Cavitation

and the
Centrifugal
Pump
Cavitation
and the
Centrifugal
Pump

A Guide
for
Pump Users

Edward Grist
 USA Publishing Office: TAYLOR & FRANCIS
325 Chestnut Street
Philadelphia, PA 19106
Tel: (215) 625-8900
Fax:(215)625-2940

Distribution Center: TAYLOR & FRANCIS

47 Runaway Road, Suite "G"
Levittown, PA 19057-4700
Tel: (215) 269-0400
Fax:(215)269-0363

UK TAYLOR & FRANCIS

1 Gunpowder Square
London EC4A 3DE
Tel: +44 171 583 0490
Fax: +44 171 583 0581

CAVITATION AND THE CENTRIFUGAL PUMP: A Guide for Pump Users

Copyright © 1999 Taylor & Francis. All rights reserved. Printed in the United States of America.
Except as permitted under the United States Copyright Act of 1976, no part of this publication may
be reproduced or distributed in any form or by any means, or stored in a database or retrieval system,
without prior written permission of the publisher.

1 2 3 4 5 6 7 8 9 0

Edited by Edward A. Cilurso and Jean Anderson. Cover design by Joseph Dieter. Printed by
Edwards Brothers, Ann Arbor, MI, 1998.

A CIP catalog record for this book is available from the British Library.
The paper in this publication meets the requirements of the ANSI Standard Z39.48-1984
(Permanence of Paper).

Library of Congress Cataloging-in-Publication Data

Grist, Edward.
Cavitation and the centrifugal pump: a guide for pump users /
Edward Grist,
p. cm.
Includes bibliographical references and index.
ISBN 1-56032-591-7 (alk. paper)
1. Centrifugal pumps - Cavitation - Handbooks, manuals, etc.
I. Title
TJ919.G75 1998 98-17168
621.6'7-dc21 CIP

ISBN 1-56032-591-7 (case)

 CONTENTS

Preface xiii

Acknowledgements xiv

Nomenclature xv

Chapter 1 Cavitation — An Unacceptable Phenomenon 1

1.1 Cavitation Defined 1
1.2 Unacceptability Defined 1
1.3 The Categories of Unacceptability — A Brief Resume 3
1.3.1 Category 1. Hydraulic performance loss 3
1.3.2 Category 2. Cavitation surging — Hydrodynamically induced 5
1.3.3 Category 3. Cavitation surging — Thermodynamically induced 6
1.3.4 Category 4. Cavitation erosion 9

Chapter 2 Independent Variables and Terminology 11

2.1 Primary Independent Variables 11
2.2 Secondary Independent Variables 12
2.3 Hydraulic Configuration Conventions 13
2.3.1 Centrifugal pump impeller design terminology 13
2.3.2 Centrifugal pump impeller design configurations 15
2.4 Performance Measurement Levels 15
2.4.1 Level 1. Impeller performance (AH, Q) 15
2.4.2 Level 2. Pump performance (AHG, QG) 17
2.4.3 Level 3. Plant performance (AHp, Qp) 17
2.5 Performance Measurement Terminology 18
2.6 Metric Units 20
2.7 Useful Conversions 21

Chapter 3 Centrifugal Pump Performance Characteristics 21

PART 1. NON-CAVITATING PERFORMANCE
3.1 A Typical Characteristic 23
3.2 Affinity Relationships and Specific Speed 23
3.2.1 Affinity laws 24
3.2.2 Specific speed 26
3.3 Performance Curves — Generic Differences 27
3.4 Other Influences 32
3.4.1 Kinematic viscosity 32
3.4.2 Impeller surface roughness 34

V
vi Contents

PART 2. CAVITATING PERFORMANCE

3.5 The Three Cavitating Flow Regimes 34
3.5.1 The stable regime 34
3.5.2 The unstable regime 37
3.5.3 The transient regime 37
3.6 Typical Cavitation Characteristics 38
3.6.1 Measurable head-drop curves 38
3.6.2 Measurable visual onset curves 41
3.6.3 The relationship between head-drop and visual performance 41
3.7 Affinity Relationships and Suction Specific Speed 45
3.7.1 Limits of validity 45
3.7.2 Affinity laws 45
3.7.3 Suction specific speed 46
3.8 Performance Curves and Generic Similarity 46
3.9 Other Influences — Thermodynamic Properties Affecting 47
Cavity Dynamics

Chapter 4 Net Positive Suction Head (NPSH)

and The Pump Operational Range 51
4.1 NPSH as a Measure of Cavitation 51
4.1.1 Role and Scope 51
4.1.2 NPSH, NPSH(A), NPSH(R) and NPSH datum values explained
4.1.3 NPSH datum conditions 51
4.1.4 Industrial NPSH tests 52
4.2 The Pump Operational Range 56
4.2.1 An overall view 56
4.2.2 The pump design duty 57
4.2.3 The pump operating point 57
4.2.4 Methods of varying the pump flowrate 59
4.2.5 Start-up conditions 61
4.2.6 Pressure changes at the liquid surface of inlet or outlet vessels 62
4.2.7 The system resistance curve equation 62
4.3 Putting a Value to NPSH 63
4.3.1 For pump testing — To establish datum values for NPSH(3pc) and
NPSH(4mm) 63
4.3.2 For plant design — To establish the value of NPSH(A) 63
 Contents vii

Chapter 5 Cavity Dynamics — A Simplified Approach 71

5.1 Background 71
5.2 Understanding the Notional Spherical Cavity Concept 71
5.3 The Rayleigh Equation 72
5.4 The Plesset Equation 73
5.5 Modelling Cavitation in a Centrifugal Pump 75
5.5.1 The driving pressure producing cavity growth 75
5.5.2 The pressure contour in impeller passages 76
5.5.3 Spherical cavities growing and collapsing in a stationary liquid 77
5.5.4 Spherical cavities in a liquid constrained to move along a duct 77
5.5.5 Spherical cavities in a liquid moving unconstrained in a
superimposed pressure gradient which increases directly as the
square of the distance travelled 78
5.6 Predicting Changes in Performance using the Model 79
5.6.1 Calculational objectives 79
5.6.2 Calculational output 79
5.6.3 The criterion for performance comparability 80
5.6.4 Numerical solutions by computer 80
5.7 Typical Calculated Volumetric Characteristics 81
5.8 Predicted and Measured Performance Compared 84
5.9 Potential Benefits — Are They Real in Practice? 85
5.10 Applying the Findings in Practice 87

Chapter 6 Hydraulic Performance Loss —

Duty Shortfall and Vapour Locking 89
6.1 Deficiencies Identified 89
6.2 Hydraulic Performance Protection — A Pump User Approach 89
6.2.1 The minimum level of protection 89
6.2.2 NPSH(3pc) cavitation test requirements 90
6.3 Pump Duty Shortfall 91
6.3.1 The underlying cause 91
6.3.2 An avoidance strategy 91
6.3.3 Learning by example 92
6.3.4 A practical method of accommodating some calculation uncertainties 93
6.4 Pump Vapour Locking 94
6.4.1 The vapour locking phenomenon and the underlying cause 94
6.4.2 An avoidance strategy 96
6.4.3 The calculation of ANPSH by the step-by-step method 97
6.4.4 Learning by example 99
6.4.5 Dealing with uncertainties and options 106
6.5 Vapour Locking Protection — A Pump User Approach 108
6.6 Ebullition in the Inlet Pipework (Vapour Bubble Formation) 110
6.7 Inlet Vessel Nozzle Cavitation 112
viii Contents

