H

i e flow
es

A. T.. Sayers, BSc (E11g), MSc, PhD, C Eng, MIMechE

Departl11el1t ofM echclnical Engineering
University of Cape To"vn

McGray/-Hili Book Company

London New York . St Louis' San Francisco Auckland Bogota
Guatemala Hamburg Lisbon Madrid Mexico Montreal
New Delhi Pan~lIna Paris San Juan Sao Paulo Singapore Sydney
Tokyo Toronto
HYDRAULIC AND COMPRESSIBLE FLOW
TURBOMACHINES
Published by
McGRA W-HILL Book Company (UK) Limited To
Shoppenhangers Road, Maidenhead, HANNAH
Berkshire, England, SL6 2QL.
Telephone Maidenhead (062H) 23432 WILLIAM
Cables MCGRA WHILL MAIDENHEAD Telex 848484
Fax 0628 35895 GILES

British Library Cataloguing in Publication Data

Sayers, A. T. (Anthony Terence), /946-
Hydraulic and compressible now turbomachines.
1. Tu rbomach inery
I. Title
621.8

ISBN 0-07-707219-7

Library of Congress Cataloging-in-Publication Data

Sayers, A. T., 1946-
Hydraulic and compressible now turbomachines / A. T. Sayers.
p. em.
Includes bibliographical references (p
ISBN 0-07-707219-7
1. Turbomachines. I. Title
TJ267.S26 1990 H9-1368H
621.406-dc20 CIP

Copyright 1990 McGraw-Hili Book Company (UK) Limited.

All rights reserved. No part of this publication may be
reproduced, stored in a retrieval system, or transmitted,
in any form or by any means, electronic, mechanical,
photocopying, recording, or otherwise, without the prior
permission of the publisher.

12345 93210

Typeset by Thomson Press (India) Limited, New Delhi

and printed and bound in Great Britain by Page Bros (Norwich) Ltd
 CONTENTS

Preface xi
Symbols xiii

1 Introduction 1
1.1 Definition 1
1.2 Units 4
1.3 Dimensional analysis 4
1.4 Prototype and model efficiency 11
1.5 Dimensionless specific speed 12
1.6 Basic laws and equations 16
Exercises 21
Solutions 23

2 lIydruulic punlps 31
2.1 Introduction 31
2.2 Centrifugal pUlnps 33
2.3 Slip factor 37
2.4 Centrifugal purnp characteristics 40
2.5 Flo\v in the discharge casing 48
2.6 Cavitation in pUlnps 53
2.7 Axial now IHlnlp 56
2.8 PUlnp and systcln 111atching 63
Exercises 68
Solutions 70

vii
viii CONTENTS CONTENTS ix

3 Hydraulic turbines 88 7 Radial now gas turbines 275

3.1 Introduction 88 7.1 Introduction 275
3.2 Pelton wheel 89 7.2 Velocity diagrams and thermodynamics of flow 277
3.3 Radial flow turbine 96 7.3 Spouting velocity 278
3.4 Axial flow turbine 103 7.4 Turbine efficiency 279
3.5 Cavitation in turbines 106 7.5 Dimensionless specific speed 281
Exercises 107 Exercises 282
Solutions 109 Solutions 283

4 Centrifugal compressors and fans 131 References 296

4.1 Introduction 131
4.2 Inlet velocity limitations 138 Index 297
4.3 Pre-whirl and inlet guide vanes 141
4.4 Mach number in the diffuser 141
4.5 Centrifugal compressor characteristic 142
Exercises 146
Solutions 148

5 Axial now compressors and fans 180

5.1 Introduction 180
5.2 Compressor stage 182
5.3 Reaction ratio 187
5.4 Stage loading 188
5.5 Lift and drag coefficients 189
5.6 Blade cascades 191
5.7 Blade efficiency and stage efficiency 199
5.8 Three-dimensional flow 200
5.9 Multi-stage performance . 202
5.10 Axial flow com pressor characteristics 204
Exercises 206
Solutions 208

6 Axial now steam and gas turbines 227

6.1 Introduction 227
6.2 Turbine stage 229
6.3 Stator (nozzle) and rotor losses 232
6.4 Reaction ratio 236
6.5 Effect of reaction ratio on stage efficiency 239
6.6 Blade types 240
6.7 Aerodynamic blade design 244
6.8 Multi-stage gas turbines 245
Exercises 247
Solutions 249
 PREFACE

This book has arisen from a collection of lecture notes compiled for an
undergraduate course in turbo machines. The subject of turbomachines covers
both hydraulic and compressible flow machines, and, while many specialized
books concerning specific types of machines are available, the author was not
able to find a book covering both hydraulic and compressible flow machines at
a suitable level.
The book is aimed at the undergraduate and diploma student, and
introduces ternlS and concepts used in turbolnachinery. It is not an exhaustive
text on the subject, nor is it intended as a design text, although many design
parallleters are given. It should rather be used as an introduction to the more
spccialized texts. Thc book aSSUlllCS that the student has followed a course of
basic fluid mechanics and thermodynamics, but mathelnatical manipulations
are minilnal. Each chapter attcrnpts to be self-contained with regard to a
particular type of machine, and it is suggested that the lecturer supplements
the text with worked exalnples during the lecture and tutorial sessions, when
the maxinlunl understanding of the subject is derived. To aid in this, some
worked exercises are supplicd at the end of each chapter.
In a book of this naturc, great patience is required in the typing and
correcting of thc lnanuscript. In this regard I would like to convey my great
appreciation to Mrs Iris von Bentheim and Mrs Lyn Scott, who jointly
performed this task. For photographic reproduction, I would like to thank Mt
Vernon Appleton, while for rnuch of the draughting of the diagrams, I thank
nly wife, Susan, \vithout whose constant encouragelnent this textbook would
not have been writtcn. Finally, nlY thanks must go to the many students who
have over the years helped to correct the notes and tutorial problems during
their undergraduate studies.
A.T.S.
xi
 SYMBOLS

A area m2
a velocity of sound m/s
b depth or height of blade m
C absolute velocity m/s
Co drag coefficient
COA annulus drag coefficient
CDC tip clearance drag coefficient
CDS secondary loss drag coefficient
CL lift coefficien t
Cp specific heat at constant pressure J/kg K
Cs spollting velocity m/s
Cv nozzle velocity coefficient
c blade chord
D diameter m
drag force N
dH hydraulic diameter m
E power per unit \veight of flow W/{N/s)
e radius of inscribed circle m
F force N
9 gravitational acceleration rnls 2
H head m
h head loss m
specific enthalpy J/kg
blade height rn
I impeller constant J/kg
incidence angle deg
xiii
 SYMBOLS XV
xiv SYMBOLS

L lift force N P blade angle deg

length n1 r circulation m 2 /s
n1 y ratio of specific heats
I span ~-... lJ deviation deg
M Mach number :r
kg e deflection deg
mass
kg/s ( enthalpy loss coefficient
n1 mass flow ra te
stagger angle deg
N nun1ber of stages
rpl11 '1 efficiency
speed (J
net positive suction head 111 camber angle deg
NPSH
dilTIensionless specific speed rev or rad ). work done factor
Ns
dilTIensionless po\ver specific speed rev or rad 11 absolute viscosity Pa s
N sp
\V p density kg/m 3
P power
p (J cavitation paralTIeter
power coefficient
Pa solidity ratio
p pressure
W (Je critical ca vi tat ion paralneter
Q heat transfer to systel11
111
3
js (Js slip factor
VOIUlTIe flow rate
Jjkg flow coefficient
q specific heat transfer
In 3 js ljJ power input factor
leakage now rate
In stage loading coefficient
R radius
111 head coefficient
eye tip radius
Jjkg K w angular velocity radjs
gas constant
Re Reynolds nUlnber
r radius In
eye hub radius 111 Subscripts
SP specific speed radjs
111 a axial
s blade pitch
Jjkg K atm atlTIosphcric
specific entropy
K av average
T temperature
s b blade
tin1e
N nl c casing
torque
ITIjS compressor
U tangential velocity
Jjkg D design concli tion
u internal energy per unit n1ass
n1js d draft tube
V velocity
n1 3 f friction
v volume
Injs fp pipe friction
velocity
W g gate
W \\'ork done per second
Injs H hydraulic
relative velocity
y
h hub
pressure loss coefficient
impeller
Z nUlnber of blades
In inlet
elevation or potential head
leakage
nl at nlean dialTIeter
Greek 111echanical
dcg N nozzle
(J. angle of absolute velocity vector
dcg nom norninal
P' angle of relative velocity vector
 :
~~:

xvi SYMBOLS

0 outlet
overall CHAPTER

p
total (stagnation) conditions
projected area
ONE
R rotor
at tip
INTRODUCTION
r radial
r runner
at hub
reI based on relative velocity
s isentropic
shaft
static
suc suction
tip
turbine
t-s total-to-static
t-t total-to-total
u unit quantities
v volumetric 1.1 \DEFINITION
vap vapour
x tangential A turbomachine can be described as any device that extracts energy from or
imparts energy to a continuously moving stream of fluid, the energy transfer
being carried out by the dynamic action of one or more rotating blade rows.
The dynamic action of the rotating blade rows sets up forces between the blade
row and fluid, while the components of these forces in the direction of blade
motion give rise to the energy transfer between the blades and fluid. By
specifying that the fluid is moving continuously, a distinction is drawn between
the turbomachine and the positive displacement machine. In the latter, the
fluid enters a closed chamber, which is isolated from the inlet and outlet
sections of the machine for a discrete period of time, although work may be
done on or by the fluid during that time. The fluid itself can be a gas or a liquid,
and the only limitations that we shall apply are that gases (or steam) are
considered perfect and that liquids are Newtonian.
The general definition of the turbomachine as used above covers a wide
range of machines, such as ship propellers, windmills, waterwheels, hydraulic
turbines and gas turbines, and is therefore rather loose for the purposes of this
text. We will limit ourselves to a consideration of only those types of
turbomachines in which the rotating member is enclosed in a casing, or
shrouded in such a way that the streamlines cannot diverge to flow around the
edges of the impeller, as would happen in the case of an unshrouded windmill
or aerogenera tor.
The types of machines falling into our defined category and which will be
considered in detail in succeeding chapters are listed in Table 1.1 and fall into
2 HYDRAULIC AND COMPRESSIULE FLO\V TURUOMACIlINES

Table 1.1 Types of turbolnachincs

Turbomachines in which

\Vork is done by lluid \Vork is done Oil lluid

Axial llow hydraulic turbinc Centrifugal pump

Radial llow hydraulic turbinc Axial llow pump
Mixed llow hydraulic turbine Ccntrifugal comprcssor
Axial llow gas turbine Axial llow comprcssor Centrifugal compressor

Pelton wheel hydraulic turbine Radial llow fan

one of t\VO classes depending on whether \vork is done by the fluid on the
rotating member or whether work is done by the rotating 111elnber on the fluid.
Types of turbolllachines can also be defined as to the l11anner of fluid
I1l0Vement through the rotating Inelnber. If the flow is essentially axial with no Kaplan turbine Radial flow fan
radial movement of the strealnlines, then the machine is classed as an axial
flow I1lachine; whereas if the flow is essentially radial, it is classed as a radial
flow or centrifugal machine. Other special types of turbolllachines exist, e.g.
the Minto \vheel or Baki turbine, but they will not be considered in this text.
Considering the two classes of Inachines listed in Table 1.1, some broad
generalizations l11ay be rnade. The first is that the left-hand colulnn consists of
machines in which the fluid pressure or head (in the case of a hydraulic
machine) or the enthalpy (in the case of a cOlnpressible flo\v nlachine)
decreases from inlet to outlet, whereas in the right-hand COIUlllI1 are listed Francis turbine Centrifugal pump
those machines which increase the head or enthalpy of the fluid flowing
through them. This decrease or increase in head, \vhen 111ultiplied by the
weight flow per unit tilne of fluid through the Inachine, represents the energy
absorbed by or extracted frolll the rotating blades, \vhich are fixed onto a shaft.
The energy transfer is effected ill both cases by changing the angular
momentUI11 of the fluid. It might therefore be reasonable to assume that
ditTerent types of turbolnachine would exhibit differing shapes of blades and Axial flow pump
rotating 1l1enlbers, and this indeed is the casc, as is S110\Vn in Fig. i.1. In
addition, because turbomachines have developed historically at different
times, nan1es have been given to certain parts of the l11achincs as well as to Steam turbine
different types of lnachines, and these are no\v dcfined.

Turbine. A nlachine that produces power by expanding a continuously

flowing fluid to a lower pressure or head; the po\ver output is usually Axial flow compressor
expressed in kW.
Figure 1.1 Types and shapes of turhomachines
PUI1Zp. A n1achine that increases the pressurc or head of a flowing liquid, and
which is usually expressed in k Pa or 111.

3
4 HYDRAULIC AND COMPRESSlilLE FLOW TURBOMACHINES INTRODUCTION 5

Fan. A term used for machines imparting only a small pressure rise to a Q p
continuously flowing gas, usually with a density ratio across the machine
of less than 1.05 such that the gas may be considered to be incompressible; I [
pressure increase is usually expressed in mm of water.
Compressor. A machine imparting a large pressure rise to a continuously
flowing gas with a density ratio in excess of 1.05.
1-:- --"----;
I t
1_-_-_-_---,,- - - -1
Flow rate. Q (mJ/s)
Speed, N (rev/s)
I ~ ) N Power, P (W)
Impeller. The rotating member in a centrifugal pump or centrifugal Energy difference
compressor. gIll
I ffilIffillILIITIIl across turhine, fI H (N mlkg)
Fluid density, p(kg/m J )
Runner. The rotating member of a radial flow hydraulic turbine or pump. Fluid viscosity, ,u(Pa s)
Rotor. The rotating member of an axial flow gas or steam turhine; sometimes , I :
Diameter, f) (m)
called a disc.
Diffuser. A passage that increases in cross-sectional area in the direction of t_~ - - -7------ ,- 2-_:
fluid flow and converts kinetic energy into static pressure head; it is
usually situated at the outlet of a compressor.
Control surface !
Draught tube. A difTuser situated at the outlet of a hydraulic turbine. Q
Volute. A spiral passage for the collection of the diffused nuid ora compressor Figure 1.2 Hydraulic turhine control volume
or pump; in the hydraulic turbine the volute serves to increase the velocity
of the fluid before entry to the runner.
volume represents a turbine or diameter D, which develops a shaft power P at a
speed of rotation N, then \ve could say that the power output is a function of all
1.2 UNITS the other varia bles, or
The units used will be those of the Sf system, the basic units being the kilogram, p= f(p,N,jl,D,Q,({j//)) (1.1)
metre and second.
In Eq. (1.1), f means a function of' and g, the acceleration due to gravity, has
4

1.3 DIMENSIONAL ANALYSIS
The large number of variables involved in describing the performance
characteristics of a turbomachine virtually demands the use of dimensionless
analysis to reduce the variables to a number of manageable dimensional
groups. Dimensional analysis also has two other important uses: firstly, the
prediction of a prototype performance from tests conducted on a scale model;
and secondly, the determination of the most suitable type of machine on the
basis of maximum efficiency for a specified range of head, speed and flow rate.
Only a brief description of the method used for forming dimensionless groups
and their application to model testing for turbomachines will be given here, the
generalities of the subject usually being covered in a course of fluid mechanics.
1.3.1 Hydraulic Machines
Figure 1.2 shows a control volume through which an incompressible fluid of
density p flows at a volume flow rate of Q, which is determined by a valve
opening. The head difTerence across the control volume is H, and if the control
be~n co.mbined with /-/ to form the energy per unit mass instead of energy per
unIt weIght. We now assume that Eq. (1.1) may be written as the product of all
the variables raised to a power and a constant, such that
(1.2)
If each variable is expressed in terms of its fundamental dimensions, mass
M, length L and time T, then, for dimensional homogeneity, each side of
Eq. (1.2) must have the same powers of the fundamental dimensions so the
indices of M, Land T can be equated to form a series of simul;aneous
equations. Thus
2
(ML /T 3 ) = const(M/L 3 t(1/T)h(M/LTY(L)d(L 3 /TY(L 2 /T 2 )f (1.3)
and equating the indices we get
M l=a+c
L 2 = - 3a - c + d + 3e + 2f
T -3= -b-c-e-2f
There are six variables and only three equations. It is therefore possible to
solve for three of the indices in terms of the remaining three. Solving for a, b
6 HYDRAULIC AND COMPRESSlBLI: F!.O\V TURIHJMACIIINI:S INTRODUCTION 7

and d in tenns of c, e and I \ve get

(/==I-c
/) == 3- c- e- 2.r
d == 5 - 2c - 3e - 2./ P/f)N:'D~)

Substituting for Q, band d in Eq. (1.2),

P == const[pl-CN3-c-e-2f/l"Ds-2c-3e-2fQt'(OJ-I)f]

and collecting like indices into separate brackets,

P == canst [(pN 3 DS ) {jLI plY D2y( QIN D 3 y(O I I I N 21)2)f] (1.4) (a) (b)
The second term in the brackets will be recognized as the in verse of the
Reynolds number and, since the value of c is unknown, this term can be Figure 1.3 Performance characteristics of hydraulic machines drawn in terms of dimensionless
inverted and Eq. (1.4) may be \vritten as groups: (a) hydraulic turbine; (b) hydraulic pump

PlplV 3D s = const[(pND 2jpY(QIND 3)t!(g}IIN 2D 2)f] (1.5)

characteristics of any other combination of P, N, Q and H for a given machine
Each group of variables in Eq. (1.5) is truly dimensionless and all are used in
or for any other geon1etrically similar machine of different diameter. Since
hydraulic turbomachinery practice. Because of their frequent use, the groups
these groups are dilnensionless, they may be divided or multiplied by
are known by the following nan1es:
themselves to forn1 other dimensionless groups depending on the type of test
PjpN 3 DS == j5 the power coefficient being carried out, and it therefore follows that while in this particular case
QIND ==3
the OO\V cocfficient solutions for Q, band d were found in terms of c, e and f, other solutions could
have been determined which give different dimensionless groups. Each set of
gHIN 2 D2 == l/J the head coefficient
groups taken together is correct, although they will of course be related by
2
The tern1 pN D IlL is equivalent to the Reynolds nun1ber Re == p VDIIL, since the differently shaped curves.
peripheral velocity V is proportional to N D. Hence Eq. (1.1) rnay be rewritten For the turbine, the hydraulic efficiency is defined as
as
Power delivered to runner
tl == ------------------- ----- ---- (1.7)
P == f(Re, , l/J) (1.6) Power available to runner
which states that the power coefficient of a hydraulic 1l1achine is a function of == PlpgQH
Reynolds nUlnber, flow coefficient and head coefficient. I t is not possible to say
what the functional relationship is at this stage, since it n1ust be obtained by Then substituting for P and rearranging gives
experiment on a particular prototype machine or model. In the case of a
tl == P(lv D 3IQ)(N 2D 2IgH)
hydraulic lnachine, it is found that the Reynolds nurnber is usually very high
and therefore the viscous action of the fluid has very little effect on the power == PIl/J (1.8)
output of the machine and the power coefficient ren1ains only a function of t/J For a pump
and . To see how j5 could vary with and l/J, let us rcturn to Fig. 1.2.
To determine the relationship bet\vecn P, l/J and 4), the head across the (1.9)
machine can be fixed, as is usually the case in a hydroelectric installation. For a
fixed value of inlet valve opening, the load on the Illachinc is varied while the
1.3.2 Model Testing
torque, speed and Oow ratc are Ineasured. Froln these lneasurelnents, the
power may be calculated, and P and plotted against l/J. Many hydraulic machines are so large that only a single unit might be
Typical dimensionless characteristic curvcs for a hydraulic turbine and required, as for example a hydraulic turbine in a hydroelectric installation
pUIUp are shown in Figs 1.3a and 1.3b, respectively. These curves are also the producing many megawatts (MW) of power. Therefore, before the full-size
8 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES INTRODUCTION 9

machine is built, it is necessary to test it in model forn1 to obtain as much Pill

T
T
information as possible about its characteristics. So that \ve may accurately
transpose the results obtained from the model to the full-size machine, three
criteria ~ust be met. The first is that the nlodel and prototype must be
geometrically similar; that is, the ratio of all lengths bet ween the model and
prototype must be the sarne. The second requirement is that of kinematic POI

similarity, where the velocities of the nuid particles at corresponding points in 02

02s 01
the model and prototype must be related through a fixed ratio. The third
requirement is that of dynamic similarity, where the forces acting at
corresponding points must be in a fixed ratio between rnodel and prototype. (a) (b)
For a geometrically similar model, dynamic similarity implies kinematic
similari ty.
In order to ensure the above criteria, the values of the dimensionless
groups in Eq. (1.5) must remain the same for both the model and the
L Figure 1.4 Compression and expansion in compressible now machines: (a) turbine; (b)
compressor

p = p/RT and it therefore hecomes supernuous since we already have T and

prototype. Therefore if the curves shown in Fig. 1.3 had been obtained for a P as variables, so deleting density, and cornbining R with T, the functional
completely similar model, these same curves would apply to the full-size relationship can be written as
prototype machine. It can then be seen that these curves apply to any size
{'02 = f{po l ' RTo l ' R T o2 , 111, N, D, II)
machine of the same family at any head, now rate or speed.
and writing P02 as a product of the terms raised to powers,
1.3.3 Compressible Flow Machines P02 = const [(PO 1Y'{RTo1 t(RTo2 r(nl)d(Nt(D)f(jl)gJ (1.11)

