Series in Thermal & Fluid

Physics & Engineering
Editor: G.F. Hewitt
Practical Thermal
Design of
Shell-and-Tube
Heat Exchangers
R. Mukherjee

3UDFWLFDO7KHUPDO'HVLJQ

RI

6KHOODQG7XEH+HDW([FKDQJHUV

by

R. Mukherjee
Heat Transfer Consultant
New Delhi, Inaia
ii
Practical Thermal Design of Shell-and-Tube Heat Exchangers
R. Mukherjee
rajiv.mukherjee¸vsnl.com
Library of Congress Cataloging-in-Publication Data in Process
Catalog record is available Irom the Library oI Congress.
This book represents inIormation obtained Irom authentic and highly regarded sources. Reprinted
material is quoted with permission, and sources are indicated. A wide variety oI reIerences are
listed. Every reasonable eIIort has been made to give reliable data and inIormation, but the author
and the publisher cannot assume responsibility Ior the validity oI all materials Ior the con-
sequences oI their use.
All rights reserved. This book, or any parts thereoI, may not be reproduced in any Iorm without
written consent Irom the publisher.
Direct inquires to Begell House, Inc., 145 Madison Avenue, New York, NY 10016.
©2004 by Begell House, Inc.
ISBN: 1-56700-205-6
Printed in the United States oI America 1 2 3 4 5 6 7 8 9 0
iii
Dedication
To the memory of my parents, who taught me to believe in myself
To my wife, Kalpana, for her unflagging patience ana support
To my aaughter, Shilpi, ana my son (-in-law), Bappa, for their faith ana conviction
Finally, to the reaaer, who maae the entire effort worthwhile
iv
Acknowledgments
I am indebted to Almighty God Ior having given me the education, intelligence, opportu-
nity, strength, and Iervor to write this book.
I am also indebted to all those Irom whom I learned heat exchanger design over the
years. I regret it is not possible to recount their names individually Ior the simple reason that
it would be too long a list!
Without the encouragement and support oI Cynthia Mascone, presently Technical
Editor at Chemical Engineering Progress, and her peer Gail Nalven, I would never have
started writing the book!
This book might not have been possible without the wonderIul exposition oI heat
exchanger technology by Heat TransIer Research, Inc. (HTRI). My long experience in the
Iield oI heat exchangers has been very largely honed on the platIorm oI HTRI whose
soItware I have been using since 1974.
Nobody can write a book on shell-and-tube heat exchangers without sourcing Irom the
Standards oI Tubular Exchanger ManuIacturers Association (TEMA), and I am no
exception.
I am grateIul to HTRI, Begell House, Cal Gavin, AIChE, TEMA, ABB Lummus Heat
TransIer, Brown Fintube, UOP, AlIa Laval and Heatric Ior some oI the diagrams and
photographs used in this book that have been duly acknowledged where they appear.
I will always be indebted to Bill Begell at Begell House Ior deciding to publish this
book.
How can I Iorget my good Iriend Graham Polley? It was he who led me to Bill in the
Iirst place.
I am grateIul to Janet Rogers at Begell House Ior managing my book, Donna Thompson
at MessagePros Ior undertaking the arduous task oI copyediting my book and doing it with
aplomb, and all those at Begell House who were responsible Ior the production oI this book.
What we are able to accomplish in our lives, whether proIessionally or otherwise, is the
result oI the Lord`s grace and the encouragement and support we receive Irom myriad
sources. This book is thereIore truly a collaborative eIIort, and the credit belongs to the
human Iraternity at large rather than to any individual.
v
About the Author
R. Mukherjee is a consultant in unIired heat transIer based in New Delhi, India. He has
over 33 years oI experience in the thermal design, revamping and troubleshooting oI air-
cooled and shell-and-tube heat exchangers, and considerable experience in the design oI
heat exchanger networks. He has written several articles and presented many papers at
technical symposia. Mukherjee has also served as Iaculty Ior numerous courses on heat
exchanger design and operation, energy conservation, and heat exchanger networks, and
presently teaches an intensive two-day in-house reIresher course in the design and opera-
tion oI heat exchangers that can be oIIered at any plant or oIIice location around the
world. He is an honors graduate in chemical engineering Irom Jadavpur University, Cal-
cutta, India.
In his spare time, Mukherjee enjoys reading (Kahlil Gibran is a big Iavorite), writing,
listening to music and collecting quotations. He lives in New Delhi with his wiIe, Kalpana.
Their daughter, Shilpi, and her husband, Bappa, live in Illinois, with their baby son, Sohum.












R. Mukherjee
vi
TABLE OF CONTENTS

PreIace........................................................................................................................................ x
Chapter 1: Introauction............................................................................................................... 1
Chapter 2: Classification of Shell-ana-Tube Heat Exchangers..................................................... 5
2.1 Components oI Shell-and-Tube Heat Exchangers ........................................................... 5
2.2 Front and Rear Heads .................................................................................................. 10
2.3 ClassiIication by Construction...................................................................................... 13
2.3.1 Fixed-tubesheet heat exchanger ............................................................................. 13
2.3.2 U-tube heat exchanger........................................................................................... 13
2.3.3 Floating-head heat exchanger ................................................................................ 14
2.4 ClassiIication by Service.............................................................................................. 16
Chapter 3: Thermal Design ana Optimi:ation of Single-Phase Heat Exchangers ....................... 19
3.1 Broad Objectives oI Thermal Design............................................................................ 19
3.2 Data to be Furnished Ior Thermal Design..................................................................... 20
3.3 Tubeside...................................................................................................................... 23
3.3.1 EIIects oI tubeside velocity ................................................................................... 24
3.3.2 Heat transIer coeIIicient ........................................................................................ 24
3.3.3 Pressure drop ........................................................................................................ 26
CASE STUDY 3.1: OPTIMIZING TUBESIDE DESIGN............................................................. 29
3.3.4 Importance oI stepwise calculations Ior viscous liquids.......................................... 30
CASE STUDY 3.2: STEPWISE CALCULATIONS.................................................................... 31
3.4 Shellside...................................................................................................................... 31
3.4.1 Shell type.............................................................................................................. 32
3.4.2 Tube layout pattern ............................................................................................... 34
3.4.3 Tube pitch............................................................................................................. 35
3.4.4 BaIIling ................................................................................................................ 35
3.4.5 Stream analysis ..................................................................................................... 39
CASE STUDY 3.3: VARIATION OF TEMPERATURE PROFILE DISTORTION FACTOR
WITH BAFFLE SPACING........................................................................................... 41
CASE STUDY 3.4: OPTIMIZING BAFFLE DESIGN ................................................................ 43
3.4.6 Reduction oI shellside pressure drop...................................................................... 46
CASE STUDY 3.5: USE OF DOUBLE-SEGMENTAL BAFFLES ................................................. 47
Chapter 4: Mean Temperature Difference.................................................................................. 51
4.1 Logarithmic Mean Temperature DiIIerence (LMTD).................................................... 51
4.2 Countercurrent Flow.................................................................................................... 51
4.3 Co-Current Flow.......................................................................................................... 52
4.4 Countercurrent and Co-Current Flow: The F
t
Factor ..................................................... 53
4.5 Temperature Cross....................................................................................................... 56
vii
4.6 Heat Release ProIiles and Zone-Wise Calculations ....................................................... 57
4.7 Temperature ProIile Distortion..................................................................................... 59
CASE STUDY 4.1: HOW A TEMPERATURE PROFILE DISTORTION PROBLEM
IS BETTER HANDLED BY TWO SHELLS IN SERIES ....................................................... 63
Chapter 5: Allocation of Siaes. Shellsiae ana Tubesiae............................................................... 67
5.1 Introduction ................................................................................................................. 67
5.2 Parameters Ior Allocation oI Sides................................................................................ 67
5.2.1 Viscosity............................................................................................................... 68
5.2.2 Corrosiveness........................................................................................................ 68
5.2.3 Fouling tendency................................................................................................... 68
5.2.4 Pressure................................................................................................................. 69
5.2.5 Flow rate............................................................................................................... 69
5.2.6 Temperature range................................................................................................. 71
CASE STUDY: 5.1 ALLOCATION OF FLUID SIDES ............................................................... 71
Chapter 6: Methoaology of the Use of Multiple Shells ................................................................ 75
6.1 Multiple Shells in Parallel ............................................................................................ 75
6.2 Multiple Shells in Series............................................................................................... 76
6.2.1 For temperature cross conditions............................................................................ 76
6.2.2 For better utilization oI allowable pressure drop..................................................... 76
6.2.3 For improving the temperature proIile distortion correction Iactor .......................... 77
CASE STUDY 6.1: USE OF MULTIPLE SHELLS IN SERIES...................................................... 77
6.3 Multiple Shells in Series/Parallel .................................................................................. 78
CASE STUDY 6.2: USE OF MULTIPLE SHELLS IN SERIES/PARALLEL..................................... 78
Chapter 7: Thermal Design ana Optimi:ation of Conaensers..................................................... 83
7.1 Introduction ................................................................................................................. 83
7.2 ClassiIication............................................................................................................... 83
7.2.1 According to construction...................................................................................... 83
7.2.2 According to layout ............................................................................................... 83
7.2.3 According to service.............................................................................................. 84
7.2.4 According to coolant ............................................................................................. 84
7.2.5 According to condensing range.............................................................................. 85
7.2.6 According to operating pressure............................................................................. 85
7.3 Mechanisms oI Condensing.......................................................................................... 86
7.3.1 Vertical in-tube condensation................................................................................. 86
7.3.2 Horizontal in-tube condensation............................................................................. 88
7.3.3 Condensation outside tubes.................................................................................... 88
7.3.4 Condensation oI mixed vapors and mixtures oI vapors and noncondensables.......... 89
7.4 Practical Guidelines Ior Thermal Design....................................................................... 90
7.4.1 BaIIling................................................................................................................. 90
CASE STUDY 7.1: ISOTHERMAL CONDENSATION WITH SINGLE-PASS SHELL
AND SINGLE-SEGMENTAL BAFFLES ......................................................................... 91
CASE STUDY 7.2 CONDENSATION WITH SINGLE-PASS SHELL AND
DOUBLE-SEGMENTAL BAFFLES ............................................................................... 93
CASE STUDY 7.3 CONDENSATION WITH DIVIDED-FLOW SHELL.......................................... 96
7.4.2 Multiple shells in series or parallel......................................................................... 98
CASE STUDY 7.4: CONDENSATION WITH MULTIPLE SHELLS IN SERIES................................ 99
CASE STUDY 7.5: CONDENSATION WITH MULTIPLE SHELLS IN SERIES/PARALLEL............. 101
7.4.3 Condensation with desuperheating and/or subcooling........................................... 104
CASE STUDY 7.6: CONDENSATION WITH WET-WALL DESUPERHEATING........................... 105
CASE STUDY 7.7: CONDENSATION WITH DRY-WALL DESUPERHEATING ........................... 106
CASE STUDY 7.8: CONDENSATION WITH INTEGRAL SUBCOOLING.................................... 109
7.4.4 Nozzle sizing....................................................................................................... 111
7.4.5 Condensing proIiles and MTD............................................................................. 112
viii
7.4.6 Low-pressure condenser design........................................................................... 113
CASE STUDY 7.9: TUBESIDE CONDENSATION................................................................. 114
7.5 Special Applications .................................................................................................. 116
7.5.1 Use oI low-Iin tubes............................................................................................ 116
CASE STUDY 7.10: USE OF LOW-FIN TUBES.................................................................... 117
7.5.2 Vacuum condenser design: Ejector condensers and surIace condensers ................ 119
CASE STUDY 7.11 EJECTOR INTERCONDENSER .............................................................. 119
Chapter 8: Thermal Design ana Optimi:ation of Reboilers...................................................... 123
8.1 Pool Boiling .............................................................................................................. 123
8.2 Parameters AIIecting Pool Boiling............................................................................. 124
8.2.1 SurIace eIIects .................................................................................................... 124
8.2.2 Mixture eIIects.................................................................................................... 126
8.2.3 Pressure eIIects ................................................................................................... 126
8.2.4 Tube bundle geometry eIIects.............................................................................. 126
8.3 Maximum Heat Flux.................................................................................................. 126
8.4 Flow Boiling ............................................................................................................. 127
8.5 Distillation Column Reboilers.................................................................................... 127
8.5.1 Internal reboilers................................................................................................. 128
CASE STUDY 8.1: LIGHT HYDROCARBON REBOILER (INTERNAL REBOILER) .................... 129
8.5.2 Kettle reboilers.................................................................................................... 130
CASE STUDY 8.2: STRIPPER REBOILER (KETTLE REBOILER) ............................................ 132
8.5.3 Horizontal thermosyphon reboilers..................................................................... 134
CASE STUDY 8.3: DISTILLATION COLUMN REBOILER (HORIZONTAL THERMOSYPHON) .... 137
8.5.4 Vertical thermosyphon reboilers.......................................................................... 138
CASE STUDY 8.4: DISTILLATION COLUMN REBOILER (VERTICAL THERMOSYPHON) ......... 145
CASE STUDY 8.5: DISTILLATION COLUMN REBOILER
(VERTICAL THERMOSYPHON/KETTLE)................................................................... 148
8.5.5 Forced-Ilow reboilers .......................................................................................... 150
CASE STUDY 8.6: DISTILLATION COLUMN REBOILER
(VERTICAL THERMOSYPHON/FORCED-FLOW) ........................................................ 151
8.6 Selection oI Reboilers................................................................................................ 153
8.7 Start-Up oI Reboilers................................................................................................. 154
Chapter 9: Physical Properties ana Heat Release Profiles........................................................ 157
9.1 Physical Properties .................................................................................................... 157
9.2 Physical Property ProIiles .......................................................................................... 160
9.3 Heat Release ProIiles ................................................................................................. 161
9.4 How to Feed Heat Release ProIiles ............................................................................ 161
Chapter 10: Overaesign........................................................................................................... 163
10.1 Mechanics oI Overdesign......................................................................................... 163
10.2 Overdesign in Single-Phase Heat Exchangers........................................................... 164
CASE STUDY 10.1: EFFECT OF OVERDESIGN ² HIGH-TEMPERATURE APPROACH CASE...... 165
CASE STUDY 10.2: EFFECT OF OVERDESIGN ² LOW-TEMPERATURE APPROACH CASE....... 166
10.3 Overdesign in Reboilers........................................................................................... 167
10.4 Overdesign in Condensers........................................................................................ 168
10.5 Overdesign Factor.................................................................................................... 169
10.6 Tube Plugging ......................................................................................................... 170
Chapter 11: Fouling. Its Consequences ana Mitigation........................................................... 171
11.1 Categories oI Fouling............................................................................................... 173
11.2 Progress oI Fouling.................................................................................................. 173
11.3 Parameters That AIIect Fouling................................................................................ 174
11.4 How to Provide a Fouling Allowance....................................................................... 176
11.5 Selection oI Fouling Resistance................................................................................ 178
ix
11.6 Design Guidelines to Minimize Fouling.................................................................... 179
11.6.1 Use heat exchanger types that Ioul less............................................................... 179
11.6.2 When shell-and-tube exchangers have to be used ............................................... 179
11.6.2.1 Dirty Iluid inside tubes ............................................................................ 180
11.6.2.2 Dirty Iluid outside tubes .......................................................................... 183
CASE STUDY 11.1: INCREASING SHELLSIDE VELOCITY FOR REDUCING FOULING .............. 184
CASE STUDY 11.2: USE OF FOULING LAYER THICKNESS.................................................. 186
Chapter 12: Jibration Analysis ................................................................................................ 189
Introduction..................................................................................................................... 189
12.1 Mechanics oI Flow-Induced Vibration...................................................................... 190
12.1.1 Natural Irequency.............................................................................................. 190
12.1.2 Flow-induced vibration phenomena ................................................................... 191
12.1.3 How and when tubes vibrate.............................................................................. 193
12.1.4 Damping ........................................................................................................... 193
12.1.5 Modes oI tube Iailure......................................................................................... 193
12.2 How to Predict Damaging Flow-Induced Vibration................................................... 194
12.3 Vital Link between Flow-Induced Vibration and Pressure Drop ................................ 195
12.4 Producing a Design that is SaIe against Flow-Induced Vibration............................... 195
CASE STUDY 12.1: PRODUCING A SAFE DESIGN USING
DOUBLE-SEGMENTAL BAFFLES IN A SINGLE-PASS SHELL ........................................ 196
CASE STUDY 12.2: PRODUCING A SAFE DESIGN USING
A DIVIDED-FLOW SHELL AND SINGLE-SEGMENTAL BAFFLES ................................... 198
CASE STUDY 12.3: PRODUCING A SAFE DESIGN USING
A DIVIDED-FLOW SHELL AND DOUBLE-SEGMENTAL BAFFLES.................................. 199
CASE STUDY 12.4: PRODUCING A SAFE DESIGN USING
A NO-TUBES-IN-WINDOW DESIGN.......................................................................... 202
12.5 Rod BaIIles.............................................................................................................. 205
12.6 Acoustic Vibration ................................................................................................... 205
Chapter 13: Enhancea Heat Transfer....................................................................................... 209
13.1 What is Enhanced Heat TransIer?............................................................................. 209
13.2 BeneIits oI Enhanced Heat TransIer.......................................................................... 210
13.3 Heat TransIer Enhancement Techniques ................................................................... 211
13.3.1 Low-Iin tubes .................................................................................................... 212
13.3.2 High-Ilux tubes.................................................................................................. 212
13.3.3 Corrugated tubes................................................................................................ 213
13.3.4 Tube inserts....................................................................................................... 213
13.3.4.1 Twisted tape inserts................................................................................. 213
13.3.4.2 Wire-Iin tube inserts................................................................................ 214
CASE STUDY 13.1: COMPARISON OF DESIGNS WITH BARE TUBES AND TUBES
WITH WIRE-FIN TUBE INSERTS............................................................................... 215
13.3.5 RODbaIIle heat exchangers ............................................................................... 217
13.3.6 Helical baIIles (Helixchangers).......................................................................... 218
13.3.7 Twisted-tube heat exchangers ............................................................................ 219
13.3.8 Plate heat exchangers......................................................................................... 220
13.3.9 Spiral plate heat exchangers............................................................................... 221
13.3.10 Plate-Iin heat exchangers ................................................................................. 221
13.3.11 Printed circuit heat exchangers......................................................................... 222
13.3.12 Hybrid heat exchangers.................................................................................... 223
13.4 Evaluation oI heat transIer enhancement techniques.................................................. 223
INDEX............................................................................................................................ 225

x
PREFACE
When I was a young boy in school, I longed to be a doctor but Iate deemed otherwise and
I ended up becoming a chemical engineer. Now chemical engineering, like all other
Iields, is a very vast Iield but I ended up in the very narrow specialization oI thermal de-
sign oI shell-and-tube and air-cooled heat exchangers. I must conIess that aIter eight to
ten years oI this activity, my soul yearned Ior a change and I sought to diversiIy into the
world oI Iired heaters. However, Ior various reasons, this did not materialize and I con-
tinue to rove the world oI unIired heat transIer. AIter a period oI another Iive years, I
Iound my interest in heat exchangers rekindled, thanks to the wonderIul exposition oI this
technology by HTRI (Heat TransIer Research, Inc.). Pinch technology came oI age
around that time and proved to be a perIect Ioil and adjunct to heat exchanger thermal
design.
I have always been inspired by these words oI George Eliot: 'What do we live Ior iI not
to make the world less diIIicult Ior each other?¨ Buoyed by a positive Irame oI mind, I
thought that it might be a good idea to share some oI the things that I had learned with
readers across the world and started writing an odd paper or two Ior journals such as
Chemical Engineering Progress and Hyarocarbon Processing. A major accident leIt me
severely handicapped and curtailed my mobility drastically. This proved to be a blessing in
disguise as Iar as my literary prowess was concerned. With a lot oI time on my hands and a
PC at home, I wrote a Iew comprehensive papers Ior Chemical Engineering Progress and
Hyarocarbon Processing and received some very appreciative and heartening Ieedback.
This gave me the conIidence that I could now write a Iull-Iledged book, an idea that Ms.
Cynthia Mascone, presently Technical Editor at Chemical Engineering Progress, supported
keenly.
My desire to write this book was precipitated by the absence oI such a book. Recent heat
exchanger design literature has been predominantly occupied by proceedings oI conIerences.
There is no book on the market that explains the logic oI heat exchanger thermal design and
gives practical suggestions, recommendations, and real-liIe case studies Ior actually
designing industrial heat exchangers. So I decided to write just such a book.
The theoretical aspects oI single-phase heat transIer, condensation, and vaporization
have been presented very well in several books. So what was really required was a practical
'how to design¨ book with numerous worked-out examples or case studies to embellish or
illustrate a particular technique, Iacet, or style oI design. The thousands oI heat exchanger
designs that I have been associated with over the last 33 years provided numerous such
opportunities. They say that one picture is more eloquent than a thousand words. II you
extend this logic, one appropriate illustration by a case study is more eminently didactic than
a long dissertation on a particular subject as a case study leaves nothing to the imagination.
xi
Throughout the book, thereIore, careIully-chosen examples are presented at strategic
locations so that the reader will have a clear understanding oI the subject matter being
discussed.
While working with HTRI soItware, I have always tried to observe the interplay oI
parameters and a basic understanding oI cause and eIIect. I have also always attempted to
understand why things happen the way they do. For example, why do viscous liquids behave
so poorly inside tubes? Why does putting shells in series reduce the penalty due to
temperature proIile distortion? Why is Ilow-induced vibration really a pressure drop
problem? And so on. While working on designs, I have always asked myselI, 'Isn`t there a
better way oI doing this?¨ Such an attitude has helped immensely in improving the quality oI
the designs and I exhort all designers to adopt a similar attitude.
This book has thereIore been written primarily Ior the heat exchanger thermal designer.
But I am conIident that it will be useIul to process engineers as well, a signiIicant part oI
whose routine job is to speciIy heat exchangers. Since operating aspects are also oIten
discussed, I trust it will be oI interest to plant operation specialists as well.
Last but not least, it is my Iond hope that even undergraduate chemical and mechanical
engineering students will Iind it interesting, inIormative, and useIul. I still remember that
when I was an undergraduate student, I used to long Ior more practical, real-liIe inIormation
about industrial practice. II one considers that many engineering graduates end up working
in the chemical process industries, there may be a lot oI merit in adding such a Ilavor to heat
transIer in the university curriculum, as indeed it is to all other Iields oI human learning. The
juxtaposition oI industrial equipment design practice with basic theory will go a long way in
making the subject more meaningIul.
Being the Iirst book I have written, there is bound to be signiIicant scope Ior
improvement. I will be very grateIul to anyone oIIering positive guidance on shortcomings
as well as inaccuracies.

R. Mukherjee
Heat Transfer Consultant
New Delhi, Inaia

xii
1
CHAPTER 1
,QWURGXFWLRQ
Shell-and-tube heat exchangers (STHEs) in their various maniIestations are undoubtedly
the most widely and commonly used unIired heat transIer equipment in the chemical
processing industries. They are also used extensively in coal- and gas-based, nuclear,
ocean thermal, and geothermal power generation Iacilities.
Although strongly challenged by the plate heat exchanger in recent years, the STHE still
remains the undisputed leader in the arena oI heat exchangers. The reasons Ior this are
maniIold:
1) STHEs are very Ilexible in size and can vary Irom less than one square meter to a
thousand square meters and even more.
2) They are mechanically robust to withstand normal shop Iabrication stresses, the
rigors oI transportation and erection, as well as the stresses oI normal and abnor-
mal operating conditions.
3) They can be cleaned relatively easily. Both mechanical as well as chemical
cleaning programs can be employed.
4) The components that are most liable to Iailuretubes and gasketscan be re-
placed easily.
5) Good thermal and mechanical design methods are widely available.
6) A very wide Iabrication base is available globally.
Besides, the development oI tube inserts, helical baIIles, and twisted tubes promises to
make the STHE even more superior as these eliminate some oI the inherent shortcomings oI
STHEs.
Evidently, since the STHE is the oldest model oI the heat exchanger, it has a well-
established methodology |1²5|. Until the late 1970s and early 1980s, this knowledge was
not esoteric but was widely understood. However, with the development oI the shellside
stream analysis model and the subsequent advent oI the personal computer and tremendous
computing speeds, powerIul soItware Ior the thermal design oI STHEs gradually evolved.
Today, several very sophisticated soItware packages are available Ior the thermal design oI
STHEs, a task now carried out by engineering contractors, Iabricators, and operating
companies all over the world, representing a wide global Iraternity. Since these soItware
packages are very user-Iriendly as well, it is now very convenient to optimize and produce a
near-perIect design Ior a given application.
However, with the availability oI such superior soItware, there has been an undue
dependence on the soItware and much oI the basic understanding oI thermal design has been
lost. In other words, these soItware packages are oIten employed as 'black boxes¨ without
the designer being truly in control oI the design process and understanding the nuances oI
2
design. It must be appreciated that soItware is only a tool and with any sophisticated
soItware, a proper and sound understanding oI the Iundamental principles and interplay oI
parameters is essential in order to exploit it successIully Ior producing an optimum design.
The principal purpose oI writing this book is to help the heat exchanger thermal designer
attain such an understanding.
As example is better than precept, several case studies are presented in this book in
order to vividly bring out a particular methodology, principle, or practice that has been
advocated.
The design oI STHEs comprises two distinct activities, viz., thermal design and
mechanical design. In thermal design, the basic sizing oI the heat exchanger is
accomplished. That is to say, parameters such as the number, outer diameter, thickness and
length oI tubes, tube pitch, number oI tube passes, shell diameter, baIIle spacing and cut,
nozzle sizes, and some other construction details are Irozen. In the subsequent activity oI
mechanical design, the thicknesses and precise dimensions oI the various components are
determined and a bill oI materials produced. Detailed engineering drawings are prepared
based upon which actual Iabrication drawings are made. In this book, as the title suggests,
we shall talk principally about thermal design.
Presently there is no book available on 'practical¨ shell-and-tube heat exchanger
thermal design. The books that are available dwell heavily or Iully on the theoretical aspects
oI unIired heat transIer as they are applicable to shell-and-tube heat exchangers. II they carry
worked-out examples, these are very simplistic and certainly not comparable to what the
commercial soItware designers employ Ior carrying out real-liIe designs. The present book is
based upon the author`s experience oI 32 years in the design oI heat exchangers Ior the oil
reIineries and chemical process industries and mirrors many real-liIe situations, which were
Iar Irom straightIorward. All these experiences have been put together in a structured,
Iocused, logical, and didactic manner and special eIIort has been made at bringing out the
interplay oI parameters Ior a thorough understanding oI basic issues.
Now, we come to the individual chapters themselves. Chapter 2, 'ClassiIication oI shell-
and-tube heat exchangers,¨ gives a detailed rundown oI the various components and
constructional Ieatures oI STHEs, as a good understanding oI these is vital to the thermal
design oI this equipment. For example, the thermal engineer must be very Iamiliar with the
various components and their relationship, know when to use which type oI STHE and be
aware oI the clearances between various components, some oI which are crucial. As such,
this chapter will be oI considerable interest to mechanical designers oI STHEs as it explains
the implications oI several constructional Ieatures on thermal design.
Chapter 3, 'Thermal design and its optimization: single-phase heat exchangers,¨ is a
very important chapter as it discusses various basic Ieatures which are relevant not just to
single-phase heat exchangers, but to condensers and reboilers as well. Shellside stream
analysis and the consequent temperature proIile distortion with its associated penalty Iactor
are explained at length. These are very basic concepts which Iorm much oI the Ioundation oI
knowledge Ior heat exchanger design. The simultaneous optimization oI shellside and
tubeside calculations is certainly not an easy task. With so many parameters (such as type oI
shell, baIIling, tube pitch, and tube layout pattern), shellside optimization is itselI quite
complex. However, with the help oI logical explanation, arguments, and case studies, the
design methodology is made easy to understand and apply. The selection oI shell and/or
baIIling styles Ior the progressive reduction oI shellside pressure drop is brought out in a
clear, step-by-step method.
Chapter 4 is entitled, 'Mean temperature diIIerence.¨ AIter discussing Iundamental
3
issues oI co-current and countercurrent Ilow, it progresses to a combination oI the two and
the resultant F
t
correction Iactor. It discusses temperature cross, the use oI multiple shells in
series, and the determination oI F
t
Ior various situations. Finally it discusses shellside
temperature proIile distortion and its associated penalty on the MTD oI a heat exchanger. A
case study demonstrates how and when to reduce this penalty Iactor by the use oI multiple
shells in series, even when there is no temperature cross.
The allocation oI sides, that is, which stream should be allocated to the shellside oI an
STHE and which stream to the tubeside, is oIten not a straightIorward process. The several
parameters that inIluence the selection process are discussed in considerable detail in
Chapter 5, 'Allocation oI sides: shellside and tubeside.¨ A case study guides the reader
through the selection process.
Chapter 6 is on the 'Methodology oI the use oI multiple shells.¨ Multiple shells are
oIten required to be used either in series or in parallel (or in a combination thereoI). In some
extreme situations, one side (say, the shellside) is connected in series while the other side (in
this case, the tubeside) in parallel. This chapter, embellished by two case studies, explores in
detail the methodology oI selection oI multiple shells. Among other things, it is clearly
brought out that multiple shells in series are not just used Ior 'temperature cross¨ situations,
but also to utilize allowable shellside pressure drop Iully, and oIten result in a lower Iirst cost
when compared to a single-shell design.
So Iar, the book has dwelt on the thermal design oI single-phase STHEs. We now move
over to services and applications involving phase change. Chapter 7, 'Thermal design oI
condensers,¨ is a comprehensive elaboration oI this subject. AIter a brieI classiIication oI
condensers according to various construction and service parameters and a brieI account oI
the mechanisms oI condensation, the chapter comes to its real intent: practical guidelines Ior
thermal design. These include the determination oI shell style and baIIling, the use oI
multiple shells, the handling oI desuperheating and subcooling, nozzle sizing, and handling
oI condensing proIiles and physical property proIiles. Low pressure condensing, the use oI
low-Iin tubes, and vacuum condenser design are also addressed. There are, in all, eleven case
studies in this chapter to highlight various issues in condenser design.
Chapter 8 is on 'Thermal design oI reboilers,¨ and begins with an account oI pool
boiling and the parameters which aIIect the same. AIter a brieI discussion oI Ilow boiling,
the reader is then taken through an analytical description oI the various types oI distillation
column reboilers which includes the principal Ieatures, advantages, and disadvantages oI
each. Among all reboilers, the design oI vertical thermosyphon reboilers is the most
elaborate and complex and Ilow regime, liquid circulation, tube size, elevation, and piping
play a more proIound role here than in other reboilers. Special considerations such as very
wide boiling range, operation near critical pressure, Iilm boiling and boiling at very low ǻT
are all discussed in a lucid manner. The chapter closes aIter oIIering a guide on the selection
oI reboilers and a discussion oI the start-up oI reboilers. There are six case studies in this
chapter on reboilers.
In Chapter 9, 'Physical properties and heat release proIiles,¨ insight is oIIered on the
various vapor and liquid physical properties which are essential Ior thermal design.These are
necessarily to be Iurnished by the process licensor. Some unusual situations regarding
variation oI physical properties with temperature are reported, one example being
hydrocarbon-hydrogen mixtures. The reader is given guidance on how to Ieed heat release
proIiles, a matter that is not as simple as it may appear.
The subject oI overdesign oI heat exchangers is perceived to be important enough to
deserve an entire chapter, hence Chapter 10. It describes why overdesign is provided and
4
discusses the modalities oI overdesign Ior single-phase services, condensers, and reboilers.
Guidelines are Iurnished regarding the optimum overdesign value Ior various situations. The
eIIect oI overdesign is brought out by case studies Ior two diIIerent situations, a high
temperature approach case and a low temperature approach case.
Chapter 11, 'Fouling: its causes and mitigation,¨ is a chapter oI considerable practical
signiIicance to the thermal designer, as Iouling is oIten a severe problem. AIter reviewing
the various categories oI Iouling and the parameters which aIIect it, suggestions are oIIered
on how to speciIy Iouling resistance. Comprehensive guidelines are then recommended in
order to minimize Iouling. Although Iouling is an extremely complex phenomenon, it is still
possible to minimize it by adopting these design practices. These range Irom the use oI
speciIic non-tubular heat exchangers in certain situations to various steps and measures the
design engineer can adopt Ior STHEs, whether the Iouling Iluid is on the tubeside or on the
shellside. One case study demonstrates how the shellisde velocity oI a dirty stream can be
increased and another case study shows the proIound inIluence oI Iouling layer thickness on
pressure drop.
Chapter 12 is on Ilow-induced vibration analysis. This is an extremely important subject
as heat exchangers must be designed so that they are saIe against Iailure oI tubes due to
Ilow-induced vibration. The mechanics oI Ilow-induced vibration and the modes oI tube
Iailure are described. Guidelines are described Ior predicting Ilow-induced vibration. Four
case studies are presented on how to produce designs that are saIe against Ilow-induced
vibration. The vital link between allowable pressure drop and Ilow-induced vibration is
brought out clearly. Finally, there is a brieI exposition oI the mechanics oI acoustic vibration
with ways and means oI preventing it.
Enhanced heat transIer is not a new subject, but it has become popular only oI late.
Chapter 13 dwells on enhanced heat transIer, the various techniques that are applied to
achieve it, and its beneIits as compared to conventional shell-and-tube heat exchangers.
References
|1| Jacob, M., Heat Transfer, Jol. 1, John Wiley and Sons, 1949.
|2| Kern, D.Q., Process Heat Transfer, McGraw-Hill Book Co., 1950.
|3| McAdams, W.H., Heat Transmission, 3ra Eaition , McGraw-Hill Book Co., 1958.
|4| Ludwig, E.E., Appliea Process Design for Chemical ana Petrochemical Plants, Jol. 3, GulI
Publishing Co., 1965 (2nd ed. 1977, 3rd ed. 1995).
|5| Coulson, J.M., and Richardson, J.F., Chemical Engineering, Jol. 6, Pergamon Press, 1954
(2nd ed. 1964, 3rd ed. 1977, 4th ed. 1990).
5
CHAPTER 2
&ODVVLILFDWLRQRI
6KHOODQG7XEH+HDW([FKDQJHUV
2.1 Components of Shell-and-Tube Heat Exchangers
In order to be able to produce optimum designs, it is essential Ior the thermal designer to
have a good working knowledge oI the mechanical Ieatures oI shell-and-tube heat ex-
changers (STHEs) and more importantly, how they inIluence thermal design.
The principal components oI an STHE are:
a) shell
b) shell cover
c) tubes
a) channel
e) channel cover
f) tubesheet
g) baIIles
h) Iloating-head cover
i) nozzles
Other components include tie-rods and spacers, pass partition plates, impingement plate,
longitudinal baIIle, sealing strips, sliding strips, supports, and Ioundation.
Tube bunale
The tube bundle is the heart oI the shell-and-tube unit and comprises tubes, tubesheet(s),
baIIles, Iloating-head cover, split ring, tie-rods, spacers, impingement baIIle, longitudinal
baIIle, and sealing/sliding strips.
The Standards oI the Tubular Exchanger ManuIacturers Association (TEMA) |1| should
be reIerred to Ior the Iollowing description oI these various components.
Tubes
Tubes represent the most vital component as it is through the tube-wall that actual heat
transIer takes place. One Iluid Ilows inside the tubes while another Ilows across or along
the outside oI the tubes. Tubes may be either seamless or electric resistance welded, but
Ior reasons oI mechanical integrity and thereby reliability, the Iormer is usually preIerred.
Tubes are usually deIined by outer diameter (OD) and wall thickness (or BWG). Since
the outer diameter is Iixed and the inside diameter varies according to the thickness, it is
6
more convenient to speciIy
tubes by their outer diameter.
Besides, outer diameter is
more important than inner
diameter as the holes in baIIles
and tubesheets have to be
drilled based on the outer
diameter oI the tubes. Wall
thickness can be either mini-
mum wall (when there is no
under-tolerance, but only over-
tolerance) or average wall,
when there is both under-
tolerance and over-tolerance.
The usual practice is to order tubes with minimum wall Ior carbon steel and low-alloy steel
tubes, and with average wall Ior non-Ierrous and high-alloy steel tubes.
Depending upon the nature oI the working Iluids, tubes are employed in various
materials oI construction, the principal being carbon steel, low- and high-alloy steels, special
stainless steels, Admiralty brass and bronze, and alloys oI copper and nickel in various
proportions, including Monel, titanium, and even exotic materials, such as Hastelloys and
tantalum. Tubes are usually bare, but in situations where the shellside heat transIer
coeIIicient is highly controlling, low-Iin tubes are sometimes used advantageously, provided
the shellside Iluid is not dirty.
Tubes are held at both ends (one end Ior U-tubes) by drilled plates called tubesheets.
Whereas Ior Iixed-tubesheet and U-tube heat exchangers, tubesheets are stationary, one oI
the two tubesheets in a Iloating-head heat exchanger is literally a Iloating tubesheet.
Tubesheet thickness can vary Irom a mere 1 in. (25 mm) Ior low-pressure and low-shell
diameter applications to over 12 in. (300 mm) in high-pressure and large-shell diameter
applications.
Depending upon the severity oI the situation, tubes are either expanded into grooves in
the tubesheet or welded to them. An expanded tube-to-tubesheet joint usually has two
grooves, as shown in Fig. 2.1a. The tubes are expanded by rotating-drill tube expanders so
that tube metal actually Ilows into the grooves. Welded tube-to-tubesheet joints (Fig. 2.1b)
aIIord higher integrity (albeit at a higher cost) and are usually employed Ior severe
conditions, such as high pressure (say, in excess oI 1140 psig or 80 kg/cm
2
g) or when
handling toxic or inIlammable Iluids where leakage is not permitted. When the hot and cold
streams cannot be allowed to mix, a special double-tubesheet construction is employed.
Baffles
BaIIles serve to support the tubes as well as to impart a suIIiciently high shellside veloc-
ity to yield a satisIactory heat transIer coeIIicient. BaIIles are held securely in place by a
combination oI tie-roas and spacers. AIter a baIIle has been guided along the tie-rods, a
set oI spacers is introduced over the tie-rods, aIter which the next baIIle is inserted, and
so on. Evidently, the length oI the spacers is equal to the baIIle spacing. Details oI baIIle
conIiguration are discussed in detail in Section 3.4.4.
A small clearance between the tube outside diameter and the baIIle hole diameter is
essential to permit tubes to be inserted through baIIles and assembly, as well as Ior tube
replacement, iI and when required. TEMA speciIies this gap as 1/64 in. (0.4 mm) Ior a
Fig. 2.1 Tube-to-tubesheet joint
7
maximum unsupported tube length larger than 36 in. (900 mm) and 1/32in. (0.8 mm) Ior a
maximum unsupported tube length smaller than 36in. (900 mm). A part oI the shellside Iluid
leaks through this small gap: this is one oI the shellside leakage streams and will be
discussed in detail in Section 3.4.5. An excessive clearance provides excessive leakage and
insuIIicient tube support with the possibility oI vibration.
The outer diameter oI the baIIle has to be less than the inside diameter oI the shell to
permit insertion oI the tube bundle into the shell and removal oI the tube bundle Irom the
shell. However, since the shellside leakage stream between the shell and the baIIles is
particularly detrimental to shellside thermal perIormance (see Section 3.4.5), this gap should
be as small as possible. Table 2.1 shows this gap Ior various shell ID ranges, as per the
TEMA standards.
Channel, channel cover, ana pass-partition plates
The channel serves to introduce the tubeside working Iluid into the exchanger as well as
to direct it out oI the exchanger. A channel may either be oI a bonnet construction
(TEMA head type B) where a dished end is welded to the channel barrel, or have a
Ilanged channel cover (TEMA head type A). Pass-partition plates inside the channels
serve to direct the tubeside Iluid along the tubes as desired by the thermal engineer (i.e.,
they vary the number oI tube passes). They Iit tightly into grooves in the tubesheet and
channel cover in order to eliminate the possibility oI leakage oI the tubeside Iluid Irom
one pass to the nextsuch leakage would evidently be highly detrimental to satisIactory
perIormance oI a heat exchanger. The actual sealing is eIIected by well-set gaskets that
must be checked periodically and replaced whenever they are Iound wanting.
The arrangement oI the pass-partition plates in multi-pass heat exchangers is somewhat
arbitrary. However, all oI them try to accomplish the Iollowing goals: maximize the number
oI tubes while employing a Iairly even distribution oI tubes in the various passes. Another
consideration in some instances is to minimize the number oI Ilow lanes in the Ilow
direction when the same has a crucial bearing on the shellside stream analysis and, thereIore,
perIormance. For example, having two pass-partition lanes in the Ilow direction instead oI
three will result in a smaller pass-partition leakage stream Ilow Iraction.
Some usual patterns oI pass-partition arrangements Ior diIIerent numbers oI tube passes
are illustrated in Fig. 2.2.
Shell ana shell cover
The shell serves to contain the shellside Ilowing stream and Iorms the outer casing oI the
tube bundle. It also serves to introduce the working Iluids into the heat exchanger as well
as to remove them Irom the heat exchanger. A shell cover is required in the case oI split-
ring pull-through Iloating-head heat exchangers.
Table 2.1: Table RCB-4.3 oI TEMA Standards: Standard cross baIIle and support plate clearances
Nominal Shell ID Design ID of Shell minus Baffle OD
6²17 in. (152²432 mm)
18²39 in. (457²991 mm)
40²54 in. (1016²1372 mm)
55²69 in. (1397²1753 mm)
70²84 in. (1778²2134 mm)
85²100 in. (2159²2540 mm)
1/8 in. (3 mm)
3/16 in. (5 mm)
1/4 in. (6 mm)
5/16 in. (8 mm)
3/8 in. (9.5 mm)
7/16 in. (11.1 mm)
8
Impingement plate
The inlet nozzle is oIten provided with an impingement plate to protect the uppermost
tubes located just below the shellside inlet nozzle against direct impingement by the
shellside Iluid. Such impingement can cause erosion, cavitation, and/or vibration. TEMA
speciIies an upper limit on ȡv
2
oI 1500 lb/It sec
2
(2232 kg/m sec
2
) Ior noncorrosive, non-
abrasive single-phase Iluids. For all other liquids, including a liquid at its boiling-point,
the limit on ȡv
2
is 500 lb/It sec
2
(744 kg/m sec
2
). For all saturated vapors and all two-
phase mixtures, an impingement plate is always required as, in
these services, it is possible to have liquid droplets traveling at
vapor velocities, with very high values oI ȡv
2
.
An impingement plate must be located suIIiciently below
the shell ID so as to leave suIIicient Ilow area between the shell
and the plate Ior the Ilow to discharge without excessive velocity
and, thereby, pressure loss. Consequently, a Iew rows oI tubes
usually have to be eliminated Irom the top oI the tube Iield (see
Fig. 2.3).
Sliaing ana sealing strips
A pair oI sliding strips is provided at the bottom oI Iloating-
head tube bundles Ior their insertion and removal to and Irom

Fig. 2.2 Usual patterns oI pass partition arrangements (Reprinted Irom the Heat Exchanger Design Hand-
book, 2002 with permission oI Begell House, Inc.)

Fig. 2.3 Impingement plate
(Reprinted Irom the Heat
Exchanger Design Hand-
book, 2002 with permission
oI Begell House, Inc.)
9
the shell. A suIIicient number oI sealing strips is required to be inserted in the gap be-
tween the shell and the outermost tubes in Iloating-head tube bundles to minimize leak-
age oI the shellside Iluid around the tube bundle.
A typical tube layout drawing is shown in Fig. 2.4, showing many oI the above
components, viz., tubes, tie-rods, sealing strips, sliding strips, and impingement plate.
Longituainal baffle
By its very name, a longitudinal baIIle is placed transversely along the centerline oI the
shell and is employed to divide the shell into two or more compartments (see Figs. 2.5a

Fig. 2.4 Tube layout diagram (Reprinted Irom the Heat Exchanger Design Handbook, 2002 with permission
oI Begell House, Inc.)

(a) Schematic (b) TypicaI arrangement
Fig. 2.5 Longitudinal baIIle (Reprinted Irom the Heat Exchanger Design Handbook, 2002 with permission oI
Begell House, Inc.)
10
and 2.5b). For example, a single longitudinal baIIle Irom one tubesheet to just short oI the
other tubesheet produces an F shell, that is, a shell with two shell passes. Longitudinal
baIIles may also be employed to produce G and H shells (see Fig. 2.6).
It will be apparent that in the case oI removable tube bundles, the longitudinal baIIle will
be a part oI the tube bundle and can thereIore not be Iixed to the shell. In order to prevent
bypassing oI the shellside Iluid Irom the Iirst pass to the second pass along the edges oI the
longitudinal baIIle in an F shell, Ilexible strips are employed at both ends oI the longitudinal
baIIle, all along the length oI the shell (Fig. 2.5b). Also, the shellside pressure drop is usually
limited to 0.35 kg/cm
2
in F shells in order to minimize the possibility oI leakage across the
longitudinal baIIle.
However, in the case oI Iixed-tubesheet heat exchangers, since the tube bundle cannot
be removed Irom the shell, the longitudinal baIIle can be Iixed to the shell by welding.
TEMA has developed a nomenclature Ior the speciIication oI various construction types.
An STHE is divided into three parts: the Iront head, the shell and the rear head. The
nomenclature Ior the various construction possibilities are shown in Fig. 2.6.
2.2 Front and Rear Heads
ReIerring to the classiIication in TEMA (Fig. 2.6), there are Iive Iront head types: A, B,
C, D, and N. There are eight rear head types: L, M, N, P, S, T, U, and W, which corre-
spond in practice to only three general construction types, namely Iixed-tubesheet, U-
tube, and Iloating-head. Rear head L is identical to a Iront head A, and rear head M is
identical to a Iront head B, while N is the same nomenclature. These three rear head types
belong to Iixed-tubesheet heat exchangers. U applies to U-tube heat exchangers while S,
T, P, and W represent various types oI Iloating-head construction. In the Iollowing sec-
tion, the diIIerent types oI heads will be discussed. The overall classiIication oI Iixed-
tubesheet, U-tube, and Iloating-head heat exchangers will be discussed in Section 2.3.
The various shell types will be discussed in Section 3.4.1.
A/M-type heaa
In this type, the channel barrel is Ilanged at both ends. The tubesheet is bolted to one
Ilange and a Ilat channel cover to the other. Thus only the channel cover has to be re-
moved Ior cleaning oI the tubes by rodding or hydro-blasting; the channel and piping are
not disturbed. However, Ior any inspection or repair oI a tube-to-tubesheet joint, the en-
tire channel will usually be required to be removed, especially Ior peripheral tubes. Re-
moval oI the entire channel is also required Ior removal oI the tube bundle.
Despite its higher cost due to the presence oI two Ilanged joints, this type oI channel is
very commonly used, especially in petroleum reIineries where dirty Iluids are handled,
necessitating Irequent bundle removal.
B/M-type heaa
In this construction, the channel barrel is Ilanged at one end only, the other end being
welded to a semi-elliptical bonnet or dished end. This type is lighter and cheaper than the
A type, especially at high pressures, as the thickness oI the bonnet is considerably less
than that oI a Ilat cover plate. Here the entire bonnet has to be removed Ior cleaning even
the inside oI tubes, which means that the channel piping connections have to be disman-
tled. For removal oI the bundle Ior cleaning the outside oI tubes, the entire channel has to
be removed anyway, as indeed has to be the A-type channel. ThereIore, this channel type
is recommended Ior services where the tubeside Iluid is clean. It will be seen in Section
11
2.3.2 that U-tubes are recommended Ior clean tubeside Iluids. Thus, a BEU construction
is commonly incorporated in heat exchangers handling clean tubeside Iluids such as
steam or reIrigerants.
For services having a high design pressure on the tubeside, a B-type channel is
advantageous as it leads to cost savings due to thin-wall bonnet construction.
A B-type channel may also be preIerred to an A-type channel Ior small diameter heat
exchangers where it is not diIIicult to remove the entire channel due to the lower weight.
C-type heaa
This is similar to the A type in that there is a Ilat channel cover. However, here the
tubesheet end oI the channel is not Ilanged to the tubesheet but welded to itthe
tubesheet is Ilanged to the shell. Evidently, this type oI construction is intended only Ior
removable bundles. It minimizes bolted joints and their dismantling on the tubeside
which would be advantageous in services handling inIlammable or lethal Iluids on the
tubeside.

Figure 2.6 TEMA construction types (Source: TEMA Standards)
12
It also enables the entire channel and tube bundle to be leIt in situ while the shell is
drawn away. This will be advantageous when the tube bundle is much heavier than the shell,
so that it becomes easier to remove the shell instead oI the tube bundle.
There is one distinct disadvantage in this channel type. AIter removal oI the cover plate,
access to the tube-to-tubesheet joints Ior any repairs is diIIicult. It thereIore becomes
necessary to speciIy a larger tube bundle-shell clearance than is otherwise required, thereby
increasing the cost. Consequently, the C-type head is not commonly used.
N-type heaa
We have just seen that in the C-type, the tubesheet is integral to the channel and Ilanged
to the shell. In the N-type, the tubesheet is integral to both the channel and the shell and
can evidently be used only Ior Iixed-tubesheet heat exchangers. It will be noticed that
here, while joints are minimized Ior the tubeside, they are totally eliminated Ior the shell-
side. Thus, they are employed in services handling toxic or lethal Iluids on the shellside.
Like the A-type head, one advantage is that the channel piping joints do not have to be
disturbed Ior cleaning the tubes Irom the inside. However, it has the same disadvantage as
the Ctype head regarding access to the peripheral tube-to-tubesheet joints.
In one variation oI the N-type head, the tubesheet is integral to the shell but Ilanged to
the channel.
D-type heaa
For channels Ior high-pressure services on the tubeside (design pressure ~ 2133 psig or
150 kg/cm
2
g), specially designed and non-bolted closures are employed. Many oI these
are patented. These special high-pressure channels are generally reIerred to as D-type
closures. Very large-diameter bolts and nuts are required to be employed so that the use
oI hydraulic bolt tensioners becomes necessary.
Conical heaa
Although not Iormally designated in the TEMA nomenclature, these heads are used quite
oIten in single tube pass heat exchangers, especially in vertical thermosyphon reboilers.
These are simply conical sections, Ilanged at both ends with the Ilange at the larger di-
ameter bolted to the tubesheet and the Ilange at the smaller diameter bolted axially to the
piping. Apart Irom the convenience oI matching the column return line elevation when
associated with a bend in a vertical thermosyphon reboiler (see Chapter 8), conical heads
are used when the nozzle is too large to be Iitted axially to the channel. Normally, the
diameter oI a nozzle is limited to halI the diameter oI the channel (or shell) to which it
has to be connected, based upon strength considerations. Thus, in services handling low
pressure gases or vapors on the tubeside, or simply very high Ilow rates, where the nozzle
size is greater than 50° oI the channel diameter, conical nozzles oIIer a convenient solu-
tion. Should the inside oI tubes require cleaning, however, both the conical channels will
have to be removed.
Sometimes only one channel is required to be conical and the other can be cylindrical.
For example, in the case oI a vertical thermosyphon reboiler, the inlet liquid line is quite
small compared to the channel diameter and, thereIore, the inlet channel can be cylindrical.
The outlet two-phase line may, however, be more than 50° oI the channel diameter and is
conveniently made conical. Similarly, in the case oI a low-pressure condenser, the inlet
vapor line can be quite large while the outlet condensate line can be rather small. Here, the
inlet channel will be conical and the outlet channel cylindrical.
13
Shell types
Shell types are discussed in detail in Section 3.4.1.
2.3 Classification by Construction
2.3.1 Fixed-tubesheet heat exchanger
A Iixed-tubesheet heat exchanger (Fig. 2.7) has straight tubes secured at both ends to
tubesheets welded to the shell. The construction may have removable channel covers
(e.g., AEL), bonnet-type channel covers (e.g., BEM), or integral tubesheets (e.g., NEN).
The principal advantage oI a Iixed-tubesheet construction is low cost, as it has the
simplest construction. In Iact, the Iixed tubesheet is the cheapest construction type, as long
as no expansion joint is required. Other advantages are:
1) permits mechanical cleaning oI the inside oI the tubes as these are accessible aI-
ter removal oI the channel cover or bonnet, and
2) oIIers maximum protection against leakage oI the shellside Iluid as there are no
Ilanged joints, an advantage Ior lethal or inIlammable services.
The disadvantages oI a Iixed-tubesheet construction are:
1) As the bundle is 'Iixed¨ to the shell and cannot be removed, the outside oI the
tubes cannot be cleaned mechanically. Since mechanical cleaning (by rodding or
jetting) is usually resorted to, this represents a limitation in that the shellside Iluid
should be clean, as otherwise, the necessary cleaning oI the outside oI the tubes
cannot be carried out. However, it should be noted that chemical cleaning is Iea-
sible. Hence, iI a satisIactory chemical cleaning program can be determined and
employed, Iixed-tubesheet construction may be selected Ior Iouling services on
the shellside. Since chemical cleaning is cumbersome and diIIicult to employ, it
is not very prevalent.
2) In the event oI a large diIIerential temperature between the tubes and the shell,
the tubesheets will be unable to absorb the diIIerential stress, thereby making it
necessary to incorporate an expansion joint on the shell. This takes away the ad-
vantage oI low cost to a signiIicant extent.
2.3.2 U-tube heat exchanger
As the name implies, the tubes oI a U-tube heat exchanger (Fig. 2.8) are bent in the shape
oI a U. Evidently, there is only one tubesheet in a U-tube heat exchanger. However, the
bending oI tubes represents an additional cost. Further, the minimum U-bend diameter is
usually three times the tube outside diameter so that the central pass-partition lane is con-
siderably larger in a U-tube heat exchanger than in one having straight tubes. Conse-

Fig. 2.7 Fixed tubesheet heat exchanger (Reprinted Irom the Heat Exchanger Design Handbook, 2002 with
permission oI Begell House, Inc.)
14
quently, Ior a given number oI tubes, a U-tube heat exchanger will have a larger shell
diameter. Evidently, this diIIerence in shell diameter will be larger Ior small shell diame-
ters where the central U-bend lane will represent a larger Iraction oI the shell cross-
sectional area and thereby result in a larger reduction in the number oI tubes that can be
accommodated.
The additional cost oI the bending oI U-tubes and the larger shell diameter more or less
oIIset the saving in cost due to the elimination oI one tubesheet. Thus, the cost oI a U-tube
heat exchanger is comparable to that oI a Iixed-tubesheet exchanger.
The advantage oI a U-tube heat exchanger is that, being Iree at one end, it permits the
bundle to expand or contract according to the diIIerential stress set up. Thus, no detailed
calculations are required to be perIormed Ior the design condition and various other
conditions (start-up, upset, shutdown, etc.) to determine the requirement oI an expansion
joint on the shell. Besides, it permits the outside oI the tubes to be cleaned, as the tube
bundle can be removed.
The disadvantage oI the U-tube construction is that the inside oI the tubes cannot be
cleaned eIIectively, since the U-bends would require Ilexible-end drill shaIts Ior cleaning.
Thus, U-tube heat exchangers should not be used Ior services which have a dirty Iluid inside
the tubes. This places a very severe limitation on U-tube heat exchangers Ior reIinery
services, which usually have dirty streams on both the tubeside and the shellside. This is
primarily the reason why Iixed-tubesheet or U-tube heat exchangers are a very uncommon
sight in oil reIineries.
2.3.3 Floating-head heat exchanger
The Iloating-head is the most versatile type oI STHE, and also the costliest. It is so chris-
tened because while one tubesheet is Iixed relative to the shell, the other is Iree to 'Iloat¨
within the shell. Thus, it permits Iree expansion oI the tube bundle and also permits the
cleaning oI both the inside and outside oI the tubes. It can thereIore be used Ior services
where both the shellside and the tubeside Iluids are dirty and, consequently, it is the stan-
dard construction type used in dirty services as in oil reIineries.
The higher cost oI the Iloating-head heat exchanger is due to the Iact that there are more
components in this type oI construction (see Figs. 2.9²2.12) than in the Iixed-tubesheet or
the U-tube types. Furthermore, the shell diameter is larger in a Iloating-head construction as
it has to be greater than the Iloating tubesheet so that the tube bundle may be pulled out.
There are various types oI Iloating-head construction:
Pull-through with backing aevice. TEMA type S
This is the most common application in the chemical process industries. ReIerring to Fig.
2.9, it is seen that the Iloating-head cover is secured against the Iloating tubesheet by

Fig. 2.8 U-tube heat exchanger (Reprinted Irom the Heat Exchanger Design Handbook, 2002 with permission
oI Begell House, Inc.)
15
bolting it to an ingenious devicea split backing ring. This Iloating-head closure is lo-
cated beyond the end oI the shell and contained by a shell cover oI a larger diameter. For
dismantling the heat exchanger, the shell cover is removed Iirst, then the split backing
ring, and Iinally the Iloating-head cover aIter which the tube bundle can be removed Irom
the 'stationary¨ end. For assembling the heat exchanger, the reverse order is Iollowed.
Pull-through. TEMA type T
In this construction, the entire tube bundle, including the Iloating-head assembly, can be
removed Irom the stationary end as the shell diameter is made large enough to match the
diameter oI the Iloating-head Ilange (see Fig. 2.10). The Iloating-head cover is bolted
directly to the Iloating tubesheet so that a split backing ring is not required.
The advantage oI this type oI construction is that the tube bundle may be removed Irom
the shell without removing either the shell or the Iloating-head cover, thus reducing
maintenance time. It is particularly suited to and usually employed only Ior kettle reboilers
having a dirty heating medium where U-tubes cannot be employed. The shell diameter is the
largest in this type oI construction, as not only the Iloating tubesheet but the Iloating-head
cover has to be removed through the shell. Hence, the cost is the highest in this type oI
construction.
Outsiae-packea stuffing-box. TEMA type P
In this construction, the shellside Iluid is sealed by rings oI packing compressed within a
stuIIing-box by a Iollower ring (see Fig. 2.11). The packing permits movement oI the
Iloating tubesheet. Since this construction is prone to leakage, its use is limited to shell-
side services, which are nonhazardous and nontoxic services, as well as having moderate
pressure and temperature (570 psig or 40 kg/cm
2
g and 570ƒF or 300ƒC).

Fig. 2.9 Floating-head heat exchanger with backing device (TEMA type S) (Reprinted Irom the Heat Ex-
changer Design Handbook, 2002 with permission oI Begell House, Inc.)

Fig. 2.10 Pull-through Iloating-head exchanger (TEMA type T) (Reprinted Irom the Heat Exchanger Design
Handbook, 2002 with permission oI Begell House, Inc.)
16
Outsiae-packea lantern ring. TEMA type W
Here, the shellside and tubeside Iluids are sealed by separate rings oI packing or O-rings
and separated by a lantern ring provided with weep holes (Fig. 2.12). Hence, any leakage
will be to the outside. The width oI the tubesheet necessarily has to be suIIicient to ac-
commodate the two packing rings and the lantern ring, plus the expansion margin. This
design is limited to 140 psig (9.9 kg/cm
2
g) and 400ƒF (204ƒC). Because oI its very con-
struction, the number oI tube passes can only be either one or two.
2.4 Classification by Service
Shell-and-tube heat exchangers may also be classiIied according to their service. Basi-
cally, a service may be single-phase (such as the cooling or heating oI a liquid or gas) or
two-phase (such as condensing or vaporizing). Since there are two sides to an STHE, this
can lead to several combinations oI services.
Broadly, services can be classiIied as Iollows:
a) single-phase (both shellside and tubeside)
b) condensing (one side condensing and the other single-phase)
c) vaporizing (one side vaporizing and the other single-phase)
a) condensing/vaporizing (one side condensing and the other vaporizing)

Fig.2.11 Outside-packed stuIIing box Iloating-head heat exchanger (TEMA type P) (Reprinted Irom the
Heat Exchanger Design Handbook, 2002 with permission oI Begell House, Inc.)

Fig. 2.12 Outside-packed lantern ring Iloating-head heat exchanger (TEMA type W) (Reprinted Irom the
Heat Exchanger Design Handbook, 2002 with permission oI Begell House, Inc.)
17
The Iollowing nomenclature is usually used Ior the sake oI convenience:
Heat exchanger ² both sides single-phase process streams (that is, not an utility)
Cooler ² one stream a process Iluid and the other stream a cold utility, such as cool-
ing water or air
Heater ² one stream a process Iluid and the other stream a hot utility, such as steam
or hot oil
Condenser ² one stream a condensing vapor and the other stream a cold utility such
as cooling water or air
Chiller ² one stream a process Iluid being condensed at sub-zero temperature and the
other stream a boiling reIrigerant or process stream (evidently cryogenic)
Reboiler ² one stream a bottom stream Irom a distillation column and the other a hot
utility (steam or hot oil) or a process stream.
Vaporizer ² one stream a vaporizing liquid and the other stream a gas or a liquid or a
condensing vapor.
It may be noted that a condensing stream may be condensing totally or partially. Simi-
larly, a vaporizing stream may also be vaporizing partially or totally. For example, a
thermosyphon reboiler eIIects partial vaporization, whereas a kettle reboiler may eIIect
either partial or total vaporization. In a distillation column having total reIlux, the entire
overhead vapor is condensed, thereby representing a total condenser. However, it oIten
happens that there is only partial reIlux back to the distillation column and the balance
overhead is a product or goes Ior Iurther processing. This represents a partial condenser.
References
|1| Tubular Exchanger ManuIacturers Association, Stanaaras of the Tubular Exchanger Manu-
facturers Association, 8th Eaition, TEMA, New York (1999).
18

19
CHAPTER 3
7KHUPDO'HVLJQDQG2SWLPL]DWLRQ
RI6LQJOH3KDVH+HDW([FKDQJHUV
The design oI shell-and-tube heat exchangers comprises two distinct activities: thermal
design and mechanical design. In thermal design, the heat exchanger is sized, which
means that all the principal construction parameters such as shell type and diameter,
number oI tubes, tube OD and thickness, tube length, tube pitch, number oI tube passes,
baIIle spacing and cut, and nozzle sizes are determined. In mechanical design, detailed
calculations are carried out to determine the dimensions oI various components such as
tubesheets, Ilanges, shell, etc. and a complete bill oI materials and engineering drawings
such as bundle assembly and setting plan drawings are generated. In this book, we shall
talk predominantly about thermal design.
The basic equations Ior tubeside and shellside heat transIer and pressure drop are well
known and are presented in several books (see reIerences). This chapter will dwell on the
application oI these and other correlations Ior the optimum thermal design oI heat
exchangers. BeIore we proceed any Iurther, let us see what the broad objectives oI a thermal
designer are when he or she sets out to produce a thermal design.
3.1 Broad Objectives of Thermal Design
The basic aims oI a thermal designer are to:
a) Produce a thermal design that has a low overall cost; the lower, the better.
The overall cost oI a heat exchanger is the sum oI the initial cost and the operating cost.
The initial cost is evidently the Iixed cost or the Iirst cost oI the heat exchanger. The
operating cost is the sum oI the pumping cost, the maintenance cost and the downtine
cost. The maintenance cost is the sum oI the cost oI periodically cleaning the exchanger,
the cost oI any anti-Ioulant treatment and the cost oI any repair or replacement.
Thus, it is not enough to produce a design having a very low Iixed cost iI its operating
cost is high due, Ior example, to Irequent Iouling and thereby the requirement Ior Irequent
cleaning. Designers oIten lose sight oI the operating cost oI a heat exchanger and should
always attempt to minimize Iouling and also minimize pressure drop. This represents a
direct conIlict because as we shall see later on in the book, the best way to minimize Iouling
is to maximize velocity (within limits oI erosion, oI course) which will directly maximize
pressue drop and thereby power consumption. Obviously, then, the designer has to optimize
the design so that while the velocity is not low enough to exacerbate Iouling, the pressure
drop is not excessively high.
In this context, the selection oI the materials oI construction is very important. The
materials should be good enough to permit the heat exchanger to Iunction Ior the liIetime oI
20
the plant (typically 20²25 years) without major repairs and without replacement oI
components (such as tubes). However, the materials should not be inordinately expensive
because then the Iirst cost oI the heat exchanger will become unnecessarily high.
b) Utilize allowable pressure drops as Iully as possible.
It will be easily appreciated that the higher the velocity oI a given stream, the higher will
be its heat transIer coeIIicient. However, accompanying the high heat transIer coeIIicient
will be a high pressure drop. So while the Iormer (high heat transIer coeIIicient) will tend
to reduce the Iirst cost oI the heat exchanger, the latter (high pressure drop) will tend to
increase the operating cost oI the heat exchanger. Thus, a very important goal Ior a good
thermal design is the best utilization oI the allowable pressure drop. This is discussed in
more detain in Section 3.2.
It sometimes happens that the permitted pressure drop is unnecessarily high to produce a
good design and iI, in such cases, the pump speciIications have not been Irozen, they can be
revised to take advantage oI the lower (than anticipated) pressure drop. However, iI the
pump speciIications have already been Irozen, the possible savings in pumping power
cannot be realized and the diIIerential pressure drop will just be wasted across a control
valve.
c) Maintain a good Ilow pattern on the shellside and adequate velocities on both sides to
minimize Iouling.
This has been discussed above and is treated in much more detail in Chapters 3 and 11.
a) Produce a design that is saIe against Iailure oI tubes due to Ilow-induced vibration
The thermal designer has to produce a design that is saIe against any possibility oI Iailure
oI tubes due to Ilow-induced vibration. Chapter 12 is a detailed exposition oI Ilow-
induced vibration in shell-and-tube heat exchangers and discusses various construction
styles (on the shellside) that can be adopted to make a design saIe against Iailure oI tubes
due to Ilow-induced vibration. Once again, the Iixed cost oI the heat exchanger must be
considered and only the lowest-cost design style should be adopted to make the heat
exchanger saIe. Thus, a no-tubes-in-window design is inherently saIer against vibration
Iailure but is more expensive than, Ior example, a divided-Ilow shell. ThereIore, iI the
latter can produce a saIe design, a no-tubes-in-window design should not be employed.
3.2 Data to be Furnished for Thermal Design
BeIore coming to the actual thermal design oI a shell-and-tube heat exchanger, let us take
a look at the data required Ior the same.
The Iollowing inIormation must be Iurnished by the process licensor Ior both streams
(wherever applicable) beIore thermal design can be taken up:
Flow rate
The complete breakdown oI vapor, liquid, steam, water, and noncondensable Ilow rates
must be Iurnished, as applicable, at both the inlet and outlet oI the heat exchanger.
Inlet ana outlet temperatures
Heat release profiles
By heat release proIiles are meant plots oI the Iollowing variables with temperature,
wherever applicable:
21
• heat duty
• weight Iraction vapor
• vapor molecular weight
Evidently, Ior single-phase services, the last two are not applicable. Besides, the plot oI
heat duty versus temperature is essentially linear so that no heat release proIile is really
required.
However, Ior any service involving phase change, heat release proIiles as deIined above
are a must. II the temperature diIIerence between the inlet and the outlet is rather small, such
as 9°F (5°C) or 18°F (10°C), a straight-line heat duty versus temperature may be speciIied as
the curvature will be minimal. Heat release and other proIiles will be discussed in detail in
Chapter 9.
Operating pressure
This is not really required Ior liquids as their properties do not vary with pressure to any
signiIicant extent. However, it is required Ior gases and vapors as their properties,
particularly gas density, vary with pressure. However, iI the physical properties are
Iurnished, the operating pressure is no longer required Ior single-phase gas and
condensing vapor streams. For reboilers and vaporizers, however, the operating pressure
is still required because boiling heat transIer coeIIicient is a strong Iunction oI operating
pressure and critical pressure. This will be discussed in detail in Chapter 8.
Allowable pressure arop
This is a very important parameter Ior heat exchanger design and the process licensor
should be aware oI the signiIicance oI the same Ior thermal design. The higher the
pressure drop, the higher will be the heat transIer coeIIicient and, thereby, the lower the
heat transIer area and Iixed cost. However, the operating cost will be higher.
Consequently, the allowable pressure drop represents the optimum balance between Iixed
cost and operating cost oI a heat exchanger such that the total cost is minimal. Generally,
Ior liquids, a value oI 7²10 psi (0.5²0.7 kg/cm
2
) is permitted per shell. A higher value is
usually warranted Ior viscous liquids, especially iI routed through the tubeside. For gases,
the usually allowed value is 0.7²2.8 psi (0.05²0.2 kg/cm
2
), a very typical value being 1.4
psi (0.1 kg/cm
2
).
It must be stated here that, whereas typical values are generally applicable, speciIic
instances must be investigated more thoroughly. For example, iI it is Iound that the
allowable pressure drop Ior a particular stream is representing a severe constraint in
producing a satisIactory thermal design, the eIIect oI a higher allowable pressure must be
examined to arrive at the optimum design based upon minimum total cost.
It may be stated here that this aspect is very important Ior good thermal design oI heat
exchangers and indeed Ior any good design: the designer must not Iollow the beaten path but
always question the various parameters speciIied and examine alternatives. It should be
remembered that some oI the parameters speciIied are not really sacrosanct, but are only
based upon hereditary engineering practice. A special situation may call Ior special
measures. The author has Iound that it always helps to keep asking oneselI: 'Is there not a
better way oI doing this? Why don't I see what happens iI I change this parameter?¨
Fouling resistance
This is another extremely important parameter and one that is unIortunately based more
22
upon experience than Iundamental understanding, thanks to the complexity oI the
phenomenon. II the Iouling resistance oI a particular stream is not Iurnished, the heat
exchanger designer should adopt the same Irom TEMA standards or Irom past operating
experience. This subject is discussed in Iar greater detail in Chapter 11.
Physical properties
Principally viscosity, thermal conductivity, density and speciIic heat, preIerably at both
inlet and outlet temperatures are required. Viscosity data must be supplied at inlet and
outlet temperatures, especially Ior liquids, as the variation with temperature is
considerable and irregular (not linear, semi-log, or log-log). Additional properties
required are latent heat and surIace tension Ior condensing/vaporizing services, and
critical pressure and temperature Ior vaporizing services.
Heat auty
Evidently, the heat duty speciIied should be consistent Ior both the shellside and the
tubeside. In Iact, it is a good idea Ior the thermal engineer to corroborate this as licensors
occasionally slip on this aspect.
Type of heat exchanger
II the type oI heat exchanger is not Iurnished, the heat exchanger designer can choose the
same based upon the characteristics oI the various types oI construction explained earlier
in Chapter 1. In Iact, the heat exchanger designer is normally in a better position than the
process engineer to choose the type oI heat exchanger.
Line si:es
It is desirable to match nozzles sizes with line sizes as no expander or reducer will then
be required. However, criteria Ior nozzle sizing (velocity and ȡv
2
) are usually more
stringent Ior heat exchanger nozzles than Ior lines, especially Ior the shellside inlet.
Nozzle sizing is based upon pressure drop, which in turn is based upon expansion and
contraction losses, whereas line sizing is based upon line pressure drop, which is
dependent upon velocity and length oI the line. Consequently, nozzle sizes are sometimes
required to be one size (or even more in exceptional circumstances) higher than the
corresponding line sizes. This is especially true Ior small line sizes where the change in
Ilow area Irom one pipe size to the next is quite considerable.
Preferrea tube si:e
By tube size is meant tube OD, thickness, and length. The customer may have a preIerred
tube OD/thickness, usually based upon inventory considerations, and the available plot
area will determine the maximum permitted tube length. Many plant owners preIer to
standardize all three dimensions, based again upon inventory considerations.
It may be mentioned here that Ior a Iixed-tubesheet heat exchanger, no extra length has
to be allocated Ior bundle removal, as must be done Ior U-tube and Iloating-head heat
exchangers. However, there may be exceptions to a standardization program. For example, a
plant may have standardized 0.75-in. (19.05-mm) OD and 16-BWG (1.65-mm) thick tubes.
However, Ior a vertical thermosyphon reboiler or tubeside condenser, it may become
unavoidable to employ 1-in. (25.4-mm) OD tubes under certain circumstances. Also, Ior
very small heat exchangers, it may be judicious to employ a shorter tube length.
As spare tubes are not usually stocked Ior Iixed-tubesheet heat exchangers, a
23
standardization program is not really warranted Ior such heat exchangers. Consequently, a
designer has Ireedom in optimizing the tube length Ior Iixed-tubesheet heat exchangers as
elaborated in Sections 3.3.1²3.3.4.
Maximum shell aiameter
This is based upon considerations oI tube bundle removal and is limited by liIting (crane)
capacities. It is evident that such limitations are applicable only to removable-tube-bundle
heat exchangers, namely U-tube and Iloating-head. For Iixed-tubesheet heat exchangers,
there is no such limitation. The only limitation Ior Iixed-tubesheet heat exchangers is the
manuIacturer's Iabricating capability as well as the availability oI such components as
Iorgings, dished ends, and Ilanges. Thus, whereas Iloating-head heat exchangers are oIten
limited to a shell ID oI 5559 in. (1400²1500 mm) and a tube length oI 20 It (6000 mm)
or 30 It (9000 mm), Iixed tubesheet heat exchangers are built up to a shell ID oI 8.2 It
(2500 mm) and tube length oI 40 It (12,000 mm), or even more.
Materials of construction (MOC)
Should the MOC oI tubes and shell be identical, all components oI a heat exchanger will
be oI this MOC. Thus, only the shell and tube MOC need be speciIied. However, iI the
shell and tubes are oI diIIerent metallurgy, the MOCs oI all principal components should
be speciIied, to avoid any ambiguity. The principal components are shell (and shell
cover), tubes, channel (and channel cover), tubesheets, and baIIles. Tubesheets and
Iloating-head covers represent special cases as they may be lined or clad on both sides or
on one side only.
Corrosion allowance
This is really an adjunct to the MOCs and should be speciIied by component in cases
where the MOC is not identical Ior all components. No corrosion allowance is applied on
tubes as the standard thicknesses recommended by TEMA already incorporate a
corrosion allowance.
For services involving multiple shells in series, the materials oI construction and
corrosion allowance oIten vary progressively Irom the hottest shell to the coldest shell.
Special consiaerations
Cycling, upset conditions, alternate cases oI operation, and whether operation is
continuous or intermittent should all be speciIied so that all demands made on a heat
exchanger during its expected liIetime can be taken into account Ior design.
Alternate cases oI operation have to be assessed careIully as a single case need not
represent the controlling case on all counts. Heat transIer area, shellside pressure drop, and
tubeside pressure drop are the broad controlling parameters. More oIten than not, a single
case is controlling Irom all points oI consideration. Occasionally, it may happen that one
case is controlling Ior the required heat transIer area, another case Ior shellside pressure
drop, and yet another case Ior tubeside pressure drop. In such situations, it is advisable to run
all cases to ensure satisIactory operation in each condition. In Iact, it is advised that unless
the designer is very experienced, all cases be run to ensure that no error oI judgment is
committed in identiIying the controlling case(s).
3.3 Tubeside
Tubeside calculations are much simpler than shellside calculations because on the tube-
24
side there is just one stream Ilowing through a circular conduit, as against the main
crossIlow and the leakage/bypass streams on the shellside, as we shall see later.
3.3.1 Effects of tubeside velocity
Both heat transIer coeIIicient and pressure drop vary with tubeside velocity, but the latter
much more strongly than the Iormer. A good design will make the best use oI the
allowable pressure drop as this will yield the highest heat transIer coeIIicient and thereby
the lowest heat transIer area and cheapest design.
Let us now see how tubeside velocity is increased. II the entire tubeside Iluid were to
Ilow through all the tubes (which is called a single tube pass), then it would result in a
certain velocity. Usually, this velocity is unacceptably low as it yields a very low heat
transIer coeIIicient and thereIore has to be increased. By incorporating pass partition plates
(and corresponding gaskets) in the channels, the tubeside Iluid is made to Ilow several times
through a Iraction oI the total number oI tubes at a time. Thus, a heat exchanger may have
200 tubes and 2 passes, so that the Iluid Ilows through 100 tubes at a time, and twice along
the exchanger, Iirst Irom leIt to right and then Irom right to leIt (or vice versa). The velocity
will evidently be twice that oI the velocity iI there were only one pass.
The number oI tube passes is usually 1, 2, 4, 6, 8, and so on. Evidently, a U-tube heat
exchanger can only have an even number oI tube passes. Floating-head heat exchangers
must also have an even number oI passes as otherwise the tubeside Iluid is required to leave
through the Iloating-head cover and the shell cover. This requires a special tailpipe
arrangement with bellowsan arrangement susceptible to leakage. Thus, a single tube pass
is recommended only Ior Iixed-tubesheet exchangers where it is a must; Ior example, where
true countercurrent Ilow is required or where two tube passes cause an excessive pressure
drop with the desired tube length. An odd number oI passes is usually avoided as the
tubeside outlet piping has to be taken back all along the length oI the heat exchanger. In
exceptional circumstances, however, a Iloating-head heat exchanger may have a single tube
pass.
A very large number oI tube passes leads to diIIiculties in Iabrication, especially iI the
shell diameter is small. Consequently, the maximum number oI tube passes is usually
restricted to 20, albeit Ior a very large shell diameter, say 5159 in. (1300²1500 mm). It
must be added here that in the vast majority oI cases, the number oI tube passes is not
required to be greater than 6. It will be evident that the number oI tube passes will be
required to be high when the number oI tubes is rather large and/or the tubeside Ilow rate
rather small. This situation is exacerbated when the tubeside viscosity is high, because then a
higher mass velocity is required to obtain a satisIactory Reynold`s number.
3.3.2 Heat transfer coefficient
The tubeside heat transIer coeIIicient is a Iunction oI the Reynold`s number, the Prandtl
number and the tube diameter.
Reynold`s number (Re) is DG/u, where
D ÷ tube ID
G ÷ mass velocity
ȝ ÷ viscosity
Prandtl number (Pr) is cȝ/k, where
c ÷ speciIic heat
25
ȝ ÷ viscosity
k ÷ thermal conductivity
II these are broken down into the Iundamental parameters, they are the physical
properties (namely viscosity, thermal conductivity and speciIic heat), tube diameter, and,
very importantly, mass velocity.
The variation in liquid viscosity being quite considerable, this physical property has the
most dramatic eIIect on heat transIer coeIIicient.
Let us take a look at the Iundamental equation Ior turbulent heat transIer inside tubes:
Nu ÷ 0.027 (Re)
0.8
(Pr)
0.33
(3.1)
or
(hD/k) ÷ 0.027 (DG/ȝ)
0.8
(cȝ/k)
0.33
(3.2)
or
h ÷ 0.027 (DG/ȝ)
0.8
(cȝ/k)
0.33
(k/D) (3.3)
Thus, heat transIer coeIIicient varies to the 0.8 power oI mass velocity. It also varies to
the -0.2 power oI tube ID, mass velocity remaining constant. That is to say, Ior the same
total tubeside Ilow area per pass (which will yield the same mass velocity), a smaller tube
size will yield a higher heat transIer coeIIicient.
With regards to the variation oI heat transIer coeIIicient with Iluid viscosity, it will be
seen that there are two opposing tendencies, one in which viscosity is a parameter oI the
Reynold`s number and the other in which it is a parameter oI Prandtl number. Thus,
h ÷ Į (ȝ)
0.33-0.8
÷Į (ȝ)
-0.47
(3.4)
which is to say, the heat transIer coeIIicient is inversely proportional to the 0.47 power oI
viscosity.
An interesting aspect to observe here is that, even Ior the same Reynold`s number, two
liquids oI Iairly diIIerent viscosities will yield Iairly diIIerent heat transIer coeIIicients due to
their diIIerent Prandtl numbersthe liquid having the higher viscosity will have the higher
Prandtl number and thereby the higher heat transIer coeIIicient. The Reynold`s numbers can
be the same iI the higher viscosity stream has a correspondingly higher mass Ilow rate.
It will also be seen that heat transIer coeIIicient is directly proportional to the 0.67
exponent oI thermal conductivity.
Curiously, it appears that the heat transIer coeIIicient is proportional to the 0.33
exponent oI speciIic heat. While this is true, it should also be realized that the heat duty will
increase directly with speciIic heat so that all other things remaining the same, a higher
speciIic heat will result in a considerably higher heat duty than heat transIer coeIIicient!
The above leads to some very interesting generalities oI heat transIer:
A high thermal conductivity promotes a high heat transIer coeIIicient. Thus, cooling
water has an extremely high heat transIer coeIIicient |thermal conductivity around 0.37
Btu/h It ƒF (0.55 kcal/h m °C)|, Iollowed by hydrocarbon liquids |(thermal conductivity
between 0.12 and 0.18 Btu/h It °F (0.08 and 0.12 kcal/h m °C)|, and then Iollowed by
hydrocarbon gases |(thermal conductivity between 0.03 and 0.045 Btu/h It °F (0.02 and 0.03
kcal/h m °C)|.
Hydrogen is an unusual gas as it has an exceptionally high thermal conductivity (same
26
as or even greater than that oI hydrocarbon liquids), as well as an exceptionally high speciIic
heat, hence its heat transIer coeIIicient is in the range oI hydrocarbon liquids.
Typical heat transIer coeIIicients oI various Iluids are as Iollows:
Cooling water: 1200 Btu/h It
2
°F (6000 kcal/h m
2
°C)
Hydrocarbon liquids: 50260 Btu/h It
2
°F (2501300 kcal/h m
2
°C)
Hydrocarbon gases: 1050 Btu/h It
2
°F (50250 kcal/h m
2
°C)
The rather large variation in the heat transIer coeIIicient oI hydrocarbon liquids is due to
the rather large variation in their viscosity, Irom less than 0.1 cp Ior ethylene and propylene
to greater than 500 cp, and even more Ior heavy hydrocarbon liquids, such as vacuum
residue and bitumen.
The large variation in the heat transIer coeIIicient oI hydrocarbon gases is attributable to
the large variation in operating pressure. With an increase in operating pressure, gas density
increases. As pressure drop is a) directly proportional to the square oI mass velocity and b)
inversely proportional to density; a higher mass velocity can be maintained when the density
is greater, Ior the same pressure drop. This higher mass velocity translates into a higher heat
transIer coeIIicient.
3.3.3 Pressure drop
So Iar, we have discussed the variation oI heat transIer coeIIicient with the variation oI
tube ID, mass velocity and physical properties. For a given Iluid, the mass velocity exerts
a very strong inIluence on the heat transIer coeIIicient. Whereas the tubeside heat transIer
coeIIicient varies to the 0.8 exponent oI tubeside mass velocity in turbulent Ilow,
tubeside pressure drop varies with the square oI mass velocity.
Further, the tubeside Iilm resistance represents only a Iraction oI the total resistance to
heat transIer (the others being tubeside Iouling resistance, shellside Iilm resistance, shellside
Iouling resistance, and tube wall metal resistance), so that Ior an increase in the tubeside
mass velocity, the increase in the overall heat transIer coeIIicient will be even less.
Consequently, there will be an optimum mass velocity above which it will be wasteIul to
increase mass velocity any Iurther. In other words, there is an optimum velocity Ior the
conversion oI pressure drop to heat transIer.
The variation oI tubeside heat transIer coeIIicient, overall heat transIer coeIIicient, and
pressure drop with tubeside velocity are shown diagrammatically in Fig. 3.1. These curves
demonstrate vividly how much quicker the tubeside pressure increases with tube side
velocity than does the overall
heat transIer coeIIicient.
Very high velocities also
lead to erosion. However, the
pressure drop limitation
usually becomes controlling
much beIore erosive veloci-
ties are attained. The mini-
mum recommended liquid
velocity inside tubes is 3.3
It/s (1.0 m/s) while the
maximum is 8.2²9.8 It/s
(2.5²3.0 m/s). In the case oI
highly viscous liquids, how-
Fig. 3.1 Variation in tubeside heat transIer coeIIicient, overall heat
transIer coeIIicient, and tubeside pressure drop with tubeside velocity
27
ever, it is oIten not possible to achieve a velocity oI even 3.3 It/s (1.0 m/s) unless a very high
pressure drop is permitted.
Furthermore, pressure drop is proportional to the square oI the velocity and the total
length oI travel. Thus, when the number oI passes is increased Irom 2 to 4 in order to
increase the tubeside velocity, the latter is doubled and so is the length oI travel, so that the
pressure drop increases by (2)
2
× 2, or 8 times. Or when the number oI passes is increased
Irom 4 to 6, the pressure drop increases by (1.5)
2
× 1.5 or 3.375 times. Thus, when the
number oI tube passes is increased Ior a given number oI tubes and a given tubeside Ilow
rate, the pressure drop increases to the third power oI this increase. In actual practice, this
increase is somewhat less because the Iriction Iactor reduces at the higher Reynold`s
number, so that the exponent may be considered to be approximately 2.8 instead oI 3.
It must be easily appreciated that the increase in tubeside pressure drop with the increase
oI the number oI tube passes is a rather steep change. Consequently, it oIten happens that
with two passes and a given number oI tubes, the pressure drop is much lower than the
allowable value, but with Iour passes it exceeds the allowable value. In such circumstances,
the tube diameter and/or the tube length may be varied so that the allowable pressure drop is
Iully utilized or, at least, utilized to a very large extent. This will evidently yield the highest
tubeside heat transIer coeIIicient in a given situation. However, it should be borne in mind
that the smallest tube diameter and longest tube length result in the cheapest heat exchanger,
essentially because the shell diameter is the smallest. See Case Study 3.1 later in this
chapter.
The Iollowing tube diameters are usually used in the chemical process industries: 3/8,
1/2, 5/8, 3/4, 1, 1
1
/
4
, 1
1
/
2
, and 2 in. OI these, 3/4 in. and 1 in. are the most popular. Tubes less
than 3/4 in. (19.05 mm) OD should not be used Ior Iouling services; in Iact, the use oI 1-in.
(25.4-mm) OD tubes is preIerable. The use oI small-diameter tubes, such as 1/2 in. (12.7
mm) OD, is really warranted only Ior small heat exchangers |say, heat transIer area less than
215 It
2
(20 m
2
)| because Ior large heat exchangers, the number oI tubes will be very large so
that the cost oI drilling tubesheets and baIIles may oIIset, to a large extent, the cost
advantage oI a smaller shell diameter.
Some plant operators preIer to standardize on tube diameter as well as length so that
inventory Ior re-tubing oI existing tube bundles or Iabricating new ones may be minimized.
In such a situation, it becomes more diIIicult to optimize the tubeside design oI heat
exchangers and sometimes, a rather high overdesign may have to be accepted only to satisIy
the allowable pressure drop. That is to say, the number oI tubes is increased not to
incorporate additional heat transIer area but just to reduce the velocity and thereby the
tubeside pressure drop. Evidently, this will lead to an unnecessary overdesign. Such a
condition is reIerred to as a pressure drop limiting design.
However, iI such a stream subsequently goes to another heat exchanger, the extra heat
duty that will be transIerred due to the oversizing should be taken advantage oI and only the
balance heat duty considered Ior the latter.
Consider a hot stream that is to be cooled Irom 248°F (120°C) to 176°F (80°C) by a
cold stream A, representing a heat duty oI 11.91 × 106 Btu/h (3.0 × 106 kcal/h), and Irom
176°F (80°C) to 104°F (40°C) by a cold stream B, representing a heat duty oI 11.19 × 106
Btu/h (2.82 × 106 kcal/h). The total allowable pressure drop Ior the hot stream is 9.8 psi (0.7
kg/cm2) out oI which 5.6 psi (0.4 kg/cm2) is allocated to the hot shell and 4.2 psi (0.3
kg/cm2) to the cold shell. While carrying out the design oI the hot shell, it is Iound that in
order to restrict the hot stream pressure drop to 5.6 psi (0.4 kg/cm2), there is an incidental
overdesign oI 20° which translates to a heat duty oI 15.82 ×106 Btu/h (3.24 × 106 kcal/h)
28
corresponding to an outlet temperature oI 170.2°F (76.8°C). As the balance heat duty is
(11.91 ¹ 11.19 12.86) or 10.24 × 106 Btu/h |(3.0 ¹ 2.82 3.24) or 2.58 × 106 kcal/h|, the
cold shell need be designed only Ior this heat duty and not Ior 11.19 × 106 Btu/h (2.82 × 106
kcal/h). OI course, the desired oversurIacing can be incorporated in each shell.
In this context, it is important to realize that the total pressure arop Ior a given stream
has to be met. The distribution oI pressure drop in the various heat exchangers Ior a given
stream in a particular circuit may be varied as Iound best to obtain good heat transIer in all
the heat exchangers.
For example, consider that the total allowable pressure drop Ior a distillation column
overhead stream is 4.2 psi (0.3 kg/cm
2
). This overhead stream has to Iirst Ilow through an
air-cooled condenser and subsequently Ilow through a water-cooled condenser. II a
satisIactory air-cooler design can be accomplished with a tubeside pressure drop oI only 1.7
psi (0.12 kg/cm
2
), the balance pressure drop oI (4.2 ² 1.7) or 2.5 psi |(0.3 ² 0.12) or 0.18
kg/cm
2
| is available Ior the trim condenser.
Another example is a cold liquid stream Ilowing through several preheat exchangers.
Normally, a pressure drop oI 10 psi (0.7 kg/cm
2
) is permitted Ior liquid streams per shell.
Consider that there are Iive such preheat exchangers having one shell each. Consequently, a
total pressure drop oI |5 × 10 (0.7)| or 50 psi (3.5 kg/cm
2
) is permitted Ior the entire circuit.
As the cold stream Ilows through the preheat exchangers, its temperature increases and
thereby its viscosity decreases. As the heat transIer coeIIicient decreases with increasing
viscosity Ior the same pressure drop, it will be prudent to allocate more pressure drop Ior the
colder heat exchangers where the cold stream is more viscous. In this way, even when the
cold stream is more viscous, it can be made to achieve a higher heat transIer coeIIicient than
would be possible by an even distribution oI pressure drop in all oI the Iive heat exchangers.
Table 3.1a: Heat exchanger service Ior Case Study 3.1
Shellside Tubeside
1. Fluid Hydrocarbon liquid Hydrocarbon liquid
2. Flow rate, lb/h (kg/h) 882,000 (400,000) 606,000 (275,000)
3. Temperature in/out, ƒF (ƒC) 437 (225)/482 (250) 590 (310)/536 (280)
4. Allowable pressure drop, psi (kg/cm
2
) 17 (1.2) 10 (0.7)
5. Fouling resistance, h It
2
ƒF/Btu
(h m
2
ƒC /kcal)
0.00342 (0.0007) 0.00293 (0.0006)
6. Heat duty, MM Btu/h (MM kcal/h) 21.81 (5.495)
7. Viscosity in/out, cp 0.8/0.6 0.4/0.5
8. Density in/out, lb/It
3
(kg/m
3
) 46.8 (750)/45.6 (730) 39.3 (630)/41.2 (660)
9. Thermal conductivity in/out,
Btu/h It ƒF (kcal/h m ƒC)
0.067 (0.1)/
0.065 (0.097)
0.06 (0.09)/
0.0625 (0.093)
10. SpeciIic heat, Btu/lb ƒF (kcal/kg ƒC) 0.6/0.63 0.75/0.73
11. Material oI construction Carbon steel
Tubes: SS 410
other parts: 5Cr
1
/
2
Mo
12. Line size, in. (mm) (nominal) 12 (300) 12 (300)
29
CASE STUDY 3.1: OPTIMIZING TUBESIDE DESIGN
Let us consider the heat exchanger service speciIied in Table 3.1a.
A TEMA AES (split-ring pull-through Iloating-head construction) is to be employed.
Tubes are to be either 0.984 in. (25 mm) OD or 0.7874 in. (20 mm) OD, 0.0787 in. (2 mm)
thick and 354 in. (9000 mm) long. However, the 0.984 in. (25 mm) OD will be preIerred.
The tube length can be reduced iI this is Iound to be advantageous.
A Iirst design was produced using 0.984-in. (25-mm) OD tubes and 29.5-It (9000-mm)
long tubes, as shown in Case A oI Table 3.1b. It will be noticed that the tubeside pressure
Table 3.1b: Details oI Iirst design produced Ior Case Study 3.1
Case A Case B
1. Type oI exchanger Floating-head (TEMA AES)
2. Shell ID, in. (mm) 36.4 (925) 31.5 (800)
3. Tube OD, in. (mm) × no. oI tubes
× no. oI tube passes
0.984 (25) × 510 × 2 0.7874 (20) × 560 × 2
4. Heat transIer area, It
2
(m
2
) 3798 (353) 3346 (311)
5. Tube pitch x tube layout angle 32, 90
o
26 , 90
o