Chapter 7 Cavitation Surging — Hydrodynamically Induced 113

7.1 Deficiencies Identified 113
7.2 The Surge Mechanism 114
7.2.1 The underlying cause 114
7.2.2 Research and development test observations — Centrifugal impellers 116
7.2.3 Research and development test observations —
Inducers and axial flow impellers 123
7.2.4 Generic analysis 126
7.2.5 Backflow recirculators 127
7.2.6 The influence on cavitation surging of pump design configuration 128
7.2.7 Summary of findings 129
7.3 Cavitation Surging (Hydrodynamically Induced) — An Avoidance Strategy 130

Chapter 8 Cavitation Surging — Thermodynamically Induced 133

8.1 Deficiencies Identified 133
8.2 The Pressure Pulsation Phenomenon 134
8.2.1 An outline description 134
8.2.2 A mathematical model 137
8.2.3 Outlet non-return valve response 138
8.2.4 Practical significance explored through numerical examples 139
8.3 The Influence of Plant Design 142
8.4 Vapour Locking Tests 144
8.4.1 Tests on high-energy pumps 144
8.4.2 Tests on small low-powered pumps 145
8.4.3 Re-establishing safe operating conditions 145
8.5 Summary of Findings 146
8.6 Cavitation Surging (Thermodynamically Induced) — An Avoidance Strategy 147
Chapter 9 Cavitation Erosion 149
9.1 Deficiencies Identified 149
9.1.1 Double shrouded impellers 151
9.1.2 Open impellers 160
9.2 Cavitation Erosion Rates 161
9.2.1 The need to know 161
9.2.2 Comparative erosion rate data 162
9.2.3 Evaluating the resistance to cavitation erosion for pump materials 164
9.2.4 Establishing a basis for cavitation performance assessment in
centrifugal pumps 166
9.3 Guidelines for Avoiding Unacceptable Cavitation Erosion 168
9.3.1 Pumps for which no cavitation test is required 168
9.3.2 Pumps for which an NPSH(3pc) test is required 169
9.3.3 Pumps for which an NPSH(4mm) test is required 174
9.4 Cavitation Erosion — An Avoidance Strategy 176
9.4.1 The tabular presentation 176
9.4.2 Pumps required for operation outside the conditional statements 178
 Contents ix

Chapter 10 Centrifugal Pump Low-Flow Protection 181

10.1 Functional Requirements 181
10.2 Choosing an Appropriate Low-Flow Protection System 183
10.2.1 Quantifying the options 183
10.2.2 The categories of low-flow protection system 185
10.3 Leak-off System Design 188
10.3.1 The basics 188
10.3.2 Aggregating theflowrate 191
10.3.3 Setting the set points 192
10.3.4 Set point values 193
10.3.5 Opening and closing times 196
10.3.6 Flowrate measurement signals 198
10.3.7 Modulating leak-off valves 198
10.3.8 The leak-off system as a heat sink 199
10.4 Leak-off System Hardware 200
10.4.1 Leak-off valves 200
10.4.2 Pressure reducing devices 204
10.4.3 Back pressure orifice plates 205
10.5 In Summary 205

Chapter 11 Cavitation Tests for Centrifugal Pumps 209

11.1 Test Specifications 209
11.1.1 Test types 209
11.1.2 National and international standards 209
11.1.3 Test quality 210
11.2 Test Basics 211
11.2.1 The test liquid 211
11.2.2 The water temperature limit 211
11.2.3 NPSH measurement reference plane 212
11.2.4 Inlet pipework configuration 213
11.2.5 The speed of the test pump 214
11.2.6 Effects made insignificant by appropriate test methodology 215
11.2.7 The limits of measurement uncertainty 216
11.3 NPSH Test Hardware 217
11.3.1 Essential requirements 217
11.3.2 Quality level C — NPSH(3pc) tests 217
11.3.3 Quality level B —NPSH(3pc) tests 220
11.3.4 Quality level B — NPSH(4mm) test layout 222
11.4 NPSH(3pc) and NPSH(4mm) Test Procedures 224
11.4.1 NPSH(3pc) tests 224
11.4.2 NPSH(4mm) tests 225
11.5 Quality Level A Tests 226
11.5.1 Transparent impeller tests 227
11.5.2 Paint erosion tests 228
X Contents

11.6 Special NPSH(3pc) and NPSH(4mm) Tests 230

11.6.1 Tests at reduced speed on clean cold water 230
11.6.2 NPSH(3pc) tests on liquids other than clean cold water 230
11.7 Cavitation Surge Testing 231
11.7.1 Identifying the need for a test 231
11.7.2 Identifying the limitations of practical tests 232
11.7.3 Cavitation surge test procedure and the criterion for concern 232
11.7.4 Making use of the test results 232

Chapter 12 New Centrifugal Pump Specifications —

Pump Selection and Cavitation 235
12.1 An Introduction to the Selection Process 235
12.2 Specific Objectives Identified 235
12.2.1 The input: Information provided by the pump user 236
12.2.2 The output: Data which will be produced by the selection process 237
12.3 The Methodology 237
12.3.1 The plant duty and the plant operating range 237
12.3.2 Selecting the pump duty and the pump operating range — Basic rules 237
12.3.3 Single fixed-speed pumps 238
12.3.4 Pumps in parallel 239
12.3.5 Variable speed pumps 240
12.3.6 Booster and main pump sets 242
12.3.7 Progressing through the evaluation process 243
12.3.8 Low flowrate protection 243
12.4 Learning by Example 243
12.4.1 The range of examples 243
12.4.2 A comment on the variety of options available 266
12.5 The Benefits of Understanding the Selection Process 266
Chapter 13 New Centrifugal Pump Offers —
Technical Assessments and Cavitation 267
13.1 Application Lists 267
13.2 Performance Data 268
13.3 Low-flow Protection 271
Chapter 14 Centrifugal Pumps in Service —
Resolving Cavitation Problems 275
14.1 Where to Start — Define an Acceptable Objective 275
14.2 The Identification of a Cavitation Problem 275
14.2.1 Listen for cavitation noise 275
14.2.2 Look for a generated head-drop at a known non-cavitating Q/AH value 276
14.2.3 Examine operational records 276
14.2.4 Examine pump internals 276
14.2.5 Confirm low-flow protection is functioning correctly 277
14.3 Examples from Experience 277
 Contents xi
14.3.1 Remedial measures — Make a change to operating procedures 277
14.3.2 Remedial measures — Make changes to plant component design or layout 279

Chapter 15 Centrifugal Pumps and Cavitation — A View of the Past ... 285
15.1 Engineering History — A Personal View 285
15.2 The Centrifugal Pump 286
15.2.1 Denis Papin — Intuitive engineering design of the highest quality 286
15.2.2 The "dark ages" of centrifugal pumping 288
15.2.3 The evolution of the modern centrifugal pump 289
15.2.4 Early uses and limitations of the centrifugal pump 295
15.2.5 Centrifugal pumps in the twentieth century 299
15.3 Cavitation 304
15.4 In Summary 307

Chapter 16 Centrifugal Pumps and Cavitation —

A View on the Future 311
16.1 Engineering Progress — A Personal View 311
16.2 Cavitation Surging 312
16.2.1 Hydrodynamically induced surging 312
16.2.2 Surging and fluid property effects 314
16.2.3 Thermodynamically induced surging 314
16.3 Cavitation Erosion 315
16.4 The Influence of Pump Design on Cavitation Performance 316
16.4.1 Constructional arrangements 316
16.4.2 Impeller design 316
16.5 Finally 317