Not all turbomachines use a liquid (hydraulic nuid) as their nuid medium. Gas Putting in the basic dimensions
turbines and axial now compressors are used extcnsivcly in the jet engines of
(M/L T 2 ) = consl [(MIL T 2 y'(L 2 /T 2 )h(L 2 /T 2 Y-(M/T)d( l/Tt(L)f{M/L T)g]
aircraft where the products of combustion and air respectively are the working
fluids, while many diesel engines use centrifugal corn pressors for supercharg- Equating indices
ing. To accommodate the compressibility of these types of nuids (gases), some
M 1 =a+d+u
new variables must be added to those already nlcntioned in the case of
hydraulic machines, and changes must be made in some of the definitions used. L - 1 = - a + 21> + 2c + j' - 0
With compressible now machines, the parameters of importance are the T - 2 = - 2a - 21> - 2c - d - e - g
pressure and tempcrature increase of the gas in a compressor and the pressure
and solving for a, hand .1' in terms of d, C, l! and g we obtain
and temperature decrease of the gas in the turbine plotted as a function of the
mass flow rate of the gas. In Fig. 1.4, the T -s charts for a compression and a=l-d-g
expansion process arc shown. h = d/2 - c - e/2 + 0/2
In isentropic flow the outlet conditions of the gas are at 02s whereas the
actual outlet conditions arc at 02. The subscript 0 refers to total conditions and
f = e - 2d - g
1 and 2 refer to the inlet and outlet points of the gas respectively. The s refers to Substitute for G, band f in Eq. (1.11), then
constant entropy. P02 = const [P6t d- 9{R To 1 )d /2 -c - el2 + g/2(R T o2 Ynl dN eDe - 2d - g jlgJ
Now the pressure at the outlet, P02' can be written as a function of the
following variables: = const X POI {(RTo2/RToIY[m(RTol)1/2/POID2Jd[ND/(RTol)1/2Je
X [,u(RToI) 1/2/PO I DJg} (1.12)
(1.10)
Now if the last term in the brackets in Eq. (1.12) is multiplied top and bottom
Here the pressure ratio P02/POI replaces the head H in the hydraulic machine,
by (RT~I)1/2 and noting that POl/RTol equals POI' then
while the mass flow rate m (kg/s) replaces Q. However, by examining Eq. (1.10)
we can see that, using the equation of state, the density may be written as jlRTol/POl {RTo1 )1/2D = jL/(RTol)1/2pOlD
10 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACIIINLS INTRODUCTION 11

But the units of (R To 1 )1 /2 are LIT, \vhich is a veloci ty, and therefore the last
term in brackets is expressible as a Reynolds nUlnber. T'hus the functional
relationship may be written as
P02/PO 1 == f((R T 02 / R To 1)' (lll(R To 1) 1/2 /po 1D2 ), (N 1)1(/~ To 1)1/2), He) (1.13)
The exact form of the function (1.13) must be obtained by experimental
measurements taken from model or prototype tests. For a particular nlachine
using a particular fluid, or for a I110del using the SaI11e fluid as the prototype, R
take place, a significant reduction in Re can occur, and this must then be taken
into account. For a particular constant-dianleter Inachine, the diameter D may
be ignored and therefore, in view of the above considerations, function (1.13)
becomes
(1.14)
where it should be noted that some of the terms are now no longer
dimensionless. It is usual to plot P02/PO 1 and T 02 /To 1 against the mass flow
rate parameter I1l rb il2 /POI for different values of the speed parameter N/T6 1/
r

is a constant and may be eliminated. The Reynolds nUlllbcr is in Inost cases so

high and the flow so turbulent that changes in this panuneter over the usual
operating range may be neglected. However, where large changes of density
for a particular machine. But for a family of machines, the full dimensionless
groups of Eq. (1.13) 1l1Ust be used if it is required to change the size of the
machine or the gas contained. The term ND/(RT01 )1/2 can be interpreted as
the Mach-number effect. This is because the impeller velocity V ex N D and the
acoustic velocity a OI ex (RTo d l/2 , while the Mach number M == V/a Ol ' Typical
performance curvcs for an axial flow compressor and turbine are shown in
Figs 1.5 and 1.6.

surgelin0
O.9~ 1.4 PROTOTYPE AND MODEL EFFICIENCY

~ I
Increasing N / T~i2
Before leaving this introduction to the use of dimensionless groups, let us look
at the relationship between the efficiency of the model and that of the

L------'
prototype, assunling that thc similarity laws are satisfied.
We wish to build a Il10del of a prototype hydraulic turbine of efficiency YIp'
Now froln siInilarity laws, denoting the model and prototype by subscripts m
and p respcctively,
(a) (b)
H p/(N pD p )2 == fl m /(N m DnJ2 or Hp/H m== (N p /N m )2(D p /D m )2
Figure 1.5 Axial flow compressor characteristics: (a) pressure ratio; (b) efficiency
Qp/NpD~ == QIll/NmDI~l or Qp/Qm == (N p /N m )(D p /D m )3
Pp/N~D~ == Pm/NI~lD~l or Pm/P p == (N m /N p )3(D m /D p )5

Now
\ Choking Choking
mass flow ' fT"' Power transferred from fluid
T ur b Inc e IClcncy == - - - - - - - - - - -

~I
J . , / Fluid power available
== P/pgQH (1.15)
; \ IncreaslI1g :I
I !
~ N /T~i2 Therefore
:I
\~
:I
:/ 11m/ l 1p == (P mI P p)(Qp/Qm)(H p/ H m) == 1

r and the efficiencies of the model and prototype are the same providing the
(b)
similarity laws are satisfied. In practice, the two are not the same due to scaling
(a) eITects, such as relative surface roughness, slight Reynolds-number changes
Figure 1.6 Axial flow gas turbine characteristics: (a) pressure ratio; (b) efficiency and Mach-number effects at higher blade speeds.
12 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES
INTRODUCTION 13

1.5 DIMENSIONLESS SPECIFIC SPEED If the diameter is elinlinated from these t\vo eq uations, then

We have seen in Sec. 1.3 that the curves showing the functional relationship D = {Ull/l/If))II2/N and (/)0 = QN 2 {t// O /[]H)3 /2
between dimensionless groups for a particular machine also apply to machines or
of the same family (similar design), providing the similarity laws are obeyed ( 1.16)
when changing to a smaller-diameter machine, at perhaps a difTerent speed
and head. It is therefore possible to obtain curves of many difTerent types of
machines, and to use these curves to select a machine design for a particular
I
I
1.
0
1
operating requirement. Typical curves that might be obtained for difTerent I

types of hydraulic pumps are shown in Fig. 1.7, where it is seen that each i 1-
machine type lies in a well-defined region of head and flow coefficients, it being
possible in some cases to choose two or more impeller types for a specific flow
coefficient. There are of course an infinite number of designs that could be
produced, but for each design only one point exists on its characteristic curve
I 0.91- > 0.63 O1
3
/s

where the efficiency is at a maximum. Thus for each design of pump unique
values of and t/J exist at the maximum efficiency point. In the case of turbines, I 0.8
the unique values would be P and at maximum efficiency.
The specifications for a pump design are usually expressed in terms of a

0'7~
~
required head H, at a flow rate of Q and speed N, the speed being specified since ;;..
u
motors are usually only available in fixed speed intervals. No mention has c
0
been made concerning the diameter or type of machine, both of which must be 'u
H:
determined. For the best design point, constant values of (Po and t/JD will exist w I
corresponding to the maximum efficiency point, or
and t/J 0 = ~J Ii/ N 2 D 2
0.6

Radial
0.5

0.4 '--'--_--'--_L-I- - l . . 1 - J - - ' - - - - - - - - L - - _ - - - L - _ - - L - - - - L - - . J _ . . L - . . . L - _ L J

0.2 0.3 0.4 0.5 2 4
Axial Dimensionless specifIc speed, N,! (rad)

Centrifugal Mixed flow Axial

Fi~ure 1.7 Characteristic curves for various pump designs Figure 1.8 Variation of hydraulic pump impeller design
14 HYDRAULIC AND COMPRESSIBLE FLOW' TURBO~tA(,lllj\;FS
INTRODUCTION 15

N s is known as the dimensionless specific speed, the units being revolution or Putting H 2 == 1 (unit head) then
radians depending on the units of IV, and I11Ust nol be confused \vith specific
speed. N 2 == NIl J-/ ~ /2 == N u (1.17)
Since D was eliminated at the InaxiiTIul11 efficiency point, the dil11ension- and this is the unit speed of the turbine. Unit quantities for Q and P may be
less specific speed acts as a design paralneter, which indicates the type of similarly obtained to give
machine that should be used for the given N, Hand Q. Equation (1.16) shows
that a pun1p with a high N s will have a low head and high flow rate, and implies (1.18)
and
an axial flow pUlnp \vith a large s\vallo\ving capaci ty. 1\ hnv IV s ilnplies a high
head and low flow rate, and a centrifugal type of pUlllp. Figure I.g shows the (1.19)
variation of N s with pUlnp in1peller type, and indicates the optil11UIn For a turbine, the din1ensionless specific speed is found by a procedure
efficiencies to be ex pected. similar to that for pUlnps except that D is eliminated from P and t/J to yield
In practice, N s is often expressed as NQ l 12/H 3 /-+, the y being dropped since what is often referred to as the power specific speed, N sp , where
it is a constant, and the resulting value of N s will therefore be ditTerent. It may
also be found that consistent sets of units are not always llsed for N, Q and H, N sp == N P l /2I p l/2(gH)5/4 (1.20)
so that when a value of IV s is expressed, it should be ensured that the definition Figure 1.9 shows typical hydraulic turbine runner shapes for difTerent specific
being used is kno\vn. In this text the SI systelll \vill be llsed and N s will therefore speeds along with their optin1um or design efficiencies.
be dimensionless. A wide range of rotor designs from low to high values of specific speed for
However, as a point of reference, conversion factors arc listed in Table 1.2 both hydraulic and cOInpressible flow machines are shown in Figs 1.10 and
so that the reader n1ay calculate the din1ensionless specific speed froln specific 1.11, where it will be noted that low-specific-speed machines have large
speeds using Q, IV and H in other units. The fluid contained is water and, where diameters and high-specific-speed machines have small diameters. In general,
quoted, gpIn are US gallons per 111inute, ft is foot, cfs arc cubic feet per second,
and hp is horsepower.
Tenns that are often used in hydraulic OO\V lnachines are those of unit
head, unit speed, unit po\ver and unit quantity. They arose fr0l11 the need to be
able to COInpare hydraulic nlachines tested under a set of standard conditions.
~ ~ !7::J
0.6 1.1 1.6
In turbine work, the speed, power output and flow rate are detennined for a
turbine operating under an assul11ed unit head of say I In or 1ft, its efficiency
0.98
renlaining constant. For instance, consider a turbine tested under a head HI
and speed N 1 rpn1. Then fron1 Eg. (1.6), for any other speed and head, Pelton wheel

fIlllvi == H 21lv ~ 0.94 I

or
. \ .

~
~
FranCIS turbmes
\
Axial flow turbines
g 0.90
v
T>
~
(J.J

Table 1.2 Conversion factors for specific speed

Specific speed Dimensionless specific speed. N s (rad)

SP\ = rpm(cfs)\j2/ftJ!~ Ns = SI\/129 0.82

SP 1 = rpm(mJ/s)\l/mJ~ N~=SP2/5J
SP 3 = rpm(gpm)\!2/ftJ~ N~ = SP 3/2730
SP~ = rpm(hp)\/2/fl~:~ Ns = SP ~/42 o 2 3 4
SP s = rpm(m~tric hp)\/2/1ll~,4 Ns = Sl\/I X7 Dimensionless specific speed. N. (rad)
Figure 1.9 Variation of hydraulic turbine runner design with dimensionless specific speed
 16 HYDRAULIC AND COMPRESSInLE FLOW TURnOMACHINES

Ns (rad)
0.05 0.10 0.20 0.50 1.0 2.0 5.0 10.0 20.0

Pelton wheel single jet II Francis turbines Kaplan

I
..I - - - - - - - - - - + i - I I.. -I
slow normal fast turbines
Pelton wheel PI~ope~l~r
I- -I
multi-jet turbines
Centrifugal pumps Mixed flow Propeller
I- 1' .1 1 -I
(radial flow) pumps pumps (I)
Radial compressors
II 'I
and fans Axial flow compressors, blowers

Axial flow and ventilators

'I _\
steam and gas turbines

Figure 1.10 Correlation or rotor designs with dimensionless specific speed (collrtesy (~r Escher
Wyss Ltd)

(4) (5) (6)

the smaller the diameter the lower will be the cost orthe rnachine, and therefore
the design usually aims for the highest possible specific speed.

1.6 BASIC LAWS AND EQUATIONS

The basic laws of thermodynamics and fluid mechanics are used in turboma-
chines although they are usually arranged into a more convenient form. All or
some may be used under any set of circumstances and each will be brieny dealt (8) (9)
with in turn.

1.6.1 Continuity
For steady flow through the control volume, the mass now rate m remains
constant. Referring to Fig. 1.12,
(1.21)
where the velocity vectors C 1 and C 2 are perpendicular to the cross-sectional (10) (11 ) (12)
areas of flow A 1 and A 2' In compressible now nlachines the mass now (kg/s) N,. Nfl,
is used almost exclusively while in hydraulic machines the volume now rate Q (I) Pelton wheel 0.05 (7' Steam turbine 0.54
(m 3 Is) is preferred. (2) Pelton wheel 0.13 (8) Steam turbine 1.07
(3)- Francis turbine 0.38 (9) Centrifugal pump 1.07
(4) Steam turbine 0.4 (10) Gas turbine 1.18
1.6.2 Steady Flow Energy Equation (First Law of Thermodynamics) (5) Centrifugal pump 0.54 (t I) Radial compressor 1.34
For steady flow through a system control volume, where the heat transfer rate (6) Radial compressor 0.54 (t 2) Axial compressor 1.6
to the system from the surroundings is Q and the work done by the system on Figure 1. t t Some rotor designs and their dimensionless specific speeds

17
 INTRODUCTION 19

YL x

(13) (14)

Control volume
/
Figure 1.12 Control volume for linear momentum

the surroundings is W, then

(16) (17) (18) Q - W = In[(P2jP2 - PI/PI) + (C~ - Ci)j2 + g(Z2 - Z 1) + (U 2 - u1)J (1.22)

where pjp = pressure energy per unit mass (Jjkg), C 2 /2 = kinetic energy per
unit mass (Jjkg), u == internal energy of the fluid per unit mass (J/kg), gZ
= potential energy per unit mass (J/kg), m = mass flow rate (kgjs), W = work
done on surroundings ( + ve) (W) and Q = heat transfer to system ( + ve) (W).
In words, Eq. (1.22) states that in steady flow through any region:
Heat added to Shaft work done Increase in Increase in
fluid per unit - by the fluid per = pressure energy + kinetic energy
(21)
mass unit mass per unit mass per unit mass
(19) (20)
Increase in Increase in
+ potential energy + internal energy
per unit mass per unit mass
The steady flow energy equation applies to liquids, gases and vapours as
well as to real fluids having no viscosity. It may be simplified in many cases
because many of the tcrms are zero or cancel with others, and this will be
(22) (23) shown in the rclevant scctions.

Ns N8
(13) Steam turbine 1.6 ( I tJ) Axial compressor 3.21 1.6.3 Newton's Second Law of Motion
(14) Francis turbine 1.72 (20) Propeller pump 3.21
This law states that the SUIll of all the forces acting on a control volume in a
(15) Francis turbine 2.14 (21) Axial blower 4.82
particular direction is cq ual to the rate of change of In0I11entum of the fluid
(16) Mixed flow pump 2.14 (22) Propeller pump 5.36
2.41 (23) Kaplan turbine 5.36
across the control YOIUI11C. With reference to Fig. 1.12, Newton's second law
(17) Kaplan turbine
2.41
may be written as
(18) Axial cOInpressor
Figure 1.11 (collld) (1.23)

18
20 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES INTRODUCTION 21

causes the power developed by a turbine to be less than the ideal isentropic
power developed and why the work input to a pump is greater than the
isentropic or ideal work input (Fig. 1.4). I n theory the entropy change might
also be zero for an adiabatic process b~t it is impossible in practice. For a
reversible process the second law is expressed as
dqjT = f1s (1.26)
---.-- where dq = heat transfer per unit mass (Jjkg), T = absolute temperature at
which heat transfer occurs (K) and f1s = entropy change (Jjkg K).
In the a bsence of motion, gravity and any other effects, Eq. (1.22) has no
potential or kinetic energy terms, and so
or dq - dw = du
where the units are Jjkg. Substituting for dq and rearranging,
Figure 1.13 Control volume for angular momentum
f1s = dqjT = (dll + dw)jT
Equation (1.23) applies for linear momentum. However, turbomachines Putting dw = p dv, where dv is an infinitesimal specific volume change, then
have impellers that rotate, and the power output is expressed as the product of
torque and angular velocity, and therefore angular momentum is the more
T ds = dll + Pdv (1.27)
useful parameter. Defining specific enthalpy as It = II + pv and substituting for du in Eq. (1.27),
Figure 1.13 shows the movement of a fluid particle from a point A to a
T ds = dh - v dp ( 1.28)
point B while at the same time moving from a radius r 1 to '2' If the tangential
velocities of the fluid are C x1 and C x2 respectively, then the sum of all the and this equation IS used extensively in the study of compressible flow
torques acting on the system is equal to the rate of change of angular machines.
momentum, Tn the following chapters, lise will be nlade of the concepts discussed in this
introduction. This chapter should have acted as a reminder of the many
IT = m('2 C x2 -'1 C x .) separate concepts learned in thermodynamics and fluid mechanics, and has
If the machine revolves with angular velocity (JJ then the power is shown how these two separate subjects combine to form the subject of
turbomach inery.
ITw = m(U 2 C x2 - U l C x.)
For a turbine
W=m(U 1 C x1 - U 2 C x2 O (1.24) EXERCISES
and is known as Euler's turbine equation. 1.1 A radial now hydraulic turbine is required to be designed to produce 30 MW under a head of
For a pump 14 m at a speed of95 rpm. A geometrically similar model with an output of 40 kW and a head of
5 m is to be tested under dynamically similar conditions. At what speed must the model be run,
W = m(U 2Cx2 - U 1 C x .) > 0 (1.25)
what is the required impeller diameter ratio between the model and prototype and what is the
which is Euler's pump equation. volume now rate through the model if its efficiency can be assumed to be 90 per cent?
1.2 The performance curves of a centrifugal pump are shown in Fig. 1.14. The impeller diameter is
127 mm and the pump delivers 2.831/s at a speed of 2000 rpm. If a 102 mm diameter impeller is
1.6.4 Entropy (Second Law of Thermodynamics) fitted and the pump runs at a speed of 2200 rpm, what is the volume now rate? Determine also the
new pump head.
This law states that, for a fluid undergoing a reversible adiabatic process, the 1.3 An axial now compressor is designed to run at 4500 rpm when ambient atmospheric
entropy change is zero, while for the same fluid undergoing an adiabatic or conditions are 101.3 kPa and 15"C. On the day when the performance characteristic is obtained,
other process, the entropy increases from inlet to outlet. It is this fact that the atmospheric temperature is 2Y'C. What is the correct speed at which the compressor must run?
22 HYDRAULIC AND COMPRESSI13LE FLO\V TURBOMACIIINES INTRODUCTION 23

turbines with a specific speed of 180 rpm are investigated. The normal running speed is to be
50 rpm in both schemes. Determine the dimensionless specific speeds and compare the two
20 proposals insofar as the number of machines are concerned, and estimate the power to be
developed by each machine. The units in either installation are to be of equal power and the
efficiency of each type may be assumed to be 0.9.
1.9 A customer approaches a salesman with a particular pump requirement and is quoted for an
16
axial now pump of rotor diameter 152.4 mm. Running at a speed of980 rpm, the machine is said to
deliver 0.283 m 3 /s of water against a head of9.1 m at an efficiency of 85 per cent. Are the claims of
the salesman realistic?
12 1.10 A Francis turbine runs at I HO rpm under a head of 146 m with an efficiency of 93.5 per cent.
Estimate the power output of the installation.

