Flow of Industrial Fluids Theory and Equatiion

Flow of Industrial FluidsTheory and Equations
Flow of Industrial FluidsTheory and Equations

By Raymond Mulley
(@RC
PRESS
fia
ISA-The
andInstrumentation,Systems,
Automation Soc~ety
Library of Congress Cataloging-in-PublicationData

Cataiog record is available from the Library of Congress
This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with
permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish
reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials
or for the consequences of their use.
Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, microfilming, and recording, or by any information storage or retrieval system, without prior
permission in writing from the publisher.
The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works.
or for resale. Specific permission must be obtained in writing from CRC Press LLC for such copying.
Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 3343 1.
Tkademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation, without intent to infringe.
Visit the CRC Press Web site at www.crcpress.com

O 2004 by CRC Press LLC
No claim to original U.S. Government works

International Standard Book Number 0-84932-767-9
Printed in the United States of America 1 2 3 4 5 6 7 8 9 0
Printed on acid-free paper
This book is dedicated to my patient wife, Ginette, to my grandchildren,

Walker, Olivia and Fiona and to Jamil who adopted me as his Tonton.
To me these children represent the future. W O

knows?
Perhaps there is a Sadi Carnot hidden in one ofthem.
m
Raymond Mulley
Flow of Industrial Fluids-Theory
and Equations
Contents
ListofTables
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .XI1
.
AbouttheAuehor
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XI11
.
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XV
.
Chapter I: Flow of Incompressible Fluids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
1-1:
Scope of Chapter - Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
1-2:
Flow of Incompressible Fluids in Conduits
1-3:
Flow Regimes - Reynolds' Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6
1-4:
Flow Profiles -Velocity Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1-5:
Fluid Flow -An "Irreversible" Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16
1-6:
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22
"Friction Losses" . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
Bernoulli Equation and the Darcy Equation Combined . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Conservation of Energy in Hydraulics Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40
Worked Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
ChapterSummary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
.
1-7:
1-8:
1-9:
1-10:
1-1 l :
1-12:
Fundamental Relationships of Fluid Flow
The Role of Viscosity
Chapter II: lncompressible Fluid Flow
.Losses of Mechanical Energy . . . . . . . . . . . . . . . . . . . . . . . .51
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .51
Reasoned Approach to Design .
A Little Personal Philosophy. . . . . . . . . . . . . . . . . . . . . . . .52
The Bernoulli Equation Revisited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53
Irreversibhties Due to Pipe and Fittings. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57
Examples of Estimations of Irreversibilities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80
Chapter Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98
11-1: Scope of Chapter - Applying Basic Concepts

11-2:
11-3:
11-4:
11-5:
11-6:
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .99
111-1: Scope of Chapter .
Pumps and Their Performance Capabilities .....................- 9 9
111-2: Functions of Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .100
111-3: A Brief History of Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104
111-4: Classification of Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
111-5: Characteristics of Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109
111-6: Inherent and Installed Characteristics of Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .130
111-7: Controlling Flow Through Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .135
111-8: Hydraulic Turbines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .141
111-9: Worked Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .143
111-10: Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .148
Chapter Ill: Pumps Theory and Equations
and Equations
Chapter IV: Compressible Fluid Flow

IV-1: Scope of Chapter
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .151
Comprehending Compressible Flow . . . . . . . . . . . . . . . . . . . . . . . . . . .151
IV-2: Differences between Compressible and Incompressible Flow . . . . . . . . . . . . . . . . . . . . . . . .152
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
.
Equations of Compressible Flow of an Ideal Gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .157
Ideal and Non-Ideal Gases .
Con~parisonof Some Equations-of-State . . . . . . . . . . . . . . . .166
Model Processes for Compressible Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .174
Choked Flow and the Mach Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
.
IV-3: UsingModels
IV-4: Treating Mixtures

IV-5:
IV-6:
IV-7:
IV-8:
IV-9: Equations for Adiabatic Flow with Irreversibilities not Involving the Mach Number .
the Peter Paige Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
.
IV-10: Equations for Isothermal Flow with Irreversibilities ...............................188
IV- l l :Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .192
.
V-l:
.Complex Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .195

Scope of Chapter .
Computations for Complicated Compressible Flow Systems . . . . . . . . 195
V-2:
Describing the Piping Network
V-3:
Describing the Flow Regime
V-4:
PlanofAttack
Chapter V: Compressible Fluid Flow
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .196
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .199
.
V-5:
.
Manifold Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .200
V-6:
Data Collection andverification
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206
Chapter Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .207
V-7:
Appendix AI: Equations of Incompressible Fluid Flow and Their Derivations . . . . . . . . . . . . . . . . . . 209
.
AI-1 : Purpose - Providing Chapter I Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .209
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .210
.
Pressure at a Point within a Fluid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
AI-2: S1 and Customary U.S. Units

AI-3:
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219
.
Friction Losses Explained . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .221
Force-Momentum Considerations forvariable Mass Systems . . . . . . . . . . . . . . . . . . . . . . . .224
Derivation of the Darcy Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .229
Derivation of the Bernoulli Equation Including Irreversibilities. . . . . . . . . . . . . . . . . . . . . .236
Larninar Flow and the Hagen-Poiseuille Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .245
AI-4: Hydrostatic Equhbrium

AI-5:
AI-6:
AI-7:
AI-8:
AI-9:
.
AI-10: Summary of Appendix AI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
Appendix All: Losses in Incompressible Fluid Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
AII-1: Purpose .
Providing Chapter I1 DetaJs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .253
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .254
AII-3: Relationship between Energy per Unit. Mass Units. Head Units and Pressure Units . . . . . . . . .256
AII-4: Churchdl-Usagi Friction Factor Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .257
.
AII-5: Pressure Drop versus "Friction Losses" . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
.
AII-6: K Factors - Loss Coefficients. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
.
AII-7: Summary of Appendix AI1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
AII-2: Relation of Valve Coefficient.
V111
to Loss Coe5cient7K
Flow o f Industrial Fluids-Theory
and Equations
Appendix Alll: Computations Involving Pumps for Liquids
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .273
AIII-1: Purpose .
Providing Chapter I11 D e d s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
.
.
AIII-2: Theory of Centrifugal Pumps. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .284
AIII-4: Real Centrihgal Pumps - Suction Lifi, Cavitation and NPSH . . . . . . . . . . . . . . . . . . . . . . . . 294
AIII-5: Positive Displacement Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .298
AIII-6: Theory and Analysis of Jet Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .299
.
AIII-7: Worked Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
.
AIII-8: Summary of Appendix AI11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
AIII-3: Pedormance of Real Centrifugal Pumps
Appendix AIV: Equations of Compressible Flow. Derivations and Applications

AIV-1: Purpose - Providing Chapter IV Equation Details
. . . . . . . . . . . . . . . .319
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .319
AIV-2: Using Thermodynamic Variables .

in Particular. Enthalpy . . . . . . . . . . . . . . . . . . . . . . . . . .320
AIV-3:Adiabatic and Irreversible Flow in Uniform Conduits
Basic Equations
. . . . . . . . . . . . . .321
AIV-4:The Peter Paige Equation. Choked Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .323
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .329
AIV-6: Adiabatic Choked Flow; F,' v. T Relationships Using the Redlich-Kwong Equation . . . . . . 350
.
AIV-7: Summary of Appendix AIV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .362
AIV-5: Choked Flow Using the Ideal Gas Equation
Appendix AV: Compressible Fluid Flow
.Complex Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .363
AV-1: Scope .
Estimating Complicated Pressure Drops and Flows . . . . . . . . . . . . . . . . . . . . . . . . .363
AV-2: Describing the Piping Network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .363
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .368
AV-4: Component Input Data. Eleven Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .369
.
AV-5:PlanofAttack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
AV-6: Irreversibilities Due to Form (and Mixing) Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .371
.
AV-7: Manifold Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
AV-8:Viscosity Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .376
AV-9: Simulation Results (Analytic Method) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .388
.
AV- 10: Summary of Appendix AV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
AV-3: Describing the Flow Regime
. . . . . . . . . . . . . . .395
.
Appendix D: Table of Caveats . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
.
Appendix E: Selected Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .a
.3
Appendix C: Table of Principle Symbols and Glossary of Principal Terms and Units
and Equations
Ix
Illustrations
Figure
1.1 .
. . . Page
Title
Reynolds apparatus
1-10
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.
Laminarflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6
Turbulentflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
Moody friction factor versus Reynolds' number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
Distribution of velocity across pipe. fully developed flow of Newtonian fluid. . . . . . . . . . . . . . 11
Velocity components downstream of close-coupled elbows in different planes . . . . . . . . . . . . . 13
Multiple sensor averaging techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14
Average. maximum and point velocities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
Developing profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
Sheardiagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23
.
1.11 .
Definition of viscosity
1.12 .
Thixotropic behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29
1.2
1.3 .
1.4 .
1.5 .
1.6 .
1.7 .
1.8 .
1.9 .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
1.13 .
Generalized viscosity chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -. 3 0
1.14 .
Interpretation of hydraulic "head"
1.15 .
11.1 .
11.2 .
11.3 .
11.4 .
11.5 .
11.6 .
.
.
11.9 .
11.7
11.8
111.1 .
111.2 .
1113.
111.4 .
111.5 .
111.6 .
111.7 .
111.8 .
IV.1
IV.2 .
IV.3
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -. 4 4
. . . . . . . . . . . . . . . . . . . . .60
Component loss coefficient correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67
Component loss coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68
Establishment of loss coefficients for a 90 degree mitered elbow. . . . . . . . . . . . . . . . . . . . . . . .71
Suddencontraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
.
Suddenexpansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73
Recovery and permanent losses across an orifice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .76
Example of system irreversibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87
Extension to system irreversibility example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94
Simple centrifugal pump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .106
Axial-flow elbow-type propeller pump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .106
Seven-stage diffuser-type pump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107
Simplified sketch of an air lift. showing submergence and total head. . . . . . . . . . . . . . . . . . . . 107
.
Ejectors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Typical head-capacity curves for centrifugal pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .111
Typical characteristic curves for reciprocating pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .124
Self-controlledpump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .136
Three models of compressible flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .174
Mass flow rate through nozzle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .178
Variation of pressure ratio with &stance from nozzle inlet . . . . . . . . . . . . . . . . . . . . . . . . . . . .178
Manometers in hydraulic practice
Moody friction factor versus Reynolds' number and pipe schedule
and Equations
.
First example of isometric sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
V.1 .
.
Second example of isometric sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .197
V.3 .
.
Manifold flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .201
V-4 .
.
Miller's K Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .202
V.2
AI.1
.
Pressure tetrahedron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .216
AI.2 .
AI-3 .
.
AI.5 .
AI.4
AI.6 .
AI.7 .
AI.8
AI-9 .
AIII.1 .
AIII-2 .
AIII-3 .
AIII.4
AIII.5 .
AIII.6
AIII-7 .
AIII.8 .
AIII.9 .
AIII.10 .
AIII.11 .
.
.
AIV.1
AIV.2
AV.1 .
AV.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .217
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Hydrostatic equilibrium
219
.
Variable mass systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Restrained system
227
.
Derivation of the Darcy equation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Development of Bernoulli equation
237
.
Horizontal laminar flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .246
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Hollow cylinder model for Poiseuille's law
249
Typical single-stage centrifugal pump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .273
.
Centrifugal pump showing Bernoulli stations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
Velocity vectors at entrance and discharge of centrifugal pump vane . . . . . . . . . . . . . . . . . . .277
.
Vector diagram at tip of centrifugal pump vane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Vector diagrams describing perfect and imperfect guidance . . . . . . . . . . . . . . . . . . . . . . . . . .285
.
Reduction in theoretical head due to various causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
Impeller and stationary vanes in dffuser pump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .287
Power versus volumetric flow in a centrifugal pump . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .288
Ejector analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .300
.
Jet pump configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .304
Suctionlift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .312
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Multiple choke points
338
Simultaneous equations - functions of two variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .346
Isometric sketch of complex vent header . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .365
.
Gas mixture viscosities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .386
Projected areas
and Equations
Tables
Figure
1.1 .
1.2 .
1.3 .
11.1 .
11.2
11.3 .
IV.1
. . . Page
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
Mechanically useful forms of specific energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40
.
Irreversibilities in turbulent flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Absolute roughness of various materials (Data from Crane) . . . . . . . . . . . . . . . . . . . . . . . . . . .59
Title
Fluid categorization by viscous behavior
Absolute roughness of various materials (Data from Miller)
. . . . . . . . . . . . . . . . . . . . . . . . . . .61
Results of computations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94

Comparison of equations-of-state for C O 2 with experimental data
. . . . . . . . . . . . . . . . . . . .169
AII.1 .
.
Kfactors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .262
AII.2 .
Classification of components
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .266
About the Author

Raymond Mulley was the original author of Control System Doc~mentation
Applying Symbols and Ident$cation (ISA, 1994). He joined ISA in 1973 and was
chairman of ISA SP5.1 the committee that developed ANSI/ISA-S5.1-1984,
Instrumentation Symbols and Identijkation.

During his more than 30 years of professional experience, he worked for Fluor
Daniel (Mississauga,Ontario, Canada, Rochester, New York, and irvine, California);
Dravo Chemplants (Pittsburgh, Pennsylvania) and International Nickel (Sudbury,
Ontario, Canada). Part of this experience was in private practice as a professional
engineer in the field of process hazards analysis and control. His experience ranges
from instrument and electrical supervisor to start-up engineer, instrument design
engineer, chief instrument engineer, director of design engineering and hazards analyst. His principal professional interests have been the flow of industrial fluids and
thermodynamics.
and Equations
XM
Preface
P - l : PURPOSE OF BOOK
INSIGHT
The purpose of this book is to supply adequate context for those who
really want an understandng of fluid flow. Every subject must be studied
within a certain context. If the context is too narrow, our knowledge of
the subject is limited, wrong assumptions are made, and errors result.
Specialization tends to narrow context. Generalization can make it so
large that we lose sight of important details. In this text, we hope to
supply enough context to give insight into the phenomena of flow.The
insight gained will hopefully aid in the solution of real problems.
This book was written with a specific audience in mind - all those who
wish to understand fluid flow.The book's purpose is to link fluid flow
theory to practice in sufficient detail to give its chosen audence an
understanding of both theory and practice. The theoretical detail is limited
to that necessary to understand practical problems - to the application of
equipment and devices, not to their design. Insight is the primary goal.
There are many texts that deal with instruments to measure and control
flow of fluids.They describe how devices work and only discuss fluid
mechanics and dynamics incidenta1ly.This text is intended to be complementary to the above mentioned ones. It takes a dfferent approach. It discusses fluid flow, so as to give the reader a clear grasp of the fundamentals
that impact his or her work.The fundamentals are then linked to entire
fluid systems.
Sometimes, the language of subject matter serves as an impediment to
insight into the subject.This is contrary to the intent of a textbook, but it
cannot be helped because of the necessity of using accepted terminology.
This book tries to make sure all terminology used is adequately explained.
The pressure losses due to fluid irreversibilities are covered.The general
characteristics of pumps, blowers, compressors, vacuum devices and other
prime movers are given.These characteristics are discussed in sufficient
detail necessary for the reader to understand some of the dynamic aspects
of fluid flow. In general, we try to show that a prime mover that adds
energy to a stream, or a device that extracts energy from a stream, should
be studied in the context of the total system through w h c h the fluid is
flowing. In particular, we analyze the limitations imposed on a system and
on its associated instrumentation by the inherent characteristics of prime
movers as these characteristics interact with those of the system.
and Equations
P - 2 : Why Study Fluid Flow i n D e t a i l ?
p
-
It seems to be taken for granted that process engineers should have a good
grasp of fluid flow. Mechanical engineers also need to understand fluid
flow. But, why should control systems personnel study the flow of industrial fluids? They could concentrate on the study of measuring instruments
and control valves. Thls introduction will attempt to answer that question.
The rest of the book will try to supply understandmg, methodology, theory
and equations useful to measure and control industrial fluid flow.
The practical importance of studying any subject is directly linked to the
frequency with which problems in the field of the study are encountered.
The most frequently encountered operation in most industrial processing
is the transfer of fluids from one place to another. Sometimes the transfer
is from point to point. Sometimes the flow &verges from one point to
many. Sometimes, the flow converges from many points to one. Each case
presents different sets of phenomena to be understood. Each case has different sets of constraints imposed upon it.
Fluid flow is always associated with a driving potential.This potential can
be constant, as when a head tank is used. It can vary with flow, as when a
pump, compressor, blower or vacuum-producing device supplies the
driving potential. Each prime mover provides its own set of characteristics
that must be understood if we are to correctly apply measurement and
control technology. Sometimes the purpose of the flowing stream is to
allow energy to be absorbed, as by a turbine.The turbine also imposes its
own characteristics on the flowing stream.
Basic characteristics of the prime movers must be understood in order to
establish the feasibility of any proposed application. The turndown capabilities of the system and those of the control devices are independent, but
they must be studied in tandem. Why pay for a flow meter with a 100:1
turndown if the system wdl operate only over a 3:l range? (There may be
other reasons than turndown for using a particular meter).
P-3: Flow i n Closed Channels

We tend to think of fluid as flowing only in pipes with circular sections.
This is the case for the majority of industrial applications. However, fluids
do flow in open channels and in ducts of non-circular sections.The pipes,
channels and ducts frequently vary in cross section.They have fittings and
transitions that cause a pressure drop different from that of the uniform
section. Each combination presents different problems to solve.
XVI
and Equations
Flow also occurs remotely from a channel wall - in the atmosphere, for
instance. Because of the necessity to put limits to all endeavors, the scope
of the book will be limited to that of fluid flowing through closed channels.This choice of scope excludes the consideration of related subjects
such as open channel flow, weirs and flumes and environmental dispersion
modeling.
Problems associated with the flow of liquids, slurries, vapors and gases
through ducts and channels are common in industry.The power requirements of pumps, blowers and compressors to cause a given flow are frequently needed. The pressure drops between two points at different elevations in a conduit is often required.The flow for a given pressure drop is
also required.
One of the most difficult flow computations associated with pressure relief
devices is that of computing the optimum size of a relief header. Here, the
driving pressure at each relief device discharge is unknown. In addition,
the pressure at the discharge pipe into the sink may be unknown because
of the possibility of choked flow. So, there is a minimum flow requirement
in a network of pipes with unknown upstream and downstream pressures.
This was the very problem that started the author on his search for a
better grasp of fluid flow phenomena.
Each fluid or mixture of fluids comes with its own pecu1iarities.Viscosities
and densities have to be estimated for pure fluids and for mixtures.Vapor
pressures of pure fluids and mixtures have to be considered if flashing is to
be avoided or accommodated and if cavitation is to be avoided.The
cholung phenomenon occurs with gases and flashing liquids. Slurries
behave differently from clean fluids. Some fluids are extremely aggressive
to industrial piping, to human beings, and to the environment.The characteristics of each fluid must be understood if we are to measure, control
and contain the fluid.
The permissible velocity of a fluid depends on the service. Allowable erosion effects, cavitation, cholung, flashing, slugging or mist flow in twophase situations must all be considered.Vortexing,pulsating flow, noise
generated by flow, flow regimes (single-, two- or three-phase flow) all
have to be predicted, if not measured and controlled.
and Equations
XW
P-4: A Word About Units

A misunderstanding of units is a common source of errors, both arithmetic and conceptual. It would be easy to decide to use only one set of
units in a book and to allow the reader to sort out the real world differences. This approach, however, would defeat the purpose of the book.
There are two sets of units prevalent in the industrial world.The set that is
given most ink is the S1 system (Systeme international d'unitks).The one
that is most common in the industrial context of North America is the
Customary US. System.
In the field of engineering and design work, mixed units are often used
due, frequently, to sectorial preferences. Conversion factors are applied
within formulae. Since most of us are faced with the problem of mixed
units on a daily basis, we have to deal with them.
In this book, we wdl use both S1 and the customary U.S. units as the basic
units. In addition, we will give some examples of mixed units. Rather than
have a section on units, we w d define them when initially presented. At
the risk of boring some, we wdl repeat the definitions from time to time
when the context seems to demand it.
P-5: Motivation
In pursuing a deeper knowledge of fluid flow, the reader should be motivated by this thought: If fluid flow is not understood, how can it be measured or controlled?
The practices of some engineering companies, in terms of division of
work, limit control systems personnel to operating on data produced by
process engineers. If such is the case, it is to be hoped that the process
engineer has access to the knowledge that ths book w d attempt to impart.
In that case, at a minimum, ths book wdl allow the control systems personnel to ask appropriate questions regarding the data that was supplied.
The best motivation for the reader to pursue h s or her quest for knowledge
of fluid flow should be that fluid flow and its measurement and control represent a fascinating study Our present knowledge of fluid flow is based on
the pioneering work of some intellectual giants. ~ e r n o u l hcarnot",
~,
coriolisiL,~ e l v i n~~e~~, n o l d s ~ , ~ e are
n t usome
r i ~ of
~ the names of people
whose contributions w d be dscussed and, hopefully, explained in ths book.
XVlll
and Equations
P-6: P l a n o f A t t a c k
The book is divided into two parts. The first part, consisting of the
chapter material, gives a common sense, logical description and explanation of fluid flow phenomena and of fluid machinery.The basic theory
and equations associated with flow and fluid machinery are given. As each
concept is introduced, an attempt wlll be made to make it relevant to
practical engineering problems.
The second part of the book, consisting of the material in the appendices,
takes a more formal approach and treats more complex problems. It gives
the detailed derivations of the equations found in the first part. It parallels
the first part in that each chapter of part one has a corresponding
appendix in part two.This approach will allow the reader to follow more
easily the logic in the first part. He or she may then use the second half
for reference or for more intensive, in-depth study.
In addition to the separation of the book into two parts based on a
common sense, logical approach, versus a more formal one with increased
complexity of detail, it is convenient to deal with so-called incompressible
fluids first and then with compressible ones.This allows a firm grasp of
more basic concepts before the more complex ones are presented. It
makes the more complex concepts easier to tackle. However, it is necessary to deal with some aspects of compressible flow during the discussions
of incompressible flow.This helps avoid the pitfalls of allowing the reader
to make false assumptions about when to apply equations that were developed for incompressible flow
The historical contributions of some of the actors in the field of fluid
mechanics, dynamics, measurement and control are of interest.They are
sprinkled throughout the book.
and Equations
XIX
Flow of
Incompressible Fluids
1-1: S C O P E O F C H A P T E R
- BASIC
CONCEPTS
Chapter I introduces basic concepts necessary to understand incompressible fluid flow in closed conduits.These concepts will be introduced and
described in logical, not necessarily historical, sequence, as they are
needed. The culmination of Chapter I will be introducing the generalized
Bernoullii equation - probably the single most useful equation for solving
fluid flow problems. Detailed computations will be left to subsequent
chapters and appendices. In particular, the concepts described in this
chapter will be developed in greater detail in Appendix AI.
Chapter I will show:
dividing fluids into incompressible ones and compressible ones is an
arbitrary construct that helps organize knowledge, but sometimes clouds
other insights;
there is a more rational way to describe mechanical energy losses than
by the term "fluid friction", which is a logical inconsistency;
in pipes and channels, there are two basic flow regimes - laminar and
turbulent - with an unstable regime between;
these regimes are well categorized by an easily computed number, the
ReynoldsVnumber;
there exists an equation, the Bernoullii equation, which can be used to
compute flow rates and pressure drop relationships;
there exists another equation, the
equation, whch relates irreversibhties ("friction losses7')to fluid viscosity, density and velocity.The
Darcy equation can be easily incorporated into the Bernoulli equation to
give a system for solving fluid flow, measurement and control problems.
arc^^"
We d start with a brief discussion of flow and what causes or influences it.
Then we w d &scuss the pioneering work of Sir Osborne Reynoldsv,introduce the Darcy equation and follow with the work of Daniel BernouILThs
chapter d consider only round ducts flowing liquid m.Greater detad wdl
be given in Appendur AI. Compressible fluids and other cross sections d
be discussed in subsequent chapters and detailed in subsequent appenhces.
and Equations
C H A P T E R
O N E
Flow o f Incompressible Fluids
1-2: FLOW OF INCOMPRESSIBLE FLUIDS I N CONDUITS

Some material in this section may seem self-evident to recent engineering graduates. However, it is worth establishing some basic concepts
as a springboard to what follows. It does not hurt to review basic material
from time to time. It is surprising what one forgets - or never learned.
Compressibility of all fluids

Many terms used in the field of hydraulics are simplified descriptions of
reality. Sometimes these didactic simplifications are useful; sometimes they
are not.They help younger people over early hurdles and hinder them in
their later comprehension.The name of the term, itself, can lead to conceptual errors. For instance, the distinction between an incompressible
fluid and a compressible one is just a matter of degree.There may be
orders of magnitude between the compressibility of a gas and that of a
liquid; nevertheless, it is wise to bear in mind that all fluids are compressible to some degree - even the ones we call "incompressible". Later, we wdl
show how compressibility helps explain another logical inconsistency fluid friction.
Mechanism of flow
It is obvious to anyone who has ever used a garden hose that water flows
when there is a higher pressure upstream than downstream.When we
open a tap, or faucet, the higher pressure in the pipe causes water to flow.
This common observation leads to the assumption that a higher pressure
upstream than downstream is always associated with flow.
The assumption is not correct. It is true a higher pressure is required
upstream to start a fluid flowing across a restriction. However, once a fluid
is flowing, it has momentum. As will be developed in greater detail when
we discuss the ~ernoulli'equation, pressure is associated with one form of
energy, static energy, and energy can take on different forms. One of these
forms is kinetic energy.When a high velocity fluid flows from a small
diameter pipe through a swage to a larger diameter pipe, the fluid slows
down because of the continuity principle. In slowing down, some of its
lunetic energy is changed to static energy.This change can cause the pressure downstream to be greater than the pressure upstream of the swage.
This phenomenon is called pressure recovery.
It is not necessary to have a higher pressure upstream in order to initiate
flow. If a pipe is connected to an open reservoir - for instance, a water
and Equations
Flow of Incompressible
Fluids
C H A P T E R
O N E
tower in a city water distribution network - flow will begin once a valve
is opened.The pressure on the surface of the water in the reservoir is
approximately equal to the pressure in a htchen sink. In this case, there is
no difference in pressure between the ends of the piping system, but flow
occurs.There is more than just pressure difference involved in the flow of
water.An unbalanced force in the direction of the flow is necessary.
Re-establishing equilibrium
The phenomenon of flow in conduits is best thought of as an attempt to
re-establish an equilibrium state in a fluid constrained by conduit walls. In
a closed system under pressure, when all valves are shut, all forces are
accompanied by equal and opposite forces.The fluid is stationary. In an
open system, when valves are open, there is an imbalance of forces and
the fluid moves in a direction that will re-establish equilibrium. If mass
and energy are added continuously to an open system, flow will be continuous.
The mass added could be water flowing to a reservoir behind a dam, or
water pumped to a head tank. Once there, the water has energy by virtue
of its position.This energy is potential energy or the potential to do
work. Upstream of a closed tap, faucet or valve, the water will be subject
to the weight (force) of the water above it. It will exert a pressure (force
per unit area) to equilibrate this force.This pressure is the result of repulsion arising when molecules approach one another. The pressure is associated with a static fluid and with a form of energy called static energy.
If the valve is opened, the water will flow through it by virtue of the
pressure difference across it.The pressure upstream of the valve will
decrease for two reasons. First, some of the pressure energy has been converted to hnetic energy, the energy a body has by virtue of its motion.
Second, some of the energy of the water has been converted to internal
energy because of what is commonly called "fluid friction" in the
upstream piping.The increase in internal energy is accompanied by an
increase in temperature - hence, the loose term, "thermal energy". This
loose term shortly will be explained more fully.
The lunetic energy is recoverable. It can be transformed back to static
(pressure) energy by closing the valve or to potential energy by connecting an open, vertical pipe to an elevation equal to that of the reservoir. The incremental internal energy is not completely recoverable to
one of the mechanical forms of energy.This is the reason the water in the
and Equations
C H A P T E R
O N E
F l o w o f incompressible F l u i d s
previously mentioned vertical pipe cannot quite reach the original elevation of the surface of the reservoir. The incremental internal energy flows
as heat through the pipe walls and is lost to practical use. It warms up the
atmosphere. In a well-insulated pipe, it remains with the fluid for a longer
period of time and is seen as a higher fluid temperature.
Logical inconsistency of fluid friction

The term "fluid friction" is used commonly to explain losses in mechanical forms of energy associated with fluid flow. A little knowledge of
physics shows it is a poorly chosen term. Friction is a macroscopic phenomenon - such as when two blocks of wood slide over one another, or
when the rubber hits the road.There is a definite resisting force to be
overcome - the force of friction.
Fluid molecules do not touch

Fluids are made up of molecules, and molecules do not touch in the fluid
state. Collisions of gas molecules are actually repulsions that prevent the
molecules from touching. Liquid molecules are less energetic than gas
molecules.They slide past one another or remain at an equilibrium distance.The equilibrium distance is established by a balance between an
attractive force and a repulsive force. If molecules do not touch, how can
there be "f?iction"?
The answer is found in the compressive forces that occur when local flow
is slowed down by obstructions of any kind.These obstructions can be
large objects such as valves or fittings; they can be edmes within the
flowing stream; they can be other molecules that have diffused between
two layers of fluid moving at different velocities.
Compression involves work

Whenever a portion of a fluid is compressed, work is done on it by an
external force.This work increases the internal energy, and the temperature increases.There has been a conversion of mechanical energy to
internal energy.Temperature increase is not uniform throughout a fluid,
especially in turbulent flow. The direct contact of different temperature
fluids is a classical case of an inefficiency: a loss in the capability to convert heat flow to useful work. In thermodynamics, the term used to
describe this natural phenomenon is "irreversibility".
The concept of irreversibility describes a law of thermodynamics which
states that work can be converted completely to internal energy or to
4
and Equations
Flow of Incompressible Fluids
C H A P T E R
O N E
heat flow. However, heat flow cannot be completely converted to work

energy in any cyclic process - no matter how perfect. Some of the heat
energy must flow to a lower temperature (the environment) or the energy
associated with it must remain in the system as internal energy. The bulk
temperature will be higher than that of the fluid before it was compressed.'
Heat and work are energy in transit across a boundary. Temperature is a
measure of the intensity of energy, not of its quantity. Once heat and
work no longer cross a boundary, we cannot recognize them as heat or
work. They simply take some other form - internal energy, for instance.
Many engineers still talk of heat as if it were a substance, like caloric, as if
it were a property of matter. It is not.
In this book, we wdl try to avoid tallung loosely about heat as if it were a
property of matter. We will also try to avoid the logical inconsistency by
using the term "irreversibility" in place of "friction loss" and the plural
"irreversibilities" in place of "losses". This concept will be more fully
developed in Appendix AI.
Equilibrium and energy transformations

To recapitulate, flow can be thought of as a means of re-establishing an
equilibrium state in energy levels. Energy exists in various forms, some of
which can be converted relatively easily to other forms. Potential energy
can be converted to static energy and to kinetic energy. Any of the three
forms of mechanical energy can be interconverted by applying suitable
constraints. Kinetic energy can also be converted to internal energy by
turbulence or any other deceleration. Once in this form, there is no
means of converting it completely back to one of the more useful forms
of mechanical energy.This is a law of nature.
The study of fluid flow is really the study of energy transformations.
Energy is indestructible; its total amount remains unchanged. Mechanical
energy converted to energy flowing as heat between two temperatures
represents a loss of useful mechanical energy, not of total energy. It is this
unidirectional change of mechanical to internal energy and the subsequent flow of heat to a lower temperature sink that explains the irreversibilities associated with flowing fluids.
and Equations
C H A P T E R
O N E
ReynoldsVperformed a series of correlations in 1883 based on the equipment shown in Figure 1-1. His apparatus consisted of a transparent tank in
which a constant level of water was maintained. The source of the makeup water is not shown in the figure.A transparent tube, open within the
tank, was used to bleed off water at controlled rates. A fine tube was used
to introduce a colored stream to the mouth of the larger transparent tube.
The apparatus allowed Reynolds to see clearly the difference between
two flow regimes and to see the transitional regime between them.
W
&g
0
.M
$2
:z
62
'2.3 g
2 ;,
U
c WF 2 2
U
G.-
273%
A . p E
g2.&2
cn.2
Figure 1-1. Reynolds'apparatus
Laminar flow
In the first flow regime, laminar flow, the colored stream occupies the very
central core of the transparent tube.There is no lateral mixing over the full
length of the tube.This regime occurs at very low flow rates (Figure 1-2).
Colored dye
Figure 1-2. Laminar flow
Turbulent flow
In the second flow regime, turbulent flow, the colored stream quickly
becomes fully mixed with the clear stream.There is obviously much lateral mixkg due to eddies in the turbulent stream.This regime occurs at
higher flow rates (Figure 1-3).
and Equations
L3
83:
C H A P T E R
O N E
Colored dye
Figure I-3.Turbulent flow
Transitional flow
The change from laminar flow to turbulent flow occurs over a very small
increment in flow.The regime within this increment is termed transitional flow. The transitional flow regime is unstable; it is a regime to be
avoided.The characteristics are not uniform nor reproducible.
Recognizable and quantifiable flow regimes
The contribution of Reynolds", h s co-workers and their successors, was in the
defimtion of recognizable and quantdiable flow regirnes.They also gave us an
index that can be computed fiom the conhtions of flow and fiom the fluid
characteristics, the Reynolds number.The Reynolds number can be used to
idenhfji the flow regime. Further, ths index can be used to calculate a fi-iction
(mechanical energy loss) factor that, in turn, allows calculating pressure drop
and irreversibhties.The Reynolds number is also important in establishng the
turndown capabhties and operating accuracies of metering devices.
Transitional flow, an unstable flow regime, occurs between the laminar
and turbulent flow regimes. In the transitional region, turbulent flow and
laminar flow alternate in an irregular manner.
The flow regime impacts our ability to accurately measure flow. It also
changes the sizes required of control valves.The pressure drop in turbulent flow is proportional to flow squared. In laminar flow it is directly
proportional to the flow rate. The equations used are different.
Reynolds number
Fortunately, the index of the flow regimes mentioned above, the
Reynolds number, NRe,allows identification of the flow regime. It enables
us to tell whether we will cross from one regime into another. It allows
us to choose the correct equations to use for both measuring devices and
control valves. It also allows us to compute with a high degree of accuracy
the mechanical energy losses in conduits and the associated pressure drop.
and Equations
C H A P T E R
O N E
The Reynolds" number is a dimensionless ratio used extensively in flow

correlations and "loss" computations. It is a very important number.
The Reynolds number is often regarded as the ratio of inertial forces to
viscous forces - the tendency to keep moving versus the tendency to
slow down. It is a useful correlating parameter for all industrial liquid
flows. It also applies to industrial vapor and gaseous flows.The Reynolds
number is usually given as:
The Reynolds number is a pure number - all dimensions cancel. D is

usually a diameter or some other length dimension, U is the average
velocity across the section, rho is density, mu is absolute viscosity (to be
defined in I-7), and G is mass velocity (velocity times density or mass
flow rate divided by area).The mean velocity across a section, U, used in
most computations, must be correlated with factors influencing the
velocity profile. These factors have been found mainly to be p, p, D, the
viscosity, the density and the internal diameter.
The numerator of the first equation could have dimensions of pipe diameter, D, in feet, average fluid velocity, U, in feet per second, and fluid density, p, in pounds-mass per foot cubed. Multiplying the dimensions and
cancelling gives units of pounds-mass over seconds times feet in the
numerator.
The denominator has units of "absolute" viscosity, which in the customary U.S. system are also given in pounds-mass over seconds times feet.
The units of the numerator and the denominator cancel leaving a dimensionless, "pure" number.
The last equation of the set, 1-1, is fi-equently seen. It is a mixed bag of older
metric (cgs or centimeter-gram-second) units and English-American units.
The viscosity is in centipoise or one hundredth of a poise.The poise is
defined as one gram divided by the product of one centimeter and one
second.We give the conversion in detail in Appendm AI.
The coefficient of the last term of Equation 1-1 has dimensions.These
dimensions cancel the mixed units ofW, lb,/h, d, pipe internal diameter,
inches, and p, centipoise.The expression is commonly used in North
American flow computations.
8
and Equations
C H A P T E R
O N E
All three equations produce the same ReynoldsVnumber for the same
conditions of flow.
Viscosity is defined in section 1-7 of this chapter and it is covered in more
detail in Appendix AI.
Different types of Reynolds numbers
It should be noted a Reynolds number can be defined for different circumstances and the appropriate dimensions should be used. The reason
there are several Reynolds numbers is the number is used for purposes of
correlation of data. For convenience, the length dimension, D, may change
in a given correlation. It is defined for flow in pipes (where D is usually
the internal diameter of the pipe). It is defined for flow past bluff bodies
(where D is usually the projected width of the body normal to the flow).
It is defined for flow past aeroplane wings (where D is usually the dimension of the wing in the direction of the flow).The velocity used is sometimes the average velocity across a pipe, sometimes the average velocity
across an orifice plate bore, sometimes a point velocity.
For instance, a Reynolds number defined for a pipe requires that D be
the pipe's internal diameter.Velocity, U, is usually the average velocity
across that section, but may be a point velocity. Density, rho, is the
average density at the section.Viscosity, mu, is the average viscosity at the
section. All units must be compatible so they cancel.
Another Reynolds number is defined in terms of hole diameter of an orifice plate. In this case, U is the average velocity across the section of the
orifice.The appropriate numbers must be used if the correct conclusions
are to be drawn from correlations. Subscripts are usually employed to differentiate among the various Reynolds numbers.
Caveat
- Reynolds numbers
Always check very carefully the definition of the Reynolds number for
the correlation being used.
In this book we will use primarily the pipe (based on the internal diameter of the pipe) Reynolds number symbolized by NReD.Thevelocity, U,
will be the average velocity across the section of pipe, not the point
velocity or the centerline velocity. We will specifically define other
Reynolds numbers when they are introduced.
and Equations
C H A P T E R
O N E
Flow o f Incompressible
Fluids
Any units can be used for the variables of the Reynoldsv number provided they are consistent, they cancel and produce a dimensionless
number. Mixed units may be used, provided the correct conversion factors are included to allow the group of numbers to be dimensionless.We
will deal thoroughly with real numbers later in the book.
Figure 1-4. ~ o o d ~ "friction

"'
factor versus Reynolds number
In circular pipe at a pipe Reynolds number below 2,000, flow is laminar.

At Reynolds numbers greater than 4,000, flow is generally considered
turbulent. However, to be sure, since laminar conditions sometimes persist
at greater flow rates, it is best to consider 10,000 the lower limit of the
turbulent zone. The range of Reynolds numbers between 2,000 and
4,000 flow is considered to be in the transitional zone. Remember that
transitional flow is not stable, it fluctuates. So, when you have control over
the design of a system, for most metering purposes, make sure you are in
and Equations
C H A P T E R
O N E
one regime or the other over the full range of flows. If you do not have
control over the design, then compute the limitations imposed on the
system by the actual range of flows.
Figure 1-4 is a plot of the ~ o o d ~ " friction
"'
factor versus the Reynoldsv
number. Relative roughness of the pipe wall is a parameter. It shows the
distinctions among the three regimes.
1-4: FLOW P R O F I L E S
- VELOCITY
DISTRIBUTIONS
A flow profile is the plot of velocities across a section of pipe. Flow profiles are influenced by the behavior of the Reynolds number and particularly by its viscosity component. Not all profiles are as smooth as those
described in Figure 1-5. As we shall see a little later, fluids are generally
characterized by their viscosities. One type of fluid, a Bingham plastic
(typical of sewage) can have a central core that flows as a plug.The surrounding annulus has a more normal velocity distribution.
OS0 0.20 (X30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Fraction of maximum velocity
Figure 1-5. Distribution of velocity across pipe, fully developed flow of

Newtonian fluid
Influences on flow profiles

Proximity to the pipe inlet from a vessel is another factor influencing the
velocity distribution. At least 30 pipe diameters of length must be allowed
to develop a steady profile in incompressible flow.
Bends and partial obstructions can distort the profile to such a degree that
the average velocity, U, cannot be inferred with any degree of accuracy
from a measuring device immediately downstream of the disturbance.
and Equations
C H A P T E R
O N E
This is the reason upstream and downstream straight lengths are specified
for almost all measuring devices. It is also the reason for the existence of
straightening vanes.
Straightening vanes are axial inserts in pipes.They are designed to cause
flow to be parallel to the axis so, shortly downstream of the insert, the
flow resumes its fully developed profile - the one for which the measuring device was designed.
Irreversibility
In turbulent flow, large eddies break up into smaller ones, and the smaller
ones into even smaller eddies, until they finally disappear. Bulk kinetic
energy is transformed into rotational energy.This is a transformation of
one form of mechanical energy to another.What is generally termed
"fluid friction" is mechanical energy transformed finally to thermal
(internal) energy by compression as the fluid is decelerated. It is this conversion from mechanically useful lunetic energy to mechanically useless
internal energy that results in the concept of6'lost energyW.Thisterm
would be better stated as "lost mechanical energyV.Wewill use the thermodynamic term "irreversibility" to represent this conversion.
In laminar flow in pipes, concentric annuli flow past one another with no
net normal flow. Molecules, however, do bounce back and forth across
the imaginary boundaries of the annuli. Compression takes place as faster
moving molecules in the bulk flow are decelerated by the proximity of
the slower moving ones. Mechanical energy is converted to internal
energy as work is done in decelerating the molecules.
The loss in mechanical energy due to irreversibhties must be made up by the
prime mover or by other means if steady-state flows are to be maintained.
Flow profiles
Some instruments are more sensitive than others to the flow profile in a
conduit.The flow profile is the plot of the axial components of the point
velocities as the conduit is traversed. Figure 1-5 is an example of fully developed flow in a circular pipe, remote from any dsturbance, including that
due to the conduit inlet.The figure depicts both laminar flow and turbulent
flow, but note they cannot both be present at the same instant. Note the
flattening of the profile in turbulent flow and note there is no attempt
made to show a transitional flow profile because transitional flow fluctuates.
12
and Equations
C H A P T E R
O N E
The point velocity at the center of the pipe is twice the average velocity
in larninar flow. It is a factor, alpha (greater than, but close to one), times
the average velocity in turbulent flow.We will develop the relationships
more exactly later in the book. The velocity profile must be taken into
consideration when measuring flow.This is done by computing the range
of Reynoldsv numbers over which flow will be measured. Controlling the
pipe geometry with straightening vanes or using minimum lengths of
straight pipe are the usual means of fixing the limits of the flow profiles.
The point velocities can be obtained by traversing a Pitot tube across the
conduit and taking point readings.
It is to be noted that both curves of Figure 1-5 have zero velocity at the
conduit wall. Both curves have a maximum velocity on the center line.
The curves are drawn so the maximum velocities coincide. Conclusions
should not be reached regarding the same maximum velocity for the same
pressure drop in both types of flow. Figure 1-6 shows that, after different
types of obstructions, the velocity profile can become very distorted. The
profile can not only be distorted over a longitudinal section, it can even
rotate axiaIly.The axis of rotation is not necessarily the pipe axis.
Influence of flow profiles on measuring instruments.

Actual
'velocity
Figure I-6.Velocity components downstream of close-coupled elbows in

different planes
Most measuring instruments are used to infer the average flow in a duct.
They do not measure flow directly.These devices are sometimes called
inferential meters.
A device such as a Pitot tube measures the impact pressure and the static
pressure at a point. Impact pressure is the static pressure at the point plus
the pressure difference created by decelerating the fluid to zero velocity.
The point velocity is computed from the difference between this impact
pressure and the static pressure at the section.This difference is often
called the velocity head.The word "head" arises from the use of open
manometers to measure differential pressure in early hydraulic work.
and Equations
C H A P T E R
O N E
Sensor location
Most of the time, we are interested in the average flow through the section; therefore, it should be obvious the location of the sensor is critical to
the accuracy of the inference. From Figure 1-6, it should be evident that a
point sensor located within a distorted profile wdl give erroneous readings.
In addition, if the profile rotates, as it will after two close pipe bends, a
pulsating signal will be generated which may not be integrated correctly.
A predictable, reproducible velocity profile is necessary for accurate measurement with most flow meters.
~ e ~ e (Flow
l " Measurement, ISA) gives a good description of the problem
of measuring flow when profiles are distorted. Figure 1-7 compares the
results of seven different measurement points across the same section of
pipe with the same flow rates but with three different profiles. He makes
the argument for using averaging techniques when dealing with distorted
flows. He also references I S 0 standards for specific techniques.
SENSOR l
7.68
7.90
9.62
SENSOR2
9.04
9.16
10.28
SENSOR3
9.04
8.98
7.79
SENSOR4
AWBAGE
7.68
8.36
7.70
8.43
5.75
8.36
Figure 1-7. Multiple sensor averaging techniques
Figure 1-8 explains the differences among mean, maximum and point
velocities. It should be noticed that the profile is uniformly smooth.
Distortion of the profile will give numbers that cannot easily be used in a
correlation.
14
and Equations
C H A P T E R
O N E
Figure 1-8. Average, maximum and point velocities
U
Developing profile
Developed profile
Figure 1-9. Developing profile
Figure 1-9 shows a profile that may not be fully developed for about
thirty pipe diameters.
Devices such as orifice plates, which infer flow rates from differential pressure, are subject to profile limitations. In other words, if the profile is not
known or controlled, the accuracy is not trustworthy. Some devices, such
as magnetic flow meters, are less susceptible to flow profile influences.
However, even a magnetic flow meter will lose some of its accuracy, due
to flow profile disturbance, if it is connected dnectly to a pipe elbow.
The extreme, non-continuous flow profile of Bingham plastic flow must
obviously be considered if we wish to obtain metering accuracy on
Bingham plastics.
A Coriolis mass flow meter, which is a true mass flow meter, is not at all
influenced by the flow profile, but is limited in available diameters.
The influence of the flow profile on metering accuracy is the reason so
much time is spent establishing the required upstream and downstream
straight runs for piping around a metering device.When these minimum
requirements cannot be met, flow straightening vanes are used to correct
the flow profile, or one simply gives up on the accuracy requirement and
uses the metered flow rate as an approximation.
and Equations
15
C H A P T E R
O N E
1-5: F L U I D F L O W
- AN
'IRREVERSIBLE' PROCESS
Thermodynamics teaches that fluid flow is an irreversible process. The

sense of the word "irreversible", in this case, is useful available mechanical
energy has been transformed into internal energy, or heat flow, without
useful work being extracted. The system and the environment cannot be
restored to their original states without additional work being done. The
irreversibility is generally the unavoidable flow of heat energy generated
by turbulence to a lower temperature sink, either in the environment or
in the same flowing fluid-There is no practical way of recovering this
incremental internal energy as mechanical energy.
Flow of a fluid through a horizontal duct of constant cross section is always
associated with a pressure drop. If the fluid flow through this duct were
"ideal", there would be no losses, and there would be no pressure drop due
to irreversibhties."Ideal" flow would not require adding energy to maintain
itself. Once established, flow would continue without external help. Real
flow requires addmg mechanical energy to make up for the constant conversion to internal energy and, ultimately, to heat flow.
If we dlvide the pressure drop, using customary U.S. units, by the fluid density, we see the units are foot-pounds-force per pound-mass or energy per
unit mass. In the S1 system, the units are newton-meters per kg or joules per
kg. In irreversible (normal) processes, the energy associated with pressure
drop due to irreversibhties is commonly described as having "dlsappeared"
or as having been "lost". It does not disappear and it is only lost to mechanical use. It is transformed to internal energy, whch results in an increase in
temperature. The increase in temperature usually causes the internal energy
to be dssipated to the environment as a heat flow.
Irreversible factors i n f l u i d flow

What follows is a list of some factors that cause losses of mechanical
energy (irreversibilities) in fluid flow.
Viscous drag forces
Force has to be applied to keep fluid layers moving past one another, even
at steady state.The reluctance of fluid layers to slide past one another is
termed viscosity. Later, we shall develop the equations that describe how
some static energy is converted to kinetic energy and how the rest is converted to incremental internal energyviscous drag forces are the principal
forces that we deal with in solving fluid flow problems.
16
and Equations
C H A P T E R
O N E
Fluid mixing
When two different streams are mixed, mechanical energy is dissipated
(transformed to internal energy) due to turbulence. Turbulence results
from energy gradients within the flowing fluid. It results from the fluid
attempting to redistribute energy, to achieve equilibrium. Ideally, this
energy could have been recovered in an imaginary turbine or a hypothetical carnotii engine, but was not. It is often stated that nature abhors a
vacuum. It would be more accurate to say than nature abhors an energy
gradient. Effectively, the natural state is one of equilibrium, as was discussed in Section 1-2.
An example of fluid mixing is when two different pressure safety valves
relieve simultaneously into the same header. Fluid is also mixed when it
first separates around obstructions and then the separated streams rejoin.
Vortex shedding sensors (bluff bodies) and control valves are examples of
such mixing processes. The mechanical energy conversion to internal
energy due to this cause is often called "form friction" as opposed to the
"skin friction" of normal pipe flows.
Pressure and temperature inequalities

Pressure gradients in a flowing stream usually cause a dissipation of useful
mechanical energy to non-recoverable internal energy.Temperature gradients usually result in loss of potential to do work. (A carnotii engine
could have recovered some of the internal energy).
Chemical reactions
Industrial chemical reactions usually are associated with irreversibilities.
Normally, chemical reactions are confined to specific reactors and we are
only concerned with the flow of the reactants and products. Sometimes,
however, the reaction takes place in a pipe or in a distillation column and
the physical properties of the flowing fluids change along the flow path.
You may be asked to calculate or measure or control the flow.
1-6: F U N D A M E N T A L R E L A T I O N S H I P S OF F L U I D FLOW
There are three fundamental relationships to which we will return time
and time again to solve fluid flow problems.These three fundamental relationships involve the conservation of mass, the conservation of energy and
the ~ernoulli'relationship. Actually, the Bernoulli relationship is a specific
application of the more fundamental relationship involving Newton's
second law.
and Equations
C H A P T E R
O N E
Conservation o f mass
Mass is not destroyed. It can flow into, accumulate within, and flow out of
a system. In the steady state - and fortunately many flow problems can be
reduced to steady state problems - what flows in equals what flows out.
There is no accumulation. If there is only one incoming flow line and
one outgoing one, the steady state situation can be described as follows,
(1-2)
m1 = m2
0
The first equation, above, simply states that the mass flow rates into and out
of the system at the steady state are equal when there is no accumulation.
The second equation is equivalent to the first, but introduces three different variables: area of the conduit, average velocity, and density. If the
density is constant, the second equation reduces to a statement that the
volumetric flow in and out is constant.This is the case with most liquid
flows, but not with gases or vapors.
The third equation defines mass velocity or mass flux as being the
product of the average velocity across a section and density (or the quotient of average velocity and mass volume). It has units of mass flow per
unit area.The fourth equation makes use of mass velocity. It is a useful
form under certain circumstances (compressible fluids) which wdl be
explained later, when we discuss compressible fluids.
Caveat
- Average or point velocities

Always be aware of the velocity that is being considered, discussed or
defined. Whether we are tallung about average or point velocities can
make a huge difference in computations of kinetic energy, friction factors,
ReynoldsVnumbers and losses of mechanical energy.We will try to be
specific each time this problem arises.
Conservation o f energy
The first law of thermodynamics deals with the conservation of energy.
It cannot be bypassed. It applies to all systems, reversible and irreversible.
However, it is not sufficient to answer all problems (This will become
clear in subsequent developments).For open (flowing), steady state systems,
18
and Equations
C H A P T E R
O N E
the first law is expressed in two forms. One form is the differential form
useful for the developing relationships.The second form is the integrated
form that is more useful when relationships already developed have to be
applied in practice. Both forms will be put on a per-unit-mass basis. All
practical equations will ultimately be in the integrated form.
Differential equations are necessary to completely understand the development of solutions.
UdU gdX
a q + 6 w n =dh+-+gc
gc
In the first equation of the set 1-3, the Greek deltas on the differential
heat and work terms signify that these differentials are not "exact", they
are dependent on the "path" between equilibrium states. The dimensional
constant, gc,has been retained for those who deal with American customary units. The units of g and gc are different - f/s2 and lb,-ft/lbfs2. If
you are in the habit of only working with S1 units, just regard the constant as equal to one and as being dimensionless.The deltas on the terms
on the right of the integrated form indicate we are dealing with differences in quantities measured between two points.The heat and work
quantities will be the corresponding heat and work flows per unit mass
between the same two points.
The above forms of the equations also assume that the user knows
enough to convert mechanical units associated with the work, velocity
and elevation terms to energy units associated with the heat and enthalpy
terms.This will be explained more fully later. In this book we will use the
convention that energy added to a body is positive, energy leaving is negative (a different convention requires that work added be negative).
The first law for open systems, as described in the set 1-3, simply states
that energy in transit across the boundaries of a system as heat flow and
work flow per unit mass results in changes in three identifiable properties
of the flowing fluid.The first is the enthalpy, which is the internal energy
of the fluid plus the pressure-volume work it took to get the fluid into
and out of the system.The second is the lunetic energy term, which is
the specific energy due to bulk motion and which will be used to derive
the equations of flow through head meters. The last term is simply the
change in energy due to change in elevation - potential energy. All terms
are on a per unit mass basis.
and Equations
19
C H A P T E R
O N E
The first law requires calculating heat energy transfer, q, per-unit-mass or,
Q, total heat energy transfer between two states.This calculation is often
not feasible in the case of an industrial piping system.Therefore, the first
law has limited direct application in fluid flow computations.
Force balance equation (Bernoulli's equation)

The ~ernoulli'equation is one of the most useful equations ever developed as far as fluid flow is concerned. Its derivation will be given in
Appendix AI. Bernoulli's original equation was for the ideal situation. It
did not include a mechanical energy loss term.
A more modern version of the Bernoulli equation is given below.
The term on the left is the difference in specific energy change between
the two points due to changes in static pressure and mass volume.The
negative sign is due to the fact that the pressure downstream is less than
that upstream. If the fluid can be considered incompressible, the term may
be integrated immediately to v(P1 - P2),mass volume times pressure difference, or the equivalent expression (P1 - P2)/p, pressure difference
divided by density. Note that in this case the pressure difference is the
upstream less the downstream pressure because of the negative sign.
The term, hf, is the mechanical energy converted to thermal energy (irreversibility).The next term represents kinetic energy.The last term is the
change in energy per unit mass associated with the change in elevation
between sections 1 and 2.
The deltas are simply the differences between the variables, u2and X, at
the two points represented by the limits of integration 1 and 2.The deltas
mean the differences between downstream and upstream conditions. This
is the normal convention.
If a pump, blower or a turbine is included between points 1 and 2, an
appropriate work term is added on the 1eft.The sign would be positive
for a pump or blower, negative for a turbine.
20
and Equations
C H A P T E R
O N E
Units of each term

Each term in the above equations has units of energy per unit mass
flowing.The term, W,, is the net work added to the fluid by a pump or
blower per unit mass of flowing fluid. It is the net work extracted in the
case of a turbine. The term "net" may be misleading. It is actually the
work added to or subtracted from the fluid. In the incompressible case the
term under the integral simply reduces to the negative of the pressure difference at the two sections divided by the density.The downstream section carries the number 2.
The first term on the right is not enthalpy. It is a loss term. It is that part
of the available energy converted to internal energy or heat flow because
of irreversibilities (shock, compression, "fluid friction"). The next term is
the change in kinetic energy and the last term is the change in elevation
head or potential energy.
The loss term, h. is always positive. Energy may be transferred in either
direction among the other terms. Energy can only be transferred to the
loss term, not from it. Equations for the loss term, hf ,have been developed experimentally for both laminar flow and turbulent flow. The
equation is the result.The Darcy equation will be described later
in this chapter and in detail in Appendix AI.
arc^^"
The ~ernoulh'equation in the form of equation 1-5 makes up for some of

the shortcomings of the first law when calculating fluid flow. From it are
derived the flow equations for head meters and control valves. It allows the
computation of condtions at each end of a conduit when the dmensions,
the change in elevation, the configuration and quahty of the wall are known.
Alternatively it allows the computation of &mensions when the flowing
quantities and routing are known. It warrants memorizing.
Caveat
- Bernoulli equation versus the first law

The resemblance of the first law equation to the Bernoulli equation and
the use of the symbol, h, to represent two different quantities, enthalpy
and lost mechanical energy, can lead to confusion.
and Equations
C H A P T E R
O N E
1-7: T H E R O L E O F V I S C O S I T Y
The ideal model for the concept of viscosity is the newtonian model that
wdl be dlscussed shortly. For the moment, the reader is invited to regard viscosity as simply a coefficient that is the ratio of two terms.The numerator is
the force per unit area applied to the moving element of two coaxial cylinders containing a fluid between them.The denominator is the velocity gradent set up in the fluid as its velocity varies between that of the moving
element and that of a stationary, concentric, inner element (zero).
Viscosity has been described as the reluctance of a fluid layer to slide over
another one. It might be better defined as a measure of the attractive forces
between molecules - forces that must be overcome if there is to be motion.
It was shown to play a key role in determining the value of the ReynoldsV
number, whch in turn determines the flow regime. Knowledge of viscosity
helps determine the mechanical energy losses in a duct or conduit.Viscosity
also influences the discharge coefficients of orifice plates and the sizes of
control valves - both through the Reynolds number. It also has an influence
on the velocity pr0file.h extreme case of this influence is the non-udorm
profile of Bingham plastic flow.Viscosity is not a thermodynamic variable.
It is, however, a property of fluid matter. It is a convenient mathematical
coefficient that varies with pressure, temperature, composition and, sometimes, with applied stress and time.
The term "viscosity" applies to fluids.We have dlscussed fluids without
defining them on the assumption everyone knows what a fluid is. From the
point of view of fluid mechanics, a fluid is a substance, gas, vapor or liquid
(or slurry) whch undergoes continuous deformation when subject to shear
stress. U&e solids, liquids and gases cannot resist small shear stresses
without relative motion occurring between the layers subject to shear.
The behavior of a flowing fluid depends on whether or not it is influenced
by a solid boundary. If it is far removed fiom a solid boundary, it is not subject to shear.The parallel planes move at the same speed. Eddles cannot
form (irrotational flow) and there is no compression and no dlssipation of
mechanical energy to incremental internal energy.
In large-scale, bulk flow, shear forces are said to be confined to the boundary
layer - the layer near the solid wall. Once turbulent flow is established in
pipes, the boundary layer fills the entire channel.Therefore, irreversibhties
are always present across the entire section, but they are greater near the wall
due to the flattened profile.
22
and Equations
Flow o f lncornpressible Fluids
C H A P T E R
O N E
flow (frequently used as a model for real flow) is flow

that can be represented by parallel vectors. Laminar flow is one-dimensional flow in which there is no gross lateral mixing between fluid layers.
In turbulent flow, there is lateral mixing, but, in a conduit, the flow can
still follow the one-dimensional model. It can still be represented by parallel' vectors in the forward direction. The eddies in the normal direction
all cancel one another.
Categorizing fluids based on viscous behavior

We will establish a general classification of fluids based on their viscous
behavior to permit a more detailed discussion. Fluids are generally categorized according to whether or not they follow the newtonian model. If
the coefficient of viscosity is independent of the shear forces and of rate
of change of shear, the fluid is newtonian. If the coefficient of viscosity
changes with the applied shear force or with the rate of change of shear,
the fluid is non-newtonian.The newtonian model makes for fairly simple
computations. These ideas will be made more concrete very shortly.
V)
Shear rate
Iduldy l
Figure 1-10. Shear diagrams
Non-newtonian fluids are further characterized as to whether their

behavior is independent of time or is a function of time. A third characterization is whether the non-newtonian fluid displays some of the characteristics of a solid such as the abhty to recover from applied stress. Non-newtonian fluids that have time-independent behavior are divided into three
subcategories dependmg on how they deviate from the newtonian model.
Bingham plastics require a minimum stress before they begin to move.
They then follow a fairly linear relationship between applied shear and rate
of change of shear.
Pseudoplastics do not require a minimum stress in order to initiate movement.Therefore, their curve goes through zero similarly to a newtonian
and Equations
23
C H A P T E R
O N E
fluid. However, their curve is not linear - it is concave downward. Less

shear stress is required to produce a rate of change of shear as the stress is
increased.
Dilatant fluids also have a curve that goes through zero. However, they
have a curve that is concave upward, opposite to that of a pseudoplastic.
It takes more shear stress to produce a rate of change of shear as the stress
is increased.
Time dependent fluids are generally divided into two subcategories: those
that "thin" when stress is applied over time, and those that "set up" when
stress is applied over time.The first subcategory is called thixotropic and
the second one is called rheopectic.
Viscoelastic fluids exhibit many of the properties of a solid.They wdl be
discussed shortly.
Table 1-1 outlines the above breakdown.
Table 1-1. Fluid categorizationby viscous behavior
1. Newtonian fluids
2. Non-newtonian fluids
A.Time independent behavior

a. Bingham plastics
b. Pseudoplastics
c. Dilatant fluids
B.Time dependent behavior
a.Thixotropic fluids
b. Rheopectic fluids
C.Viscoelastic fluids
1. NEWTONIAN FLUIDS
Much progress in engineering has been made by establishing a model against
which non-standardbehavior can be compared and quantified.The newtonian
model is a case in point.
Force
Velocity
Figure 1-11. Definition of viscosity
24
and Equations
Fluids
C H A P T E R
O N E
By way of example, consider two layers of fluid next to one another (Figure 1-1 1).
If no shear stress is applied, there is no moti0n.A shear stress can be applied by
a force acting on one of two parallel plates within the fluid.The fluid is stationary relativeto the plates a t boundaries created by the two plates, even
when one of the plates is in motion. If a force is applied to one of the two
plates and the other is fixed,a velocity gradient will be created between the
plates.The layer in contact with the moving plate moves at the same velocity
as the plate.The layer in contact with the stationary plate is stationary.
Intermediate layers have velocities depending on their relative position
between the two plates.The following relationship holds:
F=
PULl VAP6,,<
(1-6)
qz
< <
The relationship states that the force necessary to cause a velocity of a plate in
the fluid is directly proportional to the velocity, to the area of the plate to which
the force is applied and to a proportionality constant. It is inversely proportional to the distance between the plates, Z.The dimensional constant, gc,is
thrown in for good measure by chemical engineers using US. units.Various
units will be discussed under the heading, Units of Viscosity, in this chapter.The
units of this viscosity, pm,are Ibm/ft-s.
Note that equation 1-6 relates a force to a velocity, not an acceleration. If the
acceleration is zero, an equal and opposite force must arise in the fluid in order
to prevent acceleration.
If the velocity varies linearly across the gap,V/Z may be replaced by dV/dZ, the
velocity gradient. Dividing by the area gives the more usual form of the viscosity equation.
(1-7)
- P,,,d1'
F
'q,La-
=z
'C, fE
The term on the left is the force per unit area applied to move the fluid layers. It
is called the shear stress.The derivative term is called the shear rate. Mu is again
the proportionalitycoefficient and gc is the dimensional constant.The coefficient, mu, is the viscosity. It is a function of temperature, pressure and fluid composition. In a newtonianfluid, viscosity is neither a function of shear stress, nor
of shear rate.The Greek letter, tau, is just a shorthand form for the shear stress.
In one practical viscometer, the parallel plates are the curved surfaces of one
cylinder rotating inside another.The fluid whose viscosity is to be measured fills
the volume between the two cylinders.
and Equations
25
C H A P T E R
O N E
Experimentally,for newtonianfluids, it is found that the tangential force necessary to maintain a constant velocity is directly proportional to the area of the
plate to which the force is applied and to the relative tangential velocity of the
plates. It is inversely proportional to the distance separating the plates.
Tau is the force per unit area necessary to maintain a velocity gradient in a
given fluid. It is called the shear stress. It also can be thought of as the balancing stress arising in the fluid to prevent acceleration.The derivative makes
the equation more general since it covers situations when the velocity gradient
is not constant.
Some insight into the termlshear ratecan be obtained by realizing that,
nv
dZ
nx/nt
nz
nx/nz
(1-8)
nt
The change in the direction of motion, dx, with the change normal to the direction of motion,dZ, is called the shear.The time rate of change of this shear is
equal to the change in velocity with distance normal to the flow direction.
A newtonian fluid is often defined as one in which the shear rate is linearly proportional to the shear stress - pressure and temperature being fixed. In equation 1-7, the constant of proportionality between the two terms is the coefficient of viscosity divided by the dimensional constant.The use of the subscript,
m, on the viscosity will be explained under the heading, Units of Viscosity.
Gases, true solutions and non-colloidalliquids follow the newtonian model.
Many ordinary industrial fluids follow this model.This is fortunate, since it is the
simplest one.
2. NON-NEWTONIAN FLUIDS
If the coefficient of viscosity is a function of the shear stress (equivalentlyof the
shear rate) as well as of temperature, pressure and composition,the fluid is classified as being non-newtonian.Some examples of classifications of time-independent, non-newtonianfluids are seen in Figure 1-12.
To complicate the problem of viscosity, some fluids display characteristics that

vary with the length of time that they are subject to stress.This means that
non-newtonianfluids can be categorized as to whether they show time-independent or time-dependent behavior.The following categorization is essentially that given in Per$, with some commentary.
26
and Equations
Flow o f
incompressible F l u i d s
C H A P T E R
O N E
A.TIME-INDEPENDENT BEHAVIOR
a. Bingham Plastics
These fluids have a linear relationshipbetween shear rate and shear stress.
However, the line does not go through the origin. A minimum stress,T,, is
requiredto initiate flow.The curve is then reasonably linear. Equation 1-9
describes Bingham plastic shear stress versus rate of change of shear.
Note that the symbol, 7,eta, replaces the symbol, p,mu, to indicate nonnewtonian behavior.
Water suspensions of rock (slurries),grains and sewage sludge are examples of

Bingham plastics.
b. Pseudoplastics
These fluids are polymeric solutions or melts or are suspensions of paper pulp
or pigments.Two gentlemen, Ostwald and de Waelexi,established an empirical
relationshipthat describes the behavior of pseudoplastics over fixed ranges of
applied shear stress.As can be seen from the equation set I-10,this relationship
is a power law one with a coefficient, K,and an index, n.The equation set
describes how the basic power law can be manipulated so that the equation
resembles the newtonian model.The group of terms representing a viscosity
coefficient is then replaced by a single term that is an equivalent viscosity coefficient.This viscosity coefficient is not independent of applied stress, however,
and it is given a different symbol.
The shear stress is related to the shear rate as follows:
nv
(1-10 )
f?<l
and Equations
27
C H A P T E R
O N E
Flow o f lncompressible Fluids
It can be seen that equation 1-1 0, for power law fluids, is identical to equation
1-7, for newtonian fluids, with eta replacing mu.This is a mathematical convenience under fixed circumstances.The viscosity is no longer constant a t fixed
pressure and temperature. It must be computed for the flowing conditions.This
apparent viscosity decreases with shear rate (or with applied shear), hence the
direction of curvature of the shear stress - shear rate curve (1-1 2).
The coefficient, K, is called an index of consistency. Hence, in the pulp and
paper industry, one hears the term consistencymore frequently than viscositywhen talking of fiber suspensions.The index, n, is a constant that is a
characteristic of the fluid. It is less than one for pseudoplastic fluids. Both the
coefficient, K, and the index, n, must be established by experiment.
c. Dilatant Fluids
In dilatant fluids, the apparent viscosity coefficient increases with the shear
rate. Again, the constant is no longer constant. Starch, mica suspensions, quick
sand and beach sand suspensions are examples of dilatant fluids.
The same Ostwald-de WaeleXiequation, 1-10, describes dilatant fluids when the
index, n, takes on values greater than one.
B.TIME DEPENDENTBEHAVIOR
a.Thixotropic Fluids
The graphic of Figure 1-12 may be regarded as depicting the amount of shear
stress needed to maintain a shear rate, dU/dZ (dy is used in the figure). It can be
seen that if a fluid is taken fairly rapidly from point D to point A and slowly
back again, the paths will be different.
Over the return path, much less stress is needed to maintain a shear rate.
If the fluid is taken fairly rapidly from point D to point A and held a t that shear
rate, the stress must be reduced (otherwise the rate would increase).In other
words, in order to maintain the same rate of rotation of a spinning viscometer,
the power to the motor must be reduced.The amount of reduction in shear
stress required depends on the length of time the fluid is held a t the shear rate
associated with points A, B and C.
28
and Equations
C H A P T E R
O N E
!@
B
c
v)
Shear rate ldu/dyl
:&
sz
Figure 1-12.Thixotropic behavior
r: z
The return paths can be from A, B, or C.They are substantially linear. Point C is a
minimum for a given shear rate. It will always take some shear stress to produce
a shear rate.
-y L
2.. %
c
:+
23
-
v-
Mayonnaise, drilling muds, paints and inks show thixotropic behavior.
b. Rheopectic Fluids
These fluids set up.They increase their apparent viscosity on being shaken.
Bentonite SOIS,vanadium pentoxide sols, and gypsum suspensions in water are
examples of fluids that show rheopectic behavior.
C.VISCOELASTIC FLUIDS
Viscoelastic fluids exhibit many of the properties of a solid.One of these properties is an elastic recovery from deformation.Some polymeric liquids fall into
this category.
In polymer processing, mixing,extrusion, calendering,fiber spinning and sheet
forming are examples of non-newtonianflow (See Bernhardtxii,Processing of
Polymeric Materials, for details).
This book will concentrateon fluids that can be treated as being newtonian to
a close approximation.Themajority of industrial problems can be solved with
this approximation.Perf is a good source of reference material should the
reader wish more information on viscosity.
Caveat - Fluid classifications
The reader must be aware that other classifications than newtonian exist and
that they must be dealt with from time to time. In fact, in some industriesthey
must be dealt with all the time.
and Equations
C H A P T E R
O N E
F l o w o f l n c ~ r n p r e ~ ~ i bF l ue i d s
Viscosity as a fluid property

It should be evident that viscosity is a fluid property that is a function of
temperature, pressure (to a minor extent for liquids), composition and
sometimes time and shear. Published viscosities are usually correlations
based on the first three variables. Empirical relationships are required for
those fluids that are non-newtonian.
Published values for viscosities are essentially coefficients measured under
larninar flow conditions; otherwise the effect of the actual flowing conditions would have to be known. Flow conditions are taken into consideration when the Reynoldsv number is used to calculate the friction factor
and the latter is inserted into the
equation in order to compute
the mechanical energy losses.
arc^^"
Viscosity's dependency on temperature and pressure

Viscosity depends primarily on temperature. It depends weakly on pressure.
For gases, viscosity increases with temperature. For liquids, viscosity decreases with temperature.The viscosity of liquid water at 32F is about six
times its value at 212"E Note that viscosity is not linear with temperature.
Other names for viscosity
Saybolt universal
...
seconds units are just another kmematic viscosity.
See Cranex" for nomographs.
Generalized viscosity
When data are not available except for the viscosity of a liquid at one
temperature, its viscosity at another temperature may be estimated from
Figure I-13.This figure is taken from Per$. Perry is a good source of
information on the physical properties of liquids and gases.
Figure 1-13. Generalized viscosity chart
30
and Equations
Fluids
C H A P T E R
O N E
If the viscosity in centipoise is known at one temperature for a liquid, this

viscosity may be used to locate a point on the ordinate.The corresponding point on the abscissa is found.The temperature difference is then
stepped off in a direction corresponding to the sign.The new viscosity
corresponding to the new temperature is then read on the ordinate.The
method is said to be accurate within plus or minus 20%.This means that,
if information is available on a specific fluid, it should be used.
Many viscosity units
It is unfortunate there are so many different, and confusing, measures of
viscosity (for instance, Saybolt Universal Seconds). It seems that every
worker who has done research in the field of viscosity has developed his
own units.
'Absolute' viscosity
In the English spealung engineering world, two principal measures of viscosity have been used in textbooks.The first measure has already been
defined by Equation 1-7 containing g,.This equation will be repeated and
the units will be derived.
The second measure of viscosity comes from simply omitting the dimensional constant in the defining equation.
The subscripts on the viscosity terms are used simply to refer to mass and
force in the defining equations.The viscosities differ only by a multiplying factor, the dimensional constant as shown in Equation 1-13.
&l
= &?,pf
(I-13)
Since the dimensional constant is equal to 32.17, it is wise to carefully

check which viscosity is being cited before using the number associated
with it.
and Equations
C H A P T E R
O N E
Neither of the above viscosity coefficients has a name other than the general term, viscosity. Each is identified by its units. Both the above viscosities are referred to as being "absolute" viscosities. This just goes to prove
that there is nothing absolute about the use of the word, "absolute".
In the S1 system, the units of so-called "absolute" or "dynamic" viscosity
are pascal-seconds, Paes.
(I- 14)
Pascals are the equivalent of newtons per meter squared.Therefore, pascalseconds are the S1 equivalent of customary U.S. pounds-force units, pp
By far, the most popular unit of viscosity remains the poise or, more
exactly, one derived from it, the centipoise. In other words, none of the
above. The poise is based upon the older cgs (centimeter-gram-second)
system.The unit of force in this system was the dyne. The dyne was a
derived unit that in turn had units of grn-cm/s2.w h e n viscosity is given
in poises, its units are,
(I- 15)
FZ gm-cm cmas - gm
----p
p
s2 cm2ecm cm-s
It is likely that the centipoise will continue to be used, as it is a convenient decimal multiple of the official S1 viscosity unit, the pascal-second.
One Paas is equal to ten poise or 1,000 cP or 0.672 lb, /(ft.s). So, one cP
~ /(fi.s).
equals 0.000672 lb, /(ft.s) or 6.72 x I O -lb,
This book will mainly use lb, /(ft.s) mass based units and centipoise
(which are anchored firmly in usage).We will name the units being used
in each case.
'Kinematic' viscosity
The so-called kinematic viscosity is simply the "absolute" viscosity
divided by the density of the fluid at the temperature and pressure at
which the absolute viscosity was measured. It is normally given in stokes
or centistokes with units of centimeter squared per second, V = p / p .
32
and Equations
Fluids
C H A P T E R O N E
The units are,

(I- 16)
A more detailed discussion of viscosity and its units can be found in

Appendix AI.
1-8: ' F R I C T I O N LOSSES'

A ''friction loss" is mechanical energy whch was capable of performing
work (such as in moving a fluid in a pipe), but whch has been transformed
to internal energy by turbulence.This energy is usually seen as heat flow to
the surroundings, but it can remain in the fluid as an increase in internal
energy. It is considered a loss because it cannot normally be recovered as
mechanical energy and external work must be supplied to make up for the
loss in mechanical energy, otherwise the fluid would eventually stop flowing.
As has been pointed out, a friction loss corresponds to the concept of
irreversibility in thermodyna&cs. The author's preference is to avoid the
term friction loss wherever possible because it causes him to stop and
wonder what the connection to friction is. Complete avoidance is not
possible since the term is so commonly used in the literature.The term
"friction factor" is one of the most common terms in fluid flow.
Irreversibilities,the conversion of mechanical energy to internal energy, are
important in the calculation of flow through pipes.They are a function of
the flow regime that is defined by the pipe Reynolds number.The pipe
Reynolds number, in turn, is a function of velocity, density, viscosity and
internal pipe diameter. Irreversibilities are also included in the coefficient
used in orifice plate computations.Temperature influences most of the
properties considered in the ReynoldsVnumber, directly or indirectly.
Friction factor
The ~ o o d ~ friction
~ " ' factor, fM,is the coefficient of proportionality
between the irreversibility and the product of the number of pipe diameters of straight pipe and the kinetic energy (Equation I-17).The friction
factor is a function only of Reynolds number for laminar flow. It is a
function of Reynolds number and pipe roughness for turbulent flow.
In addition to straight pipe, a piping system consists of fittings (elbows,
tees, reducers, expanders), equipment (strainers, etc.) and valves. Fittings in
and Equations
33
C H A P T E R
O N E
a piping system are sometimes treated as so many equivalent lengths of

straight pipe. craneGiigives methods of calculating equivalent lengths of
various types of fittings.
Caveat
- Equivalent lengths
The equivalent length method should only be used for rough approximations.
Why must irreversibilities be computed?

Some of the reasons follow:
to establish upstream and downstream values of pressure at meters and
valves;
to calculate flow rates though piping systems;
to establish the range of the flow regime;
to size pumps, compressors, turbines, valves, and piping systems;
to size relief headers and to ensure that the backpressure on relief
devices will not prevent the devices from functioning adequately.
We will enter into more specific details in the next chapter.
The
b arc^"" equation and the friction factor
If flow were ideal, there would be no permanent mechanical energy losses

due to irreversibilities. AU of the energy terms would be completely interconvertible.The pressure would change with changes in fluid velocity and
with elevation in the system, but the total mechanical energy, static,
potential and kinetic, would remain constant. Flow, once established,
would remain constant without the need for external work. In the real
world, irreversibilities take their toll by converting otherwise useful
mechanical energy into incremental internal energy.This incremental
internal energy cannot be converted back completely to mechanical
energy. It does result, however, in an increased temperature of the fluid. If
the system is not adiabatic (not perfectly insulated), heat will flow to the
surroundings due to the temperature difference. External work is required
to maintain flow because of this conversion of mechanical energy to
internal energy and its transfer as heat energy flow to the surroundings.
The potentials involved in fluid flow are mechanical ones. An irreversible
conversion of a mechanical potential to an internal one means that less
useful work can be performed on the fluid or by the fluid.This irreversible conversion is often described as "lost head" or "lost work". The
choice of language is unfortunate and leads to confusion, but the terms
are commonly used.
34
and Equations
C H A P T E R O N E
For liquids flowing in straight, constant diameter pipe, the irreversibilities

can be predicted by the
equation,
(I- 17)
arc^^"
~ o o d ~ " friction
"'
factor
The term, fM,is the Moody friction factor. It is just a proportionality
factor between the irreversibilities and all the other terms. It is not a constant.The subscript, M, is to remind everyone that this is the Moody friction factor. The Fanningx" coefficient of friction, fF,is smaller by a factor
of four and requires that the equation be changed, that the divisor, 2, be
moved to the numerator.
The L/D ratio is simply the number of pipe diameters over which the
losses are being estimated.The term, u 2 , is the average (across the section)
fluid velocity squared.This term divided by 2g, is the kinetic energy of
the fluid at the section where the average velocity is U.The term, g,, is
the dimensional constant, 32.17 lb,
ft
lbf -l. It is not the acceleration due to gravity, 32.17 ft
the above equation uses conventional
American units, feet, seconds, pounds-force and pounds-mass. However,
since the friction factor is non-dimensional, if g, is given the value of
one, without dimensions, the same equation holds for S1 units.
The loss term, hf, has units of foot-pounds-force per pound-mass, newtonmeters per lulogram, or joules per lulogram depending on the choice of
units. It is usually defined as the losses due to "slun friction" (primarily in a
boundary layer next to the pipe wall). Remember what has been said
about irreversibilities and the fact that, in turbulent flow in pipes, the
boundary layer occupies the whole pipe section.We will show in the next
section that the loss term is a measured quantity, from experiment.
Since the velocity is also a measured quantity, the friction factor
can be correlated from experimental data. The correlations, in the form
of Moody friction charts or the churchill-usagiAi equation
(Appendix A-11), allow calculation of the friction factor and of
mechanical energy losses due to fluid irreversibilities.
Caveats
- Velocity
profiles
The reader is reminded that if the wrong assumptions are made regarding
the velocity profile or if a point velocity is used instead of the average
and Equations
C H A P T E R
O N E
Fluids
velocity errors will result in the ensuing computations.
arc^^^'
The
equation must be modified somewhat for compressible flow.
This will be done in the chapter on compressible flow.
The ~ o o d ~ friction
~ ' " factor in laminar flow is a function only of the
Reynoldsv number. It is,
(I- 18)
Fanning friction factor

The Fanning friction factor in laminar flow is,
(I- 19)
These equations result from substituting the Hagen-Poiseuillexv equation

into the ~ a r c ~ - ~ e i s b a c hequation.
'"
These equations will be developed
in Appendix AI.
This book will use the Moody friction factor exclusively. It will use the
subscript, M, to identitjr the Moody coefficient. In the case of turbulent
flow, the friction factor depends on the Reynolds number again, but not
so simply as shown in Equation 1-18 for laminar flow. It also depends on
the pipe surface roughness.The friction factor... for turbulent flow is correlated by way of friction factor charts (Cranex1' gives some excellent
charts) or by the churchill-usagiA equations. The Churchill-Usagi equations are an excellent tool when automating computations. The
Churchill-Usagi equations will be given in Appendix AII.
It is fortunate that industrial pipe manufacturing practice allows the prediction of the surface roughness and, thus, the friction factor with a high
degree of accuracy (Figure 1-4).
1-9: BERNOULLI EQUATION AND THE DARCY EQUATION C O M B I N E D

General approach to calculating flow through pipes
Most practical flow formulae are derived from the ~ernoulli'equation.
The usual integrated, steady-state form for incompressible fluids is:
36
and Equations
C H A P T E R
O N E
The ~ernoulli'equation is derived from a force and momentum balance.

