Sie sind auf Seite 1von 279

PIPELINE DESl6N fOR WATER EN6IMEERS

THIRD REVISED AND UPDATED EDITION


DEVELOPMENTSIN WATER SCIENCE, 40

OTHER TITLES IN THIS SERIES

VOLUMES 7-3 ARE OUT OIF PRINT

4 J.J. FRIED
GROUNDWATER POLLUTION
5 N. RAJARATNAM
TURBULENT JETS
6 D. STEPHENSON
PIPELINE DESIGN FOR WATER ENGINEERS
7 v. HANK AND J. SVEC
GROUNDWATER HYDRAULICS
8 J. BALEK
HYDROLOGY AND WATER RESOURCES IN TROPICAL AFRICA
9 T.A. McMAHON AND R.G. MElN
RESERVOIR CAPACITY AND YIELD
10 G. KOVACS
SEEPAGE HYDRAULICS
1i w.n. GRAF AND C.H. MORTIMER (EDITORS)
HYDRODYNAMICSOF LAKES: PROCEEDINGS OF A SYMPOSIUM
12-13 OCTOBER 1978. LAUSANNE, SWITZERLAND
12 W. BACK AND D.A. STEPHENSON (EDITORS)
CONTEMPORARY HYDROGEOLOGY: THE GEORGE BURKE MAXEY MEMORIAL VOLUME
13 M.A. MARlfJO AND J.N. LUTHIN
SEEPAGE AND GROUNDWATER
14 D. STEPHENSON
STORMWATER HYDROLOGY AND DRAINAGE
15 D. STEPHENSON
PIPELINE DESIGN FOR WATER ENGINEERS
(completely revised edition of Vol. 6 in the series)
16 W. BACK AND R. LETOLLE (EDITORS)
SYMPOSlClM ON GEOCHEMISTRY OF GROUNDWATER
17 A.H. EL-SHAARAWI (EDITOR) IN COLLABORATION WITH S.R. ESTERBY
TIME SERIES METHODS IN HYDROSCIENCES
18 J.BALEK
HYDROLOGY AND WATER RESOURCES IN TROPICAL REGIONS
19 D. STEPHENSON
PIPEFLOW ANALYSIS
20 I.ZAVOIANU
MORPHOMETRY OF DRAINAGE BASINS
21 M.M.A. SHAHIN
HYDROLOGY OF THE NILE BASIN
22 H.C.RIGGS
STREAMFLOW CHARACTERISTICS
23 M. NEGULESCU
MUNICIPAL WASTEWATER TREATMENT
24 L.G.EVERETT
GROUNDWATER MONITORING HANDBOOK FOR COAL AND OIL SHALE DEVELOPMENT
25 W. KINZELBACH
GROUNDWATER MODELLING: AN INTRODUCTION WITH SAMPLE PROGRAMS IN BASIC
26 D. STEPHENSONAND M.E. MEADOWS
KINEMATIC HYDROLOGY AND MODELLING
27 A.M. EL-SHAARAWI AND R.E. KWIATKOWSKI (EDITORS)
STATISTICAL ASPECTS OF WATER CIUALITY MONITORING - PROCEEDINGS OF THE WORKSHOP HELD AT
THE CANADIAN CENTRE FOR INLAND WATERS, OCTOBER 1985
28 M.JERMAR
WATER RESOURCES AND WATER MANAGEMENT
29 G.W. ANNANDALE
RESERVOIR SEDIMENTATION
30 D.CLARKE
MICROCOMPUTER PROGRAMS IN GROUNDWATER
31 R.H. FRENCH
HYDRAULIC PROCESSES IN ALLUVIAL FANS
32 L. VOTRUBA, 2. KOS. K. NACHAZEL, A. PATERA ANDV. ZEMAN
ANALYSIS OF WATER RESOURC_ESYSTEMS
33 L. VOTRUBA AND V. BROZA
WATER MANAGEMENT IN RESERVOIRS
34 D. STEPHENSON
WATER AND WASTEWATER SYSTEMS ANALYSIS
35 M.A. CELlA ET AL.
COMPUTATIONAL METHODS IN WATER RESOURCES, VOLUME 1 MODELING SURFACE AND SUB-SUR-
FACE FLOWS. PROCEEDINGS OF THE VII INTERNATIONAL CONFERENCE, MIT. USA, JUNE 1988
36 M.A. CELIA ET AL.
COMPUTATIONAL METHODS IN WATER RESOURCES, VOLUME 2 NUMERICAL METHODS FOR TRANS-
PORT AND HYDROLOGICAL PROCESSES. PROCEEDINGS OF THE V11 INTERNATIONAL CONFERENCE, MIT,
USA, JUNE 1988
37 D.CLARKE
GROUNDWATER DISCHARGE TESTS: SIMULATION AND ANALYSIS
THIRD REVISED AND UPDATED EDITION

DAVID STEPHENSON
Department of Civil Engineering
University of the Witwatersrand
Johannesburg, South Africa

ELSEVlER

Amsterdam - Oxford - N e w York - Tokyo 1989


ELSEYIER SCIENCE PUBLISHERS B.V.
Sara Burgerhartstraat 25
P.O. Box 2 1 1, 1000 A€ Amsterdam, The Netherlands

Distributors for the United States and Canada:

ELSEVIER SCIENCE PUBLISHINGCOMPANY INC.


655, Avenue of the Americas
New York, NY 10017, U.S.A.

ISBN 0-444-87373-2

0 Elsevier Science Publishers B.V., 1989

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system
or transmitted in any form or by any means, electronic, mechanical, photocopying, recording
or otherwise, without the prior written permission of the publisher, Elsevier Science Publish-
ers B.V./ Physical Sciences & EngineeringDivision, P.O. Box 330, 1000 AH Amsterdam, The
Netherlands.

Special regulations for readers in the U.S.A. - This publication has been registered with the
Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained
from the CCC about conditions under which photocopies of parts of this publication may be
made in the USA. All other copyright questions, including photocopying outside of the USA,
should be referred t o the publisher.

No responsibility is assumed by the Publisher for any injury and/or damage t o persons or
property as a matter of products liability, negligence or otherwise, or from any use or opera-
tion of any methods, products, instructions or ideas contained in the material herein.

Printed in The Netherlands


V

PREFACE TO FIRST EDITION

Pipelines are being constructed in ever-increasing diameters,


lengths and working pressures. Accurate and rational design bases

are essential to achieve economic and safe designs. Engineers have


for years resorted to semi-empirical design formulae. Much work
has recently been done in an effort to r a t i o n a l i z e the design of
pipelines.
This book col lates pub1 ished material on rational design methods
as well as presenting some new techniques and data. Although
retaining conventional approaches in many instances, the aim of
the oook is to bring the most modern design techniques to the c i v i l
or hydraulic engineer. It is suitable as an introduction to the
subject but also contains d a t a on the most advanced techniques in
the field. i3ecause of the sound theoretical background the book will
a l so be useful to under-graduate and post-graduate students. Many
of the s u b j e c t s , such as mathematical optimization, are s t i l l in t h e i r

infancy and the book may provide leads for further research. The
methods of solution proposed for many problems bear in mind the
modern acceptance of computers and calculators and many of the
graphs in the book were prepared with the assistance of computers.
The first half of this book is concerned with hydraulics and
planning of pipelines. In the second half, structural design and
ancillary features are discussed. The book does not deal in detail
with manufacture, l a y i n g a n d operation, nor should i t replace design
codes of practice from the engineer's desk. Emphasis is on the
design of large Pipelines as opposed to industrial and domestic
piping which are covered in other pub1 ications. Although directed
at the water engineer, t h i s book will be of use to engineers i n v o l v e d
in the piping of many other fluids as well as solids and gases.
It should be noted that some of the designs and techniques
described may be covered by patents. These include types of pre-
stressed concrete pipes, methods of stiffening pipes and branches
and v a r i o u s coatings.
VI

The S . I . system of metric u n i t s i s preferred i n the book although


imperial u n i t s a re g i ve n i n brackets in many instances. Most g r a p h s

a n d equations a r e represented i n uni versa1 dimensionless form. Worked

examples are given for many problems a n d the reader i s advised to


work through these as they often elaborate on ideas not highlighted
in the text. The algebraic symbols used in each chapter are

summarized at the end of that chapter together with specific and


general references a r r a n g e d in the order of the subject matter i n the
chapter. The appendix gives further references and standards and
other useful data.
PREFACE TO SECOND EDITION

The gratifying response to the f i r s t e d i t i o n of t h i s book resulted


in small amendments to the second impression, a n d some major a l t e r -
ations i n t h i s new edition.
The chapters on transport of solids and sewers have been
replaced by data more relevant to water engineers. Thus a new
chapter on the effects of a i r i n water pipes i s included, as well as a
chapter on pumping systems for water pipelines. The latter was
reviewed by B i l l Glass who added many of h i s own ideas.
There are additions and u p d a t i n g throughout. There is additional
information on p i p e l i n e economics and optimum diameters i n Chapter 1 .
A comparison of currently used friction formulae is now made in
Chapter 2. The sections on non-circular pipe and p a r t l y full pipes
and sewer flow are omitted. These are largely of interest to the
drainage engineer and as such are covered in the author’s book
‘Stormwater Hydrology and Drainage’ (Elsevier, 1981). A basic
introduction to water hammer theory preceeds the design of water
hammer protection of pumping a n d g r a v i t y l i n e s i n Chapter 4 .
The sections on structural design of flexible pipes are brought
together. An enlarged section on soil-pipe interaction and limit
states of f l e x i b l e pipes preceeds the design of s t i f f e n e d pipes.
Although some of the new edition is now fairly basic, it is

recognised that this i s d e s i r a b l e for both the p r a c t i c i n g engineer who


needs refreshing and the student who comes across the problem of
p i p e l i n e design f o r the f i r s t time.
VIII

PREFACE TO THIRD E D I T I O N

Recent research i n c a v i t a t i o n a n d flow control h a s prompted a d d i t i o n a l


sections on this. There a r e also new sections on supports to exposed
pipes a n d secondary stress. Additional references and a new layout
make up this edition. Some sections appearing in p r e v i o u s editions,
noteably on p i p e network systems a n a l y s i s a n d o p t i m i z a t i o n have been
ommitted as they were considered more appropriate in the alJthOr’S
p a r a l l e l book ‘Pipeflow A n a l y s i s ’ b y the same p u b l i s h e r .
IX

ACKNOWLEDGEMENTS

The b a s i s for t h i s book was d e r i v e d from my experience a n d i n the


course of my duties with the Hand Water Board a n d Stewart, S v i r i d o v
and 01 i v e r , Consulting Engineers. The extensive knowledge of
Engineers in these organizations may therefore be reflected herein
although I am solely to blame f o r any inaccuracies o r misconceptions.
I am g r a t e f u l to my wife Lesley, who, i n a d d i t i o n to looking a f t e r
the twins during many a lost weekend, assiduously typed the first
d r a f t of t h i s book.

David Stephenson
X

CONTENTS

CHAPTER 1 ECONOMIC PLANNING

lntroduct ion . . . . . . . . . . . . . 1
P i p e l i n e Economics . . . . . . . . . . . . 2
B a s i c s of E c o n o m i c s . . . . . . . . . . . 9
M e t h o d s of Ana I y s i s. . . . . . . . . . . 10
Uncertainty in Forecasts . . . . . . . . . 11
Balancing Storage . . . . . . . . . . . . 14

CHA?TER 2 HYDRAULICS

The F u n d a m e n t a l E q u a t i o n s of F l u i d F l o w . . . . . 16
F l o w H e a d Loss R e l a t i o n s h i p s . . . . . . . . . 18
Empirical Flow Formulae . . . . . . . . . 18
2 a t i o n a l Flow Formulae . . . . . . . . . 19
C o m p a r i s o n of F r i c t i o n F o r m u l a e . . . . . . . 24
M i n o r Losses . . . . . . . . . . . . . 26
P r e s s u r e and F l o w C o n t r o l i n P i p e s . . . . . . . 28
lntroduct ion . . . . . . . . . . . . . 28
T y p e s of V a l v e s . . . . . . . . . . 28
. . .
I so l a t i n g V a Iv e s . . . . . . . . . 29
Control Valves . . . . . . . . . . 29
C a v i t a t i o n in C o n t r o l V a l v e s . . . . . . . . . 32
I n t e r a c t i o n b e t w e e n C a v i t a t i o n a n d W a t e r Hammer P r e s s u r e s 34

CHAPTER 3 P I P E L I N E SYSTEM ANALYSIS AND DESIGN

Network A n a l y s i s . . . . . . . . . . . . 37
E q u i v a l e n t P i p e s for P i p e s in S e r i e s o r P a r a l l e l . . . 37
Loop Flow Correction Method . . . . . . . . 38
. .
The Node H e a d C o r r e c t i o n M e t h o d . . . . . 40
. .
A l t e r n a t i v e M e t h o d s of A n a l y s i s . . . . . 41
.
iqetwork A n a l y s i s b y L i n e a r T h e o r y . . . . . 43
O p t i m i z a t i o n of P i p e l i n e Systems . . . . . . . . 44
.
D y n a m i c P r o g r a m m i n g f o r O p t i m i z i n g Compound P i p e s 45
T r a n s p o r t a t i o n P r o g r a m m i n g for L e a s t - c o s t
A1 l o c a t i o n o f Resources . . . . . . . . . . 48
L i n e a r P r o g r a m m i n g f o r D e s i g n of l e a s t - c o s t
Open N e t w o r k s . . . . . . . . . . . . 52
XI

CHAPTER 4 WATER HAMMER AND SURGE

R i g i d Water Column Surge Theory . . . . . . 58


M e c h a n i c s o f W a t e r Hammer . . . . . . . . 60
E l a s t i c W a t e r Hammer T h e o r y . . . . . . . . 64
Method o f A n a l y s i s . . . . . . . . . . 64
Effect of F r i c t i o n . . . . . . . . . . 68
Protection of Pumping Lines . . . . . . . . 69
Pump I n e r t i a . . . . . . . . . . . 73
Pump B y p a s s R e f l u x V a l v e . . . . . . . . 76
Surge Tanks . . . . . . . . . . . . 77
Discharge Tanks . . . . . . . . . . . 79
A i r Vessels . . . . . . . . . . . . 85
In-Line Reflux Valves . . . . . . . . . 89
Release V a l v e s . . . . . . . . . . . 91
Choice o f P r o t e c t i v e D e v i c e . . . . . . . . 93

CHAPTER 5 AI3 IN P I P E L I N E S

Introduction . . . . . . . . . . . . . 97
Problems of A i r Entrainment . . . . . . . . . 97
.
A i r I n t a k e a t Pump Sumps . . . . . . . . 99
A i r Absorption a t Free Surfaces . . . . . . . . 101
. .
H y d r a u l i c Removal of A i r . . . . . . . . 102
H y d r a u l i c Jumps. . . . . . . . . . . 102
Free F a l l s . . . . . . . . . . . . . 104
A i r Valves . . . . . . . . . . . . . 105
. .
Head Losses in P i p e l i n e s . . . . . . . . 108
W a t e r Hammer . . . . . . . . . . . . . 109

CHAPTER 6 EXTERNAL LOADS

Soil Loads . . . . . . . . . . . . . . 113


Trench conditions . . . . . . . . . . . . 113
Embankment Conditions . . . . . . . . . . 116
Superimposed Loads . . . . . . . . . . . . 120
T r a f f i c Loads . . . . . . . . . . . . . 121
Stress Caused b y Point Loads . . . . . . . . 121
Uniformly Loaded Areas . . . . . . . . . . 122
Effect o f R i g i d Pavements . . . . . . . . . 123
XI1

CHAPTER 7 CONCRETE PIPES

The Effect of B e d d i n g . . . . . . . . . . 127


Prestressed Concrete Pipes . . . . . . . . . 129
Circumferential Prestressing . . . . . . . . 131
. .
C i r c u m f e r e n t i a l P r e s t r e s s a f t e r Losses . . . 131
Circumferential Stress u n d e r F i e l d Pressure . . . . 133
L o n g i t u d i n a l Prestressing . . . . . . . . . 134
. . . .
L o n g i t u d i n a l Stresses A f t e r L o s s e s . . . 136
. . . .
P r o p e r t i e s o f Steel and C o n c r e t e . . . 136

CHAPTER 8 STEEL AND F L E X I B L E P I P E

I n t e r n a l Pressures . . . . . . . . . . . . 142
.
Tension R i n g s to Resist I n t e r n a l Pressures . . . . 143
.
D e f o r m a t i o n of C i r c u l a r P i p e s u n d e r E x t e r n a l L o a d . . 146
. . . . .
Effect of L a t e r a l Support . . . . 149
. .
S t r e s s d u e to C i r c u m f e r e n t i a l B e n d i n g . . . . 151
More General Deflection Equations . . . . . . . 152
S t i f f e n i n g R i n g s to R e s i s t B u c k l i n g w i t h no
side support . . . . . . . . . . . . . 156
Tension R i n g s . . . . . . . . . . . . 156
Stiffening Rings . . . . . . . . . . . . 156

CHAPTER 9 SECONDARY STRESSES

Stresses a t Branches . . . . . . . . . . . 160


Crotch Plates . . . . . . . . . . . . 162
Internal Bracing . . . . . . . . . . . 165
Stresses a t Bends . . . . . . . . . . . . 172
T h e P i p e a s a Beam . . . . . . . . . . . 173
Longitudinal Bending . . . . . . . . . . 173
Pipe Stress a t Saddles . . . . . . . . . . 174
Ring Girders . . . . . . . . . . . . 175
Temperature Stresses . . . . . . . . . . . 176

CHA?TER 10 P I PES. F I TT I NGS AND APPURTENANCES

?ipe M a t e r i a l s . . . . . . . . . . . . . 179
Steel P i p e . . . . . . . . . . . . . 179
Cast I r o n P i p e . . . . . . . . . . . . 179
Asbestos Cement P i p e . . . . . . . . . . 180
Concrete P i p e . . . . . . . . . . . . . 180
Plastic Pipe . . . . . . . . . . . . . 180
L i n e Val ves
Sluice Valves
.. .. .. .. .. .. .
.
.
.
.
.
.
.
.
.
.
.
.
. 182
182
. .
B u t t e r f l y Valves . . . . . . . . . . 183
Globe Valves . . . . . . . . . . . . . 184
Needle a n d Control Valves . . . . . . . . . 184
Spherical Valves . . . . . . . . . . . . 185
Reflux Valves . . . . . . . . . . . . . 186
A i r Valves . . . . . . . . . . . . . 186
A i r Vent Valves . . . . . . . . . . . . 186
A i r Release Valves . . . . . . . . . . . 188
Thrust Blocks . . . . . . . . . . . . . 190
Forces Induced by Supports . . . . . . . . . 194
. .
L o n g i t u d i n a l Stress . . . . . . . . . 195
Temperature Stresses . . . . . . . . . 195
Forces at Bends . . . . . . . . . . . . 196
L a t e r a l Movement . . . . . . . . . . . 196
Forces on Supports . . . . . . . . . . . 197
Unbalanced Forces . . . . . . . . . . . 197
Flow Measurement . . . . . . . . . . . . 198
Venturi Meters . . . . . . . . . . . . 198
Nozz I es . . . . . . . . . . . . . 199
Orifices . . . . . . . . . . . . . 1 99
Bend Meters . . . . . . . . . . . . . 200
Mechanical Meters . . . . . . . . . . . 200
. .
Electromagnetic I n d u c t i o n . . . . . . . 201
Mass a n d Volume Measurement. . . . . . . . 201
Te I erne t ry . . . . . . . . . . . . . 201

CHAPTER 1 1 L A Y I NG AND PROTECTION

Selecting a Route . . .
. . . . . . . . . 205
L a y i n g a n d Trenching . .
. . . . . . . . . 206
Thrust Sores . . . .
. . . . . . . . . 208
Pipe Bridges . . . .
. . . . . . . . . 208
Underwater Pipelines . .
. . . . . . . . . 210
Joints a n d Flanges . . .
. . . . . . . . . 212
Coatings . . . .
. . . . . . . . . 21 6
Linings . . . .
. . . . . . . . . 21 7
Cat hod i c Prot ec t ion . . .
. . . . . . . . . 218
Galvanic Corrosion . . . . . . . . . . . 21a
.
Stray Current E l e c t r o l y s i s . . . . . . . . 222
Thermal I n s u l a t i o n . . . . . . . . . . . 223
XIV

CHAPTER 12 PUMP I NG I NSTALLAT I ON5

I n f l u e n c e of Pumps in P i p e l i n e Design . . . . . . 228


T y p e s of P u m p s . . . . . . . . . . . . 228
P o s i t i v e Displacement Types . . . . . . . . . 228
C e n t r i f u g a l Pumps . . . . . . . . . . . 229
T e r m s and D e f i n i t i o n s . . . . . . . . . . . 231
Head . . . . . . . . . . . . . 23 1
Total Head . . . . . . . . . . . . . 231
Net P o s i t i v e S u c t i o n Head . . . . . . . . . 232
S p e c i f i c Speed . . . . . . . . . . . . 233
Impeller Dynamics . . . . . . . . . . . . 234
Pump C h a r a c t e r i s t i c C u r v e s . . . . . . . . . 236
ivto t o r s . . . . . . . . . . . . . 239
Pumpstat ions . . . . . . . . . . . . . 240

GENERAL REFE2ENCES AND STANDARDS . . . . . . 242

BOOKS FOR FURTHER READING . . . . . . . . . 250

APPEND I X

Symbols for p i p e f i t t i n g s . . . . . . . . . . 252


Properties of p i p e shapes . . . . . . . . . . 253
P r o p e r t i e s of w a t e r . . . . . . . . . . . . 254
P r o p e r t i e s of p i p e m a t e r i a l s . . . . . . . . . 255
Conv ers i o n f a c t o r s . . . . . . . . . . . . 256

AUTHdR INDEX . . . . . . . . . . . . 258


SUBJECT INDEX . . . . . . . . . . . . 260
1

CHAPTER 1

ECONOMIC PLANNING

I NTRODUCT ION

Pipes have been used for many c e n t u r i e s for transporting fluids.


The Chinese f i r s t u s e d bamboo p i p e s t h o u s a n d s o f y e a r s ago, and lead
p i p e s were unearthed a t Pompeii. In later c e n t u r i e s wood-stave pipes
were used in England. It was only with the advent of cast iron,
however, that pressure pipelines were manufactured. Cast iron was
used extensively in the 19th Century and is still used. Steel pipes
were f i r s t introduced towards t h e e n d of t h e l a s t c e n t u r y , facilitating
c o n s t r u c t i o n of small and large bore pipelines. The i n c r e a s i n g use o f
high grade steels a n d large r o l l i n g m i l l s has enabled pipelines with
diameters over 3 metres and working pressure over 10 Newtons per
square millimetre to be manufactured. Welding techniques have been
perfected enabling longitudinally and circumferentially welded or
spiral welded p i p e s to b e m a n u f a c t u r e d . P i p e l i n e s a r e now a l s o made
in r e i n f o r c e d concrete, p r e - s t r e s s e d concrete, asbestos cement, plastics
and claywares, to suit varying conditions. Reliable flow formulae
became available for the design of pipelines this century, thereby
a l s o p r o m o t i n g t h e use of p i p e s .
Prior to this century water a n d sewage were p r a c t i c a l l y the o n l y
fluids transported by pipeline. Nowadays pipelines are the most
common means for transporting gases and oils over long distances.
Liquid chemicals and solids in slurry form or in containers are also
being pumped t h r o u g h p i p e l i n e s on ever i n c r e a s i n g scales. There a r e
now o v e r two m i l l i o n kilometres of pipelines i n service throughout the
world. The g l o b a l e x p e n d i t u r e on p i p e l i n e s i n 1974 was p r o b a b l y o v e r
5 5 000 m i l l i o n .
There are many advantages of pipeline transport compared with
o t h e r methods s u c h a s r o a d , rail, waterway and air:-
(1) Pipelines are often the most economic form of transport
( c o n s i d e r i n g e i t h e r c a p i t a l costs, r u n n i n g costs o r o v e r a l l c o s t s ) .
(2) Pipelining costs are not very susceptible to fluctuations in
2

prices, since the major cost i s the c a p i t a l o u t l a y and subsequent


o p e r a t i n g costs are r e l a t i v e l y small.
Operations are not susceptible to labour disputes as little
attendance i s r e q u i r e d . Many modern systems operate automatical-

lY.
Being hidden beneath the ground a pipeline will not mar the
n a t u r a l environment.
A b u r i e d p i p e l i n e i s reasonably secure against sabotage.
A pipeline is independent of external influences such as traffic
congestion and the weather.
There is normally no problem of r e t u r n i n g empty containers to
the source.
It is relatively easy to increase the c a p a c i t y of a pipeline by
i n s t a l l i n g a booster pump.
A b u r i e d p i p e l i n e w i l l not d i s t u r b surface t r a f f i c a n d services.
( 1 0 ) Wayleaves for pipelines are usually easier to obtain than for
roads and r a i l w a y s .
(11) The accident rate per ton - km is considerably lower than for
other forms of t r a n s p o r t .
(12) A pipeline can cross rugged terrain difficult for vehicles to
cross.
There are of course disadvantages associated w i t h p i p e l i n e systems:-
The initial capital expenditure i s often large, so i f there i s any

uncertainty in the demand some degree of speculation may be


necessary.
There is often a high cost involved in filling a pipeline
( e s p e c i a l l y long fuel lines).
Pipelines cannot be used for more than one m a t e r i a l at a time
(although there are multi-product pipelines operating on batch
bases).
There are operating problems associated with the pumping of
solids, such as blockages on stoppage.
I t i s often d i f f i c u l t to locate leaks o r blockages.

P I PEL I NE ECONOM I CS

The main cost of a p i p e l i n e system i s u s u a l l y that of the p i p e l i n e


3

itself. The pipeline cost is in fact practically the only cost for
gravity systems b u t as the adverse head increases so the power and
pumping s t a t i o n costs increase.
Table 1.1 indicates some relative costs for typical installed
p i p e l i nes.
With the economic instability and rates of inflation prevailing at
the time of writing pipeline costs may increase by 20% o r more per
year, and relative costs for different materials will vary. In
particular the cost of petro-chemical materials such as PVC may
increase faster than those of concrete for instance, so these f i g u r e s
should be inspected w i t h caution.

TABLE 1 . 1 Relative P i p e l i n e Costs

Bore mm
Pipe M a t e r i a l 150 450 1 500

PVC 6 23 -
Asbestos cement 7 23 -
Reinforced concrete - 23 80
Prestressed concrete - 33 90 - 150
M i l d steel 10 28 100 - 180
High tensile steel 11 25 90 - 120
Cast i r o n 25 75 -

<:,,-s, indicates not r e a d i l y a v a i l a b l e .


1 u n i t = d;/metre i n 1974 under average conditions
The components m a k i n g up the cost of a p i p e l i n e v a r y widely from
situation to situation but for water pipelines in open country and
t y p i c a l conditions are as follows:-
Supply of p i p e - 55% (may reduce as new
m a t e r i a l s a r e developed)
Excavation - 20% (depends on terrain,
may reduce as mechanical
excavation techniques im-
prove)
L a y i n g and j o i n t i n g - 5% (may increase w i t h la-
bour costs)
F i t t i n g s a n d specials - 5%
Coating and w r a p p i n g - 2%
Structures ( v a l v e chambers, anchors) - 2%
4
L
10
a
L
0
r
m
._
U
5

Water hammer protection - 1%


Land acquisition, access roads, cathodic .
protection, security structures, fences - 1%
Engineering and survey costs - 5%
Admi n i s t r a t i ve costs - 1%
Interest d u r i n g construction - 3%

Many factors have to be considered in s i z i n g a p i p e l i n e : For water


pumping mains the flow velocity at the optimum diameter varies from
0.7 m/s to 2 m/s, depending on flow and working pressure. It is
about 1 m/s for low pressure heads and a flow of 100 C/s increasing
to 2 m/s for a flow of 1 000 C / S and pressure heads at about 400 m
of water, and may be even h i g h e r f o r h i g h e r pressures. The c a p a c i t y
factor and power cost structures also influence the optimum flow
v e l o c i t y o r conversely the diameter for any p a r t i c u l a r flow. Fig. 1.1
i l l u s t r a t e s the optimum diameter of water mains f o r t y p i c a l conditions.
I n planning a p i p e l i n e system i t should be borne i n mind that the
scale of operation of a pipeline has considerable effect on the unit
costs. By doubling the diameter of the pipe, other factors such as
head remaining constant, the c a p a c i t y increases six-fold. On the other
hand the cost approximately doubles so that the cost per unit
delivered decreases to 1/3 of the original. It is this scale effect
which justifies multi-product lines. Whether it is i n fact economical
to install a large diameter main at the outset depends on the
f o l l o w i n g factors as well as scale:-
Rate of growth i n demand ( i t may be uneconomical to operate a t
low capacity factors during initial years). (Capacity factor is
the r a t i o of a c t u a l average discharge to design c a p a c i t y ) .
Operating factor (the ratio of average throughput at any time
to maximum throughput during the same period), which will
depend on the rate of draw-off and can be improved by
i n s t a l l i n g storage a t the consumer's end.
Reduced power costs due to low friction losses while the
p i p e l i n e i s -not o p e r a t i n g a t f u l l c a p a c i t y .
C e r t a i n t y of f u t u r e demands.
V a r y i n g costs w i t h time ( b o t h c a p i t a l and o p e r a t i n g ) .
(6) Rates of interest and c a p i t a l availability.
(7) Physical d i ffi cu l ti e s in the construction of a second pipeline i f
required.
The optimum design p e r i o d of a p i p e l i n e depends on a number of
factors, not least b e i n g the r a t e of interest on c a p i t a l loans and the
rate of cost inflation, in addition to the r a t e of growth, scale and
c e r t a i n t y of f u t u r e demands (Osborne a n d James, 1973).
In waterworks practice it has been found economic to size
pipelines for demands up to 10 to 30 years hence. For large
throughput and h i g h growth rates, technical c a p a b i l i t i e s may l i m i t the
size of the p i p e l i n e , so that supplementation may be r e q u i r e d w i t h i n
10 years. Longer planning stages are normally justified for small
bores and low pressures.
It may not always be economic to lay a uniform bore pipeline.
Where pressures are high it is economic to reduce the diameter and
consequent1 y the wal I thickness.
In planning a trunk main with progressive decrease i n diameter
there may be a number of possible combinations of diameters.
Alternative layouts should be compared before deciding on the most
economic. Systems a n a l y s i s techniques such as l i n e a r programming and
dynamic programming are i d e a l l y suited f o r such studies.
Booster pump stations may be installed along lines instead of
pumping to a high pressure head a t the i n p u t end and m a i n t a i n i n g a
high pressure along the entire line. By providing for intermediate
booster pumps a t the design stage instead of pumping to a h i g h head
at the input end, the pressure heads and consequently the p i p e w a l l
thicknesses may b e minimized. There may be a s a v i n g i n o v e r a l l cost,
even though additional pumping stations are required. The booster
stations may not be r e q u i r e d f o r some time.
The c a p a c i t y of the p i p e l i n e may often be increased b y i n s t a l l i n g
booster pumps at a later stage although it should be r e a l i s e d that
this is not always economic. The friction losses along a pipeline
increase approximately with the square of the flow, consequently
power losses increase considerably f o r h i g h e r flows.
The diameter of a pumping main to convey a known discharge can
be selected by an economic comparison of alternative sizes. The
pipeline cost increases with increasing diameter, whereas power cost
7

TOTAL
COST
5

DIAMETER

Fig. 1.2 O p t i m i z a t i o n of d i a m e t e r of a p u m p i n g p i p e l i n e .

COST I N
CENTS
P E R m3

Cl
c2
C

4 = 0, + 0;

Discharge 0
01 02 Q3

Fig. 7 -3 Optirriization of throughput for certain diaweters


8

in overcoming friction reduces correspondingly. On the other hand


power costs increase steeply as the p i p e l i n e i s reduced in d i a m e t e r .
Thus by adding together pipeline and the present v a l u e of operating
costs, one obtains a curve such as Figure 1.2, from which the
least-cost system can be selected. An example is given later in the
chapter. There will b e a h i g h e r cost the g r e a t e r the design d i s c h a r g e
rate.
If at some stage later it is desired to increase the throughput
capacity of a pumping system, it i s convenient to replot d a t a from a
diagram such as Figure 1.2 in the form of Figure 1.3. Thus for
different possible throughputs, the cost, now e x p r e s s e d i n c e n t s p e r
kilolitre or similar, is plotted a s the ordinate w i t h a l t e r n a t i v e (real)
pipe diameter a parameter.
It can be demonstrated that the cost per unit of throughput for
any pipeline is a minimum when the p i p e l i n e cost (expressed on an
annual basis) is twice the annual cost of the power in o v e r c o m i n g
friction.
T h u s t h e cost in cents p e r c u b i c metre of water is

C,(P) + C2(d)
c =
61

C, wHQ + C 2 ( d )
-
-
(1.1)
Q

= c,w + Hs) + CZ(d)/Q


2gdA

2C1 wHf/Q - C2(d)/Q2

minimum C
i.e. C Z ( d ) = 2 C,(P) (1.2)

P i s power requirement, proportional to wHQ. w i s the u n i t weight


of water, H is the total head, subscript s refers to static and f to
friction, Q is pumping rate, C2(d) is the cost of a pipeline of
diameter d, C,(P) is t h e cost of power ( a l l costs converted to a u n i t
time b a s e ) .
(In a similar manner it c a n b e shown that the power o u t p u t of a
given diameter penstock supplying a hydroelectric station is a
maximum if the friction head loss is one third of the total head

a v a i table).
9

R e t u r n i n g to F i g u r e 1.3, the f o l l o w i n g w i l l be observed:


(1) At any particular throughput Q, there is a c e r t a i n diameter at
w h i c h o v e r a l l costs w i l l b e a m i n i m u m ( i n t h i s case D
2
1.
(2) At t h i s d i a m e t e r t h e c o s t p e r ton of throughput could be reduced
further if throughput was increased. Costs would be a minimum
at some t h r o u g h p u t Q2. Thus a p i p e l i n e ' s optimum throughput is
not the same as the throughput for which it is the optimum
diameter.
(3) If Q were increased by an a m o u n t Q3 so t h a t total throughput
1
Q4 = Q1 + Q3 i t m a y b e economic not to install a second
pipeline (with optimum diameter D3 ) but to increase the flow
through the p i p e w i t h diameter D2, i.e. Q4C4 i s l e s s t h a n QIC1
+ Q3C3.
(4) At a later stage when it is justified to construct a second
pipeline the throughput through the overloaded line could be
reduced.
The power cost per unit of additional throughput decreases with
increasing pipe diameter so the corresponding likelihood of it being
most economic to increase throughput through an existing line
increases w i t h size (White, 1969).

B A S I C S OF ECONOMICS

Economics is used a s a b a s t s f o r c o m p a r i n g a l t e r n a t i v e schemes o r


designs. Different schemes m a y have different cash flows necessitating
some rational form of comparison. The c r u x o f a l l methods o f economic
comparison is the discount rate which may be in the form of the
interest rate on loans or redemption funds. National projects may
r e q u i r e a discount r a t e different from the p r e v a i l i n g interest r a t e , to
reflect a time rate of preference, whereas private organizations will
b e more interested in the a c t u a l c a s h flows, a n d consequently use the
real borrowing interest rate.
The cash flows, i.e. payments and returns, of one scheme may b e
compared with those of another by bringing them to a common time
basis. Thus a l l cash flows may be discounted to t h e i r p r e s e n t value.
For instance one pound received next year is the same as fl/l.05
( i t s present value) this year if it could earn 5% i n t e r e s t if invested
10

this year. It is usual t o meet capital expenditure from a loan over

a definite period at a certain interest rate. Provision i s made f o r


r e p a y i n g the load by paying into a s i n k i n g fund w h i c h a l s o c o l l e c t s
interest. The annual repayments at the e n d of each year required
to amount to a71 in n y e a r s i s

where r i s the i n t e r e s t r a t e on t h e p a y m e n t s i n t o t h e s i n k i n g f u n d . If
t h e i n t e r e s t r a t e o n t h e toan i s R, then the t o t a l a n n u a l payment i s

R(l+r)"+ r-R
(1.4)
( 1 + r )n-t

Normally the interest rate on the loan i s equal to the interest rate
e a r n e d b y t h e s i n k i n g fund so t h e a n n u a l p a y m e n t on a l o a n of El is

(1.5)

Conversely the present v a l u e of a p a y m e n t of S1 a t t h e e n d o f each


year over n years IS

The p r e s e n t v a l u e of a s i n g l e amount of f l i n n y e a r s i s

Interest tables are available for determining the a n n u a l payments


on loans, and the present v a l u e s of annual payments o r returns, for
various interest rates and redemption periods (lnstn. of Civil Engs.,
1962)

Methods of A n a l y s i s
Different engineering schemes r e q u i r e d to meet t h e same o b j e c t i v e s
may b e compared economically in a n u m b e r of ways. If a l l payments
and incomes associated with a scheme a r e d i s c o u n t e d to t h e i r present
value for comparison, the analysis is termed a present value or
11

discounted cash flow analysis. On the other hand if annual net


incomes of d i f f e r e n t schemes a r e compared, this i s termed the r a t e of
return method. The latter is most frequently used by private
o r g a n i s a t ions where tax returns and profits feature priminently. In
such cases i t i s suggested that the assistance of q u a l i f i e d accountants
i s obtained. Present value comparisons are most common for public
u t i I i t ies.
A form of economic analysis popular in the United States is
benefit/cost analysis. An economic benefit i s attached to a l l products
of a scheme, for instance a certain economic value is attached to
water supplies, although t h i s is difficult to e v a l u a t e in the case of
domestic supplies. Those schemes w i t h the highest benefit/cost values
a r e attached highest priority. Where schemes are mutually exclusive
such as i s u s u a l l y the case w i t h p u b l i c u n t i l i t i e s the scheme w i t h the
largest present value of net benefit is adopted. It the total water
requirements of a town f o r instance were f i x e d , the least-cost supply
scheme would be selected f o r construction.

Uncertainty i n Forecasts
Forecasts of demands, whether they be f o r water, o i l o r gas, are
i n v a r i a b l y clouded w i t h uncertainty and risk. Strictly a probability
a n a l y s i s i s r e q u i r e d f o r each possible scheme, i.e. the net benefit of
a n y p a r t i c u l a r scheme w i l l be the sum of the net benefits mul t i pl i ed
by their probability for a number of possible demands. Berthouex
(1971) recommends under-designing b y 5 to 10% f o r p i p e l i n e s to a l l o w
for uncertain forecasts, but his analysis does not account for cost
inflation.
An a l t e r n a t i v e method of a l l o w i n g f o r uncertainty i s to a d j u s t the
discount or interest rate: increasing the rate will favour a low
c a p i t a l cost scheme, which would be p r e f e r a b l e i f the future demand
were uncertain.

Example
A consumer r e q u i r e s 300 P/s of water for 5 years then plans to
increase his consumption to 600 P/s for a further 25 years (the
economic l i f e of h i s f a c t o r y ) .
12

He draws for 75% of the time every day. Determine the most
economic diameter and the number of p i p e l i n e s r e q u i r e d . The water
i s s u p p l i e d b y a p u b l i c body p a y i n g no t a x . Power costs a flat
0.5 p/kWhr, which includes an allowance for operating and
maintenance. The interest r a t e on loans ( t a k e n over 20 y e a r s ) a n d
on a s i n k i n g f u n d i s 10% p e r annum, a n d the rate of inflation in
cost of pipelines, pumps and power is 6% p.a. Pump and
pumpstations costs amount to 5300 p e r incremental kW, (including
a n allowance f o r standby p l a n t ) a n d pump e f f i c i e n c y i s 70%.
The effective discount r a t e may be taken as the interest rate less
the r a t e of i n f l a t i o n , i.e. 4% p.a., since E l t h i s year i s worth f l
x 1.10/1.06 + 51 x 1.04 next year.
The s u p p l y could be made t h r o u g h one large pipeline capable of
h a n d l i n g 600 P/s, or two smaller pipelines each delivering 300
O/s, one installed five years after the other. A comparison of
a l t e r n a t i v e diameters i s made i n the f o l l o w i n g table for a single
pipeline. Similarly an a n a l y s i s was made f o r two pipelines each
d e l i v e r i n g 300 P / s . This i n d i c a t e d an optimum diameter of 600 mm
for each p i p e l i n e a n d a total present value for both p i p e l i n e s of
5 7 500 p e r 100 m. Thus one p i p e l i n e , 800 mm diameter, w i l l be the
most economic solution.

Note that the analysis is independent of the length of the


pipeline, although it was assumed that pressure was such that a
continuous low-pressure p i p e was a l l t h a t was r e q u i r e d . Water hammer
protection costs a r e assumed incorporated in the p i p e cost here. The
a n a l y s i s i s also independent o f the c a p i t a l loan period, a l t h o u g h the
r e s u l t s would be s e n s i t i v e to change i n the interest o r i n f l a t i o n rates.
Discount factors were obtained from present v a l u e tables for 4% over
5, 25 and 30 years periods. Uncertainty was not allowed for but
would f a v o u r the two smaller p i p e l i n e s .
Another i n t e r e s t i n g p o i n t emerged from the a n a l y s i s : I f a 600 mm
p i p e l i n e was installed initially, due to a high uncertainty of the
demand increasing from 300 to 600 P/s, then if the demand did
increase, it would be more economic to boost the pumping head and
13

Sol u t i o n :
1. I n s i d e D i a .mm
600_ _ ~
_ _ 700 800 900
2. FIOW e/s N O 600 300 600 300 600 300 600
3. H e a d loss
m/100 m 0.14 0.55 0.06 0.24 0.03 0.12 0.02 0.07
4. Power loss
kw/100 m
= (3.)xQ/70 0.60 4.,72 0.26 2.06 0.13 1.03 0.09 0.60
5. Energy require-
m e n t s kW h r / y r /
100 m
= (4.1~8760
x 0.75 3900 31000 1700 13500 850 6600 590 3900
6. Annual pumping
cost
silo0 m 20 155 e 67 4 33 3 20
7. Equiv.Capita1
c o s t of p u m p i n g
over 5 years
= (6Jx4.452 90 -- 40 - 20 - 10 -

e. Equiv .capital
c o s t of p u m p i n g
o v e r 25 y e a r s .
= ( 6 . )x15.622 - 24-30 - 1060 - 520 - 310

9. P r e s e n t v a l u e of
p u m p i n g cost
= (8.)/1.170 2 070 900 440 260

10. Cost of p u m p s
etc.
f/100 m : = ( 4 . ) x300 180 1410 80 620 40 310 30 180

11. Present v a l u e of
pump cost =
(10.,)/1.170 for
second s t a g e 180 1200 80 530 40 260 30 150

12. P i p e l i n e cost
f/100 m 3600 4200 4800 5400

13. TOTAL COST


f / 1 0 0 m f o r 300
G 600 e / s
7140 5750 5560 :: 5850
7. '9. + 1 1 . t i 2
(least cost)
14

Rate - _ - -
m’ls

Reservoir empty

cumulative demand =/Qdt

Reservoir f u l l = c a p a c i t y required

24h I

Fig. 1.4 Graphical c a l c u l a t i o n of r e s e r v o i r capacity

pump the total flow through the one existing 600 mm diameter
pipeline rather than provide a second 600 mm pipeline. This is
indicated by a comparison of the present value of pumping through
one 600 mm line (57 140/100 m ) with the present value of pumping
through two 600 mm lines ( 5 7 500/100 m).

BALANCING STORAGE

An aspect which deserves close a t t e n t i o n in p l a n n i n g the p i p e l i n e


system is reservoir storage. Demands such as those f o r domestic and
industrial water fluctuate with the season, the day of the week and
time of day. Peak-day demands a r e sometimes i n excess of twice the
mean annual demand whereas peak draw-off from r e t i c u l a t i o n systems
may be six times the mean for a day. It would be uneconomic to
provide pipeline capacity to meet the peak draw-off rates, and
b a l a n c i n g r e s e r v o i r s are n o r m a l l y constructed at the consumer end ( a t
the head of the r e t i c u l a t i o n system) to meet these peaks. The storage
c a p a c i t y r e q u i r e d v a r i e s i n v e r s e l y w i t h the p i p e l i n e c a p a c i t y .
The b a l a n c i n g storage requirement for any known draw-off pattern
and pipeline capacity may be determined w i t h a mass flow diagram:
15

Plot cumulative draw-off over a period versus time, and a b o v e t h i s


curve plot a line with slope equal to the discharge capacity of the
pipeline. Move this line down till it just touches the mass draw-off
period. Then the maximum o r d i n a t e between the two lines represents
the b a l a n c i n g storage r e q u i r e d (see F i g . 1.4).
An economic comparison is necessary to determine the optimum
storage capacity for a n y p a r t i c u l a r system (Abramov, 1969) b y adding
t h e cost o f r e s e r v o i r s and p i p e l i n e s and c a p i t a l i z i n g running c o s t s f o r
different c o m b i n a t i o n s and c o m p a r i n g them, the system w i t h least total
cost is selected. It is found that t h e most economic storage capacity
v a r i e s f r o m one d a y ' s s u p p l y b a s e d o n t h e mean a n n u a l r a t e f o r s h o r t
pipelines to two day's supply for long pipelines (over 60 km).
Slightly more storage may b e economic for small-bore pipelines (less
than 450 mm diameter). In a d d i t i o n a certain amount of emergency
reserve storage should be provided; up to 12 h o u r s d e p e n d i n g u p o n
the a v a i l a b i l i t y of maintenance f a c i l i t i e s .

REFERENCES

A b r a m o v , N., 1969. M e t h o d s o f r e d u c i n g p o w e r c o n s u m p t i o n in p u m p i n g
w a t e r . I n t . W a t e r S u p p l y Assn. C o n g r e s s , V i e n n a .
Berthouex, P.M., 1971. Accommodating u n c e r t a i n forecasts. J. Am.
W a t e r Works Assn., 66 ( 1 ) 14.
I n s t n . o f C i v i l Engs., 1962. A n I n t r o d u c t i o n t o E n g i n e e r i n g Economics,
London.
Osborne, J.M. and James, L.D., 1973. M a r g i n a l economics a p p l i e d t o
p i p e l i n e d e s i g n . P r o c . Am. SOC. C i v i l E n g s . , 99 ( T E 3 ) 637.
W h i t e , J.E., 1969. Economics o f l a r g e d i a m e t e r l i q u i d p i p e l i n e s . P i p e -
l i n e News, N.J.

L I S T OF SYMBOLS

C - cost p e r u n i t o f t h r o u g h p u t

D - diameter
n - number of y e a r s
Q - throughput
r - i n t e r e s t r a t e o n s i n k i n g fund
R - i n t e r e s t r a t e on loan
CHAPTER 2

HYDRAULICS

THE FUNDAMENTAL EQUATIONS OF FLUID FLOW

The three most important equations in fluid mechanics are the


continuity equation, the momentum equation a n d t h e energy equation.
For steady, incompressible, one-dimensional flow the c o n t i n u i t y equa-
tion is simply obtained by equating the flow rate at any section to
the flow rate at another section along the stream tube. By 'steady
flow' is meant that there is no variation in velocity at any point
with time. 'One-dimensional' flow implies that the flow is along a
stream tube and there is no lateral flow across the boundaries of
stream tubes. I t also implies that the flow i s irrotational.
The momentum equation stems from Newton's basic law of motion
and states that the change in momentum f l u x between two sections
equals the sum of the forces on the fluid causing the change. For
steady, one-dimensional flow t h i s i s

AFx = pQAVx (2.1 1


where F i s the force, p i s the f l u i d mass density, Q i s the volumetric
flow r a t e , V i s velocity a n d s u b s c r i p t x r e f e r s to the 'x' direction.
The basic energy equation is d e r i v e d b y e q u a t i n g the work done
on an element of fluid by gravitational and pressure forces to the
change in energy. Mechanical a n d heat energy transfer a r e excluded
from the equation. In most systems there i s energy loss due to f r i c -
tion and turbulence and a t e r m i s included i n the equation to account
for this. The resulting equation for steady flow of incompressible
f l u i d s i s termed the B e r n o u l l i equation and i s conveniently w r i t t e n a s :

where V = mean velocity at a section


V2/2g = velocity head ( u n i t s of length)
9 = g r a v i t a t iona I acce I e r a t ion
P = pressure
17

p/y = pressure head ( u n i t s of length)


Y = u n i t weight of f l u i d
Z = e l e v a t i o n above an a r b i t r a r y datum
= head loss due to friction or turbulence between
he
sections 1 G 2

The sum of the velocity head plus pressure head p l u s e l e v a t i o n is


termed the total head.
Strictly the v e l o c i t y head should be m u l t i p l i e d b y a coefficient to
account f o r the v a r i a t i o n in v e l o c i t y across the section of the conduit.
The average v a l u e of the coefficient f o r t u r b u l e n t flow i s 1.06 and f o r
laminar flow it is 2.0. Flow through a conduit is termed either
uniform or non-uniform depending on whether or not there is a
variation in the cross-sectional velocity distribution along the
conduit.
For the Bernoulli equation to apply the flow should be steady,
i.e. there should be no change in velocity at any p o i n t w i t h time.
The flow i s assumed to be one-dimensional and i r r o t a t i o n a l . The f l u i d
should be incompressible, although the equation may be applied to
gases w i t h reservations (Albertson et a l . , 1960).
The respective heads are illustrated i n Fig. 2.1. For most p r a c -
tical cases the v e l o c i t y head i s small compared w i t h the other heads,
and i t may be neglected.

ENERGY LINE

ENTRANCE LOSS
FRICTION L O S S
COtiTRACTION L O S S

FRICTION LOSS

PRESSURE H E A D
I P/ x

E L E \ A T ION

Fig. 2.1 Energy heads along a p i p e l i n e


18

F L O W HEAD LOSS RELAT I ONSH I PS

Empirical Flow Formulae


The throughput o r c a p a c i t y of a p i p e of f i x e d dimensions depends
on the t o t a l head difference between the ends. T h i s head i s consumed
b y f r i c t i o n and other ( m i n o r ) losses.
The first friction head loss/flow relationships were derived from
f i e l d observations. These e m p i r i c a l relationships are stil I popular in
waterworks practice although more rational formulae have been
developed. The head loss/flow formulae established thus are termed
,+-
conventional formulae a n d a r e u s u a l l y i n an exponential form of the

type

v = K R~ sY or s = K*Q"/D~

where V i s the mean v e l o c i t y of flow, K and K' a r e coefficients, R i s


the hydraulic radius (cross-sectional area of flow divided by the
wetted perimeter, and for a circular pipe flowing full, equals one
q u a r t e r of the diameter) a n d S i s the head g r a d i e n t ( i n m head loss
per m length of pipe). Some of the equations more f r e q u e n t l y a p p l i e d
a r e l i s t e d below:

Basic Equation SI units f.p.s. units


1.85 1.167 = 6.84
Hazen -Wi I I iams S=K1 ( V / C w ) /D K =3.03 (2.3)
1 1
1.33 K2=2.86 (2.4)
Manning S=K2(nV)2/D K2= 6.32

Chezy K 3=13.13 K =4.00 (2.5)


s=K~(v/c~)~/D 3
Darcy S= X V /~2 g ~ Dimension- (2.6)
I ess

Except f o r the Darcy formula the above equations a r e not u n i v e r s a l


a n d the form of the equation depends on the u n i t s . I t should be borne
i n m i n d t h a t the formulae were d e r i v e d f o r normal waterworks p r a c t i c e
a n d take no account of variations in gravity, temperature o r type of
liquid. They a r e f o r t u r b u l e n t flow i n pipes over 50 mm diameter. The
friction coefficients vary w i t h p i p e diameter, type of f i n i s h a n d age
of p i p e .
19

The conventional formulae a r e comparatively simple to use as they


do not involve fluid viscosity. They may be solved d i r e c t l y as they
do not r e q u i r e an i n i t i a l estimate of Reynolds number to determine the
friction factor (see next section). The rational equations cannot be
solved directly for flow. Solution of the formulae for velocity,
diameter or friction head g r a d i e n t i s simple w i t h the a i d of a slide
rule, calculator, computer, nomograph or graphs p l o t t e d on log-log
paper. The equations are of particular use for analysing flows in
pipe networks where the flow/head loss equations have to be
i t e r a t i v e l y solved many times.
The most popular flow formula in waterworks practice is the
Hazen-Williams formula. F r i c t i o n coefficients for use in t h i s equation
are tabulated i n Table 2.1. I f the formula i s to be used f r e q u e n t l y ,
solution with the aid of a chart is the most efficient way. Many
waterworks organizations use graphs of head loss gradient plotted
a g a i n s t flow f o r v a r i o u s p i p e diameters, a n d v a r i o u s C values. As the
v a l u e of C decreases w i t h age, type of p i p e a n d p r o p e r t i e s of water,
f i e l d tests a r e d e s i r a b l e f o r an accurate assessment of C.

TABLE 2.1 Hazen-Williams f r i c t i o n coefficients c


Type of Pipe Cond i t ion
New 25 years 50 years Badly
old old corroded

PVC: 150 140 140 130


Smooth concrete, AC: 150 130 120 100
Steel, bitumen I ined,
ga I vanized : 150 130 100 60
Cast i r o n : 130 110 90 50
Riveted steel, v i t r i -
f i e d woodstave: 120 100 80 45
For diameters less than 1 000 mm, subtract 0.1 ( 1 - Dmm ) C
1 000

Rational F l o w Formulae
Although the conventional flow formulae are likely to remain in
use for many years, more rational formulae are gradually gaining
acceptance amongst engineers. The new formulae have a sound scienti-
fic b a s i s backed b y numerous measurements a n d they are uni v ers al l y
20

applicable. Any consistent units of measurements may be used and


I i q u i d s of v a r i o u s viscosities a n d temperatures conform to the proposed
formulae.
The rational flow formulae for flow i n pipes a r e s i m i l a r to those
for flow past bodies o r over flat plates ( S c h l i c h t i n g 1960). The o r i g -
inal research was on smal I-bore pipes w i t h a r t i f i c i a l roughness. Lack
of d a t a on roughness f o r l a r g e pipes has been one deterrent to use of
the r e l a t i o n s h i p s i n waterworks practice.
The v e l o c i t y in a full p i p e v a r i e s from zero on the boundary to a
maximum in the centre. Shear forces on the w a l l s oppose the flow and
a boundary layer i s established w i t h each a n n u l u s of f l u i d i m p a r t i n g
a shear force onto an inner neighbouring concentric annulus. The
resistance to relative motion of the fluid is termed kinematic
viscosity, and in turbulent flow it is imparted by turbulent mixing
w i t h t r a n s f e r of p a r t i c l e s of d i f f e r e n t momentum between one l a y e r and
the next.
A boundary layer i s established at the entrance to a conduit and
this layer gradually expands until it reaches the centre. Beyond t h i s
point the f l o w becomes uniform. The length of pipe required for f u l l y
established flow i s g i v e n by S c h l i c h t i n g , (1966).