SOLUTIONS
4

Exercise 1.1 Equating head, flow and power coefficients for the model and
OL.-.-_ _L.-.-_--I._ _ --.L._ _--.L._ _----L-_ _--.l...-_ _----L-_ _-.J prototype and noting that the density of the fluid remains the same, then, if
o 1.0 2.0 3.0 4.0 subscript 1 refers to the prototype and subscript 2 to the model,
Volume now rate, Q x 10 3 (01 3 /s)
where Pl = P2
Figure 1.14
Then
If an entry pressure of 60 k Pa is obtained at the point where the normal ambient condition mass
now would be 65 kg/s, calculate the mass now obtained in the test.
1.4 Specifications for an axial now coolant pump for one loop of a pressurized water nuclear
reactor are:
Head 85 m
Flow rate 20000 mJ/h
Speed 1490 rpm
Diameter 1200 mm
Water density 714 kg/nl"
Power 4 MW (electrical) Also
The manufacturer plans to build a model. Test conditions limit the available electric power to
500 kW and flow to 0.5 nl J/s of cold water. If the model and prototype elliciencies arc 'assumed
equal, find the head, speed and scale ratio of the model. Calculate the dimensionless specific speed
of the prototype and confirm that it is iuentical with the model. Then
1.5 A pump with an available driven speed of 800 rpm is requireu to overcome a IJD m head
while pumping 0.2 m J Is. What type of pump is required and what power is required?
1.6 A reservoir has a head of 40 manu a channel leading from the reservoir permits a now rate of
~: =(:: )1/2(Z:)= C4Y/2(ZJ
5

34 m 3 /s. If the rotational speed of the rotor is 150 rpm, what is the most suitable type of turbine to Therefore equating the diulnetcr ratios
use?
1.7 A large centrifugal pump contains liquid whose kinematic viscosity may vary between 3 and 6
(
0.266 -Nl).\i~ -_( -5 )1/2(N
- 1)
times that of water. The dimensionless specific speed of the pump is 0.1 HJ rev and it is to discharge N2 14 N2
2 m 3 /s of liquid against a total head of 15 Ill. Determine the range of speeds and test heads for a
one-quarter scale model investigation of the full size pump if the model uses water.
or
1.8 In a projected low-head hydroelectric scheme, 10000 ft J/s of water arc a vailable under a head
N 2)2/5
of 12ft. Alternative schemes to usc Francis turbines having a specific speed of 105 rpm or Kaplan ( N~ = 2.25
24 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES INTRODUCTION 25

whence and diameter D are dropped to yield Eq. (1.14). Considering first the speed
N 2 == 2.25 5 / 2 x 95 parameter,
Model speed == 721.4 rpm lV l N2
JT Ol JT 02

D 2 == 0.266 (~)3/5 273 + 25 )1 /2

D1 721.4 N 2 == 4500 (
273 + 15
Model scale ratio == 0.079
Correct speed == 4577 rpm
. Power output
M o d eI efliIClency == - - - - - - - Considering no\v the mass now parameter,
Water power input
40 x 10 3
0.9==---
pgQH
40 x 10 3 n1 2 == 65 x ( -60- )(288)1/2
--
Q==0.9 X 10 3 x 9.81 x 5 101.3 298
Mass now obtained == 37.85 kg/s
Model volume flow == 0.906 m 3Is

Exercise 1.4 Using Eq. (1.5), equate the head po\ver and now coefficien ts for
Exercise 1.2 Assuming dynamic similarity exists between the first and second the model and prototype. Then
sized pumps, we equate the flow coefficients. Thus

or
2.83 N 1 ( 20 000 ) (D 2 ) 3

2000 X 127 3 2200 x 102 3 IV 2 == 0.5 x 3600 D;

Solving we get
=11.11(~:r
Q2 == 1.61 ljs
Also
From Fig. 1.14 at Q1 == 2.83ljs (2.83 x 10- 3 m 3 Is) and 2000 rpm the head H 1 is ~=(Nl)3(~.)5(~)
14 m and equating head coefficients for both cases gives P2 N2 D2 P2

Substitute for (N II N 2); then

and substituting
4
0.5 =(11.11)3 (DD: )9(DD: )5( 1000
714)

9.81 x 14 9.81 xH 2
(2000 x 127)2 (2200 x 102)2 D 2)4 8
( D; =(11.11)J x O.714
Solving we get
Scale ratio D2 1D 1 = 0.3
H 2 == 10.9 m of water
Then
N l/lV 2 = 11.11 X (0.3)3
Exercise 1.3 The dimensionless groups of Eq. (1.13) may be used here but
Speed ratio N 2/ N 1== 3.3
since the same machine is being considered in both cases the gas constant R
26 HYDRAULIC AND COMPRESSI13LE FLOW TUR130MACIiINES INTRODUCTION 27

Also supplied to the shaft we divide by the efficiency,

Shaft power required = 3.59/0.80
Shaft power = 4.49 kW
Exercise 1.6 We have
Turbine power = pgQli
Head ratio H 2 /H I = 1.0 =1000x9.81 x34x40
= 13.3 MW
From Eq. (1.16) the dinlcnsionlcss spccific specd is given by
Power specific speed is given by Eq. (1.20)
NQI/2
N=-- Np 1 / 2
S (gH)3/4 N =----
sp P1/2(gH)5/4
For the prototype
150 x (13.3 x 10 6 )1/2
- = 2rr x 1490 x (20000)1/2(~_)3/4(~)3/4 60 X (1000)1/2 x (9.81 X 40)5/4
N sl 60 3600 9.81 85
= 0.165 rev (1.037 rad)
= 2.37 rad
From Fig. 1.10 it is seen that the Francis turbine would bc the most suitable
_.-._---------
1490 (0.5)1/2 choice for this application.
N s2 = 2rr x 60 x 3.3 x (9.81 x 85)3/4
Exercise 1.7 Since the viscosity of the liquids used in the model and prototype
= 2.35 rad
vary significantly, equality of Reynolds nUI11ber in Eq. (1.5) n1ust apply for
dynamic similarity. Let subscripts 1 and 2 apply to the prototype and model
Therefore taking rounding errors into account ~he __~ii_nlensionlcss specific respectively.
speeds of both model and prototype are thc sanlC. Equating Reynolds nUI11ber
N 1Di N2D~
Exercise 1.5 Froln Eq. (1.16)
V1 V2
NQ1/2
N
S
= ----------
(yH)J/4 !!_~ = ~ (~)2
N1 V1 D2
800 (0.2) 1/2
=-x - For the liquid with viscosity three tinlcS that of watcr
60 (9.81 X 1.83)J/4
N 2 42
= 0.683 rev (4.29 rad) - =-= 5.333
N1 3
For the given flow rate Fig. 1.8 shows that a propcller or axial flow pump is Equating flow cocfficicnts
required and that an efficiency of about 80 per ccnt can be expectcd. Therefore
the power required is
P = pyQli
= 1000 x 9.8 1 x 0.2 x I. 83
Q2=N 2 (D )J 2

Ql N I D1
= 3.59 kW
5.333
This is the power delivered to the water and to get the power that must be =43=0.0833
28 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES INTRODUCTION 29

Equating head coefficients Dimensionless specific speed of Francis turbine

= 105/42
= 2.5 rad

Dimensionless specific speed of Kaplan turbine

= 180/42
= 4.3 rad

From Eq. (1.16) These values may be checked against those values in Fig. 1.10.
N 1Qt /2 Converting to SI units
N s1 = (gH d3/4 10 000 ft3 Is = 283.17 m 3Is
0.183 (9.81 x 15)3/4 12ft=3.66m
N1 = 21 / 2
Power deli vered
= 5.47 revls Turbine efficiency = '1 bl
Power aval a e
N 2 = 5.47 x 5.33
P
Model speed = 29.16 revls 0.9 = pgQH

Q2 = 2 x 0.0833 P = 0.9 x 1000 x 283.17 x 3.66 x 9.81

Model flow rate = 0.166 m 3 = 9150kW
H 2 = 15 x 1.776 This is the total power delivered by all the turbines.
Model head = 26.67 m Now

Similarly for the case when the prototype viscosity is six times that of where N is in revls
water
For the Francis turbine
IN 2 = 14.58 revls
50 x 2 x 1t X p 1/ 2
,H 2 = 6.67m 25- - - - - - - - - - -
. - 60 x (1000)1/2(9.81 X 3.66)5/4

For one-quarter scale model whence

14.58 < model speed < 29.16 rev Is P=1761kW
6.67 m < model head < 26.67 m . . Total power required
Number of FranCIS turbInes = - - - - - - - -
Power per machine

Exercise 1.8 The dimensionless specific speed is obtained from the conversion
9150
factors for specific speed given in Sec. 1.5. In this case for the non-SI units used 1761

N sp = SPI42 = 5.19 (say 6 machines)

30 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES

For the Kaplan turbine

CHAPTER
50 x 2 x n x pl/2

4.3 == 60 x (1000)1/2(9.81 X 3.66)5/4 TWO

whence HYDRAULIC PUMPS
P == 5209.7 kW
9150
NUInber of Kaplan turbines == 5209.7

== 1.76 (say 2)

Exercise 1.9 From Eq. (1.16) for dilllensioniess specific speed

NQ1/2
N s == --3-'4 rev
(y 11) I

2 x IT x 980 X 0.283 1/ 2 "

-----~-rau
60 x (9.81 x 9.1 )3/4
2.1 INTRODUCTION
== 1.88 rad
The term 'hydraulics' is defined as the science of the conveyance of liquids
Referring to Fig. 1.10 it is seen that axial flow pUln~s only begin .at a through pipes. Most of the theory applicable to hydraulic pumps has been
dimensionless specific spced of approximately 2.0 rad. 1t IS therefore unlIkely derived using water as the fluid medium but this by no means precludes the use
that the salesman's clainls are realistic. of other liquids. Two types of pUlnps commonly used arc centrifugal and axial
A suitable pUlnp would be of the tllixed flow type which gives the stated flow types, so named because of the general nature of the fluid flow through the
efficiency at the required flow rate and calculated ditllcnsiol1less spccific speed. impeller. Both work on the principle that the energy of the liquid is increased
by imparting tangcntial acceleration to it as it flows through the pump. This
Exercise 1.10 Using Fig. 1.9 the Francis turbine has an efficiency of 93.5 per energy is supplied by the ilnpeller, which in turn is driven by an electric motor
cent at a dimensionless specific speed of 2.0 rad. Fron1 Eq. (1.20) the or SOlne other drive. In order to ilnpart tangential acceleration to the liquid,
dimensionless power specific speed is rows of curved vanes or blades move transversely through it and the liquid is
Npl/2 pushed sideways as it tnoves over the vanes as well as retaining its original
N ==---- forward component of velocity. Figure 1.1 showed typical centrifugal and
sp pi /2 (y I-I)5/4
axial flow pun1p itnpellcrs, while between these two extremes lie mixed flow
whence pumps, which are a cOlnbination of centrifugal and axial flow pumps, part of
, 2.0 x (1000) 1/2 x (9.81
X 146)5/4 x 60 the liq uid flow in the illl peller being axial and part radial.
p1 / 2 == --~------------ The centrifugal and axial flow pumps will be dealt with in turn in the
180 x 2 x n
following sections. However, before considering the operation of each type in
== 29 563 detail, we will look at a general pumping system, which is common to both
types. This is shown in Fig. 2.1 where a pump (either axial or centrifugal)
and pumps liq uid from a low to a high reservoir.
P==874MW At any point in the system, the pressure, elevation and velocity can be
expressed in tenns of a total head measured from a datum line. For the lower
reservoir the total head at the free surface is H A and is equal to the elevation of
31
32 HYDRAULIC AND COMPRESSIBLE FLOW TlJRBOMACflINI:S HYDRAULIC PUMPS 33

It,o I then

r- +
4-
D

F
Pump total inlet head = p)Prl + Vt /2q + Zj
= Po / f> Y + V l~ /2 r! +Z
--r-----~~
I P um p tot alOll tic thead 0
I
,I Total head developed by pump = [(Po - Pi)/ puJ + [( V~ - V~)/2gJ + (Zo - Zi)
= H (2.1)
Total energy line

~ This is the head that would be llsed in Eq. (I. t 2) for determining the type of
pump that should be selected, and the term ~manometric head' is often used.
The static head H s is the vertical distance bctween the two levels in the
reservoirs and from Fig. 2.1 it can be secn that for the pipeline
..I
H = Hs + I losses
I

I It is worth noting here that, for the san1e size inlet and outlet diameters, Vo
and Vi are the same, and in practice (2 0 - 2 j ) is so small in comparison to (Po
- Pi)/pg that it is ignored. It is therefore not surprising to find that the static
pressure head across th~ pun1p is often used to describe the total head
developed by the pump.

2.2 CENTRIFUGAL PUMPS

Figure 2.1 Diagram of a pumping system
the free surface above the datum line since the velocity and static gauge
pressure at A are zero. The liquid enters the intake pipe causing the head loss
hin , with the result that the total head line drops to point B. As the fluid flows
from the intake to the inlet flange of the pump at elevation Zj, the total head
drops further to the point C due to pipe friction and other losses /tfi. The fluid
enters the pump and energy is imparted to it, which raises the total head to
point D at the pump outlet. Flowing from the pump outlet to the upper
reservoir, friction and other losses account for a total head loss hfo down to
point E, where an exit loss hout occurs when the liquid enters the upper
reservoir, bringing the total head at the upper reservoir to point F at the free
surface.
If the pump total inlet and outlet heads are measured at the inlet and.
outlet flanges respectively, which is usually the case for a standard pump test,
Figure 2.2 shows the three important parts of a centrifugal pump: (1) the
impeller, (2) the volute casing and (3) the dilTuser ring.
The diffuser is optional and mayor may not be present in a particular
design depending upon the size and cost of the pump. The impeller is a rotating
solid disc with curved blades standing out vertically from the face of the disc.
The tips of the blades are sometimes covered by another flat disc to give
shrouded blades; otherwise the blade tips are left open and the casing of the
pump itself forms the solid outer wall of the blade passages. The advantage of
the shrouded blade is that flow is prevented from leaking across blade tips
from one passage to another. As the impeller rotates, the fluid that is drawn
into the blade passages at the impeller inlet or eye is accelerated as it is forced
radially outwards. In this way, the static pressure at the outer radius is much
higher than at the eye inlet radius.
The fluid has a very high velocity at the outer radius of the impeller, and, to
recover this kinetic energy by changing it into pressure energy, diffuser blades
mounted on a diffuser ring may be used. The stationary blade passages so
formed have an increasing cross-sectional area as the fluid moves through
them, the kinetic energy of the fluid being reduced while the pressure energy is
further increased. Vaneless diffuser passages may also be utilized.
34 HYDRAULIC AND COMPRESSIBLE FLOW TUR130~tACIiINES HYDRAULIC PUMPS 35

Stationar) y
diffuser
vanes Vr = C r (radial velocity of fluid)

Diffuser -

-, = C (tangential

~
VIJ I

Impeller velocity of fluid)

't----
z
y~ With diffuser

Volutc _ ....... L- _

(axial velocity of fluid) -+-----+~

\Vithout difTuser
Figure 2.2 Ccntrifugal PUIllP cOlllponcnts

Finally, the fluid 1110ves fron1 the difTuser bladcs into thc volute casing,
which collects it and conveys it to the pUInp outlct. SOll1CtilllCS only the volute
casing exists without the difTL1ser~ however, son1e pressure recovery will take
place in the volute casing alone.
In dealing with the theory of hydraulic pU1l1pS, a nUlllbcr of assuInptions
will be made. At any point within the blade passages the l1uid velocity will
in general have three components, one each in the axial, radial and angular
directions as indicated in Fig. 2.3.
The velocity may then be written as a function of the three components
V == f(r, 0, :)
However, we will assume that the following hold:
1. There are an infinite number of blades so closely spaced that avlao = O.
That is, there is no flow in the blade passage in the tangential direction and
Vo == o. . angle (Xl to the ilnpeller inlet tangential velocity vector U}, where U 1 = wr 1 , w
2. The impeller blades are infinitely thin, thus allowing the pressure ddTerence
across them, which produces torque, to be replaced by tangential forces
that act on the fluid.
3. The velocity variation across the \vidth or depth of thc ill1pcller is zero and
hence vloz = O.a
4. The analysis will be confined to conditions at the ilnpcller inlet and outlet,
and to the angular n10111enturn change bctwecn thesc t\\'o stations. No
account is taken of the condition of the fluid bctween thcse two stations.
Figure 2.3 Cylindrical coordinates for a centrifugal pump
5. It is assumed that at inlct the l1uid is moving radially after entering the eye of
the pump.
Assumptions 1 and 2 1l1Can that the velocity is a function of the radius only,
V = f(r), and now with thesc assu111ptions the velocity vectors at inlet and outlet
of the impeller can be drawn and the theoretical energy transfer determined.
Figure 2.4 shows a ccntrifugal pump impeller with the velocity triangles
drawn at inlet and outlet. The blades are curved between the inlet radius r 1 and
outlet radius r 2 , a particle of fluid moving along the broken curve shown. /31 is
the angle subtended by the blade at inlet, measured froln the tangent to the
inlet radius, while /32 is the blade angle measured from the tangent at outlet.
The fluid enters the blade passages with an absolute velocity C 1 and at an
being the angular velocity of the impeller. The resultant relative velocity of
flow into the blade passage is W} at an angle {1'} to the tangent at inlet.
Similarly at outlet the relative velocity vector is W 2 at angle /3~ from the
tangent to the blade. By subtracting the impeller outlet tangential velocity
vector U 2' the absolute velocity vector C 2 is obtained, this being set at angle (1.2
from the tangent to the blade. I t is seen that the blade angles at inlet and outlet
do not equal the relative now angles at inlet and outlet. This is for a general
case, and unless otherwise stated (see slip factor, Sec. 2.3), it will be assumed
36 HYDRAULIC AND COMPRESSrBLE FLOW TURBOMACIlINES HYDRAULIC PUMPS 37

But by using the cosine rule, W 2 = U2 + C2 - 2UCcosa, then

(fIC\ cos~\ =(Uf - vvi + Cf)/2

and
U 2C2 COSCf. 2 = (U~ - vv~ + C~)/2
and suhstituting into Eq. (2.]) gives
C2 W'l
Cr'l
E = [(U~ - Uf) + (C~ - Cf) + (Wi - W~)J/2g (2.4)
The terms in Eq. (2.4) may now be exatnined in turn.
a2
(C~ - Cf)/2(J represents the increase of kinetic energy of the fluid across
the impeller, (U~ - Uf)/2(J represents the energy used in setting the fluid into
/ circular motion about the impeller axis and (Wi - W~)/2g is the gain of static
I
I head due to a reduction of the relative velocity within the impeller. The flow
~ rate is
/
I Q = 2rrr i C,l h t = 2rrr 2 C,2 h 2
I
I

I
where C, is the radial component of the absolute velocity and is perpendicular
,
I

V'
to the tangent at inlet and outlet while h is the width of the blade (in the z
C PI
direction). It is usually the case that C 1 = C,I and hence Ct. 1 = 90. In this case
C xl = 0, where e x1 is the component of the inlet absolute velocity vector
W

resolved into the tangential direction. W xand ex are often respectively called
the relative and absolute whirl components of velocity. When 131 = f3 /l ' this
is referred to as the 'no-shock condition' at entry. In this case the fluid moves
I~ VI
tangentially onto the blade. When 132 = f3~ there is no fluid slip at the exit.
1ft I

Figure 2.4 Velocity triangles for centrifugal pump impeller 2.3 SLIP FACTOR

It was stated in the previous section that the angle at which the fluid leaves the
impeller, fi~, may not be the same as the actual blade angle f32' This is due to
that the inlet and outlet blade angles are equal to their corresponding relative
fluid slip, and it occurs in both centrifugal pumps and centrifugal compressors,
flow angles.
and manifests itself as a reduction in C x2 in the Euler pump equation. One
From Euler's pump equation (Eq. (1.25)), the work done per second on
explanation for slip is that of the relative eddy hypothesis.
the fluid per unit weight of fluid flowing is
Figure 2.5 shows the pressure distribution built up in the impeller
E=W/mg=(U2Cx2-UICxd/y (J/sperN/sorm) (2.2) passages due to the motion of the blades. On the leading side of a blade there is
a high-pressure region while on the trailing side of the blade there is a low-
where C x is the component of absolute velocity in the tangential direction. E is
pressure region, the pressure changing across the blade passage. This pressure
often referred to as the Euler head and represents the ideal or theoretical head
distribution is similar to that about an aerofoil in a free stream and is likewise
developed by the impeller only.
associated with the existence of a circulation around the blade, so that on the
Now
low-pressure side the fluid velocity is increased while on the high-pressure side
and it is decreased, and a non-uniform velocity distribution results at any radius.
Indeed, the flow may separate from the suction surface of the blade. The mean
Thus
direction of the flow leaving the impeller is therefore f3~ and not 132 as is
(2.3) assumed in the zero-slip situation. Thus Cx2 is reduced to C~2 and the
38 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACIiINES HYDRAULIC PUMPS 39

2e /sin /1 2

-- - - - Ideal
--Actual Rela ti ve edd y
,/
,/

Figure 2.6 The relative eddy hdween impeller blades

blades is 2nr 2 /Z if we have 2 blades of negligible thickness. This may be

~(tJ
approximated to 2e/sin {>2 and upon rearrangement
+
e = (nr 2/2) sin fJ2
~Cx = (U 2/2r2)(nr 2 sin f32)
Figure 2.5 Slip and velocity distribution in centrifugal pump impdler blades
= (U 2 n sin fJ 2)/2

difference ~CX is defined as thc slip. Slip factor is defined as Now referring back to Fig. 2.5 for the no-slip condition

Slip factor = C~2/Cx2 = as (2.5) Cx 2 = U2 - Cr 2 co t {1 2

Stodola 1 proposed the existcnce of a relativc eddy within the blade and substituting into Eq. (2.5) gives
passages as shown in Fig. 2.6. He proposed that if a frictionless fluid passes Slip factor = (C x2 - ~Cx)/Cx2
through the blade passages it will, by definition, have no rotation; therefore at
the outlet of the passage rotation should be zero. No\v the ilnpel1er has an
angular velocity w so that relative to the ill1peller thc fluid 111Ust have an = I - (U 2n sin {>2)/[Z( U 2 - C r2 cot fJ2)]
angular velocity - w in the blade passages to cOlnply with the zero-rotation
condition. If the radius ofa circle thatlnay be inscribed bet\\'een two successivc = I -(nsin{J2)/{Z[I-(C r2 /U 2)cot{J2]} (2.6)
blades at outlet and at a tangent to the surfaces of both blades is e, then the slip For purely radial blades, which are often found in a centrifugal
is given by compressor, {12 will be 90 and the Stodola slip factor becomes
C
'

~CX = we ()s = 1 - n/Z (2.7)

Now the ilnpeller circul11ference is 2n1' 2 and therefore the distance between The Stodola slip factor equation gives best results when applied in the
40 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINJ':S HYDRAULIC PUMPS 41

range 20 < [32 < 30. Other slip factors are named after Ruseman 2
as = [A - B( CrZ / U 2) cot [32J/[ 1 - (C r2 / U 2) cot /J 2J (2.8)
where A and B are functions of /i 2 , Z and 1'2/1'1' and are best used in the
range 30 < f32 < 80 u . The Stanitz J slip factor given by
(Js= I-O.6]rr/{Z[1 -(C r2 /U 2 )cotI1 2 J} (2.9)
is best used in the range RO < 1/2 < 90.
/
When applying a slip factor, the Euler pUlnp equation (Eq. (1.25)) Leakagc now hctwccn
blade and casing. q
becomes
(2.10)
Figure 2.7 Lcakagc and recirculation in a centrifugal pump
Typically, slip factors lie in the region of 0.9, while slip occurs even if the fluid is
ideal.

associated with the flow rate Qi through the impeller, and so the impeller
2.4 CENTRIFUGAL PUMP CHARACTERISTICS power loss is ex pressed as

In Sec. 1.3 dimensionless groups were used to express the power, head and flow (2.11 )
relationships for a hydraulic machine. A well-designed pump should operate at However, while the flow through the impeller is Qi' this is not the flow through
or near the design point and hence near its maximum efficiency, but the the outlet or inlet flange of the machine. The pressure difTerence between
engineer is often required to know how the pump will perform at off-design impeller tip and eye can cause a recirculation ofa small volume offluid q, thus
conditions. For instance, the head against which the pump is operating may be reducing the flow rate at outlet to Q as shown in Fig. 2.7, and then
decreased, resulting in an increase in mass flow rate. However, before
examining this aspect of ofT-design performance, we will look at the losses (2.12)
occurring in a pump and the differing efficiencies to which these losses give rise, If Hi is the total head across the impeller, then a leakage power loss can be
whether or not the pump is working at the design point. We will then examine defined as
the efTects of working at the off-design condition.
PI = pgHiq (N m/s) (2.13)

2.4.1 Pump Losses Equation (2.12) shows that when the discharge valve of the pump is closed,
then the leakage flow rate attains its highest value.
The shaft power, Ps or energy that is supplied to the pump by the prime In flowing from the impeller outlet to the pump outlet flange, a further
mover is not the same as the energy received by the liquid. Some energy is head loss he takes place in the difTuser and collector, and since the flow rate
dissipated as the liquid passes through the machine and the mechanism of here is Q, then a casing power loss may be defined as
this loss can be split up into the following divisions.
Pc = pgQh c (N m/s) (2.14)
1. Mechanical friction power loss, Pm due to friction between the fixed and
rotating parts in the bearing and stuffing boxes. Summing these losses gives
2. Disc friction power loss, Pi due to friction between the rotating faces of the P~ = Pm + py(hiQi + hcQ + Hiq + QH) (2.15)
impeller (or disc) and the liquid.
3. Leakage and recirculation power loss, PI due to a loss of liquid from the where the total head delivered by the pump is defined as in Fig. 2.1 and
pump or recirculation of the liquid in the impeller. Eq. (2.1).
4. Casing power loss, Pc Anum ber of efficiencies are associated with these losses:

Impeller power loss is caused by an energy or head loss hi in the impeller . Fluid po\ver developed by pump
Overall efficiency = Sf r '. t
due to disc friction, flow separation and shock at impeller entry. This loss is 1a t power Inpu
42 HYDRAULIC AND COMPRESSIBLE FLO\V TURUOMACIIINI:S HYDRAULIC PUMPS 43

or
'/0 == IJyQII/P s (2.16)
Q
Fluid power at casing outlet -q I- I'

Casing efficiency == ----~- - - - - - - - -~-

B
Fluid power at casing inlet A 0 J

Fluid po\ver at ~~~~_in_g o~t_l~_t ~~.