The complete derivation is given in Appendix AI.
The total energy at point 1 plus the energy added (pump) or subtracted
(turbine) between points 1 and 2 equals the total energy at point 2 plus
energy converted (ultimately to heat flow) between points 1 and 2.Total
energy is broken up into quantifiable types of energy.The first term on
each side is the "static" energy or mechanical energy, due to pressure. The
second term is the kinetic energy, due to average velocity in the section.
The third term is the potential energy, due to position of the flowing
fluid above a datum (any fixed datum).The last term on the left is the
work energy per unit mass added, between points 1 and 2, by a pump, for
example, or extracted (negative sign) by a turbine.The last term on the
right is the "head loss" per unit of flowing mass between the same two
points. It is the mechanical energy that has been converted to internal
energy and that is no longer available as a mechanical driving force. It will
be ultimately lost to the environment as heat energy flow. Irreversibility is
a more descriptive term than head loss or friction loss.
In the Bernoulli equation, 1-20, consistent units must be used. If you wish
to use psia for units of pressure, simply multiply the pressure terms by 144.
If you are worlung in S1 units, treat the term, gc, as equal to one, dimensionless. All the terms in the equation have units of energy per unit mass.
If S1 units are used, gc equals 1, dimensionless, but g equals 9.81
meterdsecond. If you have to switch between units, it is best to use the
above formula. It helps keep track of where the dimensional constants
should be included. The term, hf is added to the ideal Bernoulli equation
(which did not include it) to accommodate irreversibilities between
points 1 and 2. A similar looking equation, used in hydraulics practice,
will be discussed shortly.
A limitation of this form of the Bernoulli equation is that it is derived for
incompressible flow and it must be modified in order to be used for compressible flow. Another limitation is that average velocities across a section
are used in the second term on each side.A kinetic energy correction
factor is needed for more exact computations.This factor will be
described in Chapter 11.Another limitation is that the equation is derived
and Equations
37
C H A P T E R
O N E
for steady state flow.This latter limitation is usually not too serious, but it
must be borne in mind.
Experimental determination of the loss term, hf
If the ~ernoulh'equation, 1-20, is applied to steady-state flow in a long,
horizontal, constant hameter pipe in which a liquid is flowing, the velocity,
elevation and density are constant. We can choose a section of pipe that
contains no pump or turbine.The Bernoulh equation then reduces to the
statement that the difference between upstream pressure and downstream
pressure divided by the density of the fluid is equal to the irreversibilities
(losses) in foot-pounds-force per pound-mass or in joules per lulogram.
The simple relationship just described allows irreversibilities in straight
pipe to be computed from direct measurement. Once irreversibilities are
computed for different flow rates in differently sized pipes flowing different fluids, these losses can be correlated with the corresponding densities, viscosities and internal pipe characteristics (roughness, diameter). The
equation and the ~ o o d ~friction
~ " ' factor are the result of these
correlations. It is fortunate for us that so much effort has been spent
establishing what, on the surface, seem to be such simple relationships.
arc^^"
The Darcy equation

The Darcy equation is,
The equation allows estimating irreversibilities in a pipe of given linear

length and diameter for a given flow (average velocity) and a given friction factor. The last term allows irreversibilities to be estimated across fittings with a given K factor. The K factor must be associated with the pipe
diameter in which the velocity was estimated.
When the irreversibilities are estimated between two points, the values
can be plugged into the Bernoulli equation.The Bernoulli equation can
then be solved for the remaining quantities. The Darcy equation applies
to laminar and turbulent flow of incompressible fluids. Its modification for
compressible fluid computations wlll be discussed later. Both the friction
factor and the K factor are established by correlation after the irreversibilities have been computed for a known situation.They can then be used
in hypothetical situations.
and Equations
C H A P T E R
O N E
The ~ o o d ~friction
~ ' " factor, fM,is a function of the Reynoldsv number
(and through it of viscosity) and the pipe roughness.The friction factor has
been established by correlation and experiment and plotted on log/log
paper for various sized pipes and various relative roughnesses. Figure 1-4 is
an example.The friction factor may also be estimated with greater accuracy using the Churchill-Usagi XLiequations
given in Appendix AII.
Loss coefficients
For constant diameter pipe, the
equation shows the lost mechanical energy is proportional to the Moody friction factor, the pipe length
to diameter ratio, and the square of the average velocity across the pipe
section.Various fittings, valves and restrictions may be characterized by
mahng use of equivalent pipe lengths or by using an equivalent K (or
fML/D) in the Darcy equation.
arc^^^^
The loss coefficient, K, is often called the number of velocity heads lost
through a fitting. This terminology is pure jargon. The factor, K, is best
thought of as the fraction of kinetic energy that is converted to internal
energy and thermal energy by flow by the fitting.The term, u2/2gc, is
the velocity energy or the kinetic energy per unit mass. Remember the
energy is only "lost" to mechanical use. It is still present as internal energy
or it passes to the environment as thermal energy flow.
The factor, K, is associated with the diameter in which the velocity is
computed. For fittings, it is treated as being independent of Reynolds
number (therefore, of velocity, viscosity and density) and of the friction
factor. For such short lengths, it can be correlated by geometrical similarity among fitting types. We will discuss this fact in greater detail in the
next chapter. For straight pipe, K is not independent of Reynolds number
or the friction factor,
T
Experimentally, for fittings, K varies as does fM L/D for clean, commercial

pipe. cranexiiimakes use of this fact to establish some simple correlations
between the coefficient K and the types of fittings.We will discuss this in
Chapter 1I.Table 1-3 gives values of K for liquid flow in various types of
fittings and valves. The figure is taken from Per#. The lower case reference letters refer to the original source of the coefficients. Perry should
be consulted if you wish further information.
and Equations
C H A P T E R
O N E
Flow o f Incompressible
Fluids
1-10: CONSERVATION OF ENERGY I N HYDRAULICS PRACTICE

Energy is the same entity regardless of the discipline analyzing it. It is
sometimes called by different names and different groupings of variables
are used to describe it.We will summarize the treatment of energy and
include a description of how hydraulics practice treats energy.
It has already been stated that energy cannot be destroyed, but it can be
converted to various forms, one of which is not useful mechanically. This
less useful form is often described as being "lost" energy.The mechanically useful forms are to be found in Table 1-2.
TABLE 1-2. Mechanically useful forms of specific energy
Energy
J/kg (51)
ft-lbdlb, (US.)
static energy
potential energy
P/p
P/p
kinetic energy
gz
u2/2
g m c
u2/2gC
The total mechanical energy per unit mass, e, is the sum of the above
three terms.The computed value of e at a section will be the same as the
computed value at any other section in the ideal case.The ideal case gives
a starting point for mahng corrections to suit the real world.
The S1 units of specific energy are newton-meters per kilogram, N-m/kg,
or the equivalent joules per lulogram,J/kg.The U.S. units are footpounds-force per pound-mass, ft-lbf /lb,.
In hydraulics engineering it is said to be more convenient to express energy
in terms of unit weight instead of unit mass. In the author's opinion, this is
an approach that leads to confusion. Starting with the total mechanical
energy balance for the ideal case (no losses), energy per unit mass is,
40
and Equations
u2
P
e=gZ+-+-
g
P
e=-Z+-+g,
P
u2
28c
C H A P T E R
O N E
(SI)
(US)
P u2
H = Z + - - (hydraulic)
Y 22
The weight density, y, has different values and lfferent formulae in

metric and in US. practice.
and Equations
41
C H A P T E R
O N E
Additional Frictional Loss for Turbulent Flow through

f ittingo and Valves"
Type of fitting or valve
4S0 ell, standardb*"dsef

4S0 ell, long radiusC
90 ell, ~ t a n d a r d ~ * ' * ~ * ~ " ~
Long r a d i ~ s ~ * ~ * ~ * ~
Square or miterh
180 bend, close returnbJ''
Tee, standard, along run, branch blanked off
Used as ell, entering run&'
Used as ell, entering branchc*&'
Branching flowkfik
Coupline
Unione
Gate ~ a l v e ,open
~~~-~
% openn
% openn
# open"
Diaphragm valve,o open
% openn
H openn
% openn
Globe valve,8tmbevel seat, open
34 openn
Composition seat, open
34 openn
Plug disk, open
% open"
U openn
% openn
Angle valve,bteopen
Y or blowoff va~ve,b*~
open
Plug cockp (Fig. 5-42) 8 = 5O
loo
20
40
60
Butterfly valv@ (Fig. 5-45) 8 = So

100
20
40
Check vaive,km swing
60
Disk
Ball
Foot valvee
Water meter? disk

Piston
Rotary (star-shaped disk)
Turbine-wheel
Table 1-3. lrreversibilitiesin turbulent flow
42
and Equations
Additional
friction loss,
equivalent no. of
velocity heads, K
0.35
0.2
0.75
0.45
1.3
1.5
0.4
1.0
C H A P T E R
O N E
The units of the weight density as defined by gamma are force per unit
volume. The S1 units are newtons per meter cubed. The customary U.S.
units are pounds-force per foot cubed.
The use of weight density allows the hydraulic equation developed above
to have all of its terms expressed in meters or feet of fluid.This is another
source of the term "head". Physically, the hydraulic equation (last equation in the set 1-23) may be interpreted as shown in Figure 1-14.
Figure 1-14. Interpretation of hydraulic'head'
'Hydraulic head' or 'piezometric head'

In hydraulics practice, it is usual to give the name "hydraulic head" or
"piezometric head" (h) to the sum of the elevation above the datum (Z)
and the static pressure energy term (P/y).
P
h=Z+-
So, the total mechanical energy in terms of feet or meters of fluid

becomes,
The term derived from the kinetic energy term in the set of equations
1-23 is so similar to it (g replacing g 3 as to lead to confusion.This is
particularly true in metric practice when g equals 9.81 and gc equals one.
Low-pressure gases can often be treated as incompressible fluids. In
hydraulics practice, the pressures and differential pressures are often measured with manometers. The manometers are referenced to atmospheric
pressure as shown in Figure 1-15. It can be seen from the above equation
that the velocity head is equal to the difference between the total
and Equations
C H A P T E R
O N E
mechanical energy head and the hydraulic (piezometric) head. Notice

that, if the manometers are all topped up to the same levels initially, each
one will have a uniform displacement about the same center 1ine.The
total mechanical energy will equal the sum of the sum of the energy due
to velocity ("head") and the piezometric ("head").
Figure 1-15. Manometers in hydraulic practice
When the manometers that measure these two variables are referred to
atmospheric pressure, it makes no difference to the equation whether the
variables are absolute or relative.The pressure equivalent to the atmospheric pressure would have to be added to both terms. Hence, psig units
are often used in place of psia units. This is a practice that can lead to
error when dealing with compressible fluid computations.
Because the general engineering approach is less restrictive than the
hydraulic engineering one, we will use the general approach almost
exclusively.The general approach also gives a better understanding of
physical problems.
1-11: W O R K E D E X A M P L E S
We will give some very basic computations in this chapter. Examples of
more extensive computations may be found in Chapter 11.
Examples of hydraulic practice
Example 1-1: Pressure corresponding t o 50 mm WC
We can assume that 50 mm W C is the measured head, h, on a manometer
with no flow.The datum, Z, is zero. Since pressure equals the head times
the weight density or the head times the mass density times the acceleration of gravity, we can write the set 1-27.
44
and Equations
Flow of lncornpressible Fluids
C H A P T E R
O N E
The conversion from kg/m-s2 was performed by making use of Newton's

force-acceleration law: N = kg-m/s2. A pressure of 490.5 pascals is close
to two inches of water column. Since this pressure was measured with a
manometer, it is assumed that we are dealing with gauge pressure, not
absolute pressure.
Example 1-2: Pressure corresponding to 100 inches WC

The same assumptions apply as in the first example, only the units and
their values change. An additional subtlety is that we are dealing with
weight density that is numerically equal to mass density only in the U.S.
customary system.
We can write psig because of the assumptions that were accepted.
Conversion difficulties
Difficulties in conversion lie not so much in finding the right conversion
factors as in understanding what it really is that one is converting.
Appendix AI11 will discuss in detail the different uses of the word "head"
and the fact that it sometimes means a pressure difference, sometimes a
height difference and sometimes an energy per unit mass difference. In this
chapter, let us assume we have identified what it is we are talking about
and simply want to convert pressure units to head units. More exactly, we
are interested in converting differential pressure into differential manometric head (equivalent height of a column of liquid) or vice versa.
The easiest way to effect the conversion is to change all units to standard
S1 or customary U.S. units.The next step is to visualize a column of liquid
and Equations
45
C H A P T E R
O N E
Flow o f lncompressible Fluids
in tube and to ask what the height of the column would be which would
exactly balance the pressure at the bottom. It is very important to
remember the difference between pressure and force and to use the appropriate S1 or U.S. customary form of Newton's law to solve the problem.
Example 1-3: Height o f WC under maximum vacuum, S1
Suppose we want to know the absolute maximum height that we can
raise a column of water at 60F or 15.6"C under the best vacuum we can
achieve. The first point to consider is that water has a vapor pressure, so
we cannot achieve a perfect vacuum. At the temperature in question, the
vapor pressure is 0.258 psia or 1.779 kPa. We will assume the atmospheric
pressure forcing the liquid up the column is 14.700 psia or 101.350 kPa.
The vapor pressure is not insignif cant; it is 1.76% atmospheric pressure.
The differential pressure holding up the column of water will be 14.442
psi or 99.571 kPa.These are the numbers with which we must work.
The next step is to make use of the appropriate form of Newton's law but not blindly. Logically, we know that the mass of fluid represented by
the column of water would accelerate if released from an elevation equal
to that of the bottom of the column. We also know from elementary
physics that this acceleration is uniform over normal distances and is
equal for all bodies when not subject to any other forces than the force of
gravity.We now reason that in order to stop the mass from accelerating an
equal and opposite force must be applied at the bottom of the column
and that this force is given by Newton's law.
We know the pressure and we know that pressure is force per unit area.
The equation set 1-29 gives us the solution for the S1 case.
F = mg
(1-29)
Hydraulic practice makes the direct connection between pressure and

manometric head by saying that a pressure (differential) of 99.574 kPa is
equal to a head of 10.186 m.
46
and Equations
Flow o f lncornpressible Fluids
C H A P T E R
O N E
U.S. customary units

Example 1-4: Height in customary U.S. units
If we wish to solve the same problem in customary U.S. units, we transform pounds per square inch to pounds per square foot and include the
dimensional constant in Newton's law. The equation set 1-29 gives the
solution.
It is common in hydraulics practice to say that 14.442 pounds per square

inch is equal to 33.354 feet of head. What is more, as has been discussed
previously, since at fixed places on the globe, the acceleration due to
gravity and the dimensional constant are equal numerically, it is common
to lump them into the density and change the name of mass density to
weight density.
In U.S. hydraulic units, the second from last line in 1-29 would read:
(1-3 1)
P
-=z
Weight density is equal to mass density times the acceleration of gravity

divided by the dimensional constant. In U.S. practice, the numbers are the
same, but the units change.
In European metric hydraulic practice mass density and weight density is
not numerically the same.The denominator in the second from the last
line in 1-29, pg, would be replaced by y.
(1-3 2)
Y=Pg
and Equations
C H A P T E R
O N E
The weight density in European metric practice differs from the mass
density by the multiplying factor g (9.805 m/s2). In U.S. hydraulic practice, the weight and mass densities are numerically identical, but not
dimensionally.
Converting energy per unit mass units to manometric head units
When using the U.S. systems, the conversion from customary U.S. units
to U.S. hydraulic units is easy. Only the names of the dimensions change,
the numbers remain the same.Ten ft-lbf /lb, become ten feet of fluid.
When converting S1 units to metric hydraulic units, joules per lulogram
or newton-meters per lulogram have to be divided by the acceleration of
gravity, g, which equals 9.805 m/s2.
Frequently, even those of us who dislike using hydraulic units are forced
to use them - usually because a particular pump data sheet requires their
use. We will therefore give two examples of the conversion process.
Suppose that we have computed the lunetic energy, the velocity head, in a
pipe and need the equivalent feet or meters of head. How is it done?
Example 1-5: customary U.S. units
For the customary U.S. unit example we will assume an average velocity
across the pipe of 10 feet per second.The lunetic energy (still called
"head" by chemical engineers) is given by,
(1-33)
h v = -U'
- --- lo2 -1.554j-lb,/lb,,,
2g, 2(32.17)
Example 1-6: U.S. hydraulic units
In U.S. hydraulics practice this would be written,
Note the subtle change of the dimensional constant to the acceleration of

gravity and the change in units, but not in their numerical value.
Example 1-7: S1 units
In S1 units, the equivalent velocity would be 3.048 meters per second.
The lunetic energy in energy per unit mass would be,
48
and Equations
C H A P T E R
O N E
The above conversion can be checked by using the dimensions associated

with Newton's acceleration law, ~ = k ~ - m to
/ sreplace
~
seconds squared.
To convert the S1 units to meters of manometric head, one has to divide
by the acceleration of gravity, g, which equals 9.805 meters per second
squared.The formula would be,
h,=-=-u2
2g
3.048~ = 0.474 meters offluid

2(9.805)
(1-36)
The two hydraulic computations give approximately equivalent quantities,

1.544 feet equals 0.474 meters.We have used, however, four simdar but
subtly different formulae to arrive at results in different units.We can
reduce the problem substantially if we use only S1 or customary U.S. units.
1-12: CHAPTER SUMMARY

We have now defined most of the basic concepts necessary to solve problems involving circular liquid lines flowing full. The principal concepts to
be retained from this chapter (and even memorized) are:
total energy is constant, it can be transformed to dfferent forms: potential,
static, kinetic and internal.The transformation of incremental internal
energy has a unidrectional component that prevents it from being transformed to one of the other, more mechanically usehl forms, in a flowing
system.T h s explains fiiction losses or irreversibhties;
flow of liquids in circular pipes flowing full can be characterized as laminar, transitional or turbulent;
laminar flow is characterized by concentric laminae slidng past one
another, with very little lateral mixing;
turbulent flow is characterized by very small eddes and vortices with
much lateral mixing;
transitional flow is an unstable mixture of the two other flow regimes;
in a long straight pipe, at steady state, each regime develops a characteristic
flow profile or velocity dstribution across the pipe;
the profiles are parabolic for larninar flow and tend to be flattened for
turbulent flow;
and Equations
C H A P T E R
O N E
control over or knowledge of the profle is important to flow measurements in many cases. For instance, a Pitot tube measures the point velocity
in the profile. If the sensor is located at a point where the velocity profile
is unknown or distorted, there are no guarantees as to accuracy;
some measuring instruments are very sensitive to distorted profiles, others
less so;
an index exists, the Reynoldsv number, whch can prehct the flow regime
and whch can be used to calculate the irreversibllities (losses) in pipe;
to calculate irreversibhties in larninar flow, a simple formula exists for
findmg the fiiction factor;
to calculate losses in turbulent flow, graphcal methods, or the ChurchdlusagiA equations must be used to find the fiiction factor;
although the concept of viscosity being the proportionahty factor between
shear stress and shear rate is quite simple, the plethora of units is very conhsing and fi-equently leads to error.Always carefully check whch units are
being used and convert them if necessary before using them in any formula;
there are two commonly used fiiction factors, the ~ o o d ~ "and
" the
F a n n i n e factors;
the Moody coefficient is larger than the Fanning coefficient by a factor
of four.The
equation as used in this book employs the Moody
friction factor. If the Fanning factor is used, the equation has to be
adjusted accordingly;
arc^^'^
the ~ e r n o d h equation
'
is probably the most used equation in the field of
fluid mechanics. It is a force-momentum balance that equates driving
potentials in energy per unit mass to changes in kinetic and potential
energies and to lost work;
lost work is a term for irreversibhties or energy that has been "degraded"
into incremental internal energy;
the Bernoulli equation can also be derived fiom the first and second laws
of thermodynamics - a fact that establishes its generahty;
in hydrauhcs practice, a deceptively simdar equation to the Bernoulh
equation is used.
50
and Equations
Incompressible Fluid
Flow - Losses Of
Mechanical Energy
- APPLYING
BASIC CONCEPTS
This chapter will apply the basic concepts developed in Chapter I and
Appendix AI to the solution of some typical fluid flow problems. Many of
the problems relate to measurement and control applications.The reader
might be tempted to ask why he or she should invest time in learning of
things that are somewhat remote from his or her field of interest.The
answer is a general knowledge of flow phenomena helps avoid errors due
to incorrect assumptions having been made. This general knowledge may
be obtained only through broad study.
We wdl show how to apply the ~ernoulli'equation to specific problems.
These problems require an understanding of the following:
how to estimate irreversibhties and the associated pressure drop in straight
pipe;
how to estimate irreversibhties and pressure drop in various types of
fittings;
what the difference is between pressure drop and irreversibhties;
how the irreversibhties of in-line components are estimated;
what pressure recovery is.
We will take a systematic approach to the solution of flow problems in an
attempt to show it is impossible to work in isolation (or in ignorance of
other people's problems).This statement is intended to mean that one
cannot isolate "measurement" problems from system flow problems such as
those associated with pumps, compressors, pipes and piping components.
and Equations
51
C H A P T E R
T W O
lncompressib e Fluid Flow
Losses of Mechanical Energy
The chapter will concentrate on the passive elements in a piping system.

However, the chapter will prepare the way for the understanding and
analysis of complete systems.
The differences between units of energy per unit mass and "feet" or
"meters" of head will be discussed in detail in Appendix AIL
This chapter is intended to be of practical utility.Tables of data and
appropriate references will be given. When the reader has understood this
chapter, it is hoped he or she will be able to approach typical incompressible fluid loss estimation problems with confidence.
11-2: R E A S O N E D A P P R O A C H T O D E S I G N
A LITTLE PERSONAL PHILOSOPHY
Experience teaches that, in any engineering effort (a project) involving

different disciplines over time, work must be divided among the various
participants. Schedules have to be met. Plants are built only after considerable engineering co-ordination has taken place.
The opportunities for error are enormous, and new plants frequently
require considerable modification before they function to the satisfaction
of the owners. Much time is spent by managers devising quality control
methods and programs. Slogans such as, "Do it right the first time!"
abound. The impossibility of doing something right without a few iterations is not often addressed.
The pressure of schedules usually means process engineers estimate line
sizes conservatively at an early date.The sizes sometimes are not corrected
to match the final plant design. Mechanical engineers take early data from
the process engineers and add their own safety factors for the specification of the prime movers such as pumps. Control systems engineers size
their valves and other instruments based on the same conservative data
and also add safety factors.
One piece of evidence that supports the above statements is the frequent
discovery of oversized (and expensive) control valves during plant startup. A control valve is meant to minimize flow problems; an oversized
valve can create them.
and Equations
Incompressible
Fluid Flow - Losses o f Mechanical Energy
C H A P T E R T W O
A more reasoned approach, from the point of view of a control systems

engineer, would be to delay final specification of control valve sizes until
the piping layout were reasonably fixed. The control systems personnel
would then have the responsibility of performing a final hydraulic study
on a more realistic system. Real data could be used for pipe lengths, numbers and types of fittings and for pump heads. Design changes could be
accommodated more readily. Oversize pumps at least could be identified.
There might even be time to trim pump impellers.
Unfortunately, schedules sometimes get in the way of reason. It is not always
permissible to do things right the first time.They must be fixed afterwards.
Regardless of whether you are trying to do it right the first time or to fix
it afterwards, you must follow the same analytic process. In the first case
you may be acting on assumed (hypothetical) data. In the second case you
will be acting on real (as-built) data. The same engineering knowledge is
required in both cases.
11-3: T H E B E R N O U L L I E Q U A T I O N R E V I S I T E D
The ~ernoulh'equation is so important to all aspects of fluid flow that it is

worth revisiting in greater detail.The complete, integrated form of the
Bernoulh equation contains eight terms, two of whch have correction factors (alpha).It is a steady-state balance of the energy per unit mass at one
point in a piping system against that at another point. Two of the terms, W,
and hL,refer to energy per unit mass added or converted to thermal energy,
between sections 1 and 2. Not all the terms or factors are necessary to solve
all problems. For instance, in the case of incompressible flow:
if there is no pump or turbine, the fourth term on the left dsappears;

if the elevations at the two points are equal, the third term on each side
cancels;
in most cases of completely turbulent flow, the hnetic energy correction
factors, alpha 1 and 2, each can be treated as being equal to one;
if the fluid is incompressible,and the pipe at both points has the same
dameter, the second term on each side is equal and cancels (the alphas
each being made equal to one).
and Equations
C H A P T E R
T W O
Incompressible Fluid Flow
Losses o f Mechanical Energy
It is worth taking a close look at each term in the ~ernoulli'equation to

see how it is applied in practice.
The variables of the first three terms on each side have numbered subscripts.The terms represent different forms of mechanical energy per unit
mass of the flowing fluid at the sections labeled 1 and 2.The last terms on
each side do not have numbered subscripts.They are the mechanical
energy per unit mass of flowing fluid added or subtracted (on left), or
"lost" (on right) between points 1 and 2.
The word "lost" means mechanical energy has been converted to internal
energy. If the system is adiabatic, the fluid gets hotter. If it is diabatic or
isothermal, there is a transfer of heat across the system boundaries (the
pipe wall).
The complete Bernoulli equation represents a total energy balance
between two points. The word "balance" suggests each term may have its
value altered and the changes may be balanced by changes in all or some
of the other terms.There are constraints, however.The last term on the
right is always a positive increment. It is the term that accounts for lost
mechanical energy or, better stated, the conversion of mechanical energy
to internal energy and, ultimately, to heat flow to the environment.
We now will examine each term of the Bernoulli equation individually.
The word "specific" in the headings refers to the fact that the energy is
on a per unit mass basis.
Specific pressure-volume energy

The first term on each side of the Bernoulli equation is the energy per
unit mass associated with the static absolute pressure. Static pressure is the
pressure exerted normal to the flow. It can be measured by a pressure
gauge on the pipe wall if the pipe tap is flush with the pipe and has no
burrs. It is also the pressure that a hypothetical device flowing with the
fluid would measure. If the pressure gauge measures "gauge" pressure, the
atmospheric pressure must be added in order to obtain absolute pressure.
Caveat
- Hydraulic engineering practice

In hydraulic engineering practice, when deahng with liquids, gauge pressure
is often used in the above equation in place of absolute pressure.The result
is usually correct since, in the Bernouh equation, differences are being
computed. It is best always to use absolute pressures in order to avoid forgetting them when they are needed (mainly for compressible fluids).
54
and Equations
C H A P T E R T W O
The density in the denominator of the first term on each side of the
~ernoulli'equation carries a subscript because density can change in
compressible flow. In the case of incompressible flow, the densities are
equal on either side only if the temperatures, pressures and compositions
remain -the same.
Consistent units must be used. If the density is in lb,/ft3, the pressure
must be in lbf/fi2. If you wish to use psia, then multiply pressure in psia
by 144 in the equation. If the density is in kg/m3, the pressure must be in
Pa. In this case the energy units are joules per kilogram. If kilopascals are
used, the specific energy units for that term will be kilojoules/kg.The
first term on each side of the equation must be multiplied by 1,000.
ecific kinetic energy
The second term on each side is the kinetic energy.This term becomes
important when dealing with changes in pipe diameter and when there is
acceleration due to expansion of compressible fluids.The dimensional
constant, gc, is important in the customary U.S. system. In the S1 system, it
is not needed. However, it is worth being kept in the equation and simply
being treated as having a value of one, dimensionless, when S1 units are
used.This habit makes it easier for those who have to switch back and
forth between systems to remember its location in the equation.
The velocity, U, is the average velocity across the pipe at the numbered
section. It is not the point velocity. In the U.S. system the units are feet
per second. The S1 system uses meters per second.
Alpha is the kinetic energy correction factor that is close to (but greater
than) one for turbulent flow and is exactly two for laminar flow. Alpha is
necessary because the average velocity is used in most equations whereas,
in truth, the integrated point velocity should be used. The point velocity
varies across the section studied. In most industrial engineering problems
involving turbulent flow, alpha is treated as being equal to one without
too much loss of accuracy, but the assumption should always be checked.
Specific p o t e n t i a l energy
The third term on each side represents the energy per unit mass due to
position (potential energy) at sections, 1 and 2.The units of Z are feet or
meters.The dimensional constant is again 32.17 lb;ft*lb;'
* s - ~in the
customary U.S. system. It can be treated as one, dimensionless, in the S1
system.Acceleration due to gravity, g, is 32.17 f t ~ s in
- ~the U.S. system.
and Equations
C H A P T E R
T W O
Incompressible
Fluid Flow
It is 9.81
in the S1 system.These last two numbers are approximations.They are true only at fixed latitude and at sea 1evel.Their use is adequate for most engineering computations in industrial practice.
When using real numbers for Z, it is best the select the lowest point in
the system as a datum line and to measure all elevations from it.This
avoids the use of negative numbers when the pipe dips below its starting
point.The base line is selected for convenience in the computation.
Again, because we are dealing with differences, there is no absolute
height.
Caveat
- Gravity trap
Be aware of the trap of treating the gravitational constant and the dimensional constant as being equal and capable of being cancelled. This leads to
simplified equations, but conceptual difficulties, and sometimes to errors.
Specific energy added or subtracted

The fourth term on the left is the energy added per unit mass by a pump
or subtracted (negative sign) by a turbine.This term does not represent
the energy added or subtracted at a particular section. It is the energy
added or subtracted at any location between the two sections.
The energy added or subtracted per unit mass flowing is not all of the
energy delivered to the motor or removed from the shaft of a turbine per
unit mass flowing. There will be irreversibilities in both cases. More
energy is delivered to a pump motor than is transferred to the fluid and
more energy is extracted from a fluid than is delivered to a turbine shaft.
Efficiency factors have to be applied.The ~ernoulli'equation applies only to
energy changes within the fluid unless it is modified to contain an efficiency
factor. Note efficiency factors are speclfic to a physical transmission means a shaft, a pump impeller or motor, for instance.Watch the definitions!
Specific energy converted to internal energy
The fourth term on the right is the mechanical energy converted to
internal energy by "wall (skin) friction" or by "form friction". These terms
will be explained later. The fourth term is often given in terms of pressure
drop. Equating the loss term to pressure drop is a source of confusion,
because of the phenomena of pressure recovery and irreversibility. It is best
to use specific energy units so as to avoid mistakes when performing estimations.
56
and Equations
Caveat
C H A P T E R
T W O
- Bernoulli balance
The complete ~ernoulli'equation with the so-called loss term is often
called a mechanical energy balance. It is not. It is a total energy balance
when the loss term is included. The "loss" is the conversion of mechanical
to internal energy.
11-4: I R R E V E R S l B l l l T l E S D U E TO P I P E A N D F I T T I N G S
Considerable effort has been expended to establish correlations for fluid
flow through various types and sizes of pipe, valves and fittings and the
resulting permanent losses of mechanical energy (the fourth term on the
right of the Bernoulli equation). Mechanical energy losses can be categorized as being those associated with:
straight pipe of u d o r m dameter;
sudden contractions;
sudden expansions;
gradual expansions or contractions;
valves treated as fittings;
orhces;
miscellaneous fittings;
madolds with dvidng or combining flows;
the total system (all of the above).
We will discuss each of the above categories and give the basic equations
for estimating the irreversibilities. Following this discussion we will give a
series of examples of each case.The last example will treat several devices
as a system.
There are two main approaches to estimating irreversibilities. One
approach, mainly applicable to straight conduits, uses the
equation directly.The other approach makes use of the component loss coefficient, the "K" factor.We will first discuss straight pipes of uniform diameter and then explain the component loss coefficients.
Straight pipe of uniform diameter

For incompressible fluids flowing through straight, horizontal pipe of uniform diameter with no pump or turbine in the segment, the Bernoulli
and Equations
C H A P T E R
T W O
Losses o f Mechanical
Energy
equation (11-1) reduces to,
The last equation gives the means of measuring head "losses" in constant
diameter pipe.This is about the only situation in which the irreversibilities equal the pressure drop divided by density.The pressure drop divided
by the fluid density, in this case, gives the energy per unit mass converted
to internal energy. Most often, we do not have to measure this loss.We
have to predict it.
Prediction of the loss term is based on many correlations that have been
performed over the years. For uniform diameter pipe, the loss term is
equation,
given by the
arc^^"
It seems fairly logical that the conversion of mechanical energy to

internal energy should be proportional to the kinetic energy ( u ~ / ~ ~ J ,
the length (L), to a friction factor (fM)and inversely proportional to the
pipe diameter (D).
The Darcy equation applies to laminar and turbulent flow. It applies also
to incompressible and compressible flow. However, for compressible flow,
the equation must be used in differential form because velocity and the
friction factor are not constant.The compressible flow form of the equation will be developed in the chapter on compressible flow.
For pipes, the friction factor has been correlated with fluid viscosity and
density, pipe diameter, wall roughness and fluid velocity. The correlation is
presented as two charts; one for any type of commercial pipe, Figure 1-4,
and the other for clean commercial pipe, Figure 11-1. It may also be computed from the churchill-usagiAi equations given in Appendix AII.
The correlation for any type of commercial pipe gives the ~ o o d ~fric~ " '
tion factor versus the ReynoldsVnumber and the pipe relative roughness.
The relative roughness is the average height of the wall irregularities, E,
divided by the diameter, D, in the same units. In fully turbulent flow, the
friction factor increases with relative roughness. In larninar flow it does not.
58
and Equations
C H A P T E R T W O
Note that, for the same type of commercially available pipe, the absolute
roughness will be independent of diameter.The relative roughness, therefore, will vary inversely with diameter.
To use Figure 1-4, you have to know the type of pipe material.The
average absolute roughness of several pipe materials is given in Table 11-1
Simply divide the absolute roughness in feet or millimeters by the pipe
internal diameter in feet or millimeters to obtain the relative roughness.
The correlation for clean commercial pipe, Figure 11-1, gives the friction
factor as a function of the ReynoldsVnumber, the pipe diameter and
schedule number.This chart is the one most commonly used in design
engineering. Once the ~ o o d ~friction
~ ' " factor has been obtained, the
mechanical energy losses in ft-lbf/lb, or J/kg for uniform diameter pipe
may be found by plugging the appropriate data (length, diameter, velocity
equation.
and Moody friction factor) into the above
arc^^"
Table 11-2 is extracted from, Internal Flow Systems, D.S. Mdler, 2nd
Editionxv1.It is intended
...
for general hydraulic practice. The minor differences with Cranex""may be due to differences in European and North
American pipe fabrication practices. Miller states that experience with
similar systems is the best guide to selecting roughness values and deterioration tolerances. Pipes conveying water are particularly prone to deposits,
erosion, corrosion, bacterial slimes and growths and to marine and fresh
water fowling. For water, he recommends using an absolute roughness
ranging from 0.5 mm to 2.0 mm.The smaller value is for chlorinated, filtered, clear, unaggressive and non scale-forming water. The larger value is
for worst-case unchlorinated water. Both values are for water pipes or
conduits after several years of service. Industrial fluids tend to be more
predictable than water over the long term.
I TABLE 11-1.Absolute rouqhness of various materials
Material
I Drawn tubinq
Commercial steel
Asphalted cast iron
Galvanized iron
Cast iron
Wood stave
Concrete
Riveted steel
Absolute roughness, E , feet

0.000005
Data from Crane
mm
0.0015
0.0001 5
0.046
0.0004
0.1 22
0.0005
0.1 52
0.00085
0.259
0.0006 - 0.003
0.1 83-0.914
0.001 - 0.01
0.305-3.05
0.003 - 0.03
0.914-9.14
and Equations
C H A P T E R
T W O
Incompressible
Fluid Flow
Friction Factors for Clean Commercial Steel PipeA0
and Equations
Figure 11-1. ~ o o d ~friction

~ " ' factor versus Reynoldsvnumber and pipe schedule
60
Losses o f Mechanical
Energy
I TABLE 11-2. Absolute rouahness of various materials

Material
C H A P T E R
T W O
Data from D.S. Miller
Absolute roughness, E feet
mm
Smooth pipes
Drawn brass, copper, aluminum, etc.
0.0000082
0.0025
Glass, plastic, Perspex, Fiberglas, etc.
0.0000082
0.0025
Steel pipes
New smooth DiDes
0.000082
0.025
Centrifugally applied enamels
0.000082
0.025
Mortar lined, good finish
0.00016
Mortar lined, average finish
0.00033
0.05
0.1
Liaht rust
0.00082
0.25
Heavy brush enamels and tars
0.0016
0.5
1 .o
~~~~~
Water mains with general tuberculations
0.0039
1.2
0.000082
0.025
Concrete pipes
New, unusually smooth concrete with

smooth joints
Steel forms, first class workmanship with
smooth ioints
New, or fairly new, smooth concrete
and joints
Steel forms, average workmanship,
smooth ioints
I
I
o.oooo82
0.00033
0.00033
Wood floated or brushed surface in

good condition with good joints
0.00082
Eroded by sharp material in transit, marks

visible from wooden forms
0.0016
Precast pipes, good surface finish,

averaqe joints
0.00082
Segmental lined conduits in good ground
0.0033
0.025
0.1
0.1
0.25
1 .o
2.0
conditions
Other pipes
0.0000082
Sheet metal ducts with smooth ioints

Galvanized metals, normal finish
Galvanized metals, smooth finish
Cast iron, uncoated and coated
Asbestos cement
Flexible straiaht rubber DiDe with a smooth bore
Mature foul sewers
I
I
I
I
I
I
0.0025
0.00049
0.15
0.000082
0.025
0.00049
0.15
0.000082
0.025
0.000082
0.025
0.0098
3.0
and Equations
61
C H A P T E R
T W O
lncompressible
Fluid Flow
Accuracy of loss estimates for pipes - the use of safety factors

Accuracy of loss estimates depends on knowing and using the actual pipe
internal diameter, not the nominal hameter. For metering devices that
make use of piping to form part of their "run", the diameter is of great
consequence.This is why meter runs are often prefabricated so they may
be swaged into larger diameter piping. Permanent mechanical energy
losses vary inversely with approximately the fourth power of the diameter.
ille er"^' states that a 100% error in the roughness causes a 10% error in
the friction coefficient.
Caveat
- Safety factors
Do not add "safety factors" indiscriminately with each computation. It is
best to perform all computations without them and then to logically
decide where they must be applied in light of the total problem. For
instance, in choosing a pipe size one must choose a commercially available pipe.The diameter will be greater than the required diameter.This in
itself constitutes a safety factor.
Component loss coefficients, 'K' factors and 'skin friction'

When an incompressible fluid is flowing in a long, straight, horizontal
pipe, the
equation gives the mechanical energy losses. These irreversibilities are often described as "slun friction". This unfortunate term is
another of those loose and misleading terms used to describe natural phenomena. It arises from fact that when studying relative flow past profiles
(not in pipes), turbulent flow develops next to a solid boundary and then
spreads in the normal direction. In confined flow, turbulence may be
more intense near the solid boundary, but it certainly does not constitute
a "skin". In fact, this term really only applies to structures such as aeroplane wings. In pipes, the boundary layer fills the entire cross section.
arc^""
When fluid passes through a valve or a fitting, the flow profile is radcally
hsturbed and there are large edhes. Boundary layer separation occurs and
we are dealing with what is known as "form fi-iction". Form fiiction is actually the irreversibhties associated with the vortices that arise fkom major
changes in drection - obstructions, for instance. Friction factor correlations
were developed for slun fiiction, not form fiiction. K factor correlations
were developed to measure the irreversibhties associated with form fiiction.
The flow paths through some valves and some fittings change drastically
and it is not possible to derive theoretical equations for each case.This
does not stop engineers from establishing empirical relationships.
62
and Equations
The
C H A P T E R T W O
arc^"" equation and the 'K'
factor equation compared

We can compare the Darcy equation for straight pipe with the theoretic
equation for a sudden expansion and the empirical one for a sudden contraction. We can establish a relationship between the loss coefficient, K,
and the product of the friction factor and the length to diameter ratio.
We can say a given fitting would cause the same mechanical head loss as
an equivalent length of straight pipe with a diameter ratio, L/D, whose
friction factor, under specific flowing conditions, would be given by fM.
The velocities and therefore the diameters in which the velocities
occurred would be the same.
The loss coeficient, K, for a fitting with one inlet and one outlet can be
seen to be the ratio of the permanent mechanical energy losses across a
fitting and the kinetic energy in one of its legs (inlet or outlet). It is much
more a function of form friction than slun (path) friction. Therefore, it
frequently is considered to be independent of the friction factor. It is frequently considered to be dependent only on the geometry of a particular
valve or fitting.This is not quite true, so a correction must be made to the
assumption.
Caveat
- Velocity and the K factor
The numerical value of K depends upon the leg in which the velocity is
measured.The leg in which the velocity is measured must always be specified and the appropriate K factor must be used.
Geometric similarity and dissimilarity

In theory, geometrically similar valves and fittings would have the same
loss coefficient, no matter what the size.This turns out to be the case for
sudden contractions and expansions, which are only dependent on the
beta ratio, for pipe entrances, which are only dependent on the geometry
(rounding), and for pipe exits.
Flow o f I n d u s t r i a l F l u i d s - T h e o r y
and Equations
C H A P T E R
T W O
Incompressible
Fluid Flow
For pipe exits (to a tank or to atmosphere) the loss coefficient is one
(1.0), regardless of pipe size.This means the permanent mechanical energy
losses are numerically equal to what is called one velocity head (u2/2&)
across any pipe exit, based on the velocity within the pipe.The loss across
any pipe exit in energy per unit mass units can be found by multiplying
by one the average cross sectional velocity in the pipe exit squared
divided by two and the dimensional constant.The customary U.S. units
for the lost mechanical energy would be foot-pounds force per pound
mass.The S1 units would be joules per lulogram (N.m/kg).
Pressure drop not always equal to head loss

It is worthwhile pausing here to drive a point home.The ~ernoulli'equation is often manipulated to change specific energy units into head units
or even to pressure units. When pressure units are used it is particularly
difficult to keep in mind which pressure is being discussed. For instance,
the head loss across a square edged pipe exit, when expressed in pressure
units, is given as,
(11-5)
u2
AP=p--
2gc
It is very easy to assume this pressure drop means the pressure inside the
pipe at the discharge point is greater than that within the tank at the same
level. Such is not always the case. For instance, if water is flowing at 15 fps
through a pipe outlet to a tank, expressed in pounds force per square inch,
the head loss across the square edged exit would be approximately,
Expressing the loss in pressure units leads very easily to the assumption
that this is the measured pressure drop across the exit.This is not true.
The following development will prove this statement.
In this type of equation (11-6), the differential pressure term is positive
If the Bernoulli equation is left in its specific energy form, it is easier to
see the real relationship-We can state that the losses (the non recoverable
mechanical energy), across a square edged pipe outlet are,
64
and Equations
C H A P T E R
T W O
There is no confusion with either of these units.We know we are talking

about the loss term in the ~ernoulli'equation, 11-1.
In Equation 11-8, section 1 is just inside the exit from a horizontal pipe;
section 2 is in the tank.The Bernoulli equation, the first line of set 11-8,
can now be manipulated as follows (the alpha corrections each equal one).
The density of the flowing fluid will have decreased negligibly due to the
increased temperature.
The pressure inside the pipe, PI, equals the pressure, P2, at the same elevation within the tank at the point where the velocity is negligible.There is
no measurable difference between the pressure immediately inside the
pipe exit and that at the same elevation in the tank at a point where the
fluid is practically stationary.The kinetic energy has been completely converted to internal energy and the downstream fluid is warmer than the
upstream fluid.
This becomes clear if it is remembered the term associated with the K
factor represents irreversibilities. In the second from last equation of the
set 11-8, the first two terms on the left, static energy and kmetic energy,
represent the driving potential that is exactly balanced by the static pressure potential and the irreversibilities on the right. If there were no irreversibilities, the static pressure potential downstream would balance both
and Equations
65
C H A P T E R
T W O
Incompressible
Fluid Flow
Losses of Mechanical
Energy
driving potentials. For the same velocity, the pressure upstream would be
less. Recovery would have taken place. In the above example, if K is not
equal to one, the pressure will not be equal.
We will see later that recovery, a hgher pressure downstream than upstream,
is a common occurrence with certain types of fittings and with valves.
Caveat - Mechanical energy 'losses' versus pressure

Do not fall into the temptation of equating real mechanical energy losses
to pressure drop, as is very frequently done in the literature.
The loss coefficient, K2, based on the velocity within the pipe, for a sharp
edged pipe entrance is 0.5, regardless of pipe size.The irreversibilities for
the same velocity (kinetic energy) in the same sized pipe would be
exactly half those of a pipe exit.
In practice, with the exceptions given above, similar fittings are not geometrically similar over all sizes. Curvatures within fittings and valves
change as sizes change.The relative roughnesses of different sizes are not
equivalent.The value of the absolute roughness, E, generally remains the
same as the diameter increases, so the relative roughness decreases with
the increase in diameter. In any case, the effects of flow separation in fittings and valves usually outweigh the effects of normal pipe turbulence.
Useful practical correlations from cranexiii

Crane gives a useful empirical correlation among the type and size of a
valve or fitting and its loss coefficient.The equivalent K, fML/D, at fully
developed turbulent flow for 30 diameters of the same size of pipe is the
basis for the correlation.The chart is reproduced as Figure 11-2 (Fig. 2-14,
Crane). It shows that, for a given type of valve or fitting, the slopes of the
K versus valve or fitting size are very similar to the slope of the fML/D,
versus the same size of pipe 30 diameters 1ong.The pipe chosen is clean,
commercial Sch. 40 pipe.
Crane states that it is probably a coincidence the effect of geometric dissimilarity between different sizes of the same line of valves or fittings
upon the resistance coefficient, K, is similar to that of relative roughness
upon the friction factor.
66
Flow of I n d u s t r i a l Fluids-Theory
and Equations
Incompressible
Fluid Flow
Losses of
Mechanical Energy
Resistance Coefficient K, Equivalent Length

And Flow Coefficient CV
continued
0
Q
Q
4
+
d
T W O
L/&
- Resistance Coefficient
Figure 2-14, Variation of Resistance Coefficient K (=f

Symbol
C H A P T E R
LID) with Size
Product Tested
Authority
Schedule 40 Pipe, 30 Diameters Long (K = 30 f ~ ) *... . .Moody A.S.M.E. Trans.. Nov.-1944"

Class 125 Iron Body Wedge Gate Valves . . . . . . . . . . . . . . . Univ. of Wisc. Exp. Sta. Bull.. Vol. 9,No. 1, 192218
Class 600 Steel Wedge Gate Valves . . . . . . . . . . . . . . . . . . . . Crane
.
Tests
90 Degree Pipe Bends, R I D = 2. . . . . . . . . . . . . . . . . . . . . . . Pigott A.S.M.E. Tram., 19506
- 90 Degree Pipe Bends, R I D = 3.. . . . . . . . . . . . . . . . . . . . . .Pigott A.S.M.E. Trans., 19506
- 90 Degree Pipe Bends, R/D = I . . . . . . . . . . . . . . . . . . . . . . . Pigott A.S.M.E. Trans., 19506
- Class 600 Steel Wedge Gate Valves, Seat Reduced . . . . . . . .Crane Tests
- Class 300 Steel Venturi Ball-Cage Gate Valves. . . . . . . . . . . Crane-Armour Tests
- Class 125 Iron Body Y-Pattern Globe Valvea . . . . . . . . . . . .Crane-Armour Tests
- Class 125 Brass Angle Valves, Composition Disc . . . . . . . . .Crane Tests
- Class 125 Bcasa Globe Valves, Composition Diac . . . . . . . . .Crane Tests
*jr = friction factor for flowin the zone of complete turbulence; see page A-26.
Figure 11-2. Component loss coefficient correlation
In other words, the happy coincidence is that, as the pipe diameter

increases, the relative roughness, &/D,for the same material in pipes
decreases. This decrease influences the friction factor at complete turbulence in the same direction and at a similar rate as the geometric dissimilarity of commercially available valves and fittings influences the K factor.
and Equations
C H A P T E R T W O
The correlation is really between the friction factor, which is relative to

"skin" turbulence, and the K factor, which is relative to "form" turbulence.
This is a good place to point out the location of the cause of the turbulence does not necessarily coincide with the location of the irreversibilities. Irreversibilities occur downstream where the turbulence is destroyed.
...
The above correlation allows CraneX1''to establish simple relationships

among the type of valve or fitting, the size and the friction factor at complete turbulence of the same sized schedule 40 pipe.These relationships
may be found in Crane Technical Paper No. 410. They are essentially all
statements that the K factor we are interested in is equal to a friction
factor times a constant.This constant is the equivalent L/D of the fitting.
The friction factor is that which would be associated with fully developed
turbulent flow in schedule 40 pipe of the same diameter.
Note this methodology is not applicable to all types of fittings.We will
discuss these cases shortly.The method is useful when there is no better
one or for quick estimates.
Component loss coefficients
The component loss coefficient, or the K factor, is a performance parameter for a component within a system. It gives a measure of the permanent mechanical energy conversion to thermal energy due to the presence of a specific component within the system.
Figure 11-3 is an example. It depicts an orifice plate that can be removed
from a pipe without disturbing the pipe's characteristics.Total energy is
shown on the ordinate and length is depicted on the abscissa.The pipe is
horizontal in order to simplify the ana1ysis.Although an orifice plate is
used in the example, any component could have been used.
Figure 11-3. Component loss coefficients
68
and Equations
C H A P T E R T W O
The sections numbered 1 and 2 are sufficiently removed from the orifice
location so the flow profiles at these two points are substantially undisturbed when the orifice is in place.The flow profile will be disturbed on
either side of the orifice; but at the numbered locations, it is normal.
Section 1 is the reference location. Properties downstream of this location
may be compared to its properties.
The figure shows the changes when the orifice is in place.The flow rate
is the same in both cases, with or without the orifice plate.The datum
line is the bottom horizontal line. It could have been taken as the center
line of the pipe. The total mechanical energy is given by the sum of the
elevation head, the static pressure head and the velocity head.
Upstream of the orifice, the profile and the head loss will be the same with
or without the orifice except in the immediate vicinity of the plate.
Downstream, with the orifice in place, there will be a more radical change
that will become of constant slope after thirty to fifty pipe diameters.
Caveat
- Pressure recovery
When measuring irreversibilities, care must be taken to allow sufficient
distance upstream and downstream so the flow profile becomes re-established. Measuring pressure drop across an orifice, for instance, will include
both permanent pressure losses and recoverable pressure drop as can be
seen from Figure 11-3. The quantity of interest is really the irreversibilities
that occur due to the presence of the plate.These irreversibilities may
occur away from the plate as the profile recovers.
The total head or mechanical energy loss can be computed in both cases,
with and without the orifice plate. Subtracting the loss without the plate,
AHwo, from the loss with the plate, AHw, gives the loss attributable to
the plate alone, A(AH).This method of arriving at the irreversibilities
accounts for upstream and downstream disturbances due to the presence
of the component only.
Tightening up terminology
It is to be noted the difference between irreversibilities computed for two
sections is normally symbolised as AH.The loss attributed to the component is given the same symbol, but this habit lumps the upstream and
downstream irreversibilities into those due to the component. In the
figure, we have used A(AH) to try to be more specific.
and Equations
C H A P T E R
T W O
Incompressible
Looking at the figure, it can be seen that, with fully developed flow profiles in straight pipe, the total head lines can be projected to the orifice
plate 1ocation.The difference between the points of intersection with the
plate will be the same as the difference referred to previously as A(AH).
The permanent irreversibilities due to the plate or any other component
are given by A(AH).These irreversibilities will vary with the component
but will be found to be proportional to the kmetic energy, ~ ~ / 2 ~ , . ~ h e
average velocity must be associated with a specific section (of pipe, not of
the orifice) when the K factor is computed.
Note we have an equation in the form of,
The term on the left is the difference between the head losses with and
without the component. We are talking about changes in total mechanical
energy that occur slightly upstream and within 30 to 50 pipe diameters
downstream of the component.The velocity, U, must be measured or estimated at the same time.The component loss coefficient, 5,is then taken
as the ratio of the difference in projected mechanical energy across the
component to the kinetic energy at a particular point.
The coefficient must always be associated with the point at which the
velocity was measured. Ki is dimensionless. It has the same value in S1
units as it does in customary U.S. units.
Figure 11-4 gives an example of how the loss coefficient is established.
The example is that of a 90 degree mitered pipe.The example is modified
from D.S. Millerm1.The pipe is horizontal. The pipe is of uniform diameter (velocity at point 1 is equal to velocity at point 3 because the fluid is
incompressible).Point 1 is upstream of the miter and the manometer at
this point gives the static pressure at the inlet. Point 3 is sufficiently
downstream as to represent the point at which pressure recovery is complete. Point 2 represents a vena contracta that is formed due to the sudden
change in direction at the miter, the wall separation and reverse flow that
takes place immediately behind the bend.
The average velocity at point 2 will be greater than at points 1 and 3.The
series of manometers depict the interchange of static pressure energy and
kinetic energy and the partial reverse process as pressure energy is recov-
70
and Equations
C H A P T E R
T W O
ered. The permanent irreversibilities are seen to be approximately numerically equal to u2/2& or what is termed one velocity head. If the velocity
were measured at the vena contracta, it could be inferred an additional
two velocity heads had been lost. If the velocity were measured at point 3
a more realistic picture of the non-recoverable losses emerges. Point 3
should be at least 30 pipe diameters downstream of the miter.
Figure 11-4. Establishment of loss coefficient for 90 degree mitered elbow
The letters identify the terms in the ~ernoulli'equation.
A equals u12/2&, the "velocity head" at section 1,upstream of the miter.

Without flow, the liquid would reach the upper line of the A interval.
With flow, its height is indxated by the dashed curve.
B equals Pl/p, the "static head" at point 1.
C equals gZ/&, the "elevation (potential) head".
D equals hL or IWl2/2&, the permanent losses or irreversibilities.

E equals approximately 2 ~ ~ ~It represents
/ 2 ~ ~the. static head, P/p that wdl be
recovered over and above the head drop that occurs at the vena contracta.
F equals P3/p, the static head at section 3.

Note that, to obtain the true measure of the influence of the miter bend,
identical data would have to be obtained on a straight run of identical
piping.The difference between the head loss with the miter in place and
that without it gives the loss attributable to the miter alone. It is this concept that is emphasised by the symbol A(AH).
and Equations
71
C H A P T E R T W O
Incompressible
Fluid Flow
To quantify the error that would occur if pressure measurements taken at

points 1 and 2 were used to compute a head loss, the following development can be made:
If differential pressures between points 1 and 2 were transformed directly

to permanent losses by dividing by the density, there would be about a
three-fold error. Pressure recovery would have been neglected.
Note carefully the subtle distinctions drawn by the subscripts.The K factors and the permanent losses are between two points.The average
velocity (kinetic energy term) is measured at a fixed section whose location must be clearly stated.
If a similar development is followed for the loss coefficient between points
1 and 3, it wdl be seen that KI3 equals approximately 1.This is much
closer to the true value.The above example shows clearly why the differential across the normal flange taps of an orifice plate cannot be used to
represent permanent losses. Pipe taps would give a closer approximation.
Basic loss coefficients

It is convenient to establish loss coefficients at a particular ReynoldsV
number and to make adjustments to the coefficient for other values of the
Reynolds numbers. D.S. Millerxv' gives the following reasoning in picking
the standard Revnolds number:
most major iidustrial flows operate between 106< NRe< 108;
variations in K are quite small for NRe> 0.5 x 106;
commercial laboratories cannot economically run tests at NRe> 106,
therefore:
NReequals 1 X 106 is an acceptable standard Reynolds number upon

whch to base moddications (as a starting point for computation).
72
and Equations
Incompressible
Fluid Flow
C H A P T E R T W O
The use of the basic loss coefficients allows tabulation of standard K factors
that are then modified for actual conditions.We will demonstrate this by
examples. It should be noted other authors use 1 X lo5 as a standard
ReynoldsVnumber.
Sudden contraction
The worst case loss for a reduction in dimension is a sudden contraction.
It occurs when a smaller pipe is flanged drectly downstream of a larger
pipe (Figure 11-5).
Figure 11-5. Sudden contraction
Figure 11-6. Sudden expansion
It can be seen that a vena contracta occurs downstream of the entrance to

the smaller pipe with the usual turbulence and losses in mechanical
energy.When the loss is taken as proportional to the velocity head
(lunetic energy term) in the smaller pipe, the loss in mechanical energy is
established empirically as,
In other words the loss in mechanical energy per unit mass due to the contraction is proportional to the square of the average velocity at section 2
(the point of full recovery of profile) divided by two and the dimensional
constant.The factor of proportionality, K,, is the contraction loss coefficient.
It is initially established by measurement.The measurements are then correlated with pipe dimensions to establish a semi-empirical formula.The
use of the formula then obviates the necessity of malung further measurements. The irreversibilities are negligibly small for larninar flow.
For turbulent flow the loss coefficient is given by,
The beta ratio is seen to be simply the ratio of the smaller to the larger
diameter.
and Equations
C H A P T E R T W O
Incompressible
Fluid Flow - Losses of Mechanical Energy
Due to the forming of guided flow streams resembling those of a venturi,

the irreversibilities from a sudden contraction are less than those from a
sudden expansion.
To estimate losses in mechanical energy across a sudden contraction, the
average velocity in the smaller pipe at a point where the profile is fully
developed and the beta ratio are needed.These numbers are plugged into
Equations 11-1l and 11-12 to obtain the loss in mechanical energy. In ad&tion, if it is necessary to estimate the pressure drop across the same contraction, the ~ e r n o u l hequation
~
is necessary. The permanent head loss and the pressure drop divided by density will not be the same.Ths is a point that should be
kept in mind.
With all piping components, external forces must be exerted by the
piping anchors to counteract the forces developed by the flowing fluid.
Otherwise, the piping will move. These forces are easily estimated from
pressure drop and impulse-momentum data. This exercise is required of
piping engineers. Often they require input from process and control systems personnel.We will consider the estimation of these forces to be
beyond the scope of this book.
Sudden expansion
The worst case for an increase in dimension is a sudden expansion.The
gradual tapers associated with expanders have lower head losses than those
of a sudden expansion. Figure 11-6 represents a larger pipe flanged directly
to a smaller one.
The losses in mechanical energy associated with a sudden expansion are
proportional to the kinetic energy, the velocity head, at a section where
the velocity profile is fully developed in the smaller (upstream) pipe.The
loss equation is,
This equation is almost identical to that for a sudden contraction.The loss

coefficient for a sudden expansion has a different value than that for a
sudden contraction. It is,
The above equation is not entirely empirical; it can be derived from theoretical considerations (see Appendix AII) .
74
and Equations
incompressible F l u i d F l o w
Energy
C H A P T E R
T W O
It is instructive to compare the relative magnitudes of a sudden expansion

and a sudden contraction with the same beta ratios - which we will take
as 0.5.
The mechanical energy conversion at a sudden expansion is greater by a

factor of one and a half than that at a sudden contraction when the beta
ratio of each is 0.5.
Valves as fittings
Valves are treated in two ways in fluid flow engineering depending on the
purpose of the computation.They are regarded as control valves when
they are being sized or when reasonably exact estimations of overall irreversibilities and pressure drops must be performed. When the total system
is being analyzed, they are frequently regarded as just another fitting just another obstruction.When they are regarded as control valves, equations such as those of ISA-75.01 .O1 are used. Factors are used to adjust
the equations to the specific case and type of valve.
When a valve is a block valve or an on-off valve, it is usually fully open
when fluid is flowing. In this case it frequently is treated as a fitting and
standard K factors are used. In truth, treating the valve as a control valve
in all cases would give more accurate results, since the valve sizing equations apply to all valves. In one of the examples that follow we will treat
the valve as a fitting.The more accurate treatment will be given later.
Caveat
- Source of K factor correlation

It should be stated that valves from different manufacturers would have
different flow patterns. They will therefore have different K factor correlations.This means that, if the K factor method is used to estimate irreversibilities, the factors should be obtained from the specific manufacturer's literature.
Rather than use one manufacturer's K factor for another's valve, a better
method is to use L.R. Driskell's Table of Representativevalve Factors,
ISA Handbook of Control Kilverxvii,to obtain a typical valve flow coefticient, C,, for a specific valve type and then to transform it to a K value
using the method given in Appendix AII.
and Equations
C H A P T E R T W O
Orifice plates
- recovery and permanent losses across an orifice plate

There are two distinct components of pressure drop across an orifice plate
and the ratio of the two changes with the location of the pressure taps.
The components are combined and appear as one when measured
directly. However, they must be computed separately because one is
recoverable and the other is not.The pressure profile across an orifice
plate is shown in Figure 11-7.
The pressure drop associated with metered flow includes both components. If vena contracta taps are used, the measured pressure drop is considered the maximum pressure drop available across the restriction. It gives
the largest signal for a given flow rate.This pressure drop is useful to infer
flow; the greater the differential pressure, the easier it is to read the signal.
It tells us nothing about the irreversibilities. (Note the maximum available
pressure drop is really the drop between an upstream corner tap and the
vena contracta, but these locations are not standard ones.)
Figure 11-7 shows the longitudinal flow profile and pressure profile over a
length of piping corresponding to the length between pipe taps (specified
by the distances from the upstream edge of the plate).The reason for the
slight pressure build-up upstream of the plate is that kinetic energy is
converted to static energy as the fluid is decelerated in the corner. Note
this extra pressure is necessary to accelerate the fluid centripetally toward
the vena contracta. After again decelerating during the expansion, the
fluid is once again at unaccelerated steady-state.
Figure 11-7. Recovery and permanent losses across an orifice
76
and Equations
Incompressible
C H A P T E R
T W O
For pipe taps, the term dP is the pressure drop ten and a half pipe diameters downstream of the upstream pressure tap. This pressure drop dwided
by the density is often used to estimate the permanent losses even though
this term still contains some recoverable energy. The quantity AP/p estimated from pressure drop over these taps is not strictly equivalent to the
permanent losses, A(AH) or h=. It is within two or three percent of the
permanent losses in mechanical energy of an orifice plate.
It is necessary to know the irreversibilities if we wish not to compute a
flow for metering purposes, but to establish what the flow w d be under
various circumstances. In other words, we wish to consider the orifice as
just another device that creates irreversiblities.
There are two ways to estimate permanent losses across an orifice.The
first is to make use of the fact that the pressure drop across pipe taps,
2 1/2 diameters upstream and 8 diameters downstream of the upstream
face of the plate, approximate the permanent losses. In this case the permanent losses (converted to pressure drop) and the pressure drop measured across the taps coincide to within a few percent. It is assumed all the
recovery has taken place within eight dameters (remember a minimum
of 30 diameters is recommended for complete profile recovery).
The permanent pressure drop can be estimated from the formula for pipe
taps.The second method is to make use of a head loss formula simdar to
those using the K factor.We will discuss this method here.
Head loss (mechanical energy) formulae
A head loss (mechanical energy) formula for square edged, concentric
orifice plates (Simpson,L.L., Chemical Engineering,July 17, 1968miii)is:
(11-1 6)
U:,
h. = K o 2gc
Note that in Equation 11-16 the velocity is that in the pipe, not the orifice. If the beta ratio is known, Equation 11-16 can be used to estimate
K,. This factor then can be used with the average velocity in the pipe at
fully developed profile to estimate the head losses due to the presence of
the orifice.
and Equations
C H A P T E R T W O
losses of Mechanical Energy
An empirical formula for permanent loss across an orifice is given in

R.W. Millerxm.It is curve-fitted to the ASME head loss curve (ASME
Research Report on Fluid Meters).The formula is,
The loss is given in terms of the ratio of the permanent pressure drop to
the full-scale differential pressure across the orifice plate. Since the terms
on the left constitute a ratio, any units may be used as long as they are the
same for both terms.
It is sometimes necessary to use an orifice as a restriction orifice in order
to fix the maximum flow rate through a system. In this case, the total
irreversibilities necessary to fix the flow can be estimated. The irreversibilities available because of actual system configuration may then be estimated.The difference between the two gives the irreversibilities that must
be generated by the restriction orifice. Example 11-6 treats this case.
Miscellaneous fittings and manifolds

So far we have discussed rather simple fittings such as sudden enlargements or contractions.These fittings are rarely encountered. They are used
as worst-cases for conservative estimates of irreversibilities.
Since the detail involved in estimating the irreversibilities of some of the
more complex fittings goes beyond the intent of this chapter, we will
consider them more fully in Appendix AII. We will describe some of the
different cases here to remind the reader of their existence.
Gradual enlargements
These are swages from a smaller to a larger diameter (expanders).The K
factor is less than that of a sudden enlargement.A formula is given in
Appendix AI1 for the case where the included angle is equal to or less than
45 degrees.When the included angle is larger the 45 degrees, the fitting
can be treated as a sudden enlargement without too much loss of accuracy.
Gradual contractions
These are swages from larger to smaller diameters (reducers).Two formulae are given in Appendix AII.The break point for the application of
one or the other formula is 45 degrees of the included angle between the
walls of the swage.
78
and Equations
incompressible Fluid Flow
C H A P T E R T W O
Inward projecting pipe entrance

This involves a pipe fixed to a tank wall in such a manner that it projects
through it. The entrance has geometric similarity for all sizes.Therefore, K
based on average fully-developed flow in the pipe is constant (0.78).
Square edged and rounded entrances
These entrances to a pipe from a tank are flush with the tank wall and
have different amounts of rounding on the internal edge.The square
edged entrance has a K value of 0.5. If the edge is rounded the associated
K value is inversely proportional to the degree of rounding.
Manifold flows (dividing and combining flows)
This term is a general description for either a single stream that is split
into two or more streams or for two or more streams that are combined
into a single one.The loss coefficient, K, depends not only on the fitting
configuration and dimensions, but also on the relative flows through the
branches and on their proximity to one another. Some of these cases are
the most difficult to estimate with any degree of accuracy.
These and other cases will be considered in more detail in Appendix AII.
A list of common K factors and their formulae will be given.Their
sources will also be given where they are known.
Total (system) mechanical energy losses

If we have a series of fittings, pipe sizes, valves and other obstructions, sufjcientlyfar apartfrom one another, we can simply sum all of the mechanical
energy losses per unit mass of flowing fluid to get the total loss per unit
mass between two points.
The subscript, Li, stands for all the various irreversibilities such as those
due to sudden expansions and contractions, various types and sizes of fittings, valves and other forms of obstructions. It is necessary to have a correlation as a formula to obtain the loss over a certain type of obstruction.
Many of these formulae can be found in cranex"'.we have already presented some of them. D.S. Millerxv1gives extensive charts for combining
and hviding fittings of various configurations. He also discusses the
problem of combining closely connected fittings.This is a problem that is
often overlooked in the general literature on fluid flow.
and Equations
C H A P T E R
Caveat
T W O
lncompressibfe
Fluid Flow
Losses o f M e c h a n i c a l Energy
- Proximity of disturbances
When fittings, bends and devices are close to one another, their effects on
total system irreversibilities are not necessarily additive. See D.S. Mdle?'
for more details.
Once the total mechanical energy losses are obtained for a particular flow
rate, the sum may be substituted into the ~ernoulli'equation and various
manipulations may be performed, depending on the problem to be solved.
Total energy, mechanical energy and hydraulic energy grade lines

A useful tool to visuahze the energy relationships of a flowing fluid is shown
in Figure 11-8, associated with Example 11-5. It represents the spec& energy
content of a fluid and the mechanical energy converted to internal energy.
Example 11-5, System Losses, makes use of these lines and attempts to
integrate most of the cases we have discussed.
11-5: E X A M P L E S O F E S T I M A T I O N S O F I R R E V E R S l B l L l T l E S
Example 11-1: Straight pipe of uniform diameter
Suppose we have 100 feet of horizontal pipe, 6" Sch. 40, with 60F water
flowing at 900 gallons (U.S.) per minute.What is the value of the loss term
and what is the pressure drop? The pipe is clean, commercial steel pipe.
...
All the data are readily available from Crane,Technical Paper No. 410.XU1
L = 100 feet
D = 6.065/12 = 0.505 feet
U = 9.99 feet per second
as-2
gc = 32.17 1b;ft~lb;'
The use of Figure 11-1 for a 6"- Sch. 40 pipe to obtain a friction factor
requires only computation of the ReynoldsVnumber and the size and
schedule of pipe. Remember that, although the symbol for viscosity is
similar in all the terms of Equation 11-19 the units of the last two terms
are different. The last two terms use centipoise. Because of the multiplicity
of units for viscosity, this is not an uncommon occurrence. Similarly, the
fact that mixed units (gpm, lbm/h, cP, fps) must be manipulated is not
80
and Equations
Incompressible Fluid Flow - Losses o f Mechanical Energy
Caveat
C H A P T E R T W O
- Viscosity units
Pay particular attention to the viscosity units when performing computations.They are one of the most common sources of error.
The test that must always be performed is to check that the ReynoldsV
number computed with the units chosen is indeed dimensionless.
(11-1 9)
DUp - DG
W
no2 3600
N, =-- = 6.31-= 6.3lU -p P
P
@LP
4
4-b
The pipe internal diameter and the fluid velocity have already been given.
The density of 60F water is 62.365 lbm/ft3.~he
viscosity is 1.14 cl? So,
the Reynolds number is 410,699 or approximately 4.1(105).
If centipoise is the unit of viscosity the third equation of 11-19 may be
used with W in pounds-mass per hour and d in inches. Alternatively, centipoise can be converted to pounds-mass per foot-second units by multiplying it by 6.72 (IO-~)
and the first two terms can be used. In these
cases, D is in feet, U is in feet per second, rho is in pounds-mass per foot
cubed, and mu is in lbmffl *S-'.
If S1 units are required, centipoise must be converted to N*s/m2 or, what
is the same thing, Pa*s,by multiplying cP by 10". In thls case the diameter, D, is in meters, average velocity, U, is in m/s, rho is in kg/m3.
To compute the Reynolds number using S1 units, but starting from mixed
units, ANSI piping and with viscosity given in centipoise (which is not an
S1 unit), the following exercise may be performed:
D = 6.065 inches = 0.505 feet =O.l539 meters
U = 9.99 @S= 3.0449 m/s

p = 62.365 l b m / d = 998.98 kg/m3
p = 1.14 cP X 10" = 1.14 X 10" N*s/m2 (or Pa*s)
NRe = DUp/p = (0.1539m X 3.0449 m/s

~ * s / m
= 410,642
~
4.1 X I O - ~
998.98 ~ ~ / m ~ ) / 1 X. 110"
4
The formula for the Reynolds number and the units must always be carefully checked. It is a good idea to do the computation twice with two
different formulae. It is also a good idea to check the dmensions. If it is
remembered that, in S1 units, force (newtons) is a derived quantity equal
and Equations
81
C H A P T E R
T W O
to mass times acceleration, then kg*m/s2 may be substituted for N and it

will be found all the units cance1.A pure number results.
...
Figures 1-4 and 11-1 give a value for the MoodyV1llfriction factor of
0.0165.The same value is obtained from either chart.The energy converted to internal energy can be computed from the lost mechanical
energy formula as found in 11-20.
To obtain the pressure drop, simply multiply the energy term by the density.
PI - P2 = 5.07 X 62.365 = 316.2 psf = 2.2 psi
(11-2 1)
Note in this case the loss term can be converted directly to pressure drop.
In most other cases the pipe diameter changes and the elevation changes.
These factors must be included in the ~ernoulli'equation. The advantage
of keeping the complete Bernoulli equation on hand is that details such
as this are not forgotten.
Example 11-2: Sudden contraction
Take the previous example (11-1) and attach a sudden contraction to the
six-inch pipe. The contraction is constituted of a flanged connection and
four-inch diameter, schedule 40 pipe downstream.What are the head
losses and what is the pressure drop across the flange?
If we look at the terms of the Bernoulli equation, we see there is no
pump and the pipe is still horizontal; so the elevation terms cancel.We
do, however, have a change in diameter.Therefore, there is a change in
velocity and the kinetic energy terms must be left in the equation.We are
still dealing with an incompressible fluid, so the density is constant.The
Bernoulli equation reduces to,
The upstream velocity is known from the previous problem. It is 9.99 fps.
The downstream velocity can be immediately computed from the continuity equation,
82
and Equations
C H A P T E R
T W O
We have simply replaced the area ratio by the square of the ratio of the
diameters.The beta ratio is in the denominator because the downstream
diameter is smaller.
The downstream velocity, U2, equals 9.99 fps divided by the square of
0.664, the beta ratio. It is 22.67 fps. It has gone up by a factor of 2.3
because the pipe diameter decreased by a factor of 1.5. Note this velocity
is much greater than what is normally considered reasonable, but it can
occur and, therefore, it will occur.
The downstream lunetic energy or velocity head term is,
The number of velocity heads lost is given by the loss coefficient, K,.
K, = 0.5(1-
p2)=
The permanent losses of mechanical energy are given by the product of

the loss coefficient and the velocity head based on the smaller pipe.The
number is 2.24 ft-lb,/lb,
Notice loss in mechanical energy from a single fitting, in tlxs case, is almost
half that from one hundred feet of straight, six-inch pipe that was previously computed as 5.07 fi-lbf /lb,. Tlxs is typical of fittings.The bulk of the
permanent losses in a piping system normally is due to fittings and bends.
To show pressure drop is different from mechanical energy loss, we w d
take the applicable form of the ~ernoulli'equation and transform it.
Sections 1 and 2 are at the ends of the length over which the permanent
losses were computed.
and Equations
83
C H A P T E R
T W O
l n c o m p r e ~ ~ i b l Fel u i d F l o w
(11-26)
P, - P, = 540.9 psf = 3.76 psi

The value of the irreversibilities (2.24 ft-lbf/lbm) times the density
divided by 144 is 0.97 psi.The two numbers are not the same and should
not be confused even when they are expressed in the same pressure units.
The difference between the pressure drop, 3.76 psi and the permanent
losses, 0.97 psi is equal to 2.79 psi.This number represents pressure drop
that is recoverable.The difference will become clear when we introduce
the energy line and hydraulic grade line graphs.
Example 11-3: Sudden expansion
Building on the previous example, we will add a sudden expansion by
flanging a six-inch schedule 40 pipe to the four-inch pipe.The flow rates
and velocities will be those of the previous example.
The expansion loss coefficient is computed from,
K, = (1 - p2),= (1 - 0 . 6 6 4 ~ =
) ~0.313
The mechanical energy head loss is,
The change in static pressure can be found by applying the ~ e r n o u l hequa~

tion over the sudden expansion.The elevations are the same and cancel.
P, -4 = -245.4 psf = -1.7 psi

84
and Equations
CHAPTER
T W O
Notice the negative sign.The downstream pressure is greater than the

upstream pressure by 1.7 psi! This result can be disconcerting to people
who are used to thinlung that flow is due to a pressure difference. Here,
we have met the phenomenon of pressure recovery. The pressure recovery
phenomenon arises frequently in piping systems (especially safety vent
headers), in orifice plate installations, and in valve bodies.
From the above equation pressure recovery can be seen to occur when
the downstream kinetic energy is much less than the upstream kinetic
energy and their difference numerically exceeds the mechanical energy
loss term. Obviously, the assumption that pressure difference is the only
source of motive force is wrong.The rate of change of momentum must
be considered, as is done by the application of the ~ernoulli'equation.
Example 11-4: Valves
A four-inch globe valve is placed in the four-inch line of the previous
example with the same flow rates. What is the equivalent L/D ratio and
how much equivalent additional pipe does the valve add?What are the
permanent mechanical energy losses due to the presence of the valve?
We will use two different methods to solve the problem.The first uses the
second uses a more accurate
general correlation from craneXiii.~he
method involving the valve C,, which is the manufacturer's valve flow
coefficient obtained from test data.
From Crane, on the assumption the seat diameter is the same as that of
the pipe, K equals 340fT.The constant, 340, is the equivalent L/D of a
standard globe valve with a beta ratio of one. It is obtained from Crane
Technical Paper 410. The factor, fT, is the friction factor at fully developed
turbulent flow in a four inch schedule 40 pipe. It can be obtained from a
convenient tabulation in the same publication or from the flat part of the
curve for four-inch schedule 40 pipe in Figure 11-2. More details on this
methodology will be given in Appendix AII.
The friction factor at complete turbulence is 0.017.Therefore,
K = 0.017(340) = 5.78.
The first part of the question is answered above.The equivalent L/D is
340.The equivalent length is this number multiplied by the pipe inside
diameter in feet. So, the equivalent length is 114 feet for one obstruction,
a globe valve.
and Equations
85
C H A P T E R T W O
The permanent mechanical energy losses are found by applying the

equation with the above friction factor and L/D ratio and the
rated flow through the line.
arc^^"
Using data from a specific valve, we can obtain a more accurate estimation.
The valve flow coefficient, C,, can be converted to a loss coefficient, K, by
formula.The derivation for the conversion wdl be given in Appendix AII.
The formula in U.S. customary units is,
The internal diameter, d, of a four-inch schedule 40 pipe is 4.026 inches.