X
-
D = 0.7 Re'"for t u r b u l e n t flow. (2.7)

The Reynolds number Re = VD/V is a dimensionless number


i n c o r p o r a t i n g the f l u i d viscosity v which i s absent i n t h e conventional
flow formulae. Flow i n a p i p e i s l a m i n a r f o r low R e !less than 2 000)
and becomes t u r b u i e n t for higher Re ( n o r m a l l y the case i n practice!.
The basic head loss equation is derived by setting the boundary
shear force over a length of pipe equal to the loss in pressure
m u l t i p l i e d by the a r e a :

TnDL = v h nD2/4
f

= L
XE vz
-
29
21
Fig. 2.2 Moody r e s i s t a n c e d i a g r a m for u n i f o r m f l o w i n conduits
22

where X = ( 4 7 / y ) ( V Z / 2 s ) ( r e f e r r e d to as the Darcy friction factor),


is the shear stress, D is the pipe diameter and hf is the friction
head loss over a length L . X i s a f u n c t i o n of Re and !he relative
roughness e/D. F o r l a m l n a r flow, Poiseuille found that X = 64/Re i.e. X
i s independent of the r e l a t i v e roughness. L a m i n a r flow will not occur
i n normal engineering practice. The transition zone between laminar
and turbulent flow is complex and undefined but is also of little
ivterest i n practice.
Turbulenr flow conditions may occur with either a smooth or a
rough boundary. The equations for the friction factor for both
conditions are derived from the general equation for the velocity
distribution in a turbulent boundary layer, which is derived from
m i x i n g l e n g t h theory:

Integrating, w i t h k = 0.4 a n d c o n v e r t i n g logs to base 10,

where v is the velocity at a distance y from the boundary. For a


hydrodynamical l y smooth b o u n d a r y there is a laminar sub-layer, and
Nikuradse found that y‘= v / m s o that

/+= 5.75 log Y m V


+ 5.5 (2.10)

The constant 5.5 was found e x p e r i m e n t a l l y .


Where the boundary i s r o u g h the l a m i n a r sub-layer i s affected and
Nikuradse found that y’ = e/30 where e i s the boundary roughness.

Thus /%=V 5.75 log + 8.5 (2.11)

Re-arranging equations 2.10 a n d 2.11 and expressing v i n terms of


the average v e l o c i t y V b y means of the equation Q = vdA w e get

1
J-T = 210g Re fl - 0.8 (2.12)

( t u r b u l e n t boundary layer, smooth b o u n d a r y ) a n d

D
Jf -
~ 2 log - + 1.14 (2.13)

(turbulent boundary layer, rough boundary)


23

Notice that for a smooth boundary, A is independent of the r e l a -


tive r o u g h n e s s e/D and for a very rough boundary it is independent
o f t h e R e y n o l d s n u m b e r Re f o r a l l p r a c t i c a l p u r p o s e s .
Colebrook and White combined E q u a t i o n s 2.12 a n d 2.13 to p r o d u c e
a n e q u a t i o n c o v e r i n g b o t h smooth a n d r o u g h b o u n d a r i e s a s w e l l a s t h e
t r a n s i t i o n zone;

1 = 1.14 - 2 l o g e +Rx
(D 9.35)
(2.14)

Their equation reduces to E q u a t i o n 2.12 for smooth pipes, and to


Equation 2.13 for rough pipes. This semi-empirical equation yields
satisfactory results for various commercially available pipes. Niku-
radse's original experiments used s a n d as a r t i f i c i a l boundary rough-
ness. h!atural r o u g h n e s s i s e v a l u a t e d a c c o r d i n g to t h e e q u i v a l e n t s a n d
roughness. T a b l e 2.2 g i v e s v a l u e s of e f o r v a r i o u s s u r f a c e s .

TABLE 2.2 Roughness of p i p e m a t e r i a l s ( H y d r a u l i c s Research station, 1969)

V a l u e o f e ~n mm f o r n e w , c l e a n surface unles? otherwise stated

Finish: Smooth Average Rough


_-__. _____
Glas?, d r a w n metals 0 0.003 0.006
S t e r l , DVC or AC 0.015 0.03 0.06
Coat?(( strel 0.03 0.06 0.15
GaIvnr>cze-d, v i t r i f i e d c l a y 0.06 0.15 0.3
Cast i r c r o r cement l i n e d 0.15 0.3 0.6
Spun concrete or wood s t a v e 0 . 3 0.6 I .5
R i v e t e d steel 1.5 3 6
F o u l sewers, t u b e r c u l a t e d
water mains 6 15 30
Ur,Iinrd rock, e a r t h 60 150 300
__-

Fortunately A i s not v e r y sensitive to t h e v a l u e o f e assumed. e


increases linearly with age for water pipes, the proportionality
constant d e p e n d i n g on local conditions. There may also be reduction
i n c r o s s s e c t i o n w i t h age.
The v a r i o u s r a t i o n a l f o m u l a e f o r A were p l o t t e d on a single graph
::J Moody and this graph is presented as Fig. 2.2. The kinematic
viscosities of water at various temperatures are listed in the
Apprndix.
The Moody d i a g r a m i s useful for calculating manually head loss if
pipe velocity or flow rate is known. Unfortunately it is not very
amenable to d i r e c t solution for the convrse, i.e. for veloc:ity, given
head loss, and a trial and error approach is necessary i.e. guess
velocity to calculate Reynolds number, then read off A , then
r e c a l c u l s t e v e l o city etc. Convergence i s f a i r l y r a p i d however.
24

The Colebrook W h i t e e q u a t i o n i s e a s i e r to use i f h e a d loss i s g i v e n


and velocity i s to b e c a l c u l a t e d . It i s not so e a s y to solve f o r head
loss or X given velocity or flow however. The Hydraulics Research
Station at Wallingford re-arranged the variables in the
Cclebrook-White equation to produce simple explicit flow/head 1055

graphs ( H y d r s u l i c Research S t a t i o n , 1969):


E q u a t i o n 2.14 may b e a r r a n g e d i n t h e form-

(2.15)

Thus for any f l u i d at a certain temperature a n d defined roughness


e, a g r a p h may b e p l o t t e d in terms of V,D and S. Fig. 2.3 i s such a
graph for water at 15OC a n d e = 0.06. The H y d r a u l i c Research S t a t i o n
has plotted similar graphs for various conditions. The graphs are
also a v a i l a b l e for non-circular sections, b y r e p l a c i n g D b y 4R. Going
a step further, the Hydraulics Research Station re-wrote the
Colebrook-Whi te e q u a t i o n in terms o f dimensionless parameters propor-
tional to V, R a n d S, but including factors for viscosity, roughness
and gravity. Using this form of the equation they produced a
universal resistance diagram in dimensionless parameters. This graph
was also published with their charts. Fig. 2.3 as an example is
d e r i v e d on a s i m i l a r b a s i s (VVatson, 1979).

Comparison of F r i c t i o n F o r m u l a e
Diskin (1960) p r e s e n t e d a useful comparison of the f r i c t i o n f a c t o r s
f r o m t h e Hazen-Will iams a n d D a r c y e q u a t i o n s :
The D a r c y e q u a t i o n may b e w r i t t e n a s

v =JTprn (2.16)
or v = c Z &R (2.17)

w h i c h i s termed t h e Chezy e q u a t i o n a n d t h e Chezy c o e f f i c i e n t i s

cz = &p (2.18)

The Hazen-Wi I I iams equation may be rewritten for al I practical


purposes in the f o l l o w i n g dimensionless form:
0.15
S = 5 1 5 ( V / C W ) * (Cw/Re) /gD (2.19)

By c o m p a r i n g this with t h e Darcy-Weisbach equation (2.16) i t may


be deduced t h a t
26

The Hazen-Williams coefficient Cw is therefore a functionof X


and Re and values may be p l o t t e d on a Moody diagram (see Fig.
2.2). It will be observed from Fig. 2.2 that lines for constant
Hazen-Wi I I iams coefficient coincide with the Colebrook-Whi te I ines o n l y
in the t r a n s i t i o n zone. I n the completely t u r b u l e n t zone f o r non-smooth
pipes the coefficient wi II actual l y reduce the greater the Reynolds
number i .e. one cannot associate a c e r t a i n Hazen-Wil I iams coefficient
with a p a r t i c u l a r p i p e as i t v a r i e s depending on the flow rate. The
Hazen-Will iams equation should therefore be used w i t h c a u t i o n f o r h i g h
Reynolds numbers and r o u g h pipes. It will also be noted that values
of C w above approximately 155 a r e impossible to a t t a i n in water-works
6
practice ( R around 10 ) .
The Manning equation is widely used for open channel flow and
p a r t f u l l pipes. The equation i s

(2.21

where K is 1.00 in SI u n i t s and 1.486 in ft I b units, and R i s the


h y d r a u l i c r a d i u s A/P where A i s the cross-sectional area of flow and
P the wetted perimeter. R is D/4 for a c i r c u l a r p i p e , a n d i n general
for non-circular sections, 4R may be substituted f o r D.

TABLE 2.3 Values of M a n n i n g ' s ' n '

Smooth glass, p l a s t i c 0.010


Concrete, steel ( b i tumen I i n e d )
galvanized 0.01 1
Cast i r o n 0.01 2
Slimy or greasy sewers 0.013
Rivetted steel, v i t r i f i e d wood-stave 0.015
Rough concrete 0.017

MINOR LOSSES

One method of expressing head loss through f i t t i n g s and changes


i n section i s the e q u i v a l e n t l e n g t h method, often used when the con-
ventional friction loss formulae are used. Modern p r a c t i c e i s to ex-
press losses through f i t t i n g s in terms of the v e l o c i t y head i.e. he =
KV2/2g where K is the loss coefficient. Table 2.4 gives typical loss
27

coefficients although valve manufacturers may also provide sup-


plementary data and loss coefficients K which will vary with gate
opening. The v e l o c i t y V to use i s n o r m a l l y the mean t h r o u g h the f u l l
bore of the p i p e o r f i t t i n g .

TABLE 2.4 Loss coefficients f o r p i p e f i t t i n g s

Bends hBKBV2/ 2 g
Bend a n g l e Sharp r/D=l 2 6

30" 0.16 0.07 0.07 0.06


45" 0.32 0.13 0.10 0.08
60 O 0.68 0.18 0.12 0.08
90 1.27 0.22 0.13 0.08
1 80° 2.2
90" w i t h g u i d e vanes 0.2

r = r a d i u s of bend to centre of p i p e

A significant reduction in bend loss i s possible if the r a d i u s i s


f l a t t e n e d i n the p l a n e o f the bend.

Valves hv = KvV2/2g
Type: Opening: 1/ 4 1 /2 3/4 Ful I

SLuice 24 5.6 1 .o 0.2


Butterf I y 120 7.5 1.2 0.3
Globe 10
Needle 4 1 0.6 0.5
Ref I u x 1-2.5

Contractions a n d expansions in cross-section

Contract ions : Expansions


h = K V 2/2g
hc= KcVJ29 c c l

Wal I-Wal I A 2 4 Al/A2


Angle
___ 0
- 0.2 0.4
- __ __ 0.6 -0.8 -
1.0 0 0.2 0.4 0.6 0.8 1.0
7.5" .13 .08 .05 .02 0 0
150 .32 .24 .15 .08 .02 0
30 O .78 .45 .27 .13 .03 0
180" .5 .37 .25 .15 .07 0 1 .O .64 .36 .17 -04 0
28

Entrance and e x i t losses: he = KeV2/2g


Entrance Exit

Protruding 0.8 1 .o
Sharp 0.5 1 .o
Bevel l e d 0.25 0.5
Rounded 0.05 0.2

PRESSURE AND F L O W CONTROL IN PIPES

I ntroduct ion
Valves and other fittings in pipes which reduce the h e a d may be
the cause of vapour bubble formation downstream. When the pressure
is reduced water vaporizes. The bubbles may subsequently collapse
giving rise to cavitation damage to the pipe or valve (Ball; 1970,
Knapp et al, 1961). The geometry of the valve as well as the
upstream and downstream pressure, degree of valve closure and flow
r a t e affec? the c a v i t a t i o n p o t e n t i a l as well as the vapour p r e s s u r e of
the fluid. The conditions under which cavitation commences are
referred to as critical. The cavitation damage increases as flow
velocity increases and as the ratio of upstream to downstream
p r e s s u r e increases.
A number of empirical relationships for identifying the s a f e zone
of operation of control v a l v e s h a v e been p r o p o s e d b y v a l v e s u p p l i e r s .
The popular measure of potential cavitation problem is the pressure

reduction ratio, i.e. r a t i o of u p s t r e a m to d o w n s t r e a m p r e s s u r e s . Many


commercially available valves are said to operate efficiently for
pressure reducing ratios less than 3 or 4. It i s h o w e v e r more l o g i c a l
to work w i t h a c a v i t a t i o n index a s indicated Jater. Since t h e geometry
of the valve influences the critical conditions, the application of
various types of valves for pressure reducing or flow control is
described.

TYPES OF VALVES

Valves are used for two aistinct purposes in pipelines, namely


i s o l a t i n g o r flow control.
29

I s o l a t i n g Valves
Isolating valves are used to close off the flow through a pipe.
They should be operated in the fully open or fully closed position
only. They h a v e poor flow c h a r a c t e r i s t i c s i n the p a r t l y open position.
i.e. the o b s t r u c t i o n to the flow i s l a r g e l y the exposed area of a gate
and there is often l i t t l e pressure reduction o r flow control until the
gate is practically shut. At that stage cavitation is possible and
damage to the seat and v a l v e body may r e s u l t . T h i s type of v a l v e i s
not designed to reduce the pressure a p p r e c i a b l y . i.e. it i s designed
to pass the f u l l flow when f u l l y open w i t h minimal head loss.
Such valves include gate valves, butterfly valves and r o t a r y or
spherical valves. These valves are generally operated by hand and
designed to close as rapidly as possible. Although gate valves are
operated b y spindles with screw threads, butterfly valves and rot ary
v a l v e s can be operated b y t u r n i n g a h a n d l e t h r o u g h 90" to f u l l y close
the valve from the fully open position. I n order to overcome high
seating pressures however a n d sometimes to control the r a t e of closure
these valves may have reduction gear boxes in order to drive the
spindle indirectly from the handwheel. Alternatively electrical and
pneumatic control systems can be used to operate these valves.
Typical sl(iice valves, butterfly valves and spherical valves are
i!lustrated in Figs. 10.1, 10.2 and 10.4, The relationship between
open bore area a n d degree of t u r n i n g of the handwheel is generally
non l i n e a r , I t can be observed i n the case of the gate valve that the
area reduces rapidly in the final stages of closure. This is
u n f a v o u r a b l e w i t h respect to water hammer pressures as the velocity
i s reduced r a p i d l y d u r i n g the f i n a l stages of closure and this gives
rise to higher water hammer pressures than tne uniform rate of
reduction in velocity in the pipe. The problems of cavitation are
g e n e r a l l y avoided w i t h such \ ~ a l \ / e sas they a r e not operated i n p a r t l y
closed c o v d i t i o n s except i n emergencies.

Control Valves
Various types of v a l v e s h a v e been designed to control the flow of
water i n pipelines b y linearly r e d u c i n g the open a r e a a n d to c o n t a i n
the c a v i t a t i o n effect due to reduction of pressure through the v a l v e . A
30

typical valve in common use is illustrated in Fig. 10.3 ( a needle


valve). Diaphram type valves, slotted r a d i a l flow sleeve v a l v e s ( F i g .
2.4) and plunger type control v a l v e s a r e also employed a n d there is
a v a r i a t i o n of the r a d i a l flow valve which has a piston covering the
slots. Multiple expansion type valves and basket type high duty
control valves (Fig. 2.5) are even more appropriate for pressure
reducing. The former types r e l y on a s i n g l e increase i n velocity and
any head d i s s i p a t i o n due to expansion whereas the l a t t e r types which
have variable resistance trim rings or layered baskets reduce the
pressure i n a number of stages a n d a r e s a i d to s u f f e r less c a v i t a t i o n
damage as a result. This is largely because the pressure reducing
r a t i o i s l i m i t e d at each stage b u t o v e r a l l it may be r e l a t i v e l y h i g h .
It i s a l s o recognised that these v a l v e s a r e less noisy when reducing
h i g h pressures.
The needle v a l v e was once used widely i n power s t a t i o n p r a c t i c e
as it was recognised as a streamlined v a l v e which could control flow
at a l l stages of opening. I t often h a d a p i l o t needle f o r f i n a l seating
or the initial opening. The valve i s however expensive owing to its
COI struction and i s now replaced to a l a r g e extent by simpler valves
using sophisticated materials such as elastometers ( t h e sleeve type
valve) or by pistons o r diaphrams i n the case of low pressure valves
(e.g. Dvir, 1981). The latter type of valve i s not sensitive to poor
quality water as the f l e x i b l e d i a p h r a m prevents ingress of dirt into
the workings. On the other hand the diaphram cannot accommodate
h i g h pressure d i f f e r e n t i a l s and s u f f e r s wear a f t e r prolonged operation.
It may also be subject to vibration and instability as a venturi
action i s caused d u r i n g the l a s t stages of closing.
Piston type v a l v e s can b e accurately c o n t r o l l e d p r o v i d e d the flow
is towards the exposed face of the piston. In some types the
downstream face of the p i s t o n i s exposed to the downstream pressure
or the atmosphere and i s supported by a spring. I n others the down-
stream face of the p i s t o n can be subject to f l u i d pressure depending
on the operation of a p i l o t control system. Such p i l o t systems can be
operated hydraulically, that is off the water pressure. I n t h i s case
strainers a r e often r e q u i r e d to prevent the p i l o t tubes blocking. The
system can also be operated off electrical solenoids or from a
31

Fig. 2.4 F l e x f l o Sleeve v a l v e f o r f l o w control

~ i 2 . 5~ , .i i g h d u t y control v a l v e n i t n t r i p l e aasi(ets foi' heao


reduction
p n e u m a t i c s u p p o r t system.
The control system c a n be designed to limit t h e downstream pres-
sure to a c e r t a i n maximum f i g u r e . As t h e u p s t r e a m p r e s s u r e b u i l d s UP
a n d t h e sensor d o w n s t r e a m senses an i n c r e a s i n g p r e s s u r e i t w i l l open
a pilot valve held b y means o f a spring and this pilot valve will in
t u r n p e r m i t a b i g g e r p r e s s u r e f r o m u p s t r e a m t o come b e h i n d t h e p i s t o n
and partly close the ports. Alternatively the valve may b e u s e d to
control flow rate in which case the p r e s s u r e on either side of an
orifice or venturi is measured. The pressure difference in turn
controls the p i l o t v a l v e w h i c h w i l l e i t h e r p e r m i t a g r e a t e r p r e s s u r e to
get behind the piston to close the valve or exhaust the pressure
b e h i n d t h e p i s t o n to open t h e v a l v e d e p e n d i n g on w h i c h i s r e q u i r e d .

CAVITATION IN CONTROL VALVES

The mechanism whereby flow is reduced o r p r e s s u r e i s reduced in a

control valve is by converting the pressure energy upstream into a


high velocity jet and then the velocity energy is dissipated in
turbulence as it emerges into the open pipe again. The conversion
from statical pressure energy to v e l o c i t y e n e r g y i s obtained from the
Bernoul I i e q u a t i o n ;

(2.22)

Here P is the upstream pressure, W is the unit weight of water so


that P/W is the pressure head, Z is the elevation above a fixed
datum a n d V i s the w a t e r velocity. Thus if 2 i s a c o n s t a n t then t h e
velocity p o s s i b l e assuming t h e p r e s s u r e h e a d i s r e d u c e d to z e r o w o u l d
be V = w h e r e H = P/W. This however n e g l e c t s u p s t r e a m v e l o c i t y
a n d downstream p r e s s u r e a l t h o u g h the latter can be accounted f o r b y
replacing h by ah. To account for upstream velocity head a
coefficient is often introduced in the equation and in fact the p i p e
velocity i s often given in t h e f o l l o w i n g f o r m

v = c
d
JzsnF; (2.23)
where Cd is the discharge coefficient which may vary depending on
the v a l v e opening. Apart from the f a c t that t h e d e g r e e of o p e n i n g o f
the valve affects the discharge coefficient, even if the discharge
33

coefficient was separated from the degree of opening i t would be found


to v a r y owing to the d i f f e r e n t contraction r a t i o as a d i f f e r e n t area i s
open.
The actual stage of throttling at which c a v i t a t i o n commences can
only be found from experiment. Cavitation has been classified as
incipient i.e. the p o i n t at which vapour bubbles commenced to form,
a n d choking i.e. the stage a t which the d i s c h a r g e coefficient o f the

valve is appreciably affected by vaporization of the water down-


stream. The v a p o r i z a t i o n occurs because the pressure head i s reduced
and the energy i s transformed i n t o k i n e t i c energy which impinges i n t o
a low pressure downstream. A number of parameters for assessing
cavitation have been proposed (Winn and Johnson, 1970). The
significant factor in the indication of cavitation is the cavitation
index.

(2.24)

where P is the pressure a n d s u b s c r i p t d r e f e r s to downstream, u to


upstream a n d v to vapour pressure.
For streamlined valves cavitation may not occur until the cavi-
tation index drops as low as 0.1 whereas f o r p o o r l y designed v a l v e s
cavitation often occurs at a higher index i.e. nearer 1.0. In any
case i t w i l l be found on inspection that most control v a l v e s experience
a degree of c a v i t a t i o n at normal o p e r a t i n g pressure r a t i o s . Neglecting
Pv it i s seen from the c a v i t a t i o n index that if the c r i t i c a l index is
0.1 then the pressure can be reduced by a factor of nearly 10
whereas the pressure reduction ratio possible with unstreaml ined
v a l v e s i s much lower. For an average c r i t i c a l c a v i t a t i o n index of 0.33
the pressure reducing ratio i s from 2.24 neglecting vapour pressure,
4.0.
Various forms of the c a v i t a t i o n index have been proposed ( T u l l i s ,
1970). Very little is however known about the scaling down of the

c a v i t a t i o n index and much work has to be done i n t h i s f i e l d yet.


T h e method of quantitatively assessing c a v i t a t i o n i s a l s o d i f f i c u l t .
Various methods have been attempted i.e. estimation of the increase of
velocity owing to the increase in bulk volume caused by bubble
format ion, photographic methods, pressure measurements due to
collapse of the bubbles and sonic- methods. Fig. 2.6 shows the
v a r i a t i o n i n noise measured on a r a d i a l flow sleeve v a l v e at d i f f e r e n t
c a v i t a t i o n indices f o r a f i x e d degree of opening.

I n t e r a c t i o n between C a v i t a t i o n and Water Hammer Pressures

It is recognised that the collapse of vapour bubbles when the


pressure subsequently increases may cause erosion of steel. T h i s and
the accompanying noise a r e often the symptoms of a poorly selected
valve.
There is however an interesting relationship between the water
hammer wave celerity and the cavitation potential. As a valve is
closed so the b u b b l e formation increases a n d t h i s reduces the water
hammer wave c e l e r i t y downstream (Ch. 4 a n d 5 ) .
If the free gas fraction can be related to the cavitation index
which in turn is a f u n c t i o n of the upstream or downstream pressure
then it i s possible to solve sirnuttaneously the water hammer equation,
the valve discharge equation and the cavitation index equation
together with the equation relating gas release to cavitation index,
f o r free gas volume downstream, downstream pressure a n d water ham-
mer wave c e l e r i t y . It i s made d i f f i c u l t however b y the fact that p a r t
of the cavity i s not gas b u t vapour which r a p i d l y turns into l i q u i d
when the pressure again increases. Small q u a n t i t i e s of gas may be
contained in the bubbles as they are released due to the drop in
pressure. The bubbles may therefore not collapse completely. If the
equations are solved (iteratively) then it will be found that the
pressure downstream would initially drop rapidly as the valve is
closed due to the water hammer deceleration effect. Then as bubbles
begin to form the wave c e l e r i t y reduces r a p i d l y and the consequent
pressure lowering i s reduced so that the flow reduces more s t e a d i l y .
This study neglects wave reflections and so on but it does hold
promise f o r the p o t e n t i a l of cavitation to effect better control of flow
as a v a l v e closes especially d u r i n g the i n i t i a l stages of closure where
this theory indicates that flow reduction is more rapid than was
previous1 y considered the case.
35

REFERENCES

A l b e r t s o n , M.L., B a r t o n , J.R. a n d Simons, D.B., 1960. F l u i d M e c h a n i c s


f o r E n g i n e e r s . P r e n t i c e H a l l , N.J.
B a l l , J.W., 1970. C a v i t a t i o n d e s i g n c r i t e r i a . I n T u l l i s (1970). Proc.
I n s t . Colorado State Univ.
D i s k i n , M.H., Nov. 1960. The l i m i t s o f a p p l i c a b i l i t y o f t h e Hazen-
W i l l i a m s f o r m u l a e . L a H o u i l l e B l a n c h e , 6.
D v i r , Y., 1981. P r e s s u r e r e g u l a t o r s in w a t e r s u p p l y systems. Water
a n d I r r i g a t i o n Review, Water Works A s s o c i a t i o n o f I s r a e l .
H y d r a u l i c s Research S t a t i o n , 1969. C h a r t s f o r t h e H y d r a u l i c Designs o f
C h a n n e l s a n d PiDes. 3 r d E d n . . H.M.S.O., London.
K n a p p , R., D a i l y , 'J.W. a n d Hammitt, F.G.',1961. C a v i t a t i o n . McGraw-
Hill.
S c h l i c h t i n g , H., 1960. B o u n d a r y L a y e r T h e o r y . 4 t h E d n . , McGraw-Hill
N.Y.
T u l l i s , J.P., 1970. Control o f f l o w i n c l o s e d c o n d u i t s . Proc. I n s t .
Colorado State U n i v e r s i t y .
Watson, M.D., J u l y 1979. A s i m p l i f i e d a p p r o a c h t o t h e s o l u t i o n o f p i p e
f l o w p r o b l e m s u s i n g t h e Colebrook-White method. C i v i l Eng. in
S.A., 2 1 ( 7 ) , p p 169-171.
Winn, W.P. a n d Johnson, D.E. December 1970. C a v i t a t i o n p a r a m e t e r s
f o r o u t l e t v a l v e s . Proc. ASCE, HY12.

LIST O F SYMBOLS

A - cross-sectional a r e a of flow
C - Hazen-Will iams f r i c t i o n f a c t o r
C' - friction factor
Cz - Chezy f r i c t i o n f a c t o r
d - d e p t h o f water
D - diameter
e - N i k u r a d s e roughness
f - Darcy f r i c t i o n factor ( e q u i v a l e n t to A )
Fx - force
g - gravitational acceleration
he - h e a d loss
hf - f r i c t i o n h e a d loss
K - loss c o e f f i c i e n t
L - length of c o n d u i t
n - Manning friction factor
P - wetted p e r i m e t e r
p - pressure
36

Q - flow rate
R - hydraulic radius
Reynolds number
Re -
S h y d r a u l i c gradient
v mean velocity across a section
V velocity a t a point
X distance along conduit

Y distance from boundary


2 elevation

Y specific weight

P mass density
T shear stress
V k i n e m a t i c viscosity
x Darcy f r i c t i o n f a c t o r - ( f i n USA)

F i g . 2.6 Noise output from a control valve


37

CHAPTER 3

PIPELINE SYSTEM ANALYSIS AND DESIGN

NETWORK ANALYSIS

The flows through a system of interlinked pipes or networks a r e


controlled by the difference between the pressure heads a t the input
points and the residual pressure heads at the drawoff points. A
steady-state flow pattern will be established in a network such that
the f o l l o w i n g two c r i t e r i a a r e satisfied:-
(1) The net flow towards any j u n c t i o n or node i s zero, i.e., inflow
must equal outflow, and
(2) The net head loss around any closed loop is zero, i.e., only
one head can e x i s t at any point a t any time.
The l i n e head losses a r e u s u a l l y the o n l y significant head Losses
and most methods of analysis are based on this assumption. Head
loss relationships for pipes are usually assumed to be of the form
h = K?Qn/Dm where h is the head loss, P i s the p i p e length, €Ithe
flow and D the i n t e r n a l diameter of the pipe.
The calculations are simplified if the friction factor K can be
assumed t h e same f o r a l l p i p e s i n the network.

E q u i v a l e n t Pipes f o r Pipes i n Series o r P a r a l l e l


It i s often useful to know the e q u i v a l e n t p i p e which would g i v e
the same head loss and flow as a number of interconnected pipes
in series or parallel. The equivalent pipe may be used in place
of the compound pipes to perform f u r t h e r flow c a l c u l a t i o n s .
The equivalent diameter of a compound p i p e composed of sections
of different diameters and lengths in series may be calculated by
equating the total head loss for any flow to the head loss through
the e q u i v a l e n t p i p e of length equal to the l e n g t h of compound pipe:-

z
K ( Z P ) Q ~ / D= ~ K~P Q " / D ~

.. De =
(3.1)
38

( m i s 5 i n the Darcy f o r m u l a a n d 4.85 in the Hazen-Williams f o r m u l a ) .


Similarly, the e q u i v a l e n t diameter of a system of pipes i n p a r a l l e l
i s derived by equating the total flow through the equivalent pipe ' e '
to the sum of the flows through the i n d i v i d u a l pipes ' i ' i n p a r a l l e l :
Now h = h.

i.e. KP Q " / D ~ =~ K P ~ Q ~ ~ / D ~ ~
e

So Q. =

and Q =

So cance l i n g out Q , a n d b r i n g i n g D and P to the l e f t h a n d side,


e

and i f each P i s the same,

De = [ C ( D i m / n ) ] n / m (3.2)

The equivalent diameter could also be d e r i v e d u s i n g a flow/head


loss c h a r t . For pipes in parallel, assume a reasonable head loss and
read off the flow through each pipe from the chart. Read o f f the
e q u i v a l e n t diameter which would g i v e the total flow a t the same head
loss. For p i p e s i n series, assume a reasonable flow a n d c a l c u l a t e the
total head loss w i t h assistance of the c h a r t . Read off the equivalent
pipe diameter which would discharge the assumed flow with the total
head loss across i t s length.
I t often speeds network analyses to s i m p l i f y p i p e networks as much
as possible using equivalent diameters for minor pipes in series or
parallel. Of course the methods of network a n a l y s i s described below
could always be used to a n a l y s e flows through compound pipes and
this is in fact the p r e f e r r e d method f o r more complex systems than
those discussed above.

Loop F l o w Correction Method


The loop method and the node method of a n a l y s i n g p i p e networks
both involve successive approximations speeded by a mathematical
technique developed b y Hardy C r o s s (1936).
39

The steps i n ba ancing the flows i n a network b y the loop method


are:
Draw the pipe network schematically to a clear scale. Indicate
all inputs, drawoffs, fixed heads and booster pumps (if pres-
ent).
If there is more than one constant head node, connect pairs

of constant head nodes or r e s e r v o i r s b y dummy pipes represented

by dashed lines. Assume a diameter and length and calculate


the flow corresponding to fixed head loss. In subsequent flow
corrections, omit this pipe but include it n calculating head
I osses.
Imagine the network as a pattern of closed loops i n any order.
To speed convergence of the solution some of the major pipes

may be assumed to form large superimposed loops instead of


assuming a series of loops side by side. Use only as many
loops as are needed to ensure that each pipe is in at least
one loop.
Starting with any pipe assume a flow. Proceed around a loop
containing the pipe, calculating the flow i n each pipe by sub-
tracting drawoffs and flows to other loops at nodes. Assume
flows to other loops if unknown. Proceed to n e i g h b o u r i n g loops
one a t a time, on a s i m i l a r basis. I t w i l l be necessary to make
as many assumptions as there are loops. The more accurate
each assumption the speedier w i l l be the solution.
Calculate the head loss in each pipe in any loop using a
formula such as h = Kk'Qn/Dm or use a flow/head loss chart
( p r e f e r a b l e i f the a n a l y s i s i s to be done b y h a n d ) .
Calculate the net head loss around t h e loop, i.e., proceeding
around the loop, add head losses and s u b t r a c t head g a i n s u n t i l
arriving at the starting point. If the net head loss around
the loop i s not zero, correct the flows arol;nd the loop b y add-
ing the following increment in flow in the same direction that
head losses were c a l c u l a t e d :

(3.3)

T h i s equation i s the f i r s t o r d e r approximation to the differen-


t i a l of the head loss equation and i s d e r i v e d as follows:-
40

Since h = KEQn/Dm
dh = KtnQn-'dQ/Dm = (hn/Q)dQ

Now t h e t o t a l head loss around each loop should be zero, i.e.,

c ( h + dh) = 0
C h + x (hn/Q)dQ = 0

The v a l u e of h/Q i s a l w a y s p o s i t i v e ( o r zero i f h a n d Q a r e z e r o ) .

(7) I f there is a booster pump i n any loop, subtract the generated


head from Ch before m a k i n g the flow correction u s i n g the above
equations.
(8) The flow around each loop i n t u r n i s corrected thus (steps 5-71.
(9) Steps 5-8 a r e repeated u n t i l the head arourld each loop balances
to a s a t i s f a c t o r y amount.

The Node Head Correction Method


With the node method, instead of assuming initial flows around
loops, initial heads are assumed at each node. Heads a t nodes a r e
corrected b y successive approximation i n a s i m i l a r manner to the way
flows were corrected for the loop method. The steps in an a n a l y s i s
a r e as follows:-
(1) Draw the pipe network schematically to a clear scale. Indicate
all inputs, drawoffs, f i x e d heads a n d booster pumps.
(2) Assume initial arbitrary heads at each node (except i f the head
at that node is fixed). The more accurate the initial
assignments, the speedier will be the convergence of the solu-
tion.
(3) Calculate the flow in each pipe to any node with a variable
head using the formula Q = (hDm/Kt)'/" or using a flow/head
loss c h a r t .
(4) Calculate the net inflow to the specific node a n d if this i s not
zero, correct the head b y a d d i n g the am'ount

-Q
*" = C(Q/nh) (3.4)
41

T h i s equation i s d e r i v e d as follows:-

dQ = Qdh/nh

We r e q u i r e c ( Q + dQ) = 0

But dH = -dh

so

F l o w Q a n d head loss h a r e considered positive if towards the


node. H i s the head a t the node. i n p u t s ( p o s i t i v e ) a n d drawoffs
( n e g a t i v e ) a t the node should be i n c l u d e d i n E Q .
(5) Correct the head a t each variable-head node in s i m i l a r manner,
i.e. repeat steps 3 a n d 4 f o r each node.
(6) Repeat the procedure (steps 3 to 5 ) until a l l flows balance to a
sufficient degree of accuracy. I f the head difference between the
ends of a pipe i s zero at any stage, omit the pipe from the
p a r t i c u l a r b a l a n c i n g operation.

A l t e r n a t i v e Methods of A n a l y s i s
Both the loop method and the node method of b a l a n c i n g flows in
networks can be done m a n u a l l y b u t a computer i s preferred for large
networks. If done manually, calculations should be set out well in
tables or even on the pipework layout drawing if there i s sufficient
space. Fig. 3.1 i s an example analysed m a n u a l l y b y the node method.
There a r e s t a n d a r d computer programs a v a i l a b l e f o r network a n a l y s i s ,
most of which use the loop method.
The main advantage of the node method i s that more i t e r a t i o n s a r e
required than for the loop method to achieve the same convergence,
especially if the system is very unbalanced to start with. It is
n o r m a l l y necessary f o r a l l pipes to h a v e the same o r d e r of head loss.
There are a number of methods for speeding the convergence. These
include overcorrection in some cases, or using a second order
approximation to the d i f f e r e n t i a l s f o r ca I cu I at i n g correct ions.
The node head correction method i s slow to converge on account of
42

the fact that corrections d i s s i p a t e through the network one p i p e a t a


time. Also the head correction equation has an amplification factor
( n ) a p p l i e d to the correction which causes overshoot.

60m
h
3Oomm
!a
x
lQIh
2OOOm -
*26 1140 1 7

s/h
0.7
0,7
0.7
0
0 97
0.7

+ +5s 80 1.5
+ 5 4 80 1.5
tS2.7 80 1,s

0 0 0
.1.7 0 5,3
*0,7 6 9

NOTES
Heads i n metres, flows i n l i t r e s per second, diameters i n millimetres,
lengths i n metres. Arrows i n d i c a t e p o s i t i v e d i r e c t i o n of h C Q
( a r b i t r a r y assumption). Blackened c i r c l e s i n d i c a t e nodes w i t h f i x e d
heads, numbers in circles indicate order i n which nodes were
corrected. Head losses evaluated from F i g . 2.3.

1.85 Q in
H =
Q/ h

Fig. 3.1 Example of node method of network flow a n a l y s i s


43

The loop flow correction method has the disadvantage in data


preparation. Flows must be assumed around loops, and drawoffs are
defined indirectly in the assumed pipe flows. The added e f f o r t in
data preparation and interpretation often offsets the q u i c k e r conver-
gence. This i s so because of relatively low computing costs compared
with data assembly. Trial and error design is also cumbersome if
loop flows h a v e to b e changed each time a new p i p e i s added.

Network A n a l y s i s b y L i n e a r Theory
The Hardy Cross methods of network a n a l y s i s a r e s u i t e d to man-
ual methods of solution but suffer drawbacks in the effort required
in comparison w i t h computer o r i e n t a t e d numerical methods. The latter
involve the simul taneous solution of sets of equations describing
flow and head balance. Simultaneous solution has the effect that
very few iterations are required to balance a network when compared
with the number of iterations for the loop flow correction a n d node
head correction methods. O n the other h a n d solution of a l a r g e num-
ber of simul taneous equations, even if rendered linear, requires
a l a r g e computer memory a n d many i t e r a t i o n s .
Newton-Raphson techniques for successive approximation of non-
linear equations a r e mathematically sophisticated b u t the engineering
problem becomes subordinate to the mathematics. Thus Wood and
Charles (1972) l i n e a r i z e d the head loss equation, i m p r o v i n g the l i n e a r

approximation at each step and establishing equations for head


balance around loops.
Isaac and M i l l s (1980) s i m i l a r l y l i n e a r i z e d the head loss equation
as follows f o r flow between nodes i and j :

8..
IJ
= C..+(H.-H.)
IJ I J
//m (3.5)

where the term in the square root sign is assumed a constant for
each i t e r a t i o n . If the Darcy f r i c t i o n equation is employed,

Substitute equation 3.5 into the equation for flow balance at each
node :
EQ.. = Q.
i IJ J

where Q. is the drawoff at node j and Q.. i s the flow from node
J iJ
i to node j, negative if from j to i. There i s one such equation
f o r each node.
If each Q.. is replaced by the l i n e a r i z e d expression i n 3.5, one
'J
has a set of simultaneous equations (one for each node) which can
be solved for H at each node. The procedure is to estimate H at
each node initially, then solve f o r new H ' s . The procedure i s repeat-
ed u n t i l s a t i s f a c t o r y convergence i s obtained.

OPTIMIZATION O F PIPELINE SYSTEMS

The previous section described methods for calculating the flows


in pipe networks with or without closed loops. For any particular
pipe network layout and diameters, the flow pattern corresponding
to fixed drawoffs or inputs at various nodes could be calculated.
To design a new network to meet c e r t a i n drawoffs, i t would be neces-
sary to compare a number of possibilities. A proposed layout would
be analysed and if corresponding flows were just sufficient to meet
demands and pressures were s a t i s f a c t o r y , the layout would be accep-
table. If not, it would be necessary to try alternative diameters
for pipe sizes a n d a n a l y s i s of flows i s repeated u n t i l a satisfactory
solution is at hand. This trial and error process would then be
repeated for another possible layout. Each of the final networks
so d e r i v e d would then have to b e costed a n d that network w i t h least
cost selected.
A technique of determining the least-cost network directly, with-
out recourse to trial and error, would be desirable. No direct and
positive technique is possible for general optimization of networks
with closed loops. T h e problem i s that the r e l a t i o n s h i p between p i p e
diameters, flows, head losses and costs i s not linear and most rou-
tine mathematical optimization techniques require linear relation-
ships. There a r e a number of s i t u a t i o n s where mathematical optirniza-
tion techniques can be used to optimize l a y o u t s a n d these cases a r e
discussed and described below. The cases are normally confined to
45

single mains or tree-like networks for which the flow in each b r a n c h


is known. To optimize a network with closed loops, random search
t e c h n i q u e s o r s u c c e s s i v e a p p r o x i m a t i o n t e c h n i q u e s a r e needed.
Mathematical optimization techniques are also known as systems
analysis techniques (which is an incorrect nomenclature as they are
design techniques not analysis techniques), or operations research
techniques (again a name not really descriptive). The name
mathematical optimization techniques will be retained here. Such
techniques include simulation ( o r mathematical modelling) coupled with
a selection technique such as steepest path ascent or random
searching.
The direct optimization methods include dynamic programming,
which is useful for optimizing a series of events o r things, transport-
ation programming, which is u s e f u l for a l l o c a t i n g s o u r c e s to d e m a n d s
and linear programming, for inequalities (van der Veen,1967 and
Dantzig, 1963). Linear programming usually requires the use of a
computer, b u t there a r e standard optimization programs available.

D y n a m i c P r o g r a m m i n g f o r O p t i m i z i n g Compound P i p e s

One o f the simplest optimization techniques, and indeed one which


can normally be used without recourse to computers, i s dynamic pro-
gramming. the technique is i n fact o n l y a systematic way of selecting
an optimum program from a s e r i e s of e v e n t s a n d does n o t involve any
mathematics. The technique may be used to select t h e most economic
diameters of a compound pipe which may vary in diameter along its
length depending on pressures and flows. For instance, consider a
trunk main supplying a number of consumers from a reservoir. The
diameters of the trunk main may be reduced as drawoff takes place
along the line. The problem is to select t h e most economic d i a m e t e r
f o r each section of p i p e .
A simple example demonstrates the u s e of the technique. Consider
the p i p e l i n e in Fig. 3.2. Two c o n s u m e r s d r a w w a t e r f r o m t h e p i p e l i n e ,
and the head at each d r a w o f f p o i n t i s n o t to d r o p b e l o w 5 m, neither
should the hydraulic grade line drop below the p i p e p r o f i l e a t any
point. The e l e v a t i o n s of each p o i n t a n d the l e n g t h s of each section of
pipe are indicated. T h e c o s t o f p i p e i s EO.l p e r mm d i a m e t e r per m of
46

pipe. (In this case the cost is assumed to be independent of the


pressure head, although it is simple to take account of such a
variation). The a n a l y s i s w i l l be s t a r t e d at the downstream end of the
pipe (point A ) . The most economic arrangement w i l l be w i t h minimum
residual head i.e. 5 m, at point A. The head, H, at p o i n t B may be
a n y t h i n g between 13 m a n d 31 m above the datum, b u t to s i m p l i f y the
analysis, we will only consider three possible heads with 5 m
increments between them a t p o i n t s B a n d C.

Fig. 3.2 P r o f i l e of p i p e l i n e optimized b y dynamic programming

The diameter D of the pipe between A and B, corresponding to


each o f the three allowed heads may be determined from a head loss
c h a r t such as F i g . 2.3 and i s i n d i c a t e d i n Table 3.1 (1) along w i t h
the corresponding cost.
We will also consider only three possible heads at point C. The
number of possible hydraulic grade lines between B and C is 3 x

3 = 9, but one of these is at an adverse g r a d i e n t so may be dis-


regarded. In Table 3.1 (I I) a set of figures is presented f o r each
possible hydraulic grade line between 6 and C. Thus if H = 13
B
and HC = 19 then the hydraulic gradient from C to B i s 0.006 and
the diameter required for a flow of 110 P / s is 310 rnm (from Fig.
2.3). The cost of t h i s p i p e l i n e w o u l d be 0.1 x 310 x 1 000 = f31000.
Now to this cost must be added the cost of the p i p e between A a n d
B, i n t h i s case f60 000 (from Table 3.1 (I)). For each possible head
47

TABLE 3.1 D y n a m i c p r o g r a m m i n g o p t i m i z a t i o n of a compound p i p e

1 1: 1 1 I
I

.004 300 1 6 0 0 0 0

.0065 260 52000

23 .009 250 50000

111

19 1 -006 I 310 62000

L - ._
24 .0035 340 68000
8 3000
__-
I. 5 1000:'f
-
29 .001 430 86000

I --i
79000
165000
48

HC there i s one minimum total cost of p i p e between A a n d C, marked

with an asterisk. I t i s t h i s cost a n d t h e c o r r e s p o n d i n g d i a m e t e r s o n l y


which need be recalled when proceeding to the next section of pipe.
In this example, the next section between C and D is the last and

t h e r e i s o n l y one p o s s i b l e h e a d a t D, namely the r e s e r v o i r level.

In Table 3.1 (It I) the hydraulic gradients and corresponding


diameters and costs f o r Section C - D are indicated. To t h e c o s t s of
pipe for this section are added the costs of the optimum pipe

arrangement up to C. This i s done f o r each p o s s i b l e head at C, and

the least total cost selected from Table 3.1 (I1 I). Thus the minimum
possible total cost is S151 000 and the most economic diameters are

260, 310 and 340 mm for Sections A - B, B - C and C - D respec-


tively. It may be desirable to keep pipes to standard diameters in
which case the nearest s t a n d a r d diameter could b e selected for each
section as the c a l c u l a t i o n s proceed or each length could be made up
of two sections; one w i t h the next l a r g e r s t a n d a r d diameter a n d one
with t h e n e x t smal l e r s t a n d a r d d i a m e t e r , b u t w i t h t h e same t o t a l head
loss a s the theoretical result.

Of course many more s e c t i o n s o f p i p e c o u l d b e c o n s i d e r e d and t h e


accuracy would be increased by considering more possible heads at

each s e c t i o n . The cost o f the p i p e s c o u l d be v a r i e d w i t h presssures. A


booster pump station could be considered a t a n y point. i n w h i c h case
i t s cost a n d c a p i t a l i z e d p o w e r cost should be added in the tables. A
c o m p u t e r may p r ove useful if many p o s s i b i l i t i e s a r e to b e c o n s i d e r e d ,
and there a r e s t a n d a r d dynamic programming programs a v a i lable.
It will b e seen t h a t t h e t e c h n i q u e of dynamic programming reduces
the n u m b e r of p o s s i b i l i t i e s to be c o n s i d e r e d b y s e l e c t i n g t h e least-cost
arrangement at each step. Kaliy (1969) a n d B u r a s a n d Schweig (1969)
describe a p p l i c a t i o n s of t h e t e c h n i q u e to s i m i l a r a n d o t h e r p r o b l e m s .

T r a n s p o r t a t i o n programming f o r least-cost a l l o c a t i o n of resources


Transportation programming is another technique which normally

does not required the use of a computer. The technique is of use


p r i m a r i l y for a l l o c a t i n g t h e y i e l d o f a n u m b e r of sources to a number

of consumers such that a least-cost system is achieved. The cost of

delivering the resource along each route should be linearly


49

proportional to the throughput along that route and for t h i s reason


the technique is probably of no use in selecting the optimum pipe
sizes. I t i s of use, however, i n selecting a least-cost pumping p a t t e r n
through an existing pipe distribution system, provided the friction
head is small in comparison with static head, or for obtaining a
p l a n n i n g g u i d e before demands a r e accurately known.

CONSUMER M N
REQUIREMENT 10 L l S
15 L I S

SOURCE
YIELD 12 L I S

Fi9. 3.3 Least-cost a l l o c a t i o n p a t t e r n f o r t r a n s p o r t a t i o n programming


example

An example serves to i l l u s t r a t e the technique. I n t h i s example, there


a r e two sources o f water, A and B, a n d two consumers, M and N. A
and €3 could d e l i v e r 1 2 a n d 20 e / s respectively a n d M and N r e q u i r e
10 a n d 15 4 / s respectively. thus t h e r e i s a s u r p l u s of water. The cost
of pumping along routes A - M, A - N, I3 - M and B - N a r e 5, 7, 6
a n d 9 p/1 000 l i t r e s respectively.
The data are set out in tabular form for solution in Tabel 3.2
(I). Each row represents a source and each column a demand. The
unit cost of delivery along each route i s indicated i n the top right
corner of the corresponding block in the table. The f i r s t step i s to
50

make an arbitrary initial assignment of resources in such a manner


that each yield and demand is satisfied. Starting with the top left
block of the table, the maximum possible allocation is 10. This
satisfies the d e m a n d of column M and the amount is written in the
bottom left c o r n e r o f b l o c k AM. P r o c e e d i n g to t h e n e x t c o l u m n , since
the f i r s t column i s completed, the maximum possible a l l o c a t i o n in the
first row i s 2, which satisfies the y i e l d of r o w A. So t h e n e x t b l o c k
t o b e c o n s i d e r e d i s i n r o w 0 , n a m e l y c o l u m n N. Proceed through the
table making the maximum possible assignment a t each stage u n t i l a l l
resources a r e allocated (even i f to the slack column). Thus the next
allocation is the 13 in the second row, then t,he 7 in the third
column.

mi
TABLE 3.2 T r a n s p o r t a t i o n p r o g r a m m i n g o p t i m i z a t i o n o f an a l l o c a i o n

( 1 ) CONSUMER : M N SURPLUS EVALUAT ON


SOURCE Y I E L D REQUIRE- 10 15 7 NUMBER :
MENT 5 :
0
A 12

---J 2
B 20 7 13
EVALUATION NUMBER: 5 7 -2

( 1 1 )
0
A 12 ~~

2
B 20 10 7

4 7 -2

Once an initial allocation is made the figures are re-arranged


methodically until a least-cost d i s t r i b u t i o n emerges. To d e c i d e w h i c h
would be the most p r o f i t a b l e arrangement, assign a relative evalua-
t i o n n u m b e r t o e a c h r o w and c o l u m n a s f o l l o w s : -
Assign the value 0 to row 1 and work out the other evaluation
numbers such that t h e sum o f the row evaluation number a n d column
evaluation number is equal to the cost coefficient for any occupied
51

block. The v a l u e f o r column M is 5, for column N i s 7, f o r row B i s


2, and so on. Now write the sum of the row and column e v a l u a t i o n
numbers beneath the cost coefficient of each unoccupied block If this
sum if bigger than the cost coefficient of the block, it would p a y to
introduce a resource a l l o c a t i o n i n t o t h e block. T h i s i s not easy to see
immediately, but stems from the method of determining each e v a l u a t i o n
sum from the cost coefficients of occupied blocks. T h e biggest possible
r a t e of improvement i s i n d i c a t e d b y the biggest difference between the
evaluation sum and the cost coefficient. The biggest and i n fact in
our case the o n l y , improvement would b e to introduce an amount into
Block BM. The maximum amount which can be put in block BM is
determined b y drawing a closed loop u s i n g occupied blocks as corners
(see the dotted c i r c u i t i n Table 3.2 ( I ) ) . Now f o r each u n i t which i s
added to block BM, one u n i t would have to be subtracted from block
BN, added to block AN and subtracted from block AM to keep the
yields and requirements consistent. In this case the maximum
allocation to BM is 10, since this would evacuate block AM. The
maximum r e - d i s t r i b u t i o n i.e. 10 i s made, and the amount i n the block
a t each corner of the closed loop adjusted by 10 to s a t i s f y y i e l d s a n d
requirements. Only one r e - d i s t r i b u t i o n of resources should be done at
a time.
After making the best new allocation, re-calculate the e v a l u a t i o n
number and evaluation sums as in Table 3.2 (11). Allocate resource
to the most p r o f i t a b l e block and repeat the r e - d i s t r i b u t i o n procedure
until there is no further possible cost improvement, indicated by
the fact that there i s no e v a l u a t i o n sum g r e a t e r than the cost coef-
ficient in any block. In our example we arrived at the optimum
distribution in two steps, but more compl icated patterns involving
more sources and consumers may need many more attempts.
The example can o n l y serve to introduce the subject of t r a n s p o r t a -
tion programming. There are many other conditions which are dealt
with in textbooks on the subject of mathematical optimization techn-
niques such as Van der Veen (1967) and Dantzig (1963) and this
example o n l y serves as an introduction. For instance, if two blocks
in the table happened to be evacuated simultaneously, one of the
blocks could be allocated a very small quantity denoted by 'e' say.
Computat ions then proceed as before and the q u a n t i t y 'e' disregarded
at the end.
52

L i n e a r Programming f o r Design of Least-Cost Open Networks

L i n e a r programming i s one of the most powerful optimization tech-


niques. The use of a computer is normally essential for complex
systems, although the simple example given here is done by hand.
The technique may o n l y be used if the r e l a t i o n s h i p between v a r i a b l e s
is linear, so it is restricted in application. Linear programming
cannot be used for optimizing the design of pipe networks with
closed loops without resort to successive approximations. It can be
used to design trunk mains or tree-like networks where the flow in
each branch is known. Since the relationships between flow, head
loss, diameter and cost are non-l inear, the following technique is
used to render the system linear: For each b r a n c h o r main p i p e , a
number of pre-selected diameters a r e allowed and the length of each
p i p e of different diameter is treated as the v a r i a b l e . The head losses
and costs a r e l i n e a r l y p r o p o r t i o n a l to the respective p i p e lengths. The
program will indicate that some diameter pipes h a v e zero lengths,
thereby i n effect e l i m i n a t i n g them. Any other type of linear constraint
can be treated in the analysis. I t may be r e q u i r e d to m a i n t a i n the
pressure at certain points in the network above a f i x e d minimum (a
I inear inequality of the greater-than-or-equal-to-type) or within a
certain range. The total length o f p i p e of a c e r t a i n diameter may be
r e s t r i c t e d because there i s i n s u f f i c i e n t p i p e a v a i l a b l e .

DIAMETER m m 250 200 2 OD 150


UNKNOWN L E N t T H m X, x2 x3 XL
OPTIMUM L E N G T H m 50 I50 0 L 00
HEAD LOSS m 012 3.2 0 1 - 6 6 TOTAL 5 . 0

Fig. 3.4 Least-cost t r u n k main b y l i n e a r programming

The example concerns a trunk main w i t h two drawoff points (Fig.