6. BaIIle type and spacing, in. (mm)
single-segmental,
17.7 (450)
single-segmental,
15.75 (400 )
7. BaIIle cut, diameter 25 31
Shellside 3.8 (1.16) 4.86 (1.48)
8. Velocity, It/s
(m/s)
Tubeside 4.4 (1.34) 6.9 (2.1)
Shellside 385.7 (1883) 444.2 (2169)
Tubeside 234.5 (1145) 339.4 (1657)
9. Heat transIer
coeIIicient, Btu/h
It
2
°F (kcal /h m
2

°C)
Overall 69.4 (339) 77.3 (377.5)
Shellside 11.1 (0.78) 14.9 (1.05) 10. Pressure drop, psi
(kg/cm
2
)
Tubeside 2.6 (0.18) 7 (0.49)
Shellside 18.0 17.41
Tubeside 29.6 22.79
Fouling 47.95 54.74

11. Percent heat
transIer resistance
Metal wall 4.45 5.06
BaIIle-hole-to-tube (A) 0.177 0.149
Main crossIlow (B) 0.607 0.632
Shell-bundle (C) 0.112 0.121

12. Stream analysis:
Ilow Iractions
Shell-to-baIIle (E) 0.105 0.098
13. Overdesign 8.3 6.24
14. Approx. bundle wt., lb (kg ) 14,100 (6400) 11,700 (5300)
15. Approx. empty exchanger wt., lb (kg) 37,500 (17,000) 29,300 (13,300)
30
drop was only 2.56 psi (0.18 kg/cm2) when 10 psi (0.7 kg/cm2) was permitted. Further, the
tubeside heat transIer resistance was 29.6° oI the total, which meant that iI the allowable
pressure drop were better utilized, the heat transIer area would decrease. However, when the
number oI tube passes was increased Irom 2 to 4 (keeping the shell diameter the same and
decreasing the number oI tubes Irom 510 to 490 due to the extra pass-partition lanes), the
tubeside pressure drop increased to 17.8 psi (1.25 kg/cm2), which was unacceptable. It may
be added here that the shellside design was satisIactory with the allowable pressure drop
quite well utilized and an acceptable stream analysis.
As the overdesign in the 4-pass design was 28.1°, an attempt was made to decrease the
tube length in an eIIort to reduce the tubeside pressure drop. It was Iound that the tube length
could be reduced to 26.25 It (8000 mm) where the overdesign was 6.5°. However, the
tubeside pressure drop was 16.2 psi (1.14 kg/cm
2
), which was still much higher than that
permitted.
As the design with 0.984-in. (25-mm) OD tubes and 2 tube passes did not utilize the
allowable tubeside pressure drop, and that with 0.984-in. (25-mm) OD tubes and 4 tube
passes could not restrict the tubeside pressure drop to within the allowable value, it was
logical to attempt a design with 0.7874-in. (20-mm) OD tubes. This design is detailed as
Case B in Table 3.1b. It will be seen that there is a considerable reduction in the shell
diameter |Irom 36.4 in. (925 mm) to 31.5 in. (800 mm)| and an appreciable reduction in the
heat transIer surIace Irom 3798 It
2
(353 m
2
) to 3346 It
2
(311 m
2
).
Due to the much higher tubeside velocity |6.9 It/s (2.1 m/s) versus 4.4 It/s (1.34 m/s)
earlier)|, the tubeside pressure drop was much higher |7 psi (0.49 kg/cm
2
) compared to 2.6
psi (0.18 kg/cm
2
)|, as was the tubeside heat transIer coeIIicient |339.4 Btu/h It
2
°F (1657 as
against 234.5 Btu/h It
2
°F (1145 kcal/h m
2
°C)|. This led to an overall heat transIer
coeIIicient oI 77.3 Btu/h It
2
°F (377.5 kcal/h m
2
°C) as against 69.4 Btu/h It
2
°F (339 kcal/h
m
2
°C) in the Case A design. It will also be seen that there is an appreciable reduction in the
cost oI the heat exchanger as the bundle weight reduces Irom ~14,100 lb (~6400 kg) to
~11,700 lb (~5300 kg) and the empty weight Irom ~37,500 lb (~17,000 kg) to ~29,300 lb
(~13,300 kg).
The above example shows how better utilization oI allowable tubeside pressure drop led
to a more economical design. However, iI the use oI 0.7874-in. (20-mm) OD tubes had not
been permitted because oI the Iouling nature oI the Iluid or Ior inventory reasons, the Case A
design would have been required to be adopted.
3.3.4 Importance of stepwise calculations for viscous liquids
When the variation in tubeside viscosity is large, a single-point calculation oI the tubeside
heat transIer coeIIicient and pressure drop will give unrealistic results. This will be
particularly true in cases where the Ilow regime varies Irom turbulent to transition or
transition to laminar or where all three regimes exist, as the thermal perIormance is very
diIIerent Ior the various Ilow regimes. In such cases, it will be necessary to perIorm the
calculations stepwise or zone-wise. Modern soItware automatically employ zone-wise
calculations: the number oI zones is determined Irom the number oI tube passes and the
variation in the Reynold`s number.
To demonstrate the importance oI carrying out zone-wise calculations, we shall now
take a look at the design oI a kettle-type steam generator in a reIinery having a large
variation in the tubeside viscosity, and see the extent oI the variation oI results by
perIorming the calculations point-wise and stepwise.
31
CASE STUDY 3.2: STEPWISE CALCULATIONS
The principal process parameters are shown in Table 3.2a. It may be noticed that the
viscosity oI the hydrocarbon liquid varies Irom 3.07 cp at the inlet to 8.28 cp at the outlet.
Tubes / in. (19.05 mm) OD × 16 BWG (1.65 mm thick) × 20 It (6096 mm) long oI 5 Cr
½ Mo were to be used in a TEMA AKT construction. The tube bundle weight was not to
exceed 22,000 lb (10 tons).
A design was produced by perIorming the calculations stepwise which the soItware does
as a routine. Details oI this design are shown in Table 3.2b. It will be seen that two kettles
are used in parallel. This is because the tube bundle weight was exceeding the speciIied limit
in a single shell design. The allowable tubeside pressure drop has been Iully utilized; thus
the overdesign oI 7.76° cannot be reduced by having a smaller number oI tubes. As
expected, the tubeside resistance is highly controlling, contributing 77.66° oI the total.
Now suppose that the tubeside calculations were to be carried out at a single point,
namely the average temperature on the tubeside which is 402°F (205.6°C). It can be seen
Irom the tubeside monitor that the tubeside heat transIer coeIIicient at this value oI tubeside
bulk temperature is approx. 55.6 Btu/h It
2
°F (271.5 kcal/h m
2
°C) and the overall heat
transIer coeIIicient is approx. 41.77 Btu/h It
2
°F (203.9 kcal/h m
2
°C). Note this value is
signiIicantly higher than the value determined by the soItware on a zone-wise basis, namely
34.95 Btu/h It
2
°F (170.6 kcal/h m
2
°C). The ratio oI the two values is 41.77/34.95 or 1.195.
Thus, iI the thermal design calculations were perIormed on a single-point (the mean tubeside
bulk temperature), then the design would have been very conservative.
3.4 Shellside
Why the shellsiae is more complex
Consider the various parameters on the shellside:
• type oI shell
• tube layout pattern
Table 3.2a: Principal process parameters Ior Case Study 3.2
Shellside Tubeside
1. Fluid BFW/ Steam Hydrocarbon liquid
2. Flow rate, lb/h (kg/h)
16,200 (7350) (Iully
vaporized)
240,000 (108,900)
3. Temperature in/out, °F (°C) 309.2 (154)/309.2 (154) 450 (232.2)/354 (178.9)
4. Allowable pressure drop, psi (kg/cm
2
) Neg. 30.0 (2.1)
5. Fouling resistance, h It
2
°F/Btu
(h m
2
°C/kcal)
0.001 (0.000205) 0.003 (0.000614)
6. Viscosity in/out, cp 3.07/8.28
7. Thermal conductivity, Btu/h It °F (kcal/h m
°C)
0.0575 (0.086)/0.06 (0.09)
8. Density in/out, lb/It
3
(kg/m
3
) 49.87 (799)/51.5 (825)
9. SpeciIic heat, Btu/lb °F (kcal/kg °C)


Standard
0.642/0.623
10. Heat duty, MM Btu/h (MM kcal/h) 14.58 (3.674)
32
• tube pitch
• type oI baIIles
• baIIle spacing and baIIle cut
Consider also that, on the shellside, there is not just one stream, as within tubes, but one
principal crossIlow stream and Iour leakage/bypass streams. As we shall discuss in
Section 3.4.5, the Ilow Iractions oI each oI these Iive streams will have to be determined,
based upon equal pressure drop (since they are in parallel) and then the heat transIer
calculations perIormed. Besides, the presence oI Iive streams leads to a temperature
proIile distortion problem with an attendant penalty in the MTD. This will be discussed
in detail in Section 4.7. It now becomes evident why shellside design is much more
diIIicult than tubeside design.
Let us now consider the above parameters in more detail.
3.4.1 Shell type
Depending upon how a Iluid Ilows through a shell, TEMA deIines various shell patterns:
E, J, F, G, H, K, and X (Fig. 2.6).
A TEMA E shell is a single-pass shell, where the shellside Iluid enters at one end oI the
shell and leaves at the other end. This is by Iar the most common shell type that one will
come across in the chemical process industries.
A TEMA F shell (Figs. 2.5a and 2.5b) is a 2-pass shell where a longitudinal baIIle
Table 3.2b: Thermal design oI steam generator
1. Type oI exchanger TEMA AKT (kettle)
2. No. oI shells 2 (in parallel)
3. Port/kettle ID, in. (mm) 32 (813)/52 (1320)
4. Tube OD × tube thickness × tube length, in. (mm) 0.75 (19.05) × 16 BWG (1.65) × 240 (6000)
5. No. oI tubes × no. oI tube passes 680 x 10
6. Tube pitch × tube layout angle 1.0 (25.4) × 90o
7. BaIIling Full support plates only
8. Heat transIer area, It
2
(m
2
) 2 x 2595 (241.2) ÷ 5190 (482.4)
Velocity, It/s (m/s) 4.58 (1.4)
9. Tubeside
Pressure drop, psi (kg/cm
2
) 30 (2.1)
Shellside 835 (4077)
Tubeside 45 (219.7)
10. Heat transIer coeIIicient,
Btu/h It
2
°F (kcal /h m
2

°C)
Overall 34.95 (170.6)
Shellside 4.19
Tubeside 77.66
Fouling 16.18


11. Percent heat transIer
resistance
Metal wall 1.97
12. Overdesign 7.76
33
divides the shell into two passes, an upper pass and a lower pass. The shellside Iluid enters
the shell at one end in either the upper halI or the lower halI, traverses the entire length oI the
shell through one-halI the shell cross-sectional area and then turns around and Ilows through
the second pass, beIore Iinally leaving at the end oI the second pass. Evidently, the
longitudinal baIIle does not extend to the tubesheet at the Iar end but stops well short, so that
the shellside Iluid can Ilow into the second pass. This construction is used Ior temperature
cross situations, that is, where the cold Iluid leaves at a temperature higher than the outlet
temperature oI the hot stream. II a 2-pass F shell has only 2 tube passes, this becomes a true
countercurrent arrangement where a large temperature cross can be realized.
A TEMA J shell (Fig. 2.6) is a divided-Ilow shell wherein the shellside Iluid enters the
shell at the center (along the length) and divides into two halves, one Ilowing to the leIt and
the other to the right and leaving separately. They are then combined into a single stream.
This is identiIied as a J1-2 shell. Alternatively, the stream may be split into two halves and
enter the shell at the two ends, Ilow towards the center, and leave as a single stream. This is
identiIied as a J2-1 shell. Whether to use a J1-2 shell or a J2-1 shell is determined by the
convenience oI nozzle design. II the stream is a sensible liquid or gas, it does not make a big
diIIerence. However, iI the entering stream is a vapor which then condenses, the inlet line
will be much larger than the outlet line. It will be better in this case to employ a J2-1 shell as
the larger line will be divided into two smaller lines which will then enter the shell. Using a
J1-2 shell would mean a single large inlet nozzle and two much smaller outlet nozzles. For
example, iI the inlet line is 16 in. (400 mm) and the outlet 4 in. (100 mm), a J2-1 shell will
mean two 12-in. (300-mm) nozzles at the inlet and a 4-in. (100-mm) nozzle at the outlet. II a
J1-2 conIiguration were used here, it would have translated into a 16-in. (400-mm) inlet
nozzle and two 3-in. (75-mm) outlet nozzles. The two smaller inlet nozzles (12 in. or 300
mm) will result in the accommodation oI a larger number oI tubes in a given shell diameter
than the larger inlet nozzle (16in. or 400 mm). This will be particularly true iI the inlet line
size is quite large compared to the shell diameter. II the shell diameter is very large (say,
51.2 in. or 1300 mm) in the example just described, it will not make an appreciable
diIIerence whether the inlet nozzle is 16 in. (400 mm) or 12 in. (300 mm). However, iI the
shell diameter is only 31.5 in. (800 mm), it will make an appreciable diIIerence. The reverse
will be true in the case oI a vaporizer having a small inlet line size and a much larger outlet
line size, that is, it will be advantageous to employ a J1-2 shell as compared to a J2-1 shell.
A TEMA G shell is a split-Ilow shell with the conIiguration shown in Fig. 2.6. This
construction is usually employed Ior horizontal thermosyphon reboilers. There is only one
central support plate and no baIIles. Evidently, as TEMA speciIies a maximum unsupported
tube length oI about 60 in. (1500 mm), a G shell cannot be used Ior heat exchangers having
a tube length greater than 120 in. (3000 mm) as the unsupported length would then exceed
the TEMA limit oI 60 in. or 1500 mm (typically, varies with tube OD, thickness, and
material).
When a larger tube length has to be employed, a TEMA H shell is adopted (Fig. 2.6).
An H shell is really two G shells placed side-by-side, so that now there are two Iull support
plates. The description Ior this conIiguration is 'double-split¨ as the Ilow is split twice and
recombined twice. This construction is also invariably employed Ior horizontal thermo-
syphon reboilers. The advantage with TEMA G and H shells is that the pressure drop is
reduced drastically. Since there are no pumps in thermosyphon reboiler circuits, the pressure
drop has to be restricted to a bare minimum; hence, these conIigurations are employed.
Besides, and no less importantly, the natural tendency oI the two phases to separate is
minimized.
34
A TEMA X shell (Fig. 2.6) is a pure crossIlow shell where the shellside Iluid enters at
the top (or bottom) oI the shell and Ilows across the tubes and exits Irom the opposite side oI
the shell. The Ilow may be introduced through multiple nozzles located strategically along
the length oI the shell in order to achieve a better distribution. As pressure drop will be
extremely low (in Iact, there is hardly any pressure drop in the shell and whatever pressure
drop there is, is virtually all in the nozzles), this conIiguration is employed Ior cooling or
condensing vapors at very low pressure, particularly vacuum. Full support plates can be
located as required Ior structural integrity; they do not interIere with the shellside Ilow as
they are parallel to the Ilow direction.
A TEMA K shell (Fig. 2.6) is a special crossIlow shell employed Ior kettle reboilers (K
Ior kettle) having an integral vapor disengagement space in the shape oI an enlarged shell.
Here too, Iull support plates can be employed as required.
3.4.2 Tube layout pattern
There are Iour tube layout patterns (Fig. 3.2):
• triangular (30°)
• rotated triangular (60°)
• square (90°)
• rotated square (45°)
A triangular (or rotated) pattern will evidently accommodate more tubes than a square (or
rotated square) pattern. Further, a triangular pattern produces high turbulence and,
thereIore, a high heat transIer coeIIicient. However, a triangular (or rotated triangular)
pattern does not permit mechanical cleaning oI tubes Ior the normally used tube pitch
(1.25 times the tube OD) as access lanes Ior cleaning are not available. Consequently, a
triangular layout pattern is limited in use to clean services on the shellside. For services
which require mechanical cleaning on the shellside, a square (or rotated square) pitch has
to be used.
It should be noted here that chemical cleaning does not require access lanes so that a
triangular pitch may be used Ior a dirty shellside service provided chemical cleaning is
suitable and eIIective. II the tube pitch is increased suIIiciently, access lanes will become
available; but as the shell ID will increase substantially and well beyond that Ior a square
pitch, the very purpose oI accommodating more tubes will be deIeated.
A rotated triangu-
lar pattern does not
oIIer any advantage
over a triangular pat-
tern in the conversion
oI pressure drop to
heat transIer and,
hence, its use is rare.
For dirty services
on the shellside, the
usual practice is to use
a square layout pattern.
However, since this is
an in-line pattern, it
produces lower turbu-
Fig. 3.2 Tube layout pattern (Reprinted Irom the Heat Exchanger Design
Handbook, 2002 with permission oI Begell House, Inc.)
35
lence. Consequently, when the shellside Reynold`s number is low (· 2000), it is usually
advantageous to employ a rotated square pattern as this produces a much higher turbulence
which results in a higher eIIiciency oI conversion oI pressure drop to heat transIer.
It has been seen earlier that Iixed-tubesheet construction is usually employed Ior clean
services on the shellside, U-tube construction Ior clean services on the tubeside, and
Iloating-head construction Ior dirty services on both the shellside and tubeside. (For clean
services on both shellside and tubeside, either Iixed-tubesheet or U-tube construction may be
used, although U-tube is preIerable as it permits diIIerential expansion between the shell and
the tubes.) It thereIore Iollows that a triangular tube layout pattern may be used Ior Iixed-
tubesheet exchangers and a square (or rotated square) pattern Ior Iloating-head exchangers.
For U-tube exchangers, a triangular pattern may be used provided the shellside stream is
clean, and a square (or rotated square) pattern iI the same is dirty.
3.4.3 Tube pitch
This is deIined as the shortest distance between two adjacent tubes. For a triangular
pattern, TEMA speciIies a minimum tube pitch oI 1.25 times the tube OD. Thus, a 0.984-
in. (25-mm) tube pitch is usually employed Ior 0.7874-in. (20-mm) OD tubes. However,
in the case oI a welded tube-to-tubesheet joint, the minimum weld ligament must be
provided, usually
1
/
4
in. (6 mm) on a diametral basis.
For square (or rotated square) patterns, TEMA additionally recommends a minimum
cleaning lane oI
1
/
4
in. (6 mm) between adjacent tubes. Thus, the minimum tube pitch here is
|tube OD ¹
1
/
4
in. (6 mm)|. A typical example is 20-mm tubes laid on a 26-mm square pitch.
In the case oI 25-mm OD tubes, however, 31-mm tube pitch will not suIIice because,
although it provides a 6 mm cleaning lane, the ratio oI tube pitch to tube OD is less than
1.25. Consequently, 25-mm OD tubes are laid out on a square (or rotated square) pitch oI
31.25 mm or 32 mm. For
3
/
4
-in. OD tubes, the minimum tube pitch is 1 in., and Ior 1-in. OD
tubes, it is 1
1
/
4
in.
Designers preIer to employ the minimum recommended tube pitch as it leads to the
smallest shell diameter Ior a given number oI tubes. However, in exceptional circumstances,
the tube pitch may be increased to a higher value in order to, Ior example, reduce shellside
pressure drop. This is particularly true in the case oI a crossIlow shell where all other
mechanisms oI reducing the shellside pressure drop are exhausted.
3.4.4 Baffling
This is the most proIound and important aspect oI STHE design. The selection oI an
optimum baIIle design, along with the selection oI the best shell type, goes a long way in
ensuring eIIicient design and operation oI STHEs. As this is the most diIIicult aspect oI
STHE design, it is also the area where most design deIiciencies result. Thus, a good
understanding oI the complexities oI shellside design is essential Ior optimum design oI
STHEs.
Types of baffles
BaIIles are used to: a) support tubes, b) enable a desirable velocity to be produced Ior the
shellside Iluid, and c) prevent Iailure oI tubes due to Ilow-induced vibration. There are
two principal types oI baIIles: plate type and rod type. Plate baIIles may be single-
segmental, double-segmental, or triple-segmental. The various types oI baIIles are shown
in Fig. 3.3. The designer has to select the most appropriate type oI baIIles Ior the applica-
36
tion at hand. The various types oI plate baIIles and their selection are discussed in detail
later in Section 3.4.6.
Baffle spacing
BaIIle spacing is the centerline to centerline distance between adjacent baIIles. It is one
oI the most vital parameters in shell-and-tube heat exchanger design as it exerts a very
strong inIluence on the shellside thermal and hydraulic perIormance.
Minimum baffle spacing
As propounded in the standards oI TEMA, the minimum baIIle spacing is one-IiIth oI the
shell inside diameter or 4 in. (100 mm), whichever is lower. Even lower baIIle spacing
will cause
a) Poor bundle penetration by the shellside Iluid. With a reduction in baIIle spacing,
the crossIlow area decreases, thereby increasing the resistance to Ilow in this
path. Consequently, a larger Iraction oI the shellside Iluid will bypass around the
tube bundle, along the shell-wall, and through the pass-partition lanes and baIIle
hole-tube clearances, thereby reducing bundle penetration. This will be discussed
in detail in Section 3.4.5.
b) DiIIiculty oI mechanical cleaning
oI deposits on the outside surIace oI
the tubes.
Maximum Baffle Spacing
The maximum baIIle spacing is
usually the shell ID. Higher baIIle
spacing will lead to
a) Predominantly longitudinal Ilow
which is less eIIicient than crossIlow
and
b) Large unsupported tube spans
which will result in the exchanger
being prone to Iailure oI tubes by
Ilow-induced vibration.
An exception is a no-tubes-in-window
construction or a crossIlow shell (de-
scribed later in Section 3.4.6) where
the Ilow is in pure crossIlow. In Iact,
in such situations, the baIIle spacing
will necessarily be very high as the
very reason Ior their use is the reduc-
tion oI shellside pressure drop. There
is only crossIlow in these applications
and intermediate supports can be
easily incorporated, iI required, to
reduce the unsupported tube span so
that there is no possibility oI Iailure oI
tubes due to Ilow-induced vibration.
Fig. 3.3 Types oI baIIles |2| (Reproduced with permission.
Copyright 1992 AIChE. All rights reserved.)
37
Optimum baffle spacing
For turbulent Ilow on the shellside (Re ~ 1000), the heat transIer coeIIicient varies to the
0.6²0.7 power oI velocity; however, pressure drop varies to the 1.7²2.0 power. For
laminar Ilow (Re · 100), the exponents are 0.33 Ior heat transIer coeIIicient and 1.0 Ior
pressure drop. It is thereIore apparent that as baIIle spacing is decreased, the pressure
drop increases at a much Iaster rate than the heat transIer coeIIicient does. This means
that there will be an optimum baIIle spacing-to-shell inside diameter ratio which will give
the highest eIIiciency oI conversion oI pressure drop to heat transIer. This ratio is
normally between 0.3 and 0.6.
Using a very low baIIle spacing tends to increase the leakage and bypass streams. It is
easy to understand why this happens. All the Iive shellside streams are in parallel and
thereIore will have the same pressure drop. The leakage path dimensions (tube-to-baIIle hole
clearance, baIIle-to-shell clearance, pass-partition width, and shell ID to outer tube limit
clearance) are Iixed. Consequently, as baIIle spacing is decreased, the resistance oI the main
crossIlow path and, thereby, its pressure drop increases. As the pressure drop oI all the Iive
streams have to be equal, the leakage and bypass streams increase until the pressure drop oI
all the streams balance out. The net result is an increase in the pressure drop without a
corresponaing increase in the heat transIer coeIIicient, due to the Iact that the leakage and
bypass streams are much less eIIicient than the main crossIlow stream.
Baffle cut
This is illustrated in Fig. 3.4 and is expressed as the percentage oI the segment height to
the shell inside diameter. Although this is also an important parameter Ior shell-and-tube
heat exchanger design, its eIIect is less proIound than that oI baIIle spacing.
BaIIle cut can vary between 15° (or even somewhat less Ior large diameter shells) and
45° oI the shell inside diameter. The upper limit is due to the Iact that iI it were greater than
45, the central row oI tubes would not be supported by any baIIle.
How small ana large baffle cuts are inefficient
Both very small and very large baIIle cuts are detrimental to eIIicient heat transIer on the
shellside. When the baIIle cut is very small, the principal portion oI the Ilow acts like a
jet through the window and then Iollows an S-shaped pattern across the tube bundle,
thereby generating large eddies oI recirculating Iluid in the regions near the baIIle tips.
This is shown in Fig. 3.5a.
When the baIIle cut is very large, the major part oI the shellside stream bypasses the
greater part oI the bundle and Ilows between the baIIle tips, predominantly in longitudinal
Ilow. Large eddies oI recirculating Iluid are created which are ineIIicient Ior heat transIer.
This is depicted in Fig. 3.5b.
Both very small and very large baIIle cuts result in
highly nonideal Ilow and this is represented by a penalty
Iactor which may be termed nonideal window Ilow
penalty Iactor.
The ideal Ilow pattern on the shellside is crossIlow.
However, baIIles are essential Ior supporting the tubes
and increasing the shellside velocity and result in al-
tering the ideal crossIlow pattern. ThereIore, an
appropriate correction Iactor has to be employed to the
ideal tube bundle heat transIer coeIIicient. This Iactor Fig. 3.4 BaIIle cut
38
may become quite severe Ior very small and very large baIIle cuts. Only in the case oI a
well-proportioned baIIle spacing-to-baIIle cut geometry does the Ilow pattern approach that
oI an ideal tube bank as shown in Fig. 3.5c.
Recommenaea baffle cut
It is strongly recommended that only baIIle cuts between 20° and 35° be employed.
Reducing baIIle cut below 20° to increase the shellside heat transIer coeIIicient or
increasing the baIIle cut beyond 35° to decrease the shellside pressure drop usually lead
to a poor design. Other changes in tube bundle geometry should be examined to achieve
the above goals. For example, double-segmental baIIles or a divided-Ilow or even a
crossIlow shell may be used to reduce the shellside pressure drop. This will be discussed
in detail in Section 3.4.6.
Orientation of baffle cut
BaIIle cut may be either horizontal or vertical (see Fig. 3.6). For single-phase Iluids on
the shellside, a horizontal baIIle cut is recommended as this minimizes accumulation oI
deposits at the bottom oI the shell and also prevents stratiIication. However, in the case oI
a 2-pass shell (TEMA shell style F), a vertical cut is preIerred Ior ease oI Iabrication and
bundle assembly.
For condensing on the shellside, a vertical cut is invariably employed. This is discussed
in detail in Chapter 7.
Keep crossflow ana winaow velocities equal
Flow across tubes is reIerred to as crossIlow, whereas Ilow through the window area is
reIerred to as window Ilow (see Fig. 3.7the tubes have not been shown to enhance
clarity) or longitudinal Ilow, as Ilow is predominantly along the tubes. Velocities in the
crossIlow and window regions are reIerred to as crossIlow and window velocities,
respectively. The crossIlow velocity and the window velocity should be as close as
possible and preIerably within 20° oI each other. With greater ratios, repeated
acceleration and deceleration along
the length oI the tube bundle results in
ineIIicient conversion oI pressure drop
to heat transIer. This results in the
nonideal window Ilow penalty Iactor.
A large baIIle cut is worse than a
small baIIle cut. Hence, the baIIle cut
should never exceed 35 ° Ior a
sensible Iluid on the shellside.

Fig. 3.5 EIIect oI small and large baIIle cuts (Reprinted Irom the Heat Exchanger Design Handbook, 2002
with permission oI Begell House, Inc.)
Fig. 3.6 Horizontal and vertical baIIle cut
39
It is worthwhile to mention here
that the guidelines and rules oI thumb
suggested in this and the preceding
sections are just thatguidelines. The
designer should vary the baIIle spacing
and cut, and select the best or most
optimum values. With high-speed
desktop and even laptop computers,
this optimization not only takes very
little time, but it also makes the
designer`s work creative and, thereby,
enjoyable.
3.4.5 Stream analysis
Tubeside calculations Ior heat transIer coeIIicient and pressure drop are Iairly straight-
Iorward. Single-point calculations based upon average physical properties are usually
adequate. However, when the Ilow pattern extends Irom laminar to transition, and even
beyond, in a given heat exchanger, zone-wise calculations must be carried out because
heat transIer coeIIicient and pressure drop characteristics are very diIIerent Ior the two
regimes.
On the shellside, however, the situation is much more complex as there is not just one
stream but a main crossIlow stream and Iour leakage/bypass streams. This is illustrated in
Fig. 3.8a. As per the original model proposed by Tinker |1|, the various streams are as
Iollows:
• B stream ² the main crossIlow stream
• A stream ² baIIle hole-tube leakage stream
• C stream ² bundle bypass stream
• F stream ² pass-partition lane bypass stream
• E stream ² shell-baIIle leakage stream
It is a matter oI conjecture as to why the letter D had not been included by Dr. Tinker in
his nomenclature! The A and E streams are shown more vividly in Fig. 3.8b.
It will be easily appreciated Irom Fig. 3.8a that the B stream is highly eIIective Ior heat
transIer. Thus, the C stream is in contact only with the peripheral tubes around the bundle
and the F stream is in contact with the tubes only along both sides oI the pass-partition lanes.
For this reason, and also because their velocities are much lower due to the much wider Ilow
passages, these streams are Iar less eIIicient than the B stream. In U-tube heat exchangers,
the central U-bend
lane is rather large
(usually twice the
tube diameter) and,
consequently, the F
stream is even less
eIIicient Ior such
heat exchangers.
As regards the
A stream, it may
appear to be quite
Fig. 3.7 CrossIlow and window Ilow |2| (Reproduced with
permission. Copyright 1992 AIChE. All rights reserved.)
Fig. 3.8a Shellside Ilow distribution (Courtesy oI HTRI.)
40
eIIicient as the shellside Iluid is actually in contact with the tubes. However, between
separating Irom the main crossIlow stream in one baIIle space and rejoining it in the next,
the A stream does not travel any appreciable distance (tube length), so that it is not very
eIIicient.
Although the A, C, and F streams are less eIIicient compared to the B stream, they do
participate in the process oI heat transIer and contribute to the overall heat duty. However,
the E stream is by Iar the worst stream as it Ilows along the shell wall where there are no
tubes. It is the bane oI heat exchanger designers as it does not contribute to the process oI
heat transIer between the two streams and represents a total waste. UnIortunately, despite the
best Iabrication practice, a clearance between the shell and the baIIles is unavoidable. This
clearance is quite substantial Ior large-diameter Iloating-head heat exchangers. For a given
type oI heat exchanger, the larger the shell diameter, the larger is this gap. The variation oI
this clearance with shell diameter is given in Table RCB 4.3 oI the TEMA standards and
was reproduced in Section 2.1 Ior ready reIerence.
The Iractions oI these Iive Ilow streams can be determined Ior a given situation oI
exchanger geometry and shellside Ilow conditions by any specialized heat exchanger
thermal design soItware. Essentially, the Iive streams are in parallel and Ilow along paths oI
varying hydraulic resistances. Thus, these Iractions will be such that the pressure drop oI
each will be identical, since all the streams commence and terminate at the inlet and outlet
nozzles, respectively. That is to say, the stream Iractions are determined Irom an iteration oI
the pressure drop calculations. Subsequently, based upon the eIIiciency oI each oI these
streams, the overall shellside stream eIIiciency and thus the shellside heat transIer coeIIicient
is established.
It has already been mentioned that a high C stream is detrimental to good shellside heat
transIer. It has also been mentioned that while the shell-tube bundle clearance is relatively
small in Iixed-tubesheet and U-tube heat exchangers, it is quite substantial in split-ring pull-
through Iloating-head heat exchangers by virtue oI the very construction. In order to
minimize this C stream where it is high, 'sealing strips¨ should be located along the tube
rows, one pair Ior every 6-8 tube rows. It is recommended that sealing strips always be
incorporated in Iloating-head heat exchangers. For Iixed-tubesheet and U-tube heat
exchangers, however, such sealing strips are usually not necessary as the clearance between
the shell and the outer tube limit is Iar less.
As Ior the large U-bend pass-partition lane, dummy tubes, seal rods, or sealing strips
may be incorporated where this lane is parallel to the Ilow direction. For example, iI a U-
tube heat exchanger has Iour tube passes, there will be one vertical lane and one horizontal
lane. Since U-bends are usually are disposed in the vertical plane Ior better drainage oI the
tubeside Iluid upon shutdown, the large lane will be in the horizontal plane. II the baIIle cut
is vertical, this large lane will be parallel to the Ilow direction. Most sophisticated heat
exchanger thermal design soItware
packages can incorporate such
sealing strips and dummy tubes and
carry out stream analysis
accordingly.
It can be easily understood that
as the Ilow Iractions depend
strongly upon the path resistances,
variation oI any oI the Iollowing
construction parameters will aIIect
Fig. 3.8b BaIIle-to-shell and tube-to-baIIle hole leakage streams
41
the stream analysis (Ior a given shell type and baIIle type) and thereby the shellside
perIormance oI a heat exchanger handling a particular service:
1) baIIle spacing and baIIle cut
2) tube layout angle and tube pitch
3) number oI pass lanes in the Ilow direction as well as lane width
4) baIIle hole-tube OD clearance
5) shell ID-baIIle OD clearance
6) location oI sealing strips and sealing rods
The shellside Iluid viscosity also aIIects stream analysis proIoundly as the shellside
Reynold`s number has a proIound eIIect on the hydraulics oI shellside Ilow. It will be
seen later how viscous liquids on the shellside tend to have a rather high E stream and
how they are handled ingeniously.
Besides aIIecting the shellside heat transIer and pressure drop perIormance, the stream
analysis also aIIects the mean temperature diIIerence oI a STHE, oIten with adverse
consequences. This will be explained later in Chapter 4.
Many design considerations and recommendations have been expounded in detail
above, the principal oI these being stream analysis, temperature proIile distortion Iactor and
baIIle design. Two case studies will now be presented to bring these out vividly. The Iirst
case study will demonstrate how the temperature proIile distortion correction Iactor varies
with baIIle spacing, and the second will demonstrate how to optimize baIIle design Ior a heat
exchanger without a temperature proIile distortion Iactor problem.
CASE STUDY 3.3: VARIATION OF TEMPERATURE PROFILE
DISTORTION FACTOR WITH BAFFLE SPACING
Let us consider a light hydrocarbon cooler in an oil reIinery complex. The principal
process parameters are indicated in Table 3.3a: 0.7874 in. (20 mm) OD × (2.0 mm) thick
× 19.68 It (6000 mm) long tubes oI Admiralty brass and a Iloating-head construction
Table 3.3a: Principal process parameters Ior Case Study 3.3
Shellside Tubeside
1. Fluid Hydrocarbon liquid Cooling water
2. Flow rate, lb/h (kg/h) 110,230 (50,000) 298,060(135,200)
3. Inlet/outlet temperature, °F (°C) 194 (90)/113 (45) 95 (35)/109.4 (43)
4. Heat duty, MM Btu/h (MM kcal/h) 4.29 (1.08)
5. Density in/out, lb/It
3
(kg/m
3
) 56.4 (904)/56.85 (911)
6. Viscosity in/out, cp 10/29
7. SpeciIic heat in/out, Btu/lb °F (kcal/kg °C) 0.49/0.47
8. Thermal conductivity in/out,
Btu/h It °F (kcal/h m °C)
0.067 (0.1)/0.0706 (0.105)


Standard
9. Allowable pressure drop, psi (kg/cm2) 10 (0.7) 10 (0.7)
10. Fouling resistance, h It
2
°F/Btu (h m
2
°C/kcal)
0.00293(0.0006) 0.00195 (0.0004)
11. Material oI construction CS Adm. Brass
42
(TEMA type AES) are to be used.
The construction parameters (save baIIle spacing) that emerged Irom the design are
indicated in Table 3.3b. Note that the baIIle cut is 20° on diameter. We will now see the
eIIect oI varying the baIIle spacing on the perIormance oI the cooler.
Let us Iirst consider a baIIle spacing oI 13.8 in. (350 mm), which is about 34° oI the
shell ID oI 40.35 in. (1025 mm), a good starting point. The results oI the exchanger
perIormance are shown in Table 3.3c. It will be seen that, due to the low shellside Ilow rate,
the shellside cross and window velocities are rather low so that permissible shellside
pressure drop has not been properly utilized; as a consequence, the shellside heat transIer
coeIIicient is quite low as well.
The stream analysis is not very distinguished: although the main crossIlow Iraction is
passable at 0.533, the baIIle-shell leakage Iraction is quite high at 0.288. As a result oI this
and a moderately high ratio oI the shellside temperature diIIerence to the approach at the
shell outlet (4.5), the temperature proIile distortion penalty Iactor is 0.904. This represents a
direct loss in the MTD.
In an attempt to utilize the allowable shellside pressure drop better, and thereby increase
the shellside heat transIer coeIIicient, the baIIle spacing was decreased progressively Irom
13.8 in. (350 mm) to 11.8 in. (300 mm), 9.8 in. (250 mm), and 7.9 in. (200 mm). The results
oI these runs are also shown in Table 3.3c. It will be seen that with the increase in the
crossIlow velocity, there is an increase in the shellside heat transIer coeIIicient and the
shellside pressure drop. It will also be seen that there is a signiIicant deterioration in the
stream analysis with an increase in the baIIle-shell leakage Iraction Irom 0.288 to 0.398, and
a decrease in the main crossIlow Iraction Irom 0.533 to 0.417. As a direct consequence, the
temperature proIile distortion penalty Iactor reduces steadily Irom 0.904 to 0.8488.
Thus, with the progressive reduction in the baIIle spacing, there are two Iorces at play
here: (a) an increase in the shellside crossIlow velocity and, thereby an increase in the
shellside and, thereby, the overall heat transIer coeIIicient and (b) a reduction in the
temperature proIile distortion penalty Iactor. Whereas the Iormer tends to reduce the heat
transIer area, the latter tends to increase it. Depending upon the rates oI these two variations,
the overdesign Iirst decreases Irom 4.34° to 2.82° and then increases to 4.06°, and Iinally
5.38°.
Table 3.3b: Principal construction parameters Ior Case Study 3.3
Construction Parameter
1. Type oI exchanger Floating-head (TEMA AES)
2. Shell ID, in. (mm) 40.35 (1025)
3. Tubes
982 nos., 0.7874 in. (20 mm) OD × 0.0787 in. (2.0 mm) thick ×
19.68 It (6000 mm) long
4. No. oI tube passes 6
5. Tube pitch, in. (mm) 1.024 (26), rotated square (45 deg.)
6. BaIIling Single-segmental, 20 ° cut (diameter)
7. Nominal connection size, in. (mm) 4 (100) shellside; 6 (150) on tubeside
8. Heat transIer area, It
2
(m
2
) 3916 (363.9)
9. Empty exchanger weight, lb (kg) 26,300 (11,950)
43
Although the design with 7.9-in. (200-mm) baIIle spacing produces the highest
overdesign, its stream analysis is really quite unsatisIactory with a very low main crossIlow
Iraction (0.417) and a rather high baIIle-shell leakage Iraction. Such a poor stream analysis
tends to produce much heavier Iouling on the shellside. Considering that the Iouling
resistance oI the present shellside stream is Iairly high (0.00293 h It
2
ƒF/Btu or 0.0006 h m
2

ƒC/kcal), it may be prudent to aim Ior a better stream analysis. Accordingly, the design with
the 11.8 in. (300 mm) or even 13.8-in. (350-mm) baIIle spacing may be chosen as the Iinal
design. The overdesign is only marginally lower, whereas the stream analysis is Iar better.
It must be added here that as the shellside pressure drop utilization is rather poor in these
single shell designs, they are not elegant solutions and it may be much better to consider a
two-shells-in-series design, both in terms oI lower Iirst cost and lower operating cost due to
a signiIicantly higher shellside velocity and a better stream analysis. This will be examined
later in Case Study 4.1 in Chapter 4.
CASE STUDY 3.4: OPTIMIZING BAFFLE DESIGN
Consider the heat exchanger service speciIied in Table 3.4a, along with all the important
construction parameters except baIIle spacing and baIIle cut. For the sake oI demonstrat-
ing how to optimize the baIIle design, consider a Iirst design (Design A) in Table 3.4b.
Here, the baIIle spacing is 300 mm and the baIIle cut 25°. Since there are two indepen-
dent variables, baIIle spacing and baIIle cut, let us Iirst retain the baIIle cut as 25° and
vary the baIIle spacing. Subsequently, the baIIle spacing will be kept constant and the
Table 3.3c: Detailed results oI the various iterations Ior Case Study 3.3
Run 1 Run 2 Run 3 Run 4
1. BaIIle spacing, in. (mm) 13.8 (350) 11.8 (300) 9.8 (250) 7.9 (200)
baIIle hole-to-tube 0.031 0.049 0.058 0.07
main crossIlow 0.533 0.49 0.458 0.417
bundle-shell 0.089 0.081 0.076 0.069
baIIle-shell 0.288 0.326 0.358 0.398


2. Stream
Iractions
pass partition 0.059 0.054 0.051 0.046
3. Temperature proIile
distortion correction Iactor
0.904 0.887 0.8705 0.8488
velocity
cross/window
It/s (m/s)
0.43
(0.13)/0.62
(0.19 )
0.49
(0.15)/0.62
(0.19)
0.56
(0.17)/0.62
(0.19)
0.72
(0.22)/0.62
(0.19)


4. Shellside
pressure drop,
psi (kg/cm
2)

2.05 (0.144) 2.16 (0.152) 2.45 (0.172) 3.0 (0.21)
5. Shellside heat transIer
coeIIicient, Btu/h It
2
°F (kcal/h
m
2
°C)

43.5 (212.4)

43.7 (213.4)

45.6 (222.6)

48.0 (234.3)
6. Overall heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
33.5 (163.8) 33.7 (164.4) 34.8 (169.8) 36.2 (176.6)
7. MTD, °F (°C) 34.0 (18.9) 33.5 (18.6) 32.8 (18.2) 31.9 (17.7)
8. Overdesign 4.34 2.82 4.06 5.38
44
baIIle cut varied. (In real practice, both parameters should be varied simultaneously, but
by keeping one parameter constant and varying the other, it will be easier to appreciate
the inIluence oI each parameter).
Accordingly, in Designs B and C, the baIIle spacing was changed to 13.8 in. (350 mm)
and 15.75 in. (400 mm), respectively. Now consider the three designs in Table 3.4b. There is
no temperature proIile distortion Iactor problem here as the shellside temperature diIIerence
is rather low |(440.6 ² 410 ÷ 30.6ƒF) or (227 ² 210 ÷ 17ƒC)| and the approach temperature
at the shell outlet rather high |(608 ² 440.6 ÷ 167.4ƒF) or (320 ² 227 ÷ 93ƒC)| so that the
ratio oI the two is (30.6/167.4) or (17/93) or 0.1828, which is very low. Further, as the baIIle
cut is 25° and the crossIlow and the window Ilow velocities are close to each other, there is
no nonideal window Ilow penalty Iactor either.
It will be seen that as the baIIle spacing is increased Irom 11.8 in. (300 mm) to 15.75 in.
(400 mm), the main crossIlow, the bundle-shell bypass stream and the pass-partition bypass
stream increase progressively whereas the baIIle hole-tube and baIIle-shell leakage streams
decrease progressively. The overall heat transIer eIIiciency oI the shellside stream increases
progressively. There are a number oI Iactors at play here: the crossIlow and window Ilow
velocities, the number oI baIIle spaces, the stream analysis, and the overall shellside
eIIiciency. The net result is that with an increase in the baIIle spacing Irom 11.8 in. (300
mm) to 13.8 in. (350 mm), the shellside heat transIer coeIIicient drops appreciably Irom
503.4 Btu/h It
2
ƒF (2458 kcal/h m
2
ƒC) to 445.9 Btu/h It
2
ƒF (2177 kcal/h m
2
ƒC) while the
Table 3.4a: Principal process and construction parameters Ior Case Study 3.4
Shellside Tubeside
1. Fluid Hydrocarbon liquid Hydrocarbon liquid
2. Flow rate, lb/h (kg/h) 771,610 (350,000) 220,460 (100,000)
3. Inlet/outlet temperature, °F (°C) 410 (210)/440.6 (227) 608(320)/518 (270)
4. Heat duty, MM Btu/h (MM kcal/h) 14.98 (3.774)
5. Density in/out, lb/It
3
(kg/m
3
) 45.68 (732)/44.74 (717) 40.56 (650)/44.0 (705)
6. Viscosity in/out, cp 0.5/0.45 0.25/0.4
7. SpeciIic heat in/out, Btu/lb °F (kcal/kg °C) 0.635/0.65 0.77/0.72
8. Thermal conductivity in/out, kcal/h m °C 0.059(0.088)/0.058 (0.086) 0.049 (0.073)/0.055 (0.082)
9. Allowable pressure drop, psi (kg/cm2) 14.2 (1.0) 10 (0.7)
10. Fouling resistance, h It
2
ƒF/Btu
(h m
2
°C/kcal)
0.00293 (0.0006) 0.00293 (0.0006)
11. Material oI construction CS 5 Cr ½ Mo
12. Type oI exchanger Floating-head (TEMA AES)
13. Shell ID, in. (mm) 29.1 (740)
14. No. oI tubes x no. oI tube passes 300 x 4
15. Tube OD × thick × length, in. (mm) 0.984 (25) × 0.0787 (2) × 236 (6000)
16. Heat transIer area, It
2
(m
2
) 1485 (138)
17. Line size, in. (mm) 12 (300) 6 (150)
45
shellside pressure drop plunges Irom 17.6 psi (1.24 kg/cm
2
) to 11.5 psi (0.81 kg/cm
2
).
With a Iurther increase in the baIIle spacing Irom 13.8 in. (350 mm) to 15.75 in. (400
mm), while the shellside pressure drop Ialls Iurther to 0.727 kg/cm
2
, the shellside heat
transIer coeIIicient increases marginally to 470.9 Btu/h It
2
ƒF (2299 kcal/h m
2
ƒC).
As the allowable shellside pressure drop is 0.8 kg/cm
2
, Design A is ruled out as its
shellside pressure drop Iar exceeds this limit. The shellside pressure drop in design B also
exceeds the allowable value, although marginally. Thus, Design C is the bestnot only is its
shellside pressure drop within the allowable value, but its overdesign is higher and its stream
analysis better than those oI Design B.
Let us now consider the eIIect oI varying the baIIle cut, retaining the baIIle spacing as
15.75 in. (400 mm). The results are shown in Table 3.4c. Design A here is the starting point
with a baIIle cut oI 25°. It is progressively increased to 30°, 33°, and 36° in Designs B,
C, and D. In the Iinal design (Design E), the results oI a 20° baIIle cut has also been shown.
With the progressive increase oI baIIle cut Irom 25° to 36°, the Iollowing changes are
observed:
1) The main crossIlow stream Iraction increases appreciably.
2) The baIIle hole-tube, baIIle-shell and pass-partition stream Iractions decrease
steadily.
3) The bundle-shell bypass stream Iraction remains quite steady.
4) The overall heat transIer eIIiciency oI the shellside stream Iirst decreases and
then increases.
5) However, as the window velocity decreases progressively, the shellside heat
Table 3.4b: Results oI variation in baIIle spacing Ior 25° baIIle cut Ior Case Study 3.4
Design Case (Baffle spacing)
A
11.8 in.
(300 mm)
B
13.8 in.
(350 mm)
C
15.75 in.
(400 mm)
1. BaIIle hole-tube leakage stream (A) 0.1496 0.1339 0.1234
2. Main crossIlow stream (B) 0.5335 0.553 0.5668
3. Bundle-shell bypass stream (C) 0.1289 0.1327 0.1353
4. BaIIle-shell leakage stream (E) 0.1191 0.1081 0.0995
5. Pass-partition bypass stream (F) 0.069 0.0723 0.075
6. Overall shellside heat transIer eIIiciency 0.719 0.74 0.755
crossIlow 7.78 (2.37) 6.66 (2.03) 5.81 (1.77)
7. Shellside velocity,
It/s (m/s)
window Ilow 7.38 (2.25) 7.38 (2.25) 7.38 (2.25)
8. Shellside pressure drop, psi (kg/cm
2
) 17.6 (1.24) 11.5 (0.81) 10.34 (0.727)
shellside 503.4 (2458) 445.9 (2177) 470.9 (2299)
tubeside 282.4 (1379) 282.4 (1379) 282.4 (1379)
9. Heat transIer
coeIIicient,
Btu/h It
2
°F (kcal/h
m
2
°C)
overall 75.6 (396.3) 79.5 (388.1) 80.3 (391.9)
10. Overdesign, ° 8.37 6.22 7.2
46
transIer coeIIicient also decreases. The pressure drop also decreases, at almost
the same rate as the heat transIer coeIIicient.
6) At the higher baIIle cuts oI 33° and 36°, the nonideal window Ilow correction
Iactor becomes rather low (0.891 and 0.855).
7) The above Iactors are Iinally reIlected in the overdesign values. Designs E and A
have the higest overdesign values but they cannot be accepted due to their
shellside pressure drop being in excess oI the allowable value. The next highest
overdesign is in Design B which is thereIore accepted as the Iinal design.
3.4.6 Reduction of shellside pressure drop
Shellside pressure drop reduction is almost always accomplished by modiIying the
shell/baIIle style and design, and occasionally by the use oI increased tube pitch. These
will now be discussed in detail.
Table 3.4c: Results oI variation in baIIle cut Ior 400 mm baIIle spacing Ior Case Study 3.4
Design Case/Baffle cut (diameter)
A
25°
B
30°
C
33°
D
36 °
E
20°
1. BaIIle hole-tube leakage stream (A) 0.1234 0.101 0.088 0.0755 0.1516
2. Main crossIlow stream (B) 0.5668 0.6 0.6304 0.6608 0.5297
3. Bundle-shell bypass stream (C) 0.1353 0.14 0.1353 0.1336 0.144
4. BaIIle-shell leakage stream (E) 0.0995 0.09 0.084 0.0779 0.1132
5. Pass-partition bypass stream (F) 0.075 0.069 0.0623 0.0521 0.0615
6. Overall shellside heat transIer eIIiciency 0.755 0.728 0.755 0.785 0.733
crossIlow
5.8
(1.77)
5.8
(1.77)
5.8
(1.77)
5.8
(1.77)
5.8
(1.77)

7. Shellside velocity, It/s
(m/s)
window Ilow 7.38
(2.25)
5.87
(1.79)
5.18
(1.58)
4.66
(1.42)
9.74
(2.97)
8. Shellside pressure drop, psi (kg/cm
2
)
10.33
(0.727)
9.1
(0.64)
8.5
(0.595)
7.86
(0.553)
12.9
(0.91)
shellside
470.9
(2299)
435.4
(2126)
408.2
(1993)
377.3
(1842)
476.6
(2327)
tubeside
282.4
(1379)
282.4
(1379)
282.4
(1379)
282.4
(1379)
282.4
(1379)


9. Heat transIer
coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
overall
80.3
(391.9)
79.2
(386.6)
78.2
(382)
77.0
(376.1)
80.4
(392.7)
10. Correction Iactor Ior nonideal
window Ilow
0.919 0.92 0.891 0.855 0.9
11. Overdesign, ° 7.2 5.77 4.52 2.92 7.4

47
3.4.6.1 Moaifying shell/baffle style ana aesign
Single-pass shell and single segmental baffles
The Iirst baIIle alternative is the single-segmental in a single-pass (TEMA style E) shell
(see Fig. 2.6). Perhaps more heat exchangers are designed with an E shell than all other
types put together. However, in many situations, the shellside pressure drop is too high
with single-segmental baIIles in a single-pass shell even aIter increasing the baIIle
spacing and baIIle cut to the highest recommended values (see Section 3.4.4). Such a
situation may arise when handling a very high shellside Ilow rate and/or when the
shellside Iluid is a low-pressure gas. In such situations, the Iirst alternative the designer
should consider is the double-segmental baIIle (see Fig. 3.3).
Single-pass shell and double-segmental baffles
By changing the baIIling Irom single-segmental to double-segmental at the same spacing
in an otherwise identical heat exchanger, the crossIlow velocity becomes approximately
one halI since the shellside Ilow is divided into two identical parallel streams. This results
in a large reduction in the crossIlow pressure drop. However, the window velocity and
thereIore the window pressure drop cannot be reduced appreciably because the maximum
recommended baIIle cut should have been tried with single-segmental baIIles beIore
changing over to double-segmental baIIles. Nevertheless, since the crossIlow pressure
drop is invariably much greater than the window pressure drop, there is an appreciable
reduction in the total pressure drop. Evidently, there is also a reduction in the shellside
heat transIer coeIIicient, but this is considerably less than the reduction in the pressure
drop. A detailed presentation on the use oI double-segmental baIIles appears in |2|.
CASE STUDY 3.5: USE OF DOUBLE-SEGMENTAL BAFFLES
Let us consider the heat exchanger service speciIied in Table 3.5a. Tubes 0.7874 in. (20
mm) OD and 0.0787 in. (2 mm) thick and a Iloating-head (TEMA AES) construction are
to be used. The tube pitch is to be 1.024 in. (26 mm) square. The entire exchanger is to be
oI carbon steel construction.
Table 3.5a: Principal process parameters Ior Case Study 3.5
Shellside Tubeside
1. Fluid Gas Cooling water
2. Flow rate, lb/h (kg/h) 154,322 (70,000) 549,915 (249,440)
3. Inlet/outlet temperature, °F (°C) 282.2 (139)/102.2 (39) 89.6 (32)/102.2 (39)
4. Heat duty, MM Btu/h (MM kcal/h) 6.92 (1.744)
5. Operating pressure, psia (kg/cm
2
abs.) 412 (29.0) 71 (5.0)
6. Allowable pr. drop, psi (kg/cm
2
) 2.84 (0.2) 10.7 (0.75)
7. Fouling resistance, h It
2
°F
(Btu h m
2
°C/kcal)
0.00146 (0.0003) 0.00195 (0.0004)
8. Material oI construction Carbon steel Carbon steel
9. Connection size, in. (mm) (nominal) 14 (350) 10 (250)
48
A design was Iirst tried with single-segmental baIIles. It was Iound that with a baIIle
spacing oI 23.6 in. (600 mm) and a baIIle cut oI 32° (on diameter), the shellside pressure
drop could be restricted to the permissible value. However, it was observed that the tubes
were highly prone to Iailure due to Ilow-induced vibration (Iluidelastic whirling). This
phenomenon will be discussed in detail in Chapter 12. It will be seen Irom Table 3.5b that
the overdesign margin is rather high (47.5°). However, iI the number oI tubes and, thereby,
the shell diameter is reduced, the shellside pressure drop will increase to beyond the
permissible value. II the baIIle spacing is increased Iurther to contain the shellside pressure
drop, the design will become even more unsaIe against Ilow-induced vibration.
This design was thereIore abandoned and double-segmental baIIles considered. It was
Iound that the permitted shellside pressure drop could be adhered to and a design produced
that is saIe against Iailure oI tubes due to Ilow-induced vibration. The results oI this design
are also shown in Table 3.5b. The baIIle spacing is 14.2 in. (360 mm), and there is an
overlap oI Iour tube rows. Comparing the two designs, it will be seen that the double-
segmental baIIle design was able to reduce the baIIle spacing and, thereby, the unsupported
Table 3.5b: Results oI design with single- and double-segmental baIIles Ior Case Study 3.5
Single-segmental
design
Double-segmental
design
1. Type oI exchanger Floating-head (TEMA AES)
2. Shell ID, in. (mm) 37.4 (950) 33.46 (850)
3. Number oI tubes 790 635
4. Tube pitch, in. (mm)/layout angle 1.063 (27)/square
5. Number oI tube passes 4 2
Total no. oI
cross-passes
9 15
BaIIle spacing,
in in. (mm)
23.6 (600) 14.17 (360)


6. BaIIling
BaIIle cut 32° (dia.)
4 rows overlap
(35.9° on area)
crossIlow 12.96 (3.95) 12.07 (3.68) 7. Shellside
velocity, It/s
(m/s)
window Ilow 16.0 (4.88) 15.88 (4.84)
8. Tubeside velocity, It/s (m/s) 5.77 (1.76) 3.58 (1.09)
9. Shellside pressure drop, psi (kg/cm
2
) 2.56 (0.18) 2.7 (0.19)
10. Tubeside pressure drop, psi (kg/cm
2
) 10.95 (0.77) 2.42 (0.17)
Shellside 94.58 (461.8) 95.95 (468.5)
Tubeside 1144 (5586) 788.1 (3848)
11. Heat transIer
coeIIicient,
kcal/h m2 °C
Overall 64.1 (312.9) 63.1 (308.1)
12. Overdesign, ° 47.5 12.2
13. SaIe against Iailure oI tubes due to Ilow-
induced vibration
No (crossIlow velocity ~
critical velocity)
Yes
49
span considerably Ior about the same shellside velocity, pressure drop, and heat transIer
coeIIicient. The lower unsupported span was the key to a design that is saIe against Iailure oI
tubes due to Ilow-induced vibration: this will be discussed in detail in Chapter 12.
Diviaea-flow shell ana single-segmental baffles
II the allowable shellside pressure drop cannot be complied with even with double-
segmental baIIles at a relatively large spacing, or iI such a design becomes prone to
Iailure oI tubes due to Ilow-induced vibration, a divided-Ilow shell (TEMA J) with
single-segmental baIIles should be investigated next (see Fig. 2.6). Here, one halI oI the
shellside Iluid Ilows along each halI oI the shell length, when compared to the single-pass
shell. As pressure drop is proportional to the square oI the velocity and to the length oI
travel, a divided-Ilow shell will result in approximately one-eighth the pressure drop in an
otherwise identical single-pass exchanger. The superiority oI a divided-Ilow shell over
double-segmental baIIles is that it oIIers an even larger reduction in pressure drop, since
not only crossIlow velocity but even window velocity can be reduced and the length oI
travel is one-halI. The disadvantage is the increase in cost due to the additional piping
required.
As divided-Ilow construction is quite common in process condensers, a Case Study will
be presented in Chapter 7 in which we will discuss condensers in detail.
Diviaea-flow shell ana aouble-segmental baffles
II even a divided-Ilow shell with single-segmental baIIles is unable to meet the allowable
shellside pressure drop limit and ensure a design saIe against Iailure oI tubes due to Ilow-
induced vibration, it will be necessary to adopt a combination oI a divided-Ilow shell and
double-segmental baIIles. Evidently, with such a combination, a very large reduction in
shellside pressure drop is possible, as low as 4° compared to a single-pass exchanger
having the same baIIle spacing and cut. In sharp contrast, the heat transIer coeIIicient will
reduce to only about 40°.
As this construction is also much more prevalent among condensers, a Case Study will
be presented in Chapter 7.
No-tubes-in-winaow segmental baffles
This is discussed in detail later in Chapter 12.
Crossflow shell
There are some services where the pressure drop limitation is so severe that none oI the
above shell/baIIling designs can yield a satisIactory design. A steam ejector condenser
operating at a pressure oI 50 mm Hg and having an allowable pressure drop oI 5 mm Hg
is an example. The ultimate technique in shellside pressure drop reduction is applied
successIully in such a situation: the use oI the crossIlow shell (TEMA X style). Here,
pure crossIlow takes place at a very low velocity and consequently, there is virtually no
pressure drop in the shell. Whatever pressure drop there is, it is almost entirely in the
nozzles. Evidently, support plates will have to be used to meet TEMA requirements and
prevent any possible Ilow-induced tube vibration. Since the shellside Ilow is parallel to
these support plates, shellside pressure drop is not increased.
3.4.6.2 Increasea tube pitch
For a given number oI tubes, the smaller the tube pitch, the smaller the shell diameter and
50
thereIore the lower the Iirst cost. Consequently, designers tend to pack in as many tubes
as mechanically possible, limited either by ligament width (distance between adjacent
tubes ÷ tube pitch ² tube OD) Ior Iabrication purposes or cleaning requirements.
Generally, as speciIied in the TEMA standards, designers use 1.25 times the tube OD as
the tube pitch. For square or rotated square pitch, a minimum cleaning lane oI 0.25 in. or
6 mm is additionally recommended by TEMA. Thus, 0.7874-in. (20-mm) OD tubes are
laid out on a 0.984-in. (25-mm) triangular pitch or a 1.024-in. (26-mm) square pitch.
The optimum tube pitch-to-tube OD ratio Ior conversion oI pressure drop to heat
transIer is generally 1.25²1.35 Ior turbulent Ilow and around 1.4 Ior laminar Ilow.
Increasing the tube pitch to reduce pressure drop is generally not recommended as it
increases the shell diameter and, thereby, the cost. Pressure drop reduction is usually
achieved by modiIication oI baIIle spacing/cut and/or shell type in a Iar cheaper design.
However, in the case oI X shells, it may be necessary to increase the tube pitch Irom the
TEMA minimum to meet pressure drop limitations as there are no other parameters which
can be modiIied.
A Case Study showing the logic oI selecting the shell type and the baIIle type is
presented in Chapter 12 on Ilow-induced vibration, as pressure drop and Ilow-induced
vibration are inextricably linked together.
References
|1| Tinker, T, 1951, Shellsiae Characteristics of Shell ana Tube Heat Exchangers, Parts I, II, ana
III, General Discussion oI Heat TransIer, Proc. Institution oI Mechanical Engineers, London.
|2| Mukherjee, R., 1992, 'Use Double-segmental BaIIles in Shell-and-Tube Heat Exchangers,¨
Chemical Engineering Progress, 88(11), Nov., pp. 47²52.
Further Reading
1. Hewitt, G.F., (ed.), 1998, Heat Exchanger Design Hanabook 1998, Jolumes 3 ana 4, Begell
House, Inc., New York.
2. Kakac, S., et al., 1981, Heat Exchangers. Thermal ¥ Hyaraulic Funaamentals ana Design,
Hemisphere Publishing Corporation, New York.
3. McKetta, J.J. (ed.), 1992, Heat Transfer Design Methoas, Marcel Drekker, Inc., New York.
4. Rohsenow, W.M., and Hartnett, J.P. (eds.), 1973, Hanabook of Heat Transfer, Section 18,
McGraw-Hill Book Co.
5. Azbel, D., 1984, Heat Transfer Applications in Process Engineering, Noyes Publications,
New Jersey.