If we have a four-inch Masoneilan globe valve, 20000 series,ANSI Class
150, full port, the catalogue data gives the CVas 195.
This more accurate value for K contrasts with 5.78 from the more general method.
Since K is equal to fM (L/D), the L/D ratio is obtained by dividing K by
the friction factor for fully developed flow in four inch schedule 40 pipe,
0.017.The ratio is 362.4.The equivalent length is 362.4(4.026/12) =
121.6 feet.This contrasts with 114 feet from the more general method.
The permanent mechanical energy losses are given by,
These irreversibilities contrast with 46.2 foot-pounds force per pound

mass computed by the more general method.
The pressure drop across the valve can be considered to be equivalent to the
permanent losses in thls case because the valve is horizontal and the
upstream and downstream diameters are equal. It is 49.2(62.365/144) = 21.3
psi.The less accurate method gives 20 psi. The pressure drop to the vena
contracts somewhere withm the valve body will be greater than 21.3 psi.
86
and Equations
Incompressible
Fluid Flow
Losses of M e c h a n i c a l Energy
C H A P T E R
T W O
It should be noted the vapor pressure of 60F water is approximately 0.26

psia. If the upstream pressure at the valve inlet is less than about 21 psia,
the vena contracta pressure will dip below the bubble point of the water.
The water wlll vaporise to some extent. With any pressure recovery in the
valve body, the vapor bubbles will collapse and cavitation can result.
Cavitation will be dealt with more thoroughly later in the book.
Example 11-5: System irreversibilities
Figure 11-8 represents a length of piping which contains a flowing fluid,
water.The conditions at section 1 are known. We are going to explore the
changes in velocity, pressure and specific energy as the flow passes through
the following: a sudden contraction, a fully open globe valve, a 90 degree
elbow, a vertical segment, another elbow, a horizontal segment, a sudden
expansion and another horizontal segment.
Datum = Grade
NTS
Figure 11-8. Example of system irreversibilities
For the sake of continuity, we have used the data from the previous examples, but we have changed the elevations in order to show their effect. In
addtion, a datum line has been added and the pressure at section 1 is
assumed to be 100 psig. We could have performed the computations using
gauge pressures since we are dealing with water, an incompressible fluid,
but we chose to use absolute pressures, for reasons given previously.
and Equations
87
C H A P T E R T W O
The figure is not to scale. Five lines are shown on the figure.They are
defined as follows:
1. The datum line. T h s is an arbitrarily established horizontal h e fkom
~i~
which all elevation changes will be measured. Since the ~ e r n o uequation
is established in terms of the Merences between two points, we can choose
any horizontal level we wish. It is usually best to choose the lowest elevation
in the flowing system in order to avoid d e h g with negative numbers.
Since the end points of the pipe are not shown, we have assumed, arbitrarily,
that the lowest elevation is 6 feet below the pipe at section 1.
2. The elevation line. T h s line is used to measure elevations above the
datum line for flow in the pipe. If we have a tank with a liquid surface, the
surface d often correspond to the elevation. It also corresponds to the
center line of the pipe.The line represents the potential energy of the fluid
due to its position above the datum.
3. The hydraulic grade line. This line represents the energy per unit mass
attributable to static pressure plus the potential energy.With liquids, it corresponds to the height the liquid would reach if a suitable manometer
could be attached at successive points. For gases, it simply represents a
computed value.
4. The total mechanical energy line. T h s h e represents the sum of the
static (pressure) energy, the potential (elevation) energy and the lunetic
energy. In the case of liquids flowing in ulvform diameter pipe, the h e is
above the hydrauhc grade h e by a constant factor equal to the velocity
head, the kinetic energy, in the pipe. In the case of gases, the line wdl deviate
f?om the hydrauhc grade h e as the velocity increases with expansion.
5. The total energy line. T h s line is a horizontal line parallel to the datum
line. It represents the fact that, by the first law of thermodynamics,total
(mechanical and internal) energy is conserved.The dfference between the
total energy line and the total mechanical energy line represents the irreversibhties. It is the mechanical energy converted to internal energy.
Notice it corresponds to the total mechanical energy line at section 1.We
are only concerned with dfferences. Irreversibhties occurring upstream of
the starting point of the analysis are not included in the solution to a particular problem.
O n the sketch are shown the differences between the various lines and
what these differences signif):The first difference is numerically (but not
dimensionally) equal to the elevation above the datum. The second is
equal to the absolute pressure divided by the density.The third is the
klnetic energy.The fourth difference represents irreversibilities.
88
and Equations
C H A P T E R T W O
Points to bear i n mind

The following points may not be obvious at first glance.At the sudden
expansion between sections E and F, there is some pressure recovery. This
is demonstrated in the computations and in the convergence of the
hydraulic grade line and the total mechanical energy line across the
expansion. The phenomenon is quite common, but is usually unexpected.
What is even less apparent is the same phenomenon occurs in the globe
valve. Since we adopted the short cut of using an equivalent K factor, it is
possible to lose sight of the fact a valve represents a tortuous path and a
restriction.There wdl be a vena contracta at which the pressure will be
much less than valve discharge pressure. Under circumstances of low inlet
pressures and fairly pure liquids with high vapor pressures, this low pressure
can cause cavitation, slugging and possibly damage to the valve and piping.
Plan of attack
For any problem solving exercise, it is best to have a plan of attack.This
problem is fairly straightfonvard.We are given the conditions at section 1
and we have to compute the conditions sequentially downstream.
The general approach, in this case, is to compute the total energy at the
starting section by summing all the terms on the left of the Bernoullil
equation. At the first section the total mechanical energy equals the total
energy.The mechanical energy losses for a length of pipe, a valve or a fitequation or by using the appropriate K
ting are computed by the
factor equation. Each loss is subtracted from the value of the total
mechanical energy of the previous segment to obtain the total mechanical
energy at the next section.
arc^^"
Given the irreversibilities along each length and across each valve and fitting, the total mechanical energy line can be plotted. Given the velocity
heads, the hydraulic grade line can be plotted below the total mechanical
energy line. Given the irreversibilities, velocity heads and elevations, the
Bernoulli equation may be used to estimate the downstream pressures.
Frequently, engineers will convert all valves and fittings and different sized
pipes to an equivalent (theoretical) length of constant diameter pipe and
then perform the estimations. This habit is satisfactory for most simple
cases for liquids although it is not accurate as has already been demonstrated. It cannot be used for gases and vapors. It also has the disadvantage
of causing a loss of insight into what is actually taking place.We will
and Equations
C H A P T E R
T W O
follow the more general approach, but we will convert the two elbows to
equivalent pipe lengths to show what is involved.
Table 11-3 gives the results of the computations at each section. Although
the computations are made in terms of energy per unit mass, gauge pressure is frequently of immediate interest. A column for gauge pressure has
been included in the table.
Most of the computations have already been given in previously.We will
consolidate the results here for convenience.
ondition a t section 1
We can immediately fill in the first column of Table 11-3 f?om the given data.
Elevation
Pressure
Flow
Temperature
Velocity
6 feet
100 psig, 114.7 psia
900 gpm, 450,150 lbm/h
60F
9.99 fps
At section 1:
Potential energy
Static energy
Kinetic energy
Irreversibilities
Total energy
Hydraulic grade line
6.00 ftolb, /lbm

114.7(144)/62.365 = 264.80 ftlb, /lbm
1-55 ftolb, /lbm
9.992/2(32.17) =
0.00
272.35 ftelb, /lb,
270.80 ftolb, /lbm
272.35 - 1.55 =
Pipe segment 1 to A, 50 feet of six inch, Sch. 40 clean

commercial pipe
This problem has already been solved in Example 11-1 for 100 feet of
pipe. Since we are dealing with horizontal pipe of uniform diameter, the
irreversibilities are exactly one half, 2.53 ftolb, /lb,
As a reminder, the procedure is to compute the Reynoldsv number based
on known viscosity, hameter and flow rate.The ~ o o friction
d ~ factor
~ is~
then obtained from charts or from the ~ h u r c h d l - ~ s arelationship.Thls
~i~l~
factor is used in the
equation to obtain the irreversibilities,h=.
arc^""
90
and Equations
Incompressible
Fluid Flow
C H A P T E R
T W O
The irreversibilities for liquids flowing in straight, horizontal pipe can be

converted directly to pressure difference by multiplying by the density.
They are 1.1 psi in this case.
Conditions at section A
At section A, the total mechanical energy line has dropped below the
total energy line by the amount of the irreversibilities, 2.53-ft*lbf /lbm.
The total mechanical energy is now 269.8 ft*lbf/lbm. The hydraulic grade
line is below this line by the lunetic energy in that 1ine.This was computed above as 1.55 ft*lbf/lbm.The fluid is incompressible and the pipe is
uniform, so the two lines are parallel. The elevation of the pipe stayed at 6
feet, so the potential energy is still 6 ft*lbf/lbm.
Conditions at section B
The irreversibilities across the sudden contraction were given in Example
11-2 as 2.24 h*lbf/lbm.This means the total mechanical energy line dips
by this much across the fitting.The hydraulic grade line dips below the
total mechanical energy line by the amount of the kinetic energy in the
smaller downstream pipe, 7.99 ft*lbf/lbm.
Pipe segment B to C, 10 feet of four inch, Sch. 40 clean
commercial pipe
For the four-inch pipe, the mass flow rate remains the same as for the sixinch pipe.The Reynoldsv number can be computed as,
(11-3 4)
N, = 6.31 4507150 =641,387=6.4x105
4.026 (1.1)
The friction factor is obtained from Figure 11-2. It is 0.017.
The velocity in the four-inch pipe has already been computed in
Example 11-2 as 22.67 fps.We can obtain the irreversibilities from the
equation, as,
arc^"^^
The total mechanical energy line wdl dip below its starting point at section B by this amount.The hydraulic grade line will be below this
inclined line by the amount of the velocity head in the four-inch line,
7.99 ft4bf /lbm.
and Equations
91
C H A P T E R
T W O
Globe valve
The irreversibilities across the valve were given in Example 11-4 as 49.2
ft*lbf/lb,. This is the amount by which the total mechanical energy line
dips between section C and section D. Normally the hydraulic grade line
is shown parallel to this total mechanical energy line by the amount of
the lunetic energy at the inlet and discharge of the valve.We have chosen
to represent the hydraulic grade line by a dashed line to show that somewhere in the body of the valve is a vena contracta. It will serve to remind
us to look for potential problems.
Pipe segment D to E
In thls segment we have two elbows and a change in elevation.We w d convert the elbows to equivalent lengths of straight pipe and then compute the
irreversibdities for the sum of the straight pipe and the equivalent lengths.
If we assume standard 90 degree elbows, from cranexiii,K equals 2(30)fT.
It is to be noted the friction factor is that at completely turbulent flow,
not the actual friction factor. From Crane, this is also 0.017 so, in our
case, they are the same. K = 1.02 for both elbows.The equivalent L/D
would be 60 X 4.026/12, which equals 20 feet.We must add 20 feet to
the 42 feet of straight pipe to obtain an equivalent length of 62 feet.
arc^""
To obtain the permanent losses, we can use the

equation, as
before. Alternatively, since we already have the irreversibilities for ten feet
of straight four-inch pipe, we can simply ratio the lengths and multiply by
the previously computed irreversibilities. This method gives
4.047 X 62/10 25.1 ft.lbf/lb,.
The total mechanical energy line will drop by this much to section E.
The hydraulic grade line will be below the total mechanical energy line
by the kinetic energy term in the four-inch line, 7.99 fi*lbf/lb,.
The potential energy line follows the elevation changes of the pipe.
The difference between the potential energy line and the hydraulic grade
line is equal to P/p. So the pressure in psia at section E can be found by
multiplying the diEerences by the density and dividing by 144.To obtain
the pressure seen by a pressure gauge, subtract the atmospheric pressure,
14.7 psia to obtain psig.
92
and Equations
Incompressible
Fluid Flow
Energy
C H A P T E R
T W O
Sudden expansion, E to F
Example 11-3 gave the irreversibilities across this segment E to F as being
2.5 ft*lbf/lb,.The pressure drop across this segment was given as minus
1.7 psi so we talk of pressure recovery.This is not only a theoretical consideration. Pressure readings taken across such an expansion wdl show a
difference in gauge pressures. The upstream pressure will be lower than
the downstream pressure.This phenomenon is not what we expect from
common sense alone. Common sense tells us that flow should be in the
direction of decreasing pressure, not increasing pressure.
Application of the ~ernoulli'equation f?om any point upstream to section

E and then to section F will confirm what the pressure gauge tells us.
The convergence of the total mechanical energy line and the hydraulic
grade line in the figure, not only shows the phenomenon, but also shows
its 1irnits.The two lines can only get as close as the amount of the kinetic
energy, the velocity head, downstream. In addition, although pressure has
been recovered, the sum of the total irreversibilities (cumulative irreversibilities) keeps on increasing in the direction of flow.These irreversibilities are represented by h=, the distance from the total energy line
to the mechanical energy line.
Pipe segment F to 2
This segment of pipe is identical to the segment from 1 to A.The only
difference is the elevation.The irreversibilities will be identical to those
from section 1 to A: 2.53 ft*lbfAb,. The total mechanical energy line will
dip by this much.The hydraulic grade line will be below this line by
exactly 1.55 ft*lbfAb,.
It is noteworthy the pressure at section 2 is higher than the pressure

upstream at section E.
All of these numbers are tabulated in Table 11-3.
and Equations
93
C H A P T E R
T W O
Table 11-3. Results of computations

Energy distribution pipe, 60F water, 900 gpm. Upstream pressure at Section 1 :l 14.7 psia
1
Section
Sch.40, in
Elevation, ft
Velocity, fps
P.E.ft-lbf /l b,
K.E. ft-bf /l b,
S.E. ft-l bf /l b,
Total M.E. ft-lbf /Ib,
M.E. losses, ft-l bf /I b,
Cumulative M.E.
losses, ft-l bf /Ib,
Total energy,
ft-lbf /l b,
P from ~ernoulli'
psia
Conditions at section 1
Density, Ib, /ft3
Mass flow, Ib,/h
62.365
114.7
I C
45015
Example 11-6: Restriction orifice

Figure 11-9 represents an extension to the system of Figure 11-8. It is a hypothetical, but fairly frequent, case. Suppose water is to be transferred from
tank, T1, to reactor, RI, by pressure Merence. The maximum flow rate must
be limited to 900 gpm (U.S.).The minimum flow rate is not important. We
d also ignore the need for check valves for the purpose of the exercise.
Figure 11-9. Extension to system irreversibilityexample
94
and Equations
Energy
C H A P T E R T W O
The maximum pressure in T1 is 98 psig.The minimum pressure in R 1 is

35 psig. These pressures have been observed on (calibrated) pressure gauges
on the two vessels. They correspond to the operating intent. The specific
gravity of the reactor fluid is very close to 1.0, so the density of the fluid
in the reactor wdl be assumed equal to that of the water, 62.365 lb,/ft3.
The problem is the operator has no means of knowing what the actual
flow rate is. He has the habit of opening the globe valve wide open.This
is detrimental to the reaction. How can we be sure to limit the maximum
flow rate to 900 gpm with a limited budget?
A restriction orifice could be of use if there were a convenient location

for it. It would not be wise to place the orifice directly at the valve or too
close to an upstream fitting.The disturbance to the flow profile caused by
the valve or fitting would cause an uncertainty in the estimation. Suppose
a flange is located in the vertical four-inch pipe about elevation 26 feet. If
a restriction orifice were placed here, the globe valve could be fully
opened as part of the operating procedure and the flow would never
exceed 900 gpm. How do we go about sizing the restriction orifice?
Plan o f attack
Use the ~ e r n o u hequation
'
between one unit mass of fluid at the surface
of the water in tank,Tl, and one unit mass of fluid at the surface of the
reaction mixture in reactor, R I .
Compute the system irreversibhties required to limit the flow rate to
900 gpm.
Compute the system irreversibilities available without the restriction
orifice.
Subtract the irreversibhties available from the irreversibhties required.
T h s gives the irreversibhties the restriction ordice must create.
Use the orifice loss formula to compute beta iteratively.
Multiply beta by the pipe internal hameter to obtain the required
hameter of the orifice.
and Equations
C H A P T E R T W O
Required irreversibilities to l i m i t the maximum flow to 900 gpm
Customary U.S. units have been used.The velocities at the surfaces of the
liquids are zero.The factor, g/g,, equals one numerically, but is left in as a
reminder. The units of hL are ft-lbf/lb, in this case.
Computation of available irreversibilities

IRREVERSIBILITIES FROM A PIPE ENTRANCE FROM TANK T1
K = 0.5
(11-3 7)
The value of K is that value for a square edged entrance. See Appendx AI1
for details.
IRREVERSIBILITIES FROM THE PIPE EXIT T O REACTOR, R 1
K = 1.0
(11-3 8)
The value of K is that value for a square edged exit. See Appendix AII.
AVAILABLE SYSTEM IRREVERSIBILITIES
The last number on the right is fiom the previous computations. It is the
cumulative irreversibhties at section 2 of the previous system. See Table 11-3.
96
and Equations
lncomprassible
Fluid Flow
C H A P T E R
T W O
IRREVERSIBILITIES REQUIRED OF THE RESTRICTION

ORIFICE
(11-40)
121.5- 90.47 = 31.03
These mechanical energy losses have units of ft-lbf /lbm.The irreversibilities to be created by the restriction orifice must equal the total required
irreversibilities less the total available irreversibilities.
Restriction orifice sizing
The last number computed, 31.03 ft-lbf /lbm, can be used in the restriction
orifice K factor equation and a value for beta may be found by iteration, as
follows:
(11-4 1)
U:,
h, = K o 2gc
Iteration
1
1 2
I 3
1 4
Trial Beta
1
1
1
0.5
0.6
0.559
0.5467
Computed Loss Value
1
1
1
11.25
4.30
6.35
7.148
TOO HIGH
I TOO LOW
I
I TOO LOW
I
I CLOSE ENOUGH I
To obtain the bore dimension, the internal dameter, 4.025 inches, is multiplied by the d/D (beta) ratio, 0.5467, to obtain 2.200 inches (to the
nearest thousandth of an inch).
If an orifice is bored to this dimension and to this tolerance and it is
placed in the flange, the flow will be limited to 900 gpm under the given
specifications.
and Equations
C H A P T E R T W O

This chapter has taken the theoretical concepts of Chapter I and
Appendix AI and has given them practical form. In particular, the problems arising from the use of loose terminology have been addressed and
the author has had the opportunity to edit some of his own loose terminology. The reader should now have obtained some insight into the conversion of mechanical energy into less useful internal energy. He should
be able to tackle with some confidence the problems associated with
incompressible flow through pipes when fluid machinery (pumps and turbines) is not considered.
The chapter has:
analyzed the Bernouhl equation in some detd. It has dxcussed the
importance of its terms one by one.The importance of the last term
dealng with permanent losses was dscussed in detad;
described the main causes of irreversibhties in piping systems;
discussed geometric similarity;
compared the
arc^^ equation with the generalzed K factor equation;
given a general equation for permanent losses across square edged, concentric orfice plates;
dscussed the use of grade lines: total energy, mechanical energy and
hydraulic;
given examples of loss computations for the more common cases;
presented a plan of attack to be applied to systems.
Appendix AI1 will give additional related information. In particular, it will
cover :
the relationshp between energy/mass units and head units;
the ~hurchdl-usagiAfiiction factor equation;
manifold flows (combining and dvidmg flows);
permanent losses due to various instruments.
98
and Equations
Pumps:
Theory and Equations
- PUMPS AND
THEIR PERFORMANCE CAPABILITIES
Chapter 111's intent is to provide enough information about pumps to permit

the reader to understand their performance capabhties and their interactions
with the other components of a complete hydraulic system.Very little mathematical description is given in ths chapter.A few worked examples of head
computations may be found in Section 111-9. Mathematical detd can be
found in A p p e n h AIII. Mechanical design considerations, such as flange and
casing ratings, will not be considered unless the practical u&ty of the information to the reader might not seem obvious.
There are many excellent texts that deal with pump and turbine engineering, and there are many texts, such as Chemical Engineers' Handbook, and
Unit Operations, intended to help chemical engineers specify fluid
machnery. This chapter and its associated appendm are intended for those
seelung an understanding of pumps because they must work with systems
that include such fluid machinery withn a larger system.The latter personnel normally do not specify fluid machnery. They are, however, faced
with the task of deahng with the characteristics of such machinery and
controhng flow through it, or measuring that flow.The related task of
machnery monitoring and control is beyond the scope of ths book. It will
only be mentioned superficially, where it seems appropriate to the author.
Since different types of pumps have different installed characteristics and
different control requirements, a general classification will be given. The
classification will be used as the framework for further discussion.
More time will be spent on centrifugal pumps because that type of pump
is most common in industrial practice. Their connection in parallel and
series configurations will be considered.Turbines will be discussed incidentally to form a more complete picture of fluid machinery.
and Equations
C H A P T E R
T H R E E
Pumps
- Theory and Equations
111-2: FUNCTIONS OF P U M P S
Pumps move liquids through a system of conduits and equipment by
increasing the total mechanical energy of the liquid at the pump location.
Sometimes the intent is to move a liquid from one place to another transportation of the liquid. Sometimes it is simply to increase the pressure so the higher pressure can be made use of in a hydraulic machine.
Sometimes the intent is to produce kinetic energy in a stream of liquid water issuing from a fire hose, for example.
Liquid moves through conduits when the total mechanical energy
upstream is greater than the total mechanical energy downstream. Note
thermal energy does not contribute to bulk motion of a fluid, only to
molecular motion. In other words, liquid moves through a conduit from
A to B when the sum total of static, kinetic and potential energies is
greater at A than at B. If the total mechanical energies are equal, there is
no bulk movement. If the total mechanical energy at B is greater than at
A, the flow wdl be in the opposite direction, from B to A (if there are no
check valves). In a flowing system without deliberate thermal input or
extraction, the difference between the total mechanical energies at A and
B is the amount of mechanical energy converted to internal energy, and
ultimately to heat energy flow, by viscous drag, turbulence and shock.
This difference is what is termed "irreversibility" in this book.
To get liquid to move through a conduit (if we discount buoyancy due
to temperature differences), mechanical energy must be transferred to the
1iquid.There must also be some physical control over the direction of
liquid flow. Increasing the energy level at a point does no good if the
flow is allowed to go in all directions.
Means of energy transfer to liquids
The common means of transferring mechanical energy to a liquid, &Scounting buoyancy, and of giving the liquid a direction in which to flow
are by the use of:
gravity;
centrlfUga.1 force;
volumetric hsplacement;
momentum transfer;
mechanical impulse;
electromagnetic force.
100
and Equations
Pumps
C H A P T E R T H R E E
Energy transfer by gravity

Gravity is the natural means of moving liquid from a high point to a low
point - in conduit or in open channels. A head tank is a common means
of maintaining a constant pressure when flow control is critical. Head
tanks and water towers are also used to maintain reserve capacity for peak
periods and to permit smaller, less expensive pumps to be used for the
service.The direction of flow is governed by the direction of the applied
force, gravity, and the restraining walls of the conduit.The pressure at a
point is a function of the elevation hfference, the hydrostatic head, and
the fluid density.
The characteristics of transfer by gravity are:
Gravity is free, unless the liquid has to be pumped to a head tank
before gravity can be used;
in closed conduits, flow control, including shutoff, is easy;
the pressure is limited by the avadable hydrostatic head. If the conduit size
is fixed, a booster pump may be necessary to obtain a desired pressure or
flow rate.
Energy transfer by centrifugal force
Centrifugal force is a quite sophisticated way of transferring mechanical
energy to liquids (and gases). By performing work on a shaft (by muscle
power, water power, wind power or electrical power), the shaft can be
made to turn. A suitable impeller causes liquid to be forced to travel radially while it gathers velocity. The mechanical energy input to the shaft
has been transformed to hnetic energy. Some of the mechanical energy
will be transformed to thermal (internal) energy and will be lost to practical use. This transformation is a mechanical ineaciency.
If sufficient suction pressure is maintained, the liquid flow wdl be continuous. Liquid wdl move into the eye of the impeller at the same rate it is
propelled to the exterior. Once the liquid reaches the tip of the impeller,
its l n e t i c energy must be transformed into static (pressure) energy in such
a fashion as to minimize conversion of mechanical energy to internal
energy.This is usually done by carefully guiding the liquid with diffusers
or vanes.The purpose of the guidance is to smoothly slow down the
stream and to allow the lunetic energy to be converted to pressure energy.
The physical limitations imposed upon the flow rates of centrifugal
pumps are related to maximum tip speed of the impeller. A given
and Equations
101
Pumps
irnpeller design is limited to a maximum tip speed (to avoid cavitation).A

single stage is, therefore, limited to a maximum dfferential pressure. Tvical
designs cover the range fiom 70 to 100 fi-lbf/lbm (30 to 43 psi or 207 to
296 kPa with water). By carefully guiding the liquid leaving one stage, and
by carefully increasing the channel darneter, the liquid can be made to
enter another stage at its original velocity but at a hgher pressure and a
hgher total energy. Addtional mechanical energy can be added in a second,
a thrd or even a seventh stage. It should be obvious the overall mechanical
efficiency dminishes with an increased number of stages as some of the
impeller shaft energy is converted to internal energy at each stage.
Characteristics of centrifugal pumps
Centrifugal pumps have the following general characteristics:
they are relatively pulsation 6ee;
their mechanical design allows hgher capacity at (generally) lower heads
than positive dsplacement pumps;
they have fairly efficient performance over a wide range of pressures and
capacities (at least, single stage pumps do);
control can be by dscharge throtthng when consideration is given to
overheating and surging at very low flow rates;
dscharge pressure is a function of density at fixed angular velocities
(rotational speeds);
they are capable of h g h rotational speeds (up to the speed, whch produces
cavitation);
they can be made to handle slurries and corrosive or erosive liquids by
suitable coating or alloys;
they can be relatively inexpensive.
Energy transfer by volumetric displacement
Volumetric displacement was the first method invented for pumping
1iquid.The flow direction was controlled by the use of check valves or by
automatically operated valves. Volume can be displaced by mechanical
means, as by a piston, or by another fluid directly or indirectly (behind a
diaphragm), or by the same fluid directly, but in a different phase.
Some condensate pumps use direct steam as the motive force. Acid eggs
use air. In all of these devices, energy transfer is by the movement of a
force through a distance. It is easily recognizable as work.
102
and Equations
Pumps
C H A P T E R
T H R E E
With rotary pumps, a volume of liquid is trapped by a rotating member

(rotating vane or gear pumps). The design produces relative smooth, continuous flow.
Some of the general characteristics of pumps using the volumetric displacement principle are:
flow is often pulsating, requiring some external damping such as by a
pulsation-damping bottle (in both the suction and the discharge piping);
flow cannot be throttled at the discharge without stalling the driver,
bursting the equipment, relieving a pressure safety valve, or actuating a
diversion valve;
flow can be varied by varying the speed of the driver or by diverting
flow to the source or elsewhere;
volumetric pumps are often associated with low capacity, high-pressure
service;
positive displacement pumps can be designed for extremely low
flow rates.
In spite of the generalities just mentioned, positive displacement pumps
are used for "frac" jobs.These pumps pump sand slurries and emulsions at
high pressures (5,000 psig) into oil bearing formations for purposes of oil
well stimulation.
Energy transfer by momentum transfer

Jets and eductors are examples of momentum transfer devices.The deceleration of a motive fluid causes acceleration of a pumped fluid.
ne cnaracteristics or jets are:
they can pump liquids and gases;
they are usually very inexpensive to purchase;
they are usually expensive to operate as they are relatively inefficient;
being small, they can be made of exotic materials without incurring too
much first cost;
the pumped stream can contain solids.
That being said, many evaporators or crystalhzers could not be without
them.They do have characteristics that cannot be reproduced readily by
centrifugal pumps - the ability to produce vacuum, low pressure opera-
and Equations
103
C H A P T E R
T H R E E
Pumps
tion and the ease with which they handle solids and corrosive or erosive
fluids.
Energy transfer by mechanical impulse

Regenerative pumps (also called turbine pumps, but they are different
from a vertical turbine pump) use the principle of imparting a mechanical impulse in an axial as opposed to a centrifugal direction.
Energy transfer by electromagnetic force
If the fluid is a conductor, such as a molten metal, it can be made to
follow a moving electromagnetic field. This technique is highly specialized. It is used in nuclear reactors to pump heat transfer fluids.
111-3: A B R I E F H I S T O R Y O F PUMPS
Man has always had a need to raise liquids (water) through heights. If we
consider a bucket or a shn bag dropped into a well on the end of a rope as
a pump, then it is difficult to give credt to the inventor of the first pump.
The ancient Greeks had a fairly well developed hydraulic technology. They
used water power to open temple doors, for instance. Heroxx,in the second
century BC, developed a fire pump with two cylinders and a hand-operated roclung beam to supply motive force.The early pumps were all handoperated, obviously, but it did not take long for people to reahze donkeys,
camels, wind and water could supply the necessary input power.
One ingenious example of early technology is the use of a hollow rod
(bamboo?) into which was threaded a rope.The rope was connected at its
ends to make a continuous length. Rags were tied to the rope at convenient 1engths.The hollow rod was placed vertically in a well and fixed in
place.The rope was then pulled through it. Each section of the rope
between the rags would trap a quantity of water that could be made to
flow at the well head in a more-or-less continuous manner. This was a
positive displacement pump.
The Greeks used piston and cylinder arrangements, made of wood, to
force water from wells.The motivation for the development of the early
pumps was the need to supply water for drinking, irrigation and laundry.
The Romans took over and developed Greek technology.They developed
bronze pumps. All of the early pumps were positive displacement pumps.
104
low o f ~ n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
In the 16th century the German Agricolam described the extensive use
of pumps in the mining industry for mine dewatering purposes.This was
probably the first large-scale industrial use of pumps. He also described
the coupling of multiple pumps to a water wheel.This represented a
jump in technology from the use of muscle power (human or animal) to
the use of an external force (gravity).
In the Middle Ages, bilge pumps were used aboard ships.The cylinders
were often square in section - a fact we tend to forget.Another fact we
tend to forget is pumps were frequently used in a suction mode.That is,
suitable check valves allowed water to enter the cylinder on the outward
stroke, but not on the inward stroke.Water was allowed to leave on the
inward stroke either via another check valve and discharge port or simply
by leaking past the cylinder packing. In this case, the packing was devised
to fold in one direction (chevron packing).The physical limitation on
suction lift, about 33 feet or 10 meters for water was soon discovered but it was not explained too well.
With the industrial revolution, the increase in industrial activity created a
demand for pumps to move water from ever deeper coal mines and to
supply motive power to hydraulic rams. One of the first uses of the steam
engine was as the motive power to dewatering pumps.The invention of
the centrifugal pump is generally credited to Denis papin-', a French
inventor, in 1689. Its use had spread around the world by the mid 1800's.
Today, the centrifugal pump is an omnipresent part of our existence.
111-4: C L A S S I F I C A T I O N O F P U M P S
Although positive displacement pumps were the first to be used historically, in present day use, centrifugal pumps are more prevalent.Therefore
these pumps are listed first.
A lot of engineering thought has gone into developing reliable pumps
for different services.The purpose of classification is to group common
characteristics so these characteristics can be discussed with some
economy of effort. The following grouping is to allow people interested
in fluid systems involving pumps to seize the essentials - it is not meant
to be all-inclusive.
and Equations
105
C H A P T E R
T H R E E
Pumps
1. Centrifugal Pump5
Single stage pumps
a. R a d d pumps. (See Figure 111-l).These are true centrifugal pumps.
Subcategories are volute, ddkser, and turbine pumps.
b. A x d pumps. (See Figure 111-2).These pumps have a lifting or propeller

action.Ths is not really a centrhgal pump in the strictest sense of the
word "centrifugal". The liquid does not flee the center.
c. Mured flow pumps.These pumps combine both radd and axial actions.
These pumps are usually vertical and are also called "turbine" pumps but
should not be confused with the pump that combines mechanical impulse
with centrifugal force and that is called a "regenerative" pump.
Figure 111-1. Simple centrifugal pump
Figure 111-2. Axial-flow elbow-type propeller pump (Courtesy ofLawrenere pump,
Inc.)
Multistage pumps (See Figure 111-3)

These pumps overcome the tip velocity limitations of a single stage by
combining more than one stage.They can have seven or more stages on
the same shaft.
106
~ I o wo f ~ n d u s t r i a l~ I u i d s - ~ h e o r y a n d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
2. Positive displacement pumps

a. Piston and cylinder pumps.These pumps can be dnect acting, steamdriven pumps.They are classdied further as simplex, duplex, single or
double acting.
F!
b. Rotating member pumps.Ths is not really a single category. It includes

cam and piston, gear, lobular, screw, vane and peristaltic pumps.
c. Diaphragm, metering pumps. Again, ths is not a single category of pump.
.$ .$
Figure 111-3. Seven-stage diffuser-type pump
Figure 111-4. Simplified sketch of an air lift, showing submergence and

total head
and Equations
107
V1
V1
C H A P T E R
T H R E E
Pumps
3. Miscellaneous pumping devices

a. Gas, or air, lifts. (See Figure 111-4)
b.Jet pumps. (See Figure 111-5) Jet pumps are sometimes called ejectors. An
ejector is a jet device that uses a motive fluid at h g h pressure to entrain
another fluid at low pressure. It dscharges the mixed fluids at an intermedate pressure. Ejectors can use liquids or gases as motive fluids and can
pump (compress and move) liquids or gases. Figure 111-5 from Schutte
and K o e r t i n F demonstrates the large variety and the many names
given these pumps.
c. Liquid metal pumps. These pumps make use of the motor principle to
move molten metal heat exchange mecha through piping. If a magnetic
field and an electric current are mutually perpendcular and each is perpenchcular to the axis of the pipe, the molten metal will be subject to an
axial force that wdl propel it in the chrection of the force and the pipe.
The grouping is rather arbitrary as can be seen by the inclusion of axial
pumps with radial pumps in the "centrifugal" category.
Eductor
S
t
Exhauster
Steam Jet
Syphon
Water Jet
Exhauster
Steam Jet
Blower
Water Jet
Exhauster
Fume
Scrubber
ThermoCompressor
Figure 111-5. Ejectors
Notes: Eight ejectors with different designs and different names

The fume scrubber is a water jet air pump for low dfferential pressures
and large capacities.
The water jet eductor is a liquid jet pump for liquids.
The water jet exhauster is a liquid jet pump for gases.
The steam jet siphon is a steam jet pump for liquids.
The steam jet exhauster (also called a vacuum pump) is a gas jet gas
pump (two shown).
The steam jet blower is a steam jet pump for air for low differential
pressures and large capacities.
The thermocompressor is a steam jet steam compressor.
108
~ I o wo f ~ n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
111-5: C H A R A C T E R I S T I C S O F P U M P S
In this section we will generally describe the pumps classified above and
will briefly describe their characteristics.Again, the emphasis will be on
characteristics from the point of view of understanding their interaction
with the flowing system. Pump selection is the job of the specialist.
Nevertheless, a more technical description will be given in Appendix AIII.
By far, the most common pump application is one involving some type
of single stage centrifugal pump. This fact is due to their relatively simple
construction, low first cost and ease of maintenance.They can handle a
wide variety of corrosive and erosive materials (by proper choice of coatings or alloys).They are offered in many different sizes and operating
speeds in off-the-shelf designs. Generally, a single stage centrifugal pump
is the first choice unless some process requirement demands otherwise.
The engineer or technician interested in a measurement and control
problem is usually faced with an already designed system. His or her
problem is not to select a pump for the application. Rather, it is to make
it work efficiently within the system. He or she must still understand the
basic principles behind the selection and must have an overall grasp of
the operating characteristics of various pumps.
1. General characteristics of centrifugal pumps
Centrifugal pumps are usually specified on the basis of volumetric flow
rate in cubic meters per hour or U.S. gallons per minute.The differential
pressure across the pump is usually given in kilopascals or pounds per
square inch. Pressure is usually converted to "head" in meters or feet of
the fluid to be pumped.This "head" is the number of meters or feet of
fluid equal to the measured differential pressure plus the differential
kmetic energy.
A given impeller rotating at a given speed in a given casing will produce

a given head, independently of the fluid involved.This will be shown
mathematically in Appendix AIII. The actual differential pressure has to
be computed from the head and the known density.This permits the
issuance of general performance curves without specifying fluids.
The use of the concept of head has already been discussed in Chapter I,
Section 10. It is much easier to keep energy relations straight if we think
in terms of energy per unit mass and not "headV.This applies in partic-
and Equations
109
C H A P T E R
T H R E E
Pumps
ular when the term "head" is not clearly defined. However, the use of the
term is so entrenched in engineering documents, such as pump calculation sheets, that it must be clearly understood. It is worth taking the time
to establish which head is being discussed.The term should be defined
clearly on any document employing it. We wdl give examples of the use
of the head concept in Appendix AIII.
Flow control is frequently by throttling the discharge of a centrifugal
pump (not the suction, because of the risk of lowering the suction pressure below the NPSH requirements). Many centrifugal pumps are capable
of wide turndowns, but it is wise not to assume that all of them are.
Consult the pump expert when in doubt.
Priming centrifugal pumps operating in a suction mode

Liquid pumps are designed to operate on liquids, not gases.To fix ideas, let
us refer back to the statement that the total dynamic head of a single stage
is independent of the fluid being pumped.The average maximum head of
a single stage is about 85 ft-lbf /lbm in customary U.S. units. Multiplying
the head by the density will give the pressure in pounds-force per square
foot. Dividing by 144 will give the pressure in pounds-force per square
inch.The differential pressure equivalent to the total dynamic head of
60F water will be 85 X 62.34/144 = 36.8 pounds per square inch.This
will not be exactly the pressure read on a differential gauge across the
same pump, although it wdl not be far from it.The difference between the
two wdl be due to hnetic energy and elevation differences.
As long as water is available to its suction, such a pump can more than
adequately cause a negative pressure so as to ensure it remains flooded at
all times. In other words, if the pump discharge is at atmospheric pressure, the pump can cause a negative pressure at the suction that will
ensure water flow from a receiver that is also at atmospheric pressure,
provided the pump is initially primed.
If the same pump is started with 60F air in its casing and in the suction,
the merentid pressure equivalent to the same 85 fi-lbf /lbm would be
85 X 0.0764/144 = 0.045 pounds per square inch or 1.2 inches of water
column or about 32 mm of water column. By lowering the pressure in the suction piping, the pump might be able to raise water to this height, but no more.
The term "prime" refers to ensuring the presence of liquid within the
pump by means of pressure or vacuum. Pumps located above their
110
and Equations
Pumps
Theory and
Equations
C H A P T E R
T H R E E
source can be primed by using a source of vacuum (an eductor, for

example) on the pump casing in order to raise (usually,non-hazardous or
non-corrosive liquid) from the normal source. If foot valves are present,
pumps can be primed from another source while being vented.
Characteristic curves
The characteristic curve of a centrifugal pump is a useful and enlightening
tool - Figure 111-6 is an example. The figure has capacity in gallons per
minute or it may have some other volumetric rate as the abscissa and a
variety of variables on the ordinate. Units used on the ordinate can be
head, efficiency, brake horsepower or shaft power and net positive suction
head required. All of these terms wlll be defined more carefully shortly.
The principal curves of interest for the present are the head-capacity
curve and the efficiency capacity curve.The head curve shown is relative
smooth and it decreases continuously from shutoff to the maximum flow
rate.The particular efficiency curve shown starts at zero, curves up toward
70% efficiency and then dips back down rather sharply.The head curve
represents the energy per unit mass the pump is capable of transferring to
the fluid at a given volumetric flow rate.
Figure Ill-6.Typical head-capacitycurves for centrifugal pumps
The shape of the efficiency curve explains why, even though a centrifugal
pump is capable of wide turndown, it is not necessarily economical to
operate it for long periods of time away from the optimum point. Some
pumps, notably regenerative turbine pumps, are designed for low flows
with high heads per stage.The efficiency curve of regenerative pumps is
very peaked, so the useful operating range is even more restricted.
and Equations
111
C H A P T E R
T H R E E
Pumps
A plot of average design efficiencies of commercially available single-stage

pumps versus their nominal capacities will show low volumetric flow
pumps are generally less efficient than the larger pumps. The smaller
pumps have efficiencies between 10% and 37% at 20 gpm.The efficiencies increase with capacity to lie between 78% and 86% for larger pumps
at 10,000 gpm.The efficiency curve can be used to compute the extra
energy per unit mass that must be transferred to the shaft by the driver
over and above that transferred to the fluid. Dividing the head at a given
flow rate by the fractional efficiency at the same flow rate gives the
energy the driver must supply to the pump.The difference between the
two energy numbers represents the irreversibilities (within the pump
only) - the amount of mechanical energy converted to internal energy
and then to heat energy flow.
Another factor to consider is that, frequently, the head curve also has a
hump near the shutoff point. This characteristic is frequently associated
with a steep head curve and a high-speed machine.The characteristic is
similar to that of a centrifugal compressor.There are two equilibrium
points, two flow rates, for the same discharge pressure. Surging can result
at low turndowns as the pump "hunts" between the two points.
A third factor is that, even though a centrifugal pump might accept

shutoff conditions without surging, the low efficiency close to shutoff
means probable overheating of the fluid.When overheating or surging
may be a problem, minimum flow control systems that recirculate flow to
suction, sometimes via a cooler, are used.
The brake horsepower curve is simply the power, expressed in equivalent
horsepower, that w d have to be applied to the shaft to maintain a given
flow rate and head.The horsepower requirement increases with increasing
flow rate (and dminishmg head).A portion of the shaft power will be transmitted to the fluid as fluid power.The rest will heat up the fluid, any
cooling or flushmg liquid and, ultimately, the environment.The power transmitted to the fluid is equal to the shaft power ("brake" power) times the
pump efficiency.The pump has its own efficiency rating whch is separate
and hstinct fiom that of the driver.The two efficiencies (driver and pump)
must be multiplied to obtain the overall efficiency of the pump and driver.
The net positive suction head (required), NPSH(R), curve is of interest
mainly to the person who specifies the pump, but it is also important
when troubleshooting. In fact, if the pump is not operated within the
112
and Equations
Pumps
C H A P T E R
T H R E E
limitations of the minimum NPSH required, the pump may self-destruct.

The net positive suction head required is established by the pump
vendor. Basically it is the minimum head in excess of the vapor pressure
that must be available at the suction flange to prevent the pump from
cavitating. It is to be noted this required head increases with increasing
flow rate. NPSH(A) is a term used for a computed head available to the
pump.This head is dependent on the piping configuration and layout. It
must be greater than NPSH(R) over the total turndown flow range of
the pump. More details will be given later in this chapter and in
Appendix AIII.
Measures of performance of real centrifugal pumps

It is common to describe the performance of a centrifugal pump in
terms of its "capacity" and its "head". Both of these terms are rather
loosely used and they should be defined each time they are used.The
conversions between head and pressure in both S1 and customary U.S.
systems were introduced in Chapter I, Section 1-10.
Capacity can mean mass flow per unit time or it can mean volumetric
flow per unit time, both in various units. Head can mean differential
pressure, differential energy per unit mass, discharge pressure or the
height in feet or meters a column of liquid would be raised if it were
subject to an equivalent pressure at the base (but which pressure?).We
also talk of velocity head, static head, total head, and so on. Given the
potential for misinterpretation, it behoves all of us to pay particular attention to definitions.
Capacity is easily dealt with. It is flow per unit time. It can be mass flow or
volumetric flow.The mass or volumetric units and the time units should be
clearly stated so they can be translated to appropriate units by the person
using the information. Difficulties w i t h the English speaking world come
when the volume units are gallons and it is not specified whether U.S. gallons or imperial gallons are being dscussed. Imperial gallons are only about
20% larger than U.S. gallons!
Head is not so easily dealt with given the general looseness with which
the term is used. We shall begin by given the definitions of various pump
"headsV.The reader is advised to pay attention and to learn to discriminate amongst them. In addition, words like "total" and "absolute" and
"perfect" are used rather loosely in engineering. They can be misleading.
and Equations
113
C H A P T E R
T H R E E
Pumps
Fundamental definition of head

Head ultimately means energy per unit mass at a section of conduit.The
fundamental units are foot-pounds force per pound mass or newtonmeters per kilogram (joules per kilogram). Because it is not easy to
measure energy directly, units such as feet of fluid and meters of fluid
have sprung up as a substitute measure. However, ultimately in the
~ernoulli'equation and in other energy balances all of these convenience
units must be translated into energy per unit mass.
It is wise to fix the concept of head as being one or more of the different
forms of mechanical energy per unit mass of a fluid available at a section
of conduit. Computed at a cross section of conduit, it will consist of
three terms: static (pressure) energy, kinetic (velocity) energy and potential (by virtue of elevation) energy. Their sum is usually given the name
"total" head. Each of the individual terms is also called a head. In addition, each of the terms used to compute the fundamental quantities in
the head equation is also called "head": static discharge, static suction, etc.
It is worthwhile spending a little time fixing the concepts.
Total discharge head
The total discharge head, hd, is the gauge reading at the discharge flange
"corrected to the pump center line" (or to the eye of the inlet impeller in
vertical pumps), plus the barometer reading, plus the velocity head. All of
the individual terms must be summed, so they must have identical units.
In hydraulics work, the units are usually feet or meters of fluid. Note the
gauge reading could be negative and the sign must be obeyed.We will
use energy per unit mass units for computations and convert to feet or
meters where necessary.
Explanation of the term 'corrected to the pump center line'
The term, "corrected to the pump center line", can be misleading, but it is
commonly used, so an explanation is warranted. When sizing pumps, and
frequently when analyzing their pedormance, the pump center line is
chosen for the datum p1ane.The elevations of interest for potential and
static energy measurement are the center lines of the inlet and dxcharge
piping of the pump. Pressure gauges are frequently located at a different
level than the center lines. Therefore, if we apply the pressure gauge
reading in the appropriate units directly into the Bernoulli equation, we
wdl be in error.
114
~ I o wo f ~ n d u s t r i a~ ~I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
The correction is quite simple. If the gauge were on the center line of the
pipe, the rea&ng plus the atmospheric pressure would be the correct one
to use. If the gauge were higher, the pressure reading would be less and, if
it were lower, the reading would be greater. The error would be dn-ectly
proportional to the distance between the gauge center line and the pipe
center line.
Since, under steady state conhtions, the total discharge head and the total
suction head are fixed, we can correct for a Merent location of the pressure
gauge by changing the potential head (the elevation) by the negative of the
error in the static head (pressure).T h s is done automatically if the elevation,
Z, we use in the ~ e r n o u l hequation
~
is taken, not at the pipe center line, but
at the gauge center line. Doing ths automatically corrects the total head to
the pipe center line.What was lost in pressure energy is gained in potential
energy.This statement applies only to an isolated pump.When performing
computations on systems, the appropriate elevation must be used.
Since the velocity term is fixed, if we were to raise or lower the gauge to
the center line of the pump, we would have the same total energy as long
as we took the elevation term at the gauge center line. So, it makes no
difference where the gauge is as long as the elevation term is changed to
match the location of the gauge.The term, "corrected to the pump
center line", arises out of this correspondence. It would be better to state
that the reachng is corrected to the pipe center line.The three basic
energy terms are stdl present, but one of them is hidden as a "correction
to the pump center line".
Total suction head
The total suction head, hs, is the gauge readmg at the suction flange corrected to the pump center line (or to the eye of the inlet impeller in vertical
pumps) plus the barometer readmg plus the velocity head. All of the in&vidual terms must be summed, so they must have identical units. In
hydrauhcs work, the units are usually feet or meters of fluid. Note the gauge
readmg could be, and frequently is, negative and the sign must be obeyed.
Total dynamic head
The total dynamic head is the difference between the total discharge
head and the total suction head, H = hd - hs.This is a potential source of
error.We have defined the discharge and suction head terms as the sums
of the absolute pressure heads, the velocity heads and the potential heads.
We have explained in Chapter I, Section 1-10 the conversion between
and Equations
115
C H A P T E R
T H R E E
Pumps
pressure and head. If the reader is not a hydraulician and if he or she does
hydraulic computations infrequently, it is best to use energy per unit mass
units - they are easier to handle logically.
The total dynamic head is important in computing the energy requirements of a pump. It represents the total mechanical energy transmitted to
the fluid (not the pump) per unit mass of flowing fluid. Mechanical
energy converted to thermal energy is not directly considered. It appears
in the efficiency computation.
Static discharge head
The static discharge head, hsd,is the vertical distance to the free surface of
the liquid in the discharge receiver or in the discharge piping system plus the
absolute pressure on this liquid surface (gauge plus atmospheric) in consistent units. Thls head, hsd,is also corrected to the pump center line.
We have emphasised the fact that the static discharge head could be
measured within the piping because this is exactly what occurs when the
pump is started. It is also what occurs when the pipe discharges to the
top of the vapor space of a receiver.The static discharge head varies from
a minimum to a maximum and then, usually, drops to a lower 1evel.This
phenomenon occurs because the piping usually runs overhead to the
receiver. Once the pipe is filled, the siphon effect reduces the head
requirements of the pump. If a pump is sized based only on the receiver
elevation, without considering the maximum piping elevation, it may not
be able to get liquid to the receiver. Such a pump could operate if the
pipe were filled from a different source. It is always wise to check the
actual routing of piping and its maximum elevation.
Static suction head
The static suction head, h S ,is the vertical distance to the free surface of
the liquid source plus the absolute (gauge plus atmospheric) pressure corrected to the pump center 1ine.M terms in the sum must be converted
to the same units.
Total static head
The total static head, h,, is defined as the difference between the static
discharge head and the static suction heads, h, = hsd - hss.
116
and Equations
Pumps
Theory and
Equations
Velocity
We have already discussed the differences between average velocities (all
fluid assumed traveling at the same velocity) and point velocities.These
differences do not seem too important until we start to measure flow. It is
wise, however, always to remember the difference so as not to fall into
the trap of neglecting it when it is important.
Most data is given in volumetric units, meters cubed per hour, liters per
minute or gallons per minute, to name but a few sets of units.The time
base for flow rates when performing most computations is seconds and
the flow frequently has to be converted to mass units per second. It is
usually best to do all conversions on the initial statement and the final
statement of all computations and to maintain a consistent set of units.
This is particularly important for automatic calculations using computers.
To obtain average velocity from volumetric flow, we first convert all volumetric and time units to cubic meters or cubic feet and to seconds.We
then use the equality, q = AU, where the flow rate, q, is equated to the
cross sectional area, A, of the flow path times the average velocity, U.
Mixed metric units are sometimes quoted as being S1 units. For instance,
if the flow rate, Q, is given in cubic meters per hour and the pipe diameter, d, in centimeters, the following expression results:
This is an example of mixed metric units, not S1 units.The coefficient,

3.537, converts the units (hours and centimeters) to the appropriate ones
and includes the constant, IT.
A simdar expression in mixed US. units of Q gpm and d inches would be:
The opportunity for error is much greater using mixed units. It is much
better to follow the advice given above and to convert all units before
beginning a computation.
Velocity head
Velocity head represents the kinetic energy of the flowing fluid at a cross
section. It is converted to feet or meters of head units by hydraulicians. It
and Equations
117
Pumps
is easier for other disciplines to use energy per unit mass units.To make
the conversion, it is probably good to remember the lunetic energy is
numerically equal to the distance a body, starting with zero velocity, must
fall in order to reach flowing velocity.
In pump work, the velocity, U, is the average velocity across a pipe section in feet or meters per second. The coefficient, a , is close to one, and
is usually taken as such without too much error.The dimensional constant, gc, is 32.17 ft-lbm/lbf- s2 in customary U.S. units. If the units are
checked using the U.S. system, it is seen that head has the dimensions of
energy per unit mass, ft-lbf /lb, The units of this term are numerically,
but not dimensionally, equal to feet of fluid.
With S1 units the dimensional coefficient is not needed; it is equal to one
without units. The coefficient, a , is again usually taken as one. The
dimensions of the head term are velocity squared or meters squared per
second squared. Substitution of Newton's law shows these units to be
equivalent to newton-meters per hlogram. In other words, the units are
energy per unit mass. These units are not numerically equivalent to
meters. In order to obtain meters, divide by the acceleration of gravity,
9.805 m/s2 (see Chapter I, Section 10).
Viscosity effects
The resistance to motion of liquids, the viscosity, tends to decrease with
rising temperature.Viscous liquids increase the power requirements
because of the conversion of mechanical energy to internal energy in the
flowing system.Viscous liquids reduce the differential head (and differential pressure) of a given pump due to internal irreversibilities. Effectively,
very viscous fluids reduce pump capacity, head and efficiency.
Work performed i n pumping fluids
In order for liquid to flow, work must be expended.A pump converts
energy at its shaft to fluid energy.This fluid energy is used to:
1. raise liquid to a hgher elevation;
2. overcome a h g h pressure in a vessel;
3. overcome viscous resistance in conduits;
4. increase the velocity of the fluid;
5. or, all of the above.

118
and Equations
Pumps
C H A P T E R
T H R E E
Pump efficiency
Pump efficiency is usually expressed as the ratio of power out (fluid
power) to the power in (shaft power). Power is the rate of doing work.
Work is force moving through a distance.The power out is the difference
between the fluid power leaving the pump and that entering. The power
in is the shaft power to the pump.
Efficiency is of importance economically because reduced efficiency
means expenditures of work energy are leaked to the surroundings as
heat energy instead of being used to perform the intended work.
Efficiency is important in other ways. A pump with low efficiency is a
hot pump. There may be unexpected bearing failures. Increased maintenance may result.
The pump manufacturer is not normally the motor manufacturer. Pump
efficiency is not motor efficiency.When the pump manufacture specifies
the efficiency of his pump, he does not know necessarily what motor will
serve as the driver, or even if it will be an electrical motor or a steam turbine, for instance.
A general, arbitrary classification of pumps follows.

1 A. Radial pumps
Even though the single stage radial pump looks simple, it is not. It is a
carefully engineered device.
The first distinction to consider is the impeller. It can be open, meaning

the pump casing completes the channel walls, or closed or shrouded,
meaning the fluid passing through the impeller is completely contained
by the impeller.The shrouded impeller is more efficient, but has smaller
passages.The process will dictate the choice, for instance if the fluid contains solids, an open impeller will be less prone to problems because of
the extra clearance.
The second distinction is whether the pump is a single suction or double
suction pump. Normally, the impeller receives liquid along its axis, at the
eye, from one side. It accelerates this liquid radially while turning it
through ninety degrees from the axial to the radial direction.The reactive
forces involved are obviously unbalanced and thrust bearings are necessary. O n larger capacity pumps, sometimes it is advantageous to have
what are effectively back-to-back impellers on the same shaft. Equal volFlow o f Industrial Fluids-Theory
and Equations
119
Pumps
umes of liquid are fed axially from both sides.The thrust is better balanced. Discharge is to a common casing.
Casings
The casing of a pump is usually classified as circular type, volute type or
diffuser (guide vanes) type.
The circular type of casing simply encloses the impeller with some clearance.There is a single discharge point. It is the least efficient of the three
types due to the shock and turbulence of fluid leaving the impeller. It is
seldom used.
The volute type has an increasing casing radius in the direction of the
discharge point, much like a sea snail.The shape and the increment of the
radius per degree are chosen to minimize shock losses.
The diffuser type improves on the volute concept by adding stationary
guide vanes surrounding the tips of the impeller.These guide vanes
smoothly direct flow to the discharge.The diffuser type pump is the most
efficient. Diffusers are used in multistage, high head pumps to direct flow
to the suction of the next stage, as well as to the discharge.
1 B. Axial pumps
Axial pumps have a lifting or propeller action. In the strictest sense, they
are not centrifugal pumps. They move liquid axially not centrifugally.
Figure 111-2 is an example of a large axial pump used as a closed loop
recirculation pump in an evaporator.
This type of pump typically has a very high flow rate with a fairly low
differential pressure across the pump. Flow can be in excess of 450 cubic
meters per hour (2,000 gpm) and the differential head can be 15 meters
or about 50 feet (with water, about 20 psi).
In evaporators and calandrias the pump casing is often comprised of a
large piping elbow. Flow control is frequently by design of the overall
hydraulic system.The pump simply runs out along its curve until the
system resistance exactly balances the pump head.
1 C. Mixed flow pumps
Mixed flow pumps combine both radial and axial elements.They are usually vertical, multi-stage pumps and in this case are known as turbine
120
and Equations
Pumps
pumps. However, the term "turbine pump" is also used for a pump that
uses the mechanical impulse principle combined with centrifugal force to
impart momentum to the fluid.
The vertical mixed flow pump is capable of about 100 feet or 30 meters
per stage (43 psi with water). By its nature, it is a self-priming pump. The
liquid at the suction is intended to be above the eye of the bottom
impeller.The pump is frequently chosen in order to avoid bottom connections to a vessel. It is also used in sumps.
The regenerative turbine pump is a low capacity, high head pump with a
very steep, almost straight head-capacity curve. It has a very sharp peak on
the efficiency curve. Because of the turbulence generated during the
impulse, it can only be used on clean fluids. Erosion can present problems.
1 D. Multistage pumps
Multistage pumps overcome the speed (and, therefore, pressure) limitations of a single stage by combining more than one stage, usually on the
same shaft.They are used as deep well pumps, as high-pressure water
supply pumps, as boiler feed water pumps, as firewater pumps and as
charge pumps to high-pressure refinery processes.
The dscharge pressures can be as h g h as 40 MPa (6,000 psi).The more
stages a pump has, the more delicate and costly the pump is and the more
limited its turndown capabhties-The pump engineer should always be consulted for his knowledge about the specific characteristics of such pumps.
2. Positive displacement pumps
A positive dsplacement pump either has a reciprocating action or a rotary
action.The sirnplex reciprocating pump allows liquid to flow into a cylinder
during one stroke. Flow is through one set of check valves. It expels the
liquid through another set of check valves during the opposite stroke.The
rotary pump traps a quantity of liquid in a cavity at the pump suction and
squeezes the same quantity out at the dscharge. Positive dsplacement
pumps include daphragrn pumps and screw pumps. Screw pumps are rotary
positive displacement pumps whose discharge can be extremely smooth.
A reciprocating pump has an inherent pulsation associated with the passage of fixed quantities of liquid in a periodc fashion.The rotary pump
can have quite a smooth output, depending on its design.
and Equations
121
C H A P T E R
T H R E E
Pumps
One characteristic that is common to positive displacement pumps is the

fact they cannot have their discharge throttled. If flow to the user must
be controlled, it must be done by speed control or by recirculating a
quantity of fluid around the pump.Throttling would result in excessive
backpressure on the pump that could stall the driver, rupture the piping
or casing, or cause a safety valve to lift.
Pump characteristics always find their utility. In pipeline pigging operations, the positive displacement pump will force the pig through a
momentary obstruction whereas a centrifugal pump might not generate
sufficient excess pressure to overcome the obstruction.The pipe has to be
protected against overpressure by a safety valve when a positive hsplacement pump is used.
2 A. Piston and cylinder pumps
The first pumps were piston and cylinder pumps.They were made of
wood and they frequently had square section cylinders and pistons.
Bronze, cast iron and steel eventually replaced the wood. It was natural to
connect a reciprocating steam engine to the reciprocating pump for such
applications as mine drainage. In the early days, pumps were used in the
suction mode, so it did not take long to learn one could only lift water
about 10 meters or 30 feet before suction was lost. However, it was also
learned that in a discharge mode, as long as the cylinders were kept
primed and the hscharge piping and pump casing were strong enough,
water could be raised to much greater elevations.
Loss of prime
It is worthwhile to discuss "loss of prime" as a preliminary to detailed
discussions on net positive suction head. In the suction mode, the
(approximate) 10-meter or 33-foot limitation applied only as long as the
check valves were in good shape and there was no leakage of air into the
suction piping. If any of the above occurred, the pump operating in the
suction mode was observed not to be able to raise the liquid to the
usual level.
The word "prime" refers to the presence of liquid in the eye of the
impeller.Without liquid, the pump could be trying to operate on much
less dense gas or vapor.The presence of gas would be due to a leak into
the suction.The vapor could be generated by the low pressure associated
with the displacement of the piston.
122
and Equations
Pumps
When pumping technology developed to the point where there was little
or no leakage into the suction, it was found there was still a limitation on
the maximum suction 1ift.This limitation was found to be associated with
the vapor pressure of the 1iquid.The vapor pressure is essentially due to
the escaping tendency of molecules from a liquid. The vapor pressure is a
function of temperature and of the fluid.
Normally, the surroundmg pressure is greater than the vapor pressure,
thus limiting the rate at which molecules can escape.When the surrounding pressure becomes less than the vapor pressure, the escaping rate
of the molecules of liquid increases in order to maintain the vapor pressure associated with a particular temperature. Heat flow from the environment supplies the energy necessary for vaporization to occur.
In the older, slower, simple piston pump operating in a suction mode, loss
of prime signifies inability to pump, and not much else. In a centrifugal
pump, loss of prime can have devastating consequences. In particular,
problems arise as the pressure approaches the vapor pressure of the liquid
at the inlet to the pump. Subsequent acceleration of the liquid can cause
the pressure to drop below the vapor pressure and cavities (vapor bubbles) to form in the liquid. Subsequent deceleration can cause the pressure to increase and the cavities to implode, suddenly. It is this sudden
implosion that generates the enormous point pressures that are the cause
of metal erosion in a pump.The phenomenon is usually called cavitation.
Specific control requirements of piston and cylinder pumps

We will only briefly &scuss the control requirements relating to fluid
flow problems. As always, it is wise to divide control requirements into
those associated with machinery protection and those associated with
flow measurement and control.We wish to concentrate on the latter, but
one cannot ignore the former.They must at least be addressed so we are
aware of them.
Overpressure
Most reciprocating pumps are capable of discharge pressures sufficiently
high to do damage if not controlled. Above 15 psig (103 kPa), most
boiler and pressure vessel codes call for mandatory use of a pressure safety
valve.The safety valve should be set at the design pressure of the weakest
member of the protected system. It should have a capacity that allows it
to function without exceeding this pressure by a stated safety margin. In
the case of a positive displacement pump, this capacity is normally the
and Equations
123
C H A P T E R
T H R E E
~ ~ Tsh e o r y
and
Equations
rated capacity of the pump, unless there is another relieving passage in

the system.
Pulsation
All reciprocating pumps have pressure and flow pulsations in the discharge and suction lines. Pressure (and flow) dampeners are usually associated with such pumps. These dampeners are usually volume tanks.
Frequently the tanks have bladders filled with or surrounded by gas.The
function of the tank is to act as a capacity tank during the high pressure
half cycle and to relieve the accumulated volume through the system
resistance during the low pressure half cycle. Figure 111-7 shows the typical discharge characteristic of a positive displacement pump. The suction
characteristic will be similar with smaller amplitude
The purpose of most pulsation dampeners is to prevent mechanical
damage to the system. Usually residual pulsations are present that can
cause difficulties for measurement and control devices. Dampening is
applied usually for mechanical protection. This smoothing of pulsations is
frequently not sufficient for measurement and control purposes.
Figure Ill-7.Typical characteristic curves for reciprocatingtype pumps
Flow control
Flow control must be by diverting some of the volumetric flow from a
constant volume pump or by controlling the speed of the driver. Usually,
the decisions are made on the basis of process, mechanical and economic
124
F l o w o f ~ n d u s t r i a l ~ ~ u i d s - T h e o r ya n d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
considerations. If the driver is a fixed speed one, flow diversion is usually

the only solution. If the driver is a variable speed one, diversion may std
be necessary. This occurs when the combined turndown of the pump and
driver is limited or if the pressure requirements of the system prevent the
use of speed control.
2 B. Rotary pumps
Rotary pumps vary from simple peristaltic pumps, associated with laboratory dosing operations, through gear pumps to lobe pumps and screw
pumps. Their pressure and flow characteristics vary from mild pulsations
to no pulsations.They still, however, are positive displacement pumps and
as such are subject to the same overpressure and flow control requirements as mentioned above. Only the pulsation-damping requirement may
be relaxed in some cases.
Frequently, rotary pumps have internal controls that divert some flow to
suction when the backpressure becomes excessive. This sometimes leads
to the assumption that rotary pumps can have their discharge throttled.
In effect, a combination of throttling and diversion is being used. It is
well to remember when any pump operates in a recirculation mode,
energy is being added continuously to a (partially) closed system.
Inefficiencies will cause the system to heat up. If the excess heat is not
removed adequately by natural or forced cooling, problems will result.
2 C. Diaphragm pumps, metering pumps

Diaphragm pumps and metering pumps are positive displacement pumps.
They are classified together only because they use a diaphragm and check
valves. They may be large, air-driven pumps used extensively as portable
pumps in chemical plants for a large variety of services.They may be
highly precise metering pumps for dosing small quantities of fluid into a
larger process.
One problem to bear in mind is the maximum pressure the pump is
capable of delivering when the main system is shut in. It may be necessary to provide for local safety valves if this pressure is too high. The
small, innocent lookmg metering pump is often quite capable of rupturing equipment and even piping.
Miscellaneous pumping devices

There are many specialized devices that have been used for pumping.
Some are not too efficient, but are used for practical reasons.We will
and Equations
125
Pumps
Theory
and Equations
briefly describe some of them in order to complete the picture and to

supply some food for thought.The theory of such devices is frequently
enlightening.
3 A. Gas lifts
A gas lift (Figure 111-4) works on the principle that, if a column of water
exists, and, if gas of sufficient pressure can be introduced into the bottom
of the column, then that gas will reduce the overall density of the
column of liquid by virtue of mixing with it. If the column is separated
from the surroundmg fluid by a wall except at the bottom, the heavier
density surrounding fluid will force the lighter density contained fluid up
the column. The liquid and gas mixture within the column will rise to a
higher elevation than the more dense 1iquid.The length between the
normal level of the liquid and the bottom of the column is called submergence.The length between the normal surface of the liquid and the
point of discharge from the column is called lift.
The use of gas lifts (air lifts) was very common in mine drainage applications before the advent of efficient centrifugal pumps. Gas lifts are still
used in the oil extraction industry.
3 B. Jet pumps
Jet pumps (Figure 111-5) are a class of pumping device that makes use of
momentum transfer between two compatible fluids. In this chapter we
are only considering liquids, but the principle also applies to gases.
Jet pumps, like steam engines, evolved in a highly empirical manner.
Externally they have the appearance of a simple piping tee. Internally,
they consist of four parts.The first part is a nozzle that accelerates the
motive fluid to the entrance to the mixing section.The second part is an
inlet chamber that brings the driven fluid more or less smoothly to the
entrance to a mixing section.The third part is a mixing section where
the two fluids are mixed and are decelerated and accelerated by
momentum transfer. The fourth part is a diffusing section that decelerates
the mixed fluid further and increases the static pressure.
Jet pumps are given different names depending on their application, on
their user and on their manufacturer.They are called injectors, ejectors,
eductors and elevators and when they are "pumping" gases they are called
gas compressors or exhausters. In principle, they are the same. In this section, we shall refer to them as jet pumps.
126
~ I o wo f ~ n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
Jet pumps are essentially static devices. AU parts are of fixed dimensions,
therefore it can be expected that they be designed around one operating
point and that they will be less efficient when operated away from the
design point.The addition of an adjustable needle in the inlet nozzle
gives some degree of control of motive flow and efficiency in some cases.
If a sufficiently high-pressure motive fluid can be introduced to the inlet
chamber of the jet pump, the fluid can be made to accelerate through a
nozzle and thus gain kinetic energy and momentum at the expense of
the pressure energy.The static pressure drops as the fluid is accelerated.
The design of the jet pump is such that the suction inlet is connected
physically to the low pressure point of the accelerated fluid.Therefore,
the normal pressure of the fluid at the suction inlet forces it into the
accelerated stream. The mixed stream is now decelerated in the mixing
section where it gains pressure. It is further decelerated in an expanding
nozzle where it gains even more pressure.This pressure can be used to
move the mixed fluid to another location.
Efficiency considerations
At first glance, it may seem a jet pump is a wonderfully simple, inexpensive and, therefore, efficient device to use. It is true the device is simple
and inexpensive; whether or not it is efficient depends upon the alternative applications.
The first point to be made is the jet pump is not too efficient thermodynamically. It is inherently a mixing device and mixing is one of the main
causes of thermodynamic inefficiency. In addition, turbulence and shock
cannot be overcome as well as with centrifugal pumps, so the efficiency is
much less.
The second point is the motive fluid has to be pumped in order for it to
gain the pressure required at its flow rate.There are losses associated with
this pump and with the piping system to the device.The combination of
inefficiencies results in a low overall efficiency.
Utility considerations
If the jet pump is so inefficient, why is it used at all?The answer to this
rhetorical question lies in the simplicity of the device and in the possible
availability of cheap (excess) motive power.
The device is simple. It can be designed to accept solids, limited in size,
mixed with the pumped liquid at its suction inlet. It can therefore pump
slurries.
and Equations
127
C H A P T E R
T H R E E
P u m p s - Theory and Equations
It is small and can be made of exotic materials at little extra expense. It

can, therefore, handle corrosive and erosive fluids with less capital cost
than can a normal pump.
If motive power is available that cannot be economically used elsewhere, there is no need to pay for a more expensive pump.
Control considerations
The jet pump is a relatively static device.The motive fluid can be throttled only over a limited range before losing its ability to transfer
momentum within the fixed configuration of the pump. The fluid being
pumped can be throttled as a means of controlling its flow.What suction
throttling does is to create a resistance to flow into the low-pressure area
of the vena contracta of the motive fluid's jet.Throttling at this point can
be done with relative impunity.
Applications when pumping liquids
Jet pumps can pump liquids using other liquids as the motive stream, or
they can use steam as the motive stream.This section will discuss only the
use of liquids as a motive force.
When pumping liquids, jet pumps are used either in a suction mode or
in a discharge mode. In the suction mode the vapor pressure and the suction friction losses limit the height to which the liquid can be elevated.
With water, six meters (20 feet) is the approximate limit. The additional
discharge elevation is normally limited to about one half a meter when
the jet pump is operating in the suction mode. In the discharge mode,
the jet pump is located within one half a meter of the liquid to be elevated and it is capable of raising the liquid a further approximately nine
meters (30 feet). One manufacturer gives an elevation of two meters per
bar gauge (100 kPa, 14.5 psig) of motive pressure.
In spite of its low efficiency when compared to a centrifugal pump, a jet
pump is frequently used in place of a centrifugal pump. This is particularly the case when the pumping operation is infrequent such as when a
sump must be emptied periodically or when a centrifugal pump must be
primed by pulling a vacuum on its casing.The jet pump is also used
when abrasive solids would otherwise damage a more expensive pump
and when corrosive materials must be diluted or neutralised while they
are being pumped. Sometimes it is used simply because a stream of highpressure liquid is available whose energy would otherwise be wasted.
128
FIOW
o f ~ n d u s t r i a~ ~I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Pumps
Cormnercially available jet pumps are usually off-the-shelf items sold in a

choice of fixed sizes. Each size is designed around a fixed operating point.
The user picks the size closest to his application and accepts the additional
inefficiency due to the pump operating away from the design point.
One manufacturer gives performance specifications as two sets of charts:
one for the suction mode and one for the discharge mode (same jet
pump). In the discharge mode the suction elevation is limited to one half
a meter.The charts give curves for discharge elevations from two to nine
meters. Suction and motive flow rates can be read corresponding to
motive pressures. Corrections are applied for different densities and for
suction elevations greater than one half meter. In the suction mode, the
discharge elevation is limited to one half meter.The charts give curves for
suction elevations from one to six meters. Again, corrections must be
applied for non-standard conditions.
The use of such performance charts is similar to the use of performance
curves for a centrifugal pump.The major difference is the fact the motive
stream is added to the suction stream. It effectively goes along for the
ride. In Appendix AIII, we will use the above mentioned curves in order
to compare jet pump efficiencies with centrifugal pump efficiencies.
While we seem to be emphasising efficiency, it must be remembered jet
pumps are frequently chosen in spite of their inefficiency, because they fit
other requirements better than do centrifugal pumps.
3 C. Liquid metal pumps

The motor principle is that a conductor carrying an electric current at
right angles to a magnetic field will have a force imposed upon it. The
force will be perpendicular to both the magnetic field and the electric
current. In a normal electrical motor, the windings carrying the current
are distinct. In a liquid metal pump, there are no windings. The liquid
metal is the conductor.The current is made to pass at right angles to the
conduit by suitable electrodes placed opposite one another on the conduit.The magnetic field is established at right angles to both the electrodes and the conduit. The direction that is mutually perpendicular to
both the magnetic field and the imposed current is in the axis of the
conduit.This is the direction of the force and of fluid motion.
Electromagnetic pumps have been developed with differential heads of
300 psi (2 MPa) and flow rates up to 10,000 gpm (2.3 m3/h).
and Equations
129
C H A P T E R
T H R E E
Pumps
These pumps are highly specialized and are generally limited to heat
transfer applications in the nuclear industry. We will discuss them briefly
in Appendix AIII.
111-6: I N H E R E N T AND INSTALLED CHARACTERISTICS O F P U M P S

Pumps, like control valves, have inherent characteristics and installed
characteristics. It is well to remember the distinction. We will give a
rather detailed technical description in Appendix AIII.
Inherent characteristics
An inherent characteristic is one that is proper to the pump when all
external influences have been eliminated. It is what the pump manufacturer uses to describe his pump.This is done so a maximum of information can be transmitted in the most simplified form.The inherent characteristics are usually shown on the pump curve, and most of us are reasonably familiar with that of the centrifugal pump.The basic curve is that of
developed head in feet or meters of fluid versus the volumetric flow rate.
It is up to the user to convert this basic data to a more useful form.
If the reader gets into the habit of thinking of the developed head as
energy per unit mass instead of feet or meters of fluid, he or she will
avoid many potential pitfalls. The typical centrifugal pump curve gives
the energy per unit mass that the pump transfers to the fluid plotted
against the volumetric flow rate.The pump curve will give the appropriate units in feet or meters and in gallons (U.S.) per minute or cubic
meters per hour.
Since the developed head is the energy per unit mass that the pump
delivers to the fluid, this energy can manifest itself as additional kinetic
energy, additional pressure (static) energy or additional potential energy
(lifi).Depending on the physical characteristics of the pump, elevation
differences and velocity differences at the suction and discharge, the
developed head read from the pump curve might or may not correspond
directly with the measured differential pressure across the pump. The
measured differential pressure must be converted to the same units as are
used on the pump curve.The measured quantity must be corrected for
the changes due to elevation differences and to velocity differences at the
inlet and outlet of the pump.
130
FIOW
o f ~ n d u s t r i a l ~ I u i d s - ~ h e o r ya n d E q u a t i o n s
Pumps
Installed characteristics
An installed characteristic depends upon the system in which the particular pump functions. The most common installed characteristic curve
described in the literature is that of the centrifugal pump. Again,
remember centrifugal pumps have a wide variety of installed characteristic curves. Positive displacement pumps have completely different ones
and positive displacement pumps are subject to the constraint that they
cannot be directly throttled.
Constant speed centrifugal pumps

We will start with the installed characteristic curve for a constant speed
centrifugal pump. The standard figure consists of two curves.The first is
the inherent pump curve. The second is the system curve.
Usually the system is a simple one consisting of a fixed configuration, but
the resistance also can be a constant pressure with very little variable
resistance or even a mixture of the two. In analyzing a fluid system, it is
often found the system may have different configurations at different
times when subsystems are valved into and out of the main system. In
this section we will only consider the simple fixed configuration. The
other systems will be discussed in Appendix AIII.
The system curve is usually superimposed upon the inherent curve.The
inherent curve represents energy per unit mass flowing added by the
pump.The system curve represents energy per unit mass flowing required
to lift the fluid to a particular elevation, to overcome pressure differences,
and to overcome viscous resistance due to the nature of the fluid being
pumped. The requirement of overcoming viscous resistance represents
useful work transformed to internal energy.
The inherent curve is dependent on pump design. It usually curves
downward with increasing flow, but it may have a fairly sharp peak before
doing so. For the moment we will consider a monotonically decreasing
curve.This type of curve represents the fact that energy transferred to the
fluid per unit mass flowing decreases with increasing flow through the
pump.The maximum fluid energy per unit mass transferred to the fluid is
at a flow close to shut off, when there is little flow (and little cooling
effect).As the flow rate is increased, the energy transferred drops off
because of internal inefficiencies within the pump. The pump is designed
to maximize its efficiency at a fixed flow rate.
and Equations
131
C H A P T E R
T H R E E
Pumps
The system curve usually starts off above zero on the abscissa.The ordinate represents the minimum amount of energy per unit mass that must
be transferred to the fluid to lift it against gravity and the system static
pressure differential without flow. As flow increases, the system curve also
curves upward at approximately a square law rate.This is due to the
energy requirements to overcome irreversibilities generated by flow and
to create the lunetic energy being added to the static energy requirements. Note we have not drawn a distinction between pumps that
operate under positive suction head and those which must supply suction
lift. Once the pump is operating within its limits, there is no difference
between the two modes of operation.The energy transferred to overcome pressure, to create elevation and kinetic energy is recoverable.The
energy transferred to overcome irreversibilities is not.
The two curves (inherent and system) will meet at a common point. At
this point, the energy transferred to the fluid by the pump per unit mass
exactly equals the energy necessary per unit mass to produce flow and to
raise the fluid to some elevation against some pressure within the particular system.The pumping system described by this curve will operate at
this point and no other. If the pump is turned on, it will automatically
settle at the flow rate and total dynamic head associated with this point.
The difference between the inherent pump curve and the system curve
at a lower flow rate than that of the common point represents the energy
per unit mass that must be absorbed by a valve or turbine if flow is to be
controlled at this lower flow rate. Higher flow rates than the one associated with the point of convergence of the two curves are impossible you cannot get there from here.
It is conceptually important to remember that a control valve converts
useful mechanical energy into internal (heat) energy, although it is not
often apparent.The difference between the two curves at points with
flow rates lower than that of the common point represents the energy
converted to internal energy in order to maintain the given, lower, flow
rate.Throttling is not free, although when all costs and efficiencies associated with other methods of control are factored into the equation, throttling is still the most common method of controlling flow.
If an economic means could be found to recover the mechanical energy
converted to internal energy by a control valve, it would be used. Liquid
turbines, for instance, are used in some high-pressure let down situations.
They are called power recovery turbines.
132
F l o w o f I n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
Variable speed centrifugal pumps

The curves for a typical variable speed pump consist of essentially parallel
plots for different rotational speeds.At first glance, it is tempting to think
speed control is an extremely good way of adjusting the head versus
capacity requirements of a system and of saving energy.
However, there are a few points that cause many variable speed pumps to
be operated in a fixed speed mode:
From a control point of view, variable speed drives usually have much
greater inertia than control valves, so their response to a control signal is
much more sluggish. It is possible to use a variable speed drive to slowly
optirnize the response of a faster acting control valve, and ths is sometimes done. Steam turbine drivers are usually much more responsive than
drivers that use gear reducers, but steam must be avdable.Variable fiequency electrical drives may not have the problem associated with inertia,
but their cost must be considered.
From an efficiency point of view, as has been previously stated, centrhgal
pumps are designed about a fked operating point.Their efficiency drops
off as the fixed speed curve is traced in either dn-ection away fiom the
design point. Examination of variable speed curves reveals that at lower
speeds, the design maximum efficiency is not even reached.
In adltion to the pump efficiency, the driver efficiency must be taken
into account.The match between the pump and its driver may not be
perfect and, even If quite adequate at one point, the efficiency drops off
quite ra&cally at other points. Remember the overall efficiency is the
product of the pump efficiency and the driver efficiency. Both efficiencies
have curves, so the product must be plotted versus flow to obtain the true
picture.
Distinction between centrifugal pumps operating with suction l i f t and those under
a positive suction pressure
Regardless of the type of installation, the pressure at the inlet to a pump,
or within the pump, must not &p below the vapor pressure of the fluid. If
this occurs, two distinct possibilities exist.
The first one is the pump simply loses prime. It ceases to pump liquid
and simply churns whatever liquid is in the pump casing against a closed
check valve. The second possibility is flow continues and pressure is
recovered within the pump. The vapor bubbles (cavities) created when
pressure dips below the vapor pressure implode violently, creating point
Flow of industrial Fluids-Theory
and Equations
133
Pumps
pressures that can exceed 10,000 psig (69 MPa). This second possibility is
known as cavitation. Most pumps do not last too long if cavitation is
allowed to occur within them.
Since the pump manufacturer knows the inherent characteristics of his
pump, he establishes the minimum pressure that must be available at the
inlet to the pump to prevent problems with cavitation.This pressure is
given in terms of head. It is called the net positive suction head required,
NPSH(R). The system designer computes a suction head available within
a given piping configuration.The formula wdl be given in Appendx AIII.
This head is designated, NPSH(A). The NPSH(A) must always exceed
the NPSH(R) at a11 points along the pump curve.
Some pumps require more head than others do at their suction. It should
be evident a pump operating with suction lift should be designed to have
a very low NPSH(R). Once operating within its design parameters, the
pump does not "know" what mode it is operating in - lift or positive
suction head. From an operating point of view, establishing the NPSH(R)
for a particular pump and making sure the NPSH(A) always exceeds the
NPSH(R) is paramount.
Pumps operating in the suction lift mode have another problem - that of
prime.To get the pump to start pumping, liquid has to be present within
the pump casing and suction piping. Priming the pump can be done
manually, by opening valves to add liquid and vent gases, prior to starting
or it can be done by control logic. It can also be performed by pulling a
momentary vacuum on the pump casing. Priming has to take place each
time the pump is stopped more than momentarily. In some cases, foot
valves have to be present to allow the piping to be filled.Vent valves are
necessary to allow purging gases in the suction piping and pump casing.
Positive displacement pumps

The relationship that exists between differential pressure across a centrifugal pump and volumetric flow does not exist with positive displacement pumps. By its nature, the positive displacement pump, or volumetric pump, moves a fixed volume of fluid with each cycle or revolution. Increasing backpressure does cause slip to occur. Slip is leakage past
cylinders and rotary members that reduces the forward flow on increasing
backpressure. It is an incidental phenomenon; it cannot be used for control purposes.
134
F I O W o f i n d u s t r i a ~ ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Pumps
C H A P T E R
T H R E E
The constant speed positive-displacement pump discharges a fluid that is

essentially incompressible at an average rate that is fixed.The average rate
is pulsating or it can be quite smooth, depending on the type of pump.
For a positive displacement pump, the head is strictly a function of the
system backpressure. Head influences capacity only through slip.
111-7: C O N T R O L L I N G F L O W T H R O U G H P U M P S
When we talk of controlling flow through pumps, we tend to concentrate
on the flow control devices.We tend to neglect the aspects of control that
are part of the design decision-mahng process. In this section, we will discuss pumps as they interact with the system in which they are installed.
Fixed speed centrifugal pumps
When operating at a fixed speed, the volumetric flow through a centrifugal pump is a function only of the total dynamic head imposed on
the pump by the system.This mechanical energy is exactly equal to that
transferred to the fluid by the pump. The system resistance is made of
dynamic elements (such as control valves and varying pressures in vessels)
and static elements (such as pipe and equipment friction drops).
Self-regulation
As has been pointed out by Les riske ell^" (Control Valve Siting and
Selection, ISA), the ideal method of control, when feasible, is self-regulation. He also points out self-regulation does not come without thought.
Furthermore, he emphasises the importance of studying the possible
interactions of the entire system before choosing a control scheme.
Driskell gives an example of self-regulation of a centrifugal pump. The
system is that of Figure 111-8 which is taken from the above work. It is
worth analyzing, as it fixes the fundamental principles. We will put some
real numbers on the analysis in Appendix AIII.
and Equations
135
C H A P T E R
T H R E E
Pumps
F
TDH
Figure 111-8. Self-controlled pump
If the pump has a smoothly decreasing head-versus-flow curve it can be

made to operate over a wide range of flow rates without external controls.
The head imposed on the pump is due to the differences between the
suction and receiver pressures and elevations and the differences between
the inlet and outlet kinetic energies and losses due to irreversibilities associated with flow.The suction elevation varies with inflow to the vessel; the
discharge elevation is fixed. In the example, the maximum flow rate is
governed by the minimum dfference between the two elevations.The
minimum flow rate is governed by the maximum difference.
We do not have to worry too much about the maximum flow rate. It
normally, will be chosen so it is on the pump curve (this assumption may
be worth checking, however).We may have to take precautions to avoid
the loss of NPSH.These precautions can be as simple as having a lowlevel switch trip the pump (with automatic restart on a higher level).An
automatic recirculation valve would achieve the same protection without
stopping the pump.
The flow rate in this case will vary smoothly between the maximum and
the minimum depending only on the imposed head.This type of control
system will settle out at a flow rate that is in balance with the flow into
the sump. It needs no controls other than those that protect against the
suction loss mentioned above.
136
Pumps
C H A P T E R
T H R E E
Discharge throttling
A common method of flow control is discharge throttling by placing a
control valve in the pump discharge 1ine.A valve in the suction will cause
the pump inlet pressure to decrease and may cause cavitation.The control
valve in the discharge line adds to the system losses so the total dynamic
head imposed on the pump may be varied. Note the irreversibilities created by the valve must be real losses not apparent 0nes.A high recovery
valve such as a butterfly valve or an angle valve will have lower real losses
than the ones that appear to correspond to the measured differential pressure immediately across the va1ve.A high recovery valve would have to
be closed to a greater extent than would a low recovery one.
The energy per unit mass converted to internal energy by the valve can
be depicted by a straight vertical line between the inherent pump curve
and the system curve at a given flow rate. More exactly, the energy per
unit mass of flowing fluid imposed on each side of the pump by the
system pressures and friction losses govern the flow rate. Reducing the
differential head increases the flow rate; increasing the differential head
reduces the flow rate.The relationship is not linear.
There is another way of looking at the pump inherent characteristic
versus the system characteristic relationship. It is to realize the throttled
control valve simply adds to the permanent losses of the system - to the
irreversibilities that were estimated with the valve wide open.
By throttling the valve, we have created a new system with greater losses.
Since the shutoff head remains the same, and since the system curve still
follows a square law relative to flow, one can simply rotate the system
curve counter clockwise so it meets the pump's inherent curve at the
corresponding flow rate.
Unless there is a mismatch between the pump and the system, the maximum flow rate is rarely a problem with typical centrifugal pumps. Pipe
fluid velocities in suction and discharge piping have already been discussed. These are controlled by the design of the system.When all control
and manual valves are wide open, the pump simply settles down to the
point where the system curve meets the pump curve - where the energy
delivered by the pump is completely converted to static, kinetic, potential, internal energy and to system irreversibilities.
and Equations
137
Pumps
Theory and
Equations
Some centrifugal pumps require special attention. For instance, high

head, high-speed pumps frequently have a maximum in their inherent
curve. Such a pump has two dfferent flow rates that correspond to the
same total dynamic head. Since it is the total dynamic head that governs
the rate, this pump can oscillate between the two rates, much as a centrifugal compressor is subject to surging.
Such pumps are usually very expensive and very delicate. Surging can ruin
the driver's gear train and damage the assembly.Therefore, a minimum
flow recirculation, frequently through a cooler, is often required.The
pump vendor may already have a standard minimum-flow recirculation
valve designed into his pump. If no such system exists then one must be
designed.The pump vendor will have established what the minimum flow
should be. An independent safety shutdown trip may also be required.
Centrifugal pumps i n parallel