53

3.4). The p e r m i s s i b l e diameters of the f i r s t leg a r e 250 a n d 200 mm,


and of the second leg, 200 and 150 mm. There are thus four
v a r i a b I es , X,, X2, X3 and X4 which are the lengths of pipe of dif-
ferent diameters. This simple example could be optimized b y manual
comparison of the costs of all alternatives giving the correct head
loss, but linear programming is used here to demonstrate the tech-
nique.
The head losses per 100 m of p i p e a n d costs p e r m for the v a r i o u s
pipes a r e i n d i c a t e d below:-

D iameter Head loss @ 40 e / ’ s @ 14 e/s cost


mm m/1OO m m/1OO m i~100/100 m
250 0.25 5
200 0.71 0.1 4
150 0.42 3

The linear constraints on the system are expressed in equation


form below and the coefficients of the equations are tabulated in
Table 3.3 (I). Lengths a r e expressed i n h u n d r e d metres.
5
Lengths x1 + X2
x3 +
x4 = 4
Head L o s s : 0.25X1 + 0.71X2 + 0.1X3 + 0.42X4 = 5

Objective Function: 5x1 + 4x2 + 4x3 + 3xq = m i n i mum

The computations proceed by setting all real variables to zero,


so it i s necessary to introduce artificial slack variables into each
equation to satisfy the equality. The slack v a r i a b l e s a r e designated
a, b and c in Table 3.3 (I), and their cost coefficients are set
at very h i g h values designated m. To i n i t i a t e the solution, the slack
variables a, b a n d c a r e assigned the values 5, 4 and 5 respective-
l y (see the t h i r d column of Table 3.3 (I)),
The numbers i n any particular l i n e of the main body of the t a b l e
indicate the amount of the program variable which would be dis-
placed by introducing one unit of the column variable. Thus one
u n i t o f X, would displace 1 u n i t of a and 0.25 u n i t s of c.
To determine whether it is worthwhile replacing any variable
i n the program b y any other v a r i a b l e , a number known as the oppor-
54

T A B L E 3.3 L i n e a r P r o g r a m m i n g S o l u t i o n of P i p e P r o b l e m s

P-qg.
V c r able _

a
co>t
C o e_
ff.
m
Avo
_ _ unt

5
!- :l
x2
4
x3
4
x4
3
a
m
b

m
c
m
Red.

s ,

b m 4
C m 5 0.25 0.71 0.10 0.42 1 5/.71
OPPORTUNI T Y dALUE:

i x1 2 x3 x4
a b

a 5 4 4 3
4 5 1 1 1
x2 I
b m 4 1 1
1 l 4
C m 1.45 -0.46 0.1 0.42- -0.71 1 j 3.45'
lt0.46 0 4 - 1 . l m 3-1.42m 1 .71m-4 0 0

I l l a b c
x1 2 x3 x4
5 4 4 3 m m m
4 5 1 1 1 5
x2
b m 0.55 1.I 0.76 1.69 -2.38 0.5"
x4 3 3.45 -1.1 0.24 1 -1.69 2.38
11-1.lm 0 3.28-0.76m 0 1.1-0.69m 3.38-8.2

IV

4 4.5 -0.69 2.16


x2
x1 5 0.5 1 0.69 1.52 -2.16
x4 3 4 1 1 I

0 0 0.31 0 m- rn- m-

( N o f u r t h e r improvement posc iD l e 1
55

tunity number i s calculated for each column. I f one u n i t of X


1
introduced, then the cost would increase b y (5 - (1 x m) - (0 x
m ) - (0.25 x m)), which i s designated the o p p o r t u n i t y value, i.e.
the opportunity value for each column is calculated by multiplying
the entries in that column by the corresponding cost coefficients
of the program variable in the second column and subtracting the
total thus formed from the cost coefficient of the column variable.
The most p r o f i t a b l e v a r i a b l e t o introduce would be X2, since i t shows
the greatest cost reduction per unit ( o r negative o p p o r t u n i t y value).
The X2 column is now designated the key column. The key column
is that which shows the lowest opportunity value ( i n the cost mini-
mization case). Only one v a r i a b l e may be introduced a t a time.
To determine the maximum amount of the key column variable
which may be introduced, calculate the replacement ratios for each
row as follows:-
Divide the amount of the program variable for each row by the
corresponding number i n the key column. The lowest p o s i t i v e replace-
ment ratio is selected as that is the maximum amount which could
be introduced without v i o l a t i n g any of the constraints. The row with
the lowest positive replacement ratio is designated the key row a n d
the number at the intersection of the key column and key row, the
key number.
After introducing a new variable, the matrix is rearranged
( T a b l e 3.3 (11)) so that the replacement ratios remain correct. The
program v a r i a b l e and i t s cost coefficient in the key row a r e replaced
by the new variable and its cost coefficient. The amount column
as well as the body of t h e t a b l e a r e r e v i s e d as follows:-
Each number i n the key row i s d i v i d e d b y the key number.
From each number in a non-key row, subtract the corresponding
number i n the key row m u l t i p l i e d b y the r a t i o of the o l d row number
in the key column divided by the key number. The new tableau is
g i v e n as T a b l e 3.3 (11).
The procedure of studying opportunity values and replacement
ratios and revising the table is repeated until there is no further
negative opportunity value. In the example Table 3.3 (IV) shows
all positive opportunity values so the least-cost solution i s at hand
56

(indicated by the current program v a r i a b l e s and t h e i r corresponding

values).
The reader should r e f e r to a s t a n d a r d textbook on l i n e a r program-

ming (e.9. Van der Veen, (1967) a n d D a n t z i g (1963)) f o r a full des-


cription of the technique. There are many other cases which can
be mentioned below:-
If the constraints are of the 5 (less-than-or-equal-to) type and
not just equations, slack variables with zero cost coefficients
are introduced into the 1.h.s. of each constraint to make them

equations. The artificial slack variables with high cost coef-


f i c i e n t s a r e then omitted.
If the constraints are of the 2 (greater-than-or-equal-to) type,
introduce artificial slack variables with high cost coefficients
into the 1.h.s. of the constraint and subtract slack variables

with zero cost coefficients from each inequality to make them


equations.
If the objective function is to be minimized, the opportunity
value with the highest negative value is selected, but if the
function is to be maximized, the opportunity value with the
h i g h e s t p o s i t i v e v a l u e i s selected.

The opportunity values represent shadow values of the corres-

ponding variables i.e. they indicate the value of introducing

one u n i t of that v a r i a b l e i n t o the program.

If two replacement ratios a r e equal, whichever row i s selected,


the amount of program variable in the other row will be zero
when the matrix is rearranged. Merely assume it to have a
very s m a l l v a l u e a n d proceed as before.

example of two stage optimization, namely first the layout and


then the pipe diameters is given in Stephenson (1964). Non-linear
programming and search methods are also discussed. For instance

Lam (1973) described a g r a d i e n t o p t i m i z a t i o n method.


57

REFERENCES

Buras, N. and S c h w e i g , Z., 1969. A q u e d u c t r o u t e o p t i m i z a t i o n b y


d y n a m i c p r o g r a m m i n g . Proc. Am. SOC. C i v i l E n g r s . 95 ( H Y 5 ) .
Cross, H., 1936. A n a l y s i s o f f l o w i n n e t w o r k s of c o n d u i t s o r c o n d u c t -
o r s . U n i v e r s i t y of I I I i n o i s B u l l e t in 286.
D a n t z i g , G.B., 1963. L i n e a r P r o g r a m m i n g and E x t e n s i o n s , P r i n c e t o n
U n i v e r s i t y Press, Princeton.
Isaacs, L.T. a n d M i l l s , K.G., 1980. L i n e a r t h e o r y m e t h o d s f o r p i p e
n e t w o r k s a n a l y s i s . P r o c . Am. SOC. C i v i l E n g r s . 106 ( H Y 7 ) .
Kal l y , E., 1969. P i p e l i n e p l a n n i n g b y d y n a m i c computer program-
m i n g . J . Am. W a t e r W o r k s Assn., (3).
Lam, C.F., 1973. D i s c r e t e g r a d i e n t o p t i m i z a t i o n of w a t e r systems.
P r o c . Am. SOC. C i v i l E n g r s . , 99 ( H Y 6 ) .
S t e p h e n s o n , D., 1984. P i p e f l o w A n a l y s i s , E l s e v i e r , 204 pp.
V a n d e r Veen, B., 1967. I n t r o d u c t i o n t o t h e T h e o r y of O p e r a t i o n a l
R e s e a r c h . C l e a v e r Hurne, L o n d o n .
Wood, D.J. and C h a r l e s , O.A., 1972. N e t w o r k a n a l y s i s u s i n g l i n e a r
t h e o r y . ? r o c . A X E , 98 (HY7) p 1157 - 1170.

L I S T O F SYMBOLS

C - cost
D - diameter
h - head loss
H - head
K - const a n t
e - Iengt h
Q - flow
58

CHAPTER 4

WATER HAMMER AND SURGE

R I G I D WATER COLUMN SURGE THEORY

Transient pressures caused by a change of flow r a t e in conduits


are often the cause of bursts. The pressure fluctuations associated
w i t h s u d d e n f l o w s t o p p a g e c a n b e s e v e r a l h u n d r e d metres head.
Transients in closed conduits are normally classed into two cate-
gories: Slow motion mass oscillation of the fluid is referred to a
surge, whereas rapid change in flow accompanied by elastic strain
of the fluid and conduit is referred to as w a t e r hammer. For slow
or small changes in flow rate or pressure the two theories yield
t h e same r e s u l ts.
It i s normally easier to a n a l y s e a system b y r i g i d column theory
(whenever the theory is applicable) than by elastic theory. With
r i g i d column theory the w a t e r in the c o n d u i t i s t r e a t e d a s a n incom-
pressible mass. A pressure difference applied across the ends of
the column produces an instantaneous acceleration. The basic equa-
tion relating the head d i f f e r e n c e between t h e e n d s of the water col-
umq in a uniform bore conduit to the rate of change in velocity
i s d e r i v e d from Newton's b a s i c l a w o f motion, a n d is

(4.1

where h is the difference in head between the two ends, L i s the


conduit length, v is the flow velocity, g is gravitational accelera-
tion and t i s time.
The equation is useful for calculating the head rise associated
with slow deceleration of a water column. I t may be used f o r calcu-
lating the water level variations in a surge shaft following power
trip or starting up in a pumping line, or power load changes in
a hydro-electric installation fed by a pressure pipeline. The equa-
tion may be solved in steps of At b y computer, in tabular form o r
graphical ly. The f o l l o w i n g e x a m p l e demonstrates the numerical method
of s o l u t i o n o f t h e e q u a t i o n : -
59

Example
A 100 m l o n g penstock w i t h a c r o s s s e c t i o n a l a r e a , A,, o f 1 m2
is protected against water hammer b y a surge s h a f t a t the tur-
bine, with a cross sectional area, A2, of 2 m2 and an
unrestricted orifice. The initial velocity i n t h e c o n d u i t i s 1 m/s
and there is a sudden complete load rejection at the t u r b i n e .
Calculate the maximum rise in water level in the surge shaft
neglecting friction.
Take At = 1 sec. Then from Equ. 4.1, Av = -ghAt/L = -9.8h/ 100 =
-0.098h. By c o n t i n u i t y , A h = A1vAt/A2 = l v / 2 = 0.5~.

t Ah = 0 . 5 ~ h A v = -0.098h V

0.1 0.5 0.5 -0.049 0.951


1-2 0.476 0.976 -0.096 0.855
2-3 0.428 1.404 -0.138 0.717
3-4 0.359 1.763 -0.173 0.544
4-5 0.272 2.035 -0.199 0.345
5-6 0.172 2.207 -0.216 0.129
6-7 0.064 2.271" -0.223 0.094

The maximum r i s e i s 2.27 m, w h i c h may b e compared w i t h t h e a n a l y -


tical s o l u t i o n o b t a i n e d from Equ. 4.16, o f 2.26 m. The a c c u r a c y o f t h e
numerical method c o u l d be i m p r o v e d b y taking smaller time intervals
or taking t h e mean v a n d h o v e r t h e t i m e i n t e r v a l s to c a l c u l a t e A h

and Av respectively. The method c a n r e a d i i y b e e x t e n d e d to include


h e a d losses, a n d i s calculator-orientated.
Another useful a p p l i c a t i o n of t h e r i g i d w a t e r column e q u a t i o n i s w i t h
water column separation. Following the stopping of a pump at the
upstream end of a pumping line, the pressure frequently drops
sufficiently to cause vaporization at peaks along the line. In such
cases the water column beyond the vapour pocket will decelerate
slowly and r i g i d column theory is s u f f i c i e n t l y accurate for analysis.
Equ. 4.1 may be integrated twice with r e s p e c t to time t to d e t e r m i n e
the distance the water column will travel before stopping. If the
pumps stop instantaneously the volume of the vapour pocket behind
the water column o f length t w i l l be Q = A t v * / 2 9 h where v i s the
0 0
i n i t i a l flow velocity.
60

MECHAN I CS O F WATER HAMMER


Water hammer occurs, as the name implies, when a column of
water i s r a p i d l y decelerated (or accelerated). The r i g i d water column
theory would indicate an infinitely l a r g e head r i s e if a valve in a
p i p e l i n e c a r r y i n g l i q u i d were slammed s h u t instantaneously. There is,
however, a c e r t a i n amount o f r e l i e f u n d e r these c o n d i t i o n s , due to the
e l a s t i c i t y of the f l u i d a n d of the p i p e itself. Thus if a valve at the
d i s c h a r g e end of a pipeline is shut the fluid upstream of the gate
will pack against the gate, causing a pressure rise. The pressure
will rise sufficiently to stop the liquid in accordance with the
momentum l a w . The amount o f water stopped p e r u n i t time depends o n
the amount of water required to replace the volume created by the
compression of w a t e r a n d e x p a n s i o n o f t h e p i p e .
I t can thus be shown that the relationship between the head r i s e
Ah, the r e d u c t i o n in v e l o c i t y A V a n d the r a t e of p r o g r e s s of the wave
front i s A h = -CAV/g (4.2a)
where c i s t h e w a v e c e l e r i t y a n d g i s g r a v i t a t i o n a l acceleration. This
equation is often referred to a s J o u k o w s k y ’ s law. It can further be
p r o v e d from a mass b a l a n c e t h a t

c = 1/

If
.iw
3 kd
EY
the p i p e h a s a rigid lining ( e l a s t i c m o d u l u s E2 a n d
(4.3a)

thickness
y,) and r i g i d side f i l l ( m o d u l u s Es) t h e e q u a t i o n to use i s

(4.3b)

where w i s the u n i t w e i g h t of liquid, K i s i t s b u l k modulus, d is the


p i p e diameter, E i s i t s e l a s t i c modulus, y the p i p e w a l l thickness a n d
k a factor which depends on the end fixity of the pipe (normally
about 0.9). c may b e a s h i g h a s 1370 m/s for a r i g i d walled tunnel
o r as low a s 850 m/s for thin wall steel p i p e s . I n t h e case o f plastic
pipe, or when free a i r i s present (see C h a p t e r 51, c may be a s low
a s ZOO m/s.
The pressure wave caused by the valve closure referred to
earlier, thus travels upstream, superimposed on the static head, as
illustrated in Fig. 4.1(1). When the wave front reaches the open
reservoir end, the pressure in the pipe forces water backwards into
the reservoir, so that the velocity now reverses and pressure drops
b a c k to s t a t i c ( r e s e r v o i r ) p r e s s u r e a g a i n .
61

PRE5WRE NAVE

Fig. 4.1 Water hammer wave at d i f f e r e n t stages

A negative wave (2) thus travels downstream from the r e s e r v o i r .


That wave front will in t u r n reach the closed end. Now the velocity
in the entire pipeline is -v where v was the original velocity.
0 0
The negative wave of head amplitude Cv,/g below static, travels
back up the pipe to the reservoir. Upon reaching the reservoir it
sucks water into the pipe, so that the velocity in the p i p e r e v e r t s
to +v and the head reverts to static head. The sequence of waves
w i l l repeat i t s e l f i n d e f i n i t e l y unless damped b y f r i c t i o n .

---+---1

Hz g
‘with no friction

0 2 L/c 3 L/c 4L7c t


I I

Fig. 4.2 Head f l u c t u a t i o n s at v a l v e end.


62

The variation in head at the v a l v e will be as indicated i n Fig.


4.2. In the case of pumping lines, the most violent change i n flow
conditions i s normally associated with a pump trip. The water down-
stream of the pump is suddenly decelerated, resulting in a sudden
reduction in pressure. The negative wave travels towards the dis-
charge end (Fig. 4.3), where the velocity reverses and a positive
wave returns towards the pumps. The pressure at the pump alter-
nates from a pressure drop to a pressure r i s e .

Wave front

snit i d l y
Vapour pressure

Initial V = Vo

F i g . 4.3 Water hammer head drop a f t e r pump t r i p

The wave i s complicated by line friction, changes in cross sec-


tion, vaporization, or gradual flow variation. Thus if there is a
negative pressure, the water will vaporize and air will be drawn
in via air valves and from solution. The effect of f r i c t i o n can most
readily be seen from Fig. 4.4 which illustrates a wave at different
stages as it travels up the p i p e l i n e from a rapidly shut valve. The
pressure heads behind the wave are not quite horizontal, due to
the ' p a c k i n g ' effect causing some flow across the wave f r o n t .
The effect of changes i n section or b r a n c h pipes can be included
in one equation. The wave head change Ah' after r e a c h i n g a junc-
tion i s r e l a t e d to the o r i g i n a l head change Ah' by the equation

2Ah'A./c.
I I
Ah' = i4.4a)
A1/c, + A 2'=2 + A3/C3 + ...
I 2 Ah /(1 + A 2 / A 1 + A3/A1 + .. . (4.4b)

where A . i s the cross sectional area of p i p e i and c


1
. c2 c3-
63

1 2 3
Xz- F 'x "Distance from valve"
c-
vo
9

F i g . 4.4 Head a t p o i n t s a l o n g along p i p e l i n e w i t h f r i c t i o n , for


instantaneous stoppage

When a valve i n a g r a v i t y main i s closed g r a d u a l l y , the effect i s


analogous to a series of minute waves emanating from the v a l v e . The
system could be analysed numerical ly or graphically. The g r a p h i c a l
form of analysis is useful for demonstrating the principles of
simultaneous solution of the valve discharge equation and the water
hammer equation. In fact the sines d r a w n on a g r a p h a r e v e r y simi-
lar to the so-called characteristic lines adopted in numerical solu-
t ions.

Fig. 4.5 illustrates a graphical analysis of a pipeline with a


valve at the discharge end. The v a l v e i s closed over a p e r i o d equal
to 4 L/c and the valve discharge characteristics for four different
degrees of closure a r e p l o t t e d on the g r a p h . The l i n e r e l a t i n g f r i c t i o n
head loss to p i p e velocity i s also indicated.
To compute the head at the valve at time L/c after initiating
closure one applies the water hammer equation 4.2 between p o i n t s R
a n d S. Thus
64

A h -(c/g) Av - A h (4.2b)
f
where Ahf i s t h e d i f f e r e n c e i n f r c t i o n h e a d b e t w e e n R a n d S. Similarly
applying t h e w a t e r hammer e q u a t i o n f r o m S t o R ,

A h + (c/g) Av + hf (4.2~)

Computations thus proceed a l o n g l i n e s of slope + or - c/g. Ultimately


the waves peak a n d then die out d u e to f r i c t i o n . The m a x i m u m h e a d
at point S occurs at t = 4 L/c and is 182 m e t r e s a c c o r d i n g to t h e
computations on F i g . 4.5.

ELAST I C WATER HAMMER THEORY

The fundamental differential wave equations relating pressure to


velocity in a conduit may b e d e r i v e d from consideration of Newton's
l a w o f m o t i o n a n d t h e c o n s e r v a t i o n of mass r e s p e c t i v e l y , and are:

(4.5)

ah + -c 2
- - =a vo
at g ax

The last term in Equ. 4.5 accounts for friction which i s assumed
t o obey Darcy's equation with a constant f r i c t i o n factor. The assump-
tion of a steady-state friction factor for transient conditions i s not
strictly correct. Tests indicate that head losses d u r i n g t r a n s i e n t flow
are higher t h a n those p r e d i c t e d u s i n g t h e f r i c t i o n f a c t o r a p p l i c a b l e to
normal flow conditions. Energy is probably absorbed during flow
reversals when t h e v e l o c i t y i s low and t h e f r i c t i o n f a c t o r c o n s e q u e n t l y
relatively high.

Methods of A n a l y s i s
A common method of analysis of pipe systems for water hammer
pressures used to be g r a p h i c a l l y (Lupton, 1953). Fig. 4.6 i s such
a chart for the maximum a n d minimum heads a t the downstream v a l v e
for no line friction. The v a l v e area i s assumed to r e d u c e linearly to
zero over time T and the valve discharge coefficient is assumed
constant. To u s e t h e c h a r t c a l c u l a t e t h e v a l v e c l o s u r e p a r a m e t e r cT/L
65

H
( m 1
20c

6'
- 200

::m
R 5

R 2

RI

R,
-R
-. /---

_ / - -

x -t
_/--

_/--

CHART S
55

S'cVALVE

55 23

51
t'
t d t = L/c
' 0
SHUT

F i g . 4.5 Graphical water hammer a n a l y s i s f o r slow valve


closure with f r i c t i o n
66

and valve head loss parameter he/(cvo/g). Read off on the vertical
axis the maximum head parameter h ' / ( c v /g) ( f u l l lines) a n d minimum
head parameter -h/(cvo/g) (dashed lines). Multiply by cvo/g a n d the
answers are the maximum a n d minimum heads respectively above a n d
below s t a t i c head. Note t h a t the c h a r t is for linear reduction in a r e a
w i t h time, a c o n d i t i o n r a r e l y encountered in p r a c t i c e . The c h a r a c t e r i s -
tics of sluice and butterfly valves are such that most of the flow
reduction occurs a t the e n d of the valve stroke, with the result t h a t
the water hammer heads are higher than predicted by the c h a r t . A
more a c c u r a t e a n a l y s i s i s t h e r e f o r e n e c e s s a r y f o r important I ines.
T h e most economical method of solution of t h e w a t e r hammer e q u a -
tions for particular systems is by digital computer. Solution is
usually by the method o f characteristics (Streeter and Lai, 1963 a n d
Streeter a n d Wylie, 1950) w h i c h d i f f e r s l i t t l e i n p r i n c i p l e f r o m t h e o l d
graphical method. The differential water hammer equations are
expressed in finite difference form and solved for successive time
intervals. The c o n d u i t i s d i v i d e d i n t o a number of i n t e r v a l s a n d At is
set equal to ax/c. The x - t g r i d on which s o l u t i o n takes p l a c e i s
depicted in Fig. 4.7. Starting from known conditions along the
pipeline at time t, one proceeds to c a l c u l a t e the h e a d a n d velocity a t
each p o i n t a l o n g the l i n e at time t + At.
By expressing equations 4.5 and 4.6 as total differentials and
adding, one gets two simultaneous equations i n v o l v i n g dh a n d d v :

(4.7b)

Equs. 4.7a and 4.7b may be solved for and v ' at point p at
h'
P P
time t + At i n t e r m s of known h a n d v a t two other points q a n d r
at time t :

(4.8a)

(4.8b)
67

Valve closure factor


C T/L

F i g . 4.6 Maximum a n d minimum head a t downstream v a l v e f o r l i n e a r


c Io s u r e

At the terminal points, a n additional condition i s usually imposed;


either h is fixed, or v is a function of a gate opening or pump
speed. The correct Equ. 4.7a or 4.7b is solved simultaneously with
the known condition t o e v a l u a t e t h e new h a n d v at time t + A t . The
computat ions commence at known conditions and are terminated when
the pressure f l u c t u a t i o n s a r e s u f f i c i e n t l y damped b y f r i c t i o n .
Where a branch pipe s occurs or there is a change i n diameter,
t h e n Equs. 4.8a a n d 4.8b snoulci b e r e p l a c e d b y E q u s . 4.9 a n o 4.10:
68

TIME
A

/
Fig. 4.7 x - t G r i d for water hammer a n a l y s i s b y c h a r a c t e r i s t i c
method.

Effect of F r i c t i o n

Fluid friction damps the water hammer waves as they travel


along the conduit. If there is no exciting influence the waves will

gradually die away and the pressure along the conduit will tend to
s t a t i c pressure.
The characteristic method of solution by computer accurately pre-
dicts the effect of friction provided there is no discontinuity in the
wave. At a sharp wave front, it i s necessary t o r e s o r t to some o t h e r
method of analysis. Fortunately an analytical solution is feasible at
the wave front. Ludwig (1950) demonstrated that t h e a m p l i t u d e of a

water hammer wave travelling back along a line with friction

fol lowing instantaneous stoppage is indicated by the hyperbolic func-

tion (see F i g . 4.4):

xxv cv
- tanh --./p)]
( 4gd '
4 (4.11)

In pumping lines following sudden pump stopping the maximum


over-pressures at the pump will exceed the pumping head if the
friction head is greater than approximately 0.7cv / g . Fig. 4.8 indi-
0
cates minimum and maximum head envelopes along pipelines with
69

various friction heads following instantaneous stoppage at the up-


stream end. To use the chart, multiply the ordinates by cv / g and
plot the maximum a n d minimum head envelopes f o r the correct f r i c t i o n
factor hf/(cv / g ) , above o r below s t a t i c on a p i p e p r o f i l e d r a w i n g .
0
The chart is only valid provided there is no column separation
when the negative wave t r a v e l s up the p i p e l i n e i.e. the minimum head
envelope should at no point fall below 10 m below the pipeline
profile. It has, however, been found from experience that the maxi-
mum heads with column separation a r e often s i m i l a r to those without
separation.
Fig. 4.8 may also be used to determine the maximum and minimum
heads along pipelines w i th fri cti o n following sudden flow stoppage a t
the downstream end b y a closing valve. Turn the c h a r t upside down
a n d r e a d o f f maximum envelope instead of minimum a n d vice- versa.

PROTECT I ON OF PUMP I NG L I NES (Stephenson, 1972)

The pressure transients fol lowing power failure to electric motor


d r i v e n pumps a r e u s u a l l y the most extreme t h a t a pumping system w i l l
experience. Nevertheless, the over-pressures caused b y s t a r t i n g pumps
should also be checked. Pumps with steep head/flow characteristics
often induce h i g h over-pressures when t h e power i s switched on. This
i s because the flow i s small ( o r zero) when the pump i s switched on
so a wave with a head equal to the closed v a l v e head i s generated.
By partly closing the pump delivery valves during starting, the
over-pressures can be reduced.
The over-pressures caused b y closing l i ne valves or scour valves
should also be considered.
If the pumps supplying an unprotected p i p e l i n e a r e stopped sud-
denly, the flow will a l s o stop. If the pipeline profile is relatively
close to the hydraulic grade line, the sudden deceleration of the
water column may cause the pressure to drop to a value less than
atmospheric pressure. The lowest value to which pressure could drop
i s vapour pressure. Vaporization or even water column separation may
thus occur at peaks along the pipeline. When the pressure wave is
r e t u r n e d as a p o s i t i v e wave the water columns w i l l rejoin g i v i n g rise
to water hammer over-pressures.
70

Fig. 4.8 M a x i m u m and m i n i m u m h e a d e n v e l o p e s f o l l o w i n g


instantaneous pump stopping in pipelines w i t h
friction
71

Unless some method of water hammer protection is installed, a


pumping pipeline system will normally have to be designed for a
water hammer overhead equal to cv /g. I n fact t h i s i s often done w i t h
0
high-pressure lines where water hammer heads may be small in
comparison with the pumping head. For short lines this m a y be the
most economic solution, and even if water hammer protection is

i n s t a l l e d i t may be prudent to check that the u l t i m a t e s t r e n g t h of t h e


p i p e l i n e i s s u f f i c i e n t should the p r o t e c t i v e device f a i l .
The philosophy behind the design of most methods of protection
against water hammer is similar. The object in most cases is to

reduce the downsurge i n the p i p e l i n e caused b y stopping the pumps.


The upsurge will then be correspondingly reduced, or may even be
entirely eliminated. The most common method of limiting the down-
surge is to feed water into the pipe as soon as the pressure tends
to drop.

Fig. 4.9 Pipeline p ro fi l e i I lustrat ing suitable locat ions f o r various


devices f o r water hammer protection.

Suitable locations for various protective devices are i I lustrated


i n Fig. 4.9. Most of the systems i n v o l v e feeding water i n t o the pipe.
Observe that in all cases the sudden momentum c h a n g e of the water
column beyond the tank is prevented so the elastic water hammer
phenomenon is converted to a slow motion surge phenomenon. Part
of the original kinetic energy of the water column i s converted into
potential energy instead of e l a s t i c energy. The w a t e r c o l u m n g r a d u a l -
ly decelerates under the effect of the difference in heads between
the ends. If it was allowed to decelerate the water column would
gather momentum in the reverse direction and impact against the
pump to cause water hammer over-pressures. If, however, the water
column is arrested at its point of maximum potential energy, which
coincides with the point of minimum kinetic energy, there will be
no sudden change in momentum and consequently no water hammer
over-pressure. The reverse flow may be stopped by installing a re-
flux v a l v e or throttling device a t the entrance to the discharge t a n k
or air vessel, or in the pipeline. A small o r i f i c e b y p a s s to the re-
flux valve would then allow the pressures on e i t h e r s i d e to g r a d u a l -
l y equalize.
Fortunately charts are available for the design of air vessels
and for investigation of the pump inertia effects, so that a water
hammer analysis is not normally necessary. Rigid water column
theory may be employed for the analysis of surge tank action, and
i n some cases, of d i s c h a r g e tanks.
If the pipeline system incorporates in-line reflux valves or a
pump bypass valve, an elastic water hammer analysis is usually
necessary. The analysis may be done graphically or, if a number
of solutions of similar systems are envisaged, a computer program
could b e developed. Normal ly the location, s i z e and d i s c h a r g e c h a r a c -
teristics of a protective device such as a discharge tank have to

be determined by trial and error. The location and size of in-line


or bypass reflux valves may similarly have to be determined by
trial. In these instances a computer program is usually the most
economical method of solution, as a general program could be devel-
oped, and by varying the design parameters methodically, an opti-
mum s o l u t i o n a r r i v e d a t .
The following sections describe various methods o f reducing water
hammer in p u m p i n g l i n e s , a n d offer design aids.
73

Pump Inertia
If the r o t a t i o n a l i n e r t i a of a centrifugal pump and motor continue
to rotate the pump for a while after power failure, water hammer
pressure transients may be reduced. The rotating pump, motor and
e n t r a i n e d water will continue to feed water i n t o the p o t e n t i a l vacuum
on the delivery side, thereby a1 l e v i a t i n g the sudden deceleration
of the water column. The effect i s most noticeable on low-head, short
pipelines.
After the power supply to the motor is cut off, the pump will
gradually slow down until it can no longer deliver water against
the delivery head e x i s t i n g at the time. If the d e l i v e r y head is s t i l l
higher than the suction head, it will then force water through the
pump in the reverse direction, with the pump still spinning in the
forward direction, provided there is no reflux or control valve on
the delivery side of the pump. The pump will rapidly decelerate
and gather momentum in the reverse direction, and will act as a
turbine under these conditions. The reverse speed of the pump will
increase u n t i l i t reaches r u n a w a y speed. Under these conditions there
is a rapid deceleration of the reverse flow a n d water hammer over-
pressures w i I I resu I t.
I f there i s a r e f l u x v a l v e on the delivery side of the pump, the
reverse flow wi II be arrested, but water hammer overpressures wil I
still occur. The pressure changes at t h e pump f o l l o w i n g power f a i l -
u r e may be c a l c u l a t e d g r a p h i c a l l y (Parmakian, 1963) o r b y computer.
The pump speed N after a time increment A t i s obtained b y equat-
ing the work done i n decelerating the pump to the energy transferred
to the water:

Solving f o r N - N2 = L N we get
1

900wHQ A t
(4.12)
A N = n' MNFN

where M is the moment of inertia of the pump impeller and motor,


N is the speed in rprn, FN i s the pump e f f i c i e n c y a t the b e g i n n i n g
of the time interval, w i s the u n i t weight of f l u i d , H i s the pumping
head above suction head a n d q i s the discharge. As an approxima-
74

tion, FN = (N/No)Fo where s u b s c r i p t o r e f e r s to i n i t i a l conditions.


The head/discharge characteristics of the pump can be fairly
accurately represented b y the equation

t i = aN2 + bNQ - cQ2 (4.13)

a, b and c are constants which can be evaluated provided three


points on the head/discharge/speed characteristics of the pump are
known.
Substituting q = vA where A is the cross sectional area of the
pipe, and u s i n g Equ. 4.8, two equations result involving two un-
knowns which may be solved f o r h and v at the pump.
Fortunately the r e s u l t s of a l a r g e number of analyses have been
summarized in chart form. Charts presented by Parmakian indicate
the maximum downsurge and upsurge at the pumps and midway along
the pipeline for various values of pump inertia and the pipeline
parameter P. The charts also indicate the maximum reverse speed
and the times of flow reversal, zero pump speed and maximum re-
verse speed.
Kinno and Kennedy (1965) have prepared comprehensive charts
which include the effect of l i n e f r i c t i o n and pump e f f i c i e n c y ; specific
speed of the pump i s also discussed. Of interest i s a chart indicat-
ing maximum upsurge if reverse rotation of the pump is prevented.
Kinno suggests that the upsurge could be reduced considerably if
reverse flow through the pump was permitted, but with reverse r o t a -
tion prevented.
Fig. 4.10 summarizes the maximum and minimum pressures which
will occur in various pumping systems following power f a i l u r e . Val-
ues for the minimum head are taken from Kinno's paper, and the
maximum head values a r e adjusted to cover typical pump efficiencies
around 85 p e r cent. t will be observed that the minimum head often
occurs when the f i r s t wave r e t u r n s to the pump ( t = 2L/c). The v a l -
ues of the subsequen upsurge at the pump if flow reversal i s per-
mitted, are also presented in the c h a r t . These values are lower than
the magnitude of the upsurge which would occur if reverse flow was
prevented with a reflux valve. If flow reversal was prevented the
maximum headrise above operating head Ho would be approximately
equal to the lowest head-drop below H .
75

The author has a simple r u l e of thumb for a s c e r t a i n i n g whether


the pump inertia will have an effect in reducing the water hammer
pressures. If the inertia parameter I = MNZ/wALHoZ exceeds 0.01,
the pump i n e r t i a may reduce the downsurge b y at least 10 per cent.
Here M i s the moment of i n e r t i a of the pump, N i s the speed i n rpm
and AL i s the volume of water i n the pipe. The expression was d e r i v -
ed from Fig. 4.10, and it was assumed that c/g is approximately
e q u a l to 100.

F i g . 4.10 M a x i m u m and rninimiJm h e a d s at p u m p a f t e r power f a i l u r e


76

Some i n s t a l l a t i o n s have a fly-wheel f i t t e d to the pump to increase


the moment of inertia. In most cases the flywheel would have to
be impractically heavy, also i t should be borne i n mind that start-
i n g c u r r e n t s may thereby be increased.
In subsequent sections the effect of pump inertia is neglected
and the pumps a r e assumed to stop instantaneously.

Pump Bypass Reflux V a l v e

One of the simples arrangements of protecting a pumping line


against water hammer is a reflux valve installed in parallel with
the pump (Fig. 4.11). The reflux or non-return v a l v e would dis-
charge o n l y in the same d i r e c t i o n as the pumps. Under normal pump-
ing conditions the pumping head would be higher than the suction
head and the pressure difference would maintain the reflux valve
i n a closed position. On stopping the pumps, the head i n the d e l i v e r y
Pipe would tend to drop b y the amount cvo/g. The head may drop
to below the suction head, in which case water would be drawn
through the bypass valve. The pressure would therefore o n l y drop
to the suction pressure less any friction loss in the bypass. The
r e t u r n wave overpressure would be reduced a c c o r d i n g l y .

REFLJX
VALUE

S U C T I O N PIPE
J
F i g . 4.11 Pump w i t h bypass r e f l u x valve.

This method of water hammer protection cannot be used in all


cases, as the d e l i v e r y pressure w i l l often never drop below the suc-
tion pressure. I n other cases there may s t i l l b e an a p p r e c i a b l e wat-
er hammer overpressure (equal i n v a l u e to the initial drop i n pres-
sure). The method really has use only when the pumping head is
77

considerably less than cv /g. I n addition, the initial drop i n pres-


sure along the entire pipeline length should be tolerable. The suc-
tion reservoir level should also be relatively high or there may
s t i l l be column separation i n the d e l i v e r y line.
Normally the intake pipes draw directly from a constant head
reservoir. However, there may be cases where the intake pipe is
fairly long and water hammer could be a problem i n it too. I n these
cases a bypass r e f l u x v a l v e would, in a similar way to that describ-
ed above, prevent the suction pressure exceeding the d e l i v e r y pres-
sure.
It should be noted that water may also be drawn through the
pump during the period that the delivery head i s below the suction
head, especially if the machine was designed for high specific
speeds, as is the case with through-flow pumps. In some cases the
bypass r e f l u x v a l v e could even be omitted, a l t h o u g h there i s normal-
l y a f a i r l y h i g h head loss through a s t a t i o n a r y pump.

Surge Tanks
The water surface in a surge tank is exposed to atmospheric
pressure, while the bottom of the tank i s open to the p i p e l i n e . The
tank acts as a balancing tank for the flow variations which may
occur, discharging in case of a head drop in the pipe, or filling
in case of a head rise. Surge tanks are used principally at the
head of turbine penstocks, although there a r e cases where they can
be applied in pumping systems. It is seldom that the hydraulic
grade l i n e of a pumping line is low enough to enable an open tank
to be used. I t may be possible to construct a surge tank at a peak
in the pipeline profile and protect the pipeline between the pumps
and the tank against water hammer by some other means. If the
surge tank is relatively large, i t could be treated as the discharge
end of the intermediate pipeline length, and this section could be
treated as an independent p i p e l i n e shorter in length than the o r i g i n -
a l pipeline.
The f l u c t u a t i o n s of the water surface level i n a surge tank follow-
ing power failure may be studied analytically. The transients bet-
ween the pumps and surge tank are high-frequency water hammer
78

phenomena which will not affect the water level in the tank notice-
ably. The analysis concerns the slow motion surges of the water
column between the surge tank and d e l i v e r y end of the line.
R i g i d water column theory may b e used in the s t u d y , since e l a s t i c
water hammer waves will not pass the open tank. The rate of
deceleration of the water column i s d v / d t = -gh/P (4.14)
where h is the height of the delivery head above the level of the
surge tank and P i s the p i p e l i n e length between the tank and deli-
very end. The head h gradually increases as the level in the tank
drops (see F i g . 4 . 9 ) .
Another equation r e l a t i n g h a n d v may be d e r i v e d b y considering
c o n t i n u i t y of flow a t the tank o u t l e t :

A v = Atdh/t (4.15)
P
where Ap and At are the cross sectional areas of the p i p e a n d tank
respectively. Note that the flow from the pump side i s assumed to be
zero.
Solving Equs. 4.14 and 4.15 a n d u s i n g the fact that at t = 0, h =
0 and v = v to e v a l u a t e the constants o f integration, one o b t a i n s an
express ion f o r h :

(4.16)

From t h i s equation it i s apparent that the amplitude of the down-


surge in the tank is v , m and the time till the f i r s t down-
- r -

surge is (II / 2 ) J AtP/A g: Time zero i s assumed to be that i n s t a n t at


P
which the water hammer wave reaches the surge t a n k .
If there i s a g r a d u a l deceleration of the water column between the
pump a n d surge t a n k , then the above expressions w i l l not hold, and
the correct continuity equation and the rigid column equation for
deceleration will have to be solved numerically for successive time

i n t e r v a I s.
The fluctuations in tank level may be damped with a throttling
orifice, I n t h i s case the pressure v a r i a t i o n s in the l i n e may be more
extreme than for the unrestricted orifice. Parmakian ( 1 9 6 3 ) presents
c h a r t s i n d i c a t i n g the maximum upsurge f o r v a r i o u s t h r o t t l i n g losses.
A number of sophisticated variations in the design of surge tanks
have been proposed (Rich, 1963). The differential surge tank for
79

instance includes a small-diameter riser in the m i d d l e of the tank.


The tank may have a varying cross section o r m u l t i p l e shafts. Such
variations are more applicable to hydro power plant than pumping
systems, as they a r e useful f o r damping t h e surges i n cases o f r a p i d
load v a r i a t i o n on turbines.

Discharge Tanks
I n s i t u a t i o n s where the p i p e l i n e p r o f i l e i s considerably lower than
the h y d r a u l i c g r a d e l i n e i t may s t i l l be possible to use a tank, but

one which under normal operating conditions is isolated from the


pipeline. The tank water surface would be subjected to atmostpheric
pressure b u t would be below the h y d r a u l i c g r a d e line, as opposed to
that of a surge tank.

Fig. 4.12 Discharge tank

A discharge tank would normally be situated on the first rise


along the p i p e l i n e a n d possibly on subsequent a n d successively h i g h e r
rises. The tank will b e more e f f i c i e n t i n r e d u c i n g pressure v a r i a t i o n s

the nearer the level in the tank is to the h y d r a u l i c g r a d e line. It


should be connected to the pipeline via a reflux valve i n s t a l l e d to
discharge from the tank i n t o the p i p e l i n e if the p i p e l i n e head drops
below the water surface elevation in the tank. Normally the reflux
valve would be held shut by the pressure in the pumping line. A
small-bore bypass to the r e f l u x v a l v e , connected to a float valve in
the tank, should be installed to fill the tank slowly after it has
discharged. Fig. 4.12 depicts a typical discharge tank arrangement.
80

The use of discharge tanks was reported in detail by Stephenson


(1972).
The f u n c t i o n of a discharge tank i s to f i l l any low pressure zone
caused by pump stoppage, thus p r e v e n t i n g water column separation.
The water column between the tank and the discharge end of the
pipeline (or a subsequent tank) will g r a d u a l l y decelerate under the
action of the head difference between the two ends. It may be
necessary to prevent reverse motion of the water column which could
cause water hammer overpressures - t h i s could be achieved by instal-
l i n g a r e f l u x v a l v e i n the line.
A discharge tank will only operate if the water surface i s above
the lowest level to which the head in the pipeline would otherwise
drop following pump stopping. Thus the normal o p e r a t i n g head on the
first tank along a line should be less than c v /g, and subsequent
0
tanks should be successively higher. I n cases where the head on the
tank i s considerably less than cv /g, r i g i d water column theory may
be used to c a l c u l a t e the discharge from the t a n k :
Integrating Equ. 4.1 twice with respect to time, one obtains an
expression f o r the distance the water column travels before stopping.
Multiplying this distance by the cross sectional area of the pipeline
yields the volume discharged by the tank; Q = APv 2/2gh, where A is
0
the cross sectional area of the pipe, P is the distance between the
discharge tank a n d the open end of the p i p e l i n e , or the next tank if
there is one, and h is the head on the tank measured below the
h y d r a u l i c g r a d e l i n e ( o r level of the next t a n k ) .
Fig. 4.13 depicts the volume discharged from a tank expressed as
a fraction of the discharge indicated by the r i g i d column equation.
The coordinates of the g r a p h were obtained from a computer analysis
using elastic water hammer theory. I t i s i n t e r e s t i n g to note that when
the head, h, at the tank is less than approximately 0.5 cvo/g, the
discharge is accurately predicted by r i g i d column theory, and water
hammer overpressures are low. For successively higher heads rigid
column theory becomes less applicable and the water hammer
overpressures increase. I t appears, however, that the discharge from
a tank never exceeds that indicated by rigid water column theory,
p r o v i d e d the conditions discussed below a r e compl ied w i t h .
81

Fig. 4.13 i s only a p p l i c a b l e f o r a s i n g l e tank along the line. The


tank i s assumed to be a t the pump end of the p i p e l i n e , o r else there
should be an in-line reflux valve immediately upstream (on the pump
side) of the tank. Without this reflux valve, a p o s i t i v e wave would
r e t u r n from the tank towards the pumps. I n a d d i t i o n to water hammer
overpressures which would r e s u l t a t the pumps, the discharge from the
tank would exceed that f o r the assumed case where i t o n l y discharged
downstream. The maximum overpressures indicated by the c h a r t occur
along the p i p e l i n e on the d i s c h a r g e side of the tank. The overpessure
h' is measured above the discharge head, and is expressed as a
fraction of cv / g . If the upstream r e f l u x valve i s omitted F i g . 4.14
should be used.
For very long pipelines with a number of successively higher
peaks, more than one discharge tank may be installed along the
line. The tanks should be i n s t a l l e d at the peaks where water column
separation is likeliest. The lowest head which will occur at any
p o i n t beyond a tank as the downsurge t r a v e l s along the line i s that
of the water surface elevation of the preceding t a n k .
Fig. 4.15 indicates the discharge from two tanks installed
along a pipeline. The chart is also o n l y appiicable provided the
first tank is located at the pump end of the line, or else a reflux
valve should be installed immediately upstream of the tank. The
second tank should also have a reflux valve immediately upstream
of it.

Fig. 4.15 is really only applicable for cases where the distance
between the tanks is equal to the distance between the second tank
and the open end of the pipeline; however, the c h a r t s hardly vary
for different length ratios and Fig. 4.15 could be used a s a g u i d e
f o r other cases.
in Fig. 4.15 h, is the difference in elevation between the first
and second tank, and h2 i s the difference i n elevation between the
second tank and the head at the discharge end of the p i p e l i n e . The
corresponding pipe length t 1 is that between the two tanks and J?
2
is that between the second tank a n d the discharge end of the pipe-
line. The discharge from the f i r s t tank is indicated by broken lines
and the discharge from the second tank is indicated by full lines.
The relevant h and e are substituted i n the v a l u e i n d i c a t e d on the
82

tc

08

06

O L

-
9 h'
c vo
0 2

0
10

08

Q ?gh
__
A LV,'

06

OL

0.2

0
0

Fig. 4.13 D i s c h a r g e f r o m t a n k a n d m a x i m u m head r i s e w i t h


i n - l i n e r e f l u x v a l v e u p s t r e a m of t a n k .
83

F i g . 4.14 Discharge from tank a n d maximum head r i s e


(no in-line r e f l u x v a l v e ) .
F i g . 4.15 Discharge from two t a n k s a n d maximum head r i s e
( e l = @ a n d r e f l u x v a l v e upstream of each tank).
2
85

v e r t i c a l scale to o b t a i n the respective discharge Q from each t a n k .


The maximum overpressures i n d i c a t e d b y F i g . 4.15 i n v a r i a b l y occur
i n the downstream side of the second tank. I t is, however, possible to
limit the lateral extent of the overpressure b y j u d i c i o u s location of a
further reflux valve between the last tank and the delivery end of
the pipeline. In t h i s case the r e f l u x v a l v e on the pump s i d e of the
water hammer tank could possibly be dispensed with. The discharge
from the tank may then exceed that indicated b y the chart, and a
water hammer analysis should be carried out based on the theory
out1 ined e a r l i e r i n the chapter.
The best p o s i t i o n f o r discharge tanks and i n - l i n e ref l ux valves i s
selected by trial and error and experience. For simple cases the
c h a r t s presented may suffice, b u t f o r complex cases w i t h many peaks
or major pipelines with large friction heads, a complete analysis
should be c a r r i e d out, either graphically o r b y computer. In parti-
cular, a final check should be done f o r flows less than the maximum
design c a p a c i t y o f the p i p e l i n e .
Even though a number of tanks may be i n s t a l l e d along a p i p e l i n e ,
vaporization is always possible along rising sections between the
tanks. Provided there are no local peaks, and the l i ne rises f a i r l y
steeply between tanks, this I imited v a p o r i z a t i o n should not lead to

water hammer overpressures. However, air vessels should be i n s t a l led


on all rising sections. Cases are known where vaporization and
vapour bubbles t r a v e l l i n g along g e n t l y r i s i n g mains have r e s u l t e d i n
severe c a v i t a t i o n of a l i n e a l o n g the top of the pipe.

A i r Vessels
If the p r o f i l e of a pipeline is not high enough to use a surge
tank or discharge tank to protect the line, it may be possible to
force water into the p i p e b e h i n d the low-pressure wave b y means of
compressed air in a vessel (a typical air vessel arrangement is
illustrated in Fig. 4.16). The pressure i n the vessel will gradually
decrease as water i s released u n t i l the pressure i n the vessel equals
that in the adjacent line. At this stage the decelerating water
column will tend to reverse. However, whereas the outlet of the a i r
vessel should be unrestricted, the inlet should be throttled. A
86

suitable arrangement is to have the water discharge out through a


reflux valve, which shuts when the water column reverses. A
smaf I - o r i f i c e bypass would a l l o w the vessel to r e f i l I slowly.

Fig. 4.16 A i r vessel

Parrnakian ( 1 9 6 3 ) suggests that the a i r vessel inlet head loss should


be 2.5 times the outlet head loss. However, t hi s relationship i s em-
pirical a n d recent research w i t h the a i d of computers has enabled the
effect of the inlet head loss to be i n v e s t i g a t e d thoroughly. The a i r
vessel design charts reproduced by parrnakian were compiled f o r the
ratio of reverse flow head loss to forward flow head loss equal to
2.5. The author has compiled design charts covering an extended
r a n g e of p i p e l i n e s a n d v a r i o u s degrees of i n l e t t h r o t t l i n g ( F i g s . 4.17
to 4.19). The c h a r t s a r e s u i t a b l e f o r checking water column separation
a l o n g the e n t i r e p i p e l i n e length, a n d f o r r e a d i n g off the r e q u i r e d a i r
volume and degree of inlet throttling corresponding to any specified
l i m i t to the overpressures. Outflow t h r o t t l i n g i s neglected.
The c a l c u l a t i o n s for p l o t t i n g the c h a r t s were performed b y comput-
er; it was assumed that the expansion law f o r the a i r i n the vessel
1.4 -
lies between adiabatic (HS - constant) and the isothermal (HS =
Fig. 4.17 Maximum a n d minimum pressure envelopes w i t h a i r vessel.
2 = 0.5.

F i g . 4.18 M a x i m u m and m i n i m u m p r e s s u r e e n v e l o p e s w i t h a i r
vessel. P = 1.
88

constant). The relationship adopted was HS1'3 = constant. It was


further assumed t h a t t h e a i r vessel was i m m e d i a t e l y d o w n s t r e a m of t h e
pumps, t h a t the pumps would stop instantaneously and that no reverse
flow t h r o u g h the pumps was permitted.
The dimensionless parameters associated with the charts are as
fol l o w s :
Pipeline parameter P = c v /gH
0 0
A i r vessel parameter B = v 0'AL/gHoS
Throttling parameter C = Z/Ho
where S i s the initial air volume in the vessel, and 2 i s the head
loss t h r o u g h t h e a i r vessel i n l e t c o r r e s p o n d i n g to a l i n e v e l o c i t y -v
0-
The outlet head loss was assumed to be negligible. Note t h a t Ho i s
the absolute head ( i n c l u d i n g atmospheric head) i n t h i s case.

0.2 I

Fig. 4.19 Maximum a n d minimum pressure envelopes with a i r vessel.


P = 2.

The c h a r t s a r e u s e d a s f o l l o w s to d e s i g n a n a i r v e s s e l : The p i p e -
line parameter, P, i s c a l c u l a t e d from the maximum l i k e l y line velocity
89

and pumping head, and the corresponding chart selected from Figs.
4.17 to 4.19. T h e p i p e l i n e p r o f i l e is p l o t t e d on the a p p l i c a b l e set of
minimum-head curves a n d a minimum-head envelope selected such that
it does not fall below the pipeline profile at any point to cause
vaporization. The value of B corresponding to the selected line is
used to read off the maximum-head envelope a l o n g the p i p e l i n e from
the same c h a r t . The overpressures actually depend on the degree of
inlet throttling represented b y C. The normal design procedure i s to
minimize the overpressures, achieved b y s e t t i n g C approximately equal
to 10 ( o r more). ( F o r a p i p e l i n e velocity of approximately 2 m/s a n d
a pumping head of the order of 200 m, the corresponding inlet
diameter works out to be l/lOth of the main pipe diameter). I f the
overpressures indicated thus still cannot be tolerated, it will be
necessary to select a smaller value of B (corresponding to a larger
volume of a i r ) than i n d i c a t e d b y the minimum pressure requirement.
The volume of air, S, i s c a l c u l a t e d once B i s known. The vessel
c a p a c i t y should be s u f f i c i e n t to ensure no a i r escapes i n t o the pipe-
line, a n d should exceed the maximum a i r volume. This i s the volume
1/1.3
d u r i n g minimum pressure conditions a n d i s S ( H /Hmin)
The outlet diameter is usually designed to be about one-half the
main p i p e diameter. The outlet should be designed w i t h a bellmouth to
suppress vortices and air entrainment. The air in the vessel will
dissolve in the water to some extent and w i l l h a v e to replenished b y
means of a compressor.

In-L ine R e f Iux V a Ives


Normally a reflux valve i n s t a l l e d on i t s own in a pipeline w i l l
not reduce water hammer pressures, although i t may limit the later-
al extent of the shock. In fact, in some situations indiscriminate
positioning of reflux valves in a l i n e could be detrimental to water
hammer pressures. For instance if a pressure r e l i e f v a l v e was i n s t a l -
led upstream of the reflux valve the reflux valve would counteract
the effect of the other valve. It may also amplify reflections from
b r a n c h pipes o r collapse of vapour pockets.
In-line reflux valves would n o r m a l l y be used i n conjunction with
surge tanks, discharge tanks or air vessels. Following pump shut-
down, the tank or vessel would discharge water into the p i p e e i t h e r
side of the reflux valve. This would alleviate the violent pressure
90

drop a n d convert the phenomenon i n t o a slow motion effect. The reflux


valve would then arrest the water column at the time of reversal,
which coincides w i t h the p o i n t of m i n i m u m k i n e t i c energy and maximum
potential energy of the water column. There would therefore be little
momentum change in the water column when the r e f l u x v a l v e shut and
consequently n e g l i g i b l e water hammer pressure r i s e .
There a r e s i t u a t i o n s where water column separation and the forma-
t i o n of vapour pockets i n the p i p e l i n e f o l l o w i n g pump stoppage would
be tolerable, provided the vapour pockets d i d not collapse resulting
in water hammer pressures. Reversal of the water column beyond the
vapour pocket could i n fact be prevented w i t h an i n - l i n e reflux valve
at the downstream extremity of the vapour pocket. the water column
would be arrested at i t s p o i n t of minimum momentum, so there would
be l i t t l e head rise.
Vaporization would occur at peaks in the pipeline where the
water hammer pressure dropped to the vapour pressure of the water.
If the first rise along the pipeline was higher than subsequent
peaks, the vaporization wouJd be confined to the first peak. The
extent of the vapour pocket could be estimated using rigid column
theory. The decelerating head on the water column between two peaks
o r between the last peak and the discharge end would be the d i f f e r -
ence in e l e v a t i o n p l u s the intermediate f r i c t i o n head loss. Integrating
the rigid water column equation with respect to t, one o b t a i n s an
expression f o r the volume of a vapour pocket, k?AvoZ/2gh, where h i s
the decelerating head, C i s the water column l e n g t h a n d A i s the p i p e
cross sect iona I area.
In locating the reflux valve, allowance should be made f o r some
lateral dispersion of the vapour pocket. The v a l v e should be i n s t a l l e d
at a suitable dip i n the p i p e l i n e i n order to trap the vapour pocket
and to ensure proper functioning of the v a l v e doors when the water
column r e t u r n s .
A small-diameter bypass to the r e f l u x v a l v e should be i n s t a l l e d to
permit slow r e f i l l i n g of the vapour pocket, or else overpressures may
occur on restarting the pumps. The diameter of the bypass should be
of the order of one-tenth of the pipeline diameter. An air release
91

valve should be i n s t a l l e d in the p i p e l i n e a t the peak to release a i r


which would come out of solution d u r i n g the p e r i o d of low pressure.
It i s a common p r a c t i c e to i n s t a l l r e f l u x v a l v e s immediately down-
stream of the pumps. Such reflux valves would not prevent water
hammer pressures in the pipeline. They merely prevent r e t u r n flow
through the pump and prevent water hammer pressures r e a c h i n g the
pumps.
In some pump installations, automatically closing control valves,
instead of r e f l u x valves, a r e i n s t a l l e d on the pump d e l i v e r y side.
Kinno (1968) studied the effect of c o n t r o l l e d closure of a pump
delivery valve. Assuming a limited return flow through the pump
to be tolerable, he describes how water hammer overpressures can
be s l i g h t l y reduced b y c o n t r o l l i n g the r a t e of closure of the v a l v e .

