th1
Q
Cold stream
mc kg/s at Tc2
Tc1

Figure 4.5, Heat exchanger with counter-current flow

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 98

Figure 4.6 shows possible boundaries for heat balances. One end of the heat
exchanger is 1, and the other is 2. Th1 and Th2 represent inlet and outlet
temperatures of the hot stream. Tc1 and Tc2 represent outlet and inlet temperatures
of the cold stream. Three useful boundaries are marked: B1, B2, and B3. Notice
that the heat flow, Q, passing from the hot to the cold stream, crosses boundaries
B1 and B2.

B1

Th2
Hot stream
mh kg/s at Th1
Q
Cold stream
mc kg/s at Tc2
Tc1

B2

B3

Figure 4.6, Counter-current flow heat exchanger – showing boundaries for heat
balances

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 99

Selection of Heat Exchanger Type:

The prime criteria for initial selection are:

 Feasible pressure and temperature range.

 Compatibility between fluids and materials of construction.
 Size range available.
 Special features relating to the process fluids eg. fouling nature, slurries.
 Special features relating to the proposed location eg. accessibility, space
and weight limitations.

Requirements for Process Application:

 Large overall heat transfer coefficient.

 Substantial area in a reasonably compact volume.
 Reasonable pressure drops.
 Safety in operation.
 Ease of fabrication and maintenance.

Analysing heat exchangers:

 Draw a sketch.
 Do enthalpy balances as necessary to find all inlet and outlet temperatures
.
and the heat load, Q .
 Find the driving forces at each end: ΔT1 and ΔT2.
 Find ΔTLM.
.
 Use the rate equation, Q  UATLM .

You can see the main characteristics of all the heat exchanger types considered in
Table 4.1 at the end of this Topic.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 100

4.3 Mean Driving Force

Many flow patterns are used in heat exchangers. Figure 4.7 illustrates three
general kinds:

 Counter-current flow is the commonest since it gives the largest driving

force. You can heat the cold fluid up to nearly the hot inlet temperature.

 Co-current is not obviously useful - you can only heat the cold fluid up to the
hot outlet temperature.

 Cross-flow is also unusual except in carbon-block heat exchangers. These

are blocks of carbon with holes drilled at right angles and are used for very
high temperatures or with corrosive fluids.

As we shall see in shell-and-tube exchangers, real flow patterns are often

complicated mixtures of these three.

The use of log mean temperature difference is only "correct" for the parallel flow
given by strict counter-current or co-current flows. Design books give correction
factors for other arrangements.

Countercurrent flow

Cocurrent flow

Crossflow

Figure 4.7, Three simplified heat exchanger flow patterns.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 101

4.4 Temperature Profiles

Figure 4.8 shows four examples of temperature variation with position in a heat
exchanger. Can you identify counter-current and co-current flows?

One is for steam heating. Notice the short lengths (areas) of heat exchanger used
to remove superheat and to sub-cool the steam. You would neglect this variation in
steam temperature when calculating ΔTLM and take the steam temperature at inlet
and outlet to be its saturation temperature.

One of the examples is impossible in practice but is a common pit-fall in design. It

is called a temperature cross and occurs, for example, when you try to extract too
much heat from the hot stream into too small a flow of cold stream. This gives a
cold stream leaving at a higher temperature than the inlet hot stream.

A B
Temperature

Temperature

Length Length
C D
Temperature

Temperature

Length Length

Figure 4.8, Possible (and impossible) heat exchanger temperature profiles.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 102

4.5 Practical Heat Exchangers

Heat exchangers should:
 Withstand the necessary temperature and pressure.
 Be easily cleaned.
 Have a material separating the fluids that has a high thermal conductivity or
is very thin.
 Give high convective heat transfer - implying high velocities.
 Give low pressure drops - implying low velocities.
 Have a small chance of leaking if the fluids are hazardous.
 Be cheap.

The CD in Felder & Rousseau gives good heat exchanger illustrations, applications
etc. In particular, there is an animation of a plate heat exchanger.

4.5.1 Shell-and-Tube H/E

In a shell-and-tube heat exchanger a large number of tubes are fitted inside one
large tube called the shell. Figure 4.9 and Figure 4.10 show the idea. Using two or
more passes increases the fluid velocity, which improves heat transfer at the cost
of a higher pressure drop.

Figure 4.9, Shell-and-tube heat exchanger with one tube-side pass.

Figure 4.10, Shell-and-tube heat exchanger with two tube side passes.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 103

Figure 4.11 shows more examples. In (a) the shell-side ports or nozzles (inlet or
outlet connections) are 2 and 3. The tube-side ports are 6 and 7. The tube flow
goes straight through - it is one-pass. Compare this with the one in (c) where the
tube flow goes along to the end, turns, and come back - it is two-pass. In (c) the
tube bundle is held at the right hand end - it is a fixed head. In (d) the right hand
end of the tubes is free to move in the shell - it is a floating head, which is a way of
allowing the tubes to expand and contract.

In (b) there is an illustration of shell-side baffles, which force the flow back and
forward across the tubes for better heat transfer.

Figure 4.11, Shell-and-tube heat exchanger.

Shell-and-tube heat exchangers are very common. They can handle any
combination of heating, cooling, boiling and condensation. They can be designed
for high pressure and temperatures but are more expensive than plate heat

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 104

exchangers. The tube-side can be cleaned by removing the heads (ends) and
cleaning with drills, high pressure water etc. The shell-side is difficult to clean,
therefore put the dirty fluid (eg, cooling water, beer or milk) in the tubes.

4.5.2 Plate H/E

Figure 4.12 illustrates the flow pattern in a plate heat exchanger while Figure 4.13
shows more mechanical detail. The arrangement is similar to that used in plate-
and-frame filter presses. The thin (1-2 mm) metal plates are clamped together with
a small (3-6 mm) gap between them. Because the plates are so thin they can be
made of stainless steel without giving too high a heat transfer resistance -
remember that stainless steel is a poorer conductor than carbon steel. The plates
are corrugated to make them less flexible and to improve turbulence and hence
heat transfer. A rubbery gasket (seal) between the plates stops the liquids leaking
out and also channels the flow so that the two streams do not mix.

Plate heat exchangers are particularly good in the food industry since they can be
easily taken apart and cleaned. They pack more area into a given volume than
shell-and-tube exchangers.

Figure 4.12, One possible flow pattern in a plate heat exchanger.

Figure 4.13, Plate heat exchanger.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 105

4.6 Heat Exchanger Basics

The prime objective in heat exchanger (H/E) design is to find the heat transfer
surface area required to transfer the necessary amount of heat. This area, in turn,
will determine the size and cost of the H/E.

There are two approaches to heat exchanger analysis, Log Mean Temperature
Difference and NTU methods.

Example : 4.6.1

Problem:

Consider a specific case in which both streams, in a heat exchanger, are water with
mh = 1.5 kg/s and mc = 3.0 kg/s. The hot stream enters at 100oC and leaves at
80oC. The cold stream enters at 20oC. There is no phase change - only sensible
heat is involved.

a) How much heat does the hot stream lose? Use boundary B1.
b) What is the final temperature of the cold stream? Use boundary B2.
c) What is the area of heat exchanger required for the above example if the
overall heat transfer coefficient, U = 800 W/(m2K)?

Solution:

a) Systematically:
Enthalpy basis: liquid water at 0oC
.
Enthalpyin  Q  Enthalpyout
.
mhC p (Th1  0)  Q  mhC p (Th 2  0)
.
1.5  4.2  (100  0)  Q  1.5  4.2  (80  0)
.
 Q  126 kW
The negative value indicates that the heat is removed.

 Alternatively:
Heat removed from hot stream,
.
Q  mhC p (Th1  0)  1.5  4.2  100  80
.
 Q  126 kW
Positive because we stated heat removal.

b) Systematically:
Enthalpy basis: liquid water at 0oC

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 106

.
Enthalpyin  Q  Enthalpyout
.
mC C p (TC 2  0)  Q  mC C p (TC1  0)
3.0  4.2  (20  0)  126  3.0  4.2  (TC1  0)  TC1  30oC

 Alternatively:
Heat added to the cold stream,
.
Q  mhC p T  126  3.0  4.2  TC1  20
 TC1  30oC

Another way of finding the outlet temperature of the cold stream is to use boundary
B3.

 Systematically:

Enthalpy basis: liquid water at 0oC

Enthalpy in = enthalpy out

mhC p (Th1  0)  mC C p TC 2  0  mhC p (Th 2  0)  mC C p (TC1  0)

1.5  4.2  (100  0)  3.0  4.2  20  0  1.5  4.2  (80  0)  3.0  4.2  (TC1  0)
 TC1  30oC

 Alternatively:
Enthalpy lost from hot stream = Enthalpy gain by cold stream
or
Heat from hot stream = Heat to cold stream

1.5  4.2  (100  80)  3.0  4.2  (TC1  20)

 TC1  30oC

To find the heat lost or gained by a stream, if in doubt, use the systematic method.
.
If the problem is familiar and you are confident, use Q  m C p T for sensible heat
.
changes, and Q  m  for latent heat effects, e.g. condensing steam. The latent
heat, λ is the enthalpy difference (hin - hout) and it appears if both the steam and
condensate are saturated.

.
We would like to use Q  UAT where ΔT is the difference between the hot and
cold stream temperatures but the driving force is (100 - 30) at one end and (80 -
20) at the other end.

We need an average ΔT:

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 107

Take T1  100  30  70oC

T2  80  20  60oC

We use
T1  T2
TLM 
 T 
ln 1 
 T 2 

ΔTLM is called the logarithmic mean temperature driving force or simply the log
mean ΔT.
Using the values of ΔT1 and ΔT2, ΔTLM = 64.9oC.

Now using
.
Q  UATLM
126  0.8  A  64.9  A  2.43 m2

Note: if ΔT1 = ΔT2 the formula above for ΔTLM does not work.
However, suppose ΔT1 = ΔT2 = 10ºC. The average ΔT is obviously 10ºC.

........................................

4.6.1 Log Mean Temperature Difference Method

There are many types of heat exchangers, but it may be convenient at first to think
in terms of a simple concentric tube heat exchanger (see Figure 4.14).

Hot Fluid In

Cold Fluid Out Th

Tc Cold Fluid In

dA
Hot Fluid Out

Figure 4.14, Concentric tube H/E

.
Consider an element of surface area, dA, which transfers an amount of heat d Q .
.
. dQ
We can say that d Q  U (Th  Tc ) dA giving dA 
U (Th  Tc )

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 108

.
To find the total area required to transfer an amount of heat, Q , we can integrate:

.
Q .
dQ
A
0

U (Th  Tc )

In general, both U and (Th - Tc) will vary along the length of the exchanger, so
numerical integration is required. Fortunately, a number of simplifications can often
be made allowing an analytical solution. In simple heat exchange devices, Th and Tc
vary with position as shown in Figure 4.14.

Countercurrent Flow Co-current Flow

cold cold
hot hot
Temperature

Temperature

ΔT
ΔT

Q Q

Figure 4.14, Temperature profiles in co- and counter-current flow

.
In either case, provided U remains constant and T varies linearly with Q for each
fluid (e.g. no phase change and cp is constant):

.
Q  UA TLM

T1  T2
TLM 
 T 
ln 1 
 T2 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 109

ΔT1 and ΔT2 are the driving forces at the two ends of the H/E. ΔTLM is called the
logarithmic mean temperature driving force or simply the log mean ΔT.

4.6.2 Justification for Using the Log Mean as the Average Driving Force

Consider a simple case with a constant temperature on one side of the H/E.

Steam In T3

Liquid In Liquid Out

T1
T2

Condensate
Out
T3

T3
T2
T+dT
Temperature

T1

0 L
0 A

dL
dA

Figure 4.15, Temperature profile in double pipe heat exchanger.

.
Consider the heat transferred, d Q , over a small area, dA.
.
By heat transfer: d Q  U (T3  T ) dA
.
By heat gained: d Q  mL c pL ((T  dT )  T )  mL c pLdT

 U (T3  T ) dA  mL c pL dT

A T2
U dT

mL c pL 0 
dA 
T T
T 3

1

UA  T T 
  ln 3 1 
mL c pL  T3  T2 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 110

UA
m L c pL  ............................(4.1)
 T T 
ln 3 1 
 T3  T2 
By overall heat balance:

.
Q  mL c pL (T2  T1 )

Substituting for mLcpL in Equation (4.1)

.
 Q  UA
T2  T1 
 T T 
ln 3 1 
 T3  T2 

Add and subtract T3 to the numerator

.
 Q  UA
T3  T1   T3  T2 
 T T 
ln 3 1 
 T3  T2 
So:
.
Q  UA TLM

The same derivation applies when both temperatures change along the length but
the algebra is a bit more complicated. The derivation also gives the same result
with a co-current heat exchanger.
.
The assumptions behind use of Q  UA TLM are:

 U is constant over the length of the exchanger.

 For sensible heat changes, the specific heat capacity of each fluid is
independent of temperature.

 Either there is no phase change on either side or, if there is, it occurs at
constant temperature.

 Negligible heat loss or gain to the surroundings.

 No heat generation or removal by chemical reaction.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 111

4.7 Heat Exchanger Design

4.7.1 Calculations of H/E modes of operation

The following example gives a comparison of counter-current and co-current

modes of heat exchanger operation.

Example : 4.7.1

Problem:

20 kg/s of water is to be cooled from 80oC to 60oC using 25 kg/s of water entering
at 20oC. Find the areas required in counter-current and co-current heat
exchangers. See Figure 4.16.

Countercurrent Co-current

80 80

T1 =44
60 T1 =60 60
T2 =24
36 T2 =40
36
20 20

Figure 4.16, Temperature driving force with co- and counter-current flow.

Solution:

c p  4180 J/(kg K), U  2000 W/(m2 K)

A heat balance gives:

.
Q  (m c p T ) hot  (m c p T ) cold  20  4180 (80  60)  25 4180 (Tc out  20)
.
 Q  1672 kW
Tc out  36o C

Counter-current: TLM = 42ºC

Co-current TLM = 39.3ºC

From
.
Q  UA TLM
Counter-current area: A = 19.9 m2
Co-current area: A = 21.3 m2

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 112

You will see that TLM is larger for counter-current flow than for co-current flow
hence the area is smaller.

If the hot stream is cooled to 50oC rather than 60oC show that:
.
Q  2508 kW and Tc out  44o C

Counter-current area = 38.1 m2

Co-current area = 53.4 m2

If the hot stream is cooled to 40oC rather than 60oC show that:
.
Q  3344 kW and Tc out  52o C

Counter-current area = 70.3 m2

Co-current area = ? m2

This last case gives a temperature cross – see Figure 4.17.

Countercurrent Co-current

80 80
52

52
40

20 20

Figure 4.17, Possible and impossible temperature profiles.

................................................

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 113

4.7.2 Performance/Simulation mode H/E calculations

The following examples show how the performance/simulation of heat exchangers

can be calculated.

Example : 4.7.2

Problem:

Consider the counter-current heat exchanger, as calculated in example 4.7.1, using

the design found with a hot outlet temperature of 50ºC.

U = 2.0 kW/(m2 K) A = 38.1 m2

20 kg/s

80oC 50oC
Q = 2508 kW

44oC 20oC

25 kg/s

Figure 4.18, Original duty of heat exchanger.