Just because two pumps are piped in parallel does not mean they are
intended to operate in parallel.They may be spare pumps.
If two pumps are operated in parallel, their respective head curves must be
matched.The curves must be parallel and close to one another. Not only
must the two curves be matched but also the suction and dscharge piping
to the common points must be almost identical hydraulically.All of this to
make sure the differential head across each pump is as close as possible. If
the pumps and the hydraulic configurations are not matched, it is quite
conceivable a high discharge head of one pump may cause the other
pump to back up along its curve until it is shut off. It wdl overheat.
It must not be thought two pumps in parallel give twice the flow as one
pump. If an identical pump is added in parallel to an operating pump, it
initially will increase the flow.The increased flow will cause an increased
backpressure proportional to the square of the increased flow rate.The
increased backpressure is translated into increased total dynamic head that
causes both pumps to back up on their curves. Each pump, individually,
will produce a lower flow rate than the original pump did.The sum of
the flow rates of the two pumps wdl not by twice the original one; it
wdl depend upon the shape of the head versus capacity curves.
Centrifugal pumps i n series

When two centrifugal pumps are connected in series, one is frequently
considered a booster pump. For instance the pump on a fire suck takes
138
Pumps
C H A P T E R
T H R E E
its suction from the fire hydrant fed by the city water pumps.The purpose of the city water pumps is to circulate water to all hydrants at a reasonable pressure. The purpose of the fire booster pump is to increase the
pressure so the pressure energy can be converted to hnetic energy.A
high velocity jet of water results.
In pipeline work, pump stations exist along the pipeline in order to limit
the maximum pressure to which the line will be exposed. Each pump
adds the head that was lost in overcoming irreversibilities and head
changes between one station and the next. If fewer pumps were used, the
discharge pressure of each pump would increase and the pipe rating
would have to increase.
In plant work, a booster pump is sometimes added to a larger system that
contains many users, some at lower pressure, some at higher pressure.The
main problem with pumps in series is to make sure they are controlled so
their NPSH(R) requirements are always met. Each pump must have controls that prevent it from starting before the requirement for a minimum
suction pressure is satisfied and that trip the pump when the minimum
suction pressure is in danger of not being satisfied.Automatic recirculation can be used, but cooling has to be considered.The normal starting
sequence is from upstream to downstream.
Variable speed centrifugal pumps

Operating a centrifugal pump at variable speeds affords the opportunity to
change flow relatively slowly between static conditions.We have already
discussed the problems of sluggishness and the fact this inertia frequently
causes a variable speed drive to be used for trimming purposes only.The
fact the variable speed drive gives the opportunity to match the pump
discharge pressure to the system pressure requirement is often of interest.
The control characteristics of a variable speed pump are essentially those
of its driver. If the driver is a directly coupled, low inertia device (a variable frequency motor, for instance), what was said above regarding sluggishness may be ignored. Control of flow can be by control of pump
speed. If the driver has considerable mechanical inertia, or if the driver
has minimum turndown problems due to overheating at low rotational
speeds, then the characteristics of the driver may overwhelm those of the
pump. In any case, the total picture must be considered.The characteristics of each element must be known.
and Equations
139
C H A P T E R
T H R E E
Pumps
Theory a n d Equations
Start-up of centrifugal pumps

Most single stage, constant speed centrifugal pumps are started remotely
under automatic control. Limit switches are used to ensure automatic
valves are open at the appropriate time. Level switches or pressure
switches are used to ensure adequate suction head.The pump is frequently started against a closed discharge valve that is opened once the
pump is running.This is done to prevent the pump impeller from being
turned in the wrong direction because of a leaking check valve. It is also
the unloaded state for a centrifugal pump - minimum horsepower.
The more complicated the pump, the more expensive it is and the more
sophisticated the controls. The machinery expert may insist on operator
presence at start-up, even though the control systems expert may think
he can devise a perfectly safe automatic start-up sequence.The machinery
expert will usually win.
Positive displacement pumps
There are really only two means of controlling flow when positive displacement is involved. Some of the constant flow can be partially
diverted back to suction or to a receiver or the speed of the driver can be
changed. If the driver will not stall before the pipe or the casing rating is
exceeded, a pressure safety valve is necessary.
An interesting phenomenon that occurs with piston and cylinder pumps is
one that is due to the acceleration of the piston and that of the fluid in the
suction 1ine.The piston has a maximum acceleration at bottom dead center
and top dead center and zero acceleration as crosses the mid point of its
travel.The pulsating flow in the suction piping is not necessarily synchronized with the movement of the piston.When the inlet valve opens, the
fluid must be accelerated to catch up with the receding piston.The pressure required to achieve this acceleration is called "acceleration head". If
the fluid is insufficiently accelerated, there wdl be a momentary vapor
space behind the piston that wdl be eliminated suddenly somewhere in the
fill stroke - usually about two thirds of the trave1.A knock occurs and this
knock can damage the pump.The acceleration head must be supplied by
suction pressure over and above that required by a centrifugal pump.
Start-up of all types of pumps

We should always distinguish between start-up conditions and operating
conditions.This applies to all types of pumps, centrifugal or positive displacement.
140
F I O W o f ~ n d u s t r i a~ ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Pumps
Starting a positive displacement pump against a high backpressure usually

requires the pump be started in an unloaded mode - recirculation valve
open, discharge valve closed. Once the pump is operating smoothly in
the recirculation mode, the discharge valve is opened and the recirculation valve gradually closed.This allows the build up of pressure to overcome the backpressure and to open any check valves. Frequently, positive
displacement pumps are started locally so the operator can witness any
abnormal conditions.
111-8: HYDRAULIC T U R B I N E S
Hydraulic turbines are used in process work where an incoming line carries liquid at much higher pressures than is required by the local user.The
pressure may be required for transportation purposes by other users, for
instance. Letting down the liquid through a control valve means wasting
useful mechanical energy. It also might mean a special letdown valve in
order to avoid cavitation and noise.
If the flow rate and pressure are great enough, a hydraulic turbine can
convert otherwise wasted energy into useful energy. It does so while letting down the pressure to that of the final user. A suitable user of the
mechanical power would be an air compressor or a pump in continuous
service or an electrical generator.
Hydraulic turbines are sometimes described as pumps that run in reverse.
This is conceptually descriptive, but not exactly true physically. They do
reverse the action of a pump.They take energy available in a fluid stream
and convert it into mechanical energy at an output shaft.What is connected to the shaft depends on the ingenuity of the mechanical engineer.
Physically, the turbine is designed to accept high-pressure fluid smoothly
and to discharge it at some pressure that will be useful to the user.The
vector equations used to describe turbines are the same ones used to
describe centrifugal pumps. The ~ernoulli'equation also applies in an
identical fashion.
Hydraulic power recovery turbines are at least as reliable as pumps. Their
maintenance costs may be less.
Matching the turbine to the system

It must be remembered the turbine imposes its own requirements on the
process. The efficient operating range, for example, may be very narrow.
and Equations
141
Pumps
In fact, if the turbine is connected as the driver of an electrical generator

that is connected to a grid, it can be a power absorber rather than a
power producer. It can then operate as a pump during process upsets.
This is not what was intended by the designer.
In refinery service, it is not unusual to see a conventional driver connected through a one-way torque coupling to the ultimate power consumer. The conventional driver can supply power to the user until the
hydraulic turbine has developed enough torque to satisfjr the user's
requirements.The driver is then shut down.
There are no major fluid flow problems associated with hydraulic turbines different from those associated with pumps. Certainly, the computations are no different from those associated with flow caused by pumps.
The problems are those of matching the turbine to the system. Questions
that must be asked are:
Does the narrow range of the turbine fit the larger range of the process?
How do we control flow to maintain it in ths range?
Is a high-pressure letdown valve necessary to replace the turbine during
upset conditions? If so, are noise and cavitation a problem?
Is an independent conventional driver necessary during process upsets? If
so, how is it integrated into the system?
How is everythmg started up in an integrated fasluon?What happens
when we lose one element?
The technical problems associated with turbines from the point of view
of fluid flow measurement and control are of less severity, but similar, to
those associated with pumps. Net positive suction head is not normally a
problem. Turndown limitations may result in motoring (the turbine
becomes a pump). Generally, the operating limits must be clearly defined,
and the conditions imposed upon the turbine by the fluid must be controlled so the machine always remains within these limits. Once the operating limits of the specific turbine are understood, the problem usually
resolves itself to a "what if" analysis.
142
F I O W o f ~ n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Pumps
111-9: W O R K E D EXAMPLES
Example of Customary U S . units: Total discharge head
Sixty-degree Fahrenheit water is being pumped.A pressure gauge on the
discharge flange of a centrifugal pump reads 40 psig.The center line of
the gauge is 2 feet above the center line of the pump.The barometric
pressure at the pump location (usually an average number obtained from
the local airport) is 100 kPa.The discharge line is 2 inch, Sch. 40.The
flow rate is 100 U.S. gpm. What is the total discharge head in footpounds force per pound mass and in feet? It is worth remembering the
definition of total discharge head as being the sum of the absolute centerline discharge pressure and the kinetic energy both in the same units.
The first problem arises because of the use of weight density, y, in place
of mass density, p, in the hydraulic version of the ~ernoulli'equation.
Weight density has units of pounds force per unit volume. Mass density
has units of pounds mass per unit volume.The densities are numerically
identical, but the units are different. Dividing pressure in pounds force
per square foot by weight density in pounds force per cubic foot leads
directly to units of feet. Dividing the same units of pressure by mass density leads to units of foot-pounds force per pound mass, or energy per
unit mass.
The energy units are not quite so simple as the hydraulic ones, but they
are more meaningful. In the first example we will use both units, after
that we will use only energy per unit mass.
Discharge
Weight density
Mass density
Gauge pressure
Feetoffluid
Energy/mass
62.37 lbf /ft3

62.37 lb,/ft3
40 psig, 5,760 psfg
92.352ft
92.352 ft-lbf /lb,
Barometer
U.S. units
Feet of fluid
Energy/mass
100 kPa
14.5 psia, 2,088 psf
33.487 ft
33.487 fi-lbf /lb,
Atmospheric pressure
and Equations
143
Pumps
Correction to absolute units

(discharge gauge plus atmospheric conditions)
Pressure
54.5 psia, 7,848 psf
Feet
125.830 ft
Energy
125.830 ft-lb, /lbm
Correction to pump center line. If the gauge center line is above the
pump center line, the correction is positive. If it is lower, the correction is
negative. See explanation above.
(+) 2.0 ft
Distance
(+) 2.0 ft X g/gc = 2.0 ft-lbf /lbm
Energy/mass
Velocity head (a 1)
Flow rate
100 gpm,
0.2228 cfi
Flow area
0.0233 ft2
Velocity
0.2228/0.0233 = 9.56 fps
~
~ = 9.562/(2
/
2x 32.17)
~ = 1.42
~ ft-lbf /lbm
Kinetic energy
~
~ = 9.562/(2
/
2 X 32.17)
~
= 1.42 ft
Velocity head
Total discharge head, hd
feet
Gauge reading
92.352
Barometer
33.487
(+) 2.0
Correction
Velocity energy
1.42
Total
energy/mass (ft-lb, /lbm)

92.352
33.487
(+) 2.0
1.42
129.259
129.259
It is to be noticed the correction to the pump center line (potential

energy) and velocity energy each are less than two percent of the total
discharge head in this case.The percentage is even less with higher pressure pumps and, as will be shown later, is partially subtracted in subsequent computations. For this reason, these two factors are often neglected.
It is wise not to neglect them so as not to fall into conceptual traps.
Notice also the sum of the gauge pressure and the atmospheric pressure is
an absolute pressure. A source of error is the omission of the barometric
correction (because in most cases involving incompressible fluids, the
barometric pressure cancels).When dealing with compressible fluids, the
correction is mandatory.Therefore, it wdl always be included in this book.
and Equations
Pumps
Example of metric units: Total discharge head

Four hundred liters per minute of 25C water are pumped at a flange
discharge gauge pressure of 250 kPa. What is the total discharge head in
meters and in joules per kilogram or newton-meters per kilogram?The
gauge center line is two meters below the pump center 1ine.The local
barometer reading is 100 kPa.The pipe is 2 inch, Sch. 40 (not unusual,
even in metric areas).
Discharge
Mass density
Gauge pressure
Meters of fluid
Energy/mass
997 kg/m3
250 kPa ( k ~ / m ~ )
P/(pg) (1,000 X 250)/(997 X 9.805) = 25.574 m
P/p (1,000 X 250)/997 = 250.752 joules/kg
(N-m/kg)
Barometer
Metersoffluid
Energy/mass
100 kPa
(1,000~100)/(997x9.805)=10.229m
(1,000 X 100)/997 = 100.301 J/kg (N-m/kg)
Correction to absolute units

Pressure
Meters
Energy
350 kPa
35.803 m
351.053J/kg (N-m/kg)
Correction to pump (or pipe) center line

Distance
(-1 2.0 m
Energy/mass
(-) 2.0 m X g/gc = 2.0 X 9.805/1 = 19.610J/kg
(N-m/ kg)
Velocity head (a 1)
Flow rate
400 l/m,
0.00667 m3/s
Flow area
0.00216 rn2
Velocity
0.00667/0.00216 = 3.088 m/s
Kinetic energy
~
~ = 3.088~/(2
/
2 x 1)~= 4.678
~ (m/~)~
(N-m/kg) U&)
Velocity head
~
~ = 3.088~/(2
/
2 X~
9.805) = 0.486 m
and Equations
145
Pumps
Total discharge head, hd

meters
Gauge reading
25.574
Barometer
10.229
Correction
(-1 2.0
Velocity energy
0.486
Total
energy/mass (N-m/kg, J/kg)

250.752
101.301
(-1 19.610
4.678
34.289
337.121
The difference between these two examples is that, in the first, the discharge pressure gauge was above the datum line (the center line of the
pump) and the head or energy per unit mass had to be added. In the
second, the gauge was below the center line of the pump and the head
or energy had to be subtracted.
The second difference is that, in the U.S. customary system, the head and
energy per unit mass are numerically identical.The problem comes with
using these units in subsequent equations.This is the reason this book
uses energy per unit mass units. In the metric system, the head unit, the
meter, differs from the energy per mass unit, the newton-meter per kilogram or the joule per kilogram by the factor 9.805, the acceleration due
to gravity, whose units are meters per second squared.
Example of customary U.S. units: Total suction head
We will build on the previous examples.Assume the inlet piping is 4
inch Sch. 40.The measured pressure at the pump inlet, one foot above
the center line, is 6 psig.What is the total suction head?
Suction
Weight density
Mass density
Gauge pressure
Feet of fluid
Energy/mass
Barometer
U.S. units
Feet of fluid
Energy/mass
146
FIOW
62.37 lbf /ft3

62.37 lbm/ft3
6 psig, 864 psf
13.853 ft
13.853 ft-lb, /lbm
100 kPa
14.5 psia, 2,088 psf
33.478 ft
33.478 ft-lb, /lbm
o f industrial F l u i d s - T h e o r y
and Equations
Pumps
Correction to pump center line

Distance
Energy/mass
Theory and
Equations
C H A P T E R
T H R E E
(+) 1.0 ft
(+) 1.0 ft X g/gc = 1.0 ft-lbf /lbm
Velocity head (a 1)
Flow rate
100 gpm,
0.2228 cfi
Flow area
0.0884 ft2
0.2228/0.0884 = 2.520 fps
Velocity
Kinetic energy
u2/2% = 2.522/(2 X 32.17) = 0.0987 ft-lbt/lbm
Velocity head
~
~ = 2.522/(2
/
2 X 32.17)
~
= 0.0987 ft
Total suction head, hs
feet
Gauge reading
13.853
Barometer
33.487
(+) 1.0
Correction
Velocity energy
0.0987
Tota1
48.439
energy/mass (ft-lbf /lbm)

13.853
33.487
(+) 1.0
0.0987
48.439
Example of metric units: Total suction head

Assume the inlet piping is 4 inch Sch. 40 (Again, the reader is reminded
it is not unusual to see American piping in metric areas).The measured
pressure at the pump inlet gauge, one meter above the center line, is 40
kPa. What is the total suction head?
Suction
Mass density
Gauge pressure
Meters of fluid
Energy/mass
Barometer
Meters of fluid
Energy/mass
997 kg/mJ
40 kPa ( k ~ / m ~ )
P/(pg) (1,000 X 40)/(997 X 9.805) = 4.092 m
P/p
(1,000 X 40)/997 = 40.120 joules/kg
(N-m/kg)
100 kPa
(1,000 X 100)/(997 X 9.805) = 10.229 m
(1,000 X 100)/997 = 100.301 J/kg (N-m/kg)
Correction to pump (or pipe) center line

Distance
(+) 1.0 m
(+) 1.0 m
Energy/mass
g/gc = 1.0 X 9.805/1 = 9.805 J/kg

(N-m/kg)
and Equations
147
C H A P T E R
T H R E E
Pumps
Velocity head
Flow rate
Flow area
Velocity
Kinetic energy
Velocity head
400 Urn,
0.00667 m3/s
0.00821 m2
0.00667/0.00821 =: 0.812 m/s
u2/2% = 0.812~/(2X 1) = 0.330 (m/s)2
(N-m/kg) U&)
9.805) = 0.0337 m
~
~ = 0.812~/(2
/
2 X~
Total suction head, hs

Gauge reading
Barometer
Correction
Velocity energy
Total
meters
4.092
10.229
(+) 0.3
0.0337
14.655
energy/mass (N-m/kg, J/kg)

40.120
101.301
(+) 2.941
0.330
144.692

The purpose of this chapter was to give a narrative description of centrifugal pumps, positive displacement pumps, jet pumps and turbines and to
describe how they function when integrated into hydraulic systems.The
functions and the installed and inherent characteristics of pumps were discussed in fair detail. Centrifugal pumps were given much more attention
than the other types because of their importance in the process industries.
An expansion of the technical details WU be found in Appendix AIII.
Pumps were described in terms of their primary function - the transfer of
mechanical energy to a 1iquid.Transferof energy is sometimes lost sight of
in our &scussions of head and pressure.
Six d&erent means of energy transfer were described with the emphasis
given to the most common one in the process industries - centrifugal force.
A brief hlstory of pumps was given to show they indeed are a very
mature technology. "Mature" means well documented and trustworthy.
A commonly accepted, although somewhat arbitrary, classfication of
pumps was given in order to facllitate further description.
The characteristics of various pumps were described.The emphasis was
placed on what is important to those who wish to understand those characteristics that are important to operations and maintenance personnel.
It was shown that centrifugal pumps have turndown limitations and
148
low o f I n d u s t r i a l F l u i d s - T h e o r y
and Equations
Pumps
C H A P T E R
T H R E E
they are sometimes subject to surging problems. It was stressed that

positive displacement pumps cannot be throttled. Flow control must be
by varying the speed of the driver or by diverting some of the flow to
the pump suction or elsewhere.
Miscellaneous types of pumps such as gas lift pumps, jet pumps and
electromagnetic pumps were presented in order to complete the picture.
Inherent characteristics were compared to installed characteristics of
centrifugal pumps. The problem of suction lift and net positive suction
head was touched upon.
Control of flow through pumps was introduced. Emphasis was on the
most common process pump, the centrifugal pump. It was pointed out
centrifugal pumps operating in parallel must have carefully matched
head curves and one pump can actually force the other to the shut off
point if they are not well matched. Pumps in series, booster pumps,
were also discussed.
It was shown many variable speed drives operate in a fixed speed mode
because of sluggishness of the driver and because of the fact impellers
are designed to be most efficient at one flow rate.
Hydraulic turbines used for power recovery were presented briefly.
The chapter closed with a few worked examples.
and Equations
149
Compressible
Fluid Flow
I V - l : SCOPE O F CHAPTER - COMPREHENDING COMPRESSIBLE FLOW
Chapter IV will build upon the previous discussions of incompressible
flow in an attempt to render the phenomena of compressible flow comprehensible. It will discuss and explain the various conceptual models used
to simpli+ the analysis of compressible flow problems. It will attempt to
show the utility and limitations of these models, including the various
equations-of-state. The equations developed for incompressible fluid flow
wdl be shown to be applicable with modifications to many plant situations. The modifications and the range of application will be discussed.
Choked flow
The concept of choked flow wdl be introduced and explained. In particular, the important contribution of Peter Paigemv to the understandmg of
the chokmg phenomenon will be presented.The first step in any compressible flow computation should be to investigate the possibhty of
choked flow. If this is not done, the subsequent computation may be seriously in error in the unsafe &rection.The concept of choked flow will be
introduced in this chapter and specific computation methods wdl be given
in Appendix AIV
Mixed phase flow

Compressible flow includes mixed phase flow. Mixed phase flow is not
easily measured and is controlled only indirectly. It does have to be understood, however, and accommodated. Control systems professionals
encounter mixed phase flow when they size pressure relief devices and the
associated downstream piping.They also come across it when sizing control valves in which flashing flow is occurring. Extensive work has been
done on mixed phase flow resulting in some rather complicated theories.
Fortunately, Paigemv, Richte? and others have been able to develop
working equations that make many mixed phase problems more tractable.
Their work will be discussed in fair detail.
and Equations
151
C H A P T E R
F O U R
Compressible
Fluid Flow
Mach number
A new parameter, the Mach number, will be introduced.The Mach
number is important in computations for flow velocities in excess of
about 0.3 Mach and for noise computations. Mach 1 is the speed of sound
at the condtions of temperature and pressure of the fluid.We will point
out the use of the Mach number concept is limited within the terms of
reference of this book. In order to introduce the application of the Mach
number in solving higher velocity flow problems, the ideal nozzle wdl be
presented briefly. We will then develop three basic models for compressible
flow.The adiabatic model is the one most applicable to the ordinary
industrial flows covered in this book.
More complicated mathematics
In trying to explain the concepts of compressible flow, it is not possible to
avoid some rather complicated mathematics. However, to use the equations developed for compressible flow, it is not necessary to fully understand the derivations. It is necessary to have a grasp of the limitations
involved with certain assumptions so as not to fall into the trap of using
cookbook methods where they do not apply.The mathematical content in
this chapter tends to be greater than in other chapters.We wdl, however,
try to explain the logic behind the mathematics.
IV-2: DIFFERENCES B E T W E E N COMPRESSIBLE A N D INCOMPRESSIBLE F L O W

Many equations of compressible flow are molfications of those of incompressible flow.Therefore, it is important to have a grasp of the dfferences
between the two types of flow in order to be able to judge the fields of
application of the equations.
Density change
An incompressible fluid flowing through a conduit has its density changed
only slightly, often negligibly, due to temperature changes. The temperature change can be due to fluid flow irreversibilities or due to heat
exchange.The first cause results in a small temperature rise that often may
be ignored.The second cause, heat exchange, can result in viscosity
changes that alter the ReynoldsVnumber and the friction factor. Even in
the second case, problems can be solved with a fair degree of accuracy by
using average friction factors over segments of conduit.
Density variations become very important in compressible flow.The
Reynolds number remains important, but at high velocities, the Mach
152
F l o w o f ~ n d u s t r i a lF l u i d s - ~ h e o r y a n d E q u a t i o n s
Compressible Fluid Flow
C H A P T E R
F O U R
number becomes important. Acceleration effects absorb avdable energy and

can result in the cholung phenomenon both for gases and flashing liquids.
Acceleration in uniform conduits

In steady state flow of incompressible fluids in uniform conduits, there is
essentially no acceleration of the fluid. In steady state flow of compressible
fluids in uniform conduits, there is acceleration of the fluid.Although the
velocity at a point is constant, with compressible flow the velocities
increase in the direction of flow. As the molecules of the compressible
fluid separate due to the acceleration, both the temperature and the density of the fluid decrease.These effects are not found with flowing liquids.
An unaccelerated fluid in a conduit has static energy and internal energy
associated with temperature. Mechanical energy and thermal energy are
loose terms for two forms of energy. Mechanical energy is recognized as
pressure-volume energy, as lunetic energy or as energy by virtue of elevation. Thermal energy is energy associated with temperature change. It
becomes internal energy and is often dissipated due to a temperature &fference. Only mechanical energy is useful in producing flow (if we ignore
convective effects from density changes).An unavoidable consequence of
flow is the conversion of some mechanical energy to internal energy that
is no longer useful in terms of flow.
Acceleration requires energy, and this energy is obtained from the available
static energy of the fluid. The result of continued acceleration is more and
more energy is required from the available static energy and less and less is
available for the conversion process - to overcome the irreversibilities, the
losses, the "fluid friction".
Major difference between compressible and incompressible flow

A major difference between compressible and incompressible flow is, with
incompressible flow, the pressure at one extreme of the system is always
known. The problem resolves itself to computing forward or backward
from the known pressure. With compressible flow, it frequently happens
that neither the upstream nor the downstream pressure is known.That is
to say, the source pressure and the sink pressure may be known, but neither the pressure inside the inlet of a conduit nor inside the discharge are
known. A solution to this problem will be given in this chapter.
and Equations
153
C H A P T E R
F O U R
I V - 3 : USING MODELS
Models are useful conceptual tools for simplifying analysis and as an aid
for performing computations. Ideally, in the industrial situation, the use of
a model will produce a reasonable worst-case estimate of the real situation. There are many kinds of models. Some large-scale computer models
make use of many simpler smaller-scale models.
Equations-of-state
Equations-of-state are the primary models relating pressure, temperature,
volume and composition of compressible fluids. A grasp must be had of
the limitations of each equation-of-state. The ideal gas model is the easiest
to use, but it is the most inaccurate. It is fair to say that the more accurate
the equation-of-state, the more complicated it becomes. Also, the more
complicated, the greater the number of parameters needed for its solution.
Sometimes, all parameters are not available.A little knowledge allows one
to make a judgement as to when it is necessary to abandon a simpler
model for a more complicated o,ne.We will compare the ideal gas model,
o dthe
e l , virial model in this
the original ~ e d l i c h - ~ w o n g ~ ~ ~ ~ %and
chapter.
Adiabatic and isothermal models

There are two general models used for compressible flow in closed conduits - the adiabatic model and the isothermal model.These are simplifications that allow further computations.They are not equations-of-state.
For plant work involving relatively short distances with relatively large
changes in pressure in insulated conduit, the adiabatic model is useful. For
pipeline work involving relatively long distances and relatively small
changes in pressure, or for long, uninsulated, small diameter pipes, the
isothermal model is useful.
The adiabatic model is one in which the heat transfer to or from the fluid
across the conduit wall is considered negligible. In the isothermal model,
sufficient heat transfer takes place to keep the temperature of the flowing
fluid constant.
Unsatisfactory models
There are times when neither the adiabatic nor the isothermal model is
satisfactory.When flow is through heat exchangers, or when extremely
cold fluids are subject to heat transfer due to warmer ambient temperatures, changes in density and even vaporization can make the use of the
simpler models completely incorrect. In this case, a model involving heat
transfer is clearly needed.
154
~ I o wo f ~ n d u s t r i a~ ~I u i d s - ~ h e o r y a n d ~ q u a t i o n s
Compressible
Fluid Flow
C H A P T E R
F O U R
Polytropic model
In turbomachinery work, the polytropic model is frequently used. This
model is not useful for industrial piping networks because the heat flow
across the conduit walls is not sufficiently predictable. We wdl not be concerned with this model in our discussions.
Main models
The main models with w h c h to be concerned in the industrial setting are
the equations-of-state, the adiabatic model, the isothermal model and the
model for heat transfer. Whenever a model is being used in place of exact
computations, it should be examined critically to see it really applies to
the situation at hand.
The alabatic model assumes there is no energy flow to or from the fluid
under a temperature difference. This may be a reasonable assumption in
many cases when the conduit is short or is well insulated and, when the
fluid is hotter than the environment; it may even give a factor of safety.
However, when very cold or cryogenic fluids are flowing (flare headers,
for instance), there is a flow of energy from the environment to the fluid.
The enthalpy (internal energy plus Pv energy) of the fluid increases, as
does its specific volume.The friction factor is greater. Choking may occur.
The adiabatic model, in this case, can cause errors in conduit sizing in a
direction that is not conservative
Given the fact it is almost impossible to predict overall heat flow in a
piping system with any accuracy and given the fact the ahabatic model is
conservative in most cases (except the one mentioned in the previous paragraph), it is a good model for most plant situations. However, it is always
wise to check the assumptions. If the temperature lfference between the
fluid and the outside environment is not too great or if the conduit is well
insulated, the adiabatic model is usually a good one. If a gas or vapor is
deliberately being heated or cooled, the alabatic equations do not apply; at
least, not at the point of the heat exchange. Note the word "alabatic"
simply means the absence of heat energy flow.T h s situation may be
obtained by deliberately heat tracing a conduit or a vessel and controhng
the temperature of the jacket so its temperature is the same as the fluid.
In the average plant, most flows are kept below about 0.3 Mach to avoid
excessive industrial noise and potential damage due to vibration. This is
the flow range where modified compressible flow equations (adiabatic) are
most useful.
and Equations
155
C H A P T E R
F O U R
The isothermal model is a tempting one to use because isothermal equations are simpler than the adabatic ones. It is clearly incorrect to use ths
model for most plant computations involving large changes in pressure and
temperature.The fluid densities predcted by this model may be grossly in
error almost everywhere but at the starting point of the computations.
Using computer simulations

It is not possible to perform more than trivial compressible flow computations without the use of a computer. Usually, there are too many variables
that change simultaneously to allow hand computations. The simplifying
assumptions involving ideal gases do not always hold. Some mixtures (but
not all) of varying composition require some complicated equations-ofstate. A good computer simulation program can reduce an otherwise
untractable problem to manageable proportions.
It is the author's strong conviction that lack of knowledge of basic principles on the part of some users causes computer simulations to be run in a
cookbook fashion. Unchecked, incorrect assumptions can lead to major
errors - even when computers are used. A good grasp of the basic equations of compressible flow gives a comfort level that otherwise does not
exist.The careful reader should obtain this understanding of the basic
equations from this book and should achieve the ensuing comfort level.
IV-4: TREATING MIXTURES

For didactic purposes, it is convenient to deal with pure (single component) fluids. In real industrial life, such fluids are rare. Even natural gas,
which we tend to think of as a pure substance somewhat related to
methane, varies in its properties depending on its source. It varies so much
that gases from dfferent sources frequently are blended in order to obtain
a more uniform mixture.
In refineries, in particular, the gases can be mixtures of many different
components: H,, CO, CO,, O,, N,, H,S, CH,, C2H6,
C4HI0,and
many, many more. In order to measure the flows, one must know the
mixture properties as a function of the component properties. Even air is a
mixture and one tends to forget the trace components such as argon and
carbon dioxide.
The properties of gases are functions of their temperature, pressure and
composition. These properties influence the flow characteristics of the
156
F I O W o f ~ n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Compressible
Fluid Flow
C H A P T E R
F O U R
mixture and the response of measuring and controlling devices largely

through the overall viscosity and overall density.
Composition of a gas or vapor must be known or be estimated before
meaningful computations can be made.The process flow diagram and its
associated material balance are the normal sources of information on
composition.
The equation-of-state used to describe a pure component must be
modified to describe a mixture. Some equations-of-state describe pure
substances quite well and mixtures very poorly The original R e a c h ~ w o n r equation
'
works with reasonable accuracy for both pure
substances and for mixtures.
Mixing rules
A set of "mixing rules" must be decided upon.These rules give the correlation between the single component properties and the mixture properties. In many instances, these rules are laid down by an organization - a
company, a gas association, etc. In other cases, they are chosen based on
the need for accuracy or based on balancing accuracy against complexity.
We will give some general guidelines later.
IV-5: EQUATIONS OF COMPRESSIBLE FLOW OF A N IDEAL GAS

It is frequently convenient, when dealing with complicated relationships,
to start from an idealization. Once the basic equations are worked out, it is
possible to compare the ideal situation with the real one and to make
allowances for the differences.This is why we start with the ideal gas.
An ideal gas is simply one that follows the simple relationship Pv = RT,
where P is the absolute pressure, v is the volume per unit mass, R is the
particular gas constant on a mass basis (the universal constant, R,, molar
basis, divided by the molecular weight) and T is the absolute temperature.
Most gases follow the ideal gas relationship (often loosely called a law) at
sufficiently low pressures and high temperatures. Corrections are necessary
for higher pressures and lower temperatures.
Some gases such as carbon dioxide and the vapors of organic acids deviate
strongly from the above ideal gas relationship - even at low pressures and
high temperatures.With these gases and vapors, the ideal gas equation is
inaccurate. Nevertheless, it is stdl a very useful relationship, especially as a
first approximation - as a base for further computations.
and Equations
157
Fluid Flow
C H A P T E R
Usual a ssumptions for the equations of ideal gas flow

We will state the usual assumptions used in an ideal gas analysis in closed
conduit, as follows:
1. steady-state, one dmensional flow;
2. ideal gas as defined above with a constant heat capacity, c
P'
3. no shaft work (we are deahng with flow through conduits, not pumps,
compressors or turbines);
4. gravitational affects are negligible;
5. velocity gradents withm a cross section are negligible (the coefficient

alpha in the ~ e r n o u l hequation
~
equals approximately one).
Basic relationships to be developed

Some relationships are more important than others are.The ones we are
interested in are:
1. the equation of continuity;
2. the steady state, total energy balance;
3. the total energy balance with irreversibilities due to shear;
4. the equation for the velocity of sound in a fluid;

5. the ideal gas equation-of-state.
We wdl now develop the above relationships with some explanations.
Continuity
The continuity relation states that, under the assumption of steady state
flow without accumulation of matter, the mass flow rate is constant. If the
mass flow rate is measured at several points it wdl be found to be the
same. However, the mass flow rate can be factored into velocity, conduit
cross section area and fluid density.Average velocity, U, changes in compressible flow.The area of the conduit, A, may or may not change. Density
definitely changes.The product of all three terms remains constant in
steady-state mass flow.We can take logarithms in order to separate the
terms and differentiate in order to get rid of the constant terms and to
produce a differential equation as follows:
158
FIOW
o f ~ n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
C H A P T E R
F O U R
m = UAp = const
0
lnm = l n U + l n A + l n p
0
dlnm = d l n U + d l n A + d l n p = 0
dU dA dp
-+-+-=o
U
A
P
This differential equation (IV-l) allows exploring the relationships among
average velocity across a section, area and density.The only assumption
was that of a constant mass flow rate.The differential equation also allows
the development of integrated equations that are simpler to grasp.
Total energy balance, steady state
The first law of thermodynamics is a total energy balance. It is not so
dnectly useful as the ~ e r n o u hequation
'
for fluid flow work. However, it is
necessary to have an understandmg of the first 1aw.The first law of thermodynamics expressed in terms of open (flowing) systems on a rate basis is,
In S1 work, the reader is again reminded the dimensional constant, g,,

equals one and has no dimensions. In addition, the "mechanical equivalent
of heat", J, is 778.16 ft-lbf /Btu in customary U.S. units and is one and is
dimensionless in S1 units.The above equation is therefore valid for both
systems.The two enthalpy terms, h, and h2, are not hydraulic heads. All of
the terms in the brackets have the same units of energy per unit mass.
The heat energy and fluid work terms, Q and W f , are those amounts of
energy that cross the control volume during the period that the amount
of mass transits the same control volume.A dot above the three terms, Q,
Wf and m, puts them on a rate basis, heat energy flow, work flow and mass
flow per unit time. If equation IV-2 is dwided through by the mass flow
rate, the heat energy and shaft work terms are on a unit mass flowing
basis. Lower case letters would then designate them. Note fluid work is
often incorrectly called shaft work. We have used the term fluid work to
emphasize that mechanical work to or from the shaft differs from
mechanical work to or from the fluid because of inefficiencies within the
pump or turbine mechanism or because of irreversibilities in the fluid.
We can impose constraints on the above equation such as limiting the
and Equations
159
C H A P T E R
F O U R
analysis to one in which there is no change in elevation, Z2 = Z1.This

constraint usually does not introduce a major inaccuracy when the fluid is
not too dense and the elevation change is not too great. (This assumption
should always be checked with real numbers).We have already stated that
another constraint can be that there is no fluid work, Wf , done on or
extracted from the fluid, so we can drop the term containing Wf .
The above constraints allow us to simplify the first law and to put it into a
differential equation as follows:
3 -a4=- dU2 +dh

0
2gcJ
The Greek delta on the heat energy transferred, Q or q, is to remind us

thermal energy is not an exact differential, in other words it must be integrated over a particular path.The value of the integral does not depend
only on the limits of integration. The last equation states that differential
heat energy transfer per unit mass flowing is equal to the algebraic sum of
differential lunetic energy and the differential enthalpy.
Again, we have to remember the constraints or limitations we have
accepted. If necessary, we can always go back and restore variables that
have been neglected.The above equation relates changes in heat transfer
across the wall of a conduit per unit of mass flowing to changes in the
fluid's lunetic energy and enthalpy. In the case of adiabatic flow, we can
impose a further constraint by making the heat energy flow term equal to
zero.We wdl do this later in the analysis of adiabatic flows.
Mechanical energy balance modified to include 'losses'

The ~ernoulli'equation is often referred to as a mechanical energy balance. However, when the loss term, hL is included, the equation also contains a term describing mechanical energy converted to internal energy or
heat flow, or both.
(IV-4)
The above form of the equation is valid for liquids and for gases whose
density change is known.The alpha correction to each lunetic energy term
160
F l o w o f ~ n d u s t r i a lF l u i d s - T h e o r y
and ~ q u a t i o n s
C H A P T E R
F O U R
adjusts ths term to correct for velocity profiles across a section of conduit.
The velocity in the kmetic energy term is the average velocity across the
section.All terms are on a per unit mass basis.The first three terms on each
side of Equation IV-4 are functions of the particular sections.The fluid
work term represents work supplied to the fluid anywhere between points
1 and 2. Sirmlarly, the "loss" term is the accumulation of mechanical energy
converted to internal energy, enthalpy, heat flow and anythmg else, between
sections 1 and 2 - it is not "lost", in spite of the name.
'Skin friction'
Equation IV-4 can be made more applicable to turbulent, compressible
flow in a conduit by considering alpha to be equal to one, no compressor
in the segment under consideration, constant density over a short length
and all the losses to be due to "skm friction". The term "skm friction" is a
synonym for irreversibilities in a conduit, and it arises from the fact there
is a more rapid change in velocity profile close to a solid wall and a less
rapid one remote from the wall. Most of the shear resistance to flow arises
nearest a wall. However, in a conduit, shear exists across the entire cross
section. It is simply more intense near the wall.The subscript, 6,sometimes
used on the loss term emphasises this fact.
(W-5)
4-9 s(z2-zl)+u,"-~:
+hg = O
P
gc
2gc
Putting Equation IV-5 into differential form and considering elevation
changes to be negligible gives
(W-6)
P
Equation IV-6 describes changes over a very short (differential) length of
horizontal conduit.
The differential form of the d arc^"" equation, Equation 11-3, for horizontal conduits of circular section is repeated as Equation IV-7.
u2dL
dhg = fM--
D
In Equation IV-7, the average velocity, friction factor and diameter are
considered constant over the differential length of conduit.
QC
For non-circular conduits in turbulent flow, D = 4rH.The equivalent diameter equals four times the hydraulic radius.The hydraulic radius is defined
and Equations
161
C H A P T E R
F O U R
as the conduit cross sectional area dvided by the wetted perimeter.

Combining equations IV-6 and IV-7 and using the concept of the
hydraulic radius gives Equation IV-8.
This last equation is a differential equation suitable for developing relationships for compressible flow in a straight, horizontal conduit of any section. Because t h s is a differential equation, the assumption of incompressibility can be removed. If the conduit is circular, four times the hydraulic
radius may be replaced by the diameter.
Sonic velocity
Most industrial flows are deliberately kept to Mach numbers below about
0.3 Mach in order to avoid excessive noise and vibration in piping.
Therefore, it is rare the Mach number concept is needed in the context of
ordinary industrial flow in conduits.The Mach number is important when
dealing with turbo-machinery. We wdl not dwell on the subject beyond
the time necessary to grasp the concept. We will also give simpler methods
of handling related problems in a later section.
The velocity of sound in the compressible fluid becomes important when
computing the parameter called the Mach number.The velocity of sound
in a fluid was first given by Laplacem' as a correction to Newton's original hypothesis. Newton had assumed the transmission of sound in fluids
followed an isothermal process. Laplace showed that, in order to make
Newton's theory match experimental data, the process had to be considered to
be adiabatic, not isothermal.
The resulting acoustic velocity is given by the mfferential equation,
Equation IV-9.
The velocity of sound is equal to the square root of the product of the
dnnensional constant and the partial derivative of pressure with density at
constant entropy.The Mach number is defined as the ratio of the actual
fluid velocity to the above sonic velocity. Mach one is reached when a
fluid travels with a velocity equal to the velocity of sound in the fluid. It is
162
F I O W o f ~ n d u s t r i a ~~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
C H A P T E R
F O U R
usually more farmliar as the velocity of an object, such as an aeroplane,

travelling in a fluid, such as air.
It is to be noted the choked flow velocity and sonic velocity are not
always the same. More on this a little later.
Ideal gas equations
Ideal gas equations simpli6 analysis. For many industrial situations they are
not too far from reality. However, this last assumption must be re-examined
each time it is made. We wdl discuss this problem more thoroughly in
Section IV-6: Ideal and Non-Ideal Gases - Com~arisonof Some
An ideal gas is one that obeys the following equation-of-state:

Pv = R T
(W-l 0)
The first gas constant, R , is the specific gas constant for a particular gas. Its
units are energy per unit mass per degree.The numerical value of the gas
constant depends upon the units used in the formula.The second gas constant, Ro,is the so-called universal gas constant.The "universal" gas constant is on a molar basis. It must be divided by the molecular weight to
obtain the specific gas constant on a mass basis. In customary U.S. units,
R. equals 1546 foot-pounds-force per pound-mole-degree Rankm. In S1
units, it equals 8,314.3 joules per kg-mole-degree Kelvin.
Should a gas constant using a Btu as the energy unit be chosen, the
"mechanical equivalent of heat", J, must be a factor
By taking logs and differentiating, we can separate the individual terms
and get rid of the constant terms in IV-10.
dlnP = d l n p + d l n T
dP - d p dT
--+P
P
T
and Equations
163
C H A P T E R
F O U R
Compressible
Fluid Flow
An assumption frequently made for the sake of simpliqing the analysis is

the specific heat at constant pressure is independent of temperature over
the range of the analysis.When this assumption is made for an ideal gas,
the specific enthalpy becomes proportional to temperature only. We can
obtain the following dfferential equation:
That the constant heat capacity assumption must be carefully made can be
verified by looking at graphs of heat capacity versus temperature.
Assumptions must not be made arbitrarily. Fortunately, the ratio of specific
heats, cP/cv or Y,varies with temperature less than does either of the two
specific heats.This is the justification for the ratio of specific heats to be
taken as constant in what follows.
The word 'isentropic'
The term isentropic means constant entropy. Practically, it means the
process (the change being considered) is close to ideal, without noticeable
irreversibilities, and is adiabatic.There is no heat energy flow.The only
way to change the value of the quantity known as entropy in an ideal
process is by allowing energy to cross the boundary under a temperature
gradient.This flow of energy under a temperature gradient is called heat
flow. The energy involved is often loosely called "heat".
Ideal gases w i t h constant y, adiabatic, reversible processes
The reader should consult any text on basic thermodynamics if he or she
wishes a more thorough understanding of some of the derivations that
follow. For an ideal isentropic (adiabatic,reversible) process involving an
ideal gas with a constant ratio of specific heats, the following relationships
are valid:
PVY = const
= const
--
164
/y
- const
and Equations
Compressible
Fluid Flow
C H A P T E R
F O U R
The last three constants are not identical and the adiabatic, reversible constraint must be followed. A gas flowing isentropically has to follow an adiabatic process because of the definition of entropy as 6q/T.The molar
mass is introduced so as to put the gas constant on a mass basis and J, "the
mechanical equivalent of heat", is introduced because the heat capacities,
in North America, generally have Btu's as their energy unit.
We can use Laplace7sm1equation, Equation IV-9, to find the sonic
velocity of an ideal gas under the isentropic constraint as follows,
(W-14)
= const = -
py
dlnp-dlnpY = dlnconst = 0
The isentropic pressure dependence on density is seen to equal the

product of the ratio of the specific heats and the ratio of the absolute pressure to density for an ideal gas.This equivalence can be plugged into
Laplace7sequation, Equation IV-9.
Equation IV-15 can be used to obtain the velocity of sound in air taken as
an ideal gas at 60F (520 R) with the ratio of specific heats equal to 1.4, the
molecular mass equal to 29, the universal gas constant equal to 1546 filbdlbmol-Rand the dimensional coefficient equal to 32.17 lbKfi/lbfs2.~he
velocity of sound in air under these con&tions is 1117 ft/s or 340.5 m/s.
The acoustic velocity is seen to be a function only of the absolute temperature for a given ideal gas with constant ratio of heat capacities.The
Mach number for an ideal gas can be written as follows (the variable, a, is
the acoustic velocity),
and Equations
165
C H A P T E R
F O U R
Equation IV-16 might look fairly simple, but average velocity, U, temperature, T, pressure, P, and density, rho, must all be measured, or estimated at
flowing conditions at the point of interest.
I V - 6 : I D E A L A N D N O N - I D E A L GASES
EQUATIONS-OF-STATE
COMPARISON OF SOME
Ideal gas
The ideal gas model assumes gases are made up of spherical molecules
that do not interact with one another except during elastic repulsions.At
low pressures and high temperatures this model is not too bad an assumption for most gases. However, molecules do interact with one another to a
degree dependent on their size, shape and physical or chemical affinity. At
pressures and temperatures approaching the critical point (the peak of the
vapor dome), the ideal gas equation is so inaccurate as to be unusable.
Corresponding states
Corresponding states theory holds that gases, when reduced to fractions
their critical values, all behave in the same manner. The theory gives an
approximation of true behavior. A dimensionless correction often is
applied to the ideal gas law, Pv = RT, in the form of the compressibility
factor, Z.This gives Pv = ZRT. The correction factor, Z, can be estimated
from corresponding states considerations. Its use can change an accuracy
of plus or minus 20% to plus or minus 5%, as a broad example involving
highly non-ideal gases.
Cubic equations-of-state
The Dutch physicist, van der waalsmii, circa 1870, was one of the first to
try to improve the ideal gas model with a model that involved two parameters. One parameter allowed for the volume occupied by the molecules
and the other allowed for molecular interaction - simultaneous repulsion
and attraction.The van der Waals model stimulated several improvements
of which the original R e d l i c h - K w o n y equation, circa 1950, is probably still the most famous.The original R-K equation is particularly useful
for mixtures of gases. It is std only a two-parameter equation, but it gives
surprising accuracy for many gas mixtures.The van der Waals equation
now been largely replaced by other two and three parameter models.
166
F l o w o f I n d u s t r i a l ~ l u i d s - ~ h e o r ya n d E q u a t i o n s
C H A P T E R
F O U R
We will give the original R-K equation here, since it is the equation with
which we will work most frequently in this book.The gas "constant" and
the volume are molar quantities. Mass quantities could be used with a
mass gas constant.
The constants, a and b, are functions only of the critical pressures and
temperatures of a substance. In addition, fairly straightforward mixing rules
allow the constants to be altered to accommodate mixtures.
The first set of coefficients, a and b, refers to pure components.The

second set refers to mixtures.The summations involving the pure components allow computation of the corresponding coefficients for the mixtures. Mixtures are more common than pure components in the industrial
situation. Even dry air is a mixture generally given as nitrogen (0.7809),
oxygen (0.2095), argon (0.0093) and carbon dioxide and trace elements
(0.0003).Air is usually treated as a pure component that obeys the ideal
gas equation. At cryogenic temperatures, this is no longer acceptable. It has
to be treated as a mixture.
In order to improve the accuracy of the R-K equation, a third parameter
was introduced.This parameter was supposed to give a measure of the
non-spherical nature of many molecules. This is the so-called acentric
factor. Soave moddied the R-K equation using this parameter.The result
was the R-K-S equation. The Peng-Robinson equation is very simdar in
structure to the R-K-S equation. Reid, Prausnitz and PolingXXU is a good
source of the most widely used equations-of-state.
Addtional parameters called interaction parameters have been introduced
to describe the interactions between different binary pairs of molecules.
and Equations
167
C H A P T E R
F O U R
Low density gas molecules are assumed to interact between only two
molecules at a time - hence, binary interactions.All of these parameters
improve accuracy at the expense of computational effort.They make the
use of the computer mandatory.
The v i a l equation
The above mentioned equations-of-state are basically empirical.An equation-of-state with a firm theoretical background is the virial equation.
However, in spite of being preferred by many because of its theoretical
basis, it is not possible to use it over wide ranges of pressure. Nevertheless,
PrausnitzXXX
developed a solid database for its use in normal plant situations - less than several atmospheres and at temperatures between 200K
and 600K.
These equations and others are described in any good text on thermodynamics. Here we wdl draw comparisons among the more common
equations.
Comparisons among common equations-of-state
Mixed gases consist of molecules of different sizes and shapes moving at
velocities inversely proportional to their molecular mass. The pressure on
the walls of a container is proportional to the number of repulsions of
molecules per unit time.The density of a gas is dependent on the number
of molecules contained in a given volume and on the molecular masses.
The temperature of a gas is related to its internal energy. An equation-ofstate is a relationship among pressure, temperature and molar (or mass)
volume. Mixtures of gases complicate the problem by introducing molecules that have different sizes and dfferent affinities for one another.
To demonstrate the considerations that go into the choice of an equation-ofstate,Table IV-1 has been prepared. T h s table is based on a common gas, but
not necessarily a well behaved one - carbon &oxide.The table takes experimental data published in PerryXand gives the error in pressure by the application of three equations.The equations are the ideal gas equation, the virial
equation truncated after the second term and the original R-K equation.
The table covers a range of common process pressures and temperatures,
300K to 600K and one bar to 60 bar (a maximum pressure of 20 bar is
more common).We have limited the computations to three temperatures,
300K, 450K and 600K. These three temperatures should be sufficient to
show trends.
168
low o f I n d u s t r i a l ~ l u i d s - T h e o r y a n d E q u a t i o n s
Compressible
Fluid Flow
C H A P T E R
F O U R
Table IV-1. Comparison of equations-of-state-for CO, with experimental data
Actual Data
P, bar, versus
vtm31kg
Temperature
300K
450K
600K
Ideal Gas Computations

300 K
P, bar
ERR%
450 K
P, bar
ERR%
600 K
P, bar
I ERR%
Virial Equation Computations

300 K
P, bar
1.ooo
4.995
9.970
19.792
29.386
38.447
46.1 23
50.686
ERR%
0.0 1
-0.1 0
-0.30
-1.04
-2.05
-3.88
-7.75
-15.52
450 K
P, bar
1.ooo
5.000
9.999
19.987
29.993
39.996
49.812
59.536
ERR%
-0.02
0.00
-0.01
-0.06
-0.02
-0.01
-0.38
-0.77
600 K
P,bar
1.000
5.002
10.002
20.018
30.048
40.020
50.102
59.900
ERR%
-0.06
-0.1 5
-0.30
-0.61
600 K
P,bar
1.000
4.995
9.976
19.917
ERR%
-0.00
0.03
0.02
0.09
0.16
0.05
0.20
-0.17
Redlich-KwonqComputations
300 K
P, bar
1.OOO
4.999
9.993
19.945
ERR%
0.02
-0.02
-0.07
-0.28
450 K
P, bar
0.999
4.993
9.970
19.879
ERR%
-0.03
-0.1 0
-0.24
-0.41
NOTE: Actual data from Perry - Computation by author

and Equations
C H A P T E R
F O U R
Compressible
Fluid Flow
It can be seen that carbon dioxide gas approaches the ideal state at sufficiently high temperatures.AU gases do this. By "approaching the ideal
state" is meant the gas behavior can be described by the ideal gas equation-of-state. It can also be seen that even when temperatures are low,
carbon dioxide gas approaches the ideal state when the pressure is sufficiently 1ow.This is also true for all gases.
The "critical point" is a critical parameter. It is the highest point on the
liquid-vapor dome. It is the point at which it is impossible to distinguish
between a liquid and its vapor.The closer the state is to the critical point,
the more difficult it is to obtain good accuracy with an equation-of-state.
Ideal gas equation
The ideal gas equation-of-state is the simplest and potentially the most inaccurate one. Simply by including the compressibility factor, Z, the accuracy of
thls equation can be improved several fold.The compressibhty factor is
obtained by computing the reduced pressures and temperatures - the ratios
of the actual pressures and temperature to the critical ones.These ratios are
then applied to the compressibhty charts found in Per+ or in any thermodynamics textbook in order to obtain the cornpressibility. Note the conipressibhty factor is not constant; it varies with temperature and pressure.
Example I V - l : Corrected ideal gas equation
Carbon dioxide at 300K (2635C) and 20 bar has a specific volume of
0.0255 cubic meters per kg.When this specific volume is used in the ideal
gas equation without the cornpressibility correction, it has been shown
that the computed pressure is plus 11.12% in error. Recompute the pressure using the compressibility factor, Z.The critical pressure and temperature of carbon dioxide are 73.8 b and 304.1K respectively.
From the generalized compressibility charts of Nelson and Obert, the

compressibility, Z, can be read as approximately 0.89. From Table 3-166 in
Perry, it can be interpolated as 0.8871. The corrected pressure is then,
P =ZRT/v =0.8871(0.08314)300/(0.0255)(44.01)= 19.717 bar
Error = lOO(19.717 - 20.0)/20.0 = -1.417%
170
and Equations
Compressible
Fluid F
0 1
C H A P T E R
F0011
The error in the pressure has been reduced from plus 11.12%, using the
ideal gas equation, to minus 1.42% using the compressibility correction.
This type of accuracy is adequate for many problems such as sizing control valves. It permits quick results of acceptable accuracy. It is not adequate for flow measurement, when custody transfer is involved, for
instance. Also, when the pressure becomes very high or the temperature
becomes low, greater inaccuracy may be expected.
The virial equation

The virial equation-of-state is an equation in reciprocal molar volume.
The equation is given as
The virial coefficients, B (second), C (third), etc., are functions only of

temperature and composition.A great deal of effort has been expended on
developing the coefficients, but only the second virial coefficient has
extensive data available for it. By truncating the virial equation after the
second term, a very simple equation of surprising accuracy results.
Reasonable industrial accuracy is limited to densities less than about half
the critical density (Prausnitzy .
The virial equation is theoretically based and its coefficients are simply
related for mixtures. For instance, the virial equation truncated after the
second term has a mixture coefficient given by
Note inherent in the summation of only binary pairs is the assumption

that only binary interactions are possible with low density gases.
The coefficient Bii is a pure component coefficient.The coefficient Bij is
called the cross coefficient and Bij is equal to Bji.The cross coefficient is
computed using mixing rules (See Perryx 3-268 for an extensive and
easily followed discussion).
and Equations
171
C H A P T E R
F O U R
PrausnitzXXX
gives a rule of thumb for the pressure and temperature lirnitations of the truncated virial equation for use in typical process situations.
If we apply the above rule of thumb to carbon dioxide at 300 K, we find

the pressure should be limited to, P I (300/2)73.8/304.1 = 36.4 bar.This
is the point in Table IV-1 where the absolute error exceeds about two
percent.This pressure corresponds to approximately 528 psia so we see the
first column of the table gives a false impression of the virial equation at
higher pressures when not limited by the above rule of thumb.
Frequently encountered concepts
There follow two concepts frequently encountered and sometimes useful
when makmg approximations to real cases.
The asterisk condition
Equations sometimes can be simplified by referring them to the same
normalized conhtion, the sonic velocity. The asterisk condition is defined
as being one in which all properties represent those associated with the
speed of sound in the medium.
The stagnation condition, adiabatic processes

Another frequently used reference condition is the stagnation condition.
The condition is defined as the group of properties that results when fluid
traveling at high speed is brought to rest adiabatically with no developed
work. The action may be hypothetical.The first law equation can be
written for this condition as,
L
Subscript 2 represents the stagnation condition when the fluid is brought

to rest. Subscript 1 is the upstream condition. If the process is adiabatic,
Q and q equal zero. When the fluid is brought to rest, U2 equals zero.
The conversion factor,J, is the "mechanical equivalent of heat".
172
~ I o wo f i n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Compressible
Fluid Flow
C H A P T E R
F O U R
For the adiabatic case, the description of the stagnation process is
Subscript 2 is usually changed to an S (or ST) to represent the stagnation

condition. For a gas that follows the ideal gas equation, enthalpy is a function only of temperature and, if the specific heat is reasonably constant
over the temperature range, the equation is rearranged as follows,
The specific heat at constant pressure does not appear in the last equation.
For an adiabatic process,T and h (flowing conditions) will vary, but Ts and
hs (stagnation conditions) wdl be constant.
and Equations
173
C H A P T E R F O U R
I V - 7 : M O D E L PROCESSES FOR C O M P R E S S I B L E FLOW

The purpose of most models is to isolate the essentials for study and to
eliminate factors that have minimum impact on the results of the analysis.
McCabe and smith- give three basic models for compressible flow.
These models are shown in Figure IV-1.
Convergent
sect!on
Divergent section
Reservoir
fb)
Receiver
fricf ion section

/sothermu/
Now
Heufing jacket]
Figure IV-1.Three models of compressible flow: (a) lsentropic expansion in convergent-divergent nozzle. (b) Adiabatic frictional flow. (c) Isothermalfrictional flow.
174
~ I o wo f ~ n d u s t r i a~ ~~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
C H A P T E R
F O U R
Frequently, models are combined. For instance the three basic models shown
in Figure IV-1 may be combined with the ideal gas model.Ths latter
model may be further restricted to a gas that has a constant heat capacity
independent of temperature.All of these restrictions must be borne in mind
when using equations developed from the use of models.The accuracy of
the results depends upon the closeness to reihty of the assumptions.
Each model has a different utility. The first model, an isentropic, adiabatic
expansion, helps us understand flow in turbomachinery and gives us the
equations for use with near sonic, sonic and supersonic flow.The second
one, adiabatic flow with irreversibhties, is useful as the normal envelope
for plant flows. It gives us the limits within which the compressible flows
of most plants will be found.The thrd model, isothermal flow with irreversibilities, is useful for pipeline work and for work involving long, small
diameter, uninsulated conduits.
The models of Figure IV-1 will now be described.
Isentropic, adiabatic expansion

An adiabatic expansion is one that takes place without energy transfer due
to a difference in temperature between the fluid and the environment.
Isentropic means that, in addition to the adiabatic constraint, there are no
irreversibhties - no conversion of mechanical energy to internal energy,
no 1osses.The process is considered ideal. If there are no losses and no heat
energy transfer, the entropy property remains constant.The real-world
approximation is a well-insulated, well-designed, smooth nozzle.
The principal causes of entropy change are heat transfer, turbulence due
to wall asperities or to sudden expansion.The first cause is already mostly
eliminated by the use of a well-insulated conduit.The remaining causes
can be virtually eliminated by the use of a carefully shaped nozzle.
The ideal nozzle carefully contracts the flowing stream in the convergent
segment so as to accelerate it to the throat.Then, it even more carefully
expands the stream in the divergent segment to either recover the original
pressure or to accelerate the flow even more in the case of supersonic flow.
The essentials of an ideal nozzle are:
the shape;
an upstream reservoir, the source, and a downstream receiver, the sink,
where condtions are essentially stagnant;
and Equations
175
C H A P T E R
F O U R
a smoothly turned convergent segment;

a throat, the narrowest portion of the nozzle, which is the transition
between the convergent and the dwergent segments;
a divergent segment with a very small angle of divergence (less than
about seven degrees);
good insulation.
From the previous &scussion of stagnation temperature, it can be seen that
Ts is constant throughout the ideal, isolated nozzle at steady state.The
measured temperature,T, and the velocity will vary along the nozzle.
Adiabatic irreversible flow

The essentials of the model for adabatic flow with friction (conversion of
mechanical to thermal energy) are:
the same two reservoirs connected by a conduit of constant cross section;
a smoothly turned, isentropic inlet to eliminate disturbances which are
not due to conduit irreversibilities;
good insulation.
This is the model that will concern us most of the time, but one should
not assume it is the only model.
The stagnation temperature,Ts, is std constant in this model at steady
state. The flowing temperature,T, wdl vary.
Isothermal irreversible flow

This model results in equations simpler to use than the adiabatic frictional
flow model. It is therefore frequently used by default. Assumptions should
always be checked.The essentials are:
the same two reservoirs connected by a conduit of constant cross section;
a smoothly turned, isentropic inlet to eliminate disturbances which are
not due to conduit friction;
a constant temperature jacket with good heat transfer.
The stagnation temperature is not constant in this model at steady state.
The flowing temperature is constant.
176
Compressible
Fluid
Flow
C H A P T E R
F O U R
From the equation,
It can be seen that, with a constant flowing temperature,T, the stagnation

temperature must vary with changes in the velocity squared. The velocity
being measured is at the same point that the temperature is measured.
Flow of gases through ideal nozzles, qualitative discussion

Flow through an ideal nozzle is the starting point for understanding flow
through venturis, orifice plates, restriction orifices, chokes and safety valves.
The convergent segment of an ideal nozzle is usually short because separation of the flowing fluid from the conduit boundary cannot occur. Flow
separation is a major cause of irreversibilities because of the subsequent
destruction of:eddies that may occur remote from their place of generation.The divergent segment is long because the angle must be small to
avoid separation of the flowing fluid from the boundary.
In the convergent segment the velocity increases, the pressure, density and
temperature all decrease. In the divergent segment, the velocity may
decrease or increase.At the throat, the velocity can reach a maximum of
sonic velocity. Mach numbers greater than unity are not observed in a
convergent segment or at the throat.
When the velocity reaches choked velocity in the throat, the mass flow rate
is observed to remain constant in spite of decreasing downstream pressures.
The only way to increase the mass flow rate once sonic velocity is obtained
is to increase the upstream, reservoir pressure. Decreasing the receiver pressure no longer has any affect.This is the choked-flow conhtion.
This phenomenon is shown in Figure IV-2 which is given by McCabe
and Smithm. When the ratio of the receiver pressure to the reservoir
pressure is equal to one, there is no mass flow. As the receiver pressure is
decreased so that the ratio decreases, the mass flow rate increases as shown
in the figure. A point is reached, however, around a ratio of 0.5, depending
on the gas, when the mass flow rate becomes constant, independent of the
downstream pressure.The flow is said to be choked.
and Equations
177
C H A P T E R
F O U R
Compressible
Fluid Flow
Pressure rottb
p,/p,
1.0
Figure IV-2. Mass flow rate through nozzle
The nozzle flow regimes are depicted in Figure IV-3.The meaning of the
symbols must be understood.The abscissa is the length along the ideal
nozzle.The left ordinate is the ratio, P/Po, of the pressure at a point (not
the receiver pressure) to the reservoir pressure.The right ordinate is the
receiver to reservoir pressure ratio. It simply fixes the end points of the
various curves.
The reservoir pressure remains fixed at PO.The variable pressures are those
down the length of the nozzle and in the receiver.
The following phenomena occur:
Figure IV-3.Variation of pressure ratio with distance from nozzle inlet
178
and Equations
Compressible
Fluid Flow
C H A P T E R
F O U R
Line a - a'. T h line describes the situation where Pr equals Po.The

pressure, P, is constant along the entire length and is equal to both the
reservoir and the receiver pressures. No flow occurs.
Line abc. When the receiver pressure is slightly lowered, flow occurs.
The pressure, P, varies with &stance along the nozzle, L.The throat pressure is the lowest pressure in the nozzle (vena c~ntractapressure).The
line abc is representative of many lines that could be drawn so that they
terminate between points a' and e.They all have sirmlar characteristics: a
throat where the pressure is lowest and the velocity hghest and some
pressure recovery as the velocity subsequently decreases. Mass flow
increases as the line dips toward point d, at the vena contracta.
Line ade. T h s line represents the situation of choked flow with sonic
velocity at the throat but with pressure recovery and subsonic velocities
in the dwergent segment.
Line adf. If the receiver pressure is dropped to the point f, the mass flow
rate, represented by the constant mass flow rate line of Figure IV-2,
remains choked at its maximum. However, the velocity in the dwergent
segment increases. It becomes supersonic.The sound "barrier" has been
crossed.
Once sonic velocity is reached at the throat, dependmg on the
receiver pressure being at point e or point f, flow in the dwergent segment of the nozzle can be either subsonic or supersonic.The mass
flow rate is the same in both cases because of the chokmg phenomenon, even though the velocity changes.
T h s line is an isentropic one in a well-designed nozzle.There is no
heat energy flow and there are no irreversibhties. Supersonic velocity
only occurs along the line dgK; the mass flow remains choked.
If the receiver pressure is reduced below point f, for instance to point
k, no increase in velocity occurs but a sudden pressure drop occurs
afier the Qscharge of the nozzle into the receiver.
Lines adgg'i and adhh'j. If the downstream receiver pressure is

increased from point f to points j or i, the flow d be supersonic from
point d to points h or g. It wdl then suddenly drop to subsonic velocities
with an associated pressure increase to points h' or g'.The subsequent
downstream curves represent normal subsonic pressure recovery.
Further discussion of isentropic, adiabatic nozzles may be found in
Appendlx AIV
and Equations
179
C H A P T E R
F O U R
Adiabatic flow in conduits with irreversibilities, qualitative discussion

When heat transfer through a conduit wall is negligible, either due to
good insulation or to control of the external temperatures so they equal
the internal temperatures, the process can be considered a&abatic - there
is not heat energy flow to or from the system.What follows is subject to
the adiabatic constraint.
In experimental situations, in order to minimize entrance effects, it is usual
to have an isentropic segment at the beginning of a conduit.This isentropic segment is the converging segment of a nozzle. An isentropic
nozzle is necessarily adiabatic.
Experimentally, it is possible to obtain supersonic velocities within a
straight conduit by attaching the conduit to the discharge of an ideal
nozzle. In this case, the velocity will decrease with &stance along the conduit, but it wdl not cross the so-called "sound barrier" in the subsonic
direction. If the conduit is lengthened and the same upstream pressure is
applied, the mass flow rate will decrease, but the velocity will remain
supersonic in the conduit and constant at the exit from it.The phenomenon of a maximum limit to mass flow in spite of lower downstream pressure is known as "chokmg".
Sirmlarly, if the flow is subsonic and an attempt is made to increase the
velocity by lowering the pressure at the end of the long, constant section
conduit, sonic velocity can be obtained at the end of the conduit.
However, supersonic velocities cannot be obtained no matter how much
the downstream pressure is lowered. Here again, we are dealing with the
phenomenon of choking.
The sonic barrier can be crossed only in a carefully designed isentropic
nozzle as described in the previous section.
The governing differential equation for flow of a compressible fluid with
irreversibilities is given in IV-22.
This equation makes the usual simplification that elevation changes are
negligible when gases or vapors are concerned and there is no compressor
or turbine in the length of conduit. It is only when the equation is integrated
180
FIOW
o f ~ n d u s t r i a lF l u i d s - ~ h e o r y a n d E q u a t i o n s
Compressible
Fluid Flow
C H A P T E R
F O U R
that the constraints of an adiabatic or an isothermal process become important in

relating property changes.
For circular conduits, the hameter is equal to four times the hydraulic
rahus, so D replaces 4,H in the last term on the left.
In gas flow under an adabatic constraint, the temperature changes with
expansion and turbulence. The viscosity varies with temperature, so the
ReynoldsVnumber and the fi-iction factor are not constant. However, the
changes are small, and it is often possible to use average values or, more
commonly, to step up or down the conduit adjusting the values at each step.
Friction factors in supersonic flow are not well documented, but it is
thought they are about half those of subsonic flow for the same Reynolds
number.We need not be concerned with them in this book because we
limit consideration to normal plant flow rates - which are deliberately
held to less than 0.3 Mach to avoid too much noise and vibration.
Equation IV-22 wdl be developed further in Appendix AIV
Equations involving the Mach number
Even with the ideal gas assumption, the equations governing the flow of
gases with irreversibilities under adiabatic conditions at velocities greater
than about 0.3 Mach are not simple.Therefore, their development will be
moved to Appendix AIV Here, we will jump to the final, integrated result.
Equation IV-23 applies to an ideal gas flowing adiabatically with a reasonably constant ratio of specific heats. It gives the relationshp among the
length of a straight, constant section conduit, the ~ o o d ~ "fiiction
"
factor,
the ratio of specific heats and the Mach numbers at two points in the
conduit. Average values of the friction factor may be used or step computations may be made using incremental values.
The velocity of the fluid cannot cross the sound barrier, in either direction, in straight conduit. If flow is supersonic, it remains supersonic and
can only decrease to Mach 1. If it is subsonic, it remains subsonic and can
and Equations
181
C H A P T E R
F O U R
Compressible
Fluid Flow
only increase to Mach 1. At Mach 1, the mass flow rate will choke at its
maximum value.
IV-8: CHOKED FLOW AND THE MACH N U M B E R

The phenomenon of choked flow must be understood, especially by those
who size PSV's and relief headers. It is an observed fact of nature that
when a gas or vapor is allowed to flow through a conduit or restriction
from some frxed upstream pressure to a lower downstream pressure, the
mass flow rate increases with decreasing downstream pressure.
This increase in mass flow rate has a 1imit.There will come a point, when
the downstream pressure is about half the upstream pressure, when the
mass flow rate ceases to increase with decreasing downstream pressure. The
mass flow rate will remain fixed no matter how much lower the downstream pressure becomes.The flow is said to be "choked". The choked
flow condition is a common occurrence in PSV's, vent headers and control valves. The practical importance of the choked flow phenomenon is
that, for a given mass flow rate, the pressure within the exit of the conduit
at the choke point wdl be higher than irnrnedately d0wnstream.M
upstream pressures consequently wdl be higher than those computed
without considering the phenomenon.
Note the choked mass flow rate can be made to increase by increasing the
upstream pressure, but not by decreasing the downstream pressure. It should
not be assumed the choked flow rate is the same under all conditions.
In the usual explanation of the choked flow phenomenon it is stated that
small pressure disturbances in gas and vapor travel with the speed of sound
in the particular fluid. Once the molecules are themselves flowing at the
speed of sound, it is not possible for a pressure disturbance, associated with
the lowering of the downstream pressure, to be propagated upstream. It is
sometimes stated that the upstream molecules cannot "know" the downstream pressure has been 1owered.This explanation smacks of animism and
does not explain the same phenomenon in mixed phase fluids.
A more logically satisfjing explanation can be found from analyzing an

ideal nozzle. It can be shown that a velocity.equal to sound in a gas can
only be reached within the throat area of an ideal nozzle, when the
change in area along the length of the nozzle, dA, equals zero. Upstream,
the velocity is subsonic and dA is negative. Downstream, it can be subsonic or supersonic and dA is positive.
182
and Equations
Compressible
Fluid Flow
C H A P T E R
F O U R
In a conduit of constant cross section, the change in area along the length, dA, is
zero. Since increased velocity is obtained by decreased pressure, and since pressure
increases in the upstream dn-ection,the velocity must decrease upstream.The rnaxirnurn attainable velocity must be obtained at the &charge of the conduit.If the
maximm velocity attainable in the throat of an ideal nozzle is Mach 1,it cannot
be greater than this in a conduit or in any oritice. E a converging-diverging nozzle
is connected to the entrance to a conduit,supersonic velocity can exist w i h
straight conduit.
Flashing liquids
The cholung phenomenon also occurs with flashing liquids. In this case
the fluid is a mixture of liquid and gas. Sonic velocity in liquid is much
greater than that in gas.With mixed phase flashing flow it is not possible
to compute a sonic velocity. Indeed, it is not necessary to compute one to
establish the chokmg properties.
Peter Paigemv (Chemical Engineering, Aug 14, 1967) gave a more rational
explanation of the chokmg phenomenon than that generally given
involving sonic velocity. He simply reasoned from observation of the pressure profiles of flashing fluids in conduits and the fact that the length to
achieve a fixed pressure drop decreased downstream, the incremental
energy available (from pressure drop) to accelerate the fluid was converted
to kinetic energy with nothmg left to overcome addtional irreversibhties.
The flowing system is in balance at this point.The flow rate ceases to
increase, no matter how much the downstream pressure is decreased.
Examination of the differential form of the ~ernoulli'equation with irreversibilities, Equation IV-22, for horizontal flow shows there are three differential energy terms in balance. These terms are those representing pressure (static) energy, lnetic energy and irreversibilities. As the terminal
pressure is reduced, more pressure energy is converted to kmetic energy
and the length over which this conversion takes place is reduced.This can
be readily seen by examining the pressure-distance history of flow, proceeding towards the end of a conduit. At the point where the length
increment necessary to balance the equation is zero, all the pressure
energy is converted to kinetic energy and none is left to overcome any
incremental losses associated with &stance down the conduit. The energy
relationships represented by the Bernoulli equation are now in balance.
This is the choked condition.The choked-flow phenomenon applies to all
compressible fluids, and it includes mixtures of compressible and incompressible ones.We wdl develop the relationships in Appendix AIV.
and Equations
183
C H A P T E R
F O U R
Compressible
Fluid Flow
I V - 9 : E Q U A T I O N S FOR A D I A B A T I C FLOW W I T H I R R E V E R S l B l L l T l E S N O T
I N V O L V I N G T H E M A C H N U M B E R - T H E PETER P A I G E ~ E~ Q' U~ A T I O N
The importance of the Peter Paige equation is it may be used for adiabatic
flow of gases, vapors and flashing mixtures.The equation is particularly
useful for understanding choked flow for gases and flashing mixtures. It
will be given now but its development and a more detailed discussion will
be left to Appendix A N
Starting with the differential form of the ~ernoulli'equation, Equation
IV-22, for horizontal flow, Paige re-arranged the equation to solve for
incremental length.
The quantity G is the mass flux or mass flow per second per unit area and
is equal to pU.
For small, fixed increments of pressure, the following approximation can
be made.
This last equation is sometimes expressed in the equivalent form,
184
C H A P T E R
F O U R
This approximation becomes more and more exact as the pressure difference examined becomes smaller. It is therefore extremely useful for computer simulations. In these simulations, the equation is solved for the
incremental length associated with a small pressure drop and the correspondmg change in density. The lengths are then added until the actual
length of interest is reached. The ~ o o d ~friction
" ' ~ ~factor may be computed for each incremental length or an average value may be used.
Paigemv considered the driving force for flow to be the first term in the
brackets on the right of Equation IV-26. He considered the associated
acceleration term to be the second one on the right.The length over
which ths transformation takes place, due to the balance between these
competing terms, is the term on the left.
It is interesting to examine how these terms are related as we proceed
down a conduit. In an experiment performed on flashing water, Benjamin
and M d l e P U examined fixed differences of 1 psi and the corresponding
lengths. They found the length necessary to create this hfference in pressure upstream was 20.6 feet. Downstream, at the exit from the conduit, it
was only 0.57 feet.
These data allow us to reason that, for fixed pressure drops proceeding
down the conduit, the average fluid density falls and the first term on the
right in the bracket of Equation IV-26 becomes smaller.The second term
on the right becomes smaller also, but at a slower rate.The difference
becomes smaller and tends to zero.The length necessary to balance the
/
equation decreases proceeding down the conduit.
At the conduit exit, all the available pressure energy is converted to lunetic
energy and none is left to overcome the irreversibilities at the exit, so the
mass flow rate cannot increase with decreasing downstream pressure.
Explanation and use of the Peter Paige equation
Equation IV-26 gives the increment in length of a constant diameter conduit
with an inside dameter of D that w d cause a pressure drop fiom the
upstream pressure, PI, to a downstream pressure, P2.The mass velocity, G, and
the upstream and downstream densities, p, and p2, must be known and a
fiiction factor must be computed for each step.The equation is particularly
useful for computer sirnulations, but it can be used with a hand calculator.
and Equations
185
C H A P T E R
F O U R
Compressible
Fluid Flow
The equation is often used for relief system computations involving high
pressure liquid being relieved from a process vessel.The liquid may flash
on depressurization creating a two-phase mixture within the conduit. The
conduit normally vents to a knockout drum where the liquid is separated
from the vapor for recovery.The vapor frequently is vented via a flare.
To use the equation practically, fluid properties must be known as a function of absolute pressure and certain assumptions must be made (and
checked).The easiest assumption to make is that involving the use of the
alabatic model, if the conduit is insulated or reasonably short.The next
assumption is the mixture is homogeneous and there is no slip (different
velocities) between the phases. This is a reasonable assumption for most
relief situations where conduit is close to horizontal on the mixed phase
side. This assumption allows simplification of the computation of the friction factor.
The equation helps solve the peculiar problem of computing a conduit
cross section just large enough to allow evacuation of a given rate of fluid
through a known length of conduit when neither the upstream nor the
downstream effective pressures are known. The pressures in the process
vessel, the source, and in the knockout drum, the sink, w d be known, but
their lfference is not necessarily the driving lfferential pressure.
If the flow is great enough to be choked, the pressure within the conduit
at the discharge end w d be higher than that in the sink. If the flowing
fluid exits through a relief valve, the pressure on the downstream side of
the valve wdl be lower than that in the process vessel.
When a relief valve is used, an additional consideration is that the backpressure must not be great enough to prevent it functioning adequately.
Chohng can occur at any point in the system where there is an enlargement.The obvious one is at the discharge into the sink. Less obvious ones
are at junctions of smaller conduits to larger headers. Easily overlooked is
the relief valve orifice.The fluid properties at the choked (critical) condition are governed by the mass flux, G, and the upstream enthalpy. The
fluid properties immediately downstream of the choke point are governed
by the backpressure imposed by irreversibilities within the conduit to the
lscharge or to the next choke poirit.
186
Compressible
Fluid
Flow
C H A P T E R
F O U R
Requirements for solution of the Peter Paige equation

To start the solution to the Peter Paigemv equation, it is necessary to
know the specific enthalpy of the fluid at the source. If this fluid can be
approximated by an ideal gas, the enthalpy can be computed directly fiom
the temperature and the specific heat. If the fluid is liquid that wdl pass
through its bubble point, flash calculations are necessary to establish the
relationship that exists between absolute pressure and temperature and
mixture density. For non-ideal gases the enthalpy at the sources may be
estimated using the departure function and the R-K equation.
Computations involving flashing liquids are simpler than those involving
gases because the equhbrium flash strictly relates temperature and pressure. From flash calculations, the liquid and vapor densities are found as a
function of pressure.The mixture temperature is also found as a function
of pressure. From the chosen mass flow rate and the data from the flash
calculations, a ReynoldsVnumber and friction factor can be computed for
each increment.The Peter Paige equation may then be applied.
It is to be noted the method is very tedous if applied with a hand calcu1ator.The curve fitting capabilities of a computer make the solution easier.
The equation wdl give negative increments if the integration is carried
beyond the choke point.This a physical impossibility, but not a mathematical one.This inversion of sign gives a very good flag to indicate critical
(cholung) pressure. The choked pressure is the starting point for backward
(against the flow) computations.
Normally, the computation is carried out backwards from the sink. At a
chosen flow rate in a given size conduit, the equation will give positive
length increments if the flow is not choked. In this case the sink pressure
is the valid one to use and there is a loss coefficient (a K value) to be
applied at the exit.
If the flow is choked, the length increment will be negative and the pressure must be increased until the length becomes positive (zero is the cross
over point). The associated pressure is the choked flow pressure or critical
pressure that must be used in subsequent computations. In this case an exit
loss coefficient is not needed.The effective conduit length is less than that
for the non-choked flow situation.
and Equations
187
C H A P T E R
F O U R
Compressible
Fluid Flow
Once the valid pressure is established at the exit from the conduit, the
equation is applied stepwise up the conduit and the increments are
summed until the computed length equals the actual length. If the calculated upstream pressure is greater than the allowable maximum pressure,
the chosen conduit diameter was too small. If the computed upstream
pressure is too low, a smaller conduit diameter may be tried.
If the computed pressure is lower than the allowable pressure, and a restriction such as a pressure safety valve exists, the mass flow will also be choked
across the PSV T h s is the normal relief situation.The backpressure must be
checked to see it is not too great so the chosen PSV wd hnction adequately
If no restriction exists (a wide-open gate valve or a full port ball valve is
used), the mass flux wdl be greater than the assumed flux. In this case, the
assumed flux can be incremented and the computations can be done again
until the computed upstream pressure equals the actual pressure.
Peter Paigem developed another equation to calculate the choked flow
pressure directly.This equation is of theoretical interest, so its development
is given in Appendix AIV However, as R i c h t e F pointed out, by incrementing the trial choked pressure, Pc, until the difference between the
computed lengths switches from negative to positive, the critical pressure
can be found without the use of this second equation. Mulley also solved
the same problem using the Reach-KwongXXV1" equation-of-state so he
could use an analytic method instead of an incremental method for computer simulations.This equation is given in Appendix AV
Friction factor computations i n mixed flow
Extensive work has been done on the problem of findmg hiction factors for
use with various two-phase regimes @uckler,A.E. et al, "Frictional pressure
drop in two-phase flow", AIChE J., 10 44-51 (1964)lCCUii.~he
hiction factor
depends on the type of flow regime (See DeGance and Atherton-",
Chemical Engineering, March 23,1970-7. Fortunately, for most relief header
and valve sizing problems, we can use simplifjiing assumptions.
I V - 1 0 : E Q U A T I O N S FOR I S O T H E R M A L FLOW W I T H I R R E V E R S l B l L l T l E S
The isothermal model of flow with friction is only approximated in long,
relatively small dameter, uninsulated conduits that contain fluid flowing at
low Mach numbers.The equations are simpler than those for ahabatic
flow.Ths fact contributes to their popularity, but not to their accuracy.
188
Compressible
Fluid Flow
C H A P T E R
F O U R
Comparison o f isothermal and adiabatic acoustic velocities

The alabatic acoustic velocity has already been given for an ideal gas as,
If the flow is isothermal, the maximum attainable velocity is given by IV-28.
Newton first derived this formula on the assumption that sonic velocity
was due to isothermal changes. Even Newton could be wrong.
By takmg the ratio of the two equations, it can be seen that,