== F~iJ-i)-o~~er- developed by ilnpeller - Leakage loss ~
I
N
Vl

or
Vl Useful Iluid power
Vl .2 11 Hi
'Ie == PfJQII / pyQi Ii == H / fl i (2.17) .2 iU
bO
COd
~ ...:w:
u COd

Fluid po\ver at iinpeller exit 'cC'd ~

Impeller efficiency == --------~ --~~-~-- ..c:
u I
Fluid power supplied to in1pellcr 11)

:E
M L
Jz(,
Fluid power at in11?~~I~:~~~~__.. _ I K Casing loss 5% G
Fluid power developed by inlpeller + Impeller loss Impeller loss 5% 11 1
E
or F
Mechanical loss 10 %
(2.18)
0 C
Flow rate through pUInp Flow rate (m:l/s)
Volumetric efficiency == ,
Flow rate through IInpeller
Figure 2.8 Losses in a centrifugal pump
or
'Iv == Q/(Q + q) (2.19) next loss to be accounted for is the leakage loss pgHjq represented by rectangle
Fluid power supplied to the iInpeller DJKI, and finally the casing loss pghcQ represented by rectangle MLGK is
Mechanical efficiency == - - - - - , - - - - - - - - - removed. This leaves us with rectangle JBLM, which represents the fluid
Po\ver Input to the shaft
power output or power developed by the pump pgQH.
or
(2.20)
Therefore 2.4.2 The Characteristic Curve
(2.21)
Euler's pump equation (Eq, (1.25)) gives the theoretical head developed by the
A hydraulic efficiency Inay be defined as pump, but if it is assumed that there is no whirl component of velocity at entry
Actual head developed by pUInp then C xl == 0 and the act uul theoretical head developed is
l1H == Theoretical head developed by iInpeller E == U 2 C x2/g == (Hi + hi) (2.23)

== H/(H i + hJ (2.22) and if slip is accounted for, Eg. (2.23) becomes

where the theoretical head (Hi + hi) is that obtained froIn Euler's equation EN == CJsE == CJs(H i + hj) (2.24)
(Eq. (1.25)) and '1H == 'li'lc' Now
Figure 2.8 shows how each of the po\ver losses arc subtracted from the C x2 == U 2 - W x2
initial input power. The rectangle OAllC represents the total power input to
== U 2 - C r2 cot fJ 2
the shaft while OADEFC is equivalent to the ITIechanical power loss. The
impeller loss pgQihj is next relnoved and is represented by rectangle EFGI. The == U 2 - (Q/ A) co t {J 2
44 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES HYDRAULIC PUMPS 45

according to
(2.27)
where Qn is the design nO\V rate. The friction losses are accounted for in the
form
hr = K 4 Q2 (2.28)
E Equations (2.27) and (2.28) are plotted in Fig. 2.9 and the sum of them is
Hydraulic
losses
subtracted froIll the curve of Eq. (2.26) to give the final characteristic. This
'"
curve is called the head-now characteristic of the pump.
'"
'" '"
H '"
-',., EN = Err..
2.4.3 Effect of Flow Rate Variation
A pump is usually designed to run at a fixed speed with a design head and flow
rate and these conditions would normally occur at the maximum efficiency
point. I-Io\vever~ it is not always the case in practice that the operating point
lies at the design point. 'rhis rnay be due to a pipeline being partially blocked, a
valve jammed partially closed or poor matching of the pump to the piping

h8hO C'k
- ---- -, II r
system. Also in general a variable-speed rnotor is not available to correct for
any deviation from the design condition, so that in what follows it is assumed
that the speed of the pump remains constant.
Figure 2.10 shows the velocity diagrams that pertain for three possible
Figure 2.9 Centrifugal pump characteristic flow rates: normal design now rate, increased nO\V rate and decreased flow rate.
When the flow rate changes, C r2 changes, and since U 2 is constant and the
blade outlet angle fJ 2 is constant (assuming rJ~ = fJ 2)' the magnitude of W 2 and
C 2 must change along with the angle (12' Since the etTective energy transfer E
where A is the flow area at the periphery of the impeller and C, is perpendicular depends on C x2 , then E will change accordingly. Thus a reduction in Qgives an
to it. Thus from Eq. (2.23) the energy per unit weight of flow becomes increase in C x2 ' while an increase in Q gives a reduction in C x2 ' It follows that,
E = U 2[U 2 -(Q/A)cotfJ2J/g should the head against which the pump operates be momentarily increased, E
and therefore C x2 increase and Q decreases to give the new operating point at
and since U 2' fJ2 and A are constants, then the increased head. Similarly a reduction in the operating head gives an
E=K 1 -K 2 Q (2.25) increase in Q,

and this equation may be plotted as the straight line shown in Fig. 2.9,
If slip is taken into account, it is seen from Eq. (2.9) that as C,2 increases
(and hence Q) then as decreases, thus reducing the value of E in Eq. (2.25) to
(2.26)
l
The loss due to slip can occur in both a real and an ideal fluid, but in a real
fluid account must also be taken of the shock losses at entry to the blades, and
the friction losses in the casing and impeller vanes, or indeed at any point
~
where the fluid is in contact with a solid surface of the pump. At the design ~ ~
Reduced flow Design flow Increased flow
point the shock losses are zero since the fluid would move tangentially onto the
blade, but on either side of the design point the head loss due to shock increases Figure 2.10 EITect of now rate variation on outlet velocity
46 HYDRAULIC AND COMPRLSS(BLF (''L()\V TURBOMACII(l':ES HYDRAULIC PUMPS 47

H
.
-'J\
.,:,\
-- - , . ItJ -'\,
(t)

- - ---- --.- 11 = {/ (jl'l = 90)

-----------------
Vl
Reduced flow Design now Increased flow
ia
Figure 2.11 Effect of flow rate variation on inlet velocity

At the inlet the efTect of now rate change is to cause eddies on the suction
surface of the blade for a reduced nO\\1 rate and on the pressurc surface of the Q
blade for an increased flow rate. The dcsign condition is the ~no-shock'
Figure 2.13 Theoretical characteristics for varying outlet blade angle
condition, which corresponds to the now QD in Fig. 2.9. The corresponding
velocity diagrams can be seen in Fig. 2.11. In all cases it is assuI11cd that Cxt is
zero. Writing E as a head,
H = a- bQ (2.29)
2.4.4 Effect of Blade Outlet Angle and for pumps [3 2 typically lies between 15 and 90.
The characteristic curve will also be afTected by the blade angle at outlet, the Case (ii). Radial blades, [J 2 = 90
three types of blade settings being back ward-facing, forward-facing and radial
H=a (2.30)
blades. Figure 2.12 shows clearly the velocity triangles for each case with
C x ! =0.
P,H Forward-facing
Case (i). Backward-facing blades, [1 2 < 90 Head

Therefore

or
E = (U ~ / {j) - (Q U 2co t [3 2/{j A )
Radial
- Power

Backward Radial Forward

Forwarding-facing vanes Radial vanes Back ward-facing vanes
Q
Figure 2.12 Centrifugal pump outlet velocity triangles for varying blade outlet angle Figure 2.14 Actual characteristics for varying blade outlet angle
48 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIfINES
HYDRAULIC PUMPS 49

Case (iii). Forward-facing vanes, Ii 2 > 90 Cl

H = a + hQ (2.31 )
0
where fJ2 would be typically 140 for a multi-bladed centrifugal fan.
These equations are plotted in Fig. 2.13 as characteristics and they revert
to their more recognized curved shapes (for the reasons previously discussed)
as shown in Fig. 2.14.
For both radial and forward-facing blades the power is rising cont-
inuously as the flow rate is increased. In the case of back ward-facing vanes the
maximum efficiency occurs in the region of maximum power, and if, for some Impeller
reason, Q increases beyond Qn, this results in a power decrease and therefore
the motor used to drive the pump may be safely rated at the maximum power.
This is said to be a self-limiting characteristic. In the case of the radial and Volute of
forward-facing vanes, if the pump motor is rated for maximum power, then it _ increasing Figure 2.15 Simple volute or
will be under-utilized most of the time, and extra cost will have been incurred cross section scroll collector

for the extra rating, whereas if a smaller motor is employed rated at the design
point, then if Q increases above Qo the motor will be overloaded and may fail.
It therefore becomes more difficult to decide on a choice of motor for these
latter cases.
2.5 FLOW IN THE DISCHARGE CASING
The discharge casing is that part of the casing following the impeller outlet. It
has two functions: (i) to receive and guide the liquid discharged from the
impeller to the outlet ports of the pump, and (ii) to increase the static head at
the outlet of the pump by reducing the kinetic energy of the liquid leaving the
impeller. These two functions may be called collector and diffuser functions.
The former function may be used alone while the latter can occur either before
or after the collector function. In addition diffusion can take place in a vaned
or vaneless diffuser.
where P = radial force (N), /-1 = head (m), D 2 = peripheral diameter (m), B 2 =
impeller width (m) and K = constant determined from the following equation
for a particular value of Q:
(2.33)
A cross section of the volute casing is shown in Fig. 2.16. The circular
section is adopted to reduce the losses due to friction and impact when the fluid
hits the casing walls on exiting from the impeller. Of the available kinetic
energy at impeller outlet, 25-30 per cent may be recovered in a simple volute.
2.5.2 Vanelcss Diffuser
Diffusion takes place in a parallel-sided passage and is governed by the
principle of conservation of angular momentum of the fluid. The outlet

2.5.1 Volute or Scroll Collector

A simple volute or scroll collector is illustrated in Fig. 2.15 and consists of a
circular passage of increasing cross-sectional area. The advantage of the
simple volute is its low cost. The cross-sectional area increases as the
increment of discharge increases around the periphery of the impeller and it is
found that a constant average velocity around the volute results in equal
pressures around the pump casing, and hence no radial thrust on the shaft.
Any deviation in capacity (flow rate) from the design condition will result in a
radial thrust, which if allowed to persist could result in shaft bending. Values of
radial thrust are given by the empirical relationshi p 4
(2.32) Figure 2.16 Section through volute casing
50 HYDRAULIC AND COMPRESSIBLE I''LD\V TURBOMACIlINLS HYDRAULIC PUMPS 51

where b is the width of tile diffuscr passage perpendicular to the peripheral area
of the inlpeller and is usually the same as the impeller width. Letting the
subscripted variables represent conditions at the impeller outlet and the
Diffuser unsubscripted variablcs represent conditions at any radius r in the vaneless
passage difTuser, then froln continuity

f rbpC r == r 2 b 2 P2 Cr2
;I'
b / Free vortex now or
/
/
in difTuser passage (2.34)
If frictionless flow is assuillcd, then by the conservation of angular momentum
C x == C x2 r2 /r

w
Figure 2.17 Vaneless diffuser passage
tangential velocity is reduced as the radius increases, while the radial
component of absolute velocity is controlled by the radial cross-sectional area
of flow b. A vaneless difTuser passage is shown in Fig. 2.17.
With reference to Fig. 2.18 the size of the diffuser n1ay be deternlined as
follows. The mass flow rate n1 at any radius r is given by
HJ == pAC, == 2nrhpC,
But C x C r (usually) and therefore the absolute velocity C is approximately
equal to Cx or
(2.35)
From Eq. (2.35), for C to be slnall, which is what we are trying to achieve, then r
n1ust be large and thereforc, for a large reduction in the outlct kinetic energy, a
diffuser with a large radius is required.
For an incompressible fluid, the inclination of the absolute velocity vector
to the radial line remains constant at all esince at the outlet from the impeller
(Fig. 2.18)
tan (I.~ == C x2/Cr2 == constant --= tan Y.'
since rC r is constant froln the constant mass flow rate requirement, and Cxr is
constant from the conservation of angular momentulTI requirement. Thus the
flow in the diffuser remains at a constant inclination (I.' to radial lines, the flow
path tracing out a logarithlnic spiral, and if for an incremental radius dr the
fluid moves through angle dO, then from Fig. 2.18
r dO == (tan a') dr
Integrating,
(2.36)
Putting a' == 78 and (r/r 2 ) == 2, the change in angle of the diffuser is almost
180, giving rise to a long flow path, which may result in high frictional losses,
which in turn gives a low efficiency. So it is seen that the length of the diffuser
must be balanced by the pressure recovery that is required and an optimum
point is usually found based on either economic or hydraulic friction loss
considerations.

2.5.3 Vaned Diffuser

The vaned diffuser shown in Fig. 2.19 is able to diffuse the outlet kinetic energy
Figure 2.18 Logarithmic spiral now in vaneless space at a much higher rate, in a shorter length and with a higher efficiency than the
 52 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES
HYDRAULIC PUMPS 53

Collector Casing

\
Throat of diffuser passage
----.
I Diffuser
~--i-I-----i
Delivery i
I I
1/ I I

I I
..----.-+---,' __ I Impeller

/
-~~
Impeller

Figure 2.20 Head rise across a centrifugal pump

The collector and diffuser operate at their maximum efficiency at the

design point only. Any deviation from the design discharge will alter the outlet
velocity triangle and the subsequent flow in the casing. Figure 2.20 shows the
contribution of each section of the pump to the total head developed by the
pump.

2.6 CAVITATION IN PUMPS

Figure 2.19 A vaned diffuser
vaneless diffuser. This is very advantageous where the size of the pump is
important.
A ring of difTuser vanes surrounds the inlpeller at the outlet, and after
leaving the impeller the fluid moves in a logarithmic spiral across a short
vaneless space before entering the difTuser vanes proper. Once the fluid has
entered the diffuser passage, the controlling variable on the rate of diffusion is
the divergence angle of the difTuser passage, which is of the order of 8-10 and
should ensure no boundary-layer separation along the passage walls. The
number of vanes on the diffuser ring is subject to the following considerations:
1. The greater the vane number, the better is the diffusion but the greater is the
friction loss.
2. The cross section of the diffuser channel should be square to give a
maximum hydraulic radius (cross-sectional area/channel perimeter).
3. The number of difTuser vanes should have no common factor with the
number of impeller vanes. This is to obviate resonant or sympathetic
vibration.
Cavitation is a phenomenon that occurs when the local absolute static
pressure of a liquid falls below the vapour pressure of the liquid and thereby
causes vapour bubbles to form in the main body of liquid, that is the liquid
boils. When t-he liquid flows through a centrifugal or axial flow pump, the
static pressure (suction pressure) at the eye of the impeller is reduced and the
velocity increases. There is therefore a danger that cavitation bubbles may
form at the inlet to the itnpeller. When the fluid moves into a higher-pressure
region, these bubbles collapse with tremendous force, giving rise to pressures
as high as 3500 atm. Local pitting of the impeller can result when the bubbles
collapse on a metallic surface, and serious damage can occur from this
prolonged cavitation erosion, as shown in Fig. 2.21. Noise is also generated in
the form of sharp cracking sounds when cavitation takes place.
Referring to Fig. 2.1, cavitation is most likely to occur on the suction side
of the pump between the lower reservoir surface and the pump inlet since it is
in this region that the lo\vest pressure will occur. A cavitation parameter (J is
defined as
Pump total inlet head above vapour pressure
(J=------------------
Head developed by pump
54 HYDRAULIC AND COMPRESSIULE FLOW TURUOMACHINES HYDRAULIC PUMPS 55