Suppose that we use this H/E but increase the cold water inlet temperature to
25oC. What will be the outlet temperatures?

Solution:

20 kg/s

80oC TH,out
Q=?
TC,out
20oC

25 kg/s

Figure 4.19, New H/E duty with cold water inlet raised to 25ºC.

Relationships to be satisfied

.
Q  (m c p T ) hot
.
Q  (m c p T ) cold
.
Q  UA TLM

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 114

Because of the form of TLM there is no analytic solution, so use iteration.

First Estimate:

Th out  55o C (has to be higher than initial case – why?)

.
Q  20  4.18 80  55  2090 kW
Tc out  45o C
TLM  32.4o C

Giving A = 32.2 m2, which is less than the actual area of 38.1 m2. The hot stream
outlet must be less than 55ºC.

Second Estimate:

Th out  52o C

.
Q  2341kW

Tc out  47.4 o C and TLM  29.7 oC

Giving A = 39.4 m2 which is greater than the actual area of 38.1 m2.

In an exam perhaps do two trials showing that you know what to do and how to
make reasonable guesses of Th out. This hand calculation method is old-fashioned.
You could set up an Excel spreadsheet and use Goal Seek to vary Th out until a cell
containing A goes to 38.1.

Final Solution:

20 kg/s

80oC 52.5oC
Q = 2299 kW

47oC 25oC

25 kg/s

Figure 4.20, Final solution for new duty.

................................................

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 115

Example : 4.7.3

Problem:

Consider again the design in Figure 4.18. U = 2.0 kW/(m2 K), A = 38.1 m2

Suppose that the tubes foul so that U drops to 1.6 kW/(m2 K). What will be the
outlet temperatures?

Solution:

1st guess Th out = 56 gives A = 32.7 m2.

2nd guess Th out = 54 gives A = 37.2 m2.

Final solution, Th out = 53.6 gives A = 38.1 m2.

4.7.3 Maximum Possible Heat Transfer between Two Streams

The maximum possible heat transfer between two streams is achieved in a

counter-current heat exchanger with a temperature “pinch” as shown in Figure
4.21. Since the driving force at one end is zero you need an infinitely large heat
exchanger (A  ).

Tcout= Thin Thout= Tcin

mccpc < mhcph mccpc > mhcph

. .
Qmax=m
Q max mccccpc
pc(T h in–TT
(Thin c )in )
cin max  m
Qmax= mhchphc(T (Th–inTcinT)c in )
ph hin

Figure 4.21, Counter-current flow temperature pinch.

The temperatures "pinch" at the outlet end for the stream with the smaller value of
the product (m cp).

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 116

In a co-current heat exchanger as A   the temperatures will "pinch" at the exit

end, at a temperature lying between the inlet temperatures as shown in Figure
4.22.

Figure 4.22, Co-current flow temperature pinch.

The exit temperature can be found by heat balance. Make mcpΔT equal for both
streams and both streams leave at the same temperature, T. The maximum heat
transferred in co-current flow is less than the maximum transferred in counter-
current flow.

Example : 4.7.4

Problem:

Find the maximum heat transfer between two streams given a hot inlet of 80ºC and
a cold inlet of 20ºC. Both streams are water, cp = 4180 J/(kg K).
Check the following results for yourself.

a) Hot Flow = 20 kg/s , Cold Flow = 25 kg/s

Counter-current:
.
Q  5016 kW (The hot stream has the smaller mcp value.)

Co-current:
.
Q  2787 kW

b) Hot Flow = 20 kg/s , Cold Flow = 15 kg/s

Counter-current:
.
Q  3762 kW (The cold stream has the smaller mcp value.)

Co-current:
.
Q  2149 kW

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 117

4.8 Number of Transfer Unit (NTU) Method

The NTU method is an alternative to using the log mean ΔT. It solves two
problems:

 It avoids us worrying about the value of mean driving force, ΔTmean, where
the exchanger is not pure counter-current or co-current – only for these is it
strictly correct to take ΔTmean = ΔTLM. Although it is also possible to use the
correction factor, FT.

 It allows us to do "performance" calculations without trial-and-terror.

.
Within this method there are three key concepts: C*, Q max , and ε.

C is the heat capacity of a stream – the amount of heat to raise the temperature
of the stream by 1ºC. C  m c p If you work out the C value for the hot and cold
Cmin
streams, you find Cmin and Cmax and the ratio of these gives C*  .
Cmax

.
Q max is the maximum possible rate of heat transfer in a counter-current
.
exchanger – even if you are considering a co-current heat exchanger, Q max is still
based on counter-current flows.

.
Q max  Cmin (Th in  Tc in )

 If Cmin applies to the cold stream, at most it could be heated to Th in.

 If Cmin applies to the hot stream, at most it could be cooled to Tc in. Either
way the maximum change in stream temperature is Th in – Tc in)

The effectiveness of a heat exchanger is defined as:

.
Q actual
ε .
Q max

The number of transfer units is given by:

UA
NT U 
Cmin

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 118

For various configurations such as counter-current, co-current, one shell-side pass

and two tube-side passes,  has been related to the two dimensionless groups, C*
and NTU. For example, Figure 4.23 shows a typical plot of effectiveness against
NTU with varying C*.

Family of C* lines

Figure 4.23, Heat exchanger effectiveness for counter-current flow.

(C&R vol 1, p539 with minor modification)

4.8.1 Analytic solutions for the effectiveness of H/E

Analytic solutions for effectiveness, , are possible for pure co-current and counter-
current cases.

Performance:
  
.
Q 1  exp  NT U 1  C *
 
.
Q max 
1  C * exp  NT U 1  C *   Counter-current

  
.
Q 1  exp  NT U 1  C *
  Co-current
.
Q max 1  C*

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 119

Design:

NT U 
 
ln 1   C * 1     Counter-current
1  C*

NT U 
 
 ln 1   1  C *  Co-current
1  C*

Limiting Cases:

Can you derive these from the above equations? For one of them you will need to know
that e x  1  x if x is very small.

If C* = 0, one flow has such a large mCp value that its temperature stays constant.
The performance of counter-current and co-current are the same.

 countercurrent   cocurrent  1  exp(NTU)

If C* = 1
NT U
 countercurrent 
1  NT U

1  exp(2 NTU )
 co current 
2

If NTU  
 countercurrent  1

1
 co  current 
1  C*

Example : 4.8.1

Problem:

Consider again the system in Figure 4.18 in which 20 kg/s of water is to be cooled
from 80oC to 50oC using 25 kg/s of water entering at 20oC. Find heat transfer area
required for counter-current operation.

cp = 4180 J/(kg K)
U = 2000 W/(m2 K)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 120

Solution:

This is an example of “Design” calculations using effectiveness. As before:

.
Q  2508 kW
CH  20  4.18  83.6 kW/K
CC  25  4.18  104.5 kW/K

So Cmin = 83.6 kW/K and Cmax = 104.5 kW/K

Cmin 83.6
 C*    0.8
Cmax 104.5
.
Qmax  Cmin 80  20  83.6  80  20  5016 kW

2508
  0.5
5016

From chart or equation, NTU = 0.912

Cmin NTU
A  38.1 m2
U

.............................................

Example : 4.8.2

Problem:

a) How does the above countercurrent exchanger perform when the cold inlet
temperature rises to 25oC?

b) How does the above counter-current exchanger perform if, due to fouling, U
falls to 1.6 kW/(m2 K) and the cold inlet temperature is 20ºC?

Solution:

This is an example “Performance” calculations using effectiveness.

UA 2.0  38.1
a) NTU    0.912
Cmin 83.6
C 83.6
C*  min   0.8
Cmax 104.5

From chart or equation,  = 0.5

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 121

.
Qmax  Cmin 80  25  83.6  80  25  4598 kW

. .
Q   Q max  2299 kW

Th out  52.5o C and Tc out  47.0o C

b)
UA 1.6  38.1
NTU    0.7292
Cmin 83.6
Cmin
C*   0.8
Cmax

From chart or equation,  = 0.44

.
Q max  Cmin 80  20  5016 kW

. .
Q   Q max  2206 kW

Th out  53.6o C and Tc out  41.1o C

.............................................

4.8.2 Comparison of two Heat Exchanger Approaches

You should be able to work out this summary. Don't try to memorise it!

Design (Find A):

.
ΔTLM: Find: Q , Tout, ΔTLM, and then A.
. .
NTU: Find: Q , Q max , ε, C*, NTU and then A.

Performance (Know A):

.
ΔTLM: Guess one Tout and find Q , the other Tout, ΔTLM, and then A. If A is
not OK guess another Tout.
. .
NTU: Find: NTU, C*, Q max , ε, Q , and then Tout values. No trial-and-error!

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 122

4.9 Tutorials

1. Oil flows through the tube (ID = 12.7 mm) of a double pipe heat exchanger at
the rate of 0.189 kg/s. The oil is cooled by a counter-current flow of water,
which passes through the annulus. The water flow rate is 0.151 kg/s. The oil
enters the exchanger at 422 K and is required to leave at 344 K. The cooling
water is available at 283 K.
Oil side heat transfer coefficient based on inside area = 2270 W/(m2K)
Water side heat transfer coefficient based on inside area = 5670 W/(m2K)
Specific heat of oil = 2.18 kJ/(kg K)
The bit about "based on the inside area" might confuse you! In calculating the
UA value, multiply each film coefficient by the inside radius of the tube.

a) Find the outlet temperature of the water.

b) Find the heat transfer area required, ie the inside area of the tube. Neglect
the wall resistance.
c) What length of tube will be required?
d) Find the area required if both liquids passed through the exchanger in the
same direction (i.e. co-current flow).

Ans. a) 333.7 K, b) 0.269 m2, c) 6.73 m, d) 0.4 m2

2. A process liquor at 300 K is to be heated to 320 K using water at 366 K available

from another part of the plant. The flow rates of the liquor and the water will be
3.1 and 1.1 kg/s respectively. Previous experience indicates that an overall heat
transfer coefficient of 454 W/(m2K) will apply. Estimate the required area of a
counter-current heat exchanger.
Specific heat capacity of the liquor = 2.1 kJ/(kg K)
.
Ans. 6.87 m2 ( Q = 130.2 kW, ΔTLM = 41.8 K)

3. A single-pass shell-and-tube exchanger is to be used to cool a stream of oil from

125ºC to 55oC. The coolant is to be water, passing through the shell, which
enters at 21oC and leaves at 43oC. The flow pattern is counter-current. The
overall coefficient has a value of 170 W/(m2oC) based on the outside tube area.
The specific heat of the oil is 1.97 kJ/(kg oC).
For an oil flow of 24 kg/min, determine the total surface area required in the
exchanger.
If the exchanger is to be 1.8 m long, how many tubes in parallel, each 1.27 cm
OD, are required?

Ans. 5.95 m2, 83 tubes

4. A heat exchanger has an area of 100 m2 and an overall heat transfer coefficient
of 600 W/(m2 K). Use a trial and error procedure (using log-mean ΔT) to
determine the heat transferred and the outlet temperatures when operating a)
co-currently and b) counter-currently on the following streams.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 123

(Repeat using the "effectiveness" method.)

Hot fluid Cold fluid

Rate, kg/s 20 10
Inlet temperature, ºC 100 30
Cp, J/(kg K) 4000 5000
Ans: a) 1847 kW, Tcout = 66.9ºC, Thout = 76.9ºC
b) 2109 kW, Tcout = 72.2ºC, Thout = 73.6ºC

5. The feed stream to a reactor has a flow rate of 8.5 kg/s and is heated in a plate
heat exchanger by the reactor outlet stream and then by saturated steam at 1.5
bar in a shell and tube heat exchanger. The various stream temperatures are
shown in Figure 4.24. All streams have a specific heat capacity of 3.7 kJ/(kg K)
The plate heat exchanger has 50 plates with an area of 0.60 m2 per plate.

a) How many additional plates should be added to the plate heat exchanger to
raise its cold-side outlet temperature to 45oC? (The hot outlet temperature
will also have to change, of course.) The reactor temperature must be
maintained at 60oC. Assume that the overall heat transfer coefficient remains
unchanged.

Ans: 28 extra plates

b) If each heat exchanger plate costs £350 and steam costs £5.5/tonne,
calculate the time taken to recoup the cost of the extra plates through savings
in steam cost. The system operates 24 hours/day, 350 days/year with
negligible downtime.

Ans: 0.83 years of operation

30oC 40oC 60oC

10oC 60oC

Figure 4.24

6. Light oil with a flow rate of 3.0 kg/s and a heat capacity 3.2 kJ/(kg K), is to be
cooled from 85oC using 4.5 kg/s of cooling water with supply temperature 10oC
and heat capacity 4.2 kJ/(kg K).

a) Calculate the maximum amounts of heat that can be exchanged between

these streams in: i) counter-current, and ii) co-current operation
Ans: i) 720 kW ii) 477.5 kW

b) If we want as much heat transfer as possible but a minimum temperature

difference between hot and cold fluids of 7oC, calculate the two exit

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 124

temperatures and the heat transfer area required in counter-current flow.

Use an overall heat transfer coefficient of 1300 W/(m2 K).

Ans: 17oC, 44.5oC, 26.3 m2

c) Use the "effectiveness chart" or the corresponding equation, to estimate the

% reduction in heat load for the exchanger sized in (b) if the cooling water
supply temperature were increased to 15oC. Assume that the overall heat
transfer coefficient remains unchanged.

Ans: 6.7 %

7. A heat exchanger is cooling 3600 kg/h of organic fluid from 80ºC using 1500
kg/h of cooling water entering at 15ºC. The heat transferred has been calculated
to be 100 kW. The specific heat capacities of the organic fluid and water are 2.5
kJ/(kg K) and 4.2 kJ/(kg K) respectively. (Use either the log-mean ΔT method or
the effectiveness method for parts b and c; or try them with both methods.)

a) Is the heat exchanger counter-current or co-current? Explain.

b) If the overall heat transfer coefficient is 1000 W/(m2K), find the heat
exchanger area.
c) If the cooling water flow were doubled, giving an overall heat transfer
coefficient of 1200 W/(m2K), calculate the exit temperature of the organic
fluid.

Ans. 6.74 m2 25.3ºC

8. An acid is cooled by cooling water in a tubular heat exchanger. The acid flows
with an average velocity of 2.1 m/s inside the tubes, which are constructed of
two layers, the outer being steel of 3 cm od and 2 cm id, and the inner being a
corrosion resistant lining of 2 cm od and 1.6 cm id.

a) Determine the acid side heat transfer coefficient using:

Nu = 0.023 Re0.8 Pr0.33

b) Calculate the overall heat transfer coefficient for the tubes based on their
outside area.
c) What would be the overall heat transfer coefficient if the acid flow rate were
doubled?
d) Explain why this overall heat transfer coefficient is exceptionally low for
liquid/liquid heat transfer and why it is so insensitive to acid flow rate.

Density of acid = 1150 kg/m3

Heat capacity of acid = 2700 J/(kg K)
Viscosity of acid = 0.001 kg/(m s)
Thermal conductivity of acid = 0.42 W/(m K)
Thermal conductivity of steel = 45 W/(m K)
Thermal conductivity of lining = 1.1 W/(m K)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 125

Acid side fouling factor = 6000 W/(m2 K)

Water side fouling factor = 8000 W/(m2 K)
Water side film heat transfer coefficient = 3000 W/(m2 K)

Ans: a) 5210 W/(m2 K), b) 232 W/(m2 K), c) 240 W/(m2 K)

9. A stream of ethanol, 5000 kg/h, is to be cooled from 70oC to 50oC on the inside
of a countercurrent concentric tube heat exchanger, using 3900 kg/h of water
entering at 20oC as the cooling medium. Calculate the length of exchanger
required for this duty.