Appendix A The First Centrifugal Pump —

As Described by Denis Papin in 1705 ... 319

Index 323
 Xlll

PREFACE
Cavitation in a centrifugal pump that limits performance unacceptably or
compromises safe operation is evidence of an engineering failure. Numerous options
are available that avoid such undesirable consequences. Diligence in appreciating the
risks inherent in design proposals and taking appropriate action to reduce them to an
acceptable level is necessary.
From the plant failures I have seen and remedial works I have had to undertake, both
in the United Kingdom and world wide, it is clear to me that there is a need,
particularly by pump users, to acquire a comprehensive insight into the industrially
significant problems that can arise from the occurrence of cavitation in a centrifugal
pump. This book is aimed at meeting this need.
The fundamentals that must be fully understood to provide a basis for developing
commercially viable guidelines are detailed in Chapters 2, 3, 4 and 5. These are
essential to subsequent chapters where the means of dealing with the risks arising
from cavitation are enumerated. The academic content is reduced to that essential to
understanding the logic of the arguments put forward; the inclusion of unnecessary
diversions has been avoided.
The acquisition of knowledge by those working in the engineering industry,
including pump designers, pump system designers and pump users, has always rested
heavily on regular reviews of past experience. Evidence from the real world can
mercilessly denounce an unjustifiable hypothesis just as it can give added confidence
that a chosen path has viability. My earliest papers on the subject of cavitation
erosion erred, necessarily, on the side of caution. A further twenty years of evaluating
operational records leads me now to reduce some of the margins by which confidence
in securing erosion-free operation can be given. The evidence now available makes it
extremely unlikely that further reductions in these margins will take place.
My guidelines on acceptable operating conditions, both by net positive suction head
(NPSH) margins and by minimum flowrates, can, by their very nature, only be
general. Their applicability to a particular pumping system needs to be assessed
carefully. I have described the evidence and logic by which I have arrived at these
guidelines to assist those whose task it is to bear the heavy responsibility of making
such an assessment. Information is presented in a format that addresses the needs of
pump users. A method is described by which likely commercial choice and the
boundaries of risk can be evaluated before a pump supplier is approached. The same
logic helps deal with problems arising from cavitating pumps in service.
I do not profess to have covered in this book every aspect of cavitation in centrifugal
pumps. Those of significance for the majority of pump applications are included. My
hope is that by providing a greater understanding my book will significantly reduce
the incidence of unacceptable performance caused by cavitation in centrifugal pumps.
E. Grist
Congleton, England
May 1998
XIV

ACKNOWLEDGEMENTS

Illustrations and data from existing publications and other sources have proved
invaluable in compiling my presentation. All the work of individuals is referenced at
its point of inclusion. I gratefully acknowledge the permission to publish material in
this book given by the following organisations:
American Society of Mechanical Engineers, New York, NY, USA
British Pump Manufacturers' Association, Birmingham, UK
Hopkinsons Ltd., Huddersfield, UK
Ingersoll - Dresser Pump Company, Liberty Corner, NJ, USA
Institution of Mechanical Engineers, London, UK
and their publishers MEP, Bury St. Edmunds, UK
Masoneilan Division, Dresser Produits Industriels, Paris, France
National Engineering Laboratory, Glasgow, UK
Newcastle Discovery Museum, Newcastle, UK
Rutschi Pumpen AG, Brugg, Switzerland
Science Museum/Science & Society Picture Library, London, UK
Sulzer Pumps Ltd., Winterthur, Switzerland
Valtek Engineering Ltd., Pershore, UK
Weir Pumps Ltd., Glasgow, UK
Western Power Corporation, Perth, Australia
Westinghouse Electric Corporation, Pittsburgh, PA, USA
Yarway Division, Keystone Valve (UK) Ltd., Glasgow, UK

I also acknowledge the help in securing photographs and other data given by my
peers in the UK pump and valve industries particularly John Frew, Angus Grant, Ron
Palgrave, Les Statham and Ashley Crossland. Additionally, for encouraging my
efforts to produce this book, I thank C. A. Eubanks of the Westinghouse Electric
Corporation and Andrew Lichtarowicz of Nottingham University. Finally I wish to
place on record my thanks to the three directors with engineering responsibility who,
over a number of years, encouraged me to obtain the experience necessary to write the
book. They are H. R. Wilshaw of Holden & Brooke Ltd., Donald McLellan of Crane
Ltd. and Tom Leith of Weir Pumps Ltd.
 XV

NOMENCLATURE

A flow area between impeller blades (total)

B a power in the relationship NPSH <x nB
b impeller flow passage width at outlet
CLA impeller surface roughness (centre line average height)
CP specific heat of liquid
D impeller outer diameter
Dp inlet pipe diameter
f functional dependence
g gravitational constant
H head
Ht head preventing ebullition
H, inlet total head
H2 outlet total head
HT head available to suppress ebullition
AH generated head
AHG AH at duty point ("guarantee" conditions)
AHP plant aggregate AH per pumping application
5H head amplitude of a cavitation surge
h enthalpy
hp total variable ("frictional") head loss (equation 4.7)
bpi component of hF up to the pump inlet
hF2 component of hF after the pump outlet
h
sys total system head loss (equation 4.5)
hs total static head loss (equation 4.6)
h
si component of hsup to the pump inlet
h
S2 component of hs after the pump outlet

K thermal diffusivity = _k,

PLCP
k, thermal conductivity of liquid
k2 a constant for a particular system (equation 4.7)
k3 a constant for a particular pump (equation 5.8)
k4 a Thoma/specific speed constant (equation 15.3)
L thermal diffusion length
1 flow path length along impeller blade from inlet to outlet diameter
lc cavitation zone length
1, distance along impeller blade at which rmax occurs
NPSH(I) NPSH at visually observable cavitation inception
xvi Nomenclature

NPSH Net positive suction head (ISO 3555-1977 definition)

NPSH(A) NPSH available at the pump inlet
NPSH(B) NPSH at generated head breakdown conditions
NPSH(R) NPSH required to prevent unacceptable cavitation
ANPSH NPSH reduction resulting from inlet pipe transients
5NPSH pressure producing cavity growth at breakdown NPSH(I) - NPSH(B)
NPSH(xpc) NPSH at x% head-drop where x can have any value
NPSH(3pc) NPSH at 3% head-drop
NPSH(xmm) NPSH at xmm cavity length where x can have any value
NPSH(4mm) NPSH at 4mm cavity length
NPSH(spl) NPSH at unspecified sound pressure level
n shaft rotational speed
ns specific speed (defined by equation 3.4)
n
ss suction specific speed (defined by equation 3.11)
PA pump input power
PAG PA at duty point ("guarantee" conditions)
PAO PA at zero flowrate
PI pressure driving cavity growth and collapse in a stationary liquid
AP generated pressure
Ap pressure decay in inlet vessel
6p pressure amplitude of a cavitation surge
P pressure
Pb barometric pressure
Pd pressure produced by duct constraints
Pg pressure generated in an impeller by centrifugal effects
Pgc pressure generated by the end of the cavitating zone
pgl pressure generated in non-cavitating zone of the impeller
pressure producing cavity growth
Pn
Pst restraining pressure resulting from surface tension
Pth pressure equivalent of thermodynamic property effects
Pvap vapour pressure
Pvi pressure in inlet vessel
PV2 pressure in outlet vessel
Px cavity internal pressure
Pi pressure at pump inlet
P2 pressure at pump outlet
P3 pressure beyond an outlet non-return valve
Q pump volumetric flowrate
Qi Q at the pump inlet
Q2 Q at the pump outlet
Qbep Q at best efficiency point
QG Q at duty point ("guarantee" conditions)
 Nomenclature xvii