For air, the ratio of specific heats, y, is approximately 1.4.The square root
of this ratio is approximately 1.2. Laplace-" had to correct Newton's formula for the velocity of sound in air by this much, 20%, to match experimental fact.The isothermal homolog of the Mach number is Ni = U/a'.
Mass flux (mass velocity)
If the temperature change of the fluid over the length of a constant cross
section conduit is small, we can use the mass velocity in the Bernoullil
equation as follows. First we transform the Bernoulli equation by multiplying through by p2.
Then we introduce the concept of mass velocity (mass flow per unit area
or mass flux).The mass velocity along a constant cross section conduit is a
constant.The following development only applies to constant cross section
conduits.
and Equations
189
C H A P T E R
F O U R
We will again restrict our model to that of an ideal gas.
The density is mass per unit volume.We can now substitute mass velocity and
the ideal gas relationshp into the moddied ~ e r n o u hrelationshp
~
to give,
(MW)P~P--~~d~
R,T
gcp
=o
(N-33)
2gc ( 4 r ~ )
Isothermal mass flux

If the temperature and the friction factor are reasonably constant, the
above equation can be integrated to give the second of the set IV-34.
(TV-34)
wj2pdp
ROT
The density ratio in the second term may be replaced by the pressure
ratio when the ideal gas law is followed and the term 4rH may be replaced
by D in circular conduits.
m(~i
2R0T
-42)--1n-+-(~2p2 ~~f~
9, P, %cD
-&)=o
(IV-34A)
Equation IV-34 for isothermalflow is frequently used even for adiabatic

flows at low Mach numbers (below about 0.3). In fact, the equation is
often cited without identifying the fact that it was derived from
isothermal considerations only. It is wise to check assumptions. The difference between the adiabatic equation and the isothermal equation is in the
treatment of gas densities and friction factor.
190
FIOW
o f ~ n d u s t r i a~ ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Compressible
Fluid Flow
C H A P T E R F O U R
Energy transfer i n isothermal flow

As a gas expands, its temperature decreases. Flow of a compressible gas is a
process that involves expansion.Therefore, to maintain isothermal conditions, energy must flow from the environment to the gas.
Energy transfer per unit mass between two points on a conduit is equal to
the difference in stagnation enthalpy between these same two points when
the flow is isothermal.The stagnation enthalpy is the sum of the enthalpy
at flowing conditions and the kinetic energy change on bringing the fluid
to rest alabatica1ly.A value for stagnation enthalpy can be computed even
for processes that are not adiabatic since the concept is hypothetical.
The stagnation temperature is defined as the sum of the temperature
measured at the flowing conltion and the change in hnetic energy
divided by the specific heat at constant pressure on bringing the fluid to
rest adabatically.
By substituting these relationships into the first law equation for an open
system we can establish a formula for the heat transfer per unit mass. Note
that in an isothermal system the flowing temperature is constant, but not
the stagnation temperature.The constraint is that of steady state with no
fluid work and no elevation change.
Equation IV-35 gives various forms of the first law, open system energy
balance under the constraints mentioned in the previous paragraph.
and Equations
191
C H A P T E R
F O U R
We have proven that, in isothermal flow, energy transfer to the fluid per unit
mass flowing is equal to the change in lunetic energy between two points.
By substituting the mass velocity equation for the isothermal, steady state
lunetic energy change of Equation IV-35, we can obtain a relationship
involving the mass velocity and the density.The mass velocity (or mass
flux) of a compressible fluid is constant in a constant section conduit,
This is the heat transfer per unit mass necessary to maintain the isothermal
state.The only variable is the density.
I V - l l : CHAPTER SUMMARY
Chapter IV has introduced the basic concepts used in compressible flow
computations. The fact that the concepts of compressible flow were based
upon those of incompressible flow was emphasised.The important
differences between the two types of flow were shown.The use of models
was discussed and caveats were laid down as to when to question a
particular model.
The utility of the ideal gas model was outlined both as a conceptual tool,
for discussion purposes, and as a tool in real life situations. Using this
model, five basic groups of equations were presented: continuity, the total
energy balance, the mechanical energy balance, sonic velocity, and the
ideal gas equations.
The simplest non-ideal gas equation was developed using the compressibility, Z. The virial equation was discussed briefly. The very useful
R e d l i c h - K w o n y equation was given. The other equations-of-state will
be given in A p p e n h AIV.
192
C H A P T E R
F O U R
Model processes for compressible flow were discussed in the following

order: an isentropic, adiabatic expansion, adiabatic frictional flow, and
isothermal frictional flow. Their limitations were pointed out.
The Peter PaigemV equation for adiabatic flow was introduced and its
great utility for choked flow computations was remarked upon.
Isothermal flow with irreversibilities was discussed, and it was pointed out
these equations were limited to long, uninsulated pipelines.
and Equations
193
Compressible Fluid
Flow - Complex
Systems
V - l : SCOPE OF CHAPTER - C O M P U T A T I O N S FOR C O M P L I C A T E D
COMPRESSIBLE FLOW SYSTEMS
Chapter V will lay the groundwork to permit computing pressure drops
and flows of some very complicated compressible flow systems. The scope
will be again confined to the problems of the industrial plant.
We wdl choose a typical piping system and a typical problem common to
most chemical plants so the reader may then make the necessary extrapolations to his own systems.The system is a safety relief vent header. The
problem is how to get there from here or, to put it more technically: how
is flow computed when neither the upstream nor downstream pressures,
nor the fluid compositions are immediately known? The system and the
problem were chosen because they are representative of many sirmlar systems and problems involving fluid flow in industrial situations.These situations often demand solutions that go beyond experience and available
data. One must then use engineering judgment, testing and safety factors
in arriving at a solution to the problem.
Some of the main concepts have already been discussed in Chapter IV
and in Appendix A N We will put these concepts into the context of a
complicated system in this chapter and wdl demonstrate solutions to the
associated problems in the accompanying appendix.
The education engineers and technicians receive usually prepares them for
problems involving single pipes and single, pure fluids. Real life presents
problems of networks of pipes of dfferent dmensions and fluid mixtures of
different cats and dogs.The relief vent header is a typical problem consisting
of a complicated piping network in whlch, because of fluid mixing, the
fluid composition can change throughout the system of piping segments.
and Equations
195
C H A P T E R
F I V E
Compressible
Fluid Flow
Complex Systems
The methods discussed in this chapter and associated appendix evolved

over time. The search began with an isothermal model. Subsequently,an
adabatic gas or vapor model was developed. Later, an adiabatic mixed flow
model (flashing fluid) evolved.We wlll concentrate on the description of
the adabatic gas or vapor model.
The author would like to acknowledge a collaborator many years ago
when the ideas presented in this chapter and the associated appendix were
being developed. His name is Ishwar Davi. We both worked at Dravo
Chem Plants.
V - 2 : DESCRIBING THE P I P I N G NETWORK

The overall piping configuration of a vent header system generally is fixed
by the plant 1ayout.The major design variable for process engineers and
control systems engineers is the diameter of each indwidual pipe segment.
T h s variable must be chosen as a function of safety and economy. An
optimum choice must be made of design pressures of vessels, relief device
types, and pipe diameters.These elements must be integrated into a system.
A computer program requires that the physical network be described by a

simple c o l n g procedure. Any combination of possible flows can then be
imposed on the system to study the effects of changed piping configurations, changed dameters, or dfferent combinations of flows. Many design
dsciplines benefit fiom the output data fiom such a program. In addtion to
information of interest to process engineers and control systems engineers,
the program can give velocities, densities and pressures at every fitting.This
information can be used in studes of piping supports and pipe stress.
Two typical isometric figures, Figure V-l , and Figure V-2, are intended to
demonstrate the methodology.
196
~ I o wo f I n d u s t r i a l F l u i d s - ~ h e o r y
Compressible
Fluid Flow
Complex Systems
C H A P T E R F I V E
FigureV-l. First example of isometric sketch
FigureV-2. Second example of isometric sketch

and Equations
C H A P T E R
F I V E
Complex Systems
The two sketches,V-l andV-2, are meant to show complex systems can be
reduced to the essentials and, before attempting simulations, the entire
physical system has to be described.A sketch, no matter how crude, allows
description and peer review. An "as-built" sketch permits a personal verification during a "walk-down" of an actual network.
All relief devices in a system should be shown on a sketch.The network is

described by numbering each individual segment of same diameter conduit from number one, starting at the sources of relief flow, to the highest
number reached at the single segment that finally exhausts to a sink.The
sink is usually the point of lowest consistent pressure. It can be the atmosphere or a knockout p0t.A requirement in our scheme is that each intermediate segment must have its feeders numbered before it receives a
number.A segment is described as a conduit of constant section not
joined by any other conduit together with all bends included as equivalent
lengths, and including the terminal fitting. Expanders on safety valves are
treated as segments of zero length consisting of a "K" value only. Note
that, although the numbering begins at the sources, the numbers refer to
the segments. Bends may be included in a segment as a "K" factor or may
be described separately, if one requires data on forces operating on them.
The methods of Benelct et al-,
which can replace the K factor
methods, are discussed in Appenlx AV.
Once the piping network has been described on an isometric sketch, it
can be coded into a computer.The input is simply in order of the numbered segments.The required data for each segment are length, internal
lameter, pipe roughness, the sum of all "K" values for bends and fittings
except the terminal fitting, the code for the terminal fitting and, finally,
the identifjring numbers of up to four feeder segments.
V - 3 : D E S C R I B I N G THE FLOW R E G I M E
It is not necessary to size all vent headers for simultaneous relief of all
safety devices. It is recommended to divide the system into groups of
devices that may relieve together under similar conditions such as fire
exposure (by fire zone), cooling water failure, power failure, etc., and then
to investigate the groups for the worst case at each source.
If it is necessary to provide additional data for piping support or stress
analysis, segments can be described as being terminated by bends and fittings where the data are needed.
198
and Equations
Compressible
Fluid Flow
Complex Systems
It is necessary to decide on a model of fluid flow: isothermal or adiabatic

vapor or gas, or flashing flow.The most general model will involve multicomponent mixtures. The model chosen for demonstration wdl be the
multicomponent, adiabatic vapor model.
V - 4 : P L A N OF ATTACK
Each simulation project must have its plan of attack. Our plan is to:
establish the mass flow rate for each safety device based on the worst
case fdure of each vessel;
sum the mass flow rates of each successive downstream conduit;
convert mass flow rates to mass velocities (mass flux, G);
convert mass flow to mole fi-actions and total moles;
estabhsh the molar heat capacity of each mixture by using the ideal gas
heat capacity coefficients for each component and the departure functions;
establish the pseudocritical properties for each mixture;
establish the constants of the R e d h c h - K w o n p equation-of-state for
each segment;
compute the enthalpy at each source and sum for the subsequent
enthalpies in each segment;
identlfjr the choke points based on the mass velocities (mass fluxes) of
each segment (there may be more than one choke point);
estimate the downstream temperature and specific volume of the fluid
in each segment fi-om pressure and stagnation enthalpy using NewtonRaphson iteration (Ths involves computation in the backwards dn-ection from the flow starting firom sink pressure plus the pressure hfference across the exit or fi-om the choked pressure.);
compute the upstream pressure, temperature and spec& volume of the
fluid in each segment using the Peter Paigemv equation as one of the
simultaneous equations;
pass the computed data to the upstream pipe segment or use choked
downstream properties and repeat the computations upstream untll the
safety device hscharge is reached;
compare the computed pressures downstream of the relief devices with
the required maximum dscharge pressures and make any necessary
changes to segment sizes;
make changes and iterate to a satisfactory solution.
and Equations
199
C H A P T E R
F I V E
Complex Systems
Fortunately, most of the computations and the data manipulation can be

programmed to be executed automatically.
V-5: MANIFOLD FLOW

Manifold flow is defined as the splitting of a single stream into two or
more streams or the combining of two or more streams into a single one.
Safety relief vent headers contain many fittings that fall under this general
designation. Safety vent headers deal mostly with combining flows.
The ~ernoulli'equation still applies in manifold flow. If a velocity reduction across a fitting produces a negative change in hnetic energy greater
than the "friction (mechanical energy) losses", the downstream static pressure wdl increase.Ths is a fairly common, though often totally unexpected result.
The losses of mechanically useful energy are often expressed in terms of
"velocity heads", meaning a multiplying factor times the kinetic energy at
a specific location. The location must be associated with a "K" value and
the user must know whether it is a downstream or an upstream location.
Caveat
- Negative K values
Data obtained from experimental work on manifold flow consistently give
some negative values for the coefficient,"K". Some analysts have attributed
this fact to error in the use of the average velocity, U.The author believes
this phenomenon occurs because one smaller stream has been entrained
by another and has gained specific energy from it. Computations on both
streams will reveal an overall reduction in mechanical energy.
Note it is also possible that those who established the data may have neglected pressure recovery. Erroneous data may result when measurements
are taken too close to the fitting concerned, before the velocity profile has
had a chance to re-establish itself.This is why projected values should be
used.The reader is referred to the hstinction already made between AP
and A(AP) and to the projection method of overcoming the error.
Terminology
A common terminology is needed to describe flow in manifolds.The one

used here evolved over some time to be as concise as possible.There are
two types of manifold flow: combining, which is the usual case of vent
headers, and dividing. Physically, these manifold flows may be represented
200
and Equations
Compressible
Fluid Flow
Complex Systems
C H A P T E R
F I V E
by the tees of FigureV-3. Usually, the run of a contiguous tee is constant

diameter.The branch may be of different diameter than the run. The volumetric flow wdl vary depending on conditions, so a ratio, Q, of minor
flow to major flow will be set up.
%f
Tees
Combining Flows
3
Symetrical
Tees
Dividing Flows
Figure V-3. Manifold flows
Caveat
- Definition of K factors
The conventional "K" factors or "loss" coefficients are usually defined as
the change in specific mechanical energy across a path through a fitting
divided by the specific lnetic energy of the flow before it was divided or
after it was combined. Other definitions exist; therefore, when using K
factors, the reader must know which set of equations to apply. Care should
be taken with Miller'sm variation which defines the loss coefficient as the
change in "total" pressure divided by the velocity pressure at a specific
location. The velocity pressure is the product of the lnetic energy and the
density.
Benedct et alcorrelated total pressure ratios with inlet and outlet
pressure ratios to arrive at a straight-line relationship for sudden expansions and sudden contractions.The relationship between entropy increase
and total pressure ratio across an adiabatic conduit with irreversibilities can
also be shown to be 1inear.The author has reason to believe this line of
investigation wdl be profitable in the future.We wdl enlarge upon this
concept in Appendix AV.
and Equations
201
C H A P T E R
F I V E
Compressible
Fluid Flow
Complex Systems
Combining or divilng tees should have their connections numbered so

the major flow is always in connection number 3.This way, correlations
can be performed on experimental data and the data can be reduced.
Charts can be drawn of the K factors versus the fractional volumetric
K factors. Since the denominator in
flow. FigureV-4 is typical of Miller'sXUL
the definition is always the major flow, the abscissa varies from zero to one
while the ordinate can vary from about minus one to about plus 1.5.
FigureV-4 is representative of many sketches that could be drawn from
the relationships between the K factor coefficients (Miller's, in this case)
and the split in flow in a combining tee. Note the volumetric flow ratio is
used. This is probably because the relationship must represent many pipe
sizes, and venturis or orifice plates, at the time of this writing, are the most
common flow measuring devices to be used in small and large diameters.
These are essentially volumetric devices. Note also there are two curves:
one to correlate losses in mechanical energy through the run and the
other losses through the branch.A further point worthy of notice is the
sketch has the notation A2 = A3, meaning the run has a constant section.
Curves such as that depicted in FigureV-4 are useful because they can
represent many sizes and configurations of fittings. However, the constraints must be specified. Such curves can be used in curve fit programs
in simulations either directly or via a polynomial or other equation.
Figure V-4. Miller's K factors
'Loss' coefficients
Before automatically using loss coefficients, it is necessary to check that
choked flow does not exist. Choked flow is more likely in the reduced
branch of a tee rather than in the run. Should choked flow exist in a
branch, the loss coefficient for that branch is not used because the coefficient has no influence on the choked pressure.
202
and Equations
Compressible
Fluid Flow - Complex Systems
C H A P T E R
F I V E
D.S. ~ d l e F "offers insight into the differences between loss coefficients

used in incompressible flow and those used in compressible flow. He
points out the vena contracta can actually "migrate" in a fitting, and the
area of the contracted flow can change under changing relative flow rates.
He states that correcting some incompressible loss coefficients to apply to
compressible flow may not be very satisfactory. However, until experimental data are available, it is necessary to adapt incompressible coefficients to a form suitable for compressible flow.
Incompressible fluid loss coefficients

Assure yourself all the terms are understood. The subscripts are especially
important. The original "losses" were simply the differences in static, lunetic
and potential energies (dlfference in mechanical energy) across a device.A
loss coefficient was a dlmensionless term equal to the ratio of the dlfference in mechanical energy across a path through a fitting to the lunetic
energy at a specific location, upstream or downstream, of the fitting.
The original loss coefficients were ratios of mechanical energy to lunetic
energy. Mechanical energy is measured in terms of pressures, densities,
velocities and elevations.The association with pressure, in particular, leads
to misunderstanding since some pressure "losses" may be recovered, but
mechanical energy "losses" are not.
Miller's definition of a loss coefficient for incompressible flow is given inV-l.
(V-1)
Ki= c1 - c 2
put24,
This equation may be more understandable if it realized density is constant in an incompressible fluid and is the divisor of the numerator.The
equation is equivalent to the dlfference in the sums of static (pressure)
energy and lunetic energy divided by kinetic energy. It is identical to the
conventional K factor equation for incompressible fluids. The subscript, t,
means total; the subscript, i, means the coefficient refers to the section
where velocity is measured.The total pressure is the sum of the static pressure and the velocity pressure that is derived from energy considerations
and by neglecting the elevation change across the fitting, as follows:
and Equations
203
C H A P T E R
F I V E
- Complex Systems
Compressible fluid loss coefficients

For compressible flow, the density is no longer constant across the fitting and
M d e P changes the d e b t i o n of a loss coefficient to accommodate h s fact.
From the ~ernoulli'equation with losses, the original loss coefficient is
The two loss coefficients are subscripted 1 and 2 to indicate which

velocity should be used. Equation set V-4 gives the traditional definitions
of the loss coefficients.
Miller writes his definition in terms of "total" and actual pressures as follows.
For compressible fluid flow, Miller's coefficients differ from the traditional
ones by the ratio of the downstream to upstream densities applied to the
second term of the normal mechanical energy loss equation. If the density
is constant, they are identical.
204
and Equations
Compressible
Fluid Flow
Complex Systems
C H A P T E R
F I V E
The two expressions for the "loss" coefficients in the setV-5 give a means
of establishing them experimentally from the measurable quantities: pressure, velocity and density. Elevation change is neglected. It is inconsequential for compressible fluids in most vent headers.
Note each loss coefficient is established by experiment, by measuring, or
estimating from measurements, pressure, density and velocity at a point in
a specific type of fitting.The general results are published in the form of
graphs (FigureV-4) or formulae. The loss coefficients should be correlated
from specified conditions (adiabatic) and these conditions should be
stated.
The traditional correlation allows the coefficient to be used as a multiplying factor of hnetic energy at a section to obtain the mechanical energy
converted to thermal energy across the 6tting.Tl-u~"lost" energy can then
be plugged into the ~ernoulli'equation and, with the appropriate downstream quantities, the upstream quantities may be computed. ~ d l e r ' cors~~~
relation for K1 allows the coefficient to be multiplied by the kmetic energy
to produce the dfference between upstream mechanical energy and the
product of downstream mechanical energy and density ratio.
The two definitions in the setV-5 do not give identical results and the
coefficients must be used with the appropriate equations.
Combined loss coefficients

When fittings are close together, even experimental observations can be in
error due to pressure recovery and the fact data is taken without considering velocity profile.When sizing vent headers, one does not normally
have the luxury of exact data and one is normally under time constraints,
so safety factors are used. For instance, the assumption that all tees are
right-angled is usually safe. A distance between fittings of greater than
thirty pipe diameters can sometimes be specified by the piping designers.
Benedict e t a l correlations
Benedct et alm use gradents in place of K factors, or loss coefficients.
We hope to show, in subsequent work, this is the most accurate means of
estimation of "losses".
and Equations
205
Complex Systems
V - 6 : DATA COLLECTION A N D V E R I F I C A T I O N
The headmg of this section refers to data collection and verification.The
verification part must be emphasised. The author remembers vividly his
shock on "wallang down" a complex piping system and finding it had not
been installed according to his carefully conceived isometric drawing. The
field personnel had not seen the logic in the design, nor had they deigned
to inform the home office of the changes they had wrought.
Revisions are made to data and to specifications for safety devices. A verification must be made of all "as-built" data to see the data match the
design's intent.
Piping isometric map
The map of the system is the piping isometric. It identifies every device,
every segment, each hameter, each fitting and each bend in the system from
source to sink. It permits the organization of input data to the simulation.
Data sheets of all input sources
Under this heading, we include rupture disk and PSV information and
drawing and design data on vessels. Someone must check that what was
bought fits what it was designed to fit.The flowing quantities and properties at the accumulated pressure are required for each source.
Piping specifications, process and engineering flow diagrams
The piping specifications are a valuable source of data on pipe internal
chameters and pipe types.The information contained is necessary for pressure drop computations.
The process flow diagram wdl contain design data and the design material
balance.This document is a guide that must be interpreted in the light of
relieving conditions.
Engineering flow dagrams are variously called EFDS, MFD's, P&ID's, or
they may have other names.Their purpose is to conveniently present an
engineering design.They normally show every piece of equipment in a
system.They show lines and line numbers and give information that indicates
where one can find other information. In case you are tempted to thmk they
can replace the piping isometric drawing, the answer is no; they are too cluttered.Their purpose is Xerent fi-om that of the isometric drawing.
206
F I O W o f ~ n d u s t r i a~ ~I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Complex Systems
C H A P T E R
F I V E
Personal 'walk down'

Last but not least is a personal walk down of the as-built piping system to
verifjr it conforms to drawings and to make sure the isometric drawing
conforms to the physical system. Note most of the simulation work may
have been done prior to this activity, but the check has to be made.The
simulation can be revised and rerun if necessary.
We will detail further information on complex compressible flows in
Appendix AV.
V-7: C H A P T E R S U M M A R Y
This chapter has tried to outline the most important aspects of simulating
complex systems.The method chosen was to describe a common industrial situation involving a complex piping network and complex mixtures
of components. The system was a common vent header.
The importance of the following was underlined:
describing the piping network by isometric sketches;
describing the flow regimes by grouping devices that may relieve simultaneously;
having an organized plan of attack;
establishng the choke points;
understandmg madold flow and when to make use of worst-case data;
understandmg the "K" factors used and the importance of pickmg the corapproach;
rect kmetic energy term. Alternatively, using the Benedict et althe use of engineering judgement in influencing design to avoid
unknowns;
the significance of the various drawings and data sheets;
verification of data personally by a final walk down of the system.
and Equations
207
Equations Of
Incompressible Fluid
Flow And Their
Derivations
A I - l : PURPOSE
- P R O V I D I N G CHAPTER
I DETAILS
This appendx describes some of the information given in Chapter I in

much greater detail. It gives information regarding SI and customary U.S.
units and fluid properties. Pressure is defined, as is the concept of hydrostatic equilibrium. Complete derivations are given for the more important
equations. Of particular interest is the explanation of fluid irreversibilities
on a rational, thermodynamic basis. Some of the information given may
be thought to be trivial, but the author has found it useful to be reminded
about trivia from time to time.
The main derivations are for:
the steady-state force-momentum equation. Many of the other derivations are based on ths relationshp, and much insight is gained from it;
the arc^^' equation.The Darcy equation gives the permanent head or

mechanical energy loss, the irreversibhty, in terms of the length to &ameter ratio of the pipe, the lunetic energy of the fluid and the friction
factor.The fiiction factor is simply the coefficient, the ratio of the irreversibhty to the product of the L/D ratio and the kinetic energy;
the ~ e r n o u l hequation.The
~
Bernoulh equation relates all of the mechanical energy terms involved with a flowing fluid.The modern version
includes a term for the permanent head loss, the irreversibhty.Ths term
can be obtained from the Darcy equation. It also includes a term for an
external energy source or energy consumer such as a pump or a turbine.
It is probably the most widely used equation in fluid mechanics;
and Equations
209
A P P E N D I X
A I
Equations of Incompressible Fluid Flow
the thermodynamically derived ~ e r n o u l hequation.

~
We d show the
Bernoulh equation can be derived fiom first and second law considerations.Ths derivation gives insight into what happens to the so-called
"lost energym.Thefact that the Bernoulh equation can be derived fiom
thermodynamics gives the equation more legitimacy.
In addition to the more formal derivations mentioned above, we wdl discuss viscosity in some detail. In particular, the variety of units used for viscosity and the resulting potential for error wdl be pointed out - so error
may be avoided.
Laminar flow will be discussed, and the Hagen-Poiseuillem equation will
be developed. Larninar flow in circular conduits wdl be shown to have a
parabolic profile.This profile will be used to relate average velocity, U, to
point velocity,V.
A I - 2 : SI AND CUSTOMARY U.S. U N I T S

The multiplicity of units used in engineering is a source of much confusion and error.This is especially the case were different units carry the
same name, or the same nominal units may have different temperature and
pressure bases.
When writing computer programs, it is best to carefully define the units
to be used internally to the program and to do all the unit conversions
externally to the main program.This practice allows the same program to
be used by many users with different requirements. It also lessens the
chances of error.
It is fortunate the world is gradually standardizing on the S1 system.The
U.S. customary system is still very popular, however. Engineers and technicians must be conversant with at least these two systems. In adhtion to
the units of the two main systems, there are a multiplicity of mixed units
in use, many of which are falling into disuse. Since these other units usually involve only the use of conversion factors, they will not be discussed.
We now wdl briefly describe some of the engineering units.
Force
Force is that which causes a mass to accelerate. Mass is the substance of a

body. Acceleration is the rate at which its velocity changes. Newton's law
defines force.We give two forms of the law, one for S1 units and one for
210
and Equations
Equations of lntompressible Fluid Flow
A P P E N D I X
A I
U.S. customary units:

1 N = lkg 1m
llbj = llb,
1ft m ls-2 g;'
The units are newtons, lulograms, meters and seconds in the S1 system.
The newton is only a name for the group consisting of the three basic
units: kg, m and S.In the U.S. customary system, there are four basic units;
hence, we need a dimensional constant, g,, inserted in the denominator on
the right or in the numerator on the 1eft.The basic units of the force
equation are pound-force, pound-mass, feet and seconds.
The dimensional constant, g,, used in the customary U.S. system, has a
numerical value of 32.17.Ths value is identical numerically to the acceleration of gravity at sea level at a latitude of 45 degrees. It is fi-equently called a
gravitational constant, but ths terminology is incorrect.The units of acceleration are fi/s2.Those of the hmensional constant are ft-lb,/lbfs2.
The purpose of the dimensional constant is simply to allow one pound
force to equal numerically one pound mass when subject to a gravitational field that would give it an acceleration of 32.17 feet per second
squared.This acceleration is the normal acceleration of gravity at sea level
at a latitude of 45 degrees. It is supposed that since non-technical people
could not separate the concepts of weight (force) and mass, then the confusion should be compounded by giving both concepts the same name.
Lest we think users of metric systems are inherently superior, it is good to
remember that, until quite recently, the kdogram-force was used with the
Isdogram-mass for the same purpose. This use also required a dimensional
constant. In fact, the principal advantage of a metric system is it contains
repeated multiples of ten.
In switching between systems, it is wise always to write the dimensional
constant, but to regard it as equal to one and to be dimensionless when
using formulae in the S1 system.This allows its position to be kept track
of when manipulating formulae and when converting back to the U.S.
customary system.
and Equations
211
A P P E N D I X
A I
Equations o f Incompressible Fluid Flow
Example AI-l
An apple, held by its stem, exerts a force of one newton at a place where
the acceleration due to gravity is normally 9.81 m - ~ if
- ~it were released.
Its mass could be computed from
The apple would have a mass of 102 grams, 0.225 lbm,or 3.60 ounces.
We commonly say it "weighs" 102 grams or 3.6 ounces. It actually
"weighs" one newton or 0.225 pounds-force.This is the force with which
it is attracted toward the center of the earth.
Pressure
Pressure is defined as force per unit area. A more detailed discussion will
follow later in this appendix.
(AI-3)
1Pa=1N.m-2, lbar=l~~Pa=lOOkPa
, ft-2 (p$)
llb
= (1/ 144) lb,
in-' (psi)
In the S1 system, the units are pascals, newtons and meters. Since the
pascal is so small, one hundred thousand pascals are given the name "bar",
which is equivalent to one hundred kdopascals. Sirmlarly,psfa units are
replaced by psia units with a suitable multiplier of 144 to make the units
more maniable.
Psfa and pascals are still the basis of all scientific and engineering computations even though the nominal units are psia and kilopascals.
Example AI-2
At a place where the acceleration of gravity is 9.81 meters per second per
second, a mass of 102 grams of water (3.6 ounces) exerts a force of one
newton. Poured into a horizontal tray of one meter by one meter, the
water exerts a pressure of 1 ~ - m or
- ~one pascal, on the bottom of the tray.
In customary U.S. units, 102 grams is 0.225 pounds-mass.The equivalent
in pounds-force is found from Newton's law.
X
212
lb, =
0.2251bm 32.17ft s - ~
= 0.225 lb,
32.171bm f t l b y s - ~
and Equations
Equations o f Incompressible
Fluid Flow
A P P E N D I X
A I
Since one meter squared is equal to 10.764 feet squared, the pressure
exerted would be,
P=
0.225 lb,
1m2 10.764ft2/ m2
= 0.0209 lb,
ft-2
Dividing ths number by 144 gives the pressure in pounds-force per square
inch - 0.000145 psi. Multiplying psi by 27.73 gives the pressure in equivalent inches of water column - 0.004 in.W C (but at what temperature?).
The water would be spread very thinly - four thousandths of an inch or
one tenth of a d i m e t e r . T h e smallness of the pascal is the reason lulopascals, megapascals and bar (lo5 Pa) are more prevalent than pascals.
Density
Density is mass per unit volume.
At "normal" temperature and pressure, water has a density of approximately 1,000 kg-m-3 or 1.0
The ASME steam tablesmi give airfree (more dense) water as having a density of 62.3711bm-ft3at 60F
(15.5 "C). At "normal" temperature and pressure, air has a density of about
1.2 kg-m-3. At 60 F and 14.7 psia, the density of air is 0.0764 lbm-ft".
A volume of deaerated water contains 816 times the mass of an equal
volume of air at the same temperature and at 14.7 psia.
Specific weight
Specific weight is defined as the weight per unit volume or the force
exerted by a mass contained in one unit of volume.The dimensions are
newtons per meter cubed or pounds-force per foot cubed.
The dimensional constant is used with customary U.S. units.
and Equations
213
A P P E N D I X
A I
E q u a t i o n s o f incompressible F l u i d F l o w
Hydraulics engineers have a tendency to use the concept of specific

weight or weight density. Chemical engineers can get along without it.
Mass density is quite adequate for all cases. Conceptually, it is easier to
associate a mass with a volume than to associate a force with a volume.
Force is always associated with masses.
Specific weight is numerically equal to density in customary U.S. units as
long as the gravitational acceleration is close to 32.17 feet per second
squared. In S1 units, it is not the same. As an example we wdl take water
with a mass density of 1,000 lulograms per meter cubed.
For water with a density of 62.371 lb,-ft3,
Sidarly, for air at normal conditions of temperature and pressure,
kg
= 1.2,9.81,
m
= 11.8, N
S
(AI-10)
For air at 60F and 14.7 psia,
Again, the weight density is numerically the same as the mass density in
customary U.S. units, but not in S1 units.The statement that U.S. unit
weight density is numerically equal to mass density is only true as long as
the acceleration due to gravity is not too different from 32.17 ft/s2.
In customary U.S. units, the ratio of acceleration due to gravity and the
dimensional constant appears frequently. Since these numbers are equal,
they often are cancelled. It is well to remember the ratio has units.
(AI- 12)
All pounds are not the same! Different units cannot be cancelled.
214
~ I o wo f I n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Equations o f lncornpressible Fluid
Flow
A P P E N D I X
A I
In spite of appearances, this section was not meant to confuse the reader.
It was meant to make him or her aware units are often used with inadequate definitions.The reader should make it a habit of checking units each
time he or she comes across a new formula. In particular, each term of a
formula must have the same units as every other term. If it does not, the
formula is fundamentally flawed.
It might be thought the "cookbook" approach would resolve the problem
of units. However, even this method relies on:
1. the original formula being correct;
2. the formula being applicable to the problem being analyzed; and,
3. the units (includng their temperature and pressure bases) being carehlly
defined.
There is no such thing as a free lunch
even when cookbooks are used.
Energy
Energy is defined fundamentally as the potential to perform work. Work is

defined as force moving through a distance, but this is obviously only one
aspect of energy. Energy can be stored or changed from one form to
another. It cannot be destroyed; it is a conserved quantity.When we talk
about lost energy, we mean the energy is no longer available for use as
mechanical energy; it has been transformed to internal energy or has
passed to the environment as a flow of heat.The term "irreversibility" is a
better term than "lost energy" or "friction loss" .
If the force is constant over the &stance, work energy is simply the product
of the force times the distance. If it is variable, then it is integrated over
the &stance.
E=If = l N * l m
(AI-13)
E = lft-lb,
= lft. llb,
A I - 3 : PRESSURE AT A P O I N T W I T H I N A F L U I D
It is often stated that pressure at a point within a fluid is equal in all directions. If this be true, how is it that pressure decreases vertically upwards?
We wdl try to clarifjr this conundrum by a more formal derivation.
and Equations
215
A P P E N D I X
A I
In order to understand the concept of pressure within a fluid, we can

imagine a free body within a stationary fluid.The free body is at equilibrium with the surrounding fluid so all forces acting upon it are balanced.
The body does not move.
The body does not have to be different from the fluid. It can be composed of a volume of the fluid.The analysis is simpler for such a volume.
To give as much generality as possible to the analysis, the free body wdl be
chosen as a tetrahedron (following McCabe and smith-). Figure AI-1
shows the body and the forces have components in the vertical direction.
Figure AI-1. Pressure tetrahedron
Forces acting
Fluids (except pseudoplastics) cannot support shear stress without deforming

(without moving). Since the fluid is stationary,all forces acting on the body
must be perpendicular to the surfaces of the body and must also be at equilibrium, otherwise the body would move. A mathematical statement of this
is the algebraic sum of the external forces must equal zero.
In the vertical direction, the forces are:
the weight of the body acting downwards, mg/gc = R;
the force due to pressure, P , acting upwards on the bottom surface, OBC;
the vertical (downwards) component of the pressure, P, acting normal to
the inclined surface,ABC.
The forces due to the pressures acting on the surfaces,AOC, and,AOB, have
no vertical components.They enter only into the analysis of horizontal forces.
216
A P P E N D I X A I
From the common statement of Newton's second law, the sum of the
external forces equals the mass times the acceleration.The acceleration is
zero, so the forces must be balanced. The mathematical statement is,
(AI- 14)
This is a force balance with the positive direction taken downward.The

subscript, Z, on the last term is there to remind us we must find an
expression for the vertical component of this force.
Area,AOBC,can be seen fi-om Figure AI-1 to be simply AxAy/2. It d be
helpful to obtain an expression for area,AABC,in terms of the same variables, M y , so they may be cancelled. Figure AI-2 fi-om elementary geometry can be used to develop the relationship.The plane, P', passes through
the base BC of the triangle ABC and the plane, P, is a parallel p1ane.A perpendcular dropped from the apex,A, through the two planes d cut them
at a' and a. If another perpendcular, a'H, is drawn through a' and the base,
BC, we have three mutually perpendicular lines, Aa', a'H and BC. A plane
through the points A, a', H is, therefore perpendicular to the base, BC, so
the line AH is the height of the triangle BAC. The triangle Ba'C is identical to the triangle bac because it is a parallel projection of triangle bac.
Figure AI-2. Projected areas
From the figure we can see cos a equals length a'H divided by length
AH. If S is the area of the larger triangle,ABC, and S is that of the smaller
triangle, abc (or a'BC), we have,
and Equations
217
A P P E N D I X
A I
(AI- 15)
The two areas are related by the cosine of the included angle.The vertical
component of the pressure force, acting on the area BAC, is also related by
the cosine of the same angle, as can be seen from Figure AI-1.The force
balance of Equation AI-14 can now be written,
mg
AxAy
AxAy
(AI- 16)
=--P,PCOS^ = o
2,
2
2 COS a
CF,
The volume of a pyramid is given by the base area times the height
divided by three.The base area is half AxAy so the volume of the pyramid
is one sixth AxAyAz.
The mass of the free body is its volume times its density, pAxAyAz/b.The
force balance equation may now be written,
pAxAyAxg
AxAy
AxAy
(AI- 17)
- pz 2+P-=o
=
6gc
2
CFZ
PAZ~
p--
QC
=0
If the angle alpha is kept constant while the plane,ABC, (in Figure AI-l),
is moved toward the origin, the length Az becomes dz, a differential quantity. The equation becomes a differential equation,
(AI- 18)
218
and Equations
A P P E N D I X
A I
When dz becomes negligibly small (vanishes), equals P. Since the pressure P is normal to an arbitrarily inclined plane, it can be said the point
pressures must all be equal to the vertically duected pressure. Furthermore,
when analyzing horizontally directed forces, we can simply stand the free
body on its side.The first term in the equation no longer applies so all
horizontal pressures are equal.The generally accepted statement that pressure at a point is equal in all directions is only true because the point in
question is so small as to allow the differential height in Equation AI-18
to be neglected.
AI-4: HYDROSTATIC E Q U I L I B R I U M
Although pressure at a point is equal in all directions, pressure does change in
the vertical &rection.This is explained by the following analysis based on
Figure AI-3.The figure shows a column of fluid extendmg in the Z drection.We d take the upward dn-ection as positive. A t h n slice of the fluid
can be taken as a fiee body and the development in AI-19 can be made.The
sum of the external forces must be zero, otherwise, the body would move.
(AI- 19)
Figure AI-3. Hydrostatic equilibrium
The last equation of the set AI-19 is a differential equation relating change
in pressure to change in elevation. It also relates two terms that have units
of energy. If the fluid can be taken as incompressible and the gravitational
and Equations
219
A P P E N D I X
A I
acceleration is constant (if the difference in elevation is not excessive), the

equation can be integrated directly to,
In a constant density static fluid, there is an inverse relationship between

pressure and elevation. If one increases, the other decreases.
Equation AI-20 can also be integrated for compressible fluids. However, in
this case, the density varies with temperature and pressure and the acceleration of gravity varies with height, so the integration becomes more dfficult.
Note we started the analysis by considering forces acting on a free body
and ended with an expression in terms of energy per unit mass.The
expression could have been derived drectly from the ~ernoulli'equation
by realizing that, in static equhbrium, there are no losses (hf = 0), there is
no lunetic energy ( ~ ~= 0)
/ and
2 there
~ ~ is no work transfer (W = 0).
In hydradcs practice, the concept of specific weight (force per unit volume),
y, is used.The above equation is then written, for an incompressible fluid,
Density has dimensions of pounds-mass per foot cubed or kilograms per

meter cubed. Specific weight has units of pounds force per foot cubed or
newtons per meter cubed. In the author's opinion, these latter units have
resulted in much grief.
220
and Equations
A P P E N D I X
A I
A I - 5 : F R I C T I O N LOSSES E X P L A I N E D
The term "fluid fiiction loss" is a logical inconsistency used to describe
the conversion of mechanical energy to internal energy and heat flow.
Friction is a macroscopic phenomenon occurring when objects such as
brake pads are applied to brake drums or disks. When we deal with fluids,
especially with gases, physics teaches that the molecules do not touch. The
molecules of fluids approach one another until their force fields overlap.
When the repulsive force overwhelms the attractive force, the molecules
are mutually repelled.
If molecules do not touch, and if total energy is conserved, what is a more
logical explanation for the mechanical energy losses involved in fluid flow
than that of fluid fi-iction?
A more satis@ing explanation of fluid friction losses is found in a knowledge of thermodynamics. Sadi carnotii showed work (mechanical) energy
could be converted completely to thermal energy, but that, even in an
ideal cycle, thermal energy could be converted only partially to mechanical energy. In the latter case, even with an ideal heat engine (a concept
which Carnot invented), some energy must flow to a cooler sink under
the influence of the temperature difference.This valuable concept was the
source of many advances in thermodynamics - in particular the idea of
entropy as the quantity that governs natural change.
This same concept gives a more satisfactory explanation of fluid mechanical energy (friction) losses than is generally given in hydraulics texts.
Obstructions, compression and turbulence

When portions of a flowing fluid meet an obstruction such as a bend, a
valve or even the roughness of a pipe wall, compression takes place.This
phenomenon is well described in textbooks under the concept of stagnation - stagnation temperature, stagnation pressure.When compression takes
place, the local temperature increases - even with so-called incompressible
fluids.The local pressure also increases.Turbulence results as the mass of a
portion of the fluid that is at a momentarily hlgher pressure takes off along
the path of least resistance, exchanging static energy for hnetic energy.
In turbulent flow, it is easy to see that as eddies impinge upon irregularities in a pipe wall or even upon other eddies, there is compression and,
therefore, a local temperature and pressure increase.The eddles become
self-propagating.
and Equations
221
A P P E N D I X
A I
Lost work really means conversion to internal energy with possible

heat energy flow
Carnot clearly identified heat flow due to a temperature difference
without conversion of some of the thermal energy to work as a pure loss
of mechanically available energy.The energy is stdl present; it is simply in
a form that cannot contribute to bulk movement. Even with an ideal
carnotii engine, there is a flow of unconverted thermal energy to a cooler
sink.This energy is what is normally considered a loss. Carnot even
showed how to compute these 1osses.The ideal Carnot engine is simply a
mental construct that allows estimation of the maximum amount of
mechanical energy that can be obtained under a temperature difference.
Laminar flow and compression
Even in larninar flow, where it is not so obvious compression is occurring,
compression does take place. Bulk flow in a conduit is stdl in the axial
direction, but molecular motion is random and mass does move across
layers in a drrection normal (perpendicular) to the layer. Continuity
demands the normal mass flow be equal in each direction. In other words,
equal amounts of molecular mass enter as leave a given layer.
The slower molecules always slow the bulk motion of the faster molecules.This phenomenon results in compression and compression results in
increased temperature.
Once there are (many) temperature differences in a flowing fluid, there are
flows of thermal energy (heat). Some, although very little, of this energy
can be recuperated when it causes expansion of a local volume and, therefore, the production of work. However, by the second law of thermodynamics, and as demonstrated by Carnot, not all thermal energy can be
converted to work. In the real world, the higher local temperatures associated with turbulence result in heat transmission through the pipe walls to
the environment.This is obviously a loss of available mechanical energy energy that could have been used in moving the fluid down the line.
Another portion of the thermal energy remains in the fluid. The fluid
simply becomes warmer, but at a bulk temperature less than the highest
local temperature generated by compression.
The many small temperature differences cannot be measured indrvidually,
but the overall conversion of mechanical to thermal energy can be measured and correlated as pressure energy that must be supplied externally to
222
and Equations
Equations of Incompressible
Fluid Flow
A P P E N D I X
A I
maintain a constant flow rate.Ths correlation is what is done through the

~ a r c y - ~ e i s b a relationship,
ch~
the fi-iction factor and the Reynoldsv number.
The Carnot engine
The reader can get an idea of the loss mechanism by realizing the carnot"
engine was conceived of as one that would give the maximum efficiency
of conversion of heat energy flow to work.The efficiency of such an
engine is simply given by the temperature difference between the source
of heat and the sink divided by the absolute temperature of the source.
Note that a real engine does not have to exist. Carnot's purpose was
simply to establish the limits of the conversion of heat to work.
We can visualize small volumes of fluid as miniature Carnot engines that

convert some heat flow to work and pass on the remainder; the higher
temperatures that generate heat energy flow are due to fluid compression
as described above. Since we know compression of turbulent fluid only
increases its temperature by a few degrees at the most, we can get a feeling
for the order of magnitude of the conversion.
Suppose the stagnation phenomenon causes the temperature to increase
by five degrees Celsius and the bulk temperature is 298 degrees Kelvin.
The efficiency of a Carnot engine operating between these two temperatures would be lOO(303-298)/303 or about 1.7%.This represents the most
thermal energy that could possibly be converted to work under these
conditions. In reality, we do not have ideal Carnot engines conveniently
converting heat energy to work energy.
The volumes of fluid do not undergo a uniform cyclical process and much
of the heat flow wdl be drectly fiom the hot source to the cooler sink. So
the recovered work will be much less than even 1.7%. In addtion, as heat
flows, the temperature differences for subsequent cycles wdl be lower.A
one-degree Celsius difference would give only 0.34% efficiency - even i n
the ideal case.
Recapitulation
To recapitulate, a flowing fluid requires outside energy to keep it flowing.

Some of this energy is converted to the hnetic energy associated with
flow.When some of the fluid meets obstructions, wall imperfections, other
eddies or even when there is molecular diffusion in laminar flow, compression takes place. This stagnation results in locally higher pressures and
temperatures momentarily. Mechanical energy associated with the moving
and Equations
223
A P P E N D I X
A I
mass can be converted to thermal energy during ths stagnation process.

When the volume of fluid moves off into the main stream again, because
of a pressure difference, there w d be heat transfer to any lower temperature
fluid (there will also be heat transfer to a pipe wall or valve or fitting).The
volume of fluid may do net work on the surrounding fluid, but, by the
second law, the absolute maximum thermal energy that can be recovered as
work is less than 1.7% per five degrees temperature difference. It is less than
0.34% for a one-degree dfference. Since there is much direct heat transfer,
the actual useful mechanical energy recovered from heat energy flow in a
flow process is negligible. The conversion of mechanical energy to internal
energy and heat energy flow must be made up by the external source
keeping the fluid in motion. T h e dgerence between the total mechanical energy
at each end o f t h e conduit represents the irreversibilities - the losses.
AI-6: FORCE-MOMENTUM CONSIDERATIONS FOR VARIABLE MASS SYSTEMS

Newton's second law is usually written for constant mass systems. As such,
it is not a sufficiently general equation to serve as a starting point for some
analyses. For instance, the analysis of ~ernoulli's'equation given in Section
AI-8 makes use of the momentum equation without showing how it is
derived.When computing velocities and thrusts of rockets or forces and
flow in pumps and compressors, a different starting point is advisable.
Variable mass systems
Rockets, pumps and compressors may be categorized as variable mass systems.The quantity of mass either changes with time within the system or
it is flowing through the system.We w d start the analysis with a generalized system then apply it to a few specific ones.
Figure AI-4 depicts a mass, m, moving with a velocity, v overbar, at an initial time, to. A short time later, to plus At, the mass has ejected a small
mass, Am, at a velocity, u overbar-The velocity of what is left of the main
mass has increased by AV.(This could be a rocket, but it is not specified at
this point so we can derive an equation with the greatest generality.) The
overbars represent vector quantities for more generahty.They have direction as well as magnitude.
We treat the mass as a system and apply to it Newton's law relating an
external force to the rate of change of momentum.The variable, P
overbar, is the momentum vector. The force vector will have the same
dnection as the rate of change of momentum. The reader can find a sim224
A P P E N D I X
A I
ilar treatment in Halhday and Resnick.
Figure AI-4.Variable mass systems
In what follows, we have adopted the terminology of Halliday and

Resnick.The mass is in slugs, i.e. pounds-force divided by the gravitational
acceleration, g, or pounds-mass dvided by the dimensional constant, g,.
Kilograms can be used directly in the equations.
External force equals the time rate of change of linear momentum
The following development can be made.
mT + mAi? - AmT - AmAT + Am;

At
AT - A w AmAT
Am
+U&"m--v--At
At
At
At
-
L"
- mT
In the limit, as At approaches zero, several changes occur:

1. The total mass, m, becomes associated with the instantaneous mass in the
body. Effectively, the two states fuse since At is so small.
2. The quotient Am/At, a positive mass divided by a positive period, becomes
minus drn/dt since the change is associated with loss &om a body.
3. The product ArnAv becomes the product of two dfferentially small

amounts, dmdv, and may be neglected.
and Equations
225
A P P E N D I X
A I
Equations of lntompressible Fluid Flow
Most general equation for variable mass systems

The equation in the limit becomes one of four alternative expressions.
The three terms of the last equation represent the motion of the body.
They are all forces.The first term is the mass of the body times its acceleration.This term represents the actual acceleration of the body. If there is
no motion, this term is zero.The second term is the sum of all the
external forces on the body, and the third is the thrust on the body due to
the leaving mass.The thrust is made up of the relative velocity of the mass
leaving the body (negative) and the instantaneous rate of change of the
mass of the body (also negative).
Equations AI-23 are the most general equations governing variable mass
systems obeying the rules of classical mechanics (systems remote from the
velocity of light).They cannot be derived from Newton's law in its normal
form by treating mass as a variable. If such a derivation is attempted, one
term wlll be missing (the last term of the first equation), because Newton's
law in its normal form is a restricted case of a more general law.
Specific example of equations AI-23 - a rocket
Although we are not dealing with rockets in this book, the application of
the above equation to a rocket helps fix the terms of the equation clearly.
This is the only reason for discussing rockets.
1. The first term of the last equation of the set AI-23 refers to the instantaneous mass of the rocket and to the acceleration of that mass.
2. The second term represents the sum of the external forces acting on the
body of the rocket.These forces are variable.They consist mainly of the
gravitational attraction acting down and air resistance acting against the
motion.
3. The relative velocity of the thrd term is made up of two components.
The first is the velocity of the gases leaving the rocket.The second is the
velocity of the rocket.
226
A P P E N D I X
A I
4. The rate of change of mass per unit time is negative for a rocket.When
multiplied by the negative relative velocity, the result is a positive thrust in
the upward direction.This thrust must overcome the external forces @redominantly gravity) acting on the rocket if the rocket is to move fi-om its
launch pad.
Specific example of equations AI-23 - a restrained system
Figure AI-5 represents a restrained system that could be a pipe fitting or a
pressure relief valve. The general equation for variable mass systems can be
adapted to fittings, pipe bends and relief valves by simply f u n g the body
within a reference frame.The body consists of an open system through
which mass is flowing. Since the body is fixed within the reference frame,
its velocity is zero and the first term in the last equation of set AI-23 disappears.We will extend the analysis by considering mass entering as well
as mass leaving the system.
Figure AI-5. Restrained system
For a restrained system, the last term is still the thrust on the system.The
mass, m, is the mass of the body and its contents. Since, in Figure AI-5, the
velocity vector points in the positive direction and the mass leaving per
unit time is negative, the force due to leaving mass points in the opposite
dmection to flow.The force due to mass entering, however, points in the
same direction as the flow.
Steady-state simplification
We can simpliG the analysis by considering the steady-state - the most
common industrial state for most processes. In the steady-state, the mass
flow rate entering equals the negative of the mass flow rate leaving.We
can use the signs of the velocities to represent the direction of the
resultant force on the system.We have to sum two equations, one for flow
out and one for flow in, to get the resultant forces on the system (pipe fitting, valve or other open body).
and Equations
227
A P P E N D I X
A I
Note the mass flow rate is always positive. The sign on the leaving velocity
has been changed to allow the mass flow rate to be substituted for the rate
of change of the leaving mass. If the inlet velocity is greater than the
outlet velocity, the reaction will be in the negative drection.The anchor
must supply the resistance.
Equation AI-24 applies to a constant mass flow rate. It states the mass flow
rate, an inherently positive scalar, can be multiplied by the vector dfference
between incoming and outgoing velocities to find the sum of the external
forces that must be present to keep the system fixed in space. If the
incoming velocity is greater than the outgoing one, the force on the system
points in the general direction of flow and the system must be restrained
by a force (an anchor) in the opposite drection. If the outgoing velocity is
greater, the force points against the general drection of flow and the
anchor must be on the other side of the fitting.The exact drection of the
external force is given by the vector dfference of the velocities.
Caveat - Steady-state assumptions

The last equation only applies to steady-state mass flow.This fact is often
not clearly stated when the equation is cited. If this fact is not understood,
your cookbook will lead you to burn some of your creations. For
instance, the initial, instantaneous, impact forces on a pressure relief valve
are about twice the forces estimated on the basis of steady-state flow.
The external force is the sum of external forces that balance the
momentum forces. This sum includes differential pressure across the body,
viscous forces opposite flow, gravitational forces and forces due to internal
pressure acting on the walls of the fitting if they are not parallel.We will
and ~ e r n o u l hequations.
~
use this sum when we describe
arc^^^^
Because of the form of Equation AI-24, it is often stated that the sum of
the external forces on a body equals the differences between time rate of
change of momentum being convected in and out of the body. If this
statement is felt to be elegant, then it is worth remembering. If not, it is
best to simply think of the derivation in terms of relationships between
forces, masses and velocities that occur under well-defined circumstances.
228
A P P E N D I X
A I
A I - 7 : D E R I V A T I O N OF T H E DARCY E Q U A T I O N
The friction factor and t h e Darcy equation
The arc^^' or Darcy-Weisbach equation is an empirical equation given
about 1850 to express the irreversibilities due to incompressible fluid
flowing through a conduit. For a horizontal conduit of round section,
these irreversibilities are expressed as,
The first two terms in equation AI-25 are derived directly from the
~ e r n o u l hequation
~
(1-20) by considering the density to be constant and the
pipe to be horizontal and of constant section.Ths leaves only the mechanical energy loss term and the static energy terms (Ap/p is the difference
between two static energy terms for the incompressible fluid).Equation AI25 gives a way of measuring energy losses directly by measuring pressure drop in horizontal, constant diameter pipe. The pressure drop is the upstream less the downstream pressure, in ths case. Once a correlation among the length, hameter,
kmetic energy and fiiction factor is obtained for horizontal pipes, it can be
applied to all pipes, vertical, inclined and horizontal, by includmg all the
terms of the Bernoulh equation.
The last term in AI-25 is an empirical correlation of the experimentally
measured losses and known quantities. Equation AI-25 is the Darcy equation.The factor of proportionality is called the Moody friction factor. It
can be seen to be dimensionless if the units are checked. It is therefore the
same number whether S1 (g, equals one, no dimensions) or customary
U.S. units are used.The velocity, U, is the average velocity across the conduit section. The hmensional constant, g,, has units of lb,ft/lbfs2.
The relationshp is vahd completely only when a fully developed flow profile has been established. For turbulent flow, this normally occurs about 30
hameters downstream fiom a pipe entrance fi-om a vessel. The relationship
applies to both larninar and turbulent flows and to smooth or rough pipes.
The friction factor for liquids is found from experiment.The irreversibilities, equivalent to the pressure drop divided by the density in constant
diameter pipe, are divided by the terms to the right of the friction factor.
Many workers have established correlations among the friction factor, so
found, and other parameters.
and Equations
229
A P P E N D I X
A I
Fluid Flow
From dimensional analysis and experiment, the parameters that most

influence the friction factor, when flow is turbulent, are the ReynoldsV
number and the relative roughness of the pipe.The relative roughness is
the absolute roughness of the pipe divided by the diameter.The absolute
roughness is the height of the actual asperities on the pipe wall.
Fortunately, commercially available pipe is extruded on mandrels that give
a fairly reproducible relative roughness.
Laminar flow
- the Hagen-Poiseuille equation

G. H. L. HagenxY,a German hydraulic engineer, published results of
experiments on three brass tubes with diameters of 2.55 mm (0.1 inch),
4.02 mm (0.16 inch) and 5.91 mm (0.23 inch). The lengths were 47.4 cm
(18.7 inch), 109 cm (42.9 inch) and 105 cm (41.3 inch).
Louis Marie PoiseuillexY,a French physician and physicist, shortly afterwards independently presented results on flow in glass capillaries. The
flows were clearly in the larninar range.The ReynoldsVnumbers were
below 2,000. The equation resulting from the work of these two
researchers was,
The Werential pressure given in Equation AI-26 is that pressure drop

associated with losses. It is not recoverable pressure. It might be better to
divide both sides of the equation by density and state that the left-hand
side represents losses or irreversibdities.The coefficient, 32, is a pure
number. It has no dimensions.This can be seen if the appropriate dimensions for all the other terms are substituted into the equation.The viscosity units in AI-26 are 1bm-ft-l.S-' .
Equation AI-26 is sometimes transformed from one in velocity to one in
volumetric flow.
By equating the ~ a r c ~ - ~ e i s b a c equation

h"
to the Hagen-Poiseullle
equation over the same permanent pressure drop and solving for the friction factor, the larninar flow friction factor can be obtained as a function
of the Reynolds number.
230
F l o w o f I n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
A P P E N D I X
A I
When flow is in the larninar region, it is found that the effect of relative
roughness is negligible because of the viscous sublayer of fluid near the wall.
It is to be noted the ~ a r c ~ - ~ e i s b a cequation
h"~
maintains the form,
AP
h/ = - =
.P
fM
L u2
--
(AI-29)
D 2g,
for both turbulent and laminar flow. Superficially, this fact seems to contra&ct previous statements that, in laminar flow, pressure drop is hrectly proportional to average velocity and not u2.The apparent inconsistency is
resolved when one realizes the larninar flow friction factor (Equation AI28) contains a velocity term, U, in the denorninator.Ths cancels one of
the velocity terms in the numerator. The equation for laminar flow is
linear with velocity, in spite of appearances. In industrial situations laminar
flow occurs mainly with highly viscous materials - polymers and the like.
The critical zone

The range of flows involving ReynoldsVnumbers between 2,000 and
4,000 is given the name "critical" zone. Again, terms such as critical and
absolute are used rather indiscriminately.The friction factors in t h s range
of flows are indeterminate because of the fluctuations that occur between
the two principal regimes. There are large uncertainties.The critical zone
is the transition zone between two rather well defined zones.
Turbulent flow
In most industrial applications, turbulent flow is the norm.
Smooth pipes
~ l a s i u (student
s ~
of ~randtl""'~), in 1911, was the first to establish the
fact that the friction factor followed a functional relationship in ReynoldsV
number below a Reynolds number of 100,000 in hydraulically smooth
pipes. This relationship is given as,
and Equations
231
A P P E N D I X
A I
~lasius-" made use of (correlated) available data, including some from

Cornell University.
In 1914, it was concluded fi-om limited air and water tests that the relationship between the friction factor and the Reynoldsv number was independent of the fluid.This permitted exploration of irreversible energy
conversions in compressible fluids.The validity of the assumption that friction factors for gases are the same as those for liquids will be explored
more fully when the work of D.S. ~ d l e r xis~ discussed.
'
prandtlmiii developed another equation for the relationship between a
friction factor (a wall friction factor, f) and the pipe Reynolds number.
This equation correlated published experimental data over a larger range
of Reynolds numbers.
In 1858, arc^^' tested 22 pipes of cast iron, lead, wrought iron, asphalted
iron and glass. Nikuradse""" correlated Darcy's data. Nikuradse screened
sand to produce different grades based on hameter. He carefully glued the
sand on the internal walls of pipes. Using this method, he was able to vary
the relative roughnesses of pipe walls between 1/500 and 1/15, about 33:l.
colebrookx' empirically arranged the Prandtl equation and one developed
by von Kirmin to produce a formula that gave remarkably good results in
correlating published data.
Louis E ~ o o in d1944,
~ presented
~ ~ the
~ plot
~ that carries his name. His
friction factor, the one used in this book, is probably the most common
one in engineering circles.
Machine computing became simpler when the churchill-usagixli relationship (See Appendut I1.) was developed.The Churchill-Usagi equation
allows direct computation of the friction factor from basic data. Hard-toread graphs are no longer required.
Derivation o f the Darcy equation

This derivation follows that given in McCabe and Smith, Unit Operations
of Chemical Engineering, 1967.
Consider steady flow of an incompressible fluid through a horizontal, circular pipe. Consider a disc of fluid as a free body withm a section as
shown in Figure AI-6.
232
Flow of Industrial ~Iuids-iheory and ~ q u a t i o n s
Fluid Flow
A P P E N D I X
A I
Figure AI-6. Derivation of the Darcy equation
The momentum equation gives,
Because the point velocity, v, is not constant across a section, a correction

factor, beta, is included to permit the average velocity, U, to be used in
place of v. The term, F,, is the shear force (necessary to prevent acceleration - if it were not present, the fluid would accelerate) due to viscous
resistance between the layers of fluid that constitute the outside of the free
body and the surrounding fluid. It is equal to the area of the wetted
perimeter of the free body times the shear stress, T. Fg is the gravitational
force.This force wlll be zero in the horizontal direction.
In case the reader has trouble with the origin of the shear force, its presence is arrived at by logical deduction. For a constant density fluid
flowing without acceleration in a horizontal pipe of constant section, the
velocities at the two sections are equal, as are the correction factors.The
momentum term sums to zero.The gravity term in the second equation is
zero in the horizontal direction.There is a measurable pressure drop, so we
know a force exists in the forward direction. If a shear force, Fs, did not
exist to counterbalance the forward force, acceleration would result. The
two forces due to static pressure times the areas over which they act have
dfferent signs because they act on the body from dfferent directions.The
summation gives all of the external forces acting upon the body.
With steady, unaccelerated flow of a liquid in a horizontal segment of
constant dameter, the velocities and their correction factors are equal.
Since flow is unaccelerated, the external forces cancel.Their sum is zero.
and Equations
233
A P P E N D I X
A I
At a given cross section, the pressure is constant across the section, otherwise bulk flow would occur in the normal direction.The ratio of the
pressure difference, P2 - PI, between two sections a small distance apart to
the incremental length also is constant across the section.Therefore, the
first term is independent of r, the radlus, w h c h obviously varies across the
section. However, the equation must still hold, so the sheer stress, tau, must
vary directly with r. Also, because of the flat velocity profile at the center
of the pipe, tau is zero when r is zero. In other words, if shear existed at
the center line, there would be a noticeable velocity profile at the center.
Using the subscript, W,to indlcate a fixed location at the wall, we can write,
(AI-33)
The last equation ofAI-33 states that the ratio of the shear stress to the
radlus is constant across the pipe. It says nothmg about the linearity of the
relationship - the stress is more intense closer to the wall.
From the ~ernoulh'equation, head loss in steady-state flow of constant density fluid in a constant diameter, horizontal pipe is given by (AP is P2 - P,):
AP
(AI-34)
-= h&
234
F l o w o f industrial F l u i d s - T h e o r y
A P P E N D I X
A I
The subscript, fi, draws attention to the fact that the losses in mechanical
energy are mainly found in the layer of fluid closest to the wall. Note this
may be confusing because the term on the left refers to the bulk change
between two sections.
Substituting the second equation of the set AI-33, into this equation,
A P 2rw
(AI-3 5)
o=-+AL
g
The Fanning friction factor is a parameter used to correlate flow effects. It

is defined as,
(AI-3 6)
The Fanning friction factor, by the definition ofAI-36, is seen to be the

ratio of the shear stress at the wall to the product of the fluid density and
lunetic energy. It is the experimentally determined coefficient that relates
the shear stress at the wall to the product of density and lunetic energy.
The friction factor is dimensionless.When S1 units are used, gc equals one
and is dimensionless.The shear stress at the wall can be estimated with the
first of the equations in the set AI-35, not from AI-36.
The ratio of shear stress to density that appears in the above definition, AI36, also appears in the relation between head losses and wall shear, AI-35.
So, we can substitute the AI-36 into AI-35 as follows:
It is more convenient to keep the standard ktnetic energy term intact, so

the term 4fF becomes fM,the ~ o o d ~ friction
" ~ ~ ' factor.The equation with
and Equations
235
A P P E N D I X A I
this factor becomes,
Equation AI-38 is the Darcyvii equation. Obviously the ~ o o fiiction

d ~ ~
factor is four times the Fanning factor.The Moody factor is often called
the Darcy or the ~ l a s i u friction
s ~ ~ ~
factor.The subscript, S,refers to the
fact that slun "friction" is conceived of as being the main cause of head
loss when no boundary layer separation occurs. Ifformfriction (irreversibilities) comes i n to plax due to boundary layer separationfrom obstructions or changes
i n f l o w directions, h becomes greater than h . T h e symbol h L is often used to desigf
fS
nateform friction losses.
The utility of the Darcy equation is the irreversibility in units of energy
per unit mass can be found by experiment for a large variety of situations.
The ratio of the irreversibility to the correspondmg length to diameter
ratio and kinetic energy gives the Moody friction factor.This factor has
been correlated with the pipe wall roughness and the fluid densities, viscosities and flow rates.The correlation means the friction factor can be
found in the fiiction factor charts or by the ~ h u r c h d l - ~ s aequations.
~i~'
The fkiction factor charts can be found in Figures 1-4 and 11-1.The
Churchill-Usagi equation is given in Appendm 11.
A I - 8 : DERIVATION OF THE BERNOULLI EQUATION I N C L U D I N G

IRREVERSlBlllTlES
This derivation follows that originally given by Weber and Meissner,
Thermodynamics for Chemical Engineers, 1 9 5 7 ~ ~he~ derivation
.
is in customary U.S. units. It was given originally for compressible fluids, but since
it is general, it is included here for the incompressible case. Each step of
the derivation should be examined to make sure it applies to both incompressible and to compressible fluids.
This equation is for steady-state duct flow (see Figure AI-7). It is derived
for an arbitrary expansion whose elevation rises dZ for every dL along the
centerline.The cross sectional area increases by dA over the same length.
236
A I
A P P E N D I X
Figure AI-7. Development of the Bernoulliequation
In thermodynamic terms, the analysis is for a control volume over a differential expansion delineated by two normal sections, 1 and 2. Mass flow
is into section 1 and out of section 2.This is an open (flow) system.
Force balance
In what follows, letters identi$ quantities at section 1.The same letters
plus the differentials identifjr quantities at section 2.
A. Forces drected downstream:
Pressure times area at section 1,PA
Momentum force, because of the steady-state assumption, at section 1,
0
mU/g,
Axial component of pressure reaction normal to duct wall.Ths pressure is equal in all drections, so the only force iduencing flow is that
applied against the dfference between the downstream and upstream
areas, (P + dP/2)dA PdA (second order dfferential neglected).
B. Forces drected upstream:

1. Pressure times area at section 2, (P + dP)(A + dA)
0
2. Momentum force at section 2, m (U + du)/ gc
- PA + PdA + AdP
3. Viscous force, RD.This is the force that is judged to be present

because of the lack of acceleration of the fluid. It is responsible for
the irreversibilities.It is conceived of as being concentrated at the
duct walls due to the steep velocity profile at this location, but this is
a simplification.
4. The force of gravity drected downward along dZ has an upstream

component along dL.
and Equations
237
A P P E N D I X
A I
The force due to gravity is (l/gc) (volume/specific volume)(acceleration

due to gravity):
Since the cosine of the angle, $, that dL makes with dZ is dZ/dL, the
upstream force due to gravity along dL is,
So seven forces have been identified for this steady-state force balance:
two momentum forces in opposite drections;
two forces fiom external pressures in opposite drections;
one viscous force generated by the presence of the duct walls and fluid
motion, drected upstream;
one wall reaction force drected downstream;
and, one gravity force drected upstream.
If this were a horizontal parallel duct, the last two forces would have no
influence on fluid movement. They would be normal to the flow hrection.
At steady-state (constant mass flow rate), the sum of all the forces in the L
(centerline) direction is zero - they are all balanced.
a
If all the factors are multiplied and the resulting second order dfferentials
are ignored as being small relative to the first order differentials,the
238
~ I o wo f industrial ~ ~ u i d s - ~ h e o ar ny d ~ q u a t i o n s
Fluid Flow
A P P E N D I X
A I
following equation results:

0
For steady-state mass flow,
The constant term on the right of Equation AI-43 (made up of three

variables) may be substituted for the constant mass flow rate in Equation
AI-42 and the equation may be manipulated, as follows:
UdU g
-vdP = -+ -di:
+ (-),v
UdU
-vdP = -+-dZ+
2,
g,
dh,
gc
g,
g
dR,
The coefficient of the dRD term in AI-44 is a constant and it can be taken
inside the dfierential. A new dfferential is thus created whlch we have
called dhf- the differential losses in mechanical energy due to viscous drag
forces or compression.The subscript, 1, shows the values of v and A are
those at section 1 - they are constant. All the terms in Equation AI-44 have
units of force times length divided by rnass.They are energy per unit mass.
Integrating all terms between sections 1 and 2 (definite integral), but leaving
the first term under the integral sign because its integration depends on the
path (process - adabatic, isothermal, etc.), the following equation results:
and Equations
239
A P P E N D I X
A I
The units in each equation must be consistent ones. If S1 or any other system
based on three hndarnental units is used, gc becomes one and is dunensionless. It is advantageous to write the equations with ths term so as not to
forget the acceleration of gravity term, g, when switchng between systems.
Each of the terms of Equation AI-45 except the last represents changes in
energy that can be converted to changes in any of the other terms.The
last term represents energy that is converted from the other terms.The last
conversion is unidn-ectional. It comes from viscous forces worlung in the
dn-ection opposite the flow and producing local temperature increments
due to compression and irreversibilities due to dispersion of heat energy
without recovery as useful work.
Viscous forces cause compression of liquids or gases; compression results in
a local temperature increase from the work of compression; differential
temperature causes heat energy flow; heat energy flow without recuperation of mechanical energy is a loss of mechanical energy (but not of total
energy).We have used the subscripts 1 and 2 on the last term just to
emphasize the fact the integration was over a length of pipe and the irreversible energy conversion (the irreversibility) belongs to that length. The
subscripts w d be dropped subsequently.
The first term represents the difference in pressure-volume energy (static)
along the conduit.The second one is the difference in lunetic energy.The
third is difference in potential energy. The last term represents mechanical
energy converted to thermal energy (internal energy or heat flow) due to
the irreversibi1ities.M terms are on a per unit mass basis.
Another way of l o o h g at the above equation is to make the negative of
the first term equal to all the other terms. It can then be said that the
static energy may be converted to any of the other three forms of energy
including mechanical energy losses due to irreversibhties. This is a reasonable explanation because one cannot have movement without a difference
in mechanical energy.
Modifying the Bernoulli equation to include a pump or turbine

If a pump or turbine is present, the ~ e r n o u k equation
'
may be moddied as,
The second term represents pump or compressor work added to the fluid
per unit mass flowing. If a turbine is present the negative sign must be used.
240
Flow of Industrial Fluids-~heory
and Equations
Equations of Incomprcpssible
Fluid Flow
A I
A P P E N D I X
The terms on the right represent the change in energy per unit mass as it
flows through the system - kmetic energy, potential energy and irreversible changes to thermal energy.
Restricted forms of the Bernoullii equation
Equation AI-45, the differential form of the Bernoulh equation including
irreversibilities,is general.Various restrictions (constraints) can be placed
on it to approximate specific processes.
Incompressible fluids
For incompressible fluids, the first term on the left ofAI-46 can be integrated directly to -v(P2 - PI) or the equivalent -(P2 - P1)/p.
Ideal gases flowing isothermally
For gases that can be represented as ideal (most gases at low pressures), the
first term on the left can be integrated mrectly to RTln(P1/P2) when
there is no temperature change.We will discuss this form more fully in the
chapter on compressible flow. This assumption is not applicable to most
in-plant situations. It is mostly used for long, uninsulated pipelines.
Comparison of the Bernoulli equation with the first law equation for open systems
We will give two approaches to this comparison. Remember the symbol,
h, is used for enthalpy, not for irreversibilities,in the first law equation.
The first approach relies on more knowledge of thermodynamics than the
reader might possess, but it is given for completeness.The second
approach is more readdy understandable.
The first thermodynamic approach
The first law statement for open systems has already been given in
Chapter I as,
The subscript, n, on the work term means "net".Ths is the work actually
transmitted to the fluid.The subscript, f, is a synonym sometimes used for n.
and Equations
241
A P P E N D I X
A I
The chfferential form can be rewritten,
For an ideal system (thermodynamically reversible), dq equals Tds, so Tds

can be substituted for dq. The variable, S,is specific entropy.Also, one of
the fundamental equations of thermodynamics is,
(AI-49)
- V ~ P= T ~-Sdh
Substituting into the first law statement, we have,
The equation,AI-50, is identical to the ~ e r n o u l lequation

i~
except for the
term representing irreversibhties. In other words, it is identical to
Bernouh's original equation. The thermodynamically derived equation
assumes ideality (no irreversibilities), as &d Bernoulli's original equation.
All the terms are energy per unit mass.When integrated, the two terms on
the left can be regarded as driving potentials that produce the two terms
on the right.The latter terms are seen to be the lunetic energy and potential energy changes.
For the real (irreversible) case, an arbitrary term can be added which represents the mechanical energy per unit mass that is converted to thermal
energy and is no longer available to do work.
UdU gdX
-vdP awn = -+ - dh,
g,
g,
The two terms on the left of AI-51 can be regarded as the driving potentials.The three terms on the right are seen as changes resulting from the
driving potentials, lunetic energy, potential energy, and lost work (converted to thermal energy) due to irreversibilities.
This is a rather restricted viewpoint. It might be better to regard two of
the terms of Equation AI-51 as representing unidirectional changes, dhf
and dwn.The first of these terms always takes energy from the other
terms.The second one either adds energy, if it is associated with a pump
or a compressor, or removes energy if it is associated with a turbine (the
sign is changed to a negative one).
and Equations
Equations of Incompr@ssible Fluid Flow
A P P E N D I X
A I
Second thermodynamic approach

The second approach to the problem is to use the concept of "lost work"
or an even worse term, "internal heat". Although the names chosen are
dubious, the approach is probably better as it attacks the problem of useful
mechanical energy that is degraded to less useful thermal energy more
directly.The terms "lost work" and "internal heat" can be replaced by the
single term "irreversibility". The valuable piece of insight that follows was
obtained from Benedict, R.P.,Fundamentalr $Pipe Flow,Wiley 1980fii.
From thermodynamics, the general energy balance, on a unit mass basis,
for steady-state flow can be given for an open system as follows:
Equation AI-52 states that energy, ideal and non-ideal, crossing the closed
boundary of an open system, is only recognized as heat energy or work
energy. It further establishes an energy balance between changes in energy
crossing the boundary and changes inside the system: internal energy, ideal
work of the system, flow work, vdP, hnetic and potential energies.
Note equation AI-52 is usually given with the first three terms on the
right replaced by dh, the differential enthalpy.The relationship is shown in
AI-53.
(AI-53)
d h = d ( u + ~ v=
) d u + d ( ~ v =du+Pdv+vdP
)
For an ideal closed system consisting of one unit of flowing mass, the
work done by the system equals the Pdv term. In other words, if there
were no losses, the work done by the closed system on its environment
would exactly equal the integrated value of the Pdv term.The variable P
is the internal pressure and the differential dv is the change in internal
specific volume in the system under study.
"
= Pdv
The work done by a closed, irreversible system on its environment wdl

always be less than that of a reversible, ideal system.The lfference
between the ideal Pdv work and the actual work done by the system wlll
always be a positive number, not zero.This is the basis for the concept of
entropy (Greek word for "that which gives dn-ection").This lfference is
what has been termed lost work, internal heat, or simply "irreversibility".
and Equations
243
A P P E N D I X
A I
It can be written as,
6~
= pdv-6(wbY)
closed ,iweu
All terms have units of energy per unit mass.