Data: Water side film coefficient = 1600 W/(m2 K)

Water side scale factor = 6000 W/(m2 K)
Inner pipe thermal conductivity = 45 W/(m K)
Ethanol side scale factor = 7000 W/(m2 K)

The ethanol side film coefficient can be calculated from:

Nu = 0.023 Re0.8 Pr0.33

Inner pipe dimensions: id = 30 mm od = 34 mm

Ethanol properties:
Specific heat capacity = 2.7 kJ/(kg K)
Viscosity = 0.00063 Ns/m2
Density = 760 kg/m3
Thermal conductivity = 0.16 W/(m K)
.
Ans: Q = 75kW, hethanol = 2.537 kW/(m2 ºC),
U = 0.783 kW/(m2 ºC) (based on inside area), Length = 32 m

10. A stream of toluene, 4000 kg/h, is to be cooled from 80oC to 50oC using
cooling water, which enters at 20oC and will be heated to 40oC. A counter-
current concentric tube heat exchanger is to be used with water flowing in the
annulus. Calculate the length of exchanger required for this duty.

Data: Toluene side film coefficient = 2000 W/(m2 K)

Toluene side scale factor = 7000 W/(m2 K)
Inner pipe thermal conductivity = 45 W/(m K)
Water side scale factor = 8500 W/(m2 K)

The water side film coefficient, h in W/(m2 K), may be calculated from:
h = 4280 ( 0.00488 T - 1 ) ( v0.8 / de0.2 )
T = temperature, K
v = velocity, m/s
de = "effective diameter" of annulus, m

Pipe Dimensions: Inner pipe id = 30 mm od = 34 mm

Outer pipe id = 49 mm od = 54 mm

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 126

Fluid Properties:
water toluene
Density 1000 825 kg/m3
Specific heat Capacity 4.18 1.82 kJ/(kg K)

.
Ans: Q = 60.7 kW, Water film coefficient = 3127 J/(m2 ºC),
U = 0.936 kW/(m2 ºC) (based on inside area), Length =19.8 m

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 4. HEAT TRANSFER EQUIPMENT 127

4.10 Bibliography
1. "C&R vol. 1" refers to Chemical Engineering, J.M.Coulson, J.F.Richardson et
al., vol. 1, 6th ed., Butterworth, 1999

2. "C&R vol. 6" refers to Chemical Engineering Design, R.K.Sinnott vol 6, 4th ed.,
Butterworth, 2005

3. Engineering Heat Transfer, W.S.Janna, PWS Engineering, 1986.

536.2JAN

4. Heat Exchanger Design Handbook, volumes 1-5, E.U. Schlunder et al.,

Hemisphere, 1983, 621.402HEA
Everything (?) about heat exchangers.

5. Chemical Process Design and Integration, R. Smith, Wiley 2005,

660.281SMI Quite good section on heat exchangers including how to decide
how many shells to use.

6. Low Reynolds Number Flow Heat Exchangers, S. Kakac et al., Hemishere,

1983, 621.402LOW
Might be useful for plate heat exchangers.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

 Maximum Normal size ranges
Exchanger Type Temperature Range Fluid Limitation Special Features
Pressure for individual unit
30.7 MPa (for
-200 to 600ºC(higher Subject only to 1 to 1000 m2 per Very flexible; can
higher
Shell and Tube with special materials of shell; multiple shells be used for nearly
pressure, codes not
materials) construction can be used all applications
applicable)

Double Pipe > 30.7 MPa (shell) Subject only to

Standard modular
(plain materials > - 100 to 600ºC materials of 0.25 to 200 m2
construction
and finned tubes) 140 MPa (tube) construction

1.6 MPa (possible Modular

-25 to 175ºC(-40 Often unsuitable for
to construction;
Gasketed Plate to 200ºC possible gas or 2-phase
go up to 2.5 MPa 1 to 200 m2 normally the most
Exchangers for very special flow; limitation is on
depending on size economic if
types) gaskets
and temperature) applicable

High heat transfer

Subject only to efficiency;
materials of low maintenance;
Spiral Heat
1.8 MPa Up to 400ºC construction; may Up to 200 m2 self cleaning;
Exchangers
be suitable for low installation
fouling duties cost

Differential
Subject only to
pressure should
materials of
Welded Plate 3.0 MPa (higher in In excess of be <3.0 MPa;
construction; not >1000 m2
Exchangers shells) 400ºC differential
suitable for fouling
expansion should
duties
be borne in mind

©HERIOT-W ATT UNIVERSITY B48BD October 2014 v1

TOPIC 5. HEAT EXCHANGER DESIGN 129

Topic 5

Heat Exchanger Design

Contents

5.1 Introduction 131

5.2 Hydraulic Mean Diameter 132
5.3 Double Pipe (Concentric Tube) Heat Exchanger 133
5.4 Mean Temperature Difference in Multi Tube-Side Pass Exchangers 138
5.5 Shell and Tube Heat Exchanger 143
5.5.1 Tubes 145
5.5.2 Tube Side Velocity 146
5.5.3 Shells 146
5.5.4 Baffles 147
5.5.5 Fluid Allocation 148
5.6 Film Heat Transfer Coefficient 149
5.6.1 Tube-Side (Single Phase) 149
5.6.2 Shell-Side (Single Phase) 149
5.7 Heat Transfer in Vessels Using Jackets and Coils 151
5.7.1 Heat Transfer Coefficients for Coils 153
5.7.2 Heat Transfer Coefficient for Jackets 153
5.8 Transient Heat Transfer 154
5.8.1 Lumped Capacitance 154
5.8.1.1 Unsteady State Heating or Cooling 155
5.8.1.2 Time Constant for Heating and Cooling 156
5.8.2 Distributed Capacitance 158
5.9 Tutorials 163
5.10 Bibliography 165

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 130

Prerequisite knowledge

 Detailed understanding of different modes of heat transfer.

 General knowledge about heat transfer equipment.
 General knowledge about the application of shell & tube heat exchangers.
 General knowledge about the application of plate heat exchangers.
 Understanding the main concept of using mean driving forces.
 Analytic solutions for the effectiveness of heat exchangers.
 Familiarity with Performance/Simulation calculations
 Familiarity with NTU method.

Learning objectives

By the end of this topic, you should be able to:

 Distinguish between different types of heat transfer coefficients.

 Carry out a basic design of shell & tube heat exchangers.
 Understand transient heat transfer.
 Distinguish between unsteady state and time constant processes.
 Use analytic solutions for design of heat exchangers.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 131

5.1 Introduction
In the previous Topic, you were familiarised with different types of heat exchangers
and modes of operations. We learnt that shell-and-tube heat exchangers are
commonly used in industry as they can handle any combination of heating, cooling,
boiling and condensation.

The first choice in heat exchangers is usually the shell and tube exchanger.
Selecting the most appropriate pattern depends mainly on the fouling nature of the
fluids, affecting the need for cleaning, and temperature change from ambient
influencing the need for an expansion joint. The number of passes required in each
side will be found during the design process.

The double pipe exchanger may be chosen if the flowrates are small. Plate or spiral
exchangers are used where temperature difference between streams is low.

There are several methods to find the overall size of a shell and tube, each has its
own merits but the one outlined here is de to Kern.

The steps in designing a heat exchanger are:

1. Define the required duty as far as possible: heat transfer rate, fluid flowrates
and temperatures.

2. Obtain fluid physical properties (densities, viscosities, thermal conductivities) at

the relevant temperatures.

3. Select the type of heat exchanger to be used.

4. Estimate an initial trial value of U the overall heat transfer coefficient

5. Calculate the log mean temperature difference.

6. Calculate required heat transfer area A.

7. Select an appropriate exchanger layout.

8. Calculate film coefficients and fouling coefficients.

9. Calculate the overall heat transfer coefficient U based on step 8 results.

10. Compare U from step 9 with step 4. If the difference is significant, resume from
step 5 using the step 9 value of U.

11. Calculate pressure drops in each fluid stream. If these are unsatisfactory,
return to step 7 or 4 or 3 (in order of preference) and try a different choice.

12. Optimise the design by repeating steps 4 to 10 with alternative arrangements

to determine the lowest cost (usually smallest area) solution.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 132

5.2 Hydraulic Mean Diameter

 vd   hd 
When you use Reynolds   or Nusselt   it is obvious what value to use for
    k 
d if you are dealing with flow in a pipe, but what if the pipe is not circular? In this
case the hydraulic mean diameter, de, is used instead of diameter:

4  flow area
Hydraulic mean diameter: de 
EffectivePerimeter

This is simply an approximation to try and give an equivalent to diameter based on

a flow profile of the pipe or channel.

To calculate de for a circle, you should find de = d as you would expect.

For a rectangular section (L = length; B = breadth):

4LB
de 
2L  2B

For an annulus of ID = d1 and OD = d2 :

d 2 2 d12 
Flow Area    (d 2  d1 )
2 2

4 4 4

For Heat Transfer the effective perimeter =  d1 hence

d 2  d1
2 2
de 
d1

For Pipe Friction the effective perimeter   (d 2  d1 ) hence

de  d 2  d1

To a large extent this finishes the heat transfer and heat exchanger material; the
remainder of the notes illustrates design through three examples:

 Concentric Tube Heat Exchanger

 Shell and Tube Heat Exchanger

 Coil in a Tank.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 133

5.3 Double Pipe (Concentric Tube) Heat Exchanger

Basic equations for a double pipe (concentric tube) heat exchanger are:

.
Q  mh c p h (Th in  Th out )  mc c p c (Tc out  Tc in )

.
Q  U o Ao TLM

 r  
 ln o  
  i 
1 1  1 1 r 1 1 
  
U o Ao 2L  ho ro hdoro kw hi ri hdiri 
 
 

In a double pipe heat exchanger, there is a flow through an annulus. Therefore, we

use hydraulic mean diameter.

Concentric tube units are available as "hairpins" of 4 m or 5 m length, i.e. 8 or 10 m

length per unit. Figure 5.1 shows the hairpin units for a concentric heat exchanger.

Figure 5.1, Hairpin units for double pipe H/E.

Example : 5.3.1

Problem:

Toluene was used to heat 4,250 kg/h of benzene from 25oC to 45oC. For this
purpose, toluene needed to be cooled from 70oC to 40oC. The maximum allowable
pressure drop on each side was 100 kN/m2. Select a suitable heat exchanger and
design the process.

Solution:

Select appropriate standard tube sizes. You have to have a guess first. You would
choose standard tubing– see table in Fundamentals Fluids notes.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 134

Here both tubes are schedule 40, the inner tube is 1¼ inch and the outer tube is 2
inch.

Inner Tube: 3.51 cm id 4.22 cm od

Outer Tube: 5.25 cm id 6.05 cm od

 
Flow area of inside tube =    3.51  9.68 cm2
2

4

 
 
Flow area of annulus =    5.252  4.222  7.66 cm2
4

Table 5.1, Physical properties of benzene and toluene*.

T h cP   k
(oC) kJ/kg kJ/(kg K) kg/m3 cP W/(m K)
Benzene:
25 642.24 1.520 872.2 0.6050 0.1310
35 657.67 1.564 861.5 0.5287 0.1284
45 673.53 1.609 850.5 0.4654 0.1258
Toluene:
40 165.34 1.625 850.4 0.4624 0.1323
55 190.20 1.690 836.2 0.3964 0.1286
70 216.03 1.754 821.8 0.3433 0.1248
*(Data in this table are from Aspen Hysys – you will probably use this package in 3rd year.)

Heat Balance:

.
For each fluid: Q  mass flow  change in specific enthalpy
.
For benzene: Q  1.18 (673.53- 642.24) 36.9 kW

For toluene: 36.9  mass  (216.03-165.34)  mass  0.73kg/s

Mean Temperature Driving Force:

With counter-current flow:

T1 = 70 – 45 = 25ºC

T2 = 40 – 25 =15ºC

TLM = 19.6ºC

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 135

Fluid allocation:

Put the smaller flow in the annulus since it has the smaller flow area. In this case
this is the toluene flow. Physical properties, particularly viscosity, are temperature
dependent so it can be difficult to know what values to use. A simple method is to
use properties appropriate to the average temperature of the stream; for this design
the average temperatures for the hot and cold sides are 55ºC (average of 40 and
70ºC) and 35ºC (average of 25 and 45ºC) respectively. You can read more about
this if you wish in C&R vol. 6 page 661 section 12.7.5.

Inside Film Coefficient, hi:

Take benzene properties at its average temperature, 35oC.

mass flow 1.18

Volumetric flow    1.373103 m3/s
density 861

volumetricflow 1.373 103

Velocity    1.42 m/s
flow area 9.68  104

Using the Colburn Equation:

Nu  0.023Re 0.8 Pr0.33

We get h i  1314 W/(m2 K)

Outside Film Coefficient, ho:

Take toluene properties at its average temperature, 55oC

mass flow 0.73

Volumetric flow    0.873 103 m3 /s
density 836
volumetricflow 873103
Velocity    1.14 m/s
flow area 7.66 10 4

The effective diameter for heat transfer from an annulus is:

d 2
2 
 d12
 2.31 cm
d1

Using again the Colburn Equation:

Nu  0.023Re 0.8 Pr0.33

We get h o  1376 W/(m2 K)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 136

In the previous Topics we learnt that in the Colburn equation, the viscosity should be found
at the average film temperature. But that is hard work! We will look later at the validity of
just using the viscosity at the bulk fluid temperature.

Overall Heat Transfer Coefficient:

In the following equation, o refers to the outside of the inner tube that is carrying the
benzene; and i refers to the inside of the inner tube.

 r  
 ln  o  
  i  1  1 
1 1  1 1 r
 

U o Ao 2L ho ro hdo ro kw hi ri hdi ri 
 
 
Where,

kw (steel) = 45 W/(m K)

Take hdi = hdo = 5500 W/(m2 K) (typical values from C&R vol. 1)

Consider 1 m length of heat exchanger:

Ao    0.04221.0  0.1325 m2
Ai    0.03511.0  0.1103 m2

Hence, Uo = 471 W/(m2 K)

Note that for clean tubes, hdi = hdo = ∞  Uo = 579 W/(m2 K), so the dirt factors have
quite a significant effect in this case.

Area And Tube Length:

.
Q
Ao 
U o TLM

Clean Dirty
Area of outside of inner tube 36900 36900
 3.25 m 2  3.99 m 2
579 19.6 471 19.6
Area
Length, L 
 do 24.5 m 30.1 m

Select four 8 m units giving a length of 32 m.

Pressure Drops:
L 2
P  4 v
d

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 137

Where  is the friction factor = f (Re, roughness)

Take absolute roughness of steel pipe = 0.046 mm

Inside:

Re = 81200, Relative roughness = 0.046/35.1  0.001

 = 0.00275 (the relevant chart is not in these notes but see e.g. C&R vol
1 page 66 or the Foundation Fluids notes)

32
P  4  0.00275  8611.422  17 kN/m2
0.0351

Outside:

For fluid flow, de = 52.5 – 42.2 = 10.3 mm

Re = 25600, relative roughness = 0.046/10.3  0.0045

 = 0.0040

32
P  4  0.0040  836 1.142  54 kN/m2
0.0103

Both Ps are within the allowed range of up to 100 kN/m2 though the values would
be slightly higher than those calculated if bends etc were allowed for.