QH maximum Q in operating range

QL minimum Q in operating range
Q,o leak-off rated flowrate
QM pump mass flowrate
Qm inlet vessel mass flowrate
QP plant aggregate Q per pump application
Q. lower limit of transitional flowrate band during leak-off operation
Qv minimum continuous flowrate to prevent vapour locking
Vwc Q at worst case conditions
q dryness fraction (Chapter 8 only)
q heat content
Aq change in heat content
qfo forward flowrate during leak-off operation
qio leak-off flowrate during leak-off operation
T.vap flowrate of vapour by volume
R cavitation protection ratio (defined by equation 4.2)
R(3pc) R based upon 3% head-drop data (equation 4.3)
R(4mm) R based upon 4mm cavity length data (equation 4.4)
RO Universal gas constant
r cavity radius
r0 radius of cavity in equilibrium
Miiax maximum cavity radius
s specific volume
T temperature of liquid (at impeller blade entry where applicable)
TO temperature of liquid in pump inlet pipe
Tr cavity wall temperature
t time
tfo time for flowrate to change from zero to Q,0
t,o time for flowrate to change Qlo to Q^
tp transit time from inlet vessel to pump inlet (tp = 0 when t = 0)
t* time increment (Chapter 6)
Vc volume of liquid contained in pump casing
VL specific volume of liquid
vv specific volume of vapour
v, mean velocity of liquid at pump inlet
V2 mean velocity of liquid at pump outlet
W vessel liquid mass
AW change in liquid mass
Y area ratio (Fig. 3.6)
z number of impeller stages
zc cavitation site density (cavities/unit time)
Zl static head - gauge centreline to reference plane (Fig. 2.5)
Z2 static head - gauge centreline to reference plane (Fig. 2.5)
xviii Nomenclature

(3 the group [3K~| p.^2 X?

L TC J k, RO T3

X latent heat of vaporisation

* rate of heat input to fluid from pump power losses
*1 pump efficiency (%)
Tlbep r| at best efficiency flowrate (%)
9 pipichum value (defined by equation 10.4)
V kinematic viscosity
7t 3.14
PL liquid density
a Thoma cavitation parameter (defined by equation 15.1)
aT surface tension
T churned mass multiplier (defined by equation 10.5)

Special subscripts (Chapter 6)

in inflow
out outflow
(P) relating to pump inlet
(v) relating to the inlet vessel
(t) at time t
 Chapter 1

Cavitation — An Unacceptable Phenomenon

1.1 Cavitation Defined

Cavitation is defined as the occurrence of vapour-filled cavities in a liquid.

The definition describes all conditions in a centrifugal pump from the growth
and collapse of small bubbles in the pumped liquid to the occurrence of large
amounts of vapour that are sufficient to vapour lock a pump in a predominantly
liquid-filled system.
1.2 Unacceptability Defined

A cavitating centrifugal pump can be found unacceptable in a number of ways.

Those of industrial significance can be categorised as follows.

Category Unacceptable Feature

1 Hydraulic Performance Loss

2 Cavitation Surging — Hydrodynamically Induced

3 Cavitation Surging — Thermodynamically Induced

4 Cavitation Erosion

The potentially serious consequences that can arise when a centrifugal pump
cavitates are loss of economic performance, damage to expensive equipment, and,
in the extreme, loss of life. Such consequences are unacceptable.
The ways in which unacceptable behaviour can arise are presented in outline in
this chapter. This highlights the need to understand the basic pump terminology,
cavitation performance characteristics, and the limited amount of engineering fluid
mechanics presented in Chapters 2, 3, 4 and 5.
Later chapters expand on the categories of unacceptability. They give a more
detailed description of the underlying mechanisms. They also provide an
engineering methodology that identifies risk and describes how such risk can be
minimised.

1
2 Cavitation and the Centrifugal Pump

cavitation cloud

blade inlet
edge

direction
of flow

impeller rotation

Fig. 1.1 Cavitation on a centrifugal pump impeller blade

Courtesy of Weir Pumps Ltd.

"A"
system resistance
curve "B" non-cavitating
performance curve
"Frictional"
component
cavitating
performance curve

"Static" pump duty shortfall

component

QG

Flowrate Q

Fig. 1.2 Pump performance and system resistance curves

 Chapter 1 — An Unacceptable Phenomen 3

1.3 The Categories of Unacceptability — A Brief Resume

1.3.1 Category 1. Hydraulic performance loss

The presence of cavitation in a centrifugal pump impeller, such as shown in Fig.
1.1, can give rise to a loss of generated head. Whilst a modest amount of cavitation
may have no measurable effect, a large volume of cavities can result in the
generated head being reduced considerably.
Two conditions are of industrial significance:
1. Failure to meet the required duty thereby resulting in long-term
economic loss.
2. A catastrophic reduction in head caused by large volumes of cavitation,
which give rise to vapour locking of the pump and lead to an immediate
cessation of pumping. This is sometimes called vapour binding.
Mechanical damage to the rotating element of the pump nearly always
occurs as a consequence.
Pump duty shortfall. In a centrifugal pump the local pressure drops experienced
by the liquid flowing within its passages determine where cavitation occurs. This is
usually close to the inlet tips of the impeller blades. Cavitation can be suppressed
by raising the pressure in the region of its occurrence. In a near-stationary liquid a
pressure a little above vapour pressure will suffice. In a cavitating centrifugal
pump the effect of continuing to increase the pressure in the pipe immediately
prior to the pump inlet is to reduce and eventually eliminate the presence of
cavitation.
Centrifugal pump performance is usually expressed as a curve showing generated
head AH plotted against volumetric flowrate Q as shown in Fig. 1.2. Typically a
pump is tested with an excess of inlet pipe pressure to ensure cavitation does not
occur; the non-cavitating pump performance curve shown is obtained. Insufficient
provision of inlet pipe pressure to prevent cavitation leads to a reduction in
achievable generated head at higher flowrates; the cavitating performance curve
shown is then obtained.
Every pump has to overcome a system resistance which comprises fixed
("static") and variable ("frictional") components. The pump operates at the
intersection of the system resistance curve and the pump performance curve
appropriate to the particular inlet pressure conditions. The pump duty, point "A", is
where the pump is intended to operate. A test under non-cavitating conditions
would confirm its acceptability for this task since the system resistance curve can
be raised by increasing the "frictional" component, typically by the partial closure
of an outlet control valve. A cavitating pump would operate at point "B" instead of
the required point "A" resulting in a substantial shortfall in flowrate. Where pump
flowrate is directly linked to plant output this can represent an unacceptable
economic loss. Where conditions allow cavitation to persist indefinitely the
cumulative loss can be considerable.
Chapter 4 describes how to quantify the inlet conditions controlling cavitation in
a centrifugal pump at the pump duty and over the pump operating range. Chapter 6
explore how to ensure the pump duty requirement is met.
Vapour locking. As flow through a pump is reduced an impeller becomes more and
more inefficient. Much of the energy expended in the associated erratic turbulent
internal flows goes into heating the liquid. Should throughflow become very small or
 4 Cavitation and the Centrifugal Pump