51
CHAPTER 4
0HDQ7HPSHUDWXUH'LIIHUHQFH
Temperature diIIerence is the driving Iorce Ior heat transIer and, as such, is a very
important parameter in heat exchanger design. There are many issues involved in the
mechanics and determination oI the mean temperature diIIerence (MTD) oI a heat
exchanger and hence an entire chapter is devoted to this subject.
4.1 Logarithmic Mean Temperature Difference (LMTD)
As is well known, temperature diIIerence is the driving Iorce Ior heat transIer. Since the
temperature diIIerence varies across the length oI a heat transIer section, it has to be
weighed to obtain a mean value Ior a single-point determination oI heat transIer area. The
mean temperature diIIerence in a heat transIer process can be shown to be equal to the
logarithmic mean oI the two terminal temperature diIIerences, one at the hot end oI the
heat transIer section and the other at the cold end. Hence, this is reIerred to as the LMTD.
In the derivation oI the LMTD as the true mean temperature diIIerence, the Iollowing
assumptions are required to be made:
a) The overall heat transIer coeIIicient is constant over the entire length oI the path.
b) The mass Ilow rate and speciIic heat oI both the streams are constant over the
entire length oI the path.
c) There is no phase change in either Ilowing stream, that is, only sensible heat
transIer takes place.
a) Heat losses are negligible.
4.2 Countercurrent Flow
When two streams Ilow in opposite
directions across a tube wall, we
have true countercurrent Ilow (Fig.
4.1). In this situation, the only
limitation is that, at all locations, the
hot stream should be hotter than the
cold stream. The outlet temperature
oI the cold stream, however, may
well be higher than the outlet
temperature oI the hot stream.
Let us consider the Iollowing hot
and cold stream data Ior Fig. 4.1:
Fig. 4.1 Countercurrent Ilow
52
1) Hot stream: Ilow rate 27,558
lb/h (12,500 kg/h), speciIic heat,
0.64 Btu/lb ƒF (0.64 kcal/kg ƒC)
and inlet temperature 266ƒF
(130ƒC)
2) Cold stream: Ilow rate 22,046
lb/h (10,000 kg/h), speciIic heat 0.6
Btu/lb ƒF (0.6 kcal/kg ƒC) and inlet
temperature 104 ƒF (40 ƒC)
Let us also consider a minimum
temperature approach between the
hot and cold streams to be 18ƒF (10ƒC). It will be seen by a heat balance that Ior a hot
end terminal oI 18ƒF (10ƒC), the hot stream will leave at 158ƒF (70ƒC) and the cold
stream will leave at 248ƒF (120ƒC), representing a heat duty oI 1,905,024 Btu/h (480,000
kcal/h). This is represented in Fig. 4.1. The temperature diIIerence at the hot terminal is
(266 ² 248) or 18ƒF |(130 ² 120) or 10ƒC| (which was the minimum speciIied) and the
temperature diIIerence at the cold terminal is (158 ² 104) or 54ƒF |(70 ² 40) or 30ƒC|.
Thus, the LMTD will be 32.76ƒF (18.2ƒC).
4.3 Co-Current Flow
II the hot and cold streams Ilow in the same direction, we have co-current Ilow. This is
represented in Fig. 4.2. Again consider the same hot and cold stream Ilow rates and inlet
temperatures as in the countercurrent case above. It will be seen by a heat balance that
corresponding to a minimum temperature diIIerence oI 18ƒF (10ƒC), the hot stream will
leave at 204.3ƒF (95.7ƒC) and the cold stream at 186.3ƒF (85.7ƒC), thereby representing a
heat duty oI only 1,089,040 Btu/h (274,400 kcal/h) (see Fig. 4.2). This is considerably
lower than the heat duty that could be transIerred in countercurrent Ilow. Also, the
temperature diIIerence at the hot stream inlet end is (266 ² 104) or 162ƒF |(130 ² 40) or
90ƒC| and that at the hot stream outlet end is (204.3 ² 186.3) or 18ƒF |(95.7 ² 85.7) or
10ƒC|. Thus, the LMTD here will be considerably higher at 97.5ƒF (36.4ƒC).
Thus, with co-current Ilow, not only is the amount oI heat that can be transIerred lower,
the LMTD is also much lower than that Ior countercurrent Ilow, albeit Ior the same stream
temperatures. Consider a hot stream inlet/outlet temperature oI 212/158ƒF (100/70ƒC) and a
cold stream inlet/outlet temperature oI 86/140ƒF (30/60ƒC). In countercurrent Ilow, the
situation will be
212 Æ 158 100 Æ 70
140 Å 86 60 Å 30
---- ---- ---- ----
72 72 40 40
---- ---- ---- ----
Thus, the LMTD will be 72ƒF (40ƒF).
In co-current Ilow, the situation will be
212 Æ 158 100 Æ 70
86 Æ 140 30 Æ 60
---- ---- ---- ----
126 18 70 10
---- ---- ---- ----
Fig. 4.2 Co-current Ilow
53
Thus, the LMTD here is 55.5ƒF (30.83ƒC).
Co-current Ilow yields a lower LMTD Ior the same terminal diIIerences because
whereas one terminal temperature diIIerence is rather high, the other is very low. That is, the
temperature diIIerences along the path oI heat transIer are not balanced, thereby yielding a
lower LMTD. What is even more serious with co-current Ilow is that the cold stream outlet
temperature has to be somewhat lower than the hot stream outlet temperature, which is a
very big limitation. This is not true Ior the countercurrent situation where the cold stream
outlet temperature may well be greater than the hot stream outlet temperature. Consequently,
countercurrent Ilow is almost always preIerred to co-current Ilow.
Almost, because in some situations other considerations override. For example, Ior
cooling a very viscous liquid, a co-current Ilow may be employed in order to prevent the
outlet hot stream Irom encountering the coldest coolant, so that its wall viscosity is not as
high as it would be iI it did encounter the coldest coolant. The problem with the higher wall
viscosity is that it reduces the hot stream heat transIer coeIIicient. An even more important
consideration is congealing, a condition that arises when the tube-wall temperature Ialls
below the pour-point oI the congealing liquid. Evidently, a co-current arrangement will
result in a higher wall temperature at the cold end, where congealing is obviously more
likely to occur. However, considering that most industrial heat exchangers have two or more
tube passes, pure co-current does not have much practical signiIicance.
Another example oI co-current Ilow is a vertical thermosyphon reboiler which is a
single-pass vertical Iixed-tubesheet heat exchanger. In some special circumstances, co-
current Ilow is employed Ior reasons oI more eIIicient heat transIer. This will be elaborated
later in Chapter 8 which discusses reboilers.
4.4 Countercurrent and Co-Current Flow: The F
t
Factor
Let us now consider a shell-and-tube heat exchanger. Here, the shellside Iluid may either
be completely mixed due to high turbulence, or may demonstrate a selective temperature
proIile due to low turbulence. The presence oI baIIles and the consequent turbulent Ilow
across the tubes tend to promote mixing.
The Iollowing assumptions need to be made in the determination oI the true temperature
diIIerence Ior a shell-and-tube heat exchanger:
1) There is complete mixing within any shell cross pass or tube pass.
2) The overall heat transIer coeIIicient is constant throughout the exchanger.
3) The Ilow rate and speciIic heat oI both the streams are constant.
4) The number oI cross passes is greater than three.
5) The heat transIer surIace in each pass is equal.
6) There is no phase change (oI condensation or vaporization) in any part oI the
exchanger.
7) The heat loss to the surroundings, or internally across pass partition plates, is
negligible.
The very Iirst assumption is not totally true as mixing is never complete. However, the
lack oI complete mixing does not introduce a very serious error in the determination oI
the MTD.
The second assumption is oIten untenable as the overall heat transIer coeIIicient varies
signiIicantly Irom one end oI the heat exchanger to the other. In such cases, it becomes
necessary to divide the heat exchanger into several zones (which is necessary anyway due to
54
the variation oI the heat transIer coeIIicient) and perIorm MTD calculations zone-wise.
The number oI cross passes is invariably Iar greater than three.
Due to reasons already explained earlier in the sections on tubeside heat transIer
coeIIicient (Section 3.3.2) and pressure drop (Section 3.3.3), shell-and-tube heat exchangers
are invariably oI two or more tube passes. In a single-pass shell (the commonest type), the
shellside Iluid Ilows in one direction, Irom leIt to right or Irom right to leIt. Consequently,
halI the tube passes experience countercurrent Ilow and the other halI experience co-current
Ilow. The MTD Ior this situation is neither the LMTD Ior countercurrent Ilow nor that Ior
co-current Ilow, but a value in between the two. A correction Iactor called the F
t
Iactor,
which depends on the Iour terminal temperatures, has to be applied in this situation and the
same can be determined Irom charts in the TEMA standards. This Iactor is used to multiply
the LMTD Ior countercurrent Ilow to obtain the MTD. It does not matter whether there are
two, or Iour, or more passes. It should be understood that the highest value oI F
t
is 1.0, as the
MTD can never exceed the LMTD Ior countercurrent Ilow.
The F
t
Iactor is determined as Iollows.
Let
T
1
÷ hot stream inlet temperature
T
2
÷ hot stream outlet temperature
t
1
÷ cold stream inlet temperature
t
2
÷ cold stream outlet temperature
The F
t
Iactor is determined Irom
R ÷ (T
1
² T
2
)/(t
2
² t
1
) (4.1)
S ÷ (t
2
² t
1
)/(T
1
² t
1
) (4.2)
Since heat duty, Q ÷ WC(T
1
² T
2
) ÷ wc(t
2
² t
1
), (4.3)
where
W ÷ mass Ilow rate oI hot stream
w ÷ mass Ilow rate oI cold stream
C ÷ speciIic heat oI hot stream
c ÷ speciIic heat oI cold stream
it Iollows that (T
1
- T
2
)/(t
2
- t
1
) ÷ wc/WC, which is the heat capacity ratio, the ratio oI the
product oI mass Ilow rate and speciIic heat oI the cold stream to that oI the hot stream.
Thus, R is the heat capacity ratio. S is called the 'thermal eIIectiveness.¨
The F
t
correction Iactor is read Irom R and S values just determined, Irom the
appropriate TEMA chart.
Let us consider the case oI a single-pass shell with two or more passes. Let the hot
stream inlet and outlet temperatures be 302
o
F (150
o
C) and 140
o
F (60
o
C), and the cold stream
inlet and outlet temperatures be 104
o
F (40
o
C) and 136.4
o
F (58
o
C). Thus, calculating in
Celsius:
R ÷ (T
1
² T
2
)/(t
2
² t
1
) ÷ (150 ² 60)/(58 ² 40) ÷ 90/18 ÷ 5.0 (4.4)
and
55
S ÷ (t
2
² t
1
)/(T
1
² t
1
) ÷ (58 ² 40)/(150 ² 40) ÷ 18/110 ÷ 0.164 (4.5)
Evidently, the same numbers will be obtained iI the values in degrees Fahrenheit are
used.
Now, reIerring to the TEMA F
t
chart Ior a single-pass shell (reproduced here as Fig.
4.3a), we Iind that Ior R ÷ 5.0 and S ÷ 0.164, F
t
÷ 0.855.
Since each situation is diIIerent as Iar as the detailed temperature distribution is
concerned, diIIerent F
t
correction charts are presented in the TEMA standards Ior the
Iollowing conIigurations:
1) 1 shell pass, even number oI tube passes
2) 2 shell passes, 4 or multiple oI 4 tube passes
3) 3 shell passes, 6 or more even number oI tube passes
4) 4 shell passes, 8 or more even number oI tube passes
5) 5 shell passes, 10 or more even number oI tube passes
6) 6 shell passes, 12 or more even number oI tube passes
7) 1 shell pass, 3 tube passes (2 countercurrent and 1 co-current) crossIlow shell,
1 tube pass
8) 1 divided Ilow shell pass (J shell), 1 tube pass
9) 1 divided Ilow shell pass, even number oI tube passes
10) split Ilow shell (G shell), 2 tube passes
11) split Ilow shell, 4 (or multiple oI 4) tube passes
12) double split Ilow shell (H shell), 2 tube passes
The F
t
correction Iactor charts Ior the Iirst two conIigurations above are reproduced Irom
the TEMA standards as Figs. 4.3a and 4.3b, respectively.
The reader is inIormed here that since the Iirst eight oI the above TEMA charts are
stream symmetric, t and T may be taken as the cold and hot Iluid temperatures, respectively,
regardless oI whether they Ilow through the shellside or tubeside. However, the balance Iive
charts are non-stream symmetric; thereIore, t and T must be taken as the tubeside and the
shellside Iluid temperatures, respectively.
It will be noticed in the F
t
Iactor charts that the curves start rather Ilat at the upper leIt-
hand corner and rapidly increase in slope beIore ending up as steep, almost vertical lines.
Since the vertical portion represents an area oI uncertainty, it is strongly recommended that
such an operating point not be adopted. The remedy is to use an extra shell in series. For
example, iI R ÷ 8.0 and S ÷ 0.115, it is extremely uncertain what the F
t
Iactor will be with a
single shell, as the point lies on the vertical portion oI the R ÷ 8.0 line. In this situation, iI we
were to use two shells in series, the F
t
Iactor is determined Irom the corresponding chart
(Fig. 4.3b) as 0.95.
It will not be out oI place here to state that, should one oI the streams be isothermal,
there will be no correction Iactor; in other words, F
t
÷ 1.0. II both the streams are isothermal,
one condensing and the other vaporizing, the LMTD will simply be equal to the arithmetic
diIIerence between the two temperatures.
As regards the temperature approach between the hot and the cold streams, the
Iollowing guidelines may be considered:
1) In sensible heat transIer and condensing, there is no limit to the temperature ap-
proach, either lower or higherit only has to be a positive value. However, con-
56
sidering that the required heat transIer area is inversely proportional to the MTD,
a minimum temperature oI 5ƒC or 10ƒC is recommended.
2) In boiling heat transIer, however, the heat transIer coeIIicient is a Iunction oI the
temperature diIIerence across the boiling Iilm. A very low temperature diIIerence
(say, 2²3ƒC) may inhibit nucleate boiling, whereas a very high temperature
diIIerence may yield Iilm boiling. This matter will be discussed in detail in
Chapter 8.
4.5 Temperature Cross
An important limitation Ior 1²2 shells (one shell pass and 2 or more tube passes) is that
the outlet temperature oI the cold stream cannot be greater than the outlet temperature oI
the hot stream. This is because oI the presence oI one or more co-current passes. In actual
terms, a very small temperature cross is possible, but since this represents an area oI
uncertainty and the credit is very small, it is usually ignored and the equal outlet
temperature condition respected.

Fig. 4.3a LMTD correction Iactor Ior one shell pass and even number oI tube passes (Reprinted with
permission Irom Standards oI TEMA, 8th

Edition, 1999.)
57
When there is a temperature cross, multiple shells in series are required to be employed
iI pure countercurrent Ilow cannot be realized. The modalities oI this are discussed in detail
in Section 6.2.1.
An F shell has 2 shell passes, so iI there are 2 tube passes as well, it represents a pure
countercurrent Ilow situation provided the directions oI the Ilow paths are selected properly.
However, iI an F shell has 4 or more tube passes, it is no longer a true countercurrent
situation and, consequently, the F
t
correction has to be applied. An F shell having 4 or more
tube passes is represented as a 2²4 shell. The F
t
Iactor Ior a 2²4 shell is identical to that Ior
two 1²2 shells in series or 2 shell passes in TEMA parlance.
The TEMA F
t
Iactor chart Ior 3 shell passes really represents 3 shells in series, that Ior 4
shell passes 4 shells in series, and so on.
4.6 Heat Release Profiles and Zone-Wise Calculations
An important point to note here is that the LMTD and F
t
Iactor concept assumes that
there is no signiIicant variation in the overall heat transIer coeIIicient along the length oI

Fig.4.3b LMTD correction Iactor Ior two shell passes and Iour or multiple oI Iour tube passes (Reprinted
with permission Irom Standards oI TEMA, 8th

Edition, 1999.)
58
the shell. However, there are some services where this is not true, e.g., the cooling oI a
viscous liquid. As the liquid is cooled, its viscosity increases and this results in a
progressive reduction in the shellside heat transIer coeIIicient. Here, the simplistic overall
MTD approach will be inaccurate. The exchanger must be broken into several sections
and the calculations perIormed zone-wise. In Iact, commercial soItware packages
presently available carry out such detailed zone-wise calculations.
Now consider the case oI a condenser wherein a mixture oI hydrocarbons is being
condensed over a temperature range. II a plot is made oI the variation oI heat duty versus
temperature, it is usually Iound to be a curve such as shown in Fig. 4.4. This is because more
vapors are condensed at the higher
temperatures than at the lower tem-
peratures, thereby resulting in higher
heat duties being transIerred at the
higher temperatures. Such a plot oI
heat duty versus temperature is called
a heat release profile. Evidently, the
calculations will have to be broken
into several zones. The number oI
zones will depend upon the variation
oI the slope oI the heat release
proIile. The higher the variation, the
larger will be the number oI zones.
The heat release proIile will have to
be broken into a suIIiciently large
number oI zones such that the proIile
is essentially linear in each oI these
zones. This is represented in Figs.
4.5a and 4.5b wherein it will be seen
that a smaller number oI zones (two:
AB and AC) will suIIice in Fig. 4.5a,
but a larger number oI zones (Iour:
AB, BC, CD, and DE) will be
required in Fig. 4.5b.
In the case oI liquids and gases
cooled by water, the MTD usually
decreases Irom inlet to outlet. This is
because the hot end terminal tem-
perature diIIerence is almost always
higher than that at the cold end as the
temperature drop oI the hot Iluid is
invariably greater than the tempera-
ture rise oI cooling water. Consi-
dering a typical liquid cooler wherein
the liquid is to be cooled Irom, say,
158ƒF (70ƒC) to 104ƒF (40ƒC) by
cooling water |inlet/outlet tempera-
ture 86/104ƒF (30/40ƒC), it will be
seen that the temperature diIIerence
Fig. 4.4 Typical heat release proIile Ior condenser
Fig.4.5b Splitting a heat release curve into a large number oI
essentially linear zones
Fig. 4.5a Splitting a heat release curve into a small number oI
essentially linear zones
59
at the hot end is (158 ² 104) or 54ƒF |(70 ² 40) or 30ƒC| whereas that at the cold end is (104
² 86) or 18ƒF |(40 ² 30) or 10ƒC|. As a consequence, the MTD is higher at the hot end and
lower at the cold end.
In the case oI a viscous liquid cooler, the heat transIer coeIIicient is also higher at the hot
end due to the lower viscosity oI the hot liquid. Thus, both the MTD and the heat transIer
coeIIicient reduce Irom the exchanger inlet to the outlet. Consequently, Ior a given length
increment, the heat duty transIerred is much higher at the hot end and much lower at the cold
end, so that it is not uncommon Ior the last 10° heat duty to require 30° (or even more) oI
the total heat transIer surIace.
4.7 Temperature Profile Distortion
An important Iactor that has not been considered so Iar is the temperature proIile distor-
tion correction Iactor. It was discussed in Chapter 3 that there is not just one Ilow stream
on the shellside but one main crossIlow stream and Iour leakage/bypass streams (see Fig.
3.8a). The leakage and bypass streams are less eIIicient Ior heat transIer than the main
crossIlow stream. In Iact, the baIIle-to-shell leakage stream is totally ineIIective as it
completely bypasses the heat transIer surIace and, thereIore, does not enter the heat trans-
Ier process at all; consequently, its temperature proIile is horizontal. It is Ior this reason
that this stream is the most detrimental Ior heat transIer. It is like a member oI a team that
reIuses to do any work, thereby increasing the load on the other members. The other
streams are not as eIIicient as the main crossIlow stream but they do some useIul work.
Consider a case where the cold Iluid Ilows through the shellside. As the main crossIlow
stream encounters a very large Iraction oI the total heat transIer surIace, it picks up a very
large part oI the total heat duty. To illustrate this more vividly, assume that the main
crossIlow stream is 58° oI the total shellside stream but that it encounters 80° oI the total
number oI tubes. As a consequence, its temperature rises more rapidly than iI the entire
shellside stream were to pick up the entire heat duty. ThereIore, its temperature proIile will
be steeper than that oI the total stream (apparent temperature proIile) without considering the
various Ilow Iractions (see Fig. 4.6).
The temperature proIiles oI the tube-baIIle hole, bundle-shell, and the pass-partition
leakage/bypass streams will depend upon their respective Ilow Iractions and the Iractional
heat transIer area encountered. However, since the baIIle-shell leakage stream virtually does
not experience any heat transIer, the other Iour streams have to pick up the entire heat duty.
ThereIore, this combined
stream will have a tem-
perature proIile steeper
than that oI the apparent
stream. Consequently, the
temperature diIIerence be-
tween the hot and the cold
streams will be lower all
along the length oI the
heat exchanger, thereby
resulting in the reduction
oI the MTD. This reduc-
tion in the MTD is known
as the temperature profile
aistortion correction fac-
Fig.4.6 Temperature proIile distortion due to bypassing and leakage
(Courtesy oI HTRI.)
60
tor. The MTD is multiplied by this correction Iactor to obtain the eIIective MTD.
The temperature proIile oI the main crossIlow stream will not be a steady, continuous
curve as the various Ilow streams remix in the window region and again separate as
crossIlow re-commences. As Iar as the tube-baIIle hole leakage stream is concerned, it
separates Irom the main crossIlow stream and remixes with it in a complex Iashion (see Fig.
3.8a). It will be easy to understand that, since the shellside streams separate and remix
repeatedly all along the length oI the exchanger, the determination oI the actual temperature
proIile oI the various streams is extremely complicated. Thus, a sophisticated and powerIul
computer program is required to handle this situation and determine the temperature proIile
distortion correction Iactor.
The temperature proIile distortion correction Iactor is more pronounced when:
i) the leakage/bypass streams are high, especially the baIIle-to-shell leakage
streamthis is easy to understand as the greater these streams, the more distorted
will be the actual (or net) temperature proIile, and
ii) the ratio oI shellside temperature diIIerence (that is, the diIIerence between the
shellside inlet and outlet temperatures) to the temperature approach between the
hot and cold streams at the shell outlet is high. In other words, a high shellside
temperature diIIerence and a low temperature approach at the shell outlet tend to
result in a low temperature proIile distortion correction Iactor.
In order to understand the above phenomenon, let us discuss the matter in more detail. In
the Iollowing discussion, the Iour streams barring the ineIIective E stream will be
considered together as a net or eIIective stream, Ior ease oI comprehension. Whenever we
talk oI the distorted shellside stream, it is this net stream.
The larger the shellside temperature diIIerence, the greater will be the extent oI
temperature proIile distortion, as shown in Fig. 4.7. II the shellside temperature proIile is
represented by AB, it may distort to AD, so that BD is the distortion. However, iI the
shellside temperature is represented by AC, it will distort to AE, so that CE will be the
distortion. Evidently, CE will be greater than BD. Once this distortion begins, it gets greater
and greater, the longer it is allowed to proceed. Hence, the greater the shellside temperature
diIIerence, the greater the distortion.
Fig. 4.7 Variation oI temperature proIile distortion
correction Iactor
Fig. 4.8a Temperature proIile distortion in large
delta-t
s
and large delta-t
so
situation
61
Now also consider a tubeside temperature proIile GH (corresponding to AB) or FH
(corresponding to AC). When the shellside temperature proIile is AB and distorts to AD and
the tubeside temperature proIile is GH, the reduction in the MTD is not appreciable as DG is
not much less than BG. However, iI the shellside temperature proIile is AC and distorts to
AE and the tubeside temperature proIile is FH, it will be seen that there will be an
appreciable reduction in the MTD as EF is considerably less than CF.
In order to understand this even better, consider the plots in Figs. 4.8a, 4.8b, and 4.8c. In
all these cases, the hot stream is on the shellside. Further, delta-t
s
represents the diIIerence
between the shellside inlet and outlet temperatures and delta-t
so
represents the temperature
diIIerence between the shellside and the tubeside streams at the shell outlet.
Figure 4.8a represents a large delta-t
s
and a large delta-t
so
. Due to the large drop in the
shellside temperature, the temperature proIile distorts appreciably. However, since the tem-
perature diIIerence between the shellside and tubeside streams at the shell outlet is rather
large, the reduction in the temperature diIIerence at this terminal and, hence, the MTD is not
considerable.
Figure 4.8b represents a large delta-t
s
and a small delta-t
so
. It will be seen here that, due
to the Iairly large distortion oI the shellside temperature proIile and the small approach at the
shell outlet, there will be a very sharp reduction in the temperature diIIerence at the shell
outlet and, thereby, the MTD.
Figure 4.8c represents a small delta-t
s
and a small delta-t
so
. It will be seen here that, due
to the Iairly small distortion oI the shellside temperature proIile, there will not be a very
sharp reduction in the temperature diIIerence at the shell outlet, even though the temperature
approach is Iairly small. Consequently, the reduction in the MTD will not be appreciable.
It is now understood why both a high shellside temperature diIIerence ana a low
temperature approach at the shell outlet are required to produce a low temperature proIile
distortion correction Iactor. Further, a high viscosity oI the shellside Iluid and a large shell
diameter (with a corresponding large baIIle-shell clearance) will yield an appreciably high E
stream, thereby causing a more severe temperature proIile distortion. Thus, a viscous liquid
cooler being cooled through a large temperature range in a large diameter heat exchanger is
a prime candidate Ior a severe temperature proIile distortion. Vacuum residue being cooled
Irom 464ƒF (240ƒC) to 176ƒF (80ƒC) by tempered water getting heated Irom 140ƒF (60ƒC)
Fig. 4.8c Temperature proIile distortion in small
delta-t
s
and small delta-t
so
situation
Fig. 4.8b Temperature proIile distortion in large
delta-t
s
and small delta-t
so
situation
62
to 176ƒF (80ƒC) is such an example.
The leakage/bypass streams tend to be high when the shellside viscosity is high and
when the baIIle spacing is small. The latter was explained in Section 3.4.4 and demonstrated
in Case Study 3.3. Thus, care has to be exercised in the design oI viscous liquid coolers.
The minimum recommended temperature proIile distortion correction Iactor is 0.75,
below which the use oI two or more shells in series should be employed. The absolute
minimum is 0.65, below which the very methodology to determine this correction Iactor
becomes suspect. Further, this Iactor represents a direct loss in MTD, thereby requiring a
corresponding increase in the heat transIer area. Hence, it will be worthwhile to increase the
same to 90° or more, when the total heat transIer area will reduce substantially.
By using multiple shells in series, the ratio oI shellside temperature diIIerence to the
temperature approach at the shell outlet is reduced. The mixing oI the main crossIlow stream
with the bypass and leakage streams (including the E stream) aIter each shell reduces the
penalty due to the distortion oI the temperature proIile and, hence, increases the temperature
proIile distortion correction Iactor. This is shown graphically in Fig. 4.9. In a single-shell
design, the shellside temperature proIile will distort Irom AB to AE, representing an
appreciable reduction in the MTD. However, iI a two-shells-in-series design is employed,
the shellside temperature proIile will distort to AD at the end oI shell 1. At this point, all the
shellside streams will remix and the temperature will rise to F. In shell 2, the shellside
temperature proIile will distort to FG. Thus, there will be an increase in the overall MTD in
the two-shell design, corresponding to the area FGED.
Another very important gain with the use oI multiple shells in series is that with a
smaller shell diameter, the baIIle spacing can be reduced substantially. The combination oI
smaller shell diameter and closer baIIle spacing increases the shellside velocity and, thereby,
the shellside heat transIer coeIIicient considerably. Thus, not only the MTD, but also the
overall heat transIer coeIIicient increases appreciably, thereby resulting in a major reduction
in the total heat transIer area.
In many situations, a temperature proIile distortion correction Iactor is unavoidable,
such as when cooling a viscous liquid over a large temperature range, and there is no
alternative to the use oI multiple shells in
series. However, in many other situations,
improper baIIle spacing unnecessarily
imposes such a penalty where it is easily
avoidable. Designers normally tend to pack in
baIIles as close as possible to get the
maximum shellside heat transIer coeIIicient,
pressure drop permitting. In many such cases,
the use oI a somewhat higher baIIle spacing
will reduce the baIIle-to-shell leakage stream
(the principal culprit) and hence improve the
MTD penalty Iactor appreciably, thereby
producing a much better design. This author
has seen numerous otherwise good designs
tarnished by a very low baIIle spacing and
would like to strongly recommend a
minimum baIIle spacing oI 30° oI the shell
inside diameter. OIten, an even higher baIIle
spacing is strongly indicated by the penalties
Fig. 4.9 Improvement in temperature proIile
distortion correction Iactor by the use oI multiple
shells in series
63
oI baIIle-to-shell leakage stream and the temperature proIile distortion correction Iactor.
CASE STUDY 4.1: HOW A TEMPERATURE PROFILE DISTORTION
PROBLEM IS BETTER HANDLED BY TWO SHELLS IN SERIES
Let us reIer to Case Study 3.3 presented in Chapter 3. The eIIect oI baIIle spacing upon
stream analysis and temperature proIile distortion correction Iactor was brought out in
that study. The shellside temperature diIIerence was (90 ² 45) or 45ƒC. The temperature
approach at the shell outlet was (45 ² 35) or 10ƒC. Thus, the ratio oI the Iormer to the
latter was (45/10) ÷ 4.5, which is moderately high, so that the temperature proIile
distortion correction Iactor would be expected to be somewhat low. Indeed, reIerring to
Table 3.3c, it is seen that the temperature proIile distortion correction Iactor was
somewhat low, increasing progressively Irom 0.8488 to 0.904 with the increase in baIIle
spacing Irom 7.9 in. (200 mm) to 13.8 in. (350 mm).
Now let us see what happens iI two shells are used in series instead oI one. Since the
heat transIer area will double iI an identical construction is used, let us reduce the tube
length Irom 19.68 It (6000 mm) to 9.84 It (3000 mm), so that the total tube length and the
overall heat transIer area remain the same. (Another possibility is to reduce the number oI
tubes in each shell to halI, but this will entail a change in the shell diameter, itselI, and will
thereby not enable a proper comparison to be made between one shell and two shells in
series.)
Table 4.1a: Detailed results oI the various iterations Ior Case Study 4.1
(two shells in series design)
Baffle spacing, in. (mm)
13.8 (350) 9.8 (250) 7.9 (200)
BaIIle-to-tube 0.031 0.057 0.068
Main crossIlow 0.525 0.451 0.415
Bundle-shell 0.088 0.075 0.068
BaIIle-shell 0.295 0.366 0.403


Stream
Iractions
Pass-partition 0.061 0.051 0.046
Temperature proIile distortion
correction Iactor
0.9724 0.9627 0.957
Velocity cross/ window,
It/s (m/s)
0.43 (0.13)/0.62
(0.19)
0.56 (0.17)/0.62
(0.19)
0.72 (0.22)/0.62
(0.19)
Heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
43.4 (211.9) 44.4 (216.9) 45.9 (224.3)



Shellside
Pressure drop, psi (kg/cm
2
)
3.5 (0.243) 3.7 (0.262) 4.1 (0.286)
Overall heat transIer coeIIicient, Btu/h
It
2
ƒF (kcal/h m
2
ƒC)
33.5 (163.5) 34.1 (166.4) 35.0 (170.8)
MTD, ƒF (ƒC) 41.0 (22.8) 40.7 (22.6) 40.3 (22.4)
Overdesign 23.5 24.5 26.87
64
Central baIIle spacing oI 13.8 in. (350 mm), 9.8 in. (250 mm), and 7.9 in. (200 mm)
have been employed. Due to small tube length and due to the Iact that an odd number oI
baIIle spaces is usually employed, a design with a central baIIle spacing oI 11.8 in. (300
mm) could not materialize. The results are detailed in Table 4.1a. Upon comparison with
Table 3.3c, it will be seen that between the single-shell and two-shells-in-series designs, the
shellside Ilow Iractions and the shellside heat transIer coeIIicient are largely similar Ior the
same baIIle spacing. Both the shellside and the tubeside pressure drops are somewhat higher
in the two-shells-in-series design due to the extra nozzle and shell/channel entry and exit
losses. However as with the single-shell design, the permissible shellside pressure drop is
still largely unutilized, thus leading to an uneconomical design.
The temperature proIile distortion correction Iactor is considerably higher with two
shells in series than with just one shell: a direct comparison is shown in Table 4.1b. As
explained earlier in this chapter, the mixing oI the main crossIlow stream with the bypass
and leakage streams (including the E stream) aIter the Iirst shell reduces the penalty due to
the temperature proIile distortion and, hence, increases the temperature proIile distortion
correction Iactor.
Finally, it should be added here that the best design Ior the present service will be a two-
shells-in-series design, retaining a tube length oI 19.68 It. (6000 mm) so that the shell
diameter and, thereby, the Iirst cost oI the exchanger can be minimized. Such a design is
detailed in Table 4.1c and Table 4.1d. The Iormer depicts the principal construction
parameters, save baIIle spacing, and the latter the detailed perIormance results Ior Iour
diIIerent values oI baIIle spacing: 13.8 in. (350 mm), 11.8 in. (300 mm), 9.8 in. (250 mm),
Table 4.1b: Comparison oI temperature proIile distortion correction Iactor
between one shell and two shells in series (Case Studies 3.3 and 4.1)
Baffle spacing, mm
350 250 200
Case Study 3.3 (Single shell) 0.904 0.8705 0.8488
Case Study 4.1 (Two shells in series) 0.9724 0.9627 0.957
Table 4.1c: Principal construction parameters Ior the best design Ior Case Study 3.3
Construction parameter
1. Type oI exchanger Floating-head (TEMA AES)
2. Shell ID, in. (mm) 22.6 (575)
3. Tubes
292 nos., 0.7874 in. (20 mm) OD × 0.0787 in. (2.0 mm)
thick × 19.68 It (6000 mm) long
4. No. oI tube passes 2
5. Tube pitch, in. (mm) 1.024 (26), rotated square (45ƒ)
6. BaIIling single-segmental, 25° cut (diameter)
7. Nominal connection size, in. (mm) 4 (100) shellside; 6 (150) on tubeside
8. Heat transIer area, It
2
(m
2
) 2 × 1164 ÷ 2328 (2 × 108.2 ÷ 216.4)
9. Empty exchanger weight, lb (kg) 2 × 9000 ÷ 18,000 (2 × 4075 ÷ 8150)
65
and 7.9 in. (200 mm). It will be seen by comparison with Table 3.3c that the shellside
velocity, the shellside heat transIer coeIIicient, and (thereby) the overall heat transIer
coeIIicient are Iar higher in the present design. The stream analysis is also Iar better, with a
much higher main crossIlow Iraction and a much lower baIIle-shell leakage Iraction. This
translates into a signiIicantly higher temperature proIile distortion correction Iactor and,
thereby, a much higher MTD.
The design with the 7.9 in. (200 mm) baIIle spacing is the best, as it almost Iully utilizes
the permissible pressure drop, thereby producing the highest heat transIer coeIIicient and the
highest overdesign. In Iact, such a high overdesign (13.85°) is not required so that the shell
diamter and the number oI tubes can be reduced Iurther.
Note the considerably reduced total (Ior the two shells) heat transIer area in this design,
just 216.4 m
2
as against the area oI 363.9 m
2
oI the single-shell design depicted in Case
Study 3.3. The total empty weight correspondingly reduces Irom 26,300 lb (11,950 kg) to
18,100 lb (8200 kg).
Table 4.1d: Detailed perIormance results oI the best design Ior Case Study 3.3
Run 1 Run 2 Run 3 Run 4
1. BaIIle spacing, in. (mm) 13.8 (350) 11.8 (300) 9.8 (250) 7.9 (200)
BaIIle hole-to-tube 0.02 0.031 0.038 0.047
Main crossIlow 0.683 0.654 0.63 0.598
Bundle-shell 0.135 0.129 0.125 0.118
BaIIle-shell 0.161 0.186 0.208 0.237


2. Stream
Iractions
Pass-partition 0 0 0 0
3. Temperature proIile distortion
correction Iactor
0.989 0.987 0.985 0.982
velocity
cross/window
It/s (m/s)
0.69
(0.21)/3.1
(0.43)
0.8
(0.25)/3.1
(0.43)
1.0
(0.3)/3.1
(0.43)
1.25
(0.38)/3.1
(0.43)
pressure drop,
psi (kg/cm
2)

5.1 (0.36) 6.1 (0.43) 7.0 (0.49) 9.4 (0.66)



4. Shellside
heat transIer
coeIIicient,
Btu/h It
2
ƒF (kcal/h
m
2
ƒC)

66.9 (326.7)

71.3 (348.1)

72.0 (351.3)

76.8 (375.1)
5. Overall heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)

46.2 (225.4)

48.2 (235.4)

48.5 (236.8)

50.7 (247.4)
6. MTD, °F (°C) 41.6 (23.1) 41.6 (23.1) 41.4 (23.0) 41.4 (23.0)
7. Overdesign 4.53 8.91 9.35 13.85
66

67
CHAPTER 5
$OORFDWLRQRI6LGHV
6KHOOVLGHDQG7XEHVLGH
5.1 Introduction
This is a very specialized Iacet oI heat exchanger design and an area where vast
experience is particularly helpIul. It is more oI an art than a precise science because, as
we shall see, there are many Iactors which have to be considered simultaneously so that a
'Ieel¨ oI the situation is very important. For example, what is high viscosity in a given
situation? A given viscosity can be easily handled on the tubeside provided the Ilow rate
oI that Iluid is high enough, but what is high enough? A low MTD will lead to a
relatively high heat transIer area that will necessitate the use oI a rather high shell
diameter so that the given Ilow rate may be very low, but what Ilow rate is low Ior a
particular shell diameter? Heat duty, Ilow rate, viscosity, and MTD are all inextricably
linked in this complex task oI side allocation.
An expert system devised to handle such a situation would be extremely beneIicial.
However, it would be extremely diIIicult to make (considering the simultaneous variation oI
so many parameters) and would also take out a lot oI the enjoyment oI heat exchanger
design! Besides, it is a Iar happier situation when a designer learns the intricacies oI side
allocation Irom the very Iundamentals, because then she/he will be able to tackle any
situation, even one which she/he has not encountered earlier.
In many instances where the various parameters place contradictory demands upon side
allocation (as we shall see), the only way out is to design the heat exchanger both ways (hot
stream on shellside and hot stream on tubeside) and select the better design.
For many services, however, the allocation oI sides is quite straightIorward and even
established. For example, Ior cooling or condensing hydrocarbons, cooling water is
invariably placed inside the tubes. This is because cooling water is Iouling and it is easier to
clean the inside oI tubes than the outside. Besides, it is generally preIerred to have
condensing outside horizontal tubes as the condensate separates Irom the uncondensed vapor
by virtue oI gravity. However, should the process stream be corrosive and thereIore require
the use oI expensive alloy steel, an interchange oI sides may have to be examined. This is
because, with the corrosive stream on the shellside, more components will have to be made
Irom the expensive material.
5.2 Parameters for Allocation of Sides
The various Iactors that have to be considered Ior the allocation oI sides invariably place
contradictory demands on this very allocation, thereby making the selection oI sides a
68
very complicated aIIair. Let us examine each oI these Iactors in detail beIore considering
a real-liIe example Ior a better understanding.
5.2.1 Viscosity
Viscous Iluids are generally better handled on the shellside. The higher the viscosity, the
truer is this statement. This is because a considerably higher heat transIer coeIIicient is
obtained Ior the same pressure drop on the shellside (with a staggered tube layout
triangular or rotated square) than on the tubeside. The higher heat transIer coeIIicient is
essentially due to the increased turbulence which yields a higher Reynold`s number.
Evidently, an in-line tube layout will yield lower turbulence and, thereIore, a lower heat
transIer coeIIicient. The higher the viscosity oI the Iluid, the greater is the increase in the
eIIiciency oI conversion oI pressure drop to heat transIer coeIIicient Ior Ilow on the
shellside (with a staggered tube-Iield layout) as compared to Ilow on the tubeside.
The key issue here is turbulence, personiIied by the Reynold`s number. Evidently, mass
velocity, viscosity, and Ilow pattern are the crucial parameters Ior turbulence. For a given
velocity, a shellside allocation will give a much higher turbulence than a tubeside allocation,
especially with a staggered tube layout. However, iI the Reynold`s number is high enough, a
tubeside allocation may yield a satisIactory heat transIer coeIIicient.
II a high viscous Iluid is allocated on the shellside, a square or in-line tube-Iield layout
may give a satisIactory heat transIer coeIIicient, provided its Ilow rate is high enough to
yield a high Reynold`s number. II not, the use oI a staggered tube-Iield will be beneIicial.
Tubes arranged in a staggered layout are as easy to clean as tubes arranged in an in-line
layout, provided continuous cleaning lanes are incorporated.
5.2.2 Corrosiveness
II both the streams in a heat exchanger are benign and require only a base material oI
construction such as carbon steel, corrosiveness is not an issue Ior side allocation.
However, in many instances, one stream is noncorrosive and requires only a base
material oI construction while the other is corrosive and requires a superior material oI
construction such as alloy steel or stainless steel. II the corrosive Iluid is placed on the
tubeside, only the tubeside components (tubes, channel and channel cover, Iloating-head
cover, and tubeside tubesheet Iace) have to be oI the superior metallurgy. However, iI the
corrosive Iluid is on the shellside, the shellside components (tubes, shell, shell cover,
Iloating-head cover, shellside Iace oI tubesheet, and baIIles) have to be oI the superior
metallurgy. Thus, the comparison reduces to channel and channel cover Ior the tubeside
allocation versus shell, shell cover, and baIIles Ior the shellside allocation, the other
components being the same or commensurate. Considering identical geometry in both
casesidentical number oI tubes, tube diameter, tube thickness, tube length, and shell
diameterthe design with the corrosive stream on the shellside will be more expensive.
This is because the cost oI the shell, shell cover, and baIIles will be greater than that oI
the channel and channel cover. ThereIore, the more corrosive Iluid is preIerably placed
on the tubeside on this consiaeration alone.
5.2.3 Fouling tendency
As the extent oI Iouling is greater on the shellside due to the dead spaces, the diriter Iluid
is preIerably routed through the tubeside. Besides, aIter Iouling builds up to a certain
level, it becomes mandatory to clean the heat transIer surIaces. Mechanical cleaning (by
hydro-jetting, steaming, or rodding) oI the inside oI tubes is much easier than that oI the
outside oI tubes.
69
UnIortunately, dirty streams are invariably more viscous and more viscous streams are
much better handled on the shellside as explained above. Thus, in this situation, Iouling
tendency and viscosity impose contradictory demands Ior side allocation. II the dirty and
viscous stream is routed through the shellside, the initial cost oI the heat exchanger will be
lower due to the higher heat transIer coeIIicient but the operating cost will be higher due the
more Irequent (and more diIIicult) cleaning required. However, iI the dirty and viscous
stream is routed through the tubeside, the initial cost will be higher but the operating cost
lower. ThereIore, that routing which results in the lower overall cost (initial cost plus
operating cost) will be the more economical alternative. It should be realized here that the
initial or Iixed cost will depend not just the viscosities oI the streams, but the other
parameters being discussed presently.
5.2.4 Pressure
As with corrosiveness, the cost oI a heat exchanger will be less iI the higher pressure
stream is placed on the tubeside. This is because, with a tubeside allocation, Iewer higher
thickness and, thereIore, costlier parts will be required. Besides, and no less importantly,
tubes oI a given material, OD, and thickness can withstand much higher internal pressure
than external pressure. This means that when the shellside design pressure is rather high,
not only the shell and associated components, such as the shell cover, have to be thicker,
but even the tube thickness may increase Irom that required Ior the tubeside allocation.
However, it should be realized that in the low or intermediate pressure range (say, up to
356 psig or 25 kg/cm
2
g), the diIIerence in cost is not so pronounced as in the higher pressure
range. ThereIore, it is only iI the other parameters are Iairly close that pressure rating will
play a decisive role in the low-to-intermediate pressure range. However, iI the design
pressure oI one stream is low (say 213 psig or 15 kg/cm
2
g) and that oI the other high (say,
570 psig or 40 kg/cm
2
g or greater), the design pressure will play a crucial role.
5.2.5 Flow rate
For Ilow on the tubeside, the Reynold`s number is a Iunction oI mass velocity, tube
diameter, and viscosity. For Ilow on the shellside, the Reynold`s number is a Iunction oI
mass velocity, equivalent diameter, and viscosity. (The equivalent diameter is a Iunction
oI the tube diameter, tube pitch, and the layout angle.) Thus, Ior a given tube diameter (or
equivalent diameter), the Reynold`s number is a Iunction oI mass velocity and viscosity.
II a stream viscosity is high (say, 15 cp), the Reynold`s number is invariably condemned
to be low as the mass velocity cannot be increased suIIiciently to obtain a high Reynold`s
number. |The mass velocity should not be greater than a certain value, usually 400²450
lb/sec It
2
(1952²2196 kg/sec m
2
), as erosion may occur at higher velocities.| This is
particularly true on the tubeside. However, the Reynold`s number is oIten low, not because
the viscosity is high, but because the mass velocity is rather low. The two situations are quite
diIIerent: whereas the Prandtl number is quite high in the Iormer (high viscosity), it is lower
in the latter (low mass velocity). Since the heat transIer coeIIicient is a Iunction oI both
Reynold`s number and Prandtl number (see Section 3.3.2), the higher Prandtl number
situation will yield a higher heat transIer coeIIicient, provided the Reynold`s numbers are the
same in both cases.
It is well known that a suIIiciently high velocity has to be achieved in order to obtain a
satisIactory heat transIer coeIIicient. II the Ilow rate oI one oI the streams is low and iI it is
placed on the tubeside, then its velocity can be increased by incorporating a greater number
oI tube passes. OI course, this can be done up to a certain extent. For example, iI there are
70
only 100 tubes in a heat exchanger, the number oI passes can be six or eight or, at the most,
ten. The larger the shell diameter, the greater the number oI tubes and the higher the number
oI tube passes can be.
However, too many passes result in:
a) a cumbersome pass-partition arrangement with consequent diIIiculties in
Iabrication and maintenance and
b) an ineIIicient shellside Ilow pattern due to the large number oI pass-partition
lanes parallel to the Ilow direction.
Further, as the number oI passes is increased, the pressure drop also increases and may
exceed the allowable limit, so that this also becomes a limitation.
On the shellside, however, the only variables to increase the velocity Ior a given shell
diameter are baIIle spacing and baIIle cut. It was explained in Section 3.4.4 that very low
baIIle spacing and cut are detrimental to good thermal-hydraulic perIormance. Thus, the
baIIle spacing should not be less than 30²35° oI the shell diameter and should oIten be
higher. ThereIore, it is diIIicult to obtain a satisIactory velocity and, consequently, a
satisIactory heat transIer coeIIicient Ior a low-Ilow-rate stream on the shellside. II such a
stream is constrained to be on the shellside oI a heat exchanger, it sometimes becomes more
economical to have two, or even more, shells in series so that a much higher velocity and,
thereby, a much higher heat transIer coeIIicient can be obtained. This is explained in detail
with a case study each in Sections 6.2. and 6.3.
To summarize what has been said so Iar, a smaller Ilow rate requires a greater number
oI tube passes (on the tubeside) and a smaller baIIle spacing and cut (on the shellside) to
achieve a satisIactory mass velocity and, thereby, Reynold`s number. In both cases, this will
lead to a higher pressure drop so that, in certain cases, the same may exceed the allowable
value. In such cases, the number oI tube passes or the baIIle spacing/cut will have to be
restricted to such values that the pressure drop limitation is complied with.
Since turbulence exists on the shellside at much lower Reynold`s numbers than on the
tubeside, a low-Ilow-rate stream is usually better handled on the shellside. However, iI the
Ilow rate is extremely small, a tubeside allocation is oIten better because, on the shellside,
the baIIle spacing and cut cannot be decreased suIIiciently without producing an
unsatisIactory stream analysis.
It should be appreciated that Ilow rate and viscosity really go hand in hand. It may be
generalized that high-Ilow-rate streams can be handled satisIactorily on either shellside or
tubeside, intermediate Ilow-rate-streams on the shellside, and very small-Ilow-rate streams
on the tubeside. UnIortunately, high, intermediate, and low are subjective since they are
heavily dependent on the stream viscosity and exchanger size, so a more speciIic
generalization is not possible. In many instances, both designs oIten have to be carried out
and the more superior one selected aIter a careIul scrutiny oI all the relevant parameters.
There is another extremely interesting angle to this issue oI Ilow rates. Sometimes,
changing sides can result in a startling improvement in the design. Consider the Iollowing
exchanger service speciIied by a process licensor Ior an ethylene plant heat exchanger:
740,820 lb/h (336,034 kg/h) oI propylene to be cooled on the shellside Irom 102.2ƒF (39ƒC)
to 69.4ƒF (20.8ƒC) by 163,584 lb/h (74,201 kg/h) oI ethane/propane Ieed to be heated on the
tubeside Irom -45.04ƒF (-42.8ƒC) to 86ƒF (30ƒC), representing a heat duty oI 14.68 MM
Btu/h (3.7 MM kcal/h).
As speciIied above, the temperature cross condition would require the use oI two shells
in series since, due to low tubeside Ilow rate, two or more passes would be required to obtain
71
a satisIactory velocity and heat transIer coeIIicient. Since the propylene Ilow was very high,
an interchange oI Iluid sides was considered. Only one tube pass was now required to obtain
a satisIactory velocity, which meant that true countercurrent Ilow would prevail. The issue
oI temperature cross became redundant and a single shell was adequate to achieve the
speciIied duty. Needless to say, this resulted in considerable economy.
It is extremely important in the thermal design oI heat exchangers (as, indeed, in any
other creative activity) to constantly ask oneselI this simple question: Isn`t there a better way
oI doing this? Too oIten, we tread the beaten path and consequently miss all the wonderIul
opportunities to improve a design. II only we would involve ourselves more closely with the
task at hand, there would be many such opportunities in creativity.
5.2.6 Temperature range
This parameter is not very apparent to many designers, but oIten plays a very crucial role
in the allocation oI Iluids to shellside and tubeside. There are two considerations: large
temperature diIIerence and temperature proIile distortion correction Iactor.
Large temperature aifference
When a particular stream has a very large temperature range (the diIIerence between its
inlet and outlet temperatures) and the number oI tube passes is only two, it is better to
route the Iluid through the shellside. This is because, iI placed on the tubeside, this stream
will impose a very large temperature diIIerential between the two halves oI the tubesheet,
thereby making the tubesheet prone to distortion and the channel-tubesheet-shell joint
prone to leakage. The temperature diIIerence that is considered limiting is 176²194 ƒF
(80²90 ƒC) between two aafacent tube passes. II there are Iour or more passes, the
problem becomes much less severe as the temperature gradient across the Iace oI the
tubesheet is Iar more even. Also, iI there is only one tube pass, there is no problem at all
because then the entire Iace oI one tubesheet is subjected to the higher temperature and
the entire Iace oI the other tubesheet is subjected to the lower temperature.
Temperature profile aistortion correction factor
According to this consideration, a Iluid having a large temperature diIIerence is better
placed on the tubeside as it may result in a low temperature proIile distortion correction
Iactor iI placed on the shellside. The latter was discussed at length in Section 4.7. As
mentioned therein, besides a large shellside temperature diIIerence, two other Iactors
must also exist in order to produce a low temperature proIile distortion correction Iactor:
1) the Iluid is rather viscousthe higher the viscosity, the worse the situation
because the higher the baIIle-shell leakage streamand
2) the ratio oI the shellside temperature diIIerence to the temperature approach
between the hot and cold streams at the shellside outlet is low
CASE STUDY: 5.1 ALLOCATION OF FLUID SIDES
Let us consider a real-liIe example to better understand the philosophy oI allocation oI
Iluid sides. The principal process parameters Ior a heat exchanger are elaborated in Table
5.1a.
Table 5.1b lists the various side allocation parameters as they apply in this case. It will
be seen that two parameters Iavor allocation oI the cold stream on the shellside, two
parameters Iavor allocation oI the cold stream on the tubeside, while two parameters Iavor
72
neither. What will Iinally determine the Iavorable allocation oI sides is the overall cost oI the
two designs: the cold stream on the shellside and the cold stream on the tubeside. By overall
cost is meant the sum oI the initial cost and the operating cost.
It is signiIicant to note here that since the terminal viscosities oI the two streams are very
similar, it (viscosity) does not have a strong inIluence on the allocation oI sides. The
materials oI construction, design pressure, Iouling resistance, and stream Ilow rates have a
much more proIound inIluence. Whereas the Iouling resistance oI the two streams aIIects the
operating cost, the other three parameters aIIect the initial cost. The eIIects oI the materials
oI construction and the design pressures on the initial cost are quite straightIorward (as
explained earlier), but that oI the Ilow rates is not.
It has been explained earlier that with a low Ilow rate on the shellside, even the smallest
recommended baIIle spacing and baIIle cut that yield a satisIactory stream analysis may Iail
to produce a suIIiciently high velocity. Consequently, the allowable pressure drop is not
utilized and the heat transIer coeIIicient is rather low. Whether 209,437 lb/h (95,000 kg/h) is
Table 5.1a: Principal process parameters Ior Case Study 5.1
Cold stream Hot stream
1 Flow rate, lb/h (kg/h) 440,920 (200,000) 209,437 (95,000)
2. Temperature in/out, °F (°C) 455 (235)/518 (270) 644 (340)/518 (270)
3. Viscosity in/out, cp 0.2/0.1 0.2/0.3
4. Density in/out, lb/It
3
(kg/m
3
) 45.86 (735)/43.37 (695) 41.5 (665)/44.93 (720)
5. SpeciIic heat, Btu/lb °F (kcal/kg °C) 0.67 (avg.) 0.7 (avg.)
6. Thermal conductivity in/out, Btu/h It °F
(kcal/h m °C)
0.0537 (0.08)/0.05 (0.074) 0.043 (0.064)/0.048 (0.072)
7. Fouling resistance, h It
2
°F /Btu
(h m
2
°C/ kcal)
0.00293 (0.0006) 0.00195 (0.0004)
8. Heat duty, MM Btu/h (MM kcal/h) 18.61 (4.69)
9. Design pressure, psig (kg/cm
2
g) 610 (42.9) 190 (13.4)
10. Design temperature, °F (°C) 554 (290) 680 (360)
11. Nominal line size, in. (mm) 10 (250) 8 (200)
12. Material oI construction CS 5 Cr 1/2 Mo
Table 5.1b: Fluid side allocation considerations Ior Case Study 5.1
Preferred allocation
1. Viscosity No preIerence as viscosities oI both streams are very similar
2. Corrosiveness Cold stream on shellside since hot stream requires a superior metallurgy
3. Fouling nature Cold stream on tubeside as its Iouling resistance is higher
4. Pressure Cold stream on tubeside as its design pressure is much higher
5. Flow rate Cold stream on shellside as the hot stream Ilow rate is much lower
6. Temperature range No preIerence as the temperature range oI neither stream is very high
73
a low Ilow rate or not will depend upon the diameter oI the shell through which it will Ilow.
Since it is usually not possible to anticipate the shell diameter without perIorming a design,
both designs (cold stream on shellside and cold stream on tubeside) will have to be
perIormed in order to decide the cheaper design.
In the present instance, the conIlicting demands made by the diIIerent parameters as
shown in Table 5.1b make it impossible to decide on the Iavorable side allocation per se.
Hence, both designs were carried outcrude oil on shellside and crude oil on tubesideand
are shown in Table 5.1c. Tubes 0.984 in. (25 mm OD) × 0.0787 in. (2 mm) thick × 19.68 It.
(6000 mm) long were required to be used. Although there was no temperature cross, it was
Iound that two shells in series had to be used to utilize the allowable pressure drop oI
whichever stream was routed through the shellside.
The Iollowing observations can be made:
1) For the design having the cold stream on the shellside, velocities oI both the
streams are satisIactory, 3.71 It/s (1.13 m/s) Ior the cold stream and 6.33 It/s
(1.93 m/s) Ior the hot stream.
2) For the design having the cold stream on the tubeside, whereas the velocity oI the
cold stream is satisIactory at 5.77 It/s (1.76 m/s), that oI the hot stream is some-
Table 5.1c: Thermal designs Ior Case Study 5.1
Cold stream
on shellside
Cold steeam
on tubeside
1. No. oI shells in series 2 2
2. Shell diameter, in. (mm) 26.6 (675) 27.6 (700)
3. Total no. oI tubes per shell 250 280
4. No. oI tube passes 4 2
5. BaIIle spacing, in. (mm) 13.8 (350) 7.9 (200)
6. BaIIle cut (diameter), ° 25 22
7. Heat transIer area, It
2
(m
2
)
2 × 1237 (115)
÷ 2474 (230)
2 × 1400 (130)
÷ 2800 (260)
velocity, It/s (m/s) 3.71 (1.13) 5.77 (1.76)
pr. drop, psi (kg/cm
2
) 11.4 (0.8) 6.5 (0.46)

8. Cold
stream
heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
500.7 (2445) 422.9 (2065)
velocity, It/s (m/s) 6.33 (1.93) 2.4 (0.73)
pressure drop, psig (kg/cm
2
) 11.3 (0.96) 8.1 (0.57)

9. Hot
stream
heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
331.2 (1617) 277.3 (1354)
10. Overall heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
92 (449.1) 82.9 (404.5)
11. Approx. bundle/empty weight per shell,
lbs (kgs)
2500/6700 2800/7300
12. Overdesign 5.62 5.6
74
what low at 2.4 It/s (0.73 m/s). Although the cold stream is the dirtier stream and
will thereIore be better routed through the tubeside per se, it results in a
somewhat low velocity oI the hot stream, which is not a clean stream either. As
velocity plays a dominant role in Iouling, it is preIerable to have a satisIactory
velocity Ior each stream, even iI it means routing the dirtier stream through the
shellside.
3) When the design having the cold stream on the shellside is compared to that
having the cold stream on the tubeside, the Iollowing are observed:
a) The total heat transIer area is about 11.5° lower.
b) The bundle and total empty weights (Ior each shell) are 10.7° and 8.2°
lower, respectively. There are two opposing tendencies here: i) the much
higher design pressure on the shellside and ii) the reduced number oI tubes
and shell diameter. While the Iormer tended to increase the weight, the latter
tended to decrease it. Since the eIIect oI the smaller size was more
predominant, the exchanger weight reduced.
4) Thus, the Iirst design having the cold stream on the shellside is cheaper since not
only is its weight signiIicantly lower, it also has the more Iavorable designation
oI materials oI constructionthe corrosive stream is on the tubeside. Besides, it
yields satisIactory velocities Ior both the streams so that even the operating cost
will be lower. Hence, this design was adopted as the more economical one.
In many instances, as in the present, it is diIIicult to establish which allocation oI sides
will produce the cheaper design without both the designs being carried out. In some
instances, it is diIIicult to establish which design is cheaper (even aIter the thermal
designs are produced) until some mechanical design calculations or estimates are carried
out. In this context, it should be noted that, although all sophisticated heat exchanger
thermal design soItware packages currently available incorporate subroutines to carry out
calculations to determine the tube bundle and empty exchanger weights, the results are
usually approximate. ThereIore, when the weight estimates are rather close Ior the two
designs, more elaborate weight calculations should be carried out.


75
CHAPTER 6
0HWKRGRORJ\RIWKH8VH
RI0XOWLSOH6KHOOV
Multiple shells may be used either in series or in parallel, or in a combination oI both, to
handle situations that cannot be handled at all by single shells, or handled well by them.
This chapter discusses the modalities oI the use oI multiple shells with two case studies to
illustrate the same.
6.1 Multiple Shells in Parallel
Multiple shells are used in parallel primarily when the required heat transIer area is too
large to be accommodated in a single shell. This is really applicable to Iloating-head or
U-tube heat exchangers which have removable tube bundles and are, thereIore, limited by
crane handling capacities. In large modern reIineries and other chemical plants, tube
bundle weights are usually limited to 22,000 lbs (10,000 kgs), which translates to a
4850²5400 It
2
(450²500 m
2
) heat transIer area. Larger plants permit 15²20 ton tube
bundles. Since Iixed-tubesheet heat exchangers do not have removable tube bundles,
bundle weight is not a limitation Ior them and, consequently, very large heat transIer
areas (21,500 It
2
or 2000 m
2
, and even more) can be accommodated in a single shell. The
only limitations here are Iabrication capabilities, transportation Iacilities, and plot area
limitations.
Another reason to use multiple shells in parallel is Ior the sake oI control. When there is
more than one case oI operation and where the stream Ilow rates (and heat duties) Ior the
diIIerent cases vary considerably, the velocity may be unacceptably low Ior the lower Ilow
rate case. |Low velocities lead to excessive Iouling; see Sections 11.3, 11.6.2.1 (b), and
11.6.2.2. (d)|. A low velocity may also be experienced in a turndown condition. In order to
address this problem, multiple shells in parallel may be used and one or more shells by-
passed Ior the low velocity conditions.
The low velocity may be experienced on the shellside, tubeside, or both. Evidently, Ior
a turndown condition, the velocity will be low Ior both the shellside and tubeside.
Let us demonstrate this with an example. Let us consider a heat exchanger which has to
be designed Ior two cases, representing two diIIerent Ieedstocks. For the sake oI simplicity,
let us consider that Ior Case B, both the shellside and tubeside Ilow rates, as well as heat
duty, are 30° oI the values Ior Case A, the design condition. Evidently, Case A is
controlling and a suitable thermal design can be produced. Consider that the shellside
velocity is 2.3 It/s (0.7 m/s). For Case B, this will reduce to (0.3)(2.3) or 0.69 It/s |(0.3)(0.7)
or 0.21 m/s|, which is rather low and may result in excessive Iouling.
Now consider that, instead oI a single shell, the original design is conIigured with two
76
shells in parallel. Evidently, the shellside velocity will be somewhat lower than 2.3 It/s (0.7
m/s) as the net Ilow area will be somewhat higher, say 1.8 It/s (0.55 m/s). Now, since the
heat transIer area Ior Case B is rather low, one oI the two shells may be bypassed. Thus, the
shellside velocity will be (1.8)(0.3)(2) ÷ 1.08 It/s |(0.55)(0.3)(2) ÷ 0.33 m/s|, which is
signiIicantly higher than 0.69 It/s (0.21 m/s). The multiplier 2 is Ior the Iact that Ilow is now
taking place through only one shell and not two shells in parallel.
6.2 Multiple Shells in Series
Multiple shells in series are used principally to handle temperature cross conditions but
there are many situations when their use becomes advantageous in order to utilize the
permitted pressure drop more eIIectively. We will now look at both these situations.
6.2.1 For temperature cross conditions
A temperature cross condition occurs when the outlet temperature oI the cold stream is
greater than the outlet temperature oI the hot stream. This condition is quite common Ior
heat recovery applications such as Ieed/bottom heat exchangers Ior reactors, desalters,
distillation columns, etc. From elementary thermodynamics, a single 1²2 heat exchanger
(one shell pass, two or more tube passes) cannot achieve a temperature cross condition
because one pass is countercurrent and the other co-current. (A 1²1 pure countercurrent
heat exchanger can achieve a temperature cross condition but heat exchangers invariably
require more than one tube pass in order to achieve a satisIactory tubeside velocity and,
thereby, a satisIactory tubeside heat transIer coeIIicient.) The best that a 1²2 heat
exchanger can achieve is an equal outlet temperature condition; that is, the cold Iluid and
hot Iluid outlet temperatures are equal.
Thus, iI a temperature cross condition is to be achieved, multiple shells in series have to
be employed. The number oI shells in series will depend upon the extent oI the temperature
cross; that is, the diIIerence between the shellside and the tubeside outlet temperatures. The
higher the temperature cross, the greater the number oI shells required in series.
In order to determine the number oI shells required Ior a particular temperature cross
application, a simple construction oI drawing horizontal and vertical lines is all that is
required. The construction may begin Irom either the hot or the cold end and carried on till
the entire temperature proIile is covered. II such an attempt is made, it will be seen that iI the
construction is started at the cold end, the Iirst horizontal line will have to be drawn Irom the
hot stream temperature proIile. II, however, the construction is started at the hot end, the Iirst
horizonal line will have to be drawn Irom the cold stream temperature proIile. This
construction method is applicable not only to linear heat duty-temperature proIiles, but to
nonlinear ones as well. Evidently, since a Iractional shell cannot be provided (!), the next
higher whole number oI shells has to be employed in series.
This can be appreciated graphically in Fig. 6.1. In Fig. 6.1b, where the temperature cross
is moderate, two shells in series are adequate. However, in Fig. 6.1a, which has a substantial
temperature cross, Iour shells are required in series. It may be seen that each shell oI a
multiple-shells-in-series application achieves equal outlet temperatures. Thus, in Fig. 6.1b,
Shell 1 has a shellside and tubeside outlet temperature oI 446ƒF (230ƒC), Shell 2 has a
shellside and tubeside outlet temperature oI 352.4ƒF (178ƒC), and so on.
6.2.2 For better utilization of allowable pressure drop
As elaborated above, multiple shells are generally used when there is a temperature cross
77
(cold stream outlet temperature greater than hot stream outlet temperature) and pure
countercurrent Ilow cannot be achieved. However, there are occasions when the shellside
Ilow rate is so low that with a single shell and baIIles placed as close as Ieasible (25-30°
oI the shell inside diameter), the pressure drop is exceedingly low and so too is the heat
transIer coeIIicient. In such cases, it would be prudent to try two (or even more) shells in
series as it may result in a cheaper design by achieving a much higher shellside heat
transIer coeIIicient. Please see Case Study 11.1 in Chapter 11.
6.2.3 For improving the temperature profile distortion correction factor
Yet another reason Ior using two or more shells in series is to improve the temperature
proIile distortion correction Iactor. This was discussed in detail in Section 4.7 and,
thereaIter, in Case Study 4.1, it was demonstrated how a temperature proIile distortion
problem can be better handled by the use oI two shells in series.
CASE STUDY 6.1: USE OF MULTIPLE SHELLS IN SERIES
A cooler was to be designed Ior cooling a light hydrocarbon in an oil reIinery. The princi-
pal process parameters are speciIied in Table 6.1a. Carbon steel tubes 0.984-in. (25-mm)
OD × 0.0984-in. (2.5-mm) thick × 19.68-It. (6000-mm) long were required to be used.
A design was prepared, using only one shell. Even with a small baIIle spacing and cut,
the shellside velocity and, thereby, the heat transIer coeIIicient were very low, leading to a
very large heat transIer area and shell diameter. The allowable shellside pressure drop was
highly underutilized |0.85 psi versus 10 psi (0.06 kg/cm
2
versus. 0.7 kg/cm
2
)|. Any Iurther
reduction in baIIle spacing and cut resulted in an unacceptable stream analysis with large
leakage streams, especially baIIle-to-shell.
Further, since the outlet temperatures were equal, the F
t
Iactor was only 0.8. With two
shells in series, this would be substantially higher. Yet another drawback was the relatively
low temperature proIile distortion correction Iactor oI 0.907, caused by a not very
distinguished stream analysis and a relatively high ratio oI shellside temperature diIIerence
to the temperature approach at the shell outlet |(212 113)/(113 91.4) or (100 45)/(45
33) or 4.58|.

(a) (b)
Fig. 6.1 Determination oI the number oI shells in series
78
In an eIIort to increase the shellside velocity and, thereby, its heat transIer coeIIicient, a
design was carried out using two shells in series. As expected, there was a considerable
improvement in the shellside perIormance. The stream analysis was Iar better and the
shellside velocity considerably higher, thereby resulting in a 93° increase in the shellside
heat transIer coeIIicient. The overall heat transIer coeIIicient went up Irom 75.9 Btu/h It
2
ƒF
(370.6 kcal/h m
2
ƒC) to 108.6 Btu/h It
2
ƒF (530.2 kcal/h m
2
ƒC), an increase oI 43°.
Further, due to a signiIicant increase in both the F
t
Iactor and the temperature proIile
distortion correction Iactor, the MTD increased by 29°. Thus, the overall heat transIer area
reduced Irom 3034 It
2
(282 m
2
) to 1720 It
2
(160 m
2
), a reduction oI 43° and the total empty
weight Irom 33,700 lb (15,300 kg) to 18,000 lb (8000 kg), a reduction oI 48°. The shellside
pressure drop was Iar better utilized at 6.0 psi (0.42 kg/cm
2
), Iar closer to the allowable limit
oI 7.1 psi (0.5 kg/cm
2
).
Evidently, the two-shell design was adopted since its initial cost would be signiIicantly
lower than the single-shell design. The salient Ieatures oI both designs are elaborated in
Table 6.1b.
6.3 Multiple Shells in Series/Parallel
The use oI multiple shells in series or in parallel is quite straightIorward. However,
situations sometimes arise when neither a series nor a parallel arrangement can be used
and a series/parallel arrangement becomes unavoidable. This arrangement is shown in
Fig. 6.2 and is illustrated by the Iollowing real-liIe situation in a reIinery application.
CASE STUDY 6.2: USE OF MULTIPLE SHELLS IN SERIES/PARALLEL
A hydrocarbon liquid cooler had to be designed Ior the duty speciIied in Table 6.2a. As
the liquid had to be cooled through a large temperature range, it became imperative to
Table 6.1a: Principal process parameters Ior Case Study 6.1
Shellside Tubeside
1. Fluid Light hydrocarbon Cooling water
2. Flow rate, lb/h (kg/h) 145,504 (66,000) 394,039 (178,735)
3. Inlet/outlet temperature, °F (°C) 212 (100)/113 (45) 91.4 (33)/113 (45)
4. Heat duty, MM Btu/h (MM kcal/h) 8.5 (2.1417)
5. Density in/out, lb/It
3
(kg/m
3
) 38.06 (610)/42.4 (680)
6. Viscosity in/out, cp 0.35/0.7
7. Average speciIic heat, Btu/lb °F (kcal/kg °C) 0.59
8. Thermal conductivity in/out, Btu/h It °F
(kcal/h m °C)
0.0585 (0.087)/0.064 (0.095)


standard
9. Allowable pressure drop, psi (kg/cm
2
) 7.1 (0.5) 10.7 (0.75)
10. Fouling resistance, h It
2
°F/Btu (h m
2
°C/kcal) 0.00147 (0.0003) 0.00195 (0.0004)
11. Design pressure, psig (kg/cm
2
g) 206 (14.5) 92 (6.5)
12. Material oI construction CS CS
13. Nominal line size, mm 150 200
79
employ two shells in series to reduce the adverse eIIect oI the shellside leakage streams
(especially the shell-to-baIIle leakage stream) on the temperature proIile and, thereby, the
MTD. UnIortunately, while the shellside pressure drop could be contained within the
allowable limit with two shells in series, the tubeside pressure drop could not, even with
Table 6.1b: Salient Ieatures oI two designs Ior Case Study 6.1
Design with a
single shell
Design with 2 shells
in series
1. Shell ID, in. (mm) 39.8 (1010) 21.7 (550)
2. Heat transIer area per shell, It
2
(m
2
) 3034 (282) 860 (80)
3. Total heat transIer area, It
2
(m
2
) 3034 (282) 1720 (160)
4. Number oI tubes per shell 620 172
5. No. oI tube passes 6 2
6. Tube pitch, in. (mm) 1.26 (32) square
7. BaIIle spacing, in. (mm) 12.6 (320) 9.1 (230)
8. BaIIle cut, ° diameter 22 25
9. Shellside velocity, It/s (m/s) 0.76 (0.23) 2.2 (0.66)
BaIIle hole-tube leakage (A) 0.212 0.15
Main crossIlow (B) 0.471 0.536
Bundle-shell bypass (C ) 0.078 0.158
BaIIle-shell leakage (E) 0.195 0.156
Pass-partition bypass (F) 0.045 0



10. Stream
analysis
Overall shellside stream eIIiciency 0.617 0.685
11. Shellside heat transIer coeIIicient, Btu/h It
2
°F
(kcal/h m
2
°C)
125.5 (612.9) 242 (1182)
12. Shellside pressure drop, psi (kg/cm
2
) 0.85 (0.06) 6.0 (0.42)
13. Overall heat transIer coeIIicient, Btu/h It
2
°F
(kcal/h m
2
°C)
75.9 (370.6) 108.6 (530.2)
F
t
0.8 0.97
Temperature proIile distortion
correction Iactor
0.907 0.988

14. MTD
Corrected MTD, ƒF (ƒC) 37.8 (21.0) 48.8 (27.1)
15. Overdesign 2.88 6.7
16. Approx. bundle weight, lbs (kgs) 14,300 (6500)
3750 (1700) × 2
÷ 7500 (3400)
17. Approx. empty weight, lbs (kgs) 33,700 (15,300)
9000 (4000) × 2
÷ 18,000 (8000)
80
two tube passes, which is the minimum Ior a
split-ring pull-through Iloating-head construc-
tion. The cooling water Ilow rate could not be
reduced as its outlet temperature would then
exceed 113ƒF (45ƒC) which would result in
excessive Iouling (scaling).
The only Ieasible solution, thereIore, was to
have the shellside connected in series and the
tubeside connected in parallel, as shown in Fig.
6.2. Being a somewhat unusual situation which
cannot be handled in a single run on any com-
mercial heat exchanger thermal design soItware
package, it had to be designed careIully.
As it was desirable to have the two shells oI
the same size, it was evident that the hotter shell
would deliver a much higher heat duty than the
colder shell by virtue oI:
1) the higher MTD and
2) the higher shellside (and, thereby, overall) heat transIer coeIIicient, due to the
lower shellside viscosity.
Consequently, the two shells were run individually on a proprietary computer program
several times with diIIerent intermediate temperatures oI the hydrocarbon liquid between
the two shells until the two shells could be oI identical shell ID and tube length, and more
or less the same degree oI overdesign. The intermediate hydrocarbon liquid temperature
Iinally came to 235.4ƒF (113ƒC) and corresponded to 76° heat duty in the hotter shell
Fig. 6.2 Series/parallel arrangement
Table 6.2a: Principal process parameters Ior Case Study 6.2
Shellside Tubeside
1. Fluid Hydrocarbon liquid Cooling water
2. Flow rate, lb/h (kg/h) 132,300 (60,000) 1,032,400 (468,300)
3. Inlet/outlet temperature, °F (°C) 464 (240)/158 (70) 89.6 (32)/113 (45)
4. Allowable pressure drop, psi (kg/cm
2
) 10 (0.7) 10 (0.7)
5. Fouling resistance, h It2 °F/Btu
(h m
2
°C /kcal)
0.00293 (0.0006) 0.00195 (0.0004)
6. Viscosity in/out, cp 1.18/39.0
7. Density in/out, lb/It
3
(kg/m
3
) 49.1 (787)/55.3 (887)
8. Thermal conductivity in/out, Btu/h It °F
(kcal/h m °C)
0.056 (0.083)/0.07
(0.1044)
9. SpeciIic heat in/out, Btu/lb °F (kcal/kg °C) 0.636/0.556


Standard
10. Heat duty, MM Btu/h (MM kcal/h) 24.13 (6.08)
11. Design pressure, psig (kg/cm2g) 249 (17.5) 92 (6.5)
12. Material oI construction CS Admiralty Brass
81
and the balance 24° in the colder one. The detailed construction and some perIormance
parameters are shown in Table 6.2b. The principal reason Ior this was that the MTD in
the hot shell is 158° higher than that oI the cold shell; the overall heat transIer
coeIIicient was only marginally higher at 13.4°.
As already stated, the cooling water was to Ilow in parallel through the two shells.
However, iI the total cooling water Ilow rate were to be divided equally between the two
shells, the cooling water leaving the hotter shell would be much higher than 113ƒF or 45ƒC
whereas that leaving the colder shell would be much less than 113ƒF or 45ƒC (which was the
combined outlet temperature). As cooling water scaling is very pronounced at higher
temperatures, this would have led to heavy Iouling in the hotter shell. In order to avoid this
problem, the cooling water Ilow rate was divided to the two shells in the same ratio as their
heat duties, namely 3:1, so that the combined outlet temperature oI the two shells would be
113ƒF or 45ƒC.
Table 6.2b: Salient construction and perIormance Ieatures oI hydrocarbon liquid cooler
Hot shell Cold shell
1. Type oI shell Floating-head (TEMA AES)
2. Shell ID, in. (mm) 26.6 (675)
3. Tube OD × thickness × length, mm 0.7874 (20) × 0.0787 (2) × 236 (6000)
4. Tube pitch, in. (mm) 1.024 (26) rotated square
5. BaIIle spacing, in. (mm) 7.9 (200)
6. BaIIle cut, ° diameter 22
7. Number oI tubes × no. oI tube passes 418 × 2 384 × 4
8. Heat transIer area, m
2
1667 (154.9) 1531 (142.3)
9. Shellside temp. in/out, °F (°C) 464 (240)/235.4 (113) 235.4 (113)/158 (70)
10. Heat duty, MM Btu/h (MM kcal/h) 18.34 (4.62) 5.79 (1.46)
11. Cooling water Ilow rate, lb/h (kg/h) 784,400 (355,800) 248,000 (112,500)
12. Tubeside velocity, It/s (m/s) 7.78 (2.37) 5.35 (1.63)
13. Shellside crossIlow/window velocity,
It/s (m/s)
1.4 (0.42)/1.6 (0.49) 1.3 (0.4)/1.5 (0.45)
Shellside 80.9 (395) 68.4 (334)
Tubeside 1479 (7221) 1096 (5350)
14. Heat transIer
coeIIicient,
Btu/h It
2
°F (kcal/h
m
2
°C)
Overall 53.9 (263) 47.5 (232)
15. MTD, ƒC 227.2 (126.2) 88.0 (48.9)
16. Overdesign, ° 11.4 10.6
Shellside 2.8 (0.2) 4.4 (0.31)
17. Pressure drop, psi
(kg/cm
2
)
Tubeside 9.5 (0.67) 9.7 (0.68)
Shellside 6 (150) 6 (150)
18. Nominal nozzle
size, in. (mm)
Tubeside 10 (250) 6 (150)
82
Normally, it would have been necessary to employ a Ilow controller to achieve this split
oI Ilow rate to the two shells. However, it was discovered by a happy coincidence that by
having Iour tube passes in the hotter shell and two tube passes in the colder one, the cooling
water pressure drop in both the shells was virtually identical. The principal parameters oI the
Iinal design are shown in Table 6.2c. It may be noted that the hot shell has 418 tubes while
the cold shell has 384 since there are Iour tube passes in the latter and two in the Iormer.