Release Valves
There are a number of sophisticated water hammer release v a l v e s
(often r e f e r r e d to as surge r e l i e f v a l v e s o r surge suppressors) a v a i l -
a b l e commercial l y . These valves have h y d r a u l i c actuators which auto-
matically open then gradually close the valve after pump tripping.
The valves are normally the needle type, which discharge into a
pipe leading to the suction r e s e r v o i r , or else sleeve valves, mounted
in the suction reservoir. The v a l v e s must have a gradual throttling
effect over the complete range of closure. Needle a n d sleeve valves
are suitably designed to minimize c a v i t a t i o n a n d corrosion associated
with the high discharge velocities which occur d u r i n g the t h r o t t l i n g
process.
The v a l v e s a r e u s u a l l y i n s t a l l e d on the d e l i v e r y side of the pump
reflux valves and discharge directly to the suction reservoir. They
should not discharge into the suction pipe as they i n v a r i a b l y draw
a i r through the throat, and t h i s could reach the pumps.
The v a l v e s may be actuated by an e l e c t r i c a l fault or b y a pres-
sure sensor (as Fig. 4.20).
The valve should open fully before the negative pressure wave
returns to the pumps as a p o s i t i v e pressure wave. As the pressure
on the top of the piston increases a g a i n the v a l v e g r a d u a l l y closes,
maintaining the pressure within desired limits. The closing rate
may be a d j u s t e d b y a p i l o t v a l v e i n t h e h y d r a u l i c c i r c u i t .
92

Reflux
Pump valve

Acc u mu lot or

Suction
reservoir

Fig. 4.20 Water hammer release valve arrangement with simp1 i f i e d


h y d r a u l i c actuator

The valve should open fully before the negative pressure wave
returns to the pumps as a positive pressure wave. As the pressure
on the top of t h e p i s t o n increases a g a i n the v a l v e g r a d u a l l y closes,
maintaining the pressure within desired I i m i ts. The closing rate
may be a d j u s t e d by a p i l o t v a l v e i n the h y d r a u l i c c i r c u i t .
If no overpressure higher than the operating head is tolerable,
the v a l v e would be sized to discharge the f u l l flow at a head equal
to the operating head. Where reliability is of importance, and if
water hammer is likely to be a problem during partial shutdown
of the pumps, t w o or more release valves may be installed i n para-
llel. They could be set to operate at successively lower delivery
heads.
In the event of normal pump shutdown a g a i n s t throttled delivery
valves, the release v a l v e s could be disengaged to prevent t h e i r oper-
at ion.
The types of control v a l v e s a v a i l a b l e as release valves f o r pump-
ing lines n o r m a l l y cannot open in less than about f i v e seconds. Their
use is therefore limited to pipelines over two kilometers in length.
This method of water hammer protection is normally most economical
93

for cases when the pumping head greatly exceeds cv /g, since the
l a r g e r the pumping head, the smaller the v a l v e needed.
Since there is no protection against underpressures, a water
hammer analysis should be performed to check that water column
separation i s not a problem.
A less sophisticated valve than the control valves described
above, which has been used on small pump installations, is the
spring-loaded release v a l v e . The v a l v e i s set to open when the pres-
sure reaches a prefixed maximum. Some overpressure i s necessary to
open the v a l v e a n d to force water out.

Choice of Protective Device


The best method of water hammer protection for a pumping line
will depend on the hydraulic and physical characteristics of the
system. The accompanying table summarizes the ranges over which
v a r i o u s devices a r e s u i t a b l e . The most i n f l u e n t i a l parameter i n selec-
ting the method of protection i s the p i p e l i n e parameter P = cvo/gHo.
When the head cvo/g i s g r e a t e r than the pumping head Ho, a reflux
valve bypassing the pumps may suffice. For successively smaller
values of P it becomes necessary to use a surge t a n k , a discharge
tank i n combination w i t h an i n - l i n e r e f l u x v a l v e , a n a i r vessel, o r a
release v a l v e . The protective devices l i s t e d i n the t a b l e a r e a r r a n g e d
in approximate order of increasing cost, thus, to select the most
suitable device, one checks down the table until the variables are
w i t h i n the r e q u i r e d range.
It may be possible to use two o r more p r o t e c t i v e devices on the
same line. This p o s s i b i l i t y should not be ignored as the most econ-
omical arrangement often i n v o l v e s more than one method of protection.
In particular the rotational i n e r t i a of the pump often has a slight
effect in reducing the required capacity of a tank or a i r vessel. A
comprehensive water hammer analysis would be necessary i f a series
of protection devices i n combination i s envisaged.
94

TABLE 4.1

Summary of methods of water hammer protection

Method of Required r a n g e Remarks


Protection of v a r i a b I es
( i n approximate
order of increas-
i n g cost)

I n e r t i a of pump MN* Approxi mate on I y


wALHo2 > 0 * 0 1

Pump bypass Some water may a l s o


reflux valve be d r a w n through
pump

In-line reflux cv Normally used i n con-


0 > 1
va I ve __ j u n c t i o n w i t h some
gh other method of pro-
tection. Water column
separation possible

Surge tank h small P i p e l i n e should be


n e a r h y d r a u l ic g r a d e
l i n e so height of tank
i s Dractical

Automatic release P i p e l i n e p r o f i l e should


valve be convex downwards.
Water column separa-
tion l i k e l y .

D i scharge tanks cv h = pressure head


0 > '
a t t a n k , p i p e l i n e pro-
gh f i l e should be convex
upwards

A i r vessel cv P i p e i ine p r o f i l e p r e f e r -
o
__ < l
a b l y convex downwards
gHO
95

REFERENCES

Kinno, H. and Kennedy, J.F., 1965. Water hammer charts for


c e n t r i f u g a l p u m p systems, P r o c . Am. SOC. C i v i l E n g s . , 91 (HY3)
247-270.
Kinno, H. ,1968 W a t e r h a m m e r c o n t r o l i n c e n t r i f u g a l p u m p s y s t e m s ,
P r o c . Am. SOC. C i v i l E n g s . , 94 (HY3) pp 619-639.
L u d w i g , M. and Johnson, S.P., 1950 .Prediction of surge pressures
i n l o n g o i l t r a n s m i s s i o n l i n e s , P r o c . Am. P e t r o l e u m I n s t . , N.Y. 30
(5).
L u p t o n , H.R., 1953 . G r a p h i c a l a n a l y s i s o f p r e s s u r e surges in pump-
i n g s y s t e m s , J . ~ n s t . W a t e r E n g s . , 7.
P a r m a k i a n , J . , 1963. W a t e r Hammer A n a l y s i s , D o v e r P u b l i c . I n c . , N.Y.
R i c h , G.R., 1963. H y d r a u l i c t r a n s i e n t s , D o v e r p u b l i c s . I n c . , N.Y.
Streeter, V.L. and L a i , C., 1963. W a t e r h a m m e r a n a l y s i s i n c l u d i n g
f l u i d f r i c t i o n . P r o c . Am. SOC. C i v i l E n g s . , 88 (HY3) pp 79-112.
Streeter, V.L. and Wyl ie, E.B., 1967. H y d r a u l i c T r a n s i e n t s , McGraw-
Hill.
Stephenson, D., 1966. W a t e r h a m m e r c h a r t s i n c l u d i n g f l u i d f r i c t i o n ,
P r o c . Am. SOC. C i v i l e n g s . , 92 (HY5) pp 71-94.
Stephenson, D., 1972. D i s c h a r g e t a n k s f o r s u p p r e s s i n g w a t e r h a m m e r
i n p u m p i n g l i n e s . P r o c . I n t n l . Conf. o n p r e s s u r e surges,B.H.R.A.,
C r a n f ie l d.
Stephenson, D., 1972. Water hammer p r o t e c t i o n o f p u m p i n g l i n e s ,
t r a n s . S.A. I n s t n . C i v i l E n g s . , 14 ( 1 2 ) .

L I S T O F SYMBOLS

A - p i p e cross sectional area


B - a i r vessel p a r a m e t e r v 'AL/(gHoS)
c - t h r o t t l i n g p a r a m e t e r Z/H
0
d - p i p e l i n e diameter
E - modulus of e l a s t i c i t y of p i p e w a l l m a t e r i a l
F - pump rated efficiency ( e x p r e s s e d as a f r a c t i o n )
f - Darcy friction factor

9 - g r a v i t a t i o n a l acceleration
h - p r e s s u r e h e a d a t an i n t e r m e d i a t e s e c t i o n o f t h e p i p e l i n e
h' - water hammer h e a d r i s e measured above the d e l i v e r y h e a d
f r i c t i o n h e a d loss
hf -
h e a d loss t h r o u g h d o w n s t r e a m v a l v e f u l l y o p e n
he -
H - head in p i p e l i n e measured above pump suction reservoir level
(in c a s e of air vessel design, take H as absolute, i.e. plus
atmospheric head)
p u m p i n g head above suction r e s e r v o i r level
Ho -
I - p u m p i n e r t i a p a r a m e t e r MN'/WALH
0
'
J - p u m p p a r a m e t e r F M N Z c / l 80wALv gH
0 0
K - b u l k modulus of water
e - length of an intermediate p a r t of p i p e l i n e
L - pipeline length
M - moment of inertia of rotating p a r t s of pump, motor and en-
t r a i n e d water ( = mass x r a d i u s of g y r a t i o n ' )
N - pump speed i n rpm
P - p i p e l i n e parameter cv /gh
0 0

Q - volume of water discharged from discharge tank


s - volume of a i r i n i t i a l l y i n a i r vessel
t - time
T - l i n e a r v a l v e closure time
v - water v e l o c i t y i n p i p e l i n e
v - i n i t i a l water velocity in pipeline
w - weight of water per u n i t volume
x - distance a l o n g p i p e l i n e from pump

Y - wal I thickness o f p i p e
z - head loss through a i r vessel i n l e t f o r p i p e l i n e velocity =

Darcy f r i c t i o n factor
97

CHAPTER 5

A I R IN PIPELINES

INTRODUCTION

It is recognised that air is present in many water pipelines.


The a i r may be absorbed at free surfaces, or entrained in turbulent
flow at the entrance to the line. The air may thus be in solution
or i n free form i n bubbles or pockets. An air pocket implies a r e l a -
tively large volume of air, likely to accumulate on top of the pipe
cross section. The pockets may travel along the l i n e to peaks. There
they will either remain in equilibrium, be e n t r a i n e d by the f l o w i n g
water or be released through a i r valves.
Air i n solution i s not likely to present many engineering problems
It i s only when the pressure reduces s u f f i c i e n t l y to permit dissolved
air to form bubbles that problems arise. The water b u l k s a n d head
losses increase. The bubbles may coalesce and rise to the top of
the pipe to form large pockets. Flow conditions then become s i m i l a r
to those in partly full drain pipes, except that in a pipeline it

i s l i k e l y that the system, i n c l u d i n g the free a i r , w i l l be pressurized.


Air valves are frequently used to e l i m i n a t e the a i r which collects
on top of the pipe. Air valves of a slightly different design are
also used to release large quantities of air during the filling of
the l i n e a n d to draw air i n from the atmosphere d u r i n g vacuum con-
d i t i o n s i n the line.

PROBLEMS O F A I R ENTRAINMENT

Air drawn in through a pump supply can have a number of ef-


fects. Minute a i r bubbles o r air in solution can promote c a v i t a t i o n
- the formation of vacuous cavities which subsequently rapidly col-
lapse and erode the pump o r pipe. T h i s effect occurs i n pump impel-
lers in particular due to the high peripheral speed which lowers
pressures. Air drawn in in gulps can cause vibration by causing
flow to be unsteady.
98

Large quantities

intermittent1

- -__ -

a. Low level sump

I-
i
I * 1
+
Air e n t r a i n e d o ' To pump
by a falling
jet
Air enters intake

b. Free fall into sump

__1- ____
Surface dimple
barely
detectable

---+
-=-+
Well define
surface
dimple

-. 4

Air drawn
inter-
mittent ly Air core extends
from bottom into intake
of vortex (fully developed
into intake entraining vortex)

c . Vortex formation

FIG. 5.1 Inlet arrangements conducive to a i r entrainment.


99

Air i n a r i s i n g main, whether i n s o l u t i o n o r i n b u b b l e form, may


emerge from solution if ambient pressure or temperature i s reduced.
Air in free form w i l l collect a t the top of the p i p e l i n e and then r u n
u p to h i g h e r points. Here i t w i l l e i t h e r escape through a i r v a l v e s o r
vent pipes o r be washed along by the velocity of the water past the
air pocket in the pipe. The latter will result i n a head loss due to
the acceleration of the water past the a i r pocket. Even if air is in
b u b b l e form, dispersed i n the water, the f r i c t i o n head loss a l o n g the
pipeline w i l l increase. Other problems caused b y a i r present i n pipes
may include surging, corrosion, reduced pump e f f i c i e n c y , malfunction-
i n g o f v a l v e s or v i b r a t i o n s .
Air may be present in the form of pockets on the top of the
water, in bubbles, micro bubbles or in solution. Bubbles range in
size from 1 to 5 mm. Micro bubbles may be smaller, and t hei r r i s e
velocity is so slow (e.g. 1 mm bubble, 90 rnm/s, 0.1 mm bubble, 4
mm/s) that they stay in suspension for a considerable time before
rising to the surface. I n fact turbulence may create a n e q u i l i b r i u m
concentration p r o f i l e i n the same way that sediment s t a y s i n suspen-
sion i n f l o w i n g water. The r i s e velocity of small a i r bubbles i s

w = dgZ/18 v (5.1 1
where the kinematic viscosity v of water a t 20 C i s 1.1 x 10-6m2/s, d
i s b u b b l e diameter a n d g i s g r a v i t a t i o n a l acceleration.
For air bubbles to form readily a nucleus or uneven surface
should be present. Then when conditions a r e correct, air will rapidly
be released from solution. The capacity of water to dissolve air
depends on the temperature and pressure. At 20 C water may absorb
2% a i r by volume measured a t standard atmospheric pressure. This
f i g u r e v a r i e s from 3.2% at O°C down to 1.2% a t looo.

AIR INTAKE A T PUMP SUMPS

The major source of a i r i n pumping lines i s from the i n l e t sump o r


forebay. Here the water exposed to the a i r will absorb a i r a t a r a t e
depending on its temperature, pressure a n d degree of saturation of
the water.
100

a. Air pocket w i t h s u b c r i t i c a l flow past.

b. Air pocket w i t h super-critical flow past.

c. A i r pocket in e q u i l i b r i u m position.

F i g . 5 .2 Air pockets in pipelines.


101

Air in free form may also be dragged in to the conduit by the


flowing water. The formation of a vortex or drawdown outside the
suction p i p e entrance w i l l entice a i r i n t o the conduit. The h i g h e r the
entrance velocity and the greater the turbulence, the more air is
likely to be drawn in. This free air may later dissolve wholy or
partially when pressures increase beyond the pump. I n any case i t i s
carried along the conduit. As pressures again reduce, it may be
released from solution.
The configuration of the pump sump has an important b e a r i n g on
the tendency to draw in air. Extensive studies b y Denny a n d Young
(1957), Prosser (1977) and others have i n d i c a t e d how to design sumps
to minimize a i r intake. C i r c u l a t i o n i n the sump should be avoided b y
concentric approaches and straight inflow. The inlet should be
be1 lmouthed and preferably facing downwards upstream. Hoods ( s o l i d
or p e r f o r a t e d ) above the i n t a k e minimize a i r intake.
The degree of submergence should be as great as possible. Air
entraining drops . o r hydraulic jumps are to be avoided. Fig. 5.1
indicates some i n l e t arrangements to be avoided.

AIR ABSORPTION AT FREE SURFACES

The rate of diffusion of gas across a liquid interface can be


expressed i n the form

-dM_ - AK (Cs- C) (5.2)


dt

where M i s the mass r a t e of transfer per u n i t time t, A i s the area


exposed, K is a l i q u i d f i l m constant, C i s the gas concentration a n d
Cs is the concentration a t saturation. K is a f u n c t i o n of temperature,
v i s c o s i t y a n d turbulence. T h i s equation i s often w r i t t e n i n the form

(5.3)

where r i s the d e f i c i t r a t i o , Co i s the concentration a t time 0, and V


i s the volume of water per surface area A ( V / A = depth of water).
102

HYDRAULIC REMOVAL OF A I R

Air trapped in a p i p e and allowed to accumulate, will gradually


increase the air pocket size. The water cross sectional area will
therefore d i m i n i s h a n d the velocity of the water w i l l increase u n t i l at
some stage, some o r a l l of the a i r w i l l be dragged along the l i n e b y
the water. Alternatively a hydraulic jump may form in the pipe
entraining air and carrying i t away i n b u b b l e form, as depicted in
F i g . 5.2
The relationship between a i r pocket volume, washout velocity and
pipe diameter has been investigated by a number of workers, and
these r e s u l t s a r e summarized by Wisner et a l (1975).
Kal inske a n d B l i s s (1943) produced the f o l l o w i n g equation f o r r a t e
of water flow Q at which removal commences:

Q2
-
5 = 0.707 tan e (5.4)
gD

where D i s the p i p e diameter and g i s g r a v i a ional acceleration and


e i s the p i p e slope angle. The term on the l e f t h a n d side of K a l i n s k e ' s
equation is s i m i l a r to the parameter in the specific momentum a n d
energy diagrams (Figs. 5.3 and 5.4). In fact it points to the
possibility that the air is removed if the water depth drops below
critical. Then a h y d r a u l i c jump forms downstream which could e n t r a i n
air.
Wisner et al (1975) produced d a t a from which the r i s e velocity of
air pockets may be deduced. They i n d i c a t e r i s e velocity i s practically
independent of slope and is a function of Reynolds number vD/v and
r e l a t i v e volume of air pocket 4U/71D3 where U i s the a i r pocket volume
(see F i g . 5.5)

Hydraulic Jumps
A hydraulic jump draws in air in the form of bubbles. These
bubbles e x h i b i t a s u r p r i s i n g l y low tendency to coalesce a n d remain i n
free form f o r a long time. A i r may be absorbed from the bubbles by
the water, b u t t h i s takes a long time a n d many of the bubbles r i s e to
the surface before they a r e absorbed.
103
104

1
>

1o2 = VD/v

Fig. 5.5 E q u i l i b r i u m velocity of water f o r air pockets in a pipe


at l o 0 to 70° to the h o r i z o n t a l .

Kalinske and Robertson (1943) found from model experiments that the
rate of air entrainment at a hydraulic jump was given by the
equation

Qd/Q = 0.0066(F1 - 1)le4 (5.5)

where Qd is the volumetric r a t e of air entrainment, Q i s the water


I

flow rate and F1 is the upstream Froude number vl//gyl. That


r e l a t i o n s h i p was d e r i v e d f o r a jump i n a r e c t a n g u l a r channel w i t h a
free surface downstream. Experiments by the author indicate a

considerably greater air entrainment rate i n closed pipes (see Fig.


5.6)

Free F a l l s

A jet falling free into a pool of water has a s i m i l a r air intake


105

effect to a hydraulic jump. Oxygen i n t a k e at the base of free f a l l s


was studied by Avery a n d Novak (1978). For an initial oxygen defi-
ciency of 50%, they found a e r a t i o n at a r a t e of up to 1.6 k g 0 /kWh
2
for low head losses, decreasing for higher head losses, e.g. at 1
m head loss the aeration efficiency was only 1 k g 02/kWh. The kW
term represents the energy expended or lost in the jump or fall.
Assuming 21% oxygen in air, this would indicate aeration rates for
a i r u p to 8 kg/kWh.
Water flowing into inlets with a free fall e.g. morning glory
type s p i l l w a y s , o r even siphon s p i l l w a y s w i l l a l s o e n t r a i n a consider-

able amount of air (Ervine, 1976). Gravity pipelines are therefore


as much of a problem as pumping lines.

0 2 4 6 F: ~
v
J3A/B
Fig. 5.6 A i r removal a t h y d r a u l i c jumps in c i r c u l a r conduits.

A I R VALVES

Air accumulating in a pocket in a pipe may be released b y air


valves. These are normally of the 'small orifice' type (typical ori-
fices are up to 3 mm i n diameter). The v a l v e opens when sufficient
air is accumulated in a chamber to permit a ball to be released
from a seat around the o r i f i c e .
On the other hand a i r discharge d u r i n g f i l l i n g operations (before
106

This sectio parallel


t o hydraulir gradient
constitutes1 peak

Section Of pipeline running parallel t o hydraulic Section of pipeline forming peak with respect t o
gradient and constitutes peak horizontal and 0150 t o hydraulic gradient and
peok with respect t o hydraulic gradient only

No peok

Horizontal Datum
Section Of Pipeline having d w n w a r d grade Section of pipeline having upward grade and
and point of increaseof downward grade point of decrease of upward grade

Scour
Horizontal Datum
Long descending section of pipeline Long ascending section of pipeline

FIG. 5.7 P o s i t ~ o no f a i r valves along pipelines


107

pressurization) and air i n t a k e d u r i n g vacuum conditions in the p i p e


may be through 'large orifice' a i r valves, the b a l l of which i s re-
leased from the seat around the orifice when pressure inside the
pipe is low. Fig. 10.6 illustrates a double air valve with both a
small and a l a r g e o r i f i c e .
The size a n d spacing of a i r valves w i l l depend on ambient pres-
sures outside the pipe and permissible pressures inside, the size
of p i p e a n d water flow rates.
The equations for the discharge r a t e through an orifice involve
the c o m p r e s s i b i l i t y of air and are as follows (Marks, 1951). If the
pressure beyond the o r i f i c e p 2 i s g r e a t e r than 0 . 5 3 ~ ~( a l l pressures
absolute i.e. gauge p l u s atmosphere a n d p, i s i n i t i a l p r e s s u r e ) , then

(5.6)

For p 2 ' 0 . 5 3 ~ ~ then the flow becomes c r i t i c a l a n d flow r a t e i s inde-


pendent of p2. Then

(5.7)

where W i s the mass r a t e of flow of a i r , C i s a discharge coefficient,


a is the o r i f i c e area, rn is the mass density of air and k is the
a d i a b a t i c constant.
Normally whether the air valve is for letting air into the pipe
or out of it, p2 is less than 0 . 5 3 ~ ~ Then
. substituting k = 1.405
for a i r a n d C = 0.5 i n t o the last equation, i t s i m p l i f i e s to

Qa = 0.34 a /- (5.8)

where Qa is tne volumetric r a t e of flow of air ( i n c u b i c metres p e r


second if SI units are employed) measured at the initial pressure,
a i s the area of the o r i f i c e , h i s the i n i t i a l absolute head i n metres
of water a n d S is the relative density of air at initial pressure.
a
Since a i r density i s proportional to the absolute head, t h i s simp i f i e s
f u r t h e r to

Qa = lOOa ( 5 9)

where Qa is the air volume flow r a t e at initial pressure (+ O m

absolute) in m 3 / s a n d a i s in m'.
Thus t o r e l e a s e 1% a i r f r o m a p i p e f l o w i n g a t a v e l o c i t y of 1 m/s,
a n d under low head (10 m absolute being the minimum f o r no vacuum)
t h e n (5.8) r e d u c e s to
~
~ . _ _ _
0.01 x 1 m/s x A = 0 . 3 4 a J 9 . 8 x 10/1.15 x lo-’
i.e. d = 0.006D (5.10)

thus the o r i f i c e diameter should be about 1% o f the p i p e diameter to


r e l e a s e 1% a i r b y v o l u m e i n t h e p i p e .
The theory of flow through an orifice was used by Parmakian
(1950) to derive equations for the flow through air valves to f i l l a
cavity formed b y p a r t i n g water columns. If initial head (absolute) in
metres of water is hl, h2 is the head beyond the orifice, g is
gravitational acceleration, C is a discharge coefficient f o r the valve,
d is the orifice diameter, D i s the p i p e diameter, V i s the r e l a t i v e
velocity of the water column each s i d e of the air valve, k is the
a d i a b a t i c constant for a i r and S i s t h e r e l a t i v e d e n s i t y of a i r , then
a
f o r h2 > 0 . 5 3 hl,

(5.11)

a n d f o r H <C).53hl
2-

(5.12)

For air at 300 m above sea level and air temperatures of 24OC
3
p l u s 20% h u m i d i t y t h e n Sa i s 1.15 x 10- a n d atmospheric pressure i s
equivalent t o hl = 9.97 m of water. k i s 1.405. For a i r valves C is
a b o u t 0.5.

HEAD LOSSES I N P I P E L I N E S

Air suspended in b u b b l e form o r in p o c k e t s i n f l o w i n g water will


increase t h e s p e c i f i c volume. T h e mean v e l o c i t y i s consequently h i g h e r
109

to convey a c e r t a i n volume of water per u n i t time. The head loss w i l l


increase i n accordance w i t h an equation such as the Darcy- Weisbach
equation

(5.13)

The velocity v to use w i l l be v = ( 1 + f)v


W
(5.14)
where v i s the velocity of p u r e water f l o w i n g and f i s the volumet-
W
r i c concentration o f free a i r i n the pipe.
The head loss around a s t a t i o n a r y a i r pocket i s due p r i m a r i l y to
a loss of velocity head. If the size of air pocket is such that a
h y d r a u l i c jump forms, the head loss may be e v a l u a t e d from F i g . 5.4.
I f the velocity i s subcritical throughout, i t may be assumed t h a t the
difference in velocity head is lost. It could be established more
closely u s i n g momentum p r i n c i p l e s though.

WATER HAMMER

The presence of free air in pipelines can reduce the s e v e r i t y of


water hammer considerably. Fox (1977) indicates that the celerity
(speed) of an e l a s t i c wave w i t h free a i r i s

c =
/* (5.15)

For l a r g e a i r contents t h i s reduces to c =- (5.16)

where p i s the mass density of water, K i s i t s b u l k modulus, D i s the


pipe diameter, b i t s thickness, E i t s modulus of e l a s t i c i t y , p i s the
absolute pressure and f i s the free gas f r a c t i o n b y volume.
c i s reduced r e m a r k a b l y f o r even r e l a t i v e l y low gas contents. Thus
2% o f air at a pressure head of 50 m of water reduces the c e l e r i t y
from about 1100 m/s f o r a t y p i c a l p i p e l i n e to 160 m/s.
The Joukowsky water hammer head i s

A h = SAv (5.17)
9

where Av is the change in velocity of flow. There i s thus a large


reduction in Ah. I f the a i r collects a t the top of the p i p e there i s no
110

reason to see why the same equation cannot a p p l y . Stephenson (1967)


on the other h a n d d e r i v e d a n equation f o r the c e l e r i t y of a bore i n a
partly full pipe. The c e l e r i t y d e r i v e d from momentum p r i n c i p l e s i s for
smal I a i r proportions

c = J gAh/f (5.18)

where Ah i s the head r i s e b e h i n d the bore. This indicates a celerity


of 158 m/s f o r f = 0.02 a n d Ah = 50m.
There i s a school of thought which f a v o u r s the i n s t a l l a t i o n of a i r
valves in pipelines as a means of reducing water hammer over-
pressures. The intention is primarily to cushion the impact of ap-
p r o a c h i n g columns. Calculations wi II , however, indicate that an ex-
cessively large volume of air i s r e q u i r e d to produce any significant
reduction in head. The idea stems from the use of air vessels to
a l l e v i a t e water hammer in pipelines. I t will be r e a l i s e d that air in
air vessels i s under h i g h pressure i n i t i a l l y a n d therefore occupies a
relatively smal I volume. Upon pressure reduction following a pump
trip, the a i r from an a i r vessel expands according to the equation

puk = constant (5.19)

where U i s the volume of air. The size of air v a l v e s to draw i n the


necessary volume of air at low (vacuum) pressures w i l l be found on
a n a l y s i s to be excessive f o r l a r g e diameter pipelines.
An unusual problem due possibly to a i r coming out of solution in
a r i s i n g main was reported b y Glass (1980). Here a t h i n stream of a i r
along the top of the line supposedly collapsed on a pressure rise
after a pump trip. After a number of years the p i p e b u r s t along a
l i n e a l o n g the s o f f i t .

REFERENCES

Avery, S.T. and Novak, P, 1978. Oxygen t r a n s f e r at h y d r a u l i c struc-


tures. Proc. ASCE, HY11, 14190, pp 1521-1540.
Denny, D.F. a n d Young, G.A.J., 1957. The prevention o f vortices i n
intakes. Proc., 7th Con. I n t . Ass. Hydr. Res., Lisbon
E r v i n e , D.A., 1976. The entrainment of a i r i n water. Water Power and
Dam Construction. pp 27-30.
Fox, J . A . , 1977. H y d r a u l i c A n a l y s i s o f Unsteady Flow i n Pipe Net-
works. Macmillan, London.
Glass, W.L., 1980. C a v i t a t i o n of a pump p i p e l i n e . Proc. I n t . Conf.
Pressure surges. BHRA, Canterbury.
111

K a l i n s k e , A.A. and B l i s s , P.H., 1943. Removal o f a i r f r o m p i p e l i n e s


b y f l o w i n g w a t e r , C i v i l E n g i n e e r i n g , ASCE, 13, 10. p 480.
K a l i n s k e , A.A. and R o b e r t s o n , J.M., 1943. C l o s e d c o n d u i t f l o w , T r a n s .
ASCE. 108, 2205, pp 1453-1516.
M a r k s , L.S., 1951. M e c h a n i c a l E n g i n e e r s H a n d b o o k , 5 t h Ed. McGraw
H i l l , N.Y. 2235 pp.
P a r m a k i a n , J., 1950. A i r i n l e t v a l v e s f o r s t e e l p i p e l i n e s . T r a n s . ,
ASCE, 1 1 5 , 2404., pp 438-444.
P r o s s e r , M.J., 1977. T h e H y d r a u l i c D e s i g n of Pump Sumps and In-
t a k e s , BHRA, and C I R I A , L o n d o n . 48 pp.
Stephenson, 1967. P r e v e n t i o n o f v a p o u r p o c k e t s c o l l a p s e i n a p u m p i n g
l i n e . T r a n s . , S o u t h A f r i c a n I n s t . C i v i l E n g s . 9 , ( l o ) , pp 255-261.
W i s n e r , P . , Mohsen, F.M. and Kouwen, N., 1975. Removal o f a i r f r o m
w a t e r l i n e s b y h y d r a u l i c means. Proc., ASCE. 101, HY2, 11142, pp
243-257.

LIST OF SYMBOLS
a - orifice area
A - area
b - wal I thickness
c - wave celerity
c - d i s c h a r g e coef f i c i e n t
c - gas concent r a t ion
i n i t i a l concentration
co -
c o n c e n t r a t i o n at s a t u r a t i o n
cs -
d - b u b b l e diameter
D - p i p e internal diameter
E - modulus of e l a s t i c i t y
f - volumetric concentration of a i r
F - Froude number

9 - g r a v i t a t i o n a l acceleration
h - head
k - a d i a b a t i c constant
k - I i q u i d f i Im c o n s t a n t
K - bulk m o d u l u s
m - mass d e n s i t y o f a i r
M - m a s s r a t e o f t r a n s f e r o r s p e c i f i c momentum

P - pressure
Q - water discharge r a t e
volumetric r a t e of air entrainment
‘d -
r - gas d e f i c i t ratio
112

s - r e l a t i v e density o r s p e c i f i c g r a v i t y
t - time
u - a i r pocket volume
v - water v e l o c i t y
v - volume of water p e r surface area
w - r i s e velocity o f bubbles
w - mass r a t e of flow of a i r

Y - depth of flow
x - Darcy f r i c t i o n factor
e - p i p e slope a n g l e
v - k i n e m a t i c viscosity
P - mass density of water
113

CHAPTER 6

EXTERNAL LOADS

Low pressure pipes, especially sewers, gravity mains or even


large diameter pumping mains should be designed f o r e x t e r n a l loads
as well as internal loads. The v e r t i c a l soil load a c t i n g in combina-
tion with vacuum pressure inside the p i p e could cause the pipe to
collapse unless the p i p e i s adequately supported o r stiffened.

S O I L LOADS

The load transmitted to a pipe from the external surroundings


depends on a number of factors:
Rigidity of pipe: The more r i g i d a pipe is relative to the trench
sidefill the more load i t w i l l take. The s i d e f i l l tends to settle, thus
causing a l a r g e p a r t of the b a c k f i l l to rest on the pipe. T h i s occurs
w i t h f l e x i b l e p i p e too to some extent, as a p i p e i s supported l a t e r a l l y
b y the f i l l a n d w i l l not y i e l d as much as a free s t a n d i n g pipe.
Type of trench o r f i l l : Fig. 6.1 i l l u s t r a t e s v a r i o u s possible i n s t a l -
lation conditions for pipes. The load transmitted to the p i p e v a r i e s
w i t h the w i d t h a n d depth of trench since f r i c t i o n on the sides of the
trench affects the r e s u l t a n t load. Embankment f i l l s may also t r a n s m i t
different loads to a pipe, depending on the relative settlement of
s i d e f i l l and topfi I I .
Young and Trott (1984), a n d C l a r k e (1968) have developed exten-
sive equations and charts for evaluating soil loads on pipes in
various trench and embankment conditions. Although many of their
assumptions a r e subject to question, the f a c t remains t h a t there i s as
yet no other theory w i t h which to c a l c u l a t e s o i l loads on pipes so the
engineer must use t h i s theory w i t h d i s c r e t i o n (ref. Spangler, 1956).

Trench Conditions
The s o i l load transmitted to a r i g i d pipe in a trench depends on
t h e w i d t h a n d depth of trench a n d the s o i l b a c k f i l l properties.
For a normal vertical-sided trench (Fig. 6.la), fill at the sides of
114

the p i p e w i l l settle more than the pipe, a n d the s i d e f i l l support can


be neglected. On the other h a n d the f r i c t i o n of the b a c k f i l l a g a i n s t
the sides of the trench takes some of the load. The cohesion between
trench fill and the sides of the trench is neglected. The friction
developed is therefore proportional to the coefficient of friction
between the f i l l a n d the sides of the trench, a n d the r a t i o K of the

active horizontal pressure to the vertical pressure in the soil.


Equating the vertical u p w a r d forces on any horizontal slice in the
trench to the downward forces, Marston e v a l u a t e d the total load on a
p i p e at depth H ( 1 9 1 3 ) :

Upward load per unit length of Pipe a t depth h + ah below ground


s u r f ace :

W + dW = W + yBdh - 2K t a d Wdh/B

Solution g i v e s

W = CdyBZ

-2K(tane )H/B
where t h e load coefficient C =
d 2Ktan6

K = r a t i o of l a t e r a l s o i l pressure to v e r t i c a l load.
-
- 1 - s i n 9 f o r a c t i v e s o i l conditions a n d cohesionless soil
1 + sin 0
@ = a n g l e of i n t e r n a l f r i c t i o n of b a c k f i l l
e = a n g l e of f r i c t i o n between b a c k f i l l a n d sides of trench
H = h e i g h t of f i l l above p i p e
B = trench w i d t h
Y = u n i t w e i g h t of b a c k f i l l m a t e r i a l

Ktane n o r m a l l y ranges from 0.11 for soft c l a y s to 0.16 for sands and
coarse crushed stone. Cd i s given i n Fig. 6.2 for v a r i o u s values for
Ktan8. Note that f o r deep trenches, (H/B greater than approximately
l o ) , Cd approaches a l i m i t i n g v a l u e of 1/(2 K t a n e ) . T h i s implies that
the s i d e f r i c t i o n i n the trench takes more of the load the deeper t h e
trench. For very shallow or wide trenches (H/B < l ) , Cd is
approximately equal to H/B, so

W = yHB (6.2)

i.e. most of the b a c k f i l l load i s taken b y the pipe. The trench equa-
115

ALTERNATIVE SIDE
/-FORM

b : WIDE TRENCH c W I D E TRENCH


o S R R O W TRENCH
SHALLOW EMBANKMENT _-
D E E P EMBANKMENT

dSHALLOW EMBANKMELT c : DEEP EMBANKMENT f : EMBANKMENT


PLSlTlVE PROJECTION POSITIVE
-__ PROJECTION NEUTRAL PROJECTION

TA2r
- - - - - -- - - - --- -- NORMAL DENSITY FILL
__-_.__
---/
----

g SHALLOW E M B A N K M E N T h: DEEP EMBANKMENT i EMBANKMENT


N E GAT1VE PROJECT1 ON N E GAT IVE P R O JE CT ION INDUCED _
T R-E N C H

1 TUNNEL OR H E A D I N G
OR THRUST BORE

Fig. 6.1 A l t e r n a t i v e backf i I I s


116

tion indicates that the load on a r i g i d p i p e increases i n d e f i n i t e l y as


the trench width increases. This i s not the case and a t some trench
width the equation no longer a p p l i e s a n d embankment conditions (see
next section) then apply (see F i g . 6.lb and 6 . 1 ~ ) . I t i s necessary to
evaluate the load trying both the trench criterion (using Fig. 6.2)
and the embankment criterion (using Fig. 6 . 3 ) , and to select that
load which i s least.
In the case of a 'V' trench, the trench width to use i s that at
the crown of the pipe. I n t h i s case @ instead of 0 i n the formula f o r
Cd i s used since the shear plane is i n the f i l l not against the sides
of the trench. If the side-fill is well compacted and the pipe is
relatively flexible then the load on the pipe depends on the pipe
diameter a n d not to the same extent on the trench w i d t h :
W = C YBD (flexible pipe) (6.3)
d
Note that f o r small H, C d = H/B and i n t h i s case

W = YHD

Embankment Conditions
A p i p e bedded on a f i r m surface w i t h a wide embankment f i l l over
it is said to be under embankment conditions. If the crown of the
pipe projects above the original ground level it is a positive
projection (Figs. 6.ld) and 6.le). It is referred to as a complete
positive projection if the pipe is rigid and lies completely above
natural ground level, and the embankment is relatively shallow. If
the p i p e crown i s below the n a t u r a l ground level one has a negative
projection or trench condition (Figs. 6.lg and 6 . l h ) . For very deep
trenches u n d e r embankments t h e conditions may be the same as f o r the
trench condition ( r e f e r r e d to here as the complete trench c o n d i t i o n ) .
If the crown is level with the n a t u r a l ground level it i s a neutral
projection (Fig. 6.lf).
For the complete positive projection case the f i l l beside the p i p e
tends to s e t t l e more than directly above the pipe. The s i d e f i l l tends
to drag downwards on the fill directly above the p i p e and increase
the load on the pipe. The r e s u l t i n g load per u n i t l e n g t h of p i p e i s

W = CcyD2
117

where the load coefficient

2K tan@.H / D
e -1
cc = 2Ktan@

LOAD COEFICIENT C,

F i g . 6.2 Load coefficient C d f o r trench conditions.

The l i n e l a b e l l e d complete p r o j e c t i o n condition in F i g . 6.3 g i v e s the


v a l u e of C for Ktan@ = 0.19, which i s the minimum adverse f r i c t i o n
condition . For deep embankments, only a c e r t a i n h e i g h t of f i l l above
the p i p e w i l l s e t t l e at a d i f f e r e n t r a t e to the s i d e f i l l . This height,
depends on the product of the pipe projection ratio u and the
He,
settlement r a t i o s where

projection of p i p e crown above o r i g i n a l ground level


u = diameter of p i p e
118

s = settlement of adjacent f i l l o r i g i n a l l y at crown level-


settlement of p i p e crown
reduction i n thickness of adjacent f i l l below crown level

The settlement of adjacent fill includes the settlement of the o r i g i n a l


ground level a n d the compaction of the s i d e f i l l . The settlement of the
p i p e crown includes settlement of the bottom of the p i p e and v e r t i c a l
deflection of the pipe.
The v a l u e of s n o r m a l l y v a r i e s from 0.3 for soft y i e l d i n g founda-
tion soil to 1.0 for rock or unyielding soil foundation. A common
value i s 0.7 b u t 0.5 i s recommended i f the foundation a n d s i d e f i l l a r e
well compacted. The product su is used i n F i g . 6.3 to e v a l u a t e the
resulting load coefficient for v a r i o u s embankment conditions. I t i s not
necessary to e v a l u a t e H .
The equation for C for deep embankments a n d incomplete p o s i t i v e
projection conditions i s more compl icated than f o r the shal low embank-
ment condition, but it was e v a l u a t e d by Spangler f o r Ktan4 = 0.19
(i.e. minimum adverse friction) condition, and these values are
g i v e n i n F i g . 6.3.
Fig. 6.3 also gives values of C for the case when the top of
the pipe settles more than the adjacent backfill (i.e. negative 5).

Spangler evaluated C for Ktanb = 0.13 (i.e. minimum favourable


friction) for this case, and the values are indicated in Fig. 6.3,
label led complete and incomplete trench conditions since the mech-
anics a r e s i m i l a r to those f o r trench conditions.
For the negative projection (or incomplete trench) condition (see
Fig. 6.19 and 6 . l h ) w i t h the p i p e i n a trench under an embankment,

w = CnYB2 (6.6)

where Cn i s the same as C i n Fig. 6.3 f o r negative su b u t with D


replaced b y 6 i n a l l expressions. I n t h i s case the projection r a t i o i s
depth of p i p e crown below n a t u r a l ground level
u =
w i d t h of trench

a n d the settlement ratio i s

settlement of n a t u r a l ground level-settlement of level of top


s = f i l l i n trench
reduction in thickness of s o i l in trench above crown
119

F i g . 6.3 Load coefficient C f o r embankment conditions


(Spangler, 1956).

The settlement of the level of the top of the f i l l i n the trench may
i n c l u d e the settlement of the bottom of the pipe, deflection o f the p i p e
a n d compaction of the f i l l i n the trench above the crown of the pipe.
s may r a n g e from - 0.1 f o r u = 0.5, to - 1.0 f o r u = 2.0.
It i s possible to reduce the load on pipes under a n embankment b y
removing a certain amount of compacted fill d i r e c t l y above the p i p e
and replacing it by lightly compacted fill directly, i.e. inducing a
negative projection condition (Fig. 6.li). For this condition s may
r a n g e from -0.5 f o r u = 0.5 to -2.0 f o r u = 2.0.
The load per unit l e n g t h on a completely f l e x i b l e p i p e under
an embankment whether i n p o s i t i v e o r negative projection condition is

W = YHD (6.7)

The load on a r i g i d pipe i n a trench up to i t s crown level (neutral


p r o j e c t i o n c o n d i t i o n ) i s a l s o yHD.
120

C l a r k e suggests that the l o a d on a tunnel or thrust bore i s simi-

lar to that on a pipe in a trench, w th a reduction factor allowing


for t h e cohesion of t h e m a t e r i a l above. The a s s u m p t i o n s i n a r r i v i n g a t
this conclusion are not convincing and further theory would be
we1 come.
Clarke also suggests that uniform surcharges of large extent be
t r e a t e d a s embankments a n d the equivalen t depth determined. This is
preferable to using pure elastic theory, as some load transfer must
take place between soil fills at different densities and soil masses
which settle differently.
The soil density which would result in the maximum load on a
pipe is the saturated (but not submerged) density. The submerged
conditions are normally less severe than for the saturated case, as
the water pressure would produce a uniform radial pressure less
l i k e l y to c r a c k a p i p e t h a n a p u r e v e r t i c a l load.

Example - N e g a t i v e p r o j e c t i o n case
The top of a pipe in a 2 m wide trench is 1 m below natural
ground level and there is a 7 m high embankment above the
trench. E v a l u a t e the load on the p i p e .
1
Projection r a t i o u = - = 0.5
2
Settlement r a t i o s = -0.4 s a y .
H/B = 8 / 2 = 4 .
From F i g . 6.3, load coefficient C = 3.1

L o a d p e r m o f p i p e = CnyB2 = 3.1 x 18000 x 2*/1000 = 220KN/m

SUPER IMPOSED LOADS

It is customary to use elastic theory to evaluate the pressures


transmitted by surface loads to a buried pipe and to assume a semi-
infinite, homogeneous, isotropic, elastic material surrounds the p i p e .
The f a c t that t h e s i d e f i l l may s e t t l e d i f f e r e n t l y to t h e p i p e i s i g n o r e d
in the theory and suitable factors s h o u l d be a p p l i e d to decrease the
transmitted load if the p i p e is flexible. U n f o r t u n a t e l y no such f a c t o r s
a r e a v a i l a b l e yet.
If t h e l o a d a p p l i e d o n t h e s u r f a c e i s of l i m i t e d l a t e r a l e x t e n t then
the induced pressure on any horizontal plane below the surface
121

decreases w i t h depth. The deeper the p i p e , the w i d e r the spread and


consequently the lower the maximum pressure. On the other h a n d the
total v e r t i c a l force must remain equal to the superimposed load.
Formulde for evdluating the transmitted loads due to a surface
point load, or a pressure of limited or i n f i n i t e extent, are given in
the sections below. A suitable impact f a c t o r should a l s o be a p p l i e d .
The impact factor will vary with the type of load a n d depth of p i p e
below the surface, and varies from 1.3 for slow moving vehicles to
2.0 f o r fast moving vehicles i f the p i p e i s shallow. The impact factor
reduces f o r cover depth greater than the p i p e diameter and isolated
point loads, but the amount of reduction i s debatable and i t may be
safe to adopt the full value of the impact factor until conclusive

f i g u r e s a r e a v a i l a b l e (see also Compstan et a1 1978).

T r a f f i c Loads
The following loads are useful for design purposes (see a l s o
Concrete Pipe Assn. 1969).
(1) I n fields : two wheel loads each 30 kN spaced 0.9 m a p a r t .
(2) Under light roads : two wheel loads each 73 kN spaced 0.9m
apart.
(3) Under main roads and heavy traffic (BS 153 type HB load) :
eight wheel loads each 112 kN comprising two rows 1.8 m a p a r t
of four wheels each 0.9 m a p a r t l a t e r a l l y , measured from i n n e r
a x l e to inner axle.
(4) Airport runways: four 210 k N wheel loads spaced at 1.67m x
0.66 m.

Pipelines would n o r m a l l y be designed f o r f i e l d l o a d i n g a n d strength-


ened under h e a v i e r t r a f f i c . I f construction loads a r e l i k e l y to exceed
design loads e x t r a t o p f i l l should b e placed to spread the load d u r i n g
construct ion (Kennedy, 1971 ) .
Stress Caused by Point Loads
Boussinesque's e l a s t i c theory g i v e s the v e r t i c a l stress a t depth H
below p o i n t load a n d a distance X h o r i z o n t a l l y from the p o i n t load P
as
122

3P H 3
w = (6.8)
5/2
27r(HZ + X ' )

The stress at any point due to two o r more loads w i l l be the sum of
the stresses due to the individual loads. The maximum stress could
occur d i r e c t l y under e i t h e r load o r somewhere between them, and the
worst case should be selected b y p l o t t i n g the stress between them.
For large pipe diameter, the stress w i I I v a r y a p p r e c i a b l y across
the pipe diameter and some average value should be taken. The
Concrete Pipe Association (1969) l i s t s the total load i n kg/m of pipe
for surface wheel loads (l), ( 2 ) a n d (3) i n the p r e v i o u s section, and
v a r i o u s p i p e diameters and depths.

Line Loads
The v e r t i c a l stress at any depth H below a l i n e load of i n t e n s i t y
q per u n i t l e n g t h is, b y e l a s t i c theory,

2q~3
w =
n ( X Z + H 2) 2

Uniformly Loaded Areas


The stress intensity at any depth beneath a loaded area on the
surface could be e v a l u a t e d b y assuming the distributed load to com-
p r i s e a number of point loads, a n d summating the stress due to each
load. F o r t u n a t e l y Newmark has performed the i n t e g r a t i o n of i n f i n i tesi-
ma1 point loads to give the vertical stress under the corner of a
uniformly loaded rectangular area. Newmark's influence coefficients
are given i n Table 6.1. To e v a l u a t e the v e r t i c a l stress at any point
under a surface loaded area, d i v i d e the loaded area into rectangles
each with one corner directly above the p o i n t i n question. Calculate
the r a t i o s L / H a n d Y/H where L a n d Y a r e the l e n g t h and b r e a d t h of
the rectangle and read the corresponding influence factor from the
table. Evaluate the stress due to each rectangle of load by
multiplying the load intensity by the influence coefficient for the
particular rectangle, and summate the stress due to each such
rectangle. If the p o i n t i n question f a l l s outside the boundaries of the
loaded area assume the loaded area extends to above the p o i n t and
s u b t r a c t the stress due to the i m a g i n a r y extensions.
123

TADLE 6.1 Influence factors f o r v e r t i c a l pressure under the corner


of a uniformly loaded r e c t a n g u l a r area (Capper a n d
Cassie, 1969).

o 0222
00435

Similar influence coefficients a r e available for the stresses under


loaded circles and strips, but normally any shape can be resolved
into rectangles or small point loads without much e r r o r . There are
also influence charts available for evaluation of stresses under any
shape loaded area (Capper a n d Cassie, 1969).
The stress a t any depth under a uniform load of v e r y l a r g e extent
is the intensity of the surface load since there is no lateral
dispersion. T h i s could also be deduced from Table 6.1 for large L and
Y.

Effect of R i g i d Pavements
A rigid pavement above a p i p e has the effect of distributing the
load laterally and therefore reducing the stress intensity due to a
surface load. The stress at a depth H below a rigid pavement of
thickness h, is

W = C_P/R2 (6.10)

where R -
- = r a d i u s of stiffness

E = modulus of e l a s t i c i t y (28 000 N/mrnz or 4 000 000 p s i


f o r concrete)
h = pavement thickness
124

U = Poisson's r a t i o (0.15 f o r concrete)

k = modulus of subgrade reaction, which varies from


0.028N/mm3 for poor support to 0.14 N/mm'. A typical
v a i u e f o r good support i s 0.084 N/mm3
R is normally about three times the slab thickness. Cr is a
function of H/R and X/R and i s tabulated i n Table 6.2 for a
s i n g l e point load P on the surface.
A concrete slab normally has the effect of about five times its
thickness of s o i l f i l l i n a t t e n u a t i n g a p o i n t load.

TABLE 6.2 Pressure coefficients for a point load on a pavement


(Amer. Con. Pipe Assn., 1970)

W Pressure on p i p e = CrP/R2
H Depth of Top of Pipe Below Pavement
P Point Load
R Radius o f Stiffness
X Horizontal Distance from Point Load

__
H
- - -
R 00 08 1 2 20 24 28
___ -~~
(1 n l!3 ObX 032 020
03 101 (Ifil 03 1 U? i
08 089 061 03? 02:
12 076 054 032 017

16 062 047 010 022


?0 051 l!d2 026 o:?
24 043 036 llib 021
28 037 03 I 073 019

32 012 025 O? I OIR


36 027 013 019 016
40 02% 0XJ 018 015
41 020 018 015 014
48 018 Olh 013 012
52 015 014 012 01 I
56 014 01: cil? 010
60 012 Ol! on9 or19

63 011 ni; 00s 008


68 010 009 nux uni
12 009 ons 007 037
75 OOR 'I-; 00- [I!)&
gn 007 007 OOh 00'
~. -
125

REFERENCES

American Concrete Pipe Assn., 1970. Design Manual-Concrete Pipe,


Arlington.
Capper, P.L. a n d Cassie, W.E., 1969. The Mechanics of Engineering
Soils, 5th Ed., Spon., London.
Clarke, N.W.B., 1968. B u r i e d p i p e l i n e s , MacLaren, London.
Concrete Pipe Assn., 1969. Loads on B u r i e d Concrete Pipes, Tech.
B u l l . No. 2, Tonbridge.
Compston, D . G . , Cray, P., Schofield, A.N. a n d Shann, C.D., lc7C.
Design a n d Construction of B u r i e d T h i n Wall Pipes. CIRIA.
Kennedy, H., 1971. E x t e r n a l loads and foundations f o r pipes, J . Arc!.
Water Works Assn. March.
Marston, A . and Anderson, A . O . , 1913. Theory of loads on pipes i n
ditches and tests o f cement a n d c l a y d r a i n t i l e s a n d sewer p i p e .
Iowa State Univ., Engg. Res. I n s t . B u l l .
Spangler, M.G., 1956. Stresses i n pressure p i p e l i n e s a n d p r o t e c t i v e
c a s i n g pipes, Proc. Am. SOC. C i v i l Engrs., 82 ( S T 5 ) , pp 1054-1115.
Young, D.C. and Trott, J.J., 1984. B u r i e d R i g i d Pipes. Elsevier
Applied Science, London. p p 230.