(a) Excessive casing clearance wear (b) Neck ring seizure onto
impeller

Fig. 1.3 Pump damage after a vapour lock

Courtesy of Western Power
Corporation

hydrodynamically induced
cavitation surge
Direction of flow

pump inlet

Fig. 1.4 Hydrodynamically induced cavitation surging

Courtesy of Ingersoll-Dresser Pump Company
 Chapter 1 — An Unacceptable Phenomen 5
stop altogether then practically all the power absorbed by the impeller goes into
producing heat. Continuous operation in this mode ultimately leads to vaporisation
of the liquid within the pump. In low-energy pumps (e.g. domestic central heating
pumps) it can take many hours to produce sufficient cavitation for a full vapour
lock. In high-energy pumps (e.g. large power station feedwater pumps) it can take
less than one second.
The consequences of vapour locking are unacceptable. Many pumps, especially
multistage pumps, incorporate internal clearances such as neck rings, interstage
bushes, or balance drums, which act as pumped-liquid lubricated bearings. These
clearances support the radial loads on the pump rotor, which originate principally
from the unbalanced hydraulic pressure contour surrounding it. Without internal
support the resulting rotor deflections lead to contact between rotating and
stationary components and seizure ensues. Typically this appears as shown in Fig.
1.3.
Some pump designs include mechanical seals. These often depend upon the
pumped liquid being present to provide a lubricating film between rotating and
stationary components. Rapid mechanical failure can ensue if this supply of liquid
is removed by vapour locking.
Centrifugal pumps can be vapour locked when NPSH(A), a measure of
conditions suppressing cavitation defined later, falls to a very low value. This can
be the consequence of a malfunction of the low-flow protection system. It can also
be the result of a plant-induced operating transient which significantly reduces the
pressure in the inlet vessel that, as a consequence, reduces the pressure in the pump
inlet pipe.
Chapter 6 explores the problems and the available solutions.

1.3.2 Category 2. Cavitation surging — Hydrodynamically induced

At part-load flowrates, typically below about 50% of best efficiency flowrate, the
angular mismatch between incoming liquid flowpaths and impeller blades is
significant. This is usually more so at the leading edge of blades near the outer
diameter of the impeller eye. In many impellers a recirculation develops in the
low-pressure region behind each blade making it possible for flow to pass back
into the inlet pipe at its outer diameter. This recirculation becomes progressively
stronger as flowrate is reduced. Where the pressure drop behind the impeller blade
is compounded by a sufficiently low inlet pressure cavitation occurs.
Cavitation is normally suppressed in milliseconds as it enters the strongly
increasing pressure gradient associated with flows along a centrifugal pump
impeller blade at near best efficiency conditions. Where backflow recirculation
into the inlet pipe occurs the cavities find themselves in the relatively much weaker
pressure gradient of the inlet pipe. Their growth and collapse take longer, and this
allows them to be propagated impressive distances back down the inlet pipe. At
very low flowrates and with a straight inlet pipework configuration this spiralling
outer core of cavitating liquid can travel several pipe diameters against the general
direction of flow before returning to the impeller. An example of this is shown in
Fig. 1.4.
The production of vapour on the impeller blade surface is disrupted by the
returning recirculation flow of liquid-borne cavities. The whole process becomes
unstable. A periodic surging motion is set up that is predominantly dependent upon
the recycle time of the recirculating liquid. Typically this surge has a frequency of
2 Hz to 10 Hz.
6 Cavitation and the Centrifugal Pump
Throughout the region where low-pressure gradients exist in the inlet pipe such
cavitation surging can persist. This can travel through to the outlet side of low
generated-head types of impellers such as inducers and those for axial flow
machines. In centrifugal pump impellers where flow no longer enters the high-
pressure region in the normal direction of flow the growth and collapse of large
volumes of cavitation occurs.
Experience shows the intensity of cavitation surging to vary considerably for
different pump-impeller and inlet-pipework configurations. Some combinations
give rise to a leisurely "sloshing" motion. Others give rise to a violent piston-like
effect that is accompanied by a powerful "chuffing" noise — very reminiscent of a
reciprocating steam engine exhaust. The resulting pressure pulsations often
produce severe pipework vibrations. The wild variations in axial thrust and
torsional loading on the pump rotor frequently lead to thrust bearing failure and
occasionally to shaft breakage.
It is important to note that this form of cavitation surging is (a) independent of
the plant design beyond the run of inlet piping immediately preceding the pump
and (b) is more easily propagated where impeller and inlet piping are conducive to
axially symmetric motions. An end-suction pump with a long straight inlet pipe is
a common example of the latter.
Chapter 7 explores the problem and the available solution.

1.3.3 Category 3. Cavitation surging — Thermodynamically induced

The consequences of pressure containment failure are potentially so horrific it is
essential to appreciate all the underlying pump related mechanisms that could give
rise to it. A high-energy centrifugal pump, such as a power station feedwater
pump, has the capability of initiating a catastrophic failure of the pump inlet pipe.
Such a breach creates large changes in momentum. The forces that then arise can
be more than sufficient to rip pipework from supporting structures. More serious
consequences can occur when the liquid in the inlet pipe is at a temperature much
greater than that necessary to keep it as a liquid on release into the environment.
The explosive escape of the pumped liquid as it vaporises can seriously injure
individuals caught by the scalding blast.
The damage to the power station feedwater pump shown in Fig. 1.5 resulted from
a catastrophic failure of the pump inlet pipe. The explosive escape of hot
condensate caused a number of fatalities. In this particular instance the cause was
attributed to flow-accelerated corrosion in the inlet pipework. The example is used
to focus clearly on the risk which might be present should pipework over-
pressurisation occur.
Although catastrophic failures associated with centrifugal pumps are rare it is
necessary to study the features inherent in each centrifugal pump and plant design
combination to ensure that essential reliability and redundancy is identified and an
appropriate protection system is provided. Safe operating conditions for high-energy
pumps are crucially dependent upon plant design. The provision of low-flow
protection, always essential for such pumps, takes on greater importance in particular
plant design arrangements. It is necessary to understand why this is. It is also necessary
 Chapter 1 — An Unacceptable Phenomen

Displacement resulting from

forces released during
inlet pipe breach (overhead)

inlet
pipe
pump

Fig. 1.5 Catastrophic failure of pump inlet pipework

to understand the potentially serious consequences arising from the maloperation

or the operational failure of such low-flow protection.
High-energy centrifugal pumps that operate at zero or very low flowrate will
experience vapour locking. The process of vapour locking can take place in less
than one second. The amount of vapour produced in this time will approximate to
the volume of the impeller plus that of the liquid passages within the pump and
inlet pipework where liquid is churned by the impeller. In a typical boiler
feedwater pump this might be 0.5 m3. The sudden need to accommodate the
additional vapour volume needs to be addressed. In inlet pipework systems where
the pumped liquid is drawn from a vessel which has a free surface the increase in
pressure in the pump casing may be small, particularly if the inlet pipe length is
very short. Some types of shaft sealing systems may also provide relief by
allowing escape of liquid. The pipework and pump
casing also grow slightly to provide added volume as high pressures increase
stresses in pipework materials toward their elastic limit.
The pressure generated in the inlet pipe of a pump that is preceded by a non-
return valve poses a potentially more serious risk. In a system where the pump
discharges liquid into a high-pressure system through a non-return valve the
structural integrity of the pressure boundary may be challenged. This occurs when
pump speed is reduced below that which is necessary to provide a pressure
sufficient to open the outlet non-return valve. Failure to provide adequate relief
between upstream and downstream non-return valves means that the pressure
rapidly increases as the vapour being generated by the pump tries to expand in the
near incompressible trapped liquid. Two outcomes are possible:
8 Cavitation and the Centrifugal Pump

leak-off valve outlet vessel

(closed)

inlet vessel

outlet
control valve
pump direction of flow (open)