3% head
reduction

Figure 2.21 Cavitation erosion in centri- Q

fugal pump impl.:lkr

and at the inlet flange

(J = (Pi/PU + V; /2y - Pvap/PY)I II (2.37) NPSH
where all pressures are absolute. l~hc nunlerator of Eq. (2.37) is a suction head Measured
and is called the net positive suction head (IV PSH) of the pUlllp. It is a measure
~~~:;:;:;~NPSH
of the energy a vailable on the suction side of the punlp. I I I
Every pump has a critical cavitation nlJlnber (J\:, which can only be I
I

I
I
I
I
I
I
determined by testing to find the Illininlum value of N PSiJ before cavitation Figure 2.22 Critical N PSH plotted on the
occurs. Various 111cthods exist for determining the point of cavitation pump characteristic
inception, and (Jc, and therefore the minimunl N PSil required by the pump,
will depend upon the criteria chosen to define (J c as well as the conditions placed at the lower reservoir surface, then Z A is also zero and Eq. (2.38)
under which the test is carried nut. One n1ethod is to deterIlline the normal becomes
head-flow characteristic of the pUlnp and then to repeat the test with the inlet
to the pump progressively throttled so as to increasc the resistance to flow at PAl pg = p) pg + Vf 12y + ilsu\:tion
the inlet. It will be found that for difTercnt throttle valve settings the where
performance curve will fall away from the nonnal operating curve at various Z i + h in + hfi = H suction
points and one definition of the occurrence of IniniIl1Uln N PSI-I is the point at Substituting for p) pg in Eq. (2.37) gives
which the head H drops below the normal operating characteristic by some
arbitrarily selected percentage, llsually about 3 per cent. At this condition, (J == (PAl pg - Pvapl pg - H suction)1 H (2.39)
static inlet pressure Pi and inlet velocity Vi are ITIeasured, and (Je is then Providing (J is above (J c' cavitation will not occur, but, in order to achieve
calculated fronl Eq. (2.37). The mininlum required N PSH or (Je may then be this, it may be necessary to decrease Hsuction by decreasing Zi and in some cases
plotted for the difTerent degrees of inlet throttling to givc a curve of (Je versus the pump may have to be placed below the reservoir or pump free surface, i.e.
flow coefficient (Fig. 2.22). negative Zi' especially if hfi is particularly high due to a long inlet pipe. Thus
In Fig. 2.1, the energy loss bct\veen the free surface (A) and the inlet side of when the pump is connected to any other inlet pipe system, (J as determined
the pUlnp (i) is given by thc steady flow energy equation as from Eq. (2.39) may be calculated and providing (J(available) > O"e(required)
Energy at A - Energy at i = Energy lost bctween A and i then cavitation will be avoided.

Wri ting the energy in tern1S of heads

2.6.1 Suction Specific Speed
(PAl py + vi/2y + ZA) = (pJ pg + vf /20 + Zi) + (h in + hfi ) (2.38)
It is reasonable to expect that the efficiency will be dependent not only upon
where (h in + /lfi) represents the losses. No\v VA cquals zero and, if the datum is the flow coefficient but also upon another function due to cavitation. The
56 HYDRAULIC AND COMPRESSIRLE FLOW TURROMACHINES HYDRAULIC PUMPS 57

other function is the dimensionless suction specific speed and is defined as

N suc = NQl/2/[g(NPSH)]3/4 (2.40)
Thus
r, = !(, NsuJ (2.41)
2
It is found from experiments that the inception of cavitation occurs at
constant values of N slIC and empirical results sho\v that N suc ~ 3 for N in rad/s, View on
Q in m 3 /s and g(NPSH) in m 2 /s 2 The cavitation parameter may also be X-X
determined by dividing the dimensionless specific speed by the dimensionless
suction specific speed:
Ns/N suc = [NQl/2/(gH)3/4]/{NQl/2/[{J(NPSH)]3/4}
= (N PSH)3/4 / H 3/4
c~
= (J~/4 (2.42)
Also from the similarity laws
NPSH 1 /NPSH 2 =(N 1 /N 2)2(D 1 /D 2)2 = (Jl/(J2 Outlet I
guide
vane ~-~
2.7 AXIAL FLOW PUMP 1~4_ _ --- - ------ ---J
U~ = VI = V
An axial flow pump consists of a propeller type of impeller running in a casing
with fine clearances between the blade tips and the casing walls. In the absence
of secondary flows, fluid particles do not change radius as they move through
the pump; however, a considerable amount of swirl in the tangential direction
will result unless means are provided to eliminate the swirl on the outlet side.
This is usually done by fitting outlet guide vanes. The flow area is the same at
L
Figure 2.24 Axial now pump velocity triangles

inlet and outlet and the maximum head for this type of pump is of the order of
Stationary inlet Stationary outlet 20 m. It may be seen in Fig. 1.7 that the dimensionless specific speed of axial
guide vanes guide vanes
flow pumps lies at the right-hand side of the pump spectrum, its characteristics
\ 1\
1\
1\
being one of low head but high capacity. The usual number of blades lies
between two and eight, with a hub diameter/impeller diameter ratio of 0.3-0.6.
In many cases the hlade pitch is fixed but nl0st large hydroelectric units have
variable-pitch blades to allow for load variations.
Figure 2.23 shows an axial flow pump impeller. The section through the
blade at X-X is sho\vn enlarged with the inlet and outlet velocity triangles
superimposed in Fig. 2.24. Ii will be noticed that the blade has an aerofoil
section and that the inlet relative velocity vector WI does not impinge
tangentially but rather the blade is inclined at an angle of incidence i to the
relative velocity vector WI' This is similar to the angle of attack of an aerofoil
in a free stream. It is assumed that there is no shock at entry and that the fluid
leaves the blade tangentially at exit.
Changes in the condition of the fluid take place at a constant mean radius;
therefore
Figure 2.23 An axial now pump U 1 = U 2 = U = (or
58 HYDRAULIC AND COMPRESSIBLE FLOW TURBOr-.tACHINES HYDRAULIC PUMPS 59

Assuming also a constant flow area from inlet to outlet the fluid at exit frOITI the blade is relatively small, resulting in a low kinetic
energy loss. An axial flo\v pump therefore tends to ha ve a higher hydraulic
cfficiency than the centrifugal pump.
and noting that the flow area is the annulus fanned bet \veen the hub and the
blade tips, then we may write 2.7.1 Blade Element Theory
111 = pCan(R~ - R~) An axial flow pump impeller may have a large number of blades spaced closely
From Eq. (2.2), together or a few blades spaced far apart, while for mechanical strength
(2.43) considerations, the blade chord will vary from hub to tip. The peripheral
distance betwecn similar points on two adjacent blades is the pitch, and the
and for maximum energy transfer Cxl = 0, i.e. ~l = 90 and C 1 = C a , the ratio blade chord/blade pitch at a given radius is known as the solidity ratio a:
absolute flow velocity at inlet being axial for nlaxinlulll energy transfer.
Now a = cis (2.46)
It is therefore possible to have high- or low-solidity blades, an impeller
or with a low number of blades ilnplying a low solidity. Where the blades have a
C x2 = U - C a cotfJ 2
low solidity, flow interference frOlTI one blade to the next is low and the blade
Hence substituting for C x2 in Eg. (2.43) with C.d = 0, the Inaxinlunl energy may be considered to be acting alone in a free stream and is analysed as such.
transfer or head is However, for high-solidity blades implying very closely spaced blades, the flow
between the blades will be greatly influenced by the adjacent blades and we
f, = U( U - Cu cot fJ 2)/0 (2.44)
must resort to cascade data for an analysis of the forces acting on them. Since
For constant energy transfer, Eq. (2.44) applies over the whole span of the axial flow pump impellers invariably have less than six blades, it is usual to
blade from hub to tip; that is it applies at any radius r between R l and R h For E consider only isolated blade clement theory for them and this is now briefly
2
to be constant over the whole blade length it is obvious that, as U increases
with radius, so an equal increase in UCacot fJ2 nlust take place and since Ca is
constant then cot fJ 2 must increase, and the blade must therefore be twisted as
the radius changes.
Strictly speaking the work done per unit weight of flow through an
annulus of thickness dr should be considered and this then integrated across
the whole flow area from the hub to the tip
E = WlIng = U( U - Cu cot fJ2)/Y
or

But the incren1ental nlass flow rate dl1Z is

dl1z = p(2nr) (CJ dr
and U = wr. Therefore
RI
'\.
W = 2npC(JuJ r 2 (evr - C el cot f~ 2) dr (2.45) '\.

J Rh

~
Equation (2.45) can only be integrated if the relationship between fJ2 and r is
known.
For design purposes it is usual to select conditions for use in Eq. (2.44) at ..... '-.1 __
I
-1 Figure 2.25 Circulation around an isolated
the mean radius (R h + R()/2 along the blade. The whirl cOll1ponent ilnparted to A blade
 60 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES
HYDRA ULIC PUM PS 61

described, the treatment of cascade analysis being reserved for the section on
Substitution into Eq. (2.48) gives the circulation for the whole impeller as
axial flow compressors, to which machines it is more appropriately applied.
Consider the circulation r around the control surface of the isolated r = Z(S2Cx2 -.'\\ C. d )

blade shown in Fig. 2.25, where the lengths AB and CD are the blade pitches But
at inlet and outlet and AD and BC bisect the flow passages between adjacent and ZS2 = 2nr 2
blades. Therefore
The circulation is given by the line integral around ABCD and may be
evaluated by summing the individual circulations comprising the circuit, such From Eq. (2.43),
that

r AIlCO = f
where the circulation is positive anticlockwise. But
Vds= r Vds 2 + i~ Vdl+ tD

Vds l + I: Vdl (2.47)

Hence
E= wzr b/2ny
The Kutta loukowski law (L = p U or,J for lift per unit span on an aerofoil
(2.50)

may now be used, where U 0 is the free stream velocity. Dividing this by
0.5p U 6c gives the lift coefficient
= - fA V dl
while
IB e V dl
JD CL = r b /0.5U oc

I'D where c is the chord of the impeller blade.

and Jc V dS I = - eXIS I Since the appropriate free stream velocity for flo\v over the blade is the
relative velocity W instead of U 0' and since this is different at blade inlet and
Hence
outlet, the appropriate relative velocity is usually defined as
(2.48)
H/~ = Cc; + [(W.d + Wx2 )/2J 2
For a number of blades shown in Fig. 2.26 the circulation around each
blade may be determined, and the total circulation about Z blades is the sum = C; + [cot fi 1 + cot 112)/2J 2
of the individual circulations, remembering that along a line such as BG the
Therefore r h == a.5CI. W,XlC, and substituting this into Eq. (2.50)
circulation for one blade is positive (anticlockwise) while for the adjacent
blade it is negative (clockwise). Therefore for Z blades the total circulation is E = 0.5coZC L Wooc/2ng (2.51)

r ADEH == zr b (2.49) Values of lift coefficients for differing blade profiles may be determined
from readily available tables and charts 5 and an estimate for E obtained.
where r b is the circulation around a single blade.

2.7.2 Axial Flo,,, Pump Characteristics

Axial flow pump design has evolved empirically and it is only in relatively
modern times that aerofoil theory has been applied. Nevertheless, efficiencies
o C B
--7--- ---7
A
/--- --7 - - - of over 90 per cent were achieved using empirical data and it would seem that

/1)1 1/) 11)/ /1

/ /
/ I
/ /
/ /
/ I
/ / /
I aerodynanlic design has not itnproved the efficiencies by much.
Typical head--now, power and efficiency curves arc shown in Fig. 2.27. A
/ I
steep negative slope is evident on the head and power curves at low flow rates.
/ / This can be explained by considering Eq. (2.44). For a given blade design at
/ /
/ / fixed speed with axial now at inlet
/ /

L
/
.L
/
/
/
./
/ / E = U(U - CLIent /1 2 )/U
E F G H
Now Q is proportional to C a and therefore
Figure 2.26 Circulation around a number of isolated blades
dE/dC a ex dE/dQ ~ - U cot 11 2
 HYDRAULIC PUMPS 63

For axial flow at inlet, ti 2 is relatively small, and thus for a given pump at a
given speed the head- flow relationship has a steep negative slope. The power
N", = 4 rad curve is similarly very steep, the power requirement at shut-ofT being perhaps
2-2.5 times that required at the design point. This makes for a very expensive
Head electric motor to cover the eventuality of low flow rates and so the fixed-blade
,, axial flow pump is limited to operation at the fixed design point. Variable-flow
, ,,
,, machines may be designed employing variable blade stagger or setting angles.
,, Here the blade angle is adjusted so that the pump runs at its maximum
,,
,, efficiency at all loads and also reduces the shut-off power requirement. Figure
, 100 = design point
2.28 shows the efTect of changing blade stagger angle.
Power
/', ", In Fig. 2.27 the power and head curves are seen to enter a region of
,,
' ...... __
.... -
.....
instability at about 50 per cent of the design flow rate. This is due to Ca
" / I becoming increasingly slnall and thereby increasing the angle of incidence of
flow onto the blade until separation and stalling of the blade occurs. The
further head rise at even lo\ver flow rates and the consequent power increase is
Effici,ency due to recirculation of the fluid around the blade fronl the pressure side to the
suction side and then up onto the pressure side of the next blade. An increased
blade stagger angle will once again reduce this recirculation and thereby the
power req uiremen t.
o 20 40 60 80 100 120 140
Percentage of design flow rate
2.8 PUMP AND SYSTEM MATCHING
Figure 2.27 Axial flow punlp characteristics

It has been shown that a hydraulic pump has a design point at which the
overall efficiency of operation is a maximum. However, it may happen that the
pipe system in which the pump is being used is unsuited to the pump and a
different pump with a more suitable characteristic is required. This section will
Efficiency Loci of maximum examine how a pump and a pipe system may be nlatched to each other, the
lJ,H efficiencies
effect of changing the pump speed and dialneter, and finally the efTect of
\ connecting pum ps in series and parallel.
\ Consider the pipe system in Fig. 2.1. On the suction side the losses
\ \ ----~
expressed in terms of standard loss coefficients are the sum of the minor
\ \ Increasing
stagger angle losses hin
\ \
\ \ /Head hIn = ~kVtj2f1
\ \ i...J I ~

\ \ ,,-
", '--,., _/
and the friction loss
hfi = 4llj V; j2yd j

where f is the Darcy friction factor, I is the length of the inlet pipe and dj its
j

/ Increasing diaIl1eter. Thus the total head loss is

/' stagger angle

Figure 2.28 Changing blade stagger angle on an axial flow pump On the delivery side the SUlll of the bend, friction and exit losses that must be

62
64 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES HYDRAULIC PUMPS 65

overcome is It should be noted that if there is no static head rise of the liquid (e.g.
Ito = 4f/o V;/2gd o + Ik V; /2g pumping in a horizontal pipeline betwecn two rcservoirs at the same elevation)
then H s is zero and the system curve passes through the origin. This has
Finally, the liquid must be moved from the lower reservoir to the upper
implications when speed and diameter changes take place. Because of the
reservoir through the static head H s ; hence the total opposing head of the pipe
flatness of rotodynanlic pump characteristics, a poor estimate of the system
system that must be overcome in order to move the fluid from the lower to
losses can seriously affect the flow rate and head~ whereas in posi tive
upper reservoir is
displacement pumps, the ]i -Q curve is almost vertical and, even if the head
H = H s + Ito + hi (2.52) changes substantially, the flow rate stays almost constant.
No.w from th.e continuity equation (Eq. (1.21)) the flow rate through the
system IS proportIonal to the velocity. Thus the resistance to flow in the form of 2.8.1 Effect of Speed Variation
friction losses, head losses, etc., is proportional to the square of the flow rate
Consider a pump of fixed diameter pumping liquid with zero static lift. If the
and is usually wri tten as
characteristic at one speed lv 1 is known, then it is possible to predict the
System resistance = K Q2 (2.53) corresponding characteristic at speed N 2 and also the corresponding operat-
It is a measure of the head lost for any particular flow rate through the system. ing points. Figure 2.30 shows the characteristic at speed N l'
For points A, Band C the corresponding head and flows at a new speed
~fany.parameter in the system is changed, such as adjusting a valve opening, or
InsertIng a new bend, etc., then K will change. The total system head loss of Eq. N 2 are found thus. We have
(2.52) therefore becomes

H = H s + KQ2 (2.54) and

and if this equation is plotted on the head-flow characteristic, the point at Similarly
which Eq. (2.54) intersects the pump characteristic is the operating point, and
this mayor may not lie at the duty point, which usually corresponds to the and
de~ign point and maximum efficiency. The closeness of the operating and duty
pOInts depends on how good an estimate of the expected system losses has ]-f 1/ N~ = H 2/ N~ (2.56)
been made. In Fig. 2.29 the system curve is superimposed on the H-Q Applying Eqs (2.55) and (2.56) to points A, Band C and letting the
characteristic. corresponding points be A', B' and C',
and
1J,H
or
Design point H 2 oc Q~ (2.57)

H
System
resistance
c /
/

/'

Operating
point

Figur<' 2.29 System resistance and Figure 2.30 EITect of speed variation on
Q pump characteristic curves Q the operating point
66 HYDRAULIC AND COMPRESSIIlLE FLO\V TURBOMACHINES

11,H H
Operating point
for combined pumps

tSYSICI11 rcsistance
__- Single pump
operating point

Figure 2.31 EfTect of speed variation

Figure 2.32 Two similar pumps in series
Q on the efficiency Q

and we see that all corresponding points lie on a parabola passing through the
origin. This means that for an operating point at A at speed N l' it is only
necessary to apply the similarity laws directly to find the corresponding
H Single pump
operating point at the new speed since it will lie on the system curve itself.
However, if there is static lift (H s =1= 0) it will be necessary to calculate and then
plot the corresponding points A', B', C' at the new speed, since the system curve
will no longer pass through the origin. The systeITI curve is then drawn to find Combined pumps
the new operating point at its intersection with the N 2 characteristic such that
the corresponding maxin1 un1 efficiency at design poin tOP 2 remains the same Figure 2.33 Two similar pumps
as at DP 1 but at a difTerent head and flow rate as in Fig. 2.31. Q in parallel

2.8.2 Variation of PUlllP Dianlctcr

A variation of pump diameter 111ay be similarly exan1ined through the
similarity laws. For a constant speed, H
Pumps in series
Ql/Dr = Q2/D~
and

or
(2.58)
This curve does not lie on the system characteristic and therefore part of ';;:
Pumps in parallel
"

I
,~
the new characteristic 111ust be drawn through A', H' and C' at the new

/
diameter so that the new operating point ITIay bc found. 'Thc cfficiency curve ~'':.', "
moves across in a sin1ilar Inanncr to before, thc corresponding efliciencies , ~

, "
being equal. Pump I "
/\
Pump 2
2.8.3 Pumps in Serics and Parallcl
Q
Should the head or flow ratc of a single pUlnp not be sufficient for an
application, pun1ps can be cOlnbined in series to obtain an increase in head, or Figure 2.34 Two difTerent pumps combined in series and parallel

67
68 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES HYDRAULIC PUMPS 69

H ,
\
and it rotates at 1450 rpm. The now rate through the pump is 0.028 m 3 Is and a slip factor of 0.77
may be assumed. Determine the theoretical head developed by the impeller, and the number
\ of impeller blades.
\ Stable operating point
\ 2.3 An impeller with an eye radius of 51 mm and an olltside diameter of 406 mm rotates at 900
\ rpm. The inlet and outlet blade angles measured from the radial now direction are 75 and 83
\
\
\
,, \ respectively, while the blade depth is 64 mm. Assuming zero inlet whirl, zero slip and an hydraulic
efficiency of 89 per cent, calculate
(a) the volume now rate through the impeller,
"
(b) the stagnation and static pressure rise across the impeller,
(c) the power transferred to the fluid and
(d) the input power to the impeller.
2.4 The basic design of a centrifugal pump has a dimensionless specilic speed of 0.075 rev. The
Q Figure 2.35 Single axial now pump
blades are forward facing on, the impeller and the outlet angle is 120 to the tangent, with an
impeller passage width at outlet equal to one-tenth of the diameter. The pump is to be used to
pump water a vertical distance of 35 m at a now rate of 0.04 m 3 Is. The suction and delivery pipes
are each of 150 mm diameter and have a combined length of 40 m with a friction factor of 0.005.
Other losses at pipe entry, exit, hends, etc., are three times the velocity head in the pipes. If the
blades occupy 6 per cent of the circumferential area and the hydraulic efficiency (neglecting slip) is