The first law for a closed system may be written as,
6 q = d"
)closed
(AI-5 6)
There is no constraint of reversibility or irreversibhty on the term for the

work performed by the system since the first law is universal (EquationAI56 holds in either case).The work is done by the system, not on the system.
If we add the two previous equations we can eliminate the work actually
performed by the system in favor of the term Pdv. Note the actual work
is measured by the change in the system boundary against an external
force.This is the same whether the work is reversible or irreversible.
6q+6F =du+Pdv
(AI-57)
This equation brings together most of the important concepts of thermodynamics. It includes first and second law concepts. First law concepts
involve energy balances; second law concepts involve the direction of natural processes.The equation states that the sum of the heat crossing a
closed system's boundaries and the term that has been variously termed
lost work, internal heat generated, or irreversibility is equal to the sum of
the changes in internal energy and the product Pdv.The Pdv term would
have been equal to the actual work in the ideal case.
Note the equation is written as an equality between functions that depend
on the path traversed for their integration and functions that are point
functions (reproducible functions of state that only depend on the end
points for their integration). The lost work term, being positive always,
w d always serve to increase the heat energy flow from the system, the
internal energy term or the Pdv term.The amount that this term influences the two point terms w d depend on the constraints imposed on the
system - whether heat is allowed to flow or not.
So now we know what happens to "lost" energy. It is generally transformed to internal energy and Pdv energy or it flows to the environment
as heat, or it does both. Heat is defined as being energy flow caused by a
temperature difference. If the heat energy remains in the system, it
244
F l o w o f I n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
A P P E N D I X
A I
increases the internal energy and the Pdv energy.The fact that the "lost"
mechanical energy can be estimated directly from the dfferences in total
mechanical energy between two points and can be correlated to heat flow
is further confirmation of Joule's "mechanical equivalent of heat".
When the first law equation with losses is substituted into the general
steady-state equation for open systems,AI-52, the following equation results.
(AI-58)
UdU
g
6q + 6 w , = (du+Pdv) + vdp + - - dZ
8c
g,
The following observations can be made.The term on the left of the last
equation of set AI-58 is the net work (the fluid work) that crosses the
control volume and enters the fluid. The first term on the right is the
term that has been called lost work, internal heat generated, or irreversibility. It is identical to the term hf that was originally treated as an
arbitrary correction for "losses" of mechanical energy.
The above equation, derived from thermodynamic considerations, is identical to the ~ernoulli'equation with losses, derived from forcemomentum balance considerations. So, the Bernoulh equation may be
derived from force balance considerations or from first and second law
considerations. This goes a long way towards proving its validity. Practice
and experimentation do the rest. Given a well-proven equation, engineers
have established correlations that extend its usefulness.
AI-9: LAMINAR FLOW A N D THE HAGEN-POISEUILLE EQUATION

Laminar flow velocity profile
In laminar flow the velocity profile across a circular conduit is parabolic.
In defining viscosity we stated that velocity varied linearly between the
parallel plates. We will now clarifjr the apparent contradiction between the
two statements.
and Equations
245
A P P E N D I X
A I
Equations of
Figure AI-8 is a free body diagram representing laminar flow in a circular

conduit.
It defines the terms used in the following development. In unaccelerated,
steady-state, horizontal flow, a force balance on the free body requires the
sum of the external forces be equal to zero.
AU
the terms of equation AI-59, includmg RD,have units of force.
Figure AI-8. Horizontallaminar flow
The last term ofAI-59 represents the force acting externally on the
periphery of the free body.This force must be present to prevent acceleration. From the definition of viscosity for parallel plates, 1-6, we have seen
the force is proportional to the viscosity, the area affected by the force and
the velocity gradient. The units of viscosity are lb,/ft-S.
and Equations
Fluid Flow
A P P E N D I X
A I
The area affected is that of the periphery of the cylinder, 2nrdL. We can
make these substitutions into the force balance.
The negative sign on the point velocity is due to the fact that, unhke the
case of the flat plate, the velocity diminishes as the radius increases.The
distance between the plates has been replaced by the radius.
Since the pressures are those measured along the cylinder at points separated by a distance, dL, we can call their difference, dP, an infinitesimal
pressure difference.
It is to be noted that, at a section, section 1 for instance, the change in
pressure with distance is constant.The same change is seen at any point at
the same section.
We have a differential equation whose indefinite integral wdl give us a
relation between the ra&us and the velocity at a given section.
To find the constant, C, we noteV is equal to zero at the wall, so,
and Equations
247
A P P E N D I X
A I
We can substitute the value of this constant into equation AI-62.
To simplify the relationship, we note that at the center of the pipe the
velocity is a maximum and the radius is zero.These facts give,
We can now divide the last equation into the previous one to find that,
In mathematical terms, for laminar flow, the velocity profile across a circular pipe is parabolic.The point velocity,V, is a function only of the maximum velocity, the location within the pipe and the pipe diameter.
Relation between pressure drop and average velocity i n laminar flow
Poiseuille's law
To find the relationship between point velocity and the average velocity in
larninar flow we will start with Figure AI-9. This figure shows a hollow
cylinder whose wall thickness may be imagined as small as we wish. It is
differentially thick.
The circumference of the hollow cylinder is 2nr feet.The cross sectional
area is 2nrdr feet squared.The volumetric flow through the hollow cylinder
is Q =V(2nrdr) feet cubed per second.The mass flow is m = rV(2nrdr)
pounds-mass per second.
248
~ I o wo f ~ n d u s t r i a l~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
A P P E N D I X
A I
Figure AI-9. Hollow cylinder model for Poiseuille's law
To relate the average velocity, U, to the point velocity,V,we find two different expressions for the same quantity.That quantity is the mass flowing
through a pipe constituted of an infinite series of concentric hollow cylinders. It is the steady-state mass flow.The relationship developed in
Equation AI-65 is substituted for the point velocity,V.
and Equations
249
A P P E N D I X
A I
Fluid Flow
The negative sign on the pressure derivative refers to the fact that the
pressure decreases in the direction of L.The pressure derivative with
length is constant across a section.
The last equation can be transformed as follows:
If all the terms on the right are constant over a finite length of pipe, as they
would be in steady-state, laminar flow of an incompressible fluid, in a horizontal, constant diameter pipe, we can integrate the last equation directly.
(AI-69)
The last equation states that the pressure drop associated with a fluid
undergoing laminar flow in a circular pipe is directly proportional to the
viscosity, the average velocity and the length of the pipe. It is inversely
proportional to the diameter squared (also to the area).
This equation is sometimes written in terms of the volumetric flow rate,
Q, as follows,
The constants 32 and 128 are non-dimensional in the above equations. If

the term, gc, is treated as also being non-dimensional and as being equal to
one, the equations may be used with S1 units.
250
Fluid Flow
A P P E N D I X
A I
Equations AI-70 and AI-71 give a means of experimentally determining

viscosity by forcing liquids through capillary tubes of fixed lengths and
diameters.
Under the same assumptions of horizontal flow, incompressible fluid,
straight pipe, the ~ernoulli'equation reduces to,
This is Poiseuille'sx" equation for the energy per unit mass transformed
from mechanical energy to thermal energy due to fluid irreversibilities in
laminar flow.
Poiseuille's equation for mechanical energy losses in laminar flow should
be compared with ~ a r c ~equation
' s ~ ~ for
~ the mechanical energy losses
due to turbulent flow
"
u2
Clearly in laminar flow, the losses (and therefore the pressure drop in
straight pipe) are linear with the average velocity across a section. In turbulent flow, they are proportional to the average velocity squared.
Relationship of a common form of equation for NReDto AI-71
Frequently, the following equation is found for the pipe ReynoldsV
number:
The units used in AI-73 are lbm/h for W, inches for d and centipoise for
p.This really is an example of mixed units, not customary U.S. units,
although the formula is very commonly used in North America.
and Equations
251
A P P E N D I X
A I
Equations of lnaompressible
Fluid Flow
The equation can be derived as follows:
A I - 1 0 : SUMMARY O F APPENDIX A I
Appendix I has given some detailed definitions of the concepts and some
derivations of the formulae presented in Chapter I. In particular:
force, pressure, density and specific weight were defined;
point pressures were analysed mathematically, as were changes in pressure
with elevation;
Newton's law in terms of force versus rate of change of momentum
(AI-23) was ctscussed as being more fundamental than the more f a d a r
force versus mass and acceleration. It serves as the basis for many of the
later derivations in the book;
y ~
and the ~ernoulL'equation
detailed derivations of the ~ a r c equation
were given;
restricted forms of the Bernoulli equation were ctscussed;
a comparison of the BernoulL equation derived fiom momentum-force
balance considerations with one derived fiom thermodynamic considerations was made to help reinforce the equation's legitimacy;
larninar flow was described, mathematically;
a very common equation for the ReynoldsVnumber involving mixed
units was derived.
Most of the tools needed to analyse fluid flow problems have now been
presented.Their further application requires nothng more than practice
and extension.
252
~ I o wo f ~ n d u s t r i a~ ~~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Losses In
Incompressible
Fluid Flow
A l l - l : PURPOSE
- PROVIDING
CHAPTER II D E T A I L S
This appendm covers in much greater detail some of the information

given in Chapter I1 regarding loss computations. The reader is reminded
that, in the context of this book, the terms "mechanical energy
losses","losses" and "irreversibilities" are synonymous. Permanent losses
and pressure drop are not synonymous. Permanent losses are the irreversible conversions of mechanical energy to thermal energy.
Relationships are described in detail for the following:
the relationship of the valve coefficient, CV,to the loss coefficient, K;
the relationshp between head units and energy/mass units;
the ~ h u r c h d l - u s a g iequations
~
for fi-iction factor computations;
the Merence between irreversibhties and pressure drop;
the Merent types of K factors and their usefulness;
generalrzed loss coefficients and negative loss coefficients.
Fairly extensive tabulations and formulae wdl be given for the loss coefficients, the K factors.The peculiarity of negative loss coefficients will be
explained.
The valve coefficient, C,, is readily available for all control valves from
vendors' data or it can be obtained from riske ell's^^^^ table of values for
typical valve types.The valve coefficient can be used to obtain an equivalent K factor.The K factor can then be used to estimate irreversibhties.
The relationship between the valve coefficient and the loss coefficient will
be derived.
and Equations
253
A P P E N D I X
A I I
Losses I n Incompressible
Fluid Flow
The relationship between head units and energy per unit mass units wlll
will be shown in two
be clearly established.The ~ e r n o u l lequation
i~
forms; one form will use head units, the other energy per unit mass units.
The great utility of the ~hurchlll-usagiAequations is in their use during
automated computations - simulations for instance. However, they are reasonably simple and they can be programmed into a hand-held calculator.
It is much easier to read a dgital scale than to interpolate a graph.
A l l - 2 : R E L A T I O N O F V A L V E C O E F F I C I E N T , C,,, T O L O S S C O E F F I C I E N T , K
The valve coefficient is useful for makmg estimations of flow rates
through valves.The loss coefficient is useful in computations involving
flow through piping systems. It is convenient to have a relationship
between the two coefficients so published data on specific valves may be
used in estimations of losses for the total system.
Definition o f C,
By definition, the valve coefficient, CV,
is the proportionality factor
between the flow rate of water at 60F and the square root of the differential pressure across the valve.The valve coefficient was an American
invention prior to the days of S1 units, so the units of the coefficient are
gpm per
The specific gravity of water at 60F is taken as one in the above definition.
It can be seen that the valve coefficient can also be defined as the amount
of 60F water, in U.S. gpm, flowing through a valve with a one-psi pressure drop across it.
Since a gallon is a unit of volume, we can convert it to cubic feet and estimate an average velocity for a one-psi drop across a valve.
254
F I O W o f ~ n d u s t r i a l~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Losses I n lncompressible Fluid Flow
A P P E N D I X
A I I
We now have an expression for the average velocity at the inlet to a valve
in terms of its coefficient and the diameter ofthe inlet under a pressure drop
of one psi. The velocity is in ft/s. The diameter is in inches.The coefficient has units of gpm divided by the square root of pressure in psi.This is
a dimensional equation; it uses mixed units.
K as a function of C,
The generalized loss equation in terms of the original resistance coefficient, K, is often used to establish the relationship between the loss
coefficient, K, and the valve flow coefficient, C,. See Equation AII-3.
If not defined otherwise, variables are referred to the valve inlet.
The velocity at the valve inlet, U, with a pressure drop of one psi is
obtained from the previously developed equation, AII-2. Also, the pressure
drop across a valve is defined as that due to the permanent losses in
mechanical energy, the irreversibilities - recovery is not considered to be a
factor.The irreversibhties are given by the pressure drop divided by the
density if the inlet and outlet diameters are equal.The average velocity in
Equation AII-3 is that of the inlet (or outlet if the diameters are equal).
The permanent loss relationship and the K factor equation can be equated
as follows:
144(1.0)
= 62.365
= 2.309
ft - lb, / lb,
Given a valve flow coefficient established by the manufacturer, the resistance coefficient can be computed from last of equations AII-3.The permanent losses across the valve at its wide open position (its nominal C,)
may be found by computing K from the nominal C, by using the last of
equations AII-3 and by mahng use of the first of equations AII-3 with the
actual inlet velocity. Irreversibilities at fractional openings can be found by
and Equations
255
A P P E N D I X A I I
Losses I n Incompressible Fluid Flow
using the inlet velocity at the associated C,.The C, at a given opening

may be found from manufacturers' data.The fractional C, depends on the
valve characteristic. In the set ofAII-3, hf and AP should be computed as
projected values, to the valve faces.
A l l - 3 : RELATIONSHIP BETWEEN ENERGY PER UNIT, MASS UNITS,

HEAD U N I T S AND PRESSURE U N I T S
Hydraulics engineers often use "weight density" in place of mass density
and feet or meters of fluid in place of ft-lbf /lb, or N-m/kg (J/kg).This
practice can be explained formally by the following transposition of the
~ernoulli'equation with irreversibilities:
The units of each indwidual term in the last equation are now feet or
meters of fluid.The units of the group, pl (g/&), are force/length3 (lbf / ft3
or ~ / m ~This
) . is the "weight density", y, lb,/ft3.
In the customary U.S. system of units, the acceleration due to gravity and
the dimensional constant are numerically equal at sea level at a latitude of
45 degrees.The mass density is, therefore, equal numerically to the weight
density in U.S. units at this latitude. In S1 units, the dimensional constant is
replaced by one and the mass density does not equal the weight density.
If the denominator of the static energy terms is replaced by y, the last
form of the Bernouh equation becomes,
All terms have units of length (height, hence head). In hydraulics practice,
Z and P/y are often combined and are called the hydraulic (piezometric)
head. They are then written as H I and H2.The above equation becomes,
(AII-6)
'X
H,+-+-=H2+-+2g (&?/gc)
22 (&?/gc)
This latter form of the Bernoulli equation is often used to develop the
256
F I O W o f ~ n d u s t r i a i ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Losses In I n c o m p r e s s i b l e Fluid Flow
A P P E N D I X
A I I
equation for flow through an ideal, horizontal restriction.The elevations

are equal.There is no energy added or subtracted.The last term is zero
(ideal). Section 1 is upstream; section 2 is at the vena contracta.
(AII-7)
H , + -u:
= H , + -u:
2g
2g
Going one step further, pressure units can be developed by multiplying

the ~ernoulli'equation with irreversibilities by (the constant) density.
(5-~)+(...
In hydraulics practice (constant weight density),
='/h/
(NI-9)
= ( P , -P,)+
-y z 2 )
It can be seen that both of the above equations represent lost pressure as
being the sum of the dfferences between each of the three terms.The first
difference is the actual measured pressure drop, the second a velocity pressure drop and the third and pressure change due to a change in elevation.
These equations are particularly prone to errors in interpretation since it is
sometimes difficult to keep (mentally) a non-measurable "pressure drop"
separate from a measurable one.
Caveat
- Hydraulic engineers' simplification

The hydraulic engineers' approach generates simple equations for
restricted uses. The simdarity of the equations to the more formal equations leads to confusion among terms such as p and y, Z and H, and g and
g,. For generalists it is best to use the formal form of the Bernoulli equation with mass density.
A l l - 4 : CHURCHILL-USAGI F R I C T I O N FACTOR EQUATIONS

For computer computations, equations are much more convenient than
charts and tables for correlations. Stuart W. churchillAi of the University
of Pennsylvania presented a convenient method of estimating the friction
factor from the Reynoldsv number and the relative surface roughness
and Equations
257
A P P E N D I X
AII
(Chemical Engineering, Nov. 7, 1977).This method has been successfully

used in simulation programs.
The method is claimed to be good over all regimes: laminar, transitional
and turbulent.The author has used it and checked it over the laminar and
turbulent regimes. It gives excellent results.
It must be remembered the transitional regime is unstable and should be
avoided. Numerical stabhty does not equate to physical stabhty. So, even if
the churchill-usagifi equations give specific numerical results between a
Reynoldsv number of 2,000 and 4,000, the results do not represent reality.
The fiiction factor computed by the Churchill-Usagi method is one half
the FanningWV fiiction factor and one-eighth the ~ o o dfriction
~ ~ factor.
~ '
The final result must be multiplied by two or eight to obtain the correspondmg Fanning or Moody factors.We will do this at the end of this
section.
The equations can not only be computerized; they can be programmed
into a hand-held calculator-They reproduce all of the fiiction factor plots
in the literature, including the ones given in this book (from craneXLi).
The equations are,
(AII- 10)
fnr = 8 f
It can be seen that the only input terms are the Reynolds number and the
relative roughness (absolute roughness divided by the pipe internal diameter in the same dimensions).The absolute roughness of various materials
has been given in Table 11-1. Additional values can be found in many publications such as Perryx.
258
FIOW
o f ~ n d u s t r i a~ ~~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
A P P E N D I X
A I I
The dimensions of e and of D in equation AII-10 must be consistent.

Both must be in feet, meters, rmllimeters or inches.
A l l - 5 : P R E S S U R E D R O P V E R S U S ' F R I C T I O N LOSSES'
There is room for confusion between the concepts of pressure drop and
irreversibilities.The literature is full of examples where the two concepts
seem to be equated.To avoid inadvertent error, it is good to have a clear
concept of the difference.
Pressure drop in a circular pipe flowing full is due to a conversion of static
energy into other forms: potential, hnetic and thermal (internal energy)
when there is no producer or converter of energy in the 1ine.The producer would be a pump or compressor-The converter would be a turbine.
From a mechanical point of view, the irreversibilities associated with flow
constitute a loss of useful mechanical energy. In the ideal case, irreversibilities do not exist and hnetic, potential or static energy can be interconverted within physical constraints. Irreversibhties constitute a conversion
of mechanical energy that results first in a temperature increase and then
in heat flow to the environment.
The ~ernoulli'equation with no pump or turbine in the section can be
written,
The first term is left under the integral sign when compressible fluids are
involved.When incompressible fluids are involved, it can be integrated
directly to,
Written in the above fashion, the Bernoulli equation states that the available static energy hfference between two points can be converted to a
potential energy difference, a kmetic energy dfference or it can be used to
overcome viscous resistance. It can increase the elevation of a fluid mass,
increase its velocity or it can become "lost".The last term is considered a
loss of mechanical energy (conversion to internal energy) because it
cannot be converted back to one of the other three forms.This observation resulted in the concept of entropy - that which gives direction.
and Equations
259
A P P E N D I X
A I I
Fluid Flow
In a straight, horizontal, equal dameter pipe with an incompressibEe_f2uid,there

is no potential energy change and no hnetic energy change.The static
energy change equals the loss term, in ths case. By multiplying through by
the density, the irreversibhties can be expressed as a pressure drop.
=ph,
(AII- 13)
e-P,
If there exists a change in elevation, density, or velocity, the simple relationship of Equation AII-13 is no longer true and the irreversibilities times
mass density do not equal the pressure drop.
A l l - 6 : 'K' F A C T O R S
LOSS COEFFICIENTS
The K factors are a convenient means of computing the losses in mechanical energy across a commercially available fitting, orifice, valve or other
device. A K factor is the proportionality factor between the irreversibilities
and the hnetic energy of the fluid at some known section. For incompressible fluids, the K factor can be looked upon as the fraction of available hnetic energy that is converted to mechanically unusable internal
energy by some impediment to flow.The subscript, i, is meant to remind
the reader to use the average velocity at the section corresponding to the
experimental K factor.
(AII- 14)
U,'
h, = K, 28,
These factors have been correlated with the average velocity squared and
the measured losses for many devices by many workers.They have been
tabulated, put in graphical form and put into the form of equations. It is
extremely difficult to give cre&t to the original worker so we will simply
give the source from which we obtained the information. craneXLiiis an
excellent general source. Others wdl be given as each factor is discussed.
Equation AII-14 is so simple it warrants a few caveats:
Caveats
260
- Velocity
used w i t h K factor, Suffixes, K factors, C,

When using K factors, be sure to use the computed average velocity specified as being associated with the value of K. Do not use a point velocity.
The average velocity is usually that in the smaller diameter, but not always.
For instance, in the square-edged orifice equation, the velocity is usually
that computed for the pipe, not that for the orifice.
FIOW
Fluid Flow
A P P E N D I X A I I
Be aware of how suffixes are used in equations. Some authors use 1 for
upstream and 2 for downstream. Cranefiiiuses 1 for the smaller hameter
and 2 for the larger hameter.
Some authors list K factors for various types of fittings as being constant
for all sizes.When geometric similarity does not hold, and it does not in
most cases, this cannot be true. Use correlations from reputable sources,
such as the Hydraulics institutedv and Crane.
When a CVis available for a specific valve, use it to compute a K factor,
do not use a general correlation.The manufacturer has tested his valve, so
the results will be more accurate.
Although we have adopted the usual convention of using the symbols hL
or hi for the losses (conversion) of mechanical energy associated with a K
factor, remember the true losses can only be established over a length of
piping that allows normal velocity profiles at each point of measurement.
This is the reason hL was replaced by A(AH) in Chapter I1 - as a reminder.
Geometric similarity and dissimilarity

Some K factors remain reasonably constant for all sizes of fitting and some
are a function of geometry. If complete geometric similarity holds, as in
the case of pipe entrances and exits, the K factor is a constant independent
of average velocity. If geometric dissirmlarity holds, the K factor is given as
a function of the beta ratio.
Table of K factors
Note that, in what follows, all K factors are given for the smaller diameter
except for the square edged orifice.The velocity in the smaller pipe must
be used in the head loss equation, except for the orifice. If the velocity in
the larger pipe is known or is more easily computed, hviding the K factor
will give a new, larger, K factor that
computed for the smaller pipe by
can be used with the lower velocity in the larger pipe (this does not apply
to cases 1,2, 10 and 11, which are geometrically similar).
P4
and Equations
261
A P P E N D I X
A I I
Table All-1. K factors

Cause of Loss
Pipe entrance, square edged
Pipe exit, projecting, square or rounded
Sudden enlargement
Sudden contraction
K Factor
0.5
1 .O
Case
1
2
K = (I - p2)2
~=0.5(1-~~)
Source
1
14
Square edged orifice

Gradual enlargement, 0145"
0 = included angle
8
K = 2.6sin-(l2
p')'
Gradual enlargement, 0>45"

(same as sudden enlaraement)
Gradual contraction, 0145"
Gradual contraction, 0>45"
- 8 -'l2
K=0.5 sin(l-p2)
2
Inward projecting entrance

Rounded entrance
Pipe bends and flanged or

butt-welded elbows, 90"
Generalized loss coefficients and negative loss coefficients

This section wdl give a little more theory on loss coefficients. It will also
explain the peculiarity of negative loss coefficients.
The use of equivalent pipe lengths as a measure of the irreversibilities
caused by the presence of a component in a piping system is not satisfactory for other than approximations.A component can be a bend, miter,
tee, orhce plate, valve or any other obstruction to smooth flow.The use of
K factors that are established fi-om carefbl experiments is the ideal approach.
262
Fluid Flow
A P P E N D I X
A I I
A component in a piping system can be preceded and followed by long

straight lengths of pipe.This type of experimental set-up allows the flow
profiles to be well established upstream and downstream of the disturbance. It is at these sections that measurements are taken.
The pressure measurements are the total pressure measurements. By
"total" is meant the sum of the static and velocity pressures. It is not the
pressure measured by a pressure gauge.
When the component is removed and the pipes are rejoined linearly, the
same flow rate with the same fluid conditions is re-established and measurements are taken at the same sections (see Figure 11-3 and the surrounding discussion).The difference between the two total pressure dfferences, A(APT),is a measure of the loss caused by the component.This
method of measurement allows inclusion of the influence of irreversibilities that occur upstream and downstream due to the presence of the component - an influence that would be missed if pressure drop were measured drectly across the component.
Total pressure loss
For horizontal pipe plus a component, the total pressure loss for incompressible flow can be obtained from the ~ernoulli'equation.
(AII- 15)
The above equation holds for customary U.S. and for S1 units, if the
dimensional constant is treated appropriately.
Note carefully the total pressure loss is not the measured pressure drop,
PI - P2. It is the theoretical pressure loss based on density times the estimated mechanical energy loss.
The difference between total pressure differences due to the component's
presence and absence is A(APT),as mentioned above. Note this notation is
not usual, it is more often written as AF? However, we are tallung about
the difference between two differences.To be more exact we will use the
double delta and the subscript,T. If there were no extra losses due to the
and Equations
263
A P P E N D I X
AII
presence of the component, the difference, A(APT),would be zero,

whereas AP would still reflect the influence of the irreversibilities due to
the piping.
The above methodology allows the use of a technique whereby irreversibilities are estimated for the straight pipe and adhtional irreversibhties are estimated for the presence of the component in that pipe.These
irreversibilities can simply be added as part of the system irreversibilities.
K factors
To further increase the utility of the method, the irreversibhties are made
non-hmensional by dividing the total pressure drop of an incompressible
fluid by p ~ ~ / 2 ~ , . Tvelocity,
he
U, is the average velocity at the section to
which K is referred.
(AII- 16)
This equation again applies to both customary U.S. and to S1 units when
one is substituted for the hmensional constant. Hydraulics engineers use a
simdar equation where the units are feet or meters and the acceleration of
gravity is used in feet per second squared or meters per second squared.
(AII- 17)
The advantage of the K factors is they can be tabulated for a given type of
component at a fixed Reynoldsv number.They can then be adjusted for
an actual Reynolds number and multiplied by the average lunetic energy
(u2/2&)
at the location for which the average velocity was established to
give the actual irreversibilities due to the presence of the component.
About 50 pipe diameters are required to establish reliable K factors. If only
30 pipe diameters are available, the value of the K factor usually deviates
by plus or minus 2.0% from that of the longer runs.
The loss coefficients (K factors) appear to become independent of
Reynolds number once the Reynolds number is greater than about 106.It
is unlikely that they become completely independent of Reynolds number.
264
~ I o wo f i n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
L o s s e s I n incompressible F l u i d F l o w
A P P E N D I X
A I I
Negative K factors
The Werentid pressure across a fitting is taken as a positive number in
establishing the K factor. It is the upstream pressure less the downstream
pressure. Under certain conditions (combining flows),it is possible to have
negative K factors.Ths pecuharity can be understood by considering a tee
whose branch flow is much smaller than the flow through the run.There
are two K factors to be established: one for the branch and one for the run.
By examining the ~ernoulli'equation from the branch to the combined
junction, it can be seen that when the flow through the run is much
greater than that through the branch, energy is added from the main
stream to the branch stream.The downstream specific energy of the combined flow wdl be greater than that of the branch inlet flow.The pressure
will be greater downstream than upstream and a negative K factor results.
This phenomenon has been observed with surprise by many and has even
been denied by some.When both streams are considered, there is a net
loss (conversion to internal energy) as would be expected.
Classification of loss coefficients by source
It is quite difficult to obtain reliable loss coefficient data for some components. D.S. ~dlerxY'established the following categories in order to judge
the reliabhty of the data:
I.
The data are well-established experimental data from two or more

sources.The data are cross checked.
11-1. The data are experimentally derived, but have not been cross checked.
11-2. The data are estimated fiom two or more research programs.
11-3. The data are based on Class I sources, but are outside the experimental range.They are assumed reasonable.
111-1. The data are experimental but the source is thought less reliable.
111-2. The coefficients are based on extrapolations outside the range of data
from Class I and Class I1 sources.There is no information that allows
predction of the effects of being outside the range. In other words,
the data are not trustworthy, but there is nothlng available that is
more reliable.
D.S. Mdler also established a classification of components that is useful

conceptually. It is given in Table AII-2.
and Equations
265
A P P E N D I X
A I I
Fluid Flow
TABLE All-2. Classificationof ComDonents

Low Loss Components
Smooth
TURNING CONDUITS
-Bends and elbows
-c)IFFUSING COMPONENTS
Diffusers
I Moderate Losses
Transitional
Miters
I LargeLosses
Rough
90's
__
Swages - expanders
Sudden expansions
ACCELERATING JOINTS
Smooth contraction
Swages - contractions
Sudden contraction
COMBINING FllllNGS
Branch with rounded approach
Angled branch
90 branch
DIVIDING FllTlNGS
Branch with rounded path
Angled branch
90 branch
OBSTRUCTIONS
Mu1ti ple, parallel
Rotors (turbines)
Convoluted valves
Concept o f adverse gradients

When studying flow around bends and through fittings, the concept of an
adverse gradlent is useful. Potential energy (gZ/gc) and static energy (P/p)
can be converted into lunetic energy ( ~ ~ /without
2 ~ ~excessive
)
loss when
the flowing stream is restricted smoothly.The opposite process of converting
lunetic energy to one of the other two forms usually involves significant
irreversibhties.The reason, accordmg to D.S. MdeP", is to be found in the
non-udorrnity of the lunetic energy across the section at a bend or fitting
versus the uniformity of pressure energy (static) across the same section.
The kmetic energy near the wall is usually very low due to the viscous drag
force at the wall.When the fluid near the wall enters a region of rising static
pressure (in a &ser, for example), the fluid must be supplied with energy
fiom the higher energy fluid away fiom the wall, or it comes to rest. If the
static pressure energy downstream becomes sufficiently great, reverse flow
can occur 1ocally.This phenomenon causes separation of the stream fiom
the wall and contraction of the area through whlch the main body of fluid
flows. After the contraction, whlch has a variation in energy dstribution
across the section, large scale mixing occurs.The mixing, whch is thermodynamically irreversible and involves conversion of mechanical energy to
thermal energy, has the effect of equahzing the energy dstribution.
Reattachment of the stream to the conduit wall follows.
266
F l o w o f ~ n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
A P P E N D I X
A I I
If the flow is turned, static pressure gradients exist due to centrifugal

forces. Low energy fluid can escape (can be pushed) from a high-pressure
boundary towards a region of lower static pressure. Secondary flows can
be imposed on the main flow.
Example of combining flows with a negative loss coefficient
To reinforce the statement that it is possible to have a negative loss coefficient (an energy gain per unit mass of one stream) let us analyze the next
example.Take a pipe tee used to combine flows to the exit from the run.
If the branch flow into the run is less than about one third of the total
flow, energy is transferred from the main run flow to the flow from the
branch. This results in a higher spec@ energy in the combined flow than
would be present if there were no flow through the run.
The loss coefficient for incompressible fluids is defined in terms of subscripts. Subscript 1 represents the fluid entering the main run, subscript 2
is the fluid entering the branch, and subscript 3 is the combined flow
leaving the fitting.
(AII- 18)
If we pedorm computations on the tee in the case cited, K23 will be negative. If we sum the irreversibilities,X(A(APT))/p, we will find the combined loss in mechanical energy is a positive number, in spite of one
member being negative.
Example A l l - l
Water is flowing through a tee with a 6-inch Sch. 40 branch and a 12inch Sch. 40 run.The total flow is 2,000 gpm.The branch flow is 200
gpm.The flow is combining.The pressure downstream is 100 psig and the
temperature is 60F The orientation can be taken as horizontal to simplifjr the computation, but this is not necessary if the potential energy
terms are added in the ~ernoulli'equation.
The problem is to compute the irreversibilities from the inlet to the
branch to the downstream section and from the inlet to the run to the
same downstream section. Remember these irreversibilities are those due
to the presence of the fitting including irreversibilities slightly upstream
and many pipe diameters downstream caused by this presence (mainly due
to the destruction of eddies).
and Equations
267
A P P E N D I X
Caveat
A I I
Fluid Flow
- Conversion of irreversibilities to pressure drop

When the irreversibhties are converted to pressure drops, the computed
pressures might not be identical to those read on pressure gauges imrnediately upstream and downstream of the fitting.They wdl correspond,
through the ~ernoulli'equation, when measured at points sufficiently
upstream and downstream and projected to the faces of the component.
D.S. ~ i l l e igives
~ ' loss coefficients for various types of fittings in the form
of graphs. He gives the loss coefficients as a function of the ratio of flow
through the branch to the main flow and of the area of the branch to the
main area. In our case,
(AII- 19)
K,, = - 0.43
h
(5.73)2 = -0.219= K,, -- -0.432gc
p-
22,
lb,
'm
The velocity was obtained by loolung up in cranex" the velocity associated with 2,000 gpm in a 12 inch schedule 40 pipe.The K factors were
found from Mdler. Since the irreversibilities from the branch to the run
are negative, it can be assumed the fluid flowing through the branch has
gained more energy due to momentum transfer from the fluid flowing
through the run than it has lost due to the irreversibilities.
For the run, using the same approach, KI3 equals approximately 1.6, so,
h,,, = 1.6-
(5.73)2
28,
= 0.816-
p - lb,
'm
For the branch and the run, these sum to positive irreversibilities, as
expected.
Caveat
- Authors' definitions of loss coefficients

Be careful to note how the loss coefficients are quoted. Some authors are
not too careful in clearly stating which velocity is to be used. If the coefficient was established on the basis of combined flow in the run, the
268
F I O W o f ~ n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Fluid Flow
A P P E N D I X
A I I
velocity is that of the combined flow. If it was established on the basis of

flow in the branch, the velocity is that of the branch flow only.
The reference position in example AII-1 is position 3, downstream. The
datum can be taken as the center line of the pipes.The total mechanical
energy at position 3 is given by,
The pressure at the position with subscript 2, the branch inlet, is obtained
fi-om manipulating the ~ e r n o u l lequation
i~
with the potential energy term
held at zero because of the horizontal arrangement of the fitting.
(AII-22)
The pressure entering the branch is only a tenth of a psi higher than the
downstream pressure, but it is higher in this case.
We can perform a simdar exercise for the run.
The pressure entering the run is four tenths of a psi higher than the
downstream pressure.
and Equations
269
A P P E N D I X
A I I
It is instructive to look at the energy changes around the tee and to compute the total mechanical energy at each section and the irreversibilities
(mechanical energy converted to internal energy).
The flow rate of total mechanical energy at each position is given by the
mass flow rate at that position multiplied by the specific mechanical energy
at that point. Position 1 is the run inlet, position 2 is the branch inlet, and
position 3 is the combined outlet. Note that even though the symbol, h, is
used for specific mechanical energy (head),ths is not enthalpy.
m3
ft3
'm
h, = 4.456-62.371-(265.31)S
P3
ft - lb,
= 73736- f t - 2 b ,
S
'm
The rate at which total mechanical energy is being converted to thermal

energy is given by the sum of the rates of entering mechanical energy less
the leaving mechanical energy.
(AII-25)
ft - lb,
Btu
Losses = 66560.2 7367.5 - 73736.0 = 191.7= 0.246= 260 watts
Even though one of the loss coefficients was negative, which indicates an
energy gain for that stream, the overall effect was a loss. If these irreversibhties could have been recuperated, the energy could have been used
to power a low wattage light bulb.
All-7: SUMMARY O F A P P E N D I X A l l
A p p e n h AI1 has expanded upon the information given in Chapter 11. It has
concentrated only on passive components. Pumps and turbines wdl be considered in Chapter 111 and Appendut 1II.The essential points covered were:
the relationshp of the valve flow coefficient, CV,to the loss coefficient,
K, was developed and dscussed;
the relationship of head units and energy per unit mass units was developed and the limitations of the hydraulic engineering approach were
pointed out;
270
~ I o wo f i n d u s t r i a l F l u i d s - ~ h e o r y a n d ~ q u a t i o n s
A P P E N D I X
A I I
the very useful ~ h u r c h i l l - ~ s aequation

~ i ~ l ~ that is easily implemented
in computer programs was given with a warning about its use in the
transitional zone;
the easily confused relationship of measured pressure drop, which
includes recoverable and non-recoverable pressure difference, and loss
when expressed in pressure units was, hopefully, clarified;
much ofAppendur AI1 focused on the use of K factors, on the hfferent
types of K factors, on the relationship to a specific velocity at a given
location, and on the fact that there exist negative K factors;
it was shown there is a constant loss of power across a typical fitting in
spite of an apparent gain in energy across one of legs.
and Equations
271
Computations Involving
Pumps For Liquids
A l l l - l : PURPOSE
- P R O V I D I N G CHAPTER
Ill DETAILS
Appendix AI11 discusses in much greater detail, and more theoretically,

some of the information given in Chapter I11 regarding pumps.The intent
is to supply information and knowledge about pumps, including jet
pumps, in sufficient detail so those interested in the flow of fluids may
integrate both static and dynamic data into their everyday activities.
We wdl begin with the general theory of centrifugal pumps and then discuss the categories of pumps defined in Chapter 111. Only pumps that
handle liquids are considered in thls appendix.
A l l l - 2 : THEORY OF CENTRIFUGAL P U M P S
Figure AIII-1 is a cross section of a typical single-stage centrifugal pump
showing the essentials.The casing, the inlet flange and volute are shown.
The impeller turns with the shaft and the shaft is sealed against leakage by
pachng in a stuffing box. Mechanical seals (not shown) have largely replaced
paclung. Many pumps are designed so they may use either packing or
mechanical seals in a stuffing box according to the end user's wishes.
Figure Alll-1.Typical single-stage centrifugal pump
and Equations
273
6
l4
.C
W M.
M '2
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
The arrows show the dnection of flow. Sufficient pressure has to be available at the pump suction to force the liquid into the pump impeller.The
liquid is then subject to the centrifugal force of the rotating impeller. It is
projected radially to the volute where it is collected and hrected to the
pump discharge.
Total mechanical energy of the liquid i s increased by the centrifugal action
In essence, the total mechanical energy of the liquid is increased by the
centrifugal action. Energy is transferred to the shaft by the driver and to
the fluid by the impeller.The pump is simply the means of energy
transfer.The total mechanical energy at the inlet to the pump is the sum
of the static (pressure) energy, the kinetic energy and the potential energy.
A simdar summation gives the total mechanical energy at the discharge.
The difference between the two summations gives the mechanical energy
transferred to the fluid. Some of the mechanical energy transferred to the
shaft is converted to internal thermal (associated with a temperature
increase) energy by shock and turbulence.This internal energy is not
useful mechanically.The ratio of the mechanical energy transferred to the
fluid to that transferred to the shaft gives the pump efficiency.
The rate at which energy is transferred to the shaft is the shaft power, frequently called brake horsepower.The rate at w h c h it is transferred to the
fluid is the fluid power or fluid horsepower. The ratio of the two is also
equal to the pump efficiency (excluding that of the driver).The difference
between the two is the amount of power converted to internal energy
(commonly, and incorrectly, called heat) by various causes, all of which
will be hscussed shortly.
The centrifugal pump is probably the most common pump.The simple
centrifugal pump as shown in Figure AIII-1 is most frequently directly
connected to an electric motor operating at constant speed. In those areas
where 60-Hertz power is available, that speed is often 3600 rpm or 1800
rpm less the slip (resulting in 3,550 rpm or 1,750 rpm).We will start our
analysis with this pump in mind.
Pump engineering i s a very mature technology
Pump engineering is a very mature technology. Pump manufacturers,
therefore, can guarantee the performance of their pumps with a great
degree of accuracy - much more than can manufacturers of other types of
process equipment.
274
and Equations
Computations Involving Pumps for
Liquids
A P P E N D I X A I I I
When the author was a very young engineer, he heard stories from older
engineers in which they would complain about horsepower ratings of
pumps. It seemed that, in the early days, since quality control was not too
tight, when an engineer bought a five horsepower pump, a 7 % h.p. pump
was actually delivered due to the fact the manufacturer had to build in a
tolerance on the plus side.When the technology matured, the manufacturers were able to tighten their specifications and they started to deliver
five horsepower pumps. Older engineers, who had gotten used to talung
advantage of the situation, d d not appreciate the change.
Analysis of an ideal pump to which corrections will be applied

Figure AIII-2 shows the essentials of a volute pump.The flow through the
pump may be analyzed as follows:
1. Liquid enters axially at point (a). It must be under sufficient pressure to
prevent vapor cavities fiom forming withn the inlet as the pressure falls
due to acceleration.
2. Liquid turns fiom an axial direction to a radal one from point (a) to
point (1) where it enters the channel formed by the impeller and the
casing.
3. Liquid flows through the impeller channels fiom point (1) to point (2).
It is accelerated by centrfigal force.This acceleration requires work to be
supplied continuously to the impeller shafi.The work is transformed to
lunetic energy of the fluid by the impeller.
4. Liquid leaves the impeller at point (2) and moves around the volute to
the dscharge flange at point (b). Its direction has been changed again,
from a radal to a tangential one.
Pump data is usually referenced to the inlet and outlet flange centerlines.
Analysis requires that the pump be divided into three volumes: inlet, channels and outlet. The heart of the pump is the impeller, containing the
channels, so we will begin the analysis with it.
Figure AIII-3 depicts an impeller, its direction of rotation, and the velocity
vectors involved. Only one blade is shown.
At steady state, fluid enters the channel at an angle governed by the blade
angle, PI. T h s is the angle between the (usually) backward leaning blade
and the tangent to the circle described by its leading edge.The liquid has
an instantaneous velocity, relative to the blade, of vl. At the same time the
and Equations
275
A P P E N D I X
A I I I
Computations
Involving Pumps for Liquids
inner tip of the blade is moving with a tangential velocity of ul.A fluid
particle entering a channel has a resultant velocity at this point of V1.
The resultant velocity is given the unfortunate name "absolute" velocity.
The angle between the absolute velocity vector, vl ,and the tangent to the
circle at the inner blade tip is al.
The absolute velocity is simply the velocity that a particle has relative to
the frozen reference fi-ame of the observer. In other words, if we could take
high speed photos of the fluid particles separated by rmlliseconds, a particle
would appear to be moving in the direction of the absolute velocity vector
and the instantaneous absolute velocity could be computed.
Figure Alll-2. Centrifugal pump showing 8ernoullii stations
276
F l o w o f ~ n d u s t r i a lF l u i d s - T h e o r y
Computations Involving Pumps for Liquids
A P P E N D I X
A I I I
Figure Alll-3.Velocity vectors at entrance and discharge of centrifugal pump

vane.
Leaving the channel, the fluid particles have been accelerated; they have
gained kmetic energy from the work of the impeller. Since they are
flowing along the curved blade, the direction of the relative velocity, v2, is
controlled by the angle at the tip of the blade.The angle between the tangent to the circle at the tip and the relative velocity is P2.The tangential
velocity of the tip is u2. Since the particle is moving relative to the blade
and the tip is moving tangentially, the resultant "absolute" velocity can be
obtained by the parallelogram law. The vector,V2,is the result.
VectorV2 is the instantaneous velocity that a particle has as photographed
by a high-speed camera talung pictures in a parallel plane to that of the
plane of the impeller.The angle between the absolute velocity and the
tangent to the circle at the tip is a2.
I
r.?
J-
Figure Alll-4.Vector diagram at tip of centrifugal pump vane

and Equations
277
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
Simplifying assumptions
For the sake of the analysis, simplifjring assumptions (to be corrected later)
are made. These are:
1.The point velocity withm all channels is identical across a cross section of
the channel.This assumption is equivalent to assuming no slip between
layers of fluid and no recirculation w i t h the channel.
2. The angle P is the actual vane angle.This assumption is equivalent to
assuming what is called perfect guidance. It would be true only if there
were an infinite number of curved vanes. In truth, the fluid leaves a
channel at varying angles because of varying velocities in the same
channel.
For fluid leaving the periphery, the velocity vectors and the tip radius can
be depicted as in Figure AIII-4.The absolute velocity vector,V2,has been
resolved into a radial vector,Vr2,and a tangential one,Vu2.We are now
ready to proceed with the analysis.
Analytical principle
In physics and engineering, much economy of effort can be made by
finding a universal principle that is always true and then by applying that
principle lrectly without concern for its development.
Newton's law
In linear motion, such a universally true principle is Newton's law, which
relates the acceleration of a body (a mass) to the sum of the external
forces applied to it.The simplified version of this law is stated as force
equals mass times its acceleration, Equation AIII-1. The more fundamental
version, which also applies to systems of changing mass, states that the sum
of the external forces is equal to the time rate of change of linear
momentum, Equation AIII-2. The dimensional constant may be ignored if
S1 units are used.
(AIII- 1)
It should be noted force is intimately associated with mass through these

equations; we recognize the presence of forces by their influence on masses.
278
and Equations
Computations
Involving Pumps f o r
Liquids
A P P E N D I X
A I I I
Rotary analog of Newton's law

In rotary motion, the analog of Newton's law relates torque to the time
rate of change of angular momentum, Equation AIII-3.
The units of torque are foot-pounds-force or newton-meters.The angular

momentum is the term in parentheses.The mass of a particle is m. The
radms, r, is that of the instantaneous circle of revolution of the particle of
mass.The tangential velocity isV, (whch is calledVu2in Figure AIII-4), the
tangential component of the absolute velocity. The torque required to effect
a time rate of change of the terms in parenthesis is ofien given without the
summation sign; summation of all external forces being understood.
Note torque is not energy in spite of having the same units. Torque represents force applied at a distance, not force travehng through a distance.
If we draw a control boundary so as to include the volume withn a
channel of Figure AIII-2, mass is seen to cross the boundary at two sections, (1) and (2). No mass crosses the boundaries other than at the inlet
and outlet.The rotary speed of the channel and the mass flow rate wdl be
assumed constant.
The external torque associated with the passage of a particle at each of the
sections is given by Equations AIII-4.These equations are the rotary
analogs of equations AI-24.
The tangential velocities and the radii are constant.The quantity of mass
changes as it enters (disappears into) and leaves (emerges from) the control
volume.This change is equal to the mass flow rate and is the same for
both stations under the steady state assumption. If the rate of change of
angular momentum is greater at the exit than at the inlet, the external
torque is positive.Torque has to be applied to cause the change. If the
torque were negative, we would be dealing with a turbine, and energy
would be extracted.
and Equations
279
A P P E N D I X
AIII
Computations
On the assumption of perfect guidance, all particles have the same mass
flow rates, so all of the mass flowing in a channel, and indeed all mass
flowing in the impeller, can be lumped into the mass flow rate term.The
sum of the external torque applied to each particle is the total torque
applied to all the mass flowing through the pump.
Simplification
By design of the blade at the inlet,for a givenJow rate, the absolute velocity
vector at the inlet can be made to point radially, at ninety degrees to the
tangential vector. This means its tangential component is zero and
Equation AIII-4 reduces to Equation AIII-5.
The power (energy per unit time) required to maintain a body in circular
motion against a resisting torque is given by equation AIII-6. We have
dropped the subscript, ext, for convenience.
P=ZW
(AIII-6)
The units of torque are foot-pounds-force or newton-meters.The angular
velocity, m, has units of inverse time. So, the units of power are footpounds-force per second or newton-meters per second or joules per
second, depenlng on the quantities of measure chosen. Equation AIII-7 is
used with customary U.S. units to express power as horsepower.
(AIII-7)
p=- zw
550
There are 550 ft-lbf/s in one horsepower (the average rate at which a
theoretical horse can do work).
The work done per unit mass is obtained by dividing power by the mass
flow rate (energy per unit time dwided by mass per unit time equals
energy per unit mass).
The units are foot-pounds-force per pound-mass or, if S1 units are used
and gc equals one and is dimensionless,they are newton-meters per lulogram or joules per lulogram. Note the work and power are the work and
280
~ I o wo f I n d u s t r i a l ~ I u i d s - ~ h e o r ya n d E q u a t i o n s
Computations
Liquids
A P P E N D I X
A I I I
power imparted to the fluid. Keep in mind the same names are used to
express quantities that may differ by efficiency factors. The brake horsepower or the shaft horsepower wdl be greater than the fluid horsepower.
Equation AIII-8 applies to the case when the inlet tangential velocity of a
particle is zero.
Relating channel hydraulics to inlet and outlet hydraulics
The above equations refer to flow through the channels from points (1) to
(2) of Figure AIII-2.We have to relate the inlet (a) and outlet (b) flows to
these equations.The ~ernoull?equation can be written between points
(a) and (1) and between points (2) and (b). Elevation (potential energy)
changes can be neglected and there is no pump work in these sections.
The alpha correction factors are applied to the flows in the conduits
because of the turbulent profile within these conduits and the use of
average velocity, U.They are not applied to the channels because of the
assumption of perfect guidance (parallel flow).The velocities represented
by U1 and U2 are those perpendicular to the entrance to and exit from
the ideal channels.
Within the channel, when the inlet tangential velocity is zero, a Bernoulli
equation can be written linlung theoretical work per unit mass flowing to
the fluid energy terms, temporarily neglecting elevation changes and losses
but including the work per unit mass done on the fluid.
(AIII- 10)
We can now substitute equations AIII-9 into equation AIII-10 to give,

(AIII- 11)
r2V,,,m
P, a b v , '
0, a a u f
g,
2g
'.=-=[P+lg)-[p+
So, the external work per unit mass of fluid passing through the pump can
be measured by the radius of the pump impeller at the tip, the tangential
velocity component of the flowing fluid at the impeller tip and the rotational speed in radians per second. It can also be measured by the dfferFlow of Industrial Fluids-Theory
and Equations
281
A P P E N D I X
A I I I
Computations
ences of the sums of the static (pressure) energies and the kinetic energies
measured at the outlet and inlet flanges. From a practical point of view,
the latter form is more useful. The hnetic energies have a correction factors, alpha, because the average velocities, U, were used for the pipe flow.
The inlet tangential velocity is zero.
The first sum in parenthesis on the right of AIII-l l is called the "dscharge
head" and the second one is the "suction head".The hfference is called the
total dynamic head or developed head.The indvidual terms are often
called pressure heads and velocity heads. It is important to keep the dstinctions between pressure and energy clear and not to fall into the trap of
using the word "head" indiscriminately and to assume it simply means discharge pressure. In our terms, head is energy per unit mass. It is the author's
contention that it is easier to make the distinctions among the various head
terms when energy per unit mass is the conceptual tool rather than feet or
meters.This is especially true when the fluid is compressible.
Expressed in terms of head in foot-pounds-force per pound-mass or in
newton-meters per hlogram, the total dynamic head is given as,
The subscript, (ideal),is to remind us we have not yet applied correction

factors for the assumptions that were made.The suction and dwharge heads
can be computed fiom measurable pressure, flow rate and pump inlet and
outlet internal &ameters.They give the ideal total dynamic head drectly.
Relation of t o t a l dynamic head to volumetric flow i n an ideal pump
The volumetric flow through a pump is given by the following equation.
(AIII- 13)
%deal
= '2nd
Ap
The units are either cubic meters per second or cubic feet per second.
The velocity is the velocity perpendicular to the circumference of the
impeller.The area is the sum of the areas of the channel cross sections at
the periphery.
and Equations
Computations
Involving Pumps for
Liquids
A P P E N D I X
AIII
From Figure AIII-4, the tangential component of the absolute velocity is

given in Equation AIII- 14.
(AIII- 14)
From Equation AIII-12 combined with AIII-13 and AIII-14 and the
assumption of zero inlet tangential velocity, we can derive the following
relationship.
(AIII- 15)
It can be inferred from Equation AIII-15 that the ideal total dynamic
head, TDH, is proportional to the tangential kinetic energy.The equation
also shows the ideal pump, at a constant rotational speed, has a linear relationship between the total dynamic head (energy transferred to fluid by
the pump per unit of flowing mass) and the volumetric flow rate. More
importantly, it shows this relationship is independent of the fluid. Since we
have not made any assumptions regarding the nature of the fluid, the
above equation is also valid for real fluids (in an ideal pump).
The relationship between head and volumetric flow rate given above is
the main reason pump manufacturers use head and not pressure on their
curves. An economy of description is gained. One curve fits all fluids
pumped through the same pump. If manufacturers used pressure on their
curves, they would have to specifi the fluid density.
Caveat
- Conversion of 'head'
to pressure
Another word of warning is warranted here. To obtain pressure units, it is
simply necessary to multiply head as defined above by density in compatible
units. However, total dynamic head includes both the energy per unit mass
associated with pressure and that associated with velocity. So multiplying by
density does not give the Merential pressure across the pump. It gives a
number that differs fiom the dd5erentia.l pressure by the dfference in
velocity pressures (the pressure generated in reducing the velocity to zero).
Since thls number is not greatly Merent fiom the Merentid pressure, it is
often assumed the two numbers are one and the same, but this is not so.
and Equations
283
APPENDIX AIII
The blades on pumps are usually curved backwards to the direction of

rotation so the blade angle, P, is less than 90 degrees and the tangent of
the angle is positive.With some fans and blowers, the blade angle may be
curved in the forward dlrection.This is not often done with pumps
because it produces problems of flow instabhty.
Equation AIII-15 shows that if the tangent of the angle is positive, the
(ideal) head curve will slope downwards with increasing flow. If it is infinite, the slope d be horizontal for all flows. If the tangent is negative,
the slope of the TDH curve versus q d be positive-These characteristics
correspond to blades whose leaving angles are less than 90 degrees, equal
to 90 degrees and greater than 90 degrees when measured in the clockwise direction from the tangent line.
Alll-3: PERFORMANCE OF REAL CENTRIFUGAL P U M P S

The developed head (the total dynamic head) of a real centrifbgal pump is
much less than that of its ideal model.The rate at which mechanical
energy is transmitted to the fluid is less than the rate at which it is transmitted to the shaft.The fluid horsepower is less than the brake horsepower. The efficiency of the pump excluding the driver, q, is less than
one. Efficiency here is being defined as mechanical energy per unit mass
flowing transmitted to the fluid dwided by energy per unit mass flowing
transmitted to the shaft.
Head losses versus power losses
"Losses" are frequently classhed as head losses or power 1osses.The dstinction comes about because some deviations fiom the ideal pump model cause
a lower total dynamic head without causing a decrease in the energy per unit
mass pumped.Thls fact is of interest theoretically and for ddactic reasons.
Head losses: Correcting the previous assumptions - imperfect guidance
In the real pump, guidance is not pe&ect. Flow is not parallel through the
channels.There may even be reverse flow in portions of the channel
causing a circulating pattern to be imposed on the overall flow.
Two vector diagrams are given in Figure AIII-5.They may help explain
the real situation.These hagrams represent a fixed volumetric flow rate
and a fixed pump speed.The right hand diagram (dashed lines) represents
the case of perfect guidance; the left-hand diagram (solid lines) represents
a real case.
Computations
Involving
Pumps for
Liquids
A P P E N D I X
A I I I
The bases of the hagrams are identical in length because the peripheral
velocity, u2, is identical in both cases. Sindarly, the altitudes of the triangles
represent the ideal radial velocity on the right and the actual r a l a l
velocity on the left. Both of these velocities are related to the volumetric
flow rate and are equal.
The so-called absolute velocity of the fluid decreases in the real case due
to the lack of perfect guidance.The stream leaves the impeller at a more
backward angle so that the angle, P, is decreased.The tangential component of the absolute vel~city,V~,~,
also decreases (Vu2+ Vfu2).Since by
Equation AIII-15, the total dynamic head is proportional to the tangential
velocity, the total dynamic head decreases in the real case.
The energy requirement per unit mass is proportional to the tangential

velocity by Equation AIII-g.Therefore, power decreases with the decrease
in tangential velocity and the efficiency is not affected by this isolated
cause (of imperfect guidance).
The reduction in total dynamic head due to less-than-perfect guidance is
quite severe.The reduction is greatest at low flow rates and least at the
design point when the flow streams become more nearly parallel. If we
plot ideal head versus volumetric flow as in Figure AIII-6, we can depict a
new theoretical head discounted for guidance losses by a line below the
ideal head curve.This line will have a slightly less negative slope. The area
between the two lines represents the losses.
and Equations
bi,
-'B
g .E
g' n.2
Figure AM-5.Vector diagrams describing perfect and imperfect guidance
5-
285
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
V O L U M E T R I C FLOW RATE, q
Figure AM-6. Reduction in theoretical 'head' due to various causes
Head losses: Correcting the previous assumptions - irreversibilities

According to the complete ~ernoulli'equation, fluid irreversibilities
decrease the developed head (TDH). Useful mechanical energy is converted into unrecoverable internal energy, which manifests itself by a temperature rise. According to the arc^"" equation, these losses increase with
the square of the flow rate. Figure AIII-6 shows the line representing head
discounted for so-called friction losses. It has the same origin as the
second line above, but it is concave downward. It diverges increasingly
with higher flow rates from the previous curve.
Head losses: Correcting the previous assumptions - shock losses
The father of thermodynamics, Sadi carnot', had a father, General
Hypolite Carnot.The grandfather of thermodynamics not only was an
army general under Napoleon, he was an important politician (member of
the Directorate) during the French revolution and he was a hydraulician
who wrote extensively on such things as shock losses in turbomachinery.
Hypolite Carnot was well aware that if a fluid did not enter a channel at
the same angle as the blade or if it left at an angle different from that of
the surroundmg fluid, a loss in available mechanical power occurred due
to "shockfl.The maximum mechanical power could not be generated
unless shock was minimized. Some have suggested the concern of
Hypolite Carnot for efficiency and its relationshp to shock losses was the
stimulus that lead Sadi Carnot to develop his theories on ideal thermodynamic engines and on the second law of thermodynamics.
286
F I O W
o f I n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
APPENDIX
AIII
For our purposes, it should be realized a sudden change in flow direction

causes turbulence, loss of head and conversion of mechanical energy into
internal thermal energy.This change in direction can be due to a mismatch between the incoming fluid velocity and that of the leadmg edge
of the impeller blades. It can also be due to the sudden introduction of a
leaving stream at an angle to a circulating stream. It results in a lower
mechanical efficiency, q .
Corrective measures - diffusers
Large volute pumps have huge losses due to shock. Frequently, stationary
dffusers are added to guide the fluid more smoothly to the pump exit as
shown in Figure AIII-7. Diffusers, as well as guiding the fluid smoothly,
reduce its velocity and increase its pressure by providing a smoothly
expanding path.
DiGsers perform two functions-They reduce shock losses and they convert kmetic energy to static (pressure) energy.
stafuonory
diffuser
Figure Alll-7. lmpeller and stationary vanes in diffuser pump
Best efficiency point

A fixed speed centrifugal pump is designed to be most efficient at a fixed
flow rate. Moving away from t h s flow rate in either direction causes the
efficiency to decrease. More energy is consumed per unit mass being
pumped.The point at which the efficiency is greatest, the best efficiency
point, corresponds to the design flow of the pump and the design head
(TDH).
The shock losses increase on either side of the best efficiency point. In
Figure AIII-6, the shock losses are shown to discount the head versus flow
and Equations
287
A P P E N D I X
A I I I
Computations
curve.The combination of mechanical energy losses produces the curve

that is typical for a particular pump design.
Other power losses i n a pump
Internal circulation
The pump impeller has clearance between it and the casing.The high
pressure at the discharge and low pressure at the suction causes liquid to
circulate internally between the impeller and the casing. This results in
"disk friction" (again, mechanical energy converted to internal energy).
The internal circulation is maintained at the expense of energy that would
otherwise appear as energy within the fluid.
Bearing losses
Mechanical "friction7'is present in pump shaft bearings. "Friction7' is a
force against the drection of motion that must be overcome by a force
supplied by the driver.Additiona1energy per unit mass must be supplied
to the shaft to overcome all of the above mechanical energy 1osses.The
rate at which energy is supplied to the shaft constitutes power.
Theoretically, the following equation is true.
(AIII- 16)
So, the theoretical power is close to a straight-line function of the mass

flow rate or, since liquids can be considered constant density fluids, of the
volumetric flow rate.The rate at which energy is transmitted to the fluid
is the hydraulic power.The rate it is transmitted to the shaft of the pump
is the shaft or brake power.The ratio of the fluid power to the shaft power
is the pump efficiency exclusive of any driver's efficiency, 11.
Power versus volumetric flow - a schematic description
Figure AIII-8 is a plot of power versus volumetric flow in a pump.
VOLUMETRIC FLOW RATE.q
Figure Alll-8. Power versus volumetric flow in a centrifugal pump

and Equations
Computations
A P P E N D I X
A I I I
The uppermost line represents theoretical power supplied to the shafi frequently termed brake horsepower.The line starts at some positive value
corresponding to all of the losses occurring when the pump is deadheaded (operating against a closed valve).As flow increases, the power to
the shaft increases more or less linearly.
Several lines are drawn parallel to and below the brake horsepower line.
The vertical distance between each line represents the power lost for various reasons: bearing friction, disc friction, leakage and internal circulation.
Below the lowest straight line is drawn a curved 1ine.The intervening
space represents irreversibhties due to shock. The line is convex upward.
This line starts at zero flow where all of the remaining irreversibilities are
those due to shock. It comes close to the lowest straight line at the design
operating point where the shock losses are minimized. It diverges from the
lower straight line from there onwards as the shock losses increase again.
Below the curved line representing power discounted for shock, leakage,
disc friction and bearing losses, there is another curved line that represents
power actually transmitted to the fluid as mechanical power.The space
between this curve and the previous one represents losses due to flow
through the pump.The two lower curves diverge at a rate roughly proportional to the flow rate squared.This causes the fluid power curve to have a
definite maximum at its design operating point.
Head, efficiency and power versus volumetric flow
In customary U.S. units, the head in feet (which is really foot-poundsforce per pound-mass), the efficiency and the brake horsepower are frequently plotted versus volumetric flow. Each plot wdl be read against a
different ordmate scale.The horsepower scale usually starts above zero.
The brake horsepower curve usually slopes upward from its minimum
value in a fairly linear fashion.The efficiency curve curves upward from
zero and peaks at the design operating point. O n most process pumps, the
head curve slopes downwards following a curve that is concave downward.The curve can drop off quite radically past the design operating
point giving rise to the expression,"the pump ran out on its curve".
Pumps with high heads and high speeds often have a peaked head curve.
There is usually a definite hump close to shutoff.This type of pump has
limited rangeability. If the pump is throttled too far back on its curve it
and Equations
289
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
can become unstable because there are two flow rates that can be in equilibrium with the same total dynamic head.The pump will hunt between
the two equilibrium states and the surging flow can cause damage to the
driver and the pump.
Maximum head per stage and multistage pumps
The tip speed of a centrifugal process pump limits the maximum practical
total dynamic head to between 70 and 100 foot-pounds-force per poundmass or in terms of feet of fluid 70 to 100 feet.This is about 20 to 30
meters or 210 to 230 newton-meters per lulogram (or J/kg).
If we take 85 foot-pounds-force per pound-mass or 85 "feet" as the
average, such a pump working on 60F water w d produce 85 ft-lbf/lbm
times 62.4 divided by 144 or 36.8 psi of differential pressure. Increasing the
pressure substantially requires multistaging, or a different design, or both.
In multistaging, the developed heads per stage add to give the overall
head. The overall efficiency is found by multiplying the stage efficiencies.
If they are all equal, then the overall fractional efficiency is equal to the
single stage fractional efficiency raised to a power equal to the number of
stages. A seven-stage pump with an efficiency per stage of 90% would
have an overall efficiency of 47.8%. Better than half the energy input to
the shaft is wasted in the conversion to internal energy whose first manifestation is an increase in the temperature of the fluid.
Use of the ~ e r n o u l l equation
i~
with pumps
We will repeat the most general form of the Bernoulli equation and then
apply it to a system containing a pump, Several examples of computations
wdl be given.A simple system consists of an open tank, a pump and associated piping.The discharge piping fiom the pump feeds an open receiver.
The surface of the liquid of the first tank is designated as point 1.The open
dscharge pipe above the receiver is point 2. The following can be written:
The term, wiz, is the energy the pump transmits to the fluid per unit mass
of flowing fluid. The term, h12,is the energy per unit mass of flowing fluid
converted to internal energy.The suffix, 12, on each term is meant to
show these terms refer to mechanical energy per unit mass that is added
or converted to thermal energy within the system somewhere between
sections 1 and 2.The subscripts, 1 and 2, indicate the energy per unit mass
is associated with the specific sections.
290
FIOW
o f ~ n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Computations
Liquids
A P P E N D I X
A I I I
The term, hlz, is usually called a loss term.These irreversibilities occur in

the fluid in the piping, not in fluid in the pump. The term, w12,refers
only to energy transmitted to the fluid.The pump also has losses not
included in the piping loss term.These losses occur due to friction in
stuffing boxes and bearings and in fluid irreversibilities within the pump.
The mechanical energy transmitted to the fluid is equal to that transmitted to the shaft less the sum of the losses within the pump.
w12= wS-hp vwS
(AIII-18)
The fluid energy per unit mass transmitted by the pump is given in terms
of the shaft energy per unit mass multiplied by an efficiency factor that is
always less than one.
The ~ernoulli'equation with pump losses and pump shaft work taken
into consideration takes on the following form.
This equation is applicable to customary U.S. units and to S1 units if the

dimensional constant is taken as one with no units.The subscript, S,means
the work is that transmitted to the shaft.
Note equation AIII-19 describes energy changes between sections 1 and 2
external to the pump.The shaft work is transformed to hydraulic work by
the efficiency factor whch takes into consideration pump irreversibilities.
The mechanical energy losses given by h12 are those external to the pump.
Total heads
The concept of "total" head refers to a summation of indwidual heads at

the pump inlet flange and at the discharge flange.These locations are designated in the following equation by points 1 and 2.These locations are
not the same as those in equation AIII-19.The difference in head is also
given the designation "total" (total dynamic head), so the adjective "total"
becomes meaningless through overuse. The total dynamic head, TDH, is
the mechanical energy imparted to the fluid by the pump impeller per
unit mass of flowing fluid.
and Equations
291
A P P E N D I X
A I I I
Computations
When analyzing a pump, we draw an imaginary boundary around it cutting through the inlet and outlet flanges.The rest of the physical system is
considered only through its influence at the boundaries. Only the external
influences at the flanges (pressure, velocity and elevations) are considered.
Since we have reduced the system to what exists between these two
points, the piping loss term, h12, drops out of the ~ernoulli'equation.The
efficiency factor remains with the shaft work term.
~ w =,h, - h, = AH = (TDH)
The first line of Equation AIII-20 is a specific energy balance. It states that
the specific (per unit mass) energy transmitted to the fluid is added to the
total incoming specific energy, subscript 1.The sum equals the outgoing
specific energy, subscript 2. The arrangement of the equation shown allows
computation of the energy requirement at the shaft of a pump in order to
effect the change shown on the right.
Analysis o f terms
It is worth analyzing the above set of equations.The first term on the left
in the first and second equations is the energy per unit mass flowing delivered to the fluid by the pump. In the last equation, it is the energy per unit
mass of flowing fluid delivered to the shaft. The units are foot-pounds-force
per pound-mass or, if gc is treated as one, no units, and S1 units are used
elsewhere, they are joules per lulogram or newton-meters per lulogram.
The first term on the right in parenthesis ofAIII-20 is the total discharge
head, which can be replaced by the sum, h2.The units of both terms in
parenthesis are the same as those of the input term.The second term
parenthesis on the right is the total suction head, which can be replaced
by the sum, hl.
The only arbitrary terms in the equations are the elevations, Z1 and Z2.
They are arbitrary because they depend on a datum selected for convenience. Since the results of the Bernoulli equation depend upon dfferences,
the datum selected will not change the results.The values do depend
upon the system being analyzed.The subscripts, 1 and 2, must be used
consistently in a given problem.
and Equations
A P P E N D I X
A I I
The difference between the total lscharge and the total suction heads,
AH, is called the total dynamic head. It is the actual energy per unit
flowing mass transmitted to the fluid at a particular flow rate. It wdl
change over the operating range of the pump.
The three components of each of the total heads are also called "heads".
The first component is the static or pressure head, pressure divided by
mass density in the above equation - pressure lvided by weight density in
the hydraulic equations.The second component is elevation head.The
third component is the velocity head.The first component is numerically
the largest contributor to total head.The other two are relatively small
when their differences are considered.
The use of the word "head" to describe at least six different components,
the dfferences between them and the dfferences of their "total" sums
obviously can lead to confusion and error. The author has made such
errors himself and has observed that, even in very respected publications,
such conceptual errors have been made. It is very easy to use the word
"head" in a general sense and then to f d into the conceptual trap of misapplying it in specific computations. If you use the word, you are advised
to qualify it always with words such as "total discharge7',"total suction",
"total dynamic", "velocity", "static suction", and "static dischargeW.The
point has been belabored sufficiently.
Power to the shaft

The power to the shaft in terms of horsepower is called the "brake"
horsepower. Presumably, the term grew out of widespread use of the
Prony brake to measure power.The horsepower unit was a natural development at the time when horses were a common means of motive power.
It has little meaning today. It, nevertheless, retains its popularity in many
sectors, but is a confusing term when conversions are made. For instance
one horsepower is not exactly equal to one cheval vapeur.
Power to the shaft is the rate at which energy is transmitted to the shaft.
Fluid power is the rate at which it is transmitted to the fluid.The ratio is
the efficiency - power in divided by power out or energy in lvided by
energy out. Power to the shaft is distinct from power to the driver.The
input power to a shaft is the output power from a driver.The input power
to the driver wdl be somewhat greater.The driver has its own efficiency.
The set of equations in AIII-21 describes shaft (subscript S) and fluid (subscript f) power and horsepower.The latter is in U.S. units only.
and Equations
293
A P P E N D I X
A I I I
The first three equations in the set AIII-21 may use S1 or customary U.S.
units. In both cases the units are energy transmitted per second - footpounds-force per second or joules per second. Eta is the kactional efficiency.The last equation uses customary U.S. units only. The coefficient,
1/550, is the conversion between foot-pounds per second and horsepower.
A l l l - 4 : REAL CENTRIFUGAL P U M P S - SUCTION LIFT,

CAVITATION AND NPSH
Equation AIII-19 gives the general relationship describing systems containing pumps.The pump work term and the piping loss term are not
point functions.The loss term is distributed across the system external to
the pump and the pump term does not take into account its specific location.Within limits, the pump can be anywhere within the piping system.
In this section we will establish the limits.
The suction pressure can be below atmospheric pressure, but the total
mechanical energy at the suction cannot be below the energy associated
with the vapor pressure of the liquid. If it is, flashing may result and perhaps cavitation. Flashing w d not occur if the sum of the velocity head
and the pressure head, in other words if the sum of the kinetic energy and
the static energy, is sufficiently greater than the energy associated with the
vapor pressure of the liquid at the flowing temperature.
The criterion to avoid flashmg is stated more specifically in equation AIII-22.
In this equation the subscript, f, is used simply to indicate the location is a

specific one anywhere in the fluid stream.The subscript, fv,indicates vapor
294
F l o w o f ~ n d u s t r i a~ ~~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Computations
A P P E N D I X
A I I I
pressure at the temperature of the fluid at the 1ocation.The velocity profile

coefficient, a f , can be ignored in most cases (can be equated to one).The
equation applies to the S1 system if the dimensional constant is made
equal to one and is dimensionless. It is to be noted the pressure at the
location can be below the vapor pressure as long as the kinetic energy
term is sufficiently great.
An explanation of equation AIII-22 is vapor pressure is an inherent characteristic of matter and is a function of temperature.The tendency to
vaporize that a liquid possesses is inhibited by externally imposed forces.
The term on the right of the inequality represents the static energy the
fluid has inherently, because of its temperature.Ths energy tends to separate the molecules.
If there is no motion, there is no bulk hnetic energy and only the
external pressure energy opposes the internal inherent energy.When the
sum on the left dips below the term on the right, equilibrium is dsturbed
and vaporization occurs.The temperature drops to a new equdibrium
temperature.A portion of the liquid is vaporized.
NPSH
The excess of the sum of the hnetic energy and the static energy over the
inherent energy represents a measure of safety in the avoidance of flashing
and potential cavitation. It is given the name "net positive suction head".
This head is computed from conditions measured at the inlet flange of a
pump. Equation set AIII-23 describes the concept.
(AIII-23)
aus2
P+
% ?PS2
+ $-$v
NpS&4-HSv =
L
1---28,
PS
P,
2g,
P,
The subscripts reflect the fact we are referring to conditions at the suction
(inlet) flange of the pump. The letter A added to NPSH means available.
Suction l i f t
The concept of net positive suction head applies to pumps, even when
they are operated in a suction lift mode. Pumps operating in t h s mode
must be primed before they are started (see Section 111-5, Priming
Centrifugal Pumps Operating in a Suction Mode).
and Equations
295
A P P E N D I X
A I I I
Computations
The ~ernoulli'equation can be written for the suction piping of a system

operating under suction lift and the NPSHA equation, AIII-23, can be
combined as follows:
(AIII-24)
9,
2g,
g,
Assumptions
The assumptions in the above development are: the datum elevation is the
liquid level in the feed tank or sump, the velocity at that point is essentially zero, there was no additional pump in the pipe being considered and
the pump was already primed and ready to go when started.The terms
involving elevation and hnetic energy at point 1, and pump energy
between point 1 and the pump suction are all zero.The density is assumed
constant between point 1 and the pump suction.
Analysis
The development started with the general Bernoulli equation, it eliminated
the terms that were made equal to zero by the stated assumptions and then
rearranged the equation so it could be substituted into the NPSHA equation, AIII-23.This allowed the NPSHA equation to be written in terms of
variables that are known to be true for the stated problem.
The last equation in the set AIII-24 states that the NPSHA, or the theoretical safety factor, in energy units per unit mass, is equal to the algebraic
sum of three terms.
The first term is difference between the pressure on the liquid surface in
the feed tank and the vapor pressure of the liquid at the suction flange
dtvided by the liquid density.This is the most important term.Vapor pres296
FIOW
o f ~ n d u s t r i a~~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Computations
Involving Pumps for liquids
A P P E N D I X
A I I I
sure and the pressure on the surface are expressed in absolute units.The
vapor pressure and the density are measured at the flowing temperature at
the suction flange.
The second term represents the irreversibilities in the inlet section. It must
be in the same units of energy per unit mass and it must include all losses
- pipe entrance, any foot valves, other valves, pipe bends or strainers. It is
subtracted from the first term because it represents a loss of safety factor.
The last term represents the loss of safety factor due to the elevation of
the pump suction being higher than the surface of the liquid. It is also
subtracted from the first term. If the pump is below the surface, this factor
will become positive.The sign will be changed automatically if distance
below the datum is considered to be negative.
The last equation of the set AIII-24, as written, has units of foot-poundsforce per pound-mass or joules per lulogram depending on the units
chosen and the use of the dimensional constant. If the same equation is
multiplied through by the density, the units of each term become those of
pressure. If the original equation is dlvided through by the ratio g/gc, the
units of each term become those of manometric head, feet or meters.
U t i l i t y o f NPSHA and NPSHR
As long as a pump is operating within its design constraints, little attention
need be paid to the net positive suction head available or required. In
addltion, operations and maintenance personnel have little say in the initial
choice of a system, so they tend to assume a lack of interest is justified.
It is only when a pump has been rnisapplied, or the fluid temperature has
changed radically and troubleshooting is necessary, do operating or control
personnel become involved with NPSH problems. Problems involving
NPSH occur when the NPSH required by the pump is greater than that
available.When this happens, it is usually accompanied by symptoms such
as unusual noise at the pump or loss of pumping ability. If it is allowed to
continue, cavitation damage can occur in the impeller channels and in the
casing.
Troubleshooting involves makmg sure there is no unnecessary blockage in
the pump suction lines caused by, for example, plugged strainers or partially closed valves.The next step is usually to check the pump data versus
the operating conditions.The pump data should give the design condiFlow of Industrial Fluids-Theory
and Equations
297
A P P E N D I X
A I I I
Computations
tions and an NPSHR cuwe.The actual operating condtion can be used

to compute the NPSHA, at least at one point. It is important the data be
exact, not old nor incorrect due to measurement errors.Temperature is
especially important, or more exactly the vapor pressure is important.
We will give some examples of computed NPSHA following the summary of equations to follow.
Summary of equations using different units for NPSHA
The equation set AIII-25 summarizes transformations between units used
to compute NPSHA.
In hydrauhc practice, the denominators of the pressure terms in the last two
equations are called specific weight, y.The units are force per unit volume.
A l l l - 5 : P O S I T I V E DISPLACEMENT P U M P S
We have already discussed the principal difference between a positive displacement pump and a centrifugal pump.Thls difference is due to the
incompressible nature of the liquids and the volumetric displacement of
positive displacement pumps. It is not possible to throttle these pumps as a
means of controlling flow. Some path for fluid flow must always be open.
The size of the section of this path depends on the backpressure that can
be tolerated by the pump and piping system.
298
F l o w o f ~ n d u s t r i a l~ I u i d s - ~ h e o r y a n d ~ q u a t i o n s
Computations
Involving
Pumps for
Liquids
A P P E N D I X
A I I I
Suction l i f t and cavitation

Positive displacement pumps, if their sealing mechanism and check valves
are adequate, can elevate liquid to close to the theoretical limit of static lift
without the presence of a foot valve, once primed. It has already been
pointed out in Chapter I11 that a reciprocating pump requires an additional acceleration head in order to make sure the liquid is always in contact with the receding piston.
Positive displacement pumps are not immune to cavitation.The amount of
damage done will depend on their construction and speed.
Discharge characteristics
The discharge characteristics of positive displacement pumps vary from
very smooth with screw pumps to pulsating with reciprocating pumps. The
initial concern for pulsating pumps is purely mechanical. Pulsations are
damped in order to reduce the consequences to the piping and connected
equipment. Spring hangers and adequate supports are used on piping.
Flexible couplings and shock absorbing supports are used on equipment.
Positive displacement pumps are not normally damped with the measurement or control of the flowing fluids in rnind.The primary concern of
the pump engineer or equipment engineer is the longevity of the equipment. Damping for the purpose of measurement and control usually is left
to control systems personnel.
A l l l - 6 : T H E O R Y A N D A N A L Y S I S O F JET P U M P S
Jet pumps appear simple, but are not. They consist of what looks like a tee
whose branch is the suction and whose run consists of two connections,
the driving fluid connection and the discharge connection. The names
given to the different types ofjet pumps in an attempt to describe the
various possible combinations of fluids and the various operating configurations, in particular, lead to confusion. In this appendix we are only concerned with jet pumps that have liquids at each connection.
Principle of operation
The way a jet pump functions is as follows:
1. The motive fluid is accelerated to a h g h velocity in the nozzle. Some of
the static (pressure) energy is converted to kmetic energy.The pressure at
the vena contracts drops radically, most often below atmospheric pressure.
and Equations
299
A P P E N D I X
A I I I
Computations
2. The driven fluid is forced into the entrance to the mixing section by the
difference in pressure between the source and the low pressure at the
nozzle discharge.Thls point is common to both fluid streams.
3. The two fluids mix in a highly irreversible manner and momentum is
transferred between them.The mixed fluid leaves the mixing section at
an intermediate pressure, hgher than the suction inlet pressure but lower
than the motive pressure.
4. A ddlixing section decelerates the mixed fluids and converts some of

their kmetic energy back to static (pressure) energy.
If we compare a jet pump to a centrifugal pump, we see one essential dlfference is the motive fluid is added to the driven fluid. More fluid reaches
the downstream sink than left the suction source.This extra fluid is not
present in the case of a centrifugal pump.The centrifugal pump simply
adds energy, not mass, to the suction fluid.Therefore, in comparing efficiencies, only the mass flow to the suction of the jet pump should be considered.The motive liquid could have been let down dlrectly to the sink
without passing through the pump - it had sufficient pressure. When
comparing efficiencies, we should only consider the energy necessary to
move suction fluid from the source to the sink.
Theory of the mixing section - force and rate of change of momentum balance
The following analysis is based on that originally given by Weber and
~eissnerl'. Figure AIII-9 is a simplhed longitudinal section of a jet pump.
We can create a control volume for the purpose of the analysis by talung a
section across the entrance and one at the exit of the mixing chamber. For
simplicity, consider the mixing chamber to have parallel sides.
The external forces acting upon the control volume are due either to
static pressure acting on the cross sectional area, to the forces generated by
the rate of change of momentum associated with the flowing streams or
to viscous drag.
C
Figure Alll-9. Ejector analysis
300
and Equations
Computations
Involving
Pumps for
Liquids
A P P E N D I X
A I I I
The pressure, PI, at the entrance to the control volume can be taken as
constant across the entire section. It is common to both the fluid discharging from the nozzle and the fluid entering from the suction inlet.
The force is the product of this pressure and the total area at section 1. It
acts to the right.The pressure, P2, at the hscharge from the control
volume also acts across the entire area.The associated force is the product
of this pressure and the cross sectional area at section 2. It acts to the left.
The viscous drag force, RD, resists fluid motion and acts to the left.
The sum of the inlet areas can be taken to be equal to the outlet area if
we ignore the nozzle wall thickness.The sum of the inlet mass flow rates
is exactly equal to the discharge mass flow rate.
A steady state force balance can be set up as follows:
Using customary U.S. units, we will assume both streams contain water at
60F, 62.5 lbm/ft3.~hereis no viscous drag (RD is to be neglected) and
the process is adiabatic (no heat transfer to or from the boundary).
Data
A, = 0.25 fi2
Ab = 0.10 fi2
Ac = 0.35 f?
U, = 10.0 fi/s
Ub = 50.0 fi/s
PI = 40.0 psia
The velocity at the discharge from the control volume can be computed
from the continuity equation.
( ~ A P+ P
)A
~ P ) ,= ( ~ A P ) ,
and Equations
301
A P P E N D I X
A I I I
Computations
The mass flow rates can be obtained from the velocities, densities and
cross sectional areas.
If we assume the viscous drag force to be negligible in the force balance

equation, AIII-26, the only unknown is the discharge pressure, P2.
Note the increment in pressure over the mixing section is 4.41 psi.This is
another example of pressure recovery. This increment is before the diffuser
where the pressure wdl be increased even more. A real jet pump wdl be
analyzed shortly in order to compare jet pump efficiency with that of a
centrifugal pump.
Effect of temperature on density

In the above example, an assumption of constant density was made. It is
worthwhile checlung to see if the assumption was warranted. In order to
do this, we wdl make use of the first law of thermodynamics for open
(flowing) systems.The expressions, dh, h, and Ah refer to specific enthalpy,
not irreversibility;H is total enthalpy and J is the conversion factor to
mechanical energy units from thermal units (778.16 foot-pounds-force
per Btu - one in S1 units).
302
and Equations
Computations Involving Pumps for
UdU
6 q + 6 w 5 =dh+-+-dZ
Assume
Liquids
A P P E N D I X
A I I I
gc
gc
dZ =6 q =6w, =0
(h, - ha+b
) = l l . 6 3 / 468.75 = 0.02 Btu / lb,
Ah = CpdT (1)AT
AT
- 0.02OF
This very small change in temperature along the mixing section was due
to the change in enthalpy that, in turn was due to the change in velocities.The actual change in a real case would be slightly greater because
irreversibilities were ignored in the above analysis. However, one can see
the order of magnitude is quite small.The assumption of constant density,
therefore, is not bad. Note that, if the temperatures of the suction and
driving streams are radically different, the density will change along the
mixing section.
and Equations
303
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
Analysis of j e t pump efficiency

We wdl establish the equations necessary to compute the efficiency of a
jet pump on the same basis as that of a centrifugal pump so we may compare the relative efficiencies.The affects of fluid irreversibkties, suction
and discharge elevations and motive pressures wdl be investigated.
Datum
Suction mode
01
Discharge mode
Figure AM-10. Jet pump configurations
Figure AIII-10 gives examples of the two common modes of jet pump
operations.
The following assumptions wdl be made: 1)A strainer and poppet foot
valve are used, adding to the losses in the suction piping.This is done to
prevent back flow of water to the suction tank and to limit the size of
solids so as not to plug the jet pump. 2) The suction line must be at least
0.5 m longer than the suction lift. 3) The suction tank and the discharge
line are at atmospheric pressure.Ths simplifies the analysis and it is quite a
common situation. (When jet pumps are used as injectors, however, they
do operate with different suction and mscharge pressures.)
Definition of efficiency
In the case of a centrifugal pump, efficiency is the rate at which mechanical energy is transferred to the fluid divided by the rate at which it is
transferred to the shaft. An efficiency of one would mean one hundred
percent of the mechanical power to the shaft was transferred to the fluid.
Efficiency is sometimes stated equivalently as the mechanical energy per
unit mass flowing that is transmitted to the fluid divided by the mechanical energy per unit mass flowing that is transmitted to the shaft.
Note a distinction has been made between the two forms of energy transmitted. One form is mechanical energy and the other is thermal energy.
Only the former is considered useful from the point of view of producing
Computations
Involving Pumps f o r Liquids
A P P E N D I X
AIII
work. Note carnot" efficiency is defined as the ratio of the mechanical

energy obtained from an engine to the total thermal energy delivered to
it. In contrast, pump efficiency is defined in terms only of mechanical
energy, fluid energy divided shaft energy.
A jet pump has additional mechanical energy in the discharge stream

because of the presence of liquid from the motive stream.This additional
energy should not be counted in the efficiency computation. The suction
stream is the stream we wish to elevate to a higher energy level-Therefore,
in order to compare a jet pump to a centrifugal pump, we should ratio the
mechanical energy difference projected to the discharge and the suction
flanges of a mass flow equivalent to the suction stream to the mechanical
energy present at the motive fluid inlet.
The rate of transfer of energy from the driving stream is the energy per
unit mass at the motive inlet times the mass flow rate of the motive
stream. In what follows, the subscripts are 1) motive, 2) suction and 3) discharge. Mass flows 2) and 3) are equal if we discount the presence of the
motive stream in the discharge.
Note the definition of efficiency given here does not correspond to any
of the definitions previously given for centrifugal pumps. Our definition is
simply a way of measuring the ratio of the useful energy output to energy
input for purposes of comparison with the efficiencies of centrifugal
pumps. We have assumed the alphas equal to one.
The purpose of establishing the comparison is to try to put an objective
number on efficiency from the point of view of energy used in accomplishing a specific task.We have already pointed out the choice of a jet
pump is made usually on grounds other than efficiency.
and Equations
305
A P P E N D I X
A I I I
Computations Involving
Pumps f o r
Liquids
The ratio of the increment in suction stream mechanical energy to the

motive stream mechanical energy gives the relative efficiency as follows:
The development in equation set AIII-32 shows the relative efficiency of a

jet pump for comparison with a centrifugal pump is equal to the ratio of
the mass flow rates of the suction fluid to motive fluid times the ratio of
the difference in specific energy across the pump to the specific energy in
the motive stream.The elevation, Z1,in the case of a jet pump operating
in the discharge mode, is practically zero. The elevation enters into the
efficiency computation only when the suction mode is used.
Computing the terms of the efficiency equation

Since the discharge pressure and the inlet pressure at the jet pump are
rarely given in manufacturers' data sheets, they must be computed from
the flow rates that are usually given versus motive pressures.The pressure
at the discharge of the jet pump, P?, must be back computed fiom known
condtions at the end of the 1ine.The pressure at the suction inlet to the
jet pump, P2, must be computed from the known conditions at the surface
of the liquid in the suction tank.
The average sectional velocities can be computed fiom the given flow rates
and internal dimensions of the connected piping.We have to make certain
assumptions fiom manufacturers' recommendations. These recommendations are to minimize the number of elbows, especially in the dscharge
piping, and to include a strainer and check valve in the suction piping.
306
Computations
Involving
Pumps for
Liquids
A P P E N D I X
AIII
We will write the ~ e r n o u l hequation

~
with irreversibilities using subscripts
(U) for upstream and (d) for downstream so as to fix our thnlung.The loss
coefficients will be identified by the summation sign.The terms that do
not apply to the specific case will then be eliminated and the equation
will be applied to the suction and discharge lines of the jet pump.
The result is
The last term in each equation is the energy per unit mass transformed to
internal energy. These terms are identities.When they are equated, the
arc^^ equation is recovered.
The subscripts on the loss coefficients, K;, refer to the diameter in which
the velocity is measured. The velocity in the kinetic energy term must be
associated with the same diameter. Equation AIII-33 is a general equation
that would allow computations for work done by a jet pump to be
equated to dfferences in static energy, kinetic energy, potential energy and
to conversion of energy caused by piping components.
We will use the equation only for the suction or the discharge piping, so
the work terms drop out of the equations.
For the two piping configurations shown in Figure AIII-10, the differences are one additional elbow in each of the suction and the discharge
lines when in the discharge mode configuration plus the dfferent lengths.
The configurations dffer by the numerical values of the various terms and
by the fact the suction lift becomes dscharge lift.
We will perform computations on a hypothetical case based on the
second sketch of Figure AIII-10, the discharge mode, in order to compare
approximate efficiencies of centrifugal pumps and jet pumps.
Suction piping i n the discharge mode
For the suction piping in the discharge mode, the losses can be identified
with the suction strainer and poppet foot valve, one standard elbow and
the length of suction pipe.The loss coefficient for a typical poppet foot
and Equations
307
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
valve can be obtained fiom craneXLU

as 420fT,where fT is the standardzed
friction factor for nominal piping. For 25 mm (1 inch) nominal piping, the
standardized fiiction factor is 0.023, so the K factor for the check valvestrainer combination is 9.66.That for the standard elbow is 0.023(30) or
0.69.The K factor for the piping is given by fM (L/D).The ~ o o d ~ "fric"
tion factor must be obtained from the fluid velocity, viscosity, pipe diameter
and roughness. It can be looked up in charts or obtained from the
~hurchdl-usagix"relationshp. For the purpose of ths exercise, we wdl use
the standardized friction factor assuming hlly turbulent flow.
For the suction piping between points 0 and 2, there is no pump.The
liquid simply is forced by pressure dfferential from approximately zero
velocity at the liquid surface to the velocity at the inlet to the jet pump.
When the zero level is taken as the liquid level in the sump, the Bernoullil
equation reduces to
A habit of many is to use lulopascals for pressure units and to divide all
the other terms by 1,000 in order to obtain consistency.This leads to the
possibility of forgetting the conversion unit.
Units in AIII-34 are pressure in pascals or psf, density in kilograms per
meter cubed or lb,/ft3, velocity in meters per second or fi/s, acceleration,
g, in meters per second per second or ft/s2, elevation in meters or feet,
pipe linear length in meters or feet and pipe diameter in meters or feet.
The pipe length is not identical to the elevation change. In this case it is
one half a meter 1onger.With S1 units, the dimensional constant, g,, is one
and dimensionless.The K factor for the check valve is dimensionless.
Discharge piping in the discharge mode
The presence of an adhtional elbow in the discharge mode configuration
should be noted.The upstream section is point 3 and the downstream section
is taken through the stream of water as it emerges fiom the piping, section 4.
The loss coefficients are those associated with flow through the length of
pipe, fM (L/D), with the two elbows, 2(30)fT,(once more from cranex"')
and the coefficient of the exit loss from the pipe, KEXIT= 1.
308
~ I o wo f I n d u s t r i a l F l u i d s - ~ h e o r y a n d E q u a t i o n s
Computations
Involving Pumps for
Liquids
A P P E N D I X
A I I I
The velocities at points 3 and 4 are identical, so the two kinetic energy
terms drop out of the equation.The specific ~ernoulli'equation with
Z4 Z3 becomes
Equation AIII-35 can be made more general by replacing the bracketed

" ~ ' ~
term by the sum of the Ki for the discharge piping.The ~ o o d ~friction
factor wdl be replaced by fT for convenience.
General expression for comparative jet pump efficiency
Starting with AIII-32, we can substitute the last of equations AIII-34 and
AIII-35 into the efficiency relationshp to eliminate the unknown pressures,
P2, and P3.
0, U:
m-+-------
P2
U;
(AIII-32)
(AIII-3 6)
incompressible Juids, discharge mode P, = P,, U, = U,, Z, = Z, = 0
and Equations
309
A P P E N D I X
A I I I
Computations
Involving
Pumps for
Liquids
It is worthwhile to analyze the terms of the last equation of the set AIII36.The first term on the right, the ratio of the mass flow rates, suction
flow to driving flow, reflects the fact we are only interested in the fraction
of mass equivalent to the suction flow.The second term, g/gc(Z4),is the
energy per unit mass correspondmg to the height fiom zero to Z4
through which a unit mass will be raised in the discharge pipe; the thrd
term is the mechanical energy necessary to make up for discharge line
irreversibhties; the fourth is the mechanical energy converted to thermal
energy in the suction line.
In case there should be confusion regarding the multiplication of the suction mass flow rate by the discharge irreversibilities,the discharge losses per
unit mass are a function of the actual velocity in the pipe.The friction
factor, diameter and Reynoldsv number must be based upon the actual
velocities in the dscharge and suction piping.Ths is in accord with our
intention to compare useful output with input in both jet pumps and
centrifugal pumps.We are trying to compare apples and apples.
Equation AIII-36 is a model for a worlung equation for computing the
efficiency of a jet pump in the configurations given in Figure 111-10. It
can be modified for other situations.
Example of a real case based on a j e t pump operating i n the discharge mode
This is an example of a deep well ejector.We will use a manufacturer's
published data for a 25-mm jet pump in our example.The size in millimeters is the nominal orifice size corresponding to a nominal one inch
in dameter. For the piping we will use U.S. nominal Sch. 4 0 , l inch pipe
for the suction and dscharge and U.S. nominal Sch. 40, % inch pipe for
the driving fluid.
We wdl use nominal factors, from craneU' for both
...
the Moodyv" factor and the factor at complete turbulence to simplify the
computations.The data is summarized in the table that follows.
G,
Inlet
25 mm, 1 inch
Pipe, sch. 40
nominal diameters
1 m, 3.28 feet
Flow rate
Flow velocity
1 5.83 fps
Pressure
310
F I O W
2.2 m3/h,9.69 gpm
1 3.5 bar(g), 65.4 psia
I Discharqe
1 25 mm, 1 inch
1 6 m, 19.68 feet
1 ell 1 strainer,
1 check valve
2 els, 1 exit
10.52
2.1 m3/h, 9.23 gpm
7.56
4.3 m3/h,
18.92 gpm
3.43 fps
7.03 fps
Computations
Involving Pumps for
Liquids
A P P E N D I X
A I I I
When a jet pump is operating in the hscharge mode shown by the right
hand sketch ofAIII-10, the discharge piping and suction piping coefficients become
The subscript, i, refers to dscharge and the subscript,j, refers to suction.