(A) heat from pump losses

carried away by liquid flow

inlet non-return
valve (open) System 1

leak-off valve outlet vessel

(closed)

inlet vessel

outlet
control valve
no flow in trapped (closed)
pump liquid volume

(B) heat from pump losses

retained in expanding volume
of trapped liquid
inlet non-return
valve (closed by overpressurisation at weakest point causes breach
expanding liquid) (likely to be in the inlet pipework)

Fig. 1.6 Over-pressurisation of pump inlet pipework resulting from

thermodynamically induced cavitation surging

1. Where the inlet pipework is sufficiently strong the pressure in the pump and
the inlet pipework rises to the pressure in the pipework beyond the outlet non-
return valve.
As the cavity continues to expand this valve opens. The pump then tries to
commence normal pumping. A small amount of new liquid is drawn into the
impeller, the generated head then increases, and some of the vapour/liquid mixture
is displaced toward the outlet pipe thereby allowing the inlet pressure to fall. This
relieves the outlet pressure and allows the outlet non-return valve to close again.
The cycle then repeats itself. The pipework vibrations associated with such
thermodynamically induced surging are often frightening to witness.
 Chapter 1 — An Unacceptable Phenomen 9
2. Where the inlet pipework is not strong enough the pressure containment
boundary fails catastrophically. Rupture occurs at the weakest point and this can
reveal a design deficiency other than in the pump. The role of thermodynamically
induced cavitation surging in providing the pressure pulse to finally breach the
pressure boundary in such circumstances can become overlooked.
In a system such as that shown in Fig. 1.6 where flow through the pump is
stopped — typically by the rapid closure of an outlet control valve — the outcome
is one in which the mechanical integrity of the pump, pump pipework, and inlet
non-return valve can be challenged. In practice this scenario occurs when the time
to close the outlet control valve is much less than the time to open a flowpath
through the low flow protection "leak-off line. Mechanical damage is inevitable
unless relief of the pressure produced by the expanding liquid/vapour mixture is
available.
Chapter 8 explores the problem and the available solution.

1.3.4 Category 4. Cavitation erosion

When cavitation occurs cavitation erosion might take place — or it might not! It
is important to appreciate this lack of inevitability when considering the risk of
cavitation erosion and when attempting to quantify its intensity by extrapolating
from the damage observed on other impellers. Equally, it is necessary to
understand why such uncertainty exists.

Fig. 1.7 Typical cavitation erosion damage on an impeller blade

Courtesy ofSulzer Pumps Ltd.
10 Cavitation and the Centrifugal Pump
Cavitation erosion is caused as vapour-filled cavities enter a region of higher
pressure within a pump and collapse. When this violently implosive cavitation
process takes place in close proximity to a metal surface a damage mechanism that
produces local pitting can result. The uncertainty surrounding both the onset of
erosion and the intensity of attack when cavitation does occur in a commercially
manufactured pump arises from the lack of awareness of how near the cavities get
to the pump internal surfaces. Collapse of cavities within the body of liquid
flowing through the pump that are distant from an internal surface may not result
in cavitation erosion damage. However, as the point of low pressure that initiates
cavitation is nearly always at such a surface (e.g. close to the impeller blade tip at
inlet), damage is more often than not observed. A characteristic audible crackling
often accompanies such a damaging cavity collapse.
The probability of damage increases where a pump is used over a range of
flowrates since the changing flowpaths of cavitation within the liquid in the
machine are more likely at some point to impinge upon an internal surface.
Interestingly, examination of
damaged pumps shows that sometimes when cavitation erosion does occur it may
be inconsequential to the safe running of the pump even after cavitation attack has
persisted for thousands of hours. Conversely cavitation damage, such as that
shown in Fig. 1.7, is sometimes observed after a matter of a few hours. Where
cavitation erosion damage is so severe or cracks are initiated the structural failure
of an impeller can result.
Chapter 9 explores the problem and the available solutions.
 Chapter 2

Independent Variables and Terminology

The independent variables that are used to describe centrifugal pump
performance come from a background where choice has been exercised for
particular beneficial reasons. Also, custom and practice in the pump industry have
led to a number of terms being used that have a specific meaning when applied to
the technology. It is useful to review these independent variables and terminology.

2.1 Primary Independent Variables

Non-cavitating pump performance is, for a majority of applications, found by test

to be completely defined by four independent variables:
(1) Generated head AH
(2) Volumetric flowrate Q
(3) Shaft rotational speed n
(4) Impeller outer diameter D

Cavitating pump performance is, for a majority of applications, found by test to

be completely defined by four independent variables:
(5) Net positive suction head NPSH
6 (2) Volumetric flowrate Q
7 (3) Shaft rotational speed n
8 (4) Impeller outer diameter D

From the bracketed numbers it can be seen that five independent variables
between them describe both non-cavitating and cavitating performance. These are
considered in turn.
(1) Generated head AH. Often referred to as "pump total head" or more simply
as "head", this is the rise in head between the inlet (suction) and outlet (discharge)
branches of a pump. It is a primary variable.
The choice of generated head rather than generated pressure is significant and
intentional. Centrifugal pump performance is determined by the kinetic energy of
the pumped liquid as it leaves the impeller at a velocity close to its peripheral
velocity. The generated head, which has a value directly proportional to this
energy, is constant for a given shaft rotational speed irrespective of the liquid
density. The use of this single independent variable to describe the added energy
aids analysis. It also mirrors the way in which system losses — which determine
the pump duty — are calculated.
In a strict analytical sense the product of generated head and gravitational constant

11
12 Cavitation and the Centrifugal Pump
is the correct measure of pump performance. However, industrial practice is to
assume the gravitational value is unchanging and to describe pump performance by
head measurement alone. This has the obvious perceptual advantage — which was
the original basis for the choice — of being relatable to physical heights in
pumping plant design.
It must be remembered that when generated head is used in the dimensional
analysis of pump performance its product with the gravitational constant has to be
included.
(2) Volumetric flowrate Q. Often referred to simply as "flowrate", this is a
primary variable. The choice of volumetric flowrate rather than mass flowrate is
significant and intentional. Centrifugal pump performance is controlled by the
velocity of liquid passing through the impeller outlet flow passage area so making
the volumetric flowrate a constant for a given speed irrespective of the liquid
density. The use of this single independent variable to describe flowrate aids
analysis and also mirrors the way in which system losses — which determine the
pump duty — are calculated.
(3) Shaft rotational speed n. Often referred to simply as "speed", this is a
primary variable.
(4) Impeller outer diameter D. This is a primary variable. As a length
measurement it characterises the dimensional proportions of a pump. It also
provides the basis for performance analysis.
(5) Net positive suction head NPSH. Usually referred to simply by the initials
NPSH, this is a primary variable. A measure of conditions suppressing cavitation is
the difference between the inlet total head (which includes the velocity head in the
inlet pipe) and the head equivalent of the vapour pressure of the liquid being
pumped. This head difference is termed "net positive suction head".