. - Instability 76 per cent, what must be the diameter of the pump impeller?
2.5 When a laboratory test was carried out on a pump, it was found that, for a pump total head of
36 m at a discharge of 0.05 mJ/s, cavitation began when the sum of the static pressure plus the
velocity head at inlet was reduced to 3.5 m. The atmospheric pressure was 750 mmHg and the
vapour pressure of water 1.8 k Pa. If the pump is to operate at a location where atmospheric
pressure is reduced to 620 mm Hg and the vapour pressure of water is 830 Pa, what is the value of
the cavitation parameter when the pump develops the same total head and discharge? Is it
necessary to reduce the height of the pump above the supply, and if so by how much?
Q Figure 2.36 Axial flow pumps in parallel 2.6 The inner and outer diameters of an axial now pump are 0.75 and 1.8 m respectively. Fixed
stator blades lie downstream of the rotor with an inlet angle of 40 (at the mean diameter)
measured from the direction of blade motion. The rotor blade outlet angle (at the mean diameter)
in parallel for an increase in flow rate. The combined pumps need not be of the also measured from the direction of blade motion is 30 and the rotor rotates at a speed of250 rpm.
same design. If the whirl velocity upstream from the rotor is zero at all radii, determine
(a) the axial velocity if the now onto the stator blade occurs at zero incidence,
Figures 2.32 and 2.33 show the combined }-I -Q characteristic for the cases
(b) the rotor torque if the axial velocity is constant across the now annulus, and
of identical pumps connnected in series and parallel. It will be observed that (c) the root and tip rotor blade angles for zero incidence and zero inlet whirl.
the operating point changes in both cases. In Fig. 2.34 the combined 2.7 A twelve-bladed axial now fan has a hydraulic efficiency of 0.92, a mean radius of 0.93 m and
characteristics of two different pumps connected in parallel and series are rotates at 450 rpm. Air enters the blades axially at a speed of 40 mls and the head developed across
drawn. the blades is 35 m of air. If the chord length at the mean radius is 0.33 m, lind the blade angles at
When axial flow pumps are connected in parallel, care must be taken to inlet and outlet if the hlades may he considered to act as isolated aerofoils. If the blades are
aerofoils with the following characteristics, lind also the angle of incidence of the blades and the
ensure that the system characteristic does not cut the pump characteristic in
blade stagger angle. All calculations are to be carried out at the mean radius.
two places, otherwise instability may result. This arises due to the lateral
spreading of the pump characteristic illustrated in Figs 2.35 and 2.36.

Angle of incidence (deg) - 4 -2 a 4 8 12

EXERCISES
Lift coefficient -0.18 -0.02 0.13 0.46 0.77 1.025
3
2.1 A centrifugal pump of 1.3 m diameter delivers 3.5 m Imin of water at a tip speed of 10 mls and
a flow velocity of 1.6 m/s. The outlet blade angle is at 30 to the tangent at the impeller periphery.
Assuming zero whirl at inlet, and zero slip, calculate the torque delivered by the impeller. 2.8 A centrifugal pump is used to pump water from a low to a high reservoir having a water level
difference of 32 m. The total length of pipe is 1000 m with a friction factor of 0.005 and its diameter
2.2 The impeller of a centrifugal pump has backward-facing blades inclined at 30 to the tangent
is 200 mm. Neglecting all losses except friction, determine the rate of flow between the reservoirs
at impeller outlet. The blades are 20 mm in depth at the outlet, the impeller is 250 mm in diameter
70 HYDRAULIC AND COMPRESSIBLE FLOW TUR130MACHINES HYDRAULIC PUMPS 71

and the power required to drive the pump. The characteristics of the pump are as follows: But C xl == 0 since there is no inlet whirl component. Therefore
, U 2Cx2
"
f. = - -
Q (m /h)3
0 46 92 13R IH4 230 9
H(m) 68 64 54 42 26.4 8 U2
=-(U 2 -Wx2 )
1/(%) 0 49.5 61 63.5 53 10 9

- ~ ( 10- ~1.~ )
2.9 The characteristics of a 0.7 In diameter centrifugal pump impeller running at 750 rpm are as 9.81 tan 30t)
follows:
== 7.36 m or W /(N/s)
Power delivered == E x (lng)
Q (n1 3lInin) o 14 21 28 35 42 49 56
7.36 x 10 3 x 3.5 x 9.81
H(m) 40 40.6 40.4 39.3 38.0 33.6 25.6 14.5 o
11(%) o 41 60 74 83 83 74 51 o 60
== 4211.8 W
(a) If the pump is initially used to transfer water from one reservoir to another at the same . Power
level, determine the pump operating point if the head lost to frictional and other resistances is 35 m
Torque delIvered == - - - - - - -
Angular velocity
at 25 m 3 /min.
(b) The pump is then used to transfer water between two reservoirs having difTerence in 4211.8 x 0.65
levels of 15 m through a pipe of 0.45 m diameter. If the pipeline is 93 m long with a friction
factor of 0.004 and pipe entry and exit loss coefficients of 0.5 and 1.0 respectively, find the 10
volume flow rate and power absorbed.
== 273.7N m
(c) If the pump is now changed to one of 0.51 m diameter and the motor is changed to
one running at 975 rpm, what is the llew volume flow rate and power absorbed?
2.10 A single axial flow watcr pump has the following charactcristics: Exercise 2.2 Consider first the no-slip condition in Fig. 2.5.
Assuming blades of infinitesimal thickness the flow area may be
calculated as
o 0.136 0.233 0.311 0.388 0.466 0.608 Flow area == Impeller periphery x Blade depth
12.6 9.45 9.07 7.61 5.45 o == IT x 0.02 x 0.25
== 15.7 x 10- 3 m 2
A pumping system in which the resistance to flow is purely resistive \vith no static lift exhibits the Flow velocity C' r2 = Q/ It
samc volume flow rate whcn two of the pumps are connccted ill parallel as when they are
connected in series. What would be thc flow ratc and head that a singlc pump would deliver when 0.028
connected to the same system? 15.7 x 10- 3
= 1.78 nl/s
From the outlet velocity triangle
SOLUTIONS
C
r2
W\"2 =------
Exercise 2.1 Referring to Fig. 2.5, since there is no slip 1>2 == I>~. , tan 30
The Euler head is given by Eg. (2.2):
1.78
E == W == _u_'2 _C_x_,2 _-_U C_X_l
_l_ tan 30
ilIff 9 = 3.08 mls
 72 HYDRAULIC AND COMPRESSIRLE FLOW TURROMACHINES
HYDRA ULIC PUMPS 73

Now
and p/I == PI where
U 2 =rrDN160
n x 0.25 x 1450 /51==90-75
60 == 15
= 19 mls At inlet, tangential impeller velocity is
Absolute whirl component C x2 = U 2 - Wx2 VI == wr 1
= 19 - 3.08
= 15.92 mls
= C006~ 2n )to'OS!)
From Eq. (2.2) the Euler head is == 4.81 mls
E= U 2 C x2 - UIC xI From Fig. 2.4
9
tan Ii 1 == -C,I == -C 1 (since zero wh'Ir I)
and assuming C x1 = 0 (no whirl at inlet) VI VI
19x15.92 C I = 4.81 tan 15
E= --- ---
9.81 == 1.29 mls
= 30.83 m Volume now through the pump is
From Eq. (2.5) Q == A 1 VI
== 2nr 1bC,1
Therefore the theoretical head with slip is == 2n x 0.051 x 0.064 x 1.29
== 0.0265 m 3 Is
EN = 0.77 x 30.83
= 23.74m (b) Using the continuity equation (Eq. (1.21))

The Stodola slip factor is given by Eq. (2.6) r 1 C,t == r 2 C,2

0.051 x 1.29
nsinlJ 2 C --------
a = 1- --- --- . ,2 - 0.203
s Z[I-(C'2IU 2)cot/52J
== 0.324 m/s
Then
n sin 30 u At outlet, tangential irnpeller velocity is
0.77= 1 - - - - - - -
Z[I-(1. 78/ 19)cot300J U 2 == wr 2
whence
Z= 8.15 = C006~ 2n )to.203)
number of blades required =_8
S ay, - - - - - - - == 19.13 mls

Exercise 2.3 (a) Figure 2.5 may be used and in this case IJ~ = f32 where Total head developed by pump
Hydraulic efficiency == - - - - - - - - - - -
Theoretical head developed
IJ 2 = 90 0
- 83
H
'111 ==- (from Eg. (2.22))
E
74 HYDRAULIC AND COMPRESSIBLE i"LO\V TURBOMACIIINES HYDRAULIC PUMPS 75

If the change in potential head across the pUlnp is ignored, the total 2 2
= 28.6 _ (16.49 - 1.29 )
head developed by the pump is given by Eq. (2.1): 2y

H=e 2;/1 )+(c;~Cl) P2 - Pl = 143.5 kPa

(c) Power given to fluid = pgQH

and for an incompressible fluid the total pressure head difference is
= 10 3 x 9.81 x 0.0265 x 28.6
P02 - Po 1 =
pg
(!!2py +C~)
2g
_(~+ Cf) = H
pg 2g = 7.13 kW

Now (d) Input power to impeller = 7.43/0.89

E = 2
U C x2
= U2 (U 2 - Wx2 ) = 8.35kW
9 9
19.13 ( 19.13---
C,2 )
=-- Exercise 2.4 The velocity diagram is shown in Fig. 2.4. From the continuity
9.81 tan 7 equation, the velocity in the pipes is
= 19.13(19.13_ 0.324) V= Q/A
9.81 tan 7
0.04 x 4
= 32.1501
nx0.15 2
Therefore = 2.26m/s
J-I = 0.89 x 31.9 1
Total losses = Pipe friction losses + Other losses
= 28.6m
4flv 2 3v 2
whence =--+-
2gd 2g
P02 -POI =28.6 X 10 3 x 9.81
= 278.5 kPa = (4 x 0.005 x 40 + 3) ~2
0.15 2y
At impeller exit
g.333 x 2.26 2
C,2 2 x 9.81
C ., = U ------
x_ 2 t an fJ 2
= 2.1601
0.324 Total required head = 35 + 2.16
= 19.13 - - _ .
tan 7 =37.16nl
= 16.49 m/s
From Eg. (1.16) the speed of the pump is determined,
C 2 = (C;2 + C: 2 )1/2
= (0.324 2 + 16.49 2 )1/ 2 NQ1/2
N = _._-
= 16.49 01/S S (gH)3/4

Solving for the static pressure head 0.075(9.81 X 37.16)3/4

N=-------
(0.04)1/2
P2-P1=H_(C;-C i)
pg 2y = 31.28 rev/s
 76 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES
HYDRAULIC PUMPS 77

Flow area perpendicular to impeller outlet periphery Exercise 2.5 Cavitation began when
D
= nD x - x 0.94
10
~ + V~= 3.5
PO 2g
= 0.295 D2
and at this condition PI = Pvnp' the vapour pressure. Therefore
Now
Q Vi 1.8 x 10 3
- = 3 . 5 - - - -3
C'2=---- 2g 9.81 x 10
Flow area
0.04 = 3.317m = NPSH
Now
0.295D 2 V2
(JTH = __
1
= 0.136 rn/s 2gH
D2
3.317
Also
U 2 = NnD 36

= 31.28 x nD (JTH = 0.0921

= 98.3Dmjs
From the steady flow energy equation (Eq. (2.38)) taking the reservoir
From Eq. (2.22) the hydraulic efficiency is level as datum (2 0 = 0) we get for case (1)
Fluid power developed by pump V2
~ + -) + Z 1 = ~~ - (Sum of head losses)
'1H = Fluid power supplied to impeller PrJ 2y P{}
H Po PI
(2 1 +hr1)=--(JTH H - -
E PO pg
or
=(0.75 x 13.6)-3.317-0.1835
c _ gH
x2 - U 2'1 = 6.7m
For case (2)
9.81 x 37.16 I
98.3D x 0.76
-P2 + --
V;
+ Z 2 = -Po - lr2
pg 2y py
= 4.87jD mjs (2 2 + hf2) = (0.62 x 13.6) - 3.317 - 0.0846
The outlet velocity triangle gives = 5.03 m
C,2 Since the flow rate is the same, hr1 = hr2 and the pump must be lowered a
tan600=---
C x2 - U 2 distance (2 1 - 2 2 ) = 1.67 m at the new location.

0.136 Exercise 2.6 Referring to the velocity triangles of Fig. 2.24, i = 0, lI. 2 = 40 and
2
D (4.87jD - 98.3D) {J2 = 30 at the mean radius
whence 1.8 + 0.75
Mean diameter Dm = - - - -
D3 = 0.0495D - 0.0008 2
Impeller diameter D=0.214m
= 1.275 m
78- HYDRAULIC AND COMPRESSIULE FLOW TURUOMACIJINES HYDRAULIC PUMPS 79

(a) From Eq. (2.43) 1.362 X 10 6 x 60

250 X 2n
= 52025Nm
nND
U=- (c) At the root
60
n x 250 x 0.75
250 x n x 1.275 U =-----
r 60
60
= 9.8 m/s
= 16.69 m/s Since Cxl = 0,
Ca tan 13Ir = Ca/U r
and C x2 = - - (at the Inean radius) = 5.71/9.8
tan 40
Now At the root 13Ir = 30.2

At the tip
n x 250 x 1.8
U =-----
16.69 = 2.92C a t 60
and
Ca = 5.71 m/s (at the 11lean radius) = 23.56m/s
tan /J It = 5.71/23.56
(b) Flow rate Q through the annulus is
At the tip /J 1 t = 13.6
----------
Q = CaA
= 5.;1 ][(1.82 _ 0.75 2) Exercise 2.7 Since the density change across an axial flow fan is so low,
it may be considered to be operating with an inconlpressible fluid and
therefore the saIne equations as apply for axial flow pumps may be used.
At the mean radi us Hydraulic efficiency = H/E
C -~ From Eq. (2.51), E=0.5wZC L W oo c/2ng and putting w=2UD and
x2 - tan 40
Z = nD/s then
= 6.8In/s
16.69
E = ~ x 6.8 W/(N/s)
Now
nND
Then power transferred = pgQE u=-- at the mean diameter
60
3
10 x 9.81 x 12 x 16.69 x 6.8
n x 450 x 1.86
9.81
60
= 1.362 MW
= 43.83 m/s
Power Also
Torque = -----
Angular velocity
80 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACllfNES HYDRAULIC PUMPS 81

and substituting for E 1.2

H UC x2 NACA 2J06
1.0
rf" g
0.8
35 x 9.81 CL
C 2=----
x 0.92 x 43.83 0.6
= 8.51 m/s 0.4
Therefore from Fig. 2.24
Wx2 = U - C xl
= 43.83 - 8.51
= 35.32m/s
Now Figure 2.37 Lift coefficient versus
(J = cis = cZ/rrD -0.4 incidence angle
0.33 x 12
rr x 1.86
tan fJ 1 = Ca/U
= 0.678
fJ 1 = tan - 1 (40/43.83)
But since C xl = 0, and therefore Wxl = U, then = 42.38
W~ = C; + [(Wx2 + Wxl )/2J 2
tan f32 = tan -1 (C a/Wx2 )
= 40 2 + [(35.32 + 43.83)/2J2
f32 = tan - 1 (40/35.32)
= 3165.78 m/s
= 48.55
Woo = 56.26 m/s
Now Blade stagger angle = i + fJoo
2g 1 H = 3.8 + 45.3
CLW = - x - x -
00 rfH U (J = 49.1
35 x 2 x 9.81
Exercise 2.8 From Eq. (2.54)
0.92 x 43.83 x 0.678
System resistance = H s + KQ2
= 25.1 m/s
from which Head loss due to pipe friction only is given by
C L = 25.1/56.26 h = 4flv
2
r
= 0.446 2gd
The aerofoil data are plotted in and the corresponding incidence angle and putting
found from Fig. 2.37 V= Q/A
then

Now
i = 3.8
hr = G:~) C:2 Y Q2

sin fJa.> = Ca/W00 = (4 x 0.005 x 1000 x 16)Q2

fJoo = sin - 1 (40/56.26) 2 x 9.81 X rr 2 x 0.2 5
= 45.3 = 5164Q2 m (Q in m 3 /s)
 82 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACIIlNES HYDRAULIC PUMPS 83

70 70 . pgQH
Power to dnve pump = - -
Etllciency Yf
3
60 60 10 X 9.81 x 145 x 40.2
60 x 60 x 0.625
0

50 = 25.4kW
50

Exercise 2.9 (a) Figure 2.39 shows the head-flow and efficiency characteristics
__ 40
t::0 plotted for the speed of 750 rpm. Since water is being transferred between
40 (J

g reservoirs of the same water level, then from Eq. (2.53),

"'0
~
0
!~, System resistance = KQ2
:r: (j
c
30 30 'u
it)

~
Solving for K at the point given:
iJ.l
K = 35/25 2 = 0.056
20 // 20 Therefore the system head loss at the difTerent flow rates may be calculated:
Pump characteristic

Q(m 3 jmin) o 7 14 21 2X 35 42 49 S6
10 10 System loss (m) o 2.74 1I 24.7 43.9 6S.6

The system resistance curve is now drawn (note that it passes through
OIL..---......L.----..L l...--_ _---l...-_ _---J
o zero) and the head and now read ofT at point A. The corresponding efficiency
o 50 100 150 200 250 is read off at point n.
Volume flow rate. Q (m:l/h) At the operating point
Figure 2.38 Pump and system characteristics
Q == 26 In 3/ n1 in
Ii == 38.3 m
1/ == 81 per cent

System resistance H = 32 + 5164Q2 m (b) Sum of the head losses and static head is given by Eq. (2.54):
H==H s +KQ2
o 46 92 138 184 230
The head losses may be written as
System resistance, H (m) 32.0 32.8 35.4 39.6 45.5 53.1
4flv 2 v2 v2
Head losses == - - + k exil - + kenlrY-2--
The operating point is at the intersection of the pump characteristic and 2gd 2g 9
system resistance curves. At the operating point A in Fig. 2.38
4 x 93 x 0.004 ) v2
Q = 145 m 3 /h = ( + 1 +0.5 -
0.45 2g
v2
H = 40.2m == (3.31 + 1 + 0.5) -2
9
The efficiency corresponding to the flow rate of 145 111 3/h is
v2
'7 = 62.5 per cent == 4.81 -
2g
84 HYDRAULIC AND COMPRESSIBLE FLOW TURROMACIfINES HYDRAULIC PUMPS 85

50
,100
Including the static Jift

System characteristic
System Joss = 15 + 9.69Q2 m
with zero static head The head Joss is now dctcrrnincd for thc various flo\v rates.
90

40 I 80
I
I Q(m 3 jmin) o 7 14 21 28 35 42 49
I

~
I lI(m) 15 15.1) 15.52 16.18 17.11 IH.) 19.75 21.47
... --- .../.....
I
I
", \
,
I
\
____________ 70
t
The new system resistance curve is drawn noting that it begins at H = 15 m.
30 I
I
I
I

---------j 60
The operating point is at point C and the corresponding efficiency at point D.
,, At the operating point
I

~
t
c
d)
U
g I

,,
I

"0
50 ],
t >-.
~
:t t
I u
c::
d)
H = 20.4m
I Tj
~
I
20 40
'1 = 68.4 per cent
I
E\ ~ 30
\ I
, I
Power absorbed = pgQH
--,I
\
\ ,
I
'1
975 rpm \ \ 3
10 I -20 10 X 9.81 x 45 x 20.4
I
,
I
0.684 x 60
I
\ 975 rpm
.....---
I 10 =219.4kW
I
I
I
I
I
I (c) Since we have static lift, it is necessary to construct part of the
o '.-_----L_ _----l....-_------L'..L.-_--L_ _-L-_..l.....-..l.....----.J o characteristic at the new speed of 900 rpm. The corresponding points for the
o 10 20 30 40 50 60
Volume flow rate (m 3 /min) new impeller and the new speed are found from Eg. (1.6):
Figure 2.39 Pump characteristics at 750 and 900 rpm
and

whence
975) (0.51)3
Now Q2 = Ql ( 750 0.7
v= Q/A
and substituting for v = 0.503Ql
4.81 X 4
Head losses = [ 2 x 9.81 (n x 0.45 2)2
2
l Q
2
H =}-I
2
975 x 0.51)2
1 ( 750 x 0.7

= 9.69Q2 m
86 HYDRAULIC AND COMPRESSIBLE FLOW TURUOMACHINES HYDRAULIC PUMPS 87

QI 0 7 14 21 28 35 42 49 56
30
Q2 0 3.5 7.1 10.6 14.2 17.7 21.3 24.8 28.3
HI 40 40.6 40.4 39.3 38 33.6 25.6 14.5 0
H2 36 36.5 36.4 35.4 34.2 30.24 23.0 13.1 0 25

The new characteristic is dra\vn and also the efficiency curve by nloving the
corresponding values of efficiency horizontally across. The operating point 20 Pumps in series
is at E and the corresponding efficiency at G.
At the operating point
g
:t:
-d 15
~
0
::r: ,
11== 16.51n
\
, ..... A
10 '-- .....
"-
'/ == 62.4 per cen t " '\B
'\
Pumps in parallel
5 \
'\
23.75 x 16.5 x 10J x 9.81 System '\

Power absorbcd == -- - ---- curve Single ~

0.624 x 60 pump "
0
== 102.7 kW 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Volume flow rate, Q (m 3 /s)

Exercise 2.10 The single pun1p characteristic is plotted in Fig. 2.40 along Figure 2.40 Axial now pump characteristics when connected in series and parallel
with the characteristics for the pun1ps connccted in parallel and series.
Since the san1e pUlnp is used in both cascs, for the series connection the flow
rate through the two pumps remains the saIne while the head is doubled From Eq. (2.54) for the system curve
and for the parallel connection the head across the purnps rcrnains the same H==H s +KQ2
while the flow rate is doubled.
But H s == O. Therefore
Series connection H == 0.48 2 K
and
Q(m 3 js) 0 0.136 0.233 0.311 0.388 0.466 0.608
K == 42.32
H(m) 25.2 18.9 19.58 18.14 15.22 10.9 0
The system characteristic is parabolic and may now be drawn in for various
heads and flow rates and the point B gives the operating point for the single
Parallel connection
pump within the system.
Q(m 3 /s) 0 0.272 O.4()6 0.622 0.776 0.932 1.216
System now rate (m 3 js) 0 0.1 0.2 0.3 0.4 0.5 0.6
H(m) 12.6 9.45 9.79 9.07 7.61 5.45 0
System head loss (01) 0 0.42 1.69 3.81 6.77 10.58 15.23

At point A both connections give the sanlC head and flow and the systen1
characteristic must pass through this point and zero since there is no static lift. The single pump operates at point B:
At operating point A in Fig. 2.40
Q == 0.41 m 3 Is
3
Q == 0.48 n1 /s
H == 9.75 III H == 7m
 HYDRAULIC TURBINES 89

Table 3.1 Operating range of hydraulic turbines

CHAPTER
Francis Kaplan
THREE Pelton Wheel turbine turbine

Dimensionless srccific 0.05-0.4 0.4-2.2 1.8-4.6

HYDRAULIC TURBINES speed (rad)
Head (01) 100-1700 80-500 up to 400
Maximum rower (k\V) 55 40 30
Best efficiency (%) 9J 94 94
Regulation Spear nozzle and Guide Blade
mechanism dcncctor rlate vanes stagger

types of axial now turbines exist, these being the propeller and Kaplan
turbines. The former has fixcd blades whereas the latter has adjustable blades.
Table 3.1 summarizcs the head, power and efficiency values that are typical
but by no means maxima for each type of turbine.
A reversiblc pump-turbine can operate as either a pump or a turbine
and is used in pump-storage hydroelectric schemes. At times of low electricity
demand (e.g. during the night) cheap electricity is used to pump water from
3.1 INTRODUCTION the low- to the high-level rescrvoir. This watcr may then be used during the
day for power generation during peak periods, when the unit runs as a
Turbines are used for converting hydraulic energy into electrical energy. The turbine in the reverse direction.
capital costs of a hydroelectric power scheme (i.e. reservoir, pipelines, turbines, One of the largest pump-storage schemes in the world is at Cabin Creek
etc.) arc higher than thermal stations but they have many advantages, some in Colorado, where each turbine generates 166 MW with a head of 360 m.
of which are: In the sections that follow, each type of hydraulic turbine will be studied
separately in terms of the velocity triangles, efficiencies, reaction and method
1. High efficiency of operation.
2. Operational flexibility
3. Ease of maintenance
4. Low wear and tear 3.2 PELTON WHEEL
5. Potentially inexhaustible supply of energy
6. No atmospheric pollution The Pelton wheel turbine is a pure impulse turbine in which a jet of fluid
issuing from a nozzle impinges on a succession of curved buckets fixed to
The main types of turbines used these days are impulse and reaction the periphery of a rotating wheel, as in Fig. 3.1, where four jets are shown.
turbines. The predominant type of impulse machine is the Pelton wheel, The buckets deflect the jet through an angle of between 160 and
which is suitable for a range of heads of about 150-2000 m. One of the largest 165 0 in the same plane as the jet, and it is the turning of the jet that causes
single units is at the New Colgate Power Station, California, with a rating the momentum change of the nuid and its reaction on the buckets. A bucket
of 170 MW. Reaction turbines are of two types: is therefore pushed away by the jet and the next bucket moves round to be
similarly acted upon. The spent water falls vertically into the lower reservoir
1. Radial or mixed flow or tailrace and the whole energy transfer from nozzle outlet to tailrace takes
2. Axial flow place at constant pressure. Figure 3.2 shows a large Pelton wheel with its