Table 6.2c: Principal parameters oI Iinal design Ior Case Study 6.2
Number of shells 2 (Shellside in series, Tubeside in parallel)
1. Heat transIer area, It
2
(m
2
)
3174 (295)
hot shell: 1614 (150), cold shell: 1560 (145)

2. Tube details
0.7874 in. (20 mm) OD, 0.0787 in. (2 mm) thick,
19.68 It. (6000 mm) long; 404 tubes
in hot shell and 392 tubes in cold shell.
3. Number oI tube passes
2 in hot shell
4 in cold shell
4. BaIIle spacing, in. (mm)
7.0 (178)
5. BaIIle cut
22° diameter (horizontal)
6. Cooling water velocity, It/s (m/s)
8.0 (2.43) in hotter shell
5.54 (1.69) in colder shell
7. Cooling water pressure drop, psi (kg/cm
2
)
10 (0.7)
8. Hydrocarbon liquid pressure drop, psi (kg/cm
2
)
9.0 (0.63) |3.6 (0.25) in hot shell
and 5.4 (0.38) in cold shell)
9. MTD, ƒF (ƒC)
Hot shell: 219.4 (121.9), cold shell: 87.3 (48.5)
10. Overall heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
Hot shell: 57.6 (281.1), cold shell: 49.2 (240)
83
CHAPTER 7
7KHUPDO'HVLJQDQG2SWLPL]DWLRQ
RI&RQGHQVHUV
7.1 Introduction
The thermal design oI condensers is a Iascinating subject as there is a considerable
variation in service, construction, operating pressure, condensing range, and even the
choice oI coolant. Due to large-scale research carried out in recent years, the phenolme-
non oI condensing is now quite well understood and this sophisticated knowledge is
embodied in several proprietary soItware packages. UnIortunately, there is a lot to be de-
sired in the application oI these soItware packages in the practical design oI condensers,
both with their speciIication by process licensors and their actual design by engineering
companies and/or Iabricators. As with any complex subject, it is important to grasp the
Iundamentals and the interplay oI parameters in order to not only enjoy the activity but to
produce eIIicient and optimized designs as well. The accent in the Iollowing sections will
thereIore not be on a plethora oI equationsoI which there is a preponderance in the
published literature (see ReIs. |16|)but rather on their application Ior optimum design.
7.2 Classification
7.2.1 According to construction
It is usually preIerred to have condensation on the shellside oI a heat exchanger. There
are two important reasons Ior this. The Iirst is that it is easier to separate the condensate
Irom the uncondensed vapor on the shellside by the gravitational Iorce as the cross-
sectional area is much higher and the number oI streams much less than on the tubeside.
The second is that water is the cooling medium in the vast majority oI condensers
(although air has become increasingly popular, the use oI water is still more prevalent)
and is preIerably routed through the tubeside as it is more corrosive as well as Iouling.
However, in the petrochemical and specialty chemical industries, the process Iluid is
oIten corrosive, thereby requiring alloy steel construction. In such instances, it is more
economical to route the process stream through the tubeside Ior reasons oI economy, as
explained in Section 5.2.2. In some other instances, the process Iluid is at high pressure;
consequently, it is again better routed through the tubeside as explained in Section 5.2.4.
A comprehensive selection guide is presented in |1|.
7.2.2 According to layout
Condensers, as with other shell-and-tube heat exchangers, are preIerred to be mounted
horizontally Ior ease oI maintenance, especially Ior exchangers having removable tube
bundles. It is much easier to remove a horizontal tube bundle than a vertical one, espe-
84
cially iI it is large. However, in certain special situations, it is Iar better to have a vertical
construction. A common example is a reIlux or a knock-back condenser. Here, vapor
Ilows up inside tubes and condensate Ialls down progressively, thereby eIIecting better
separation oI condensate Irom uncondensed vapor or noncondensables. II desirable (Ior
reasons mentioned above) to have condensation on the shellside, a vertical construction
can still be employed. Care will have to be taken in such situations to prevent bypassing
oI the condensing vapors around the tube bundle. This is usually achieved by the
judicious application oI a sealing arrangement.
It should be realized that a vertical construction results in a considerable saving in plot
area. Consequently, when a condensing stream is routed through the shellside and is
relatively non-Iouling so that no cleaning is required, a Iixed-tubesheet construction can be
employed in a vertical position. A vertical construction can also be employed Ior condensing
vapors inside tubes.
In either case, iI the condensation is total, vapor can enter at the top oI the condenser and
Ilow downward. Thus, it is extremely common to have steam condense on the shellside oI
vertical thermosyphon reboilers and Ilow Irom top to bottom. For a vertical partial
condenser, however, it will be essential to have the vapor Ilow up and the condensate Ilow
down to eIIect separation.
7.2.3 According to service
Condensation can either be total or partial. This will depend upon the service and the
operating conditions oI pressure and temperature. A pure component will condense
isothermally and will invariably be totally condensed. A mixture oI components or a
mixture oI a condensable and a noncondensable can be condensed either totally or
partially, depending upon the process requirements.
7.2.4 According to coolant
As stated earlier, water is the most commonly used cooling medium due to its historical
easy and plentiIul availability. However, due to statutory environmental pollution
abatement requirements, the cost oI water treatment has increased considerably. Besides,
the availability oI water has dwindled over the recent past, so that air is now invariably
the Iirst choice as a cooling medium Ior condensation.
It should be mentioned here that, should the condensing stream be at a suIIiciently high
temperature, it can be condensed against another suitable process stream that has to be
heated or vaporized. AIter process-to-process condensing has been maximized, or rather,
optimized, it becomes essential to reject any residual latent heat oI condensation to the
atmosphere.
As ambient temperature is higher than that oI water, water can condense a stream to a
lower temperature than can air. Thus, it becomes logical to employ air Iirst and then water as
a cooling medium. The break-temperature between air and water cooling has to be
established by a detailed study wherein the total cost (equipment cost and operating cost) is
minimized. Usually, the break-temperature is 2732ƒF (1518ƒC) higher than the design
ambient temperature. This will depend upon the total cost oI water (including the cost oI
treatment, make-up, piping/valving, and apportioned cost oI a cooling tower), the cost oI
equipment, and the cost oI energy.
In many situations, such as in ethylene plants, condensation is required to be carried out
at temperatures below ambient. Evidently, air or water cannot be used in these situations and
the use oI a reIrigerant becomes essential. In Iact, several levels oI reIrigeration exist in
85
ethylene plants, depending upon the temperature level oI the process streams.
7.2.5 According to condensing range
II a condensing vapor is a pure component, condensation will evidently be isothermal.
However, iI the condensing stream is a mixture oI various components, the condensation
will be accomplished over a temperature range. The wider the mixture (in terms oI
volatility), the wider is the condensing range.
Another case is when a pure component has a noncondensable associated with it, in
which case condensation oI the condensing component is initially quite rapid (meaning over
a small temperature drop), slowing down progressively as the percentage oI noncondensable
increases. OI course, there could be an initial desuperheating zone iI the stream is not
saturated. An example is a steam-jet ejector condenser where a mixture oI predominantly
steam associated with some heavy hydrocarbon vapors and leakage air enters a condenser.
AIter a large desuperheating zone, condensation oI steam is very rapid but Ialls abruptly as
the concentration oI noncondensables increases dramatically.
Quite a diIIerent situation is encountered in an air compressor intercooler, where a very
small percentage oI atmospheric water vapor enters the condenser along with bulk air which
has to be cooledas the air is cooled, some associated water vapor condenses. Vapor shear
is usually very high in such cases due to the high vapor weight Iraction.
Evidently, the determination oI the heat transIer coeIIicient is much easier Ior pure-
component isothermal condensation than Ior condensation oI a mixture through a
temperature range. In the latter case, sensible vapor cooling and diIIusion enter the picture
and complicate matters considerably. Pure-component condensation yields a much higher
heat transIer coeIIicient than condensation oI a mixture oI diIIerent components over a
temperature range, as (a) no sensible vapor cooling is involved in the Iormer and (b) there is
no mass transIer resistance or diIIusion. This aspect is discussed in Section 7.3.
7.2.6 According to operating pressure
The operating pressure oI a condenser can vary Irom high vacuum to hundreds oI
atmospheres. For example, in an oil reIinery vacuum column steam-jet ejector condenser,
the operating pressure can be as low as 50 mm Hg, whereas in an ammonia plant or a
polyethylene plant, condensation can take place at over 300 atmospheres. Since vapor
shear plays a very important role in non-isothermal condensation and since vapor density
is directly proportional to the operating pressure, the latter exerts considerable inIluence
in condenser design.
Evidently, it is much more diIIicult to handle low-pressure services, primarily because
oI the low vapor densities ana the low allowable pressure drop. This is really a double
penalty since a low vapor density produces a high pressure drop, all other things remaining
constant. Thus, handling a low-pressure condensing service on the tubeside necessitates the
use oI one or more oI the Iollowing: a lower number oI tube passes (oIten only one), a larger
tube diameter, and/or a smaller tube length. On the shellside, special Ieatures such as a
divided-Ilow shell and/or double-segmental baIIles are oIten required to be employed in
order to limit the pressure drop to within the permitted value. Common to all the above
Ieatures, whether on the tubeside or on the shellside, are a larger Ilow area and a shorter Ilow
length.
86
7.3 Mechanisms of Condensing
Condensation occurs when a vapor comes in contact with a surIace which is at a
temperature below its dew point. The normal mechanism Ior heat transIer in commercial
condensers is Iilmwise condensation. Although much academic investigation has been
devoted to dropwise condensation, there has been very little application oI this Ior
commercial purposes, since special surIaces are required to maintain this mode oI
condensation. Furthermore, beneIicial results are demonstrated only at low liquid
loadings. Consequently, we shall limit this discussion to Iilmwise condensation only.
When condensation occurs, a Iilm oI liquid covers the heat transIer surIace. The
thickness oI the Iilm depends upon the rate oI condensation and the rate oI removal oI the
condensate. The latter depends upon the actions oI shear and gravity. ThereIore, it is very
important Ior condenser calculations to determine the Ilow regime, shear-controlled or
gravity-controlled, as diIIerent correlations have to be employed in the two regimes.
When the condensing stream is a mixture oI various components, with or without
noncondensables, the condensation process becomes Iar more complex than Ior a pure vapor
since it then involves mass transIer eIIects, which creates additional thermal resistances,
thereby reducing the heat transIer coeIIicient considerably.
7.3.1 Vertical in-tube condensation
Let us consider condensation inside vertical tubes,
as this is the simplest situation where the
mechanisms can be easily visualized. Figure 7.1
represents the situation, wherein it will be seen
that the condensate Iilm Ilows under the inIluence
oI both the shear Iorce and the Iorce oI gravity,
and is restrained by the shear Iorce (Iriction) at the
wall. Based upon the balance oI these Iorces and
the liquid and vapor Ilow rates and their physical
properties, the Iilm thickness can be determined.
Gravity-controllea conaensation
Here, the Iorce oI gravity is considerably higher
than that oI shear. This is the classical Nusselt
mode oI condensation wherein the Iilm is laminar
and heat transIer is considered to take place totally
by conduction through the condensate Iilm.
In practical applications, however, the liquid
Iilm does not remain laminar but, aIter passing
through a transition region, becomes turbulent aIter
a critical Reynold`s number is reached. When this
happens, the heat transIer coeIIicient increases.
The general variation oI gravity-controlled
condensate heat transIer coeIIicient is depicted
graphically in Fig. 7.2. This curve is general in that
it portrays the typical behavior, whether conden-
sation is inside the tubes or outside the tubes, Ior
both vertical and horizontal layouts.
Fig. 7.1 Condensation on a vertical surIace
in the absence oI vapor shear (Reprinted
Irom the Heat Exchanger Design Handbook,
2002 with permission oI Begell House, Inc.)
87
Shear-controllea conaensation
In this regime, the vapor velocity is very high and, thereIore, the gravity component
becomes negligible. The Ilow pattern is annular and is shown in Fig. 7.3a. The variation
oI the heat transIer coeIIicient is essentially linear and is represented graphically in Fig.
7.3b. Evidently, Ior the same Reynold`s number, the heat transIer coeIIicient will be
higher when the Prandtl number oI the condensate is higher.
The main Ilow pattern oI vertical
intube downIlow is the existence oI
annular Ilow Ior a wide range oI Ilow
conditions. In Iact, Ior a low Ilow
rate case, annular Ilow persists
throughout the entire tube.
Reflux conaensers
A reIlux condenser is a special kind
oI vertical tubeside condenser
wherein vapor travels up vertical
tubes, condenses, and Ialls back in
an annular Iilm while noncon-
densables leave Irom the top. It is
also known as a 'knock-back¨
condenser, as the condensables are
condensed and knocked back.
ReIlux condensers are employed
in the chemical process industries Ior
condensing the maximum amount oI
a vapor Irom a noncondensable oII-
gas or Irom vapor streams leaving
vapor-liquid entrainment separators.
For distillation column applications,
a reIlux condenser sits directly on top
oI the column, thereby eIIecting
major savings in piping and struc-
ture. Coolant usually enters at the
top, thereby providing maximum
cooling to the vapors.
The major disadvantage oI reIlux
condensers is their capacity limita-
tion due to Ilooding which occurs
when the vapor velocity is so high
that it prevents the Iree drainage oI
condensate. To provide better con-
densate drainage and to minimize
condensate blockage oI the vapor at
the tube inlet, tubes can be made to
extend beyond the lower tubesheet
and have their ends taper-cut.
Because oI the relatively low velo-
Fig. 7.2 Typical variation oI gravity-controlled condensate
Iilm heat transIer coeIIicient with liquid Reynolds number
(Redrawn Irom HTRI.)

Fig. 7.3a Annular Ilow in shear-controlled condensation
inside vertical tubes (Courtesy oI HTRI.)
Fig. 7.3b Typical variation oI shear-controlled condensate
Iilm heat transIer coeIIicient with vapor Reynolds number
(Redrawn Irom HTRI.)
88
cities in reIlux conden-
sers, they are not well
suited Ior condensing
mixtures which have a
large amount oI noncon-
densable gases.
A number oI studies
have been carried out on
Ilooding in reIlux con-
densers. The correla-
tions which are usually
employed to check the
possibility oI Ilooding
are |7| and |8|.
7.3.2 Horizontal in-tube condensation
Horizontal tubeside condensation is commonly employed in air-cooled heat exchangers
as well as in kettle-type chillers Ior vaporizing a reIrigerant on the shellside. It represents
a more diIIicult situation because oI the more complex nature oI the Ilow patterns,
especially in the gravity-controlled regime.
For both high and low total mass velocities, condensation begins in the annular regime.
Subsequently, in the case oI high total mass velocity, the condensation progresses to slug
Ilow whereas in the case oI low total mass velocity, it progresses to wavy or stratiIied Ilow.
This is represented in Fig. 7.4a. Cross-sectional views oI stratiIied and annular Ilow are
shown in Fig. 7.4b.
In stratiIied Ilow, the heat transIer coeIIicient is a strong Iunction oI the height oI the
stratiIied layer, which is diIIicult to predict. In the shear-controlled regime, however, the
conditions are much less complicated and, in Iact, the heat transIer coeIIicient is virtually the
same as in vertical intube condensation.
7.3.3 Condensation outside tubes
Condensation outside horizontal tubes is the most common situation Ior shell-and-tube
process condensers in the chemical process industries with cooling water or even a
single-phase or boiling hydrocarbon coolant inside tubes. One common application oI
condensation outside vertical tubes is in steam-heated vertical thermosyphon reboilers,
where the process stream vaporizes
inside tubes and steam condenses
on the shellside.
For condensation outside a sin-
gle tube without vapor shear, the
Nusselt equation again applies. For a
tube bundle, however, there are ad-
ditional eIIects oI condensate drip-
ping Irom row to row (inundation)
and the eIIect oI vapor shear. The
tube spacing and tubeIield pattern
(whether in-line or staggered) are
important variables Ior these eIIects.
Fig. 7.4a Typical Ilow patterns in horizontal in-tube condensation (Reprinted
Irom the Heat Exchanger Design Handbook, 2002 with permission oI Begell
House, Inc.)
Fig. 7.4b StratiIied and annular Ilow in horizontal in-tube
condensation (Reprinted Irom the Heat Exchanger Design
Handbook, 2002 with permission oI Begell House, Inc.)
89
With the introduction oI baIIles, the
situation becomes even more com-
plex. Thus, condensation on the
outside oI tube bundles is Iar more
diIIicult to predict than that inside
tubes due to the much more complex
geometry.
When condensing is on the
shellside oI horizontal condensers,
baIIles are invariably cut vertically
and have triangular drain notches at
the bottom. A vertical baIIle cut per-
mits the separation oI condensate
Irom uncondensed vapor due to the
eIIect oI gravity. It should be realized that two Iorces act on the condensate Iilm on tubes:
the Iorce oI gravity and that oI vapor shear. In the case oI a horizontal shell and a vertical
baIIle cut, the shear Iorce acts horizontally whereas the gravitational Iorce acts vertically
downward. Thus, depending upon the balance oI these two Iorces, the condensation mode
will be either gravity-controlled or shear-controlled. Actually, the demarcation is not so Iine
and there is a Iairly wide transition region as well. Initially, when the vapor enters the shell,
the vapor shear can be rather high so that condensation may well be in the shear-controlled
mode. However, with the progress oI condensation, vapor Ilow rate and, thereIore, vapor
shear decreases rapidly and the mode alters Iirst to transition and then to gravity-controlled.
DiIIerent correlations are employed depending upon whether the condensation is in the
shear-controlled or gravity-controlled regime. DiIIerent situations exist, depending upon
whether the tubes are horizontal or vertical. The general nature oI the variation oI heat
transIer coeIIicient with liquid Reynolds number is the same as shown in Fig. 7.2. The eIIect
oI vapor shear is vividly shown in Fig. 7.5.
7.3.4 Condensation of mixed vapors and mixtures of vapors and noncondensables
So Iar, we have discussed condensation oI pure vapors. However, a Iar more practical
application is the condensation oI a mixture oI vapors, with or without noncondensables.
Here, we may have total condensation oI a multicomponent mixture or partial
condensation oI only some oI the components.
This situation is rather complicated and the Iollowing Ieatures will have to be considered
Ior design:
a) Since the heavier components will condense Iirst and the lighter components
later, condensation will take place over a temperature range.
b) Because the condensation is not linear, there will be sensible heat duty besides
condensing duty. Any vapor that has not been condensed will have to be cooled
to its dew point beIore it can condense. Similarly, any vapor that has condensed
will have to be cooled to the Iinal outlet temperature. Since vapor cooling heat
transIer coeIIicient is considerably lower than that oI condensing, it reduces the
overall heat transIer coeIIicient appreciably. It Iollows, thereIore, that the wider
the range oI components and, thereby, the wider the condensing temperature
range, the lower will be the heat transIer coeIIicient, all other things remaining
constant.
Fig. 7.5 EIIect oI vapor shear on condensation outside a bank
oI tubes (Courtesy oI HTRI.)
90
c) As the composition oI the vapor and the liquid vary continuously along the path
oI condensation, so will their physical properties. Particularly signiIicant is the
liquid viscosity.
a) The molecules oI the heavy components must diIIuse through the barrier oI the
molecules oI the lighter components to reach the condensing surIace. ThereIore,
the condensation rate will be controlled by the rate oI diIIusion as well as the rate
oI heat transIer.
The heat transIer rate across the condensate Iilm is calculated in the same way as Ior pure
components. For the vapor cooling process, however, both the processes oI diIIusion and
the cooling oI the vapor phase to the condensation temperature oI the vapor mixture pre-
vailing at the interIace must be considered.
Pressure arop
The determination oI pressure drop in a condenser is a very complex task since the velo-
city and Ilow pattern change constantly along the Ilow path. The various components are
a) inlet and exit losses (contraction and expansion) in nozzles and headers
b) two-phase Iriction loss
c) static head
a) momentum change
The static head is usually insigniIicant in condensers.
The momentum change results in a pressure gain since there is deceleration oI the vapor
as its Ilow rate decreases with the progress oI condensation. However, this is insigniIicant
unless the condenser operates under vacuum, in which case the pressure gain could be
substantial.
The two-phase Iriction is usually the largest component oI the overall pressure drop and
is determined stepwise along the length oI the condenser, using Martinelli`s or other
correlations. The Martinelli correlation is particularly accurate Ior condensing inside
horizontal tubes.
7.4 Practical Guidelines for Thermal Design
In the case studies presented in the Iollowing sections, various aspects oI condenser
design will be illustrated, namely, condensing range, operating pressure, shell
type/baIIling, multiple shells, desuperheating, subcooling, and low-Iin tubes.
7.4.1 Baffling
OI all the shellside parameters, baIIling is the most crucial. By baIIling is meant the type
oI baIIles (single or double segmental, etc.), baIIle spacing, and baIIle cut. BaIIle cut
means not only the value oI the cut, but its orientation as well. The reader should reIer to
Section 3.4.4 Ior a detailed discussion on baIIling.
As with single-phase heat transIer, the designer's job is to best utilize the allowable
pressure drop. The value oI pressure drop usually permitted in a condenser is 0.72.84 psi
(0.050.2 kg/cm
2
), and may be even lower at very low pressures. For example, only 5 mm
Hg is permitted in a steam-jet ejector pre-condenser Ior an oil reIinery vacuum column
operating at a typical pressure oI 50 mm Hg. Pressure drop higher than 2.84 psi (0.2 kg/cm
2
)
is usually not required to obtain a reasonably high condensing heat transIer coeIIicient.
Since vapor shear decreases progressively along the length oI a condenser, the designer
91
can decrease the baIIle spacing progressively. By doing so, the heat transIer coeIIicient does
not decrease as rapidly as it would iI uniIorm baIIle spacing were employed. This is
particularly true in a wide-range condenser where the vapor-phase resistance plays a more
predominant role. However, the beneIit obtained Irom such an arrangement in terms oI a
higher heat transIer coeIIicient Ior the same pressure drop is not very signiIicant and,
consequently, the use oI variable baIIle spacing is not very prevalent.
Condensers usually have single-pass shells (TEMA type E) and single-segmental
baIIles. However, as explained in Section 3.4.6, a suitable combination oI single/double-
segmental baIIles and single-pass/divided-Ilow shell is oIten required to be employed when
there is a problem oI limiting the shellside pressure drop to within the allowed value.
Since condensers predominantly handle vapors, the possibility oI a Ilow-induced
vibration problem is clearly high. While the subject oI Ilow-induced vibration is discussed in
detail in chapter 12, it will suIIice to say here that, with a proper selection oI shell type and
baIIling, the possibility oI Ilow-induced vibration can be totally eliminated. No-tubes-in-
window design plays a particular useIul role in this area.
Let us illustrate some oI the various combinations oI shell type and baIIling.
CASE STUDY 7.1: ISOTHERMAL CONDENSATION
WITH SINGLE-PASS SHELL AND SINGLE-SEGMENTAL BAFFLES
Consider a light hydrocarbon condenser having the principal process parameters detailed
in Table 7.1a. CS tubes 0.7874-in. (20-mm) OD, 0.0787-in. (2-mm) thick, and 19.68-It
(6-m) long were to be used in a Iixed-tubesheet construction (TEMA type AEL).
An initial thermal design was prepared and is shown in Table 7.1b. This had a single-
pass shell (the simplest type), a baIIle spacing oI 11.8 in. (300 mm), and a baIIle cut oI 25°
on diameter. In order to demonstrate the eIIect oI baIIle spacing on such a design, the baIIle
spacing was altered to 10.6 in. (270 mm), 9.84 in. (250 mm), 9.06 in. (230 mm), and 7.9 in.
(200 mm), progressively. The results oI this study are also shown in Table 7.1b.
Looking at the table, it will be seen that, when the baIIle spacing is decreased Irom 11.8
in. (300 mm) to 7.9 in. (200 mm), the shellside (condensing) heat transIer coeIIicient
Table 7.1a: Salient process parameters Ior Case Study 7.1
Shellside Tubeside
1. Fluid Light hydrocarbon Cooling water
2. Flow rate, lb/h (kg/h) 79,366 (36,000) 895,244 (406,080)
3. Operating pressure, psia (kg/cm
2
a) 242 (17.0) 71 (5.0)

4. Temperature in/out, °F (°C)
123.3 (50.7)
|dew point and bubble point ÷
123.3°F (50.7°C)|

91.4 (33)/102.2 (39)
5. Allowable pressure drop, psig (kg/cm
2
g) 2.8 (0.2) 10 (0.7)
6. Fouling resistance, h It
2
°F/Btu
(h m
2
°C/kcal)
0.00049 (0.0001) 0.00195 (0.0004)
7. Heat duty, Btu/h (kcal/h) 9,659,300 (2,433,800)
8. Design pressure, psig (kg/cm
2
g) 341 (24.0) 85 (6.0)
9. Material oI construction CS CS
92
increases by 12.9°. As the shellside Iilm resistance is only a small part oI the overall
resistance to heat transIer and the balance resistances do not alter due to a change in the
baIIle spacing, the overall heat transIer coeIIicient increases by only 5.3°. In stark contrast,
however, the shellside pressure drop increases by an enormous 107°. This is explained by
the Iact that, in isothermal condensation, the heat transIer coeIIicient is rather high even with
a moderate shear rate so that the increase in the same with higher shear is not considerable.
However, the pressure drop is a very strong Iunction oI the shellside velocity and the
number oI cross-passes. Hence, with an increase oI both these parameters as a result oI
reduced baIIle spacing, the shellside pressure drop increases dramatically.
In the present instance, the design with a baIIle spacing oI 7.9 in. (200 mm) may be
selected as the Iinal design as the shellside pressure drop is just within the permissible value
and the condenser is adequately surIaced (2.58° overdesign).
Table 7.1b: Thermal design oI light hydrocarbon condenser
Baffle spacing, in. (mm)

11.8
(300)
10.6
(270)
9.84
(250)
9.06
(230)
7.9 (200)
1. Type oI exchanger Fixed tubesheet (TEMA AEL)
2. Shell ID, in. (mm) 31.5 (800)
3. Number oI tubes 643
4. Tube OD × thickness × length, in. (mm) 0.7874 (20) × 0.0787 (2) × 236 (6000)
5. Number oI tube passes 2
6. Tube pitch, in. (mm) 1.024 (26) square
7. BaIIle cut, ° (diameter) 25
8. Heat transIer area, It
2
(m
2
) 2561 (238)
velocity, It/s (m/s) 5.81 (1.77)
pr. drop, psi (kg/cm
2
) 6.0 (0.42)

9. Tubeside
heat transIer
coeIIicient, Btu/h It
2

°F (kcal/h m
2
°C)

1147 (5600)
heat transIer
coeIIicient, Btu/h It2
°F (kcal/h m
2
°C)
323
(1577)
333.9
(1630)
340
(1660)
349.4
(1706)
364.8
(1781)

10. Shellside
pr. drop, psi (kg/cm2)
1.4
(0.1)
1.73
(0.122)
1.95
(0.137)
2.3
(0.163)
2.9
(0.207)
Shellside, in/out 10 (250)/10 (250)
11. Nominal
connection
size, in. (mm) Tubeside, in/out 10 (250)/10 (250)
12. Overall heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
140
(684)
142
(693.8)
143.2
(699.1)
144.9
(707.3)
147.5
(720)
13. Overdesign, ° - 2.77 - 1.74 - 1.25 - 0.52 2.58
93
CASE STUDY 7.2 CONDENSATION WITH SINGLE-PASS
SHELL AND DOUBLE-SEGMENTAL BAFFLES
Consider a distillation column overhead condenser service whose principal process
parameters are elaborated in Table 7.2a. As its operating temperature level is rather high,
its energy is recovered Ior the preheating oI demineralized water. It may be noted that the
vapors are partially condensed at the inlet and that its pressure is rather low (39.8 psia or
2.8 kg/cm
2
abs.) and so, consequently, is the vapor density.
Although the shellside service is rather clean, so that a Iixed-tubesheet construction
could have been used, a Iloating-head construction was speciIied. However, this is
immaterial in the present context. Carbon steel tubes 0.7874-in. (20-mm) OD, 0.0787-in. (2-
mm) thick, and 19.68-It (6000-mm) long were to be used.
A design was attempted with an E (single-pass) shell and single-segmental baIIles.
However, it was Iound that, in order to contain the shellside pressure to the speciIied limit oI
1.5 psi (0.105 kg/cm
2
), there could be only eight cross-passes representing a baIIle spacing
oI 25.6 in. (650 mm). Consequently, the unsupported tube span came to 110°, 82.4°, and
97.7° oI the maximum recommended by TEMA at the inlet, central, and outlet regions
Table 7.2a: Salient process parameters Ior Case Study 7.2
Shellside Tubeside
1. Fluid
Distillation column
overhead
Demin. water
2. Flow rate, lb/h (kg/h) 143,300 (65,000) 300,000 (136,100)
3. Operating pressure, psia (kg/cm
2
a) 39.8 (2.8) 142 (10.0)
4. Temperature in/out, °F (°C)
287.6 (142)/248 (120)
bubble point 248 (120)
185 (85)/246.2 (119)
5. Mol. wt. vapor 90 -
liquid 0.2/0.22 4
6. Viscosity in/out, cp
vapor 0.0093/0.0087 -
liquid
0.065 (0.096/0.0706
(0.105)
0.378 (0.563)/0.383
(0.57)

7. Th. cond. in/out, Btu/h It °F
(kcal/h m °C)
vapor 0.01 (0.015)/0.0095 (0.014) -
liquid 0.51/0.495 1.0/1.0
8. Sp. heat in/out, Btu/lb °F
(kcal/kg °C)
vapor 0.5/0.485 -
liquid 40.4 (648)/42.1 (675) 60.3 (967)/59.1 (947)
9. Density in/out, lb/It
3
(kg/m
3
)
vapor 0.44 (7.05)/0.45 (7.25) -
10. Allowable pressure drop, psi (kg/cm
2
) 1.5 (0.105) 7.0 (0.5)
11. Fouling resistance, h It
2
°F/Btu (h m
2
°C/kcal) 0.00098 (0.0002) 0.00098 (0.0002)
12. Heat duty, MM Btu/h (MM kcal/h) 18.42 (4.641)
13. Design pressure, psig (kg/cm
2
g) 71 (5.0) 284 (20.0)
14. Material oI construction CS CS
94
when it is prudent to restrict this to 80°. More importantly, the tubes were prone to Iailure
by Ilow-induced vibration (Iluidelastic whirling). See Chapter 12 Ior a detailed discussion on
Ilow-induced vibration.
As a design with single-segmental baIIles was not Ieasible, the baIIling was modiIied to
double-segmental. It was now possible to produce a design which had a shellside pressure
drop within the permitted value and which was saIe against Iailure oI tubes due to Ilow-
induced vibration. Both the single-segmental and double-segmental baIIle designs are
detailed in Table 7.2b.
For a better appreciation oI the double-segmental baIIle design, Table 7.2c presents the
variation oI some oI the signiIicant shellside and overall parameters along the length oI the
shell. As there were 18 increments in the computer printout, and as it would have been
Table 7.2b: Salient Ieatures oI single- and double-segmental
baIIle designs Ior deheptanizer condenser
Design with single-
segmental baffles
Design with double-
segmental baffles
1. Type oI shell Floating-head (TEMA AES)
2. Shell ID, in. (mm) 40.4 (1025)
3. Heat transIer area, m
2
3830 (356)
4. Number oI tubes 966
5. Tube OD × thickness × length, mm 0.7874 (20) × 0.0787 (2) × 236 (6000)
6. No. oI tube passes 4
7. Tube pitch, in. (mm) 1.024 (26) square
8. BaIIle spacing, in. (mm) 25.6 (650) 12.8 (325)
9. BaIIle cut, ° diameter 35 6 rows overlap
Shellside 167.5 (818) 161.6 (789)
Tubeside 877 (4283) 877 (4283)

10. HTC, Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 104.6 (510.9) 102.3 (499.4)
11. MTD, °C 50.9 (28.3) 49.9 (27.7)
12. Overdesign, ° 10.9 6.44
Shellside 1.5 (0.104) 1.3 (0.091)
13. Pressure drop, psi (kg/cm
2
)
Tubeside 2.84 (0.2) 2.84 (0.2)
14. Tubeside velocity, It/s (m/s) 2.7 (0.82) 2.7 (0.82)
15. Shellside crossIlow/window velocity,
It/s (m/s)
12.9 (3.92)/13.6 (4.14) 13.1 (3.99)/13.0 (3.96)
Propylene 18 (450)/6 (150)
16. Nozzle size in/out, in. (mm)
Demin. Water 6 (150)/6 (150)
17. SaIe against Iailure oI tubes due to
Ilow-induced vibration
No Yes
95
cumbersome to display them all, two increments were lumped up as one region: thus there
are nine regions. The Iollowing observations are interesting:
1) The condensing heat transIer coeIIicient decreases steadily Irom inlet to the
penultimate zone and then drops somewhat sharply in the last zone. This is only
to be expected since, with the progress oI condensation, the Ilow rate oI the
uncondensed vapor reduces and, thereIore, so does the vapor shear. However, the
reader may have observed that the condensing heat transIer coeIIicient in region
1 is lower than that in region 2. This is because the inlet baIIle space is much
larger than the central baIIle space, thereby resulting in much lower shear and,
consequently, a lower condensing heat transIer coeIIicient.
2) Along with the condensing heat transIer coeIIicient, the overall heat transIer
coeIIicient also reduces Irom the inlet to the outlet oI the condenser. Further, the
MTD also reduces Irom inlet to outlet as a result oI the temperature proIiles oI
the two sides. As a result, Ior the equal heat transIer area increments oI 425 It
2

(39.5 m
2
) in regions 2²8, the incremental heat duty reduces steadily Irom 2.754
MM Btu/h (0.694 MM kcal/h) to 1.31 MM Btu/h (0.33 MM kcal/h).
Table 7.2c: Variation in condensing pattern along tube length
in double-segmental baIIle design Ior distillation column overhead condenser
Region
1
Region
2
Region
3
Region
4
Region
5
Region
6
Region
7
Region
8
Region
9
1. Vapor weight
Iraction
0.752 0.572 0.458 0.357 0.267 0.187 0.113 0.0496 0.018
2. Shellside heat
transIer coeIIi-
cient, Btu/h It
2
ƒF (kcal/h m
2

ƒC)

164.9
(805.3)

197.7
(965.4)

189.6
(925.5)

176.1
(859.9)

163.3
(797.4)

157.7
(770.1)

158.3
(772.8)

145.4
(710)

144.8
(707.2)
3. Flow regime Shear Shear Shear Trans Trans Trans Gravity Gravity Gravity
4. Overall heat
transIer
coeIIicient,
Btu/h It
2
ƒF
(kcal/h m
2
ƒC)

102.7
(501.5)

113.4
(553.5)

110.4
(539.1)

105.9
(516.8)

101.4
(495)

99.1
(484)

99
(483.2)

91.9
(448.9)

93.8
(458.2)
5. MTD, ƒF (ƒC)
68.0
(37.8)
61.6
(34.2)
56.25
(31.25)
51.75
(28.75)
47.8
(26.55)
44.2
(24.55)
40.6
(22.55)
35.7
(19.85)
33.1
(18.4)
6. Incremental
heat duty, MM
Btu/h (MM
kcal/h)

3.74
(0.943)

2.75
(0.694)

2.43
(0.613)

2.18
(0.55)

1.905
(0.48)

1.71
(0.43)

1.59
(0.4)

1.31
(0.33)

0.81
(0.204)
7. Cumulative
heat duty, MM
Btu/h (MM
kcal/h )

3.74
(0.943)

6.5
(1.637)

8.93
(2.25)

11.1
(2.8)

13.0
(3.28)

14.7
(3.71)

16.3
(4.11)

17.62
(4.44)

18.43
(4.644)
8. Incremental
area required,
It
2
(m
2
)
575.7
(53.5)
423.9
(39.4)
425
(39.5)
423.9
(39.4)
425
(39.5)
(39.4) 425
(39.5)
423.9
(39.4)
288.4
(26.8)
9. Cumulative
area required,
It
2
(m
2
)
597
(55.5)
1000
(92.9)
1425
(132.4)
1849
(171.8)
2274
(211.3)
2698
(250.7)
3144
(290.2)
3546
(329.6)
3835
(356.4)
96
3) The condensing heat transIer coeIIicient here is in the range oI 144.8197.7
Btu/h It
2
ƒF (707.2965.4 kcal/h m
2
ƒC) whereas in Case Study 7.1, it was in the
range oI 323364.8 Btu/h It
2
ƒF (15771781 kcal/h m
2
ƒC). There are two
principal reasons Ior this. First, in Case Study 7.1, the condensing Iluid was a
pure component condensing isothermally whereas, in the present case, the
condensing Iluid is a mixture oI hydrocarbons condensing over a temperature
range. Second, as the operating pressure and thereby the vapor density is much
lower in the present case, only a much lower velocity can be tolerated within the
permissible pressure drop. The lower vapor velocity translates into a lower
condensing heat transIer coeIIicient.
CASE STUDY 7.3 CONDENSATION WITH DIVIDED-FLOW SHELL
Consider the service speciIied in Table 7.3a Ior a splitter overhead condenser.
Note that the operating pressure is very low, 24.2 psia (1.7 kg/cm
2
abs). The condensing
range is small |194 ² 174.2 ÷ 19.8ƒF (90 ² 79 ÷ 11ƒC)|, hence a linear heat duty versus
temperature proIile was speciIied by the licensor.
A single-pass shell (TEMA E shell) design with single-segmental baIIles was consi-
Table 7.3a: Salient process parameters Ior Case Study 7.3
Shellside Tubeside
1. Fluid Splitter overhead Cooling water
2. Flow rate, lb/h (kg/h) 220,460 (100,000) 1,417,600 (643,000)
3. Operating pressure, psia (kg/cm
2
a) 24.2 (1.7) 71 (5.0)

4. Temperature in/out, °F (°C)
194 (90)/174.2 (79)
dew point 194 (90)
bubble point 174.2 (79)

89.6 (32)/111.2 (44)
5. Mol. wt. vapor 76 -
liquid 0.3/0.25 0.775/0.616
6. Viscosity in/out, cp
vapor 0.011/0.011 -
liquid 0.06 (0.093)/0.063 (0.094) 0.364 (0.542)/0.368 (0.548)
7. Th. cond. In/out,
Btu/h It °F (kcal/h m °C)
vapor 0.0084 (0.0125)/0.008 (0.012) -
liquid 0.56/0.55 1.0/1.0 8. Sp. heat in/out, Btu/lb °F
(kcal/kg °C)
vapor 0.47/0.465 -
liquid 45.2 (725)/45.7 (732) 62 (993)/61.7 (989)
9. Density, lb/It3 (kg/m
3
)
vapor 0.25 (4.05)/0.24 (3.79) -
10. Allowable pressure drop, psi (kg/cm2) 3.6 (0.25) 10 (0.7)
11. Fouling resistance, h It
2
°F/Btu
(h m
2
°C/kcal)
0.00146 (0.0003) 0.00195 (0.0004)
12. Heat duty, MM Btu/h (MM kcal/h) 30.58 (7.7057)
13. Design pressure, psig (kg/cm
2
g) 57 (4.0) 100 (7.0)
14. Material oI construction CS CS
97
dered. As expected, due to the very low operating pressure on the shellside and, thereby, the
very low vapor density, it was not possible to satisIy the shellside pressure drop limitation.
Next, a single-pass shell design with double-segmental baIIles was considered. By
employing a baIIle spacing oI 14.6 in. (370 mm) and only 3 rows overlap (which yielded a
40.7° baIIle cut on area), it was possible to restrict the shellside pressure drop to the
permitted value; however, the tubes were Iound to be highly prone to Iailure by Ilow-
induced vibration due to Iluidelastic whirling and, as such, this design could not be adopted.
A divided-Ilow shell (TEMA J shell) was then investigated and it was Iound that a
design could be produced that respected the shellside pressure drop and, at the same time,
was saIe against Iailure due to Ilow-induced vibration. This design was adopted. The baIIle
spacing was 9.45 in. (240 mm) and the baIIle cut 30° on diameter.
The results oI both the E and J shells are presented in Table 7.3b. Because oI the large
Table 7.3b: Salient Ieatures oI single-pass and divided-Ilow shell
designs Ior splitter overhead condenser
Design with
single-pass shell
Design with
divided-flow shell
1. TEMA type AES AJS
2. Shell ID, in. (mm) 36.4 (925)
3. Heat transIer area, It
2
(m
2
) 3346 (311)
4. Number oI tubes 764
5. Tube OD × thickness × length, in. (mm) 0.7874 (20) × 0.0787 (2) × 236 (6000)
6. No. oI tube passes 2
7. Tube pitch, in. (mm) 1.024 (26) square
8. Type oI baIIles double-segmental single-segmental
9. BaIIle spacing, in. (mm) 14.6 (370) 9.5 (240)
10. BaIIle cut, ° diameter
3 rows overlap, vertical
(40.7° on area)
30, vertical
Hydrocarbon 249 (1215) 267.5 (1306)
Demin. water 1368 (6677) 1368 (6677)
11. Heat transIer coeIIicient,
Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 112.3 (548.2) 116 (566.2)
Shellside 3.4 (0.24) 3.6 (0.25)
12. Pressure drop,
psi (kg/cm2)
Tubeside 8.4 (0.59) 8.4 (0.59)
13. MTD, °F (°C) 82.8 (46.0) 82.4 (45.8)
14. Overdesign, ° 1.79 4.79
Shellside 22 (550)/10 (250) 2 x 16 (400)/10 (250)
15. Nozzle size (nom.) in/out,
in. (mm)
Tubeside 12 (300/12 (300)) 12 (300)/12 (300)
16. SaIe against Iailure oI tubes due to
Ilow-induced vibration
No (High probability oI
Iailure due to Iluidelastic
whirling)

Yes
98
shell inlet nozzle size, it will be advantageous to incorporate a vapor belt (or distribution
belt) at the shell inlet in both designs. A 2-1 divided-Ilow design (two inlet nozzles and one
outlet nozzle) was adopted as the inlet nozzle was much larger than the outlet nozzle
diameter; iI a 1-2 design were adopted, the shell inlet nozzle would be much larger (22 in. or
550 mm instead oI 16 in. or 400 mm, all nominal).
7.4.2 Multiple shells in series or parallel
Multiple shells in parallel
As in the case oI single-phase services, condensers are also oIten required to be oI multi-
ple shell construction. When the required heat transIer area is too large to be incorporated
in one shell oI a removable-bundle condenser (U-tube or Iloating-head), multiple shells
are used in parallel. Thus, iI the maximum permitted bundle weight is 22,000 lbs. (10
tons) and iI the weight oI the tube bundle in a single-shell design comes to 33,000 lbs.
(15 tons), two shells will have to be used in parallel. An additional advantage with the use
oI multiple shells in parallel is that it aIIords the Iacility oI running the unit at part-load
even when one shell is down Ior maintenance. One oil reIining company the author
knows oI has a standard practice oI employing (at least) two shells in parallel Ior every
condenser (and cooler), each designed Ior 60° oI the total heat duty, even iI the total heat
transIer area required is small and can easily be accommodated in one shell. Thus, the
condenser (or cooler) can deliver 60° heat duty even with one shell out oI commission.
Sea water is the coolant in this plant and the incidence oI bio-Iouling is very heavy,
thereby requiring Irequent cleaning. Hence the above design philosophy.
Multiple shells in series
The reasons Ior using multiple shells in series are maniIold:
1) There is a temperature cross between the condensing stream and the coolant. For
example, iI the overhead vapors oI a distillation column are to be cooled and
condensed Irom 212ƒF (100ƒC) to 104ƒF (40ƒC) by cooling water entering at
91.4ƒF (33ƒC) and leaving at 113ƒF (45ƒC), two shells in series will be required.
2) Even iI the outlet temperatures oI the two streams are equal and a single shell
may be used, there is a Iairly big advantage in going Ior two shells in series. The
MTD correction Iactor F
t
will increase appreciably Irom around 0.8 in the single-
shell design to over 0.92 and even up to 0.95 (depending upon the terminal tem-
peratures) with two shells in series, thereby representing a 1519° reduction in
the heat transIer area on this count alone.
This is particularly advantageous when two shells have to be used anyway
because the required heat transIer area is too large to be incorporated in a single
shell. Instead oI connecting the two shells in parallel, they are connected in
series. OI course, this arrangement will produce a much higher shellside pressure
drop so that it can be employed only iI the allowable pressure drop can be
adhered to. The use oI double-segmental baIIles and/or a divided-Ilow shell helps
considerably in achieving this.
With two shells connected in series, however, much more extensive
piping/valving is required to take out an individual shell Ior maintenance.
3) The permitted shellside pressure drop cannot be utilized in a single design, even
with the smallest baIIle spacing and cut Ieasible; consequently, the heat transIer
coeIIicient will be low and, thereIore, the heat transIer area high. With two shells
99
in series resulting in a much smaller shell diameter, the permitted shellside
pressure drop can be utilized much better with a corresponding increase in the
shellside velocity and, thereby, the heat transIer coeIIicient. Thus, there is an
appreciable reduction in the overall heat transIer area.
CASE STUDY 7.4: CONDENSATION WITH MULTIPLE SHELLS IN SERIES
The salient process parameters oI a stabilizer overhead condenser are speciIied in Table
7.4a. It will be noted that there is no temperature cross, so that multiple shells in series is
not a must. The allowable shellside pressure drop is rather high, at 6.4 psi (0.45 kg/cm
2
).
The usual value Ior such services is 1.42.8 psi (0.10.2 kg/cm
2
).
Tubes oI carbon steel 0.984-in. (25-mm) OD and 0.0984-in. (2.5-mm) thick were to be
used in a Iloating-head construction (TEMA AES). The required heat transIer area is Iar too
high to be incorporated in a single shell, so that the use oI two shells is essential. Evidently,
the choice is between two shells in series and two shells in parallel. As just explained in
point 2in the previous section, 'Multiple shells in series,¨ the MTD is usually signiIicantly
higher when two shells are connected in series than when they are connected in parallel.
ThereIore, we shall explore the possibility oI such a conIiguration, Iirst, and accept the
Table 7.4a: Salient process parameters Ior Case Study 7.4
Shellside Tubeside
1. Fluid Stabilizer overhead Cooling water
2. Flow rate, lb/h (kg/h) 110,230 (50,000) 770,070 (349,300)
3. Operating pressure, psia (kg/cm
2
a) 170.6 (12.0) 71.1 (5.0)
4. Temperature in/out, °F (°C)
149 (65)/114.8 (46)
dew point ÷ 149 (65)
89.6 (32)/111.2 (44)
5. Weight Iraction vapor, in/out 1.0/0.013 -
6. Allowable pressure drop, psi (kg/cm2) 6.4 (0.45) 10 (0.7)
Vapor 0.01/0.0093 -
7. Viscosity in/out, cp
Liquid 0.12/0.18 0.775/0.616
Vapor
0.0094 (0.014)/0.0091
(0.0135)
-
8. Thermal conductivity,
in/out, Btu/h It °F
(kcal/h m °C)
Liquid 0.0605 (0.09)/0.0618 (0.092) 0.364 (0.542)/0.368 (0.548)
Vapor 1.385 (22.19)/1.42 (22.76) -
9. Density, in/out, lb/It
3
(kg/m
3
)
Liquid 32.45 (520)/33.2 (532) 62.0 (994)/62.34 (999)
Vapor 0.44/0.435 -
10. SpeciIic heat in/out,
Btu/lb °F (kcal/kg °C)
Liquid 0.59/0.58 1.0
11. Fouling resistance, h It2 °F/Btu
(h m
2
°C /kcal)
0.00146 (0.0003) 0.00195 (0.0004)
12. Heat duty, MM Btu/h (MM kcal/h) 16.61 (4.185)
13. Design pressure, psig (kg/cm
2
g) 228 (16.0) 100 (7.0)
14. Material oI construction CS CS
100
design iI it is Ieasible; that is, (a) iI the condenser is adequately surIaced, (b) the shellside
pressure drop and the tubeside pressure drop are within the allowable limits, and (c) the
design is saIe against the possibility oI Iailure oI tubes due to Ilow-induced vibration.
A design was made with two shells in series and the principal construction and
perIormance parameters are detailed in Table 7.4b. The design was acceptable on all oI the
above counts. In the stream analysis, it will be seen that the tube-baIIle hole is Iairly high, at
21.7°. However, this stream is quite eIIective Ior heat transIer and is oIten quite high in
shellside condensers.
Table 7.4b: Salient Ieatures oI stabilizer overhead condenser
1. Type oI heat exchanger Floating-head (TEMA AES)
2. Shell ID, in. (mm) 37.8 (960)
3. No. oI tubes 590
4. Tube OD × thickness × length, in. (mm) 0.984 (25) × 0.0984 (2.5) × 236 (6000)
5. Heat transIer area, It
2
(m
2
) 2 × 2937 (273) ÷ 5874 (546)
6. No. oI tube passes 2
7. Tube pitch, in. (mm) 1.26 (32) square
8. Type oI baIIles Single-segmental
9. BaIIle spacing, in. (mm) 11.0 (280)
10. BaIIle cut, ° diameter 22
Shellside 209.9 (1025)
Tubeside 737.9 (3603)

11. Heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
Overall 96.7 (472.9)
12. MTD, °C 29.9 (16.6)
13. Overdesign 2.06
Shellside 6.1 (0.43)
14. Pressure drop, psi (kg/cm
2
)
Tubeside 4.4 (0.31)
Main crossIlow 0.48
BaIIle hole-to-tube 0.254
Bundle-shell 0.091
Pass-partition 0.05


15. Stream analysis
BaIIle-shell 0.126
16. Cooling water velocity, It/s (m/s) 3.5 (1.06)
Shellside In/interconnecting 10 (250), out 6 (150)
17. Nominal nozzle size, in. (mm)
Tubeside In 10 (250) Out 10 (250)
18. SaIe against Iailure oI tubes due to Ilow-induced vibration Yes
101
Table 7.4c presents a break-up in the shellside perIormance oI the two shells as the same
is rather interesting. As expected, with decreasing vapor weight Iraction with the progress oI
condensation, the Ilow regime changes Irom shear-controlled to transition in the hot shell
and Irom transition to gravity-controlled in the cold shell. As a direct consequence, the
condensing heat transIer coeIIicient Ialls steadily. The hot shell has an appreciably higher
shellside and thereby overall heat transIer coeIIicient as well as a higher MTD than the cold
shell. Consequently, it delivers about 59° oI the total heat duty.
The shellside pressure drop is also higher in the hotter shell since it handles a higher
vapor Ilow rate. In Iact, the pressure drop in the hotter shell is almost 2.6 times that in the
colder shell. As the permitted shellside pressure drop is quite high in this case, the baIIle
spacing and baIIle cut in the two shells are kept identical. However, iI the permitted shellside
pressure drop had been much less, an unequal baIIle spacing and cut could have been
employedhigher in the hotter shell handling the higher vapor Ilow rate, and lower in the
colder shell handling the lower vapor Ilow rate.
CASE STUDY 7.5: CONDENSATION WITH MULTIPLE SHELLS IN SERIES/PARALLEL
Let us consider the condensing service speciIied in Table 7.5a. This is a distillation
column overhead trim condenser which Iollows an air-cooled condenser, with 58° oI the
vapors already condensed in the air-cooled condenser. It is an existing heat exchanger
which had been designed with Iour shells in parallel since the heat transIer area is very
large. The construction details and perIormance parameters are elaborated in Table 7.5b.
It will be observed that the permitted shellside pressure drop was not Iully utilized so that
the use oI double-segmental baIIles appears somewhat imprudent. The stream analysis is
not very satisIactory with a relatively low main crossIlow Iraction and overall
eIIectiveness. The condenser is operating totally in the gravity-controlled regime and,
thereIore, the condensing heat transIer coeIIicient is extremely low.
Let us thereIore attempt to reduce the size oI the condenser by utilizing the permitted
shellside pressure drop better. Two changes can be considered Ior this:
Table 7.4c: Break-up in perIormance in two shells oI stabilizer overhead condenser
Hot shell Cold shell
1. Heat duty, MM Btu/h (MM kcal/h) 9.79 (2.467) 6.82 (1.718)
2. Shellside temperature in/out, °F (°C) 149 (65)/128.7 (53.7) 128.7 (53.7)/114.8 (46)
3. Cooling water temp. in/out, °F (°C) 98.5 (36.93)/111.2 (44) 89.6 (32)/98.5 (36.93)
4. Shellside vapor weight Iraction, in/out 1.0/0.413 0.413/0.013
5. MTD, °F (°C) 32.4 (18.0) 26.8 (14.9)
6. Shellside pr. drop, kg/cm
2
4.4 (0.31) 1.7 (0.12)
7. Shellside heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
276 (1349) to 161 (785)
avg. 256 (1250)
208 (1013) to 149 (729),
avg. 173 (842)
8. Overall heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
106 (517.3) 87.8 (428.5)
9. Condensing Ilow regime
Iirst 12 regions: shear
next 6 zones: transition
Iirst 5 zones: transition
Next 13 zones: gravity
10. Overdesign, ° 2.85 1.13
102
a) change in baIIle type Irom double-segmental to single segmental
b) change in the conIiguration Irom Iour shells in parallel to two shells in series and
two shells in parallel
First, let us consider a revised design with single-segmental baIIles. The revised per-
Iormance is also shown in Table 7.5b. Please note that, although the shellside velocity
has increased signiIicantly and so has the shellside pressure drop, the shellside heat
transIer coeIIicient has hardly improved. Please also notice that the condenser is still
totally in the gravity-controlled regime and has an even poorer stream analysis. Although
the shellside velocity has increased, the increase has not been large enough to liIt the
condenser out oI the gravity-controlled regime.
Now, let us consider the second optionemploying two shells in series and two shells
in parallel. There are two drastic improvements. First, the MTD increases Irom 20ƒF
(11.1ƒC) to 23.5ƒF (13.08ƒC), an increase oI 17.87°. Second, the shellside velocity has
increased substantially which pushes the condenser predominantly into the transition regime,
thereby increasing the condensing heat transIer coeIIicient Irom 96 Btu/h It ƒF (469 kcal/h
m
2
ƒC) to 151.8 Btu/h It ƒF (741 kcal/h m
2
ƒC), an increase oI 58°! The overall heat transIer
coeIIicient increases 30.9°. Thus, the heat Ilux (which is the product oI the MTD and the
heat transIer coeIIicient) increases by 54° and results in a substantial reduction in the heat
Table 7.5a: Salient process parameters Ior Case Study 7.5
Shellside Tubeside
1. Fluid
Distillation column overhead Cooling water
2. Flow rate, lb/h (kg/h) 187,400 (85,000) 1136,000 (515,300)
3. Operating pressure, psia (kg/cm
2
a) 42.7 (3.0) 71 (5.0)
4. Temperature in/out, °F (°C) 149 (65)/102.2 (39) 91.4 (33)/102 (38.9)
5. Vapor weight Iraction, in/out 0.42/0.2 -
6. Vapor molecular weight 62.7 -
Liquid 0.25/0.3
7. Viscosity in/out, cp
Vapor 0.01/0.0091
Liquid 0.067 (0.1)/0.069 (0.102)
8. Th. cond. in/out, Btu/h It °F
(kcal/h m °C)
Vapor 0.0114 (0.017)/0.011 (0.016)
Liquid 41.1 (658)/41.7 (668)



Standard

9. Density in/out, lb/It
3

(kg/m
3
)
Vapor 0.39 (6.28)/0.43 (6.87) -
10. Allowable pressure drop, psi (kg/cm
2
) 1.14 (0.08) 14.2 (1.0)
11. Fouling resistance, h It
2
°F/Btu
(h m
2
°C /kcal)
0.00147 (0.0003) 0.00195 (0.0004)
12. Heat duty, MM Btu/h (MM kcal/h) 14.84 (3.04)
13. Design pressure, psig (kg/cm
2
g) 71 (5.0) 100 (7.0)
14. Material oI construction CS CS

103
transIer area. Also note that the stream analysis has improved considerably due to the larger
baIIle spacing and cut. This has also contributed towards the increase in the condensing heat
transIer coeIIicient.
The shell connections had to be increased Irom 10 in. (250 mm) to 16 in. (400 mm),
Table 7.5b: Salient Ieatures oI multiple shell designs Ior
distillation column overhead trim condenser
Original 4P design Revised 4P design New 2S 2P design
1. Type oI heat exchanger Floating-head (TEMA type AES)
2. Shell ID, in. (mm) 38.6 (980) 38.6 (980) 30.3 (770)
3. Heat transIer area, It
2
(m
2
)
4 × 2937 (273)
÷ 11,750 (1092)
4 × 2937 (273)
÷ 11,750 (1092)
4 × 1700 (158)
÷ 6800 (632)
4. Number oI tubes 590 590 340
5. Tube OD × thickness × length,
in. (mm)
0.984 (25) × 0.0984 (2.5) × 236 (6000)
6. No. oI tube passes 8 8 2
7. Tube pitch, in. (mm) 1.26 (32) square
8. BaIIle type double segmental single segmental double segmental
9. BaIIle spacing, in. (mm)/no.
oI cross passes
11.8 (300)/18 11.8 (300)/18 15.75 (400)/12
10. BaIIle cut, ° diameter
6 rows overlap
(29° area)
25
2 rows overlap
(41° on area)
Shellside 94 (459) 96 (469) 151.8 (741)
Tubeside 986 (4814) 986 (4814) 877 (4282)
11. Heat transIer
coeIIicient,
Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 63 (307.7) 64 (312.4) 83.8 (409)
Shellside
(0.08)/0.7
(0.05)
1.1 (0.08)/1.3
(0.09)
1.1 (0.08)/1.1
(0.08)
12. Pressure drop
allow/calc,
psi (kg/cm
2
)
Tubeside
14.2 (1.0)/13.5
(0.95)
14.2 (1.0)/13.5
(0.95)
14.2 (1.0)/5.6 (0.4)
Tubeside 5.1 (1.56) 5.1 (1.56) 4.4 (1.35)
13. Velocity,
It/s (m/s)
Shellside,
Cross/window
5.25 (1.6)/5.5
(1.69)
10.7 (3.26)/8.2
(2.49)
13.1 (3.98)/16.5
(5.04)
Main crossIlow 0.545 0.466 0.747
BaIIle-to-shell 0.17 0.121 0.073

14. Stream
analysis
Overall
eIIectiveness
0.693 0.643 0.833
15. MTD, °F (°C) 20 (11.1) 20 (11.1) 23.5 (13.06)
16. Overdesign 23.06 25.0 11.0
Shellside 10 (250)/10 (250) 10 (250)/10 (250) 16 (400)/16 (400) 17. Nominal
nozzle size,
in/out, mm
Tubeside 8 (200)/8 (200) 8 (200)/8 (200) 10 (250)/10 (250)
18. Condensing regime Gravity in all zones Gravity in all zones Shear/Transition
104
since there are now two shells in parallel instead oI Iour. A shell inside diameter oI 30.3 in.
(770 mm) was Iound adequate Ior this situation aIter incorporating a vapor belt at the shell
inlet.
The total weight oI the Iour bundles has reduced Irom approx. 22.7 tons in the original
design to approx. 13.2 tons in the 2S 2P (two shells in series and two shells in parallel)
design, and the total empty weight Irom approx. 41.4 tons to approx. 26.3 tons. Thus, the
revised design will be considerably cheaper.
All oI the above designs were saIe against Iailure oI tubes due to Ilow-induced vibration.
7.4.3 Condensation with desuperheating and/or subcooling
The vapor that enters a condenser is usually saturated, so that condensation begins right
away. This is always the case when the vapor comes Irom the top oI a distillation column.
However, it sometimes happens that the vapor is not a distillation column overhead and is
superheated. In such a case, there will be desuperheating until the vapor becomes
saturated at the operating pressure, Iollowed by condensation.
The condensate leaving a condenser is usually a saturated liquid but is sometimes
required to be subcooled Ior process reasons (e.g., the condensate is volatile and has to be
cooled to prevent Ilashing, it has valuable heat which can be recovered, or a distillation
column requires subcooled condensate as reIlux).
Thus, we may have any oI the Iollowing combinations in a condensing service:
1) only condensation
2) desuperheating and condensation
3) desuperheating, condensation, and subcooling
4) condensation and subcooling
The principal diIIerence between condensation and the other two phenomena are the
Iollowing:
1) Desuperheating involves the sensible cooling oI a vapor and consequently, its
heat transIer coeIIicient is extremely low.
2) Due to the low allowable pressure drop, the baIIle spacing in a condenser is Iairly
high.
When the vapor has totally condensed, however, the density increases considerably so
that iI the same baIIle spacing is employed, the velocity and thereIore heat transIer
coeIIicient will be rather low in the subcooling regime.
Let us consider desuperheating and subcooling separately.
Desuperheating
The tube-wall temperature will always be lower than the bulk vapor temperature, since
coolant Ilows on the other side. True desuperheating will exist only as long as the tube-
wall temperature is greater than the vapor saturation temperature. II the tube-wall
temperature is less than the dew point oI the vapor, it is said to be a 'wet wall condition.¨
However, iI the tube-wall temperature is greater than the dew point, it is said to be 'dry
wall condition.¨
The tube-wall temperature will depend upon both the heat transIer coeIIicient oI the
desuperheating vapor and that oI the coolant. As the Iormer will invariably be much lower
than the latter, the tube-wall temperature will tend to be close to the coolant bulk
temperature. Consequently, only a part oI the desuperheating heat duty will be transIerred as
105
gas cooling, while the rest will be transIerred as condensing. Evidently, how much oI the
heat duty will be transIerred as gas cooling (dry-wall) and how much as condensing (wet-
wall) will depend upon the extent oI superheat and the desuperheating and coolant heat
transIer coeIIicients. This will be observed in the case study presented later in this section.
In actual practice, the phenomenon is somewhat complicated since, even with wet-wall
desuperheating, the bulk vapor is still superheated. Thus, although condensate Iorms at the
tube wall, the uncondensed superheated vapor re-Ilashes some oI the condensate. This pro-
cess continues until the bulk vapor cools down to the dew point when true condensation
begins. For all practical purposes, however, the heat transIer coeIIicient in the wet-wall con-
dition is virtually the same as in the true condensing mode; in reality, it is somewhat lower.
All sophisticated heat exchanger thermal design soItware can handle desuperheating in
both the dry-wall and wet-wall modes described above.
The penalty (in the Iorm oI additional heat transIer area) associated with desuperheating
is usually rather small since the loss in the heat transIer coeIIicient is largely compensated by
the increase in MTD. The greater the superheat, the greater is the decrease in the heat
transIer coeIIicient but then, the greater is the increase in MTD. Consequently, Ior the same
total heat duty, the diIIerence in heat transIer area between only condensing and condensing
preceded by desuperheating is not appreciable. However, the pressure drop may increase
appreciably, especially iI the operating pressure is low.
Occasionally, a desuperheater and a condenser are separate shells. One advantage oI
such an arrangement is that the coolant may be heated to a temperature higher than the
saturation temperature, provided dry-wall desuperheating is ensured in the desuperheater.
For this purpose, the entire superheat is not removed in the desuperheater; a small part oI it
is carried on to the condenser.
Instead oI a separate desuperheater, an integral desuperheater may also be employed in
the same shell, shrouded in order to ensure dry-wall in the desuperheater. Such a
construction is commonly employed in power plant Ieedwater heaters wherein the Ieedwater
is heated to a temperature greater than the saturation temperature oI the condensing steam.
CASE STUDY 7.6: CONDENSATION WITH WET-WALL DESUPERHEATING
Let us consider the light hydrocarbon condenser presented in case Study 7.1. In that
example, the hydrocarbon entering the condenser was saturated. In order to demonstrate
the eIIect oI superheat on the perIormance oI a condenser, we shall consider Iour
arbitrary cases oI hydrocarbon superheat: 176ƒF (80ƒC), 194ƒF (90ƒC), 212ƒF (100ƒC),
and 248ƒF (120ƒC). The saturation temperature is 123.3ƒF (50.7ƒC). The cooling water
Ilow rate was increased in the same ratio as the heat duty, thereby retaining the same inlet
and outlet temperatures, 91.4ƒF (33ƒC) and 102.2ƒF (39ƒC). The existing design oI the
light hydrocarbon condenser in Case Study 7.1 (with the baIIle spacing oI 200 mm) was
rerun Ior these cases oI superheat and the results are reported in Table 7.6a, along with
that Ior the saturated case, itselI. The Iollowing will be observed in Table 7.6a:
a) With the gradual increase in the desuperheating temperature range and heat duty,
the shellside heat transIer coeIIicient reduces moderately Irom 364.8 Btu/h It
2
ƒF
(1781 kcal/h m
2
ƒC) to 307.4 Btu/h It
2
ƒF (1501 kcal/h m
2
ƒC), a drop oI about
19°.
b) The overall heat transIer coeIIicient reduces marginally Irom147.5 Btu/h It
2
ƒF
(720 kcal/h m
2
ƒC) to 142.1 Btu/h It
2
ƒF (694 kcal/h m
2
ƒC), a decrease oI only
3.6°. This is because (a) the shellside resistance contributes around 43° oI the
106
total resistance and so does the Iouling resistance and (b) there is an increase in
the tubeside heat transIer coeIIicient Irom 1147 Btu/h It
2
ƒF (5600 kcal/h m
2
ƒC)
to 1653 Btu/h It
2
ƒF (8069 kcal/h m
2
ƒC).
c) The MTD increases sharply Irom 26.5ƒF (14.7ƒC) to 34.4ƒF (19.1ƒC), thereby
more than compensating Ior the loss in the overall heat transIer coeIIicient.
Thus, there is a steady increase in the the heat Ilux (U × MTD) Irom 3906 Btu/h
It
2
(10,584 kcal/h m
2
) to 4891 Btu/h It
2
(3,255 kcal/h m
2
).
a) The overdesign changes substantially Irom 2.58° to (-)19.06°, principally due
to the increase in heat duty Irom 9.66 MM Btu/h (2.434 MM kcal/h) to 15.47
MM Btu/h (3.897 MM kcal/h) with increasing superheat oI the inlet hydrocarbon
stream.
CASE STUDY 7.7: CONDENSATION WITH DRY-WALL DESUPERHEATING
In Case Study 7.6 above, wet-wall conditions prevailed as the tube-wall temperature was
rather low. This was a direct consequence oI the high tubeside heat transIer coeIIicient
Table 7.6a: Results oI light hydrocarbon condenser with varying degrees oI superheat
Saturated Superheated
t
in
÷ 123.3
ƒF (50.7 ƒC)
t
in
÷ 176 ƒF
(80 ƒC)
t
in
÷ 194 ƒF
(90 ƒC)
t
in
÷ 212 ƒF
(100 ƒC)
t
in
÷ 248 ƒF
(120 ƒC)
1. Heat duty, MM BTU/h
(MM kcal/h)
9.66
(2.434)
12.2
(3.076)
13.0
(3.282)
13.83
(3.485)
15.47
(3.897)
2. CW Ilow, lb/h (kg/h)
895,240
(406,080)
1,131,400
(513,200)
1,207,000
(547,500)
1,281,500
(581,300)
1,433,000
(650,000)
3. Shellside (condensing) heat
transIer coeIIicient, Btu/h It
2

°F (kcal/h m
2
°C)
364.8
(1781)
337.7
(1649)
324.6
(1585)
317.7
(1551)
307.4
(1501)
4. Tubeside heat transIer
coeIIicient, Btu/h It
2
°F
(kcal/h m
2
°C)
1147 (5600) 1372 (6697) 1443 (7047) 1513 (7387) 1653 (8069)
5. Overall heat transIer
coeIIicient U, Btu/h It
2
°F
(kcal/h m
2
°C)
147.5
(720)
145.7
(711.5)
144
(702.8)
143
(699)
142.1
(694)
6. U × MTD, kcal/h m
2
10,584 11,171 11,526 11,955 13,255
7. Shellside pr. drop,
psi (kg/cm
2
)
2.94 (0.207) 3.44 (0.242) 3.5 (0.245) 3.6 (0.252) 3.7 (0.26)
8. Tubeside pr. drop,
psi (kg/cm
2
)
6.0 (0.42) 8.5 (0.6) 9.7 (0.68) 10.8 (0.76) 13.2 (0.93)
9. MTD, °F (°C) 26.5 (14.7) 28.3 (15.7) 29.5 (16.4) 30.8 (17.1) 34.4 (19.1)
10. Overdesign, ° 2.58 -13.3 -16.34 -18.3 -19.06
shell 40.43 43.16 44.35 45.08 46.24
tube 12.86 10.63 9.97 9.46 8.6
Iouling 43.2 42.69 42.17 41.94 41.64

11. ° thermal
resistance
metal 3.509 3.527 3.506 3.508 3.525
107
which resulted in a low temperature drop across the tubeside Iilm, thereby pulling down
the tube-wall temperature towards the tubeside bulk temperature. As wet-wall
desuperheating is not very inIerior to true condensing, the reduction in the condensing
heat transIer coeIIicient was not appreciable. In Iact, the MTD improved much more than
the overall heat transIer coeIIicient reduced. Thus, reIerring to Table 7.6a and comparing
the saturated hydrocarbon case with that oI a superheated inlet (248ƒF or 120ƒC), it will
be seen that the duty in the latter case is 60° higher whereas the diIIerence in overdesign
is 1.0258/(1 ² 0.1906) or 1.267. In other words, the overall perIormance was 1.6/1.267 or
1.263 times better.
Let us now consider an entirely diIIerent application, one in which superheat will result
in inIerior perIormance. This is a Iuel oil heater, with viscous Iuel oil being heated by steam.
The principal process parameters are presented in Table 7.7a.
A design was prepared with saturated steam and the salient construction and
perIormance parameters are indicated in Table 7.7b.
The steam condition is now changed to superheated with an inlet temperature oI 536ƒF
(280ƒC). When the same design was checked, it was Iound to have an overdesign oI 3.32°,
when a minimum overdesign oI 5° is required. Thus, the reduction in perIormance is 3.5°,
Irom an overdesign oI 7.08° to an overdesign oI 3.32°. The desuperheating zone has a
considerably lower heat transIer coeIIicient and, although the MTD is appreciably higher,
the heat Ilux is signiIicantly lower. The relevant details are presented in Table 7.7c. It will be
seen that, although the desuperheating heat duty is only 10.26° oI the total, it requires a heat
transIer area oI 16.6° oI the total. In situations where the desuperheating heat duty is higher,
the reduction in perIormance could be even more.
Table 7.7a: Salient process parameters Ior Case Study 7.7
Shellside Tubeside
1. Fluid Fuel oil Steam
2. Flow rate, lb/h (kg/h) 66,138 (30,000) 2463 (1117)
3. Operating pressure, psia (kg/cm
2
a) 284 (20.0) 142 (10.0)
4. Temperature in/out, °F (°C) 251.6 (122)/316.4 (158) 361.8 (183.2)/361.8 (183.2)
5. Viscosity in/out, cp 91/19.8
6. Thermal conductivity, Btu/h It °F
(kcal/h m °C)
0.0739(0.11)/0.0672 (0.1)
7. SpeciIic heat in/out, Btu/lb °F (kcal/kg °C) 0.48/0.5
8. Density in/out, lb/It
3
(kg/m
3
) 58.03 (930)/56.47 (905)


Standard
9. Allowable pressure drop, psi (kg/cm
2
) 10 (0.7) 1.4 (0.1)
10. Fouling resistance, h It
2
°F/Btu
(h m
2
°C/kcal)
0.0098 (0.002) 0.00049 (0.0001)
11. Hat duty, MM Btu/h (MM kcal/h) 2.1 (0.5292)
12. Design pressure, psig (kg/cm
2
g) 356 (25.0) 55 (13.0)
13. Material oI construction CS CS
108
Subcooling
This is a much more diIIicult situation than desuperheating. Usually some degree oI
subcooling is desired to prevent Ilashing in the condensate discharge and, thereby, cause
problems when the condensate is pumped. A small degree oI subcooling is accomplished
in a total condenser even without specially catering to it, but a larger degree oI
subcooling is generally not recommended in the same shell due to the Iollowing:
a) The heat transIer coeIIicient is very low due to the low velocity.
b) The true liquid level and MTD are very diIIicult to determine.
II subcooling is required in the same shell as the condenser, the best conIiguration is the
vertical intube. The worst conIiguration is the horizontal intube. The most common
condensation conIiguration is horizontal shellside. While this is not very well suited to
accomplish subcooling, it can be made to Iunction very eIIectively and economically by
incorporating a methodology described below.
The principal problem with condensation outside horizontal tubes is that the condensate
Table 7.7b: Salient Ieatures oI design oI Iuel oil heater
1. TEMA type BEU
2. Shell ID, in. (mm) 17.25 (438)
3. Heat transIer area, It
2
(m
2
) 704 (65.4)
4. Number oI tubes 186 (93 U-tubes)
5. No. oI tube passes 2
6. Tube OD × thickness × length, in. (mm) 0.7874 (20) × 0.0787 (2) × 217 (5500)
7. Tube pitch, in. (mm) 1.024 (26) rotated square
8. BaIIle spacing, in. (mm) 5.9 (150)
9. BaIIle cut, ° diameter 25
10. Nominal shellside cross/window velocity, It/s (m/s) 1.4 (0.43)/1.54 (0.47)
Main crossIlow 0.685
BaIIle to shell 0.267

11. Stream analysis
Overall eIIectiveness 0.71
Shellside 89.3 (436.1)
Tubeside 3243 (15,834)

12. HTC, Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 45.2 (220.6)
Shellside 5.6 (0.39)
13. Pressure drop, psi
(kg/cm
2
)
Tubeside 0.5 (0.032)
14. MTD, °C 70.7 (39.3)
15. Overdesign 7.08
Shellside 4 (100)/4 (100)
16. Nozzle size in/out, in.
(mm)
Tubeside 2 (50)/1 (25)
109
separates and Ialls to the bottom oI the shell by virtue oI its high density. Thus, the
condensate does not encounter the requisite heat transIer surIace.
The shell may be partially Ilooded by using a dam baIIle or a loop seal to provide
contact Ior condensate with tubes. However, subcooling cannot be predicted accurately since
the actual liquid level may be quite diIIerent, depending upon the pressure gradient in the
shell. Further, since all the tubes will be Ilooded along the entire length, the initial length oI
the Ilooded tubes will be unavailable Ior condensing. Thus, iI the shell is divided into two
parts, one Ior condensing and the other Ior subcooling, the Ilooded tubes in the Iirst part
(Irom the shell inlet) will be unavailable Ior condensing and the un-Ilooded tubes in the
second part will be unavailable Ior subcooling. This is represented in Fig. 7.6. Evidently,
such a design is very ineIIicient.
One way out is to employ two separate shells, one Ior condensing and the other Ior
subcooling. Considering that condensing and subcooling are two entirely diIIerent
phenomena, this is technically sound since each shell can then be made very eIIicient.
However, the cost oI the equipment increases considerably. II the condenser and the
subcooler are Iairly large units so that two shells will be required anyway, such an
arrangement will be acceptable. However, this is usually not the case so that an integral
condenser/subcooler represents a more economical design.
By using a special design, it is
possible to incorporate an eIIicient
condenser and an eIIicient subcooler
in the same shell. This is described in
the Iollowing case study.
CASE STUDY 7.8:
CONDENSATION WITH
INTEGRAL SUBCOOLING
A shell-and-tube condenser was
required Ior condensing distillation
column overhead. The principal
Table 7.7c: Analysis oI design oI Iuel oil heater with superheated steam
Desuperheating zone Condensing zone
1. Heat duty, MM Btu/h (MM kcal/h) 0.0531 0.4762
2. ° heat duty 10.26 89.74
3. Tubeside heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
20.2 (98.8) to 36.8 (179.9) 3872 (18,907) to 5057 (24,691)
4. Overall heat transIer coeIIicient,
Btu/h It
2
°F (kcal/h m
2
°C)
11.5 to 18.2 (56.2 to 88.7) 46.6 to 44.2 (227.5 to 215.7)
5. MTD, °F (°C) 193.4 (107.7) to 67.1 (37.3) 345.7 (25.4) to 109.6 (60.9)
6. Heat Ilux, Btu/h It
2
(kcal/h m
2
)
2235 to 1227
(6059 to 3326)
2855 to 4686
(7739 to 12,705)
7. Heat transIer area required, It
2
(m
2
) 116.8 (10.85) 587.5 (54.6)
8. ° heat transIer area oI total 16.6 83.4
Fig. 7.6 EIIect oI Ilooding in a horizontal shellside condenser
110
process parameters are listed in Table 7.8a. The inlet vapor is saturated so that
condensing begins at 203ƒF (95ƒC) and is over at 197.6ƒF (92ƒC). Subcooling proceeds
Irom 197.6ƒF (92ƒC) to the outlet temperature oI 114.8ƒF (46ƒC), so that the outlet
condensate is subcooled. The high viscosity oI the condensate may be noted: it will result
in an extremely low subcooling heat transIer coeIIicient.
Since the phenomena oI condensing and liquid subcooling are vastly diIIerent, it would
have been a normal practice to employ two shells in series, the upper one Ior condensing and
the lower one Ior subcooling. However, in the present instance, the allowable pressure drop
Ior the condensing/subcooling stream is only 0.14 psi (0.01 kg/cm
2
). As the use oI two shells
in series would have resulted in additional pressure drop in the extra nozzles and piping, it
became necessary to employ a single shell.
In power plant Ieedwater heaters, it is common to employ separate zones Ior de-
superheating and subcooling. This idea was applied to the subject condenser. A longitudinal
baIIle was used to divide the shell into two separate zones: an upper condensing zone and a
lower subcooling zone. As condensing was nearly isothermal and vapor shear was very high
due to the extremely low operating pressure, baIIles were not required to achieve a high
condensing heat transIer coeIIicient. Only cross-Ilow was employed with the required
support plates to reduce pressure drop to a minimum while still achieving a rather high heat
transIer coeIIicient.
In the subcooling zone, however, it was imperative to use baIIles to increase the
condensate velocity to a suIIiciently high value so as to obtain a reasonably high heat
Table 7.8a: Salient process parameters Ior Case Study 7.8

Shellside Tubeside
1. Fluid Distillation column overhead Cooling water
2. Flow rate, lb/h (kg/h) 3086 (1400) 88,200 (40,000)
3. Operating pressure, psia (kg/cm
2
a) 0.5 (0.035) 57 (4.0)

4. Temperature in/out, °F (°C)
203 (95)/114.8 (46)
dew point ÷ 203 (95)
bubble point ÷ 197.6 (92)

95 (35)/109.4 (43)
5. Vapor molecular weight 79.0 -
6. Viscosity liquid in/out, cp 3.1/13.1
7. Thermal conductivity liq/vap,
Btu h It °F (kcal/h m °C)
0.114 (0.17)/0.0114 (0.017)
8. SpeciIic heat liq/vap, Btu/lb °F (kcal/kg °C) 0.667/0.5
9. Sp. gravity liquid 0.99 ¸ 203 °F (95 °C)


Standard
10. Allowable pressure drop, psi (kg/cm
2
) 0.14 (0.01) 7.1 (0.5)
11. Fouling resistance, h It
2
°F/Btu (h m
2
°C /kcal) 0.00098 (0.0002) 0.00195 (0.0004)