Within an impeller the local pressure controlling cavitation is related to the

interaction between rotating impeller blades and liquid flow paths. The head drop
on the downstream side of a blade results from the kinetic energy loss due to the
mismatch of the two. A true measure of pump performance is the product of NPSH
and the gravitational constant. Industrial practice again assumes the gravitational
constant is unchanging and cavitating pump performance is usually described by an
NPSH measurement alone. It must be remembered that when NPSH is used in the
dimensional analysis of pump performance its product with gravitational constant
has to be included.
The choice of NPSH to measure cavitation has a perceptual advantage where a
pump draws from a vessel containing a liquid at its boiling point and the inlet
velocity head is insignificant. A large number of pump installations where
cavitating problems occur match this description. In these circumstances the
physical height of the liquid surface level to pump shaft centreline is a close
approximation to the value of NPSH at the pump inlet.

2.2 Secondary Independent Variables

A number of other independent variables are known to influence the performance of a

centrifugal pump in particular circumstances. For the majority of pump applications
 Independent Variables and Terminology 13
where cold water or liquids having properties similar to those of cold water are
pumped they are insignificant. They are regarded as secondary independent
variables.
The circumstances in which these secondary independent variables need to be
taken into account and a methodology for dealing with the risks and uncertainties
they bring to engineering applications are discussed in Chapter 3.
The secondary independent variables fall into two groups: Group 1 for those
having their principal effect on non-cavitating performance and Group 2 for those
having their principal effect on cavitating performance.
Group 1
Kinematic viscosity v
Impeller surface roughness CLA
Group 2 (Thermodynamic properties affecting cavity dynamics)
Vapour pressure pvap
Latent heat of vaporisation A,
Specific volume of vapour Vv
Liquid density pL
Thermal conductivity of liquid k,
Specific heat of liquid Cp

2.3 Hydraulic Configuration Conventions

2.3.1 Centrifugal pump impeller design terminology

A typical centrifugal pump impeller is shown in Fig. 2.1. Flow enters the
impeller through a single "eye" and passes to blades contained within front and
back shrouds.
These blades "impel" the pumped liquid through the pump whilst at the same time
moving it radially outward from the pump shaft. Liquid leaving the periphery of
the impeller does so at a higher circumferential velocity. Energy is gained by the
liquid as, relative to the inlet, it accelerates rotationally with the impeller blades.
This energy is transformed into an increase in pressure by being constrained to
slow down by the geometry of a collecting chamber.
The impeller is the head generating component within a centrifugal pump. The
combined outlet geometry of the impeller and the collecting chamber surrounding
it determines the value of the generated head rise and the best efficiency point
flowrate. The impeller inlet geometry determines the cavitation performance
characteristics of the pump.
The outlet blade tip usually finishes straight and parallel to the shaft axis as
shown in Fig. 2.1. By exception, it finishes at an angle to the axis or is shaped. This
latter design feature is provided on high speed pumps to reduce the pressure pulse
that occurs as the blades pass close to stationary parts in the collecting chamber.
This has little or no influence on cavitation performance except where, for good
manufacturing reasons, it affects the design of the blade at inlet.
The blade shape at inlet is of crucial importance in preventing cavitation. Ideally the
blade should be shaped so that the angle at which the blade meets the flow is the same
at all points across the flowpath. This reduces the propensity of an impeller to local
recirculations behind the blade at or near best efficiency flowrate. Where local pressure
14 Cavitation and the Centrifugal Pump

direction impeller blade

of rotation 'twisted'
impeller blade
'plain'
neck ring
front shroud

back shroud
A [B~
shaft axis
impeller
eye
INLET

blade tip
parallel to shaft
axis
blade inlet
edge OUTLET

impeller hub

Fig. 2.1 Impeller terminology

drops are low it delays the onset of cavitation. To achieve a common flow
impingement angle across the blade inlet edge there has to be a three-dimensional
twist since with a uniform axial flow into the impeller eye the circumferential
velocity at the neck ring (Fig. 2.1 point "A") is much greater than that at the
impeller hub (Fig. 2.1 point "B"). Such twisted blades can, as shown in Figs. 1.1
and 1.7, exhibit a near flat surface when seen from the impeller eye. Cavitation and
cavitation damage is thus spread uniformly across the blade. Repair of damaged
areas becomes a possibility.
Manufacturing twisted blades is difficult (Ref. 2.1) and costly. For pumps where
cavitation is not a problem and for those pumps where the cavitation can be easily
avoided by small changes to plant design three-dimensional twisted blades are an
unnecessary added expense. The use of simple-to-manufacture two-dimensional
blades, usually referred to as "plain" vanes, then makes good commercial sense.
Blades that spiral radially outward in a clockwise direction, like those in Fig. 2.1,
are termed right-handed impellers (because they are easier to hand-file in a bench
vice by a right-handed person). Those that spiral anticlockwise are termed left-
handed the impellers. This "handing" of an impeller is of crucial importance where
the prime mover driving
 Independent Variables and Terminology 15
the pump is unidirectional. An example of this is a pump driven by a steam turbine
that has internal nozzles that are not suited to operation in both directions.

2.3.2 Centrifugal pump impeller design configurations

There are three basic centrifugal pump constructional configurations. They are (i)
the single-entry pump, (ii) the double-entry pump and (iii) the multistage pump.
They are shown in cross-sectional view in Fig. 2.2. The illustration shows
variations based on shrouded impellers. In some pumps one or both of the shrouds
are omitted to give an easier-to-manufacture "open" impeller design. The
multistage assembly also comes with a variety of design possibilities. Figure 2.2(c)
shows a multistage pump made up of single-entry impellers. In practice a
combination of single-entry and double-entry impellers is often used to maximise
the benefits inherent in these particular hydraulic designs.
A multitude of variations in centrifugal pump construction has been derived from
the three basic configurations. A fuller description of them is outside the scope of
this book. Karassik and Carter (Ref. 2.2) give a good pictorial guide to many of the
pump and pump impeller designs that have been used.

2.4 Performance Measurement Levels

Centrifugal pump performance is described at three levels. The context that data
are used in makes it self-evident which level is being referred to.

2.4.1 Level 1. Impeller performance (AH, Q)

To be able to compare impeller performance genetically it is necessary to define
a common basis for all the impeller design variants. The convention adopted by the
pump industry is to use a single-entry impeller as a basis for performance analysis.
Performance is always quoted at best efficiency flowrate unless otherwise stated.
The use made of best efficiency based data as a measure of impeller performance,
as will be seen in later chapters, demands no great accuracy. It is therefore common
practice to include losses in the collecting chamber surrounding the impeller (e.g.
volute or diffuser ring) in determining the generated head. For example, the head
measured across the inlet/outlet branches of an end-suction pump is used.
Similarly, the head across the inlet/outlet to branches divided by the number of
stages is used where two or more impellers of equal diameter are mounted in
series. Double-entry impellers are treated as two identical impellers mounted back-
to-back.
The data for impeller performance comparisons are obtained as follows:
1. the best efficiency flowrate Q^,
2. the generated head per impeller stage at Q^
3. the volumetric flowrate per impeller eye* at Q^.
* Flowrate per impeller eye is strongly recommended. Occasionally in the
technical literature the pump performance for double-entry pumps has been
described using total flowrate. Some manufacturers have considered this
advantageous when correlating performance of double-entry volutes. The flowrate-
per-eye used here has greater hydraulic consistency and follows the method
described by Karassik and Carter (Ref. 2.2), Stepanoff (Ref. 2.3) and Anderson
(Ref. 2.4).
16 Cavitation and the Centrifugal Pump

Qbep i AH
generated head AH

flowrate Qbep flowrate_QbeP flowrate_Qbep

per eye ~ "2" per eye 2

(a) single entry (b) double entry

generated head per stage = AH

(where z = number of stages)

Qbep AH

Qbep

(c) multistage

Fig. 2.2 Impeller configurations (cross-sectional views)

 Independent Variables and Terminology 17

2.4.2 Level 2. Pump performance (AHG, QG)

This describes the "guaranteed" duty of a pump. Level 2 is, without exception,
used as the basis for procurement specifications and contractual acceptance tests.
The pump duty is usually close to the best efficiency point, but it seldom has this
value.
When selecting a new pump to meet the guaranteed duty the numerous
permutations of hydraulic configuration that provide diversity in economic choice
are further amplified by the option of running off best efficiency flowrate. This
strategy can be used to secure a different operational advantage (e.g. a steeply
falling AH/Q curve) whilst still meeting the essential requirement of eliminating
the risk of unacceptable cavitation.