buckets.
Of the radial flow type the Francis turbine predominates, a single unit at A diagram of a Pelton wheel hydroelectric installation is shown in
Churchill Falls having a power output of 480 MW with a head of 312 m. Two Fig. 3.3. The water supply is from a constant-head reservoir at elevation H 1
88
90 HYDRAULIC AND COMPRESSIBLE rU)\V TURBOMACIII~LS

Figure 3.2 Pelton wheel (courtesy of

Escher Wyss Ltd)

Figure 3.1 Elements of a Pelton wheel turbine (('(wries)' (~j' /:"sc},el' ~V)'ss Lid)

above the centre-liIle of the jet. A shallow-slope pressure tunnel extends from
the reservoir to a point alITIOst vertically above the location of the turbine.
A pipe of almost vertical slope called the penstock joins the end of the
Surge tan_;-- ~ i h,

pressure tunnel to the nozzle, while a surge tank is installed at the upper
end of the penstock to damp out flow control pressure and velocity transients.
It is emphasized that, compared with the penstock, the pressure tunnel could / I
I
!
I
I

Pressure tunnel
be extremely long, its slope is extrcI11cly shallo\v and it should undergo no ,I
II
I
large pressure fluctuations caused by inlet valve flow control. The penstock
H' H
must be protected against the large pressure fluctuations that could occur :I
between the nozzle and surge tank, and is usually a single steel-lined concrete Penstock

1J
pipe or a steel-lined excavated tunnel. At the turbine end of the penstock
is the nozzle, which converts the total head at inlet to the nozzle into a
water jet with velocity C 1 at atll10spheric pressure. Datum
The velocity triangles for the flow of fluid onto and off a single bucket are
shown in Fig. 3.4. If the bucket is brought to rest, then subtracting the bucket
speed U 1 from the jet velocity C 1 gives the relative fluid velocity It'l onto the
bucket. The angle turned through by the jet in the horizontal plane during its Figure 3.3 Pelton wheel hydroelectric installation
passage over the bucket surface is Y. and the relative exit velocity is W z . If the
91
92 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES HYDRAULIC TURBINES 93

t' "
I "

t
)
V
-r
WI

)
-1:I \
\
/ Theoretical

j. Ct
J /7 Actual

zJ
V
C2
Figure 3.4 Velocity triangles for a
Pelton wheel

0.5

bucket speed vector U 2 is added to W2 in the appropriate direction, the Figure 3.~ Efficiencies and jet speed ratio of a Pelton wheel
absolute velocity at cxit, C 2' results. It should be rcalizcd that the component
C x2 of C 2 can bc in the positive or negative x direction depending on the
magnitude of U. Eq. (3.1) becorl1cs
From Euler's turbine equation (Eq. (1.24))
E = U(C I - U)(t - kcosa)/y
W/nl = U IC xl - U 2 C x2
where k is the relative velocity ratio W 2 /W I
and since in this case C x2 is in the negative x direction, If the hydraulic efficiency is defined as
W/m = U{(U + WI) + [W2 cos(180 - a) - UJ} Energy transferred
'1 - -_.--
Assuming no loss of relative velocity due to friction across the bucket surface II - Energy available injet
(WI = W2 ), then
= E/(Ci/ 2g) (3.4)
W/n1 = U(W1 - WI cosa)
then if C( = 180, the maximum hydraulic efficiency is 100 per cent. In practice,
Therefore the deflection angle is in the order of t 60-165 to avoid interference with
E= U(C I - U)(1-cosa)/g (3.1) the oncoming jet and 1JH is accordingly reduced. Figure 3.5 shows the
theoretical efficiency as a function of speed ratio. The overall efficiency is
the units of E being watts per newton per second weight of flow.
lower than the theoretical as well as having a reduced speed ratio at maximum
Equation (3.1) can be optimized by differentiating with respect to U. Thus
efficiency. This is due to pipeline and nozzle losses, which will be discussed
dE/dU = (1 - cos a)(C I - 2U)/g = 0 in a later section.
for a maximum, and then
3.2.1 Pelton Wheel Load Changes

or Hydraulic turbines are usually coupled directly to an electrical generator and,

since the generator must run at a constant speed, the speed U of the turbine
(3.2)
must remain constant when the load changes. It is also desirable to run at
Substituting back into Eq. (3.1) gives maximum efficiency and therefore the ratio U/C I must stay the same. That
is, the jet velocity must not change. The only way left to adjust to a change
E max = Ci( 1 - cos a)/4y (3.3)
in turbine load is to change the input water power.
In practice, surface friction of the bucket is present and W 2 #- WI' Then The input water power is given by the product pgQH' but H' is constant
94 HYDRAULIC AND COMPRESSIBLE I'LO\-\, TURBOMACIIINLS HYDRAULIC TURllINES 95

when expressed as kinetic energy per unit weight of flow. Now

Fl0~g_~_~ __ *_~~_ or
.. " . Energy at end of pipeline
PipelIne tranSrl11SSl0n effiCiency = - - -- -- -- -----.
Energy available at reserVOIr

High load Low load I/lrans = (H 1 - hdlH 1 = HIH 1 (3.5)

and
.. . Energy at nozzle outlet
Nozzle enlclency = ------- ---- - - ----

Deflector plate in
normal Position~",
" '
,,
< Bu~kel
trajectory
\\ Fully J<:lkcl<:J j<:l
~ \", .
~ '" '" /
,~
II
or

So
Energy at nozzle inlet

, -
-----~~--'==~~-~--.-==-~~~-.: --
-
~::-,:~~",,- ~-=--
~~\~~
Nozzle and pipe transrnission efficiency = (HI H d(H'1H) = Ci/ 2g H 1

~"~"~,
Also
Actualjet velocity
Nozzle velocity coefficient = --------- -------
Figure 3.6 Load control by a spear valve and deflector plate Theoretical jet velocity
or
C v = C 1/(2{jH)l i 2 (3.7)
Therefore the nozzle efficiency becomes
(and therefore C 1), and the only renlaining variable is Q. The change in flow
rate is effected by noting that Q = C 1 A, where A is the nozzle area. Since C 1 lIN = C~/2gH = C; (3.8)
is constant, then the cross-sectional area of the nozzle In ust change. This is
The characteristics of an impulse turbine are shown in Figs 3.7 and 3.8.
accomplished by a spear valve, which alters the jet cross-sectional area as
These curves are drawn for a constant head and it is seen that peak power
illustrated in Fig. 3.6. The position of the spear is controlled by a
and efficiency occur at about the same speed ratio for any gate opening and
servo-mechanisn1 that senses the load change. For a sudden loss of load, a
that the peak values of efficiency do not vary much. This is due to the nozzle
deflector plate rises to renlove the jet totally froIll the buckets and to allow
velocity remaining constant in nlagnitude and direction as the flow rate
time for the spear to nlove slowly to its ne\v load position. This prevents
changes, giving an optinlum value of UIC 1 at a fixed speed. Windage,
excessive overspeeding.

3.2.2 Pelton Wheel Losses and Efficiencies

Head losses occur in the pipelines conveying the water to the nozzle and are 90
Constant head Nozzle setting
composed offriction and bend losses. Losses also occur in the nozzle an? these 4
80
are expressed in ternlS of a velocity coefficient C". Final,ly tI~ere are \v!nda~e
and friction losses in the wheel itself. l~he total head llne IS shown In FIg.
3.3, \vhere the water supply is froIn a reservoir at a head II, above the nozzle.
As the fluid Illoves through the pressure tunnel and penstock up to entry to 60
the nozzle, a frictional head loss hr occurs. A further head loss hin due. to
losses in the nozzle takes place so that the head available for po\ver generatIon 50
at exit froIn the nozzle is H':
N
H' = j-J 1 - (h r + hin )
= C~/2{j Figure 3.7 Efficiency versus speed at various nozzle settings
 "

l'

96 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES

HYDRAULIC TURBINES 97

fluid st rcanll ines a rOll nd t he edge of t he blades. The essential difference

Nozzle setting
P Constant head between the reaction rotor and itllpulse \vheel is that in the former the water,
1
under a high static head, has its pressure energy transformed into ki~etic
2 energy in a nozzle, \vhich in itself forms part of the rotor. Therefore, SInce
3 a static pressure drop occurs across the turbine rotor, part of the work done
4 by the fluid on the runner is due to reaction from the pressure drop, ~nd
part is due to a change in kinetic energy, which repre.sents a.n impulse.functlon.
Figure 3.10 shows a cross section through a FrancIs turbIne. A typIcal rotor
is also shown.
The total head of the radial flow turbine ranges from about 500 down
to 30 m, and the dilncnsionless specific speed froln 0.3 to 2.5 rad. The water
first enters a volute or spiral casing. It then passes through a row of fixed
guide vanes follov/ed by adjustable guide vanes, the cross-sectional area
between the adjustahle vanes being varied for OO\V control at part load. The
Figure 3.M Pelton wheel power output
N versus speed a t va rious nozzle settings
water then passes ilnnlediately into the rotor where it moves radially through
the rotor vanes and exits frolll the rotor blades at a smaller diameter, after
which it turns through 90') into the draft tuhe. The function of the draft tube
is to bring the \vatcr pressure back to the pressure of the tailrace, and during
100 this process to rel110VC the kinetic energy still existing at the runner outlet.
In some rotors the work transfer is acconlplished not only while the water
80 is moving radially but also in a part axial direction. This is done by a
Constant head judicious choice of rotor design.
~ 60 In considering the now through the runner of a radial flow turbine and
the theoretical analysis with respect to the velocity triangles, the same
40 assumptions will be made as for the centrifugal pump. Figure 3.11 shows
the total energy line through the complete system from supply reservoir to
20
tailrace. The free surface of the reservoir is often called the 'head water'. The
fluid experiences a frictional head drop hfp in the pipeline up to the inlet
o 25 50 75 100 Figure 3.9 Variation of Pelton flange of the turbine at point O. At this point the fluid enters the volute, the
Percentage of full load wheel efficiency with load fixed inlet guide vanes and also the adjustable guide vanes, where a further
frictional head loss hJl. is experienced. Upon leaving the inlet guide vanes,
the fluid moves into the runner where a frictional head loss hr takes place
mechanical losses and variations in loss coefficients cause the small variations.
while energy is supplied to the runner and shaft. The total energy line
In practice, one is usually more interested in the fixed-speed condition since
decreases to the point 3 where the fluid exits from the draft tube with a
the generators run at constant speed. Figure 3.9 shows that the
frictional head loss ltd occllrring in the draft tube, and a residual kinetic
variation of efficiency with load is slight, except at low loads where the
energy loss V 2 /2g at exit from the draft tube. In considering the overall
decrease is due to changes in nozzle efficiency, and at high loads where the
turbine efficie;cy, the inlet volute and draft tube are considered to be parts
increased jet diameter gives rise to higher bucket losses.
of the turbine.
The inlet and outlet velocity triangles for the runner are shown in Fig.
3.12. This shows the water emerging from the inlet guide vanes at radius r i
3.3 RADIAL FLOW TURBINE
with absolute velocity C l ' at an angle at to the direction of rotation. The
tangential velocity at inlet is U I' and by subtracting U I from C I the relative
The radial flow or Francis turbine is a reaction machine and, to achieve
velocity vector WI is obtained at an angle fJ 1 to the direction of rotation.
reaction, the rotor must be enclosed by a casing to prevent deviation of the
PI is also the inlet blade angle for shock-free entry. At the outlet radius r 2 ,
 f

I
I. Volule
y~. Fixed \luitl' .
I
J. A ' . e- C vanes
dJuslablc gu' 1
4. Runner It e vanes ____ TOlal i
e.nergy I
_ ~ . ,C), Dral"llube hne
2 - - t-(L Support b''
C.lllllg I
IH

Draft tubc

Tailracc t
'/ ------- f
J ';/2f~
( a)
Figure 3.11 Energy d', '
lstnbulion tl Hough a h Yd' '
rdulrc rC'1
l C't'IOn turhine

Figure
(b) 3.10 A F'
turbin . turbine (.
r.lI1CIS Escher
e runncr (OUre es.r 0./' Figure 3.12 Vel . triangles
for a F . oClty
rancis turbine
98
99
100 HYDRAULIC AND COMPRESSIIJLE FLOW TURIJOMACHINES HYDRAULIC TURBINES 101

the water leaves the blade at angle /32 to the tangential velocity vector, the
resultant of W 2 and U 2 being the absolute outlet velocity C 2 The flow
velocities C r1 and Cr2 are directed towards the axis of rotation and are given
by Q/2rrr 1 b 1 and Q/2rrr 2 h 2 respectively, where h is the height of the runner.
Euler's turbine equation (Eq. (1.24)) gives
length. As the length increases, hd also increases, and since the angle of
divergence should not exceed about 8, to ensure that separation of the
boundary layer docs not occur, the draft tube could be very long. There is
also an upper limit on the value of 2 2 due to cavitation, and this is discussed
in a later section.

(3.9)
3.3.2 Turbine Losses
and E is a maximum when C x2 is zero, that is when the absolute and flow
velocities are equal at the outlet. The losses can once again be related in terms of an energy balance through the
turbine:

3.3.1 Net Head Across Reaction Turbine (3.14)

The net head H across the turbine is the difference in the total head where Ps = shaft power output, Pm = mechanical power loss, P r = runner
between the inlet flange and the tail water level. Gross head HI should not power loss, Pc = casing and draft tube loss, PI = leakage loss and P = water
be confused with net head. Thus power available. Together, Pr + Pc + PI is the hydraulic power loss.
Runner power loss Pr is due to friction, shock at impeller entry and flow
Turbine total inlet head = Po/py + V~/2{J +20 separation, and results in a head loss hr associated with a flow rate through the
and runner of Qr:
Turbine total outlet head = P3/P{] + V;/2r} +23
(3.15)
where the pressures are gauge pressures. Summing these,
Leakage power loss PI is caused by a flow rate q leaking past the runner
Total head across turbine = (Po - P3)P{} + (V~ - V;)/2g + (2 0 - 2 3) and therefore not being handled by the runner. Thus
=H
(3.16)
But in the tailrace P3 is atmospheric and 2 3 is zero. Therefore
and, with a total head /-1 r across the runner, the leakage power loss becomes
(3.10)
Also (3.17)
H = H 1 - hfp - V;/2g (3.11) Casing power loss l'c is due to friction, eddy and flow separation losses
and the energy given to the runner by the water per unit weight of flow is in the casing and draft tube. If this head loss is he then

W/mg = H - hd - hg - hr (3.12) Pc = poQh c (N m/s) ( 3.18)

If the water discharged directly into the tailrace from the runner outlet, The total energy balance of Eq. (3.14) thus becomes
the kinetic energy lost would be high. By fitting the draft tube between runner P{}Q 1-/ = Pm + pg(hrQr + hcQ + Hrq + Ps) (3.19)
outlet and tailrace, a continuous stream of water is formed between the two.
The tailrace velocity is reduced because of the increase in cross-sectional Then we have
area of the draft tube, and, because the tailrace pressure is atmospheric, the . Shaft output power
runner outlet pressure must now be below atmospheric pressure. Applying O vera 11 e fliICIency = - - - - - - - - - - - - -
Fluid po\vcr a vailable at inlet flange
the energy equation between the runner outlet and tailrace gives
or
P2/pg + V~/2g + 2 2 = P3/pg + V~/2g + 2 3 + hd
(3.20)
and putting P3 and 2 3 equal to zero and
P2/pg = (V; - V;)/2g - 2 2 + hd (3.13) m'
H yd rau IIC e ICIency = - - Power
- - -received
- - -by-runner
----
There is a limit to the amount that V3 can be reduced because of tube Fluid power available at inlet flange
 102 HYDRAULIC AND CO~tPRLSSIBLL HJ)\V TURBOMACIIIl\:I:S
or
(3.21)
The term (P s + P m)/ pgQH of Eq. (3.21) is the theoretical energy transfer per
unit weight of fluid flo\v. Therefore for n1axilllU111 efficiency
'111 = U I C.'d/u H
3.3.3 Characteristic Curves
Curves of water power input, torq ue exerted by the wheel, flow rate, efficiency
and brake power output for a constant gate opening are drawn in Fig. 3.13.
Inc0111 pa rison \v it h the i 111 PuIse whee I, the fl 0 \V rat cis no Ionger
independent of the wheel speed since there is now an unbroken flow of water
down to the tailrace and changes \vithin the runner afTecl the flow. Of most
interest, however, is the behaviour of the turbine at constant speed, since the
generator runs at a fixed speed. As the electrical load changes, so the flow
rate is changed by variation of the gate opening. FrOll1 a set of characteristic
curves at different gate openings, the constant-speed characteristic of Fig.
3.14 is constructed. It is seen that the head increases sligh tly as the load is
decreased due to the friction head loss, \vhich, being proportional to Q2, is
less at lighter loads. I t will also be noted tha t the efficiency curve at constant
speed is not as flat as for an in1pulse turbine.
When a reaction turbine operating at constant speed experiences a load
decrease, the cross-sectional area between the inlet vanes changes and angle a 1
Torque
Fluid power input
./
Flow rate. Q
Brake power
I output '
Emciency
N(rpm)
Figure 3.13 Reaction turhine characteristics at full gate opening
HYDRAULIC TURBINES 103
EllIciency
I lead
\
Water powcr input
Brake powcr output
Figure 3.14 Reaction turbinc characteristics at constant speed but varying gate opening
decreases. The inlet runner area is constant and therefore to satisfy continuity
the relative velocity WI ITIUSt decrease. The result is that the flow onto the
runner is no longer shock-free and at exit C 2 may increase. This gives a
higher kinetic energy loss at runner exit as well as increasing the whirl
component C x2 down the draft tube. The flow is then spiral in nature, which
decreases the draft tube efficiency. The efficiency of a reaction turbine at light
loads therefore tends to be less than that of the Pelton wheel, although the
design maximum efficiency ITIay be greater.
3.4 AXIAL FLOW TUI~llINE
The energy distribution through an axial flow hydraulic turbine is the same
as in Fig. 3.11. However, the velocity triangles are markedly different since
the fluid is assumed to flow from blade inlet to outlet at a constant radius.
A Kaplan turbine is illustrated in Fig. 3.15. The inlet guide vanes are fixed
and are situated at a plane higher than the runner blades such that the fluid
0
must turn through 90 to enter the runner in the axial direction. Load changes
are effected by adjustment of the runner blade angle. The function of the
guide vanes is to impart whirl to the fluid so that the radial distribution of
velocity is the same as in a free vortex. Since this type of turbine is used for
low heads and high flow rates, the blades must be long and have large chords
so that they are strong enough to transmit the very high torques that arise.
Pitch/chord ratios of 1-1.5 are typical for axial flow turbines and this results
in a four-, five- or six-bladed runner.
 104 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIfINES HYDRAULIC TURBINES 105

Figure 3.16 Velocity triangles for an axial now hydraulic turbine

I. Spiral casing with fixed guidc vancs
2. Adjustable inlet guide vanes
J. Transition passage
If E is constant along the blade radius, and C a is constant over the
4. Runner cross-sectional area, then as U 2 increases from hub to tip, U cot f31 must
5. Draft tubc decrease to keep Eq. (3.23) constant. That is, fJ 1 must increase from hub to
tip and the blade must therefore be twis.ted. ., .
Figure 3.15 Axial flow Kaplan turhinc (courtesy of Escher Wyss Ltd)
The characteristic curves for the aXIal nO\V Kaplan turbIne a~e sI~llar
to those of the radial now turbine, and a comparison of the efficIencIes of
The velocity triangles are usually drawn at the mean radius, since
conditions change from hub to tip, and are shown in Fig. 3.16. The flow 100
velocity is axial at inlet and outlet and hence Crl = C r2 = Ca' The blade Kaplan
velocity vector U 1 is subtracted from the absolute velocity vector C 1 (which
90
is at angle (.(1 to U.) to yield the relative velocity vector WI' For shock-free
entry onto the runner, W 1 is at the blade angle 1/ t. For maximum efficiency, Pclton wheel
the whirl component C x2 is zero, in which case the absolute velocity at exit 80
-.
is axial, and then C 2 = C r2 . Euler's turbine equation (Eq. (1.24)) gives ~
~

Ii. = U(C x1 - Cx2 )lg (liN) u

~
4)
70
'0
and for zero whirl at exit ~ Prorel1cr (fixed blade)
L.LJ
60
(3.22)
Now

and therefore
40 60 80 Figure 3.17 Comparison of hy-
20
(3.23) Perccntage of design power draulic turbine efficiencies
 HYDRAULIC TURBINES 107
106 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACIIINLS

the impulse and two reaction \vheels at part load is shown in Fig. 3.17. Of 2..--------------,
Propeller
the three, the impulse wheel efficiency curve is nluch nattcr although the
maximum efficiency is lower. The Francis turbine peaks at the highest Kaplan
efficiency but falls off rapidly at part load, while the Kaplan turbine has a
much flatter curve than the Francis and exhibits a siInilar Inaxin1uIn efficiency.
0.4