For clean commercial, one inch, schedule 40 steel pipe, fT is 0.023.The
suction length is one meter (3.28 ft) and the discharge pipe length is 6
meters (19.68 ft), so the total lift is 19.68 + 3.28/2 = 21.32 feet (one half
the suction piping is submerged).
The nominal K factors for the suction and discharge piping are respectively (0.023)3.28/(1.049/12) + 9.66 =10.52 and (0.023)19.68/(1.049/12)
+ 2(30)0.023 + l = 7.56.The number 9.66 is the K factor for the suction
strainer-foot valve combination in the suction pipe. It may come as a surprise to find that the suction piping (called inlet piping) has a greater
mechanical energy loss than the hscharge piping when the discharge
piping is 1onger.This is because the form factor influence of the strainer,
elbow and foot valve is greater than that of the discharge fittings.
With a driving gauge pressure of 3.5 bar (65.4 psia), the flows at the suction (1 inch Sch. 40), at the driving fluid source (3/4 inch Sch. 40) and at
the combined discharge (1 inch Sch. 40), from vendor? data, are 2.10
m3/h (9.23 gpm, 3.43 fps), 2.20 m3/h (9.69 gpm, 5.83 fps) and 4.30 m3/h
(18.92 gpm, 7.03 fps).
The mass flow rate ratio has been replaced by the volumetric flow rate
ratio because the densities are the same in this case.
and Equations
311
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
It can be seen that the efficiency of a centrifugal pump which is normally

around 60% compares with 18.2% in this case.This confirms the statement that the decision to use a jet pump is always made on the basis of
practicahty, not efficiency.
Alll-7: WORKED PROBLEMS

Examples of computed NPSHA i n different units
Figure AIII-11 shows a centrhgal pump operating in a suction M mode.
The elevation change fiom the water surface to a pump center line is 20
feet.There is a foot check valve at the suction pipe Inlet that is 3 feet below
the water surface.The suction fittings include two screwed elbows and a
screwed tee. Plug valves are used to valve in one sump or another.We will
assume the worst case (greatest height, most resistance).The suction pipe
size is 4 inch Sch. 40.We d assume clean commercial steel pipe. NPSHA
is compared to NPSHR to see if the chosen pump d function.
Figure Alll-11. Suction lift
The problem of priming is taken care of by filling the suction piping and
pump casing with liquid, before starting the pump.We wlll consider this
to have been done. We wlll do the first computation twice, at two different temperatures, to show the importance of the temperature variable.
The first computation wdl then be repeated using S1 units.
312
and Equations
Computations
A P P E N D I X
A I I I
DATA, Case 1
Fluid = water
q = 100 U.S. gpm
T = 60F
Vapor pressure = 0.257 psia
p = 1.1 CP
p = 62.37 1bm/ft3
NRe = 50.6qp/dp = 50.6(100)62.37/4.026(1.1)= 71,262(Common
mixed unit formula from craneXLU)
From the friction factor charts or, more accurately from the Church&
usagiA relationships, the fiiction factor for the above ReynoldsVnumber
in 4-inch schedule 40 pipe, is obtained.
f = 0.0242
(From Churchdl-Usagi, relative roughness = 0.00015)
We d use Crane's fully turbulent coefficients for the fittings and valves,
since we are dealing with a hypothetical case.The first thing to do is to list all
the components and to compute the resistance coefficients and to sum them.
Component
Pipe
Els
Tee (branch)
Foot valve
Plug valve
Computations
O.O242(37/4.026)12
=
2 X 30(0.017)
60(0.017)
420(0.017)
18(0.017)
Sum K =
Losses
h, = (Sum K ) u ~ /(2jgC) =
12.155(2.52~)/(2X 32.17)
1.200 ft-lb, /lbrn.
Using the last equation in the set AIII-24, the NSPHA can be computed as,
The numbers show that, for cold water, the first term dominates and the
second term is very small.The NPSHA depends on the piping configuration. If the NPSHR of the pump chosen is greater than the NPSHA, the
pump is unsuitable for that elevation.
and Equations
313
A P P E N D I X
A I I I
Computations
Involving
Pumps for
Liquids
Comparison of the first and fifth equations ofAIII-25 shows them to give
numerically identical results. Only the units change. Hydraulicians would say
we have computed net positive suction head available of 12.146 feet.Ths is
the number we must compare with the stated NPSHR of the pump.
DATA, Case 2
In Case 2, we will change the temperature.This changes the viscosity, the
density, the ReynoldsVnumber and the vapor pressure.
Fluid = water
q = 100 U.S. gpm
T = 120F
Vapor pressure = 1.695 psia
p = 0.53 CP
p = 61.713 lbm/ft3
NRe = 50.6qp/dp = 50.6(100)61.713/4.026(0.53)= 145,345
f = 0.0228
K,,
= fL/D = 0.0228(37/4.026)12 = 2.514
Sum K = 11.944
= 1.184 ft-lbf/lbm.
h, = 12.0(2.52~)/(2~32.17)
HSv= 144(14.7 - 1.695)/61.713 - 1.184 - (g/gJ20 = 30.346 - 1.184 - 20
= 9.162 ft-lbf /lbm.
Note for a temperature change fi-om 60F to 120F we have lost about
3 fi-lbf/lbm (or 3 feet) of available NPSH. If the computation is repeated
for 180F water, a not uncommon water temperature in pharmaceutical
and biotechnology work, it wdl be found the computed NPSHA becomes
negative. No pump would function at that elevation and temperature.The
piping layout would have to be changed.
DATA, Case 3
In Case 3, we will repeat Case 1 with S1 units.
Fluid = water
q = 100 U.S. gpm = 6.309 X 10" m3/s
T = 60F = 288.7"K
Vapor pressure = 0.257 psia = 1.772 kPa
p = 1.1 CP = 1.1~10"Pa-S
p = 62.37 lbm/ft3 = 999.04 kg/m3
D = 4.026/12 ft = 0.3355 ft = 0.10226 m
U = 2.52 fps = 0.768 m/s
~ 71 328
NR, = DUp/p = 0.10226(0.768)999.04/1.1 X I O -=
(The 0.1% discrepancy is due to rounding errors)
314
and Equations
Computations
Involving Pumps for
Liquids
A P P E N D I X A I I I
f = 0.0242
(Same number from ~ h u r c h i l l - ~ s a ~Absolute
i ~ l ~ . roughness and ReynoldsVnumbers are pure numbers)
Computations
Pipe
Els
Tee (branch)
Foot valve
Plug valve
Components
0.0242(11.2776/0.10226)
2 X 30(0.017)
60(0.017)
420(0.017)
18(0.017)
Sum K
(pure numbers are the same as Case 1)
Losses
hL = (Sum K ) u ~ / ~
=
=
=
=
=
=
K
2.669
1.020
1.020
7.140
0.306
12.155
= 12.155(0.768~)/2
= 3.586 J/kg
H, = (101 350 - 1 772.0)/999.04 - 3.586 - 9.805(6.096) = 99.674
- 3.586 - 59.771 = 36.317 J/kg
To obtain meters of manometric head from energy per unit mass units,
divide by g = 9.805 m/s. This gives an equivalent metric manometric
head of 3.704 meters or 12.152 feet.The minor discrepancy with Case 3
is due to rounding errors.
A few typical problems
We wdl use the data associated with Figure AIII-11 and the data of Case 1
to solve a few typical problems.
Problem statement 1
Given the piping configuration of Figure AIII-11, compute the shaft work
to the pump, the fluid work, the pressure drop across the pump and the
suction and discharge pressures. We take the general ~ernoulli'equation of
AIII-19 and simplifjr it based on known facts.The first problem will consider only fluid work. We make the following assumptions:
Assumptions
1. Point 2 is the cross section of the stream as it emerges f?om the pipe.
2. The mechanical energy "losses", h12,are the combined suction and &Scharge piping irreversibhties, 1.2 + 8.8 = 10.0 fi-lbf /lbw
The pressure at point 1 is the atmospheric pressure as is that at point 2.
The fluid densities can be taken as the same at both points (the thermal
energy generated by the pump as conversion of mechanical energy will
change the density a little, but engineeringjudgement wdl be used).
and Equations
315
A P P E N D I X
A I I I
Computations
Involving Pumps for
Liquids
5. The velocity profile correction, alpha, will be approximate one because

the flow is fully turbulent and we are not loolung for ultimate accuracy
6. The fluid velocity at point 1 is essentially zero (the tank is sufficiently
large that the velocity at ths point can be taken as zero).
7. The elevation at point 1 dbe taken as zero (the elevation is not

important per se, the ddference in elevation is important; therefore, the
elevation can be chosen to satisfy the requirements of the problem).
8. The centrihgal pump has an efficiency of 6O%.Ths is a common guess
when the real number is not known.
The ~ernoulli'equation can be simplified, under the above assumptions to,
This is the work transmitted to the shaft of the pump by the driver per
unit of flowing mass.The driver wdl require more energy than t h s due to
its own efficiency rating.
Problem statement 2
Find the dfferential pressure developed by the pump at the operating
point given by the data of Figure AIII-11 and Case 1.
Assumptions
1. The discharge is on the same center line as the inlet.

2. Since only the pump is being considered we can set points 1 and 2 as
being the inlet and dscharge flanges of the pump.The losses are those of
the pump itself, not of the piping.
3. The alpha correction factors for kinetic energy are again equal to one.
316
and Equations
A P P E N D I X
A I I I
The ~ernoulli'equation can be simplified for this case to,
Note the hfference between lunetic energy terms in the above equation
has a value of only 0.539.This fact leads many to ignore the terms altogether and to state that the shaft work is equal to the difference in pressure divided by the fluid density.Ths habit can lead to serious error when
compressible fluid problems are considered.
Problem statement 3
Compute the power supplied to the pump (not the motor) in foot-poundsforce per second and in horsepower using the above-developed data.
There are no assumptions.The power supplied to the pump shaft is equal
to the shaft work per unit mass of flowing fluid times the mass flow rate.
Pump efficiency is not considered.This is not the power supplied to the

driver or to the fluid.The power to the driver is this number divided by
the driver efficiency.The power to the fluid is this number times the
pump efficiency.
and Equations
317
A P P E N D I X AIII
Computations
A l l l - 8 : S U M M A R Y OF A P P E N D I X A l l l
Appendix I11 has concentrated on giving the basic theory and equations
of pumps in order to provide an understanding of their interactions with
fluid systems.This knowledge wdl be expanded in later chapters and
appendices.
The centrifugal pump was treated in greater detail than any other pump,
and the turbine was not treated.The turbine has a simdar theory to the
pump and is less prone to such problems as cavitation.
It was pointed out the essence of any pump is the transfer of energy per
unit mass to a fluid. It was also recommended to think in terms of these
units instead of feet and meters of head units. Not only are energy units
more logically satisfjang, their use can help avoid pidalls for the unwary.
Enough theory was developed so the fairly simple governing equations
could be accepted with some degree of codort.This statement applies
in particular to centrifugal pumps. Positive displacement pumps are best
understood by the limitations imposed upon their application by the
incompressible nature of liquids.
Once the governing equations were developed, the differences between
some of the different types of centrifugal pumps were more easily discussed.The importance of efficiency in determining the characteristics
of some of the pumps was pointed out. Efficiency is not normally a
concern to operations or control personnel, but it is worth considering
because it does affect dramatically pump performance in more ways
than just an economic one.
The commonly used measures of performance were given with a
number of examples.This was done in an attempt to clear away some
of the misconceptions due to the use (and misuse) of units.
The all important ~ e r n o u l hequation
~
was again stressed as being one
of the most important tools for the solution of fluid flow problems including those associated with pumps and turbines.
The net positive suction head problem was dealt with in a fair amount
of detail. Samples were given in different units.This was done for two
reasons.The first reason is operations or control personnel do not normally spend too much time thinlung about NPSH, and yet it is
extremely important when trouble shooting pumping systems.The
second reason is the drgerent units used lead to confusion - they must
be clearly defined and understood.
318
and Equations
Equations Of Compressible
Flow, Derivations
And Applications
AIV-1: PURPOSE
- PROVIDING CHAPTER
I V EQUATION DETAILS
This appendix covers in much greater detail some of the equations of

compressible flow given in Chapter IV. Relationshps are described in
detail for the following:
the use of thermodynamic variables, in particular the enthalpy;
adiabatic flow with irreversibilities,basic equations. (Equations that do
not involve the concept of the Mach number d be M y developed.
They wiU be applied to flow of fluid in straight conduit with irreversibilities.Ths group of equations is the most important withn the context of
ths book.);
the Peter Paigemv equation (Ths equation is an excellent tool to help
understand flashing flow and the chokmg phenomenon. It wdl be developed and covered in some detail.);
adabatic flow with irreversibhties involving equations using the Mach
number (Equations that involve the concept of the Mach number d be
developed.They wdl be applied to flow of fluid in straight pipe with
irreversibllities.T h s development is simply for completeness to give
information to those who are interested and who may need it in their
work. Most computations performed in the context of ths book do not
require the Mach number.);
the concept of choked flow.Ths is critical to an understanding of compressible flow.The use of the R e d h c h - K w o n y equation in order to
obtain greater accuracy than one can obtain fi-om the ideal gas equation
wdl be described in detd.
and Equations
319
A P P E N D I X
A I V
Equations of Compressible
Flow
A I V - 2 : U S I N G T H E R M O D Y N A M I C VARIABLES
I N PARTICULAR, ENTHALPY
Enthalpy is a concept. It is a variable that is composed of the internal

energy plus the pressure-volume energy. In other words, h = u + Pv, or
H = U + PV in total units. It stems from the study of open (flowing) systems. The variable, enthalpy, appears fiequently in flow formulae.
Like other thermodynamic variables, enthalpy is not an absolute number,
since it is obtained by integration. Enthalpy is measured as the difference
between two quantities whose absolute values are unknown. Therefore,
one of the quantities is fixed in an arbitrary reference state.The arbitrariness of the state makes no difference to the result, so long as we all agree
on the reference state.
The reference state for a pure (single component) substance is chosen for
convenience. It differs with different substances. For instance, the reference
quantity for the internal energy of pure water is usually set at zero for
liquid water at 0.01 degrees Celsius and 0.6113 kPa (the triple point is the
reference state) for the steam tables, but this choice is not the only possibility.
Since the enthalpy concept is derived from the internal energy, its reference point for water is also O.OlC and 0.61 13 kPa. Its reference value is
not zero, however, since it has the value of the product Pv for the liquid.
Caveat
- Reference states for thermodynamic properties

When comparing enthalpies or any other thermodynamic property from
dfferent sources, it is important to check the reference state. Otherwise,
the numbers can be misleading.
The choice of different reference states for different substances does not
present a problem for pure substances. For mixtures, it is more convenient
to have a common reference state mainly for purposes of automatic computation. Prausnitz,Anderson and Grens use the ideal gas state at 300K as
the reference state for the enthalpy of mixtures.
Liquid enthalpies are found by integration from this ideal gas state to the
liquid state.Vapor enthalpies are found by integration from the same reference state to the temperature of interest using the ideal gas specific heats
and then using a departure function fiom the hypothetical ideal gas state
to the real state.
Vapor enthalpies are important in fluid flow computation; liquid
enthalpies are less important.
320
Equations o f Compressible Flow
A P P E N D I X
A I V - 3 : A D I A B A T I C A N D I R R E V E R S I B L E FLOW I N U N I F O R M C O N D U I T S
BASIC EQUATIONS
A I V
Bernoulli equation including 'lost work'

The ~ e r n o u l hequation
~
(including the lost work term, Equation IV-22)
can be transformed by multiplying all terms by density squared.We can
constrain it by assuming horizontal conduit and no compressor or turbine
in the piping.This eliminates the potential energy term and the work
term. With most changes in elevation in an industrial plant and with most
gas or vapor densities, t h s is acceptable (but the assumption must be
checked).We w d use the assumption in order to reduce the complexity
of the mathematics.
Mass velocity (flux)

Then we introduce the concept of mass velocity (mass flow per unit area
or mass flux). Even with compressible fluids, the mass velocity along a
constant cross section conduit is a constant in steady-state flow as can be
seen by examining the units (lb,/ft2-S).
The following development only applies to constant cross section conduits.
(AIV-2)
C P U
and Equations
321
A P P E N D I X
A I V
Equations of Compressible Flow
Transformed Bernoulli equation

The above relationsbp is useful when deahng with compressible fluids, as it
simplifies integration. At steady-state flow, the mass flux is constant
throughout the length of a constant section conduit, even when the velocity
changes, because it is the mass flow rate hvided by the cross sectional area.
T b s should be borne in mind for use with rectangular or other shaped
ducts.We wdl concentrate on conduits of circular section.Therefore,4rH in
equation AIV-1 wd1 be replaced by the hameter, D. The velocity, U, can be
e h n a t e d fiom the transformed ~ e r n o u l l equation,
i~
as follows
In Equation AIV-3, since the mass velocity, G, is constant at steady state in

a constant section conduit, even with compressible fluids, it can be taken
outside an integral.The friction factor can be assumed constant over small
steps. An average value can be used by estimating the initial and final value
of the step and dividing by two.The above approach reduces the number
of variables from 6 (p, P, U, fM,L, D) to three (p, P, L). In case it is felt too
many assumptions are being made, it is possible to check computations
against actual tests or published results
Partially integrated, transformed Bernoulli equation

The partially integrated form of the transformed Bernoulh equation is given
below.The first term is left under the integral until it is decided what h r ther constraints to impose. In particular, it must be decided wbch equationof-state will give the optimal accuracyThe relationshp of density to pressure over a path must be known in order to complete the integration. For
now, we will use ths equation to analyze flow in conduits quahtatively.
The first term is derived from the driving differential pressure energy. The
second term is derived from the kinetic energy term in the governing
Bernoulli equation. The third term contains the irreversibilities.All terms
are related at the same end points, sections 1 and 2, of the integration. The
terms, although derived from specific energy, no longer have units of
energy per unit mass.
322
Flow
A P P E N D I X
A I V
Perhaps a more common form of the equation is obtained by multiplying

and dividing the second term by two. Note the change in signs and the
change in subscripts on the densities.
Equations AIV-4 and AIV-5 are equivalent. Their use requires the choice
of an equation-of-state to relate density, pressure, temperature and composition and the choice of a sufficiently short length of conduit, or sufficiently turbulent flow, so an average friction factor may be estimated.
These two equations are useful for computer simulations.
A I V - 4 : THE PETER PAlGE EQUATION, CHOKED FLOW

Peter Paigemv transformed Equation AN-5 to solve for the length of conduit necessary to balance a given change in the other two terms. He
assumed, if the length were short enough, or if the pressure hfference were
small enough, the average density could be used in the first term above to
s i m p Q the integration.The equation is general. Constraints such as the
ahabatic one arise only when estimating densities and the friction factor.
(m-6)
1
The driving differential pressure establishing flow is essentially the pressure

Merence in the first term on the right of the last equation. The values of
the densities in the second term on the right are strictly related to P andT
at the corresponding sections designated by the subscripts on P, L, and p.
The average density and the average friction factor are also averages
between the two corresponding sections. Note the pressure hfference is
the upstream less the downstream pressure - what is normally called
differential pressure.
and Equations
323
A P P E N D I X
A I V
Equation AIV-6 can be used to estimate the length necessary to effect a

given change in pressure and density. It can also be used to estimate a
chohng pressure. It is not only useful for gases and vapors, but it can be
used for mixed phase fluids as well. In the latter case, the densities and
friction factors must be those of the mixture.
Choked flow occurs with gases, vapors and mixed flows. It is the phenomenon whereby for fixed reservoir conditions, in spite of a lower downstream pressure, the mass flow rate in a conduit remains unchanged. The
pressure, temperature and density within the exit from the conduit also
remain fixed, in spite of the lower downstream pressure.
Note the expression "choked flow" refers to the uncoupling of the
dependence of the mass flow rate in a segment and the downstream pressure, nothing more.The pressure at the most downstream section of a segment of conduit is higher than, and is independent of, that in the next
downstream segment.The choked mass flow rate will be a function of all
the usual variables, but the pressure in the most downstream section of a
segment is substituted for the upstream pressure of the next segment or
the sink pressure. Exit losses do not play a role in the flow computations
for the segment in which chohng occurs.
Derivation of choked flow equations based on observation

In an experiment performed on flashing water in a pipe, Benjarnin and
Mdle?
used 1 psi as a constant pressure drop.They performed measurements on the lengths necessary to create this pressure drop in a pipe.
The length necessary to create one psi of pressure drop upstream in the
pipe was 20.6 feet. Downstream, close to the exit, it was only 0.57 feet.
These facts, when observed often enough with different fluids, allow us to
reason that, for fixed pressure drops proceeding down the pipe, the average
fluid density falls, causing the first term on the right of the last of
Equations AIV-6 to become smaller.The second term on the right
becomes smaller also, but at a slower rate as we proceed downstream.The
length necessary to complete the length balance decreases towards zero.
The driving energy is shared between the lunetic energy and the conversion
(the "losses") to internal energy and heat energy flow.The previous statement can be verlfied by considering the first equation of the set AIV-1.
Since the distance over which the conversion takes place tends to zero and
the location of the zero point is at the end of pipe, we can reason all of
324
Equations o f Compressible
Flow
A P P E N D I X
A I V
the driving energy is converted to kmetic energy and none is left for conversion to internal energy (losses in mechanical energy) at the end of the
pipe. Note the first law of thermodynamics equates energy change in a
body to the sum of work energy flows and heat energy flows.
Joule clearly established the equivalence of mechanical energy and thermal
energy. Mechanical energy is considered to be the only form of "useful"
energy, and mechanical energy converted to thermal energy is considered
to be a "loss".
The ~ e r n o u l lequation
i~
with irreversibhties in differential form, Equation
AIV-1, can be transformed to the following equation:
At steady-state mass flow, compressible fluid still accelerates as it flows

down a conduit. If we arbitrarily fix the differential pressure increment at
some low value, say 1 psi or 1 kPa. We can examine the above equation
for each segment of the conduit that has the same pressure drop.
Starting at the upstream section, it wlll be noticed the densities decrease
progressively as we proceed down the conduit.The friction factor
increases a little, but not much.
Equation AIV-7 can be integrated over a one unit pressure drop using
average values of density, friction factor and velocity over the length of the
increment in the last term to give equation AIV-8.
The first term in Equation AIV-8 has units of pressure, as do all the other
groups of terms.The difference between the squares of the outlet and inlet
velocities of a segment is much smaller than the square of the average
velocity in the length over which the pressure drop takes place.To maintain the validity of the equation as we proceed down the conduit, the difference in length has to be reduced because the average velocity squared is
increasing at a greater rate than the difference of the squares.This conclusion is confirmed by observations (Benjamin-Mille?).
and Equations
325
A P P E N D I X
A I V
Flow
The dfference in length, associated with a fixed pressure drop necessary to

effect a change in lunetic energy and to overcome permanent conversion
of mechanical energy, decreases as we move down the conduit in the
drection of flow. If the upstream pressure is sufficiently high, or the
downstream pressure sufficiently low, and the conduit sufficiently long, the
length increment can be reduced to approximately zero.The equation wdl
stdl balance. Physically, we cannot go beyond this state - we cannot create
negative length.
The choking phenomenon, which is the condition we are describing, will
always take place at the end of the conduit.There is less resistance when
the fluid suddenly leaves the conduit.The acceleration is greatest at ths
point and the conversion of mechanical to thermal energy is greatest for a
minimal length increment.
Criterion for choked flow

A mathematical switch for the incremental method of establishing choked pressure
Mathematically, we could keep on integrating equation AIV-7 with negative differences in lengths. Physically, we cannot. This fact gives a nice
switch for computer and calculator computations.We can approximate the
point at which flow chokes by setting dL,in equation AIV-7 equal to zero.
At the end of a constant cross section line, the change in pressure with
change in velocity is equal to a constant, minus G/gc when flow is
choked. T h s conclusion stems from the observation that the length increment for a fixed pressure drop tends to zero at the end of the line or at a
sudden expansion when flow is choked.
If we can write a general equation for the derivative in AIV-9 and equate
it to the term on the right, we can solve for the cholung (so-called critical) conditions.Tlus "critical pressure" is the pressure immediately inside
326
and Equations
Equations
o f Compressible Flow
A I V
A P P E N D I X
the conduit exit when the length increment to cause an infinitely small
pressure drop tends to zero. It is not the thermodynamic "critical pressure". Specific volume and velocity at this point are also of interest.
Note we now have a criterion for choked flow derived from thermodynamic considerations and observations of experimental facts.The criterion
relates changes in two variables, pressure and velocity, to a constant mass
flux. T h s criterion applies to all fluids under all conditions, not only the
adiabatic one.
First law w i t h constraints (adiabatic flow, horizontal conduit, no work)
Starting with the first law energy balance for open (flow) systems (true for
all fluids, reversible or not) and assuming horizontal conduit (no elevation
change), an ahabatic process, and no fluid work, we have equations AIV-10.
The units of each term are thermal (Btu/lb,) in the customary U.S. system.
(AN-10)
du2
6q+6w=dh+-
2gcJ
The dh term represents differential enthalpy.We can integrate the second

differential equation of AIV-l0 to obtain the following:
Adiabatic 'stagnation enthalpy'

The sum of the enthalpy and the lunetic energy is the same anywhere in
the conduit under the adiabatic constraint. The lunetic energy terms and
the enthalpy terms have identical units when the J factor is included.The
sum of the flowing enthalpy and the lunetic energy is given the name
"stagnation enthalpy". Stagnation enthalpy is the enthalpy that would be
measured or estimated if the fluid were brought to rest adiabatically.At
steady state in adiabatic flow, it is constant.
and Equations
327
A P P E N D I X
A I V
Flow
If there is no heat transfer across the conduit walls, the stagnation enthalpy
is constant.We can use this fact to derive terminal conhtions from known
starting conditions. In particular, if the system is close to adiabatic (good
insulation), the reservoir (source) enthalpy at the point where the velocity
is close to zero can be used.
To simplify the development, we will use the ideal gas approximation with
a constant heat capacity ratio, y. In the set AIV-12, we define enthalpy, h,
in Btu/lbm in terms of the ratio of specific heats, y, pressure, P, and specific
volume, v. Since the term containing the ratios of specific heats is constant, we call it B to reduce the number of symbols.We then substitute for
enthalpy in the stagnation enthalpy equation. Lastly, we convert to mass
velocity and a differential equation is derived.
The differential equation gives the change in pressure with specific
volume of an ideal gas under conditions of constant stagnation enthalpy.
The ratio of specific heats, y, of an ideal gas is reasonably constant over
normal adiabatic changes in conduits. For instance, the specific heat of
steam (as an ideal gas) varies by 0.8% from 368F to 317F a difference in
temperature of 51EThis corresponds to an adiabatic pressure drop from
170 psia to 39 psia or 131 psi in the example that w d follow. For the sake
of simpliGing the differentiation, we can consider the ratio a constant. In
what follows, the subscript, S,refers to the source or the stagnation condition; the subscript, o, refers to the reference temperature (which drops out
of the equation).
From Equation AIV-12, the flowing temperature,T, wdl be lower than the
stagnation temperature,T,, because of kinetic energy's contribution to total
energy.The gas constant, R', has units of Btu/lb,R
in the customary
U.S. version of the ideal gas equation and is so used here.The dimensional
conversion factor,J ft-lbf /Btu, has been included and can be made equal
to one, dimensionless,if S1 units are used.
Also, if the gas constant must be replaced by the difference between the
specific heats at constant pressure and at constant volume, it must be converted from mechanical to thermal units, because the specific heats are
normally in thermal units in the customary U.S. system.
328
~ I o wo f ~ n d u s t r i a l ~ I u i d s - ~ h e o r y a n d ~ q u a t i o n s
Flow
A P P E N D I X
A I V
A I V - 5 : CHOKED FLOW U S I N G THE IDEAL GAS EQUATION

Constrained first law including stagnation enthalpy (adiabatic model)
For an ideal gas flowing adiabatically, the constrained first law equation
including the stagnation enthalpy is as shown as one of the last three
equations of the set AIV-12.
Initially at least, we are interested in average velocity, U, at a section of
conduit rather than specific volume, v, so we wdl effect a change in variables by using the mass flux relationship.The assumption of steady state
makes the mass flux constant in a straight conduit.
The equation set AIV-12 makes use of "thermal" energy terms (BtuAb,)
and mechanical energy units, ft-lbf Ab,, hence the conversion factor,J.
The last equation of the set AIV-13 is a first law equation for ideal gases.
It simply states that the negative of the pressure change with velocity anywhere along a conduit with constant section is inversely proportional to
velocity squared at a fixed mass flux. It uses mechanical energy units.
Note the first equation of the set AIV-12 imposes an adiabatic constraint
on the development.
and Equations
329
A P P E N D I X
A I V
Flow
The fourt:h line ofAIV-13 is derived by malung substitutions for dv an(

in the last line of AIV- 12.
Computing ideal gas choked quantities (adiabatic conduit)

Equation AIV-9 can be used to make AIV-13 specific to conditions of an
ideal gas at the end of the line when flow is choked.
( A N -14)
l=
gccpJT - B - --
330
*2
y -l- y+l
y-l
2 2(y-1)
~ I o wo f ~ n d u s t r i a i~ i u i d s - ~ h e o r ya n d ~ q u a t i o n s
A P P E N D I X
A I V
The mechanical equivalent of heat,J, is only required when using customary

U.S. units.The expression for U* in the equation set ofAIV-14 gives the
velocity at the conduit exit under choked conditions in an adabatic conduit
for an ideal gas.The expression for v* gives the spec& volume under the
same condtions for a given mass flux as a hnction of the stagnation temperature.We now need the critical (choked) pressures. Note the mass flux at
the choked condtion is imposed by the upstream fluid properties and conduit sizes and configuration.The asterisk/star (*) as a superscript has no real
sigmficance other than correspondence to the choked condtion.
Choked pressures, ideal gas (adiabatic conduit)

Choked variables are the values of those variables within the exit from a
conduit when the mass flow rate becomes constant in spite of lower
downstream pressure. Choked variables are independent of the downstream variables.
The use of the ideal gas equation-of-state leads to relatively simple expressions for variables at the choked condition. Be warned that, although simpler to use than the equations for real gases, these expressions may be grossly
in error for gases that are not ideal and for regimes that are not adabatic.
We start with the third from last of the equations of the group AIV-12
and immediately transform it and differentiate it.
and Equations
331
A P P E N D I X
A I V
Flow
Equations AIV-14 gave an expression for the choked specific volume, v*.
This can be substituted into the last equation of the set AIV-15.
A check of the units of Equation AIV-16 will show the choked pressure
has units of pounds-force per foot squared.
Equation AIV-16 gives the choked flow pressure of an ideal gas as a function of the choked mass flux (the mass flux at the same point that choked
pressure is established) and the stagnation enthalpy (specific heat times the
absolute temperature at the source).The mass flux is usually obtained from
the required design conditions, for instance, when it is necessary to relieve
an amount of fluid under emergency conhtions. If the design conhtions
do not impose the mass flow, it must be estimated, usually by iteration,
after careful logical analysis of the situation.We will give the criteria
shortly.The stagnation enthalpy can be estimated from the quiescent conditions in the source vessel - where the velocity is close to zero.
These equations are useful for quick estimates, for an understandmg of gas
behavior, or as the starting point in an iteration. If the estimated choked
flow pressure for the assumed or estimated mass flux is less than the sink
pressure, the flow is not choked under these assumed or estimated con&tions and the downstream pressure for further flow computation is the
sink pressure. If it is greater, flow is choked and the downstream pressure
(within the conduit exit) to be used is the estimated choked pressure. If
the flow is choked, the conduit exit loss coefficient is not used in conduit
pressure drop computations. If the flow is not choked the exit loss coefficient must be included.
Proof t h a t equation AIV-16 for choked pressure corresponds to the traditional
equation for sonic velocity i n an ideal gas undergoing an adiabatic process
The velocity of sound in an ideal gas is given traditionally as the square
root of the product of the ratio of specific heats, the dimensional constant,
the gas constant and the absolute temperature.This is true, but how do we
obtain the temperature that would permit us to estimate the sonic velocity
if the flow is choked? The above development gives one method.
Starting from the concept of stagnation enthalpy in AIV-11, we can
develop the concept of stagnation temperature.This is the temperature an
332
A P P E N D I X
A I V
ideal gas would reach if brought to rest adiabatically from its flowing conhtion. It is a temperature that can be approached in carefully designed
experiments using a stationary thermal sensor in the flowing stream.The
stagnation temperature is also equal to the reservoir temperature in a carefully insulated system.
For an ideal gas, the enthalpy is a function only of temperature.We can
substitute the functional relationship into equation AIV-11 as follows:
The mechanical equivalent of heat, J, equals 778.16 ft-lbf /Btu and has
been included to remind the users of customary U.S. units to convert
thermal energy units (Btu) to mechanical energy units (ft-lbf ) .This constant is one in S1 units.The numerical subscripts refer to (1) a fixed
upstream section and (2) a fixed downstream section.The duplication of
the term containing the average velocity squared is permitted because the
stagnation temperature of an ideal gas is a constant in an adiabatic process,
just as is the stagnation enthalpy.
The combination of terms involving velocity squared can be verified to
have units of temperature.The kmetic energy corrections are assumed
equal to one.
Traditional form of equation for speed of sound
The temperature form of the equation for acoustic velocity in terms of
the Mach number is
u2= g c y ~ ~ ~ & a
and Equations
333
A P P E N D I X AIV
Flow
Equation AIV-18 uses a gas "constant" with mechanical energy units

(ft-lbf /lbm). We can substitute the velocity, U, from AIV-18 into AIV-17.
(AIV- 19)
Y=c.=
c,,
CP
cp-R/J
The equation developed in AIV-19 gives the relationship for an ideal gas
between the temperatures at two sections in a straight length of adiabatic
conduit and the Mach numbers at the same two sections. For an ideal gas,
we can arbitrarily pick point one as the source where the velocity, therefore the Mach number, is zero and we can pick point two as the choked
flow point where the Mach number is one,
Equation AIV-20 shows the relationship between the temperature in the

vessel, the source, and the temperature at the choke point, the "critical"
temperature, is a very simple function of the ratio of specific heats for an
ideal gas undergoing an adiabatic expansion.
334
and Equations
Flow
A P P E N D I X
A I V
Link to other thermodynamic variables

Equation AIV-20 shows, if we can assume ideal gas behavior and an adiabatic process in an ideal, choked nozzle, we can estimate the temperature at
the choke point.This may sound like a lot of ifs, but the second part, at
least, is not so far fetched.An adabatic process in an ideal, choked nozzle is
relatively easy to approximate. Ideal gas behavior is more dfficult to ensure.
Given the temperature at the choke point and the temperature at the source,
we can relate other thermodynamic variables such as pressure and density to
the temperature ratio during the same adabatic process. For a detailed development of what follows, see any text on hndamental thermodynamics.
Essentially, any fluid undergoing an ideal, adabatic process, as ideahzed by
an insulated flow nozzle at the entrance to a pipe (or by a PSV), has its
flowing properties related by T2/T, = [p2/p,]y-l = [ P ~ / P ~ ] ()l?.
~ - AS
' we
already have fixed conditions in the reservoir, and we can estimate the
choked temperature by AIV-20, we can estimate the choked density and
pressure fiom the above property relationship.The choked average velocity,
U, at the section is given by the fifth equation of the set AIV-14.The
choked mass flux, G* lbm/s-ft2,may be obtained fiom the product U*p*.
The stagnation temperature,Ts,from equation AIV-20, is substituted into
equation AIV-16 and the mechanical energy gas constant, R , equals R'J.
and Equations
335
A P P E N D I X
A I V
Flow
The last equation in the group AIV-21 is the conventional one for sonic
velocity in an ideal gas. It was derived by equating "critical" (sonic) pressure estimated from the ideal gas "law" with choked pressure estimated
fiom the choked flow criterion, minus dP/dU equals G*/g,.
The "mechanical equivalent of heat", J, is a conversion factor required
only when the "constant" R is in "thermal" energy units of Btu/lbKR.
The equations for properties of choked flow were derived without the use
of the conventional equation for sonic velocity.The fact they reduce to
the conventional equation for sonic velocity of an ideal gas helps establish
their validity.They are also more useful than the canned equation, since
they permit computations based on known upstream conditions. The conventional sonic velocity equation for an ideal gas simply tells you what the
velocity is fi you know the corresponding temperature.
Conclusions to be drawn for the adiabatic, choked flow condition

The following concepts apply quantitatively to ideal gases and qualitatively
to real gases.
1. From AIV-14, the choked velocity of an ideal gas flowing adabatically is
a function only of condtions in the reservoir (source) - fundamentally of
the enthalpy, practically of the temperature.
2. From AIV-17, the temperature at the choke point is also a function only
of the condtions in the reservoir, but, in order to estimate temperature,
the velocity must first be estimated. If there are multiple choke points,
due to enlargements, the velocity and the temperature wdl be the same at
each choke point fiom the velocity-stagnation temperature relationshp
of AIV- 17.
3. The variables P* (AIV-16) and v* (AIV-14) are also functions of the

reservoir condtions, but, in addtion, the mass flux first must be estab1ished.The choked pressure, P*, is proportional to G and the choked mass
volume, v*, is inversely proportional to G.
4. If flow is choked across an inlet restriction to the conduit (a PSV; for
example), the mass flux at the throat is a function only of the reservoir
condtions and the effective throat area.The flow might choke again at the
exit fiom the conduit, but mass flow rate remains the same - it is governed only by the first choke point. In ths latter case, a new value of G is
estimated by dividmg the constant mass flow rate by the area of the cross
section.Ths value of G is used to estimate an upstream pressure to make
336
Flow o f Industrial F l u i d s - ~ h e o r y and ~ q u a t i o n s
A P P E N D I X
AIV
sure ths pressure is below either the critical pressure for a balanced PSV
or below the manufacturer's recommended limit for an unbalanced one.
5. To have multiple choke points, it is necessary to have adhtional conduits

with increased cross sections.The smaller value of G in each addtional
segment means the spec& volume at the choke point of each segment
d be greater (AIV-14) and that the pressure will be less (AIV-16).
Although G is Merent at each section and at the throat of any inlet
nozzle, the mass flow rate is constant throughout all sections, if there are
no branches.
6. If flow is choked only at a downstream point in a constant dameter pipe,
then G is a function of the choked pressure, the reservoir pressure, the
conduit L/D and relative roughness and the variable fluid density and
viscosity. The exit loss is not considered.
7. If flow is not choked at all, then G is a fbnction of the fluid, the reservoir
conhtions, the sink conhtions, the conduit cross sectional area and L/D,
the relative roughness and the exit loss.
8. The enthalpy at the source is a function only of the temperature at the
source for an ideal gas.
The importance of the chokmg phenomenon should be evident from the
above discussion.
Example AIV-1: Multiple choke points with ideal gases, adiabatic models
Figure AIV-1 shows a pressure relief valve with three horizontal lengths of
different diameter conduit. If the conduit is insulated, or if it is short
enough, the adiabatic model can be used. If the gas is hotter than the surroundings, the adiabatic model gives conservative results for safely sizing
relief headers. If the gas is colder, the results are not conservative and heat
transfer must be taken into consideration.
The example in Figure AIV-1 was chosen to demonstrate the cholung
phenomenon and for purposes of comparison with a real gas model. It is
hypothetical.The use of an ideal gas model permits an analytic solution to
the problem as opposed to an iterative one.
Note that, with an ideal gas flowing adiabatically with no heat energy
flow to the environment, the temperatures at the choke points will be
identical by Equation AIV-20. This temperature w111 be less than the
source temperature for gamma greater than one. If the deceleration in the
sink is adiabatic and isentropic, the source temperature wdl be recovered.
and Equations
337
A P P E N D I X
A I V
rt
Figure AIV-1. Multiple choke points
Example AIV-1: (Continued from previous page)

With multiple choke points, one choke point (in ths case it must be the
PSV) d set the mass flow rate.This rate will be constant throughout the
conduit.The mass flux d be Merent in each conduit segment, Li.When
sizing or checkmg relief header piping, one first sizes the relief valve to
relieve the estimated mass flow rate based on the worst case for the vessel
being protected - heat input, power loss, blocked discharge, etc.Then the
smallest, commercially available relief valve that will pass t h s mass flow rate
is ch0sen.A new mass flow is estimated based on the actual orifice size and
the accumulated pressure in the vessel.The allowable accumulated pressure
depends upon the vessel code.The allowable accumulated pressure will be a
multiple of the design pressure such as 1.1 or 1.25, dependmg on the safety
design basis.The new, greater, mass flow rate is the one that must be used
to verlfjr that the backpressure at the valve is not greater than that allowable
for the type of valve, balanced or unbalanced.
In a situation such as the one depicted in Figure AIV-1, we cannot just

assume flow wdl be choked only at the relief valve.The backpressure at
the relief valve must be checked, because, if it is too great, it will prevent
the valve from passing the required flow rate.We will show that, in the
example, flow can choke at four locations: the PSV nozzle, the two
expander inlets and the exit to the sink. Because flow chokes at the inlet
to an expansion, the effective length of the downstream conduit is greater
for pressure drop computations than the actual length by the equivalent
length of the inlet swage.
The governing mass flux is GI.This is the mass flow rate required of the
PSV divided by the effective area of the PSV nozzle. The other mass fluxes
are related to GI by the ratios of the internal nozzle to internal conduit
areas or the squares of the internal nozzle to internal pipe diameters.
338
F I O W
o f ~ n d u s t r i a~ ~I u i d s - ~ h e o r y a n d ~ q u a t i o n s
Flow
A P P E N D I X
A I V
From Equation AIV-16 we can estimate the choked pressure at each

expansion using the ideal gas assumption,
Since the ratios of the mass fluxes are smaller sequentially, subsequent
downstream choked pressures are lower. The choked pressure at the
nozzle, P*I,is the first choked pressure. Subscripts 1 to 3 give the subsequent ones.
If the estimated backpressure at Segment 1 exceeds the maximum allowable backpressure for the particular PSV, then the pipe diameter must be
increased for one or more segments of conduit. It is assumed the noise
associated with "sonic" velocity can be tolerated for the length of time
relief flow occurs.
and Equations
339
A P P E N D I X
A I V

The temperature of the fluid at the choke point can be estimated from
Equation AIV-20. It is the same for all choke points for an ideal gas. It is
proportional to the stagnation temperature and to a function of gamma,
the ratio of specific heats.
The temperature also can be estimated for the nozzle by the following
relationship. Both formulae give the same results for ideal gases. Note the
presence of the exponent signifies an adiabatic process. See most elementary thermodynamic texts for a derivation.
The above equation is applicable only to the first choke point, the nozzle.
Subsequent choke points have lower pressures but the same temperature.
The densities at the choke points will dffer because the choked pressures
change.
Caveat - Integration with constraints
When performing an integration under an isothermal or an adiabatic constraint, it is well to remember the difference in the mechanisms. In the
first, energy is exchanged to maintain isothermality. In the second, the
energy remains with the mass flowing so the equilibrium properties are
different due to a higher temperature in the adiabatic case.
Data for the example (ideal gas, pressure relief)
PSV. 1/2-inch ideal nozzle.
Pipe: Schedule 40, = 3/4 inch, 4 1 ~= 1 inch, +3 = 1 1/2 inch, L1 =
8 feet, L2 = l foot, L3 = 8 feet of equivalent length (lengths wdl hffer
due to entrance effects).
Source fluid: Ideal air at 30 atmospheres, 1000 R, ratio of spec& heats is 1.4.
The mass flow through a choked nozzle is a function of the reservoir conditions and of the throat area. This may be obtained by multiplying the
mass flux by the effective area of the nozzle throat. The mass flux is found
from the development following AIV-20.
340
~ I o wo f ~ n d u s t r i a l F l u i d s - ~ h e o r y a n d ~ q u a t i o n s
A P P E N D I X
A I V

Choked pressure
The choked pressure for the first choke point, the PSV, the asterisk pressure, can be estimated directly for ideal gases &-omthe equation that fol1ows.The derivation d not be given because it is lengthy and it involves
considerations that go beyond the intended scope of this book. Equation
AIV-16, which was derived fully, could also have been used.The hnctional
relationshp is among the reservoir (source) pressure, the choked pressure
and the ratio of specific heats only for ideal gases (ideal air, in this case).
P,*
= 0.528(30) = 15.840 atm
The nozzle pressure is 15.84 atmospheres, 232.8 psia or 33,530 psfa.
Choked temperature for an ideal gas

The temperature at the choke point can be estimated fiom the following
equation.The derivation of this equation wdl not be given for the same
reasons cited above. This temperature is a function of the critical pressure
ratio and the reservoir (source, subscript S) temperature. It can also be estimated by using equation AIV-17 that was derived fully.
Choked densities
In order to estimate the densities at the choke points, we must first obtain
the density of the gas in the reservoir from the ideal gas relationship.
and Equations
341
A P P E N D I X
A I V

The density at the first choke point is obtained fiom the following equation.
Choked velocity (speed)

The speed of sound in an ideal gas is given by the well-known "sonic"
velocity equation as a function of the temperature and the ratio of specific
heats only. It can also be found from equation AIV-21 by redsing RT*
equals P*/p*.
Choked mass flux

The mass flux, GI*,at the first choked flow point can be found from its
definition. It must be remembered this mass flux is not constant
throughout the nozzle and mass flux is not mass flow (which is constant at
steady state).
G,* = U,*p,* = 1414.5(0.755)= 1067.8 lb,
/S
- ft2
The internal diameters of the Sch. 40 pipes are 0.824, 1.049 and 1.610
inches. The mass fluxes downstream are 393.1,242.5 and 103.0 lb,/s-fi2
respectively from the equation on page AIV-18.
342
F I O W o f ~ n d u s t r i a~ ~~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
A P P E N D I X
A I V
Example A I M : (continued from previous page)

Choked mass flow
The mass flow rate is constant throughout the different diameters of conduit.The mass flux equals the mass flow divided by the area.
Choked pressures
The pressures at the downstream sections when flow is choked at each
point may be estimated from AIV-16.The specific heat of air at constant
pressure, c,,, is taken as 0.24 Btu/lb,F
and gamma, the ratio of specific
heats, is taken as 1.4.
(2 [p;:
~ ; = q *-
~ 5 . 8 4- = 3.6 atm (7625.3 p$a)
e*=p2-
(1.049J
=3.6 - =1.53 atm (3237.1psf)
1.610
*(:l
and Equations
343
A P P E N D I X A I V
Flow
Reynolds numbers, for pressure drops i n conduit

The ReynoldsVnumber for the ideal gas case wdl depend on the mass
flux at the choke points and will vary with varying temperature upstream.
In what follows, we ignore that variation for the sake of brevity.
To estimate the Reynolds number at the choke point, the viscosity at
833.33 R is required. Sutherland's general formula for viscosity is given by
cranex"' for several gases.
For air the reference temperature can be taken as 459.67 R.The corresponding reference viscosity is 0.016 centipoise. Crane gives the value of
the constant, C, as 120. Using these values in the above formula gives a
viscosity at 833.33 R of 0.0252 centipoise.The temperature and the viscosity at the choke points are constant for an ideal gas.
Crane gives a formula for the Reynolds number (which can be checked
by the formulae given in this book) as follows:
Using the two preceding formulae, we obtain the following Reynolds

numbers corresponding to the conduit sections at the choke points:
NReYI
= 1.597E6
NRez= 1.254E6
NRe, = 0.817E6
From the churchill-Wag@ equation, Equation AII-10, the correspondmg
Moody friction factors are:
fM, = 0.0241
fM2= 0.0225
fM3= 0.0206
344
and Equations
Flow
A P P E N D I X
A I V
Limitations of the Peter Paige equation: Two unknowns, second equation needed
The Peter Paigemv equation, Equation AIV-6, can be used for gases as well
as for mixed flow. It is a force balance, second law (of thermodynamics).The
equation is not limited to ideal gases. It can be rearranged as a functional
relationshp equal to zero.
Starting at the end of the conduit, fixed length increments can be chosen.
The downstream choked conditions are known, if it exists.We have two
unknowns in each length increment, the upstream pressure and the
upstream density, P1 and pl.The mass flux can be fixed from the relief
requirements.
If we start at the beginning of the conduit, the upstream conditions must
be known and the unknowns are subscripted with a 2.
A second independent equation is needed in order to solve for the

unknowns.The last equation of the set AIV-12 gives an independent first
law relationship among the same variables for adiabatic flow. It can be
integrated and the specific volumes can be transformed to densities.This
equation is limited to ideal gases.
The functional relationships, E(Pl ,pl) and F(Pl,pl) are derived respectively
from a first law energy balance and a second law force balance.The first
law equation is independent of considerations of losses (irreversibilities),
the second law equation includes them.
Newton-Raphson iteration, an algorithm from Balzhiser, Samuels and Eliassen
A very useful little book no longer in print, Chemical Engineering
Thermodynamics, by Balzhiser, Samuels and Eliassen ( B S & E ) gave
~ ~ an
algorithm for the solution of simultaneous equations of two variables
when the partial derivatives are available.
and Equations
345
A P P E N D I X
A I V
Flow
As always when there are two unknowns, two independent equations

must be solved, simultaneously.The two equations can be arranged as two
functional relationships, each equal to zero.The functions can be represented as two curves in three dimensional space as in Figure AIV-2.
Figure AIV-2. Simultaneous equations -functions of two variables
It is to be noted that setting the value of the function to zero only applies
to a single point. Each pair of values of the independent variables w111
result in an associated value of the function. The expression "each pair"
refers to the values at the same cross section of conduit. The single point
at which the two curves cross the horizontal plane gives the simultaneous
values of the two independent variables.The two partial derivatives of
each of the two functions (four partial derivatives) give approximations to
the change in &stance along their respective axes.The approximations
become more and more exact as the curves approach the zero transition.
The BS&E algorithm gives the zero transition point.
346
F I O W o f ~ n d u s t r i a~ ~~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
A P P E N D I X
A I V
The two vector triangles in Figure AIV-2 represent the partial derivatives
along their respective planes.
Det = xq - yp
A P = W', P )p - F P , P )x
Det
The two functions equal zero when the values of P and p are correct
simultaneously.The new values of P and p for the next iteration are P +
AP and p + Ap. The algorithm is remarkably robust for well-behaved,
smooth functions.We wdl use it for the solution of the example problem.
Note the partial derivatives will have different values dependmg on
whether the upstream or downstream values are held constant.
The term, Det, means "determinant".
Partial derivatives of t h e equations E(PIp) and F(PIp)

The terms X,y, p and q of the above set are required in order to implement the algorithm.Those terms used for an ideal gas when the upstream
quantities change are given below. It should be apparent, and it wdl
become more apparent, solutions to complex problems involving multicomponent mixtures only can be obtained on a computer.
and Equations
347
A P P E N D I X
A I V
Flow
Tabulated results of simulation of Example AIV-1 (ideal gas, adiabatic model)

Following are the tabulated results of the simulation of Example AIV-1. The
choked properties are those of the downstream sections: P2(1,l), P2(2,l ) ,
P2(3,1),T2(1,1),T2(2,1),T2(3,1),
U2(1,1),U2(2,1),U2(3,1), and
R H 0 2 (1,l) RH02(2,1) RHO2(3,l).The numbering system used is that
the first letter or letter series sigmfies the variable; the first number indicates
upstream, 1, or downstream, 2; the first number in brackets gives the conduit
segment and the second one gives the increment withn the segment.
As has already been pointed out, for an ideal gas, the velocities and temperatures, U2(i,l) and T2(i,l) at each choke point are identical to one
another, the pressures, the densities and the temperatures f d as we proceed downstream. The velocities increase.
One phenomenon not clear in the results of the simulation is that the sink
temperature, on the basis of an ideal gas undergoing an adiabatic process, is
the same as the source temperature, in spite of the lower, intermediate
temperature at the choke points.
Each conduit segment was divided arbitrarily into twenty lengths
(counting upstream) for the simulation. A comparison wdl show the
upstream pressures of each segment are lower than the corresponding
upstream choked pressures of the next, upstream, segment; that is
P1 (1,20)<PI (32336<33530), P1 (2,20)<P2(1, l ) (11857<12258),
P1 (3,2O),<P2(2,1)(7198<7625) and Patm<P2(3,1) (21l6<3237).
The simulation shows how the variables behave throughout the header.
348
Simulation of Examde AIV-l

Pipe segment counting upstream from choke point
2
J->
1
3
Upstream pressures, PSFA
921
871
907
928
876
914
10
11
12
13
14
15
16
17
18
19
20
934
880
920
939
884
924
943
887
946
890
932
949
892
935
952
895
938
954
897
940
956
899
942
958
901
945
960
902
946
961
904
948
962
906
950
964
907
951
965
908
953
966
910
954
Downstream oressures. PSFA
Downstream temperatures, R
T2(1,J)
T2(2,J)
T2(3,J)
834
834
834
894
857
882
911
865
897
Downstream densities, LBmlFT3
929
APPENDIX A I V
A I V - 6 : ADIABATIC CHOKED FLOW; P, v, T RELATIONSHIPS U S I N G THE

REDLICH-KWONG EQUATION
Example AIV-2: Multiple choke points with a real gas, incremental solution
We wdl give an example based on a real gas.The gas chosen is air and the
choice was made in order to permit a comparison of two dfferent
methods of estimation of a similar problem.The piping in example,AIV-2,
has the same configuration and diameters as that given in Figure AIV-1
and the condtions in the reservoir (source) and the mass flow rate
through the PSV are identical.The major difference is air, although almost
an ideal gas, is treated as consisting of three components: nitrogen, oxygen
and argon, in order to test the R-K EOS for mixtures against the previous
results for air as an ideal gas.The trace amounts of water vapor and carbon
dioxide have been lumped into nitrogen.
This section is concerned with developing relationships for choked flow
using the ~edlich-~wong-" equation-of-state. We are no longer dealing
with an ideal gas.The original R-K EOS is one of the simpler ones but it
gives surprising accuracy even for mixtures. The derivations necessary for
its implementation are far from simple. These derivations will be avoided
by using the Peter PaigeXavequation and the fact the incremental length
to balance this equation tends to zero at the end of a straight segment
when the flow are choked.The derivations of the analytical choked flow
equations, however, are given in the paper, Choked Flow by R. Mulley
(Available from ISA).
The Peter Paige equation was given in Section AIV-4 as Equation AIV-6.
The use of the ~ernoulli'equation, from which the Peter Paige equation
was derived, to create a mathematical switch was discussed in the same
section.The mathematical switch gives the signal for incipient choked
flow and avoids the necessity of some onerous mathematical derivations.
The isenthalpic flow relationship wdl be used to estimate changes from
the starting point, the source. It is,
The enthalpy of an adabatic process with zero flow is symbolized by h,

or hs (or hST),the total or stagnation enthalpy (mass or molar basis). It is
equal to the enthalpy at the source in the case of an amabatic network.
350
F I O W o f ~ n d u s t r i a ~ ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Flow
A P P E N D I X
A I V

It also is equal to the enthalpy at a downstream section plus the kinetic
energy at the same section.The lunetic energy can be expressed either as a
function of the average velocity across a section or as a function of the
mass flux and specific volume. The customary U.S. units of enthalpy are
Btu/lb, and J is 778.16 ft-lbf/Btu.The S1 units resolve to N-m/kg or
J/kg and the coefficients J and g, equal one and are dimensionless.The
coefficient,J, is also equal to one and it is dimensionless when the units of
enthalpy are foot-pounds-force per pound-mass.
The enthalpy at the source is estimated from the departure function in the
next section.
Enthalpy departure function

The enthalpy departure function estimates the change in enthalpy from
the ideal to the real gas state at the system temperature. It gives a means of
using an ideal gas relationship to estimate the enthalpy change from the
ideal gas state at a base temperature to the system temperature and then it
adds the departure function to obtain the real enthalpy.The departure
function may be established in terms of any of the equations-of-state.
The reader is reminded once more that enthalpy is not absolute. Its value
d depend not only on the units chosen, but also on the temperature
chosen as a base.
The original R-K equation (Reid, Prausnitz, and P o l i n F , Fourth Edition)
gives the enthalpy of non-ideal gas mixtures as follows in equation AIV-23.
i=J,
cR,,d T + (departurefunction)
]/ g - mole
The first four terms on the right of the second equation of the set AIV-23
give the ideal gas mixture molar enthalpy in J/g-mole.The last three
terms give the departure of the enthalpy from the ideal gas state of the
and Equations
351
A P P E N D I X
A I V

mixture.The four coefficients given in Reid, Prausnitz and Poling are the
coefficients of the isobaric heat capacity for ideal gases.The units of the
heat capacity coefficients give enthalpy in J/g-mole.The units of all the
l e pressure
.
is in bars
terms of the departure function are b a r - ~ m ~ / ~ - m oSo,
and molar volume is in ~ m ~ / ~ - m oMultiplying
le.
the departure function
by the ratio of the two appropriate "universal" molar gas constants converts the departure function to the same units as the ideal gas function,
joules/g-mole. The units of the conversion factor (the ratio) are (J/gmole-K) / (bar-~m~/~-mol-K)
or jouledbar-cm3.
The author apologizes for harping on about terminology, but he has found
that the habit of otherwise intelhgent people of playing fast and loose with
logic and common sense serves as an impediment to intellectual progress.
Note in the above book by Reid, Prausnitz and Poling, the term,^'.^, in
the denominator of the R-K equation is included in the coefficient, a.
This leads to errors if one is not fully aware of the fact. In this book we
wdl separate the system temperature from the term, a.The coefficient of
attraction, a, of the original R-K equation is 0.427 48
T ~ . ~and
/ Pthe
~
coefficient of repulsion, b, is 0.08664 R Tc/Pc. Neither coefficient is a
function of the system temperature,T. The reference temperature,To, is
usually 298.2OK (25OC, the usual temperature in laboratories).The units of
R underbar,Tc and Pc are bar-~rn~/~-mole-K,
K and bars. Be aware of
the sudden switch to a molar basis (g-moles). O n a mass basis (grams),
h =h/(MW).
The coefficients of the terms, a and byare pure numbers, but the other components of the terms possess units. In the set AIV-23, a and b are found
usingT in degrees K and P in bars with R underbar in bar-~m~/~-mole-K.
Equation AIV-23 gives a way of estimating a real enthalpy at a fixed temperature starting from ideal conditions at a reference temperature. Note
enthalpy, in this equation, is a function of P,T, v and composition.
The numerical value of the enthalpy depends upon the units chosen and
on the reference state. In this case, the reference state is an ideal gas at
some stated temperature, subscripted with a zero. The reference temperature can be any convenient temperature. Enthalpy is not "absolute".
The units chosen for the accompanying simulations were customary U.S.
units.
352
~ I o wo f I n d u s t r i a l ~ I u i d s - ~ h e o r y a n d ~ q u a t i o n s
A P P E N D I X
A I V
All terms in equation AIV-23 are on a per unit g-mole basis.The coefficients of the ideal gas heat capacity terms,Ai, Bi, Ci and Di are those of
individual components of a mixture. They can be found for a large
number of components in the Property Data Bank of Reid, Prausnitz and
Poling.The moles referred to are g-moles for the rnixture.The R-K coefficients, a and b, refer to the mixture as do y and R . The reader is
reminded to check each term for consistency of units.
Equivalent terms for equation AIV-23
O n the last line ofAIV-23, the first two terms, Py - ET, the difference
between the real molar PI energy and the ideal energy, can be replaced by
their equivalent difference as expressed in AIV-24.
The use of one or the other expression depends upon whether or not one
wishes pressure to appear explicitly. Equation AIV-24 can be proved by
substituting the R e d l i c h - K w o n y equation into Pv-RT and simplifing
the result. By substituting equation AIV-24 into the set AIV-23, the
enthalpy of the source for an adiabatic system can be found as a function of
T and y.The velocity or the mass flux, in this case, is equal to zero.
A more accurate heat capacity correlation ( i f needed)
A more accurate equation than that used in AIV-23 for the ideal gas heat
capacity between the strict limits of 200K and 600K (the normal industrial temperature range) is,
The coefficients for this equation can be found in Prausnitz,Anderson and

Grens.The units of specific heat in this case are J/g-mole-K.
To convert enthalpy in J/g-mole to Btu/lb-mole, multiply h by I = 0.43021.
Equation AIV-25 is given only for completeness. It was not used in the
simulations associated with the problems given in this book.
and Equations
353
A P P E N D I X
A I V
Flow

Minimum number of simultaneous equations required
For a fluid of known or fixed composition, a thermodynamic relationshp
is one among three variables: P,T and 1.From Wlard Gibbs' phase rule, we
only need to specie any two of the thermodynamic variables of a single
phase, single or fixed component system, before all other variables become
fixed.We need two independent equations for simultaneous solution. For
uniform section segments, the first independent equation comes from the
first law of thermodynamics for flowing systems.T h s law defines the
energy relationships of a fluid.The second independent equation comes
fiom the second law, whch describes the direction in whch first law
changes tend naturally. In our case, we d use the a&abatic constraint to
simplifjr the mathematics and to conform more closely to industrial reality.
Energy relationship among variables, E(T,v), for adiabatic flow (first law)
The first of the two equations needed is derived by substituting equation
AIV-23, the R-K description of enthalpy, into AIV-22, the isenthalpic
(adabatic) flow relationship. It is transformed to a functional relationship
for later use in a Newton-Raphson iteration, E(T,v).
All terms have to be on a consistent basis.We choose a mass basis (lb, or

kg).The reader would be wise to check all the conversions before using the
algorithrn.The energy units will be Btu's or joules. We must check equation
AIV-26 to see all the algebraic summations are BtuAb, or joules/kg.
The last term of equation AIV-26 has units of BtuAb, in the customary
U.S. system.The conversion factor,J, is necessary in the denominator to
change mechanical energy units to thermal units. In the S1 system, two
factors, g and J, in the denominator can be made equal to one, &mensionless, to give joules/kg.
We have already established the units of the terms in the large parentheses
as joules per gram-mole due to the source of the data used. In the customary US. system, we have to convert both the energy and the mass
terms. Using the factor, I, in the numerator converts joules to British
thermal units. The factor (MW) (mixture molecular mass, g/g-mole) in
the denominator converts molar units to mass units (grams).The factor L
in the denominator converts grams to lb,. In the S1 system, K equal to
1,000 g/kg is necessary in the numerator. Other terms that must be consistent are: the molar gas constant in joules/(g-mole-K); the temperature is
in degrees Kelvin; and the molar volume is in ~ m ~ / ~ - m o l e .
354
and Equations
A P P E N D I X
A I V

E(T,v) = 0 =
bRT
a
+[v-b
(,+b)T0.5
-
3a - ]1:nI v + b 8.3144
2 b ~ O . ~ v 83.144
~~v~
+-
28,l
U.S. I = 0.43021 Btu/J, gc = 32.17 ft-lbm /lbf -s2,
Equation AIV-26 is a functional relationship for adiabatic flow down a

conduit. It relates v andT at a section to mass flux, G, and to the total
enthalpy (stagnation enthalpy) or the enthalpy at the source.The equation
is written in terms of enthalpy, not irreversibhties. Because it does not
account for irreversibhties, it cannot be used dn-ectly to solve for losses in
mechanically usefd energy. The accuracy is as good as the accuracy of the
~ e d l i c h - ~ w o nequation.
r~~'
Note both molar and mass units are used in Equation AIV-26.The overall
molecular mass (mixture) has to be included to give consistent units.The
equation gives the values of variables at a section.
The advantage of using the mass flux, G, is it simphfies the derivations of
compressible flow. A molar basis is helpfd, in the general case, when
dealing with mixtures. Most of the available data wrll be on ths basis.
The fhctional relationship ofAIV-26 is a good example of the &fficulties
engineers face in the real world where data comes from different sources.
It is well to analyze each combination of algebraically addtive terms.The
J)Ghas
term G ~ v ~ / ( ~
, units of Btu/lbm in customary US. units and
joules/kg in S1 units (after manipulation, with both gc and J equal to unity
with no units).
and Equations
355
A P P E N D I X
A I V
Flow

Force-balance relationship among variables, F(T,v), (second law)
The other equation used in the solution of the simultaneous equations
will be the Peter Paige"" equation.The Peter Paige equation is a second
law equation. An analytical method for finding choked pressure for real
gases using the Redlich-Kwong"N"1 equation is given in the paper Choked
Flow by R. Mulley (Available from ISA).
Viscosity computations to establish loss of mechanical energy
Viscosity is a naturally arising resistance that liquid and gaseous fluids offer
to flow. It is one of the most difficult to grasp concepts in science and
engineering probably due to the way in which it is taught.We have
already pointed out the logical inconsistency of the concept of "fluid friction". How can the behavior of molecules that never touch be described
by the macroscopic phenomenon of friction? Nevertheless, eminent scientists and engineers still try to find ways of using the term "fluid friction"
to describe energy relationships in flow phenomena.
The author prefers to consider the phenomenon to be due to the compression of gases and liquids during flow and to the subsequent local
increase in temperature.When the temperature differential dissipates by
energy transfer, the result is a loss in available mechanical energy. As was
clearly established by Sadi carnot'', transfer of energy under a temperature
gradient without recovery of mechanical energy is an irreversible loss of
mechanical energy.
Viscosity is nothmg more than an experimentally established ratio between
the shear stress (force per unit area in the hrection of the force) applied to a
layer of fluid and the velocity grahent (change in velocity per unit &stance
perpenhcular to the force).This ratio, which is called viscosity, is measured
practically by various means and, therefore, receives various names and units,
all equally confusing.Viscosity is a dynamic quantity as opposed to a static
one such as pressure, temperature or specific volume. It is sometimes called a
non-equhbrium quantity, but the term "equhbrium" is applied to dynamic
equihbrium as well as to static equhbrium so viscosity could be described
as an equhbrium property under certain circumstances.
Viscosity values are a means to an end in terms of irreversibilities associated with fluid flow.They must be correlated with such quantities as the
ReynoldsVnumber in order to estimated the mechanical energy losses that
occur with flow. In addition, viscosities of pure components are normally
356
F l o w o f ~ n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
Flow
A P P E N D I X
A I V

avadable as functions of temperature, but mixture viscosities are not available. Mixture viscosities must be estimated by the use of various algorithms.
Algorithm t h a t uses Chapman-Enskog theory

The Chapman-Enskog theory gives a means of estimating viscosities for
low pressure (undefined) multicomponent gas mixtures. The book by Reid,
of Gases and Liquids, gives various
Prausnitz and ~ h e n v o o d ~ ~ " ' ,Properties
~he
means of estimating mixture viscosities.We wdl use a method from this
book (RPS). It is based on estimating the pure component viscosities in
micropoise and then using a weighting method to estimate the mixture
viscosity.When compared to experimental data, the error is usually less
than 2% if one omits data on mixtures of wildly varying molecular masses.
The pure component viscosity in micropoise is first estimated using one
of the three qualifications: (l), non polar, (2), polar, non hydrogen bonding
and (3),polar, hydrogen bonding, then the weighting equation is applied:
j=1
The term, phi, is found by Wilke's approximation:
Phi is found for each binary pair by interchanging the subscripts.The

theory requires the gas be of low density and the molecules be so far apart
only binary interactions take place.
For a binary system with subscripts 1 and 2, the first equation becomes,
v, =
y1v1
Y1+ ~2@12
Y272
Y2
+~ 1 @ 2 1
The above equation can be manipulated to show the value of phi with
identical subscripts (ll , 22,33, etc.) is one.
There are more complicated and more accurate approximations for viscosity given in the subsequent revision of the above book, but it was felt
the above set of equations was adequate for the task at hand.
and Equations
357
A P P E N D I X
A I V
Flow
Analysis of simulation of example AIV-2

The first six sets of row information in the table giving the results of the
simulation were established for the purpose of checlung the input data to the
Fortran simulation of example AN-2. The remaining sets of row information
were intended to allow comparison with the tabulation of the results of the
simulation of Example AIV-1, the ideal gas simulation.The choked pressures
of example AIV-2 should be compared with the first column of downstream
pressures of AIV-1. The choked temperatures, volumes and average velocities
should be compared with the correspondmg columns.
Analysis of simulation o f example AIV-2
The real case choked pressures vary horn minus 1.1% to plus 1.6% of the
ideal gas pressures.The choked temperatures are about 4% higher and are
not constant.The choked velocities are about 4% higher and are not constant.These dfferences are to be expected and are in the right dnection.
The results of the simulation in example AIV-2 should be verified to see
the upstream pressures of each segment are lower than the choked pressures upstream.The sink pressure should also be lower than the choked
pressure of segment 3.The sink pressure is 2,117 psfa and the choked
pressure of segment 3 is 3,288 psfa. The pressure upstream in segment 3 is
3,337 psfa and the choked pressure in segment 2 is 7,711 psfa.The pressure upstream in segment 2 is 7,731 psfa and the choked pressure in segment 1 is 12,280.The pressure upstream in segment 1 is 25,000 psfa and
the choked pressure in the PSV nozzle is about 33,530 psfa.This latter
pressure was not computed for the real case. It was taken hom the ideal
nozzle calculation.The difference in pressure was considered sufficient to
justify this shortcut.
358
F l o w o f ~ n d u s t r i a l ~ I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Flow
A P P E N D I X
A I V
Results of simulation of example AIV-2, customary U.S. units

WMOL
CPA
.7440E+01
.6713E+01
1CPB
-.l
800E-02
-.4883E-06
PCM
I TCM
1287E-05
.3809E-08
-.4362E-09
0.000E+01
SUMYA
SUMYB
SUMYC
1813E-05
SUMYD
-A65 1E-09
ENTHO
PSRCE
TSRCE
1 000E+04
VSRCE
VCM
SUMZJ
TSlNK
998.3
CHOKED P
.l
228E+05
.7711E+04
.3288E+04
CHOKED T
.8636E+03
.8686E+03
.8698E+03
CHOKED V
.3759E+01
.6018E+01
.l413E+O2
P, SEC 1
.2500E+05
P, SEG 2
.773 1E+04
P, SEG 3
.3337E+04
V, SEC 1
.l827E+00
V, SEG 2
.l
059E+00
VUP SEC 3
A50 1E-01
Leqend
WMOL
R
PCM
TCM
VCM
SUMZJ
ARKM
BRKM
CG(i)
CP A,B,C,D
SUMYi
ENTHO
PSlNK
TSlNK
VSlNK
,P, SEG
CHOKED U
.1478E+04
SEC 1
mixture molar mass, Ibm/lbm-mole

mixture gas constant, ft-lb,/I
-, ...
. bR
critical pressure, mixture, psfa
critical temperature, mixture, R
critical volume, mixture, ft3/lbm
sums of critical compressibility times mole fraction
Redlich-Kwonq coefficient, a, mixture
Redlich-Kwonq coefficient, b, mixture
mass velocity in throat or segment, Ibm/s-ft2
component heat capacity coefficients, cp in ft-lbf/lbm-R
mole fraction weighted sum of heat capacity coefficients
source or stagnation enthalpy, ft-lbf/Ibm
sink pressure, psfa
sink temperature, R
specific volume gas in sink,ft3/lbm
upstream pressure in segment 1,2 or 3, psfa
and Equations
359
A P P E N D I X
A I V
Flow
Analysis of simulation of example AIV-2

The simulation ofAIV-2 was much more complicated than AIV-1 in that it
considered three components of a non-ideal (real) mixture.The enthalpy
computations and the fiiction factor (fi-om viscosity and ReynoldsVnumber)
computations were more detded. Also, the choked conditions were estimated fiom a constraint on the Peter Paigemv equation, not fiom a theoretical equation based on an ideal gas.The use of the R-K EOS makes the
second simulation more usehl generally, especially for gas mixtures.
The reader is invited to make his own analysis and to compare the results
with experimental data.
Example AIV-3: Real gas, analytic solution
The author finally resolved the problem of finding an analytic solution to
the choked flow phenomenon.The solution is described in the paper
Choked Flow by Raymond Mulley and the Fortran code is found in the
paper Simulations by the same author (Available from ISA).The tabulated
results are given below. If the tabulations of examples AIV-2 and AIV-3 are
compared, it wdl be seen they are very close.
The analytic solution is inherently more accurate. It d be demonstrated
in Appendix AV for a very complex problem.
360
Flow
A P P E N D I X
Real Gas Example AIV-3

Analytic Method
Results o f simulation o f analytic method, real gas, configuration o f Fiqure AIV-1
WMOL
R
PCM
. TCM
. VCM
. SUMZJ
76950.0
28.96
53.39
237.0
.04757
.2896
ARKM
1.377E+04
BRKM
1.436E-02
C W
1068.0
CG(11
393.2
CG(2)
242.6
103.0
CP A
.7440E+01
.6713E+O1
.4969E+0
CP B
-.l
800E-02
-.4883E-06
-.4261 E-05
CP C
.l
975E-05
.l287E-05
.3809E-08
CP D
-.4784E-09
-.4362E-09
0.000E+01
SUMYA
.7265 E+O 1
SUMYB
-.l 406E-02
SUMYC
1813E-05
SUMYD
-.465 1E-09
ENTHO
.8752E+05
PSRCE
.6350E+05
TSRCE
.l
000E+04
VSRCE
.8472E+00
PSlNK
21 17.0
TSlNK
998.3
VSlNK
25.18
CHOKED P
1.24E+04
7.66E+03
3.25E+03
CHOKED T
8.37E+02
8.37E+02
8.37E+02
CHOKED V
3.60E+00
5.83E+00
1.37E+01
CHOKED U
1.42E+03
1.42E+03
.l 414E+04
Pup SEC 1
2.5 1E+004
Pup SEC 2
7.68E+003
Pup SEC 3
3.30E+003
,V, SEC 1
1.83E-01
Vup SEC 2
1.06E-01
,V
, SEC 3
4.50E-02
l PCM
SUMZJ
ARKM
BRKM
CG(i)
CP A,B,C,D
SUMYi
ENTHO
PSINK
TSINK
N2
02
AR
SEGl
SEG2
SEC3
mixture molar mass, Ib,/lb,-mole

mixture gas constant,ft-lbf/lb,-R
critical pressure, mixture, psfa
critical temperature, mixture, R
critical volume, mixture, ft3/lb,
sums of critical compressibility times mole fraction
Redlich-Kwong coefficient, a, mixture
Redlich-Kwongcoefficient, b, mixture
mass velocity in throat or segment, Ib,/s-ft2
component heat capacity coefficients, cp in ft-lbf/Ibm-R
mole fraction weighted sum of heat capacity coefficients
source or stagnation enthalpy, ft-lb, Ilb,
sink pressure, psfa
sink temperature, R
specific volume gas in sink, ft3/lb,
upstream pressure in segment 1,2 or 3, psfa
and Equations
A I V
A P P E N D I X
A I V
Flow
AIV-7: SUMMARY O F APPENDIX AIV

Appendix AIV has given a good deal of theoretical and practical information on the basic equations of compressible flow.
The very useful and very practical Peter PaigeXXIv
equations were developed and the Benjamin and ~illetcu"'experiment on the choking effects
of flashing water was referenced. It was shown the Peter Paige equation is
equivalent to, and more general than, the commonly used equation to
describe sonic flow.
An extensive set of equations involving the R e d l i c h - K w o n y equation-of-state was developed.An algorithm for computer implementation
of the R-K equation was given.
Finally, a real gas example using the analytical method was given. A complete development of the real gas equations wdl be found in the paper
Choked Flow by R . Mulley and in the paper Simulations by the same
author (Available hom ISA).
A Glossary of terms used in this appendix wdl be found at the end of this
book.
362
F I O W
Compressible Fluid Flow Complex Systems