Estimation of mean film temperatures:

(to check validity of our "short cut" with respect to viscosity in Reynolds Numbers):

 
.
Q  ho Ao Toutside film  36900 1376 3.99  55  Ttol film  Ttol film  48.2oC

A sort of average toluene film temperature is, therefore (55 + 48.2)/2 = 51.6ºC

di 3.51
Ai  Ao  x 3.99  3.319m2
do 4.22

 
.
Q  hi Ai Tinside film  36900 1314 3.319 Tbenz film  35  Tbenz film  43.5o C

A sort of average benzene film temperature is, therefore (35 + 43.5)/2 = 39.3ºC.
You could go back and do the calculation for the outside film coefficient again – but
only if you are ludicrous! The correlations are only roughly correct in the first place.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 138

5.4 Mean Temperature Difference in Multi Tube-Side Pass

Exchangers

Figure 5.2 shows the temperature profile in a countercurrent shell-and-tube heat

exchanger with 1 shell-side pass and 1 tube-side pass.

T1 T1

t1

t2 T2 t2
T2
t1

Figure 5.2, Shell-and-tube temperature profile; shell-side dotted.

As a good approximation, since the flow is more or less counter-current, the

appropriate mean driving force is TLM .

Tmean = TLM

Figure 5.3 shows the situation with 1 shell-side pass and 2 tube-side passes.
Notice that, in one direction, the tube-side flow is counter-current and in the reverse
direction it is co-current. The bottom diagram shows how a temperature cross can
arise in this arrangement, which you would not expect if you just looked at the end
temperatures from which you would see that the cold outlet is colder than the hot
inlet, and the hot outlet is hotter than the cold inlet.

T1 T1
t2

T2
t1
T2
t1

T1
T1 t2

t2 t1
T2
T2
t1

Figure 5.3, Shell-and-tube temperature profile; two pass tube-side.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 139

For cases like this (i.e. cases in which the flows are not simply counter-current or
co-current), Tmean < TLM and a correction factor FT can be used such that,

Tmean  FT Tlog mean, countercurrent

FT is the same for both of the configurations in Figure 5.9 and depends on two
ratios of temperature differences (R and S) defined in C&R vol. 6 as:

R
T1  T2 
t2  t1 

S
t2  t1 
T1  t1 
T refers to the shell side and t to the tube side. 1 refers to the stream inlet and 2 to
(mCp ) tube
the outlet. You may be able to see that R is the ratio and S is the ratio of
(mCp ) shell
heat gained by the tube-side fluid divided by the heat transfer if it reached the shell
inlet temperature.

Some books use a different nomenclature, for example C&R vol. 6 defines R and S
in terms of tube-side and shell-side fluids but other books do this in terms of hot
and cold fluids. The value of FT is the same whatever you do so long as 1 = inlet
and 2 = outlet.)
For some arrangements, such as 1 shell pass and 2 tube passes, there is an
analytic solution for FT (see C&R vol. 6 p656) but it is generally more convenient to
obtain it from charts such as Figure 5.4. For a good design, look for FT greater than
0.85 and certainly not less than 0.75.

Figure 5.4, Temperature correction chart for one shell pass; even number of tube
passes. Source: C&R vol. 6 p657

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 140

Example : 5.4.1

Problem:

Duty
A hot fluid is to be cooled from 100oC to 50oC in a 1 shell-pass, 2 tube-pass
exchanger. Cooling fluid is available at 20oC and will exit at 40oC. Figure 5.5 shows
a sketch of the arrangement with an indication of the temperature profile.

100 100oC

40oC

40 50

20 20oC
50oC

Figure 5.5, 1 shell-pass, 2 tube-pass exchanger with temperature profile

a) Determine the mean temperature difference for this exchanger.

b) Repeat for the same coolant temperature rise but an inlet temperature of
i) 30oC, ii) 40oC.
c) Repeat for double the coolant flow (half the temperature rise) and an inlet
temperature of 40ºC.

Solution:

Temperature ratios R and S:

100  50
R  2.5
40  20

40  20
S  0.25
100  20

We can find FT from Figure 5.10: ⇒ FT = 0.9

Mean temperature difference:

(100  40)  (50  20)

Tmean  FT TLM  0.9  39.0 ºC
 100  40 
ln 
 50  20 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 141

Table 5.2 gives the other results – based on the equation from C&R vol. 6 p656 –
just check that the values for FT are reasonable from the chart.

Table 5.2, Effective mean temperature driving force for the four cases
Cold in Cold out R S FT Tlog mean Tmean
t1 t2
a) 20 40 2.5 0.250 0.901 43.3 39.0
b) i 30 50 2.5 0.286 0.807 32.7 26.2
b) ii 40 60 2.5 0.333 <0.5 21.6 "small"
c) 40 50 5.0 0.167 0.817 24.9 20.7

You should try plotting the temperature profile for each case as was done in Figure
5.11 for case (a). You will see that in case (b) there is a “pinch” at 50ºC giving a
very small mean driving force – there may even be a temperature cross at some
point in the exchanger. A higher coolant flow rate, as in c), removes this problem
and gives a better FT value.

Another way to avoid the pinch or temperature cross is to use an alternative shell-
side configuration such as the 4 tube-side pass, 2 shell-side pass exchanger
shown in Figure 5.6.

If you look at the temperature profile in this figure you may think that there is a
temperature cross but for any point on the cold stream you will find that it is
exchanging heat with a higher temperature hot stream.

For example, when the cold stream is at 60ºC it is in contact with the hot stream at
100ºC.

T1=100oC
t2= =60oC

100

60
t1=40oC
50
T2=50oC
40

Figure 5.6, 4 tube-side pass, 2 shell-side pass exchanger

Using charts in C&R vol. 6 p657/658, we can find FT values for alternative shell-
side configurations. Still for R = 2.5, S = 0.333.

Two shell passes: FT = 0.89 (C&R vol. 6 Fig. 12.20 – 4 tube-side passes)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 142

Divided flow shell: FT too small (C&R vol. 6 Fig. 12.21)

Split flow: FT = 0.89 (C&R vol. 6 Fig. 12.22)

Heat exchanger design involves the iterative solution of:

.
Q  UA Tmean

Solution is iterative because U depends on the inside and outside film coefficients,
which in turn depend on flow velocities and hence on the heat exchanger flow
patterns on both the shell and tube side.

To start the procedure it is useful to have an approximate value for U. C&R vol. 6
Table 12.1 and Figure 12.1 gives typical values for shell-and-tube exchangers, one
or two of which are given in Table 5.3.

Table 5.3, Typical overall heat transfer coefficients

Hot fluid Cold fluid U W/(m2 K)
Water Water 800 – 1500
Organic solvent Organic solvent 100 – 300
Gas Gas 10 -50

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 143

5.5 Shell and Tube Heat Exchangers

Shell and tube exchangers comprise an outer cylindrical vessel, the shell. The shell
contains one of the fluids and, immersed in the fluid, a bundle of tubes which
contain the second fluid.

The shell has two nozzles for inlet and outlet of the shell side fluid. Similarly, there
are two nozzles for the tube side. The tube sheets at each end of the shell seal the
shell side fluid in the shell. The tubes are sealed into the tube sheets. The heads
feed tube side fluid into and out of the tubes. The baffles serve both to support the
tubes inside the shell and also to achieve a high velocity of the shell side fluid
across all the tubes’ surface.

This pattern is “fixed head” since both heads are rigidly attached to the shell. This
is cheap to manufacture but means the shell side cannot be mechanically cleaned
and if the operating temperature is significantly different from ambient then
differential expansion of the shell and tubes requires an expansion joint.

In many cases, the fluid velocity must be adjusted to achieve the required heat
transfer coefficient. This can be achieved by altering the number of passes by the
fluid through the tubes or through the shell.

One head is divided in two and the inlet and outlet nozzles are on opposite sides of
that head. Tube side fluid flows in one direction in half the tubes and in the opposite
direction in the other half. More tube side passes can be achieved by dividing both
heads into more compartments suitably aligned. A longitudinal baffle can be used
to achieve 2 passes on the shell side. It is unusual to have more than two shell side
passes. In both of these configurations in Figure 5.7, the tube bundle is
permanently fixed within the shell so mechanical cleaning of the outside of the
tubes is impossible.

Figure 5.7, Fixed head single pass shell-and-tube heat exchangers.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 144

Cleaning inside the tubes is readily achieved by removing a cover plate from one
head. To facilitate cleaning outside the tubes, the tube bundle is secured to only
one fixed tube sheet. The other end of the tube sheet is attached to either a floating
head or the tubes are in a U-tube bundle (Figure 5.8).

Figure 5.8, Fixed head single pass shell-and-tube heat exchangers with a U-tube.

The U-tube bundle is the cheapest way to implement access to the outside of the
tubes. But cleaning inside the tubes is less simple.

Internal floating head type exchangers are more versatile than fixed head or U tube
exchangers. Since the tube bundle can be cleaned right through inside and can be
removed to clean the outside of the tubes, fouling liquids can be accommodated.
But they are significantly more complex and costly.

The Pull Through floating head design (Figure 5.9) is simpler but the large
clearance required between the outermost tubes and the shell allows some shell
side fluid to bypass the tubes.

Figure 5.9, Pull Through floating head; Split Ring floating head

Both these designs require gasketted joints inside the shell. These are prone to
leaking which may go undetected. The external floating head arrangement is
outside the shell. The joint between head and shell is a sliding gland with a stuffing
box. This is also prone to leaking. It is not suitable for pressures over 20 bar or for
flammable or toxic shell side fluids.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 145

5.5.1 Tubes

Tubes for heat exchangers are normally specified according to their true outside
diameter to a strict tolerance. C&R vol. 6 p645 gives some information on the
standard lengths and diameters. Although they may be quoted nowadays in
millimetres they are based on feet and inches.

The commonly used sizes are 5/8th inch (16mm) to 2inch (50mm). Up to 1 inch
diameter is generally preferred as these sizes give smaller exchangers but larger
sizes are easier to clean so they are selected where fouling is a serious problem.
Standard lengths are 6 feet (1.8m), 8 feet (2.4m), 12 feet (3.7m), 16 feet (4.9m), 20
feet (6.1m) and 24 feet (7.3m). For a fixed heat transfer area, longer tubes reduce
the shell diameter which reduces cost especially of a high pressure exchanger.
Optimum tube length to shell diameter ratio is normally in the range from 5 to 10.
This sort of ratio is often called an “aspect ratio”.

Tubes are available in several metals for high thermal conductivity and suitable
corrosion resistance. Wall thickness will depend on the working pressure. There
can be no compensation of the tube wall thickness for the shell pressure since the
tubes must not fail even if all pressure is lost from the heat exchanger shell.

There is a limit on the closeness of tubes in the bundle since the tube sheet must
be able to withstand the pressure in the shell.

The shortest distance between tubes is the “clearance”. “Pitch” is the distance
between centres of adjacent tubes. Tubes are laid on either a square or triangular
pitch.

Square pitch allows external cleaning and gives lower shell-side pressure drop
when flow is in line with the pitch. Rotated square pitch and triangular pitch give
enhanced heat transfer.

External cleaning of triangular pitch tubes can be achieved if the clearance is large
enough to leave open cleaning lanes.

Figure 5.10, Fixed head single pass shell-and-tube heat exchanger.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 146

5.5.2 Tube Side Velocity

Tube side velocity can be influenced by selection of the number of tube side
passes. The passes are produced by dividers or "pass partitions" in the tube
headers. Table 5.4 explores the effect of number of tube-side passes on various
parameters. You can fill in the values for six passes yourself.

A single tube-side pass gives almost pure counter-current (or co-current) operation
but a multipass arrangement (usually an even number) allows the designer more
choice of tube-side velocity. Probably most designs are at least two-pass on the
tube side.

Table 5.4, Effects of tube-side passes taking a single pass as the basis.
Passes Flow Area Tube-side Path length Tube-side Tube-side
velocity heat transfer pressure
coefficient drop
1 1 1 1 1 1
2 0.5 2 2 1.74 8
4 0.25 4 4 3.03 64
6

The last two columns are estimated, assuming turbulent flow, from:

Heat transfer coefficient hi  Re 0.8

Pressure drop P  L v 2

Effective heat transfer design is generally a trade-off between heat transfer

coefficient and pressure drop. Fluid velocity, both on tube and shell side, is
generally a good guide to whether these factors are well balanced.
Typical tube-side velocities are 1 to 2 m/s for liquids and 30 m/s for gases– see
C&R vol. 6 p660 for more details.

5.5.3 Shells

Shells are either large pipes or, above about 610 mm, made from rolled plate. The
diameter of tube bundle (Dbundle) depends on the number of tubes (NT), outside
diameter of tubes (do), tubes arrangement (triangular or square), and the number of
tube passes and pitches.

C&R vol. 6 (p648 and Table 12.4) gives an approximate relationship between the
number of tubes and the diameter of the bundle:
n1
D 
Nt  K1  bundle  where K1 and n1 are constants.
 do 
Dshell depends on Dbundle and the type of head - see C&R vol. 6 Figure 12.10.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 147

Shell types are categorised based on the pass arrangements (see Figure 5.11).

E-1 E-2 E-6 E-5

I-1 I-5

(a) (d)

E-10 E-9 E-7 E-8

I-7 I-6

(b) (e)

E-12 E-11

I-8

(c)
Figure 5.11, Shell types (pass arrangements); (a) one-pass shell (E shell), (b) split
flow (G shell), (c) divide flow (J shell), (d) two-pass shell with longitudinal baffle (F
shell), (e) double split flow (H shell). Source: C&R vol.6 p650.

5.5.4 Baffles

The use of baffles on the shell-side influences shell-side velocity and hence the
shell-side film heat transfer coefficient and pressure drop. Liquid velocities are
about 0.3 to 1 m/s for liquids. Gas velocities are similar to those used on the tube
side. Figure 5.12 shows some typical baffle arrangements.

(a)

(b)

(c)

Figure 5.12, baffle arrangements; (a) segmental baffles (typically 0.3 to 0.5 shell
diameters apart), (b) segmental and strip baffles, (c) disc and doughnut baffles.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 148

5.5.5 Fluid Allocation

Selecting which fluid to put in the shell and which in the tubes depends on several
factors. The choice often involves compromising between these factors.

The reasons for choosing a tube side fluid include:

 Corrosion. Putting more corrosive fluid in tubes reduces cost of corrosion

resistant metal.

 Fouling. Higher velocity in tubes gives some self-cleaning. Tube insides

are easier to clean.

 High temperature. If the temperature is so high that special metal is

required, cost is kept down by putting it inside tubes. Putting hot fluid in the
tubes reduces shell surface temperature, reducing the need for lagging.

 High pressure. High pressure tubes are cheaper than a high pressure
shell.

 Gases. If the fluid contains vapour or gas, heat transfer coefficient is

greater on the tube side; gas will not gather in pockets on the tube side as
they may do on the shell side.

Allocate a fluid to the shell side if:

 Pressure drop. For a given pressure drop, heat transfer coefficients will be
higher on the shell side; fluid with lower allowable pressure drop should be
on the shell side.