L I S T O F SYMBOLS

trench w i d t h
load coefficient, embankment condj tion, positive projection

case
load coefficient, trench condition
load coefficient, embankment condition, negative projection
case
load coefficient, r i g i d pavemerot
d i ameter
effective modulus of e l a s t i c i t y of s o i l
modulus of e l a s t i c i t y
pavement thickness or v a r i a b l e depth
depth of cover
depth of f i l l below p l a n e of equal settlement
momen t of inert ia
modulus of subgrade reaction
modulus of subgrade reaction
r a t i o of l a t e r a l to v e r t i c a l s o i l pressure
l e n g t h of loaded area
bending moment
bending or deflection coefficient (subscripts t, b, s, x

and y refer to top, bottom and side moments a n d hori-


zon t al and v e r t i ca I d i stort ions respective I y )
126

pressure
p o i n t load
l i n e load per u n i t length
radius
r a d i u s of stiffness
settlement r a t i o
wal I thickness
project ion r a t i o
v e r t i c a l pressure
W l a t e r a l soil pressure
W permissible r i n g load
W permi ssi b l e v e r t i c a I load
w vertical load per u n i t length
horizon t a I distance
w i d t h of load area
a n g l e of bottom support
def I ect ion
u n i t weight of soil
a n g l e of friction between backfill and sides of
trench
a n g l e of f r i c t i o n of b a c k f i l l m a t e r i a l
Poisson's r a t i o
127

CHAPTER 7

CONCRETE PIPES

THE EFFECT OF BEDDING

Non-pressure sewer or drain pipes are designed to withstand


external loads, not internal pressures. Various standards specify
the design load per unit run of pipe for different classes and the
pipes are reinforced accordingly. The main stresses due to vertical
external loads are the compressive stress at the haunches and the
bending stresses at the crown, the bottom and the haunches. The
main stresses are caused by live loads, vertical and horizontal soil
loads, self weight and weight of water ( i n t e r n a l water pressure a n d
transient pressures a r e neglected f o r non-pressure pipes).
Concrete pipes a n d the other r i g i d non-pressure pipes a r e normal-
ly designed to withstand a vertical line load while supported on
a flat rigid bed. The load per unit length r e q u i r e d to f r a c t u r e the
pipe loaded thus is called the laboratory strength. Although the
laboratory strength could be calculated theoretically, a number of
practical factors influence the theoretical load and experimental
determination of the load is more reliable. (The tensile strength
of concrete is very uncertain and the effect of lateral constraint
of the supports may be a p p r e c i a b l e ) .
Alternative standard testing arrangements f o r pipes a r e i I lustrat-
ed in Fig. 7.1. The strength is defined in B r i t i s h Standard 556 as
that load which will produce a c r a c k 1/100 inch (0.25 mm) f o r unre-
inforced concrete pipes or 60% of the ultimate load for reinforced
pipes, or 90% of the 1/10000 inch (0.025 mm) wide crack load for
prestressed c y l inder type concrete pipes.
The strength u s i n g the 3-edge bearing test is usually 5 to 10%
in excess of that for the 2-edge bearing test. Recently the 3-edge
b e a r i n g test was standardized and t h i s i s now defined as the labora-
t o r y strength.
Pipes laid in trenches are usually supported over a relatively
wide length of arc by the bedding. The strength is consequently
higher than the laboratory strength as the bending moments are
128

less. F i g . 8.2 summarizes the theoretical b e n d i n g moment c o e f f i c i e n t s


for pipes supported over various angles (defined i n Fig. 8 . 4 ) (CPA, 1962).

0 mm

15&brmm
rubber

2 -EDGE 3-EDGE 3-EDGE


FOR ANY OIA UP TO l8OOrnrn

Fig. 7.1 S t a n d a r d c r u s h i n g test b e a r i n g s f o r r i g i d p i p e s

The ratio of field strength to laboratory strength is defined as


the bedding factor. Various types of bedding and the corresponding
b e d d i n g f a c t o r s a r e l i s t e d b e l o w (CPA, 1967; ACPA, 1970):-

TABLE 7.1 Bedding factors

Class A 120" R.C. c o n c r e t e c r a d l e o r a r c h 3.4


Class A 120° p l a i n c o n c r e t e c r a d l e o r a r c h 2.6
Class B Granu Ia r bedding 1.9
Class C H a n d s h a p e d t r e n c h bottom 1.5
Class D H a n d trimmed f l a t bottom t r e n c h 1.1

For rigid concrete pipes the lateral support of the sidefill in


the t r e n c h does n o t a d d n o t i c e a b l y to t h e s t r e n g t h .
A factor of safety K of 1.25 to 1.5 i s normally used w i t h u n r e i n -
forced concrete pipe. When p i p e i s p r e s s u r i z e d , i t s h o u l d be designed
so t h a t : -

external load internal pressure < -


1
f i e l d strength
+ bursting pressure (7.1
K

Unreinforced or reinforced (not prestressed) concrete pipes are


129

normally only used for non-pressure pipes such as sewers and


drains. The concrete may be stressed in tension if the pipe was
subjected to internal pressures, and even though concrete has a
certain tensile strength, c r a c k s may develop a n d leaks a r e likely if
the tensile stress i s maintained (Kennison, 1950).
Concrete pipes normally do not require lining or wrapping.
Special precautions may be necessary for certain liquids, for in-
stance sewers a r e often made w i t h limestone aggregate to m a i n t a i n the
r a t e of corrosion of the aggregate a t the same r a t e as the cement,
thus m a i n t a i n i n g an even surface. The f r i c t i o n coefficients of concrete

pipes, especially when c e n t r i f u g a l l y cast, can compare favourably


w i t h those of l i n e d steel pipes.

PRESTRESSED CONCRETE PIPES

Prestressed concrete is becoming a popular medium for large-bore


pressure pipes. Prestressed concrete competes economically with steel
for long pipelines over approximately 800 mm diameter. It has the
advantage that the prestressing steel can be stressed to h i g h e r stres-
ses than for plain-walled pipes. The wall thickness o f p l a i n walled
steel pipes must be reasonably thick to prevent buckling, collapse
and d i s t o r t i o n even if the thickness i s not r e q u i r e d to r e s i s t i n t e r n a l
pressures. Consequently the use of high-tensi le steels is restricted
when manufacturing plain-walled steel pipes, b u t t h i s i s not the case
w i t h prestressed concrete pipes which a r e more r i g i d than steel pipes.
Prestressed concrete pipes are formed by winding pretensioned
wires onto a core. The helical winding is subsequently coated. The
core forming the b a r r e l of each p i p e may be concrete cast vertically
i n a mould o r c e n t r i f u g a l l y i n a h o r i z o n t a l position. The concrete is
then steam cured to ensure rapid strength increase before winding.
Alternatively the core may be formed w i t h a steel c y l i n d e r lined with
mortar. The steel cylinder improves watertightness, distributes the
concentrated load under the wire, and acts as longitudinal rein-
forcing. Plain concrete cores are often reinforced with bars or pre-
stressed longitudinally as well as c i r c u m f e r e n t i a l l y . The longitudinal
reinforcing resists l o n g i t u d i n a l bending i n the trench as well as local
130

bending stress i n the core w a l l s d u r i n g c i r c u m f e r e n t i a l winding. The


longitudinal wires are pretensioned and released once the core has
set.

Fig. 7.2 Prestressed concrete p i p e

The prestressing wire i s wound h e l i c a l l y onto the core a t a stress


close to y i e l d p o i n t . It loses a proportion of i t s stress due to creep
and s h r i n k a g e before the p i p e i s put into service. A concrete c o a t i n g
i s a p p l i e d as soon as possible a f t e r w i n d i n g to get as much stress as
possible t r a n s f e r r e d to the c o a t i n g d u r i n g the s h r i n k a g e of the core.
The tensile strength of the core is neglected in accounting for
f i e l d pressures b u t i f t r a n s i e n t stresses i n the f i e l d o r i n the process
of manufacture do cause cracking, the cracks soon seal due to the
inherent h e a l i n g p r o p e r t i e s of concrete. A h y d r o s t a t i c test is usually
done on the pipe immediately after winding but before c o a t i n g and
some seepage of water through the core may be permitted at this
stage. Further load tests may be done immediately before dispatch of
t h e pipes from the works, and i n the field.
The pipes a r e u s u a l l y joined with spigots a n d sockets sealed w i t h
rubber insertions. The socket may be of steel o r concrete b u t care i s
needed in the design of integral concrete sockets to ensure no
stressing or spalling of the socket will occur. Specials (bends and
tees) may be f a b r i c a t e d from steel o r cast iron, and are f i t t ed with
matching spigots and sockets. Small angle bends are n o r m a l l y made
up over a number of joints which may take up to 2" deflection,
depending on the type of j o i n t .

C i rcurnferent i a l Prestressing
The tensile stress in the prestressing wires immediately a f t e r the
wire is wound onto the core i s less than the a p p l i e d tensile stress.
This i s because the core deforms under the p r e s t r e s s i n g force. At the
instant of prestressing any section, the pipe to one side of the
section i s under stress whereas the other side i s free. The s t r a i n in
the core at the point of winding is half its final strain. The
relaxation of the steel stress a f t e r winding wi II therefore correspond
to h a l f the t o t a l s t r a i n of the concrete core and

Afs = i f n
co 1
where n, i s the r a t i o of the modulus o f elasticity of steel to that of

the concrete a t the time, fco i s t h e core stress a f t e r w i n d i n g and f

i s the steel stress, hence

fso = f
SI
. - i f n
co 1

where f .
51
i s the initial steel stress and f
so i s the steel stress a f t e r
winding. Since f o r e q u i l i b r i u m of forces

f
so
s = f t
co c

(7.2)

(7.3)
and fSo = fcotc'S
where S i s the steel cross sectional area per u n i t length o f p i p e , a n d
tc i s the core thickness.

Circumferential Prestress a f t e r Losses


After winding the core, the p i p e is usually subjected to a hydro-
static test to detect cracks. Any creep which the prestressing steel
may undergo has usually taken place by this time, and a certain
amount of concrete creep has occurred. The concrete stress j u s t before
the test becomes

f . = (7.4)
CI fco
132

where u is the combined steel a n d concrete creep coefficient between


winding and testing. The corresponding steel stress i s :

fsi = fCitC/S (7.5)

The stress for any particular test loading may be calculated using
Equs. 7.9 - 7.15 b u t r e p l a c i n g subscripts 2 b y 1 a n d s e t t i n g the core
thickness tb equal to zero. The c o a t i n g should be a p p l i e d and cured
as q u i c k l y as possible a f t e r testing. This ensures t h a t the s h r i n k a g e
and creep of the concrete core transfer as much compression to the
c o a t i n g as possible.
The c o a t i n g adds to the cross sectional area a n d reduces stresses
thereby a l s o l i m i t i n g the creep of the core.
After the pipe has been cured and i s r e a d y for service (usually
up to three months a f t e r m a n u f a c t u r i n g ) the c i r c u m f e r e n t i a l tension is

taken by the core and the coating:

fS,SCf t + f t (7.6)
c2 c b2 b
Equating the movement due to e l a s t i c compression, creep and s h r i n k -
age of the core a n d coatings to that of the steel:
Deformation of the core:

w f .+f
+ - c1
-fc2-f c CI c2
+ v =
sl -fs2
(7.7)
2 c2
Ec2 Ec2 ES

( E l a s t i c ) (Creep ) (Shrinkage) ( E l ast ic-steel )

Deformat ion of coating :

Sl - f 52
b2
b2 b2 ES

( E l a s t i c ) (Creep) ( S h r i n k a g e ) ( E l a s t i c deformation of steel)

Now if the properties of the concrete core and coating are s i m i l a r ,


the e l a s t i c modulus E -
c2 - Eb2
and the s h r i n k a g e coefficient v = vb2 = v say.
c2
Solving Equs. 7 . 5 - 7 . 8 f o r f
sl - f52’
133

2+ w
v E c 2 ( t c + t 6 K b ) + tcfcl w c
fsl - fS2 = S ( l + w-
c ) + (tc+tb2+wc)/n
__ (7.9)
2
2 2+Wb

Hence from 7.8

(7.10)

(7.11)

Circumferential Stress Under F i e l d Pressure


There are a number of field loading conditions which should be
examined i n c l u d i n g :-
(1) In open trench with internal test or operating pressure p l u s
self weight a n d weight of water.
(2) I n b a c k f i l l e d trench with i n t e r n a l pressure p l u s l i v e load p l u s
self weight a n d weight of water.
(3) I n backfilled trench with live load p l u s self weight and pipe
empty.
(4) In backfilled trench with internal pressure, self weight,
weight o f water a n d t r a n s i e n t pressures.
The most highly stressed sections are usually at the crown (due to
prestressing plus bending under external loads plus internal pres-
sures), at the haunches ( d u e to prestressing p l u s bending p l u s v e r t i -
cal load p l u s i n t e r n a l pressure) a n d a t the support ( d u e to prestres-
sing plus bending plus vertical plus horizontal loads plus internal
pressure).
Compressive and tensile forces a r e taken d i r e c t l y on the effective
area per unit length of wall comprising the p i p e core pl us coating
p l u s transformed steel section:

A = t + tb + ( n 2 - 1 ) S (7.12)

Bending moments due to soil loads, external live loads and


weight of pipe plus weight of water are resisted by the effective
moment of inertia I per unit length of wall. The distance of the
c e n t r o i d of t h e effective section from the i n s i d e of the core i s :
134

(7.13)

where d i s the diameter of the prestressing wire. The corresponding


distance to the outside of the c o a t i n g i s :

e = t c + t b - e (7.14)
0

The moment of i n e r t i a of the effective section i s :

e3 + e
0
+ ( t c + ds/2 - e I 2 ( n 2 - 1 ) S.
I=____ (7.15)
3

Long itud in a I Prestressing


Some prestressed pipes are prestressed longitudinally as we1 I as
circumferential l y . The longitudinal prestressing is added to prevent
tensile c r a c k s i n the concrete d u r i n g w i n d i n g of the core a n d due to
the longitudinal bending moment i n service. The l o n g i t u d i n a l b a r s or
wires a r e p l a c e d i n the. mould f o r the concrete core a n d pretensioned
before casting the core. When the wire is wound onto the core, it
causes local bending stress and shear i n the core walls. The tension
in the longitudinal bars assists in reducing the resulting principle
stresses b y i m p a r t i n g a compressive stress to the concrete core.
As for the c i r c u m f e r e n t i a l wires, the longitudinal bars lose some
stress due to creep of the steel a n d concrete. I f the b a r s a r e i n i t i a l -
ly stressed to f . their total cross sectional area i s AS, the p i p e
SlL’
core cross sectional area is A and the creep r e l a x a t i o n coefficient is
uL, then immediately a f t e r release of the l o n g i t u d i n a l b a r s , the stress
i n the core becomes:

u f . A
L SlL s
(7.16)
f c o ~- + nlAS
a n d the l o n g i t u d i n a l steel stress becomes

fsoL = fcoLAc’As (7.17)

The core is normal l y prestressed circumferential l y immediately after


release of the longitudinal bars. The h e l i c a l wire i s wound from one
end to the other. The r a d i a l shear stress i n the core, at the p o i n t of
application of the winding, is derived from elastic theory for
pressurized c y l i n d e r s (Timoshenko and Woinowsky-Krieger, 1959) and i s
135

approxima tel y

q = 0.54 S f S O / v (7.18)

where f i s t h e circumferential steel stress a f t e r winding, and D is


so
the e x t e r n a l diameter of the core.
The maximum local longitudinal tensile and compressive stresses
due to bending i n the core w a l l d u r i n g w i n d i n g h a v e been determined
experimental l y to be approximately

fcmL = 0.3fco (7.19)

where fco i s the core c i r c u m f e r e n t i a l stress a f t e r winding.


The effect of the l o n g i t u d i n a l steel on the cross sectional a r e a has
been neglected in Equs. 7.18 and 7.19 as the steel is near the
neutral a x i s a n d the expressions a r e approximate anyway. The rnaxi-
mum tensile stress in the core wall may be d e r i v e d w i t h the assis-
tance of a Mohr d i a g r a m ( F i g . 7 . 3 ) :

I f fcmL > fcoL, which i s u s u a l l y the case, then

fcml - f ( 7.20a )
c t L =/qz +
-
( ""4 2- + COL
2 2

and i f f c o L > fcmL

then f c t L = h2 + ( f C 0 L - FcmL
2
2 -
fcoL
2
- cmL
(7.20b)

I
SHEAR
STRESS

Fig. 7.3 Mohr c i r c l e f o r stress i n core d u r i n g w i n d i n g .


136

It will be seen from Equs. 7.20 a and b that the larger the
longitudinal prestress, which i s proportional to f the smaller is
COL
the p r i n c i p l e t e n s i l e stress in the core f
ctL'

Longitudinal S t r e s s e s A f t e r L o s s e s
The longitudinal bars act to resist longitudinal bending of the
pipe in the f i e l d . The p i p e a c t s a s a beam s p a n n i n g between uneven
points in the bedding. At this stage a c e r t a i n amount of shrinkage
and creep will have occurred in the concrete and some longitudinal
compressive stress will h a v e been t r a n s f e r r e d to the c o a t i n g . As the
longitudinal stress is not high, the concrete creep can usually be
neglected. S h r i n k a g e of t h e c o r e does r e d u c e t h e t e n s i l e s t r e s s i n t h e
b a r s and t h e c o m p r e s s i v e s t r e s s i n t h e c o r e becomes:

u f A - v ESAS
L siL s
(7.21 )
fc2L =
Ac + n2AS
Once the pipe is in service, longitudinal bending stresses are
added to the stress due to prestressing. The extreme fibre bending
stresses a r e

-
-
MD
-
fc3L 21 (7.22)

where M i s t h e b e n d i n g moment and t h e e f f e c t i v e moment o f inertia of


the section i s

l (d + 2t + d 5 ) *
I = -64 (l D 4- d
4 + (n2 - 1 ) AS (7.23)
8

P r o p e r t i e s of Steel and C o n c r e t e
The steel used for w i n d i n g prestressed concrete p i p e should h a v e
a s h i g h a y i e l d s t r e s s a s p o s s i b l e and a y i e l d s t r e s s o f 1650 N/rnmz
(240 000 p s i ) i s n o t uncommon. This will e n s u r e t h a t a f t e r c r e e p and
shrinkage have taken place, the r e m a i n i n g compressive stress in the
concrete is fairly high. The steel stress w i l l d r o p a f t e r w i n d i n g , and
it is normal to confine working stresses to less than 50% o f yield
stress.
Steels with high yield stresses are often b r i t t l e and difficult to
work w i t h and in f a c t the u l t i m a t e s t r e n g t h may not be much h i g h e r
137

than the yield stress. Care is therefore necessary in selecting the


prestressing w i r e . T h e modulus of e l a s t i c i t y of steel i s approximately
6
200 000 N/mm2 (30 x 10 psi).
The steel may creep s l i g h t l y a f t e r prestressing but it is difficult
to distinguish between concrete a n d steel creep between w i n d i n g and
works test, so they are usually considered together. This loss in
stress in the steel is typically around 5% a n d occurs w i t h i n a few
hours a f t e r w i n d i n g .
Concrete cores are cast under vibration or centrifugally and 28
day cube c r u s h i n g strengths of 60 N/mm2 ( 8 700 p s i ) a r e f r e q u e n t l y
achieved. High e a r l y strength (e.9. 50 N/mm2 o r 7 200 p s i ) i s desir-
able for w i n d i n g as e a r l y as possible. Compressive stresses u p to 50%
of the cube strength at the time a r e permitted d u r i n g w i n d i n g (any
cracks will most p r o b a b l y h e a l ) . Working compressive stresses should
be confined to less than 1/3 of the 28 day cube strength.
The bending tensile stress in the core d u r i n g w i n d i n g should be
less than about 10% of the cube strength at that age, and the
calculated bending tensile stress in the core due to longitudinal
bending in the field should be limited to about 5% of the cube
strength i n order to a l l o w f o r unknowns a n d to prevent any p o s s i b i l -
i t y of cracks developing. No c i r c u m f e r e n t i a l tensile stress i s permitted
i n the core i n the f i e l d f o r normal o p e r a t i n g conditions b u t tension i s
sometimes p e r m i t t e d under transient conditions. The tensile stress in
the coating should be less than about 10% of the cube strength.
(This i s for b u r i e d pipe. Exposed pipes may develop cracks a n d the
tensile stress should be less t h a n t h i s . )
BS 4625 does not permit tensile stress in the core for normal
operating plus backfill pressures, but permits a tensile stress of

0 . 7 4 7 m if water hammer pressure i s included, a n d a tensile stress of


0.623 / F during works hydrostatic test. (Where F is the crushing
strength of 150 mm cubes a t 28 d a y s , a l I i n N/mm2 . )
The modulus of elasticity of concrete increases w i t h the s t r e n g t h
and v a r i e s from 20 000 N/mm2 ( 3 000 000 p s i ) to 40 000 N/mm2
( 6 000 000 p s i ) (BSCP 2007, 1970; Ferguson, 1958).
The creep coefficient, w, of concrete i s defined b y the equation

creep s t r a i n = AL/L = w fc/Ec


138

where fc is the average compressive stress during the time that


creep occurs a n d E i s the f i n a l modulus of e l a s t i c i t y . w varies with
time a n d the method of c u r i n g . Creep i s high f o r ' g r e e n ' concrete and
the r a t e of creep reduces w i t h time. w i s approximately 0.3 two days
after casting (at the time of factory testing the core) and is 1.3
three months a f t e r casting. Hence f o r the coating, the coefficient 1.3
i s used to c a l c u l a t e creep, b u t f o r the core, the creep coefficient f o r
the time between f a c t o r y test a n d f i e l d conditions i s 1.3 - 0.3 = 1.0.
I t may be 30% h i g h e r f o r pipes not p r o p e r l y cured o r exposed to the
elements i n the f i e l d .
Shrinkage of concrete also depends l a r g e l y on the method of c u r -
ing, typical values of s h r i n k a g e per u n i t length, v, v a r y i n g from
to

Example
Calculate circumferential concrete and steel stresses during
winding, after curing and under field conditions for the
prestressed concrete p i p e described below:

Effective p i p e length L = 4 m, bore d i = 2 000 rnm, core thickness


tc = 75 mrn, coating tb = 25 mm, steel winding 5 mm diameter at
15 mrn c/c. S = 0.00131 m2/m of pipe. L o n g i t u d i n a l wires 24 N o . x

8 mm diameter, A S = 0.00121 m2. Maximum internal pressure 0.8


N/mmz. Vertical loading due to soil and live load 0.04 N/mrn2,
lateral soil pressure 0.5 x vertical = 0.02 N/mmz. Bottom support
e f f e c t i v e l y over 30" of arc. Neglect selfweight a n d weight of water
inside.
Steel: L o n g i t u d i n a l prestress = 400 N/mm2 .Winding prestress =
1000 N/mmz. ES = 200 000 N/mm2. Creep coefficient u = 0.95.
Concrete: E
= 30 000 N / m m 2 , Ec2 = 38 000 N/mmz, Creep w = 1.0
CI
from f a c t o r y test to f i e l d test. Wb = 1.3, shrinkage v = 10
4
:
.
n = 200 000/30 000 = 6.7
1
n = 200 ooo/38 000 = 5.3
2
Stresses due to w i n d i n g

- 1 000 x 0.0131
= 16.5 N/mm2 (7.2)
fco - 0.075 i 0.00131 x 6 . 7 / 2
139

0.075
= 16.5 = 880 N/mmz (7.3)
so 0.00131

Works t e s t o f c o r e :

= 0.95 x 16.5 = 15.7 N/mm2 (7.4)


fcl

fSl
= 0.95 x 880 = 840 N/mm2

a f t e r curing :
2+1 .o
' (0.075+0.025m)+O. 075x15.7xl.O
1 O-4x38x1 0
f -f = (7.9)
s2 0.00131 (ld$)+(0.075+0.02*)/5.3 2+1 0

= 76 N/mmz

= 840-76 = 764 N/mmz


s2
76 38 Oo0 = 6.4 N/mmz (7.10)
fb2 = (200 000 2+1 .3

- 764~0.00131-6.4x0.025
= 11.2 N/mm2 (7.11)
c2 0.075

Under f i e l d loading:

Transformed section A = 0.075 + 0.025 + (5.3-1)0.00131


= 0.105 m'/m (7.12)
C e n t r o i d to
0.12/2+(5.3-1)(0.075+0.005/2)(0.00131)
inner surface e =
0.1+0.00131(5.3-1)

= 0.0515 (7.13)

Cent r o id t o
outer surface e = 0.1 - 0.0515 = 0.0485 (7.14)

Moment o f inertia I = 0*05'53+0*04853 + (0.075+0.005/2-0.0515)2


3
x(5.3-1)(0.00131) = 85.9~10-~m~/m (7.15)
Stresses a t b a s e B : ( t e n s i o n - v e f o r concrete stresses)
T e n s i o n d u e to n e t i n t e r n a l p r e s s u r e :

0.8 x 2 - 0.02 x 2.2 - -7,4 N/mmz


ft =
2 x 0.105

B e n d i n g moment c o e f f i c i e n t s f r o m F i g . 8.2:
For vertical load, Nb = 0.235, and f o r h o r i z o n t a l l o a d , NS =
0.125
Net b e n d i n g moment = 0.235 x 0.04 x 2.2/2 - 0.125 x 0.02 x 2.2/2
= 0.00755
140

Stress on o u t e r face f - - f + -MeO


I
b3 - fb2 t

0.00755 x 0.0515= -IN/mm2


Stress on i n n e r face f = 11.2-7.4-
c3
85.9 x
Steel s t r e s s : e = 0.075 + 0.005/2-0.0515 = 0.026 (on outer side
of centroid)

0.00755 x 0.026) = 791 N/mm2


= 764 + 5.3 (7.4
85.9 x

Although the stresses a t t h e b a s e a r e u s u a l l y t h e most s e v e r e , the


stresses at other points on the circumference should be checked
similarly, and c h e c k s s h o u l d b e d o n e b o t h w i t h and w i t h o u t t r a n s -
ient winding, t e s t i n g and i n t h e f i e l d .

REFERENCES

Am. C o n c r e t e P i p e A s s n . , 1970. D e s i g n M a n u a l - C o n c r e t e P i p e , A r l i n g -
ton.
BSCP 2007 P a r t 2, 1970. R e i n f o r c e d and P r e s t r e s s e d C o n c r e t e S t r u c -
t u r e s , BSI, L o n d o n .
C o n c r e t e P i p e Assn., 1962. L o a d s o n B u r i e d C o n c r e t e P i p e s , Tech.
B u l l e t i n No.2, T o n b r i d g e .
C o n c r e t e P i p e Assn., 1967. B e d d i n g and J o i n t i n g o f F l e x i b l y J o i n t e d
C o n c r e t e P i p e s , Tech. B u l l e t i n No. 1 , T o n b r i d g e .
F e r g u s o n , P.M., 1958. R e i n f o r c e d C o n c r e t e f u n d a m e n t a l s , Wiley,N.Y.
K e n n i s o n , H.F., 1950. D e s i g n of p r e s t r e s s e d c o n c r e t e c y l i n d e r p i p e , J.
Am. W a t e r W o r k s Assn., 42.
T i m o s h e n k o , S.P. and W o i n o w s k y - K r i e g e r , S., 1959. T h e o r y o f P l a t e s
and S h e l l s , 2 n d Edn., McGraw H i l l , N.Y.

L I S T O F SYMBOLS

A - cross sectional area


d - inside diameter
D - outside diameter
e - distance from centre of g r a v i t y
E - modulus of e l a s t i c i t y
f - stress (compressive o r tensile)
F - c r u s h i n g s t r e n g t h o f 150 mrn c u b e s a t 28 d a y s
141

I - moment of inertia
K - factor o f safety
L - length
M - bending moment
n - e l a s t i c modular r a t i o E /E c
s
q - shear stress
s - steel area per u n i t length of p i p e
t - thickness
u - creep coefficient f o r steel and concrete before f a c t o r y test
v - s h r i n k a g e coefficient
w - creep coefficient

Subscripts
b - coating
c - core
s - steel
m - bending

9 - shear
t - tensile
L - longitudinal
I - initial

0 - after winding
1 - at time of f a c t o r y test
2 - a t time of laying
3 - under f i e l d pressure
142

CHAPTER 8

STEEL AND FLEXIBLE P I P E

1 NTERNAL PRESSURES

The highest pressure a pipe has to resist are normally those


due to internal fluid pressure. The pressure is uniform around the
pipe and there are no bending stresses except for water and self

weight effects. The general equations for the resulting stresses


in a hollow cylinder are:
Circumferential wal I stress
p.d.' - podo' d i 2 doZ( P-,P )
I I
(8.1 )
Fw = do2 - diL doL - d.'

p.d.'
I 1 - podo' d i z do2 ( P,-P 1
Radial stress
Fr = do2 - d.'
+ doZ - d i z
(8.2)

where p i s pressure, d i s diameter at w h i c h the stress i s sought, d.


is internal diameter and d i s external diameter. The e q u a t i o n s a r e
f o r p l a i n stress, i.e. a cylinder free to expand longitudinally.
For the p a r t i c u l a r case of n o e x t e r n a l p r e s s u r e the c i r c u m f e r e n t i a l
s t r e s s i s a m a x i m u m on t h e i n n e r s u r f a c e a n d i s

di2 + do2 P
F max = -
W di + do 2t

In practice the wall thickness is normally small in comparison


with the diameter, a n d the w a l l thickness of a steel p i p e designed to
resist i n t e r n a l pressure i s obtained from the formula

where t is t h e w a l l t h i c k n e s s
p i s the i n t e r n a l pressure
d i s the e x t e r n a l diameter
f i s t h e y i e l d s t r e s s o f t h e steel
G i s the design factor
J i s the j o i n t factor.
E q u . 8.4 overpredicts the true stress by some 1% for every 1%
of t/d. On t h e o t h e r h a n d , just u s i n g d instead of (d-t) would under-
3 r e d i c t s t r e s s b y 1% f o r e v e r y 1% t / d .
143

The design factor G allows a safety margin. What factor to use


will depend on the accuracy with which loads a n d t r a n s i e n t pressure
have been assessed, the working pressure, economics and the conse-
quences of a burst. I f the p i p e l i n e i s protected a g a i n s t water hammer
over-pressure, a factor of 0.6 o r even 0.7 i s reasonable, b u t i f there
are many unknown loads, use of a factor of the order of 0.3 to 0.5
would be wiser.
The j o i n t factor J allows for imperfections i n welded seams, and
varies from 0.85 for furnace butt-welded joints to 1.0 for seamless
pipes.
Smal I-bore, high pressure pipes are designed to resist internal
pressures only, but the larger the diameter and the thinner the
wal Is, the more important become the e x t e r n a l loads. Vaporization of
the fluid in the pipe may also be possible, in which case this
internal pressure should be considered acting in conjunction with
e x t e r n a l loads.
If the p i p e i s restrained longitudinally, a longitudinal tension of
magnitude VF is induced where v is Poisson's ratio and F is the
circumferential wal I tension.

TENS I ON R I NGS TO RESIST I NTERNAL PRESSURES

Pipes which have to resist very high internal pressures may be


strengthened b y b i n d i n g w i t h hoops a n d s p i r a l s . It i s often d i f f i c u l t
to form pipes in the thick p l a t e s which would be r e q u i r e d to r e s i s t
some h i g h pressures. On the other hand it may be possible to form
the p i p e of t h i n n e r p l a t e than would be r e q u i r e d f o r an unstrengthen-
ed pipe, a n d w i n d i t w i t h s t r a p s o r rods.
The r i n g s r e q u i r e d to r e s i s t internal pressures do not perform the
same function as the stiffening rings required to resist col lapse
against external loads. The w i n d i n g to r e s i s t internal pressures acts
in tension and does not need to be as prominent as the rings to
resist external load, which should increase the moment of i n e r t i a of
the l o n g i t u d i n a l section through the p i p e a n d therefore be as h i g h as
possible. I n fact the tension r i n g s , as they w i l l be c a l l e d , should be
144

as f l a t a n d b r o a d as possible to keep the distance between them to a


minimum. There will be high longitudinal bending stresses and
circumferential stresses in the pipe wall between rings if the r i n g s
are f a r apart. I n comparison, s t i f f e n i n g r i n g s to r e s i s t e x t e r n a l loads
are usually spaced a number of diameters a p a r t . Tension r i n g s may
also be used to strengthen old pipes which are to be subjected to
h i g h e r pressures than o r i g i n a l l y designed for.
Using the stress-strain r e l a t i o n s h i p and the d i f f e r e n t i a l equations
of equil i b r i u m f o r c y l i n d e r she1 I s under the action of r a d i a l pressure,
Timoshenko (1959) d e r i v e d a n equation f o r the radial displacement of
a cylinder subjected to a radial pressure a n d w i t h tension r i n g s at
equi-spaci n g .
12(1-v2)
Put b4 = (8.5)
dZt'

and a = bs (8.6)

where v i s the P o i s s o n ' s r a t i o (0.3 for s t e e l ) , d i s the p i p e diameter,


t i s the w a l l thickness a n d s the centre-to-centre r i n g spacing.

Let
-
- cosh a + cos a
x1 sinh a + sin a (8.7)

- sinh a - sin a
x2 sinh a + sin a (8.8)

-
- cosh a - cos a
x3 sinh a + sin a

Then the r i n g stress, Fr, i s g i v e n by

(8.11)

where A = cross sectional area of the r i n g .


The circumferential pipe wall stress, Fw, mid-way between rings is
given by

= @[l+ 2 ( F -
2t - 1 ) X 4 ] (8.12)
FW 2t r Pd

a n d the longitudinal bending stress in the p i p e under the rings,


Fb
i s given by
145
g!
r i
i
a,
D
C
3
m
a,
L
c
m
[r
C
.-
U
C
a,
m
I I I I 1 I I
2 -0
. <
0
P

m
In
a,
L
I
m
a
.-C
LL
D
m
a
m
a .-C
L
.-C
> L C
N
d .-0
m
-
C
a,
r.-
5
D
.-
E
-
U
.-Um
m
QJ
m
m
c
m m
-
- m
a,
5
3
- -
vi
- a,
i
.-5 m
C
a,
L
a,
-
c m
5 .-a
.-i LL
V
m
146

(8.13)

The stresses are indicated by Fig. 8.1, for v = 0.3. It may be


shown that a s s i s d e c r e a s e d t h e r i n g s t r e s s Fr tends to f pds/ (ts +
A) w h i c h c o u l d h a v e been anticipated. N o t e t h a t f o r s m a l l A, Fr tends
to equal the circumferential wall stress of a p l a i n pipe, pd/2t. Also,
f o r s m a l l s, the circumferential p i p e wall stress F
W
tends to 4 pds/(ts
+ A ) , w h i c h w o u l d b e e x p e c t e d , and t h e l o n g i t u d i n a l beding stress Fb
tends to zero.

For l a r g e r i n g spacing ( > approx. 2-1,

F tends to 1 -
pd (8.14)
1 + 0.9lA/t& 2t

Fb tends to 1.65 -
pd (8.15)
tJtd/A + 0.91 2t

and F t e n d s to @ i.e. t h a t f o r a p i p e w i t h o u t r i n g s .
W 2t
I t may be observed b y comparing F F r and Fb f r o m F i g s . 8.la,
W’
8.lb and 8 . l c t h a t t h e m a x i m u m s t r e s s f o r most p r a c t i c a l r i n g s i z e s i s
in fact Fw, the circumferential pipe wall stress. Also, Fw is only
reduced if s / J y z i s l e s s t h a n a p p r o x i m a t e l y 2.0. I n other words, the
r i n g spacing should be l e s s than 2 J l d and the r i n g cross sectional
area A should be of the same order of magnitude as the p i p e w a l l
longitudinal cross sectional area between rings, ts, to enable the
r i n g s t o b e o f use.

DEFORMATION O F CIRCULAR P I P E S UNDER EXTERNAL L O A D

For large diameter and flexible pipes under low internal pres-
sures, the external load is frequently t h e c r i t i c a l one. Pipes may f a i l
under external load by buckling, overstressing due to arching or

bending, excessive deflect ion.


For elastic rings under plane stress subjected to vertical loads
only, Spangler (1956) e v a l u a t e d t h e b e n d i n g moments and d e f l e c t i o n a t
critical points around the circumference. The worst b e n d i n g moments
occur at the crown, the i n v e r t or t h e s i d e s . T h e b e n d i n g moments p e r
147

Angle of bottom suwor? p degrees

Angle of tmttom support p degrees

Fig. 8.3 D e f l e c t i o n c o e f f i c i e n t s for loaded pipe.


148

u n i t length a r e g i v e n b y a n equation of the form


M = NWR (8.16)

where R is the pipe radius, W i s the v e r t i c a l load per unit length


and N is a coefficient. (Nt, Nb and NS f o r top, bottom and sides
respectively). The vertical and horizontal changes in diameter are
p r a c t i c a l l y equal a n d opposite a n d a r e of the form

A = NnWR3/EI (8.17)

where N i s a coefficient.
A
The moment of i n e r t i a I per u n i t length of p l a i n p i p e w a l l is

I = t3/12 (8.18)

Figs. 8.2 and 8.3 g i v e values of Nt, Nb a n d Ns f o r the bending


moments at the top, bottom and sides respectively, and N for the
A
vertical deflections of the pipe diameter. The coefficients are for a
line load or a load d i s t r i b u t i o n across the w i d t h of the p i p e ( W per
unit length), and for different angles of bottom support 6 , as
indicated i n Fig. 8.4.
Collapse of a steel p i p e w i l l p r o b a b l y not occur u n t i l the diameter
has been distorted some 10 o r 20 percent. I n p r a c t i c e deflections up
to 5 percent of the diameter are sometimes tolerated. The deflection
should normally be limited to about 2 percent to prevent damage to
linings, a n d for pipes w i t h mechanical j o i n t s .
The h y d r a u l i c properties, i.e. the cross sectional area a n d wetted
perimeter a r e not affected noticeably f o r normal d i s t o r t i o n s .

LOAD W /UNIT LENGTH OF PIPE

UPTHRUST

F i g . 8.4 Pipe l o a d i n g and deflections


149

Effect of L a t e r a l Support
The lateral support of sidefill in a trench increases the s t r e n g t h
of flexible pipes considerably and reduces deformations. Without
lateral support to a p i p e the r i n g bending stresses a t the s o f f i t a n d
haunches o r deflections would l i m i t the v e r t i c a l e x t e r n a l load the p i p e
could carry. But a pipe in a compacted fill will deflect outwards
l a t e r a l l y as it is loaded v e r t i c a l l y thereby i n c r e a s i n g the pressure of
the sidefill against the sides of the pipe. An equilibrium condition
may be established with the vertical load being transferred to the
haunches b y a r c h action as well as b y r i n g action. The stress due
to the a r c h action i s compressive so that the load which the p i p e can
carry i s considerably h i g h e r than i f the p i p e were a c t i n g in bending.
In the extreme case, the lateral stress will equal the v e r t i c a l load
stress a n d the p i p e wal I w i l l be i n p u r e compression, w i t h the stress
equal to wd/2t

where w = W/d (8.19)

If the p i p e underwent no noticeable lateral distortion the load it

could support would be determined by the bending strength plus


whatever arch strength i s given to the p i p e b y the s o i l pressure on
the sides of the pipe, i.e. the permissible v e r t i c a l load per u n i t area
on the p i p e i s

w = w + w (8.20)
b a
where i s the permissible
w bending load (limited either by ring
b
bending stress or more l i k e l y by deflection). w i s the a r c h i n g load,
a
equal to lateral soil pressure on the sides of the p i p e , which, con-
servatively, may be taken as the a c t i v e soil pressure, but i s usually
greater than this as indicated in subsequent equations. For sand,
the a c t i v e lateral pressure i s approximately one t h i r d of the v e r t i c a l
pressure due to soil dead load o n l y and for clay i t i s approximately
h a l f the v e r t i c a l pressure.

I f the v e r t i c a l load i s g r e a t e r than the sum of the r i n g load p l u s


active lateral soil pressure, the pipe wall will deflect out laterally
and increase the lateral pressure. The h o r i z o n t a l stress w i I I decrease
away from the pipe and Barnard (19571, using elastic theory,
150

suggests assuming a triangular stress d i s t r i b u t i o n w i t h the horizontal


pressure equal to total vertical pressure minus r i n g load a t the p i p e
wall decreasing l i n e a r l y to zero at 2.5d away from the p i p e w a l l . The
corresponding l a t e r a l deflection of each side of the p i p e i s

AX/2 = 1.25 (w-w,) d/ES (8.21 )

where E i s the effective modulus of e l a s t i c i t y of the s o i l . The factor


1.25 should be increased as the l a t e r a l deflection increases, since the
radial pressure increases as the r a d i u s of c u r v a t u r e decreases. The
factor becomes 1.4 for a deflection of 2 percent of the diameter and
1.7 for a deflection of 5 percent. The deflection also increases w i t h
time, and an a d d i t i o n a l 25 to 50 percent of the i n i t i a l deflection can
be expected e v e n t u a l l y .
If the creep deflection of the soil is large, lateral support can
decrease. On the other hand pipe materials which exhibit creep,
e.g. plastics, can compensate f o r this, a n d could even shed v e r t i c a l
load.
The relationship between stress and s t r a i n f o r the s o i l should be
determined from laboratory triaxial consol i d a t i o n tests. The effective
modulus of elasticity of soil varies widely depending on soil type,
degree of compaction o r n a t u r a l density, c o n f i n i n g pressure, duration
of l o a d i n g and moisture content. F o r example it may be as low as 2
N/mmz for loose c l a y or as high as 20 N/mm2 f o r dense sands. The
modulus is approximately 3 N/mmzfor loosely compacted f i l l , 5 N/mrn2
for fill compacted to 90% Proctor density a n d 7 N/mmz f o r f i l l to 95%

Proctor density. Values h i g h e r than 100 have been recorded for moist
compacted sands.
The r i n g load on the p i p e i s p r o p o r t i o n a l to t h e e l a s t i c deflection
as indicated by Equ. 8.17. T a k i n g N, equal t o 0.108, which corres-
ponds to bottom support over 30' and the load over the entire pipe
width and p u t t i n g

AY = AX = A , then s o l v i n g f o r A/d from 8.17 and 8.21,

A
- -- 0. 108wd3
(8.22)
d 8E I +0.043ESd3

For p l a i n pipe I = t3/12, so one has an equation f o r deflection as


a function of diameter, loading, soil modulus and the ratio wall
th i ckness/di ameter.
A
- -- 0.108~
d 0.67E(t/d)' + 0.043Es (8.23)
151

The r e l a t i o n s h i p between deflection a n d wal I thickness f o r p l a i n


p i p e i s p l o t t e d i n F i g . 8.5. I t w i l l be observed that a steel p i p e
wall thickness as low as 4% of the diameter will be sufficient to
restrain distortion to 2% p r o v i d e d the soil modulus is greater than
5MPa a n d soil load i s less than 50 kPa.

Pipes are sometimes strutted internally during backfilling of the


trench to increase the vertical diameter and reduce the horizontal
diameter. The lateral support increases when the s t r u t s a r e removed
a n d the p i p e tends to r e t u r n to the r o u n d shape. The v e r t i c a l deflec-
tion a n d tendency to b u c k l e a r e consequently reduced considerably.

STRESS DUE TO C I RCUMFERENT I A L BEND I NG

It i s possible to compute w a l l stresses due to bending a n d a r c h i n g


(Stephenson, 1979). If i t can be assumed that the load i s spread over
the f u l l w i d t h of the p i p e ( a = 180') and the bottom support i s over
60' ( 6 = 60") f o r f l e x i b l e p i p e , then from (8.17) a n d F i g . 8.3,
4
AY = 0.103wbd /8E1 (8.17b)

Now from (8.21), A X = 2.5(w-wb)d/ES (8.21b)

Equating AY a n d A X , a n d s o l v i n g for w b the r i n g load i n terms of w


the total load,
w I /d3 (8.24)
w =
b I/d3 + 0.006ES/E

The bending moment in the wall, M, is due to ring load a n d is a


maximum a t the base a n d from (8.16)

M = 0 . 1 9 ~ d 2 / 2 hence b e n d i n g stress f = Mr/l (8.25)


b b
where r is the distance from the centre of g r a v i t y of the section to
the extreme f i b r e (t/2 for p l a i n wall pipe). The balance of the load
i s taken in a r c h action of the p i p e according to (8.20) so
0. 006wES/E
W -
-
(8.26)
a I /d' + 0. 006ES/E
The bottom w a l l hoop stress i s

fa = wad/2a (8.27)

where a is the cross sectional area of wall per unit length ( t for
plain wall pipe).
152

The total lateral compressive stress i n the w a l l at the base i s f


=
fb + fa, therefore permissible l o a d i n g p e r u n i t area i n terms of
permissible stress f is

(l/d3 + 0.006ES/E)f
(8.28)
w =
0.1 r/d + 0.003 dE /Ea

Thus the permissible v e r t i c a l load w i s a f u n c t i o n of the permis


s i b l e stress, the r e l a t i v e thickness t / d a n d the r a t i o o moduli of soi
ES to steel E. The r e l a t i o n s h i p i s plotted i n Fig. 8.5 f o r p l a i n wal
pipe, w i t h E = 210 000 N/rnmz, a n d f = 210 N/mrn2.
For t h i c k pipes, the permissible load increases w i t h wal I thickness
as more load can be taken i n r i n g bending. For t h i n n e r pipes, the
deflection becomes large so that the soil side-thrust increases until
the pipe is i n p u r e compression, and the limit i s due to the wall
hoop stress. Actually side w a l l hoop stress exceeds that a t the base,
so (8.27) p l u s (8.25) i s not the l i m i t i n g stress f o r h i g h a r c h i n g .
On the same c h a r t i s p l o t t e d b u c k l i n g load

W = /32ESEl/d3 (8.29a)

The b u c k l i n g equation, proposed b y ClRlA (1978) allows for lateral


support. This may be compared w i t h an a l t e r n a t i v e equation investi-
gated by the Transport and Road Research Laboratory (which is
found to o v e r p r e d i c t w ) :

W = ( 1 6 E s 2 E l / d 3 )1 / 3 (8.29b)

The load w d g i v i n g a deflection of 2% i n the diameter i s also p l o t t e d


on the c h a r t from (8.23). The c h a r t thus y i e l d s the l i m i t i n g c r i t e r i o n
for any p a r t i c u l a r w a l l thickness r a t i o . The lowest permissible load w
is selected in each case from the chart by comparing deflection,
overstressing and b u c k l i n g lines f o r the r e l e v a n t t / d and E .
Similar charts should be p l o t t e d where stiffening rings a r e used
a n d where a l t e r n a t i v e m a t e r i a l moduli and permissible stresses a p p l y .
By careful design it is possible to reach the limit in two or more
c r i t e r i a a t the same load. T h i s i s c a l l e d balanced design.

More General Deflection Equations


Spangler (1956) allowed for lateral support to the pipe due to
the backfill in a more theoretical way than Barnard. The equation
he d e r i v e d f o r v e r t i c a l deflection i s :
153

UZWd3
A = (8.30)
8 E l + 0.06Esd”

T h i s corresponds to ( 8 . 2 2 ) i f UZ i s substituted for N and a v a l u e of


A
0.15 i s assumed f o r N A i n the denominator ( i n s t e a d of 0.108 i n 8.22)
Here U = s o i l consolidation time l a g factor ( v a r i e s from 1.0 to 1.5).
Z = bedding constant (varies from 0.11 for point support to
0.083 f o r b e d d i n g the f u l l w i d t h of p i p e ) , n o r m a l l y taken as
0.1.
1 = moment of i n e r t i a of p i p e w a l l per u n i t length.
E = passive resistance modulus of s i d e f i l l .
The pressure inside a pipe may also contribute to its stiffness.
Due to the fact that the vertical diameter i s compressed to s l i g h t l y
less t h a n the h o r i z o n t a l diameter, the v e r t i c a l u p t h r u s t due t o i n t e r n -
al pressure becomes g r e a t e r than the s i d e t h r u s t b y an amount of 2@
which tends to return the p i p e t o a c i r c u l a r shape. A more general
expression f o r v e r t i c a l deflection thus becomes

UZWd3 (8.31)
A =
8 E I + 0.05ESd’ + 2 UZpd3

W
Q
0.002 0.005 0.01 0.02 0.03
Relative thickness t / d

F i g . 8.5 Permissible load on p l a i n p i p e


154

ST FFENING RINGS TO RESIST BUCKLING W I T H NO SIDE SUPPORT

Morley (1919) developed a theory for the buckling of stiffened


pipes under uniform external pressure. The theory often indicates
stiffening ring spacings wider than i s considered necessary in p r a c -
tice. The theory also neglects the possibility of failure under com-
bined internal a n d e x t e r n a l pressures and bending. The equations do,
however, y i e l d an i n d i c a t i o n of s t i f f e n i n g r i n g spacing.
Using an analogy with a strut, Morley developed an equation
which indicates the maximum v e r t i c a l external pressure, w, which a
cylindrical she1 I can take without b u c k l ing. The equation a l lows f o r
no axial expansion, and assumes the wall thickness is small in
comparison w i t h the diameter. v i s Poisson's r a t i o ,

24 El
w = (8.32)
1 - v 2 d7
For p l a i n p i p e , I = t3/12, so

2E t
w =
1 - v2
(3) (8.33)

Experiments i n d i c a t e d a p e r m i s s i b l e stress 25% less than the theoreti-


c a l , so f o r steel,
t
w = 1.65 E (8.34)

For thick-wal led tubes, the col lapse pressure wi II be that which
stresses the w a l l material to i t s elastic limit ( w = 2ft/d) whereas f o r
intermediate wal I thickness, f a i l u r e w i I I be a combination of b u c k l i n g
and elastic yield. An empirical formula indicating the maximum
permissible pressure on a p i p e of intermediate thickness i s

(8.35)

where f i s the y i e l d stress.


The e x t e r n a l load may be increased i f s t i f f e n i n g r i n g s a r e used to
resist buckling. It was found by experiment t h a t the collapse load,
w, i s i n v e r s e l y p r o p o r t i o n a l to the distance between s t i f f e n i n g r i n g s ,
s, if s i s less than a c e r t a i n c r i t i c a l length, L.
155

From experiment L = 1 . 7 3 E (8.36)

so i f w = 2E ( t / d ) 3 for s t i f f e n i n g pipe, then f o r s t i f f e n e d p i p e ,

w = L2E ( t / d j 3 = 3.46 (8.37)

The actual permissible stress is less than the theoretical due to


imperfections i n the m a t e r i a l a n d shape of p i p e , so the p r a c t i c a l r i n g
spacing i s g i v e n by

(8.38)

If the full elastic r e n g t h of he p i p e i s to be developed o resist


vertical external pressures i.e. w = -
'Jt, then the ring spacing

shou I d be

(8.39)

t
Rings w i l l o n l ybe of use i f w >1.65E ( - ) 3 . I t should also be ascer-
d
2f t
tained that w z - , which i s the e l a s t i c y i e l d p o i n t .
d
If w = Pft/d, then r i n g s w i l l o n l y be of use i f

t/d<f/E = 1/30 (8.40)

f o r m i l d steel. F u r t h e r studies of b u c k l i n g of s t i f f e n e d p i p e a r e g i v e n
b y Stephenson (1973) a n d Jacobsen (1974).

Fig. 8.6 Pipe w i t h s t i f f e n i n g r i n g s


156

Example 1 :

Tension Rings
A 500 mm bore p i p e i s to take an internal pressure of 5 N/mm2.
The maximum permissible working stress in the wall is to be 100
N/rnm*, and the maximum wall thickness which can be rolled is 10
mm. Assume a 10 rnm diameter w i r e is av ai l abl e for b i n d i n g the p i p e
and calculate the required spacing of the binding. An unstiffened
p i p e c o u l d take an i n t e r n a l pressure of

“ow pd
2t
=
2 x
500
100
= 125, s o we r e q u i r e F/@
2t
= 100
125
= 0.8

1 82A - 1 . 8 2 ~ ~ ~= 1o 0
. 2~
Tension r i n g area parameter -
tr‘td
-
10 ~ ‘ 1 0 x 5 0 0

By inspection of Figs. 8.la and 8.1b, the r i n g spacing r e q u i r e d to


reduce both the c i r c u m f e r e n t i a l p i p e w a l l stress a n d the r i n g stress to
the desired v a l u e i s g i v e n b y

s / m = 0.45, :. s = 0.45 J10 x 500 = 32 mm.

Notice t h a t the amount of steel required i s s l i g h t l y more than if


the p i p e were merely thickened to take the e x t r a pressure. T h i s effect
becomes more noticeable the wider the r i n g spacing. For wider spacing
it is usually the circumferential wall stress which is the limiting
stress in the system, so there i s l i t t l e point i n using high- tensile
r i n g s i f the p i p e w a l l i s of m i l d steel.

Example 2:

St i f f e ning R i n g s
Design a 3 m diameter steel pipe with stiffening r i n g s to support
a uniformly distributed load of 60 kN/m2 with a soil modulus of 3
N/mm2.
Maximum steel stress = 103 N/mm2, delection 2%.
Select t = 10 mm
Then s / d = (210 000/503) ( . 0 0 3 3 ) 3 / 2 (8.39)
157

= 0.39 :. s = 1177 mm; Select 1200 mm.


Use 100 mm h i g h r i n g s w i t h h/b = 10 m a x f o r no b u c k l i n g i.e. ring
t h i c k n e s s b = 10 mm

(h+t)bh
Centroid from c e n t r e w a l l c = = 4.25 mm (8.41)
2( bh+ts)
h3b + S ( t c ) ' + -
t3 + tC2
Moment o f i n e r t i a I = - (8.42)
12s hb 12

= 3104 mm3/mm

Extreme f i b r e d i s t a n c e r = h +
= 100.8 mm (8.43)

0.1~0.060~3~
Deflection: A/d = (8.22b)
8x210 0 0 x 3 1 0 4 ~ 10-9+0.05x3x33

= 0.017 = 1.7%

Maximum w a l l s t r e s s

(0.1x0.10/3.0+0.003x3x3/(210000x0.011 ) ) 0.06
f =
3 1 0 4 ~ 1 0 - ~ /+0.006x3/210000
3~
= 1.7 N/mm2.

The section could thus be modified. By v a r y i n g the r i n g p r o p o r -


tions, an optimum or balanced design may be possible i.e. both
deflection l i m i t a n d stress l i m i t a r e attaine d simultaneously.

REFERENCES

B a r n a r d , R.E., 1957. D e s i g n a n d d e f l e c t i o n c o n t r o l o f b u r i e d steel


p i p e s u p p o r t i n g e a r t h l o a d s a n d l i v e l o a d s , Proc. A m . SOC. f o r
T e s t i n g M a t e r i a l s , 57.
C l R i A ( C o n s t r u c t i o n I n d u s t r y Research a n d I n f o r m a t i o n A s s o c i a t i o n ) ,
1978. D e s i g n a n d C o n s t r u c t i o n o f B u r i e d T h i n - d a l l P i p e s , Report
78, L o n d o n , 93 p p .
Jacobsen, S., 1974. B u c k l i n g o f c i r c u l a r r i n g s and c y l i n d r i c a l t u b e s
u n d e r e x t e r n a l p r e s s u r e , Water Power, 26 ( 1 2 ) .
Morley, A . , 1919. S t r e n g t h o f M a t e r i a l s , L o n g m a n s Green E. Co., pp
326-333.
S p a n g l e r , M.G., 1956. Stresses i n p r e s s u r e p i p e l i n e s and p r o t e c t i v e
c a s i n g p i p e s . Proc. Am. SOC. C i v i l E n g s . , 82(ST5) 1054.
Stephenson, D., 1973. S t i f f e n i n g r i n g s f o r p i p e s , P i p e s and P i p e l i n e s
I n t l . , 18 ( 4 ) .
Stephenson, D. ,1979. F l e x i b l e p i p e t h e o r y a p p l i e d to t h i n - w a l I nPVC
p i p i n g , P i p e s a n d P i p e l i n e I n t l . , 24 ( 6 ) , 9-17.
158

Timoshenko, S.P. and Woinowsky-Krieger, S., 1959. Theory of Plates


a n d Shells, 2nd Ed., McGraw H i l l , N.Y., Ch. 15.