12. Heat duty, MM Btu/h (MM kcal/h)
Total ÷ 1.27 (0.32)
Condensing ÷ 1.1 (0.277)
Subcooling ÷0.17 (0.043)
13. Design pressure, psig (kg/cm
2
g) Iull vacuum/21 (1.5) 100 (7.0)
14. Design temperature, °F (°C) 252 (122) 158 (70)
111
transIer coeIIicient. The construction is akin to a TEMA G shell; however, as will be seen
later, the number oI tubes in the upper (condensing) and lower (subcooling) zones is quite
diIIerent. Thus, it became necessary to run the condensing and subcooling sections
individually as TEMA F shells and consider the results oI the appropriate pass only Ior each
run. Several runs had to be taken in order to optimize the distribution oI tubes in the
condensing and the subcooling zones so that the overall number oI tubes and, thereby, the
cost were the minimized.
The Iinal key results, as well the interesting construction Ieatures, that Iinally emerged
are elaborated in Table 7.8b. It will be noticed that, although the subcooling heat duty
(170,000 Btu/h or 43,000 kcal/h) was only 13.4° oI the total (1,270,000 Btu/h or 320,000
kcal/h), its required heat transIer area was as high as 71.4° oI the total. This was attributable
to the much lower MTD and overall heat transIer coeIIicient. Consequently, the condensing
zone had only 74 tubes whereas the subcooling zone had 184.
7.4.4 Nozzle sizing
Nozzles are sized on the basis oI ρv
2
, which determines pressure drop and the tendency to
Table 7.8b: Salient Ieatures oI very low pressure overhead condenser design
Condensing zone Subcooling zone
1. Type oI heat exchanger Fixed-tubesheet
2. Shell ID, in. (mm) 23.6 (600)
3. Heat transIer area, It2 (m2) 114 (10.6) 284 (26.4)
4. Flow pattern on shellside CrossIlow BaIIled (up-and-over)
5. Number oI tubes 74 184
6. Tube length, It (mm) 8.2 (2500)
7. Tube OD × thickness, in. (mm) / (19.05) × 16 BWG (1.65)
8. No. oI tube passes 2 4
9. Tube pitch, in. (mm) 1.0 (25.4) square
10. BaIIle spacing, in. (mm) Only support plates 4.0 (100)
11. BaIIle cut, ° diameter Nil 25 ° (hor.)
Shellside 273 (1333) 16.7 (81.5)
Tubeside 958 (4677) 958 (4677)
12. Heat transIer coeII.,
Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 116.1 (567) 15.4 (75)
13. MTD, °F (°C) 97.2 (54) 43.7 (24.3)
14. Overdesign, ° 17 11.9
Shellside Negligible
15. Pressure drop,
psi (kg/cm
2
)
Tubeside 5.0 (0.35)
16. Cooling water velocity, It/s (m/s) 4.4 (1.33)
Shellside 12 (300)/4 (100)
17. Nominal nozzle size, in/out,
in. (mm)
Tubeside 4 (100)/4 (100)
112
erode. The latter assumes added signiIicance Ior condensers, especially Ior saturated
vapors. This is because there may be a Iew droplets oI condensate in the vapor stream
entering a condenser which will then travel at the vapor velocity, produce an extremely
high ρv
2
and thereby tend to erode tubes in the top row. Thus, it is always advisable to
incorporate an impingement plate just below the shell inlet nozzle oI a condenser.
It is a usual practice to limit the pressure drop in the inlet and outlet nozzles oI a heat
exchanger to about 20° oI the total allowed Ior either side (shellside or tubeside) and the
same holds true Ior condensers. This is to leave suIIicient pressure drop Ior the shell or tubes
proper where heat transIer actually takes place. Since there is expansion at the shell (or
channel) inlet and contraction at the shell (or channel) outlet, the pressure drop in the inlet
nozzle is always greater than that in the outlet nozzle Ior the same ρv
2
. A ρv
2
oI 1344-2016
lb/It sec
2
(20003000 kg/m sec
2
) is usually considered Ior the sizing oI nozzles Ior gas or
vapor-liquid services. For liquids, a ρv
2
oI 3360 lb/It sec
2
(5000 kg/m sec
2
) is quite normal
and 4030 lb/It sec
2
(6000 kg/m sec
2
) an upper limit. This is Ior the simple reason that a
higher pressure drop is permitted Ior liquids than Ior gases, condensing vapors, and
vaporizing liquids. It must be evident that the higher the value oI ρv
2
permitted, the smaller
will be the size oI a nozzle.
It must be stated here that it is very important to size condensate nozzles Ior proper
drainage so as to provide weir-type Ilow rather than a Ilooded drain pipe. In the latter
situation, Ilooding may result with a deleterious eIIect on the perIormance oI the condenser.
It is a common practice to size steam condensate nozzles Ior a velocity oI 2.0 It/s (0.6 m/s)
and hydrocarbon condensate nozzles Ior 3.3-4.0 It/s (11.2 m/s) in order to achieve weir
type Ilow.
7.4.5 Condensing profiles and MTD
In single-phase applications, the variation oI heat duty versus temperature, or the heat
release proIile, is essentially linear. Essentially, because vapor or liquid speciIic heat does
increase with temperature, but this variation is not large enough to impart an appreciable
curvature to the heat release proIile.
However, in the case oI condensers, there is usually an appreciable curvature in the heat
release proIile. This is particularly true when the condensing stream is a multicomponent
mixture. Due to the eIIect oI partial pressures, the less volatile components will condense
Iirst, and the more volatile components later. Thus, since more heat duty will be transIerred
per unit oI temperature diIIerence at the hotter end oI the condenser than at the colder end,
the slope oI the curve will be steeper at the hotter end than at the colder end, as shown in
Fig. 7.7a. However, a pure compo-
nent condenser will have an isother-
mal heat release proIile, as shown in
Fig. 7.7b.
The situation is Iurther compli-
cated when desuperheating and/or
subcooling zones are also present. A
typical desuperheating, condensing,
and subcooling situation is shown in
Fig.7.7c. Evidently, iI the condensing
stream is a pure component, the
proIile in the condensing zone will
be linear. Fig. 7.7a Condensing proIile oI a multicomponent mixture
113
It is extremely important to Ieed the condenser heat release proIile to the thermal design
soItware. The entire proIile will have to be divided into a suIIiciently high number oI
virtually straight-line segments and the calculations perIormed segment-wise. This is
important because not only the MTD, but the shellside condensing heat transIer coeIIicient
will vary signiIicantly Irom inlet to outlet. The higher condensing duty at the inlet end will
produce a correspondingly higher heat transIer coeIIicient.
Besides heat duty, the vapor weight Iraction will also have to be Ied against temperature.
This can be easily understood as the relative amounts oI vapor and liquid in the condensing
stream will determine the heat transIer coeIIicient and pressure drop in various segments
(increments oI tube length).
7.4.6 Low-pressure condenser design
The thermal design oI condensers operating at low pressures represents a problem:
restricting the pressure drop to within the permitted value. There is no hard-and-Iast
deIinition oI low pressure: 2.8²4.2 psia (2²3 kg/cm
2
abs.) or lower may be considered to
be low pressure. The lower the operating pressure, the greater the diIIiculty in complying
with the permissible pressure drop. This is essentially due to the low vapor density oI the
vapor being handled.
Conaensing on the tubesiae
When the condensing stream is on the tubeside, the variables are the diameter and length
oI the tubes and the number oI tube passes. In the case oI Iixed-tubesheet condensers, the
number oI tube passes is either one or two. Thus, the number oI tube passes is a big step
change Ior pressure drop because,
Ior a given number oI tubes oI a
given diameter and thickness,
reducing the number oI passes Irom
two to one results in an ap-
proximately seven-Iold reduction in
pressure drop.
In U-tube condensers, the num-
ber oI tube passes is invariably two.
In Iloating-head condensers, the
number oI tube passes is also nor-
mally two. Every additional tube
pass involves exit oI the two-phase
stream Irom the tubes into the
header, turn-around, and entry Irom
the header into the tubes oI the next
passes. Since this represents an in-
creased possibility oI Ilow maldis-
tribution, a larger number oI tube
passes should not be employed.
Besides, an increase in the number oI
tube passes does not result in an
appreciable increase in the tubeside
heat transIer coeIIicient in conden-
sers, unlike in single-phase applica-
Fig. 7.7b Condensing proIile oI a pure component
Fig. 7.7c Desuperheating, condensing, and subcooling proIile
114
tions, whereas the pressure drop in-
creases appreciably.
In a Iloating-head condenser, a
single tube pass construction entails
potential mechanical problems in that
the tubeside outlet connection has to
begin at the Iloating-head cover and
pierce the shell cover (see Fig. 7.8),
consequently requiring the use oI an
expansion bellows or expansion joint
on the tailpipe. This makes the
Iloating-head cover/shell cover prone
to internal leakage. In a Iixed-tubesheet condenser, a single tube pass does not represent any
problem.
CASE STUDY 7.9: TUBESIDE CONDENSATION
Let us consider the overhead condenser presented in Case Study 7.2. Both oI the streams,
distillation column overhead and demineralized water, are clean and the service has a low
temperature diIIerence between shell and tube sides; hence, a Iixed-tubesheet con-
struction can be used and the streams switched Irom shellside to tubeside without any
problem.
Let us study the eIIect oI routing the condensing stream through the tubeside. For the
Iirst run, the basic condenser geometry was retained identical to the original geometry with
the Iollowing essential changes:
a) connection sizes interchanged between shellside and tubeside
b) number oI tube passes reduced Irom Iour to one, as tubeside pressure drop will
be excessive with more than one tube pass
c) the baIIles were changed Irom double-segmental to single-segmental and the
baIIle cut to 25° on diameter
a) shell type changed Irom AES to AEL, i.e., a Iixed-tubesheet construction (a
Iloating-head construction is contraindicated Ior a single tube pass and besides, it
is not even required here)
e) The number oI tubes could have been reduced in view oI the Iixed-tubesheet
construction and the smaller shell inlet nozzle, but was deIerred as this was only
the Iirst run.
The construction and perIormance data oI the original design having condensing on the
shellside, and the revised design having condensing on the tubeside are shown in Table
7.9a. It will be seen that the perIormance oI the condenser is not much diIIerent between
the two designs: condensation on the shellside and condensation on the tubeside, except
Ior a much lower pressure drop oI the condensing stream in the latter. ThereIore, an
attempt was made to reduce the heat transIer area by taking advantage oI this margin in
pressure drop. Increasing the number oI tube passes Irom one to two yielded an
overdesign oI 17° but the tubeside (condensing) pressure drop shot up to 3.7 psi or 0.26
kg/cm
2
, well in excess oI the permissible value oI 1.4 psi or 0.1 kg/cm
2
.
An attempt was then made to reduce the number oI tubes and, correspondingly, the shell
ID. It was Iound that the shell ID could be reduced Irom 40.4 in. (1025 mm) to 36.4 in. (925
mm) and the number oI tubes Irom 966 to 890 and still have an overdesign margin oI about

Fig. 7.8 Single tube pass Iloating-head condenser with tailpipe
and expansion bellows (Reprinted Irom the Heat Exchanger
Design Handbook, 2002 with permission oI Begell House,
Inc.)
115
5°. The tubeside pressure drop went up marginally Irom 0.67 psi (0.047 kg/cm
2
) to 0.75 psi
(0.053 kg/cm
2
). Thus, due to the constraint oI having to use 0.787-in. (20-mm) OD, 0.0787-
in. (2-mm) thick, and 236-in. (6000-mm) long tubes, the permissible condensing stream
pressure drop could not be utilized eIIectively.
To demonstrate the eIIect oI tube orientation, the tubes were then changed Irom horizon-
tal to vertical, everything else remaining identical, and the results are again shown on Table
7.9a. It will be seen that the results are quite similar, with a small decrease in the tubeside
Table 7.9a: Salient construction and perIormance Ieatures
oI various designs Ior overhead condenser
Condensation on tubeside
Horizontal unit

Original design,
condensation on
shellside
Geometry
identical to
original design
Reduced
heat transfer
area
Vertical
unit
1. Type oI heat exchanger Floating-head Fixed-tubesheet Fixed-tubesheet Fixed-tubesheet
2. Shell ID, in. (mm) 40.4 (1025) 40.4 (1025) 36.4 (925) 36.4 (925)
3 Heat transIer area,
It
2
(m
2
)
3830 (356) 3830 (356) 3540 (329) 3540 (329)
4. Tube OD × thick. ×
length × tube pitch, in.
(mm)
0.7874 (20) × 0.0787 (2) × 236 (6000) × 1.024 (26) square
5. Number oI tubes 966 966 890 890
6. No. oI tube passes 4 One One One
7. BaIIle type double segm. single segm. single segm. Single segm.
8. BaIIle spacing, in.
(mm)/no. oI cross
passes
11.8 (300)/18 11.8 (300)/18 11.8 (300)/18 11.8 (300)/18
9. BaIIle cut, ° diameter
2 rows overlap
(44.3° area)
25 25 25
10. Nominal connection
size in/out, in. (mm)
Ovhd. 16 (450)/6 (150), Demin. water 6 (150)/6 (150)
Overhead 160.8 (785) 156 (761) 166.5 (813) 160 (779)
Demin.
water
877 (4283) 987 (4817) 1127 (5502) 1127 (5502)
11. Heat
transIer coeII.,
Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 101.9 (497.6) 96 (469) 107 (522.6) 104 (508.5)
Overhead 1.4 (0.1) 0.67 (0.047) 0.75 (0.053) 0.8 (0.056)
12. Pressure
drop, kg/cm
2
Demin.
water
2.8 (0.2) 2.4 (0.17) 2.7 (0.19) 2.7 (0.19)
13. MTD, °F (°C) 49.9 (27.7) 50.8 (28.2) 50.9 (28.3) 50.9 (28.3)
14. Overdesign 5.86 6.88 4.84 3.09
15. Vapor Reynold`s no. 107,124 to 3278 128,216 to 6085 139,054 to 6384 137,902 to 7972
116
(condensing) heat transIer coeIIicient and a small increase in the tubeside (condensing)
tubeside pressure drop.
Conaensing on the shellsiae
Let us now consider low-pressure condensing on the shellside, which is Iar more
prevalent than condensing on the tubeside. In Section 3.4.6 (Reduction oI shellside
pressure drop), various methodologies were discussed Ior reducing shellside pressure
drop. Essentially, the variables are type oI shell and type oI baIIling. The Iollowing order
should be Iollowed progressively in order to reduce shellside pressure drop:
TEMA E shell and single-segmental baIIles
TEMA E shell and double-segmental baIIles
TEMA J shell and single-segmental baIIles
TEMA J shell and double-segmental baIIles
TEMA E shell and NTIW single-segmental baIIles
TEMA X shell
That is, Iirst an E shell should be attempted with single-segmental baIIles. Should it not
be possible to produce a satisIactory design with it, the E shell should be retained but the
baIIling changed to double-segmental. Thus, the designer should proceed through a J
shell to NTIW baIIles and Iinally, an X shell.
There is a strong invisible link between shellside pressure drop and Ilow-induced
vibration. The pressure drop problem really translates into a vibration problem. By making
the baIIle spacing suIIiciently high, the shellside pressure drop can be contained within the
allowable limit. However, this increases the unsupported tube span to such an extent that the
condenser becomes prone to Iailure oI tubes due to Ilow-induced vibration. Thus, not only
must the allowable shellside pressure drop be respected, but the condenser tubes should also
be saIe against Iailure due to Ilow-induced vibration. This is explained in detail in Chapter
12 on Ilow-induced vibration.
In the extreme case oI steam-jet ejector condensers operating at around 5060 mm Hg, it
becomes imperative to employ a TEMA X construction to satisIy the extremely low
allowable pressure drop oI 57 mm Hg. See Section 7.5.2 Ior a case study.
7.5 Special Applications
Some special applications in condenser design are described in the Iollowing sections.
7.5.1 Use of low-fin tubes
Finned tubes are categorized into low-Iin and high-Iin tubes, depending upon Iin density.
Low-Iin tubes range Irom 32 Iins/in. (1260 Iins/m) and 0.032-in. (0.81-mm) Iin height to
11 Iins/in. (433 Iins/m) and 0.117-in. (2.97-mm) height. The smaller Iin heights are Iar
more popular. High-Iin tubes have 5²11 Iins/in. (197²433 Iins/m), with the Iin height
being 1/2 in. or 5/8 in. Low-Iin tubes are employed in shell-and-tube heat exchangers
while high-Iin tubes are invariably employed in air-cooled heat exchangers.
Bare tubes are generally used Ior condensation (as indeed Ior all shell-and-tube heat
exchanger applications) because they are the cheapest, but in certain condensing
applications, the use oI low-Iin tubes can result in an even cheaper design. The Iins not only
increase the surIace area, but also introduce surIace tension eIIects that can play a signiIicant
role in reducing the thickness oI the condensate Iilm, thereby increasing the heat transIer
coeIIicient. As per the Gregorig eIIect, condensation on a low-Iin tube occurs primarily at
117
the top oI the Iins. SurIace tension Iorces then pull the condensate into the grooves, where it
runs oII. However, whereas only a small wedge oI liquid exists at the top oI the tube, there is
'Ilooding¨ at the bottom oI the tubes. The resultant heat transIer coeIIicient is appreciably
greater than that Ior a uniIorm condensate Iilm thickness.
Evidently, Iin geometry (namely spacing, thickness, and height oI Iins) will play a
signiIicant eIIect in the perIormance oI low-Iin tubes. As Iin spacing decreases, there will be
more Ilooding and, at a critical Iin spacing, it is possible Ior the entire tubes to Ilood and,
thereby, heat transIer will suIIer. Consequently, the Iin spacing must be optimized Ior best
perIormance.
Low-Iin tubes are advantageously used when the condensing heat transIer coeIIicient is
controlling in the heat transIer process. Evidently, these Iin tubes Iind use in shellside
condensers. However, the use oI these tubes is contraindicated Ior Iouling services which
will clog the gap between the Iins and have a deleterious consequence on the perIormance oI
the heat exchanger. Another contraindication is that the surIace tension oI the condensate
should not be very high because then condensate may not run oII the ridges, but tend to
build up, even at the top oI the tubes. Consequently, low-Iin tubes should not be used Ior the
condensation oI steam. Fortunately, as the heat transIer coeIIicient oI condensing is high,
even with bare tubes (due to the very high thermal conductivity oI its condensate), there is
really no need to use low-Iin tubes Ior this service. Thus, the best application Ior low-Iin
tubes is clean hydrocarbon streams, such as propylene.
CASE STUDY 7.10: USE OF LOW-FIN TUBES
A light hydrocarbon condenser was to be designed Ior the salient process parameters
shown in Table 7.10a. Tubes /-in. (19.05-mm) OD × 14 BWG (2.108mm thick) × 236-
in. (6000-mm) long oI carbon steel were to be used. A TEMA AEL (Iixed-tubesheet)
construction was to be employed in view oI the clean service on the shellside.
A design was produced with a heat transIer area oI 2055 It
2
(191 m
2
), having a shell ID
oI 26 in. (660 mm). The shellside resistance was 48.88° and the tubeside only 11.63°, a
ratio oI 4.2. It was thereIore Ielt that it may be a good idea to use low-Iin tubes in order to
reduce the cost oI the condenser.
Table 7.10a: Salient process parameters Ior Case Study 7.10
Shellside Tubeside
1. Fluid Light hydrocarbon Cooling water
2. Flow rate, lb/h (kg/h) 48,500 (22,000) 679,000 (308,000)
3. Operating pressure, psia (kg/cm
2
a) 240 (16.9) 71 (5.0)
4. Temperature in/out, °F (°C)
172.4 (78)/123.1 (50.6)
dew point and bubble point
÷ 123.1 (50.6)

91.4 (33)/102.2 (39)
5. Allowable pressure drop, psi (kg/cm
2
) 2.84 (0.2) 7.1 (0.5)
6. Fouling resistance, h m
2
°C/kcal 0.00049 (0.0001) 0.00195 (0.0004)
7. Heat duty, MM Btu/h (MM kcal/h) 7.326 (1.846)
8. Design pressure, psig (kg/cm
2
g) 313 (22.0) 100 (7.0)
9. Material oI construction CS CS
118
Using Wolverine-type low-Iin tubes having 0.0535-in. (1.36-mm) Iin height and 16
Iins/in. (630 Iins/m), the shell ID could be reduced to 20.7 in. (525 mm). The approximate
empty weight oI the condenser reduced Irom 5200 kg to 4100 kg. Depending upon the cost
oI bare and low-Iin tubes, the cost oI this condenser might be reduced by the use oI low-Iin
tubes.
As the light hydrocarbon condensing heat transIer coeIIicient was eIIectively enhanced
by the use oI low-Iin tubes, Iouling became the major resistance (48.2°) in the low Iin-tube
design. II the cooling water Iouling resistance were lower, the reduction in cost would have
been appreciable. Thus, low-Iin tubes are ideal Ior services that have clean streams both
inside and outside tubes.
The salient Ieatures oI both the bare tube and the low-Iin tube designs are shown in
Table 7.10b. It is interesting to note that, on the shellside, both the inlet and the outlet
connections are 8 in. (200 mm) nominal. Usually, a condensate line is smaller than the
corresponding line due to its higher density; however, a 6-in. (150-mm) nominal line, in this
case, was going to result in condensate Ilooding in the shell, and the same was thereIore
Table 7.10b: Salient Ieatures oI bare-tube and low-Iin designs Ior light hydrocarbon condenser
Design with bare tubes Design with low-fin tubes
1. Type oI heat exchanger Fixed tubesheet (TEMA AEL)
2. Shell ID, in. (mm) 26 (660) 20.7 (525)
3. Heat transIer area, It
2
(m
2
) 2055 (191) (bare OD) 2841 (264) (Iinned)
4. Number oI tubes 542 350
5. Tube OD × thickness × length,
in. (mm)
/ (19.05) × 16 BWG (1.65 mm) × 236 (6000)
6. No. oI tube passes 2 One
7. Tube pitch, in. (mm) 0.984 (25) triangular
8. BaIIle spacing, in. (mm) 11.8 (300) 7.9 (200)
9. BaIIle cut, ° diameter 25 35
Shellside 265 (1295) 369 (1803)
Tubeside
(on ID)
1114 (6582) 1637 (7994)
10. Heat transIer
coeIIicient,
Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 632.8 92.4 (451) (on Iinned area)
11. Overdesign, ° 3.26 4.5
12. MTD, °F (°C) 29.2 (16.2) 29.2 (16.2)
Shellside 2.0 (0.14) 2.84 (0.2) 13. Pressure drop, psi
(kg/cm
2
)
Tubeside 6 (0.42) 5.1 (0.36)
14. Cooling water velocity, It/s (m/s) 5.5 (1.68) 4.5 (1.97)
Shellside 8 (200)/8 (200) 15. Nozzle size,
in/out,
in.(mm)
Tubeside 8 (200)/8 (200)
16. Approximate empty weight, lb (kg) 11,500 (5200) 9040 (4100)
119
increased to 8 in. (200 mm) nominal.
7.5.2 Vacuum condenser design: Ejector condensers and surface condensers
Ejector condensers and surIace condensers are special cases oI condensers operating in
vacuum. These invariably employ crossIlow shells as the vapor density is exceptionally
low. A typical design is illustrated in the Iollowing case study.
CASE STUDY 7.11 E1ECTOR INTERCONDENSER
Consider the Ejector Intercondenser service depicted in Table 7.11a. A mixture oI steam
and air (with a negligible amount oI hydrocarbons) is to be condensed at 5.26 psia (0.37
kg/cm
2
absolute) by cooling water at 93.2ƒF (34ƒC). In view oI the relatively clean
service (mostly steam) on the shellside, a Iixed-tubesheet construction was permitted.
Thus, the tube length was not Iixed and could be optimized. The tubes were to be oI
Admiralty Brass, 0.984-in. (25-mm) OD and 0.0787-in. (2-mm) thick.
As the shellside operating pressure is extremely low, and the allowable pressure drop
only 0.14 psi (0.01 kg/cm
2
), a baIIled construction employing a single-pass or divided-Ilow
shell is simply not Ieasible. A no-tubes-in-window design was examined, but even that could
not limit the shellside pressure drop to the permitted value. Hence, a pure crossIlow (TEMA
X) shell was employed and a design produced. The principal construction and perIormance
parameters oI this design are shown in Table 7.11b.
The Iollowing construction Ieatures may be noted:
1) The shell inlet nozzle area was very high and, thereIore, three inlet nozzles oI 12
in. (300 mm) nominal were employed on the shellside. These were to be located
along the length oI the shell Ior an even distribution. This was possible only
because oI crossIlow on the shellside.
2) The tube length was optimized at 14.8 It (4500 mm) since this produced an
acceptable tubeside velocity oI 4.8 It/s (1.46 m/s) and an acceptable tubeside
pressure drop oI 5 psi (0.35 kg/cm
2
).
Table 7.11a: Salient process parameters Ior Case Study 7.11: Ejector intercondenser
Shellside Tubeside
1. Fluid
Air: 14° by wt.
Steam: 86° by wt.
Cooling water
2. Flow rate, lb/h (kg/h) 19,840 (9000) 1136,500 (515,500)
3. Operating pressure, psia (kg/cm
2
a) 5.26 (0.37) 71 (5.0)
4. Temperature in/out, °F (°C)
266 ((130)/113 (45)
dew point 158 (70)
93.2 (34)/109.4 (43)
5. Allowable pressure drop, psi (kg/cm
2
) 1.42 (0.01) 0.5
6. Fouling resistance, h It
2
°F/Btu (h m
2
°C
/kcal)
0.00146 (0.0003) 0.00195 (0.0004)
7. Heat duty, MM Btu/h (MM kcal/h) 18.42 (4.64)
8. Design pressure, psig (kg/cm
2
g) 50 (3.5) 100 (7.0)
9. Material oI construction CS CS
120
3) A rotated triangular tube pitch (60ƒ) was used in order to eliminate the possibility
oI acoustic vibration which was indicated with a triangular tube pitch (30ƒ).
4) A total oI three Iull support plates were employed. Within each 'baIIle¨ space,
three support plates were used to eliminate any possibility oI Iailure oI tubes due
to Ilow-induced vibration.
The Iollowing perIormance Ieatures are interesting to note:
1) The desuperheating heat duty is approx. 11.6° oI the total, and the desuper-
heating heat transIer area is only about 8.3° oI the total. This is because,
although the desuperheating heat transIer coeIIicient is signiIicantly lower than
that in the condensing zone, MTD is Iar greater.
2) The shellside pressure drop is predominantly in the nozzles, about 65° in the
inlet nozzle and about 28° in the outlet.
3) The rate oI condensation, as shown in Fig. 7.9, is very steep initially, Irom the
dew point oI 158ƒF (70ƒC) to about 149ƒF (65ƒC), aIter which it decreases
rapidly and then Ilattens out, due to the partial pressure exerted by the
noncondensables.
Table 7.11b: Salient Ieatures oI design oI ejector intercondenser
1. TEMA type AXL
2. Shell ID, in. (mm) 48.2 (1225)
3. Heat transIer area, It
2
(m
2
) 4239 (394)
4. Number oI tubes × tube OD × thickness, in. (mm) 1140 × 0.984 (25) × 0.0787 (2)
5. Tube length, It (mm) 14.8 (4500)
5. No. oI tube passes 4
6. Tube pitch, in. (mm) 1.25 (31.25) rotated triangular (60°)
7. No. oI tube supports/support spacing, in. (mm) 3/43.3 (1100)
8. No. oI intermediate supports in each cross-pass 3
9. Cooling water velocity inside tubes, It/s (m/s) 4.8 (1.46)
Main crossIlow 0.934
10. Stream analysis
Bundle to shell 0.066
Steam ¹ air 163.4 (798)
Cooling water 997 (4866)

11. HTC, Btu/h It
2
°F (kcal/h m
2
°C)
Overal 90.8 (443.5)
Steam ¹ air 1.42 (0.01)
12. Pressure drop, psi (kg/cm
2
)
Cooling water 5.0 (0.35)
13. MTD, °F (°C) 50 (27.8)
14. Overdesign 4.56
Shellside 3 × 12 (300)/10 (250)
15. Nozzle size in/out, in. (mm)
Tubeside 14 (350)/14 (350)
121
4) Condensation is entirely in
the gravity-controlled re-
gime because the Ilow area
was very large and the
vapor velocities very low.
The condensing heat trans-
Ier coeIIicient Ialls very
sharply Irom around 1229
Btu/h It
2
ƒF (6000 kcal/h m
2

ƒC) at the onset oI conden-
sation to as low as 30.7
Btu/h It
2
ƒF (150 kcal/h m
2

ƒC) at the lowermost tubes.
This is a direct result oI the sharp reduction in vapor Reynold`s number and,
thereby, the sensible vapor cooling heat transIer coeIIicient. As the MTD also
decreases appreciably Irom top to bottom, the heat Ilux reduces proIoundly.
Thus, Ior the last 14° heat duty, the heat transIer area required is as high as 42°,
and Ior the last 3° heat duty, as high as 17°.
5) As the shellside experiences pure crossIlow, there are only two streams: the main
crossIlow stream and the bundle-shell bypass stream. As the shell ID is very
high, the main crossIlow stream is also very high and there is very little
bypassing around the tube bundle.
References
|1| Hewitt, G.F., 1998, Heat Exchanger Design Hanabook, Volume 3, Section 3.4.2. Begell
House, Inc., New York.
|2| Butterworth, D., and Hewitt, G.F., 1977, Two-Phase Flow ana Heat Transfer, OxIord
University Press.
|3| Collier, J.G., 1972, Convective Boiling ana Conaensation, McGraw-Hill.
|4| Taborek, J., Hewitt, G.F., and AIgan, N. (eds.), 1983, Heat Exchangers. Theory ana Practice,
Hemisphere Publishing Corp.
|5| Kakac, S., (ed.), 1991, Boilers, Evaporators ana Conaensers, John Wiley and Sons, Inc.
|6| Kakac, S., Bergles, A.E., and Mayinger, F., 1981, Heat Exchangers. Thermal-Hyaraulic
Funaamentals, Hemisphere Publishing Corp.
|7| English, K.G., Jones, W.T., Spillers, R.C., and Orr, V., 1963, 'Flooding in a Vertical UpdraIt
Partial Condenser,¨ Chemical Engineering Progress, 59(7), pp. 51²53
|8| Diehl, J.E., and Koppany, C.R., 1969, 'Flooding Velocity Correlation Ior Gas-Liquid
CounterIlow in Vertical Tubes,¨ Chemical Engineering Progress Symp., Series 65, No. 92,
pp. 77²83.
Fig. 7.9 Variation oI weight Iraction vapor versus temperature
Ior Case Study 7.11
122

123
CHAPTER 8
7KHUPDO'HVLJQDQG
2SWLPL]DWLRQRI5HERLOHUV
A reboiler is a heat exchanger whose primary purpose is to vaporize the bottom oI a
distillation, absorption, or stripping column and return the vapor to the same column.
BeIore commencing a discussion on reboilers, let us Iirst brieIly discuss the phenomenon
oI boiling, itselI.
Boiling is an extremely complex heat transIer phenomenon and one that is currently the
most researched. What makes boiling particularly diIIicult and yet Iascinating is that it is the
only heat transIer phenomenon where the heat transIer coeIIicient is a Iunction oI the
temperature diIIerence across the Iilm.
Two types oI boiling are encountered in reboilers: pool boiling and Ilow (convective)
boiling. Kettle and internal reboilers embody the Iormer while thermosyphon reboilers
represent the latter. OI course, Iorced-Ilow 'reboilers¨ do not experience boiling at all.
8.1 Pool Boiling
Let us Iirst consider pool boiling. A typical pool boiling curve is shown in Fig. 8.1. This
curve is the starting-point Ior all discussions on pool boiling and is one oI the Iew rela-
tions in the boiling phenomenon that is well understood.
The pool boiling curve is a plot oI heat Ilux as a Iunction oI temperature diIIerence, ǻT.
The heat Ilux is the amount oI heat transIerred per unit area oI heat transIer surIace.
Consequently, it is also the product oI the heat transIer coeIIicient and ǻT as
Q ÷ UAǻT (8.1)
where
Q ÷ heat duty
U ÷ heat transIer coeIIicient
A ÷ heat transIer area
ǻT ÷ temperature diIIerence
There are Iive clearly-deIined re-
gimes in the pool boiling curve:
Regime A-B. Natural convection
A certain degree oI liquid superheat
is required beIore nucleation can
begin and this corresponds to a cer-
Fig. 8.1 Pool boiling curve
124
tain ǻT. When ǻT is below this value, there is no nucleation and only sensible heat
transIer by natural convection takes place.
Regime B-C. Incipient boiling
In this regime, ǻT is suIIiciently high Ior nucleation to commence and heat transIer takes
place by a combination oI nucleation and natural convection. The heat transIer coeIIicient
is approximately the sum oI the natural convection and nucleate boiling heat transIer co-
eIIicients. This regime may be considered to be the transition between natural convection
and true nucleation.
Regime C-D. Nucleate boiling
This is the regime oI true nucleate boiling, which is the desired mode oI boiling heat
transIer due to its high rate oI heat transIer. The bubbles Iorm Iar more rapidly and Irom
more centers oI nucleation and exert an appreciable agitation on the liquid. Here, heat
transIer is a strong Iunction oI ǻT, as demonstrated by the steep increase oI the slope oI
the pool boiling curve.
Regime D-E. Unsteaay-state film boiling
Above a certain ǻT, the heat Ilux begins to decrease with Iurther increase in ǻT as an
unsteady-state Iilm boiling phenomenon develops. What really happens here is that, due
to the extremely high rate oI nucleation, the bubbles cannot escape Irom the hot surIace
and a vapor Iilm blankets the tube surIace. However, this Iilm is blown oII the surIace
and nucleation starts until a vapor Iilm Iorms again. This is a regime oI unstable
perIormance and designers should ensure that no part oI a reboiler lies in this regime.
Operationally, it is very poor due to its reverse-control characteristics, namely, a decrease
in the heat Ilux with increasing ǻT. (In all the other regions, heat Ilux increases with ǻT).
Regime E-F. Steaay-state film boiling
Here, ǻT is so high that a permanent vapor Iilm Iorms on the heat transIer surIace since
no liquid can exist on the tube surIace. Naturally, the heat transIer coeIIicient reduces
considerably due to the Iormation oI the vapor Iilm and so does the heat Ilux. With
Iurther increase in ǻT, the heat Ilux begins to increase again due to the increase in ǻT
since it is the product oI the heat transIer coeIIicient and ǻT.
Due to the low heat transIer coeIIicient, designers do not like a reboiler to operate in this
region. However, it is perIectly stable and, thereIore, acceptable Ior a reboiler to operate in
this regime when it is unavoidable due to process reasons (Ior example, the heating medium
temperature cannot be reduced).
8.2 Parameters Affecting Pool Boiling
Since the nucleate boiling regime is the most relevant and practical to design engineers,
the parameters aIIecting nucleate boiling will be discussed here. These are the eIIects oI
surIace, pressure, boiling range, and tube bundle geometry.
8.2.1 Surface effects
Nucleate boiling is a surIace phenomenon and microscopic pits are essential Ior the Ior-
mation and growth oI bubbles. A certain degree oI superheat is also required Ior bubble
125
growth and the degree oI this superheat is a Iunction oI the cavity radius. The size
distribution oI microscopic pits on the surIace oI the tubes is an important determinant oI
the heat transIer coeIIicient, especially at low temperature diIIerences.
In general, rough surIaces yield a
higher heat transIer coeIIicient and
polished surIaces a lower heat trans-
Ier coeIIicient. ArtiIicially roughened
surIaces, such as high-Ilux tubes,
induce nucleate boiling at very low
ǻT where bare tubes would only
experience natural convection or
incipient boiling, thereby producing a
considerable increase in the boiling
heat transIer coeIIicient. High-Ilux
tubes have a special artiIicially
sintered surIace which incorporates
stable reentrant sites and are so called
as they increase the boiling heat
transIer coeIIicient and, thereby, the
heat Ilux, especially at low ǻT`s (see
Fig. 8.2). Thus, these tubes oIIer a
distinct advantage in services where
it is advantageous to have a low
temperature diIIerence driving Iorce.
Various types oI boiling sites are
shown in Fig. 8.3. A stable reentrant
site is one which does not permit
wetting (which would make the site
useless as a boiling site) but which
allows the vapor bubble produced to
be released and Iresh liquid to enter
the site, thus producing sustained
eIIicient boiling.
The mechanisms oI boiling Irom
a bare surIace and a sintered porous
surIace are shown schematically in
Fig. 8.4. Whereas bubble generation
Irom a bare surIace occurs Irom ran-
dom pits and scratches, numerous
stable bubble nucleation sites are
incorporated in the sintered porous
surIace.
Low-Iin tubes (Fig. 8.5) also
yield a higher heat transIer coeIIi-
cient under such circumstances, but
the augmentation is much lower than
that oIIered by high-Ilux tubes. An
important limitation in the use oI
Fig. 8.2 Variation oI heat Ilux with temperature diIIerence Ior
bare and high-Ilux tubes
Fig. 8.3 Boiling sites (Reprinted Irom the Heat Exchanger
Design Handbook, 2002 with permission oI Begell House,
Inc.)

Fig. 8.4 Mechanisms oI vapor generation Irom bare and por-
ous surIaces (UOP Advanced Heat TransIer Technology with
HIGH FLUX Tubing. ‹ 1998 UOP LLC. All rights reserved.
Used with permission.)

Fig. 8.5 Low-Iin tube
126
high-Ilux and low-Iin tubes is that the service should be clean, as any Iouling will evidently
negate the augmentative eIIect.
8.2.2 Mixture effects
When a pure component boils, the only heat transIer is by phase change (boiling) which
produces a high heat transIer coeIIicient, provided ǻT is suIIiciently high. However,
when a mixture oI diIIerent components boils, there is an additional sensible heating duty
besides the boiling heat duty. Consequently, there is a reduction in the net heat transIer
coeIIicient as the sensible heating has a lower heat transIer coeIIicient, especially in pool
boiling where it will be accomplished by natural convection.
Evidently, the greater the number oI components in the boiling liquid, the greater will be
the reduction in the net heat transIer coeIIicient. A good measure oI the mixture eIIect is the
boiling range, that is, the diIIerence between the dew point and the bubble point. In a pure
component, this is zero since the dew point and the bubble point are the same. The greater
the number oI components and the greater the diIIerence in their volatility, the greater will
be the boiling range and the lower the net boiling heat transIer coeIIicient.
This eIIect is akin to that oI condensation oI a multicomponent mixture, where the vapor
cooling duty reduces the heat transIer coeIIicient quite sharply. The reduction is greater in
multicomponent condensation than in multicomponent boiling because vapor cooling heat
transIer coeIIicients (encountered in condensation) are even lower than liquid heating heat
transIer coeIIicients (encountered in boiling).
8.2.3 Pressure effects
In pool boiling, the boiling heat transIer coeIIicient increases with pressure, all other
things remaining constant. This is because oI the Iollowing eIIects:
1) As pressure increases, liquid surIace tension decreases so that smaller stable
bubbles can be Iormed. This enables more sites to be active and increases the rate
oI generation oI bubbles Irom each site, thereby promoting the heat transIer
coeIIicient.
2) As pressure increases, the latent heat oI vaporization decreases. This means that,
Ior a given heat duty, more vapor is generated, thereby causing an increase in the
turbulence level with a resultant increase in the boiling heat transIer coeIIicient.
As the critical pressure is reached, both surIace tension and latent heat approach zero, so
that the boiling heat transIer coeIIicient becomes extremely high.
8.2.4 Tube bundle geometry effects
Because oI the presence oI surrounding tubes in a tube bundle, there is an increase in the
eIIective velocity oI the two-phase Ilow past a given tube. This results in a decrease in
the liquid Iilm thickness. Consequently, the heat transIer coeIIicient Ior pool boiling in a
tube bundle is greater than that Ior single-tube boiling.
8.3 Maximum Heat Flux
The prediction oI the maximum heat Ilux Ior a boiling application is extremely important,
Ior it represents a limit which cannot be surpassed even by increasing the temperature
level oI the heating medium. (Remember, heat Ilux is the product oI the heat transIer
coeIIicient and ǻT.) This means that an undersized reboiler can be made adequately sur-
127
Iaced by an increase in the heating medium temperature level and, thereby, ǻT only if the
resultant heat flux is less than the maximum heat flux.
The variables controlling the maximum heat Ilux are pressure and tube bundle
geometry. Usually, the heat Ilux increases with increasing reduced pressure (ratio oI
operating pressure to critical pressure) until a value oI about 0.3, aIter which it decreases all
the way to a reduced pressure oI 1.0.
As compared to a single tube, a tube bundle hinders the release oI vapor and,
consequently, there is an appreciable decrease in the maximum heat Ilux. Increasing the tube
pitch Iacilitates vapor release and thereby increases the maximum heat Ilux in a tube bundle.
8.4 Flow Boiling
Many oI the principles that apply to pool boiling apply to Ilow boiling as well. The
principal diIIerence between the two is that, unlike in pool boiling, Ilow is restricted to a
channel in Ilow boiling. As a direct consequence, hydrodynamic eIIects are much more
pronounced in Ilow boiling than in pool boiling.
There are two distinct groups oI Ilow regimes in Ilow boiling:
a) when liquid is at the wall (wet) ² bubble, slug, churn, and annular
b) when vapor is at the wall (dry) ² mist and Iilm
Since the heat transIer coeIIicient is much higher under wet conditions, the designer
always tries to operate in this region.
Under the majority oI operating conditions encountered, both nucleate boiling and
convective boiling occur in the heat transIer process and the contributions by the two
mechanisms are additive. Thus, the net heat transIer coeIIicient can be expressed as:
h
b
÷ h
conv
¹ ah
nb
(8.2)
where
a ÷ nucleate boiling suppression Iactor
h
b
÷ overall boiling heat transIer coeIIicient
h
conv
÷ convective heat transIer coeIIicient
h
nb
÷ nucleate boiling heat transIer coeIIicient
The nucleate boiling suppression Iactor accounts Ior the decrease in the nucleate boiling
rate caused by the two-phase convection mechanism.
The nucleate boiling mechanism tends to predominate in the heat transIer process Ior
narrow boiling-range mixtures and at high operating pressures, whereas the convective
mechanism tends to predominate Ior wide-boiling mixtures and at low pressures.
Since the heat transIer coeIIicient Ior convective boiling varies appreciably with the
boiling regime, calculations usually have to be perIormed stepwise over small increments oI
tube length.
8.5 Distillation Column Reboilers
Distillation column reboilers are oI the Iollowing construction types:
1) Internal
2) Kettle
3) Vertical thermosyphon
4) Horizontal thermosyphon
5) Forced Ilow
128
In the Iirst Iour types, there is no pump and Ilow is by means oI natural density diIIerence
oI the working Iluid between column and reboiler. In Iorced Ilow, however, as implied by
the name itselI, there is a pump Ior Iorcing Ilow through the reboiler.
Let us now consider the salient Ieatures, advantages, and disadvantages oI these
individual types.
8.5.1 Internal reboilers
Principal features
Internal reboilers employ a tube bundle which is inserted directly in the column as shown
in Fig. 8.6. This is the cheapest construction since there is no shell or shellside piping.
However, it is evident that only a very small heat transIer area can be incorporated in an
internal reboiler because both the number oI tubes and the tube length are limited by the
column diameter. The tube length is limited directly and the number oI tubes is also
limited because the maximum diameter oI the opening to accommodate the internal
reboiler is usually one-halI the column diameter.
ReIerring to Fig. 8.6, it will be seen that a part oI the straight length oI tubes is in the
column nozzle. Since the entry oI liquid into this section as well as the release oI vapor Irom
it are hampered, it is advisable to ignore this part oI the straight length when determining the
heat transIer area since it is not very eIIective. Usually U-tubes are used, but iI the heating
medium is dirty, straight tubes and a Iloating-head construction (TEMA type T) should be
employed.
Since internal reboilers also experience pool boiling, like kettles, design procedures are
precisely the same. Here, the space in the column above the reboiler bundle, itselI, acts as
the vapor disengaging space.
Aavantages
1) The advantages Ior this construction are the same as Ior kettles (see section 8.5.2).
2) Additionally, the cost is considerably lower due to the absence oI the shell and
shellside piping. In Iact, this is the cheapest construction.
Disaavantages
1) Again, these are the same as Ior kettles, except Ior the high cost.
2) Additionally, as already explained, the surIace that can be incorporated in a given
distillation column is very limited. Consequently, internal reboilers are rather rare,
the most common applications
being in petrochemical plant units
such as ethylene glycol, benzene,
butadiene, linear alkyl benzene,
etc.
3) For any maintenance work
even on the distillation column,
the reboiler has to be removed.
Let us now consider a case study
which will demonstrate pool
boiling oI a pure hydrocarbon.

Fig. 8.6 Internal reboiler (Courtesy oI HTRI.)
129
CASE STUDY 8.1: LIGHT HYDROCARBON REBOILER (INTERNAL REBOILER)
Consider the service elaborated in Table 8.1a. This is a light hydrocarbon reboiler where
the light hydrocarbon is vaporized outside tubes at 334.2 psia (23.5 kg/cm
2
abs.) and
163ƒF (72.8ƒC) by low pressure steam at 35.55 psia (2.5 kg/cm
2
abs.) and 338ƒF (170ƒC),
having a saturation temperature oI 259.5ƒF (126.4ƒC). Since the heat duty is very small, it
was expected that the required heat transIer area could be easily installed in an internal
reboiler and the unit was thereIore speciIied as such. The bundle was to be 'stabbed¨ in
the column having an internal diameter oI 33.5 in. (850 mm).
Tubes oI 0.78974-in. (20-mm) OD and 0.0787-in. (2-mm) thick carbon steel were to be
used. Since the tubeside stream is clean, U-tubes could be employed.
Thermal design was carried out and the principal construction and perIormance
parameters oI the resultant design are detailed in Table 8.1b. Since a part oI the
straight length oI the tubes would be ineIIective since it would be extending beyond
the wall oI the drum, the straight length oI tubes indicated is the eIIective value and
the actual straight tube length will have to be increased by the length by which the
inner Iace oI the tubesheet extends beyond the wall oI the drum.
It will be seen that the eIIective straight length oI the tubes is only 17.7 in. (450 mm)
and the U-bend diameter will be approximately 5.9 in. (150 mm), the port diameter being
13.27 in. (337 mm). Consequently, even the outermost U-bends will be well clear oI the
column wall since the column has an internal diameter oI 33.5 in. (850 mm).
The desuperheating oI steam on the tubeside hydrocarbon stream was completely in the
wet-wall mode due to the low tube-wall temperature. Hence, the temperature proIiles oI both
the streams are linear so that the MTD is the arithmetic diIIerence oI 259.5ƒF (126.4ƒC) and
163ƒF (72.8ƒC), i.e., 96.5ƒF (53.6ƒC).
It has been stated earlier that boiling is the one phenomenon where the heat transIer
coeIIicient is a Iunction oI the heat Ilux. Strictly speaking, this heat Ilux is the product oI the
heat transIer coeIIicient and ǻT. Now, since Q ÷ UAǻT, (UǻT) is equal to (Q/A), which is
how heat Ilux is popularly perceived as. The important thing to realize here is that, as long as
Table 8.1a: Salient process parameters oI light hydrocarbon reboiler
Shellside Tubeside
1. Fluid circulated Light hydrocarbon L P Steam
2. Flow rate, lb/h (kg/h) 7915 (3590) 765 (347)
3. Temperature in/out, °F (°C) 163 (72.8) 338 (170)/259.5 (126.4)
4. Operating pressure, psia (kg/cm
2
abs.) 334.2 (23.5) 35.55 (2.5)
5. Permitted pressure drop, psi (kg/cm
2
) negl. 1.4 (0.1)
6. Fouling resistance, h It
2
°F/Btu (h m
2
°C /kcal) 0.001 (0.0002) 0.0005 (0.0001)
7. Heat duty, Btu/h (kcal/h) 706,400 (190,000)
8. Design pressure, psig (kg/cm
2
g) 427 (30.0) 100 (7.0)
9. Design temperature, °F (°C) 212 (100) 428 (220)
10. Material oI construction Carbon steel
11. Nominal line size, in. (mm) Stab-in 2 (50)/1.5 (40)
130
the heat transIer area provided is exactly equal to the required heat transIer area, the heat
Ilux is the same either way. However, when the heat transIer area provided is in excess oI
the required value, there is a diIIerence between the two and that is the extent oI
oversurIacing.
This brings us to an interesting point. Some oI the older heat exchanger soItware used to
determine the boiling heat transIer coeIIicient as a Iunction oI the heat Ilux which was
considered as (Q/A). ThereIore, iI the heat transIer area speciIied was much in excess oI that
required, the boiling heat transIer coeIIicient would be unduly low and, to that extent, the
design would be conservative. In such situations, the prudent thing to do would be to let the
program determine the heat duty (simulation mode) and then to change the heat transIer area
as required to arrive at an optimum design.
In the present instance, the boiling heat transIer coeIIicient was evaluated as 2547 Btu/h
It
2
ƒF (12,435 kcal/h m
2
ƒC) and the overdesign as 12.4°. The other perIormance parameters
are reported in Table 8.1b.
Effect of ǻT
In order to demonstrate the eIIect oI ǻT on the boiling heat transIer coeIIicient, the steam
pressure was varied progressively to 21.33 psia (1.5 ata), 28.4 psia (2.0 ata), and 42.7 psia
(3.0 ata). The resultant values oI boiling heat transIer coeIIicient, along with that oI the
base value at 35.55 psia (2.5 ata) are reported in 8.1c. It will be seen that the boiling heat
transIer coeIIicient varies appreciably with ǻT.
8.5.2 Kettle reboilers
Principal features
A kettle reboiler is so called because oI its appearance due to the enlarged shell (Fig. 8.7).
This is provided Ior the disengagement oI vapor. Consequently, kettles can Iully vaporize
the column bottoms, thereby representing an additional theoretical stage Ior the distilla-
Table 8.1b: Principal construction and perIormance parameters oI light hydrocarbon reboiler
1. Type oI reboiler Internal, U-tube
2. Port diameter, in. (mm) 13.27 (337)
3. No. oI tubes × no. oI tube passes 40 U`s × 2
4. Tube OD × thickness, in. (mm) 0.7874 (20) × 0.0787 (2)
5. EIIective straight length, in. (mm) 17.7 (450)
6. Heat transIer area, It
2
(m
2
) 27.7 (2.57)
Shellside Nil
7. Pressure drop, kg/cm
2

Tubeside Negligible
Shellside 2547 (12,435)
Tubeside 1356 (6618)

8. Heat transIer coeIIicient, Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 317.5 (1550)
9. MTD, °F (°C) 96.5 (53.6)
10. Overdesign, ° 12.4
131
tion process. The liquid level is maintained a little above the top oI the tube bundle by
means oI a vertical weir located beyond the bundle. Sometimes, a more sophisticated
level control arrangement is incorporated Ior better controlhere, the liquid level may
even be lowered, should entrainment become excessive.
The enlarged shell is called the kettle and the reduced shell at the tubesheet is called the
port. A kettle reboiler may be oI Iixed-tubesheet, U-tube, or Iloating-head construction, as
discussed in Chapter 2. Essentially, iI the heating medium (which is inside the tubes) is clean
(e.g., steam), a U-tube kettle is the most preIerred construction due to the lower cost and its
capability to permit diIIerential expansion between the shell and the tube bundle. However,
iI the heating medium is dirty, U-tubes cannot be used and the Iollowing selection is
indicated:
1) use Iixed tubesheet construction iI the boiling Iluid is clean
2) use Iloating-head construction iI the boiling Iluid is dirty
It may be noted that, since a dished-end construction has necessarily to be used Ior a
kettle reboiler, TEMA S style cannot be used and a TEMA T style becomes necessary
(TEMA AKT), as mentioned in Section 2.3.3. Evidently, since the entire Iloating-head
assembly, including the Iloating-head Ilange, will have to be removed Ior maintenance
due to the lack oI access Irom outside the reboiler, the port diameter will have to be
signiIicantly higher than the outer tube limit.
For a U-tube construction, the number oI tube passes will evidently be two (or a higher
even number Ior clean heating mediums). It can be any even number Ior a Iloating-head
construction and can additionally be one Ior a Iixed-tubesheet construction, iI required.
The boiling liquid enters at the
bottom oI the shell and the vapor leaves
Irom the top oI the shell. For wide-
boiling mixtures, it is recommended that
the liquid enter at multiple locations
along the length oI the shell Ior better
distribution. For large tube lengths (say,
above 13.12 It or 4 m), two outlet nozzles
are recommended to reduce the hori-
zontal vapor velocity in the vapor dis-
engagement space, so as to minimize
entrainment. II there is a liquid draw-oII
(distillation column bottom product), its
nozzle will evidently be beyond the weir
plate (see Fig. 8.7). In extreme situations,
demisters are installed in the outlet noz-
Table 8.1c: EIIect oI variation oI ǻT on boiling heat transIer coeIIicient
Steam Pressure, psia (kg/cm
2
abs.)
21.33 (1.5) 28.4 (2.0) 33.55 (2.5) 42.7 (3.0)
Boiling heat transIer
coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)

2248 (10,977)

2623 (12,806)

2910 (14,208)

3145 (15,356)
MTD, ƒF (ƒC) 67.7 (37.6) 83.7 (46.5) 96.5 (53.6) 107.5 (59.7)

Fig. 8.7 Kettle reboiler (Courtesy oI HTRI.)
132
zle(s) to virtually eliminate mist entrainment. One possible example oI this is a chiller where
a liquid reIrigerant is vaporized and has to be returned to the compressor.
Aavantages
1) The use oI a kettle reboiler incorporates an additional stage Ior distillation. Where the
number oI distillation stages is not very high, this represents a distinct advantage.
2) Very high heat Iluxes can be achieved.
3) Due to pool boiling, a kettle reboiler is relatively insensitive to hydrodynamics.
Consequently, operation is much better than thermosyphons Ior the Iollowing:
a) when there are Iluctuations in operating conditions
b) at very low pressure
c) at high pressure (near critical)
4) Kettle reboilers operate eIIiciently even at low ǻT.
5) Kettle reboilers can Iully vaporize the column bottoms stream.
6) Since there is no external two-phase piping Irom the reboiler to the column, the
complicated process oI determining the outlet line pressure drop is eliminated. The
advantage is that this aIIords reliability. The sizing oI two-phase outlet piping and the
determination oI pressure drop in it is Irequently an area oI major uncertainty in
thermosyphon systems.
7) At very high ǻT, high heat Iluxes can be sustained by locating tubes at a higher pitch,
thereby providing vapor release lanes.
Disaavantages
1) Kettle reboilers are relatively costly due to the enlarged shell. When a Iloating-head
construction (TEMA AKT) is required, because oI a dirty heating medium, the cost is
even higher.
2) Since the accurate determination oI the MTD is diIIicult and uncertain, a conser-
vative approach is adopted and the MTD is based upon the shellside outlet tem-
perature only. That is, the inlet temperature oI the boiling stream is also considered to
be its outlet temperature. The lower MTD results in the incorporation oI additional
heat transIer area which Iurther increases the cost. Evidently, the higher the boiling
temperature range, the greater will be the penalty. This is demonstrated very clearly
in Case Study 8.2.
3) Since there is very little turbulence and heavy material residue accumulates con-
tinually on the tubes, kettle reboilers have the tendency to Ioul readily, even with a
Iairly high draw-oII rate. The high residence time may even result in degradation oI
the boiling liquid. Hence, kettle reboilers are contraindicated Ior Iouling services.
CASE STUDY 8.2: STRIPPER REBOILER (KETTLE REBOILER)
Let us consider the design oI a stripper reboiler, the principal process parameters oI
which are indicated in Table 8.2a. It will be noticed that the boiling Iluid is a wide-
boiling mixture because it has a boiling range (bubble point minus dew point) oI 161.1ƒF
(89.5ƒC). Consequently, the boiling heat transIer coeIIicient will not be particularly high.
Only 48° oI the liquid entering the reboiler is vaporized, and the remaining 52° is
drawn oII as a product. A 187 It
3
(5.3 m
3
) surge capacity is to be incorporated in the
kettle.
133
Carbon steel tubes /-in. (19.05-mm) OD and 14 BWG (2.11-mm) are to be used. A
minimum overdesign oI 10° is to be provided. A TEMA BKU construction is recom-
mended. As the heating medium is clean, the use oI U-tubes is perIectly acceptable.
Since the boiling stream in this application is clean, pool boiling is acceptable. However,
the boiling range is high and contraindicates the use oI pool boiling due to both the lower
boiling heat transIer coeIIicient and the loss in MTD; this will result in a larger and costlier
unit. However, this reboiler was speciIied as a kettle reboiler and a design will be produced
as such.
Thermal design was perIormed and the principal construction and perIormance
parameters are detailed in Table 8.2b. It will be noticed that the boiling heat transIer
coeIIicient is only 274 Btu/h It
2
ƒF (1338 kcal/h m
2
ƒC), principally due to the large boiling
range. As has been pointed out earlier, the MTD was determined considering the inlet
shellside temperature also as 350.6ƒF (177ƒC). Thus, the heat transIer area is rather large at
3067 It
2
(285 m
2
), due to the low
overall heat transIer coeIIicient and the
low MTD.
In order to demonstrate the eIIect
oI the boiling range, the same was
artiIicially decreased in steps Irom the
actual value oI 161.1ƒF (89.5ƒC) to
1.8ƒF (1.0ƒC). The variation oI the
boiling heat transIer coeIIicient is
shown in Table 8.2c. For a better ap-
preciation oI this variation, Fig. 8.8
shows the variation oI the boiling heat
transIer coeIIicient with boiling tem-
Table 8.2a: Salient process parameters oI stripper reboiler (kettle reboiler)
Shellside Tubeside
1. Fluid circulated Hydrocarbon mixture M.P. Steam
2. Flow rate, lb/h (kg/h) 159,800 (72,500) 19,580 (8880)
3. Temperature in/out, °F (°C)
269.6 (132)/350.6 (177)
bubble point ÷ 269.6 (132)
386.6 (197)
4. Boiling range, °F (°C) 161.1 (89.5) -
5. Operating pressure, psia (kg/cm
2
abs.) 341.3 (24.0) 213.3 (15.0)
6. Permitted pressure drop, psi (kg/cm2) negl. 1.4 (0.1)
7. Fouling resistance, h It2 °F/Btu
(h m
2
°C /kcal)
0.00098 (0.0002) 0.00049 (0.0001)
8. Heat duty, MM Btu/h (MM kcal/h) 16.11 (4.06)
9. Design pressure, psig (kg/cm
2
g) 398 (28.0) 284 (20.0)/Full vacuum
10. Design temperature, °F (°C) 419 (215) 464 (240)
11. Material oI construction carbon steel
12. Line size in/out, in. (mm) 12 (300)/14 (350) 8 (200)/3 (80)
Fig. 8.8 Variation oI boiling heat transIer coeIIicient with
temperature range in stripper reboiler
134
perature range graphically. It will be seen
that, with decreasing boiling temperature
range, the heat transIer coeIIicient re-
duces gradually at Iirst, but much more
sharply later.
8.5.3 Horizontal
thermosyphon reboilers
Principal features
A horizontal thermosyphon reboiler has
vaporization on the shellside while the
heating medium Ilows inside the tubes
(Fig. 8.9). Since there are no pumps in
thermosyphon reboiler circuits, the pres-
sure drop has to be restricted to a bare
minimum. As already discussed in Sec-
tion 3.4.1, a split-Ilow shell (TEMA G
shell) conIiguration or a double split-
Ilow shell (TEMA H shell), both shown
in Fig. 2.6, are usually employed Ior
horizontal thermosyphon reboilers. The
advantage with TEMA G and H shells is
that the pressure drop is reduced drasti-
Table 8.2b: Principal construction and perIormance parameters
oI stripper reboiler (kettle reboiler)
1. Type oI reboiler BKU
2. Port/shell ID, in. (mm) 37.4 (940)/53.15 (1340)
3. No. oI tubes × no. oI tube passes 448 U`s × 2
4. Tube OD × BWG × st. length, in. (mm) /

(19.05) × 14 (2.11) × 197 (5000)
5. Tube pitch, in. (mm) 1.0 (25.4), square
Shellside 12 (300)/2 × 10 (250) 6. Nominal nozzle size
in/out, in. (mm)
Tubeside 8 (200)/3 (80)
7. EIIective heat transIer area, It
2
(m
2
) 3067 (285.0)
Shellside Nil 8. Pressure drop,
psi (kg/cm
2
)
Tubeside Negligible
Shellside (boiling hydrocarbon) 274 (1338)
Tubeside (condensing steam) 2099 (10,248)
9. Heat transIer
coeIIicient, Btu/hIt
2
°F
(kcal/h m
2
°C)
Overall 166.6 (813.5)
10. MTD, °F (°C) 20.0
11. Overdesign, ° 14.2
Table 8.2c: Variation oI boiling heat transIer
coeIIicient with boiling range (kettle reboiler)
Boiling range,
ƒF (ƒC)
Boiling heat transfer coefficient,
Btu/h ft
2
ƒF (kcal/h m
2
ƒC)
161.1 (89.5) 274 (1338)
144 (80) 284 (1387)
126 (70) 296.4 (1447)
108 (60) 311.1 (1519)
90 (50) 328.9 (1606)
72 (40) 351.7 (1717)
54 (30) 381.4 (1862)
45 (25) 400.2 (1954)
36 (20) 423.1 (2066)
27 (15) 451.6 (2205)
18 (10) 489.1 (2388)
9 (5) 543 (2651)
3.6 (2) 592.5 (2893)
1.8 (1) 654.2 (3194)
135
cally as there are no cross baIIles. By
using a central Iull support plate in a
G shell and limiting the tube length to
9.84 It (3 m), the maximum unsup-
ported tube span is restricted to within
the TEMA limit. Similarly, by using
two Iull support plates in an H shell
and restricting the tube length to
19.68 It (6 m), no cross baIIles are
required to comply with the TEMA
maximum unsupported tube span
limit.
The longitudinal baIIle also serves
to minimize phase separation and pro-
mote mixing oI the light and heavy
components as vaporization proceeds.
Another advantage oI the split Ilow or the double split Ilow shell is that hydrodynamic
problems associated with baIIled shellside Ilow such as phase separation (with vertically cut
baIIles) and pulsating Ilow (with horizontally cut baIIles) are eliminated. In Iact, this is a
general limitation Ior vaporization on the shellside. Whereas a vertical cut is recommended
Ior shellside condensing because it enables separation oI the condensate Irom the
uncondensed vapor, a vertical cut is highly detrimental Ior shellside vaporization. Since
liquid has a considerably higher density than vapor, it will occupy only a small cross-
sectional area at the bottom oI the shell and, thereby, not contact the majority oI the heat
transIer surIace. Consequently, even iI the required heat transIer surIace is provided, a large
part oI it will not be eIIective.
Besides G and H shells, E and J shells are occasionally used in horizontal thermosyphon
reboilers when a higher velocity is desired. Such a requirement may arise when vaporizing a
wide-boiling mixture. For very low pressure applications, it may become necessary to
employ pure crossIlow in a TEMA X shell.
As elaborated in Section 2.3.2, U-tubes may be employed Ior clean heating mediums,
e.g., TEMA BHU. However, iI the heating medium is dirty, straight tubes are required to be
employed. Here, iI the boiling Iluid is clean, a Iixed-tubesheet (e.g., TEMA BHM)
construction may be used but iI it is dirty, a Iloating-head construction (e.g., TEMA AHS)
will have to be employed.
Circulation in horizontal thermosyphon reboilers is provided by the diIIerence in
elevation between the column and the bottom oI the shell.
Vaporization is usually limited to 2030° Ior best operationand never more than
50°in order to avoid the penalties oI low heat transIer coeIIicient, excessive Iouling due
to high skin temperature, and reduction in MTD.
Aavantages
1) High circulation rates impart a high convective heat transIer component and, thereby,
yield a high boiling heat transIer coeIIicient.
2) A horizontal thermosyphon reboiler oIIers a distinct advantage in MTD, as Iollows:
a) As compared to a kettle reboiler, the MTD is higher because the penalty Ior
MTD in kettles described in Section 8.5.2 is not applicable here since a regular
convective Ilow pattern is established. Besides, since the liquid is only partially
Fig. 8.9 Horizontal thermosyphon reboiler (Courtesy oI
HTRI.)
136
vaporized (typically 2030°), the outlet temperature oI the vaporizing stream is
even lower, thereby leading to a Iurther increase in the MTD.
b) As compared to vertical thermosyphon reboilers, the increase in boiling point is
less due to the much lower diIIerence in elevation between the column and the
reboiler. This can become signiIicant in low-ΔT applications, especially at low
operating pressures.
3) The above is particularly true Ior wide-boiling mixtures having high liquid viscosity,
where the high degree oI mixing produced by the presence oI baIIles in an E or J
shell yields a higher heat transIer coeIIicient.
4) The Iouling potential is reduced by virtue oI the higher velocity and the lower exit
vapor Iraction. This minimizes the accumulation oI heavy components in the reboiler.
Thus, horizontal thermosyphon reboilers can handle dirty streams much better than
kettle reboilers.
5) Horizontal thermosyphon reboilers are less sensitive than vertical thermosyphon
reboilers to Iluctuations in operating conditions. Kettles, oI course, are the least
sensitive.
Disaavantages
1) Whatever Iouling occurs is on the shellside, which is more diIIicult to clean than
Iouling inside tubes. Consequently, vertical thermosyphon reboilers are superior to
horizontal thermosyphon reboilers Ior vaporizing dirty streams.
2) Besides the usual Iloating-head construction, multiple nozzles and maniIold piping
increase the cost. The cost oI a horizontal thermosyphon reboiler is almost as much
Table 8.3a: Salient process parameters oI distillation column reboiler (horizontal thermosyphon)
Shellside Tubeside
1. Fluid circulated Column bottoms LP Steam
2. Flow rate, lb/h (kg/h) 110,200 (50,000) 12,974 (5885)
3. Fraction vaporized 0.12 -
4. Temperature in/out, °F (°C) 278.4 (136.9) 298.8 (148.8)
5. Operating pressure, psia (kg/cm
2
abs.) 100 (7.0) 67.5 (4.75)
6. Permitted pressure drop, psi (kg/cm
2
)
As per
thermosyphon circuit
1.0 (0.07)
7. Fouling resistance, h It
2
°F/Btu (h m
2
°C /kcal) 0.001953 (0.0004) 0.000488 (0.0001)
8. Heat duty, MM Btu/h (MM kcal/h) 12.3 (3.1)
9. Critical pressure, psia (kg/cm
2
abs.)
10. Critical temperature, °F (°C)
Standard as per
water/steam
Standard as per
water/steam
11. Design pressure, psig (kg/cm
2
g) 135 (9.5) 92 (6.5)
12. Design temperature, °F (°C) 338 (170) 392 (200)
13. Material oI construction carbon steel
14. Line size in/out, in. (mm)
As per
thermosyphon circuit
6 (150)/2 (50)
137
as that oI a kettle reboiler iI we compare U-tube to U-tube or Iloating-head to
Iloating-head.
3) Two-phase Ilow in the exit piping requires careIul analysis and design. II the exit
piping is too small, choke Ilow may result, thereby leading to instability. On the other
hand, iI the outlet piping is too large, phase separation may result. In a wide-boiling
mixture, this will result in the accumulation oI heavy components in the reboiler with
its associated penalties oI aggravated Iouling and increase in the boiling temperature,
thereby decreasing the MTD. Consequently, the outlet piping has to be designed
careIully to Iully exploit the advantages oIIered by this construction.
CASE STUDY 8.3: DISTILLATION COLUMN REBOILER
(HORIZONTAL THERMOSYPHON)
A horizontal thermosyphon reboiler had to be designed Ior the service detailed in Table
8.3a. The boiling Iluid is a dilute solution, so that the physical properties were virtually
that oI water. Carbon steel tubes 0.7874-in. (20-mm) OD × 0.09-in. (2.3-mm) thick ×
16.4-It (5000-mm) long were to be used. An oversurIacing oI 10° was to be
incorporated.
Thermal design was carried out and the principal construction and perIormance
parameters are reported in Table 8.3b. Based upon indicative piping isometrics, the inlet and
outlet piping lengths were considered as 115 It (35 m) and 154 It (47 m), respectively. While
Table 8.3b: Principal construction and perIormance parameters oI
distillation column reboiler (horizontal thermosyphon)
1. Type oI reboiler Horizontal thermosyphon
2. Shell diameter, in. (mm) 34.65 (880)
3. No. oI tubes × no. oI tube passes 774 × 2
4. Tube OD × BWG × length, in. (mm) 0.7874 (20) × 0.09 (2.3) × 197 (5000)
5. EIIective heat transIer area, It
2
(m
2
) 2712 (252)
Shellside 0.82 (0.058)
6. Pressure drop, psi (kg/cm
2
)
Tubeside 0.54 (0.038)
Shellside 926 (4522)
Tubeside 3812 (18,611)

7. Heat transIer coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
Overall 232.5 (1135)
Shellside 25.09
Tubeside 7.92
Fouling 60.12

8. Heat transIer resistance, °
Metal 6.87
9. Maximum heat Ilux, Btu/h It
2
(kcal/h m
2
) ~88,560 (240,000)
10. Design heat Ilux, Btu/h It
2
(kcal/h m
2
) 4466 (12,102) to 4717 (12,784)
11. MTD, °F (°C) 21.4 (11.9)
12. Overdesign, ° 10.4
138
perIorming the thermal design, the main lines were sized as 8 in. (200 mm) Ior the inlet and
14 in. (350 mm) Ior the outlet, and the nozzle piping was sized as 4 in. (100 mm) Ior the
inlet and 10 in. (250 mm) Ior the outlet. The required static head was Iound to be 7.55 It (2.3
m) Irom the Iollowing pressure drop in the circuit, in psi (kg/cm
2
):


Pressure drop in reboiler ÷ 0.825 (0.058)
Pressure drop in inlet piping ÷ 1.01 (0.071)
Pressure drop in outlet piping ÷ 1.195 (0.084)
Total pressure drop ÷ 3.03 (0.213)
It will be seen Irom Table 8.3b that the
values oI the condensing and boiling heat
transIer coeIIicients and, thereby, the over-
all heat transIer coeIIicient are very high. In
Iact, the major resistance to heat transIer is
Iouling, which accounts Ior 60.12° oI the
total resistance. Even the tube metal resis-
tance is 6.87° oI the total, which is about
as high as it ever gets.
The critical heat Ilux Ior the present
service is oI the order oI 88,560 Btu/h It
2

(240,000 kcal/h m
2
) in the various incre-
ments, whereas the calculated heat Ilux varies
Irom 4466 to 4717 Btu/h It
2
(12,102 to
12,784 kcal/h m
2
).
8.5.4 Vertical thermosyphon reboilers
BeIore commencing a discussion oI vertical
thermosyphon reboilers themselves (Fig.
8.10), it may be pointed out that several
column conIigurations may be employed to
control the column-reboiler operation better
as well as to eIIect cost economy, both Iixed
and operating. This is really a matter oI
process design since it addresses nuances
oI column operation. The reader is reIerred
to some inIormative papers in this regard
|13|.
In the present context, it will suIIice to
say that there are basically two column-re-
boiler conIigurationsnamely, once-through
and recirculatingas shown in Fig. 8.11.
Although these have been shown Ior vertical
thermosyphon reboilers, they are applicable
to horizontal thermosyphon reboilers, as well.
The key diIIerence between the two con-
Iigurations is that the vapor generated in the
reboiler is not allowed to mix with the
column bottom (product) in the once-through