 Viscous fluid. Higher heat transfer coefficient can be obtained with viscous
fluid on the shell side, provided turbulent flow is achieved (Re>200).

 Gases. Pure gases or boiling liquids achieve higher heat transfer

coefficients on the shell side

 Low flowrate. A lower cost exchanger is generally achieved by allocating

the lower flowrate stream to the shell side.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 149

5.6 Film Heat Transfer Coefficient And Pressure Drop

See C&R vol. 6 Section 12.8 p662. The following notes clarify (perhaps) the
section in C&R vol. 6.

5.6.1 Tube-Side (Single Phase)

In our concentric tube heat exchanger example we used different values of

"hydraulic mean diameter, de” for heat transfer and for pressure drop.

Laminar flow would not normally be expected in a process heat exchanger, so the
turbulent flow correlations are much more likely to apply.

Correlations using Nu or the "jh factor v Re" plot are provided.

For the film heat transfer coefficient equation hi (t w  t )  U (T  t ) is strictly correct

only if "U" is "Ui".

For calculation of the pressure drop the method is based on:

 L  u
2
P  8 j f  
d  2
which is a version of
 L
P  4   v 2 .
d 

Adding an extra 2.5 "velocity heads" per tube pass to allow for expansions,
contractions and flow reversals gives:

  L    m  u 2
P  N P 8 j f     2.5 t
   
  di   w   2
 

The viscosity correction term (μ/ μw) is often neglected.

5.6.2 Shell-Side (Single Phase)

Concentrate on Kern’s method in C&R vol. 6 Section 12.9.3 p671. Use jh and jf
factor charts again.

Shell-side velocity, vs, is taken to be the velocity at the centre-line of the shell
shown in Figure 5.13. If Lb is the distance between the baffles, the cross-sectional
area for flow in an empty shell would be DsLb.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 150

But, of course the tubes get in the way and reduce the area. C&R vol. 6 p672 (p674
in 3rd ed.) gives the formula to account for this.

Shell-side hydraulic mean diameter = 4 x (free area/ perimeter)

vs

Figure 5.13, Idealised flow round segmental baffles.

Looking at Figure 5.14 and square pitch, the area for shell-side flow is the area of
the square less four quarter-circles. The wetted perimeter is four quarter-
circumferences.

Pt   d o
2 2
de 
 do

which leads to C&R vol. 6 equ. 12.22 page 672 (p674 in 3rd ed).

do

Pt

Figure 5.14, Square and triangular tube pitches.

Pressure Drop:
0.14
D  L  u 2   
P  8 j f  S    
 de  lb  2   w 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 151

5.7 Heat Transfer in Vessels using Jackets and Coils

Reaction vessels, fermenters and storage tanks often use heating coils or jackets
as shown in Figures 5.15 and 5.16.

TC,in TC,out
Coolant in

Process
Fluid in TP,in

TP,out
TV

dcoil

Figure 5.15, Vessel with heating or cooling coil.

Steam Coolant out

Condensate Coolant in

Figure 5.16, Jacketed vessels.

.................................................

Example : 5.7.1

Problem:

In a reacting vessel, the fluid flow is 5 kg/s and the inlet temperature of the fluid TP,in
is 30oC. Reaction process takes place at TV = 80ºC. Having the following
information about the process, carry out the energy balance and determine the
mean driving force and the water flowrate required for this process.
 Heat evolved by reaction = 600 kW
 Cooling water is supplied at 25ºC and is heated to 10ºC below reactor
temperature.
 Specific heat capacity of process fluid = 1800 J/(kg K).
 Specific heat capacity of cooling water = 4180 J/(kg K).
 Assuming perfect mixing in tank, TP,out = 80ºC

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 152

Solution:

Systematically: Basis: 0ºC and liquid.

Enthalpy balance for the process stream:

. .
Inlet enthalpy Q  Qr  Outlet enthalpy
.
5 1.8 30 - 0  Q  600  5 1.8  80  0
.
 Q  150 kW
The negative value means that the heat is removed from the process fluid.

For the CW:

.
Inlet enthalpy Q  Outlet enthalpy
m  4.18 25 - 0  150  m  4.18 70  0
 m  0.8 kg/s

Alternatively:
.
Heat removed by cooling coil Q = Reaction heat – Heat removed by the process fluid
.
 Q  600 - 5 1.8 80 - 30  150 kW

150
Water flow required   0.8 kg/s
4.18 (70  25)
Heat transfer driving force at:

CW inlet = 80 – 25 = 55ºC

CW outlet = 80 – 70 = 10ºC
Hence,
55  10
TLM   26.4 ºC
 55 
ln 
 10 
...............................................

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 153

5.7.1 Film Heat Transfer Coefficients For Coils

Inside coefficient:  
hi coil  hi straight 1 3.5 d pipe / dcoil 
Notice that the inside film coefficient is larger in a coil than a straight pipe. Is that
intuitively reasonable?

Outside coefficient: This depends on the fluid properties but also on the degree of
agitation and on the geometry of the impeller and coil.
0.14
 
Nu  0.87 Pr0.33 Re 0.62  
 s 
L2 N
"Agitation" Reynolds number Re 

dv = vessel diameter
L = impeller diameter
N = impeller speed (revs/unit time)
s = viscosity at surface temperature

Note that the tip speed of the impeller is πLN. This is the circumference of the circle
travelled by the tip of the impeller times the revolutions per second.

5.7.2 Film Heat Transfer Coefficients For Jackets

Inside coefficient:
Correlations are similar in form to those for the outside of coils,

d
c
Nu  a Pr Re  
b

 s 

Outside coefficient:
The geometry and flow path make this difficult to estimate for a liquid coolant. The
film heat transfer coefficient for condensing steam will be very large, so 1/ho will be
very small, and will not contribute significantly to U.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 154

5.8 Transient Heat Transfer

So far we have used steady-state analysis – there is no change with time. Now we
look at unsteady-state or transient analysis. There are two approaches: lumped
capacitance where there is no spatial variation (easy); and distributed capacitance
(difficult).

5.8.1 Lumped Capacitance

Look at Figure 5.17. A sphere is being heated (or cooled) by convection. The
temperature of the sphere is T everywhere – there is no spatial variation.

T Ts

Figure 5.17, Sphere with convection.

Analysis of heating or cooling of the sphere is easy. Suppose the sphere is being
heated with steam at temperature Ts. If we neglect radiation, over a short period (an
element) of time, dt:

Change in the sphere = Heat transfer by convection

mCp dT  hATs  T dt

mCp dT
 dt
hA Ts  T
mCp Tf dT t

hA 
Ti Ts  T 
 dt
0

mCp  T T 
ln s i t
hA  Ts  T f 

In the above:
t=0 T = Ti
t=t T= Tf

We could include radiation heat transfer and/or heat generation in the sphere but
the integration would be more difficult – perhaps we would need to do it
numerically.

You should think about when this lumped capacitance approach would be valid:

 If the contents of the sphere were mixed.

 If the convection resistance is much more than the conduction
resistance.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 155

You do not have to know this but one way of judging where the lumped capacitance
approach is valid is to work out the Biot Number.
hL
Biot number is Bi  and if Bi<0.1 the approach is valid. The characteristic length, L, is
k
the ratio of volume/surface area. Let’s repeat the analysis for heating or cooling a
tank.

5.8.1.1 Unsteady State Heating or Cooling

Consider the batch heating process shown in Figure 5.18. We will assume that the
contents of the vessel are well-mixed and that there is no heat transfer to or from
the surroundings.

Steam

Condensate

Figure 5.18, Batch heating.

m = the mass of liquid in the vessel.

cp = liquid specific heat capacity.
T = liquid temperature at any time.
TS = steam temperature.

Heating is applied at time, t = 0, when liquid temperature, T = T0. The temperature

response will be something like that shown in Figure 5.19.
Temperature, T

Ts
T+dT
T

To

0 t t+dt Time, t

Figure 5.19, Temperature response of vessel contents with batch heating.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 156

Suppose in time dt, the liquid heats up by an amount dT.

Heat transferred from steam = Heat gained by liquid

UA(Ts  T )dt  mcp dT

t T
UA dT
mcp 0 
dt  
T T
T s0

UA  T T 
t  ln s o  ................................................................(5.1)
mcp  Ts  T 

UA(T-To ) mcp
Multiply both sides by (T  To )   (T  To )
 Ts  To  t
ln 
 Ts  T 

Add and subtract T3  UA((T3  To )  (T3  T ))

m c p T
UATLM  .............................................................. (5.2)
t

Note that the U, A, m, and cp have been assumed to be constant.

Depending on the situation, equation (5.1) or (5.2) could be the more useful.
Equation (5.2) is interesting since it is similar to the form for continuous flow:

UATLM  mcp T

5.8.1.2 Time Constant for Heating and Cooling

An important chemical engineering concept, especially in control, is that of time

m cp
constants. The time constant in this batch process is . Check that this group
UA
has units of time. (Using equation (5.1), can you show that when t equals the time
constant, 63.2% of the possible change of temperature of the batch has taken
place?)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 157

Example : 5.8.1

Problem:

Figure 5.20 shows a thermocouple junction. Find:

 The junction diameter to give a time constant of 0.5 s.

 The time for the junction to reach 199oC starting at 25oC.

Figure 5.20, Thermocouple (Incropera et al., 2007) – with permission.

Solution:

Time constant,
mCp
tc 
hA
4
m  V  r 3 
3
A  4r 2

Substituting these into the equation gives r = 1.765 × 10-4 m, or d = 3.529 × 10-4 m

mCp  T T 
t ln s i 
hA  Ts  T f 


You could work out m = 1.957 × 10-7 kg and A = 3.913 × 10-7 m2 but we already know
that tc = 0.5 s so use that.

 200  25 
t  1  ln   2.58 s
 200  199 

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 158

5.8.2 Distributed Capacitance

Look back at Figure 5.17. If we were cooling the sphere the surface temperature
would be less than the temperature at the centre. This would matter if the thermal
conductivity were poor and the convection good – the lumped parameter method
would not be valid. Analysis is difficult and often requires a finite-element approach.
Luckily (for all of us!) we won’t be doing this.

The direction of heat transfer is not important. Therefore, to keep it simple we will
always refer to heat loss, rather than heat gain, from the surface through the fins.

In order to increase the rate of convective heat transfer from a surface you could
increase the area or use a cooler fluid but these may not be practical or economic.
An alternative way of increasing the area is to use fins. Have a look at the radiators
on the bridge over to the Coulson building. The fins on the radiators will be obvious.
Figure 5.21 shows various kinds of fin.

Figure 5.21, Fin configurations (a) straight fin of uniform cross section, (b) straight
fin of non-uniform cross section, (c) annular fin, (d) pin fin.
(Incropera et al., 2007) – with permission.
f_03_14

To keep the analysis simpler we will only look at fins with a constant CSA such as
the cylindrical fin shown in Figure 5.22.

Figure 5.22, Cylindrical fin (or pin fin) on a planar wall.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 159

For a particular fin design we might reasonably want to calculate:

 The heat transfer rate for given length.

 The heat transfer rate for infinite length
 The temperature distribution along the length.
 The effectiveness of the fin.

Figure 5.23 shows the cylindrical fin from Figure 5.22 but simplified and with added
notation.

Tf

Tb
d
x=0 x=L
x

qconv

qx T qx+dx

dx

Figure 5.23, Nomenclature for fin analysis. The insert shows an expanded element
of the fin.

Since the temperature will vary along the fin you can guess that the solution will
involve taking an element and integrating along the length of the fin.

For a heat balance over the element:

Heat into the element at x = heat out at (x+dx) + heat loss by convection

qx  qx  dx  qconv ................................(5.3)

qx  dx  qx  qconv

There are two areas in the element:


A is the CSA and equals d2
4
Ac is the area for convection and equals d dx

dT dqx d 2T
Now qx  kA from which  kA 2 if A is constant.
dx dx dx
dq
And qx  dx  qx  x dx and qconv  h dAc (T  T f )
dx

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 160

So from Equation (5.3):

d 2T dA
 kA 2
 h c (T  T f ) ................................(5.4)
dx dx

To avoid a lot of writing in the solution, it is useful to let θ = T – Tf.

Since Tf is a constant,

d dT d 2 d 2T
 and  2 .
dx dx dx2 dx

Hence from Equation (5.4):

d 2
 m2  0 ...............................................(5.5)
dx2

hP dA
Where m 2  and P  c (P is the perimeter of the fin).
kA dx

Equation (5.5) is a linear (all the important bits are to the power one) homogeneous
(No term only in x) 2nd order differential equation.

The general solution to Equation (5.5) is:

  C1emx  C2emx

If you differentiate this twice, you should be able to show that this gives Equation
(5.5).

There are two constants of integration and so we need two boundary conditions
(BCs).

BC1: x=0 T = Tb (ie. θ = θb)

BC2: x = L

Heat conducted to the end of the fin = Heat lost from the end by convection

dT
 kA xL  hA(TL  T f )
dx

d
or - k x  L  h L
dx
BC1 gives: θb = C1 +C2

BC2 gives: km (C2emL  C1emL )  h (C2emL  C1emL )

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 161

Solving for C1 and C2 gives:

h
cosh m( L  x)  sinh m( L  x)
 mk
 ........................................(5.6)
b cosh mL 
h
sinh mL
mk

Equation (5.6) allows us to get the temperature distribution along the fin. The total
.
heat loss, Q , from the fin is the heat entering the fin at x = 0.

. dT
Q  kA x 0
dx
This gives:

h
. sinh mL 
cosh mL
Q  hPkA b mk ............................................(5.7)
h
cosh mL  sinh mL
mk

The fin effectiveness is the heat transfer with the fin divided by the heat transfer
without the fin.
Without the fin, the heat transfer is achieved from the equation QW  hAT  T f  and
.

.
Q
effectiveness from   .
.
Qw
Obviously the effectiveness should be large (more than 2, usually) or there is no
point in using fins.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 162

Example : 5.8.2

Problem:

A 50 mm copper rod with diameter 5 mm has one end maintained at 100oC. The
surface of the rod is exposed to air at 25oC with a convection coefficient of
100 W/(m2K). Find the temperature half-way along the rod, the heat transfer from
the rod and the effectiveness of the fin. The thermal conductivity of copper is
400 W/(m.K).

Solution:

P = fin perimeter = πd = 0.016 m

x = 0.025 m (half way along the rod)

hP 100 0.016
m   14.14
kA 400 A

b  100  25  75o C

From Equation (5.6):

h
cosh m( L  x)  sinh m( L  x)
 mk

b cosh mL 
h
sinh mL
mk
100
cosh (0.353)  sinh (0.353)
 14.14  400
  0.8395    62.96oC
75 cosh (0.707)  100
sinh (0.707)
14.14  400

 T    25  87.96oC

From Equation (5.7)

h
. sinh mL 
cosh mL
Q  hPkA b mk  5.16 W
h
cosh mL  sinh mL
mk
 
.
And QW  hA T  T f  0.147 W

.
Q 5.16
Effectiveness :   .
  35.1
0.147
Qw
.................................................