LIST OF SYMBOLS

a - bs, o r area per u n i t length

A - cross sectional a r e a of r i n g

b -

C - distance from centre of w a l l to c e n t r o i d of w a l l section


d - p i p e diameter
d. - i n t e r n a l p i p e diameter
e x t e r n a l p i p e diameter
modulus of e l a s t i c i t y
passive resistance modulus of s i d e f i I I
permissible w a l l stress
F - wal I stress
bending stress
circumferential stress
bending stress under r i n g
r i n g stress, o r r a d i a l stress i n p i p e w a l l
circumferential wal I stress
design factor
r i n g height
depth of b a c k f i l l above pipe
I - moment of inertia
J - j o i n t factor
K - r a t i o of a c t i v e l a t e r a l pressure of f i l l to v e r t i c a l
pressure
c r i t i c a l r i n g spacing
bending moment
moment or deflection coefficient
i n t e r n a l pressure
pipe radius
distance from c e n t r o i d to extreme f i b r e of beam
r i n g spacing
wal I thickness
159

U - deflection time l a g factor


V - Poisson's r a t i o
W - vertical pressure, subscript a refers to arching, b to

bending a n d d to deflection
W - vertical load p e r u n i t length ( W = wd)
Y - r i n g thickness
z - bedding constant

A - def I ect ion


160

CHAPTER 9

SECONDARY STRESSES

STRESSES AT BRANCHES

If a hole i s cut in the w a l l of a pipe, the circumferential stress


in the wall will increase on each side of the hole. The tangential
stress on the horizontal axis at the side of a small elliptical hole
i n a p l a t e i s (Tirnoshenko a n d Goodier, 1951),

S' = S(l + 2 a/b) (9.1

where S is the uniform vertical stress applied to the plate, a is


the horizontal axis length and b is the vertical axis length. For
a circular hole a = b and the stress S ' i s consequently three times
the applied stress. In the case of a circular hole o r branch on a
pipe the circumferential stress on each side of the hole o r branch
may also increase to three times that in the p l a i n pipe. The stress
distribution beside a hole in a plate is given in Fig. 9.1. If the
pipe wall stress is at all significant it will be necessary to rein-
force the w a l l where a b r a n c h p i p e occurs. Reinforcing i s also desir-
able to minimize distortion. A common practice especially with low
pressure pipework i s to use s k i r t s at the b r a n c h connection to m i n i -
mize the entrance head loss. In such cases the size of hole cut in
the main pipe is larger than the bore o f the b r a n c h p i p e and r e i n -
forcing may be especially necessary. For small b r a n c h pipes at r i g h t
angles to a barrel a s u i t a b l e method of strengthening the p i p e con-
nection is to weld a collar onto the main pipe around the branch
pipe (Fig. 9.2). The w i d t h of c o l l a r may be c a l c u l a t e d f o r any p a r t i -
cular thickness by assuming the col t a r takes the circumferential
force which would otherwise have been taken across the diameter
of the piece cut out of the main pipe, i.e. 2t'w' = tD, where t'
is the collar thickness, w' the width, D the branch p i p e outside
diameter and t the thickness of the wall of the main pipe. This
is assuming that the permissible stress in the c o l l a r and main pipe
walls are the same, and that the collar width is small compared
w i t h the b r a n c h p i p e diameter.
161

s
t t f t t t t t t

Fig. 9.1 Stress d i s t r i b u t i o n beside a hole i n a p l a t e under


a uniform stress S.

Fig. 9.2 Collar reinforcing at a branch pipe

7-
lp-Sornetimes used
LI

Fig. 9.3 L a t e r a l w i t h e x t e r n a l crotch plates


162

For l a r g e r diameter b r a n c h pipes and branches at v a r i o u s angles


to the m a i n pipe, crotch p l a t e s a r e p r e f e r a b l e to c o l l a r s ( B l a i r , 1946).

Crotch P l a t e s
Various types of s t i f f e n i n g p l a t e s have been proposed f o r r e i n f o r c -
ing laterals. The Swiss f i r m Sulzer Bros. (1941) pioneered the design
of crotch plates, and a design a i d i n the form of a nomograph was
l a t e r p u b l i s h e d b y Swanson et a l (1955). The design was f o r e x t e r n a l
plates which would be welded i n t o the c r o t c h a t the j u n c t i o n of the
pipes. The nomograph was p r e p a r e d b y a n a l y s i n g a l a r g e number of
cases. Trial sections were selected and r e v i s e d u n t i l forces balanced
and deflections were compatible. Only r e c t a n g u l a r sections f o r crotch
p l a t e s were considered although T sections may have g r e a t e r r i g i d i t y .
Figs. 9.3 to 9.7 a r e design c h a r t s f o r e x t e r n a l crotch p l a t e s based
on Swanson's results, but expressed in dimensionless parameters f o r
use in any system of units. The design c h a r t s a r e f o r s i n g l e c o l l a r
plates (normally for small diameter laterals) and for double crotch
p l a t e s ( f o r l a t e r a l diameter n e a r l y equal to the m a i n p i p e diameter).
The method of design u s i n g the c h a r t s i s as follows:
C a l c u l a t e Pd / f t knowing t h e maximum i n t e r n a l pressure P, the
1
main p i p e diameter dl, the permissible p l a t e stress f a n d an
assumed p l a t e thickness t.
Read off from Fig. 9.5 the crotch plate width expressed in
terms of the main p i p e diameter dl. bl and cl a r e the crotch
plate widths for a 90" lateral with the same diameter as the
main pipe.
If the branch is at an a n g l e to the b a r r e l , or the diameters
are unequal, read Kb and K from Fig. 9.6 and correct the
crotch p l a t e w i d t h b y m u l t i p l y i n g bl b y K b a n d c1 b y Kc.
The plate widths are now b=d (bl/d,)
1
K
b
a n d c=dl(c,/dl)K .
Read the p l a t e top depth r a t i o a/dl corresponding to b/dl or
c/dl and the angle of the lateral from Fig. 9.7 and hence
c a l c u l a t e a, the top depth of each plate.
Check whether either plate width i s g r e a t e r than 30 times the
plate thickness. If so, the plate is liable to buckle, so
increase the plate thickness and try again. The shape and
weight of p l a t e s could be optimized b y t r i a l .
163

Fig. 9.4 E x t e r n a l crotch p l a t e s

The shape of the p l a t e should then be drawn for fabrication by


developing the line of intersection of the pipes. Single plates may
often be bent i n t o shape whereas double p l a t e s w i l l be f i l l e t welded
together at the intersection point, and if t h i n enough, may a l s o be
bent to follow the intersection of the b r a n c h a n d the b a r r e l .

b /d,
8
c, dtJ,

1.0

0.5

0 0.5 20 pd 1.5
-1
f t

Fig. 9.5 Crotch p l a t e side w i d t h


164

Fig. 9.6 Effect of l a t e r a l diameter and angle on c r o t c h p l a t e s i z e .


165

For large installations a t h i r d p l a t e may be r e q u i r e d , placed at


right angles to the a x i s of the main pipe (see F i g . 9.3). Although
design c h a r t s f o r 3-plate designs a r e not a v a i l a b l e , Swanson suggests
merely a d d i n g a t h i r d p l a t e t o a 2-plate design. Discretion should be
used i n r e d u c i n g the thicknesses o r widths of the two c r o t c h p l a t e s i n
such circumstances. The deflections are reduced considerably by
a d d i n g a t h i r d p l a t e though.
Whether crotch p l a t e s should be used at all i s also a matter of
judgement. The author advises using crotch plates if the internal
design pressure P i s g r e a t e r than y t a"/135Gd2 where G i s the factor
1
of safety, t, is the wall thickness of the main p i p e , d2 i s the
diameter of the b r a n c h p i p e a n d a i s the a n g l e between the a x i s of
the b r a n c h p i p e a n d b a r r e l . y i s y i e l d stress.

Internal Bracing (Stephenson, 1971)


Certain disadvantages in external crotch plates prompted Escher
Wyss L t d . (Suss a n d Hassan, 1957) to i n v e s t i g a t e i n t e r n a l b r a c i n g for
pipe 'Y's. External crotch plates act in bending, consequently use
material inefficiently and require more steel than internal crotch
p l a t e s which could be i n p u r e tension.
An internal web may also act as a guide vane. T h i s effect, com-
b i n e d w i t h a g r a d u a l taper of the pipes a t the j u n c t i o n , reduces head
losses for some flow configurations. For the case of water flowing
through two pipes to a confluence, the head loss w i t h internal webs
is considerably less than with a standard 'T' junction. The
arrangement of the b r a n c h p i p e j o i n i n g the main p i p e a t a skew angle
with both pipes gradually flared at the confluence, together with
i n t e r n a l b r a c i n g f o r the acute a n g l e (see F i g . 9.8) i s recommended f o r
pcmping station delivery pipes. For water f l o w i n g from a main p i p e
into two branch pipes, such as pumping station suction pipes and
penstocks f o r hydro-electric stat ions, h y d r a u l i c model tests should be
performed to determine the best angles of divergence a n d taper, and
the arrangement of the i n t e r n a l b r a c i n g .
The o v e r a l l weight of steel for internal bracing is less than f o r
external c r o t c h plates. The compact arrangement f a c i l i t a t e s transport
a n d reduces excavation costs.
Unless the pipes are designed with cone-shaped flares at the
confluence, internal bracing will p r o t r u d e i n t o the flow paths, caus-
ing increased velocities and higher head losses than with external
crotch p l a t e s o n l y . On the other h a n d i f the expansion i s too a b r u p t
there w i l l be a h i g h head loss. The head loss coefficient f o r conical
diffusers increases rapidly with the a n g l e of flare once the angle
between the w a i l a n d a x i s exceeds about 7; degrees (Rouse, 1961). A
s u i t a b l e compromise f o r the angles of f l a r e f o r a 45-degree confluence
appears to be 74 degrees to the axis for the main pipe and 15
degrees to the a x i s f o r t h e b r a n c h p i p e .
The use of internal bracing confines the angle between the
branch and main pipes to a practical range. For confluence angles
approaching 90 degrees ( I T ' j u n c t i o n s ) the b r a c i n g web would obstruct
the flow in the main pipe. For very small angles of confluence the
hole cut in the main p i p e becomes l a r g e r a n d the b r a c i n g would be
impractically heavy. T h e design c h a r t s presented here a r e confined to
a 45-degree confluence, and for any other angle of approach the
branch pipe would have to be constructed with a bend immediately
before the f l a r e . Note that the use of i n t e r n a l b r a c i n g i s confined to
the acute angle, as internal bracing at the obtuse angle would
obstruct flow. The obtuse a n g l e must be strengthened with external
crotch plates.
To minimize head losses, the diameters of the main incoming p i p e ,
the b r a n c h p i p e a n d the confluent p i p e should be such t h a t the velo-
c i t i e s i n a l l pipes a r e approximately equal.
If all branches of the confluence are tangential to a sphere
whose centre l i e s a t the intersection of t h e i r axes, the intersection of
the surfaces w i l l be i n planes. T h i s means t h a t the b r a c i n g webs a r e
flat. There may be a k i n k i n the obtuse a n g l e intersection l i n e where
the main pipe flare meets the confluent pipe (see Fig. 9.8). To
eliminate the resulting problems of torsion, plate 'B' may be
continued beyond the point where it should bend. The pipe walls
would be strong enough to t r a n s f e r the loads to t h i s p l a t e , and the
r e s u l t i n g arrangement simp1 i f i e s f a b r i c a t i o n .
To determine the p r o f i l e s of the lines of intersection of the cone
surfaces, equations were set up r e l a t i n g the cone r a d i i a t any point
and the horizontal distance along the i n t e r s e c t i n g p l a n e to the dis-
tance from the intersection of the axes along the main a x i s . Using the
fact that the elevations of the surface of both cones a r e equal on the
intersection I ine, the coordinates of the intersection I ine were derived.
The a n g l e of the intersection p l a n e to the main a x i s , 8 , i s given by
the equation t a n B = (cos Q - cos a cos e ) / s i n a cos e , where a i s the
confluence angle, @ is the angle of the branch cone surface to its
a x i s and e i s the a n g l e of the main cone surface to i t s a x i s .
The intersection profiles w e r e obtained b y computer a n a l y s i s pro-
ceeding in steps along the i n t e r s e c t i n g plane. T h i s was done simul-
taneously w i t h the stress a n a l y s i s .
Although it i s possible to study the stresses in the bracing for
simple pipe 'Y's analytically, the general case i s v e r y complex a n d
must be solved by computer, proceeding i n small increments a l o n g the
intersection line. In fact previous studies have relied on model
studies to o b t a i n the stresses a t the intersecting plane.
The analysis was based on membrane theory i.e. the pipe walls
can take no b e n d i n g stress. The c i r c u m f e r e n t i a l w a l l stress in a cone
is PR/(cos 0 ) per u n i t w i d t h a n d the l o n g i t u d i n a l w a l l stress PR/(2
cos e ) per unit width where P i s the l i q u i d pressure a n d R i s the
r a d i u s of the cone p e r p e n d i c u l a r to the a x i s .
The forces due to the cone wall stresses on an element of the
intersection I i n e were resolved i n t o three components, p e r p e n d i c u l a r to
the intersection plane, v e r t i c a l l y a n d h o r i z o n t a l l y in the plane.
The results indicated that the perpendicular forces on the inter-
section p l a n e were i n a l l cases zero. The p o s i t i o n and d i r e c t i o n of the
resultant force in the plane, to the s i d e of four points along the
intersection line, was plotted g r a p h i c a l l y . It was observed that the
l i n e of action of the r e s u l t a n t was constantly i n s i d e the intersection
p l a n e a n d a c t i n g i n tension on the i n t e r n a l b r a c i n g p l a t e .
The most economical internal b r a c i n g p l a t e would be one i n pure
tension. T h i s s t a t e could be assured i f the r e s u l t a n t force f e l l on the
centre line of the plate. Thus the width of the widest internal
bracing plates indicated in Figs. 9.8 to 9.10 at the centre, or
crown, is twice the distance from the crown to the r e s u l t a n t force.
Hence the thickness of the plate may be derived in terms of the
permissible plate stress, and the width anywhere else calculated
knowing the p l a t e thickness and position and magnitude of the r e s u l -
tant force on the intersecting l i n e of the cones. I n some cases i t w i l l
be found that the most economical bracing plate thickness is
impractically thin, in which case a thicker p l a t e may be selected,
and the corresponding shape interpolated from the c h a r t s . The n a r -
rower plates are i n combined bending and tension and consequently
contain more steel than the p l a t e in p u r e tension. The w i d t h of the
plate perpendicular to the resultant force at various points was
obtained from the equation for extreme f i b r e stress, f = 6F (u-w/Z)
/tw2 + F/tw, where f i s permissible p l a t e tensile stress, F i s magni-
tude of resultant force which acts a t a distance u from the p o i n t in
question on the intersection line, w is plate width and t is plate
thickness. A similar analysis was performed for the e x t e r n a l crotch
p l a t e s b r a c i n g the obtuse angle.
The charts, Figs. 9.8 to 9.10, were compiled and f a c i l i t a t e the
design of crotch plates for a 45 degree ' Y ' and various angles of
flare to the main pipe and branch pipe. The most suitable flare
angles should be selected from h y d r a u l i c considerations. Model tests to
determine head losses could be conducted i f warranted.
169

I e
Lc=!23 tcn-
..L 2 2
6; /
/
j G.03303 \ 0
"
-.---

WITH R E C O M M E N D E D A N G L E S

b TOP H A L F OF EXTERNALCROTCH PLATE B

c TOP HALF OF INTERNAL C R O T C H ~ P L ,TE 4

F i g . 9.8 Crotch p l a t e s f o r 7io/15O confluence.


170

b TOP H A L F OF EXTERNAL CROTCH PLATE 6

c TOP HALF OF INTERNAL CROTCH PLATE A

Fig. 9.9 Crotch plates f o r p l a i n p i p e confluence


171

OL 0.2 0 02 0.L X / D 0.6

b TOP HALF O F EXTERNAL CROTCH PLATE B

c TOP HALF OF INTERNAL CROTCH PLATE A

Fig. 9.10 Crotch p l a t e s f o r plain pipe with 15" taper


confluence pipe.
172

The design procedure f o r the crotch p l a t e s would then be to select


a trial plate width, say the maximum indicated, and compute the
p l a t e thickness, t, from the corresponding formula. To ensure that the
crotch plate for the obtuse angle is rigid enough to withstand
buck1 i n g under bending compression, the p l a t e thickness should not be
less than w/30. The plate thicknesses should also p r e f e r a b l y not be
less than the wall thickness of the main p i p e , to ensure d u r a b i l i t y

a g a i n s t wear. After selecting the p l a t e thicknesses, the corresponding


p l a t e shapes a r e interpolated from the a p p l i c a b l e c h a r t .

STRESSES AT BENDS

The wall stresses in a fabricated or cast iron bend are higher


than those i n a p l a i n p i p e w i t h the same w a l l thickness. Longitudinal
stresses are induced as the bend tends to straighten out, local
bending stresses are caused by anchor blocks thrusting against the
outside wall, and circumferential stresses on the inside of the bend
a r e m a g n i f i e d because the l e n g t h of w a l l on the i n s i d e i s less than
on the outside of the bend. The increase in c i r c u m f e r e n t i a l stress is
normally more severe than the increase i n l o n g i t u d i n a l stress a n d it
is safe to design only for the worst circumferential stress on the
inside of the bend. The wall thickness on the outside of the bend
could a c t u a l l y be reduced i f account i s taken o n l y of circumferential
stress, but it i s safe to keep the same thickness as on the i n s i d e of
the b l o c k . Circumferential stresses a r e c a l c u l a t e d below.
Using membrane theory i.e. assuming the p i p e w a l l takes no bend-
i n g stress, the c i r c u m f e r e n t i a l force on a length of w a l l on the inside
of the bend i s equated to the t h r u s t due to the i n t e r n a l pressure on
the i n s i d e segment of p i p e shown shaded i n F i g . 9.11,

ft (R-D/2) 0 = P(R-D/4) 0D/2

f = -PD (R-D/4)
(9.2)
2t (R-DJ2)
i.e. the normal w a l l thickness at the bend should be increased i n the
ratio (R-D/4) / (R-D/2) where R i s the r a d i u s of the bend a n d D is
the p i p e diameter.
173

Fig. 9.11 Stress concentration a t a bend

In addition to the so-called secondary stress at bends, junctions


etc., discussed above, there a r e sometimes t e r t i a r y stresses in pipe-
work caused by relative movements, elastic strains or temperature
variations in piping systems. These stresses a r e r e a l a n d should be
designed for in exposed chemical and plant pipework or at large
pipeline interconnections if no movement joints or thrust anchors a r e
installed. A pipe system can be treated as a s t r u c t u r a l system w i t h
axial and lateral thrusts and movements. The system could be
analysed using structural design techniques such as moment-distri-
bution, slope-deflection or f i n i t e element methods, many o f which a r e
available in the form of standard computer programs. (See also
Crocker a n d K i n g , 1967).

THE P I P E AS A BEAM

Longit udi na I Bend ing


Pipes should normally be designed to r e s i s t some bending i n the
longitudinal direction even if they are to be buried. Unevenness o r
settlement of the bedding could cause sections to c a n t i l e v e r o r span
between supports, Three possible bending patterns together with the
174

corresponding critical b e n d i n g moments a r e i n d i c a t e d i n F i g . 9.12 .


The maximum span which a simply supported pipe could accom-
modate i s c a l c u l a t e d below. The m a x i m u m f i b r e s t r e s s i s

Fb = M/Z (9.3)

where M is t h e b e n d i n g moment and Z is t h e s e c t i o n m o d u l u s . For a


p i p e whose w a l l thickness, t, is small in comparison w i t h the i n t e r n a l
d i a m e t e r D,

Z = nD2 t / 4 (9.4)

So if M = WL2/8 (9.5)

Fb = W L Z / ( 2 n D 2t ) (9.6)

o r L2 = 2nD2 t Fb/W

and if a pipe of specific weight y is conveying water at a specific


weight yw

L = J8DtFb/(ywD + 4ySt) (9.7)

Thus f o r D = 1 m, t = 0.012 m, F b = 100 N/mm2, - 10 000 N/m 3 ,


yw -
ys = 80 000 N/m then the maximum permissible simply supported
s p a n L = 26 m or 86 f t .
For low internal water pressures, b u c k l i n g of the w a l l s may also
b e a p r o b l e m and t h e p o s s i b i l i t y s h o u l d b e i n v e s t i g a t e d ( e s p e c i a l l y a t
t h e s u p p o r t s w h e r e t h e s h e a r s t r e s s is h i g h ) . See also P e a r s o n , 1977.

P i p e S t r e s s at Saddles
The m a x i m u m local stress in a continuous o r s i m p l y supported p i p e
supported in a s a d d l e i s (Roark, 1954):

F = (0.02 - 0.00012 ( B - 90°))(Q/t' )Loge (f) (9.8)

where Q i s the total saddle reaction. Note t h a t f o r a s i m p l y s u p p o r t e d


span, Q i s the r e a c t i o n d u e to two ends r e s t i n g o n the support, i.e.
due to the weight of pipe of length equal to the span, as for a
continuous pipe. If Q i s i n Newtons a n d t i n mm t h e n F i s i n N / m m 2 .
175

WlUNlT LENGTH WlUNlT LENGTH WlUNlT LENGTH

REACTION WL12 Wll2 WLI2 WLl2 0 WL

MAXIMUM
BENDING wL2/a
MOMENT.

DEFLECTION 5 x 2
3 z El

a. Simply
Supported b. Fixed Ends c. Cantilevered

F i g . 9.12 Beam moments and deflections

The a n g l e of bottom support, 6 , i s n o r m a l l y 90" or 120' for con-


crete or strap saddles. The stress is practically independent of the
width of the support provided it is small compared with the pipe
diameter. To t h i s local stress a t the saddle must be added the rnaxi-
mum l o n g i t u d i n a l bending stress a t the support f o r the p i p e a c t i n g as
a continuous beam, unless a j o i n t occurs a t the support.

R i n g Girders
Steel pipes may be laid above ground over ravines, marshy
ground etc. B u c k l i n g of t h e p i p e a n d stresses i n these circumstances
may be more severe than under the ground, as there i s no s i d e f i l l
support and the p i p e saddles cause stress concentrations. Rigid r i n g
g i r d e r s at each support a r e useful i n these circumstances. The rings
may be plain rectangular in cross section, or T o r H shaped, with
the back fixed all r o u n d the pipe. The local p i p e bending stress in
the vicinity of the ring may be evaluated using the ring theory
developed in Chapter 8. The longitudinal bending stress i n the p i p e
w a l l under a r i n g is

1.65 Pd
0.91 + h J%T/A 2h (9.9)
176

where A i s the cross sectional area of the r i n g . To t h i s stress must


be added any longitudinal stress in the pipe due to beam action
between supports.
T h e r i n g stress i n the r i n g g i r d e r i s

1 Pd
(9.10)
1 + 0.91A/(hj hd) 2 7

In addition there will be local stresses due to t h e method of fixing


the r i n g girder to the supports. Legs a r e securely fixed to the r i n g
g i r d e r a n d rest on a s l i d i n g or r o l l e r b e a r i n g on a p i e r . The weight
of the p i p e a n d contents i s t r a n s f e r r e d to the bottom h a l f of the r i n g
g i r d e r a n d thence through the legs to the p i e r .
A more comprehensive a n a l y s i s of the stresses i n r i n g girders is
presented in Crocker and King (1967). See also Cates (1950) and
Scharer ( 1933).

TEMPERATURE STRESSES

Pipes with high differences in temperature between inside and


outside will be subject to r a d i a l a n d c i r c u m f e r e n t i a l stresses due to
temperature differences. The circumferential and l o n g i t u d i n a l stresses
have their maximum value on the inner a n d outer surfaces, being
compressive on the i n n e r surface and tensile on the outer surface if
the temperature on the inner wall is above t h a t on the outer wall,
and the heat flow i s steady. I f the w a l l is thin i n comparison w i t h
the p i p e diameter, the stresses may be approximated b y the f o l l o w i n g
equations:

a ET
on the i n n e r surface F = (9.11)
Fe = -20
aET
a n d on the outer surface F (9.12)
c = F P = -2(1-u)

where a i s the thermal coefficient o f expansion, E i s the modulus of


elasticity, T i s the temperature difference, v i s Poisson’s r a t i o , Fc i s
c i r c u m f e r e n t i a l stress and Fe i s l o n g i t u d i n a l stress. The r a d i a l stress
i s compressive a n d is always less than the maximum c i r c u m f e r e n t i a l
and l o n g i t u d i n a l stresses, b e i n g zero on both boundaries.
For normal c I imatic temperature differences the stresses are rela-
177

tively low. Thus f o r T = 10°C, E = 200 000 N/mmz, a = 12 x per


‘C, v = 0.3 f o r s t e e l p i p e , t h e n F m a x = 17 N/mmz.
A longitudinal stress may also be induced in a pipe restrained
longitudinally as a result of temperature change after instal lation.
The m a g n i t u d e o f t h i s stress i s aET (tensile if the temperature drops
after installation) (see Ch. 10 f o r an analysis including effects of
supports).

REFERENCES

B l a i r , J.S., 1946. R e i n f o r e c e m e n t o f b r a n c h p i p e s . E n g i n e e r i n g , 162.


Cates, W.H., 1950. D e s i g n s t a n d a r d s f o r l a r g e d i a m e t e r s t e e l w a t e r
p i p e , J. Am. W a t e r W o r k s Assn., 42.
C r o c k e r , 5. and K i n g , R.C., 1967. P i p i n g H a n d b o o k , 5 t h E d . , M c G r a w
H i l l , N.Y.
P e a r s o n , F.H., 1977. Beam b e h a v i o u r o f b u r i e d r i g i d p i p e l i n e s . P r o c .
Am. SOC. C i v i l E n g r s . , 108 ( E E 5 ) 767.
R o a r k , R.J., 1954. F o r m u l a s f o r s t r e s s and s t r a i n , McGraw H i l l , N.Y.
Rouse, H., 1961. E n g i n e e r i n g H y d r a u l i c s , W i l e y , N.Y., p . 418.
S c h a r e r , H., 1933. D e s i g n of l a r g e p i p e l i n e s . T r a n s . Am. SOC. C i v i l
Engs., 98 ( 1 0 ) .
Stephenson, D., 1971. I n t e r n a l b r a c i n g f o r p i p e c o n f l u e n c e s . T r a n s . ,
S.A. I n s t n . C i v i l E n g r s . , 13 (11).
Sulzer, 1941. P a t e n t e d s t i f f e n i n g c o l l a r s o n t h e b r a n c h e s o f high-
pressure p i p e l ines f o r hydroelectric power works. Sulzer Technical
R e v i e w , 2 ,lo.
Suss, A. and H a s s a n , D.R., 1957. R e d u c t i o n o f t h e w e i g h t and loss o f
energy in d i s t r i b u t i o n p i p e s f o r h y d r a u l i c p o w e r l a n t s . Escher
Wyss News, 30 (3) 25.
Swanson, H . S . , C h a p t o n , H.J., W i l k i n s o n , W.J., King, C.L. and N e l -
son, E.D., 1955. D e s i g n o f w y e b r a n c h e s f o r s t e e l p i p e s . J. A m .
W a t e r W o r k s A s s n . , 47 ( 6 ) 581.
Timoshenko, S. and G o o d i e r , J.N., 1951. T h e o r y o f E l a s t i c i t y , McGraw
H i l l , N.Y.

L I S T OF SYMBOLS

A - a r e a of r i n g g i r d e r
a,b - h o r i z o n t a l and v e r t i c a l a x e s o f an e l l i p s e
a,b,c - c r o t c h p l a t e t o p and s i d e w i d t h s
D or d- p i p e diameter
E - modulus of e l a s t i c i t y
f - permissible crotch p l a t e stress
F - r e s u l t a n t force on c r o t c h p l a t e a t a n y point, o r stress
G - factor o f safety
h - wal I thickness
178

span o r length
bending moment
water pressure (same u n i t s as f )
saddle reaction
radius
stress a p p l i e d to a p l a t e
crotch p l a t e thickness
temperature difference
distance from p i p e w a l l to F
w i d t h of crotch p l a t e at any p o i n t
load p e r u n i t l e n g t h
horizontal distance from crown along centre-I ine of
crotch p l a t e
Y - vertical distance above mid-plane or crown of crotch
plate
section modulus
a n g l e of confluence of b r a n c h p i p e w i t h main p i p e
coefficient of temperature expansion
a n g l e between crotch p l a t e and main p i p e a x i s , o r a n g l e
of support
specific weight
angle of flare, or taper, of main pipe (measured from
a x i s to cone)
a n g l e of f l a r e , o r taper, of b r a n c h p i p e
Poisson's r a t i o
CHAPTER 10

PIPES, FITTINGS AND APPURTENANCES

PIPE MATERIALS

Steel Pipe
Steel i s one of the most versatile materials for pipe walls, as it
i s d u c t i l e yet has a high tensile strength. It is relatively easy to
work, a n d the welded j o i n t f r e q u e n t l y used w i t h steel i s the strongest
type of j o i n t .
Steel grades used f o r p i p e s i n the U.K., and their corresponding
minimum y i e l d stresses a r e as follows:
BS 4360 Grade 40 : 230 N/mm2
Grade 50 355 N/mm2
Grade 55 : 450 N/mm2
US 572 Grade 4 2 290 N/mm2
Grade 50 345 N/mm2
Grade 60 : 414 N/mm2
Grade 65 : 448 N/mm2
The higher grades are preferred for high-pressure pipelines but
for low pressure p i p e l i n e s the lower grade steels a r e more economical,
they a r e easier to weld, a n d on account of the e x t r a wall thickness
they a r e more r e s i s t a n t to e x t e r n a l loads.
Small-bore (less than 450 mm) steel pipes are rolled without
seams, but larye-bore pipes are made from steel plate, bent and
welded either horizontally or spirally. Pipes a r e made in lengths up
to 10 m and more a n d j o i n t e d on site. Steel pipe is more expensive
than concrete, asbestos cement and plastic pipe for small bores 2nd
low pressures, and it requires coating and sheathing to prevent
corrosion.

Cast Iron Pipe


Cast iron is more corrosion-resistant than steel, but more
expensive and more r i g i d . I n fact actual sand c a s t i n g i s now r a r e l y
used f o r p l a i n p i p e , a l t h o u g h i t i s used f o r 'specials' such a s bends,
tapers and flanges. Plain p i p e s a r e n o r m a l l y formed of grey iron
180

or ductile iron by centrifugal spinning. Standard pressure classes


and dimensions of ductile iron pipes are specified in BS 4772 and
of grey i r o n p i p e s i n 854622.

Asbestos Cernent Pipe


Asbestos cement p i p e i s made of cement and asbestos f i b r e , which
is able to resist relatively high tensile stresses. Although asbestos
cement is relatively cheap, strong and corrosion-resistant, it is
susceptible to shock damage and cannot be used f o r 'specials'.
Joints a r e made w i t h a sleeve f i t t e d w i t h r u b b e r sealing r i n g s .

Concrete Pipe
Reinforced or prestressed concrete pipes a r e s u i t a b l e and econom-
ical for large diameters. They are able to resist external buckling
loads easier than steel on account of their extra wall thickness,
and are corrosion resistant. Their main weaknesses appear to be
in jointing and making later connections. There i s as yet not much
long-term experience w i t h prestressed concrete pipes.

Plastic Pipe

Technological advances in plastics in the 1960's were r a p i d and


this has been o f special interest in pipeline engineering. Unplasti-
cized p o l y v i n y l c h l o r i d e pipes a r e now used extensively f o r chemicals,
gases, drains, waterpipes, irrigation and sewer pipes i n the United
Kingdom and Europe. Polyethylene is also used to a limited extent
for small bore pipes although it is usually more expensive than
UPVC for large diameters. High density polyethylene has recently
been used w i t h some success f o r l a r g e diameter p i p e l i n e s .
Glass fibre and resin is also used to a l i m i t e d extent for pipe-
lines. It is comparatively expensive but used with an inert filler
it is suitable for some specialised applications as it i s corrosion-
resistant, r i g i d , l i g h t and strong.
Although pla sti c pipe has not yet stood the r i g o r o u s test of time
and is still not acceptable under many by-laws, i t s many advantages
a r e causing i t to g a i n r a p i d acceptance amongst engineers.

The working stress of plastic i s less 14 N/mrnz so p l a s t i c cannot


be used f o r l a r g e diameter high-pressure mains. I t i s also subject to
limited strength deterioration w i t h time a n d its flexibility can cause
buckting and collapse. It i s resistant to many corrosive f l u i d s a n d
can be used in difficult locations such as undersea o u t f a l l s a n d gas
pipes. The friction loss i s lower than f o r most m a t e r i a l s a n d i t can
be used beneficially as a l i n i n g to other types of pipes. It is less
subject to encrustation than other types of p i p e , accommodates g r e a t e r
ground movements a n d i s easier a n d l i g h t e r to l a y . On the other h a n d
it has a high coefficient of thermal expansion, so joints may be
loosened i f the p i p e contracts on cooling. I t i s susceptible to damage
and may distort or even collapse under load. Ribbed and stiffened
pipes a r e b e i n g developed to overcome these l i m i tations.
P l a s t i c p i p i n g i s n o r m a l l y made of thermoplastic which i s r e l a t i v e -
ly easy to extrude when heated. Polyethylene and PVC are thermo-
plastics. As PVC i s f l e x i b l e a n d e x h i b i t s creep, it i s generally used
in an u n p l a s t i c i z e d form (UPVC) f o r the manufacture of pipes. Poly-
ethylene is easier to work than UPVC and for this reason earlier
developments were with polyethylene. With improvement i n production
techniques UPVC i s now cheaper than polyethylene a n d i t holds c e r t a i n
other advantages. It is stronger and has a smaller coefficient of
thermal expansion than polyethylene, although recent ly-developed
high-density polyethylene (HDPE) has improved properties. Polyethy-
lene i s more d u c t i l e than UPVC a n d f o r t h i s reason i s often p r e f e r r e d
for gas p i p i n g . Rubber toughened p l a s t i c s such as ABS have a l s o been
developed for flexible pipework. Properties of plastics are tabulated
in the appendix. Because of the low elasticity of plastics, water
hammer pressures a r e less severe than f o r other m a t e r i a l s .
Plastic pipes may be j o i n t e d by solvents of fusion welding, al-
though these methods are difficult for large diameters. Screwed o r
factory fitted flanges or spigots and sockets with rubber rings are
also used.
The manufacture of moulded bends, tees a n d branches is difficult
i n UPVC a n d even more so i n HDPE so steel o r C.I. f i t t i n g s a r e fre-
quen t I y used.
The p r i c e of PVC pipes i n the U.K. dropped b y more than 30% in
the 1960's, whereas other p i p e m a t e r i a l s a r e c o n t i n u a l l y i n c r e a s i n g i n
182

cost. In the U.K., PVC pipes are cheaper than asbestos cement and
cast iron pipes for diameters less than 450 mm and pressures less
than 1 N/mmz. I t i s r a p i d l y r e p l a c i n g salt-glazed and asbestos-cement
pipes f o r sewers a n d low pressure systems.
At least 5700 m i l l i o n was spent i n 1972 on p l a s t i c p i p i n g through-
out the world. The figure is likely to r i s e considerably as plastic
pipes are used in ever increasing proportions and new discoveries
may yet r e v o l u t i o n i z e pipeline engineering (see also Boucher, 1948;
BVMA, 1964; P a u l , 1954).

LINE VALVES

There a r e many types of v a l v e s f o r use i n p i p e l i n e s , the choice of


which depends on the d u t y . The spacing of v a l v e s a n d the size w i l l
depend on economics. Normally v a l v e s a r e sized s l i g h t l y smaller than
the pipe diameter, and installed with a reducer on either side. In
waterworks practice i t i s p r e f e r a b l e to keep the s o f f i t of the v a l v e at
the same level as the soffit of the pipe, to prevent air being
trapped, whereas in sewerage a n d sol i d s t r a n s p o r t , the i n v e r t s should
be lined up. I n choosing the size, the cost of the v a l v e should be
weighed against the cost of the head loss through it, although in
certain circumstances it may be desirable to maintain the full pipe
bore ( t o prevent erosion o r b l o c k a g e ) .
Isolating valves are frequently installed at intervals of 1 to 5
km, the spacing being a function of economics and operating
problems. Sections of the p i p e l i n e may have to be isolated to r e p a i r
leaks and the volume of water which would h a v e to be d r a i n e d to
waste would be a f u n c t i o n of the spacing of i s o l a t i n g valves. Scour
valves are i n s t a l l e d at the bottom of each major d i p i n the p i p e l i n e
profile.
It i s sometimes a d v i s a b l e to i n s t a l I smal I-diameter by-pass valves
around in-line v a l v e s to equalize pressures across the gate a n d thus
facilitate opening ( w h i c h may be manual o r by means of an e l e c t r i c
o r mechanical a c t u a t o r ) .

Sluice Valves
Sluice v a l v e s , o r g a t e valves, a r e the normal type of v a l v e s used
183

for isolating or scouring. They seal well under high pressures a n d


when f u l l y open o f f e r l i t t l e resistance to f l u i d flow.
There are two types of spindles for raising the gate: A rising
spindle which is attached to the gate and does not r o t a t e w i t h the
handwheel, and a non-raising s p i n d l e which i s rotated i n a screwed
attachment i n the gate ( F i g . 10.1). The r i s i n g s p i n d l e type i s easy to
I u b r ica t e.
The gate may be para1 lel-sided or wedge-shaped. the wedge-gate
seals best b u t may be damaged b y g r i t . For low pressures r e s i l i e n t o r
gunmetal sealing faces may be used b u t f o r high pressures s t a i n l e s s
steel seals a r e preferred.
Despite sluice valves' simplicity and positive action, they are
sometimes troublesome to operate. They need a big force to unseat
them a g a i n s t a high unbalanced pressure, a n d l a r g e v a l v e s t a k e many
minutes to t u r n open o r closed. Some of the problems can be overcome
b y i n s t a l l i n g a v a l v e w i t h a smaller bore t h a n the p i p e l i n e diameter.

Fig. 10.1 Sluice v a l v e

Butterfly Valves ( F i g . 10.2)


Butterfly valves are cheaper than sluice valves for larger sizes
a n d occupy less space. The sealing i s sometimes not as e f f e c t i v e as
for sluice valves, especially at high pressures. They also offer a
fairly h i g h resistance to flow even i n the f u l l y open state, because
the thickness of the d i s c obstructs the flow even when i t i s rotated
184

90 degrees to the fully open position. Butterfly valves, as well as


s l u i c e valves, a r e not s u i t e d f o r operation in p a r t l y open positions as
the gates and seatings would erode rapidly. As both types require
high torques to open them against high pressure, they often have
geared handwheels o r power d r i v e n actuators.

Fig. 10.2 Butterfly valve

Globe Valves
Globe v a l v e s have a c i r c u l a r seal connected a x i a l l y to a v e r t i c a l
spindle and handwheel. The seating is a ring perpendicular to the
pipe axis. The flow changes d i r e c t i o n through 90 degrees twice thus
r e s u l t i n g i n h i g h head losses. The v a l v e s a r e n o r m a l l y used i n small-
bore pipework a n d as taps, a l t h o u g h a v a r i a t i o n i s used as a control
valve.

Needle and Control Valves ( F i g . 10.3)


Needle v a l v e s a r e more expensive than s l u i c e a n d b u t t e r f l y v a l v e s
but are w e l l suited for t h r o t t l i n g flow. They have a g r a d u a l throttl-
ing action as they close whereas sluice and butterfly valves offer
l i t t l e flow resistance u n t i l p r a c t i c a l l y shut and may suffer c a v i t a t i o n
damage. Needle valves may be used with counterbalance weights,
springs, accumulators or actuators to maintain constant pressure
conditions e i t h e r upstream o r downstream of the v a l v e , o r to m a i n t a i n
a constant flow. They a r e streamlined i n design a n d r e s i s t a n t to wear
even at high flow velocities. The method of sealing is to push an
axial needle o r spear-shaped cone i n t o a seat. There i s often a p i l o t
needle which operates f i r s t to balance the heads before opening.
A v a l v e w i t h a s i m i l a r h y d r a u l i c c h a r a c t e r i s t i c to the needle v a l v e
185

is the sleeve valve which is suitable for use discharging into the
atmosphere or an open b a y . They seal b y a sleeve which s l i d e s over
orifices around the tube and spread the flow i n a n umbrella-shaped
spray, thereby dissipating the energy. Needle a n d sleeve valves are
a l s o occasionally used as water hammer release v a l v e s when coupled
to an electric or hydraulic actuator (see Chapter 4). They require
dismantling for maintenance as the working parts are inside the
valve. The cone valve is a variation of the needle v a l v e but the
sealing cone rotates away from the pipe axis instead of being
withdrawn a x i a l l y .

Fig. 10.3 Needle v a l v e

Spherical Valves ( F i g - 10.4)


Spherical valves have a rotary p l u g with an axial hole t h r o u g h
it. When the v a l v e is fully opened there i s no resistance to flow as
the bore i s equal to that of the pipe. The v a l v e s close b y rotating
the sphere, a n d normally have a n offset action to unseat them before
r o t a t i n g the sphere into the open position.
186

Fig. 10.4 Spherical valve

Reflux Valves (Fig. 10.5)


Reflux, or non-return, or check valves as they a r e also known,
are used to stop flow automatically in the reverse direction. Under
normal flow conditions the gate i s k e p t open by the flow, and when
the flow stops, the horizontally h i n g e d g a t e closes b y g r a v i t y o r w i t h
the a i d o f s p r i ngs. A counterbalance can be f i t t e d to the gate s p i n d l e
to keep the gate fully open a t practically all flows, or it could be
used to assist in r a p i d c l o s i n g when the flow does stop. Springs are
a l s o sometimes u s e d to a s s i s t c l o s i n g . Larger reflux v a l v e s may h a v e
multi-gates, in which case the thickness of e a c h g a t e w i l l partially

obstruct the flow, and swabbing of the pipeline may be difficult.


Mounted in h o r i z o n t a l pipes, t h e g a t e s of some t y p e s o f r e f l u x v a l v e s
tend to flutter at low flows, and for this reason an offset hinge
w h i c h c l i c k s t h e g a t e open is sometimes used.

AIR VALVES (see a l s o Lescovich, 1972; Parmakian, 1950; Sweeton,


1943).

Two types of air release valves are normally used i n pipelines.


One is a small-orifice automatic air release v a l v e a n d the other i s a
large-orifice a i r vent v a l v e .

A i r Vent V a l v e s
When a p i p e l i n e i s f i l l e d , a i r could be trapped a t peaks along the
profile, thereby increasing head losses a n d r e d u c i n g the capacity of
187

the pipeline. Air vent valves are normally installed at peaks to


permit air in the pipe to escape when displaced by the f l u i d . They
also let air into the p i p e l i n e d u r i n g scouring o r when emptying the
pipeline. Without them vacuum may occur a t peaks a n d the p i p e could
collapse, or i t may not be possible to d r a i n the p i p e l i n e completely.
It is also u n d e s i r a b l e to have a i r pockets in the p i p e as they may
cause water hammer pressure fluctuations during operation of the
pipeline. T h e s e a l i n g element i s a buoyant b a l l which, when the p i p e
is full, i s seated a g a i n s t a n opening a t the top of the v a l v e . When
t h e pressure i n s i d e the p i p e f a l l s below the e x t e r n a l pressure the b a l l
drops thereby p e r m i t t i n g a i r to be drawn i n t o the pipe. When the p i p e
is being f i l l e d the valve will remain open u n t i l the water f i l l s the
pipe and l i f t s the b a l I a g a i n s t the seating. The v a l v e w i l l not open
a g a i n u n t i I the pressure fal I s below the e x t e r n a l pressure.

Fig. 10.5 Reflux v a l v e

The valve will tend to blow closed at high air velocities. It


should be r e a l i s e d t h a t a differential pressure of o n l y 0.1 N/mmz (13
psi) will cause sonic velocity of air through the valve (300m/s).
Pressure differences around 0.027 N/mrn2 ( 4 p s i ) h a v e been found to
close some air vent valves. The slamming closed may damage the
hollow b a l l seal.
In practice a diameter approximately l/lOth of the p i p e diameter
188

is used, although larger diameters and duplicate installations are


preferable. The v a l v e i s r e f e r r e d to b y the size of the i n l e t connec-
t i o n diameter, a n d the o r i f i c e diameter i s u s a l l y s l i g h t l y smaller.
The size of air vent valve w i l l depend on the r a t e of filling or
desired scouring rate of the pipe. The volume r a t e of flow of air
through an o r i f i c e i s approximately 40 times the flow of water f o r the
same pressure difference, but ful I vacuum pressure should not be
a1 lowed to develop.
A i r vent v a l v e s should be i n s t a l l e d a t peaks i n the p i p e l i n e , both
relative to the horizontal and relative to the hydraulic gradient.
Various possible hydraulic gradients including reverse gradients
during scouring, should b e considered. They are normally fitted in
Combination with an air release valve, as discussed in the next
sec t ion.
Equations for sizing air valves are derived in Chapter 5. As
a rule of thumb the orifice area in m2 should be Q/lOO where Q

is the r a t e of flow of air in m 3 / s at initial pressure. The r a t e of


flow of free air is the most difficult factor to determine though.
During fitting operations, the r a t e of evacuation of air through all
vent valves will equal the r a t e of filling the line. Care should be

taken as a l l the a i r i s evacuated as the flow of water w i l l suddenly


be stopped a n d water hammer i s possible.

A i r Release Valves
Air i s e n t r a i n e d in water in many ways; by vortices in the pump
suction r e s e r v o i r o r merely b y absorption a t exposed surfaces, at air
pockets or in the suction reservoir. The air may be released when
there i s a drop i n pressure, e i t h e r a l o n g a r i s i n g main o r where the
velocity i s increased through a r e s t r i c t i o n such as a p a r t i a l l y closed
valve. An increase i n temperature w i l l also cause a i r to be released
from solution. Small orifice a i r release valves are i n s t a l l e d on the
p i p e l i n e to bleed o f f the a i r which comes out of solution.
Smal I orifice air valves are designated by their i n l e t connection
size, usually 12 to 50 mm diameter. This has n o t h i n g to d o w i t h the
a i r release o r i f i c e size which may be from 1 to 10 mm diameter. The
larger the pressure in the pipeline, the smaller need be t h e o r i f i c e
size. The volume of a i r to be released w i l l be a f u n c t i o n of the a i r
entrained which is on the average 2% of the volume of water (at
189

atmospheric p r e s s u r e ) , (see Chapter 5 ) .


The small o r i f i c e release v a l v e s a r e sealed b y a f l o a t i n g b a l l , o r
needle which i s attached to a f l o a t . When a c e r t a i n amount of a i r has
accumulated i n the connection on top of the p i p e , the b a l l will drop
or the needle valve will open and release the air. Small orifice
release v a l v e s a r e often combined w i t h large orifice a i r vent valves
on a common connection on top of the pipe. The arrangement i s c a l l e d
a double air valve, (see Fig. 10.6). An isolating sluice valve is
normally f i t t e d between the p i p e a n d the a i r valves.

LAREE ORIFICE
AIR VENT VALVE
n
SMALL ORIFICE FIlTINQ

Fig. 10.6 Double a i r v a l v e

Double a i r valves should be installed at peaks in the pipeline,


b o t h w i t h respect to the h o r i z o n t a l a n d the maximum h y d r a u l i c g r a d -
ient. They should also be installed at the ends and intermediate
points along a l e n g t h of p i p e l i n e which is parallel to the h y d r a u l i c
grade line. It should be borne in mind that air may be dragged
a l o n g i n the d i r e c t i o n of flow i n the p i p e l i n e so may even accumulate
in sections falling slowly in relation to the hydraulic gradient.
Double a i r v a l v e s should be f i t t e d every 4 to 1 km along descending
sections, especially a t p o i n t s where the p i p e d i p s steeply.
A i r release v a l v e s should a l s o be i n s t a l l e d on a l l long ascending
190

lengths of pipeline where air is likely to be released from solution


due to the lowering of the pressure, again especially at p o i n t s of
decrease in g r a d i e n t . Other places a i r v a l v e s a r e r e q u i r e d a r e on the
discharge side of pumps and at high points on large valves and
upstream of o r i f i c e p l a t e s a n d r e d u c i n g tapers.

THRUST BLOCKS !See also Morrison, 1969)

Unbalanced thrust results at a bend in a pipeline due to two


act ions:
(1) The dynamic thrust due to the change i n d i r e c t i o n of flow. The
force in any direction is proportional to the change in
momentum in that d i r e c t i o n a n d is:

Fxl = oqAVx (10.1)

where F is the force, p is the fluid mass density, q i s the


flow and A Vx the reduction of the component of velocity i n the
direction x. T h i s force is, under normal conditions, negligible
compared with the force due to the internal pressure in the
pipe.
(2) The t h r u s t i n the d i r e c t i o n of each leg of the bend due to the
pressure i n the p i p e i s :

Fx2 = PA (10.2)

where p i s the i n t e r n a l pressure, A i s the cross sectional area


of p i p e flow a n d 0 i s the a n g l e of d e v i a t i o n o f the p i p e ( F i g .
10.7). The resultant outward thrust is the vector sum of the
forces i n both d i r e c t i o n s of the p i p e a x i s , and is

Fr = 2pA s i n 0 / 2 (10.3)

The unbalanced t h r u s t may be counteracted b y l o n g i t u d i n a l tension


in an all-welded pipeline, or by a concrete thrust block bearing
against the foundation m a t e r i a l . I n the case of a j o i n t e d p i p e l i n e the
size of the block may be c a l c u l a t e d u s i n g s o i l mechanics theory. In
addition to frictional resistance on the bottom of the thrust block
and the circumference of the pipeline, there is a lateral resistance
against the outer face of the p i p e a n d block. The maximum r e s i s t i n g
191

pressure a soi mass w i l l o f f e r i s termed the passive resistance and


i s (Capper and Cassie, 1969)

f =
1 + sin@ + 1 + sin<
P ~ s 1h - s i n @ 2cJ 1 - s i n @

where f i s the r e s i s t i n g pressure a t depth h, y s i s the s o i l d e n s i t y ,


P
@ is the effective i n t e r n a l a n g l e of f r i c t i o n of the soil a n d c i s i t s
cohesion. If the thrust block extends from the surface to a depth H
below the surface, a n d i t s l e n g t h i s L, the total r e s i s t i n g t h r u s t i s

P =
HZ
Ys-2
'
1
+

-
sin6
sin6
+ kHLJ
1 + sin$
1 - sin
(10.5)

This maximum possible resistance will only be developed if the


thrust block i s a b l e to move i n t o the soil mass s l i g h t l y . The corres-
ponding maximum soil pressure is termed the passive pressure. The
minimum pressure which may occur on the t h r u s t block i s the a c t i v e
pressure, which may be developed if the thrust block were free to
y i e l d away from the s o i l mass.
This i s

fa = ysh 1
I- sin4 -
+ s i n @ "1
$-si+ nmsin4
(10.6)

which i s considerably less than the passive pressure and w i l l o n l y be


developed i f the force on which i t i s a c t i n g i s free to move away from
the soi I e x e r t i n g the pressure.

Fig. 10.7 T h r u s t a t a bend


192

Fig. 10.6 Thrust block

I n p r a c t i c e the pressure which may be r e l i e d upon i s s l i g h t l y less


than the p a s s i v e pressure, as f u l l movement of the p i p e i s usually
not permissible. A factor of safety of at least 2 should be used w i t h
the passive resistance formula. The i n w a r d movement of the top of a
thrust block to activate full passive pressure varies from H/20 for
soft c l a y s to H/200 f o r dense sand, (CP 1972), where H i s the depth
of the excavation. These movements are seldom permissible, and the
"at rest" coefficients a r e p r o b a b l y more r e a l i s t i c . T h e "at rest" r a t i o
of horizontal soil stress to vertical soil stress varies from 0.4 for
loose sands to 0.7 for very p l a sti c clays. An impact factor should be
used in a l l o w i n g for water hammer pressures. T y p i c a l values of a n g l e
of internal friction $ a n d cohesion c for v a r i o u s soils a r e indicated
i n Table 10.1.
193

TABLE 10.1 Strength of Soils - t y p i c a l values

Type of soil Angle of f r i c t i o n 4 Cohesion


N/rnmz

Gravel 35 O 0
Sand 30 0
Silt 28 O 0.007
Dense c l a y 5" 0.035
Soft s a t u r a t e d c l a y 0 0.15

Exarnp I e
Calculate the size of a thrust block to r e s i s t the u n b a l a n c e d
pressures a t a 45O b e n d i n a 1 000 rnrn d i a m e t e r p i p e o p e r a t i n g a t
a p r e s s u r e o f 2.0 N/rnrnz. L e n g t h of 1 p i p e l e n g t h = 10 r n , depth =
1 m, @ = 30°, C = 0.005 N/mmz. L a t e r a l f o r c e F = 2 x 2.0 3 x -If
4
s i n 2 2 i 0 = 1.20 MN.
T r y a t h r u s t b l o c k 3 x 3 ~ 3rn:
3
Weight of t h r u s t b l o c k 3 x 3 ~ 3 ~200x9.8/10
2 = 600kN
Weight of 10 rn l e n g t h of p i p e = 40kN
Weight of w a t e r i n p i p e 0.785x1zx10x9.8x1 OOO/lO' = 80kN
Weight o f s o i l a b o v e p i p e l x l x 7 x l 8O0x9.8/1O3 = 130kN
Total weight = 850kN
F r i c t iona I resistance 0 . 3 ~ 8 5 0 = 250kN
L a t e r a l r e s i s t a n c e o f soi I a g a i n s t b l o c k :

1 + sin$
~- - 1 + 0.5 =
1 - sin$ 1 - 0.5
32
F = 18 0 0 0 ~ ~ ~ 3O3
~+32/ ~1 5 ~' v3 3~ 3 = 889kN
L a t e r a l resistance of soil a g a i n s t p r o j e c t i o n of p i p e :
(18 000x1 . 5 x 3 / 1 0 3 x 2 x 5 j 3 ) 1x6.5 = 640kN
Total l a t e r a l resistance = 1770kN

F a c t o r o f s a f e t y 1.77/1.20 = 1.48

A t h r u s t b l o c k s h o u l d b e d e s i g n e d so t h a t t h e l i n e o f a c t i o n o f the
resultant of the resisting forces coincides w i t h the l i n e of t h r u s t of
the pipe. This w i l l prevent overturning or u n b a l a n c e d stresses. This
may best be done g r a p h i c a l l y or by taking moments a b o u t the centre
of the p i p e .
194

Thrust blocks are needed not only at changes in vertical or


horizontal alignment of the pipeline, b u t also a t f i t t i n g s which may
not be able to transmit longitudinal forces, such as flexible
coupl ings. V i k i n g Johnson o r s i m i l a r coupl i n g s a r e frequently i n s t a l led
on one side of a valve i n a v a l v e chamber, to f a c i l i t a t e i n s t a l l a t i o n
and removal of the valve. The opposite wall of the valve chamber
would thus h a v e to be designed as a t h r u s t block.