Fig. 8.10 Vertical thermosyphon reboiler (Courtesy
oI HTRI.)
Fig. 8.11 Once-through and recirculation reboilers
139
style, whereas it is allowed to do so in the recirculating style.
However, it is evidently very important that whatever style has been considered Ior the
process design should be speciIied clearly by the process designer so that the reboiler
thermal design is perIormed accordingly. In Iact, the process designer should interact with
the heat exchanger designer so that the key elements oI process design and reboiler design
may be considered in arriving at the most optimum conIiguration in a given situation.
Going one step Iurther, the reboiler designer must not design the reboiler in isolation as
it is a part oI the column-reboiler-column circuit, but must design the reboiler as a part oI the
whole system. Since piping plays a very important role in reboiler operation, especially the
outlet piping (reboiler to column), the Iinal design must be veriIied with the actual piping
isometrics Iinally developed.
Principal features
Vertical thermosyphon reboilers (Fig. 8.10) are invariably oI Iixed-tubesheet and single
shell pass (TEMA E) construction, having boiling on the tubeside. Boiling inside vertical
tubes is a Iavorable proposition, unlike boiling on the outside oI vertical tubes, and is
very well researched. Boiling outside vertical tubes is complicated by the necessary
presence oI baIIles and is, hence, used only in special circumstances, as explained later.
Since boiling is inside tubes, it is essential to have a single tube pass Ior a thermosyphon
arrangement because the Ilow oI a two-phase mixture against gravity is Iraught with
jeopardy. It has already been explained in Sections 3.3.1 and 7.4.6 that a single-pass
arrangement (see Fig. 7.8) is a very undesirable Ieature Ior Iloating-head construction.
Consequently, Iixed-tubesheet construction is invariably employed Ior vertical
thermosyphon reboilers. However, this imposes a limitation on the heating medium: either it
should be clean, or iI it is dirty, the Iouling caused by it should be amenable to chemical
cleaning. As long as the heating medium is steam or hot oil, there is no problem. However, iI
the heating medium is dirty, the Iollowing options present themselves:
1) implement a chemical cleaning program
2) use a Iloating-head construction
3) have boiling outside tubes
Flow Orientation
Countercurrent Ilow (boiling Iluid Ilowing up, heating medium Ilowing down) is
employed Ior condensing heating mediums as well as sensible heating mediums with
wide boiling-range mixtures.
For sensible heating mediums and narrow boiling-range mixtures, it may be a good
practice to use co-current Ilow (both streams Ilowing up). This provides a higher ǻT and
more nucleate boiling at the bottom oI the tubes, thereby improving circulation. This is not
recommended Ior wide boiling-range mixtures since the loss in MTD will usually be greater
than the gain in boiling heat transIer coeIIicient.
Flow Regimes
Various Ilow regimes occur inside the tubes, depending upon the extent oI vaporization
as well as various physical properties, as Iollows:
• Non-boiling zone, subcooled boiling, bubble and slug Ilow, churn Ilow, annular
Ilow, mist Ilow and Iilm boiling.
• The two-phase Ilow leaving a vertical thermosyphon reboiler is usually in the an-
140
nular or slug regime and occasionally in the bubble regime when the exit vapor
Iraction is low.
• Mist Ilow occurs at very high weight Iraction vaporized and is very ineIIicient
Ior heat transIer. A rule oI thumb Ior avoiding it is to limit the weight Iraction
vaporized to 0.5.
• A combination oI nucleate boiling and Iilm boiling should also be avoided. This
is discussed in detail later on.
Tube Diameter
Tube diameter assumes greater signiIicance Ior vertical thermosyphon reboilers than Ior
sensible heating/cooling and condensing services. Generally, /-in. or 20-mm OD tubes
are used. However, when the Iraction vaporized tends to be high (~ 35°), the use oI
bigger diameter tubes (1in., 1
1
/
4
in., and even 1
1
/
2
in.) will help.
When the Iraction vaporized tends to be low (· 10°), the use oI smaller diameter tubes
will be oI advantage
3
/
4
in. or 20 mm is usually the smallest tube OD, based upon cleaning
considerations. However, iI the tubeside Iluid is very clean, the use oI 5/8-in. or 16-mm OD
tubes may be considered.
Tube Length
Tube length may be varied Irom 3.28 It (1000 mm) to 19.68 It (6000 mm), depending
upon the heat transIer area required and the tubeside Ilow area desired. Since these are
Iixed tubesheet exchangers, it is not necessary to use standard tube lengths.
However, when using a large tube length, it should be ensured that suIIicient headroom
exists at the bottom oI the reboiler Ior routine maintenance. In some extreme cases, the
column skirt height may have to be increased on this account.
Elevation
In a vertical thermosyphon reboiler, the column bottoms stream enters at the bottom,
rises, and partially vaporizes as it Ilows up the tubes, and the two-phase mixture
discharges into the column. The driving Iorce is the density diIIerence between the liquid
in the column and the two-phase Iluid in the tubes. The diIIerence in elevation between
the liquid level in the column and the lower tubesheet represents the driving head. The
liquid level in the column is usually maintained at the level oI the top tubesheet since this
normally represents the optimum balance between driving Iorce and circuit resistance. In
the case oI reboilers operating at vacuum, however, it is common to elevate the reboiler
so that the driving head is 6080° oI the tube length.
In order to understand this, consider the variation oI pressure drop with mass velocity
and two-phase density. Pressure drop varies with the square oI mass velocity and inversely
with density. Now, Ior a given heat duty, there can be either a higher circulation rate and,
consequently, a lower Iraction vaporized, or a lower circulation rate and a higher Iraction
vaporized. The lower Iraction vaporized will mean a higher mixture density and vice versa.
Thus, we have two cases:
1) a higher circulation rate and a lower mixture density
2) a lower circulation rate and a higher mixture density
As per the variation oI pressure drop with mass velocity and mixture density already
stated, the higher circulation rate case will result in a higher pressure drop, and the lower
circulation ratey case will result in a lower pressure drop.
141
However, in an operating thermosyphon reboiler, there is a natural equilibrium
condition when the driving head is balanced by the pressure drop in the column-reboiler-
column circuit. Consequently, when the driving head is higher, there will be a higher
circulation rate, and when the driving head is lower, there will be a lower circulation rate.
Due to the static head oI the liquid in the column, the liquid entering a vertical
thermosyphon reboiler has a higher boiling point than that in the distillation column.
Consequently, there will be a region oI subcooled liquid heating beIore the liquid attains its
boiling point. Due to the single tube pass, the liquid velocity will be rather low and,
consequently, so will be the heat transIer coeIIicient. Now, when the operating pressure is
moderate, the eIIect oI this phenomenon is only marginal because the static head represents
only a small Iraction oI the operating pressure. At vacuum, however, the situation is much
worse, and there will be a substantial subcooled liquid heating zone. In order to minimize
this penalty, the reboiler is elevated so that the driving head is reduced. The lower driving
head will result in a lower elevation in boiling point and, thereby, a smaller subcooled liquid
heating zone.
Piping
The sizing oI outlet piping or reboiler-to-column piping is a very important aspect oI
vertical thermosyphon reboiler design. In Iact, unduly small outlet piping is the biggest
single cause oI maloperation oI vertical thermosyphon reboilers.
Depending upon the diameter oI the outlet pipe and the diIIerence in elevation between
the upper reboiler channel and the return nozzle to the column, a closed-coupled or a
conical/mitered arrangement is employed Ior the outlet piping. The Iormer is preIerred
because it minimizes the pressure drop in the outlet piping. When the elevation diIIerence
between the upper tubesheet and the vapor return line to the column is small, use a close-
coupled arrangement. When it is large, use a conical channel with a suitable bend.
In either case, since the pressure drop in the outlet piping is critical, the length oI the
outlet piping and the number oI bends shoula always be minimi:ea. Further, a valve should
never be placed in the outlet piping. However, a valve placed in the inlet liquid line is useIul
to stabilize Iluctuations.
It is generally recommended that the ratio oI the outlet piping Ilow area to the total
tubeside Ilow area should be 1.0 Ior high heat Iluxes. This can be reduced Ior lower heat
Iluxes but should be at least 0.4. Evidently, the lower the operating pressure and the lower
the vapor molecular weight, the larger will be the outlet piping diameter.
The reboiler design may be perIormed on the basis oI an assumed number oI bends in
the outlet piping, but once the piping isometrics are Iinalized, the reboiler design must be
veriIied with the actual piping conIiguration, especially Ior low-pressure operation (say, 43
psig or 3 kg/cm
2
gauge and less). There are several instances when a vertical thermosyphon
reboiler has not perIormed satisIactorily due to a larger number oI bends in the outlet piping
than catered to in its design.
Liquia circulation/vapori:ation consiaerations
It may not always be possible to design a vertical thermosyphon reboiler to have
precisely the same circulation and the Iraction vaporized speciIied in the process data
sheet. Should the variation be appreciable (say, 25° or more), the matter should be
discussed and resolved with the process licensor.
Tubeside pressure drop is always higher Ior a higher circulation and lower weight
Iraction vaporized. The Iraction vaporized should not be too high as it:
142
a) causes excessive Iouling
b) reduces the tubeside heat transIer coeIIicient
c) reduces the MTD due to the higher outlet temperature
The maximum weight Iraction vaporized recommended as a rule oI thumb is 3035°.
However, in the case oI vacuum operation, the same may be relaxed to 50°, as
otherwise, the subcooled liquid heating zone may impose a severe penalty. The minimum
weight Iraction vaporized generally recommended is 10° because lower vaporization
may not produce good pumping or 'thermosyphoning¨ action.
Maximum allowable aesign heat flux
The maximum allowable design heat Ilux is a very strong Iunction oI the reduced
pressure, p/p
c
Thus, Ior a light hydrocarbon reboiler, the maximum allowable design heat
Ilus is 12,425 Btu/h It
2
(33,672 kcal/h m
2
) at a reduced pressure oI 0.3 but Ialls abruptly
to 4484 Btu/h It
2
(12,152 kcal/h m
2
) as the reduced pressure increases to 0.6. With Iurther
increase in the reduced pressure to 0.8 and 0.9, the maximum allowable design heat Ilux
plummets to 1774 Btu/h It
2
(4808 kcal/h m
2
) and 783 Btu/h It
2
(2123 kcal/h m
2
),
respectively.
Aavantages
1) Due to the Iixed-tubesheet construction, the cost is the lowest Ior this type oI
reboiler.
2) Fouling is less pronounced due to the high circulation and, thereby, high shear rates.
Besides, whatever Iouling occurs is easier to clean since it is on the tubeside.
However, due to the vertical disposition, cleaning is not as convenient as Ior
horizontal tubes unless the reboiler is dismantled and placed horizontally.
3) The MTD is the highest, due to the pure countercurrent situation, and straightIorward
to determine.
Disaavantages
1) The maximum heat Ilux that can be supported is lower than Ior shellside boiling, due
to the reduced boiling-side Ilow area. This is particularly relevant Ior operation near
the critical pressure.
2) The operation oI these reboilers is rather sensitive to changes in operating conditions.
Consequently, unless a sophisticated soItware package is available Ior design, these
reboilers are not recommended, especially at low pressure (subatmospheric) and
pressure near critical.
3) Two-phase Ilow in the exit piping requires careIul analysis and design. II the exit
piping is too small, choke Ilow may result, thereby leading to instability. On the other
hand, iI the outlet piping is too large, phase separation may result. In a wide-boiling
mixture, this will result in the accumulation oI heavy components in the reboiler with
its associated penalties oI aggravated Iouling and increase in the boiling temperature,
thereby decreasing the MTD. Consequently, the outlet piping has to be designed
careIully.
4) For large-diameter reboilers, uniIorm distribution oI the boiling Iluid into the various
tubes may be diIIicult to achieve unless special measures are adopted.
143
5) The heat transIer area is likely to be much higher Ior low-pressure services due to the
elevation oI the boiling point and consequent subcooled liquid heating, as explained
earlier.
Special aesign consiaerations
Very wide boiling range
As already mentioned in section 8.4 above, the convective component tends to predomi-
nate in such applications. In Iact, nucleation may be largely or Iully suppressed, thereby
resulting in a poor heat transIer coeIIicient due the low mass velocity caused by the sin-
gle tube pass. Should the liquid viscosity be relatively high, matters will be even worse.
The use oI an increased static head by lowering the reboiler may help in such situations
by improving the circulation and, thereby, the heat transIer coeIIicient. Nevertheless,
vertical thermosyphon reboilers may not be economically attractive Ior such applications.
As already explained in Section 8.5.2, kettle reboilers are not very good either Ior such
conditions. Horizontal thermosyphon reboilers can handle such services much better.
Operation at low pressure
Once again, the boiling heat transIer coeIIicient will tend to be low Ior vertical
thermosyphon reboilers due to several Iactors:
1) Suppression oI nucleate boiling. The lower the operating pressure, the larger is
the cavity size required Ior eIIective boiling. This shortcoming can be overcome
by the use oI high-Ilux tubes. However, this is proprietary technology with its
associated disadvantages.
2) Subcooled liquid heating. This was explained earlier in the section entitled
'Elevation.¨
3) Low liquid circulation rate. This is a Iall-out oI the smaller-than-usual elevation
diIIerence between the column and the reboiler necessitated by the need to
reduce subcooled liquid heating. The lower diIIerence in elevation translates into
a lower liquid circulation rate and, consequently, a lower liquid mass velocity
and, consequently, a lower heat transIer coeIIicient.
Thus, there is a conIlict between two opposing tendencies in 2 and 3 above: a lower head
reduces subcooled liquid heating but also reduces the liquid circulation rate. While the
Iormer will tend to improve the boiling heat transIer coeIIicient, the latter will tend to
reduce it. Thus, a balance will have to be struck between these two opposing tendencies.
This is demonstrated in Case Study 8.6 in a later section.
Operation near critical pressure
As the operating pressure increases and approaches the critical pressure, the diIIerence
between the liquid density and the vapor density decreases and, consequently, so does the
driving head. This tends to make 'thermosyphoning¨ more diIIicult. Besides, as a conse-
quence oI the reduced diIIerence in density between the liquid and the vapor, the poten-
tial Ior vapor-liquid separation also decreases, thereby reducing the maximum heat Ilux.
Kettle reboilers operate much better in such circumstances, as demonstrated in Case
Study 8.5. OI course, the limitation here is that the boiling Iluid should not be dirty. II that
were to be the case, the best option will be to adopt a Iorced-Ilow reboiler with no
vaporization.
It must be mentioned here that authentic physical properties are very important in this
region and the same are diIIicult to predict. Especially important are vapor and liquid density
144
and critical pressure and temperature. Pseudo-critical pressure and temperature are oIten
Iurnished and are not accurate enough. For example, the true critical pressure is usually
greater than the pseudo-critical pressure.
Film boiling
Film boiling occurs at very high temperature diIIerences and is undesirable as the heat
transIer coeIIicient is low, usually 82123 Btu/h It
2
ƒF (400600 kcal/h m
2
ƒC) and it can
result in heavy Iouling and hot spot problems.
Film boiling can be avoided by using either or both oI the Iollowing:
a) Decrease heating medium temperature level ² this is the direct approach.
b) Increase tube diameter ² oIten, this change alone is suIIicient to overcome Iilm
boiling.
Sometimes, Iilm boiling cannot be helped because increasing the tube diameter is not
suIIicient and the temperature level oI the heating medium cannot be decreased.
(Evidently, this is applicable to sensible heating mediums because with a condensing
medium, such as steam, the pressure level can be reduced to eliminate Iilm boiling.) In
such instances, as long as the entire reboiler is in Iilm boiling, the design may be
accepted.
However, in no instance should a design be accepted when a part oI the tubes is in
nucleate boiling and the balance in Iilm boiling; and this includes operation at part-load or
even start-up. This is because operation is diIIicult to predict since changes in operating
conditions can result in extreme changes in perIormance. Besides, reverse control charac-
teristics prevail here: increasing ǻT decreases the vapor rate and heat duty, which is the
reverse oI the normal response, namely, an increase in ǻT results in an increase in heat duty.
In such situations (part nucleate and part Iilm boiling), the heating medium temperature
level should be changed so that the entire reboiler is either in the nucleate boiling regime (by
decreasing the inlet temperature) or in the Iilm boiling regime (by increasing the outlet
temperature).
The temperature diIIerence Ior the beginning oI Iilm boiling is a very strong Iunction oI
the reduced pressure (p/p
c
), whereas the temperature diIIerence Ior the end oI Iilm boiling
does not vary signiIicantly.Thus, Ior a light hydrocarbon reboiling service, the ǻT Ior the
commencement oI Iilm boiling increases Irom 20ƒF (11.1ƒC) to 59ƒF (32.8ƒC) to 112ƒF
(62.2ƒC) to 169.9ƒF (94.4ƒC) as the p/pc reduces Irom 0.9 to 0.8 to 0.6 to Iinally 0.3.
However, the corresponding values oI ǻT Ior the completion oI Iilm boiling undergo a Iar
smaller change, Irom 392.9ƒF (218.3ƒC) to 401.0ƒF (222.8ƒC) to 406.0ƒF (225.6ƒC) to
432ƒF (240ƒC).
Very low ǻT
A very low ǻTless than 7.2ƒF (4ƒC)is sometimes speciIied in order to eIIect energy
economy or due to process limitations. As has been said earlier, nucleate boiling heat
transIer coeIIicient is very low at a low ǻT and is very sensitive to microscopic surIace
conIiguration. The most eIIective solution in such circumstances is to employ special
porous surIaces (high-Ilux tubes) which augment the nucleate boiling heat transIer
coeIIicient several-Iold. This is proprietary hardware which necessitates the design to be
in the vendor's scope. In Iact, low ǻT and porous boiling surIaces oIten go hand-in-hand.
Low-Iin tubes also greatly improve perIormance by virtue oI increased nucleation and
the extended surIace. However, the use oI these tubes (porous surIace or low-Iin tubes)
should be restricted to clean boiling Iluids.
145
A low ǻT translates into a low heat Ilux and a low vaporization rate which means that
the 'siphoning¨ eIIect is subdued. Consequently, thermosyphon reboilers are inherently
unsuitable Ior low ǻT operation and kettle reboilers operate much better.
Boiling on the shellside
Since vertical thermosyphon reboilers have a single tube pass, they are invariably oI Iixed
tubesheet construction. Thus, they are not suitable Ior dirty heating mediums unless
chemical cleaning can be resorted to.
Should the heating medium be very dirty, a vertical thermosyphon reboiler can have
boiling on the shellside. However, due to the presence oI baIIles in such a construction, it
becomes very diIIicult to simulate the shellside perIormance. The use oI rod baIIles is much
better suited than that oI segmental baIIles, especially single-segmental baIIles which cause
vapor trapping at the baIIle roots. The use oI vertical thermosyphon reboilers having boiling
on the shellside is extremely rare.
The Ieatures oI vertical thermosyphon reboilers that have just been discussed will be
demonstrated by two case studies.
CASE STUDY 8.4: DISTILLATION COLUMN REBOILER
(VERTICAL THERMOSYPHON)
The principal process parameters oI a vertical thermosyphon reboiler are elaborated in
Table 8.4a. Being a vertical thermosyphon, a Iixed-tubesheet construction was to be
employed. Tubes were to be either 0.7874-in. (20-mm) OD and 0.0787-in. (2-mm) thick
Table 8.4a: Salient process parameters oI a distillation column reboiler 1 (vertical thermosyphon)
Shellside Tubeside
1. Fluid circulated Hot oil Hydrocarbon
2. Flow rate, lb/h (kg/h) 83,300 (37,800) 140,430 (63,700)
3. Fluid vaporized - 47,000 (21,300)
4. Temperature in/out, ƒF (ƒC)
435.2 (224)/337.6
(169.8)
192.7 (89.3)/199.6
(93.1)
5. Operating pressure, psia (kg/cm
2
abs.) 85.0 (6.0) 340.0 (23.9)
6. Permitted pressure drop, psi (kg/cm
2
) 8.5 (0.6)
As per
thermosyphon circuit
7. Fouling resistance, h It
2
ƒF/Btu (h m
2
ƒC /kcal) 0.00195 (0.0004) 0.00098 (0.0002)
8. Heat duty, MM Btu/h (MM kcal/h) 4.96 (1.25)
9. Mol. wt. vapor out - 47.6
10. Critical pressure, psia (kg/cm
2
abs.) - 540 (38.0)
11. Critical temperature, ƒF (ƒC) - 262 (128)
12. Design pressure, psig (kg/cm
2
g) 142 (10.0) 384 (27.0)
13. Design temperature, ƒF (ƒC) 572 (300) 266 (130)
14. Material oI construction Carbon steel
15. Line size in/out, in. (mm) 4 (100)/4 (100) 8 (200)/10 (250)
146
or 0.984-in. (25-mm) OD and 0.0984-in. (2.5-mm) thick.
Since use oI smaller diameter tubes leads to a cheaper design, a Iirst design was made
with 0.7874-in. (20-mm) OD tubes. The results are shown in Table 8.4b, run 1. The most
striking aspect oI this design is that the entire reboiler is in the Iilm-boiling region due the
very high temperature diIIerence between the heating medium and the boiling Iluid. As has
been stated earlier, while there is nothing wrong with a Iilm boiling design, a designer
always preIers to have a nucleate boiling design, as it usually leads to a more economical
design. Besides, Iilm boiling can result in heavier Iouling and hot spot problems.
As has also been stated earlier, one way to alter the operation oI a thermosyphon
Table 8.4b: Thermal design oI distillation column reboiler (vertical thermosyphon)
Run 1
Base design
with 0.7874
in. (20 mm)
OD tubes
Run 2
with
0.984 in.
(25 mm)
OD tubes
Run 3
with
reduced
hot oil
temp.
Run 4
with Iurther
reduced
hot oil
temp.
Run 5
with still
Iurther
reduced hot
oil temp.
1. Hot oil Ilow rate,
lb/h (kg/h)
83,300
(37,800)
83,300
(37,800)
91,880
(41,675)
128,700
(58,380)
199,300
(90,390)

2. Hot oil temp. in/out,
°F (°C)
435.2
(224)/
337.6
(169.8)
435.2
(224)/
337.6
(169.8)
360
(182.2)/
270 (132.2)
300
(148.9)/
230 (110)
285.1
(140.6)/
230 (110)
Circulation,
lb/h (kg/h)
140,430
(63,700)
147,700
(67,000)
155,200
(70,400)
147,600
(66,950)
157,300
(71,350)

3. Hydro-
carbon
°
vaporized
33.4 31.0 29.5 32.4 29.9
4. Shell ID, in. (mm ) 17.3 (438) 20.1 (510) 25 (635) 28.5 (725) 28.5 (725)
5. No. oI tubes 200 190 280 360 360
6. Tube OD × thickness,
in. (mm)
0.787 (20)
×
0.0787 (2)
0.984 (25)
×
0.0984 (2.5)
7. Tube length, It (mm)
11.8 (3600) 9.84 (3000) 9.84 (3000) 9.84 (3000) 11.5 (3500)
8. Driving head, in. (mm)
11.8 (3600) 9.84 (3000) 9.84 (3000) 9.84 (3000) 11.5 (3500)
9. Heat transIer area, It
2
(m
2
) 487 (45.3) 481 (44.7) 709 (65.9) 912 (84.8)
1064 (98.9)
Hot oil 221 (1081) 173 (844) 143 (699) 141 (687) 173 (844)
Hydro-
carbon
114 (555) 114 (555) 283 (1380) 367 (1792) 387 (1889)
10. Heat
transIer
coeIIicient,
Btu/h It
2
°F
(kcal/h m
2
°C)
Overall 59 (288) 55.9 (273) 64.7 (316) 71.9 (351) 83.6 (408)
11. MTD, °F (°C)
185.8
(103.2)
185.2
(102.9)
113.8
(63.2)
63.2
(35.1)
58.1
(32.3)
12. Overdesign, ° 7.6 1.7 5.5 -18.5 4.0
Bottom Film Film Bubble Bubble Bubble
Middle Film Film Film Bubble Bubble

13. Flow regime
Top Film Film Film Film Slug

14. Design acceptable
Yes, Iully
Iilm boiling
Yes, Iully
Iilm boiling
No, partial
Iilm boiling
No, partial
Iilm boiling
Yes, Iully
nucleate
boiling
147
reboiler Irom Iilm boiling to nucleate boiling is to increase the tube diameter, as it provides
more Ilow area on the tubeside Ior a given heat transIer area and a given tube length.
Accordingly, in run 2, 0.984-in. (25-mm) OD and 0.0984-in. (2.5-mm) thick tubes were
used. Further, the tube length was reduced Irom 11.8 It (3600 mm) to 9.84 It (3000 mm).
However, it was discovered that the reboiler still remained in the Iilm boiling mode. This led
to the conclusion that the existing ǻT was too high to permit nucleate boiling.
The next run (run 3) was taken with a reduced hot oil inlet temperature, 360ƒF (182.2ƒC)
instead oI 435.2ƒF (224ƒC). There were two signiIicant changes:
1) The reboiler went partly into the nucleate boiling regime at the inlet oI the
reboiler and, as a direct consequence, the boiling heat transIer coeIIicient
increased Irom 113.7 Btu/h It2 °F (555 kcal/h m2 °C) in run 2 to 282.6 Btu/h It2
°F (1380 kcal/h m2 °C). The overall heat transIer coeIIicient increased Irom 55.9
Btu/h It2 °F (273 kcal/h m2 °C) to 64.7 Btu/h It2 °F (316 kcal/h m2 °C), an
increase oI 15.8°.
2) The MTD reduced Irom 185.2°F (102.9°C) to 113.8°F (63.2°C), a steep Iall oI
about 39°.
Since the Iall in the MTD was much greater than the rise in the overall heat transIer
coeIIicient, the heat transIer area was higher at 709 It
2
(65.9 m
2
) as compared to the 481
It
2
(44.7 m
2
) oI the previous design.
However, as explained earlier, this design is not acceptable because it is partly in the
nucleate boiling regime and partly in the Iilm-boiling regime.


Run 4 was taken with a Iurther decrease in the hot oil inlet temperature, down to 300ƒF
(148.9ƒC). Here, a greater part oI the reboiler was in the nucleate boiling regime but, since
part Iilm boiling still existed at the outlet oI the reboiler, the design was still not acceptable.
Finally, in run 5, a totally nucleate boiling design could be produced with a hot oil inlet
temperature oI 285.1ƒF (140.6ƒC). Although the overall heat transIer coeIIicient increased
appreciably, the MTD had plummeted to 58.1ƒF (32.3ƒC). Hence, the heat transIer area went
up to 1064 It
2
(98.9 m
2
), considerably higher than the 487 It
2
(45.3 m
2
) in the original Iilm
boiling design.
It was Ielt prudent to retain the original design due to the Iollowing:
a) It was much cheaper.
b) Totally Iilm boiling design is perIectly acceptable Ior this application, as the
temperature level is not high enough Ior hot spots to occur and the boiling Iluid is
clean enough to preclude any heavy Iouling.
II Iilm boiling were not acceptable, it would have become necessary to modiIy the hot oil
circuit in order to reduce the hot oil inlet temperature to the reboiler.
The above example demonstrates the nuances oI the boiling mode in a vertical
thermosyphon reboiler vis-a-vis some key operating variables.
The reader may ask here as to why the hot oil inlet temperature was so high in the Iirst
place. The answer is that there must have been other users in the plant which were to operate
at a much higher temperature level. However, when two or more users operate at rather
diIIerent temperatures, a suitable series/parallel hot oil arrangement is usually implemented
so that the ǻT is optimum in all the users.
A signiIicant observation to be made here is that a high heating medium heat transIer
coeIIicient (such as with steam or hot water) results in a higher tube-wall temperature and,
thereby, a higher ǻT across the boiling Iilm. This inIluences the boiling heat transIer
148
coeIIicient, depending upon the position oI the initial operating point on the heat Ilux versus
ǻT curve. For example, iI an initial operating point is in the lower nucleate boiling regime, it
is possible to increase the boiling heat transIer coeIIicient signiIicantly by increasing the heat
transIer coeIIicient oI the heating medium, thereby increasing ǻT across the boiling Iilm.
Steam is the preIerred heating medium Ior reboilers as compared to hot oil because it is
more economical and oIIers better control by throttling. It is only when the required steam
temperature and, thereby, the steam pressure become excessive that steam heating is no
longer economical. The upper limit oI a process Iluid temperature Ior steam heating is
usually around 455464ƒF (235240ƒC) so that saturated steam at 569 psia (40 kg/cm
2
abs.)
and 480.6ƒF (249.2ƒC) may be used.
CASE STUDY 8.5: DISTILLATION COLUMNLIGHT REBOILER
(VERTICAL THERMOSYPHON/KETTLE)
The principal process parameters oI another distillation column reboiler are elaborated in
Table 8.5a. Once again, being a vertical thermosyphon reboiler, a Iixed-tubesheet
construction was to be employed. Tubes were to be 0.984-in. (25-mm) OD and 0.0984-in.
(2.5-mm) thick.
A thermal design was produced and the principal construction and perIormance
parameters are detailed in Table 8.5b. However, although the design appeared to be
acceptable, the Iollowing observations were interesting:
Table 8.5a: Salient process parameters oI a distillation column reboiler (vertical thermosyphon)
Shellside Tubeside
1. Fluid circulated
HP Steam Hydrocarbon
2. Flow rate, lb/h (kg/h)
37,370 (16,950) 661,400 (300,000)
3. Fluid vaporized, lb/h (kg/h)
- 161,600 (73,300)
4. Temperature in/out, °F (°C)
451.0 (232.8)/451.0 (232.8)
362.5 (183.6)/435.4 (224.1)
5. Operating pressure, psia (kg/cm
2
abs.)
426.6 (30.0) 378.3 (26.6)
6. Permitted pressure drop, psi (kg/cm
2
) 1.4 (0.1)
As per
thermosyphon circuit
7. Fouling resistance, h It
2
ƒF/Btu
(h m
2
ƒC /kcal)
0.00049 (0.0001) 0.00098 (0.0002)
8. Heat duty, Btu/h (kcal/h)
28.6 (7.2)
9. Mol. wt. vapor out
71.2
10. Critical pressure, psia (kg/cm
2
abs.)
429.4 (30.2)
11. Critical temperature, ƒF (ƒC)
457.2 (236.2)
12. Design pressure, psig (kg/cm
2
g)
480.6 (33.8) 455 (32.0)
13. Design temperature, °F (°C)
500 (260) 473 (245)
14. Material oI construction
Carbon steel
15. Line size in/out, in. (mm)
8 (200)/4 (100) 14 (350)/16 (400)
149
a) A satisIactory design could not be produced with the speciIied 0.984-in. (25-mm)
OD tubes because a part oI the reboiler would be in the nucleate boiling mode
and the balance in the Iilm boiling mode. The tube outside diameter had to be
increased to 1.25 in. (31.75 mm).
b) The maximum heat Ilux was 7467 Btu/h It
2
(20,237 kcal/h m
2
) whereas the
design heat Ilux was 6180 Btu/h It
2
(16,747 kcal/h m
2
).
c) Although the reboiler was entirely in the nucleate boiling regime, any attempt to
reduce the heat transIer area, either by reducing the number oI tubes or the tube
length, resulted in part oI the tubes going to Iilm boiling.
a) The ǻT Ior the commencement oI Iilm boiling was 27.1ƒC and ǻT at the inlet
end (bottom) oI the reboiler was very close to it.
e) Thus, the situation appeared to be extremely sensitive, so that minor diIIerences
in the operating conditions could easily result in perIormance upsets.
f) Finally, the overall heat transIer coeIIicient was rather low at 108.3 Btu/h It
2
ƒF
(529 kcal/h m
2
ƒC) so that the required heat transIer area was rather high at 4831
It
2
(449 m
2
).
It was Ielt that a vertical thermosyphon design Ior this service was not acceptable.
ThereaIter, the entire process simulation was revised based upon a kettle type reboiler.
The salient process parameters oI the kettle type reboiler are given in Table 8.5c.
It will be noticed that the heat duty was essentially the same. The hydrocarbon operating
pressure was identical but, since its composition was diIIerent (lighter) due to the diIIerent
operation, its temperature level was much lower. Superheated steam was available at 597
psia (42 kg/cm
2
abs.) and 518ƒF (270ƒC), the saturation temperature being 485.6ƒF (252ƒC).
The construction and perIormance parameters oI the kettle design are shown in Table
Table 8.5b: Principal construction and perIormance parameters oI
distillation column reboiler (vertical thermosyphon)
1. Type oI reboiler Vertical thermosyphon
2. Shell diameter, in. (mm) 53.2 (1350)
3. No. oI tubes × no. oI tube passes 900 × 1
4. Tube OD × BWG × length, in. (mm) 1.25 (31.75) × 12 (2.77 mm thick) × 197 (5000)
5. EIIective heat transIer area, It
2
(m
2
) 4831 (449)
Shellside Negligible
6. Pressure drop, psig (kg/cm
2
)
Tubeside 3.6 (0.255)
Shellside 1814 (8859)
Tubeside 152 (742)
7. Heat transIer coeIIicient, Btu/h
It
2
ƒF (kcal/h m
2
ƒC)
Overall 108.3 (529)
8. Maximum heat Ilux, Btu/h It
2
(kcal/h m
2
) 7467 (20,237)
9. Design heat Ilux, Btu/h It
2
(kcal/h m
2
) 6180 (16,747)
10. MTD, °F (°C) 58.1 (32.3)
11. Overdesign, ° 6.3
150
8.5d. The heat transIer area reduced Irom 4831 It
2
(449 m
2
) to 1173 It
2
(109 m
2
) due to a
very large increase in the MTD (117°) by virtue oI the revised boiling temperature and a
very large increase in the overall heat transIer coeIIicient (93°). The ratio oI the design to
maximum heat Ilux was 0.616 as against 0.828 in the vertical thermosyphon reboiler design,
indicating a much more stable design. The kettle design was adopted.
8.5.5 Forced-flow reboilers
In special situations, it becomes prudent to suppress vaporization, e.g., Iouling liquids
having a rather high viscosity or a wide-boiling mixture operating at a low pressure. In
the Iormer, the heat transIer coeIIicient could be inordinately low and in the latter, the
reduction in MTD due to the head oI liquid could be dramatic. In such situations, it is
oIten better to have no vaporization by having a pump circulate the liquid at a high circu-
lation rate and at a pressure above saturation (Fig. 8.12). The liquid is sensibly heated and
Table 8.5c: Salient process parameters oI distillation column reboiler (kettle reboiler)
Heat duty, MM Btu/h (MM kcal/h) 28.6 (7.2)
Flow rate, lb/h (kg/h) 297,600 (135,000)
Pressure, psia (kg/cm
2
abs.) 378.3 (26.6)
Hydrocarbon Fraction vaporized 0.443
Bottom drawoII, lb/h (kg/h) 165,800 (75,195)
Temp. in/out, ƒF (ƒC) 269.6 (132)/359.6 (182)
Pressure, psia (kg/cm
2
abs.) 597 (42.0)
Steam
Temperature in/out, ƒF (ƒC) 518 (270)/485.6 (252)
Table 8.5d: Principal construction and perIormance parameters oI
distillation column reboiler (kettle reboiler)
1. Type oI reboiler Kettle
2. Shell/port diameter, in. (mm) 47.2 (1200)/27.6 (700)
3. No. oI tubes × no. oI tube passes 168 U`s × 2
4. Tube OD × BWG × straight length, It (mm) 0.984 (25) × 0.0984 (2.50 × 13.1 (4000)
5. Heat transIer area, It
2
(m
2
) 1173 (109)
Shellside Negligible
6. Pressure drop, psi (kg/cm
2
)
Tubeside Negligible
Shellside 2198 (10,732)
Tubeside 4133 (2017)

7. Heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
Overall 209 (1020)
8. Maximum heat Ilux, Btu/h It
2
(kcal/h m
2
) 39,833 (107,948)
9. Design heat Ilux, Btu/h It
2
(kcal/h m
2
) 24,521 (66,452)
10. MTD, °F (°C) 126 (70)
11. Overdesign, ° 8.1
151
then Ilashed across a valve just beIore
entry into the distillation column.
Evidently, very large Ilow rates have to
be circulated as only sensible heatand
no latent heatis imparted.
Thus, these 'reboilers¨ are just like
ordinary single-phase heat exchangers
and the same correlations apply. The
column bottom could be either on the
shellside or on the tubeside, as explained
in Chapter 5.
The disadvantage oI this type oI 're-
boiler¨ is the additional cost oI the pump
and the high operating cost. Consequent-
ly, these are generally preIerred and employed only in the special situations described above.
CASE STUDY 8.6: DISTILLATION COLUMN REBOILER
(VERTICAL THERMOSYPHON/FORCED-FLOW)
A distillation column reboiler was to be designed Ior a petrochemical plant. The principal
process parameters are Iurnished in Table 8.6a. It may be noticed that the operating

Fig. 8.12 Forced-Ilow reboiler (Courtesy oI HTRI.)
Table 8.6a: Salient process parameters oI distillation column reboiler
(vertical thermosyphon/Iorced-Ilow reboiler)
Shellside Tubeside
1. Fluid circulated Steam
Distillation Column
Bottoms
2. Flow rate, lb/h (kg/h) 3350 (1520) 29,540 (13,400)
3. Fluid vaporized, ° - 23.8
4. Temperature in/out, °F (°C)
340 (171.1)/340
(171.1)
302.7 (150.4)/308.5
(153.6)
5. Operating pressure, psia (kg/cm
2
abs.) 115 (8.1) 3.56 (0.25)
6. Permitted pressure drop, psi (kg/cm
2
) 1.0 (0.07)
As per
thermosyphon circuit
7. Fouling resistance, h It
2
ƒF/Btu (h m
2
ƒC /kcal) 0.00049 (0.0001) 0.003 (0.000614)
8. Heat duty, MM Btu/h (MM kcal/h) 2.92 (0.736)
9. Mol. wt. vapor out 63.1
10. Critical pressure, psia (kg/cm
2
abs.) 1067 (75.0)
11. Critical temperature, ƒF (ƒC) 700 (372)
12. Design pressure, psig (kg/cm
2
g) 156 (11.0) 85 (6.0)
13. Design temperature, °F (°C) 374 (190) 338 (170)
14. Material oI construction CS SS 304
15. Line size in/out, in. (mm) 4 (100)/1 (25) 6 (150)/16 (400)
152
pressure is extremely low3.56 psia (0.25 kg/cm
2
abs). Thus, there will be a signiIicant
subcooled liquid heating zone due to the eIIect oI the static head.
A vertical thermosyphon reboiler construction was considered Iirst. A condensing steam
heat transIer coeIIicient oI 6100 kcal/h m
2
ƒC was employed. Tubes /-in. (19.05-mm) OD
were Iound unsuitable and 1-in. (25.4-mm) OD tubes were used. It was Iound that the design
was very sensitive to the liquid driving head and a design was Iinalized with a tube length oI
12 It (3.657 m) and a driving head oI 7 It (2.133 m). The salient construction and
perIormance parameters oI this design are detailed in Table 8.6b.
It will be noticed that the boiling heat transIer coeIIicient is rather low, only 148.7 Btu/h
It
2
ƒF (726 kcal/h m
2
ƒC). This was due to:
a) large subcooled heating zone due to the static head
b) the low tubeside liquid velocity
The design heat Ilux was only 1938 Btu/h It
2
(5253 kcal/h m
2
).
In order to demonstrate how sensitive the perIormance oI this reboiler is with liquid
driving head, Table 8.6c presents the perIormance oI this reboiler with varying driving
heads: 4 It (1.22 m), 5 It (1.52 m), 6 It (1.83 m), 7 It (2.13 m), and 8 It (2.44 m).
The Iollowing eIIects are observed with increasing static head:
a) higher circulation and lower Iraction vaporized
b) higher liquid velocity at the bottom oI the reboiler (entry to tubes) as a direct
result oI a
c) increasing length oI subcooled liquid heating zone
Table 8.6b: Principal construction and perIormance parameters
oI distillation column reboiler (vertical thermosyphon/Iorced-Ilow reboiler)
1. Type oI reboiler
Vertical
thermosyphon
Suppressed
vaporization
2. Shell diameter, in. (mm) 30.7 (780) 25.2 (640)
3. No. oI tubes × no. oI tube passes 478 × one 274 × 4
4. Tube OD × BWG × length, It (mm) 1.0 (25.4) × 12 (2.77) × 12.0 (3657)
5. EIIective heat transIer area, It
2
(m
2
) 1506 (140) 839 (78)
6. Tubeside inlet liquid velocity, It/s (m/s ) 0.075 (0.023) 6.07 (1.85)
Shellside Negligible negligible
7. Pressure drop, psi (kg/cm
2
)
Tubeside Negligible 6.54 (0.46)
Shellside
1249 (6100)
(speciIied)
1249 (6100)
(speciIied)
Tubeside 148.7 (726) 572 (2794)

8. Heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
Overall 73.7 (360) 132.9 (649)
9. Maximum heat Ilux, Btu/h It
2
(kcal/h m
2
) 14,524 (39,360) -
10. Design heat Ilux, Btu/h It
2
(kcal/h m
2
) 1938 (5253) -
11. MTD, °F (°C) 26.3 (14.6) 27 (15.0)
12. Overdesign, ° Nil 2.8
153
a) decreasing boiling heat transIer coeIIicient due to c
A lower head was Iound to be capable oI transIerring a higher heat duty. However, since
the heat duty is Iixed, this will translate into a lower heat transIer area. However, the
Iraction vaporized was signiIicantly higher and would, thereby, cause more severe
Iouling. Thus, a static head oI 7 It (2.13 m) was considered to be optimum.
However, even this design was not considered very appropriate in view oI the extremely
low liquid velocity which would result in heavy Iouling. An alternate design employing
'suppressed vaporization¨ was examined using 1-in. (25.4-mm) OD tubes. The bottoms cir-
culation was Iixed at 338,400 lb/h (153,500 kg/h), which corresponded to an outlet tempera-
ture oI 314.6ƒF (157ƒC). The salient construction parameters are Iurnished in Table 8.6b.
It will be noticed that, due to the considerable increase in the tubeside heat transIer
coeIIicient, the heat transIer area reduced Irom 1506 It
2
(140 m
2
) to 839 It
2
(78 m
2
). Even
more signiIicant is the high tubeside velocity oI 6.07 It/s (1.85 m/s) which will minimize
Iouling. It is true that the pumping cost will be high, but the total operating cost may actually
be less in view oI the diminished Iouling. It was Iinally decided to adopt the 'suppressed
vaporization¨ design. Although a vertical disposition can be adopted, a horizontal
disposition will be better Ior ease oI maintenance.
8.6 Selection of Reboilers
Now that the principal Ieatures, advantages, and disadvantages oI the various types oI
distillation column reboilers have been elaborated, we are in a position to discuss the
methodology oI selection oI a reboiler type Ior various applications and situations.
Table 8.6c: Variation in perIormance oI distillation column reboiler
with driving head (vertical thermosyphon)
Driving head, ft (m)
4 (1.22) 5 (1.52) 6 (1.83) 7 (2.13) 8 (2.44)
1. Heat duty, MM Btu/h (MM
kcal/h)
4.59 (0.941) 4.28 (0.876) 3.93 (0.804) 3.59 (0.736) 3.29 (0.673)
2. Liquid circulation,
lb/h (kg/h)
17,341
(7866)
21,850
(9912)
25,410
(11,526)
29,430
(13,350)
33,000
(14,970)
3. Weight Iraction vaporized 0.531 0.39 0.305 0.238 0.19
4. Boiling heat transIer
coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
182.7 (892) 171.4 (837) 160.2 (782) 148.7 (726) 136.4 (666)
5. Overall htc, kcal/h m
2
ƒC 86.4 (422) 82.5 (403) 78.2 (382) 73.7 (360) 69.2 (338)
6. MTD, °F (°C) 28.6 (15.9) 27.9 (15.5) 27.0 (15.0) 26.3 (14.6) 25.6 (14.2)
7. Length oI liquid heating
zone, It (m)
1.44
(0.44)
2.17
(0.66)
2.89
(0.88)
3.61
(1.1)
4.56
(1.39)
Bottom slug bubble bubble bubble bubble
Middle slug slug Slug slug slug

8. Boiling
regime
Top annular annular annular annular slug
9. Liquid velocity at tube inlet,
It/s (m/s)
0.044
(0.0135)
0.056
(0.017)
0.066
(0.02)
0.075
(0.023)
0.085
(0.026)
154
Since there are several parameters, based upon which the selection must be made, and
since some oI them may be indicated Ior a particular type and others contraindicated Ior the
same type, an optimum selection must be made in many situations. In other words, the type
that scores the maximum points Ior all the parameters should be adopted. This is similar to
the selection oI sides (shellside and tubeside) described in Chapter 5. It must be highlighted
here that it is the total cost (Iixed cost plus operating cost) that determines the optimum
selection. However, satisIactory operation is a very important prerequisite and must never be
compromised in the selection process.
Since kettles and internal reboilers are identical in Ieatures and operation, comments
made below Ior kettles apply to internal reboilers too, unless otherwise noted.
1) Fouling tendency
Dirty Iluids are better handled in vertical thermosyphons, as Iouling is minimized
and cleaning oI the Iouled deposits is easier. Kettle reboilers are the worst type
Ior dirty services.
2) Mixture boiling range
For wide-boiling mixtures, thermosyphon reboilers are best because with kettles,
there is a drastic reduction in the MTD. Horizontal thermosyphons score over
vertical thermosyphons. For very wide-boiling mixtures, however, a "suppressed
vaporization" design becomes Iavorable.
3) Operating pressure
At low pressure (vacuum) and at high pressures near critical, vertical thermo-
syphons are severely limiting. Horizontal thermosyphons are better, but kettles
are best. In extreme situations, Iorced-Ilow "reboilers" are required to be
employed.
4) Temperature diIIerence
At high ǻT where the possibility oI Iilm boiling is high, kettles are the best.
Vertical thermosyphons should be avoided.
8.7 Start-Up of Reboilers
When a clean reboiler is started up, it is oversurIaced since there is no Iouling.
Consequently, the ǻT across the boiling Iilm will be higher than that designed Ior, and so
will be the heat Ilux. Normally, this should pose no problem, as the reboiler will tend to
deliver a larger heat duty. However, as pointed out earlier, the operating point may shiIt
Irom stable nucleate boiling to unstable nucleate boiling whence there may be a reduction
in perIormance. Consequently, it is always advisable to start-up reboilers gradually.
II the heating medium is steam or any other condensing medium, reduction oI pressure
oIIers excellent control. A control valve decreases the Ilow rate and saturation pressure.
Consequently, the saturation temperature and the MTD reduce to the desired extent. This is
the most preIerred method oI control.
In some cases, the requisite pressure drop across a control valve is not available. In such
cases, the condensate Ilow is pinched so as to Ilood the bottom part oI the shell, thereby
reducing the eIIective heat transIer area and oIIering the required control. However, this
method has disadvantages and should only be employed as a last resort.
II the heating medium is a sensible liquid, Ilow is bypassed around the reboiler. This
oIIers a twoIold control by reducing (a) the heating medium heat transIer coeIIicient and (b)
155
the MTD, as the outlet temperature oI the heating medium reduces. A disadvantage with this
method is the accelerated Iouling caused by the reduced tubeside velocity.
References
|1| Sloley, A.W., 1997 'Properly Design Thermosyphon Reboilers,¨ Chemical Engineering
Progress, March, pp. 52²64.
|2| Martin, G.R., and Sloley, A.W., 1995 'EIIectively Design and Simulate Thermosyphon
Reboiler Systems,¨ Hyarocarbon Processing, July, pp. 67²78.
|3| McCarthy, A.J., 1995, 'Reboiler System Design: The Tricks oI the Trade,¨ Chemical
Engineering Progress, May, pp. 34²47.
Further Reading
1. Hewitt, G.F., 1998, Heat Exchanger Design Hanabook, Begell House, Inc., New York.
2. Love, D.L., 1992, 'No Hassle Reboiler Selection,¨ Hyarocarbon Processing, Octr, pp. 41²47.
3. Yilmaz, S.B., 1987, 'Horizontal Shellside Thermosyphon Reboilers,¨ Chemical Engineering
Progress, Nov, pp. 64-70.
4. Thome, J.R., |1988|, 'Reboilers with Enhanced Boiling Tubes,¨ Heat Transfer Engineering,
9(4), pp. 45²62.
156

157
CHAPTER 9
3K\VLFDO3URSHUWLHVDQG
+HDW5HOHDVH3URILOHV
The importance oI Ieeding authentic physical properties cannot be overemphasized. The
author has seen numerous instances wherein the Ieeding oI incorrect properties has
produced incorrect results. Besides physical properties themselves, heat release proIiles
are also very important and must be Ied accurately and meaningIully in order to produce
realistic and consistent results.
It must be stated here that when the Ieeding oI an incorrect physical property produces
an absurd or exaggerated heat transIer coeIIicient or pressure drop, the experienced designer
may oIten be able to sense it because these values will be well beyond the expected range.
But not so with the inexperienced or less experienced designer. Consequently, it is strongly
recommended that the input values be checked careIully in order to eliminate any probability
oI incorrect results.
9.1 Physical Properties
Please reIer to Section 3.2 where a brieI discussion oI physical properties was presented.
Viscosity, thermal conductivity, speciIic heat, and density are the Iundamental properties
Ior single-phase applications. Additionally, Ior condensing/vaporizing services, critical
pressure, critical temperature, and surIace tension also assume signiIicance.
It must be stated at the outset that physical properties should be Iurnished at both the
inlet and outlet temperatures, especially iI the diIIerence between the two temperatures is
highwhen the values oI these properties are likely to vary appreciably.
The physical property that varies the most with temperature is liquid viscosity,
especially in the case oI heavy liquids. To illustrate this, iI the inlet temperature oI a stream
to a heat exchanger is 140ƒF (60ƒC) and the outlet 104ƒF (40ƒC), the variation is not large
and even average physical properties may be Ied without entailing any serious error.
However, iI the inlet temperature is 284ƒF (140ƒC) and the outlet 140ƒF (60ƒC), it would be
inadvisable not to Ieed physical properties at both these temperatures.
In many designs, the heat transIer coeIIicient varies considerably along the length oI the
exchanger, primarily due to the variation oI liquid viscosity. Coupled with this is the Iact that
the MTD may also vary signiIicantly along the length oI the exchanger, due to the nature oI
temperature proIiles oI the two streams. (That is, one stream has a considerably larger
temperature change Irom inlet to outlet than the other.) II one considers the example oI a
viscous liquid cooler, both the heat transIer coeIIicient and the MTD will reduce
considerably Irom inlet to outlet along the length oI the exchanger, so that the heat transIer
area required per unit heat duty will increase Irom exchanger inlet to outlet. Consequently, iI
158
average physical properties are Ied, the results will be unrealistic because a constant heat
transIer coeIIicient will be applied.
It is the responsibility oI the process licensor to Iurnish authentic physical properties and
this usually does not represent a problem since the output oI any standard process simulator
includes all relevant physical properties. Even in cases where comprehensive physical
property data is not Iurnished, it is not diIIicult to obtain speciIic heat, density, and thermal
conductivity data oI hydrocarbons, as the same is very well documented |1²3|.
The speciIic heat, viscosity, and thermal conductivity data oI hydrocarbon vapors are a
Iunction oI molecular weight and temperature. The values oI all these properties increase
linearly with temperature. Hydrogen has a considerably higher speciIic and thermal
conductivity than hydrocarbon vapors.
The density oI a hydrocarbon vapors is expressed as (pM)/(zRT), where
p ÷ operating pressure
M ÷ molecular weight
z ÷ compressibility Iactor
R ÷ universal gas constant
T ÷ absolute temperature
At moderate temperatures and pressures, the compressibility Iactor may be considered to
be 1.0 without entailing any serious error.
The speciIic heat, speciIic gravity, and thermal conductivity oI a hydrocarbon liquid
vary with temperature and API gravity. While speciIic gravity and thermal conductivity
decrease with temperature, speciIic heat increases with temperature. The viscosity oI
hydrocarbon liquids varies irregularly with temperature and the same cannot be represented
on any conventional scale (linear, semi-log, or log-log). This variation Ior various
hydrocarbon liquids is represented in special plots by ASTM (Fig. 9.1). It will be seen that

Fig. 9.1 ASTM plot oI variation oI hydrocarbon liquid viscosity with temperature (Reprinted with
permission Irom Standards oI TEMA, 8th

Edition, 1999.)
159
the variation becomes extremely large Ior heavy liquids at low temperatures.
While all other physical properties may be obtained Irom various sources iI they are not
Iurnished in the process data sheet, it will be prudent to insist upon the liquid viscosity
values Irom the process licensor. OI course, this excludes standard pure components Ior
which data is available in |13|.
It should also be stated here that, excepting liquid viscosity, all other physical properties
vary linearly with temperature Ior all practical purposes and, hence, they need be Iurnished
to the heat exchanger soItware only at the inlet and outlet temperatures. However, as already
mentioned, the variation oI viscosity with temperature is highly nonlinear and, consequently,
iI the variation in liquid viscosity between the inlet temperature and the outlet temperature is
high, it is advisable to Ieed the viscosity values at intermediate temperatures as well.
The reason Ior this is that even a sophisticated thermal design soItware package may not
have the provision oI being able to evaluate the intermediate values accurately. Evidently,
the number oI intermediate viscosity values that should be Ied will depend upon the vari-
ation in viscosity. The important thing to do is to Iurnish a suIIicient number oI values so
that iI a straight-line interpolation is implemented between any two points Ied, the repre-
sentation will not be erroneous.
Thus, iI the inlet viscosity is 2.0 cp at 248ƒF (120ƒC) and the outlet viscosity is 5.0 cp at
140ƒF (60ƒC), Iurnishing an intermediate viscosity value oI 3.5 cp (say) at the mean
temperature may not be unreasonable. However, iI the inlet and outlet viscosities at these
same temperatures are 10.0 cp and 54.0 cp, more points will have to be Ied Ior a proper
representation. Evidently, the greater the number oI points Ied, the more accurate will be the
results. Most thermal design soItware packages permit the entry oI values up to a maximum
oI ten temperature points. It is suggested that the intermediate points be so Ied that the ratio
between any two viscosity values is more or less the same. Many designers Ieel tempted to
Ieed the intermediate points along roughly equal temperature increments. Thus, in the above
case, the temperature variation between inlet and outlet is (248 ² 140) or 108ƒF |(120 ² 60)
or 60ƒC|. Thus, the designer may Ieed intermediate viscosity values at two 36ƒF (20ƒC)
intervals, i.e., at 212ƒF(100ƒC) and 176ƒF (80ƒC). However, as the variation oI liquid
viscosity is not linear but exponential, this will not result in a proper representation. The
variation is much steeper at the lower temperature range, so there should be more points at
the lower temperature end.
In the case oI a condenser where there is no liquid at the inlet and condensation begins at
a slightly lower temperature (that is, the inlet vapor is somewhat superheated), a licensor's
data sheet oIten indicates only the outlet liquid physical properties. This is because, in the
data sheet, there is provision Ior speciIying physical properties only at the inlet and outlet
temperatures. For example, iI the inlet temperature is 212ƒF (100ƒC) and the outlet
temperature is 104ƒF (40ƒC), liquid physical properties are Iurnished only at 104ƒF (40ƒC).
A common mistake in such cases is to Ieed the liquid physical properties only at the outlet
temperature. What most heat exchange thermal design soItware does in such situations is to
assume that the liquid physical properties at the inlet temperature are identical to those at the
outlet temperature. As already explained, this may result in error. What should be done in
such situations is to obtain the liquid physical properties at the inlet temperature Irom
standard physical properties charts or sensible extrapolation and Ieed the same to the thermal
design soItware package.
160
9.2 Physical Property Profiles
Besides heat duty and weight Iraction vapor, any other parameter that does not vary
linearly with temperature will also have to be speciIied Ior an accurate thermal design.
Physical properties such as speciIic heat, thermal conductivity, and density oI both vapor
and liquid essentially vary linearly with temperature, as does vapor viscosity. However,
liquid viscosity does not vary linearly with temperature and, iI the diIIerence between the
inlet and outlet temperatures is not low, there could be a signiIicant variation in this
parameter Irom the inlet to the outlet oI the condenser. In such a situation, the variation in
liquid viscosity with temperature should also be Ied as a part oI the heat release proIiles.
Another interesting case is observed in the condensation oI streams containing both
hydrogen and hydrocarbons in hydrogen plants. When such a mixture is cooled, only the
hydrocarbon will condense so that the concentration oI hydrogen will increase. As the value
oI speciIic heat is the same in British and metric units, only the values will be mentioned
hereaIter in this chapter. Now, hydrogen has a much higher speciIic heat (typically 3.5 as
compared to 0.5 oI hydrocarbons) so that as the hydrogen concentration increases, the
speciIic heat oI the mixture increases. There is another eIIect on the speciIic heat oI the
mixture, namely that oI temperature. The speciIic heat oI both hydrogen and hydrocarbons
decreases with a reduction in temperature. However, the eIIect oI the increase in the
concentration oI hydrogen usually Iar outweighs that oI the reduction in temperature, so that
the speciIic heat oI the mixture increases as the hydrogen concentration increases with the
reduction in temperature.
In many such condensers, there is an initial desuperheating zone, aIter which the
condensation oI hydrocarbon starts. Thus, Irom the inlet temperature to the dew point, there
will be a decrease in the total vapor speciIic heat as the hydrogen concentration remains the
same and the temperature reduces. However, once condensation oI hydrocarbon begins, the
total vapor speciIic heat will begin to increase and usually ends being signiIicantly higher
than that at the inlet. In such cases, thereIore, the variation oI vapor speciIic heat with
temperature must be Iurnished to the thermal design soItware package as well.
It should be understood that, besides accurately determining the heat transIer coeIIicient
zone-wise, the vapor speciIic heat proIile is also required in the above case Ior reconciling
the heat duty in each zone. An actual case study Ior a hydrogen plant condenser is shown in
Fig. 9.2 below. A mixture oI naphtha and hydrogen at 78.2 psia (5.5 kg/cm
2
abs.) is con-
densed Irom 680ƒF (360ƒC) to 104ƒF (40ƒC). The variation in the total vapor speciIic heat is
represented by the curve ABCD. It will be seen that the speciIic heat decreases Irom 0.698 at
680ƒF (360ƒC) to 0.51 at the dew point oI 320ƒF (160ƒC), where aIter it increases sharply to
1.29 at 104ƒF (40ƒC). II this curve were not Ied, a linear variation (shown by the straight line
AD) would be considered by the soIt-
ware. As a consequence, it would Iail
hopelessly to reconcile the heat duty oI
each zone speciIied in the heat duty ver-
sus temperature proIile because it would
consider much higher values oI mixture
speciIic heat!
The above phenomenon is true oI the
thermal conductivity oI hydrogen-hydro-
carbon mixtures as well. Thus, at 212ƒF
(100ƒC), the thermal conductivity oI
hydrogen is 0.121 Btu/h It ƒF (0.18 kcal/h
Fig. 9.2 Variation oI vapor speciIic heat with
temperature in a hydrogen plant condenser
161
m ƒC), whereas a typical hydrocarbon thermal conductivity at the same temperature is
0.0148 Btu/h It ƒF (0.022 kcal/h m ƒC). II the intermediate values oI the thermal conduc-
tivity oI the vapor mixture are not Ied, a linear interpolation would be employed between the
values Ied at the inlet and outlet temperatures. Consequently, much higher values oI thermal
conductivity would be considered, thereby leading to an unrealistically optimistic design.
9.3 Heat Release Profiles
A heat release proIile is a plot oI heat duty and weight Iraction vapor versus temperature
and is an essential part oI the process data sheet.
In single-phase services, the heat release proIile is essentially linear. However, in
condensing and vaporizing services, the heat release proIile is usually not linearthe slope
varies Irom inlet to outlet. II the temperature range is low-to-moderate (say, · 3654 ƒF or
2030 ƒC), the variation in slope is usually small and a linear proIile may be considered.
II a heat release proIile is linear, the same is stated in the process data sheet and a plot
need not be Iurnished to the computer program or considered Ior hand calculations.
II a heat release proIile is not Iurnished and it is also not stated that the same is linear, a
straight-line proIile may be assumed, provided the temperature range is low-to-moderate (·
54ƒF or 30ƒC). However, iI the temperature range is higher, the licensor may be asked to
either Iurnish the heat release proIile or conIirm that a linear proIile may be considered.
When a heat release proIile is not linear, there will be an error in the determination oI
the MTD by considering it to be linear. Depending upon the nature oI the proIile, it could be
higher or lowerusually it is lower. The greater the curvature oI the proIile, the greater will
be the diIIerence in the MTD determined Irom the actual proIile and an assumed linear
proIile.
The heat duty versus temperature plot is essential Ior the determination oI MTD in the
various zones. The weight Iraction vapor condensed versus temperature data is essential Ior
two reasons:
a) To reconcile the heat duty oI each regionthis includes the phase change duty and
the sensible (vapor and liquid cooling/heating) duty. Thermal design soItware
packages usually evaluate the vapor and liquid cooling duties Irom the respective
Ilow rates, inlet/outlet temperatures, and speciIic heats. They then subtract the total
sensible heating/cooling duty Irom the total heat duty, to obtain the vapo-
rizing/condensing duty. The latent heat is then determined Irom the vapori-
zing/condensing duty and the amount oI liquid vaporized/vapor condensed. A
negative latent heat or an unusually high or low latent heat indicate an error in the
data, which should then be examined and rectiIied. This is a very important step in
the design oI condensers and reboilers/vaporizers.
b) The vapor and liquid Ilow rates in the various locations (zones) are required to
compute the heat transIer coeIIicient and pressure drop.
9.4 How to Feed Heat Release Profiles
The number oI points should be so chosen that each segment oI the curve is virtually a
straight line. ReIerring to Fig. 9.3, the Iollowing points must be specifiea as a minimum:
A (80ƒC), E (48ƒC), F (44ƒC), and G (42ƒC)
Points B, C, and D need not be Ied because they lie on the straight line AE. Additionally,
it is recommended that one intermediate point may be Ied in sector EF and another in
sector FG Ior extra accuracy, as these sectors have some curvature. However, points E
and F are absolutely essential because, iI not Ied, the proIile will alter considerably.
162
No extra accuracy is achieved by
Ieeding several points on the straight por-
tion oI a curve. On the other hand, by
Ieeding unnecessary points, the proba-
bility oI a mismatch in heat duty and a
consequent negative latent heat increases,
especially iI the temperature increments
and/or the amount vaporized or con-
densed are small. It is, thereIore, recom-
mended that the minimum number oI
points required to represent the data
authentically be Ied.
When the divisions on the axes are quite large, it is sometimes diIIicult to read some
intermediate points Irom plots. In such cases, it is helpIul to construct intermediate lines or
plot the same data on graph paper having more intermediate divisions.
The temperature points to be Ied to the thermal design soItware should be so chosen that
the heat release curve is represented authentically. The values oI weight Iraction vapor
should be Ied at these temperatures.
Even Ior services where only one stream has a nonlinear heat release proIile, it is
beneIicial to plot both the proIiles together, as useIul insight may be obtained Irom such a
composite plot, e.g., the existence oI a 'pinch¨ or a 'near-pinch.¨ Impossible situations with
the cold stream getting hotter than the hot stream at an intermediate temperature point can
also be detected and avoided without wasting any time in thermal design.
In the event oI a temperature cross (outlet temperature oI cold stream greater than outlet
temperature oI hot stream), the number oI shells in series can be determined Irom a
composite heat release plot. This method is described in detail in Section 6.2.1.
For a Ieed-eIIluent exchanger, or any other service where both streams have a nonlinear
plot, the number oI shells required in series can still be determined Irom a composite heat
release plot. In tight situations, the next higher number oI shells in series should be adopted,
as this oIIers considerable Ilexibility Ior cases oI operation other than design. For example, iI
the number oI shells Irom such a plot is determined to be just three, it is advisable to employ
Iour shells in series. Besides oIIering Ilexibility Ior alternate cases oI operation, it will also
permit the use oI identical shells with a nominal overdesign margin in each. This may not be
possible iI only three shells are used in seriesto employ identical shells, some shells may
have a very high overdesign. This is because no transIer oI duty is possible Irom one shell to
another as the 'equal outlet temperature¨ condition prevails in all the individual shells.
Sometimes, a heat release plot is Iurnished in a tabular Iashion. Since it is not possible to
assess the curvature oI a plot Irom tabulated heat duty versus temperature data, it is strongly
recommended that a curve be plotted Iirst, either manually or by using a computer. Once this
curve is plotted, it is relatively simple to decide as to how many points should be Ied Ior a
proper representation. The same methodology described above should then be Iollowed.
References
|1| Gallant, R.W., and Railey, J.M., (1984), Physical Properties of Hyarocarbons, Vols. 1 and 2,
2nd Edition, McGraw-Hill.
|2| Dean, J.A., (ed.), Lange´s Hanabook of Chemistry, Eaition 14, McGraw-Hill.
|3| Green. D.W., (ed.), (1997), Perrys Chemical Engineers Hanabook, 7th Eaition, McGraw-
Hill.

Fig. 9.3 How to Ieed heat release proIiles
163
CHAPTER 10
2YHUGHVLJQ
10.1 Mechanics of Overdesign
As the name implies, overdesign is the extra margin incorporated in a design. Unless
otherwise speciIied, it implies the margin on the heat transIer surIace. Thus, iI a heat
exchanger service requires 1076 It
2
(100 m
2
) heat transIer area, and iI 1291 It
2
(120 m
2
) is
provided, we say that the overdesign is 20°. Designers preIer to incorporate an
overdesign as an insurance towards uncertainties which could be in Ilow rates,
temperatures, physical properties, and even the soItware itselI.
Evidently, iI a heat exchanger is relatively small, a 10° or even a 20° overdesign may
not be objectionable because it is the extra absolute cost and not the percentage that is
signiIicant. Besides, a larger margin may be required to be retained in cases where going to
the next lower pipe size results in underdesign or very little overdesign. For larger heat
exchangers, a smaller overdesign such as a 5° margin may be more prudent. Evidently, the
higher the accuracy oI the process data and the higher the conIidence level oI the design
method (soItware), the lower the margin oI overdesign that needs be incorporated.
For very large heat exchangers, a smaller margin may be retained (down to 5°) Ior
reasons oI economy. Thus, Ior heat exchangers up to 2152 It
2
(200 m
2
), a 10° margin may
be retained which may be reduced progressively to 5° Ior 43044842 It
2
(400450 m
2
) and
beyond. It will be evident that these values are oI indicative nature only.
Clients occasionally insist on standardization oI heat exchanger design Ior the heat
exchangers in a plant or complex, evidently with the intention oI inventory control. For
example, iI spare bundles are required and several designs are identical, a single spare
bundle will be adequate Ior all oI them since cleaning or other maintenance oI the several
exchangers can be carried out progressively. Considering the various construction
parameters oI shell-and-tube heat exchangers, standardization oI designs is a rather diIIicult
task. The designer may be compelled to compromise with rather high overdesign Ior some
oI the heat exchangers.
It has been mentioned earlier that up to 18 in. (450 mm) nominal size, pipe shells are
employed. II a standard tube OD and length have to be used, it may so happen that a given
pipe size (say, 14 in. or 350 mm NB) shell yields a rather high overdesign margin. However,
iI the shell size is lowered to the next size (12 in. or 300 mm, in this case), the design
becomes undersized. Since this is not acceptable, the designer is compelled to accept the
large overdesign oI the 14-in. or 350-mm shell diameter design. One may argue that even
with the larger shell, only the requisite number oI tubes may be incorporated so that the
overdesign is not unnecessarily high. This logic is certainly sound, as tubes may cost up to
30° oI a heat exchanger. However, care should be taken to ensure a proper tube layout and,
thereby, a proper stream distribution on the shellside. What this means is that there should
164
not be a large gap between the shell ID and the outer tube limit because this will result in a
large C stream. II there is a large gap, sealing strips should be employed to minimize the
shell-bundle leakage stream. However, it would be Iar better iI the gap between the shell and
the outer tube limit is maintained at the normal value Ior the given type and diameter oI the
heat exchanger, and tubes eliminated Irom the top and the bottom rows oI the shell,
assuming a horizontal baIIle cut and up-and-over Ilow.
10.2 Overdesign in Single-Phase Heat Exchangers
It will be easily appreciated that an overdesigned heat exchanger will deliver more than
the design heat duty. In the case oI sensible heat transIer, the cold stream will get heated
more and the hot stream will get cooled more. For the Iinal conditions oI terminal
temperatures and heat duty, the overdesign will obviously have to be nil.
The increased heat duty that an overdesigned heat exchanger can perIorm, expressed as
a percentage oI the design heat duty, is called overdesign on perIormance. Thus, iI a stream
has to be cooled by 90ƒF (50ƒC) but actually gets cooled by 99ƒF (55ƒC), it has a 10°
perIormance overdesign.
The value oI overdesign on perIormance is always less than the value oI overdesign on
surIace. This is because, as the heat duty increases in an overdesigned heat exchanger, the
MTD will decrease.
While overdesign on surIace indicates the percentage oI tubes that can be plugged in
case oI leaking (provided, oI course, there is a corresponding margin in the tubeside pressure
drop), overdesign on perIormance indicates how much better an exchanger may be expected
to perIorm. Heat exchanger services are usually analyzed and interpreted in the Iouled
Table 10.1a: Principal process parameters Ior Case Study 10.1: Low temperature approach case
Shellside Tubeside
1. Stream
Liquid hydrocarbon Liquid hydrocarbon
2. Flow rate, lb/h (kg/h)
672,400 (305,000) 749,600 (340,000)
3. Temperature in/out, ƒF (ƒC)
415.4 (213)/381.2 (194) 219.2 (104)/257 (125)
4. Heat duty, MM Btu/h (MM kcal/h)
16.03 (4.04)
5. Allowable pr. drop, psi (kg/cm
2
)
10 (0.7) 10 (0.7)
6. Fouling resistance, h It
2
ƒF/Btu
(h m
2
ƒC /kcal)
0.00146 (0.0003) 0.00293 (0.0006)
7. Density in/out, lb/It
3
(kg/m
3
)
39.9 (640)/40.8 (654) 49.4 (792)/48.05 (770)
8. Viscosity in/out, cp
0.22/0.31 1.74/1.4
9. Thermal conductivity,
Btu/h It ƒF (kcal/h m
2
ƒC )
0.058 (0.086)/0.059 (0.088) 0.0685 (0.102)/0.0645
(0.096)
10. SpeciIic heat, Btu/lb ƒF (kcal/kg ƒC)
0.705/0.695 0.58/0.56
11. Connection size, in. (mm) (nominal)
12 (300) 12 (300)
12. Design pressure, psig (kg/cm
2
g)
192 (13.5) 427 (30)
13. Design temperature, ƒC
455 (235) 284 (140)
14. Material oI construction
Carbon steel Carbon steel
165
condition. However, when a heat exchanger is new and, thereby, cleanor aIter it has been
cleaned when used Ior some length oI timeits overall heat transIer coeIIicient is much
higher than that in the Iouled condition, so that it is oversurIaced, over and above the
overdesign the designer has incorporated. Evidently, the extent oI over-perIormance will
depend upon the extent to which the total Iouling resistance is controlling the heat transIer
process. Consequently, a new or cleaned heat exchanger will deliver a heat duty higher than
the design value.
The Iollowing two case studies are presented to illustrate the Ioregoing.
CASE STUDY 10.1: EFFECT OF OVERDESIGN-
HIGH-TEMPERATURE APPROACH CASE
Consider the case oI a heat exchanger having the principal process parameters speciIied
in Table 10.1a.
A thermal design was made Ior this service and the principal construction and
perIormance parameters are detailed in Table 10.1b. The overdesign on surIace is 20.2°,
that is, the heat transIer area provided (1382 It
2
or 128.4 m
2
) is 20.2° more than that
required to deliver the speciIied heat duty.
Table 10.1b: Principal construction and perIormance parameters
Ior Case Study 10.1: Low temperature approach case
1. Shell ID, in. (mm)
28.5 (725)
2. No. oI tubes
278
3. Tube OD × thickness × length, in. (mm)
0.984 (25) × 0.0984 (2.5) × 236 (6000)
4. No. oI tube passes
2
5. Tube pitch, in. (mm)
1.26 (32) rotated square
6. Type oI baIIles
single-segmental
7. BaIIle spacing, in. (mm)
15.16 (385)
8. BaIIle cut, °
35
9. No. oI shells
One
10. Heat transIer area, It
2
(m
2
)
1382 (128.4)
Velocity, It/s (m/s)
4.07 (1.24)
Pressure drop, psi (kg/cm
2
)
6.0 (0.42)

11. Shellside
Heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
403.5 (1970)
Velocity, It/s (m/s)
9.0 (2.75)
Pressure drop, psi (kg/cm
2
)
9.0 (0.63)

12. Tubeside
Heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
293.3 (1432)
13. MTD,

ƒC
158.9 (88.3)
14. Overall heat transIer coeIIicient, kcal/h m
2
ƒC
88.2 (430.5)
15. Overdesign, °
20.8
166
Since the exchanger is oversurIaced, the hot stream will cool somewhat more and the
cold stream will get heated somewhat more until the required and provided heat transIer
areas match. A 'perIormance run¨ was taken Ior this exchanger to ascertain its actual
expected operating perIormance. Table 10.1c gives a comparative statement oI the design
and expected duties.
It is seen that the exchanger will actually deliver a heat duty oI 18.62 MM Btu/h (4.692
MM kcal/h) instead oI the design value oI 16.11 MM Btu/h (4.06 MM kcal/h). ThereIore,
the overdesign on perIormance is (4.692/4.06), or 1.156, or 15.6°. What has happened is
that the heat duty has increased and the MTD decreased so that the heat transIer area
required is equal to that provided. The overall heat transIer coeIIicient has virtually remained
the same, but this may change as well, depending principally upon the change in temperature
and, thereby, the physical properties, principally viscosity.
It will be seen Irom Table 10.1c that, while there is no change in the overall heat transIer
coeIIicient, the MTD is lower by 3.85° and the heat duty is higher by 15.6°. Since the heat
transIer required is directly proportional to the heat duty and inversely proportional to the
MTD, the heat transIer area required has gone up by |1.156/(1 ² .0385)| 20.2°, which is
precisely the overdesign on surIace in the base design.
In this case, the diIIerence between the overdesign on surIace (20.2°) and the
overdesign on perIormance (15.6°) is unusually low. This is because the hot and cold
stream outlet temperatures are rather wide apart, so that there is considerable scope Ior more
heating (oI the cold stream) and cooling (oI the hot stream). Usually, however, the diIIerence
between the hot stream outlet temperature and the cold stream outlet temperature is lower, so
that an increase in heat duty is accompanied by a sharp Iall in the MTD. Consequently, the
overdesign in perIormance is much lower than the overdesign on surIace. Typically, a 10°
overdesign on surIace translates to a 34° overdesign on perIormance
CASE STUDY 10.2: EFFECT OF OVERDESIGN-
LOW-TEMPERATURE APPROACH CASE
Table 10.2a elaborates the principal parameters oI a liquid-liquid heat exchanger. It will
be noticed that there is a temperature cross which will require the use oI two shells in se-
Table 10.1c: Comparative statement oI design and expected perIormance duties
Ior Case Study 10.1: Low temperature approach case
For design duty Expected
performance
Percent
change
Inlet temperature, ƒF (ƒC) 415.4 (213) 415.4 (213)
Hot stream
Outlet temperature, ƒF (ƒC) 381.2 (194) 375/8 (191)
Inlet temperature, ƒF (ƒC) 219.2 (104) 219.2 (104)
Cold stream
Outlet temperature, ƒF (ƒC) 257 (125) 262.9 (128.3)
Heat duty, MM Btu/h (MM kcal/h) 16.03 (4.04) 18.62 (4.692) 15.6
MTD, ƒF (ƒC) 158.9 (88.3) 184.8 (84.9) 3.85
Overall heat transIer coeIIicient, kcal/h m
2
ƒC 88.2 (430.5) 88.2 (430.6) Nil
Overdesign 1.202 Nil
167
ries. A thermal design was prepared Ior this service and the principal construction and
perIormance are presented in Table 10.2b. The overdesign on surIace is 20.14°.
A 'perIormance run¨ was taken next, and the expected perIormance oI the heat
exchanger is reported in Table 10.2c. It will be seen that the heat duty is 35.79 MM Btu/h
(9.017 MM kcal/h), as against the design heat duty oI 33.22 MM Btu/h (8.371 MM kcal/h).
The overdesign on perIormance is, thereIore, 9.016/8.371 or 7.7°. Thus, the overdesign on
perIormance is considerably less than that on surIace.
It will be seen Irom Table 10.2c that, while there is no change in the overall heat transIer
coeIIicient, the MTD is lower by 10.4° and the heat duty is higher by 7.7°. Since the heat
transIer area required is directly proportional to the heat duty and inversely proportional to
the MTD, the heat transIer area required has gone up by |1.077/(1 ² 0.104)| or 1.202, which
is precisely the overdesign on surIace in the base design, 20.2°.
10.3 Overdesign in Reboilers
One area where this could lead to a problem is a kettle or thermosyphon reboiler. Unlike
sensible heating/cooling or condensing, boiling heat transIer coeIIicient is a strong
Iunction oI the temperature diIIerence across the boiling Iilm. When a reboiler is clean, it
will have a much higher temperature diIIerence across the boiling Iilm than when it is
Iouled (and designed Ior). As a result, it may easily happen that, instead oI nucleate
boiling as per the design in the Iouled condition, a reboiler may experience Iilm boiling,
which is much less eIIicient than nucleate boiling. What would be even worse is iI a part
oI the tubes would experience nucleate boiling and the rest Iilm boiling. This will lead to
unstable perIormance and diIIiculty in control. Consequently, thermosyphon reboilers are
generally started up gradually: iI the heating medium is steam, its pressure is increased
Table 10.2a: Principal process parameters Ior Case Study 10.2: High temperature approach case
Shellside Tubeside
1. Stream
Liquid hydrocarbon Liquid hydrocarbon
2. Flow rate, lb/h (kg/h)
654,300 (296,800) 837,750 (380,000)
3. Temperature in/out, ƒF (ƒC)
381.2 (194)/311 (155) 255.2 (124)/320 (160)
4. Heat duty, MM Btu/h (MM kcal/h)
29.89 (7.53)
5. Allowable pr. drop, psi (kg/cm
2
)
14.2 (1.0) 17 (1.2)
6. Fouling resistance, h It
2
ƒF/Btu
(h m
2
ƒC /kcal)
0.00195 (0.0004) 0.00293 (0.0006)
7. Density in/out, lb/It
3
(kg/m
3
)
40.6 (650)/42.4 (680) 50.5(810)/49.3 (790)
8. Viscosity in/out, cp
0.3/0.4 1.5/0.9
9. Thermal conductivity, Btu/h It ƒF (kcal/h
m
2
ƒC)
0.052 (0.077)/0.055 (0.082) 0.0524 (0.078)/0.049 (0.073)
10. SpeciIic heat, Btu/lb
o
F (kcal/kg ƒC)
0.65 (average) 0.55 (average)
11. Connection size, in. (mm) (nominal)
14 (350) 12 (300)
12. Material oI construction
carbon steel carbon steel
168
gradually, as the reboiler gets Iouled progressively. This has been discussed in detail in
Chapter 8.
Since boiling heat transIer coeIIicient is dependent upon temperature diIIerence across
the boiling Iilm, the diIIerential Iouling resistance will alter the boiling heat transIer
coeIIicient. Hence, care should be taken to limit the diIIerential Iouling resistance to a
reasonable value so that the boiling heat transIer coeIIicient and, thereIore, the thermal
design are authentic. For this purpose, it is suggested that the overdesign (excess area) be
maintained at around 10°.
10.4 Overdesign in Condensers
In the case oI total condensers, overdesign on surIace will translate into subcooling. Since
subcooling is rather ineIIicient, as compared to condensing in an integral shell, the
increase in heat duty will be very small. However, in the case oI partial condensers, there
could be an appreciable increase in heat duty due to overdesign.
The most important parameter which controls overdesign on perIormance is the
temperature proIile oI the hot and cold streams. As explained in an earlier section on
Table 10.2b: Principal construction and perIormance parameters
Ior Case Study 10.2: High temperature approach case
1. Shell ID, in. (mm)
49.2 (1250)
2. No. oI tubes
958
3. Tube OD × thickness × length, in. (mm)
0.984 (25) × 0.0984 (2.5) × 236 (6000)
4. No. oI tube passes
4
5. Tube layout angle, degrees
90
6. Type oI baIIles
single-segmental
7. BaIIle spacing, in. (mm)
15.75 (400)
8. BaIIle cut, °
20
9. No. oI shells
2 (in series)
10. Heat transIer area, It
2
(m
2
)
2 × 4670 (434) ÷ 9340 (868)
11. MTD, ƒF (ƒC)
54.9 (30.5)
Velocity, It/s (m/s) 2.43 (0.74)
Pressure drop, psi (kg/cm
2
) 6.1 (0.43)

12. Shellside
Heat transIer coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
256.4 (1252)
Velocity, It/s (m/s) 5.74 (1.75)
Pressure drop, psi (kg/cm
2
) 16.1 (1.13)

13. Tubeside
Heat transIer coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
192.9 (942)
14. Overall heat transIer coeIIicient, Btu/h It
2
ƒF (kcal/h m
2
ƒC)
69.4 (338.8)
15. Overdesign, °
19.9
169
multiple shells, a single shell cannot deliver a temperature cross perIormance. Thus, iI the
outlet temperatures oI the two Iluids in a single shell are equal, there is no possibility oI any
increase in the heat duty, no matter how high the overdesign on surIace. The smaller the
diIIerence between the outlet temperatures oI the two streams, the smaller the scope Ior
increase in heat duty due to oversurIacing. Case Study 10.1, shown above, had a very large
diIIerence in the two temperatures: the hot stream design outlet temperature was 381.2ƒF
(194ƒC), while the cold stream outlet temperature was 257ƒF (125ƒC). Consequently, the
overdesign on perIormance was rather high.
10.5 Overdesign Factor
An overdesign Iactor (margin) is oIten speciIied Ior the design oI heat exchangers Ior one
or more oI the Iollowing reasons:
a) Iuture increase in capacity
b) plant control Ilexibility
c) upset conditions
a) alternate Ieedstocks
e) uncertainty in results oI process simulation
This overdesign Iactor may be on heat transIer surIace only (excess area) when the
thermal design will have to be perIormed on the speciIied Ilow rates. Usually, however,
the overdesign Iactor is speciIied Ior both Ilow rates and heat duty. Since a heat
exchanger can be designed only Ior a consistent set oI parameters, this overdesign Iactor
should be the same Ior shellside Ilow rate, tubeside Ilow rate, and the heat duty.
Sometimes, a process data sheet speciIies diIIerent Iactors Ior Ilow rate and heat duty.
For example, the shellside Ilow Iactor may be 120°, the tubeside Ilow Iactor 110°, and the
heat duty 115°. In such cases, the matter should be reconciled with the process licensor so
that the same Iactor is speciIied Ior all the three parameters. II necessary, the licensor may
speciIy an additional multiplier Ior Ilow rates Ior pressure drop purposes only iI the same is
expected to go up under certain conditions.
In some instances, the process licensor may speciIy diIIerent multipliers Ior the shellside
Table 10.2c: Comparative statement oI design and expected perIormance duties
Ior Case Study 10.2: High temperature approach case
For design
duty
Expected
performance
Percent
change
Inlet temperature, ƒF (ƒC) 381.2 (194) 381.2 (194)
Hot stream
Outlet temperature, ƒF (ƒC) 311 (155) 306.1 (152.3)
Inlet temperature, ƒF (ƒC) 255.2 (124) 255.2 (124)
Cold stream
Outlet temperature, ƒF (ƒC) 320 (160) 324.5 (162.5)
Heat duty, MM Btu/h (MM kcal/h) 29.89 (7.53) 31.95 (8.05) 6.9
MTD, ƒF (ƒC) 54.9 (30.5) 49.1 (27.3) 11.7
Overall heat transIer coeIIicient, kcal/h m
2
ƒC 69.4 (338.8) 69.5 (339.3) negligible
Overdesign 19.9 Negligible
170
and tubeside Ilow rates Ior an alternate case oI operation, and request that Ior the thermal
design (geometry) Iinalized, the heat exchanger designer indicate the outlet temperatures oI
both streams Ior certain speciIied inlet temperatures. This situation is acceptable since it is
realistic and there is no inconsistency. An alternate perIormance run need only be taken to
indicate the desired data.
II the process engineer speciIies overdesign on Ilow and duty, and also speciIies excess
area, both should be complied with. The excess area required could be towards uncertainties
in simulation, whereas the margin on Ilow rates could be towards Iuture increase in plant
capacity.
In order to maintain a satisIactory velocity inside the tubes, the overdesign oI a heat
exchanger may sometimes be on the higher side since reduction in the number oI tubes (to
reduce overdesign) may increase the tubeside pressure drop beyond the allowable limit. This
is acceptable.
10.6 Tube Plugging
Although not usually speciIied Ior the same, and oIten not realized directly, an
overdesign margin is oIten useIul in that it permits a certain extent oI tube plugging
beIore a heat exchanger has to be taken out Ior repair. Evidently, the throughput oI the
tubeside Iluid can be sustained only iI there is suIIicient margin in tubeside pressure drop.
Thus, even iI only 5° oI the tubes in a heat exchanger are plugged, the tubeside pressure
drop will increase by about 10°. We all know that, despite the best practices in material
selection and exchanger design/Iabrication/testing, tubes do Iail occasionally in real liIe.
ThereIore, it is very useIul iI the permitted tubeside pressure drop is not utilized
completely, but a small margin leIt unutilized.