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 163

5.9 Tutorials
1. 2.0 kg/s of hot liquor, heat capacity 4.0 kJ/(kg K), is to be cooled from 80oC to
45oC using 3.0 kg/s of a cold liquid with supply temperature 6oC and a heat
capacity of 3.45 kJ/(kg K).
a) Calculate the heat transfer area requirement for a multi-tube-pass, single
shell pass countercurrent exchanger with an overall heat transfer coefficient
of 1200 W/(m2 K). Use the correction factor, FT.
b) By calculating FT again, comment on the suitability of such an exchanger if
cooling of the hot liquor to 35oC (instead of 45oC) was required.

Ans: a) Tcout = 33.05ºC, FT = 0.907 (I used the formula), 6.0 m2

b) FT = 0.7, not good. (Just for your interest - a two shell pass design gives FT =
0.95)

2. Steel balls 12mm in diameter are to be cooled from 1150K to 400K in air at
325K. Estimate the time required. (You will use the lumped capacitance
model. Check that it is valid by working out the Biot number. See page 155.)
Film heat transfer coefficient = 20 W/(m2K)
Steel thermal conductivity = 40 W/(m K)
Steel density = 7800 kg/m3
Steel heat capacity = 600 J/(kgK)

Ans. 1122s

3. A tank containing 5000 kg of liquid of heat capacity 3.0 kJ/(kg K) currently

takes 25 min to heat from 30oC to 70oC. Dry saturated steam is supplied at 2
bar absolute to a heating coil of outside area 10 m2 and exits at its condensing
temperature. The tank is stirred with an agitator rotating at 1.6 Hz, at which
rate the liquid-side heat transfer coefficient is estimated at 1000 W/(m2 K).
Assuming that this coefficient varies in proportion to (Agitation Reynolds
Number)0.65, determine the heating time that would be achieved if the agitator
speed was increased to 2.5 Hz and the steam pressure to 3 bar absolute.

Hint: Use the data on the existing heating operation to evaluate its overall heat
transfer coefficient (586 W/(m2 K)). Then find the new U at the increased
agitator speed (687 W/(m2 K)) and hence the new heating time.

Ans: 17.8 min

4. A heat exchanger made with stainless steel tubes, 30 mm id, 40 mm od, uses
cooling water inside the tubes to condense steam on the outside.

Steam side film coefficient = 14200 W/(m2 K)

Water side film coefficient = 11400 W/(m2 K)
Steam side dirt coefficient = 8.33 kW/(m2 K)
Water side dirt coefficient = 5.56 kW/(m2 K)
Thermal conductivity of stainless steel =16 W/(m K)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 164

Calculate the overall heat transfer coefficient based on outside area, i) for
clean tubes and ii) including the dirt factors.

Ans: i) 1828 W/(m2 K) ii) 1103 W/(m2 K)

5. A brass rod 100 mm long and 5 mm in diameter extends from a casting at

200oC. The rod is in air at 20oC. If the convection coefficient is 30 W/(m2K)
what is the temperature of the rod at 25 mm, 50 mm, and 100 mm from the
casting? The thermal conductivity of brass = 133 W/(m.K)

Ans. P = 0.016 m, m = 13.433, 156.3oC , 128.0oC , 106.7oC

6. Consider a rectangular fin with length 10 mm, thickness 1 mm and width 2 mm.
The temperature at the base of the fin is 100oC and the fluid temperature is
25oC. The fin is made of an alloyed aluminium with k = 180 W/(m.K). The
convection coefficient = 100 W/(m2K).
Find the temperature at the end of the fin, the heat loss from the fin, and the fin
effectiveness.
Ans. P = 6 x 10-3 m, m = 40.825, 93.8oC, 0.439 W, 29.3

7. A double pipe heat exchanger is a unit that has one tube inside another, with
hot fluid passing through the inner tube, and cold fluid passing between the
annulus of the inner and outer tube.
On an offshore pipeline, a double pipe exchanger is used to quickly heat oil from
20°C to 30°C using hot water taken from the gas turbine system. If the water
velocity is 0.23 m/s, and the oil velocity is 0.5 m/s. Calculate the following:

a) Film heat transfer coefficient of the oil passing through the inner pipe of the
exchanger.
b) Film heat transfer coefficient of the water inside the annulus section.
c) Overall heat transfer coefficient
d) The heat load, using the answer from part c
e) Use Q=mCpΔT to work out the heat load from the change in temperature
for the oil and water side. Compare this answer to the one from part d. What
does this say about the insulation on the exchanger.

Additional data
Inside Pipe – inside diameter 1”, outside diameter 1.2”
Outside Pipe – inside diameter 2”, outside diameter 3”
Pipe length 4.4m
Thermal conductivity of pipe wall 45 W/mK

Fluid Properties Oil Water

Density (kg/m3) 990 1000
Viscosity (Ns/m2) 3x10-3 1x10-3
Specific heat capacity (kJ/kgK) 2.5 4.18
Thermal conductivity (W/mK) 0.25 0.58

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 165

5.10 Bibliography
1. "C&R vol. 1" refers to Chemical Engineering, J.M.Coulson, J.F.Richardson et
al., vol. 1, 6th ed., Butterworth, 1999

2. "C&R vol. 6" refers to Chemical Engineering Design, R.K.Sinnott vol 6, 4th ed.,
Butterworth, 2005

3. Engineering Heat Transfer, W.S.Janna, PWS Engineering, 1986.

536.2JAN

4. Heat Exchanger Design Handbook, volumes 1-5, E.U. Schlunder et al.,

Hemisphere, 1983, 621.402HEA
Everything (?) about heat exchangers.

5. Chemical Process Design and Integration, R. Smith, Wiley 2005,

660.281SMI Quite good section on heat exchangers including how to decide
how many shells to use.

6. Low Reynolds Number Flow Heat Exchangers, S. Kakac et al., Hemishere,

1983, 621.402LOW
Might be useful for plate heat exchangers.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 5. HEAT EXCHANGER DESIGN 166

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 167

Topic 6

Boiling and Condensation

Contents

6.1 Introduction 169

6.2 Condenser Theory & Design 171
6.2.1 Desuperheating and Subcooling 174
6.2.2 Mixtures 174
6.2.3 Pressure Drop in Condensers 176
6.3 Reboiler Theory & Design 181
6.3.1 Types of Reboiler 181
6.3.2 Choice of Reboiler Type 183
6.4 Boiling Heat Transfer 184
6.4.1 Pool Boiling 184
6.4.2 Critical Heat Flux 185
6.4.3 Film Boiling 185
6.4.4 Convective Boiling 185
6.4.5 Convective Heat Transfer Coefficients 186
6.5 Design of Reboiler 188
6.5.1 Forced Circulation Reboilers 188
6.5.2 Thermosyphon Reboilers 188
6.5.3 Kettle Reboilers 191
6.6 Bibliography 194

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 168

Prerequisite knowledge

• Engineering mathematics.
• Engineering chemistry.
• Mass and energy balances for both closed and open systems.
• Detailed understanding of the ideal gas law.
• Thermodynamic principles and definitions.
• Detailed understanding of both the steady-flow energy equation and the
non-flow energy equation.

Learning Objectives

By the end of this topic, you should be able to:

• Identify the 4 configurations that can exist for a condenser.

• Be familiar with design equations for the 4 configurations.
• Understand the principles of desuperheating and subcooling in condensers.
• Design a condenser.
• Know the different designs of reboiler that are available together with their
advantages and disadvantages for different duties.
• Understand pool and film boiling and be able to calculate the boiling regime
in a reboiler.
• Be able to design a reboiler.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 169

6.1 Introduction
Boiling and condensing are special cases of convection in which there is a phase
change. Boiling and condensing give high heat transfer. A tempting explanation is
that in boiling, a small mass of fluid can take up a lot of heat (latent heat) from the
surface as it changes phase and then carry that heat away into the bulk fluid.
Similarly in condensing, a small mass can deposit a lot of heat at the surface as it
changes phase. However tempting this explanation is, it is the turbulence caused
by the bubbles that is the most important factor in giving high heat transfer
coefficients with boiling (C&R vol 16th ed, p490).

• Condensation

Condensation occurs by two mechanisms, filmwise and dropwise:

The filmwise mechanism is the usual one in industrial equipment. Vapour

condenses on a cold surface and the condensate runs off but a film of condensate,
which restricts the heat transfer, always covers the surface.

In dropwise condensation, drops form, coalesce and fall or run off leaving the
surface dry. Heat transfer is better than with filmwise condensation but it is difficult
to arrange for dropwise condensation to be maintained.

Perhaps you can see parallels between nucleate boiling and dropwise
condensation, and film boiling and filmwise condensation.

The presence of a non-condensing gas can severely reduce the rate of

condensation. Consider a mixture of steam and air in contact with a cold surface.
At first steam condenses on the surface so that air is carried towards the surface
and accumulates there. Steam must now find a way past the air to the cold surface
by diffusion or convection. This gives an extra resistance to the condensation of the
steam.

• Boiling

Molecules in a liquid at a liquid-vapour interface are pulled back into the liquid by all
the other molecules in the liquid. This produces the phenomenon of surface
tension. Surface tension is higher in some liquids than others (effects of hydrogen
bonding) but generally the effect of surface tension is greater the smaller the
“object”. If you cool a vapour to its dew point it is difficult to form tiny droplets of
liquid because these have a higher vapour pressure than bulk liquid would have
and so tend to re-evaporate.

Similarly if you cool a solution to its saturation point it is difficult for small crystals to
form because they have a higher solubility than the bulk solid would have so they
tend to re-dissolve. If you heat a liquid to its boiling point and produce a tiny bubble
of vapour, the liquid surrounding the bubble has a lower vapour pressure than you
would expect so the bubble tends to collapse.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 170

All of this mean that you have to superheat a liquid (take it above its bubble point)
in order to get its vapour pressure high enough to allow bubbles to grow.

Think of a heating element (eg a kettle element) immersed in water that is at its
boiling point (100oC). If the element is at 101oC, the water touching it will also be at
101oC. Natural convection will carry the hot water from the element into the bulk
water and evaporation will occur at the water surface. There are no bubbles. If the
temperature of the element is increased, the water next to it will get hot enough for
bubbles to form, break off and float up through the liquid - this is nucleate boiling.
Further increase in temperature gives more vigorous boiling. Eventually there will
be so many bubbles that the element will be covered in vapour. Heat transfer now
will be by radiation and conduction through the vapour. Heat transfer will be
reduced.

This is film boiling. Heat transfer can be increased but only at very high element
temperatures. If the element is heated by steam or hot oil, its temperature is limited
but if it is heated by electricity it may keep heating up until it melts.

Most industrial boiling is in the nucleate boiling region.

Figure 6.1 illustrates these ideas. The x-axis gives the temperature driving force -
the difference in temperature of the heating element and the liquid. The y-axis give
heat flux. Notice the "critical heat flux", the highest heat flux possible in the
nucleate boiling regime, and the unstable transition region between nucleate and
film boiling. You would design for operation somewhat to the left of the critical heat
flux.

Figure 6.1, Boiling regimes

The above mechanisms of boiling are for pool boiling where you have a submerged
heating surface in a pool of liquid. Boiling can also occur as liquid flows through hot
tubes. The bulk liquid may be subcooled while the superheated liquid touching the
tube walls boils. The bubbles collapse as they are cooled by the bulk liquid.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 171

6.2 Condenser Theory and Design

Condensers are shell and tube heat exchangers. Four configurations are possible:

1. Horizontal with condensation in the shell and cooling medium in the tubes.
2. Horizontal with condensation in the tubes.
3. Vertical, with condensation in the shell.
4. Vertical, with condensation in the tubes.

For condensation in the shell – i.e. outside horizontal tubes – the heat transfer
coefficient is determined by the following equation:

(hc )1 = 0.95k L [ρ L (ρ L − ρv )g / µ LΓ]0.33 Eq (6.1)

Where:
(hc)1 = mean condensation film coefficient for a single tube (W/m2K)
kL = condensate thermal conductivity (W/mK)
ρL = condensate density (kg/m3)
ρv = vapour density (kg/m3)
µL = condensate viscosity (Ns/m2)
g = gravitational acceleration (m/s2)
Γ = tube loading – i.e. condensate flow per unit length (kg/ms)

When there is a bank of tubes, the condensate from the upper tubes will add to that
of the lower tubes. If there are Nr tubes in a vertical row and the condensate
undergoes steady laminar flow from row to row, then the mean coefficient is related
to that of the top tube by:

(hc )Nr = (hc )1 Nr −0.17 Eq (6.2)

In practice, though, the condensate does not generally flow smoothly from one tube
to the next. Kern (1950) suggests reducing the index to 0.17. Frank (1978)
suggests the single tube coefficient by 0.75.
We take Nr to be two thirds of the number of tubes in the central row.

For condensation inside or outside vertical tubes, the Nusselt model gives the
following:

(hc )v = 0.926k L [ρ L (ρ L − ρv )g / µ LΓv ]0.33 Eq (6.3)

Where:
(hc)v = mean condensation coefficient (W/m2K)
Γv = vertical tube loading – i.e. condensate rate per unit tube perimeter (kg/ms)

Equation (6.3) applies up to a Reynolds number of 30. Above this value, waves on
the condensate film become important. Reynolds number for a condensate film is
given by:

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 172

4Γv
Re c = Eq (6.4)
µL

Waves increase the heat transfer coefficient. This means that the use of the
equation for Reynolds number over 30 will give a conservative estimate. At
Reynolds numbers in excess of 2000, the condensate film becomes turbulent. The
following diagram illustrates this:

Pr = 10

5
1
 v2  3 3
h l 
g
  2
kl
1

Wave-free Wavy
Laminar Laminar
0.1
10 30 100 1000 1800 10,000
Re

Figure 6.2, Non-dimensionalised heat transfer coefficient for the wave-free laminar
and turbulent flow of condensate on vertical plates.

[ ]
vt = µ L2 / ρ L (ρ L − ρv ) Eq (6.5)

Prandtl Number: Pr = CP µ L / k L

If the inlet stream is saturated vapour and the vapour is totally condensed in the
condenser the Boyko & Kruzhlin (1967) correlation may be used:

(hc ) = h't (1 + (ρ L / ρv ) )2

 
 Eq (6.6)
 

Where:

h't = 0.021(k L / di )Re0.8 Pr 0.43 Eq (6.7)

It is important to avoid flooding in vertical tubes. Hewitt & Hall-Taylor (1970) give a
condition which must be satisfied if flooding is to be avoided:

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 173

[u ρ
0.5 0.25
v v ]
+ u L0.5 ρ L0.25 〈 0.6[g di (ρ L − ρv )]
0.25
Eq (6.8)

Where:
uL and uv are the velocities of the liquid and vapour (m/s)
di = internal diameter of tubes (m)
The critical condition occurs at the bottom of the tubes so the velocities should be
evaluated there.

For condensation inside horizontal tubes, the heat transfer coefficient will depend
on the flow regime at that point. Flow will vary from a single-phase vapour at the
inlet to a single phase liquid at the outlet. In between, the full range of possible
flows will exist – i.e. annular flow, slug flow and bubbly flow. This is simplified by
considering the flow to be in one of two regimes – stratified flow and annular flow.
Figure 6.3 shows the different types of flow that are possible:

Bubbly

Slug

Annular

Stratified

Figure 6.3, Different types of flow.