2.4.3 LevelS. Plant performance (AHp, Qp)

This describes the overall plant performance where a number of pumps, either in
series or in parallel, are used at a particular location in aggregate to meet the
requirements of the plant. Selecting combinations of new pumps from the
numerous permutations of pump set configurations provides further diversity in
economic choice whilst meeting the requirement of eliminating the risk of
unacceptable cavitation. Pump selection is considered in some detail in Chapter 12.
The basic layouts that can be used are shown in Fig. 2.3.

(a) single pump

(b) pumps in series

(c) pumps in parallel

Fig. 2.3 Pump layout configurations

18 Cavitation and the Centrifugal Pump
A number of useful options are available when pumps are arranged in series.
Examples of those frequently used in power plant are shown in Fig. 2.4. The
usual purpose of a series arrangement is to enable a slow running "booster" pump
to provide sufficient pressure at the inlet of a high-speed pump to suppress
cavitation. The method of attaining the high speed determines the configuration.

typical prime mover

electric motor

** .
typical prime mover
steam turbine

Fig. 2.4 Pumps in series — Some pump set configurations

2.5 Performance Measurement Terminology

Total head H.
The notation used is illustrated in Fig. 2.5.

1. Inlet total head H, = z, + _E, + v, 2

P,g 2
§ 2.1

2. Outlet total head H2 = z2 + ^2 + y/

P2§ 2
§ 2.2
 Independent Variables and Terminology 19
Pressures p, and p2 are "gauge" values. For cavitation performance assessments
the liquid densities p, and p2 are assumed to be constant and have a value p,. This
assumption has validity where the inlet performance of the pump for a given liquid
at a given temperature is being evaluated.
pressure gauge
reading p2

pressure gauge
reading n

Fig. 2.5 Pictorial view of notation conventions

Generated head AH.

The word "generated" and the symbol AH are used to provide an unambiguous
reminder that they represent a change in value.

AH = R - H, -2.3

ISO 3555-1977 (Ref. 2.5) uses "pump total head H" instead of "generated head
AH".

Net positive suction head NPSH. This is the inlet total head, plus the head
corresponding to the atmospheric pressure, minus the head corresponding to the
vapour pressure.

NPSH = H, + p_t - p^
P,,g P,.g -2.4

The definition and industrial usage of NPSH are dealt with in detail in Chapters 3
and 4. The choice of NPSH measurement reference plane is discussed in Chapter
11.
20 Cavitation and the Centrifugal Pump
Pump input power PA. This is the power absorbed measured at the pump
coupling.

Best efficiency flowrate Qbef. Cavitation performance data in pump user

assessments of performance are always related to best efficiency conditions. That
is, they correspond to the volumetric flowrate at which the maximum value of
efficiency is attained for a centrifugal pump operating at a specified shaft rotational
speed.

Pump Efficiency r|.

Pump efficiency r\ = £>k-S- Q AH x 100%
(as a percentage) PA 2.5

Where a significant change in liquid density occurs — such as in a boiler

feedwater pump where a considerable compression of the pumped liquid is made
before discharging it at a very high pressure — the additional useful work added to
the pumped liquid needs to be included where energy utilisation assessments are
being made. In cavitation performance calculations where pump efficiency is used
only to define the value of Q^, the error arising by assuming the liquid density
does not change from inlet value pL is insignificant.

Pump duty. The "duty" of a centrifugal pump is described by the specified values
of generated head AH0 and volumetric flowrate Q0, which the pump is required to
deliver when it runs at a specified shaft rotational speed. The values AHG and QG are
used in specifications to define the contractual "guarantee" point in performance
acceptance tests. QG is usually close to Q^ but is seldom of the same value.

2.6 Metric Units

The units of measurement for pump selection and specifications are taken from
the System International (SI) with reference, where appropriate, to ISO 31 —
Quantities, units and symbols. SI allows choice in the use of subdivisions. Those
used in the following text are listed in Table 2.1.

Table 2.1 Metric units

Variable Notation Units Abbreviation
Flowrate Q litres per second mVs x 10"3
see Note 2
Head H]
Generated head AHj metres m

Pump shaft speed n revolutions per minute rpm

see Note 3

Impeller/pump D millimetres mm
dimensions
 Independent Variables and Terminology 21
Table 2.1 continued
Variable Notation Units Abbreviation
Pressure pl bar bar
Generated pressure APj

Liquid density PL kilogram per cubic metre kg/m3

Note 1. 1 bar = 1 x 105 newtons per square metre.

Note 2. The SI unit of first choice for flowrate is the cubic metre per second
(mVs). Numerically this is only convenient for measuring extremely large flows
such as those of rivers. For most pump applications pump duty is better expressed
as one-thousandth of this by the litre per second (mVs x 10"3). Unfortunately the
abbreviation for "litres per second" (1/s) is too easily confused with that for "one
per second" (1/s) so litres per second is written in full where the expression mVs x
10~3 is not convenient or is not appropriate. The alternative abbreviation dm3/s is
also used.
Note 3. The practice of measuring rotational speed in rpm is widespread. Choice
of this unit results from strong user requirement for consistency with the
established method of measuring the rotational speed of other rotating machinery.

2.7 Useful Conversions

Mass flowrate/volumetric flowrate. Pumped product output is usually measured

by mass in the chemical process and related industries.

Q = QM
PL" 2.6

Pressure/head. Whilst pump performance and associated hydraulic calculations

are normally expressed as head measurements there are occasions when the
pressure at a particular point has to be determined. A pipework pressure
containment integrity assessment is an example of this.

The pressure p corresponding to a value of head H is given by

p = pL g H -------------------------------------- 2.7

Care should always be taken to make the distinction between static head and
generated head. (n.b. generated pressure AP = pL g AH).
The conversion of a cold water (20°C approx.) pressure (bars) to head (metres) is
frequently required. Using equation 2.7 with p = 1 bar = 1 x 10s N/m2 ,
pL = 1000 kg/m3 and g = 9.81 m/s2 it can be seen that 1 bar of cold water pressure
is equivalent to a head 10.2 m. This value appears frequently in calculations. It
must be remembered however that it only applies to a liquid with a density of
about 1000 kg/m3.
22 Cavitation and the Centrifugal Pump
References
2.1 Church, A. H., Centrifugal Pumps and Blowers. Wiley, New York, 1956.
2.2 Karassik, I. and Carter R., Centrifugal Pumps. McGraw-Hill, New York,
1960.
2.3 Stepanoff, A. J., Centrifugal and Axial Flow Pumps, second edition, Wiley,
New York, 1957.
2.4 Anderson, H. H., Centrifugal Pumps and Allied Machinery. p58,
fourth edition, Elsevier, Oxford, 1994.
2.5 ISO 3555 - 1977, Acceptance Tests for Centrifugal. Mixed Flow
and Axial Pumps. 1977.