171
CHAPTER 11
)RXOLQJ,WV&RQVHTXHQFHV
DQG0LWLJDWLRQ
Fouling may be deIined as the deposition oI undesirable matter on a heat transIer surIace
and is an inevitable consequence oI the process oI heat transIer between two streams
across a metal wall. Taborek called it the 'major unresolved problem in heat transIer¨
way back in 1972 |1|, and the situation does not appear to have changed signiIicantly.
Since Iouling has a direct impact on the energy eIIiciency oI a heat exchanger, on both
heat transIer and pressure drop, it is essential Ior the designer to be Iully aware oI the
phenomenon, its consequences, and its mitigation.
Evidently, Iouling can take place both inside and outside tubes, depending upon the
nature oI the Iluids Ilowing. The deposition oI the Ioulant results in:
1) a reduction in the overall heat transIer coeIIicient due to the extra resistance to
heat transIer, thereby resulting in a larger heat transIer area
2) a reduction in the area oI the Ilow passages, resulting in increased pressure drop
oI the Ilowing streams
The extent oI Iouling varies markedly with the nature oI the Iluids being handled.
Consequently, exchangers that handle clean Iluids may remain largely Iree oI Iouling,
whereas exchangers that handle dirty streams may be constantly aIIlicted by it.
A Iact that is usually not recognized is that the increase in pressure drop is usually a
more serious consequence oI Iouling than the increase in the thermal resistance and, thereby,
the reduction in the heat transIer coeIIicient. When a heat exchanger is taken out oI service
Ior cleaning, it is invariably due to the reduced throughput as a result oI partial blockage oI
Ilow areas, rather than reduced thermal perIormance.
The adverse eIIects oI Iouling are:
1) Increased capital cost due to the reduced overall heat transIer coeIIicient
2) Additional energy requirement to make up Ior the loss in perIormance
3) Maintenance costs Ior anti-Ioulant, chemical treatment, and cleaning oI Iouled
surIaces
4) Downtime cost associated with the outage oI the heat exchanger Ior cleaning
5) Reduction in throughput
In order to vividly demonstrate the extent oI Iouling and its repercussion in a heat ex-
changer, Table 11.1 presents the percentage oI the combined shellside and tubeside Ioul-
ing resistance in the 25 heat exchangers in an oil reIinery crude preheat train. Here, crude
oil is progressively heated by the products and circulating reIlux streams Irom the atmos-
172
pheric and vacuum distillation columns. For a better representation, the same data is
presented graphically in Fig. 11.1. It will be seen that the lowest combined (crude and
heating medium) Iouling resistance, as a percentage oI the total resistance to heat
transIer, is 23.1°; the highest is 49.73°. The average works out to 31.03°. Thus, oI the
total 9052 m
2
oI heat transIer area provided Ior all oI the 25 heat exchangers, 2809 m
2

had to be provided to account Ior Iouling alone!
In a shell-and-tube heat exchanger, there are Iive resistances to heat transIer, namely:
shellside Iilm, shellside Iouling, tubeside Iilm, tubeside Iouling, and tube wall. In order to
appreciate the percentages oI these resistances and their variation, Fig. 11.2 shows these
values Ior a typically dirty service (both shellside and tubeside) and a typically clean service
(both shellside and tubeside).
Table 11.1: Fouling in a crude preheat train
Item no.
Crude oil fouling resistance
as percent of total resistance
Heating medium fouling
resistance as percent of total
resistance
Total fouling resistance
as percent of total
resistance
1. E-1 20.58 12.86 33.44
2. E-2 18.06 15.05 33.11
3. E-3 22.68 14.18 36.86
4. E-4 17.25 13.8 31.05
5. E-5 21.24 13.28 34.52
6. E-6 19.74 16.45 36.19
7. E-7 30.6 19.13 49.73
8. E-8 17.85 9.52 27.37
9. E-9 21.98 8.79 30.77
10. E-10 14.4 11.52 25.92
11. E-11 21.38 17.1 38.48
12. E-12 9.3 24.8 34.1
13. E-13 20.22 20.22 40.44
14. E-14 15.48 19.35 34.83
15. E-15 16.8 6.3 23.1
16. E-16 28.88 11.55 40.43
17. E-17 17.55 9.36 26.91
18. E-18 15.3 10.2 25.5
19. E-19 25.88 10.35 36.23
20. E-20 17.33 13.86 31.19
21. E-21 10.13 27.0 37.13
22. E-22 22.08 14.72 36.8
23. E-23 19.13 15.3 34.43
24. E-24 8.03 21.4 29.43
173
11.1 Categories of Fouling
There are six principal categories oI Iouling:
1) Precipitation Iouling (scaling) is the precipitation oI dissolved substances on the
heat transIer surIace. II the dissolved solutes have inverse solubility charac-
teristics, the precipitation will take place on superheated surIaces. Precipitation
Iouling is oIten reIerred to as scaling, e.g., calcium sulIate in water.
2) Particulate Iouling is the accumulation oI particles suspended in a Iluid on the
heat transIer surIace. It is caused by gravity in some applications, where it is
reIerred to as sedimentation Iouling. A common example oI sedimentation
Iouling is that oI rust particles in cooling water.
3) Corrosion Iouling occurs when the heat transIer surIace, itselI, reacts to produce
corrosion products which then Ioul the heat transIer surIace. Rusting oI carbon
steel is the most common example oI corrosion Iouling. As the heat transIer
surIace becomes rougher due to corrosion Iouling, it produces nucleation sites Ior
precipitation and particulate Iouling.
4) Chemical reaction Iouling is the Iormation oI deposits by chemical reaction
among the diIIerent constituents oI the Ilowing stream. The surIace material,
itselI, does not enter into reaction. Common examples are polymerization,
coking, and cracking oI hydrocarbons.
5) Bio-Iouling is the attachment oI biological micro- or macro-organisms to the heat
transIer surIace. A common example is Iouling by barnacles and algae in sea
water services.
6) SolidiIication Iouling is
the solidiIication oI a
pure liquid or particular
constituents oI a liquid
solution on a subcooled
heat transIer surIace. A
common example is ice
Iormation at sub-zero
temperatures.
In a given heat exchanger
application, depending upon the
situation, one or more oI the
above modes oI Iouling may
occur in conjunction. This is
what makes Iouling a very com-
plex and unpredictable pheno-
menon.
There are several parameters
which aIIect the degree oI the
various types oI Iouling.
11.2 Progress of Fouling
The buildup oI Iouling on a heat
transIer surIace is ideally repre-
Fig. 11.1 Heat exchangers in a crude preheat train Iouling
resistance as a percentage oI total resistance to heat transIer

Fig. 11.2 Contribution oI Iouling towards overall heat transIer
coeIIicient typical scenarios
174
sented by an asymptotic curve as shown in Fig. 11.3. There are three distinct regions in
this curve:
a) an initial period 01 where there is very little evidence oI any Iouling and may be
considered to be an initiation period which may range Irom a Iew hours to
several weeks. This will depend upon several parameters which are discussed in
Section 11.3.
b) Period 12 wherein a steady increase in Iouling deposition is observed. This
period could vary Irom a Iew hours to several months.
c) Period 23 where the rate oI increase oI Iouling resistance decreases Irom that in
the period 12 and the curve attains a plateau or asymptotic Iouling resistance
value
In real liIe, however, a Iouling curve may vary considerably Irom this idealized curve so
that in some cases there may be no initiation period, and in others the increase in Iouling
resistance may be virtually linear. Whatever the nature oI a Iouling curve, there will come
a time, with most Iluids, when the thermal and hydraulic perIormance oI a given heat
exchanger will deteriorate to such an extent that the heat transIer surIace will have to be
cleaned to restore the original (or close to original) perIormance.
Evidently, the higher the Iouling resistance considered Ior a particular stream, the longer
will a heat exchanger be able to operate beIore being subjected to a shutdown. However, the
higher Iouling resistance will also mean a costlier heat exchanger. ThereIore, the selection oI
a design Iouling resistance will have to be made on the basis oI optimization oI the total cost
(Iixed plus operating cost).
There is another Iactor that has to be considered in this context, and that is the normal
turnaround oI a plant. All plants have to be shut down periodically Ior inspection and over-
hauling oI equipment, piping, instrumentation, etc. This period usually varies Irom 12 to 36
months. Thus, it is a normal practice to consider this normal turnaround period Ior the
selection oI design Iouling resistances oI all Ilowing streams so that the heat exchangers can
be cleaned in the same period.
11.3 Parameters That Affect Fouling
The principal parameters are:
1) The nature oI the Ilowing Iluidwhether clean or dirty. This is the starting point
Ior all discussions on Iouling, and it is only Ior dirty streams that all the subse-
quent considerations assume importance. For heat exchangers handling clean
streams, such as steam and very
light hydrocarbons, Iouling is not a
problem at all. On the other hand,
heavy hydrocarbon streams, such as
long residue or vacuum gas oils,
Ioul readily and cause considerable
deterioration in perIormance.
2) Flow velocity and temperature
are perhaps the most crucial vari-
ables that control the Iouling pro-
cess. A high velocity minimizes vir-
tually all types oI Iouling. Fouling is
a dynamic process wherein deposi-
Fig. 11.3 An idealized Iouling curve
175
tion and removal oI Ioulant occur simultaneously. The net Iouling at any given
instance represents the equilibrium balance between these two opposing
phenomena.
With increase in velocity, the viscous sub-layer close to the tube wall be-
comes thinner, thereby resulting in a reduction in the resistance to diIIusion Irom
the bulk to the wall. This permits a comparatively higher rate oI deposition Ior a
diIIusion-controlled Iouling process. However, at the same time, the higher vel-
ocity increases the shearing Iorces that tend to remove the Iouled deposit. The net
rate oI Iouling will, thereIore, depend upon these two opposing eIIects oI velo-
city. Usually, the rate oI decrease oI Iouling due to the increase in the shear Iorce
is greater than the rate oI increase oI deposition due to the reduction oI the vis-
cous sub-layer. Consequently, higher velocities invariably result in less Iouling.
The general nature oI the degree oI Iouling versus Ilow velocity is
represented in Fig. 11.4.
3) The temperature oI the Iluid-Ioulant interIace strongly inIluences the extent oI
various modes oI Iouling. Bulk Iluid temperatures and their heat transIer
coeIIicients, as well as the Iouling and tube wall resistances, will determine this
interIace temperature. When normal-solubility salt solutions are heated, they do
not exhibit Iouling since the solubility increases at higher temperatures.
However, cooling such solutions will obviously result in Iouling problems.
The reverse is true Ior inverse-solubility salts such as calcium sulIate,
calcium carbonate, and magnesium hydroxide. II the temperature at the surIace
increases above the solubility limit, the dissolved salts will come out oI solution
and deposit on the tube surIace. However, as Iouling increases, the interIace tem-
perature reduces as a result oI the extra thermal resistance oI the Ioulant;
consequently, the net rate oI Iouling reduces. Thus, an asymptotic curve results.
The rates oI chemical reaction are a strong Iunction oI temperature. Thus, iI a
Iouling process involves a chemical reaction, the extent oI Iouling will depend
upon temperature. The rate oI increase or decrease oI Iouling with time will be
related to the rate constant oI the chemical reaction itselI.
Corrosion is basically a chemical reaction. Consequently, the Iouling oI
surIaces by the products oI corrosion will be dependent upon temperature.
Bio-Iouling is also a strong Iunction oI temperature. At higher temperatures,
the rates oI chemical and enzyme reactions increase with a resultant increase in
the rate oI cell growth. However, iI the temperature increases even Iurther, some
heat-sensitive cell consti-
tuents may deactivate. For a
particular organism, there is
a temperature below which
reproduction and growth are
retarded and a temperature
above which the organism
ceases to thrive, and may
even be destroyed. Between
these two temperature levels
is a range where cell acti-
vity is the highest. This
temperature range may be
Fig. 11.4 Variation in Iouling resistance with Ilow velocity
176
rather large, Irom below Ireezing to about 176ƒF (80ƒC), depending upon the
organism in question.
4) Material oI construction and surIace Iinish ² The roughness, size, and density oI
cavities aIIect crystalline nucleation, sedimentation, and the sticking tendency oI
deposits. It is generally believed that very smooth surIaces are less likely to
receive and retain dirt than are rough surIaces. However, it may be argued that
this will be true only Ior the initial Iouling because thereaIter the surIace will no
longer be rough. However, the practical experience is that polished or smooth
surIaces Ioul signiIicantly less than rough ones.
In such a complicated scenario where there are various modes oI Iouling, as
well as several Iactors which determine the degree oI the various modes oI Ioul-
ing, it is evident that the prediction oI the extent oI Iouling is extremely diIIicult.
Consequently, it becomes unavoidable to adopt a qualitative approach and rely
on past experience Ior the selection oI Iouling resistances Ior various services.
TEMA (Tubular Exchanger ManuIacturers` Association) speciIies point and
range values oI various streams encountered in the chemical processing indus-
tries |2|. These values are Ior guidance, only, and should be modiIied based upon
actual operating Ieedback, wherever available. II no actual data is available, the
TEMA values should be adopted. The selection oI appropriate Iouling resistances
contributes signiIicantly in ensuring satisIactory operation oI heat exchangers.
11.4 How to Provide a Fouling Allowance
Many heat exchangers operate satisIactorily Ior several years without cleaning. Others
are constantly aIIlicted by Iouling. However, most heat exchangers experience some
Iouling so that it becomes necessary to provide suIIicient heat transIer area to enable
them to operate Ior a reasonably long period oI time (usually one or two years) beIore
requiring shutdown and cleaning.
The extra heat transIer area that must be provided to account Ior Iouling is usually
determined in one oI the Iollowing three ways:
1) Assign a Iouling resistance to each stream, that is, a resistance to heat transIer
caused by the Iouling layer deposited on the tube wall. This is oIten incorrectly
reIerred to as a 'Iouling Iactor.¨ A Iactor is something one multiplies with,
whereas a resistance is something one adds.
2) Apply an overall 'cleanliness Iactor¨ which is the ratio oI the Iouled overall heat
transIer coeIIicient to the clean overall heat transIer coeIIicient, usually between
60²90°. This practice was developed speciIically Ior steam surIace condensers
in the power industry where Iouling is predominantly on the cooling water side.
Since the cleanliness Iactor does not distinguish between Iouling on the two sides
oI the tube wall and since it is arbitrary, its use is not common in the chemical
process industries.
3) A third approach is to add a certain percentage oI additional heat transIer surIace
(usually 2030°) to that determined Irom the clean overall heat transIer coeIIi-
cient. This methodology also does not distinguish between shellside and tubeside
Iouling, but treats them together as a lump sum. Further, iI the number oI tubes is
increased to add heat transIer area, the tubeside heat transIer coeIIicient will de-
crease due to the lower velocity inside the tubes. The shellside heat transIer coeI-
Iicient will also decrease because oI the reduced velocity caused by the increased
177
shell diameter, unless compensated Ior by reduced baIIle spacing and/or baIIle
cut. One way out is to incorporate an 'overdesign¨ or 'excess area¨ oI the
speciIied amount (2030°) in the Iinal design geometry, based upon the actual
overall heat transIer coeIIicient.
Since there are innumerable heat exchanger services (combinations oI hot Iluid and cold
Iluid) possible in the chemical process industries, the logical and scientiIic methodology
to account Ior Iouling is to use the Iouling resistance method. In Iact, this is the method
that is Iollowed universally.
Apply fouling allowance for fouling only
It must be mentioned here that the application oI Iouling resistances should be aimed only
at making an allowance Ior anticipated Iouling. It should not cater to uncertainties in
design methodology (whether in process simulation or in the thermal design soItware),
the prediction oI physical properties, or Iuture increase in plant capacity. II such
possibilities exist, they should be catered to by speciIic allowances Ior each. For example,
a 10° margin may be retained Ior uncertainties in physical properties and/or a 10°
multiplier may be applied to Ilow rates and heat duty in view oI Iuture capacity
expansion. The advantage oI this methodology is that there will be a much lower
probability oI overspeciIying the 'overall¨ margin that is applied on the heat transIer
surIace, than there would be by applying a single 'overdesign¨ Iactor based on an
'overall¨ Ieel oI the situation to cater to all the uncertainties and/or requirements.
Do not overspecify fouling resistance
Designers oIten consider it prudent to apply conservative Iouling resistances so as to
obtain a longer turnaround period (period between successive cleanings). However,
unnecessarily high Iouling resistances may actually cause more harm than good because
oI the Iollowing:
1) The application oI unduly large Iouling resistances will result in large heat
exchangers, whereby it may not be possible to maintain suIIiciently high
velocities within the pressure drops permitted. For example, Ior the same
tubeside velocity, a larger number oI tubes will require a greater number oI tube
passes; this will result in a higher pressure drop. II the resultant pressure drop
exceeds the allowed limit, the number oI passes will have to be reduced. This
lower velocity will actually result in heavier Iouling!
On the shellside, a larger number oI tubes will result in a larger shell
diameter. In order to maintain the shellside velocity, the baIIles will have to be
brought closer, which again will result in a higher pressure drop. II this pressure
drop exceeds the allowable limit, the baIIle spacing will have to be increased;
this will result in a lower velocity and, thereby, greater Iouling. It is Ior this
reason that Iouling is oIten reIerred to as a 'selI-IulIilling prophecy.¨
There is another problem when baIIles are brought closer: the leakage and
bypass streams increase, thereby increasing the ineIIiciency oI the shellside Ilow.
(This was discussed in detail in Section 3.4.4.)
Thus, it is imperative Ior the designer never to compromise on Ilow velocity.
2) The application oI high Iouling resistances will also result in a large diIIerence
between the clean and the dirty overall heat transIer coeIIicients. Thus, such heat
exchangers will be highly oversurIaced when clean (at start-up), whence they
178
may be diIIicult to control. In this context, it is always a good idea to examine the
perIormance oI a heat exchanger under clean conditions to detect and address any
potential problems. This is particularly true oI reboilers and was discussed in
Chapter 8.
ThereIore, it is strongly recommended that only realistic Iouling resistances
be considered Ior heat exchanger design.
11.5 Selection of Fouling Resistance
As has been mentioned earlier, even aIter so many years oI research on Iouling, the heat
exchanger designer still has to invariably Iall back on the values oI Iouling resistances
Iurnished in the TEMA standards. This is not surprising, considering the variety oI Iluids
being handled in the chemical process industries, along with the variation in Iluid
velocities, wall temperatures, and materials oI construction.
Thus, it becomes important Ior plant operators to monitor Iouling Ior the various Iluids
handled. While this is a tedious task, it is well worth the eIIort in terms oI the realistic
Iouling resistance data that will emerge. One problem here is that, by monitoring a heat
exchanger perIormance in terms oI heat duty and terminal temperatures, it is possible to
determine only the overall Iouling resistancethat is, the sum oI the tubeside and the
shellside Iouling resistances. The breakup oI the overall Iouling resistance into the individual
Iouling resistances can be accomplished only by careIul analysis oI the combined Iouling
resistance data Ior various services, many oI which have common Iluids.
The TEMA Iouling resistance selection guide is not without its limitations. Apart Irom
crude oil and water, Iouling resistances oI other Iluids are Iurnished without considering
variations in the same due to:
1) side allocation ² shellside or tubeside
2) Ilow velocity
3) temperature level oI the Iluid, as well as the temperature level oI the other Iluid
4) material oI construction
Nevertheless, the TEMA values oI Iouling resistances are the only ones usually
available and most designers use TEMA values with a high level oI success!
Also, while the earlier TEMA editions Iurnished only point values oI Iouling resistance
oI various Iluids handled in the chemical process industries, the latest (1999) edition
Iurnishes range values Ior certain Iluids. TEMA values are evidently indicative and should
be used with discretion in the absence oI any other authentic in-house data. Where range
values are given, selection must be made on the basis oI speciIics.
For example, the Iouling resistance oI kerosene has been speciIied as 0.0020.003 h It
2

ƒF/Btu (0.000410.000614 h m
2
ƒC/kcal). This can be interpreted in the Iollowing manner.
For kerosene produced in the crude distillation unit oI an oil reIinery processing a light
crude, the Iouling resistance may be considered to be 0.002 h It
2
ƒF/Btu (0.00041 h m
2

ƒC/kcal), whereas Ior kerosene produced in the delayed coking unit oI a reIinery processing
a heavy crude, the Iouling resistance may be considered to be 0.003 h It
2
ƒF/Btu (0.000614 h
m
2
ƒC/kcal). The Iinal selection should also depend upon the velocity oI kerosene in the
given heat exchanger, as well as the temperature level oI the kerosene and the temperature oI
the heating/cooling medium.
It is interesting to note that in the 1999 edition oI the TEMA standards, the Iouling
resistance oI certain Iluids have been increased, while those oI certain other Iluids have been
179
decreased Irom the values in the last edition. Thus, the Iouling resistance oI compressed air
has been reduced Irom 0.002 to 0.001 h It
2
ƒF/Btu (Irom 0.00041 to 0.000205 h m
2
ƒC/kcal)
while that oI reduced crude has been increased Irom 0.005 to 0.007 h It
2
ƒF/Btu (Irom
0.00102 to 0.00143 h m
2
ƒC/kcal). Evidently, these revisions have been incorporated on the
basis oI Ieedback received Irom plant operators, as well as Irom a better understanding oI the
phenomenon oI Iouling.
To summarize, the selection oI Iouling resistance has to be done careIully and should be
based upon past experience. Values speciIied in the TEMA standards are Ior guidance only,
and should be tempered by operating Ieedback and engineering judgment. A proper
selection oI Iouling resistance will go a long way in ensuring the satisIactory operation oI
heat exchangers.
11.6 Design Guidelines to Minimize Fouling
Although it is diIIicult to determine Iouling resistances accurately, there are several
qualitative as well as quantitative rules oI thumb which should be Iollowed as they
minimize the extent oI Iouling:
11.6.1 Use heat exchanger types that foul less
Fouling is always more pronounced in dead spaces due to the lack oI turbulence. On the
shellside oI shell-and-tube heat exchangers, adjacent to each baIIle, there are dead spaces
which are inherently prone to Iouling. However, there are other types oI heat exchangers
(such as plate and spiral heat exchangers) in which there are no dead spaces and which,
thereIore, Ioul less. These exchanger types should, thereIore, be adopted wherever
Ieasible. UnIortunately, plate and spiral heat exchangers have a limitation in the design
pressure as well as design temperature. Fluidized-bed heat exchangers also minimize
Iouling inside the tubes by the scouring action oI small rough particles on the tube wall.
Many oI these heat exchangers are discussed in some detail in Chapter 13.
11.6.2 When shell-and-tube exchangers have to be used
Due to the limitations oI 'alternative¨ heat exchangers such as plate and spiral plate,
shell-and-tube heat exchangers still remain the most commonly used heat exchangers in
the chemical process industries. Let us, thereIore, now consider design strategies to
minimize Iouling in shell-and-tube heat exchangers.
For easier maintenance, the dirtier stream should be routed through the tubes oI a shell-
and-tube heat exchanger as it is much easier to clean the inside than the outside oI tubes.
However, iI the dirtier stream is viscous as well, placing it inside the tubes will result in a
poor heat transIer coeIIicient due to laminar Ilow.
A viscous stream Ilowing through the shellside yields a much higher heat transIer
coeIIicient due to the much higher turbulence. The higher heat transIer coeIIicient translates
into a smaller and cheaper heat exchanger. However, cleaning will be much more diIIicult.
Thus, there is a conIlict between the initial cost and the operating cost oI the exchanger.
Allocating the dirty viscous stream through the tubeside will result in a higher initial cost,
but a lower operating cost. The reverse is true iI this stream is routed through the shellside.
Thus, the overall cost (initial cost plus operating cost) will decide which allocation oI sides
should be adopted. Since the dirty stream may be on the shellside, on the tubeside, or on
both the sides, let us consider both the possibilitiesdirty Iluid inside tubes as well as
outside tubesand see what measures should be adopted to minimize Iouling.
180
11.6.2.1 Dirty fluid inside tubes
a) Use large diameter tubes
For a given heat exchanger service, a smaller tube diameter results in a smaller shell
diameter and, thereIore, a cheaper heat exchanger. However, very small diameter tubes
are extremely diIIicult to clean internally, especially Ior relatively long lengths.
Consequently,
3
/
4
-in. (20-mm) and 1-in. (25-mm) tubes are commonly used Ior heat
exchanger design. Smaller diameter tubes (
1
/
2
-in. or 12.7-mm OD and
5
/
8
-in. or 16-mm
OD) may be used Ior clean services. However, Ior dirty services (Iouling resistance
greater than 0.0004 h m
2
ƒC/kcal), it is recommended that 1-in. (25-mm) OD tubes be
used. This is because the increase in tubeside pressure drop due to the same degree oI
Iouling (say, a 0.02-in. or 0.5-mm thick deposit) will be higher with smaller diameter
tubes. Furthermore, smaller diameter tubes tend to get plugged much more readily than
larger diameter ones, and are more diIIicult to clean.
b) Maintain high velocity
A high velocity suppresses all types oI Iouling. For cooling water as well as other Iouling
liquids, a minimum velocity oI 3.3 It/s (1.0 m/s) is generally recommended, although 5
It/s (1.52 m/s) is preIerable. The worse the quality oI the liquid, the higher should be the
minimum velocity to restrict Iouling to an acceptable level.
Occasionally, a proper velocity cannot be maintained within the limitation oI allowable
pressure drop because a large number oI tube passes is required to achieve this velocity. In
such cases, either a higher pressure drop should be allowed or the tube velocity best
maximized by varying the tube diameter and length.
A problem usually arises Ior viscous and Iouling liquids (which are very common in oil
reIineries) where a very high pressure drop is required to sustain a satisIactory velocity. This
could go up to even 110140 psi or 810 kg/cm2 (with multiple shells in series) which
prima Iacie may appear to be unacceptable, but which should be examined Irom an overall-
cost-optimization point oI view. The standard values oI pressure drop permitted Ior various
heat exchanger services (typically 1.42.8 psi (0.10.2 kg/cm2) Ior gases and
vaporizing/condensing services, and 10 psi (0.7 kg/cm2) Ior liquid services) are based upon
such an optimization. It is only that 10 psi (0.7 kg/cm2) is not the proper value Ior viscous
liquids and 1442 psi (1.03.0 kg/cm2) per shell, depending upon the viscosity and the side
allocation, is more authentic. A viscous Iluid Ilowing on the tubeside requires a much higher
pressure drop Ior the same velocity than when it is Ilowing on the shellside.
It will be easily realized that the higher the velocity in the Iinal design, the higher will be
the heat transIer coeIIicient and the lower will be the incidence oI Iouling. It is, thereIore,
suggested that in no case should the velocity be allowed to Iall below 1.6 It/s (0.5 m/s). The
magnitude oI this problem can be Iully appreciated when it is considered that plants oIten
run in turndown conditions due to various constraints when the Ilow velocities are even
lower than the design values.
c) Allow suIIicient margin in pressure drop
When heavy Iouling is anticipated, it would be wise to leave suIIicient margin (say, 30
40 °, and even more) between the allowable and calculated values oI pressure drop. (A
more scientiIic approach is to consider an appropriate Iouling layer thickness as discussed
in the next section.) The idea is to permit an exchanger to operate at design load even
181
with the increased pressure drop due to Iouling. II the margin is not provided, the Ilow
rate will reduce in the Iouled condition such that the available pressure drop is not
exceeded, thereby limiting plant capacity.
a) Use Iouling layer thickness
Although heat exchanger designers employ a Iouling resistance to account Ior Iouling in
terms oI extra thermal resistance to heat transIer, they do not generally translate the
expected Iouling into a Iouling layer thickness to account Ior increased pressure drop.
Evidently, as a Iouling deposit builds up, it will reduce the Ilow area and, consequently,
result in an increased pressure drop. It is quite straightIorward to translate the Iouling
resistance into a Iouling layer thickness, iI the thermal conductivity oI the deposit is
known. UnIortunately, the thermal conductivities oI Iouling deposits Irom hydrocarbon
liquids are not readily available. However, the thermal conductivity oI asphalt is 0.43
Btu/h It ƒF (0.64 kcal/h m ƒC). Since Iouling layer thickness will be signiIicant only Ior
very Iouling Iluids, such as vacuum gas oils, Iuel oils, reduced crude (long residue), short
residue, and asphalt, a value oI 0.34 Btu/h It ƒF (0.5 kcal/h m ƒC) may be considered Ior
the deposits Irom heavy and dirty hydrocarbon liquids other than asphalt and the Iouling
resistance converted to the corresponding Iouling layer thickness. Thus a Iouling
resistance oI 0.01 h It
2
ƒF/Btu (0.002 h m
2
ƒC/kcal) will yield a Iouling layer thickness oI
0.0394 in. (1.0 mm), and so on.
Most sophisticated thermal design soItware packages have a provision Ior incorporating
a Iouling layer thickness. A case study is presented later, demonstrating the eIIect oI
applying Iouling layer thicknesses Ior both shellside and tubeside.
e) Use wire-Iin tube inserts
Wire-Iin tube inserts (Fig. 13.3, Ch. 13) were developed to increase heat transIer
coeIIicient Ior laminar Ilow inside tubes. However, it was discovered that the use oI these
inserts also resulted in a proIound reduction in Iouling |3, 4|. This is only to be expected,
considering that the principal action oI these inserts is to prevent boundary layer
separation and promote radial mixing Irom the tube wall to the center. The churning
action not only increases the heat transIer coeIIicient by increased convection, but also
minimizes the deposition oI Ioulants due to the extra turbulence. Wire-Iin inserts have
been demonstrated to reduce tubeside Iouling to a dramatic level Ior many dirty services
in reIineries, and appear to be an excellent application. However, Ior reasons unknown to
this author, their use has not really lived up to their potential.
For a detailed discussion on wire-Iin tube inserts including a case study, see Chapter 13.
f) Use twisted-tube heat exchangers
This is a recent innovation in shell-and-tube heat exchangers and is addressed in some
detail in Section 13.3.7.
g) Use on-line cleaning
Since prevention is better than cure, it would be better iI Iouling could be prevented in
the Iirst place. One such methodology is the use oI on-line cleaning in which brushes or
rubber balls are circulated through the tubes periodicallythe scouring action prevents
any buildup oI Iouling. On-line cleaning is particularly applicable to cooling water
Ilowing inside tubes in steam surIace condensers, as the cooling water Iouling resistance
is by Iar the predominant resistance to heat transIer in these applications.
182
Two systems are commonly in use: the brush cleaning system and the ball cleaning
system. In the brush cleaning system (Fig. 11.5a), a polypropylene brush is made to Ilow
along each tube periodically. Retaining cages (Fig. 11.5b) are inserted at each end oI every
tube. A computer-controlled Iour-way valve (Fig. 11.5c) reverses the Ilow direction
periodically at a prespeciIied interval. (The periodicity oI the operation depends upon the
Iouling nature oI the water; the dirtier the water, the more Irequently the Ilow direction will
be reversed.) When the Ilow reverses, the brushes travel Irom one end oI the tubes to the
other, thus removing the dirt or debris and maintaining tube cleanliness. Here, they are
retained in the cages until the next Ilow reversal makes them travel again to the other end.
The brushes usually consist oI nylon bristles held in place by titanium wires and nose
cones made oI special polypropylene. The brush diameter is slightly larger than the tube
internal diameter to create an interIerence Iit that ensures positive brushing action across the
entire internal surIace oI the tube. The retaining cages are made oI special polypropylene,
which is resistant to various types oI water.
In the ball cleaning system (Fig. 11.6), sponge rubber balls are circulated in a closed
loop and made to travel along the length oI the tubes. Since the balls are slightly larger in
diameter than the ID oI the tubes, they are compressed as they travel along the tubes. This
constant rubbing action keeps the tube walls clean and Iree oI any deposits. By selecting the
balls properly, a uniIorm distribution can be ensured so that all tubes are kept clean.
The balls are circulated in a closed loop and
are caught at the discharge end in a screen, aIter
which they are rerouted through a collector and
back to the condenser.
h) Use a spare tube bundle or even a spare
heat exchanger
For services where heavy Iouling is un-
avoidable, it is beneIicial to have a spare tube
bundle. Whenever the operating tube bundle
gets Iouled heavily and the heat exchanger
perIormance reduces beyond acceptable
limits, it is taken out and the spare tube bundle
inserted. The Iouled tube bundle is then
cleaned and kept ready Ior the next replace-
ment, and so on. However, Iixed-tubesheet

Fig. 11.5b Cages Ior retaining brushes (Courtesy
oI WSA Engineered Systems, Milwaukee, WI,
www.wsaes.com)

Fig. 11.5c Four-way valve Ior reversing Ilow
direction (Courtesy oI WSA Engineered Systems,
Milwaukee, WI, www.wsaes.com)

Fig. 11.5a Brush (Courtesy oI WSA Engineered
Systems, Milwaukee, WI, www.wsaes.com)

Fig. 11.6 On-line ball cleaning system (Courtesy oI
WSA Engineered Systems, Milwaukee, WI,
www.wsaes.com)
183
exchangers with heavy tubeside Iouling necessarily have to have a spare unit since their
tube bundles are not removable.
For extremely heavy and rapid Iouling applications such as polymerizing services (e.g.,
reboilers in a butadiene plant), a total spare unit is desirable. Here, the incidence oI Iouling is
so heavy that the tube bundle has to be cleaned very Irequently (such as, once every 24
weeks). Since the unit has to be shut Ior removing a tube bundle and replacing it, it becomes
advantageous to have a spare unit altogether.
11.6.2.2 Dirty fluid outside tubes
a) Use Iloating-head or U-tube design
This is a minimum requirement because Iixed-tubesheet exchangers cannot be cleaned on
the shellsideexcept by chemical means, which is not very convenient. II the tubeside
Iluid is clean, a U-tube construction may be used. However, iI the tubeside Iluid is also
dirty, a Iloating-head construction should be used.
b) Minimize dead spaces
Using a well-proportioned baIIle spacing-to-shell diameter ratio and a moderate baIIle
cut produces a Iavorable shellside stream analysis, thereby minimizing dead spaces
adjacent to each baIIle. This is depicted graphically in Fig. 11.7. As a rule oI thumb,
baIIle spacing should be between 35²60° oI the shell ID. The optimum baIIle cut is
25°, and it should never exceed 35°. A good baIIle design results in a good stream
analysis, that is, one with a high main crossIlow stream and low leakage and bypass
streams. Such a design will have minimum dead spaces.
c) Use larger tube pitch Ior very dirty services
Tube pitch is usually 1.25 times tube OD, as recommended in the TEMA standards.
However, iI the shellside Iluid is very dirty, it will be advantageous to employ a higher
tube pitch since the increase in shellside pressure drop with the progress oI Iouling will
be slower. However, care should be taken to maintain a Iairly high shellside velocity
because, otherwise, its very purpose will be deIeated. A satisIactory stream analysis must
also be maintained.
a) Maintain high velocity
As on the tubeside, a high velocity will minimize all categories oI Iouling. A minimum
velocity oI 1.5 It/s (0.46 m/s) is usually recommended Ior liquids. For low-Ilow-rate
streams on the shellside, the
velocity is oIten very low, even
with the smallest baIIle spacing and
the permitted pressure drop highly
unutilized. Since the shellside heat
transIer coeIIicient will be low, the
heat transIer area will be large.
Thus, not only will the heat
exchanger be costly, there will be
pronounced Iouling as well.
In such situations, it is usually
better to employ two (or even more)
shells in series, especially iI the
shellside heat transIer coeIIicient is

Moderate baffle cut and baffle spacing

Wide baffle spacing and large baffle cut
Fig. 11.7 EIIect oI baIIle spacing and cut on shellside Iouling
(Courtesy oI HTRI.)
184
controlling. Due to the lower shell ID and, consequently, the lower baIIle spacing (which is
related to the shell diameter), the shellside velocity will be appreciably higher, thus resulting
in reduced Iouling, as well as a smaller heat transIer area. Case Study 11.1 demonstrates this
vividly.
CASE STUDY 11.1: INCREASING SHELLSIDE VELOCITY FOR REDUCING FOULING
Consider the case oI a heat exchanger service having the principal process parameters
elaborated In Table 11.1a.
A thermal design was produced with a single shell. Tubes 1-in. OD (25.4-mm) × 12
BWG (2.769-mm thick) were to be used. It was Iound that the shellside velocity was only
0.62 It/s (0.19 m/s) and the shellside pressure drop only 0.74 psi (0.052 kg/cm
2
), even with a
baIIle spacing oI 9.65 in. (245 mm) in a 39.8 in. (1010-mm) ID shell. The stream analysis
was not particularly distinguished, with an E stream oI 0.21 and a B stream oI only 0.394.
The temperature proIile distortion penalty Iactor was 0.896. The shellside heat transIer
coeIIicient was only 152.6 Btu/h It
2
ƒF (745 kcal/h m
2
ƒC) and, as a result, the heat transIer
area was 3163 It
2
(294 m
2
). The shellside resistance was 40.1° oI the overall resistance to
heat transIer.
A revised design with two shells in series was produced. The shellside velocity
increased to 1.64 It/s (0.5 m/s), its pressure drop to 3 psi (0.21 kg/cm
2
), and its heat transIer
coeIIicient to 254 Btu/h It
2
ƒF (1241 kcal/h m
2
ƒC). As a result oI the corresponding increase
in the overall heat transIer coeIIicient and also the signiIicant increase in the MTD due to the
use oI two shells in series, the total heat transIer area reduced Irom 3163 It
2
(294 m
2
) to 1958
It
2
(182 m
2
), a reduction oI 38°. The total empty weight reduced Irom 34,800 lb (15,800 kg)
to 21,000 lb (9500 kg), a reduction oI 40°. This revised design also had a much better
Table 11.1a: Principal process parameters Ior Case Study 11.1
Shellside Tubeside
1. Fluid
Liquid hydrocarbon Liquid hydrocarbon
2. Flow rate, lb/h (kg/h)
112,000 (50,800) 668,000 (303,000)
3. Temperature in/out, ƒF (ƒC)
408.2 (209)/239 (115) 204.8 (96)/237.2 (114)
4. Heat duty, MM Btu/h (MM kcal/h)
11.9 (3.0)
5. Viscosity in/out, cp
0.18/0.32 1.0/0.8
6. Density in/out, lb/It
3
(kg/m
3
)
38.7 (620)/43.4 (695) 48.8 (782)/47.2 (757)
7. Thermal conductivity in/out,
Btu/h It ƒF (kcal/h m ƒC)
0.0585 (0.087)/ 0.0663
(0.0986)
0.0672 (0.1)/
0.0665 (0.099)
8. SpeciIic heat in/out, Btu/lb ƒF (kcal/kg ƒC)
0.64/0.62 0.54/0.56
9. Fouling resistance, h It
2
ƒF/Btu (h m
2
ƒC/kcal)
0.00195 (0.0004) 0.00293 (0.0006)
10. Allowable pressure drop, kg/cm
2

10 (0.7) 17 (1.2)
11. Design pressure, kg/cm
2

200 (14.0) 327 (23.0)
12. Design temperature, ƒF (ƒC)
437 (225) 266 (130)
13. Nominal connection size, in (mm)
6 (150) 10 (250)
14. Material oI construction
Carbon steel Carbon steel
185
stream analysis, with the main crossIlow stream up Irom 0.394 to 0.467 and the overall
stream eIIectiveness up Irom 0.559 to 0.655. The temperature proIile distortion penalty
Iactor improved 0.896 to 0.985 due to (a) the use oI two shells in series and (b) a lower E
stream. Thus, the revised design will not only result in lower Iouling because oI the higher
shellside velocity and the better stream analysis, but will also be lower in Iirst cost.
The principal Ieatures oI the two designs are elaborated in Table 11.1b.
e) Check pressure drop with a Iouling layer thickness
Just as Ior the tubeside, a Iouling layer thickness should be employed Ior the shellside as
well. Here, the situation is somewhat more complicated since it will aIIect the main
crossIlow and the leakage/bypass streams unequally. Thus, the baIIle hole-tube stream
will be aIIected the most and may be totally blocked since this clearance is 1/64 in. (0.4
mm) or 1/32 in. (0.8 mm), depending upon the total unsupported tube span and the tube
diameter. The shell-baIIle clearance is also aIIected appreciably and may even be consi-
Table 11.1b: Principal Ieatures oI two designs Ior Case Study 11.1
Single shell 2 Shells in Series
1. Shell ID, in. (mm)
39.8 (1010) 24.6 (625)
2. No. oI tubes × tube length, in. (mm)
638 × 236 (6000) 234 × 197 (5000)
3. No. oI tube passes
4 2
4. Tube pitch, in. (mm) × tube layout angle
1.25 (31.75) × 90
o

5. BaIIle spacing, in. (mm)
9.65 (245) 9.06 (230)
6. BaIIle cut, ° (diameter)
21 20
7. Heat transIer area, It
2
(m
2
) 3163 (294)
2 × 979 (91)
÷ 1958 (182)
Velocity, It/s (m/s) 0.62 (0.19) 1.64 (0.5)
Pressure drop, psi (kg/cm
2
) 0.74 (0.052) 3.0 (0.21)
Heat transIer coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
152.6 (745) 254 (1241)
B/E stream 0.394/0.21 0.467/0.177
A/C stream 0.311/0.085 0.23/0.127



8. Shellside
Temp. proIile distortion penalty Iactor 0.896 0.985
Velocity, It/s (m/s) 7.2 (2.2) 9.78 (2.98)
Pressure drop, psi (kg/cm
2
) 12 (0.84) 17 (1.2)

9. Tubeside
Heat transIer coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
285.3 (1393) 361 (1763)
10. Overall heat transIer coeIIicient, Btu/h It
2
ƒF (kcal/h m
2
ƒC)
61.3 (299.4) 80 (391)
11. MTD, ƒF (ƒC)
63.7 (35.4) 81.5 (45.3)
11. Overdesign, °
4.84 4.4
12. Empty weight, lb (kg) 34,800 (15,800)
2 × 10,500 (4750)
÷ 21,000 (9500)
186
dered to be totally blocked Ior very dirty services such as bitumen. Even the width oI the
lanes between tubes, through which the main crossIlow takes place, will be reduced.
Thus, depending upon the reduction oI the various Ilow widths, the stream analysis will
change. OI course, since the net result will be a reduction oI Ilow area, both the shellside
heat transIer coeIIicient and the pressure drop will increase. The pressure drop will
increase more than the heat transIer coeIIicient because the Iormer is a stronger Iunction
oI velocity.
Let us consider the Iollowing case study in order to have a better Ieel oI the eIIect oI the
Iollowing layer thickness on the perIormance oI a heat exchanger.
CASE STUDY 11.2: USE OF FOULING LAYER THICKNESS
Consider a exchanger having the principal process parameters elaborated in Table 11.2a.
Tubes 0.7874-in. (20-mm) OD × 0.0787-in. (2.0-mm) thick × 236-in. (6000-mm) long
were to be used. Although there was no temperature cross, a 3-shells-in-series design was
produced to make good use oI the allowable pressure drop oI the short residue stream. It will
be noticed that there is a high overdesign (20.48°). This is because the speciIied tubes oI
0.7874-in. (20-mm) OD, 0.0787-in. (2-mm) thick, and 236-in. (6000-mm) long only were to
be used; otherwise, the tube length could have been reduced to 197 in. (5000 mm) and the
exchanger would have been just adequately surIaced. Despite the high Iouling resistances on
both sides, no Iouling layer thicknesses were used. The principal construction parameters
and the relevant thermal perIormance parameters are speciIied in Table 11.2b.
Now, let us see what happens when appropriate values oI Iouling layer thickness are
speciIied Ior the thermal design run. Fouling layer thicknesses oI 0.04 in. (1.0 mm) and 0.02
in. (0.5 mm) were speciIied Ior the shellside and tubeside streams, respectively. The results
are detailed in Table 11.2b. It will be noticed that both the shellside and tubeside velocities,
pressure drops, and heat transIer coeIIicients, as well as the overdesign, have increased
Table 11.2a: Principal process parameters Ior Case Study 11.2
Shellside Tubeside
1. Fluid
Liquid hydrocarbon Liquid hydrocarbon
2. Flow rate, lb/h (kg/h)
207,000 (94,000) 271,000 (123,000)
3. Temperature in/out, ƒF (ƒC)
571 (299.4)/463 (239.4) 379 (192.8)/465 (240.6)
4. Heat duty, MM Btu/h (MM kcal/h)
13.65 (3.44)
5. Viscosity in/out, cp
7.6/23 5.65/3.25
6. Density in/out, lb/It
3
(kg/m
3
)
54.6 (875)/56.2 (900) 54.3 (870)/53.3 (854)
7. Thermal conductivity in/out,
Btu/h It ƒF (kcal/h m ƒC)
0.052 (0.077)/0.056 (0.083) 0.061 (0.09)/0.0565 (0.084)
8. Fouling resistance, h It
2
ƒF/Btu
(h m
2
ƒC/kcal)
0.0098 (0.002) 0.0049 (0.001)
9. Allowable pressure drop, psi (kg/cm
2
)
21.3 (1.5) 20 (1.4)
10. Design pressure, psig (kg/cm
2
g)
356 (25) 242 (17.0)
11. Design temperature, ƒF (ƒC)
626 (330) 48200 (250)
12. Nominal connection size in/out, in. (mm)
6 (150) 6 (150)
187
substantially. In Iact, the pressure drops on both sides have exceeded the allowable margins.
This problem would not have arisen iI the original design had incorporated Iouling layer
thicknesses and had been designed accordingly in the Iirst place.
f) Use helical baIIle exchangers
For a detailed account oI Helixchangers, the reader is reIerred to Section 13.3.6.
g) Use twisted-tube heat exchangers
For a detailed account oI twisted-tube heat exchangers, the reader is reIerred to Section
13.3.7.


Table 11.2b: Comparison oI heat exchanger perIormance with and without Iouling layer thickness
Without fouling
layer thickness
With fouling
layer thickness
1. No. oI shells
3 (in series)
2. Shell ID, in. (mm)
35.4 (900)
3. No. oI tubes × no. oI tube passes
728 × 4
4. Tube pitch, in. (mm) × tube layout angle
1.024 (26) × rotated square (45
o
)
5. BaIIle spacing, in. (mm)
7.9 (200)
6. BaIIle cut, ° (diameter)
20
7. Heat transIer area, It
2
(m
2
)
3 × 2862 (266) ÷ 8586 (798)
8. MTD, ƒF (ƒC)
93.2 (51.8) 93.1 (51.7)
A (baIIle hole-tube leakage) 0.097 0
B (main crossIlow stream) 0.546 0.525
C (bundle-shell bypass) 0.09 0.12
E (baIIle-shell leakage) 0.266 0.356


9. Stream analysis
F (pass partition bypass) 0 0
CrossIlow 1.54 (0.47) 2.2 (0.67)
10. Shellside velocity,
It/s (m/s)
Window-Ilow 1.5 (0.45) 1.64 (0.5)
11. Shellside heat transIer coeIIicient, Btu/h It
2
ƒF (kcal/h m
2
ƒC)
74.7 (364.8) 89.2 (435.6)
12. Shellside pressure drop, psi (kg/cm
2
)
1.05 1.82
velocity, It/s (m/s) 1.08 1.23
heat transIer coeIIicient, kcal/h m2 °C 53.5 (261) 63.5 (309.9)

13. Tubeside
pressure drop, kg/cm2 1.4 1.83
14. Overall heat transIer coeIIicient, kcal/h m
2
ƒC
20.7 (101) 22.4 (109.3)
15. Overdesign, °
20.48 30.0

188
References
|1| Taborek, J. et al, 1972, 'Fouling ² The Major Unresolved Problem in Heat TransIer,¨ Parts I
and II, Chemical Engineering Progress, 68(2), pp. 59²67 and 68(7), pp. 69²78.
|2| Tubular Exchanger ManuIacturers Association, 1999, Stanaaras of the Tubular Exchanger
Manufacturers Association, 8th Eaition, TEMA, New York.
|3| Gough, M.J., and Rogers, J.V., 1987, 'Reduced Fouling by Enhanced Heat TransIer Using
Wire-Matrix Radial Mixing Elements,¨ AIChE Symposium Series, Vol. 83, No. 257, pp. 16²
21.
|4| Gough, M.J., and Rogers, J.V., 1991. 'Getting more perIormance Irom heat exchangers,¨
Processing, July, pp. 15²16
Further Reading
1. Epstein, N., 1978, 'Fouling in Heat Exchangers,¨ Proc. Sixth International Heat Transfer
Conf., Keynote Papers, Vol. 6, pp. 235²253, Toronto, Hemisphere Publishing Corp., New
York.
2. Garret-Price, B.A., et al, 1985, Fouling of Heat Exchangers, Characteristics, Costs,
Prevention, Control ana Removal, Noyes Publications, Park Ridge, New Jersey.
3. Hewitt, G.F. (ed.), 1998, Heat Exchanger Design Hanabook 1998, Vols. 3 and 4, Begell
House, Inc., New York.
4. Knudsen, J.G., 1984, 'Fouling oI Heat Exchangers: Are We Solving the Problem?¨, Chemical
Engineering Progress, February, pp. 63²69.
5. Melo, L.F., Bott, T.R., and Bernardo, C.A (eds.), 1988, Fouling Science ana Technology,
Kluwer Academic Publishers.
6. Somerscales, E.F.C., and Knudsen, J.G., (eds.), 1981, Fouling of Heat Transfer Equipment,
Hemisphere Publishing Corp., New York.
7. Kakac, S., et al., 1981, Heat Exchangers. Thermal-Hyaraulic Funaamentals ana Design,
Hemisphere Publishing Corp., New York.
8. Joshi, H.M., 1999, 'Mitigate Iouling to improve heat exchanger reliability,¨ Hyarocarbon
Processing, Jan., pp. 93²95.
9. Bott, T.R., 1995, Fouling of Heat Exchangers, Elsevier.
189
CHAPTER 12
9LEUDWLRQ$QDO\VLV
Introduction
With the increase in the size oI plants in the chemical process industries, and a corre-
sponding increase in the size oI shell-and-tube heat exchangers, tube Iailure due to Ilow-
induced vibration has become a more visible reality. The prime reason Ior this has been
the scale-up oI heat exchanger sizes without a proper understanding and analysis oI Ilow-
induced vibration. The principal culprit has been the use oI larger and larger unsupported
tube lengths employed to handle the very large Iluid Ilow rates on the shellside. Today,
the situation is much better with a Iairly good working knowledge oI producing designs
saIe against Iailure oI tubes due to Ilow-induced vibration. The term 'working
knowledge¨ is very signiIicant in that Ilow-induced vibration is an extremely complex
phenomenon, and what the designer really requires is a practical avoidance technique.
Flow-induced vibration analysis oI a shell-and-tube heat exchanger is an integral
element oI its thermal design. A proper design is one that it is absolutely saIe against Iailure
oI tubes due to Ilow-induced vibration. Most sophisticated thermal design soItware packages
carry out vibration analysis as a routine ingredient oI thermal design. This is essential since
it is during thermal design that the geometry oI a heat exchanger is Iinalized and it is this
same geometry, along with Ilow and physical property parameters, that determine whether a
given heat exchanger is saIe against Iailure oI tubes due to Ilow-induced vibration.
Flow-induced vibration is a very complex subject and involves the interplay oI several
parameters, many oI which are not very well established. Although many cases oI Iailure oI
tubes due to Ilow-induced vibration have been reported in the past several years, an
understanding oI the Iactors responsible Ior these Iailures leave much to be desired. The
literature depicts several interesting studies on speciIic Iacets oI the vibration problem;
however, very Iew investigations have considered the speciIic problems associated with
shell-and-tube heat exchangers. Experimental data on Ilow-induced vibration are usually
obtained under controlled conditions using either single tubes or ideal tube banks employing
crossIlow or parallel Ilow, conditions hardly representative oI actual conditions in industrial
heat exchangers.
Consequently, extrapolation oI such data obtained under idealized test conditions to
real-liIe heat exchangers is Iraught with peril. It is, thereIore, prudent to be somewhat
conservative in one's approach Ior producing a design that is saIe against Iailure oI tubes due
to Ilow-induced vibration. Usually, Ior the design oI new heat exchangers, this conservatism
does not result in a signiIicantly costlier design since a saIe design can be produced by
modiIying the shell type and/or the baIIle style and baIIle design.
It will not be in the scope oI this book to dwell in detail on the mechanism oI Ilow-
induced vibration. However, a brieI description is oIIered below so that the reader can easily
190
understand the subsequent discussion. For a detailed understanding oI the phenomena
associated with Ilow-induced vibration and their mechanisms, please reIer to ReIs. |17|.
12.1 Mechanics of Flow-Induced Vibration
Tubes are the most Ilexible part oI a heat exchanger and, consequently, the most
vulnerable to Ilow-induced vibration caused by the Ilow oI Iluid past them. Heat
exchanger tubes oIten vibrate, but at such low Irequencies that there is no damage. It is
only when the Irequency oI the tube vibration becomes appreciable that there is a danger
oI Iailure. The lowest Irequency at which tubes vibrate is called the Iundamental natural
Irequency. The higher natural Irequencies are known as the second mode, third mode,
and so on. The Iundamental Irequency is what designers are usually concerned with, and
this is simply reIerred to as the natural Irequency oI a tube.
12.1.1 Natural frequency
The natural Irequency oI tubes depends upon the Iollowing:
a) the way in which they are Iixed (whether clamped or supported)
b) the nature oI the intermediate supports (whether clamped, pinned, or supported)
c) the tube unsupported span
a) the number oI spans
e) the tube material and thickness/OD
Ordinary shell-and-tube heat exchangers with plate baIIles have tubes that are simply
supported at the tubesheets and the baIIles.
Since the Iirst and last baIIle spaces are invariably larger than the central baIIle spacing
in order to accommodate the nozzles and shell Ilanges, the unsupported span varies along the
length oI the tubes and is largest at the inlet and outlet regions. The determination oI the
natural Irequency oI a tube supported at diIIerent spans along its length is rather
complicated. A somewhat less accurate method, but one that is normally adequate Ior
engineering calculations, is that proposed by MacDuII and Felgar |1|, which assumes that all
spans are oI equal length. This method can consider diIIerent types oI end and intermediate
support; however, it is customary to consider clamped ends and simple intermediate
supports Ior predicting the natural Irequency oI heat exchanger tubes. Hence, vibration
analysis Ior shell-and-tube heat exchangers is usually carried out individually Ior three
regionsthe inlet, central, and the outletand the potential Ior vibration damage evaluated.
The equation proposed by MacDuII and Felgar is:
I
n
÷ 0.04944 C
n
(EIg
c
/W
e
L
4
)
0.5
(12.1)
where
I
n
÷ natural Irequency oI straight tubes, Hz
C
n
÷ Frequency constant
E ÷ modulus oI elasticity, kg/m
2

I ÷ sectional moment oI inertia, m
4

g
c
÷ gravitational constant, 9.81 m/s
2

W
e
÷ eIIective weight per unit tube length including the weight oI the
Iluid inside tubes, kg/m
L ÷ tube unsupported span, m
191
For the determination oI the natural Irequency oI straight tubes and U-tubes, the reader
may reIer to the TEMA Standards, 8th Edition (1999, Section 6, pp. 97103). It will be
seen Irom Table V-5.3 (TEMA Standards, p 99) that there is a correction Iactor Ior any
axial stresses that may be present in heat exchanger tubes, either compressive or tensile,
due to the manuIacturing process itselI as well as due to operating temperatures. Com-
pressive axial loads decrease the tube natural Irequency, whereas tensile axial loads in-
crease it. During the Iabrication oI bundles, tubes are sometimes placed under compres-
sion as a result oI over-rolling; this results in a decrease in the natural Irequency.
Looking at Eq. (12.1), it is seen that the natural Irequency oI a tube varies:
a) inversely with the square oI the tube unsupported span
b) to the 0.5 power oI the elastic modulus oI the tube material
c) to the 0.5 power oI the moment oI inertia oI the tube cross section (the latter
varies to the power oI 4 oI the tube outside and inside diameters)
a) inversely to the 0.5 power oI the eIIective tube weight per unit length, which
includes the weight oI the Iluid inside the tubes
Hence, Ior a given tube material, outside diameter and thickness (which are speciIied Ior
a given service, and over which the heat exchanger thermal designer has no control), the
only variables are the tube unsupported span and the density oI the tubeside Iluid. This
means that, Ior a given service, the only variable is the tube unsupported span. In Iact,
this is the most crucial parameter in vibration analysis.
The determination oI the natural Irequency oI a U-tube is Iar more diIIicult than Ior
straight tubes. Usually, a simpliIied procedure is employed that considers the straight portion
and the bent portion separately. Please reIer to the above reIerred section oI the TEMA
standards Ior details.
12.1.2 Flow-induced vibration phenomena
The various Ilow-induced tube vibration phenomena are vortex shedding, turbulent
buIIeting, and Iluidelastic whirling. Additionally, there is also the phenomenon oI
acoustic vibration. Let us consider tube vibration Iirst.
Jortex sheaaing
When a Iluid Ilows across a single tube, it produces a series oI vortices in the
downstream wake (Fig. 12.1). This alternate shedding oI vortices produces alternating
Iorces, the Irequency oI which varies directly with the velocity oI Ilow. This Irequency is
usually described by the dimensionless Strouhal number:
S ÷ I
s
d
o
/v (12.2)
where
S ÷ Strouhal number
I
s
÷ vortex shedding Irequency
v ÷ crossIlow velocity
d
o
÷ outside diameter oI tube
Vortex shedding also occurs when Ilow
takes place across a tube bank. Here, the
Strouhal number is a Iunction oI the longi-

Fig. 12.1Vortex shedding resulting Irom Ilow across
a tube (Reprinted Irom the Heat Exchanger Design
Handbook, 2002 with permission oI Begell House,
Inc.)
192
tudinal and transverse spacing oI the
tubes as shown in Fig. 12.2, where
the Strouhal number is plotted
against the ratio oI tube pitch to tube
diameter Ior 30ƒ (triangular), 90ƒ
(square), 45ƒ (rotated square), and
60ƒ (rotated triangular) layout
angles. Interestingly, the values oI
the Strouhal number do not vary
signiIicantly with the Reynold`s
number. Thus, it will be seen that
the tube layout angle plays a
signiIicant role in Ilow-induced
vibration.
It will also be seen Irom Eq. (12.2) that, since the Strouhal number is Iairly constant Ior
a certain tube layout angle and tube pitch, the higher the crossIlow velocity, the higher will
be the vortex shedding Irequency.
Turbulent buffeting
Highly turbulent vapor or gas Ilow on the shellside oI a heat exchanger contains a wide
range oI Irequencies distributed around a central dominant Irequency. When such Ilow
buIIets the tubes, they selectively extract energy Irom the turbulence at their natural
Irequency. The central dominant Irequency increases with the crossIlow velocity. Thus,
the higher the crossIlow velocity, the higher the turbulent buIIeting Irequency.
Fluiaelastic whirling
This was propounded by Connors |3| and is characterized by tubes vibrating in an orbital
or 'whirling¨ manner, once suIIicient energy is available Ior resonance to occur. This
motion is produced when the shellside Ilow across the tubes causes both liIt and drag
movement oI the tubes at their natural Irequencies. It can lead to a 'runaway¨ condition iI
the energy supplied to the tubes cannot be absorbed by the system damping and, thereby,
lead to Iailure oI tubes due to Ilow-induced vibration. Such a Iailure is likely to occur iI
the crossIlow velocity is greater than a critical velocity.
The critical velocity is given by the expression
v
crit
÷ ȕI
n
D
o
|(M
e
d
0
)/(ȡ
s
D
o
2
)|
0.5
(12.3)
where
v
crit
÷ critical velocity, m/s
ȕ ÷ Iluidelastic instability threshold constant, dimensionless
I
n
÷ natural Irequency oI tubes, Hz
d
0
÷ log decrement Ior the tube bundle in the shellside Iluid under no-
Ilow conditions, dimensionless
D
o
÷ tube outside diameter, m
M
e
÷ eIIective mass per unit tube length, kg/m
ȡ
s
÷ density oI shellside Iluid, kg/m
3

The Iluidelastic instability threshold constant ȕ is a Iunction oI the tube pitch ratio and
Fig. 12.2 Strouhal numbers Ior tube layouts (Reprinted Irom
the Heat Exchanger Design Handbook, 2002 with permission
oI Begell House, Inc.)
193
the tube layout angle. Fluidelastic instability has been interpreted to be a major cause oI
tube Iailure due to Ilow-induced vibration.
12.1.3 How and when tubes vibrate
The essential phenomenon is the resonance between the Iorcing Irequency (oI vortex
shedding Ior liquids, and vortex shedding and turbulent buIIeting Ior gases and vapors)
and the natural Irequency oI the tubes. It will be seen Irom the above discussion that both
the natural tube Irequency and the Iorcing Irequency (vortex shedding and turbulent
buIIeting) are strong Iunctions oI the baIIle spacing/unsupported tube span.
1) The tube natural Irequency varies inversely as the square oI the tube unsupported
span.
2) The Iorcing Irequency (oI both vortex shedding and turbulent buIIeting) varies
directly with the crossIlow velocity. Since crossIlow velocity varies inversely
with the baIIle spacing, Iorcing Irequency also varies inversely with baIIle
spacing. The baIIle spacing is directly related to the tube unsupported span. Thus,
Iorcing Irequency varies inversely with unsupported span.
Hence, when the baIIle spacing oI a heat exchanger is reduced, the natural tube Irequency
increases signiIicantly more than the Iorcing Irequency. Consequently, a lower baIIle
spacing always makes a heat exchanger saIer against Iailure oI tubes due to Ilow-induced
vibration.
12.1.4 Damping
When adequate energy is Ied to tubes by the Ilowing Iluid, they are excited into vibration.
The intensity oI this vibration is evidenced by the extent oI periodic movement oI the
tube, with the greatest displacement (amplitude) usually at the mid-span oI the tube.
The energy oI vibration is dissipated by internal and external damping. Internal damping
is oIIered by the mechanical properties oI the tube material, the geometry oI the intermediate
supports (baIIles), and the tube-baIIle clearance. External damping is oIIered by the shellside
Iluid, with the viscosity being the determining parameter. Thus, liquids oIIer considerable
external damping, but gases and vapors much less.
System damping has a strong inIluence on the variation oI amplitude with time. Once
the energy ceases to excite vibration, the amplitude oI vibration decays with time. The rate at
which Ilow-induced vibration dampens out is usually exponential. The diIIerence in
successive amplitude peaks is called the log decrement and is a measure oI the degree oI
damping. The higher the value oI the log decrement, the greater is the degree oI damping.
Heat exchanger tubes are normally lightly damped structures with a low value oI log
decrement.
II the energy that is input cannot be dissipated through damping, the resultant amplitude
oI vibration will increase with time and lead to a 'runaway¨ condition. This is described
later under 'Iluidelastic whirling.¨ II there is a balance, the amplitude remains constant.
However, iI the system damping is greater than the energy input, the amplitude decays with
time. This is represented in Fig. 12.3, which also quantiIies and depicts the log decrement.
12.1.5 Modes of tube failure
Prolonged vibration oI tubes leads to their Iailure by one or more oI the Iollowing modes:
194
a) BaIIle damage ² Since baIIle holes are drilled larger than the outside diameter oI
tubes to Iacilitate their insertion into the baIIles during assembly, tubes are Iree to
move at the baIIles. The repeated cutting action at the baIIle wall may result in
the cutting oI the relatively thin tube wall.
b) Tube joint leakage ² The joint between the tube and the tubesheet, welded or
expanded, can also Iail due to vibration.
c) Collision damage ² Adjacent tubes may impact with one another iI the amplitude
oI vibration is large enough. Such repeated impacting can reduce the tube
thickness over a period oI time and eventually split the tube open.
a) Fatigue ² Repeated bending oI the tubes over a period oI time can result in
Iatigue Iailure due to the stresses created.
Each tube in a shell-and-tube heat exchanger is held at the tubesheet at both ends, and by
the baIIles along its length. With U-tubes, however, there is a single tubesheet. The
distance between any two adjacent points oI support is the unsupported span which is
subject to vibration due to the action oI the Iluid Ilowing across the tubes. Although tube
Iailure due to Ilow-induced vibration have been reported Irom many parts oI a heat
exchanger, the areas oI high unsupported spans and high local velocities have the greatest
susceptibility, especially the tubes within the Iirst Iew tube rows oI the baIIle tips since
these experience the highest local velocities.
12.2 How to Predict Damaging Flow-Induced Vibration
The tube unsupported span is really the key to the prevention oI damaging Ilow-induced
vibration in shell-and-tube heat exchangers. II this span can be reduced adequately, the
tube natural Irequency will be suIIiciently higher than the Iorcing Irequency, so that no
resonance will occur. Usually, the ratio oI the Iorcing Irequency to the tube natural
Irequency is kept below 0.8 to produce a saIe design.
In order to prevent the occurrence oI Iluidelastic whirling, it is recommended that the
ratio oI crossIlow velocity to critical velocity be less than 0.8, and this includes crossIlow
velocity in the central rows, as well as the rows near the baIIle tips.
It will be reiterated here that mere matching oI Iorcing Irequency (due to vortex
shedding or turbulent buIIeting) and natural Irequency oI tubes is not enough to result in
damage oI tubes due to Ilow-induced vibration. The energy content oI the Ilowing stream
must be high enough to produce a suI-
Iiciently large amplitude oI vibration.
Generally, an amplitude oI over 25° (a
conservative value) oI the gap between
tubes (in other words, tube pitch minus tube
OD) is considered to be unsaIe. This is ap-
plicable not only Ior collision damage, but
Ior baIIle damage, Iatigue, and tube joint
leakage because the greater the amplitude
oI vibration, the greater is the cutting action
at the baIIles and the greater the Iatigue
stresses produced. As long as the amplitude
is less than this value, a heat exchanger
need not be considered unsaIe against
damage due to Ilow-induced vibration, even

Fig. 12.3 Damping oI a vibrating tube aIter excitation
stops (Reprinted Irom the Heat Exchanger Design
Handbook, 2002 with permission oI Begell House,
Inc.)
195
iI the ratio oI Iorcing Irequency to natural Irequency is greater than 0.8 and less than 1.2.
Thus, the important checks to be carried out to ensure a saIe design against the
possibility oI Iailure oI tubes due to Ilow-induced vibration are:
1) The crossIlow velocity at all locations in the tubeIield should be less than 80° oI
the critical velocity.
2) The ratio oI the vortex shedding Irequency (and turbulent buIIeting Irequency Ior
gases or vapors) to the natural Irequency oI the tubes should be less than 0.8,
provided the amplitude oI vibration is greater than 25° oI the gap between tubes;
iI the amplitude is less, the Irequency ratio can be greater than 0.8.
12.3 Vital Link between Flow-Induced Vibration and Pressure Drop
We have come to realize that the unsupported tube span is the most signiIicant parameter
in Ilow-induced vibration. We also know that the baIIle spacing (and, thereby, the
unsupported tube span) is really linked to the shellside pressure drop, which is also linked
to the density oI the shellside Iluid. The baIIles in a heat exchanger can be brought as
close as the allowable shellside pressure drop will permit.
The allowable pressure drop is usually much lower Ior vapor or condensing services
than Ior liquid heating or cooling services, typically 1.42.8 psi (0.10.2 kg/cm
2
) Ior the
Iormer and 10 psi (0.7 kg/cm
2
) Ior the latter. To make matters worse, vapor density is much
lower Ior gases and vapors than Ior liquids. As a direct repercussion oI these two realities,
baIIle spacing and, thereby, tube unsupported span are much larger in gas coolers/heaters
and condensers than in sensible liquid heating/cooling services. The lower the operating
pressure, the larger will be the unsupported span, other parameters remaining constant.
Consequently, Ilow-induced vibration damage is Iar more prevalent in gas coolers/heaters
and condensers. Evidently, in the above discussion, the gas or vapor stream is on the
shellside and this will continue to be the premise in the subsequent discussion.
Consider what typically happens in the design oI a gas cooler. The allowable shellside
pressure drop is rather low and so is the gas density. With a relatively low baIIle spacing
(say, 40° oI the shell inside diameter) and cut (say, 25°), the calculated pressure drop Iar
exceeds the allowable pressure drop; however, Ilow-induced vibration may not be a
problem. Since the permitted shellside pressure drop must be complied with, the designer
increases the baIIle spacing and the baIIle cut until he or she manages to reduce the shellside
pressure drop to the permitted value. However, the baIIle spacing and, thereby, the
unsupported tube span is now so high that Iailure oI tubes due to Ilow-induced vibration is a
distinct possibility.
The real culprit, thereIore, is the combination oI the low allowable shellside pressure
drop and the low vapor/gas density. Since gas or vapor density is a direct Iunction oI the
operating pressure and the molecular weight oI the gas or vapor, the most diIIicult situation
is a low molecular weight gas at a low operating pressure. Should the gas or vapor have a
high molecular weight and/or be at a high pressure, producing a design saIe against Iailure
due to Ilow-induced vibration will be Iar less diIIicult.
12.4 Producing a Design that is Safe against Flow-Induced Vibration
The designer has the twin objectives oI satisIying the shellside pressure drop limitation
and producing a design that is saIe against the Iailure oI tubes due to Ilow-induced vibra-
tion. As stated earlier, the real ploy is in reducing the unsupported tube span while still
restricting the shellside pressure drop to within the permitted value. Consequently, the
196
methodology is the same as employed Ior reducing the shellside pressure drop, as that is
the crux oI the matter. This was discussed in detail in Chapter 3.
Single-pass shell ana single-segmental baffles
The Iirst shell style is evidently the TEMA E shell, that is, a single shell pass (Fig. 2.6).
Here, the entire shellside Iluid Ilows along the entire length oI the shell. Also, the Iirst
baIIle style is the single-segmental (Fig. 3.3). The baIIle spacing and cut are increased
progressively until the allowable shellside pressure drop limitation is met. However, iI in
so doing, the design becomes unsaIe against Ilow-induced vibration, the design is
unacceptable and the next alternative should be attempted.
Single-pass shell ana aouble-segmental baffles
In the second design, the TEMA E shell is retained but the baIIles are changed to double-
segmental (Fig. 3.3). What happens now is that the shellside Iluid is made to split and
recombine as it Ilows along the length oI the shell. Thus, the crossIlow velocity is
approximately one halI oI that with single-segmental baIIles at the same spacing. The
window velocity does not reduce, since the entire shellside Iluid has to Ilow through the
total window area. However, the major shellside pressure drop with single-segmental
baIIles is in crossIlow, so that there is an appreciable reduction in the pressure drop Ior an
otherwise identical heat exchanger (including baIIle spacing) with single-segmental
baIIles. The use oI double-segmental baIIles permits the baIIles to be brought closer Ior
the same shellside pressure drop. This translates into a design that is saIer against Iailure
oI tubes due to Ilow-induced vibration.
CASE STUDY 12.1: PRODUCING A SAFE DESIGN USING
DOUBLE-SEGMENTAL BAFFLES IN A SINGLE-PASS SHELL
Let us reIer to Case Study 7.2, presented in Chapter 7. The salient process parameters are
elaborated in Table 7.2a. Two designs were presented, a single-pass shell with single-
segmental baIIles and a single-pass shell with double-segmental baIIles. The construction
and perIormance parameters oI both oI these designs were presented in Table 7.2b. The
design with single-segmental baIIles had been Iound prone to Iailure oI tubes due to
Ilow-induced vibration while the design with double-segmental baIIles was Iound to be
saIe.
The Ilow-induced vibration analysis had not been presented because the same had not
Table 12.1: Results oI vibration analysis oI Case Study 12.1
Single-pass shell with
single-segmental baffles
Single-pass shell with
double-segmental baffles
1. Critical velocity, It/s (m/s)
14.4 (4.39) 30.2 (9.21)
2. Average crossIlow velocity, It/s (m/s)
12.7 (3.87) 9.78 (2.98)
3. BaIIle tip crossIlow velocity, It/s (m/s)
12.7 (3.87) 10.5 (3.21)
4. Vortex shedding ratio
2.844 0.985
5. Unsupported span, in. (mm)
51.2 (1300) 25.6 (650)
6. (Unsupported span)/TEMA maximum
0.824 0.412
7. CrossIlow amplitude, in. (mm)
- 0.0026 (0.067)
197
been discussed up to that stage. Table 12.1 details the vibration analysis results oI both oI the
designs Ior the central region. It will be seen that, in the Iirst design, the crossIlow velocity is
88° oI the critical velocity and the vortex shedding ratio is unacceptably high. In the second
design, however, the crossIlow velocity is only 32.4° oI the critical velocity. Although the
vortex shedding ratio is 0.985, the crossIlow amplitude is extremely low at 0.0026 in. (0.067
mm), so as to render the design perIectly saIe against Iailure oI tubes due to Ilow-induced
vibration. The principal constructional diIIerence between the two designs is the
unsupported span, 51.2 in. (1300 mm) in the Iirst and 25.6 in. (650 mm) in the second.
Diviaea-flow shell ana single-segmental baffles
II, with an E shell and double-segmental baIIles, it is still not possible to achieve the twin
objectives oI producing a design saIe against vibration Iailure and satisIying the shellside
pressure drop limitation, the next construction to explore is a divided-Ilow (TEMA J
style) shell (see Fig. 2.6) with single-segmental baIIles. Here, with an identical baIIle
spacing, not only the crossIlow but even the window-Ilow velocity is reduced since the
Ilow rate itselI is reduced to one halI. Thus, Ior the same shellside pressure drop, the
baIIles can be brought even closer than is possible with an E shell and double-segmental
baIIles.
Table 12.2a: Principal process parameters Ior Case Study 12.2
Shellside Tubeside
1. Fluid
Hydrocarbon gas Cooling water
2. Flow rate, lb/h (kg/h)
113,000 (51,260) 357,000 (162,000)
3. Temperature in/out, ƒF (ƒC)
174.2 (79)/102.2 (39) 89.6 (32)/102.2 (39)
4. Operating pressure, psia (kg/cm
2
a)
250 (17.6) 71 (5.0)
5. Allowable pressure drop, psi (kg/cm
2
)
3.27 (0.23) 10 (0.7)
6. Fouling resistance, h It
2
ƒF/Btu (h m
2
ƒC/kcal)
0.00098 (0.0002) 0.00195 (0.0004)
7. Density in/out, lb/It
3
(kg/m
3
)
0.75 (12.0)/0.82 (13.1)
8. Molecular weight
16.1
9. Inlet/outlet viscosity, cp
0.01/0.01
10. Average speciIic heat, Btu/lb ƒF (kcal/kg ƒC)
0.56
11. Average thermal conductivity, Btu/h It ƒF
(kcal/h m ƒC)
0.013 (0.019)



Standard
12. Heat duty, MM Btu/h (MM kcal/h)
4.5 (1.135)
13. Design pressure, kg/cm
2
g


300 (21.1) 100 (7.0)
14. Design temperature, ƒF (ƒC)
200 (93) 158 (70)
15. Material oI construction
Carbon steel Carbon steel
16. Nominal line size, in. (mm)
16 (400) 6 (150)
198
CASE STUDY 12.2: PRODUCING A SAFE DESIGN USING
A DIVIDED-FLOW SHELL AND SINGLE-SEGMENTAL BAFFLES
Let us consider the service speciIied in Table 12.2a. Carbon steel tubes 0.7874-in. (20-
mm) OD × 0.0787-in. (2-mm) thick × 19.68-It. (6000-mm) long were to be used Ior the
design. Note the moderately high gas Ilow rate on the shellside and the low gas density.
An attempt was Iirst made to produce a thermal design with and "E" shell and single-
segmental baIIles. The results are shown in Table 12.2b. The vibration analysis results are
Ior the central region. It will be seen that Ior a shell ID oI 27.6 in. (700 mm), the baIIle
spacing had to be as high as 37.4 in. or 950 mm (which is well above the TEMA limit) in
order to contain the shellside pressure drop within the allowable limit. As expected, this
Table 12.2b: Results oI Case Study 12.2

E shell with
single-segmental
baffles
E shell with
double-segmental
baffles
1 shell with
single-segmental
baffles
1. Shell ID, in. (mm)
27.6 (700) 27.6 (700) 30.5 (775)
2. Total no. oI tubes × no. oI tube passes
564 × 4 564 × 4 726 × 4
3. Heat transIer area, It
2
(m
2
)
2249 (209) 2249 (209) 2894 (269)
4. Tube OD × thick, in. (mm)
0.7874 (20) × 0.0787 (2)
5. Tube length × tube pitch, in. (mm)
236 (6000) × 0.984 (25) ǻ
6. BaIIle spacing, in. (mm)
37.4 (950) 17.7 (450) 13.8 (350)
7. BaIIle cut, ° diameter 35
27.3
(1 row overlap)
25
8. Shellside pressure drop, psi (kg/cm
2
)
3.1 (0.22) 3.1 (0.22) 2.6 (0.18)
9. Tubeside pressure drop, psi (kg/cm
2
)
10.0 (0.7) 10.0 (0.7) 6.7 (0.47)
10. Shellside heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/m
2
ƒC)
135.2 (660) 128.2 (626) 94.0 (459)
11. Overall heat transIer coeIIicient,
Btu/h It
2 o
F (kcal/h m
2
ƒC)
83.4 (407) 80.7 (394) 64.7 (316)
12. Overdesign, °
18.5 11.5 13.3
Vibration Analysis (central region)
1. Critical velocity, It/s (m/s)
8.6 (2.61) 32.4 (8.64) 37.9 (11.55)
2. Average crossIlow velocity, It/s (m/s)
24.4 (7.43) 27.8 (8.48) 19.5 (5.95)
3. Vortex shedding ratio
5.3 1.824 0.96
4. Turbulent buIIeting ratio
8.584 2.953 1.55
5. Unsupported span, in. (m)
74.8 (1900) 35.4 (900) 27.6 (700)
6. (Unsupported span)/TEMA maximum
1.205 0.571 0.444
7. CrossIlow ȡv
2
,

lb/It sec
2
(kg/m sec
2
)