The Boyko-Kruzhilin equation (earlier) can be used to estimate the coefficient for
annular flow. For stratified flow, we can use:

(hc )s = 0.76kL  ρ L (ρ L − ρv )g µ
0.33

 Eq (6.9)
 Γ
L h

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 174

6.2.1 Desuperheating and Subcooling

If the vapour entering the condenser is superheated to a significant degree and/or

the condensate leaving the condenser is substantially subcooled, the temperature
profile will be as shown in Figure 6.4. If there is a large degree of desuperheating
and/or subcooling we must divide the temperature profile into sections and
estimate the mean temperature difference and heat transfer coefficient separately.

Tsuperheat

Tsat
Temperature

Tsubcool

Desuper- Condensation
Heating Subcooling

LMTDDesuperheating LMTDCondensation LMTDSubcooling

Heat Transfer →

Figure 6.4, Temperature profile of a condenser.

6.2.2 Mixtures

All of the correlations given so far are for single component systems – for example
an essentially pure component condensing from a column. Design for a mixture of
components is more complicated. The condensation will not be isothermal, since
the heavy component will condense first and the dew point of the remaining vapour
will change. There will also be sensible heat transfer between the two phases and
changes in the physical properties along the length of the tubes.
To estimate multicomponent condensation heat transfer coefficients, we must first
develop a temperature-enthalpy diagram for the condensation. This allows us to
determine the true temperature difference (i.e. driving force). There are two limiting
conditions of condensate-vapour flow:

1. Differential Condensation: This is where the liquid separates from the

vapour from which it has condensed.
2. Integral Condensation: This is where the liquid remains in equilibrium with
the uncondensed vapour.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 175

We normally assume that integral condensation takes place. An example of a

temperature-enthalpy diagram for a condenser is given in Figure 6.5:

A
Desuperheating

Tdew B Condensation
Temperature

E
ΔTa ΔTb Subcooling

Tbulb C
D

F
0 QH ΔQ Heat → QT
hin ← Enthalpy hout

Figure 6.5, Temperature-Enthalpy diagram.

Once we have plotted a temperature enthalpy graph, we can obtain the required
area through numerical integration of the graph:

Qt
dQ
A= ∫ U (T
0 v − Tc )
Eq (6.10)

Where:

Qt = total heat transferred (W)

U = overall heat transfer coefficient (W/m2K)

Tv = local vapour temperature (K)

Tc = local cooling medium temperature (K)

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 176

U is determined by assuming a value for the cooling water side coefficient, fouling
factor and wall coefficient and taking the process side coefficient as:

1 / h'cg
1 / h'cg = 1 / h'c + Z / h'g
∆H s  dT 
Z= = xC pg  
 Eq (6.11)
∆H t  dH t 

and
ΔHs/ΔHt = ratio of the change in sensible heat to total enthalpy change
dT/dHt = slope of the temperature enthalpy graph
x = mass fraction of vapour
Cpg = vapour specific heat (J/kgK)

6.2.3 Pressure Drop in Condensers

The pressure drop on the condensing side is difficult to predict since two phases
are present and the vapour mass velocity is changing through the condenser. We
have two alternatives:

1. Calculate as if for single phase flow and apply a correction factor. Frank
(1978) suggests taking the pressure drop as 40% of the value based on
inlet vapour conditions. Kern (1950) suggests a factor of 50%.

2. The method of Gloyer (1970). Calculate a vapour flowrate as a function of

the ratio of vapour flowrate in and out and the temperature profile. A chart
on p.895 of Coulson & Richardson Volume 6 gives average vapour flowrate
as a function of inlet vapour flowrate.

………….…………………………………………

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 177

Example 6.2.1:

Condenser Design

Design a condenser for the following duty:

14 kg/s of propane vapour is to be condensed. The condenser will operate at
10bar. The vapour enters the condenser at its dew point at 27°C. Chilled water is
available at 10°C. We will use tubes of admiralty brass, 20 mm o.d., 16.8 mm i.d.,
and 4.88 m long. The vapours must be completely condensed. No subcooling is
required.

Latent heat of condensation of propane at 10 bar = 356 kJ/kg

Heat transferred from vapour = (14)(356)(1000) = 4984 kW

Assuming an approach of 10°C, chilled water flow = 4984/[(17-10)×4.18]

Chilled water flow = 170.3 kg/s

Assumed overall coefficient = 900 W/m2K (from Coulson & Richardson, Vol 6)

Try a horizontal exchanger with condensation in the shell and 4 tube passes. Using
Figure 12-19, p.840, Coulson & Richardson, Vol 6),

S=
(17 − 10) = 0.41
(27 − 10)
R = 0 and Ft = 1.0 approximately.

∆TLM =
(27 − 17 ) − (27 − 10) = 13.2o C
ln (27 − 17 ) 
 (27 − 10)
Trial area =
(4984)(1000) = 420 m 2
(900)(13.2)
Surface area of one tube = 0.305 m2 (ignoring tube sheet thickness)

Number of tubes = 420/0.305 = 1377 tubes

Use square pitch, Pt = (1.25)(20) = 25 mm

Tube bundle diameter = 20×(1377/0.158)0.44 = 1083 mm

Number of tubes in centre row = 1083/25 = 44 tubes

Shell-Side Coefficient

Estimate tube wall temperature, assuming a condensing coefficient of 1500 W/m2K

Shell-side = 27°C

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 178

Tube-side = (10+17)/2 = 13.5°C

(27-Tw) × 1500 = (27-13.5) × 900 → Tw = 18.9°C

Mean temperature of condensate = (18.9 + 27)/2 = 23°C

At 23°C,
µL = 9.5 × 10-5 Ns/m2,
ρL = 495.7 kg/m3,
kL = 0.094 W/mK,
ρV = 21.7 kg/m3

Wc 14
Γh = = = 2.08 × 10-3 kg/s m
LN t (4.88)(1377 )

Nr = 0.67×44 = 30

hc = 0.95 × 0.094 × (495.7 × (495.7 − 21.7 ) × 9.81)

0.33
 
× 30− 0.17
 9 .(5 × 10 −5
× 2 . 08 × 10 )
−3 


hc = 1536 W/m2K

This is near enough to the assumed value of 1500 W/m2K so no correction to Tw is

needed.

Tube-Side Coefficient
2
 16.8 × 10−3 
Total tube cross-sectional area = 1377 × π ×   = 0.305 m2

 2 
Assume density of water between 10°C and 17°C is 1000 kg/m3

Water temperature = 13.5°C

Tube velocity = (170.3)/ (1000 × 0.305) = 0.56 m/s

Using Eagle & Ferguson (1930) correlation for water

hi = 4200 (1.35+0.02t) ut0.8/di0.2

where: hi = inside coefficient for water (W/m2K)

t = water temperature (°C)
ut = water velocity (m/s)
di = tube internal diameter(mm).

4200 × (1.35 + (0.02 × 13.5)) × (0.56 )

0.8
hi = = 2431 W/m2K
16.80.2

As neither fluid is heavily fouling we use 6000 W/m2K for both sides.
Assume kw = 50 W/mK

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 179

Overall Coefficient
1
=
1
+
1
+
20 × 10−3 ln 20 (
16.8 + 20
) ( 1 )(
+ 20 ) (1 )( )
U 1536 6000 2 × 50 16.8 6000 16.8 2431

U = 650 W/m2K

This is much lower than the original estimate of 900 W/m2K so repeat calculation
using assumed value of 600 W/m2K.

Shell-Side Pressure Drop

Use a pull-through floating head so no need for close clearance.

From Coulson & Richardson Vol 6, Clearance = 95mm

Shell i.d. = 1083 + 95 = 1178mm

Crossflow area, A = [(1178)(25-20)/(25)](1178 × 10-6) = 0.278 m2
Mass flux based on inlet conditions = 14/0.278 = 50.36 kg/sm2

For a square pitch arrangement, the equivalent diameter;

de =
1.27 2
d0
( )
pt − 0.785d 02

In this case, de = (1.27/20) × [(25)2-(0.785)(20)2] = 19.7 mm

Vapour viscosity = 7.6 × 10-6 Ns/m2
Volumetric flowrate = 14/21.7 = 0.645 m3/s
Vapour velocity = 0.645/0.278 = 2.32 m/s.
Re = (21.7)(19.8 × 10-3)(2.32)/(7.6 × 10-6) = 131159

Take pressure drop as 50% of that calculated using inlet conditions.

From Coulson & Richardson Vol 6, assuming a 45% baffle cut,

jf = 2.2 × 10-2

( )
8 j f  s  L  ρ u 2
D
 d e  lb 
∆P = (neglecting viscosity correction)
2

Where:
lb = baffle spacing (m) take 1 shell diameter

Taking 50% of this,

ΔP = (0.5)[(8)(2.2 × 10-2)(1178/19.7)(4.88/1.178)(21.7)(2.32)2]/2
ΔP = 1.27 kN/m2

This is negligible.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 180

Tube-Side Pressure Drop

Viscosity of water = 10-3 Ns/m2

Re = (1000)(16.8 × 10-3)(0.56)/10-3 = 9408

From Coulson & Richardson Vol 6,

jf = 5 × 10-3

Again, neglecting viscosity correction,

(
8 j f  L  ρ u 2
 di 
)
∆P =
2

ΔP = (8)(5 × 10-3)(4.88/16.8 × 10-3)(1000)(0.56)2/2 = 1822 N/m2

This is negligible.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 181

6.3 Reboiler Theory and Design

Reboilers (and condensers) are vital to the operation of distillation columns. The
reboiler evaporates a fraction of the bottom product of the column whilst the
condenser condenses a fraction of the top product. Without reboilers and
condensers any distillation column would operate as a single stage! In this section,
we will look at reboilers.

6.3.1 Types of Reboiler

Thermosyphon Reboiler

Here, no pump is used. The reboiler is either vertical or horizontal. In a vertical

thermosyphon the liquid evaporates in the tubes whilst with a horizontal
thermosyphon vaporisation takes place on the shell side. In all cases the fluid
moves by natural convection caused by the density difference between the liquid at
the base of the column (high density) with the two-phase mixture in the exchanger
(low density).

Figure 6.6 shows a vertical thermosyphon reboiler.

Vapour
+ Liquid

5-10 min Floating

holdup Head
exchanger

Bottoms

Liquid

Figure 6.6, Thermosyphon reboiler.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 182

Forced Circulation Reboiler

In this design, fluid is pumped through the heat exchanger. This is shown in Figure
6.7:

Vapour + Liquid

Vapour
Liquid
level
Reboiler
Liquid

Liquid Bottom
product

Figure 6.7, Forced circulation reboiler.

Kettle Reboiler

In this type of reboiler there is no circulation of liquid through the exchanger.

Evaporation takes place on the surface of a bundle of tubes – like a kettle! The
diagram below illustrates this.

Figure 6.8, Kettle reboiler.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 183

6.3.2 Choice of Reboiler Type

Three factors determine the optimum type of reboiler for a given duty:

1. The nature of the process fluid – its viscosity and propensity to fouling.

2. Operating pressure.

3. Equipment layout – available space, etc.

Forced Circulation Reboilers:

Viscous and heavily-fouling fluids, low vacuum operations, low vaporisation rates.
Disadvantages: Running cost of pump, reliability of pump, leakage from pump
seals.

Thermosyphon Reboilers:

Most economical but not suitable for viscous fluids or high-vacuum operations.
Disadvantages: Column must be elevated to provide the hydrostatic head
necessary for the thermosyphon effect.

Kettle Reboilers:

Suitable for vacuum operation and high evaporation rates. Can be built in to the
column. Disadvantages: - low heat transfer coefficient due to lack of circulation.
Not suitable for fouling materials. Generally high capital cost.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 184

6.4 Boiling Heat Transfer

In order to be able to design reboilers we must understand the basics of boiling
heat transfer. Heat transfer from a submerged surface to a pool of liquid depends
on the temperature difference between the surface and the liquid. When the first
bubbles of vapour appear, they agitate the surface and cause the heat transfer
coefficient to rise rapidly. As the heat input increases, this reaches a peak and then
the heat transfer coefficient falls rapidly as dry patches start to appear on the
surface. Up to the peak is known as pool (or nucleate) boiling and after the peak
we have film boiling. The critical excess temperature at which the transition from
pool boiling to film boiling occurs is approximately 75°C above the bulk temperature
of the liquid. Figure 6.9 illustrates the different boiling regimes.

Figure 6.9, Boiling Heat Transfer Coefficients.

6.4.1 Pool Boiling

Mostinski(1963) gives an empirical correlation:

 0.17 1.2 10

hnb = 0.104 (Pc )
0.69
(q )0.7 1.8 P P  + 4 P  + 10 P   (Eq 6.12)
  c   Pc   c 
P
where:

hnb = pool (or nucleate) boiling coefficient (W/m2K)

P = operating pressure (bar)
Pc = liquid critical pressure (bar)
q = heat flux (W/m2) = hnb = (Tw − Ts )

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 185

6.4.2 Critical Heat Flux

In our calculations, it is important that we check that no calculated heat flux

exceeds the critical heat flux – otherwise the calculated result would be
meaningless. Mostinski gives an empirical correlation for the maximum critical heat
flux:
0.35 0.9
  P 
qc = 0.000367 Pc  P  1 −  Pc  (Eq 6.13)
 Pc 

6.4.3 Film Boiling

For film boiling on tubes we can use another empirical relationship given by
Bromley(1950):

[
h fb = 0.62 k v3 (ρ L − ρ v )ρ v gλ / µ v d 0 (Tw − Ts ) ]
0.25
(Eq 6.14)
where:

hfb = film boiling heat transfer coefficient (W/m2K)

λ = latent heat of vaporisation (J/kg)
kv = vapour thermal conductivity (W/mK)
µv = vapour viscosity (Ns/m2)
g = gravitational acceleration (m/s2)
ρ = density (kg/m3)
d0 = outside diameter of tubes (m)

It should be emphasised here that reboilers are always designed to operate in the
pool, or nucleate, boiling regime. Film boiling is to be avoided.

6.4.4 Convective Boiling

Usually, the situation is more complicated than a simple pool boiling throughout the
heat exchanger. If the liquid is boiling in tubes then at the start of the tubes it will
not be boiling at all. Then as we move along the tube, small bubbles start to
appear. These get larger as we move further along until the end of the tubes
where we could be approaching the film boiling regime. The following diagram
illustrates this;

Subcooled Bubbly Plug Annular Mist Superheated

Liquid Flow Flow Flow Flow Vapour

Figure 6.10, Convective boiling.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 186

Four regimes are defined in this scenario:

1. Single-phase flow region – this corresponds to the “subcooled liquid”

region above.
2. Sub-cooled boiling – this is the “bubbly flow” region.
3. Saturated boiling – this is “plug flow” and “annular flow”.
4. Dry-wall region – this – as the name suggests – is the area to the right
where the wall is dry.

Naturally, such a chaotic system is impossible to model rigorously. Chen (1966)

developed an empirical method for the determination of convective boiling heat
transfer coefficients.

6.4.5 Convective Heat Transfer Coefficients

Chen’s method evaluates the convective heat transfer coefficicents. In forced

convective boiling, the total heat transfer coefficient, hcb, can be thought of as
consisting of two components – a convective component and a nucleate boiling
component:
hcb = h' fc + h'nb (Eq 6.15)

The convective coefficient, h’fc, can be calculated by firstly calculating the single
phase coefficient and then modifying it to account for the effects of two-phase flow
using a factor fc:
h' fc = h fc . f c (Eq 6.16)

The single-phase coefficient, hfc, is calculated assuming that the liquid phase is
flowing alone in the conduit. The two-phase correction factor, fc, is obtained from
Figure 6.11.