-li;

Fig 10.9 Segment geometry s t a n d a r d med i urn-rad i us 900 f a b r i ca t ed


bend.

FORCES INDUCED BY SUPPORTS

The i n t e r a c t i o n of a p i p e and i t s supports i s a complex subject.


There are many secondary and tertiary forces and displacements
involved. The effects are general ly greatest in pipes supported
above ground, on plinths or by other isolated supports. There is
a tendency of the pipe to expand or contract longitudinally and
circumferent i a l l y under the effects of temperature change, differen-
tial temperatures, secondary strains due to the Poisson r a t i o effect,
195

and partial or total f i x i t y of supports which impose forces on the


pipe vertically, longitudinally and laterally. iio support is
completely r i g i d a n d the more f l e x i b l e the support the less force or
reaction i t imposes on the pipe.
The stresses on the p i p e may be loosely c l a s s i f i e d as follows:
Primary; circumferential due to internal pressure across diameter;
longitudinal due to internal pressure across the cross section.
Secondary; bending a t elbows, branches, changes i n section a n d
across supports. Temperature effects. Poisson ratio counter-
strains at right angle to primary strain. Due to reaction of
supports.
Tertiary; l a t e r a l deflection due to l o n g i t u d i n a l expansion.
Al though the secondary and tertiary stresses are often not
accounted f o r i n p i p e design, they can be very s i g n i f i c a n t .

Longitudinal Stress
If the end of a p i p e i s b l a n k e d o f f a n d u n h e l d the l o n g i t u d i n a l
stress i n the p i p e w a l l is

i .e. 4 the circumferential stress.

The Poisson ratio effects tries to reduce the length by vLpd/2tE


and if the ends are held the wall tensile stress becomes Ee =
Wpd/2t and corresponding force is Vp-irdz/2 where E is the elastic
modulus, t the wall thickness, e the strain and w is Poisson's
r a t i o (about 0.3 for s t e e l ) .

Temperature stresses
A pipe w i l l expand or contract by aAT where a i s the coefficient
of expansion a n d LT the temperature change i n time.
A pipe restrained longitudinally will undergo a stress aEAT
-6
where the coefficient of expansion for steel, a , is 12 x 10 per
"C a n d for p l a s t i c , 50 x
If there is a temperature gradient across the pipe wall a
Circumferential stress i s created equal to +_ aEAT/2(1-v).
196

Forces at Bends
The force due to water pressure along each a x i s at a p i p e bend

is PA. The net resultant force acting outwards i s 2pA sin (8/2)
where 8 is the bend angle. I f the r e s u l t a n t force i s not countered
with a thrust block it creates a longitudinal force in the pipes
equal to 2pA s i n 2 ( 8 / 2 ) .

Lateral Movement
A straight pipe of length x fixed at both ends will buckle
outwards a relatively big amount under expansion due to temper-
a t u r e etc., b y an amount

dy (10.8)

since dy2 + (x/2)* = (x/2 + d ~ / 2 ) ~

dy =
-6
e.9. x = loom, d x = xaAT = 100 x 12 x 10 x 20 = 0.02m

dy = 400 x 0.02/2 = lm.

I f the pipe is restrained laterally it could try to buckle


between supports or twist supports or exert a lateral force equal
to that required to prevent the buckling. For example, the deflec-
t i o n of a simple supported ( l a t e r a l l y ) p i p e beam, dy, under a force
F is
dy = - FL3/EI
48

= 1 f o r above example.
1
:. F = 1 x 210 x lo9 2
8
13x.01/(1003x-~
48
= 40 kN
The l a t e r a l deflection is referred to as snaking and can cause
pipes to fa1 I o f f the side of supports, or bend the supports.
A pipe fixed at both ends with an elbow will bend outwards
under expansion by an amount:
dy = d x / ( y / a + y/b)
where dx is the net expansion in length of pipe with length
( a + b ) = c due to temperature, pressure etc.
197

Forces on Supports
The forces on the support blocks or p i l l a r s or hangers w i l l be
equal and opposite to the forces exerted by the supports on the
pipe.
Free to Slide: The longitudinal force of a freely sliding or highly
h e l d p i p e equals the f r i c t i o n force on the block. This i s pW, where

IJ is the f r i c t i o n coefficient and W i s the normal force ( w e i g h t ) of


the pipe on the block. There is also a lateral force to resist
snaking .
Fixed to Block: The resulting force is the smaller of i ) maximum
force which can be imposed on block b y p i p e e.g. f r i c ion between
pipe and block, or (ii) sliding resistance of block, or ( i i i ) resist-
ance of support a t the other end, or ( i v ) maximum net total longi-
tudinal force in pipe due to temperature change, Poisson ratio
effect a n d bends.
There i s thus a d i s t i n c t i o n between the f r e e to s l i d e case a n d the
fixed case in that ( i ) or (ii) will apply to a "free to s l i d e " pipe
a n d case ( i i i ) a n d ( i v ) w i l l g e n e r a l l y a p p l y to a p i p e r i g i d l y f i x e d
to the support.
Thus i f one end of a p i p e i s free (see F i g . 10.9) then there w i l l
obviously be no force at that end, a n d beyond the f i r s t support
the force i s computed as above, a n d beyond the second support also
as above etc. F o r the f i r s t few lengths case ( i ) or ( i i ) may apply
a n d then case ( i i i ) or ( i v ) w i l l take over as one of these is less.
For case ( i v ) the "force a t the other end" may i n fact be that due
to other support blocks.

Unba Ianced Forces


Generally it can be expected that forces in welded pipework
balance out. That i s forces i n one direction a t one end a r e equal
a n d opposite to those at the other end. Along the length of the p i p e
however there may be bending moments. For instance a t elbows or
bends there a r e net forces of l i q u i d on the p i p e w a l l inside a c t i n g
to t r y to s t r a i g h t e n the bend.
198

During transient flow conditions there may however be large


unbalanced forces which could pull the pipe off supports. For
instance following closure of a vapour pocket in a pipe water
hammer pressure at one end of a line may force the pipe in one
direction. Although the pressure wave may travel along the pipe
length and eventually re-balance forces, there are momentary
unbalanced forces which can cause s i g n i f i c a n t movement s t a r t i n g at
bends.

Support
4 resistance
length x + dx

l a t e r a l movement
dy

Fig. 10.10 L a t e r a l movement due to temperature expansion

Temperature
total force loose c o u p l i n g
F1 + F2 + F3 5" E L T Movement

I
f r i c t i o n on
supports

Fig. 10.11 L o n g i t u d i n a l forces i n p i p e w a l l created b y f r i c t i o n


on supports

FLOW MEASUREMENT

Venturi Meters (Fig. 10.12)

The Venturi meter offers an accurate method of flow measurement


with minimum head loss. It is used mainly for large-diameter
pipelines. The v e n t u r i throat has a converging section and g r a d u a l l y
d i v e r g i n g section to minimize head losses.
Flow measurement is based on the Bernoulli equation which is
r e a r r a n g e d as
199

v22
-
29
--29V 1 2 - h , - h 2 = h (10.9)

Solving f o r discharge

Q = CdA2 ,/= (10.10)

(10.11 )

where d i s the diameter and C i s a velocity coefficient which allows


V
for flow separation at the throat a n d includes a f a c t o r f o r upstream
and downstream conditions in the pipe. A venturi meter should be
calibrated in place a n d should have a s t r a i g h t l e n g t h of a t least 5
pipe diameters f o r very smal I t h r o a t / p i p e diameter r a t i o s , increasing
to 10 o r even 20 p i p e diameters f o r l a r g e r throats, o r from branches
or double bends, is required (BS 1042). F u l l bore branches, p a r t l y
closed valves a n d f i t t i n g s which impart s w i r l i n g motion may need up
to 100 p i p e lengths downstream before a meter. The head loss in a
v e n t u r i meter i s o n l y approximately O.1V ‘/2g.
2

Fig. 10.12 Venturi meter

Nozz 1 es
A contracted form of v e n t u r i meter i s the nozzle meter, which has
a short bellmouth rounded i n l e t a n d a b r u p t expansion beyond i t . The
discharge equation is similar to that for the venturi meter but the
v e l o c i t y head recovery beyond i s very small.

O r if ices
An o r i f i c e consisting of a thin plate with a central ori f i c e i s a
200

popular method of measuring flow in large pipes. There is an


appreciable orifice contraction coefficient which may be as low as
0.61. The discharge formula i s s i m i l a r to Equ. 10.10 where
-

The velocity head recovery is low but the orifice has the
a d v a n t a g e that it i s short a n d can be i n s t a l l e d i n a short l e n g t h of
straight pipe. The lengths of straight p i p e r e q u i r e d on each side of
the o r i f i c e a r e s i m i l a r to those f o r v e n t u r i meters. The o r i f i c e should
be c a l i b r a t e d i n place f o r accurate results.

Fig. 10.13 O r i f i c e meter

Bend Meters
A g r e a t e r pressure i s exerted on the outside of a bend i n a p i p e
than on the inside, due to c e n t r i f u g a l force. The pressure difference
can be used to measure the flow. The meter should be c a l i b r a t e d in

place, b u t the discharge coefficient f o r E q u . 10.10 i s approximately

Cd = r/2d (10.13)

where r i s the r a d i u s of c u r v a t u r e of the centreline.

Mechanical Meters
Flow to water consumers is invariably measured by mechanical
meters which i n t e g r a t e the flow over a p e r i o d of time. A common type
has a r o t a t i n g t i l t e d disc w i t h a s p i n d l e which rotates a d i a l . Other
types have r o t a t i n g wheels, lobes o r propellers.
201

There is also a type of meter which reacts to the drag on a


deflected vane immersed i n the flow. The meter records f o r p a r t l y f u l l
conduits as well.
The rotometer consists of a calibrated vertical glass tube which
has a taper increasing in diameter upwards. A float is suspended
in the upward flow through the tube. The float positions itself so
that the d r a g on it ( w h i c h depends on the flow p a s t i t a n d the tube
diameter) e q u a l s the submerged weight.
The object may be e l t h e r a sphere o r a tapered shape w i t h vanes
to make i t r o t a t e a n d centre i t s e l f i n the tube.

Electromagnetic Induction
By creating a magnetic field around a pipe of non-conducting
material and ionizing the I i q u i d by i n s e r t i n g electrodes, an electro-
motive force is induced and can be measured. The method has the
advantage that there is no l o s s of head and a variety of liquids

including sewage can be measured by this means. A similar tech-


nique which also does not obstruct the flow and i s based on sonic
velocity of an impressed shock wave, i s also being developed. The
difference between upstream and downstream sonic velocity is
measured.

Mass and Vol ume Measurement


The most accurate methods of measurement of flow are by mass
or volumetrically. To measure the mass or weight of fluid flowing
over a certain time the flow i s diverted i n t o a weighing tank. The
volume flow may be measured from the volume filled in a certain
t ime.

TELEMETRY

Automatic data transmission a n d control i s often used on p i p e l i n e


systems. Telemetry systems a r e tending to replace manual d a t a hand-
ling which is more expensive and less reliable than automatic
methods. Cables may be l a i d simultaneously w i t h the p i p e l i n e .
Data which may be required include water levels in reservoirs,
pressures, flows, opening of valves, temperature, qua1 i t y or pro-
202

perties of the fluid, pump speed or power consumption. The infor-

mation is read by a standard instrument linked by a mechanical


device to a local d i a l or c h a r t , or to a coder a n d transmitter.
S i g n a l s may be sent out from a t r a n s m i t t e r i n the form of contin-
uous analogue waves or in a series of digits. Analogue messages
are usually less accurate (up to 2 percent accuracy) but also less
expensive than digital. The transmitter may repeat messages contin-
uously or be interrogated at regular intervals, termed scanning.
With digital systems data are converted to binary code electric
pulses by relays, magnetic cores or contacts. Every possible level
of the measurand i s represented by a d e f i n i t e combination of binary
coded numbers. Signals are transmitted in series and in groups
known as bits. The signals may be transmitted via telephone lines
or private multiplex cables. One pair of conductors is needed for
each bit, plus a common feed. A number of u n i t s of data may be
sent in r o t a t i o n v i a the same conductors.
Alternatively hydraulic or pneumatic signals may be conveyed
over short distances by pipe. Radio transmission becomes economic
f o r systems w i th long transmi ssion distances.
The transmission system will send the messages to a receiver
linked to a decoder or digital comparater. S i g n a l s can be decoded
from analogue to digital and vice versa. The decoder feeds the
information to a data bank or by means of a servo motor can dis-
play the readings on dials, charts or lights or sound alarms at
extreme values of the measurand. Data is frequently displayed on
a mimic diagram (a diagramatic picture) of the system with tights
or dials at the p o s i t i o n of the u n i t being observed. These diagrams
a r e mounted i n c e n t r a l control rooms.
The information may also be fed to a data storage system and
stored on p a p e r or magnetic tapes, discs or cards. The information
may be fed directly or at a later stage to a computer to be used
for decision-making or manipulat i n g the data. L a r g e computers can
be programmed to optimize a system, for instance b y selecting those
pumps which would result in minimum pumping cost, or turbines
which would match power demand. The computer could send signals
to close valves or operate hydraulic machinery or water hammer
r e l i e f mechanisms.
203

Mini digital computers are now the most popular form of control
of telemetered systems. They are sI ight ly cheaper than analogue
computers but considerably cheaper than larger digital computers.
They can be programmed to perform certain tasks or computat ions
automatically but cannot be used easily for other computations for
which they are not programmed. They can be increased in c a p a c i t y
or l i n k e d to l a r g e r computers as t h e need a r i s e s .

REFERENCES

Boucher, P . L . , 1948. Choosing v a l v e s , K i Imarnock.


B r i t i s h V a l v e M a n u f a c t u r e r s Assn. 1964. V a l v e s f o r the C o n t r o l of
Fluids.
BS 1042, Flow measurement, BSI, London.
C a p p e r , P.L. a n d Cassie, W.F., 1969. The M e c h a n i c s o f E n g i n e e r i n g
Soils, 5 t h E d . , Spon, London.
CP 1972. L a t e r a l S u p p o r t i n S u r f a c e E x c a v a t i o n s , S.A. Instn. C i v i l
Engs.
L e s c o v i c h , J.E., 1972. L o c a t i n g a n d s i z i n g a i r r e l e a s e v a l v e s . J. Am.
Water Works Assn., 64 ( 7 ) .
M o r r i s o n , E.B., 1969. Nomograph f o r the d e s i g n of t h r u s t b l o c k s , C i v i l
Engg., Am. SOC. C i v i l Eng.
P a u l , L . , 1954. Selection of v a l v e s f o r w a t e r s e r v i c e s . J. Am. Water
Works Assn., 46 ( 1 1 ) 1057.
P a r m a k i a n , J., 1950. A i r i n l e t v a l v e s f o r steel p i p e l i n e s . T r a n s . Am.
SOC. C i v i l Engs., 115 ( 4 3 8 ) .
Sweeten, A.E., 1943. A i r i n l e t v a l v e d e s i g n f o r p i p e l i n e s . E n g g .
News Record, 122 ( 3 7 ) .

L I S T OF SYMBOLS

A - cross sectional a r e a
C - soi I cohesion
- con t r a c t i o n c o e f f i c i e n t
cC
- d i sc h a r g e coef f i c i en t
‘d
- velocity coefficient
cV
d - i n t e r n a l diameter
e - strain
E - e l a s t i c modulus
f - stress
F - force
9 - gravitational acceleration
h o r H - d e p t h below s u r f a c e
f l u i d pressure
flow r a t e
r a d i u s of c u r v a t u r e
wal I thickness
temperature
velocity
normal force
length
direct ions

I a t e r a I d i s p I acement
temperat u r e c o e f f i c i e n t
soil density
f r i c t i o n coefficient
Poisson's r a t i o
s o i l mass d e n s i t y
a n g l e of i n t e r n a l f r i c t i o n of s o i l
205

CHAPTER 1 1

LAYING AND PROTECTION

SELECTING A ROUTE

The selection of a suitable route for a p i p e l i n e has an important


bearing on the capital cost and operating costs. A pipeline route

is selected from aerial photos, topographical and cadastral plans,


on-site inspections and any other data available on the terrain,
obstacles and local services. In selecting a route, the costs and
prcticability have to be considered. Care should be taken to ensure
the ground profile is below the hydraulic grade line. (Low-flow
conditions should be considered as well as peak rates, as the
hydraulic gradient i s flattest for low flows). If there were a peak
above the grade line between the input and discharge heads,
obviously pumps would have to be designed to pump over t h i s peak.
Peaks may also be p o i n t s of p o s s i b l e water column separation which
result i n water hammer overpressures. On the other h a n d the general

level of the pipeline r o u t e should be kept as near to the h y d r a u l i c


grade lines as possible to minimize pressures a n d consequently pipe
costs.
Once a preliminary route is selected, it is pegged out by a
surveyor. Pegs are put in along the centre-line at 10 to 100 m
regular intervals and at changes i n g r a d e a n d at horizontal deflec-
tions. Offset pegs, 2 to 5 m from the centre l i n e pegs a r e also put
in for use d u r i n g laying. The l e v e l s of pegs a r e observed a n d the
profile is then plotted in the drawing office, with ground levels,
bed levels and depth indicated at each peg. Holes may be augered
along the r o u t e to identify m a t e r i a l s which w i l l have to be excavated.
Strength tests may be done on s o i l samples to decide whether trench
s h o r i n g w i l l be necessary.
Underground and overhead services should be accurately located
at this stage or at least before excavation to avoid them. (Drains
and underground cables may have to be supported on bridges, or
a heading c o u l d be hand-excavated under them).
206

L A Y I NG AND TRENCH I NG

The type of p i p e to be used for any job will l a r g e l y be a matter


of economics, although the facilities for laying, life of the pipe
materials and other factors affect the decision. The lengths of
delivered pipes will depend on the type of pipe, the weight, the
rigidity, size of transport vehicles and method of jointing. Concrete
pipes are often s u p p l i e d in 2 to 4 m lengths, w h i l e steel pipes may

be over 10 or even 20 m long, and thin walled p l a s t i c p i p e s may


be supplied in rolls. Pipes should be handled carefully to avoid
damage to coatings or pipe or distortion. They should be stored
and transported on padded c r a d l e s or sandbags. Pipes may be low-
ered i n t o the trench by b e l t s l i n g s from mechanical booms or t r i p o d s .
For small diameter thick w a l l e d welded pipes, j o i n t i n g and wrapping
may be done at the side of the trench, a n d the p i p e then 'snaked'
i n t o the trench. The working width required for construction of pipe-
lines varies with the method of excavation and laying. If the
excavation and laying is done by hand, 3 m may be wide enough
for small diameter pipes. On the other hand for large mains where
excavation, laying and possibly even s i t e coating i s done mechanic-
ally, working widths up to 50 m may be required. Reserve widths
of 20 to 30 m a r e common, a n d allow f o r one or two a d d i t i o n a l pipe-
l i n e s at a l a t e r stage.
Once the trench i s excavated, sight rails may be set up across

the trench with the centre line of the pipe marked thereon. The
frames a r e set a d e f i n i t e height above the bed a n d the pipe l e v e l l e d
w i t h the assistance of these (BSCP, 1970).
Pipes are normally laid with a minimum cover of 0.9 m to 1.2
m to avoid damage by superimposed loads. A minimum cover of 1
m is necessary in the U.K. to prevent frost penetration. The depth
may be reduced i n rocky o r steep ground to minimize costs i n which
case a concrete cover may be necessary. The width of trench is
normally 0.5 m to 0.8 m more than the pipe diameter, with local
widening and deepening for joints. (0.8 m is used for pipes over
300 mm diameter which are difficult to straddle while laying and
jointing).
207

The sides of the open trench will h a v e to be supported i n unco-

hesive or wet soils and for deep trenches, for safety reasons. The
shoring should be designed to r e s i s t the a c t i v e l a t e r a l s o i l pressure.
Alternatively the sides of the trench may be b a t t e r e d back to a safe
angle. Spoil should be kept well away from the excavation to a v o i d
slips. At least one trench depth away is advisable. D r a i n s may be
r e q u i r e d i n the bottom of a trench i n wet ground to keep the trench
dry while working, to reduce the p o s s i b i l i t y of embankment s l i p s a n d
to enable the backfill to be compacted p r o p e r l y . D r a i n p i p e s should
be open jointed and bedded in a stone filter with frequent

cross-drains to lead away water. In very wet ground cutoff drains


beside the trench may be necessary.
Pipes may be bedded on the flat trench bottom, but a proper
bedding is preferable to minimize deflections of f l e x i b l e pipes o r to
reduce damage to rigid pipes. The b e d of the trench may be pre-
shaped to fit the profile of the pipe. In rocky ground a bed of
sandy material may be b u i l t up a n d trimmed to shape. Beds of sand,
gravel or crushed stone up to 20 mm in size, brought up to the
haunches of the pipe, a r e often used, as they do not e x h i b i t much
settlement and a r e easy to compact. Concrete beds and haunches o r
surrounds a r e sometimes used f o r low-pressure r i g i d p i p e s a n d sewers.
Bedding should be continuous under joints but haunches should stop
short of j o i n t s to f a c i l i t a t e r e p l a c i n g damaged pipes. Fig. 11.1 shows
typical types of bedding. (Refer to Chapter 7 to bedding factors
associated w i t h d i f f e r e n t types of b e d d i n g ) (CPA, 1967; ACPA, 1970).
The best method of b a c k f i l l i n g a n d compaction depends on the type
of soil. Soil properties may be described in terms of the p l a s t i c l i m i t
(PI), liquid limit (LL) and plasticity index ( P I = LL-PL). For soils
with low PI'S fill may often be satisfactorily compacted merely by
inundation. For high LL soils, dynamic compaction is required i.e.
stampers o r vibrators. B a c k f i l l densities may be specified i n terms
of the Proctor Standard Density. For large pipes it is practice to
compact the bed w i t h a w i d t h of at least h a l f the p i p e diameter under
the p i p e to 95 to 100 % Proctor density. The soil f i l l around the p i p e
should be compacted in 100 to 150 mm l a y e r s to 90 or 95% Proctor
density up to haunch level or to 3 / 4 of the depth of the pipe. I t may
208

be necessary to s t r u t f l e x i b l e p i p e i n s i d e to prevent d i s t o r t i o n d u r i n g

t h i s operation. The fill around the top half of the pipe, up to 300
mm above the crown should be compacted to 85 to 90% Proctor density.
For b a c k f i I1 under roads, the top l a y e r s may have to be compact-
ed to 90-95% Proctor density, while for runways 110% may be re-
quired. 85% is normally sufficient in open country and densities
as low as 80% a r e obtained i f control i s b a d ( R e i t r , 1950).
To a v o i d damage to p i p e o r coatings, fill around the p i p e should
not h a v e stones in it. Surplus spoil may be spread over the trench
width to compensate for settlement or carted away. Reinstatement
of topsoi I and vegetation, p r o v i s i o n of land drainage, and fencing
o f f of w o r k i n g areas should be considered (see also Sowers, 1956).

THRUST BORES

To avoid excessively deep excavations, or to avoid disrupting


traffic on a r o a d over a pipe, the p i p e may be j a c k e d through the
soil for short lengths. A hole i s dug at each end of the length to
be jacked. A jack i s set up in one hole w i t h i t s back against a
thrust block or timber a n d the p i p e set up on r a i l s i n front of the
jack. Small pipes (less than 600 rnrn diameter) may have a sharp
head or shoe f i t t e d to the f r o n t to ease the j a c k i n g load a n d main-
tain the d i r e c t i o n of the p i p e , as the s l i g h t e s t obstacle w i l l deviate
the pipe. Occasionally the outside of the pipe is lubricated with
bentonite to reduce f r i c t i o n . Pushing should never stop for long or
the pipe may stick. With larger pipes the soil ahead of the pipe
is dug out by auger or hand and removed through the pipe. The
pipe is then pushed i n f u r t h e r , the j a c k r e t r a c t e d a n d another pipe
inserted behind the pipes in the bore and the process repeated.
The jacking force required to push a pipe with a shoe may be up

to 1 200 tons, but if an auger i s used, a 100 ton j a c k may suffice.

PIPE BRIDGES

Pipes f r e q u e n t l y h a v e to span r i v e r s o r gorges. For short spans,


rigid jointed pipes may have sufficient strength to support them-
209

Class A bedding
Reinforced A,: I 0% M= 4 8
RamforcedA, 0 4%M. 3
: 4
Dt 200 v m mln PI"," M.28

0 IZ5H
150 mm l l i "
0 25 d
I O O m m mm
Plain or Reinforced
bncr.1e 1 5 0 m k mi
Compocted
Gronulor t4
Y

CONCRETE ARCH

Class B bedding
M= 19

Dens@
Compacted
Bochf~l1

Class C bedding
M : I ~ 01251'

Lightly
tO"'p0ct.
Bocuftll

GRANULAR IFOUNDATION
CiKANULAK WUNUAII~

Class D bedding

NOTE For rock or other incompresrible malert316,the trench


should be ovaexcovoled 0 minimum of 150mm and
refi1l.d wrlb gronulor mat.rm

M-bedding factor

Fig. 11.1 Types of beddings

selves plus the fluid. For larger spans, it may be necessary to

support the pipe on trusses or concrete bridges, or h a n g the pipe


from an existing traffic bridge. If a truss bridge is to be con-
210

structed for the pipe, the p i p e could act as the tension member at
the bottom, or the compression member a t the top. A n a t t r a c t i v e form
of bridge is one with the pipe supported from suspension cables.
Two cables a r e p r e f e r a b l e to one, as they may be spaced a p a r t and
the hangers attached at an angle to the v e r t i c a l p l a n e through the
pipe. T h i s arrangement helps to support the p i p e against wind forces
and reduces wind vibrations. The pipe could also be designed to
act as an arch, but again lateral support would be necessary -
either two p i p e s could be designed to act together w i t h cross brac-
ing, or cable stays could be erected. In all cases except if sup-
ported on an independent bridge, the pipe should be r i g i d jointed;
preferably welded steel. T h i s means there w i l l h a v e to be some form
of expansion joint along the p i p e to prevent thermal stresses devel-
oping.

UNDERWATER P I PEL I NES

The laying of pipes underwater is very expensive, nevertheless


it is often inescapable. In addition to river crossings, undersea
laying is becoming common. Gas and oil lines from undersea beds
or off-shore t a n k e r b e r t h s a r e frequent. There a r e also many indust-
ries and towns which discharge effluent or sludge into the sea
through undersea pipes. The method used for underwater laying will
depend on conditions such as currents, wave height, type of bed
a n d length of under-water pipe:

(1) Floating a n d Sinking: In calm water the p i p e may be j o i n t e d


on shore, floated out and sunk. Lengths up to 500 rn have
been laid thus. Care is needed in sinking the pipe and
winches on barges should be positioned along the line to
assist.
(2) Bottom Towing: Lengths of pipe up to a few kilometres may
be assembled on land i n strings, each up to 1 km long, and
towed along the bed of the sea by a winch of a barge
anchored at sea. Strings are jointed together on the shore.
The pipe is bouyed up by air in the pipe or by buoys tied
to the pipe, to reduce f r i c t i o n on the sea bed. If the pipe
211

i s f i l l e d with air i t may h a v e to be weighted down by a con-


crete coating to get the correct buoyancy. Too light a pipe
would be susceptible to side currents and may bend out of
line. The stresses i n the p i p e due to the tension and bending
caused by underwater currents or waves may be critical.
The lateral drag per unit length due to waves and currents
is
Cow ( u + v ) ~D/2g (11.1)

where C i s the d r a g coefficient (about 0 . 9 ) , w i s the s p e c i f i c


D
weight of the water, u is the wave velocity and v is the
current velocity (or their components perpendicular to the
pipe), D is the outside diameter of the p i p e a n d g is gravi-
t a t i o n a l acceleration.
(3) Lay Barge: Pipes may be taken to sea in a pipe barge,
jointed on a special lay-barge a n d lowered continuously down
an arm (or stringer) into place. The barges proceed f o r w a r d
as l a y i n g proceeds.
(4) Underwater Jointing: Underwater jointing is difficult as each

pipe has to be accurately aligned. A diver w i l l be r e q u i r e d ,


which l i m i t s the method to shallow depths (Reynolds, 1970) .

It i s usually necessary to b u r y the p i p e to a v o i d damage due to

currents. Dredging i s d i f f i c u l t as sediment may f a l l or be washed into


the trench before the pipe is laid. The trench normally has to be
over-excavated. A novel way of l a y i n g the p i p e i n soft beds i s w i t h a
bed fluidizer (Schwartz, 1971). High-pressure jets of water fluidize
the bed under the pipes a n d the p i p e s i n k s into the bed. A number of
passes of the f l u i d i z e r a r e r e q u i r e d so that the p i p e s i n k s g r a d u a l l y
over a long length. Too deep a trench may cause bending stresses i n
the p i p e at the head of the trench.
Most underwater pipes are steel with welded joints. Pipes and
joints should be well cwted and lined and cathodic protection is
d e s i r a b l e under the sea.
212

JOINTS AND FLANGES

The type of joint to use on a pipeline will depand on the type


of pipe, the facilities available on site, cost, watertightness of
joint, strength and flexibility requirements. The following types
of j o i n t s a r e used ( E E U A , 1968) :-

(I) Butt-welded: The ends of the steel pipe are trued and
bevelled approximately 30' back, leaving a 2 mm root. The
joint is clamped and/or tacked in places and then welded
right around with one or more runs. Large pipes with thick
walls may also require a weld pass inside, although it is
difficult to make good the internal lining in pipes less than
600 mm bore.
(2) Sleeve welded: One end of a steel pipe may be flared out
to form a socket and the leading edge fillet welded to the
barrel of the other pipe. If possible the joint should be
welded inside too. (see F i g . 11.2). The e x t e r n a l p i p e coating

and the internal lining, if possible, should be made good


at the weld a f t e r a l l s l a g has been wire-brushed off.

Two methods of welding are available: Oxy-acetylene gas


welding and metallic arc welding. Top-class welding is

essential to achieve high weld strength and avoid cavities.


A thorough testing programme should accompany the welding.
Destructive testing should be confined where possible to the
factory and to sample welds on short lengths of pipe but
occasional field destructive tests may be necessary. Strips
including the weld are cut from the p i p e w a l l a n d bent over
a predetermined radius. The weld is ground and etched to
reveal air pockets or slag inclusions. Normally field tests
w i I I be non-destructive: e i t h e r by X-ray, Gamma-ray, magnetic
or ultra-sonic methods. X-rays are preferred for hi gh defini-
tion and contrast but Gamma-ray equipment i s more portable
and useful for otherwise inaccessible corners. With large-bore
pipes, the X-ray equipment may be inserted in the p i p e and
a film is wrapped around the p i p e to receive the image from
a single wall. For smaller pipes, double, overlapping expo-
213

o m i t t e d for small
bore pipes

Fig. 11.2 Sleeve-welded joint

Fig. 11.3 Push-in type j o i n t

sures are made from outside, thereby exaggerating defects


i n the near w a i l .
Welding of high tensile steels require better craftmanship
that for low-tensile steels. Pre-heating of the pipe ends,
and post-heating f o r stress r e l i e f a r e n o r m a l l y desirable.
(3) Screwed: Threaded and screwed ends and sockets are only
used on small steel pipes.
(4) Spigot and socket: Concrete, cast iron and p l a s t i c pipes fre-
quently have one end socketed and the other end plain. The

male end has a rubber sealing gasket f i t t e d over i t a n d the


socket is forced over the ring. (see Fig. 11.3). Various
shaped r u b b e r r i n g s a r e used. Another way of sealing the j o i n t
is to caulk i t w i t h bitumen compound, cement mortar, lead or

resin. Joints f i t t e d w i t h r u b b e r r i n g s a r e often c a u l k e d as well


to hold the ring in place, improve s e a l i n g and prevent dirt
getting into the joint. Spigot and socket joints with rubber
rings can normally accommodate one or two degrees of
def I ect ion.
(5) Clamp-on joints: Various proprietry joints are available for
jointing plain ended pipes. They normally involve clamping
a rubber seal between each pipe barrel and a cover sleeve.
Fig. 11.4 illustrates one such type of joint (also referred
to as a Viking Johnson or Dresser coupling). These joints
accommodate movement easily, and may be designed to take
several degrees of deflection between pipes or longitudinal
movement. If the joint is to resist tension, tiebolts may be
fixed to the barrel of each pipe by brackets to take the
tension. The b r a c k e t s cause local bending a n d tension stresses
in the walls of the pipe a n d may cause damage unless care-
fully designed. Normally four or even two tie bolts are
sufficient. The diameter of the bolts is selected to take the
tension and the brackets designed so that the bolts are as
close as possible to the barrel to minimize e c c e n t r i c i t y . The
brackets are normal l y "U" shaped in plan with the legs
welded parallel to the pipe axis. The length of the legs of
the bracket should be such that longitudinal shear a n d bend-
ing stresses are acceptable. In fact it i s impossible to e l i m -
inate all bending stresses on the pipe barrel with brackets
and for this reason rings, like flanges, may be preferable
for transferring the tie-bolt load to the pipe wai I. The
Victaulic coupling used f o r cast iron pipes, uses l i p s on the
ends of the pipes to t r a n s f e r tension to the coupling.
215

Fig. 11.4 Slip-on type coupling.

(6) Flanged: Steel and cast iron pipes are often flanged at the
ends, especially if p i p e s or fittings are likely to be removed
frequently. Faces are machined and bolted together with 3
rnrn rubber or other insertion gasket between. Flanges are
drilled to standard patterns according to the diameter and
working pressure.
Ful I face gaskets, with holes for bolts a n d the bore, are

used for high-pressure joints and cast iron or soft metal


flanges. Joint rings with an outside diameter slightly less
than the inside diameter of the b o l t holes a r e used f o r thick
steel flanges and low-pressure pipes. Some flanges have a
raised face inside the bolt circle for use with joint rings.

Flanges used with joint rings tend to dish when the bolts
are tightened though. The method of attaching the flanges
to the pipe varies. The simplest method i s to s l i p the f l a n g e
over the pipe but to leave approximately 10 mm extending
beyond the end of the pipe. A fillet weld is then carried
around the front of the pipe and at the back of the flange
around the pipe barrel. Alternatively both or one of the
inside edges of the flange may be bevelled for welding.
A method preferred for high pressure work is to have a lip
216

on the flange projecting over the end of the pipe so that


the bore at the j o i n t w i l l be f l u s h a f t e r welding.
Jointing and thrust flanges are fairly thick, as indicated by
the standard codes of practice, and must be securely welded to the
p i p e or cast i n t e g r a l l y . Two other types of f l a n g e s a r e common:-
(1) Puddle flanges are used on pipes which are cast in walls
of water retaining structures. The object of these flanges
is to prevent leakage of water and they are not necessarily
as thick as flanges which are designed to transfer longi-
tudinal pipe stress, although the flange should be securely
attached to the p i p e w a l l .
(2) Blank flanges are, as the name implies, i n s t a l l e d at blank
ends. Their thickness should be calculated to resist bending
moments at the edges and centre. For a c i r c u l a r disc rigidity
clamped at the edges the maximum radial bending stress
occurs at the perimeter bolts, and is equal to 3 pdZ/16t2
where p is the fluid pressure, d i s the p i t c h c i r c l e diameter
and t is the flange thickness. If the edges were not rigidly
clamped the maximum stress would occur at the centre and
would be 50% i n excess of t h i s amount.

COAT I NGS

Buried steel pipes are subject to corrosion and damage unless


the p i p e i s coated. Coatings should i d e a l l y be r e s i s t a n t to s c r a t c h i n g
during transport and laying of the pipe, to moisture, chemical and
biological attack, electric currents and temperature variations. They
should be h a r d enough to prevent damage during h a n d l i n g a n d due
to stones in the trench, yet sufficiently adhesive to adhere well
to the pipe wall and flexible enough to withstand the flexing of
the p i p e w a l l ( P P l , 1975; Cates, 1953).
The most common coating for pipes is a thin adhesive coat
followed by a coating reinforced with fibres and then possibly an
outer wrapping. The pipe surface is initially cleaned by wire
brushing, sandblasting or acid pickling, a n d the prime coat is then

applied by spray, brush or dipping the pipe in a bath. Bitumen


217

or coal t a r enamel i s the p r e f e r r e d prime coating. After the p r i m a r y


coat the pipe may be spirally wrapped with impregnated felt or
woven glass fibre matting. This i s sometimes followed by paper or
asbestos f e l t impregnated w i t h bitumen or coal tar. The p i p e i s then
whitewashed to assist in detecting damage a n d to s h i e l d the c o a t i n g
from the sun. The coating may be a p p l i e d in the f i e l d a f t e r welding
the pipe joints, or in the factory, in which case the ends a r e left
bare for jointing and coated in the field. The total thickness of
coating should be at least 5 mm.
Other types of coatings include an asbestos f i b r e bitumen mastic 3
to 6 mm t h i c k , coal tar pitch, epoxy paints, PVC or polythene tapes
(either self adhesive o r bedded on an adhesive), resins or plastics,
cement mortar and zinc applied by galvanizing. Exposed p i p e s may
be primed and painted with bitumen based aluminium or enamel
(AWWA, 1962).
Cement mortar coatings offer additional resistance to buck1 i n g
in the case of l a r g e bore t h i n walled pipes. Cement mortar coatings
are usually 12 to 20 mm ($ to 3/4 inch) thick (AWWA, 1954, 1962).
Finished coatings may be checked for flaws, pin holes etc. by
means of a Holiday detector. An electrical conductor in the form of
metal brushes or rolling springs is run along the p i p e coating. An
electrical potential is applied across the coating and a current is

observed when f l a w s in the c o a t i n g a r e de ected.

LININGS

Steel pipes a r e l i n e d to r e s i s t interna corrosion and minimize the


friction losses. Unlined steel pipe may be oxidized by corrosive
substances in the fluid. In the case of solids-conveying systems in
particular, the oxide is rapidly scraped off leading to further
corrosion. Corrosion of water-conveying pipes may be inhibited by
m a i n t a i n i n g a h i g h pH e.g. by adding lime. Lime on the other h a n d
could cause carbonate s c a l i n g .
The most popular linings are bitumen (3 to 5 mm t h i c k ) o r coal
tar enamel (2 to 3 mm thick). Bitumen, which is a by-product of
petroleum, i s t h e cheaper of the two.
Before applying the lining the pipe wall is cleaned by sand-
218

b l a s t i n g or other methods and the lining i s then a p p l i e d by brush,


spray, dipping or s p i n n i n g to o b t a i n a smooth surface. Spun enamel

in particular provides a smooth f i n i s h . Coal tar enamel i s also more


resistant to moisture than bitumen, although it is more b r i t t l e and
consequently subject to damage by impact and flexing of the pipe.
Plasticised coal tar enamels have, however, now been developed to
overcome the problems of b r i t t l e n e s s .

Epoxy paints are also used successful l y for I inings, although


careful appl ication is necessary to ensure successive coats adhere
to each other. The recommended thickness of the lining varies with
the type o f paint, but it i s n o r m a l l y of the o r d e r of 0 . 3 mm a p p l i e d
in 2 to 4 layers. Coal tar epoxy linings should be avoided for
portable water pipes as they taint the water. Lead based p r i m e r s
should also be avoided as they a r e toxic.

Cement mortar lining, applied centrifugally by spinning the p i p e ,


is also used for large bore pipes. The applied lining is usually

6 to 12 mm ( 1 / 4 to 3 inch) thick. Mortar l i n i n g s have been a p p l i e d


successfully to pipes in the f i e l d . The o l d surface i s f i r s t thoroughly
cleaned b y wire-brushing or sand b l a s t i n g , and then coated c e n t r i f u -

gally by a machine drawn or propelled slowly along the line (Cole,


1956).
On pipes over about 600 mm bore, most factory linings may be
made good at field joints manually or by mechanical applicators.

CATHOD I C PROTECT I ON

Despite the use of protective coatings, corrosion of the walls


of steel pipes often occurs through flaws, pin-holes or at exposed
fittings. Corrosion i s due p r i m a r i ly to two causes; g a l v a n i c corrosion
and stray current electrolysis (Uhlig, 1948).

Ga I vanic Corrosion
When two dissimi l a r materials are connected through an electro-
lyte, current may flow from one m a t e r i a l to the other. The r e s u l t i n g
electric current flows from the anode to the cathode through the
electrolyte. Particles or ions, leave the anode causing corrosion.
T h e cathode i s not attacked.
219

Such action may take place when two dissimilar metals are in
contact in conductive soil (for instance fittings of a different
material to the p i p e ) . Another more frequent form of gal v ani c corr-
osion occurs w i t h pipes i n corrosive soils. The effect is particularly
marked in soi I w i t h v a r y i n g characteristics, differential oxygen con-
c e n t r a t ions or in water w i th high chemical content , especial l y
sulphur. Corrosion i s alSo caused by biochemical action in the s o i l
which is a type of galvanic a c t i o n r e s u l t i n g from b a c t e r i a . A method
which has been used successfully to counteract soil corrosion was
to add lime to the trench backfill. Stress concentration in the steel
in an e l e c t r o l y t e may also lead to corrosion. The amount of corrosion
due to the l a s t named two influences i s u s u a l l y small.
Potential corrosive areas may be detected by s o i l resistance tests,
pipe - soil potential tests or measurement of the c u r r e n t in the pipe.
If the s o i l i s a t a l l suspect, or as a s t a n d a r d p r a c t i c e f o r major
pipelines, a soi I resistivity survey should be conducted. Soi I r e s i s t -
ivity i s measured i n ohm - centimetres. Readings a r e n o r m a l l y taken
in-situ using a bridge circuit or by measuring a current and the
associated volt drop. Standard probes are available for these
measurements, but any major surveys should be done by an exper-
ienced corrosion engineer (Schneider, 1952).
A highly conductive soil may have a resistivity of 500 ohm cm,
a n d a poorly conductive s o i l , more than 10 000 ohm cm.
The potential difference between a buried pipe and the soil is
important in evaluating corrosive conditions. The potential is
measured by connecting a voltmeter between the pipe and a special
electrode in contact with the soil. A copper sulphate half cell is
frequently used as the electrode under normal conditions. The poten-
tial of a p i p e i s 0.5 to 0.7 v o l t s below that of the s u r r o u n d i n g s o i l .
If the pipe is at a higher voltage than 0.85 volts below that of
the soil, currents are likely to flow from p i p e to s o i l , thereby cor-
r o d i n g the pipe.
To prevent t h i s corrosion, a sacrificial anode may be connected
by a conductor to the pipe in the vicinity of possible corrosion
(Fig. 11.5). The s a c r i f i c i a l anode i s b u r i e d i n a conductive s u r r o u n d ,
preferably below the water table and currents will tend to leave
220

from the anode instead of the pipe, thereby limiting the corrosion
to the anode. If the anode is sufficiently dissimilar from the steel
pipe to cause galvanic action no external electrical potential need
be applied. In fact, a resistor may sometimes be installed in the
connection to I imi t the current. Common sacrificial anodes are
magnesium, zinc and aluminium, or alloys of these metals. Mag-
nesium has the greatest potential difference from iron, has a high
electro-chemical equivalent (ampere hours per k i logram of material)
and is fairly resistant to anodic polarization. ( A l a y e r of hydrogen
ions may replace the ions leaving the anodes, thereby actively
insulating the metal against further attack. This is termed polar-
ization). Magnesium anodes give 200 to 1 200 ampere hours per kg.
Magnesium anodes are normally designed for about a 10 year life
b u t zinc anodes often l a s t 20 or 30 years.

Junction BOI Wilh


R o a l ~ t o ri f Required

Fig. 11.5 S a c r i f i c i a l anode i n s t a l l a t i o n f o r cathodic protection

T h e spacing a n d size of anodes will be determined by the c u r r e n t


requirements to bring the pipe to a safe potential (at least 0.85
volts below the soil potential along the entire length of the pipe).

The spacing of sacrificial anodes w i l l vary from 3 m in poorly con-


ductive soil and poorly coated pipe to 30 m in highly conductive
soil provided the p i p e i s well coated. The most r e l i a b l e way of est-
imating the required current is to actually drain current from the
pipe by means of a ground bed and d.c. source and measure the
r e s u l t i n g p i p e to s o i l p o t e n t i a l a l o n g the pipe.
221

As a rule of thumb, the current required to protect a pipe


a g a i n s t corrosion i s i = 10/r per square metre of b a r e d p i p e surface,
where i is the current in amps, r i s the s o i l resistance i n ohm cm
and a factor of safety of 3 is incorporated. The area of exposed
pipe may be as low as 0.5 per cent for a good coating, r i s i n g to
20 p e r cent f o r a poor coating.
Once the required draw-off current per km of pipe is known,

the total anode size may be calculated. In highly conductive soils


(less than 500 ohm cm) large anodes (20 k g ) may be used but in
soils with lower conductivities (greater than about 1 000 ohm cm)
a number of smaller anodes should be used to ensure an adequate
c u r r e n t output.
If an electricity supply is available, it i s usually cheaper to
use an impressed current type of protection (Fig. 11.6) instead of
a sacrificial anode. A transformer-rectifier may be i n s t a l led to pro-
v i d e the necessary dc c u r r e n t .
With impressed current i n s t a l l a t ions, the anode need not be self-
corrosive. Scrap iron, or graphite rods buried in coke fill, are
frequently used. Steel anodes will quickly corrode in highly con-
ductive soils and for this reason high s i l i c o n cast i r o n anodes a r e
preferred.
The type of protection to use w i l l depend on the long-term econ-
omics. Impressed current i n s t a l Iations are frequently the cheapest
in the long run, as maintainance costs are low and fewer instal-
lations are required than for sacrificial anodes. In the case of

impressed current installations a relatively high voltage may be


applied, which in turn protects a longer length of pipe than a
sacrificial anode could. The pipe-soil potential fa1 I s off rapidly
away from the applied voltage though, consequently a high voltage
may have to be applied to protect a long length. The pipe-to-soil
potential in the immediate vicinity of the point should not, in
general exceed 3 volts (the pipe potential b e i n g below that of the
soi I ) . Larger voltages may damage coatings. Impressed current
installations may be at 1 to 100 km spacing depending on the
qua1 i t y of the p i p e coating. Impressed c u r r e n t installations are pre-
ferable to sacrificial anodes if the soil resistivity is higher than
about 3 000 to 5 000 ohm cm.
222

Considerable economic savings are often achieved by protecting


only "hot spots", or points subject to aggressive attack, and if
occasional shutdowns can be tolerated this should be considered.

Stray Current E l e c t r o l y s i s
The most severe form of corrosion is often caused by stray dc
currents leaving a pipe. Railways and other users of dc current
r e t u r n c u r r e n t s through rails, but if there i s another conductor such
as a steer p i p e nearby, a p r o p o r t i o n of the c u r r e n t may flow through
the conductor instead of the r a i l . Where the c u r r e n t leaves the p i p e ,
steel will be corroded at a rate of 9 kg (20 Ibs) per year per
ampere of current. The current may be detected by actual current
measurements or from pipe/soi 1 potent iat measurements. Continuous
recordings should be taken over a day, as the currents may fluc-
t u a t e w i t h time.

B o r e d or S u b m r r g e d
Structure Hod8 Cothodic
Rolotlv8 t o E l w . t r o l y t 8

Iron or grophite 1

Fig. 11.6 Impressed c u r r e n t corrosion protection.

The corrosion associated with stray currents may be prevented


by connecting the pipe at the point where the current leaves it,
to the destination of the current, with a conductor. The current
may leave the pipe along some length, in which case a current
would have to be impressed on the pipe to maintain the voltage
sufficiently low to prevent current escaping. A ground bed and
transformer r e c t i f i e r may be r e q u i r e d f o r t h i s protection.
223

When a pipe i s cathodically protected, c a r e should be taken that

the mechanical joints are electrically bonded, by welding a cable


across them if necessary. Branch pipes may have to be insulated
from the main p i p e to control the currents.

THERMAL I NSULAT I ON

Fluid in the pipeline may often be heated or cooled by the


surroundings. Ice formation i n a r c t i c climates a n d h e a t i n g i n t r o p i c a l
climates are sometimes a serious problem. Heat is transferred to
or from the interior of an exposed pipe in a number of ways: by
conduction through the pipe wall and wrapping and a boundary
layer of fluid inside the pipe wall, by r a d i a t i o n from the e x t e r n a l
face of the pipe and by convection and wind currents in the air
s u r r o u n d i n g the pipe.

TABLE 1 1 . 1 Thermal c o n d u c t i v i t i e s

Therma I c o n d u c t i v i t y
Mater ia I k cal B t u in
m sec C" sq f t hr F o

Water 0.000 14 4
Air 0.000 005 0.15
Steel 0.014 420
B i tum i n ized Wrapping 0.000 01 0.3
Concrete 0.0002 6
Slag wool 0.000 01 0. 3

EEUA, 1968.

The heat loss by conduction through an homogeneous pipe wall


is proportional to the temperature gradient across the wall a n d the
heat t r a n s f e r r e d per u n i t area of p i p e w a l l per u n i t time i s

-Q - - k
AT
T A0 (11.2)
224

where Q is the amount of heat conducted in kilocalories or Btu, A


is the area of pipe surface, T is the duration time, A0 is the
temperature difference across the wal I , t is the wal I thickness and
k is the thermal conductivity, tabulated in Table 11.1 for various
materials.
An equation was developed by Riddick et at (1950) for the total
heat transfer through the wall of a pipe conveying water. The
equation has been modified to agree with relationships i n d i c a t e d by
EEUA (1964) a n d i s converted to metric u n i t s here:
Rate of heat loss of water kcal/kg/sec

(0.I - ' o ) / 250D


_- _ _ _
t, + t o + 1 + 1
(11.3)

and since the specific heat of water is unity, the r a t e of drop in


temperature in degrees Centigrade per sec equals the heat loss in
k c a I/kg/sec , where

temperature of water C "


ambient temperature outside C "
diameter m
thickness of p i p e m
c o n d u c t i v i t y of p i p e w a l l kcal/m sec C o
thickness of coating m
c o n d u c t i v i t y of c o a t i n g kcal/m sec C o
heat loss through boundary layer =
0.34(1 + 0.020i)v0'8/D0'2 kcal/m2 sec C"
water velocity m/s
heat loss by r a d i a t i o n = 0.054 E('O + 273 1 3
100
kcal/m2 sec C o
e m i s s i v i t y factor = 0.9 f o r b l a c k a s p h a l t a n d concrete
0.7 for cast i r o n a n d steel
0.4 for aluminium p a i n t
'i -90 0.25
heat loss by convection = 0.000 34f1 + 0.78V (- 1
D
kcal/m2 sec C"
v = w i n d velocity km/hr
225

The h e a t generation by fluid friction is usually negligible


although i t theoretically i n c r e a s e s the t e m p e r a t u r e of t h e f l u i d .

Example
An uncoated 300 mm diameter pipeline 1 000 m long with 5 mm
thick steel walls, exposed to a 10 km/hr wind, conveys 100 C / s of
water at an initial t e m p e r a t u r e of 10°C. The a i r t e m p e r a t u r e i s 30°C.
Determine t h e e n d t e m p e r a t u r e of t h e w a t e r .

E = 0.7
V = 0.1/0.785 x 0.32 = 1.42 m/s
= 0.34 x 1.42°'8/0.30.2 = 0.67
Kf
= 0.054 x 0.7 ( 30 + 27313 10-3 = 0.001
Kr

tl
- = 0.005
~- - o.35
0.014
kl
(30 - 1 0 ) / ( 2 5 0 x 0 . 3 )
= 0.00103°C/sec
= 0.35 + 1/0.67 + (0.001 + 0.0029)
T e m p e r a t u r e r i s e o v e r 1 000 m = 1 000 x 0.00103 = 0.7"C
1.42

The insulation of industrial pipework carrying h i g h or low temp-

erature fluids i s a subject on i t s own. The cost of t h e heat t r a n s f e r


s h o u l d be b a l a n c e d a g a i n s t t h e cost of t h e l a g g i n g .
Heat transfer to or from buried pipelines depends on the temp-
erature of the surroundings, which in turn is influenced by the
heat transfer. The temperature gradient and consequently rate of
heat transfer may vary with time and are difficult to evaluate. If
the temperature of the surroundings is known, Equ. 11.3 may be
used, o m i t t i n g the term l / ( K + Kc).

REFERENCES

American Concrete P i p e Assn., 1970. Concrete P i p e Design M a n u a l ,


Arlington.
AWWA S t a n d a r d C602, 1954. Cement M o r t a r L i n i n g of Water P i p e l i n e s
in P l a c e , N.Y.
226

AWWA S t a n d a r d C203, 1962. C o a l T a r Enamel P r o t e c t i v e C o a t i n g s f o r


Steel W a t e r P i p e 30 Ins. and O v e r , N.Y.
AWWA S t a n d a r d C205, 1962. Cement M o r t a r P r o t e c t i o n L i n i n g and
C o a t i n g f o r Steel W a t e r P i p e 30 I n s . and O v e r , N.Y.
BSCP 2010, 1970. P a r t 2, D e s i g n and C o n s t r u c t i o n of Steel P i p e l i n e s
in Land, B S I , L o n d o n .
Cates, W.H., 1953. C o a t i n g f o r s t e e l w a t e r p i p e , J. Am. W a t e r W o r k s
Assn. 4 5 ( 2 ) .
Cole, E.S., 1956. D e s i g n o f s t e e l p i p e w i t h cement c o a t i n g and l i n -
ing, J . Am. W a t e r W o r k s Assn., 4 8 ( 2 ) .
C o n c r e t e P i p e Assn., 1967. B e d d i n g and J o i n t i n g of F l e x i b l y J o i n t e d
C o n c r e t e P i p e s , Techn. B u l . No. 10, T o n b r i d g e .
Engineering E q u i p m e n t U s e r ' s Assn., 1964. Thermal I n s u l a t i o n of
P i p e s and Vessels, H a n d b o o k No. 12, C o n s t a b l e , L o n d o n .
Engineering E q u i p m e n t U s e r s Assn., 1968. P i p e J o i n t i n g Methods,
H a n d b o o k No. 23, C o n s t a b l e , L o n d o n .
P i p e s and P i p e l i n e s I n t e r n a t i o n a l , 1975. P i p e l i n e p r o t e c t i o n r e v i e w ,
20(4).
Reitz, H.M. 1950. S o i l m e c h a n i c s and b a c k f i l l i n g p r a c t i c e , J . Am.
W a t e r W o r k s Assn., 4 2 ( 1 2 ) .
R e y n o l d s , J.M., 1970. S u b m a r i n e p i p e l i n e s , P i p e s and M a n u a l . 3rd
Ed. S c i e n t i f i c S u r v e y s , L o n d o n .
Riddick, T.M., L i n d s a y , N.L. and T o m a s s i , A., 1950. F r e e z i n g of
w a t e r in e x p o s e d p i p e l i n e s , J. Am. W a t e r W o r k s Assn., 42(11).
S c h n e i d e r , W.R., 1952. C o r r o s i o n and c a t h o d i c p r o t e c t i o n o f p i p e l i n e s ,
J. Am. W a t e r Works Assn., 44(5).
Schwartz, H. I., 1971. H y d r a u l i c t r e n c h i n g of s u b m a r i n e p i p e l i n e s ,
P r o c . Am.Soc. C i v i l E n g r s . 9 7 ( T E 4 )
Sowers, G.F., 1956. T r e n c h e x c a v a t i o n and b a c k f i l l i n g , J. Am. W a t e r
W o r k s Assn., 4 8 ( 7 ) .
U h l i g , H.H., 1948. The C o r r o s i o n H a n d b o o k , W i l e y , N.Y.