AV-1: SCOPE
- E S T I M A T I N G COMPLICATED
PRESSURE DROPS AND FLOWS
Appendu:AV lays the groundwork for detailed estimations of pressure drops

and flows of some very complicated compressible fluid systems. By "complicated compressible fluid systems" is meant vent headers with flows of multicomponent mixtures fiom multiple, Uerent sources.An example of a computer simulation of an adabatic vapor relief problem wdl be given.The scope
of the a p p e n b wdl be confined to the problems of the industrial plant.
The system chosen for demonstration is a complex safety relief vent
header. Some of the main concepts have already been discussed in Chapter
IV and in A p p e n k AIV. We wdl put these concepts into the context of
the network shown in Figure AV-1. This example is taken from a real case.
The piping network existed and had to be analyzed.The example was
chosen to demonstrate real-life problems. The flow rates were estimated for
the worst-case scenario.The input data was changed a little to emphasize
certain phenomena and to protect the original client.The results fiom a
run of the simulation w d be given with suitable commentary.
AV-2: DESCRIBING THE P I P I N G NETWORK

Figure AV-1 describes the piping network.The plant layout and modifications to it generally fix the overall piping configuration of a vent header
system.The design configuration generally is changed due to addtions and
deletions, as vessels are put to different use or added or eliminated over the
course of time. Figure AV-1 is based on a real chemical plant. In this case, it
was necessary to confirm or infirm the safe operation of the system.
and Equations
363
A P P E N D I X
Compressible
Fluid Flow
Complex Systems
It was stated in ChapterV that a computer simulation requires the physical

network be described by a simple coding procedure.Any combination of
possible flows can then be imposed on the system to study the effects on
the relief devices of changed piping configurations, changed diameters, or
different combinations of flows. Many design hsciplines benefit from the
output data from such a program.The program can give information of
interest to process engineers, piping design personnel and control systems
engineers (and many others) such as velocities, densities, temperatures and
pressures at every fitting.This information can be used in piping support
studies and in pipe stress studies, as well as for a general analysis of how
the system wiU perform.
A sketch is meant for a human, not a computer.The sketch is a record of
what exists and what the basis of the simulation was. It is used during the
walk-down of the system for checlung purposes. It must be up-to-date,
even if it exists only in marked-up form.
The isometric sketch was described simply in ChapterV Figure AV-1 is a
more complex version of such a sketch. All relief devices in a system
should be shown on an isometric sketch.
The original model allowed the possibility of three conduit feeders (a
connection to a single segment with a cross) to each segment.We have
not changed this option even though the author has not observed its use.
Most junctions resolve themselves to simple pipe tees (two possible feeders
to any particular segment).
Once the piping network has been described by an isometric sketch, it
can be coded into a computer. If output data is not needed for individual
bends, the fittings of constant diameter pipe can be included as an effective length in the data for the segment.The author prefers not to use
effective lengths, but to describe each segment individually so that data for
each fitting and segment may be generated.
364
and Equations
Figure AV-1. Isometric sketch of complex vent header
A P P E N D I X
and Equations
Complex Systems
A P P E N D I X
Compressible F l u i d Flow
Complex Systems
Examination of the isometric sketch of Figure AV-1 reveals four vertical

12-inch diameter segments: numbers 25,38,49 and 72.These segments
are treated as any other segment but with zero 1ength.The segments are
essentially modeled as small vessels, with the K factor for the inlet fitting
attributed to the upstream segment and only that for outlet fitting attributed to the nominal segrnent.The K factors are transformed to delta factors internally using a method that adapts the work of Benedict et al-.
The column with the heading Transition type identifies each of the eight
types of transitions of the paper Changes to Adiabatic Ent Header Simulation.
These transitions are listed after the piping configuration table.
I
SEC #
Pipe Config uration for Vent Header

SEGMENTS: 79
Length, l I.D.
ft
inches
SOURCES: 11
Roughness,
ft
Nominal
terminal K
0.238
0.238
0.0
0.1 96
0.21 9
0.1 96
0.1 96
366
FIOW
1
]
Transition
type
2
6
4
3
7
2
2
FEEDERS
J
K
20
21
19
18
14
25
26
o f ~ n d u s t r i a l ~ ~ u i d s - ~ h e o ra yn d ~ q u a t i o n s
0
0
22
23
24
0
0
1
I
0
0
0
0
0
0
0
End
Connection
L4x4
B-T4x4x4
R-T4x2x4
E4X12
Contract.
L8x8
L8x8
Compressible F l u i d Flow
II
I
SEG #
SEGMENTS:79
Complex Systems
A P P E N D I X
Pipe Configuration for Vent Header

SOURCES: 11
I.D.
inches
ft
Nominal
terminal K
28
4.2
7.981
1.50E-04
0.0
27
38
39
40
41
42
43
44
45
46
47
48
49
1.O
0.6
2.5
15.5
4.0
18.0
10.5
2.8
2.0
2.5
1.O
1.2
12
3.068
3.068
7.981
2.067
2.067
3.068
3.068
3.068
3.068
3.068
12
1.50E-04
1.50E-04
1.50E-04
1.50E-04
1.50E-04
1SOE-04
1SOE-04
1.50E-04
1.50E-04
1.50E-04
1.50E-04
1.50E-04
0.469
0.252
0.252
0.266
0.266
0.266
0.252
0.252
0.252
0.0
0.879
0.459
7
2
6
5
2
8
2
2
6
4
3
7
37
38
39
28
8
42
43
44
45
7
46
48
0
0
0
40
0
0
0
0
0
0
47
0
Roughness,
Transition
type
FEEDERS
J
Length,
ft
L
Connection
0
0
0
0
0
0
0
0
0
0
0
0
Contract.
L3~3
B-T~x~x~
O-T~X~X~
L2~2
L2~3
L3~3
L3~3
B-T3~3x3
R-T~x~x~
E3~12
Contract.
and Equations
A P P E N D I X
I
SEC #
SEGMENTS79
Length,
ft
I.D.
inches
Complex Systems
Pipe Configuration for Vent Header

SOURCES: 11
FEEDERS
Roughness,
ft
End
Connection
0.879
Contract.
L3x3
B-TBx3x8
L8x8
L8x8
L8x8
The transition types are:

1. the exit to atmospheric pressure or to the lowest sink pressure;
2. bends;
3. pipe enlargements other than the exit to the smk;
4. tees with one inlet contiguous with the run;
5. tees with two opposed inlets at 90 degrees to the run;

6. tees with one branched inlet;
7. contractions in the exits fi-om knockout pots;
8. combination bends and enlargements that were combined and called
transition 8 in the program.
AV-3: D E S C R I B I N G THE FLOW R E G I M E

Recommendations made in ChapterV will be repeated. It is not necessary
to size all vent headers for simultaneous relief of all safety devices. It is
recommended to divide the system into groups of devices that d relieve
together under similar conditions such as fire exposure (by fire zone),
cooling water failure, power fdure, etc., and then to investigate the groups
for the worst case backpressure at each source.The worst case backpressure can then be compared to the maximum allowable backpressure,
whlch is a hnction of the type of relief device - rupture disk, conventional PSV or balanced bellows PSI? In adhtion, the safety device d
have been chosen based on engineeringjudgment and the maximum
allowable backpressure may be lower than that required by the vessel code.
368
~ I o wo f ~ n d u s t r i a~~I u i d s - ~ h e o r ya n d ~ q u a t i o n s
Compressible
Fluid Flow
Complex Systems
A P P E N D I X V
If it is necessary to provide data for piping support or stress analysis, segments can be described as being terminated by bends and fittings where
the data are needed.This technique is favored by the author and was followed in the example case.
It is necessary to decide on a model of fluid flow: isothermal, adiabatic
vapor or gas, or adiabatic flashng flow.The most general model d
involve multicomponent mixtures. The model chosen for demonstration is
the multicomponent, adiabatic vapor model.
AV-4: COMPONENT I N P U T DATA, ELEVEN SOURCES

I
Properties of Components at Sources

I
I Psink ~ s i aPatm,
l
Psia 1 # comm. I
14.7
Comp.# MW
Pc, Psia
Tc, R
VC,ft3/lb,
Zc
Coeff. B
Coeff. C
3.20E+01
1.l 7E+03
9.23E+02
5.90E-02
2.24E-01 1.23E+02
2.29E-01
4.63E-05
-2.84E-08
1.80E+01 3.20E+03
1.l
7E+03
4.98E-02
2.29E-01 3.23E+02
1.l
OE-02 3.36E-05
-6.36E-09
9.31 E+01 7.70E+02
1.26E+03
4.64E-02
2.47E-01 -8.08E+01 7.08E-01
14.7
IMET HA NO^
l WATER 1
3
Coeff. A
COMPONENTS
ANILINE
Coeff. D
-3.1 6E-04 5.59E-08
VISCOSITY CODE
3
Data m Sources
Source
1
2
3
4
5
6
7
8
9
10
11
Max*
P acc. Rated T acc. Rated P. Rating Rated
permissible
by Code, Psig by Code, F
by Code, Psig
backpressure.
Psig
60
60
30
30
60
18
20
60
60
60
60
60
60
30
60
60
18
60
60
60
60
60
50
50
25
50
50
15
50
50
50
Component flow rates at

sources, Ibmlh
I
MeOh
9
6
6
6
6
0.6
0
0
0
H,O
6000
6000
889
3640.1
6000
975
2973.2
13.3
3008
Analine
40
9
6
8041.9
93.5
9.2
447.8
4.6
0
* The maximum permissible backpressure is arrived at by engineering judgment. It is always equal

to or lower than backpressure permitted by any vessel code. It sometimes is lower due to known
equkment connections that are not rated according to the vessel code.A system is only as good as its
weakest member.
and Equations
369
A P P E N D I X
Compressible
Fluid Flow
Complex Systems
AV-5: PLAN O F ATTACK

This simulation was originally written using customary U.S. units. The
units have not been changed but it is not difficult to do so.The iteration
variables were chosen to be temperature and specific volume. For fixed
composition mixtures and for pure fluids, any two thermodynamic variables may be chosen. Even though composition in the network varies due
to mixing, once mixed the composition remains constant until the next
mixing point.
Each simulation project must have its plan of attack. Our plan is to:
1. establish the mass flow rate (normally the mass flow at accumulated pressure) for each safety device based on the worst-case failure of each vessel;
2. sum the mass flow rates of feeders to each successive downstream segment;
3. convert mass flow rates to mass velocities for each segment;

4. for each segment, convert mass flow to mole flow, mole &-actionsper segment and total mole flow;
5. establish the molar heat capacity of each mixture in each segment by

using the ideal gas heat capacity coefficients,A, B, C and D, for each
component;
6. establish the real gas enthalpies using the departure hnctions fiom the
R-K equation;
7. establish the pseudocritical properties for each mixture.These are, at segment "n":
(Z,,),
= Z,Y,(Z,)~,
= critical
mixture compressibdity
The subscript,m, refers to a property of the mixture, c, refers to a

critical property and, i, to the component.
(vcm), = Eiyi(vC),,fi3/lbm-mol, critical mixture molar volume in
segment n
(M,),
= Eiyi(Mc)i,lbm/lbm-mol,mixture molar mass in segment n
(T,,),
= Ziyi(TC),,
R , critical mixture temperature in segment n
(P,),
= Zi(yi(Pci)),,lbf/fi2 critical mixture pressure in segment n
(R,), = 1545/ (M,),,

segment
mixture gas "constant", lbf -ft/lb,-R
8. establish the constants, a and b, of the Redlich-Kwong""

state for each new pipe segment;
in
equation-of-
9. estimate the enthalpy at each source and sum the feeder enthalpies for
the subsequent enthalpies in each pipe segment;
370
and Equations
Complex Systems
A P P E N D I X
identlfjr the choke points (they may be multiple) based on the analytic
method;
estimate the downstream temperature and specdic volume of the fluid in
each pipe segment fiom pressure and stagnation enthalpy using
Newton-Raphson iteration on simultaneous equations for T and v (Start
fiom sink pressure plus the pressure Merence across the exit or fiom
the choked pressure.);
estimate the upstream pressure, temperature and spec& volume of the
fluid in each pipe segment using the Peter Paigemv equation;
pass the estimated data to the upstream pipe segment or use choked
properties and repeat until a safety device dscharge is reached;
compare the estimated dscharge pressures with the maximum permissible dscharge pressures and make any necessary changes, and;
iterate to a satisfactory solution.
'No one said it was going to be easy! Fortunately, the estimations and the
data manipulation can be programmed to be executed automatically.
AV-6: I R R E V E R S I B I L I T I E S D U E T O F O R M ( A N D M I X I N G ) EFFECTS
In simulations of safety vent headers, permanent changes in mechanical
energy across a fitting or some other piping transition may have to be estimated for incompressible fluids (liquids), compressible fluids (vapors and
gases) and for mixed phase fluids (flashing fluids). Permanent changes in
mechanical energy, in general, fall into two categories: those due to the
wall effect of conduits ("skin fiiction") and those due to an obstruction or
a change in fluid path ("form fiiction"). Form effects are usually more
substantial than skin effects, and yet there is much less formal analysis of
the former than of the latter.
As has been pointed out throughout ths book, there is no such thing as
fluid fiiction. It is a misnomer for an irreversible change in kmetic energy
to thermal energy.This is why the author prefers the term "irreversibilities" to "fiiction" although it is impossible to avoid the term "fiiction"
consistently since it is so ingrained in the literature.
There have been reams of information published on wall effect "fi-iction"
factors. The K factors used to estimate mechanical energy conversions to
thermal energy due to form effects have been treated more summarily.
and Equations
371
A P P E N D I X
Complex Systems
Approach to treatment of 'form friction'

There are three different practical approaches to estimating form effects:
l. the use of conventional K factors;
2. the use of Miller's K factors;
3. the Benedct et al-
approach.
The first two approaches bear the same name, K factor, but the factor is
only identical in the case of incompressible fluids (liquids).Both
approaches ratio permanent changes in mechanical energy across a transition to kinetic energy at some section.
Approach number 1, conventional K factors, is the most widely used and
most widely documented. It is the most inaccurate for compressible fluids,
however, and requires judicially chosen safety factors.We use it for incompressible fluids.
The Mder approach is more accurate than the traditional one, if good
data is available, but it is not so well documented.
The Benedict et a1 methods abandon use of the ratios of permanent
changes in mechanical energy to lunetic energy at a section in favor of a
more general, empirical relationship related to entropy increase across a
transition.
The Benedct et a1 methods have great potential, but they are not well documented.The author believes the Benedct et al methods to be potentially
the most accurate of all the methods, but ths remains to be proven. The
original simulation used conventional K factors exclusively.The simulation
described in ths appendut uses a method based on Benedct et al.
Data reduction
If we study the numerous fitting and piping transitions of the isometric
sketch of Figure AV-1, we see one sudden enlargement at the sink and five
more at the entrances to the four vertical segments. Each sudden enlargement, includng those at the branch inlets to pipe tees and those at changes
in pipe dameter must be examined for the possibhty of choked flow.
The four vertical segments also have sudden contractions at the inlets to
the downstream segments.
372
F l o w o f ~ n d u s t r i a~ ~I u i d s - ~ h e o r y a n d ~ q u a t i o n s
Complex Systems
A P P E N D I X
The many elbows present represent single input-single output devices.

A few have different, larger dmneters at their exits. Many are of Merent size.
The tees represent mixing devices.They may have different configurations,
sizes and connections.
Categories of causes of permanent changes i n mechanical energy
We can reduce the number of categories of transitions and conduit for
analysis as follows:
Non-mixing categories
1. sudden expansion;
2. pure bend;
3. enlargement;
Mixing categories
4. contiguous inlet tee;
5. opposed inlet tee;

6. branched inlet tee;
Non-mixing categories
7. contraction;
8. bend plus expander.

Estimation methods
We use our own adaptation of the methods of Benedct et alm. In postulating ths adaptation, we were not too concerned with absolute accuracy.
Establishing the computer simulation was the primary concern. It is
expected someone else wdl verifjr the method against reahty and will
signal any errors. Note the computer simulation has been used successfully
in simpler form (incremental method of finding the choked pressure and
original K factors), so we think this approach is reasonable.
The categories of conduit and transitions require different estimation
methods. For the straight conduit, we use conventional techniques and
"friction" factors. For sudden enlargements and contractions, the Benedct
et al methods are used. For elbows and tees, we use methods that were
modified from those of Benedict et al.The nominal K factors referred to
in the printouts are simply to give a starting point for corrections.
and Equations
373
A P P E N D I X
Compressible
Fluid Flow
Complex Systems
In what follows, remember the simulation calculates in the direction of

the flow for mass balances but in the opposite direction for pressure drops.
For elbows, bends and sudden enlargements, we use a method based on
the first sketch from ~ptimal-systemfbiifrom the paper by Benedict et
alX"".This method estimates the ratio of stagnation pressures across a fitting
(downstream to upstream) when the upstream ratio of flowing to stagnation pressures is known as seen by equation (l).The method demands iteration on the unknown stagnation pressure ratio and the unknown
upstream pressure to stagnation pressure ratio. The input is the pseudo K
value for translation to a value of the slope, 6, internal to the program.
For sudden contractions and combining tees, we use equation (2), which
can be solved directly since downstream conditions are known.This
method is based on the second sketch fiom ~ ~ t i m a l - ~ ~ s t e m f ~ ~ ~ ~ .
The first ratio in equations [l] and [2] is known as a;the second ratio, R ,
is either R I , the upstream flowing pressure to stagnation pressure ratio, or
R2, the downstream flowing pressure to stagnation pressure ratio. When
pedorming tests, the projection method of eliminating recovery effects
should be adhered to.
AV-7: MANIFOLD FLOW

Manifold flow is defined as splitting a single stream into two or more
streams or combining two or more streams into a single one.The combining fittings of safety relief vent headers f d under the general designation of madold flow.
The ~ e r n o u l l equation
i~
still applies in manifold flow. If a velocity reduction across a fitting produces a negative change in kinetic energy whose
absolute value is greater than the "friction losses" (permanent changes in
mechanical energy), the downstream static pressure will increase.This is a
fairly common, though often totally unexpected result.
The losses of mechanically useful energy are ofien expressed in terms of
"velocity heads" meaning a multiplying factor times the lunetic energy at
374
and Equations
Complex Systems
A P P E N D I X
a specific 1ocation.The location is associated with a K value and the user

must know whether it is a downstream or an upstream location.
The K factors or the "loss" coefficients are usually defined as the change
in specific mechanical energy across a path through a fitting divided by
the specific kinetic energy of the flow before it was dwided or after it was
combined. ~ d l e t defines
c ~ ~ the coefficients somewhat differently and the
coefficients of the two methods are not strictly interchangeable for compressible fluids although they are the same for incompressible fluids.
Benedct et aldo not use the K factors at all.
The numbering system used for piping networks in ths book identifies
the connections of combining or dividmg tees, so the major flow is always
number 3.This makes it easier to perform correlations on experimental
data and the data can be reduced. D.S. Miller has developed charts of his K
factors versus the fkactional volumetric flow.
Caveat
- Verifying definitions
Given the propensity of otherwise intelhgent people to use the same
name for different things or to use different names for the same thng, it is
wise always to check definitions carefully.
The correlations of Benedct et al can replace the K factor correlations for
conversion of mechanical to thermal energy in compressible flow when
greater accuracy is needed.The accuracy obtained will depend on that of
the correlation.These correlations are used to predict loss in mechanical
energy when a fluid traverses a fitting.
Estimating losses of mechanical energy across a pipe tee, three methods

The problem of estimating mechanical energy conversion to thermal
energy across a pipe tee is attacked by either one of three methods. In
each case, appropriate safety factors must be included to minimize error in
the wrong direction.The three methods are:
1. the traditional K factor method used for compressible and incompressible
fluids;
2. Miller's K factor method for compressible fluids.The FigureV-4 is used as
the basis fkom experimental data when Miller7.K factors are used.Ths
sketch shows the loss coefficients plotted against the ratio of volumetric
flow through the branch to volumetric flow in the combined leg.Two
curves are shown: one for the run and one for the branch.These curves
are established experimentally;
and Equations
375
A P P E N D I X
Complex Systems
3. a method based on the Benedct et al-
correlations for compressible
fluids.
At the beginning of an estimation, only the downstream con&tions are
known. Of the upstream condhons, only the source properties and the mass
flows in each leg are known unless one or more flows are choked.These
facts suggest that iteration is necessary.
For the tradtional K factor method, the iteration is started by assuming
the ratio of volumetric flows is equal to that of the mass flows. With this
initial ratio, the trial loss coefficients can be found, and the trial losses can
be estimated from the known downstream conditions.The trial losses are
~ ~ with
u i referring
~ ~ to~ upstream
/ ~ and~k ~
estimated from hik = K
downstream. The subscripts ik mean use the one that is appropriate.
The sum of the trial losses, the downstream kinetic energy and the downstream static energy equals the trial sum of the upstream lanetic and static
energies.The unknowns can be transformed into functions of T and v
using the Redlich-KwongXrvlU equation and U=Gv.The total energy
equation also is given as E(Tl,vl).We now have two equations for
Newton-Raphson iteration on trial values of upstream temperature and
specific volume.The next trial of the volumetric flow ratio can be estimated from the mass flow rates and the specific volumes.The algorithm
used in our original program made use of an external "false position" loop
to achieve convergence. Once converged on temperature and specific
volume, pressure can be found from the R-K equation.
Note for both K factor methods, losses across the exit to a sink are based
on the upstream velocity (lanetic energy).The velocity in the sink is
essentially zero.
AV-8: V I S C O S I T Y C O N S I D E R A T I O N S
Viscosity is a property of a fluid used to estimate its resistance to flow.
Viscosity is particularly important in estimating pressure drop through
uniform section pipe segments. It is not usually a factor in estimating
changes in mechanical energy through fittings.
The treatment of the subject of viscosity in most texts is entirely unsatisfactory.The use of concepts such as "fluid friction" flies in the face of
common sense. Even a high school student knows molecules do not rub
376
F l o w o f ~ n d u s t r i aF~l u i d s - ~ h e o r y a n d ~ q u a t i o n s
Compressible
Fluid Flow
Complex Systems
A P P E N D I X
together. Friction is a macroscopic phenomenon. The concept is applied

well to solid materials such as brake pads and drums. It falls flat on its face
when applied to molecules that have force fields that repel other molecules when they come too close.
The problem seems to be that a satisfactory set of equations that describe
viscosity theoretically is extremely chfEcult to formulate. However, viscosity
can be measured experimentally quite readily. Rather than abandon, or at
least postpone, theoretical considerations in favor of practical ones, most
authors seem to compound the d~fficultyby using confusing terminology
and by ignoring some fundamental thermodynamic considerations.
Friction factor
It is not possible to avoid using the term, fiiction, in all cases - especially
since it is ...
incorporated into textbooks and much reference material, such
as Cranem.The friction factor is a case in point.
The fiiction factor helps give a measure of the resistance to flow of fluids
that extends the utility of the concept of viscosity. It really only applies to
conduits of constant internal section.Viscosity is one of the properties
influencing the magnitude of the fiiction factor.When considering
changes in mechanical energy through fittings or expansions and contractions, the concept of "form friction" is more useful and the K factors
come into play. Friction factors are well established and have been thoroughly correlated for liquids, at least, with the various flowing quantities.
is
The K factors are less well established-The work of Benedict et alpromising as a more accurate method that replaces the use of K factors.
Caveat
- Using correct K factors

When using the various coefficients, the reader is reminded to use them
only with their applicable equations.Two definitions for K factors, the traditional one and that of Mdler, should serve as an example of Werences in
the governing equations,
Viscosity plays a role in &iction factor estimations. It is a fluid property that
serves to give a measure of the resistance to motion of the fluid.Viscosity
arises out of the forces, attractive and repulsive, that exist between molecules.
When fluid is put into bulk motion by an external force, the intermolecular
forces move with the molecular movement. Forces moving through a
distance in the chrection of the forces constitute the definition of work.
Work is done and the mechanical work energy of compression is partially
converted to thermal energy.
and Equations
377
A P P E N D I X
Complex Systems
The difference in mechanical work energy along a pipe or across a fitting

constitutes a loss of available mechanical energy and is treated as lost
energy even though first law considerations teach that total energy is conserved. One has to remember the word "losses" applies only to mechanical
energy, not to mechanical plus thermal energy (the total energy remains
constant).
Three indices for viscosity
In the case of uniform cross section pipe, estimations for the friction
factor require an index of the type of molecular interaction that wdl
occur.There are three major types of interaction that have been modeled:
that between hard spheres with no attractive or repulsive forces; that
between nonspherical molecules that can form electrical &poles due to
charge redistribution; and, that between molecules with isolated positive
hydrogen atoms that can attract or bond with the negative portions of
simdar or different molecules.
Our simulation relies on the work of Reid, Prausnitz and ~ h e n v o o d ~ in
"'
the use of these indices.The methods used are corresponding states
methods.The assumptions for mixtures of gases are the molecules are sufficiently separated and their velocities are so great only two-body interactions need be considered.The bodies can be s i d a r or different molecules.
The first model used for gases was the ideal gas model.The ideal gas
model concerns itself with spherical molecules that have no intermolecular forces except at the time of closest approach.This model subsequently
was modified to take into consideration polar molecules that definitely
were not spherical and that could form dipoles due to weak van der Waals
forces.An additional modification considered those molecules that form
stronger bonds between some of their positive hydrogen atoms and a negative kernel of a simdar or a different type of molecule.
Three formulae for two-body interactions resulted for:
nonpolar gases;
polar gases, nonhydrogen bonding;
polar gases, hydrogen bonding.
The first category, nonpolar gases, includes the vast majority of pure gases.
Examples are: acetylene, benzene, isobutane, n-butane, carbon dioxide,
carbon disulphide, carbon tetrachloride, chlorine, cyclohexane, ethane, ethylene, methane, n-pentane, propane, propylene, sulphur dioxide and
toluene.
378
~ I o wo f ~ n d u s t r i a l F l u i d s - ~ h e o r y a n d ~ q u a t i o n s
Complex Systems
APPENDIX
The second category, non-hydrogen bondmg polar gases, includes

dimethyl ether and chloromethane.
The thrd category, hydrogen bondmg polar gases, includes water and most
alcohols, ethers and phenols.The main manifestation of hydrogen bondmg
is a hgher b o h g temperature than that of close molecular weight hydrocarbons.This category includes water, methanol, &methyl ether, ethanol,
n-propanol, i-propanol, n-butanol, i-butanol, n-pentanol, n-hexanol,
n-heptanol and acetic acid. Note the positive hydrogen has to bond to an
electronegative component either in a similar or in a different molecule.
The simulation makes use of a simple index, M = 1 , 2 or 3, that is applied
to each individual component of a gas mixture to instruct the program on
the formula to apply.
Gas viscosity
The theory of gas viscosity depends heavily on that of incompressible

fluids.The following assumes larninar flow.
Gas molecules in a container move randomly, recoil from one another,
transfer momentum, and recoil from the walls of the container.The
changes in momentum at the walls are measured as pressure.
If a fluid is put into motion by movement of a solid wall oriented parallel to
the motion, the molecules vvlll have a drected motion superposed on their
random motion.Ths directed motion is fluid flow. The drected motion has a
velocity that is the same as that of the moving element close to the element,
and that falls off perpenddar to the element. The directed flow can be separated into layers.The random movement of molecules s d takes place
between 1ayers.Ths random movement now has a hected component that
imparts negative or positive momentum to the receiving layer of fluid.
The faster moving layer wdl receive negative momentum; it will be
slowed down.The slower moving layer wdl receive positive momentum; it
will have its velocity increased in the dn-ection of flow. Even in the ideal
case, the velocities wdl tend to equalize.
Practically, an external force is required to produce flow. It is known that a
force produces an acceleration. Therefore, if an external force produces a
steady-state velocity without acceleration, other forces, equal and opposite
to the external force, must arise in the fluid in order to maintain the
steady state. These forces are called viscous forces.
and Equations
379
A P P E N D I X
Compressible
Fluid Flow
Complex Systems
Usual description of viscosity

The usual description of viscosity starts by showing two parallel plates, one
of which is in motion. The fluid under investigation is between the plates.
Practical viscometers are either formed of capillary tubes, or of a rotating
drum around a stationary cylinder, or vice versa. If the cylinders and drums
were sufficiently large, they could approximate the two parallel plates.
The analysis usually starts with a moving plate of known area to which a
force is applied causing the motion. From experiments with real viscometers, the force is proportional to the area of the plate in motion and to the
velocity gradient.The velocity gradient is the rate of change of the velocity
in the direction of motion with the change in distance in the perpen&cular direction. It is important to distinguish the dn-ections and to
remember the property, viscosity, is measured under laminar flow conditions. Further correlations are required to make it usefd for turbulent flow.
The equations describing the above are:
An equal and opposite internal force is necessary to prevent acceleration

fiom the steady state.Velocity,v, is perpendicular to area, S.
Flow through tubes

J. S. PoiseuilleXV
(1844) established the basic practical equations for larninar,
incompressible flow through tubes. At each velocity or flow rate, external
force is necessary to overcome the internal force or resistance to flow.To
maintain steady-state flow, thls external force must be applied constantly.
For flow to occur in tubes or pipes, there must be a pressure difference. In
a uniform hameter pipe, the force in the direction of flow is the differential pressure times the cross sectional area of the pipe.To prevent acceleration, an internal force must arise and must be equal to and opposite from
the external force.
380
and Equations
Compressible
Fluid Flow
Complex Systems
A P P E N D I X
Viscosity is the name given to the coefficient of proportionality, eta, of the

above relationship.
The units of viscosity may be obtained from the set AV-3.
(AV-3)
Functional relationship between point velocity and radius of pipe

From equation AV-2, the relationshp between point velocity and ralus
requires an indefinite integral.The stationary plate is a cylindrical reference area in this case.
v=Jdv+c=j
(P, -P,)rdr
2r71
+c
When v = 0, r = R.
So that
and Equations
381
A P P E N D I X V
Compressible
Fluid Flow
Complex Systems
Total volumetric flow, incompressible fluids

The point velocity in steady state, incompressible, laminar flow is proportional to the product of the differential pressure and the difference of the
squares of the pipe radius and that of the point velocity. It is inversely proportional to the product of viscosity and flow length.
To obtain the total volumetric flow, we must integrate the point velocity
across the normal cylindrical cross sections of the pipe. The term 2nrdr is
an annulus of area to be multiplied by the point velocity, v, in order to
give volumetric flow through the annulus.
The last equation in the above development is Poiseuille'sx" equation for

the laminar flow of incompressible fluids through uniform pipes or tubes.
It shows laminar flow is proportional to differential pressure and the
fourth power of the radius of the conduit. It is inversely proportional to
the viscosity. The equation permits the experimental determination of the
viscosity of incompressible fluids from measurable quantities
382
and Equations
Compressible
Fluid Flow
Complex Systems
A P P E N D I X
Total volumetric flow, compressible fluids, standardized condition

To extend the use of the PoiseudleXV
equation to compressible fluids, it is
usual to assume ideal gas behavior.The subscript zero refers to the standardized condition. The terms V, andVo represent average and standard
total volumes.
The last equation of the above development permits the experimental

measurement of the viscosity of compressible ideal gases using standardized quantities.
'Elementary' kinetic theory

Transport properties are properties used in the description of mass,
momentum and energy fluxes. Flux is the transport across a unit area
under a density &fferential.The density is mass per unit volume in the
case of mass flux; it is momentum density (nmvy) in the case of moment
flux and it is energy density (cpT) in the case of energy flux. The three
associated transport coefficients are D, for mass,q, for viscosity and h for
energy.We d only be concerned with viscosity at this point.
Depending on the model used, a value can be estimated for the viscosity
coefficient. For the noninteracting, rigid-sphere model, the coefficient is
estimated from AV-9.
5' = 26.69-
(AV-9)
(MT)"~
o2
and Equations
383
A P P E N D I X
Complex Systems
The viscosity coefficient, eta, has units of micropoise; M is the molecular

mass (weight) in grams per g-mo1e;T is the temperature in degrees K, and
sigma is the hard-sphere diameter in angstrom units.
The theoretical treatment of the effect of intermolecular forces is usually
handled by Chapman-Enskog equations. These equations involve an intermolecular potential function and are extremely &fficult to grasp. Practical
equations usually use corresponlng states methodology. Thodos and coworkers developed the equations (slightly modified by Reid, Prausnitz and
herw wood^^^^) that follow:
Nonpolar gases
V = (4.6101;O.~~~
- 2.04e-0.4949T.
+ 1.94e-4.058T,
+ 0.1)/{
(AV-10)
Non-hydrogen-bonding polar gases a t reduced temperatures below 2.5
V = ((1.91; - 0 . 2 9 ) ' / ~ ~ ; ~ / ~ ) / 5
(AV-11)
Hydrogen-bonding polar gases a t reduced temperatures below 2.0
11 = ((0.7551; - 0.055)~;~")/{
In the above equations, M is molecular mass in grams/g-mole, Pc is the
critical pressure in atmospheres,Tr is the reduced temperature,T/Tc, with
temperatures in K and eta is in micropoise. Epsilon is given by AV-13.
5 = T,'/6M-l/2pc-2/3
(AV-13)
The Thodos equations should not be used with hydrogen, helium or with
diatomic halogens. Reid, Prausnitz and sherwoodhi suggest using the
Golubev equations in ths case, AV-14. The Golubev equations do not
work for gases that associate significantly.
384
F I O W
Average error
Complex Systems
A P P E N D I X V
Reid, Prausnitz and herw wood^"^ gave average percentage errors compared with experimental values of 18 nonpolar gases using the Thodos
and co-workers method as 2.2%. For eight polar gases, the error was 3.1%.
There may be more accurate methods but, given the ease of programming
the above equations and the acceptable accuracy for industrial purposes,
the Thodos method will be used in this work.
Viscosities of low-pressure gas mixtures

The experimental viscosities of gas mixtures are shown in Figure AV-2,
taken from Reid, Prausnitz and Sherwood.There is sometimes a maximum as shown by the ammonia-hydrogen binary mixture and to a lesser
extent by the ethylene-ammonia binary mixture. Minimums are not
encountered. An approximation to the rigorous solution is,
(AV-15)
" Yiqi
qmix
i=l
Wilke approximated the binary coefficient, p h , by the equations shown in

AV-16.
The error varies with the mixture, but is usually less than 2.0%.
and Equations
385
A P P E N D I X V
Complex Systems
Mole fraction first component
Figure AV-2. Gas mixture viscosities

Temperature, K
No
System
Hydroqen sulfide-ethyl ether

Hydroqen sulfide-ammonia
Methane-n-butane
Ammonia-hydrogen
2
3
4
33 1
33 1
293
306
The expansion of equation AV-1 6 is
~ 1 @ 3 1+ ~ 2 @ 3 2+ ~ 3 @ 3 3
p=
386
micropoise
Y3V3
/ 10000 cenfipoise
and Equations
Complex Systems
A P P E N D I X
The ReynoldsVnumber, NRe,is given in AV-1 8.
Wn= Ib,,, / h
d = id
inches
p = viscosity coeficien t, centz'poise

The fiiction factor, fM,can be estimated with the ~ h u r c h d l - ~ s a gequai~~
tions.
Fixed length increments can be chosen such as five feet and the upstream
pressure can be found by using the Peter Paige equation for the mechanical energy balance with losses and the total energy equation to solve by
Newton-Raphson iteration for T and v. The Reach-Kwongm"U equation will permit a solution for pressure once the upstream values of temperature and specific volume are known.
and Equations
387
A P P E N D I X
Complex Systems
A V - 9 : S I M U L A T I O N RESULTS ( A N A L Y T I C M E T H O D )
388
F l o w o f I n d u s t r i a l ~ l u i d s - ~ h e o r ya n d ~ q u a t i o n s
Compressible
SEG
Fluid Flow
Complex Systems
A P P E N D I X
Simulationof Complex Pipi.ngNetwork- Multicomj onent Vapors

FLOW
Ibm/h
40
7084.3
P1
T1
psig
v1
ft3/lbm
15.4
296.1
14.6
U1
ftls
560.1
P2
I T2
v2
U2
VMACH
psig
ft3/lbm
ftls
ratio
15.3
282.3
14.5
567.9
0.341
and Equations
389
A P P E N D I X V
Complex Systems
A V - 1 0 : S U M M A R Y OF A P P E N D I X AV
This appendm has tried to outline the most important aspects of simulating complex piping systems relieving gaseous mixtures. The method
chosen was to describe a common industrial situation involving a complex piping network and complex mixtures of components.The system
was a vent header, common to most industrial plants and refineries.
The importance of the following was emphasized:
defining the piping network by isometric sketches;
defining the flow regimes by grouping devices that may relieve simultaneously;
having an organized plan of attack;
establishmg the choke points;
understandmg madold flow and when to make use of worst-case data;
understandmg the K factors used and the importance of piclung the correct lunetic energy term;
the Merences between the three methods of estimating permanent
mechanical energy changes across a transition:
1. Conventional K factors
2. Mdler's K factors
3. Benechct et alXXXV
correlations
using engineeringjudgment in influencing design to avoid unknowns;

the si@cance
of the various drawings and data sheets;
personally v e r e n g data by a final walls-down of the system using an

isometric sketch as a basis for recordmg "as-built" mformation.
The results of a Fortran simulation of multicomponent gas flow in a very
complex piping network were given.The simulation was based on most of
the ideas presented in ths book.
390
Flow of Industrial ~luids-Theory and Equations
Endnotes
Bernoulh, Daniel, 1700-1782. Swiss Mathematician and Physicist. Developed the ideal
Bernoulli equation that was later moddied to include irreversibilities. Historical reference.
ii
Carnot, Sa&. French Engineer. 'Father ofThermodynamics'.The father of thermodynamics, Sadi Carnot, had a father, General Hypolite Carnot.The grandfather of thermodynamics not only was an army general under Napoleon, he was an important
politician (member of the Directorate) during the French revolution and he was a
hydrauhcian who wrote extensively on such things as shock losses in turbomachnery.
Fox, Robert., "Riflexions sur la puissance motrice du feu7',Vrin,Paris. Historical reference.
...
Coriolis, Gustave-Gaspard, Circa 1835. French engineer-scientist who described the

force named afier h that is seen as a force perpendicular to an object's motion.
Historical reference.
iv
Kelvin, Lord,Wfiam Thompson, 1824-1907. Enghsh scientist who, along with

Clausius, German, was active in promoting Carnot's ideas. Responsible for the
absolute temperature scale named afier him. Historical reference.
Reynolds, Osbourne, 1842-1912. Enghsh physicist and mechanical engineer.

Historical reference.
vi
Venturi, Giovanni Battista, Circa 1800. Italian physicist and hydrauhcian. Historical reference.
vii
Darcy, Henry Phhbert Gaspard, 1803-1858. French hydrauhcian. Known for the
equation of the same name that relates irreversibhties to fluid viscosity, density and
velocity.The equation is sometimes called the Darcy-Weisbach equation. Historical
reterence.
viii
Moody, Louis E, Circa 1944. Developed the 'fiiction factor' charts used extensively in
this book.
and Equations
391
A P P E N D I X
Endnotes
ix
Kegel (Flow Measurement, ISA)
Perry, Chemical Engineers' Handbook, 50 th edrtion, McGraw-Hdl. General

reference.
xi
Ostwald (1853-1932)and de Waele, Ostwald was born in Latvia but was professor in
Leipsig and is claimed by Germany. Nobel prize for catalysis in 1909.Taught
Arrhenius, van't Hoff and Nernst among others.The name Ostwald-de Waele is given
to a power law function describing apparent viscosity. Historical reference.
xii
Bernhardt, Processing of Polymeric Materials, Reinhold, N.Y,1959. Source of further

information.
xiii
Crane,Technical Paper No. 410. Seventeenth Printing, 1978 - but see latest printing.
General reference.
xiv
Fanning, Circa 1880. Responsible for the 'fkiction factor' named after hlrn. Historical
reference.
xv
Hagen-Poiseudle. G. H. L. Hagen, German hydradc engineer and Louis Marie

Poiseudle, French physician and physicist ,Circa 1844. Historical reference.
xvi
Mdler, D.S. Internal Flow Systems, 2nd Edtion, BHRA Cranfield, UK, 1990.
xvii
Driskell, L.R., ISA Handbook of Control Valves, Table of Representative Valve Factors
xviii
Simpson, L.L., Chem. Eng.,July 17,1968, Head loss (mechanical energy) formula for
square edged, concentric ordice plates, Equation 11-16
xix
Mdler, Richard.W, Empirical formula for permanent loss across an ordice curve-fitted
to the ASME head loss curve (ASME Research Report on Fluid Meters). Source of
further information: Flow Measurement Engineering Handbook.
XX
Hero, Greek, Second century BC, Developed a fire pump with two cylinders and a
hand-operated roclung beam to supply motive force. Historical reference.
XXi
Agricola, German, 16th century. Described the extensive use of pumps in the mining
industry for mine dewatering purposes. Historical reference.
xxii
Papin, Denis, French inventor, Circa 1689.The invention of the centrlhgal pump is
generally credrted to hlm. Its use had spread around the world by the mid 18005.
Today, the centrifugal pump is an omnipresent part of our existence. Historical
reference.
xxiii
Schutte and Koerting, Manufacturers of ejectors, amongst other thmgs.
xxiv
Paige, Peter, Engineer responsible for the equation of the same name.
(Chemical Engineering,Aug 14,1967)
XXV
392
Richter, S.H., Size Relief Systems for Two Phase Flow, Hydocarbon Processing,
July 1978.
and Equations
Endnotes
A P P E N D I X
xxvi
Laplace, Simon Laplace, 1749-1827. French scientist. Corrected Newton's hypothesis

for isothermal flow in computing the speed of sound in air to an adabatic hypothesis (18 16).Historical reference.
xxvii
van der Wads, Dutch physicist, circa 1870, one of the first to try to improve the ideal
gas model with a model that involved two parameters. Historical reference.
xxviii
Reach-Kwong equation, original, circa 1950,is probably std the most famous EOS.
xix
Reid, Prausnitz and Poling,The Properties of Gases and Liquids, McGraw-Hill,

Fourth Edtion, 1987. Successor to an equally useM and prestigeous book by the
same name but authored by Reid, Prausnitz and Sherwood. General reference for
physical-chemical data.
xxx
Prausnitz,Anderson and Grens, Computer Calculations for MulticomponentVaporLiquid and Liquid-Liquid Equihbria. Prentice-Hall, 1980, out of print but &l1 of
good ideas. Developed a solid database for its use in normal plant situations.
xxxi
McCabe and Smith, Unit Operations of Chemical Engineering, McGraw-Hdl,

1967. Gerneral reference.
xxxii
Benjamin and Miller,The Flow of a Flashng Mixture ofVlrater and Steam through
Pipes, ASME Transactions, October 1942. Cited by Peter Paige who used the
Benjamin and Miller Data to arrive at some of his conclusions. Historical reference.
xxxiii
D~ck1er~A.E.
et al,"Frictional pressure drop in two-phase flow",AIChE J., 10 44-51
(1964)
xxxiv
DeGance and Atherton, Published useful information on two phase flow systems.
Ch. Eng, March 23,1970
xxxv
Benedct et al (Benedct, R.P., N.A. Carlucci and S.D. Swetz, Flow Losses in Abrupt
Enlargements and Contractions,Journal of Power Engineering,January 1966). Cited
as a source of data and ideas by Optimal-Systems.
xxxvi
ASME steam tables, General reference.
xxxvii Blasius (student of Prandtl), circa 1911. First to establish that the fkiction factor followed a functional relationship in Reynolds" number below a Reynoldsv number of
100 000 in hydradcally smooth pipes. Equation AI-30. Historical reference.
XXXViii Prandtl, circa 1911, Prandtl developed another equation for the relationshp between
a fkiction factor (a wall fkiction factor, superscripted 6) and the pipe Reynoldsv
number. Historical reference.
Nikuradse, circa 1900, correlated ~ a r c y ' data,

s ~ He carefully glued the sand on the
internal walls of pipes. Using ths method, he was able to vary the relative roughnesses of pipe walls between U500 and 1/15, about 33:l. Historical reference.
and Equations
393
A P P E N D I X
Endnotes
xl
Colebrook,circa 1911, empirically arranged the Prandtl equation and one developed
by von Kh-mh to produce a formula that gave remarkably good results in correlating
published data. Historical reference.
xli
Churchd-Usagi. Stuart VI! Churcld of the University of Pennsylvania presented a

convenient formula of estimating the fiiction factor fiom the ReynoldsVnumber and
the relative surface roughness (Chem. Eng. Nov. 7,1977)
Xhl
Weber and Meissner,Thermodynamics for Chemical Engineers, 1957
x
l
l
l
l
Benedict, R.P., Fundamentals of Pipe Flow,Wiley 1980.
xliv
Hydraulics Institute. Source of reputable data
xlv
Chemical Engineering Thermodynamics, Balzhiser, Sarnuels and Eliassen (BS&E).

No longer in print.
xlvi
Reid, Prausnitz and Shenvood,The Properties of Gases and Liquids, gives various
means of estimating mixture viscosities. See successor book by Reid, Prausnitz and
Poling.
xlvii
Optimal-Systems,Website - http://optimal-systems.demon.co.uk.Commercial
organisation, source of good data.
xlviii Golubev equations for estimations of mixture viscosity. Used by Reid, Prausnitz and
Sherwood.
394
and Equations
Table of Principal Symbols and

Glossary of Principal Terms
and Units
Note that most of the symbolism, including the subscripts and superscripts, used in this
book is found in the general literature. A small amount is the author's invention.The
author only included in these tables that which he felt important. If the reader wants further enlightenment, it can usually be found in the narrative surrounding the equation in
which the symbol was first used in the main text.
Caveat -Always
check the units of each term of each equation at the start of a computation. Each term should have the same units as every other term of the same equation.The
author has attempted to be consistent and to define each term carefully, but he has been
constrained by common useage in some cases.
Description
Cross sectional area, conduit

Calculated term used in Churchill-Usagi
equation
'Acoustic' velocity
R-K re~ulsioncoefficient
Pump impeller inlet blade angle
Correctionfactor (value of ratio of a point
velocity to the average velocity across a
section)
Ratio yl(y-l)
Calculatedterm used in Churchill-Usagi
equation
R-K attraction coefficient
Value of ratio of a small diameter to a
larger one
Pump impeller outlet blade angle
Constant point velocity
lsobaric'heat1capacity1mass
l~obaric'heat'capacity~
molar
lsometric'heat'capacity
Isometric 'heat' capacity, molar
F
Fig.Alll-5
AV-4
IV-l2
not used
IV-l3
not used
C,
All-l
c,
c,
c,
Valve coefficient - proportionality between

flow rate of water at 60F (US.gpm) and
(~p)l/~(psi)l/~
Customary
U.S. Units
feet2
SI Units (or
submultiples)
m2
feet/s
degrees
dimensionless
mls
(J/~Q)KO.~
degrees
dimensionless
dimensionless
dimensionless
ft3/lb,
dimensionless
m3/kg
dimensionless
degrees
ft3/s
Btullb
R
-,
Btullbm,-R
Btullbm-R
Btull
R
-,,,b,
degrees
NA
Jlkg-K
Jlkg-mole-K
Jlkg-K
Jlkg-mole-K
I (ft-lbfllbm)R0a5 I
~ . ~ . g p m l p s i ' ~ NA
and Equations
395
A P P E N D IX
Term
d
d
rl
Example
EqJNote
1-1
1-3
Alll-18
AV-9
1-3
6
D
D
l of AV
1-1
All-10
1-3
1-23
All-10
Alll-31
AV-13
,;f
1-17
f
F
F
All-10
1-6
AI-57
A
e
E
E dot
Y
g
g,
G
h
h
hfl(hL)
I
I
1-24
IV-13
1-3
1-3
1-2
1-3
1-4
h, h,
H
J
K
AIV-22
1-23
IV-2
1-21
AIV-4
1-1
1-1
1-6
1-2
1-1
P
&p
Pm
m
NR,
396
Table o f Principal Symbols a n d Glossary o f Principal Terms a n d Units
Description
I
I
Customary
U.S. Units
Diameter
inches
Differentialoperator
dimensionless
Pump efficiency, energy transmitted to fluid1 dimensionless
enerqy transmitted to shaft
Viscosity coefficient, eta
Poise11o6
Path dependent operator, differential or small dimensionless
but finite
Gradient
I dimensionless I
Diameter or length dimension
feet
Pipe ID
feet or inches
Finite change between limits
dimensionless
Total energy per unit mass at a section
ft-lbf/lbm
Absolute rouqhness
feet or inches
Total mechanical transferred per unit time
ft-lbf/s
Epsilon, a calculated coefficient in the Thodos et al relationships
Moody friction factor (subscript F refers to
dimensionless
Fanninq)
-dimensionless
Frictionfactor in Churchill-Usagi equation
Force
lbf
Mechanical energy converted to
ft-lbfllbm
irreversibilities
'Weightldensity - synthetic unit substituted 1bf/ft3
for mass densitv
I
I
dimensionless
Ratio of 'heat capacities: c&
feet/s2
Acceleration of gravity
32.1 7 1b,-Wl bf -2
Dimensional constant
Ib,/s-ft2
Mass velocity (flux)
Btullb,
Enthalpy (flow energy) (mass basis)
Btullb-mole
Enthalpy (flow energy) (molar basis)
ft-lbf,lbm
Mechanical energy converted to thermal
energy ('losses')
Btullb,
Stagnation enthalpy
'Head1- synthetic unit substituted for energy feet
776.1 6 Ibf-WBtu
Conversion factor mechltherm
'Loss'coefficient of an obstruction (fraction of dimensionless
kinetic energy converted to thermal energy).
ft
Length or position
Ib,/s-ft
'Absolutelviscosity
0.01glcm-s
viscosity (centipoise)
Ib,Ift-s
Viscosity (no name, US. units)
Mass (a dot signifies per second)
Ib,
dimensionless
Reynolds number (a further subscript, D,
refers to pipe ID - o refers to orifice bore)
and Equations
S1 Units (or
submultiples)
NA
dimensionless
dimensionless
Poise11o6
dimensionless
dimensionless
meters
meters of mm
dimensionless
Jlkg
meters or mm
Jls
dimensionless
dimensionless
N
JIkg
N/m3
dimensionless
m/s2
l,none
kgls-m2
Jlkg
Jlkg-mole
Jlkg
Jlkg
meters
l,none
dimensionless
m
Kgls-m
0.01 glcm-s
kglm-s
kg
dimensionless
Table o f P r i n c i p a l Symbols a n d Glossary o f P r i n c i p a l Terms a n d Units
Term
Example
Eq./Note
IV-16
AV-15,16
,N
,
P
P
P
1-4
All-l
AM-21
11-19
1-3
71:
111-1
AM-l
111-1
Q
Q
1111-2
Q dot IV-2
IV-23
r,
R
IV-10
IV-10
R,
AIV-24
R
AI-59
R,
q
q
1-1
AI-49
p
S
'C
T
T,
U
v
1-7
1-8
IV-l0
IV-20
AI-57
I-1
AV-4
1-2
AIV-26
V
1-6
o
I AM-6
W
1-3
1-1
W
W dot1 IV-2
X
1-8
X
1-3
Z
1-6
v
-
Description
Mach number
Phi, calculated coefficients in the Thodos
relationships
Pressure (absolute)
Differential pressure
Power (rate of transfer of energy)
Constant eaual to 3.1 416.. ...
Mechanical energy per unit mass that has
flowed under a temperature difference
Liquid volumetric flow rate
Flow rate in U.S.gpm of water at 600F
Liquid volumetric flow rate
Flow rate in U.S.gpm of water at 60F
Energy flow rate, thermal units
Hydraulic radius
Gas constant, mass
'Universalrgas constant, molar
Gas constant, molar
Resisting force developed by parallel layers
of fluid
Density (mass)
Entropy per unit mass (energylmasstemperature)
Stress (force per unit area)
Variable time
Temperature, absolute
Stagnation temperature
Internal energy
Velocity (averaqe across section)
Point velocity (as opposed to average
velocity, U)
Mass volume
Molar volume
Linear velocity
I Anqular velocity
Net work energy per unit mass
Mass flow rate
IWork enerqy flow rate, thermal units
Variable distance
Vertical dimension
Distance
Customary
U.S. Units
dimensionless
dimensionless
S1 Units (or
submultiples)
dimensionless
dimensionless
Ibf/ft2
Ib,/inch2
foot-lbf Is or hp
dimensionless
ft-lbf/lbm
N/m2 (Pa)
NA
J/s or CV
NA
Jlkg
ft3/s
Ibf
NA
NA
NA
NA
J/s
m
N-mlkg-K
N-mlkg-mole-K
N-mlkg-mole-K
N
lbm/ft3
ft-lbf/lbm-R
kg/m3
N-m/kg-K
gPm
ft3/h
gPm
Btuls
I ft-lbfl lbm-R
ft-l bf/lbmo16R
ft-lbf/lbm,,dR
ft
A P P E N D I X C
Ib, /ft2
N/m2
R, (459.67 + F)
R, (459.67 + F)
ft-l b, /lb,
ft/s
ft/s
ft3/lbm
ft3/lb-mole
ft/s
I l/t
ft-lbf/lbm
Ibm/h
1 Btuls
feet
feet
feet
I
1
K, (273.1 5+C)
K, (273.15+C)
N-mlkg
m/s
mls
m3/kg
m3/kg-mole
m/s
l/t
Jlkg
NA
J/s
meters
meters
meters
and Equations
397
A P P E N
D IX C
Table o f P r i n c i p a l Symbols a n d Glossary o f P r i n c i p a l Terms a n d U n i t s
Major subscripts and superscripts used i n equations

Subscript or superscript
Example equ./note
Meaning
-X
AIV-15
Qualifier to indicate value at choked section
Re, ReD
1-l,!-18
Reynolds number, number base on diameter D
CP
1, 2, (0
n, f
1-3'1-4
1-3
Centipoise
Section specifier
Net or fluid (energy transferred)

Qualifier to indicate a constant orlcritical'
cl e
m
11-15
Qualifier to indicate compression or expansion
1-6
Qualifier to indicate mass based quantity
11-16
Qualifier to indicate orifice
' (prime)
1
I
1-9
IV-28
I Qualifier to indicate minimum

I Qualifier to indicate an ideal, perfect gas.
1-10
Qualifier used on shear stress of pseudoplastics
1-10
Ostwald and de Waele index
1-12
Qualifier to indicate force based quantity
fs
IV-5
Qualifier to indicate'skin friction' ( double misnomer)
1-19
Fanning
IV-8
Qualifier to indicate hydraulic
Example AIV-1
11-1
1-18
Qualifier to indicate quantity at initial choke point

Qualifier to indicate a 'loss' (conversion) in mechanical
energy
Moody
Ma
Mav
I IV-16
I AIV-4
IV-12
I Mach
I Moody, average (over integration)
I
Qualifier to indicate constant pressure

Qualifier to indicate units are Poise, glcm-s
1-15
rad
Alll-13
Radial
Example IV-l
Qualifier to indicate relative or divided by critical value
IV-9
Qualifier to indicate constant entropy
AIV-12
Qualifier to indicate stagnation
AIV-22
Qualifier to indicate source or reservoir condition
tan
Alll-3
Tangential
11-18
Qualifier to indicate total
IV-13
Qualifier to indicate constant volume
398
1-1
1-2
11-17
I Qualifier to indicate wall
and Equations
Table of Caveats
The word 'Caveat' has been used often in this book to warn unsuspecting readers to pay
particular attention to a common intellectual trap.The word comes from the Latin 'Caveat
Emptor' which roughly translated means 'Let the buyer beware!'The author feels strongly
that highly intelhgent people have a propensity to use words loosely, in a Mad Hatter from
Ahce in Wonderland fashion. It is supposed that their intent is to find a simple way of
expressing complex concepts.They only succeed in confusing issues.There is no such thing
as a free lunch.
Caveat Subject
Page
Reynolds numbers
Average or point velocities
Bernoulli equation versus the first law
Fluid classifications
Equivalent lengths
Velocity profiles
Hydraulic enqineerinq practice
Gravity trap
Bernoulli balance
Safety factors
Velocity and the K factor
Mechanical enerqy 'losses'versus pressure
Pressure recovery
Source of K factor correlation
Proximity of disturbances
Viscosity units
Negative K values
Definitions of K factors
Steady-state assumptions
Hydraulic enqineers'simplification
Velocity used with K factor, Suffixes, K factors, C,
18
21
29
34
35
54
56
57
62
63
66
1
I
75
80
81
200
201
228
257
260
Conversion of irreversibilities to pressure drop

Authors'definitions of loss coefficients
Conversion of 'head'to pressure
Reference states for thermodynamic properties
lnteqration with constraints
Verifyinq definitions
Using correct K factors
69
268
268
283
320
340
375
377
and Equations
399
Selected Bibliography
Benehct, R.F? Fundamentals
of Pipe Flow. Wiley, 1980.
The Crane Company. Flow of Fluids Through klves, Fittings and Pipe, Technical Paper No. 41 0.
Crane, 1988 (reprint of 1942 document).
Driskell, Les. Control klve Selection and Sizing. ISA, 1983.
Driskell, L.R. "Sizing ControlValves, Part I: Sizing Theory and Applications." ISA Handbook
of Control Wlves. Second Edition. Pp. 180 - 205. ISA, 1976.
Kegel, Thomas. "Chapter 22: Insertion (Sampling) Flow Measurement," Flow Measurement.
Second Edition. D.W. Spitzer, Edtor, pp. 597 - 638. ISA, 2001.
Kyle, Benjarnin G. Chemical and Process Thermodynamics. T h r d edtion. Prentice Hall international series in the physical and chemical engineering sciences. Prentice Hall PTR, 1999.
Mdler, Donald S. Internal Flow Systems. Second ehtion. BHRA (Information Systems), 1990.
W e r , Richard W Flow Measurement Engineering Handbook.T h r d Editon. McGraw-Hdl, 1996.
Perry, Robert H., Don W. Green, and James 0 . Maloney, Ehtors. Perry's Chemical Engineers'
Handbook. Seventh Edition. McGraw-Hdl Professional, 1997.
Pohng, Bruce E, Jon Prausnitz, and John F? 07Connell.The Properties of Gases and Liquids.
Fifth Edition. McGraw-Hill, 2001.
Smith,Juhan C., Peter Harriot, and Warren L. McCabe. Unit Operations of Chemical Engineering.
Sixth edition. McGraw-Hdl Science/Engineering/Math, 2000.
Hydraulic Institute - www.pumps.org
9 Sylvan Way
Parsippany NJ, 07054
Phone: (973) 267-9700
and Equations
401
A P P E N D I X
Selected B i b l i o g r a p h y
Optimal Systems Limited - optimal-systems-demon.co.uk

20 - 22 Bedford Row
London
W C l R 4JS
England
Great Britain
Schutte and Koerting - www.s-k.com
2233 State Road
Bensalem, PA 19020
Phone: (215) 639-0900
Fax: (215) 639-1597
402
F I O W
o f ~ n d u s t r i a l~ I u i d s - ~ h e o r y a n d ~ q u a t i o n s
Index
?Index page numbers identzj?ed with an asterix are key citations, and as such,
the reader may wish to consult these pages$rst.
A
Absolute pressure, 45,54*, 87,115,144,
157,165,186
Adiabatic model, 152, 154*, 155, 186,329,
337,348
Adverse gradients, 266
Analytic method, 371,388
As-built sketch, 198
Assumed (hypothetical) data, 53
Asterisk condition, 172
Atmospheric pressure, 43,46,54*, 92,
110,115,143,294,299,304,315,368
Audience, IX
Average velocity, 8,9,11,12,13,18*, 35,
37,39,48,117*, 118,158,159,161,
166,200,210, 254,255,260,261,264,
281,325,329,333,335,351
Axial pumps, 106*, 108,120
B
Balzhiser, Samuels and Eliassen, 345
Bearing losses, 288,289
Bernoulli, XII, l , 2,17,20*, 21*, 36,37,
38,50, 51,53,54,56,57,64,65,71, 82,
89,99,114,115,141, 143, 159, 160,
183,184,189,190, 200,204,205,209,
210,224,229,240,241, 242,245,251,
252,256,257,259,265,267,268,269,
276,281,286,290,291,292, 296,307,
308,309,315,316,317,318,321,322,
374
Bingham plastic, 1l*, 15,22,23,24,27
Blowers, IX, XI, 284
Brake horsepower, 293
Brief History of Pumps, 104
C
Carnot, XII, 221,356
Casings, 120
Cavitation, XI, 87, 102,123,134,137,
141,142,294*, 295,297,299,318
Channel wall, XI
Characteristic curves, 11l*, 124, 131
Chemical reactions, 17
Choked densities, 341
Choked flow, XI, 151,163,179,182*,
184,187,188,193, 202,319,323,324,
326,327,329,332,334,336,342,350,
356,360,362,372,376
Choked Flow and the Mach Number, 182
Choked mass flow, 182*, 324,343
Choked mass flux, 332*, 335,342
Choked pressures, 331*, 339,340,343,
348,358
and Equations
403
I N D E X
Choked temperature, 335,341

Cholung, XI, 151,153,155,179,180,
183*, 186,187,319,324,326,337,362
Churchill-Usagi, 36,39,254,257*, 258*,
271,308,313,315,344
Classification of Pumps, 105
Closed Channel, X
Compressible flow, XIII, 36,37,55,58,
151*, 152,153,154,155,156,157,158,
161,162, 174,175, 192,195,203,204,
241,319,355,362,375
Compressible fluids, XIII, 1
Computer simulations, 156*, 185, 188,
323
Conceptual models, 151
Conservation of energy, 17,18
Conservation of mass, 17
Control considerations, l28
Control systems personnel, X, XII, 53*,
74,299
Cooling water failure, 198,368
Coriolis, XII, 15
Corrected to the pump center line, 114,
115
Corresponding states, 166*,378,384
Critical parameter, 170
Critical zone, 231
Cubic equations-of-state, 166
Customary U.S. System, XI1
D
Darcy, 1,21,36,38,39,209,223,229*,
230,231,232,236,251,252,286,307
Density, 1, 8*, 9, 16, 18,20,21,32,33,38,
39,41,43,44,45,47,48,55,58,65,72,
74,77,81,82,84, 88,91,92,94,95,
101, 102, 109,110, 126,143, 145, 146,
147,152,153,154,157,158,159,160,
161,162,165,166,168,171,177,185,
187,190,192,201,203,204,205,213,
214,218,220,229,230,233,234,235,
252,255,256,257,260,263,283,288,
293,296,297,302,303,308,314,315,
317,321,322,323,324,325,335,337,
341,342,345,357,383
Departure function, 187,320,351*, 352,
370
Describing the Flow Regime, 198,368
Describing the Piping Network, 196*,
207,363
Diaphragm pumps, metering pumps, 125
Didactic simplifications, 2
Differential pressure, 13, 15,45,46,64*,
76*, 78, 102,109, 110, 113,118, 120,
130,134, 137,186,228,230,254,265,
283,290,316,322,323,325,380,382
Diffusers, 101*, 120,266,287
Dilatant fluids, 24,28
Discharge characteristics, 299
Discharge throttling, 102, 137
Dispersion modeling, XI
Division of work, XI1
Dravo Chem Plants, l96
Driving potential, X, 50,65
qndex page numbers identz3ed with an asterix are key citations, and as such,
404
INDEX
E
Efticiency, 56, 102, I l l * , 112*, 116, 118,
119*, 121,127*, 128,129*, 131,133,
223*, 274,281,284,285,286,287,288,
289,290,291,292,293,294,302,304,
305,306,309,310,312,316,317,318
Ejector, 310
Ejectors, 108, 126
Energy,IX,X, 1,2,3,4,5*,7,12,16, 17,
18, 19,20,21,22,30,33,34,35,37,39,
40,43,44,45,48,49,50,51,52,53,54,
55,55-56,57,58,59,60,61,62,63,64,
65,65-66,67,68,69,70,70-71,72,73,
74,75,76,77,78,79,80,81,82,83,84,
85,86,87,88,89,90,91,92,93,94,95,
96,97,98, 100, 101, 102, 103, 104, 109,
110,111, 112,113,114,115,116, 117,
118, 119, 123, 125, 127, 128, 130, 131,
132,133,135,137,139,141,143,144,
145,146,147,148,153,155,158,159,
160, 161, 163, 164,165, 168, 175, 176,
179,180,183,185,191,192,200,201,
202,203,204,205,207,209,210,215,
220,221,222,223,224,229,235,236,
239,240,241,242,243,244,245,251,
253,254,255,256,257,259,260,261,
264,265,266,267,268,269,270,271,
274,275,277,279,280,281,282,283,
284,285,286,287,288,290,291,292,
293,294,295,296,297,299,300,302,
304,305,306,307,309,310,315,316,
317,318,320,321,322,324,325,326,
327,328,329,333,334,335,336,337,
340,345,351,353,354,355,356,371,
372,373,374,375,376,377,378,383,
387,390
Energy transfer by centrifugal force, 101
Energy transfer by electromagnetic

force, 104
Energy transfer by gravity, 101
Energy transfer by mechanical impulse,
104
Energy transfer by momentum transfer,
103
Energy transfer by volumetric
displacement, 102
Engineering companies, XI1
Enthalpy, 19*, 21,155,159,160,161,164,
173, 186, 187, 191, 199,241,243,270,
302,303,319,320,327,328,329,332,
333,336,337,350,351,352,353,354,
355,360,370,371
Estimations of Irreversibilities, 80
Fanning, 35*, 36,50,235,236,258

Fire exposure (by fire zone), 198,368
Flashing, XI, 151,153, 183*, 184, 185,
187,196,199,294,295,319,324,362,
369,371
Flow control, 101, 110*, 112, 120*, 124*,
125*, 135*, 137,149
Flow profile disturbance, 15
Flow profile influences, 15
Flow regimes, XI, 1,6*, 7*, 49,178,207,
390
Fluid flow, IX
Fluid flow theory, IX
Fluid friction, 1,2,3,4*, 12*, 21,371,376
Fluid systems, IX, 105,318,363*
Fluid viscosity, 1,58
Tndex page numbers identi$ed with an asterix are key citations, and as such,
and Equations
405
I N D E X
Force, 3,4,16,20,22,23,25,26,31,32,
35,37,38,40,43,45,50,64,74,81-82,
85,86,100,101,102,104,105,106,
108, 110, 114, 119,121,122,126, 128,
129,143, 148,149, 163,185, 198,209,
210*, 211,212,213,214,215,216,217,
218,219,220,221,224,225,226,227,
228,233,237,238,239,240,245,246,
247,252,256,266,267,274,275,278,
279,280,282,288,289,290,292,294,
297,298,300,301,302,317,332,345,
351,356,377,379,380
Form friction, 17*, 56,62,371,377
Friction factor, 10, 11,30,33*,34,35,36,
38,39,50,58,59,61,62,63,66,67,68,
80,82,85,86,90,91,92,98,152,155,
161, 181,185, 186, 187,188,190,209,
223,229,230,231,232,235,236,253,
257,258,308,309,310,313,322,323,
325,360,377,378,387
Friction loss, 5*, 33*, 37,215
Fundamental Relationships, 17
G
Gas lifts, l26
Gauge pressure, 54
Geometric similarity and dxsirnilarity, 63,
261
Gradual contractions, 78
Gradual enlargements, 78
H
Hagen-Poiseuille, 36,210,230*, 245
Head loss (mechanical energy) formulae,
77
Head losses versus power losses, 284
Head tank, X, 3,101

Hydraulic head, 43
Hydraulic practice, 44*, 46,47,48
Hydraulic radius, 161*, 162,181
Hydraulic Turbines, 141*, 142, 149
Hydrostatic equilibrium, 209,219
1
Ideal gas model, 154,166*, 175,192,337,
378
Impeller,56,101,102,109*, 114,115,
119, 120, 121, 122,140,273,274,275,
277,280,281,282,285,287,288,291,
297
Impellers, 119, 149
Incompressible flow, XIII, l*, 11,37,53,
55,98,151,152,153,192,203,263,
380
Incompressible fluid, 1 , 2
Incremental internal energy, 3*, 4,16,22,
34,49,50
Incremental method, 188*, 326,373
Inherent characteristics, IX, l30*, 134,
148,149
INSIGHT, IX
Installed Characteristics of Pumps, 130
Internal circulation, 288,289
Inward projecting pipe entrance, 79
Irreversibdities, IX, 1,5,7,12,16,17,21,
22,33,34,35,37,38,42,49,50,51,
56*, 57*, 58*, 62,65,66,68,69,70,71,
73,74,75,76,77,78,79,80,84,86,87,
88,89,90,91,92,93,95,96,97,98,
112,118,132,136,137,139,152,153,
158, 159,161,164,175,176,177,179,
Yndex page numbers identijed with an asterix are key citations, and as such,
the reader may wish to consult these pagesjrst.
406
and Equations
I N D E X
180, 181,183, 184,185, 186, 188, 193,

201,209,224,229,230,236,237,240,
241,242,251,253,255,256,257,259,
260,262,263,264,266,267,268,270,
286,289,291,297,303,304,307,310,
315,319,322,325,345,355,356,371
Irreversibility, 4,5, 12*, 16,20,33,37,56*,
94,209,215,236,240,243,244,245,
302
Irreversible, 16*, 18,34, 176,232,240,
241,242,243,244,253,266,300,321,
356,371
Isentropic, adiabatic expansion, 175, 193
Ishwar Davk, 196
Isothermal model, 154*, 155,156,188
Logical inconsistency, 1,2,

Loss of prime, 122, 123
Loss term, hf, 21*, 35,38
Mach number, 152,162,165,181,182,

184,189,319,333*, 334*
Magnetic flow meters, 15
Manifold Flow, 200*, 207,374,390
Manifold flows (dividing and combining
flows), 79
Mature technology, 148,274
Means of energy transfer to liquids, 100
Measurement and control, X, XII, XIII, 1,
51*, 109,123,124,142,299
Measures of performance, 113,318
Jet pumps, 108*, 126,127,128,129,148,

149
K
Kegel, 14
Kelvin, 163*, 223,354
Measuring instruments, X, 13,50

Mechanical energy conversion, 17,68*,
75,375
Mechanical engineers, X, 52
Meter runs, 62
Minimum flow requirement, XI
Miscellaneous fittings and manifolds, 78
Miscellaneous pumping devices, 108,125
L.R. Driskell's Table of Representative

Valve Factors, ISA Handbook of
Control Valves, 75
L/D ratio, 35*, 85,86
Laminar, 1,6,7, 10,12, 13,21,23,30,33,
36,38,49,50,55,58,73,210,222,223,
229,230,231,245,246,248,250,251,
252,258,379,380,382
Language of subject matter, IX
Liquid metal pumps, 108,129
Mist flow, XI
Mixed flow pumps, 106,120
Mixed phase flow, 151
Mixed units, XII, 8*, 10,80,81, 117,210,
251,252,255
Mixing rules, 157*, 167, 171
Mixture,VI, XI, 49,95, 126,131, l56*,
157*, 167,171,183,186,187,199,320,
324,347,350,351,352,353,354,355,
357,360,370,379,385,386
*Index page numbers identified with an asterix are key citations, and as such,
and Equations
407
I N D E X
Moody, 35*, 36,39,50,59,181,185,229,

236,258,308,309,310,344
Motivation, XI1
Multiple choke points with ideal gases,
adiabatic models, 337
N
Negative K factors, 265,271
Newtonian f l ~ d11,23*,
,
25*, 26*
Newton-Raphson iteration, 199,345*,
354,371,376,387
Non-newtonian fluids, 23*, 24*, 26
NPSH, 110,112,113,134,136,139,294*,
295*, 297,314,318
0
Orifice plate, 262
Orifice plates, 15,22,51,76*, 77,98, 177,
202
Original Redlich-Kwong equation, 157
Ostwald and de Waele, 27
Overpressure, 122,123*, 125
P
Peter Paige, 151,183*, 184*, 185*, 187,
188,193,199,319,323,345,350,356,
360,362,371,387
Piezometric head, 43
Pipe taps, 72,76*, 77
Pipe, valves and fittings, 57
Piping system, 3,20,33,34, 116,127, 155,
195*,206,207,262,263,294,298
Piston and cylinder pumps, l07*, 122,
123,140
Polytropic model, 155

Positive displacement pumps, 102*, 103,
104,105,107,121,122,125,131,134,
140,141,148,149
Power failure, 198,368
Power to the shaft, 288*, 289*, 293,304
Prausnitz, 167*, 168,171, 172,320,351,
352,353,357,378,384,385
Pressure, IX, X, XI, 1,2,3,7,13,15, 16,
17, 19,20,21,22,25,26,28,30,32,34,
37,38,43,44,45,46,51,52,54,55,56,
58,64,65,66,69,70,71,72,74,75,76,
77,78,80,82,83,84,85,86,87,88,89,
90,91,92,93,94,95,100,101,102,
103,108, 109,110,112,113,114,115,
116,118, 120,121, 122,123,124, 125,
126, 127,128, 130,131, 132, 133,134,
137,139,140,141,142,143,144,145,
146, 147, 148,151, 152,153,154,156,
157,162,164,165,166, 167,168, 170,
171,172,173,175,177,178,179,180,
182, 183,184, 185,186,187,188,190,
191,195,196,198,199,200,201,202,
203,204,205,206,209,210,212,213,
214,215,216,218,219*, 220*, 221,
222,224,227,228,229,230,231,233,
234,237,240,243,247,248,250,251,
252,253,254,255,256,257,259,260,
263,264,265,266,267,268,269,271,
274,275,282,283,287,288,290,292,
293,294,295,296,297,298,299,300,
301,302,304,306,308,310,311,313,
314,315,316,317,320,322,323,324,
325,326,327,328,329,331,332,335,
336,337,338,339,340,341,343,344,
348,352,353,356,357,358,363,364,
368,369,370,371,373,374,379,380,
382,384,387
?Index page numbers identi$ed with an asterix are key citations, and as such,
408
~ I o wo f ~ n d u s t r i a l ~ I u i d s - ~ h e o r y a n d ~ q u a t i o n s
I N D E X
Pressure and temperature inequalities, 17
Repulsions, 4,166,168
Pressure drop, IV, X, XI, 51, 179, 183*,

185, 188,229,230,231,233,248,250,
251,315,324,325,326,328,376
Restriction orifice, 78,94*, 95,97

Re~nolds,XII, 6*, 7,8,9,10, 11,22,33,
36,39,387
Pressure losses, IX, 69
Rheopectic, 24,29
Pressure relief devices, XI, 151
Rotary analog of Newton's law, 279
Prime, 110*, 122,123,133,134
Prime movers, IX, X, 52

Priming, 110*, 121,134,295,312
Safety relief vent header, 195,363
Process engineers, X, XII, 52*, 196,364

Prony brake, 293
Self-regulation, 135
Shock losses, 120*, 286,287,289
Pseudoplastics, 23*, 24,27
SI, XII, 16, 19,32,35,37*, 40,43,45,46,

48,49,55,56,64,70,81, 113, 117, 118,
159,163,209,210, 211,212,214,224,
229,235,240,250,254,256,263,264,
278,280,291,292,294,295,302,308,
312,314,328,333,351,354,355, 370
Pulsation, 102, 103, 121*, 124*, 125

Pump efficiency, 112,119*, 133,274,288,
302,304,305,309,317
Pumps, IX, XI, 34,51,52,53,98,99*,
100,101,102,103,104,105,106,107,
108,109,110, 111,112, 113,114,115,
119,120,121, 122,123, 124,125,126,
127,128,129,130,131, 132,133,134,
135,138,139,140, 141,142,144,148,
149,158,224,273,274,275,284, 287,
289,290,294,295,298,299,304,305,
307,310,318
Pure fluids, XI, 370
Skm friction, 17*, 35,62,161,371

Slugging, XI, 89
Smooth pipes, 61,231
Specific weight, 213*, 214,220,252,298
Square edged and rounded entrances, 79
Stagnation conhtion, adiabatic
processes, 172
Static discharge head, 116
Purpose of book, IX
Static suction head, 116

Straight pipe of uniform dameter, 57,80
R
Radial pumps, 106,108,119*
Real (as-built) data, 53
Reciprocating pump, 121*, 122,299
Recovery, 51,56,66,69*, 70,72,73,76,
77,85,87,89,93
Reach-Kwong model, 154
Relief header, XI, 188,338
Suction lift, 105, 123*, 132, 133,134,149,

294,295,296,299,304,307,312
Sudden contraction, 63,73*, 74,75,82,
87,91
Sudden expansion, 63,73*, 74,75,84,87,
89,93,175,326,373
Swage, 2,78*, 338
rndex page numbers identz3ed with an asterix are key citations, and as such,
and Equations
409
I N D E X
T
Terminology, IX, 39,69*, 98,200,211,
225,352,377
Thermal energy is not an exact
differential, 160
Utility, 52*, 99, 122, 127, 151, 175,192,

193,236,254,264,297,377
Thixotropic, 24
Vacuum devices, IX
Valve coefficient, C, 253,254
Time dependent behavior, 28
Valves as fittings, 75
Time dependent fluids, 24

Time-independent behavior, 27
Total (system) mechanical energy losses, 79
Total dynamic head, ll0,115*, 116,132,
135, 137, 138,282,283,284,285,290,
291,293
Van der Waals model, 166

Velocity, XI, 1,2, 8*, 9,11,12,13,14,18,
19,22,25,26,33,34,35,36,37,38,39,
43,44,48,49,50,55,58,59,63,64,65,
66,69,70,71,72,73,74,77,81,82,83,
87,88,89,90,91,93,94,101,102,106,
113,114,115,117,118, 130,139,144,
145,146, 147,148, 152, 153,158, 159,
161,162,163,165,166,172,176,177,
179,180,181,182,183,189,190,192,
200,201,203,204,205,210,224,226,
227,228,229,230,231,233,234,237,
245,246,247,248,249,250,251,254,
255,256,257,259,260,261,263,264,
268,269,271,275,276,277,278,279,
280,281,282,283,285,287,292,293,
294,295,296,299,301,307,308,310,
316,321,322,325,326,327,328,329,
331,332,333,334,335,336,342,351,
356,359,361,374,376,379,380,381,
382
Venturi, XII, 74,177,202
Virial equation, 168*, 171, 172, 192
Total energy, mechanical energy and

hydraulic energy grade lines, 80
Total static head, 116
Total suction head, 115*, 146,147,148,
292
Turbine, X, 17,20,21,37,38,53,56,57,
99,104,106,111,119,120,121,132,
133,141*, 142*, 159,180,209,240,
242,259,266,270,279,318
Turbulent, 1,4,6*, 7,10, 12,13,21,22,
23,33,35,36,38,42,49,50,53,55,58,
62,66,68,73,85,92,161,221,223,
229,230,231,251,258,281,308,313,
316,380
Turndown,X, 7,111,113,121,125,139,
142,148*
Two-phase, XI, 186,188
U
Units, XII, 159, 163*, 210,253,359
Unsatisfactory models, 154
Unstable, 1,7*, 49,258, 290
Viscous drag forces, 16,239
W
Walk-down, 198
Weirs and flumes, XI
qndex page numbers identzj?ed with an asterix are key citations, and as such,
the reader may wish to consult these pages _first.
410

Flow of Industrial Fluids Theory and Equatiion

Uploaded by

Copyright:

Available Formats

Flow of Industrial Fluids Theory and Equatiion

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Flow of Industrial Fluids Theory and Equatiion

Uploaded by

Copyright:

Available Formats

Flow of Industrial FluidsTheory and Equations

Flow of Industrial FluidsTheory and Equations

Library of Congress Cataloging-in-PublicationData

Visit the CRC Press Web site at www.crcpress.com

No claim to original U.S. Government works

This book is dedicated to my patient wife, Ginette, to my grandchildren,

To me these children represent the future. W O

Flow of Industrial Fluids-Theory

Scope of Chapter - Basic Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1

Flow of Incompressible Fluids in Conduits

Flow Regimes - Reynolds' Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

Flow Profiles -Velocity Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

Fluid Flow -An "Irreversible" Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16

Fundamental Relationships of Fluid Flow

The Role of Viscosity

Chapter II: lncompressible Fluid Flow

.Losses of Mechanical Energy . . . . . . . . . . . . . . . . . . . . . . . .51

11-1: Scope of Chapter - Applying Basic Concepts

Chapter Ill: Pumps Theory and Equations

Flow of Industrial Fluids-Theory

Chapter IV: Compressible Fluid Flow

Comprehending Compressible Flow . . . . . . . . . . . . . . . . . . . . . . . . . . .151

IV-2: Differences between Compressible and Incompressible Flow . . . . . . . . . . . . . . . . . . . . . . . .152

IV-4: Treating Mixtures

.Complex Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .195

Describing the Piping Network

Describing the Flow Regime

Chapter V: Compressible Fluid Flow

Data Collection andverification

AI-2: S1 and Customary U.S. Units

AI-4: Hydrostatic Equhbrium

Flow o f Industrial Fluids-Theory

Appendix Alll: Computations Involving Pumps for Liquids

AIII-3: Pedormance of Real Centrifugal Pumps

Appendix AIV: Equations of Compressible Flow. Derivations and Applications

AIV-2: Using Thermodynamic Variables .

AIV-4:The Peter Paige Equation. Choked Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .323

Appendix AV: Compressible Fluid Flow

.Complex Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .363

AV-3: Describing the Flow Regime

Flow of Industrial Fluids-Theory

Thixotropic behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29

Generalized viscosity chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -. 3 0

Interpretation of hydraulic "head"

Moody friction factor versus Reynolds' number and pipe schedule

Flow of Industrial Fluids-Theory

Flow of Industrial Fluids-Theory

Absolute roughness of various materials (Data from Miller)

Results of computations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94

About the Author

Instrumentation Symbols and Identijkation.

Flow of Industrial Fluids-Theory

P - 2 : Why Study Fluid Flow i n D e t a i l ?

P-3: Flow i n Closed Channels

Flow of Industrial Fluids-Theory

Flow of Industrial Fluids-Theory

P-4: A Word About Units

Flow of Industrial Fluids-Theory

Flow of Industrial Fluids-Theory