102 fc 1.0 Sc

0.8
101
0.6

0.4
1 1
0.2
1/X
0
10-1 1 10 102 104 105 ReL fc1.25 108
(a) (b)

Figure 6.11, Two-phase correction factor fc.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 187

(a) using the Lockhart-Martinelli two phase flow parameter, X, which is given by:

ρ L  µv 
0.9 0 .5 0.1
1 =x 
X  (1 − x )
(Eq 6.17)
 ρ v   µ L 

Where x is the mass fraction of vapour.

Note that capital X and small x are two different variables.

N.B. For this equation to work there must be turbulent flow in both phases.

The nucleate boiling coefficient can be similarly calculated using the pool boiling
coefficient cited earlier modified by a factor fs to account for the fact that in a
turbulent liquid it is more difficult to obtain stable boiling nuclei:

h' nb = hnb . f s (Eq 6.18)

The suppression factor, fs, is obtained from the second chart (b) below. ReL is
calculated assuming that liquid only is flowing in the conduit. It is given by:

Re L = (1 − x ) Gd e / µ L (Eq 6.19)

Where G is total mass flowrate per unit flow area and de is effective hydraulic
diameter.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 188

6.5 Design of Reboilers

6.5.1 Forced Circulation Reboilers

This is the easiest type to design. This is because they are usually designed with
no vapour in the tubes. A throttling valve is placed after the tubes and vaporisation
takes place here. This means that the design required is for a liquid-phase heat
exchanger and a pump. You already know how to design these! If a significant
amount of boiling occurs in the tubes, then Chen’s method is employed. We usually
use conventional shell and tube design for the heat exchanger:

• Process fluid on shell side – one shell pass and two tube passes.
• Process fluid on tube side – one shell and one tube pass.
• We usually have high tube velocities (3 – 9 m/s) to reduce fouling.

6.5.2 Thermosyphon Reboilers

The design of thermosyphon reboilers is more complicated. This is mainly because

we cannot accurately determine the fluid circulation rate. This means that we must
use an iterative process to determine that heat transfer coefficient.
First, we model the system as a u-tube. One leg is the heat exchanger while the
other leg is the column base. The driving force for mass transfer is the density
difference between the liquid leg in the column base and the lighter, two-phase
mixture in the heat exchanger leg.

The iterative process is as follows:

1. Calculate the vaporisation rate required.

2. Assume a value for overall heat transfer coefficient.
3. Calculate the required heat transfer area using this coefficient using
maximum heat flux of 39700 W/m2 (vertical) or 47300 W/m2 (horizontal).
4. Assume a value for circulation rate.
5. Decide the exchanger layout and pipe dimensions.
6. Calculate the pressure drop in the inlet piping assuming single phase.
7. Assume that density decreases linearly along tube.
8. Calculate the pressure drop at the outlet of the tubes. This will be mostly or
all vapour.
9. Compare the total ΔP with available differential head. If they are compatible,
move on to 10. If not, return to step 4 and assume a different value for
circulation rate.
10. Calculate the heat transfer coefficient section by section up the tubes. Use
Chen’s method for the sections in which boiling is occurring.
11. Calculate the overall heat transfer coefficient from the sectional heat
transfer coefficients. Use this figure to calculate the rate of vaporisation.
If this is sufficiently close to that calculated in 1, proceed to 12. If not, return
to step 2 and assume a different heat transfer coefficient.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 189

12. Check that the critical heat flux is not exceeded anywhere in the tube using
Mosinski’s equation.
13. Repeat the procedure as necessary to optimise the design.

Clearly, this type of laborious calculation is best carried out on a computer.

…………..……………………………….

Example 6.5.1

Thermosyphon Reboiler Design

A C4 splitter column requires a reboiler. The required vapour rate is 2000 kg/hr. The
heat transfer coefficient is 1500 W/m2K. Bottoms composition are all C4. Operating
pressure is 8 bar. Bubble point of mixture is approximately 90°C. Assume that the
feed is all liquid at its bubble point.

Design a vertical thermosyphon reboiler:

1. Vaporisation rate is 2000 kg/hr (given).

2. Assume heat transfer coefficient = 1500 W/m2K

3. Latent heat of vaporisation of butane is 386 kJ/kg so heat load is

Heat Load = (2000) × (386) = 772000 kJ/hr = 214.4 kW
Area required = 214400/39700 = 5.4m2

4. Assume circulation rate is 100 kg/hr.

5. Use 25 mm i.d, 2.5 m long tubes. This is popular for vertical thermosyphon
reboilers.
Area of one tube is 0.196 m2.
Therefore number of tubes required = 5.4/0.196 = 28 tubes.

6. Circulation rate is 100 kg/hr. Density of liquid is approximately 600 kg/m3

Plus condensed vapour = 2100 kg/hr.
So circulation rate = 2100/600 = 3.5 m3/hr.
Cross-sectional area of tube = 0.00049 m2
Total cross-sectional area of bundle = (28)×(0.00049) = 0.01372 m2.
Velocity through tubes = 3.5/(3600)×(0.01372) = 0.07m/s.
Pressure drop will be negligible.

7. At tube exit, mass of vapour will be 2000 kg/hr = 0.56 kg/s.

Density of vapour at 8 bar at its bubble point = (2.78)×(8)×0.4 = 6.4 kg/m3

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 190

Assume that this is the density at the top of the tubes – i.e. liquid
contribution is negligible.

8. Velocity through tubes = (0.56/6.4)/0.01372 = 6.38 m/s.

Viscosity of vapour = 8.1 x 10-6 Ns/m2
Re = ρdū/µ = (6.4)(0.025)(6.38)/8.1 x 10-6 = 126025.
Assume relative roughness of tubes = 0.001.
From friction factor chart, Φ = 0.0055

ΔPf = 4Φ(L/d)(ρu2) = (4)(0.0055)(2.5/0.025)(6.4)(6.38)2 = 573 N/m2 or 573/2.5 =

229N/m2 per m.

Assuming that the pressure drop increases linearly from 0 N/m2 per m at the
inlet to 229 N/m2 per m at the outlet, mean pressure drop = 229/2 = 114.5 N/m2.
Frictional pressure drop = (114.5)(2.5) = 286N/m2.
Hydrostatic pressure in tubes assuming linear density gradient:

ΔPh = gL/(v0 – vi) × ln(v0/vi)

Vi = 1/600 = 0.00167 m3/kg; v0 = 1/6.4 = 0.156 m3/kg
ΔPh = (9.81)(2.5)/(0.156 – 0.00167) × ln(0.156/0.00167) = 721 N/m2
Total ΔP = 721 + 286 = 1007 N/m2

9. Total available head (driving force) = ρLgL = (600)(9.81)(2.5) = 14715 N/m2

This is easily enough to drive the circulation.

10. Using Mostinski’s equation to calculate nucleate boiling coefficient:

 0.17 1.2 10

hnb = 0.104 (Pc )
0.69
(q )0.7 1.8 P P  + 4 P  + 10 P  
  c   Pc   Pc  

= 0.104 (38) (37900)0.7 1.8(8 38) ( ) ( )  = 4095 W/m 2 K

0.17 1.2 10
+4 8 + 10 8
0.69

 38 38 

Assuming that vapour fraction varies linearly from 0 at the inlet to 1 at the
exit of the tubes, mean x = 0.5.
Applying Chen’s method:
ρ L  µv 
0.9 0.5 0.1
1 =x 
X  (1 − x )  ρ v   µ L 

Liquid viscosity = 2.82 x 10-3 Ns/m2

[600 6.4]
0.9 0.1
= 0.5  8.1×10 − 6 
0.5
1
X  (1 − 0.5)  2.82 ×10 −3 
1 = 0.028 Enhancement factor will be negligible.
X

ReL = (600)(0.025)(0.0044)/(2.82 × 103) = 23.4 Laminar. Lockhart-Martinelli

inapplicable.
Very low. Take suppression factor to be 0.9 from Chen’s chart.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 191

Calculating, h’nb = (4095)(0.9) = 3686 W/m2K.

Calculation of hfc.
Single phase Nu = 0.023 Re0.8 Pr0.33 (neglect wall viscosity difference).

h (0.025/0.12) = 0.023(23.4)0.8[(1675)(2.82 x 10-3)/0.12)]0.33

hfc = 4.615 W/m2K
Mean heat transfer coefficient = (3686 +4.615)/2 = 1845 W/m2K.

11. Overall heat transfer coefficient, neglecting wall resistance and assuming
steam side coefficient is 8000 W/m2K,

1/U = 1/8000 + 1/1845 = 6.67 × 10-4

U = 1500 W/m2K. Fits perfectly!

12. We used the maximum allowable heat flux according to Kern (1950).

……………………………………………

6.5.3 Kettle Reboilers

The kettle reboiler operates entirely in the pool boiling regime. Bubbles rise from
the lower tubes and pass the upper tubes. This has two opposing effects:

1. The vapour tends to blanket the upper tubes.

2. The movement of the bubbles agitates the water adjacent to the upper
tubes.

This means that the maximum heat flux for stable nucleate boiling will be less fur a
tube bundle than for a single tube. The following modified Zuber equation (by Palen
& Small (1964)) takes account of this:

 d 0

qcb = K b  t  λ
p
N

 [
 σ g (ρ L − ρ v )ρ v 2 ]
0.25
Eq (6.20)
 t 

where:
qcb = maximum (critical) heat flux for the tube bundle (W/m2)
Kb = 0.44 for square pitch = 0.41 for equilateral triangular pitch
pt = tube pitch
d0 = tube outside diameter
Nt = total number of tubes.

For U-tubes, Nt will be twice the number of actual tubes.

Palen & Small(1964) suggest applying a safety factor of 0.7 to the maximum flux
obtained from the modified Zuber equation. This still gives values well above the
traditionally-used standard values – e.g. 37900 W/m2 (Kern (1950)).
A tube pitch of between 1.5 and 2 times the tube outside diameter should be used
to avoid vapour blanketing.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 192

Depending on the heat flux, the required shell diameter is as follows:

Heat Flux (W/m2) Shell Φ/ Bundle Φ

Less than25000 1.2 – 1.5
25000 – 40000 1.4 – 1.8
Greater than 40000 1.7 – 2.0

The freeboard between the liquid level and shell should be at least 0.25m.
To avoid excessive entrainment, the maximum liquid velocity, ȗv, at the liquid
surface should be less than:

u v 〈 0.2[(ρ L − ρ v ) / ρ v ]
 0.5
Eq (6.21)

The equations for nucleate boiling coefficients can be used for close boiling
mixtures – i.e. < 5°C but will overestimate the coefficient if used for mixtures with a
wide boiling range. Palen & Small (1964) give the following correlation for mixtures:

(hnb )mixture = f m (hnb )single component Eq (6.22)

where:
f m = e [−0.0083(Tbo −Tbi )]
Tbo = temperature of the vapour mixture leaving the reboiler (°C)
Tbi = temperature of the liquid entering the reboiler (°C)

………………………………………………

Example 6.5.2:

Kettle Reboiler Design

Design a kettle reboiler for the duty used for the thermosyphon reboiler. Steam is
available at 1.7 bar.

Heat load = 214.4 kW (from previous example)

Add 5% for losses = (214.4) × (1.05) = 225 kW

Using Figure 12.1 in Coulson & Richardson Volume 6 (Sinnott & Towler (2009)) we
get the following:

Assumed U value = 1000 W/m2K

Boiling point of butane = 90°C
Steam saturation temperature = 115°C

Assume both sides are isothermal:

ΔTm = 115 - 90 = 25°C
Tube area required = 225 × 103/(1000)(25) = 9 m2

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 193

Select 30 mm o.d. U-tubes.

Nominal length = 4.8 m
i.d. = 25 mm
Number of u-tubes = 9/ (30 × 10-3)(π)(4.8) = 20 tubes

Using Mostinski’s equation;

Heat flux based on estimated area (q) = 225/9 = 25 kW/m2
hnb = 0.104 (38) (25000 ) 1.8 8
0.69 0.7

( )
38
0.17
( )
+4 8
38
1.2
( )
+ 10 8
10

38 
 = 3051 W/m 2 K

Taking steam coefficient as 8000 W/m2K, fouling coefficient 5000 W/m2K, butane
fouling coefficient 10000 W/m2K (i.e. clean) and the tube wall is carbon steel, kw =
55 W/mK
1 =1 +1 +
[( ) (
30 × 10− 3 ln 30 )] ( )( )
(2 × 55) + 25 5000 + 8000 ⇒ U 0 = 1152 W/m K
25 30 1 1 2
U0 3051 10000

This is slightly greater than the original estimate of 1000W/m2K so the design can
stand.

Layout

Draw a tube diagram. This indicates that the bundle diameter is approximately
350mm.
Taking shell diameter as twice the bundle diameter, i.e. 700 mm.
If liquid level is 450 mm from base, freeboard = 700 – 450 = 250 mm. Adequate.
Width at 450 mm is approximately 0.6 m.
Surface area of liquid = (2.4)(0.6) = 1.44 m2
Vapour flowrate = (2000/3600)/(6.4) = 0.087 m3/s (from thermosyphon example)
Velocity = 0.087/1.44 = 0.06 m/s
Maximum allowable velocity = 0.2[(600 – 6.4)/(6.4)]0.5 = 1.93 m/s

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2

TOPIC 6. BOILING AND CONDENSATION 194

6.6 Bibliography
Boyko, L.D; Kruzhilin,G.N.(1967) Int J. Heat Mass Transfer 10,361. Heat transfer
and hydraulic resistance during condensation of steam in a horizontal tube and in a
bundle of tubes.

Bromley, L.A., (1950) Chem. Eng. Prog. 46, 221. Heat transfer in stable film boiling.

Chen, J.C., (1966) Ind. Eng. Chem. Proc. Des. Dev. 5, 322. A correlation for boiling
heat transfer to saturated liquids in convective flow.

Eagle, A; Ferguson, R.M. (1930) Proc. Roy. Soc. A. 127, 540. On the coefficient of
heat transfer from the internal surfaces of tube walls.

Frank, O. (1978) Simplified design procedure for tubular exchangers.in Practical

Aspects of Heat Transfer. Chem. Eng. Prog. Tech. Manual. (Am. Inst. Chem. Eng.)

Gloyer, W. (1970) Hydro. Proc. 49 (July) 107. Thermal design of mixed vapour
condensers.

Hewitt, G.F; Hall-Taylor,N.S. (1970) Annular two-phase flow. Pergamon press.

Kern, D.Q., (1950) Process Heat Transfer McGraw-Hill.

Mostinski, I.L. (1963) Teploenergetika 4, 66.English abstract in Brit Chem Eng 8,

580 (1963). Calculation of boiling heat transfer coefficients, based on the law of
corresponding states.

Palen, J.W; Small, W.M (1964) Hyd. Proc. 43 (Nov) 199. A new way to design
kettle reboilers.

Sinnott, R; Towler, G. Chemical Engineering Design. Coulson & Richardson Vol 6. I

Chem E publications 2009, pp878-900.

Sinnott, R; Towler, G. Chemical Engineering Design. Coulson & Richardson Vol 6. I

Chem E publications 2009, pp900-925.

©HERIOT-W ATT UNIVERSITY B48BD October 2015 v2