L I S T OF SYMBOLS

A - area
- a r e a of steel
AS
B - w i d t h of t r e n c h
- drag c o e f f i c i e n t

d - inside diameter
D - outside diameter
E - thermal emissivi ty factor
f - stress
9 - gravitational acceleration
H - backfi I I a b o v e t o p o f p i p e
I - c u r r e n t in amps
227

K heat loss by convection


C
heat loss through water f i l m
Kf
heat loss b y r a d i a t i o n
Kr
c o n d u c t i v i t y of p i p e wal I
kl
c o n d u c t i v i t y of coating
k2
M bedding factor

P pressure
Q amount of heat
r resistance i n ohms
T time
thickness of p i p e w a l l
tl
t h i ckness of c o a t i n g
t2
u component of wave v e l o c i t y p e r p e n d i c u l a r to p i p e
V mean water velocity, or component of current velocity

p e r p e n d i c u l a r to p i p e
v w i n d velocity
W s p e c i f i c weight of water

Y depth of bedding m a t e r i a l below p i p e


0. temperature i n s i d e p i p e
0 temperature outside p i p e
0
228

CHAPTER 12

PUMP I NG I NSTALLAT IONS

INFLUENCE OF PUMPS I N P I P E L I N E DESIGN

The pipeline engineer i s frequently concerned with the design of


pumping lines. That is, the fluid is forced through the pipe by a
pump as opposed to a gravity fed line. The design of the pipe is
inter-related with the pumping equipment selected and vice versa.
Pumes must overcome line friction losses in addition to net static
head, hence t h e p u m p i n g h e a d a n d p o w e r r e q u l .ement a r e a function
of the pipe diameter. The pipe wall thickness is in turn based on
interna I pressures. Therefore friction head a n d f l u i d velocity (which
a f f e c t s w a t e r hammer h e a d ) a l s o a f f e c t t h e w a l l hickness of the pipe.
The engineer therefore cannot design one v, thout considering the
other. The system must be designed in a n integrated manner. Pipeline
engineers should therefore be aware of the various types of pumps,
their I i m i t a t ions, their characteristics and costs. Stepanoff (1957),
IWt (1969) and Addison (19551, for example, give more complete
information. The pump size and suction head requirements w i l l also
have to be considered in the pumpstation design. The d e s i g n o f the
pump, i.e. the optimum casing or impeller shape and material, is
more the concern of the manufacturer. In fact it i s covered in books

such as t h o s e b y K a r a s s i k and C a r t e r (1960), K o v a t s (1964) and KSB


(1968). Similarly electrical power supply, motor design a n d control
equipment a r e so s p e c i a l i z e d that they should be selected o r designed
by electrical engineers. It is from the pipeline engineer’s point of
view that the following i n f o r m a t i o n i s presented.

TYPES O F PUMPS

P o s i t i v e D i s p l a c e m e n t Types
Least used in pipeline practice nowadays, but still used in
some industrial a p p l i c a t ions, are positive displacement pumps.
229

The reciprocating pump i s one such type. A p i s t o n or ram i s forced


to move to and fro i n a cylinder. One-way v a l v e s p e r m i t the liquid
to enter on a reverse stroke and exit into the receiving conduit
on a f o r w a r d stroke of the piston.
The p i s t o n can be replaced by a flexible diaphragm. There are
also rotary positive displacement pumps. These include s pi ral types
a n d intermeshing gears.

Centrifugal Pumps
Rotodynamic pumps move l i q u i d by i m p a r t i n g a velocity and hence
pressure to it. This action may be in the form of an axial flow
like through a propeller. Liquid is forced along a tabular casing
by rotating blades. There may be guide vanes in the casing to
reduce s w i r l . T h i s type of pump i s most common f o r l a r g e flows a n d
low heads. For high heads the centrifugal pump i s preferred. Here
r o t a t i o n of an impeller creates an o u t w a r d ( r a d i a l ) c e n t r i f u g a l flow.
The latter form is the most common, and the centrifugal pump
in various forms is universal ly accepted as practical ly standard
in waterworks engineering. This type of pump i s capable of hand-
l i n g flows up to 20 cumecs. For h i g h flows a n d to p e r m i t f l e x i b i l i t y ,
pumps in parallel are common. The head which can be generated
by an impeller is limited to a l i t t l e over 100 metres of water, but
by assembling a number of impellers i n series a wide r a n g e of heads
i s possible.

lcllcr
,nes

Fig. 12.1 Sectional e l e v a t i o n of v o l u t e t y p e c e n t r i f u g a l pump


(Webber, F l u i d Mechanics f o r C i v i l Engineers, 1971)
230

The inlet c o n d i t i o n must be controlled to a v o i d c a v i t a t i o n or air

entrainment and the permissible maximum suction pressure i s r e i a t e d


to the delivery pressure of the first pumping stage. In some cases
instead of connecting impellers i n series on a common shaft a number
of independent pump u n i t s a r e i n s t a l l e d i n series. T h i s permits more
flexibility particularly if one of the units is to be driven by a
v a r i a b l e speed motor.
The driving device i n a c e n t r i f u g a l pump i s termed the impeller.
I t may be a series of vanes shaped to f l i n g the water from a c e n t r a l
eye or inlet towards the outer periphery. Here it is directed by
a volute casing into the discharge nozzle. The efficiency of this
type of pump may be as much as 90 percent although pump and
motor units together rarely h a v e efficiences above 80 percent except
in large sizes. Pumping installations for unusual duties such as
pumping solids in suspension can have efficiencies as low as 30
percent.
Usually the inlet to the impeller eye is on one side, and the
d r i v i n g shaft to the motor on the other side of the impeller. i n some
s i t u a t i o n s an i n l e t may be on each side i n which case i t i s r e f e r r e d
to as a double e n t r y pump. This involves a more expensive casing
b u t i t r e s u l t s i n a balanced t h r u s t i n the d i r e c t i o n of the shaft.

Fig. 12.2 Section through a s i n g l e e n t r y horizontal spindle centrifugal


pump
231

Pump u n i t s may be mounted w i t h the s h a f t vertical or horizontal.


The horizontal shaft with a split casing is the most practical from
the maintenance point of view. Where space is limited, or a deep
suction well is required, a vertical spindle arrangement may be
preferable. The motor is then mounted on a p l a t f o r m above the pump.

TERMS A N D DEF I N I T I ONS

Head
The head in i t s general sense is the energy per u n i t weight of
water. It comprises elevation above a certain datum, z, p l u s pres-
sure head p / w p l u s velocity head v2/2g.

Total Head
Total head usually refers to the excess of discharge head to
inlet head and it is the head generated by the pumps (see F i g .
12.3).

Fig. 12.3 D e f i n i t i o n of net p o s i t i v e suction head.


232

Net Positive Suction Head

In order to avoid cavitation, air entrainment and a drop-off


in pumping efficiency a pump needs a c e r t a i n minimum suction head.
The net positive suction head (NPSH) r e q u i r e d can u s u a l l y be indi-
cated by a pump manufacturer as a function of r a t e of pumping.
The requirement is referred to as the net positive suction head.
It i s generally expressed as the absolute head above vapour pres-
s u r e r e q u i r e d a t the pump inlet. I t is c a l c u l a t e d as follows:

NPSH = Hp - Pv/W ( 2.1

= H r + H a - H s - H - H - PV/W ( 2. 2)
f e
where H = head at inlet, relative to centre line of inlet, and
P
equal to absolute pressure head p l u s velocity head.
= vapour pressure
PV
W = u n i t weight of water

= suction r e s e r v o i r head above datum


Hr
= atmospheric pressure head (about 10 m of water
Ha
at sea l e v e l )
H = elevation of centreline of pump suction above datum
S
= f r i c t i o n head loss between r e s e r v o i r a n d pump
Hf
= turbulence head loss between r e s e r v o i r a n d pump
He
NPSH requirement is a subjective figure as there i s not a sharp
drop-off in efficiency or increase in noise level at this value. It
increases with discharge. It i s p r e f e r a b l e to be on the conservative
side when considering NPSH. The NPSH requirement i s of the order
of 10 percent of the head of the f i r s t stage of the pump b u t a more
re1 i a b l e figure should be obtained from the manufacturer's tests.
NPSH can be assessed by observing c a v i t a t i o n o r from performance
tests (Grist, 1974). As the NPSH i s reduced cavitation commences,
firstly at flowrates away from best efficiency flowrate but eventually
at al I flowrates. At an NPSH just less than that associated with
maximum erosion due to cavitation the head generated by the
impeller s t a r t s t o decrease. The head r a p i d l y decreases w i t h decrease
in NPSH. A l i m i t of 3% decrease in head i s recommended for estab-
lishing the NPSH. The NPSH r e q u i r e d to avoid cavitation should be
at least 3 times the 3% head drop NPSH, though. For this reason
some degree of c a v i t a t i o n i s normally acceptable.
233

Specific Speed

The specific speed Ns of a pump i s a useful i n d i c a t i o n of the


pump's capabilities. There are two slightly different definitions of
s p e c i f i c speed a n d the more general one i s g i v e n f i r s t :
Specific speed i s a c h a r a c t e r i s t i c velocity of the r o t a t i n g element
relative to a characteristic velocity of the water, i.e. it is the
v e l o c i t y of the impeller r e l a t i v e t o that of water.
The speed of the periphery of the impeller is proportional to
ND where N is the impeller rotational speed and D is i t s diameter
(units will be introduced later). From B e r n o u l l i ' s theorum the water
velocity i s proportional to JgH where H i s the total head. Therefore
Ns is proportional t o ND/ JgH. But discharge Q is proportional to
vDb where b is the width of the impeller. For a n y g i v e n geometric
p r o p o r t i o n b i s p r o p o r t i o n a l t o D, so

Qa,&8D2 (12.3)
E l i m i n a t i n g D one a r r i v e s a t

NS
O- NQf / (12.4)

This is a dimensionless form of specific speed p r o v i d e d the u n i t s


are consistent. Unfortunately most catalogues and text books omit
the g term and express N in revolutions per minute (rpm) not
r a d i a n s p e r second. One encounters the expression

NS = NQi/H3/4 (12.5)

In imperial units Q is in gallons per minute and H i n feet. In


S.1. u n i t s Q i s i n cumecs ( c u b i c metres p e r second) a n d H in metres.
The alternative definition of NS i s the speed in rpm of a geo-
metrically similar pump of such size that i t would d e l i v e r one cumec
(or gal. per min) against 1 metre (or 1 foot) head. Then by sub-
stitution one arrives directly at the latter dimensionally dependent
expression for N .
By reviewing the former definition of NS, one obtains an idea
of the size of impeller relative to pumping head. In fact a high
Ns is associated with a low head and fast rotational speeds,
whereas a low NS implies a high head and low speed relative to
that of water (water thus has a high radial component relative to
t a n g e n t i a l component).
234

Fig. 12.4 indicates the r e l a t i o n s h i p between head per stage and


s p e c i f i c speed f o r a number of pumps selected from r e a l applications.
N is i n r p m a n d w i n radians/sec (w = 271 N/60).
7 Lo
Lo 0

\ W2 v u l t i - s t a g e
74-
50
t@
Ln
0,
L
$20 - \-- n \
E
I

\ 7 T e d flow

1
I
\ '
5

20
I
<
50 100
I
200
I Ns

500 1000'

Fig. 12.4 Relationship between head and specific speed for pumps

IMPELLER DYNAMICS

The shape of the impeller blades i n a c e n t r i f u g a l pump influences


the radiai flow pattern which in t u r n gives the pump c e r t a i n char-
acteristics. A basic understanding of relative velocities inside the
pump i s therefore useful in selecting a type of pump f o r a p a r t i c u l a r
duty.
Referring to Fig. 12.5 the circumferential speed of the impeller
is u, the velocity of the water relative to the impeller is w and
the absolute water velocity i s v.

Torque T o n the water i s force x r a d i u s


But force = change in momentum
235

T h e r e f o r e i m p a r t e d T = c h a n g e i n moment of momentum

= A(mvr) (12.6)
= 1 r2v2 cos a dm - l r l v l cos a dm (12.7)
2 1

Fig. 12.5 V e l o c i t y v e c t o r s a t i m p e l l e r of c e n t r i f u g a l pump.

where dm is an element of mass discharge per unit time, and the


integral i s o v e r a l l the vanes.
Power P = T w where i s t h e r o t a t i o n a l speed (12.8)
.'. P = A (mvr) w (12.9)
Now r w = u1 a n d r w = u 12.10)
1 2 2
.'. P = 1 u 2 v 2 c o s a dm -
2
/ulvl cosa drn
1
12.1 1 )
A l s o P = pgHQ where Q i s the d i s c h a r g e r a t e 12.12)
.'. P = gHJdm s i n c e p Q = Jdm 12.13)
.*. Head H = ( u v c o s a - UIVl
2 2
cosal)/g 12.14)
The term u,vl coscll i s n o r m a l l y r e l a t i v e l y s m a l l .
Now v cosa2 = u 2 - t 2 cot B2 (12.15)
where t 2 i s t h e r a d i a l component of v
2
.*. H (u, - t 2 cot B 2 ) / g (12.16)
a n d Q = nDbt (12.17)
2
... H A - BQ cot B2 (12.18)
where A and 0 are constants, g is gravity and p is water mass
density. The blade angle B2 therefore affects t h e shape of the pump
H-Q c h a r a c t e r i s t i c c u r v e , as i l l u s t r a t e d in Fig. 12.6.
236

1
Q

Fig. 12.6 Effect of b l a d e a n g l e on pump c h a r a c t e r i s t i c s

PUMP CHARACTER I ST I C CURVES

The head-discharge characteristic of pumps i s one of the curves


usually produced by the manufacturer and used to design a viable
pumping system. This relationship is useful for pipe size selection
and selection of combinations of pumps in parallel as will be
discussed l a t e r .

-BH

Fig. 12.7 Pump c h a r a c t e r i s t i c curves.


237

Break horse power (BHP) or shaft power required by the pump


and pump efficiency E a r e often i n d i c a t e d on the same d i a g r a m (see
Fig. 12.7). The duty H-Q c u r v e may a l t e r with the d r i v e speed a n d
could also be changed on any pump by fitting different impeller
diameters. The BHP a n d E curves would also change then. The r e s u l -
tant head is proportional to NZ a n d D2, whereas the discharge is
p r o p o r t i o n a l to N a n d D.
The required duty of a pump i s most e a s i l y determined g r a p h i c -
ally. The pipeline c h a r a c t e r i s i t c s a r e p l o t t e d on a head - discharge
graph. Thus Fig. 12.8 illustrates the head (static plus friction)
required for different a's, assuming (a) a h i g h suction sump water
level and a new, smooth pipe, and (b) a low suction sump level
and a more pessimistic f r i c t i o n factor applicable to an older pipe.
A line somewhere between could be selected for the duty line. The
effect of para1 lel i n g t w o or more pumps can also be observed from
such a g r a p h .

I 2 in parallel

llow sump and old pipe /


0flurves/ pipeline

a
\\
high sump level and new pipe

Fig. 12.8 P i p e l i n e a n d pump c h a r a c t e r i s t i c s .


238

Pumps in series are more difficult to control. Frequently,


however, a booster pump station (Fig. 12.9) is required along a
pipeline to increase c a p a c i t y or to enable the f i r s t section of pipe-
line to operate within a lower design pressure. In such cases the
pump curves are added one above the other, whereas for pumps in
parallel, the discharge at any head is added, i.e. the abscissa
i s m u l t i p l i e d by the number of pumps i n operation.
Booster pumps may operate in-line, or with a break pressure
r e s e r v o i r or sump on the suction side of the booster. I n the latter
case a certain balancing volume is required in case of a trip at
one station. The hydraulic grade l i n e must also be drawn down to
reservoir level. In the former case water hammer due to a power
f a i l u r e i s more d i f f i c u l t to p r e d i c t a n d allow f o r .

Fig. 12.9 Location of booster pump station.


239

MOTORS

The driving power of a pump could be a steam turbine, water


wheel or most commonly nowadays, an electric motor. The r o t a t i o n a l
speed of the pump is controlled by the frequency of alternating
current. The pump speed i s approximately N = 60Hz divided by the
number of p a i r s of poles, where N i s the pump speed in r e v o l u t i o n s
per minute and Hz is the e l e c t r i c a l frequency in cycles per second
or Hertz. There i s a correction to be made f o r s l i p between the r o t o r
and starter. Slip can be up t o 2 o r 5 percent ( f o r smaller motors).
Three types of AC motors a r e commonly used for driving pumps:
a) s q u i r r e l cage induction motors (asynchronous)
b ) wound r o t o r ( s l i p - r i n g ) i n d u c t i o n motors (asynchronous)
c) synchronous i n d u c t i o n motors.
Squirrel cage motors are simple in design, robust, economic a n d
require minimum maintenance. They a r e general l y more e f f i c i e n t than
the other two types. Their disadvantage is that they require large
starting currents. However if this can be tolerated, a s q u i r r e l cage
motor with d i rect-on-1 ine starter offers the most economical and
reliable combination possible. If direct-on-I ine starting is not
permitted, it i s usual to employ either star-delta or auto-transformer
starters. A l t e r n a t i v e s t a r t e r s compare as follows:-

Type of Starter S t a r t i n g Current S t a r t i n g Torque


( p r o p o r t i o n of f u l l load) ( f u l l load = 1 )

Direct - on 4.5 to 7.0 0.5 to 2.5


Star - Delta 1.2 t o 2.0 0.3 to 0.8
Auto - transformer as r e q u i r e d as r e q u i r e d

Slip-ring induction motors are started by means of resistance


starters and h a v e the a d v a n t a g e that they c a n be smoothly brought
up to full speed with relatively low starting current. They are
normal l y used when s q u i r r e l cage motors a r e impermissible.
Synchronous motors are used on large installations. They have
the a d v a n t a g e of a high load factor. They may be e i t h e r induction
240

motors or salient pole motors. They run at constant synchronous

speed. Starting can be by direct-on-I ine, auto-transformer, resist-


ance o r reactor starters depending mainly on the permitted s t a r t i n g
current.
Variable speed motors are more expensive a n d less e f f i c i e n t than
constant speed motors. Belt drives on smaller units are possible
although commutator motors, or motor resistances on s l i p - r i n g motors
are used as well. Thyristor starters and v a r i a b l e speeds are now
being used with savings in cost for small motors. Pole c h a n g i n g
can also be used for selecting two speeds for squirrel cage motors.

PUMPSTAT IONS

The cost of the structures to house pumps may exceed the cost
of the machinery. While the design details cannot be covered here
the pipeline engineer will need to concern himself with the layout
as he will h a v e to install the suction and delivery of pipework to
each pump with control valves. The capacity of the suction sumps
will effect the operation of the system, and there may be water
hammer protection incorporated in the stat ion. A more complete
discussion of pumpstation layouts is given by Twort et al (1974).
The design of the sump and pump inlet pipework has an import-
ant b e a r i n g on the capacity of a pipeline. Head losses i n t o the pump
can effect its operational efficiency. Air drawn in by vortices or
turbulence could reduce the capacity of the pipeline considerably.
The awareness of the requirements of the entire pumping system
therefore should be the d u t y of the p i p e l i n e engineer.

REFERENCES

Addison, H., 1955. C e n t r i f u g a l a n d Other Rotodynamic Pumps, 2nd


Ed., Chapman a n d H a l l , London, pp 530.
Grist, E . , 1974. Nett p o s i t i v e suction head requirements f o r avoidance
of unacceptable c a v i t a t i o n erosion i n c e n t r i f u g a l pumps, Proc.
C a v i t a t i o n conference, Instn. Mech. Engrs. London, p 153-162.
I n s t i t u t i o n of Water Engineers, 1969. Manual of B r i t i s h Water Engin-
eering Practice. V o l . I I , Engineering Practice, 4th ed. London,
K a r a s s i k , I . J . a n d Carter, R., 1960. C e n t r i f u g a l Pumps, F.W. Dodge,
N.Y. pp 488.
241

Kovats, A., 1964. Design a n d Performance of C e n t r i f u g a l a n d A x i a l


Flow Pumps a n d Compressors, Pergamon Press, Oxford, p p 468.
KSB, Pump Handbook, 1968. K l e i n , Schanzlin 0 Becker, F r a n k e n t h a l ,
p p 183.
Stepanoff, A.J., 1957. C e n t r i f u g a l a n d A x i a l Flow Pumps, Theory,
Design a n d Application, 2nd Ed., Wiley, N.Y., p p 462.
Twort, A.C., Hoather, R.C. a n d Law, F.M., 1974. Water Supply,
Edward Arnold, London, p p 478.
Webber, N.B., 1971. F l u i d Mechanics f o r C i v i l Engineers, Chapman
a n d H a l l , London, p p 340.

L I S T OF SYMBOLS

A,B - constants
BHP - break horsepower
b - width
D - di arneter
E - efficiency

9 - g r a v i t a t ional acceleration
H - head ( s u b s c r i p t a - atmospheric, e - turbulence,
f - friction, p - inlet and s - elevation).

mass per u n i t time


pump r o t a t i o n a l speed
net p o s i t i v e suction head
specific speed
pressure
power
discharge r a t e
radius
torque
r a d i a l component of velocity
p e r i p h e r a l velocity
velocity
unit weight of water p 9. (9800 Newtons per c u b i c metre)
u n i t mass of water
a n g I es
242

GENERAL REFERENCES AND STANDARDS

BR I T I SH STANDARDS

Steel Pipes
BS 534 Steel Pipes, F i t t i n g s and Specials
778 Steel Pipes a n d Joints
1387 Steel tubes and t u b u l a r s f o r screwing
1965Ptl But-welding pipe fittings. Carbon steel
2633 Class 1 a r c welding of f e r r i t i c steel pipes
291 0 Radiographic ex. welded circ. butt joints in steel
pipe
3601 Steel pipes a n d tubes - carbon steel - ordinary duties
3602 Steel pipes and tubes for pressure purposes
Carbon steel : h i g h duties
3603 Steel pipes a n d tubes f o r pressure purposes
Carbon a n d a l l o y steel pipes
3604 Low and medium a l l o y steel p i p e s
3605 Steel pipes a n d tubes for pressure purposes
Austeni t i c stainless steel

Cast I r o n Pipe
78-1 C I spigot a n d socket pipes - p i p e s
78-2 C I spigot a n d socket pipes - f i t t i n g s
143 Malleable C I screwed pipes
41 6 C I spigot a n d socket s o i l , waste a n d vent pipes
437 C I spigot a n d socket d r a i n p i p e s a n d f i t t i n g s
1 130 C I d r a i n f i t t i n g s - spigot a n d socket
121 1 Cast ( s p u n ) i r o n pressure p i p e s
1256 Malleable C I screwed pipes
2035 C I f l a n g e d pipes a n d f i t t i n g s
4622 Grey i r o n p i p e s and f i t t i n g s

Wrought I r o n Pipe
788 Wrought i r o n tubes a n d t u b u l a r s
243

1740 Wrought steel p i p e f i t t i n g s

D u c t i l e I r o n Pipe
L772 Ductile iron pipes a n d f i t t i n g s

Asbestos Cement Pipe


BS L86 A C pressure p i p e s
582 A C soil, waste and ventilating pipes and fittings
201 OP t4 Design a n d construction of A C pressure pipes in l a n d
3656 A C sewer p i p e s a n d f i t t i n g s

Concrete
556 Concrete pipes a n d f i t t i n g s
4101 Concrete unreinforced tubes a n d f i t t i n g s
4625 Prestressed concrete pipes ( i n c . fittings)

Clay Pipe
65 Clay d r a i n a n d sewer pipes
539 Clay d r a i n a n d sewer pipes - f i t t i n g s

540 Clay d r a i n a n d sewer pipes


1143 Salt glazed ware pipes with chemical l y resistant
properties
1196 Clayware f i e l d d r a i n pipes

P l a s t i c a n d Other Pipe
1972 Polythene p i p e ( t y p e 3 2 ) for c o l d water services
1973 Spec. polythene pipe (type 425) for general purposes
3284 Polythene p i p e ( t y p e 50) f o r c o l d water services
3505 UPVC pipes ( t y p e 1420) f o r cold water supply
3506 UPVC p i p e f o r i n d u s t r i a l purposes
3796 Polythene p i p e ( t y p e 5 0 )
3867 D i m s . of p i p e s O.D.
4346 Joints a n d f i t t i n g s for UPVC pressure p i p e
4514 UPVC s o i l a n d v e n t i l a t i n g p i p e

4660 UPVC underground a n d d r a i n p i p e


244

4728 Resistance t o c o n s t a n t i n t e r n a l p r e s s u r e of
thermopl a s t i c p i p e
2760 Pitch f i b r e pipes and couplings
CP312 P l a s t i c pipework

l n s u l at i o n
1334 Thermal insulation
4508 Thermal l y i n s u l a t e d u n d e r g r o u n d p i p i n g systems
CP3009 Thermal l y i n s u l a t e d u n d e r g r o u n d p i p i n g systems

Valves
8s 1010 Taps a n d v a l v e s f o r water
1212Pt2 8 a l l valves - diaphragm type
1218 &
5163(m) S I ui ce v a I ves
1415 V a l v e s f o r domestic p u r p o s e s
1952 &
51 54( m) Copper a l l o y g a t e v a l v e s
1953 Copper a l loy check v a l v e s
2060 Copper a l l o y stop v a l v e s
2591 G l o s s a r y of v a l v e s
3464 &
5150(rn) C I wedge a n d d o u b l e d i s c v a l v e s
3948 &
5151 ( m ) C I parallel slide valves
3952 &
5155(m) C I butterfly valves
3961 &
5152(rn) C I stop and check v a l v e s
4090 &
51 53( rn) C I check v a l v e s
4133 &
5157(rn) F l a n g e d steel p a r a 1 l e l s l i d e v a l v e s
431 2 Stop a n d check v a l v e s

5156(m) Screwdown d i a p h r a g m v a l v e s
5158(rn) Plug valves
51 59(m) Bal I valves
245

Jointing

10 Flanges a n d b o l t i n g f o r pipes etc.


21 Threads
1737 Jointing m a t e r i a l s a n d compounds
1821 Oxy-acetylene welding of steel p i p e l i n e s
1965 B u t t we I d i ng
2494 Rubber j o i n t rings
3063 Dimensions of gaskets

4504 Flanges a n d b o l t i n g f o r pipes etc.

Miscellaneous
1042 Flow measurement

1306 Non-ferrous pipes f o r steam


1553 Graphical symbols
1710 I d e n t i f i c a t i o n of p i p e l i n e s
205 1 Tube a n d p i p e f i t t i n g s
29’17 Graphical symbol s
3889 Non destructive testing of pipes a n d tubes
39 74 Pipe supports
4740 Control v a l v e c a p a c i t y

SOUTH AFRICAN BUREAU OF STANDARDS

Steel Pipe
SABS 62 Steel pipes a n d f i t t i n g s up to 150 mm
71 9 E l e c t r i c a l welded low carbon steel p i p e s
720 Coated a n d l i n e d m i l d steel p i p e s

Cast I r o n Pipe
50 9 Malleable C I pipe f i t t i n g s
746 C I soil, waste, water a n d vent pipes
81 5 Shouldered end pipes, f i t t i n g s and j o i n t s

Asbestos Cement Pipe


546 C I f i t t i n g s f o r A C pressure p i p e s
72 1 A C soil, waste and vent p i p e s a n d f i t t i n g s
246

946 A C pressure pipe - constant internal diameter type


286 A C pressure pipes - constant outside diameter type
819 A C sewer p i p e s

Concrete Pipe
676 R C pressure p i p e s
677 Concrete non-pressure pipes
975 Prestressed concrete pipes
902 Structural design and i n s t a l l a t ion of precast concrete
pipelines

Glazed Earthenware Pipe


559 Glazed earthenware d r a i n a n d sewer pipes a n d f i t t i n g s

P l a s t i c Pipe
SABS 791 UPVC sewer a n d d r a i n p i p e f i t t i n g s
92 1 P i t c h impregnated f i b r e pipes
966 UPVC pressure p i p e
967 UPVC s o i l , waste and vent pipes
997 UPVC pressure p i p e s f o r i r r i g a t i o n
01 12 I n s t a l l a t i o n of PE a n d UPVC p i p e s
533 Black polyethylene p i p e s

Va I ves
144 C I s i n g l e door r e f l u x v a l v e s
191 Cast steel g a t e v a l v e s
664 C I gate v a l v e s

AMER I CAN WATER WORKS ASSOC I AT I ON

A W W AC20 1 F a b r i c a t i n g e l e c t r i c a l l y welded steel water p i p e


c202 M i l l type steel water p i p e

C203 Coal-tar enamel protective coatings for steel water


pipe
C205 Cement mortar protective coatings for steel water
p i p e of sizes 30" a n d over
247

C206 F i e l d w e l d i n g o f steel water p i p e j o i n t s


C207 Steel p i p e f l a n g e s
C208 Dimensions f o r steel w a t e r p i p e f i t t i n g s
C300 R e i n f o r c e d c o n c r e t e w a t e r p i p e - Steel c y l i n d e r
type - not prestressed
C301 R e i n f o r c e d c o n c r e t e w a t e r p i p e - Steel c y l i n d e r
t y p e - prestressed
C302 R e i n f o r c e d c o n c r e t e w a t e r p i p e - Non-cyl i n d e r
type - not prestressed
C600 -

54 T I n s t a l l a t i o n of C I watermains

C602 Cement m o r t a r l i n i n g o f w a t e r p i p e l i n e s i n p l a c e
( 1 6" and o v e r )

AMER ICAN PETROLEUM INSTITUTE

API Std.5A Spec. f o r Casing, T u b i n g and D r i l l P i p e


Std.5AC Spec. f o r G r a d e C - 75 and C95 C a s i n g and T u b i n g
Std.5AX Spec. f o r H i g h S t r e n g t h C a s i n g and T u b i n g
Std.5L Spec. for Line Pipe
Std.5LA Spec. f o r S c h e d u l e 5 Alum. Alloy Line Pipe
Std.5LP Spec. for Thermoplastic L i n e Pipe
Std. 5LR Spec. for Glass Fibre Reinforced Thermosetting Resin
Line Pipe
Std.5LS Spec. f o r S p i r a l Weld L i n e P i p e
Std.5LX Spec f o r H i g h Test L i n e P i p e
RP5C I C a r e and Use o f C a s i n g , T u b i n g and D r i l l P i p e
RP5L I R a i l r o a d t r a n s p o r t of L i n e Pipe
RP5L2 I n t e r n a l C o a t i n g of L i n e P i p e f o r Gas T r a n s m i s s i o n
RP5L3 Conducting Drop Weight Tear Tests o n L i n e P i p e
Bul.5C2 Performance Properties of Casing, Tubing and DrilI
Pipe
Bul.5T1 Non-destructive Testing Terminology
248

AMER I CAN SOC I ETY FOR TEST I NG MATER I ALS

Concrete P i p e s
ASTMC14 Concrete Sewer, Storm D r a i n a n d C u l v e r t P i p e
C76 Reinforced Concrete Culvert, Storm Drain and Sewer
Pipe
C118 Concrete P i p e f o r I r r i g a t i o n or Drainage
C361 R e i n f o r c e d Concrete Low-Head Pressure Pipe
C412 Concrete Cjrain T i l e
c443 Joints for Circular Concrete Sewer and Culvert Pipe
w i t h Rubber gaskets
c444 P e r f o r a t e d Concrete P i p e
c497 Determining Physical Properties of Concrete Pipe or
T i le
C505 Non-reinforced Concrete Irrigation Pipe and Rubber
Gasket J o i n t s
C506 Reinforced Concrete Arch Culvert, Storm Drain and
Sewer P i p e

C507 Reinforced Concrete El I i p t i c a l Culvert, Storm Drain


a n d Sewer P i p e
C655 Reinforced Concrete D-load Culvert, Storm Drain and
Sewer P i p e

Steel P i p e s
ASA 836.10 Steel p i p e s

Cast i r o n Pipes
A1 42 C I Pipes
A377 C I pipes
A121.1 C I pipes

Asbestos Cement P i p e s
C296 A C Pressure Pipes
C500 A C Pressure Pipes
C428 A C Pipes and Fittings for Sewerage and Drainage
249

Plastic Pipes
D2241 U n p l a s t i c i s e d PVC p i p e s

USDI BUREAU OF RECLAMATION


S t a n d a r d Spec. f o r Reinforced Concrete P r e s s u r e P i p e
250

BOOKS FOR FURTHER READING

Addison, H., 1964. A treatise on Applied Hydraulics, Chapman H a l l ,


London
Albertson, M.L., Barton, J.R. and Simons, D.B., 1960. F l u i d Mech-
a n i c s f o r E n g i n e e r s , P r e n t i c e - H a l I, 1960.
American Concrete Pipe Association, 1970. Design Manual-Concrete
Pipe, Arlington.
Am.Soc.Civi I Engs. and W a t e r Pol In. Control Federation, 1970. Design
and C o n s t r u c t i o n o f S a n i t a r y and Storm Sewers, M a n u a l 37, N.Y.
Am.Water Works Assn., 1964. Steel Pipe - Design and Installation,
M a n u a l M11, N.Y.
Bell, H.S., 1963. (Ed.), Petroleum Transportation Handbook, McGraw
H i l l , N.Y.
Benedict, R.P., 1977. F u n d a m e n t a l s of Pipe Flow, Wiley Interscience,
531 pp.
Bureau of Public Roads, 1963. Reinforced Concrete Pipe Culverts -
Criteria for Structural Design and Instal lation, U.S. Govt. Printing
Of f i ce, Was hi n g t o n , DC.
Clarke, N.W.B., 1968. Buried Pipelines - A Manual of Structural
D e s i g n a n d I n s t a l l a t i o n , M a c L a r e n and Sons, London.
Colorado State Univ., 1971. Control of Flow in Closed Conduits,
Proc. Inst., Fort Collins.
Crocker, S. and K i n g , R.C., 1967. P i p i n g Handbook, 5 t h Ed., McGraw
Hill.
Davis, C.V. and Sorensen, K.E., 1969. Handbook of Applied Hyd-
raulics, 3 r d Ed., McGraw H i l l , N.Y.
Holmes, E., 1973. Handbook of Industrial Pipework Engineering,
McGraw H i l l , N.Y.
Instn. Water Engs., 1969. Manual of British Water Engg. Practice,
4 t h Ed., London.
Littleton, C.T., 1962. Industrial Piping, McGraw H i l l , 349 pp
Martin, W.L., 1961. H a n d b o o k of I n d u s t r i a l Pipework, Pitman, London.
Nolte, C.B., 1978. Optimum Pipe Size Selection, Trans Tech Publi-
cations, 297 p p .
Rouse, H., 1961. Engineering Hydraulics, Wiley, N.Y.
251

Stephenson, D., 1984. Pipeflow Analysis, E l s e v i e r , 204 pp.

Streeter, V.L., 1961. (Ed), Handbook of Fluid Dynamics, McGraw


Hill, N.Y.
Twort, A.C., Hoather, R.C. and Law, F.M., 1974. Water Supply,
2 n d Ed., A. A r n o l d , London.
Walski, T.M., 1984. Analysis of Water Distribution Systems. Van
Nostrand Reinhold, N.Y. 275 pp.
Walton, J.H., 1970. Structural Design of Vitrified Clay Pipes, Clay
P i p e Development Association, London.
Watters, G.Z., 1984. Analysis and control of unsteady flow in
pipelines. 2 n d Ed. Butterworths.
Young, D.C. and Trott, J.J., 1984. Buried Rigid Pipes. Elsevier
A p p l i e d Science, London, 230 pp.
252

APPEND I X

SYMBOLS FOR P I P E F I T T I N G S

- -
GENERAL

ELBOW Jo. BEND

JPtKETED
A-
SLEEVED

TEE
FILLET WELDEOTEE
1 1
FRONT VIEW OF TEE

BACK VIEW OF TEE

CROSS OVER I HANGER

I YMPLE SUPWRT
BELLMOUTH
D CHANGE IN D I A
TAPER

JOINTS

E L E CT RICALLY BONDED
FLANGED

ELECTRIULLY
INSULATED
BMT W E L D --+-
FLEXIBLE SCREWED _3_

SWIVEL SOCKET
-4-
E X PANS ION SPIGOT 0 SOCKET
_f_

S L E E V E COUPLING
END CAP
D

__
VALVES

EUl TERFLY
ISOLATING

NEEDLE
WEDGE GATE

GLOBE
REFLUX

DIAPHRAGM
ROTARY PLUG

AIR
RELIEF
w
MISCELLANEOUS

STRAINER
HYDRANT

FLOW INDICATOR
SPRAY
--+
VENT
J--
+
SURFACE BOX

DRAIN P L A T E BLIND
t
M 0 TOR
0 HANDWHE E L T
253

PROPERTIES OF P I P E SHAPES

-
Ring
AREA - F(D2-d2kTDt fa -11 I

MOMENT OF INERTIA ABOUT DIAMETER = fi (0'- d') =TD"

MOMENT O F INERTIA ABOUT CENTRE g ( D b - d L ) Z s D a t


PROPERT I ES OF WATER

Temperature Specific Mass Kinematic Viscosity B u l k Modulus Vapour Pressure

C O F" kg/rn3 Ib/cu f t rn2/ s s q ft/sec N/rnm2 psi N/rnmz psi

0 32 1 000 62.4 1.79~10-~ 1 .93x 1 0-5 2 000 290 000 0.6~10-~ 0.09
10 50 1 000 62.4 1 .31XI
0-6 1.41~10-~ 2 070 300 000 l.2~lO-~ 0.18
6
20 68 999 62.3 1.01x10- 1.09~10-~ 2 200 318 000 2 . 3 ~ 1 0 ~ ~ 0.34
3
30 86 997 62.2 0.81 x10-6 0.87~1
0-5 2 240 325 000 4.3~10- 0.62
PROPERTIES OF P I P E MATERIALS

Coef. of Density Modulus of elasticity Y i e l d Stress Tensile Strength Poisson's


exp.per O C kg/m' N/mrn2 psi N/rnrn2 psi N/mm2 psi ratio

6
Clay 5x10
Concrete 1ox10-6 2 600 14 000-40 000 2x106-6x106 -70 -10 000 -2.1 -300 0.2
6
Asbestos cernen t 8.5~
1 0-6 2 500 24 000 3.5~10 17 2 500
6
Cast Iron 8 . 5 ~ 0-6
1 7 800 100 000 15x10 150 22 000 225 23 000 0.25
M i l d Steet 1 1 .9x10-6 7 850 210 000 3 1 ~ 1 0 ~ 210 30 000 330 48 000 0.3
High Tensile
6
Steel 1 1 .9x10-6 7 850 210 000 31x10 1 650 240 000 1730 250 000
Pitch Fibre 40x 1 0-6
Polyethylene 1 60x 1 0-6 900 138 20 000 3 400 14 2 000
Po I y e t h y l e n e
(high density) 2oox 1 o-6 955 240 35 000 22 3 200 32 4 600 0.38
PVC 50~10-~ 1 300 8 1 100 17 2 500
6
UPVC 50~10-~ 1 400 3 500 0.5~10 40 + 6 000 52 7 500
Po I y p r o p y I ene 1 8ox10-6 91 2 150 22 000 25 3 600 30 4 500
256

CONVERSION FACTORS

Length 1 i n c h = 25.4 mm
1 ft = 0.3048 m
1 m i l e = 1.61 k m
Area 1 sq i n c h = 644 mrn'
1 a c r e = 2.47 ha
1 sq ft = 0.0929 mz
1 ha = lo4 mz
Vol ume 35.31 c u f t = 1 m3
1 gal.(imperial) = 4.54 I i t r e s
1 US g a l . = 3.79 I i t r e s
1 barrel = 42 US gal. = 35 i m p . g a l .
= 159 I i t r e s
Speed 1 ft/sec = 0.3048 m/s
1 m p h = 1.61 k m / h r .
Acceleration 32.2 f t / s e c 2 = 0.981 m/s' = g
Discharge 1 mgd (imperial) = 1.86 cusec
35.3 c u s e c = im3/s
13.2 g p m ( i m p e r i a l ) = 1 P / s
Mass 1 I b = 0.454 k g
32.2 I b = 1 s l u g (US)
Force 1 Ib. f o r c e = 4.45 N
( 1 Newton = 1 k g x 9.81 m / s ' )
Pressure 145 p s i = 1 MPa = 1 MN/m2 = 1 N/mmz
14.5 p s i = 1 b a r
Energy 778 f t 16. = 1 B t u
1 Btu = 252 c a l o r i e s
1 c a l o r i e = 4.18 J o u l e s
Power 550 f t Ib/sec = 1 HP
1 HP = 0.746 kW
Kinematic Viscosity 1 sq ft/sec = 929 s t o k e s
= 0.0929 m z / s
Absolute o r dynamic viscosity 1 c e n t i p o i s e = 0.001 k g / m s
Temperature F o = 32 + 1.8C"
257

A b s o l Ute temperature R "


(Ranki ne) = FO(Fahrenheit1 + 460
A b s o l Ute temperature KO
(Kelvin) = Co(Centigrade) + 273
71 = 3.14159
e = 2.71828
258

AUTHOR INDEX

Abrarnov, N. 15 J a c o b s o n , S. 155, 157


A d d i s o n , H. 228, 241, 250 Johnson, S.P. 95
A l b e r t s o n , M . L . 17, 35, 250 Joukowsky , 60
A n d e r s o n , O.A. 125
A v e r y , S.J. 105, 110 K a l i n s k e , A . A . 102, 104, 111
K a l l y , E. 48, 57
B a l l , J.W. 28, 35 K a r r a s s i k , I.J. 228, 240
B a r n a r d , R.E. 149, 157 K e n n e d y , H. 121, 125
B a r t o n , J.R. 17, 35 K e n n e d y , J.F. 74, 95
B e l l , H.S. 250 K e n n i s o n , H.F. 129, 140
B e n e d i c t , R.P. 250 K i n g , C.L. 177
B e r n o u l I i , 16 K i n g , R.C. 173, 177
B e r t h o u e x , P.M. 10, 15 K i n n o , H. 74,95
B l i s s , R.H. 101, 111 K n a p p , R. 28, 35
B o u c h e r , P . L . 182 203 K o v a t s , A . 228, 241
B u r a s , N. 49, 57
L a i , C . 66, 95
C a p p e r , P . L . 123, 125, 191, 203 L a m , C.F. 56, 57
C a s s i e , W.E. 123, 125, 191, 203 L a w , F.M. 240, 250
C a t e s , W.H. 176, 77, 216, 226 L e s c o v i c h , J.E. 186, 203
C h a p t o n , H.J. 177 L i t t l e t o n , C.J. 250
C h e z y , 18 L u d w i g , H. 68, 95
C l a r k e , N.W.B. 113, 125, 250 L u p t o n , H.R. 64, 95
Cole, E.S. 218, 226
Col e b r o o k , 23 M a n n i n g , 18
C r o c k e r , S. 173, 177, 250 M a r k s , L.S. 107, 1 1 1
Cross, H. 38,57 Marston, A. 114, 125
M a r t i n , W . L . 250
D a n t z i g , G.B. 51, 57 M i l l s , K.G. 32, 43, 57
D a r c y , 21 Moody, 21
D a v i s , C.V. 250 M o r l e y , A . 154, 157
D e n n y , 9 . F . 101, 110 M o r r i s o n , E.B. 190, 203
D i s k i n , M.H. 24, 35
D v i r , Y . 30, 35 N e l s o n , E.D. 177
N e w m a r k , 122
Ervine, D.A. 105, 110 Newton, 16, 58
N i k u r a d s e , 21
F o x , J.A. 109, 110 N o l t e , C.B. 250
F e r g u s o n , P.M. 137, 140
Osborne, J.M. 7, 15
Glass, W.L. 110
G o o d i e r , J.N. 177 P a r m a k i a n , J. 73, 95, 108, 111,
G r i s t , E . 232, 240 186, 203
P e a r s o n , F.H. 177
H a s s a n , D . R . 165, 177 P a u l , L. 182, 203
Hazen-VJ i I I i ams , 18 P o i s s o n , 176, 195
H o a t h e r , R.C. 240, 250 P r o c t o r , 150
Holmes, E. 250 P r o s s o r , 1vl.J. 101, 111

Isaacs, L.T. 43, 57 R e i t z , H.M. 208, 226


259

R e y n o l d s , 20
R e y n o l d s , G.M. 211, 226
R i c h , G.R. 78, 95
R i d d i c k , T.M. 224, 226
R o a r k , R.J. 174, 177
R o b e r t s o n , J.M. 104, 111
Rouse, H. 166, 177, 250

S c h a r e r , H. 176, 177
S c h l i c h t i n g , H. 20, 35
S c h n e i d e r , W.R. 219, 226
S c h w a r t z , H . I . 211, 226
S c h w e i g , 2. 48, 57
Sirnons, D.B. 17, 250
Sorenson, K.E. 250
Sowers, G.F. 208, 226
S p a n g l e r , M.G. 113, 125, 146, 152,
157
S t e p a n o f f , A.J. 228, 241
Stephenson, D. 56, 57, 69, 80, 9 5 ,
110, 1 1 1 , 151 TSSYJ, 157, 165, 177,
25 1
S t r e e t e r , V.L. 66, 95, 251
Suss, A. 165, 177
Swanson, H.S. 162, 177
Sweeton, A.E. 186, 203

Tirnoshenko, S . P . 134, 140, 158, 160, 177


T r o t t , J.J. 113, 125, 251
T u l l i s , J.P. 35
T w o r t , A.C. 240

Uhlig, H.H. 218, 226

Van d e r Veen, B. 51, 57

W a l s k i , T.M. 251
Walton, J.H. 250
Watson, M.D. 24, 35
W a t t e r s , G.Z. 251
Webber, N.B. 229, 241
Weisbach, 21, 109
W h i t e , 23
W h i t e , J.E. 9, 15
W i l k i n s o n , W.J. 177
W i n n , W.P. 33, 35
W i s n e r , P. 102, 111
W o i n o w s k i - K r i e g e r , S. 134, 140, 157
Wood, D.J. 43, 57
W y l i e , E.B. 66, 95

Young, D.C. 113, 125, 251


Young, G.A.J. 101, 110
260

SUBJECT INDEX

A b s o r p t i o n 99 Chemicals 1
A c t i v e s o i l p r e s s u r e 114, 149, 191 C l R l A 152
A c t u a t o r 92, 185 Clamp 212
A d i a b a t i c 86, 107 C l a y 255
Aesthetics 2 C o a t i n g , bi tumen 21 6
A i r pocket 89, 100 c o a l t a r 217
v a l v e s 106, 189 epoxy 217
vessel 87 g l a s s f i b r e 217
American Concrete P i p e Assn. 122, m o r t a r 23, 217
125, 140 t a p e 217
Anode, s a c r i f i c i a l 219 Cohesion 191
A p p u r t e n a n c e s 182 Col l a p s e 148
Asbestos cement 180 C o l l a r 143, 161
A x i a l e x p a n s i o n 176, 195 Concrete p i p e s 127, 180
p r e s t r e s s e d - 129 180
B a c k f i l l 113, 192 p r o p e r t i e s 137, 255
B a r g e , 211 Compound p i p e s 37
Beam 213 Compressibi I i t y 16, 60
B e a r i n g t e s t 128 Computers, a n a l o g u e 202
B e d d i n g 115. 127 d i g i t a l 41, 202
B e d d i n g f a c t o r 128, 207 Concrete P i p e Ass. 122, 125
Bend loss 27 C o n d u c t i v i t y 219, 223
meter 200 Conductor 219
s t r e s s 173 Cone v a l v e 30, 184
B e n d i n g 144, 196, 213 Consol i d a t i o n 149
B i n a r y n u m b e r 202 C o n s t r a i n t 52
B i t u m e n 217 Container 2
B o n i n g 206 C o n t r a c t i o n c o e f f i c i e n t 200
B o u n d a r y l a y e r 22 Convect i o n 224
B r a c i n g 161 Conversion f a c t o r s 256
B r a n c h 62, 68, 161 Core 129
B r i d g e 208 C o r r o s i o n , g a l v a n i c 218
B u b b l e 99 Cost, p i p e 3
Buck1 i n g 152, 168 power 4, 12
B u t t e r f l y v a l v e 27, 32, 184 p u m p i n g 4, 8
B y p a s s 76, 79, 182 C o u n t e r b a l a n c e 186
C o u p l i n g 230
Cant i l e v e r 174 C r a c k 127
Capacity factor 4 Creep 130
Cash f l o w 1 1 Cross section 253
C a s i n g 217 C r o t c h p l a t e 161
Cast i r o n 180 C r u s h i n g 137
C a t h o d i c p r o t e c t i o n 218 Current, electric- 219
C a v i t a t i o n 30, 97 impressed- 21 9
C e l e r i t y 60 stray- 220
C e n t r i f u g a l f l o w 200 C y l i n d e r 129
f o r c e 190
pump 229 D e f l e c t i o n 146, 150
C h a r a c t e r i s t i c method 66 Design f a c t o r 142
pump 70, 73, 229 O i ameter 4
261

D i s c h a r g e c o e f f i c i e n t 200 H a u n c h 128, 148


t a n k 79 Head, loss 16, 37, 199, 232
Discount r a t e 9 v e l o c i t y - 16, 232
D i s s o l v e 99 w a t e r hammer- 62
D r a i n 207 H e a d i n g 208
D y n a m i c p r o g r a m m i n g 45 Heat loss 224
History 1
Hole 161
Economics 2
E l a s t i c i t y 64, 255 Hol i d a y d e t e c t o r 21 7
E l e c t r o d e 220 H y d r a u l i c g r a d i e n t 18, 68, 238
E I ec tro-chemi c a l e q u i v a l e n t 220 H y d r a u l i c j u m p 104
E l e c t r o l y t e 218
E l e c t r o m a g n e t i c i n d u c t i o n 201 I m p a c t f a c t o r 121
El l i p s e 253 I m p e l l e r 229
Embankment 1 1 6 I n e r t i a 73, 94
Ernrnissivity 224 I n f l u e n c e c o e f f i c i e n t 122
E m p i r i c a l 18 I n s u l a t ion, t herrna I 224
E n e r g y 16 I n t e r e s t 10
E n t r a n c e l o s s 27 I sotherrnal 86
Epoxy 217
E q u i v a l e n t d i a m e t e r 37 J o i n t , b u t t - w e l d e d 212
Escher Wyss 177 clamp-on 214
E x c a v a t i o n 206 f a c t o r 142
E x t e r n a l l o a d 115, 146 f l a n g e d 215
screwed 213
F a b r i c a t e d b e n d 194 sleeve w e l d e d 212
F a c t o r o f s a f e t y 143 s p i g o t a n d socket 130, 214
F i e l d p r e s s u r e 133
F l a n g e 212 K i n e t i c e n e r g y 72
b l a n k 216
p u d d l e 216 L a g g i n g 224
F l e x i b l e p i p e 142 L a m i n a r f l o w 20
F l o a t i n g 210 L a t e r a l s u p p o r t 149
F i t t i n g s 182 L a y i n g 205
Flow measurement 198 L i n e l o a d 122
b e n d meter 200 L i n e a r p r o g r a m m i n g 52
e l e c t r o m a g n e t i c 201 L i n i n g , b i t u m e n 217
mass 201 c o a l t a r 217
m e c h a n i c a l 200 epoxy 218
n o z z l e 199 m o r t a r 21 8
o r i f i c e 199 L i v e l o a d 120
v e n t u r i 198 L o a d c o e f f i c i e n t 114
volume 202 soi I 113
Flow r e v e r s a l 75, 91 superimposed 120
F l y w h e e l 76 Longitudinal stress 134, 198,
F r i c t i o n 18, 68, 74 21 3
F r o u d e n u m b e r 104 Loop 38
Loss c o e f f i c i e n t 27
G a l v a n i c a c t i o n 218 Losses 26
Gasket 215
Gate v a l v e 182 Mass, c o n s e r v a t i o n of 16
Globe v a l v e 184 M a t e r i a l s 255
G r a p h i c a l a n a l y s i s 14, 65 M e c h a n i c a l meter 200
262

Membrane t h e o r y 167 P r e s s u r e , e x t e r n a l 110, 146


Meter b e n d 200 i n t e r n a l 142
rnkchan i c a l 200 w a t e r hammer 58
n o z z l e 199 P r e s t r e s s e d c o n c r e t e p i p e 129
o r i f i c e 199 P r o c t o r d e n s i t y 150, 207
r o t 0 200 Profile, pipeline 72, 106, 205,
v a n e 200 238
v e n t u r i 198 P r o t e c t i o n , c a t h o d i c 218
Model 165 w a t e r hammer 69
M o d u l u s of e l a s t i c i t y 137, 255 P u d d l e f l a n g e 216
M o d u l u s of s u b q r a d e r e a c t i o n 124 Pump, b o o s t e r 2, 9
Mohr d i a g r a m 136 c e n t r i f u g a l 75, 229
Momentum 16, 90 i n e r t i a 73
Motor 239 power 5
t y p e s 228
Needle v a l v e 91, 184 P u m p i n g c o s t s 3, 12
Network 38
Net p o s i t i v e s u c t i o n h e a d 232 R a d i a l s t r e s s 142
Node 39 R a d i a t i o n 224
Non-l i n e a r p r o g r a m m i n g 56 R a d i o 202
Normal t h r u s t 190 R a d i u s o f s t i f f n e s s 124
Nozzle 199 Rail 1
Re