477 (709) 619 (921) 304 (452)
8. CrossIlow amplitude, in. (mm)
- - 0.006 (0.154)
199
design was not acceptable since it had a high probability oI Iailure oI tubes due to Ilow-
induced vibration with: (a) the average crossIlow velocity 2.84 times the critical velocity and
(b) very high vortex shedding and turbulent ratios.
Successive designs were made in the Iollowing order:
E shell with double-segmental baIIles
J shell with single-segmental baIIles
The results oI these designs are also shown in Table 12.2b. Let us take a look at them.
First the E shell design with double-segmental baIIles. Although this design was much
saIer against Iailure due to Ilow-induced vibration than the previous design with single
segmental baIIles, it was not saIe enough because:
1) The average crossIlow velocity (27.8 It/s or 8.48 m/s) is 98° oI the critical
velocity (28.35 It/s or 8.64 m/s)
2) The vortex shedding and the turbulent buIIeting ratios are still unacceptably high
This design was also, thereIore, not acceptable.
The third design tried was a J shell with single-segmental baIIles. This design appears to
be saIe against Iailure oI tubes due to Ilow-induced vibration since the average crossIlow
velocity is only 51.5° oI the critical velocity. The vortex shedding and turbulent buIIeting
ratios, 0.96 and 1.55, are not absolutely saIe; however, considering that the crossIlow
amplitude is only 0.006 in. (0.154 mm), the unsupported span is only about 44° oI the
TEMA maximum and the crossIlow ȡv
2
is only 304 lb/It sec
2
(452 kg/m sec
2
), this design
may be accepted as a saIe design.
It may be noted in the above table how the unsupported span has been progressively
reduced Irom 74.8 in. (1900 mm) in the Iirst design to 35.4 in. (900 mm) in the second
design and to 27.6 in. (700 mm) in the Iinal design. Consequently, the ratios oI
1) crossIlow velocity to critical velocity
2) vortex shedding Irequency to natural Irequency oI tubes
3) turbulent buIIeting Irequency to natural Irequency oI tubes
have progressively been reduced.
Diviaea-flow shell ana aouble-segmental baffles
Should even a divided-Ilow shell construction with single-segmental baIIles be unable to
produce a satisIactory design, the baIIle type should be altered to double-segmental. That
is, a combination oI a divided-Ilow shell and double-segmental baIIles should be
examined.
CASE STUDY 12.3: PRODUCING A SAFE DESIGN USING
A DIVIDED-FLOW SHELL AND DOUBLE-SEGMENTAL BAFFLES
Let us consider the service speciIied in Table 12.3a Ior a hydrocarbon gas cooler. Tubes
oI carbon steel 0.7874-in. (20-mm) OD, 0.0787-in. (2-mm) thick, and 19.68-It. (6-m)
long were to be used in a Iixed-tubesheet construction. A minimum overdesign oI 10°
on surIace was to be incorporated.
Since the gas Ilow rate was very high at 463,000 lb/h (210,000 kg/h), a single-pass shell
(TEMA E) was quite hopeless in limiting the shellside pressure drop to the permissible value
and had to be discarded.
A divided-Ilow shell with single-segmental baIIles was considered and the principal
200
construction and perIormance parameters that emerged are shown in Table 12.3b. Although
the shellside pressure drop could be contained within the allowable limit oI 4.3 psi (0.3
kg/cm
2
), it was only with a very large baIIle spacing oI 27.9 in. (709 mm). This led to a tube
unsupported span that was 90° oI the TEMA maximum and a Ilow-induced vibration
analysis that was Iar Irom satisIactory: the crossIlow velocity was well above the critical
velocity and the ratios oI vortex shedding/natural Irequency and turbulent buIIeting/natural
Irequency were unacceptably high. This design also had to be relegated and the next attempt
was made with a J shell and double-segmental baIIles.
A satisIactory design emerged with this construction. The salient construction and
perIormance Ieatures oI this design are shown in Table 12.3b as well. Not only could the
shellside pressure be limited to 4.3 psi (0.3 kg/cm
2
), the design was saIe against Iailure oI
tubes due to Ilow-induced vibration. The crossIlow velocities were much lower than the
critical velocity and, although the ratios oI vortex shedding/natural Irequency and turbulent
buIIeting/natural Irequency were 0.95 and 1.362, respectively, the crossIlow amplitude was
so low as to render the design saIe. This design was, thereIore, accepted.
The principal diIIerence between the two designs, as Iar as Ilow-induced vibration is
concerned, is the unsupported span: 55.9 in. (1419 mm) in the single-segmental baIIle
design and 23.6 in. (600 mm) in the double-segmental baIIle design
No-tubes-in-winaow segmental baffles
There will be many situations when even a divided-Ilow shell with double-segmental baI-
Iles will not yield a satisIactory design. In such situations, a 'no-tubes-in-window¨ (see
Table 12.3a: Principal process parameters Ior Case Study 12.3
Shellside Tubeside
1. Fluid
Hydrocarbon gas Cooling water
2. Flow rate, lb/h (kg/h)
461,900 (209,500) 354,300 (160,700)
3. Temperature in/out, ƒF (ƒC)
122 (50)/104 (40) 91.4 (33)/104 (40)
4. Operating pressure, psia (kg/cm
2
a)
412 (29.0) 71 (5.0)
5. Allowable pressure drop, psi (kg/cm
2
)
4.3 (0.3) 10 (0.7)
6. Fouling resistance, h It
2
ƒF/Btu (h m
2
ƒC/kcal)
0.00098 (0.0002) 0.00195 (0.0004)
7. Density in/out, lb/It
3
(kg/m
3
)
1.47 (23.5)/1.49 (23.9)
8. Molecular weight
21.7
9. Inlet/outlet viscosity, cp
0.011/0.0105
10. Average speciIic heat, Btu/lb ƒF (kcal/kg ƒC)
0.538
11. Average thermal conductivity, Btu/h It ƒF (kcal/h m ƒC)
0.021 (0.031)



Standard
12. Heat duty, MM Btu/h (MM kcal/h)
4.5 (1.134)
13. Design pressure, psig (kg/cm
2
g)
512 (36) 100 (7.0)
14. Design temperature, ƒF (ƒC)
160 (70) 158 (70)
15. Material oI construction
CS CS
16. Nominal line size, in. (mm)
24 (600) 6 (150)
201
Fig. 3.3) design should be considered. This is a special design where, as the name
implies, there are no tubes in the window region; that is, tubes are only in the region
common to all the baIIles. Consequently, the unsupported tube span is now equal to the
baIIle spacing itselI, since all the baIIles support all the tubes. With normal single-
segmental baIIles, the unsupported tube span is twice the baIIle spacing, since tubes in
the window region are supported by every alternate baIIle.
Should it become necessary to use an extremely large baIIle spacing in a no-tubes-in-
window design to satisIy the shellside pressure drop limitation, the use oI intermediate
supports increases the natural Irequency oI the tubes, thereby ensuring a design that is
absolutely saIe against Iailure oI tubes due to Ilow-induced vibration. With the addition oI
intermediate supports, there is no appreciable increase in the shellside pressure drop because,
essentially, we have crossIlow at virtually the same velocity. This is the true Iorte oI the no-
tubes-in-window design.
The minimum number oI cross-passes in a no-tubes-in-window design should be three
Table 12.3b: Results oI Case Study 12.3
1 shell with single-
segmental baffles
1 shell with double-
segmental baffles
1. No. oI shells
One One
2. Shell ID, in. (mm)
38.6 (980) 39.4 (1000)
3. Total no. oI tubes
1086 1134
4. No. oI tube passes
6 6
5. Heat transIer area, It
2
(m
2
)
4282 (398) 4465 (415)
6. BaIIle spacing, in. (mm)
27.9 (709) 11.8 (300)
7. BaIIle cut, ° diameter
30 4 rows overlap
8. Overdesign, °
16.8 9.3
9. Shellside pressure drop, psi (kg/cm
2
)
4.3 (0.3) 4.3 (0.3)
10. Tubeside pressure drop, psi (kg/cm
2
)
9.4 (0.66) 8.7 (0.614)
11. Shellside heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/m
2
ƒC)
197.2 (963) 164.5 (803)
12. Overall heat transIer coeIIicient,
Btu/h It
2
ƒF (kcal/h m
2
ƒC)
101.2 (494) 91.6 (447)
Vibration Analysis (central region)
1. Critical velocity, It/s (m/s)
13.9 (4.22) 39.3 (11.98)
2. Average crossIlow velocity, It/s (m/s0
19.7 (6.01) 20.5 (6.24)
3. Vortex shedding ratio
2.6 0.95
4. Turbulent buIIeting ratio
3.725 1.362
5. Unsupported span, in. (mm)
55.9 (1419) 23.6 (600)
6. (Unsupported span)/TEMA max.
0.9 0.38
7. CrossIlow amplitude, in. (mm)
- 0.17
202
since this a basic assumption in the determination oI the mean temperature diIIerence
(MTD).
A no-tubes-in-window design is really the last recourse Ior a designer since such a
design requires a signiIicantly greater shell diameter Ior a given heat transIer area, thereby
increasing its cost, usually by about 1215°. This is because no tubes can be located in the
window regions oI the tube bundle. This deIiciency is oIIset, to some extent, by the higher
shellside heat transIer coeIIicient in a no-tubes-in-window design when compared to a
conventional tubes-in-window design because the Iormer employs pure crossIlow which is
more eIIicient than a combination oI crossIlow and window Ilow that is employed in the
latter.
CASE STUDY 12.4: PRODUCING A SAFE DESIGN USING
A NO-TUBES-IN-WINDOW DESIGN
Consider the service elaborated in Table 12.4a. This is a gas-gas heat exchanger and you
will notice that the MTD is very small and that there is a temperature cross. Since both
the streams are very clean, a Iixed-tubesheet construction may be employed. The
expected heat transIer area being very large, 39.4-It (12-m) long tubes oI 0.7874-in. (20-
mm) OD and 0.063 in. (1.6-mm) thick were to be used. The tube material is 304L
stainless steel and the shell material is carbon steel. It may be noted that the shellside
operating pressure is rather low at 132 psia (9.3 kg/cm
2
abs.). The Ilow rates oI both the
streams are very high; consequently, the line sizes are also high.
A single shell pass (TEMA E) design was Iound utterly inIeasible due to excessive
Table 12.4a: Principal process parameters Ior Case Study 12.4
Shellside Tubeside
1. Fluid
Hydrocarbon gas Hydrocarbon gas
2. Flow rate, lb/h (kg/h)
390,000 (177,000) 463,000 (210,000)
3. Temperature in/out, ƒF (ƒC)
55 (12.8)/82 (27.8) 95 (35)/77 (25)
4. Operating pressure, psia (kg/cm
2
a)
142 (10) 782 (55)
5. Allowable pressure drop, psi (kg/cm
2
)
4.3 (0.3) 2.8 (0.2)
6. Fouling resistance, h It
2
ƒF/Btu (h m
2
ƒC /kcal)
0.00098 (0.0002) 0.00098 (0.0002)
7. Density in/out, lb/It
3
(kg/m
3
)
0.5 (8.0)/0.48 (7.7) 3.5 (56.1)/3.6 (57.7)
8. Molecular weight
18.9 22.5
9. Average viscosity, cp
0.01 0.01
10. Average speciIic heat, Btu/lb ƒF (kcal/kg ƒC)
0.5 0.6
11. Average thermal conductivity, Btu/h It
o
F (kcal/h m
o
C)
0.012 (0.018) 0.01 (0.0155)
12. Heat duty, MM Btu/h (MM kcal/h)
5.16 (1.3)
13. Design pressure, psig (kg/cm
2
g)
206 (14.5) 867 (61.0)
14. Design temperature, ƒF (ƒC)
130 (54) 140 (60)
15. Material oI construction
Carbon steel SS 304L
16. Nominal shellside/tubeside line size, in. (mm)
24 (600) 16 (400)
203
shellside pressure drop.
A divided-Ilow shell (TEMA J) with single-segmental baIIles was considered next, and
the Iinal design that emerged is detailed in Table 12.4b. Although the shellside pressure drop
could be contained within the allowable limit, it was at the expense oI a very large baIIle
spacing, 38.5 in. (978 mm), which is considerably higher than the TEMA limit. Conse-
quently, the design had a very high probability oI Iailure oI tubes due to Ilow-induced
vibration and was, thereIore, rejected.
In the next design, the divided-Ilow (TEMA J) shell was retained and the baIIles
changed Irom single-segmental to double-segmental. The Iinal construction and perIor-
mance parameters are shown in Table 12.4b. This design was much better in that the
shellside pressure drop could be complied with despite a much lower baIIle spacing, 25 in.
(636 mm). The maximum unsupported span was a borderline 80° oI the TEMA limit. Since
Table 12.4b: Results oI Case Study 12.4
1 shell with single-
segmental baffles
1 shell with double-
segmental baffles
1. No. oI shells
One One
2. Shell ID, in. (mm)
47.2 (1200) 45.3 (1150)
3. Total no. oI tubes × no. oI tube passes
1470 × one 1431 × one
4. Tube pitch, in. (mm)
1.1 (28) ǻ 1.024 (26) ǻ
5. Heat transIer area, It
2
(m
2
)
11,664 (1084) 11,363 (1056)
6. BaIIle spacing, in. (mm)
38.5 (978) 25 (636)
7. BaIIle cut, ° diameter
35 5 rows overlap
pressure drop, psi (kg/cm
2
) 4.0 (0.28) 4.1 (0.29)
8. Shellside
heat transIer coeIIicient, Btu/h It
2
ƒF (kcal/m
2
ƒC)
86.4 (422) 88.8 (433.7)
9. Tubeside pressure drop, psi (kg/cm
2
)
1.3 (0.09) 1.3 (0.09)
10. Overall heat transIer coeIIicient, Btu/h It
2
ƒF
(kcal/h m
2
ƒC)
40.4 (197.2) 41.3 (201.5)
11. MTD, ƒF (ƒC)
11.7 (6.5) 11.5 (6.4)
12. Overdesign, °
7.03 5.49
13. Weight oI empty exchanger, lb (kg)
70,800 (32,100) 66,500 (30,150)
Vibration Analysis (central region)
1. Critical velocity, It/s (m/s)
14.04 (4.28) 26.4 (8.04)
2. Average crossIlow velocity, It/s (m/s)
28.2 (8.58) 29.1 (8.86)
3. Vortex shedding ratio
8.177 2.923
4. Turbulent buIIeting ratio
9.9 4.191
5. Unsupported span, in. (mm)
77 (1956) 50 (1272)
6. (Unsupported span)/TEMA max.
1.24 0.806
7. CrossIlow ȡv
2
,

lb/It sec
2
(kg/m sec
2
)


376 (558) 401 (596)
204
the average crossIlow velocity was greater than the critical velocity, this design too was
unsaIe against Iailure oI tubes due to Ilow-induced vibration.
Next, a no-tubes-in-window design was attempted with the shell type retained as
divided-Ilow (TEMA J). The Iinal construction and perIormance parameters are shown in
Table 12.4c. Notice that this design is perIectly saIe against the possibility oI Iailure oI tubes
due to Ilow-induced vibration. The shell inside diameter is somewhat high (53.15 in. or 1350
mm) since no tubes can be located in the window regions.
Take a comparative look at the three designs. The unsupported tube span has reduced
Table 12.4c: More results oI Case Study 12.4
1 shell with NTIW
single-segmental
baffles
E shell with NTIW
single-segmental
baffles
1. No. oI shells
One One
2. Shell ID, in. (mm)
53.2 (1350) 47.2 (1200)
3. Total no. oI tubes × no. oI passes
1538 × one 1200 × one
4. Tube pitch, in. (mm)
1.024 (26) ǻ 1.063 (27) ǻ
5. Heat transIer area, It
2
(m
2
)
12,191 (1133) 9523 (885)
6. BaIIle spacing, in. (mm)
38.3 (974) 154 (3912)
7. No. oI intermediate supports in each baIIle span
One 6
8. BaIIle cut, ° diameter
18 18
pressure drop, psi (kg/cm
2
) 4.1 (0.29) 4.3 (0.3)
9. Shellside
heat transIer coeIIicient, Btu/h It
2
ƒF
(kcal/m
2
ƒC)
77.2 (376.8) 56.1 (273.7)
10. Tubeside pressure drop, psi (kg/cm
2
)
1.2 (0.083) 1.54 (0.108)
11. Overall heat transIer coeIIicient, Btu/h It
2 o
F
(kcal/h m
2
ƒC)
37.7 (184) 34 (165.7)
12. MTD, ƒF (ƒC)
11.7 (6.5) 16.7 (9.3)
13. Overdesign, °
4.0 4.57
14. Weight oI empty exchanger, lb (kg)
76,300 (34,600) 65,000 (29,500)
Vibration Analysis (central region)
1. Critical velocity, It/s (m/s)
172.6 (52.6) 147.4 (44.92)
2. Average crossIlow velocity, It/s (m/s)
23.9 (7.28) 16.14 (4.92)
3. Vortex shedding ratio
0.367 0.365
4. Turbulent buIIeting ratio
0.526 0.476
5. Unsupported span, in. (mm)
19.2 (487) 22.0 (559)
6. (Unsupported span)/TEMA max.
0.309 0.354
7. CrossIlow ȡv
2
,

lb/It sec
2
(kg/m sec
2
)


272 (403) 126 (186)
8. CrossIlow amplitude, in. (mm)
0.00016 (0.004) 0.00016 (0.004)
205
Irom 77 in. (1956 mm) in the Iirst design to 50 in. (1272 mm) in the second, to a mere 19.2
in. (487 mm) in the third.
All three designs indicated a strong possibility oI acoustic vibration in the Iirst or
Iundamental mode; consequently, a set oI deresonating baIIles will have to be employed to
eliminate acoustic vibration. We will discuss acoustic vibration in the Section 12.6.
So Iar, a very important point has not been mentioned which the discerning reader may
have noticed: as there is a temperature cross, a TEMA J shell will impose a penalty in the
MTD oI the heat exchanger since one halI oI the tube length is in countercurrent Ilow and
the other halI in co-current Ilow. In the present case, as there is a signiIicant cross, the F
t

Iactor comes to just 0.7, which is a huge penalty.
Now, as long as a regular tubes-in-window design was being considered, a J shell had to
be used because oI the shellside pressure drop limitation. However, once we move to a no-
tubes-in-window (NTIW) design, this limitation is no longer valid. Logically, thereIore, an E
shell should be considered here, as it will oIIer a big improvement in the MTD. Accordingly,
such a design was tried and a much cheaper design emerged. The parameters are presented
in Table 12.4c. Three cross-passes with six intermediate supports provided a design that is
absolutely saIe against the possibility oI Iailure oI tubes due to Ilow-induced vibration.
Let us compare this design with the previous one, that is, the NTIW design with the J
shell. As a result oI the much lower crossIlow velocity, the shellside and, thereby, the overall
heat transIer coeIIicient are signiIicantly lower. However, thanks to the large increase in the
MTD, the heat transIer area has reduced Irom 12,191 It
2
(1133 m
2
)

to 9523 It
2
(885 m
2
) and
the empty exchanger weight Irom 76,300 lb (34,600 kg) to 65,000 lb (29,500 kg) when
compared to the J shell NTIW design. Thus, the E shell NTIW design was selected as the
most economical design.
12.5 Rod Baffles
Rod baIIles are another means oI eliminating Iailure oI tubes due to Ilow-induced
vibration. In this concept, all the tubes in a heat exchanger are supported in all Iour
directions by a matrix oI rods called a baIIle set (see Fig. 13.4). The diameter oI the rods
is equal to the gap between the tubes, so that the rods actually touch the tubes and pin
them on all Iour sides, thus restraining any movement.
Because oI the inherent unobstructed Ilow path created by the rods, the Ilow pattern is
predominantly longitudinal. Since the Ilow direction is not repetitively changed, as it is with
plate baIIles, velocity proIiles are more uniIorm. Furthermore, stagnant regions are virtually
eliminated. As a result oI the above, the Iollowing additional beneIits can be realized by the
use oI rod baIIles in appropriate situations:
a) A higher conversion oI pressure drop to heat transIer on the shellside. This can
be employed to obtain a lower pressure drop Ior the same heat transIer
coeIIicient, or a higher heat transIer coeIIicient Ior the same pressure drop.
b) Reduced Iouling due to the absence oI stagnant regions
Rod-baIIle exchangers have been employed predominantly Ior services handling gases or
condensing/vaporizing streams, since such services have a high possibility oI Iailure oI
tubes due to Ilow-induced vibration.
12.6 Acoustic Vibration
Acoustic vibration in shell-and-tube heat exchangers involves exciting the acoustic
Irequency oI the shell cavity to create noise. It is a consequence oI a Irequency match be-
206
tween an exciting Irequency, either due to vortex shedding or turbulent buIIeting, and the
acoustic Irequency oI the shell. When this happens, resonance occurs and the kinetic
energy oI the Ilow stream is converted into sound pressure waves.
Acoustic resonance occurs due to gas column oscillation which is a molecular
phenomenon. The oscillation occurs perpendicular to both the direction oI Ilow and the tube
axis, and produces acoustic vibration oI a standing wave typeprecisely what happens to
the air column inside a Ilute when it is played.
The characteristic Irequency oI acoustic vibration in a shell-and-tube heat exchanger
depends upon a characteristic dimension (which is the shell inside diameter) and the velocity
oI sound in the shellside Iluid medium. A correction has to be added Ior the presence oI
tubes in the shell, as well as the tube layout angle.
The acoustic resonant Irequency I
ac
can be expressed by the Iollowing equation
I
ac
÷ m u
sound
/2d (12.4)
where
m ÷ the mode number, a dimensionless integer
u
sound
÷ velocity oI sound
d ÷ shell diameter
It will be seen that the acoustic resonant Irequency is directly proportional to the velocity
oI sound and inversely proportional to the shell diameter. The velocity oI sound is
inversely proportional to the square root oI the gas or vapor molecular weight.
When m ÷ 1, the lowest acoustic Irequency is achieved and the characteristic length is
equal to the shell inside diameter. This is known as the Iundamental tone. Higher overtones
(m ÷ 2, 3, or 4) vibrate at Irequencies two, three, or Iour times the Iundamental Irequency, as
shown in Fig. 12.4.
In a shell Iilled with tubes, the tube layout angle deIines the open lanes where the
standing wave Iormation is most probable. Figure 12.5 shows a Iirst-mode standing wave in
both an in-line (square) and staggered (rotated square) tube arrangement. The shaded
portions represent areas where resonance can occur.
The acoustic Irequencies oI a heat exchanger can be excited by either vortex shedding or
turbulent buIIeting when these Irequencies are
between 0.8 and 1.2 oI the acoustic
Irequency. When this happens, a loud noise
may be produced. Acoustic vibration is not
Fig. 12.4 Typical wave Iorms (Reprinted Irom the
Heat Exchanger Design Handbook, 2002 with
permission oI Begell House, Inc.)

Fig. 12.5 First-mode standing wave Iormation |5|
(Reproduced with permission. Copyright 1973
AIChE. All rights reserved.)
207
destructive unless the acoustic Irequency is in
resonance with the natural Irequency oI the
tubes, which is very rare. However, the noise
produced is very unpleasant, so that designers
should strive to eliminate the possibility oI
acoustic vibration in the design stage itselI.
II the probability oI acoustic vibration is
indicated while carrying out a thermal design,
the designer should examine an alternative
tube layout angle. The use oI a 60
o
tube layout
angle oIten eliminates the probability oI
acoustic vibration. Although not well under-
stood, most problems oI acoustic vibration
have occurred in heat exchangers employing a
45
o
tube layout angle. Hence, the use oI this
conIiguration should be avoided.
II the probability oI acoustic vibration cannot be eliminated otherwise, deresonating
baIIles should be incorporated in all the cross passes since acoustic vibration can occur at
any length oI the shell. These are Ilat plates set parallel to both the centerline oI the tubes and
the plane oI crossIlow, and are welded to the baIIles/tubesheet on either side. Thus, iI a heat
exchanger is horizontally mounted and has horizontally cut baIIles, the deresonating baIIles
will be in the vertical plane.
The most eIIective location oI deresonating baIIles is, theoretically, at the antinode oI
the displacement wave (where the gas particle displacement is maximum). However, actual
experience has shown that the deresonating baIIle need not be located precisely at the
antinode in order to be eIIective, but can be located within a Iew tube rows oI the antinode.
The advantage with this practice is that a single deresonating baIIle can be employed to
prevent more than one mode oI acoustic vibration. An example oI this is shown in Fig. 12.6
where several resonant modes (m ÷ 1, 2, and 3) are eliminated in a heat exchanger
employing a rotated square pitch.
With a square tube pitch, the prevention oI Iirst-mode vibration is quite straightIorward,
as shown in Fig. 12.7a. However, the situation is somewhat more diIIicult Ior preventing
acoustic vibration in the higher modes. Placing the Iirst deresonating baIIle as shown in Fig.
12.7b results in the generation oI an intermediate wave length and Irequency, so that a
second deresonating baIIle has to be located as shown. Thus, two deresonating baIIles will
have to employed. The idea is to create acoustic Irequencies, quite diIIerent Irom the Iorcing

Fig. 12.6 Location oI deresonating baIIles Ior
preventing acoustic vibration in a rotated square
tube layout conIiguration |5| (Reproduced with
permission. Copyright 1973 AIChE. All rights
reserved.)

(a) (b)
Fig. 12.7 Location oI deresonating baIIles Ior preventing acoustic vibration in a square tube layout conIigu-
ration |5| (Reproduced with permission. Copyright 1973 AIChE. All rights reserved.)
208
Irequencies due to vortex shedding or turbulent buIIeting, in each segment Iormed by the
deresonating baIIles.
For a detailed account oI acoustic vibration in shell-and-tube heat exchangers, please
reIer to |5| and |6|.
References
|1| MacDuII, J.N., and Felgar, R.P., 1957, 'Vibration Design Charts,¨ Trans. ASME, 79, (Oct)
pp. 1459²1474.
|2| Chen, Y.N, 'Flow-induced Vibration and Noise in Tube-bank Heat Exchangers Due to Von
Karman Streets,¨ Trans. ASME, J. Eng. Ind., 90, pp. 134146.
|3| Connors, H.J., 1970, 'Fluidelastic Vibration oI Tube Arrays excited by Cross Flow,¨ Symp.
on Flow Inaucea Jibration in Heat Exchangers, ASME Annual Winter Meeting, New York.
|4| Tubular Exchanger ManuIacturers Association, (1999), Stanaaras of the Tubular Exchanger
Manufacturers Association, 8th Eaition, TEMA, New York.
|5| Barrington, E.A., 1973, 'Acoustic Vibrations in Tubular Exchangers,¨ Chemical Engineering
Progress, 69(7), pp. 62²68.
|6| Barrington, E.A., 1978, 'Cure exchanger acoustic vibration,¨ Hyarocarbon Processing, July,
pp. 193²198.
|7| Hewitt, G.F. (ed.), 1998, Heat Exchanger Design Hanabook 1998, Vols. 3 and 4, Begell
House, Inc., New York.
209
CHAPTER 13
(QKDQFHG+HDW7UDQVIHU
So Iar in this book, we have dwelt on heat transIer in conventional shell-and-tube heat
exchangers having bare tubes oI circular cross section and plate baIIles perpendicular to
the axis oI the tubes (hereaIter only called conventional shell-and-tube heat exchangers).
Now, conventional shell-and-tube heat exchangers are so pervasive in the chemical
process industries that the reader may well assume them to embody eIIicient heat
transIer. UnIortunately, this is Iar Irom the truth because both oI these elements, the bare
circular tube and the perpendicular plate baIIle, invariably represent signiIicant
limitations as Iar as eIIicient heat transIer is concerned. This is because, near the tube wall,
there is a thermally ineIIicient boundary layer where there is very little or no mixing. On the
shellside, the Iormation oI eddies, crossIlow Irom one cross pass to another, and dead spaces
adjacent to each baIIle are ineIIicient Ilow phenomena which result in poor heat transIer
eIIiciency and aggravated Iouling. By modiIying the bare circular tube and/or the perpen-
dicular baIIle, it is possible to realize signiIicant improvement in heat transIer. In this
chapter, we will look at enhanced heat transIer as applicable to shell-and-tube heat
exchangers and discuss, brieIly, enhanced heat transIer in non-tubular heat exchangers.
13.1 What is Enhanced Heat Transfer?
BeIore answering this question, let us Iirst talk about eIIicient heat transIer. By its very
nature, convective heat transIer is always achieved at the expense oI Iluid Ilow and,
thereby, pressure drop. Now, the higher the heat transIer coeIIicient, the smaller will be
the heat transIer area and the lower the Iirst cost oI the heat exchanger. At the same time,
the higher the pressure drop, the higher will be the operating cost. ThereIore, it is vitally
important to minimize pressure drop Ior a given heat transIer coeIIicient so that the
overall cost oI the heat exchanger, which is the sum oI the Iixed cost and the operating
cost, is low. The eIIiciency oI heat transIer may, thereIore, be deIined in terms oI the
conversion oI pressure drop to heat transIer.
Because oI their inherent deIiciencies described above, conventional shell-and-tube heat
exchangers do not have a high eIIiciency oI heat transIer. This is not to detract Irom the
useIulness and versatility oI these heat exchangers since they are still oIten the best bet Ior
many applications in the chemical process industries, especially Ior high-pressure and high-
temperature services. Nonetheless, there are many services where alternate heat exchanger
constructions deserve serious consideration.
We now come to the term 'enhanced heat transIer.¨ This is commonly understood to be
an enhancement or increase in the heat transIer coeIIicient. Now, an increase in the heat
transIer coeIIicient can be achieved even in a conventional shell-and-tube heat exchanger by
an increase in Iluid velocity, but that would hardly qualiIy as heat transIer enhancement
210
since the increase in pressure drop will be Iar greater than the increase in the heat transIer
coeIIicient: consequently, the eIIiciency oI conversion oI pressure drop to heat transIer will
actually decrease.
What enhanced heat transIer really implies is an enhancement oI the heat transIer
coeIIicient for a given pressure arop. In other words, enhanced heat transIer produces a
higher eIIiciency oI conversion oI pressure drop to heat transIer. The key to enhanced heat
transIer is higher turbulence which is produced not by an increase in the Iluid velocity, but
by an improvement in the basic Ilow pattern Ior the application at hand. The increased
turbulence leads to not only enhanced heat transIer, but reduced Iouling as well. This is not
surprising when one considers that the ineIIicient Ilow phenomena mentioned at the
beginning oI this chapter are the root cause oI both ineIIicient heat transIer and excessive
Iouling. In Iact, in many troublesome situations, a reduction in Iouling can be even more
desirable than an enhancement in heat transIer.
As enhanced heat transIer produces a 'higher¨ eIIiciency oI conversion oI pressure drop
to heat transIer, it is a relative term and a datum has to be deIined. As the reader may easily
understand by now, the conventional shell-and-tube heat exchanger Iorms the datum with
respect to which enhanced heat transIer is assessed.
One important point to bear in mind here is that although an enhanced heat transIer
product will produce a higher heat transIer coeIIicient and, thereby, a lower heat transIer
area Ior a given pressure drop (oI one or both oI the streams), the Iirst cost oI the enhanced
heat transIer heat exchanger will depend not just upon the heat transIer area, but also upon
its cost per unit area. For the Iirst cost to reduce as compared to a conventional shell-and-
tube heat exchanger, the reduction in heat transIer area should be greater than the increase in
the cost per unit area.
The term heat transfer augmentation is oIten employed Ior enhanced heat transIer or
heat transIer enhancement.
13.2 Benefits of Enhanced Heat Transfer
The principal beneIits oI enhanced heat transIer are:
1) Reduce the Iirst cost oI a heat exchanger
By virtue oI the higher overall heat transIer coeIIicient, the size and weight oI the
heat exchanger will be lower with corresponding saving in structural steel,
thereby reducing the Iirst cost.
2) Save pumping power
The cost oI a heat exchanger embodying enhanced heat transIer can be kept the
same as that oI a conventional STHE, at the expense oI a much lower pressure
drop on the side which has been enhanced (tubeside, shellside, or both). This will
result in a direct saving in pumping power oI the corresponding stream/s.
3) Reduced Iirst cost and pumping power
By not taking Iull advantage oI the possible enhancement in the heat transIer
coeIIicient, it is possible to obtain an advantage in both the Iirst cost and the
pumping cost. Evidently, the reduction in the Iirst cost will be less than that in
beneIit 1 above and the saving in pumping power will be less than that in beneIit
2 above.
211
4) Use lower MTD and, thereby, cheaper utility
Since the overall tubeside heat transIer coeIIicient is higher with enhanced heat
transIer, the MTD can be correspondingly lower Ior the same heat transIer area.
Consequently, in a steam-heated exchanger, lower pressure steam (which is
cheaper) can be used. This will have the added advantage oI reducing the tube-
wall temperature, which is a bonus Ior a Iouling application since the extent oI
Iouling will be less. Similarly, in the case oI reIrigerants, cheaper cold utility can
be used Ior low-temperature coolers such as reIrigerated coolers.
5) Reduce number oI shells
As discussed in detail in Chapter 6, services which have a temperature cross
(where the cold stream is heated to a temperature greater than the outlet tempera-
ture oI the hot stream) require two or more shells in series, depending upon the
extent oI the temperature cross. II the use oI enhanced heat transIer results in a
reduction in the number oI tubes to such an extent that only one tube will suIIice,
true countercurrent Ilow will be realized and, consequently, only one shell
required. This will result in considerable economy.
6) Cost-eIIective in revamp applications
A very potent beneIit oI enhanced heat transIer is in revamps (handling higher
Ilow rates and heat duty) where, instead oI supplementing or replacing existing
heat exchangers by bigger ones, only the existing tube bundles are replaced by
new tube bundles embodying enhanced heat transIer (Ior example, a new tube
bundle with twisted tubes or helical baIIles). This would eliminate the cost oI
structural and piping modiIications and associated downtime, which can be very
signiIicant.
7) Reduce Iouling
Due to the elimination oI the plain circular tube and the perpendicular plate baI-
Ile, and due to the increase in turbulence produced by the enhancement device,
the extent oI Iouling will be appreciably less than that in a conventional STHE.
13.3 Heat Transfer Enhancement Techniques
In order to produce more eIIicient and, thereby, cheaper heat exchangers than
conventional shell-and-tube heat exchangers, three broad methodologies Ior heat transIer
enhancement have been pursued.
The Iirst methodology has been to retain the basic shell-and-tube conIiguration and
develop more eIIicient tubes or baIIling. Some oI the more common examples oI this are:
• low-Iin tubes
• high-Ilux tubes (Ior low-temperature-diIIerence boiling applications)
• corrugated tubes
• tube inserts
• rod baIIles
• helical baIIles
• twisted-tube heat exchangers
The second methodology has been to move over Irom the shell-and-tube heat exchanger
and produce basic innovations in the exchanger type itselI. Examples oI this are:
212
• plate heat exchangers
• spiral plate heat exchangers
• plate-Iin heat exchangers
• printed-circuit heat exchangers
The third and Iinal methodology, the most recent one, has been to create a hybrid oI the
STHE and the plate heat exchanger (PHE) so as to derive the advantages and eliminate
the limitations oI each.
As mentioned earlier, this book being on shell-and-tube heat exchangers, we will take a
detailed look at the Iirst methodology (tubular exchangers) and a brieI look at the other two
(non-tubular exchangers).
13.3.1 Low-fin tubes
Gases exhibit extremely low heat transIer coeIIicients due to their low thermal
conductivity and low density. (An exception is hydrogen, which has a thermal
conductivity oI the order oI hydrocarbon liquids.) Viscous liquids also exhibit poor heat
transIer coeIIicients due to very poor turbulence. When a bare tube heat exchanger has
widely diIIerent heat transIer coeIIicients on the two sides, the use oI Iinned tubes
increases the overall heat transIer coeIIicient considerably. Evidently, the Iins will have
to be on the side that has the poor heat transIer coeIIicient. The usual situation is a low
heat transIer coeIIicient on the shellside which makes it highly advantageous to employ
externally Iinned tubes. The most common example oI this is the air-cooled heat
exchanger which has air Iorced across high-Iinned tubes having aluminum Iins varying
Irom 1/2 in. (12.7 mm) to 5/8 in. (16 mm) in height. The surIace ratio (which is the ratio
oI total Iinned and bare tube area to bare tube area) typically varies Irom 8.5 Ior 1-in.
(25.4-mm) tube OD, ½-in. (12.7-mm) Iin height, and 5 Iins/inch (197 Iins/m) to 23.5 Ior
1-in. (25.4-mm) tube OD, 5/8-in. (16-mm) Iin height, and 11 Iins/inch (433 Iins/m).
Low-Iin tubes (Fig. 8.5) come in a staggering array oI Iin height and density with a
surIace ratio varying Irom just under 3.0 to about 4.5. These are employed in shell-and-tube
heat exchangers and are advantageously employed in applications where:
a) the shellside Iluid is relatively clean
b) the shellside heat transIer coeIIicient is relatively low
c) the tubeside heat transIer coeIIicient is high
a) the tubeside Iouling resistance is low, because otherwise, an increase in the
eIIective tubeside Iouling resistance will neutralize the increase in the eIIective
shellside heat transIer coeIIicient. The major applications are gas coolers and
condensation oI low surIace tension Iluids such as reIrigerants.
As low-Iinned tubes undergo severe cold working, it is important to ensure that the base
tube material is oI a suitable grade.
Internally-Iinned tubes can be employed advantageously in situations where the tubeside
heat transIer resistance is predominant.
13.3.2 High-flux tubes
Boiling heat transIer coeIIicient is a very strong Iunction oI the temperature diIIerence
across the boiling Iilm, which in turn depends upon the temperature diIIerence between
the hot and the cold streams. At very low temperature diIIerences across the boiling Iilm,
213
such as 3.65.4ƒF (23ƒC), bare tubes cannot sustain nucleate boiling and heat transIer
occurs by natural convection, which is very ineIIicient. Special sintered tube surIaces
have been developed which incorporate stable reentrant cavities. These surIaces produce
nucleate boiling, even at extremely low temperature diIIerences, thereby augmenting the
boiling heat transIer coeIIicient severalIold.
With such tubes, it is possible to accomplish a given heat duty with a much lower heat
transIer surIace than is possible with bare tubes. Alternatively, with the same heat transIer
area, a much lower temperature diIIerence can be employed Ior achieving the same heat
duty. For chiller services involving the boiling oI reIrigerants in ethylene or other plants, this
translates into savings in compression power.
It oIten happens that the boiling heat transIer coeIIicient is enhanced so much by the use
oI high-Ilux tubes that the heating medium heat transIer coeIIicient becomes controlling. In
such cases, it becomes advantageous to employ doubly-enhanced tubes. For example, in a
vertical thermosyphon reboiler having a condensing heating medium, OD Iluted tubes and
ID high-Ilux coated tubes can be employed.
High-Ilux tubes are also covered in Chapter 8.
13.3.3 Corrugated tubes
Corrugated tubes (Fig. 13.1) induce turbulence on both the tubeside and shellside Iluids,
thus ensuring high Reynold`s numbers, even at low velocity. This results in signiIicantly
enhanced perIormance over smooth tubes, with the resulting advantages oI lower size and
space, reduced Iouling, shorter residence time and lower hold-up, closer temperature
approaches, and uniIorm temperature distribution, even Ior viscous liquids. Surprisingly
these heat exchangers are not commonly employed in the chemical process industries at
present.
13.3.4 Tube inserts
13.3.4.1 Twisted tape inserts
The twisted tape metal insert (Fig. 13.2) is easy to make and Iit. By creating a swirling
Ilow pattern, they result in increased exposure oI the Iluid to the inner heat transIer
surIace with a minimum pressure drop penalty. However, iI the Iit is too loose, the Iin
eIIect and some oI the swirling action oI the Iluid is lost. Damage to the tubes may also
be caused by being repeatedly hit by
the tapes. II the Iit is too tight,
insertion oI the tapes into the tubes
may be diIIicult.
Since they are very simple in
construction, twisted tape inserts

Fig. 13.1 Corrugated tubes (a) spirally indented and (b) spirally Iluted |8| (Courtesy oI ESDU International,
www.esdu.com)
Fig. 13.2 Twisted tape tube insert
214
might appear to be cost-eIIective. However, perIormance data oI twisted-tape inserts are not
readily available in terms oI heat transIer augmentation, pressure drop, and Iouling
propensity and, as such, evaluation is diIIicult. Besides, whereas these inserts only disrupt
the laminar Ilow regime near the tube walls, the much more scientiIically designed wire-Iin
inserts not only disrupt the laminar Ilow regime near the tube-walls, but also mix the bulk
Iluid. Since both are needed Ior eIIective heat transIer enhancement, wire-Iin tube inserts are
superior, even Irom a Iundamental standpoint.
13.3.4.2 Wire-fin tube inserts
Wire-Iin tube inserts (Fig. 13.3) are manuIactured Irom Iormed wire loops, spaced
radially and axially within the tube, and supported Irom a central core. The inserts
continually remove low-velocity or stagnant Iluid Irom the tube-wall and replenish it with
Iluid Irom the center oI the tube. By doing this, wire-Iin tube inserts minimize the eIIect
oI Irictional drag, thereby preventing the Iormation oI a stable boundary layer and
resulting in dramatically higher rates oI heat transIer Ior a given pressure drop. This
increase can vary Irom 2 to even 25 times, depending upon the Reynold`s number in the
bare tube situation: the lower the Reynold`s number, the greater the enhancement. In
addition, shorter residence time and reduced temperature diIIerence due to higher
tubeside heat transIer coeIIicient can eliminate thermally dependent causes oI Iouling.
Now, Ior the same length oI travel and velocity, these inserts will evidently result in an
increased tubeside pressure drop that may be unacceptable. However, the number oI tube
passes can be decreased appropriately such that the heat transIer coeIIicient is appreciably
higher than that obtained with bare tubes, while the pressure drop is still within the allowable
limit. Thus, the eIIiciency oI conversion oI pressure drop to heat transIer is higher with these
inserts under laminar Ilow conditions.
When wire-Iin tube inserts were Iirst applied, the principal beneIit was expected to be
the improved heat transIer characteristics. However, it was Iound that the extent oI Iouling
reduced dramatically. This is not surprising, considering that the boundary layer separation
and the absence oI eIIective mixing (between the Iluid at the tube-wall and that at the center
oI the tube) are the principal culprits responsible Ior both ineIIicient heat transIer and
aggravated Iouling. The improved mixing results in:
1) a shorter residence time oI the Iluid within the tube
2) a lower temperature diIIerence between the tube metal and Iluid Iilm
thereby reducing the Iouling propensity appreciably, especially Ior thermally-dependent
applications.
Aavantages of wire-fin tube inserts
Summarizing the above, wire-Iin tube
inserts oIIer several advantages.
1) They can be used to reduce the Iixed
cost oI a heat exchanger or to save pumping
power on the tubeside.
2) As the tubeside heat transIer coeIIicient
will be higher with inserts, the MTD can be
correspondingly lower Ior the same heat
transIer area. Consequently, in a steam-
heated exchanger, lower pressure steam
Fig. 13.3 Wire-Iin tube insert (© Cal Gavin Ltd.
Reprinted with permission)
215
(which is cheaper) can be used. This will have the added advantage oI reducing
the tube-wall temperature, which is a bonus Ior a Iouling application, since the
extent oI Iouling will be less.
3) II the use oI tube inserts results in such a reduction in the number oI tubes, to
such an extent that only one tube pass will suIIice, true countercurrent Ilow will
be realized and, consequently, only one shell used, even in temperature-cross
situations.
4) A very potent beneIit is in revamps where instead oI supplementing or replacing
existing heat exchangers, they are modiIied to have a lower number oI tube
passes and Iitted with appropriate inserts, thereby eliminating the cost oI
structural and piping modiIications and associated downtime which can be very
signiIicant.
5) Another very signiIicant advantage oI using tube inserts is the reduction or even
elimination oI thermally-dependant causes oI Iouling by virtue oI radial mixing,
shorter residence time, and reduced temperature diIIerence across the tube. In
Iact, where excessive Iouling plagues the perIormance oI a heat exchanger,
mitigation oI Iouling may be the prime advantage sought. However, since dirty
liquid streams are generally also viscous, the beneIits oI enhanced heat transIer
and reduced Iouling in such situations usually go hand-in-hand. A notable
description oI the mitigation oI tubeside Iouling is described in |1|.
CASE STUDY 13.1: COMPARISON OF DESIGNS WITH BARE TUBES
AND TUBES WITH WIRE-FIN TUBE INSERTS
Let us consider the thermal design Ior a service whose principal process and other para-
meters are speciIied in Table 13.1a. This heat exchanger is Ior a crude oil-crude oil ser-
Table 13.1a: Principal process parameters Ior Case Study 13.1
Service Shellside Tubeside
1. Type oI exchanger
Floating-head (TEMA type AES)
2. Fluid
Crude oil Crude oil
3. Flow rate, lb/h (kg/h)
625,000 (283,500) 568,500 (257,850)
4. Temperature in/out, ƒF (ƒC)
86 (30)/145.4 (63) 194 (90)/122 (50)
5. Heat duty, MM Btu/h (MM kcal/h)
20.48 (5.16)
6. Total allowable pressure drop,
shellside ¹ tubeside, psi (kg/cm
2
)
44 (3.1)
7. Viscosity in/out, cp
200/43 20/80
8. Density in/out, lb/It
3
(kg/m
3
)
55.2 (885)/54 (865) 53.1 (850)54.5 (/873)
9. Average speciIic heat, Btu/lb ƒF (kcal/kg ƒC)
0.55 0.5
10. Avg. thermal conductivity, Btu/h It ƒF
(kcal/hr m ƒC)
0.072 (0.107) 0.07 (0.104)
11. Fouling resistance, h It
2
ƒF/Btu (h m
2
ƒC/kcal)
0.00195 (0.0004) 0.00195 (0.0004)
12. Material oI construction
Carbon steel Carbon steel
216
vice Ior a crude oil desalter project. It will be noticed that the viscosity oI the crude oil
streams is rather high. As has been reasoned earlier, the lower viscous crude oil stream is
routed through the tubeside. Since a Iairly dirty service is involved, a Iloating-head con-
struction was to be employed. Carbon steel was to be the material oI construction Ior the
entire heat exchanger. Since both oI the crude oil streams were to be pumped by the same
pump, a total (shellside ¹ tubeside) allowable pressure drop was speciIied: 44.1 psi (3.1
kg/cm
2
).
A thermal design was produced using bare tubes. The total heat transIer surIace was
69,984 It
2
(6504 m
2
) and twelve shells had to be used, three in series and Iour in parallel. As
Table 13.1b: Results oI Case Study 13.1

Bare tube design
Design with wire-fin
tube inserts
In series 3 2
1. No. oI shells
In parallel 4 4
2. Shell inside diameter, in. (mm)
43.3 (1100) 987
3. Tube OD × thickness × length, in. (mm)
0.984 (25) × 0.0984 (2.5) ×
354 (9000)
1.0 (25.4) × 12 BWG
(2.77) × 217 (5500)
4. No. oI tubes per shell
814 736
5. No. oI tube passes
4 2
6. Tube pitch, in. (mm)
1.25 (31.25) rotated square
7. BaIIle spacing, in. (mm)
12.6 (320) 13.6 (346)
8. BaIIle cut, ° (diameter)
25 25
9. Heat transIer area per shell, It
2
(m
2
)
5832 (542) 3389 (315)
10. Total heat transIer area, It
2
(m
2
)
69,984 (6504) 27,115 (2520)
Shellside 15.6 (1.1) 6.5 (0.46)
11. Pressure drop,
psi (kg/cm
2
)
Tubeside 28 (1.93) 27 (1.9)
Shellside 44.6 (217.9) 52.8 (258)
Tubeside 9.9 (48.4) 43.2 (211)
12. Heat transIer
coeIIicient,
Btu/h It
2 o
F
(kcal/h m
2

o
C)
Overall 7.8 (38.1) 21.4 (104.3)
A, baIIle hole-to-tube
leakage
0.018 0.021
B, Main crossIlow 0.561 0.693
C, bundle-shell bypass 0.07 0.08
E, baIIle-shell bypass 0.351 0.205


13. Shellside Ilow
Iractions
F, pass-partition bypass 0 0
14. Temperature proIile distortion correction Iactor
0.98 0.98
15. MTD, ƒF (ƒC)
39.6 (22.0) 36.7 (20.4)
16. Overdesign, °
5.88 4.07
217
the tubeside heat transIer coeIIicient was extremely low, an alternate design was investigated
using wire-Iin tube inserts. As expected, the total heat transIer area reduced considerably
Irom the bare tube design, Irom 69,983 It
2
(6504 m
2
) to 27,115 It
2
(2520 m
2
), a reduction oI
about 61°. AIter incorporating the cost oI the inserts, the cost oI the design with inserts was
only 55° oI the bare-tube design. The salient construction Ieatures oI both the designs are
indicated in Table 13.1b.
The principal diIIerence between the two designs is, oI course, the tubeside heat transIer
coeIIicient. In the bare tube design, it is only 9.7 Btu/It
2
ƒF (47.4 kcal/h m
2
ƒC), whereas in
the design employing wire-Iin inserts, it is 44.2 Btu/It
2
ƒF (216 kcal/h m
2
ƒC), which is 4.56
times higher. The overall heat transIer coeIIicient increases Irom 7.64 Btu/It
2
ƒF (37.3 kcal/h
m
2
ƒC) to 21.7 Btu/It
2
ƒF (106 kcal/h m
2
ƒC), which is 2.84 times higher.
Although the design with inserts has a better stream analysis, it has a lower temperature
proIile distortion correction. This may appear to be anomalous, until the designer observes
that the number oI shells in series is three in the bare tube design but two in the design with
inserts.
13.3.5 RODbaffle heat exchangers
The Iirst RODbaIIle heat exchanger (RBHE) (Fig. 13.4) tube bundles were built in the
early 1970s to replace existing double segmental, plate-baIIled tube bundles that had
Iailed due to Ilow-induced vibration. The RBHE is a shell-and-tube heat exchanger
employing an array oI rods Ior supporting tubes.
These rods, with a diameter equal to the clearance between the tube rows, are welded to
baIIle support rings so that alternating tube rows are supported by the rods in each baIIle.
(Tubes can only be laid out on a square or a rotated square pattern.) Thus, a set oI Iour
baIIles , two having vertical support rods and two having horizontal support rods ,
constitutes one baIIle set which supports the tubes in all Iour directions and, thereby,
eliminates vibration oI the tubes. There will be several such sets to cover the entire tube
length. Individual support baIIles are typically spaced on 152 mm centers, such that the
maximum unsupported tube span over a Iour-baIIle set is 608 mm.
Although originally developed to eliminate tube Iailures due to Ilow-induced vibration,
RBHEs were later Iound to possess additional beneIits, the principal one being an increased
ratio oI heat transIer to pressure drop, as compared to ordinary plate-baIIle STHEs. This is
attributable to the essentially longitudinal Ilow on the shellside oI an RBHE. Another
advantage with RBHEs is lower Iouling on the shellside due to the elimination oI low Ilow

Fig. 13.4 RODbaIIle heat exchanger (a) baIIle arrangement and (b) bundle detail (Courtesy oI
ConocoPhillips.)
218
or stagnant areas associated with plate baIIles.
Thus RBHEs are used where a low pressure drop is permitted on the shellside, and
conventional baIIles cannot produce a design that is saIe against Iailure oI tubes due to Ilow-
induced vibration.
Occasionally, it becomes impractical to make these heat exchangers long enough to
consume the allowable shellside pressure drop. This is because oI the longitudinal Ilow
pattern on the shellside, which tends to produce low velocity and thereby low pressure drop.
In such cases, a 2-shellpass, F-shell RODbaIIle exchanger design should be considered.
Major industrial applications oI RBHEs include gas coolers, Ieed/eIIluent exchangers,
condensers, steam generators, and reboilers.
13.3.6 Helical baffles (Helixchangers)
A conventional shell-and-tube heat exchanger uses perpendicular plate baIIles to support
the tubes, as well as to direct the shellside stream to Ilow across the tubes at a desirable
velocity. The tortuous zigzag Ilow on the shellside gives a relatively high pressure drop
per unit oI heat transIer, since a good part oI the pressure drop is wasted in reversing the
Ilow direction rather than in generating good heat transIer. As the Ilow passes a baIIle,
the high crossIlow pressure drop Iorces the Iluid to leak through the gaps between the
shell and the baIIle (E stream) and the baIIle holes and the tubes (A stream). This has two
adverse eIIects: the MTD is lowered due to the distortion oI the shellside temperature
proIile, and the shellside heat transIer coeIIicient is also lowered due to reduction in the
Ilow rate oI the main crossIlow stream.
Besides, because oI the very nature oI the Ilow on the shellside, there will be dead
spaces on either side oI every baIIle where Iouling will be pronounced. II these dead spaces
can be eliminated, the perIormance oI a heat exchanger will improve signiIicantly.
The Helixchanger (Fig. 13.5) was developed with this goal in mind. Each baIIle
occupies one quadrant oI the shell and has a certain inclination to the centerline oI the
exchanger, successive baIIles touching at the outside edges so as to Iorm a continuous helix.
The helical Ilow reduces shellside turning losses and, thereby, gives a higher shellside heat
transIer coeIIicient than a normal segmentally-baIIled heat exchanger Ior the same pressure
drop. Thus, in a heat exchanger where the shellside Iilm resistance is controlling, there will
be an appreciable increase in the overall heat transIer coeIIicient Ior the same pressure drop,
thereby resulting in a smaller and cheaper heat exchanger. Alternatively, Ior a given heat
transIer area, the shellside pressure drop will be appreciably lower than that in a conven-
tional segmentally-baIIled heat exchanger, thereby resulting in reduction in operating cost.
Aavantages of Helixchangers
Compared to conventional segmentally
baIIled heat exchangers, Helixchangers
oIIer the Iollowing advantages:
1) Higher eIIiciency oI conversion oI
pressure drop to heat transIer. This leads
to substantial saving in pressure drop,
resulting in reduced pumping power.
2) For Ieed preheat trains, such as Ior
crude, long residue, and vacuum gas
oils, the considerably lower Ieed pres-
Fig. 13.5 Helixchanger (© ABB Lummus Heat TransIer.
Reprinted with permission)
219
sure drop translates into reduced mechanical design pressure on the Ieed side oI
the preheat exchangers. This results in a reduction in the initial cost oI the
preheat exchangers.
3) Reduced investment costs Ior services where the shellside heat transIer coeI-
Iicient is controlling.
4) Reduced Iouling, since the shellside Ilow is virtually along the baIIles, thereby
eliminating dead spaces.
5) Reduced maintenance costs.
Best Applications of Helixchangers
Helixchangers are best suited Ior services in which the shellside heat transIer coeIIicient
plays a determining role and/or when there is an incentive to reduce the shellside pressure
drop or Iouling. Shellside media may range Irom H
2
-rich gas to viscous Iluids with
Iouling tendencies, in single or two-phase. Thus, Helixchangers can be advantageously
employed Ior various services in reIineries and petrochemical plants. Over 750 Helix
-
changers have been supplied across the world and are reported to be operating satisIac-
torily. Please reIer to |2| and |3| Ior more detailed inIormation on the Helixchanger.
13.3.7 Twisted-tube heat exchangers
We have seen that perpendicular plate baIIles contribute to poor heat transIer and heavy
Iouling and corrosion by creating no-Ilow and recirculation areas. High turnaround
pressure drop on the shellside causes leaks (tube-baIIle and shell-baIIle) that result in low
thermal eIIectiveness as regards both the shellside heat transIer coeIIicient and MTD.
Bare tubes oI constant cross section are not very eIIicient Ior heat transIer and result in
poor conversion oI pressure drop to heat transIer.
The twisted-tube heat exchanger (Fig. 13.6) eliminates the baIIles altogether by letting
the tubes support themselves. The tubes are Iormed into an oval cross section with a super-
imposed twist in a special manuIacturing process that ensures good mechanical integrity.
Swirl Ilow inside tubes creates turbulence, even at low velocities and/or high viscosities.
On the shellside, complex interrupted swirl Ilow (which is predominantly longitudinal)
maximizes turbulence while minimizing pressure drop. The Ilow distribution and velocity
are homogenous. Even with a triangular pitch, the twist arrangement Ior baIIle-Iree support
provides cleaning lanes that permit complete mechanical cleaning by hydroblasting. Thus
the twisted-tube heat exchanger is a doubly-enhanced heat exchanger which enhances the

(a) Twisted tubes (b) Finished heat exchanger
Fig. 13.6 Twisted-tube heat exchanger (© Brown Fintube Co. Reprinted with permission.)
220
perIormance both inside and outside tubes.
Aavantages of twistea-tube heat exchangers
1) Better thermal-hydraulic perIormance. These heat exchangers typically yield 40°
higher heat transIer coeIIicient Ior the same pressure drop, or 50° pressure drop Ior
the same heat transIer coeIIicient, as in conventional shell-and-tube heat exchangers.
2) Reduced Iouling and better cleanability by virtue oI: (a) elimination oI dead spots, (b)
high turbulence, and (c) uniIorm velocity and constant Ilow distribution
3) No vibration since each tube extensively supported at multiple contact points
4) Simultaneous heat transIer enhancement on shellside and tubeside.
5) Very cost-eIIective because heat transIer is enhanced on both tubeside and shellside.
Twisted-tube heat exchangers represent a superior technology and are becoming in-
creasingly popular in many countries. Over 600 units have been supplied and reported to
be operating satisIactorily.
13.3.8 Plate heat exchangers
The gasketed plate heat exchanger PHE (Fig. 13.7) was originally introduced in the
1930s to meet the hygienic demands oI the dairy industry. The Iirst big advance in PHE
design came in the early 1950s and incorporated lateral corrugations to provide greater
plate strength and promote turbulence.
The herringbone pattern conIiguration was introduced towards the end oI the 1950s and
imparted even greater strength oI plates, thereby leading to thinner and lighter plates, as well
as higher levels oI turbulence, which resulted in higher heat transIer coeIIicients. Recent
advances now permit operation at pressures oI up to 427 psig (30 kg/cm2 g) and
temperatures up to 356°F (180°C). Semi-welded and Iully-welded types are also available
that permit operation at pressures oI above 569 psig (40 kg/cm
2
g) and 662
o
F (350
o
F).
PHEs have the Iollowing advantages over STHEs and are, consequently, preIerred over
the latter iI the service is within the allowable temperature and pressure:
a) much higher heat transIer coeIIicients resulting in lower cost, space requirement,
and hold-up
b) considerably lower Iouling
c) greater Ilexibility and the ability to handle widely diIIerent Ilow rates
a) capability oI achieving very close tem-
perature approaches and high degree oI
temperature crossover

Fig. 13.7 Plate heat exchanger (Courtesy oI AlIa Laval.)

Fig. 13.8 Spiral heat exchanger (Courtesy
oI AlIa Laval.)
221
13.3.9 Spiral plate heat exchangers
The spiral plate heat exchanger (Fig. 13.8), oIten reIerred to as the spiral heat exchanger
(SHE), was originally developed in Sweden to recover waste energy Irom contaminated
water eIIluent in pulp mills. Since then, the spiral heat exchanger has Iound its way into
several diverse industries including the chemical processing industries.
A spiral heat exchanger takes the Iorm oI two concentric channels Iormed by winding
two strips oI metal around a cylindrical core, one channel Ior each oI the two Iluids which
are to exchange heat. The Iluids are constrained between the plates by end covers on either
side oI the windings. At the periphery oI the SHE, nozzles and headers are employed to
direct the Iluids into and out oI the spiral channels. Flow can be either spiral or axial.
Due to the continuous curved channel, turbulent Ilow is obtained at a much lower
Reynold`s number than in a tubular heat exchanger. The single Ilow passage yields good
Iluid distribution with Iew dead spots. It is possible to achieve almost true countercurrent
Ilow. By virtue oI these Ieatures, SHEs yield a high eIIiciency oI conversion oI pressure
drop to heat transIer. Further, as the gaps between the channels can vary Irom 0.16 in. (4
mm) to 1 in. (25 mm), a wide range oI Ilow ratio can be handled.
SHEs are typically employed Ior slurries and viscous liquids. They tend to Ilush away
scale because: (a) the single cross section Ior each Iluid provides high shear rates and (b)
turbulence is induced by the swirling path. Consequently, Iouling resistances are considered
to be only a Iraction oI the TEMA values Ior STHEs. SHEs are now being extended to
condensing and reboiling applications.
Recent advances in SHE design have resulted in exchangers that can operate at 430 psig
(30 bar) and temperatures up to 752ƒF (400ƒC), depending upon size and material oI
construction.
13.3.10 Plate-fin heat exchangers
Plate-Iin heat exchangers (PFHEs) (Fig. 13.9), also called brazed aluminum heat
exchangers, are constructed Irom alternating layers oI corrugated and Ilat parting sheets.
The stacked arrangement is then brazed, yielding the exchanger core. Headers and
nozzles are Iixed to route the Iluids in and out oI the core. Heat is exchanged between
Iluids through both sheets. The corrugations, or Iins, serve not only as additional heat
transIer area but also provide mechanical support Ior the core.
PFHEs were initially developed Ior heat exchangers in the aerospace industry in the
1940s where prime requirements were low weight and low volume. Later, their use spread to
the chemical process industries, particularly the oIIshore industry where also weight is oI
prime importance.
The advantages oI close temperature approaches, true countercurrent Ilow, and a unique
ability to exchange heat with multiple streams make PFHEs viable alternatives to traditional
STHEs.
At low temperatures, aluminum is the ideal material oI construction, thanks to its easy
machinability, light weight, and high thermal conductivity. For temperatures which are too
high Ior aluminum, manuIacturers have developed PFHEs in stainless steels, titanium, and a
Iull range oI copper alloys.
Although PFHEs are generally employed Ior pressures less than 290 psi (20 bar),
designs have been developed Ior higher pressures.
PFHEs are susceptible to Iouling. Waxes, asphaltenes, and hydrates can all Ireeze out
and the small passages oI PFHEs are especially vulnerable. The metal wall temperature must
be kept above the wax or asphalt point or the hydrate Iormation temperature under all
222
conditions oI operation.
PFHEs are Iragile, especially those oI aluminum alloy construction. Pipework loads on
PFHE nozzles must, thereIore, be avoided. PFHEs are not as tough as STHEs, but will give
reliable service iI not abused.
13.3.11 Printed circuit heat exchangers
Printed circuit heat exchangers (PCHEs) (Fig. 13.10) are super-compact exchangers with
high heat exchange surIace densities and high heat transIer coeIIicients. They are made
Irom Ilat metal plates which have Iluid Ilow passages chemically milled into them, a
technique similar to that used in the manuIacture oI electronic printed circuit boards,
hence the name. The plates are then diIIusion bonded together into blocks to Iorm a heat
exchanger core to which are then attached nozzles and headers. One manuIacturer claims
that the Ilow passages in a PCHE are typically 0.0120.059-in. (0.31.5-mm) deep,
which results in a volumetric heat transIer area oI as high as 762 It
2
/It
3
(2500 m
2
/m
3
).
This manuIacturer also claims that PCHEs can be employed Ior liquid-liquid (even vis-

Fig. 13.9 Plate-Iin heat exchanger |9|
223
cous liquids), gas-liquid, gas-gas, con-
densing, and vaporizing duties.
13.3.12 Hybrid heat exchangers
The latest heat exchanger innovation is the
hybrid heat exchanger (HHE) (Fig. 13.11)
that successIully combines the high
perIormance oI the PHE with the high
pressure and temperature capabilities oI the
STHE. Here, a pressed plate oI a special
conIiguration is mated to a similar sheet to
Iorm an element. Several elements are then
stacked and Iully welded on alternating
edges to Iorm the core oI the hybrid. The
edges oI the core are then attached to headers, and the exchanger is shrouded with two
pressure Irames to Iorm the Iinished hybrid exchanger.
The materials oI construction vary Irom stainless steel through Hastelloy, Incoloy, and
even titanium. Hybrids exhibit a very high volumetric density oI heat transIer, up to 76 It
3

per It
3
(250 m
2
per m
3
), and are best suited Ior vaporizing and condensing duties.
13.4 Evaluation of heat transfer enhancement techniques
The evaluation oI heat transIer enhancement techniques is rarely straightIorward, as nu-
merous Iactors are involved. The Iollowing Iactors should be considered Ior a meaningIul
evaluation:
1) Heat transIer coeIIicient
2) Pressure drop
3) Capital cost oI the exchanger, including the cost oI the augmentation device
4) Space requirement
5) Operating cost oI the exchangerIluid pumping, maintenance, and downtime.
Maintenance costs include the cost oI opening the heat exchanger Ior inspection
and cleaning. The downtime cost is the cost oI lost production.


(a) PIates before bonding (b) Finished exchanger
Fig. 13.10 Printed circuit heat exchanger (Courtesy oI Heatric.)
Fig. 13.11 Hybrid heat exchanger element
(Courtesy Hydrocarbon Asia |4|.)
224
References
|1| Gough, M.J., and Rogers, J.V., 1987, 'Reduced Fouling by Enhanced Heat TransIer Using
Wire-Matrix Radial Mixing Elements,¨ 24th Natl. Heat Transfer Conf., AIChE Symp. Series,
Pittsburgh, Aug, Series No. 83, pp. 16²21.
|2| Butterworth, D., 1992, 'New Twists in Heat Exchangers,¨ The Chemical Engineer, 10 Sept.,
pp. 23²28.
|3| Stehlik, P., et al. 1994. 'Comparison oI Correction Factors Ior Shell-and-Tube Heat
Exchangers with Segmental or Helical BaIIles,¨ Heat Transfer Engineering, 15(1), pp. 55²65.
Further reading
1. Mascone, C.F., 1986, 'CPI Strives to Improve Heat TransIer in Tubes,¨ editorial survey,
Chemical Engineering, Feb 3, pp. 22²25.
2. Bergles, A.E., 1978, 'Enhancement oI Heat TransIer,¨ Dept. oI Mechanical Engineering and
Engineering Research Institute, Iowa State University, Paper No. KS-9, 6th Intl. Heat
Transfer Conf, Aug 7-11, Toronto, Canada.
3. Marner, W.J., and Bergles, A.E., 1978, 'Augmentation oI Tubeside Laminar Flow Heat
TransIer by Means oI Twisted-Tape Inserts, Static Mixer Inserts and Internally-Finned
Tubes,¨ Paper No. FC(a)-17, 6th Intl. Heat Transfer Conf., Aug 7-11, Toronto, Canada.
4. Hewitt, G.F., 1998, Heat Exchanger Design Hanabook, Volume 3, Sections 3.7 and 3.9,
Begell House, Inc., New York.
5. Kakac, S., et al., 1981, Heat Exchangers. Thermal-Hyaraulic Funaamentals ana Design,
Hemisphere Publishing Corp., New York.

225
INDEX
A
air compressor intercooler, 85
B
baIIle cut, 32, 37, 38, 40²48, 70, 72, 89, 90,
91, 97, 101, 114, 164, 177, 183, 195
very small and very large, 37, 38
baIIle cut, orientation oI
horizontal cut, 38
vertical cut, 38, 89, 135
baIIle cut, recommended, 38

baIIle spacing, 2, 6, 19, 32, 36²39, 41²45,
47, 48²50, 62²65, 70, 72, 77, 90²93, 97,
98, 101, 103²105, 116, 177, 183, 184,
190, 193, 195²198, 200, 201, 203
maximum baIIle spacing, 36
minimum baIIle spacing, 36, 62
optimum baIIle spacing, 37
variable baIIle spacing, 91
baIIles
deresonating, 205, 207
double-segmental, 38, 47²49, 85, 91, 93,
97, 98, 101, 116, 196, 197, 199, 200
helical, 1, 211
rod baIIles, 145, 205, 211
single-segmental, 47²49, 91, 93, 94, 96,
102, 116, 145, 196²199, 201, 203
baIIle-shell leakage (see also leakage and
bypass streams), 42, 43, 44, 59, 65, 71
boiling
boiling range, 3, 124, 126, 132, 133, 143,
154
convective boiling, 127
critical pressure, 3, 21, 22, 126, 127, 142²
144, 157
critical temperature, 157
Iilm boiling, 3, 56, 124, 139, 140, 144,
146, 147, 149, 154, 167
Ilow boiling, 3, 127
incipient boiling, 125
maximum heat Ilux, 126, 127, 142, 143,
149, 150
natural convection, 124²126, 213
nucleate boiling, 56, 124, 125, 127, 139,
140, 143, 144, 146²149, 154, 167, 213
pool boiling, 3, 123, 124, 126²128, 132,
133
pool boiling curve, 123, 124
porous surIace, 125, 144
rough surIace, 125, 176
sintered surIace, 125
steady-state Iilm boiling, 124
subcooled liquid heating, 141²143, 152
unsteady-state Iilm boiling, 124
wide-boiling mixture, 127, 131, 132,
135²137, 142, 150, 154
broad objectives oI thermal design, 19
C
channel, 5, 7, 10²13, 23, 64, 68, 71, 112,
127, 141, 221
channel cover, 5, 7, 10, 11, 13, 23, 68
co-current Ilow (see, Ilow co-current)
condensation
inside horizontal tubes, 86
inside vertical tubes, 86
oI mixed vapors,
outside horizontal tubes, 108
outside vertical tubes, 88
dropwise, 86
Iilm-wise, 86
gravity-controlled, 86, 88, 89, 101
isothermal , 85, 92
shear-controlled, 86, 88, 89, 101
condenser
ejector condenser, 49, 85, 116
reIlux condenser, 87, 88
surIace condenser, 119, 176, 181
condensers, 2, 3, 4, 49, 83, 86²91, 98, 100,
226
112, 113, 116, 117, 119, 160, 161, 168,
176, 181, 195, 218
condensing
on the shellside, 38, 114, 116
on the tubeside, 114, 116
condensing range, 83, 85, 90, 96
conical head, 12
conversion oI pressure drop to heat transIer,
26, 34, 35, 37, 38, 50, 68, 205, 209, 210,
214, 218, 219, 221
corrosion allowance, 23
corrosiveness, 68, 69
corrugated tubes, 211
countercurrent Ilow (see Ilow,
countercurrent)
critical pressure, 3, 21, 22, 126, 127, 142²
144, 157
critical temperature, 157
D
dam baIIle, 109
dead spaces, 68, 179, 183, 209, 218, 219
desuperheating, 3, 85, 90, 104²108, 110,
112, 120, 129, 160
dry-wall desuperheating, 105, 106
wet-wall desuperheating, 105, 107
dummy tubes, 40
E
enhanced heat transIer, 4, 209²211, 215
beneIits oI, 210, 215
equal outlet temperature, 56, 76, 162
expansion joint, 13, 14, 114
F
Ieed-eIIluent exchanger, 162
Iilm boiling see boiling, Iilm,
Iinned tubes, 116, 212
Ilooding, 87, 88, 112, 117, 118
Ilow
co-current , 52²54, 139, 205
countercurrent, 3, 24, 51²54, 57, 71, 77,
205, 211, 215, 221
laminar, 37, 50, 179, 181, 214
longitudinal, 36²38, 217, 218
turbulent, 26, 37, 50, 53, 221
window, 37, 38, 44, 46, 202
Ilow rate
high Ilow rate, 12
low Ilow rate, 72, 87
Ilow regime
annular Ilow, 87, 88, 139
choke Ilow, 137, 142
churn Ilow, 139
mist Ilow, 139
slug Ilow, 88, 139
Ilow velocity, 175, 177, 178, 197
Ilow-induced vibration
acoustic vibration, 4, 120, 191, 205²207,
208
damping, 192, 193
Iluidelastic whirling, 48, 94, 97, 191, 193,
194
modes oI tube Iailure, 193
natural Irequency oI tubes, 190, 192, 194,
199
tube vibration, 49, 190, 191
turbulent buIIeting, 191²195, 199, 200,
206, 208
unsupported span, 49, 190, 191, 193²195,
197, 199, 200, 203
vortex shedding, 191²195, 197, 199, 200,
206, 208
Iouling
adverse eIIects oI Iouling, 171
bio-Iouling, 98
categories oI Iouling, 4, 173, 183
cleanliness Iactor, 176
corrosion Iouling, 173
dead spaces, 68, 179, 183, 209, 218, 219
design guidelines to minimize Iouling,
179
Iouling resistance, 4, 22, 26, 43, 72, 106,
118, 165, 168, 171, 174, 176²181, 186,
212, 221
higher tube pitch, 183
on-line cleaning, 181
parameters which aIIect Iouling, 166
particulate Iouling, 173
selection oI Iouling resistance, 179
spare tube bundle, 182
spare unit, 183
TEMA Iouling resistance, 178
Iouling layer thickness, 4, 180, 181, 185,
186
H
head
Iront head, 10-12
rear head, 10-12
heat exchangers
Iixed tubesheet , 13, 23, 35, 40, 75, 113,
131
Iloating-head , 7, 10, 14-16, 22, 23, 24, 40
U-tube, 6, 10, 13, 14, 35, 39, 40, 75, 113
heat Ilux, 102, 106, 107, 121, 123²127, 129,
130, 132, 138, 141²143, 145, 148²150,
227
152, 154
heat release proIiles, 3, 20, 21, 157, 160
heat transIer augmentation, 210, 214
heat transIer coeIIicient
condensing , 90, 95, 96, 101, 102, 107,
110, 113, 117, 118, 121
overall, 26, 30, 31, 51, 53, 57, 62, 89, 95,
101, 138, 147, 149, 150, 165²167, 171,
176, 177, 184, 205, 210, 212, 217, 218
shellside, 6, 38, 40, 44²47, 58, 62, 64, 65,
77, 78, 102, 105, 176, 183, 184, 186,
202, 212, 218, 219
tubeside, 24, 25, 26, 27, 30, 31, 54, 76,
106, 113, 142, 153, 176, 211, 212, 214,
217
heat transIer eIIiciency, 44, 45, 209
helical baIIle exchangers, 1, 211
helixchangers, 218, 219
hybrid heat exchangers, 223
hydrogen, 3, 160, 212
I
impingement plate, 5, 8, 9, 112
L
lanes
pass-partition lane, 7, 13, 30, 36, 39, 40,
70
leakage and bypass streams
baIIle hole-tube, 39
baIIle-shell, 42²45, 59, 61, 65, 71
pass partition, 5, 24, 53
LMTD, 51²55, 57
longitudinal baIIle, 9, 10, 33, 135
loop seal, 109
low-Iin tubes, 3, 6, 90, 116²118, 126, 144,
211
M
mean temperature diIIerence
F
t
Iactor, 54, 55, 57, 77, 78, 205
MTD, 51, 53, 54, 56, 58, 59, 61, 62, 67, 98,
99, 101, 105²108, 111²113, 132, 133,
135, 137, 139, 142, 150, 154, 157, 161,
211, 214, 218
multiple shells
in parallel, 75, 98
in series, 3, 23, 57, 62, 76²78, 98, 99,
101, 180
in series/parallel, 78, 101
multiple shells:, 3, 23, 57, 62, 75²78, 90, 98,
99, 101, 169, 180
N
natural convection, 124²126, 213
noncondensables, 84²87, 89, 120
nozzles, 5, 12, 22, 33, 34, 40, 49, 90, 98,
110, 112, 119, 120, 131, 136, 190, 221,
222
O
on-line cleaning, 181
overdesign, 3, 27, 162²170, 177
on perIormance, 164, 166²168
on surIace, 164²169
P
pass-partition plates, 7, 24
physical properties, 3, 21, 25, 26, 39, 86, 90,
137, 139, 143, 157²159, 163, 166, 177
plate heat exchangers, 212, 220
plate-Iin heat exchangers (brazed aluminium
heat exchangers), 212, 221
Prandtl number, 24, 25, 69, 87
pressure drop
allowable (permitted), 4, 20, 21, 24, 27,
28, 30, 49, 72, 73, 76, 85, 90, 98, 104,
110, 116, 119, 180, 195,
shellside, 2, 3, 10, 23, 35, 36, 38, 42, 43,
45²49, 64, 77²79, 91, 92, 94, 97²102,
116, 119, 120, 183, 195²201, 203, 205,
218, 219
tubeside, 23, 26²28, 30, 31, 64, 79, 100,
114²116, 119, 164, 170, 180, 214
utilization oI, 43
printed-circuit heat exchangers, 212, 222,
223
R
reboilers
exit piping, 137, 142
Iorced-Ilow, 143
horizontal thermosyphon, 33, 134²138
internal, 123, 128, 129, 154
kettle, 15, 17, 34, 130²133, 135, 136,
137, 143, 145
selection oI reboilers, 153
start-up oI reboilers, 154
static head, 90, 138, 141, 143, 152, 153
vertical thermosyphon, 3, 12, 22, 53, 84,
88, 136, 138²141, 143, 145, 147, 148,
150, 152, 213
reduced pressure, 127, 142, 144
reIlux condensers, 87, 88
Reynold`s number, 24, 25, 27, 30, 35, 41,
68, 69, 70, 86, 87, 121, 192, 213, 214, 221
228
RODbaIIle heat exchangers, 217, 218
S
seal rods, 40
sealing strips, 5, 8, 9, 40, 41, 164
shear
vapor shear, 85, 88²90, 95, 110
shell
crossIlow shell, 34²36, 38, 49, 55, 119
divided-Ilow shell, 20, 33, 49, 85, 91, 96²
98, 119, 198²200, 203
double-split Ilow shell, 33
single-pass shell, 32, 47, 49, 54, 55, 91,
96, 97, 196, 199
split-Ilow shell, 33, 134
TEMA E, 32, 96, 116, 196
TEMA F, 32, 111
TEMA G, 33, 111, 134
TEMA H, 33, 134
TEMA J, 33, 97, 116, 205
TEMA K, 34
TEMA X, 34, 116, 135
two shells in series, 55, 63, 64, 70, 73,
76²79, 98²100, 102, 104, 110, 167,
184
sliding strips, 5, 8, 9
spare tubes, 22
speciIic heat, 22, 24²26, 51, 52²54, 112,
157, 158, 160, 161
spiral heat exchangers, 179
stepwise calculations, 30
stream analysis, 1, 2, 7, 30, 39²45, 63, 65,
70, 72, 77, 78, 100²103, 183²186, 217
subcooled boiling, 139
subcooling, 3, 90, 104, 108²112, 168
T
TEMA, 5²8, 10-12, 14²17, 22, 23, 31²36,
38, 40, 50, 54²57, 116, 176, 178, 179,
183, 188, 191, 196²200, 202²205, 208,
221
temperature cross, 3, 33, 56, 57, 70, 73, 76,
98, 99, 162, 166, 169, 186, 202, 205, 211,
220
temperature proIile distortion, 2, 3, 32, 41,
42, 44, 59²65, 71, 77, 78, 184, 185, 217
thermal conductivity, 22, 25, 117, 157, 158,
160, 181, 212, 221
tube diameter, 24, 25, 27, 39, 68, 69, 85,
144, 147, 180, 185, 192
tube inserts
twisted-tape inserts, 214
wire-Iin, 181, 214, 215, 217
tube layout
tube layout angle, 41, 192, 193, 206, 207
tube layout pattern, 2, 31, 34, 35
tube length, 7, 19, 22²24, 27, 30, 33, 40, 80,
85, 113, 119, 127²129, 131, 135, 140,
189, 205, 217
tube pitch, 2, 19, 32, 34, 35, 41, 46, 47, 49,
50, 69, 120, 127, 183, 192, 194, 207
tubes
bare, 117, 125, 209, 213²216
corrugated, 211
dummy, 40
Iinned, 212
Iluted, 213
high-Ilux, 125, 143, 144, 211, 213
large diameter tubes, 180
low-Iin, 3, 6, 90, 116²118, 126, 144, 211
tubeside velocity, 24, 26, 27, 30, 76, 119,
153, 155, 177
twisted, 1, 211
unsupported span, 49, 190, 191, 193²195,
197, 199, 200, 203
U-tubes, 6, 11, 14, 15, 128, 129, 131, 133,
135, 191, 194
tubesheet, 5-7, 10-16, 23, 71, 87, 140, 141,
194, 207
turbulence, 34, 35, 53, 68, 70, 126, 132, 179,
181, 192, 210²213, 219, 220, 221
twisted-tube heat exchangers, 181, 187, 211,
219, 220
U
U-bends, 14, 40, 129
V
vapor density, 85, 93, 96, 97, 113, 119, 143,
195
velocity
low, 49, 74, 75, 108, 213, 218
shellside, 6, 37, 43, 49, 62, 65, 75²78, 92,
99, 102, 177, 183, 184
tubeside, 24, 26, 27, 30, 76, 119, 153,
155, 177
viscosity, 22, 24²26, 28, 30, 31, 41, 53,
58, 59, 61, 62, 67²72, 80, 90, 110, 136,
143, 150, 157²160, 166, 180, 193, 216

ISBN 1-56700-205-6

Sign up to vote on this title
UsefulNot useful