The advantage of adjustable blades on the Kaplan turbine is shown by
comparing it with the curve for a fixed-blade propeller turbine (also shown
0.2 Safe region
in Fig. 3.17).
b(.l 0.1

3.5 CAVITATION IN TUl~BINES

Turbine cavitation occurs on the suction surfaces of the blades, at the runner
outlet, where the static pressure is a nlinimunl and the absolute velocity high.
Although it has little if any effect on the perfornlance of the turbine since it
occurs after the runner, it should be avoided if at all possible. Referring again
to Eq. (3.13), as the outlet velocity V2 increases then P2 decreases and has its
lowest value when the vapour pressure is rcached. At this pressure cavitation
begins, and putting P3 equal to Palm and P2 equal to [Jvap' Eq. (3.13) becolnes
[(V~ - V~)/2fJJ - Ill! == (Palm -"Pvap)/PO - 2 (3.24)
Division of Eq. (3.24) by the net head across the turbine gives the Thoma
cavitation parameter for the turbine:
a == [(Palm - Pvap)/PY - Z2J/J-I == (NPSII)/fI (3.25)
The critical value of N PSH at which cavitation occurs is determined
from a test on a ITIodel or full-size 111achine in which P2 is decreased until
the minimunl value at which ca vitation begins or the efficiency suddenly
decreases is found. Knowing Z 2 and H it is easy to cOInpute the critical
value o"c' which is the value below which 0", as given by Eg. (3.25), for any
other siInilar nlachine of the SHIne hOInologous series lnust not fall.
Equation (3.25) shows that the n1aximum elevation of the turbine above
the tailrace is given by
Z 2 == (Palm - Pvap)/ PY - a cJ-I (3.26)
Equation (3.26) indicates that, as the net head is increased, so the turbine
elevation above the tailrace Inust be decreased. For an exccssive net head,
2 2 might be negative, which ilnplics that excavation would be needed to
place the turbine below the level of the tailrace.
The dependence of aeon thc dilllcnsioniess specific speed and thus on the
design of the turbine is shown in Fig. 3.18 where turbincs of high N s have
a high o"c and nlust therefore bc sct lower than thosc of smaller N s The
similarity relationships used for pllInps in connection with cavitation may
also be used for turbines.
0.04
Francis
0.02
0.01 L-.4--1.--1.--L-- - - - 1 .- -.. . . .--''---'8
0 l 2 4
Ns x 10- 1 (rev) Figure 3.18 Cavitation limits for reaction turbines
EXERCISES
3.1 A generator is to be driven by a small Pelton wheel with a head of 91.5 m at inlet to the nozzle
and discharge of 0.04 m 3 /s. The wheel rotates at 720 rpm and the velocity coefficient of the
nozzle is 0.98. If the efficiency of the wheel (based on the energy available at entry to the nozzle)
is 80 per cent and the ratio of bucket speed to jet speed is 0.46, determine the wheel-to-jet-diameter
ratio at the centre-line of the buckets, and the speed of the wheel. What is the dimensionless
power specific speed of the wheel?
3.2 A reservoir with a height of 280 In is connected to the powerhouse of a hydroelectric plant
through three pipes each 2.5 km long and with friction factor 0.006, in which the head loss is
not to exceed 34 m. It is a requirement that a total shaft output of 18 MW be developed, and
to achieve this it is decided to install a number of single-jet Pelton wheels, each with a
dimensionless specific speed not exceeding 0.23 rad. The ratio of bucket speed to jet speed is
0.46, while the wheel speed is to be 650 rpm. If the nozzles have a discharge coefficient of 0.94
and velocity coefficient of 0.96, and assuming that each wheel has an overall efficiency of 87
per cen t, find
(a) the number of Pelton wheels required,
(b) the wheel diameter,
(c) the jet nozzle diameter and
(d) the diameter of the supply pipes.
3.3 The buckets of a Pelton wheel deflect the jet through an angle of 170, while the
relative velocity of the water is reduced by 12 per cent due to bucket friction. Calculate
the theoretical hydraulic efficiency from the velocity triangles for a bucket/jet speed ratio of
0.47. Under a gross head of 600 m the wheel develops 1250 kW when the loss of hend due to
pipe friction between the reservoir and nozzle is 48 m. The bucket circle diameter of the wheel is
 HYDRAULIC TURBINES 109
108 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES

(b) Determine the outlet angJe of the rotor hlade if tl~e whi.rl component at .outlet is zero.
900 mm and there arc two jets. The nozzle velocity coefficient is O.9R. Find the speed of rotation
(c) Calculate the theoretical power output if the whirl at II1let and outlet IS the same at
of the wheel and the diameter of the nozzles if the actual hydraulic efficiency is 0.9 times that
calculated above. all radii.
3.4 An electricity generating installation uses a Francis turbine with a rotational speed of t t A hydf'\ulic turhine is designed to run at 300 rpm under a net head of 50 m and to prdo~uc.e
3. ' . f 1 n' t '" .. 'ted to be 104 mls an hIt IS
1260 rpm. The net head across the turbine is 124 m and the volume now rate is 0.5 m 3 Is. The 2 MW of power. The runner outlet velOCity 0 t le Ull IS expel:. . .
roposed to situ'\te the runner outlet at a height of 6 m ahove the tal1ra~e. The atmosp .erIC
radius of the runner is 0.6 m, the height of the runner vanes at inlet is 0.03 m and the angle of
p .'. u 'Iv,\'lent t() 101m of water and the saturation pressure of water IS 4 kPa. Determme
the inlet guide vanes is set at 72 from the radial direction. Assuming that the absolute flow pressure IS eq ( ..
velocity is radial at exit, find the torque and power exerted by the water. Calculate (a) whether cavitation is likely to occur. ' . . . '. . , 'voided and
the hydraulic efficiency. (b) the limiting height of the runner outlet If cavitatIOn IS to .be ~llst.,\.
'l I n cavitation IS Just about to occur.
3.5 An inward now radial turbine has an overall efficiency of 74 per cent. The net head H
(c) the head loss between runner outlet an d tal race w 1e
across the turbine is 5.5 m and the required power output is 125 kW. The runner tangential Critical cavitation parameters arc given below:
velocity is 0.97(2gH)1/2 while the now velocity is 0.4(2gH)'/2. If the speed of the runner is 230 rpm
with hydraulic losses accounting for IR per cent of the energy a vailable, calculate the inlet o 0.049 0.096 0.144 0.192 0.24
guide vane exit angle, the inlet angle of the runner vane, the runner diameter at inlet and the o 0.04 0.1 0.18 0.28 0,41
height of the runner at inlet. Assume that the discharge is radial.
3.6 A Francis turbine has a diameter of 1.4 m and rotates at 4JO rpm. Water enters the
runner without shock with a now velocity (Crt) of 9.5 mls and leaves the runner without
whirl with an absolute velocity of 7 m/s. The dilTerence between the sum of the static
and potential heads at entrance to the runner and at the exit from the runner is 62 m. If the
turbine develops 12250 kW and has a flow rate of 12 m 3 Is of water when the net head is SOLUTIONS
115m, find
(a) the absolute velocity of the water at entry to the runner and the angle of the inlet Exercise 3.1 Overall eOiciency
guide vanes,
(b) the entry angle of the runner blades and POVJcr developed
(c) the head lost in the runner. 110 == -I~ower availabl;
3.7 An inward flow vertical shaft reaction turbine runs at a speed of 375 rpm under an available
net total head from inlet flange to tailrace of 62 m. The external diameter of the runner is 1.5 m p == fJOQ H1 10
and the dimensionless power specific speed based on the power transferred to the runner is == t 0 3 X 9.81 x 0.04 x 91.5 x 0.8
0.14 rev. Water enters the runner without shock with a flow velocity of 9 mls and leaves the
runner without whirl with an absolute velocity of 7 m/s. It discharges to the tailrace with a Power developed == 28.72 kW
velocity of 2.0 m/s. The mean height of the runner entry plane is 2 m above the tailrace level
while the entrance to the draft tube is 1.7 m above the tailrace. At entrance to the runner the Velocity coefficient
static pressure head is 35 m above atmospheric pressure, while at exit from the runner the static C1
pressure head is 2.2 m below atmospheric pressure. Assuming a hydraulic efficiency of 90 per C == - ------
v ( 2(/ Ii) 1/2
cent, find
(a) the runner blade entry angle, C == 0.9R(2 x 9.81 x 91.5)1/2
t
(b) the head lost in the volute casing and guide vanes, in the runner and in the dr~ft tube and
(c) the draft tube entry diameter. == 41.52 m/s
3.8 An axial flow hydraulic turbine has a net head of 23 m across it, and, when running at a Therefore
speed of 150 rpm, develops 23 MW. The blade tip and hub diameters are 4.75 and 2.0 m u == O.46C 1
respectively. If the hydraulic efficiency is 93 per cent and the overall efficiency 85 per cent,
== 0.46 x 41.52
calculate the inlet and outlet blade angles at the mean radius assuming axial now at outlet.
3.9 A Kaplan turbine operating under a net head of 20 m develops 16000 kW with an overall == 19. t m/s
efficiency of 80 per cent. The diameter of the runner is 4.2 m while the hub diameter is 2 m and
the dimensionless power specific speed is 3 rad. If the hydraulic efficiency is 90 per cent, calculate
the inlet and exit angles of the runner blades at the tip and at the hub if the now leaving the Also
runner is purely axial. U == (I)D where D is wheel diameter
3.10 An axial flow turbine with tip and hub diameters of 2.0 and 0.8 m respectively rotates at 2
250 rpm. The turbine is fitted with fixed stator blades upstream of the rotor and at the mean
2x19.1x60
diameter these are set at 42 to the direction of blade rotation. Also, at the mean diameter and D==------
measured from the direction of blade rotation, the blade angle at inlet is 148. 720 x 2rr
(a) Assuming a constant axial velocity across the annulus, what is the now rate for which the
angle of incidence of the rotor blades is zero? == 0.507 m
110 HYDRAULIC AND COMPRESSIBLE FLO\V TURUOMACHI 'ES HYDRAULIC TURBINES III

Jet area
P == (0.23 _x 6~)2(9.81 x 246)5/2 x 1000
A == QIC 1 2n x 650
0.04 == 3266 kW
41.52
~ Total power output
== 0.963 x 10- 3 m 2 Number of wheels == -p- -- h'
ower per mac Ine
and jet dian1eter
4A)1/2 18 X 10 6
--- --
d== ( -
n 3.266 x 10 6
== 5.51 (say 6 machines)
= (4 x 0.96~ X IO-JY I
2

(b) Nozzle velocity coefficient

== 0.035 111
Diameter ratio Cl
C == --- -----
D 0.507 'v (2gH)1/2
d 0.035 Jet velocity
== 14.5 C 1 == 0.96(2 x 9.81 x 246) 1/2
== 66.7 m/s
Dimensionless power specific speed is given by Eg. (1.20): Bucket speed
N pl/2 U == 0.46 x 66.7
N =-------
sp pl/2(yH)5/4 == 30.7 m/s

== 720 x (28.72 x 10
3) 1/2 X( ~__ )5/4
Wheel diameter
60 x U
60 10 3 9.8 1 x 91. 5 D==--
nN
==0.0131 rev
60 x 30.7
== 0.0131 x 2n rad
n x 650
Power specific speed == 0.082 rad
== 0.901 m
A check with Fig. 1.10 shows this value of N sp to be within the range for a
Pelton wheel. (c) Overall efficiency

Exercise 3.2 (a) From Eg. (1.20) we get the power specific speed for each wheel Shaft power developed
Np l / 2 110 == Power available per wheel
N == -- ---- where N is in rplll
sp pl/2(glJ)5/4 18 x 10 6
Available head Power per wheel == - - -
0.87 x 6
H == Gross head - Head loss
== 3.448 MW
== 280 - 34
Also
== 246 III Power a vailable per wheel == 0.5 l1zCi
Available power per wheel == 0.5pQ(2gH)C;
112 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES HYDRAULIC TURBINES 113
U WI
Thus
3.448 X 10 6 = 10 3 X Q X 9.81 X 246 X 0.96 2 I 1 ~\

0
and flow rate
I.. e'l
J
Discharge coefficient

C __ ~ctual nozzle discharge

d - Theoretical nozzle discharge
e~

nd 2 (2gH)1 /2 Q
where d is the nozzle diameter u
4 Cd
Figure 3.19 Inlet velocity triangle
1.55 X 4) 1/2 [ I ]
d = ( 0.94 x n x (2 x 9fu i46)li~i5
Nozzle diameter d = 0.174 m Exercise 3.3 Figure 3.19 illustrates the system with the velocity triangles.
From Eg. (3.4)
(d) Total discharge for six machines = 1.55 X 6
3
Power output
=9.3m /s Hydraulic efficiency = - - - - - - - -
Energy available injet
Total discharge or flow per pipe = 9.3/3
W
=3.lm 3 /s
0.5mC~
The frictional head loss in the pipe is given by
At entry to nozzle
4flv 2
hr = - - H = 600 - 48
2gd
= 552m
where v is the flow velocity and d the pipe diameter. Whence
Nozzle velocity coefficient
d= 4 X0.006 X 2500v 2
2 x 9.81 x 34 1 C
C =-------
v Theoretical velocity
= 0.09v 2 m
But C1
2
nd
Q=vA = v -
4 Thus
and substituting for v C 1 = 0.98(2 x 9.81 x 552) 1/2
= 102 m/s
d = 0.09(3.1
nd
X
2
4)2 Now
and W/nl= UICxl-U2Cx2
d S = 1.4 = U { (U + W d - [U - W 2 cos (180
0
- ex) J}
d = 1.07 m = U[(C 1 - U)(1 - k cos a)J
114 HYDRAULIC AND COMPRESSIULE FLOW TURUOMACIiINES HYDRAULIC TURBINES 115

where W 2 = kW l . Substituting for the symbols Torque

W/nl=0.47C l (C l -0.47C l )(I-0.88cos 170) T= m(r 2 Cx2 -'1 C x1 )
W = 0.465111Ci
But since the flow is radial at outlet, Cx2 = 0 and therefore
Theoretical hydraulic efficiency = 0.465/0.5
T= - mr l C x1
= 0.93 = -pQr l C x1
Actual hydraulic efficiency = 0.9 x 0.93 = - 10 3 x 0.5 X 0.6C x1
= 0.837 = - 300C x1 Nm
Wheel bucket speed The inlet area A is
U=0.47 x 102111/S
Wheel rotational speed A = 2nr 1 b 1 where b 1 is the inlet runner height
0.47 x 102 x 60 = 2n x 0.6 x 0.03
N=------
0.45 x 2n = 0.113 m 2

= 1017rpm Now flow velocity Cr1 is given by

. . Actual power developed Crl = Q/A
Actual hydraulIc efficIency = - - - - - - - - -
Energy available injet 0.5
1250 x 10 3 0.113
0.837 = 2
0.5111C 1 = 4.42m/s
Substituting for eland solving for the mass flow rate From Fig. 3.12

1250 x 10 3 Cxl = Crl tan 72

111 = ([837 x 0.5 x 102 2 = 4.42 x 3.08
= 287 kg/s
= 13.6m/s

Hence for one nozzle Substituting for C xl gives

T= - 300 x 13.6
111 = 143.5 kg/s
Also from continuity = -4080Nm
nl=pC l A where A is the nozzle area This is the torque exerted on the control volume (i.e. on the fluid). The
= pC l nd
2 torque exerted by the fluid is + 4080 N m and is the torque exerted on the
where d is the nozzle diameter runner.
4
and hence Torque exerted by water on runner = 4080 N m
d2 = 143.5x4
Power exerted
n x 102 x 10 3

= 1.792 X 10- 3 m 2 W=Tw

d = 42.3 111m 4080 x 2n x 1260

60
Exercise 3.4 The angular momentum equation may be used. = 538 kW
116 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES HYDRAULIC TURBINES 117

Hydraulic efficiency from which P1 = - 36.2 to give the blade angle Pl as (180 - 36.2) = 143.8
with V 1 > Cxl '
Power exerted Runner speed
rJH=------
Power available
nD1N
538 X 10 3 V l =--
60
pgQH
60 x 0.97 x (2 x 9.81 x 5.5)1/2
538 x 10 3 Dl = -----------
n x 230
10 3 x 9.81 x 0.5 x 124
Runner inlet diameter = 0.836 m
= 0.885
= 88.5 per cent Overall efficiency
Exercise 3.5 Figure 3.12 shows the velocity diagrams at inlet and outlet. Power output
Hydraulic efficiency rJo= P. ower aval'1 a bl e

Power given to runner or

rJH = -------- 125 X 10 3
Power available pgQH=---
0.74
m(V1C xl - V 2C x2 )
whence now rate
pgQH
125 X 10 3
But since the flow is radial at outlet, C x2 is zero and m equals pQ. Therefore Q=--- -----
0.74 X 10 3 x 9.81 x 5.5
V1C xl
rJH=--
gH = 3.13 m 3 /s
But also
0.82 = 0.97(2gH)1/2C Xl Q = nD l b 1 Cr1
gH Therefore
whence b _ 3.13
C xl = 0.423(2gH)1/2 1 - n x 0.836 x 0.4(2 x 9.81 x 5.5)1/2
Now
Height of runner = 0.287 m
tan (Xl = Crl /C x1
0.4/0.423
=
Exercise 3.6 Use the notation of Fig. 3.12.
from which inlet guide vane angle (a) Runner tip speed
(Xl = 43.4 nND
V l =--
60
n x 430 x 1.4

C xl - Vl 60

0.4 = 31.5 m/s

0.423 - 0.97 Power given to runner
= - 0.731 W=m(V1C xl - V 2C x2 ) (from Eq. (1.24))
118 HYDRAULIC AND COMPRESSIULE FLO\V TURUOMAClliNES HYDRAULIC TURBINES 119

But Cx2 = 0 since there is zero whirl at outlet. Hence Hence loss of head in the runner
J
C 1= 12250 x 10 x 60
x 10J x 12 x n x 1.4 x 430
= 32.4 m/s But
Guide vane angle
Cl~P2)+(ZI -Z2)=62m

2
33.8 _72) (31.5 x 32.4)
Head loss in runner = 62 + ( 2 x 9.81 - 9.81
= tan-l(~)
32.4 = 13.69 m
= 16.3
Exercise 3.7 (a) From Eq. (3.21) the hydraulic efficiency is given by
Inlet velocity Power transferred to runner
C 1 = (C;1 + C;1)1/2 Y/H = Power available
= (9.5 2 + 32.4 2)1/2 Power transferred to runner = pgQH Y/H
= 33.8 m/s = WY/H

(b) Crl But from Euler's turbine equation (Eg. (1.24))

tan fJ 1 = - - -
C xl - VI
W UICxl-U2Cx2
E=-=------
9.5 mg 9
32.4 - 31.5 Therefore
= 10.55
whence
Runner blade entry angle {J 1 = 84.6 and Cx2 = O. Hence

0.9 x 9.81 x 62
C xl = - - - - -
(c) Total head across runner VI
= Energy (head) transferred to runner + Head lost in runner Now
nND
At inlet V I =--
60
H1
PI
= ---- + -C7 + Z I
py 2g n x 375 x 1.5
At outlet 60
= 29.45 mls
Therefore
Now for zero whirl at outlet 0.9 x 9.81 x 62
C -------
xl - 29.45

"'f} 9 = 18.58 m/s

 HYDRAULIC TURBINES 121
120 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES

The velocity triangle is therefore as shown in Fig. 3.20 with U 1 > C x1 ' Substituting values

Wxl = U 1 C x1 426.2 )
- 62 = ( 35 + 2x9.81 + 2 + hI 055, I
= 29.45 - 18.58
whence
= 10.86 mjs hIOS 5 , I = 62 - 58.72
tan (fil - 90) = W xl jC,1 Head lost in casing = 3.27 m
10.86
(ii) For the loss in the runner, apply the steady flow energy equation
9
between points 1 and 2:
= 1.21 mjs
V2 V2
~ + ~ + 2 1 = P2 + ~ + 2 2 + hlos s ,2 + Work (head) given to runner
fil - 90 = 50.4
Entry angle pg 2g pg 2g

Now

(b) (i) For the volute casing and guide vane loss, apply the steady flow mg g
energy equation between points 0 and 1: 29.45 x 18.58
2 2
V V 9.81
Po + ~ + 2 0 =~+ --l + 2 1 + h
pg 2g pg 2g 1055,1
= 55.8 m
Now VI = C 1 and Therefore 2

Ci = C;I + C;I h 2 = (35 + _426.2 + 2) - (- 2.2 + 7 + 1.7) - 55.8

10!iS, 2x9.81 2x9.81
= 18.58 2 + 9 2
58.72 - 2.0 - 55.8
=
= 426.2 m 2 j s2
Runner head loss = 0.92 m of water

(iii) Applying the steady now energy equation between points 2 and 3:
V2 V2
P2 + ~ + 2 2 = P3 + ~
2 + 2 3 + hloss,3
29.45 PO 2g pg fJ
But P3 is atmospheric pressure (zero gauge) and 2 3 is the datum level, whence
18.58 10.86
2.02 )
h
loss,3
= 2.0 -
(
0+
2 x 9.81
+0

9
= 2.0 - 0.204
Head loss in draft tube = 1.8 m

(c) Dimensionless power specific speed from Eq. (1.20) is

Np l / 2
N =----
~p P 1/2(gH)5/4
Figure 3.20 Velocity triangle at inlet
122 HYDRAULIC AND COMPRESSIBLE FLO\V TURllOMACHINES
HYDRAULIC TURBINES 123

p 1 / 2 = 0.14 X (10 3 )1/2 X (9.81 X 62)5 14 X 60 Exercise 3.8 Mean diameter

375
d =D+~
= 2140 m 2
P=4578kW (this is the power delivered to the runner) 4.75 +2
Thus 2
3
0.9 = 4578 x 10
pgQH = 3.375 m
Flow rate Overall efficiency
4578 X 10 3 Power developed
Q = 10 3 X 9.81 x 62 110 = Power available
=7.53m 3 js 23 X 10 6
Available power = - - -
Flow rate 0.85
FI ow area =-----
Flow velocity =27MW
Q Also
C2 Available power = pgQH
and at exit from the runner the flow area may be written in terms of the runner 27 x 10 6 = 10 3 x 9.81 x 23Q
exit diameter and runner height b2 : Whence flow rate
nd 2 b2 = QjC 2
where d2 is the draft tube entry diameter. Now the runner height at entry hi is Rotor speed at mean diameter
given by nNd m
U =--
m 60

n x 150 x 3.375
7.53 60
n X 1.5 X 9 = 26.5 mjs
= 0.178 m Power given to runner == Power available x 11H
Also
== 27 X 10 6 x 0.93
b1
b 2 == 2 + - - 1.7 =25.11 MW
2
= 2 + 0.089 - 1.7 But theoretical power given to runner is from Eq. (1.24)
= 0.389 m ~v = pQU m C x1 (C x2 = 0)
6 3
Substituting for b 2 we get 25.11 X 10 == 10 x 119.7 X 26.5C x1
6
7.53 25.11 X 10
d 2 = -- ----.-- ---_. C x1 =------
3
n x 0.389 x 7 10 x 119.9 x 26.5
Draft tube dialneter = 0.88 In = 7.9 mjs
 124 HYDRAULIC AND COMPRESSIRLE FLOW TURBOMACHINES HYDRAULIC TURBINES 125

Axial velocity Therefore

C = Qx4 16000 X 10 3
a n(D2 _ d2) Available power = - - - -
0.8
119.7 X 4 = 20000kW
n(4.75 2 - 22 ) Also
= 8.21 rn/s Available power = pgQH
From the inlet velocity triangle Therefore
20000 X 10 3
tan (180 - PI) = Ca Q = 10 3 x 9.81 x 20
Urn - C x1
= 101.9m 3 js
8.21
Power given to impeller = Power available x tlH
26.5 -7.9
= 20000 X 10 3 x 0.9
Inlet blade angle fJ 1 = 156.2
= 18 000 x 10 3 k W
Now from Eq. (1.24) the power transferred per unit mass flow is
At outlet
tan fJ2 = C a/Wx2 Wjm == U 1 C xl - U 2 C x2

But Wx2 equals Urn since Cx2 is zero. Hence and putting C x2 = 0 since at exit flow is purely axial, and writing U 1 in terms
of the radius at the blade tip, i.e. at 2.1 m,
R 8.21
tanp2 = - 18 000 x 10 3 = 17.41 x 2. 1 x 10 3 X 10 1.9 x CxI
26.5
C xl = 4.8 mjs
Ou tlet blade angle P2 = 17.2
Mean flow velocity

Exercise 3.9 Using Eq. (1.20) for power specific speed, C = Qx4 where D and d are tip and hub diameters
a n(D 2 _ (/2)
Np l / 2
N =---- where N is in rad/s 101.9x4
sp pl/2(gH)5/4 2
n(4.2 2
- 2 )

)1 12(9 81
3
N =3x ( 10 0
X 20)5/4 = 9.51 mls
16000 X 10 3
From Fig. 3.16
= 17.41 rad/s
W x1 = U - C xl at the blade tip
17.41 x 60
2n rpm
= (17.41 x 2.1) - 4.8
= 31.8 mls
= 166.3 rpm and
Overall efficiency tan (180 - fJ d = C a jWx1
Power developed 9.51
tlo = Power available 31.8
 HYDRAULIC TURBINES 127
126 HYDRAULIC AND COMPRESSIllLE FLOW TURBOMACHINES

At r = 2.1 m Exercise 3.10 (a) The flow rate is calculated at the mean diameter and the
Inlet angle f31 = 163.4 0 velocity triangles of Fig. 3.16 are used.
Mean diameter
d =D+d
Atr=2.1m m 2
Wx2 =U (since flow is axial at cxit) 2 + 0.8
= 17.41 x 2.1 2
= 36.6 m/s
= 1.4m
and
Radial velocity
tan f3 2 = CalWx2
u = nNd m
9.51 m 60
- 36.6
n x 250 x 1.4
Atr=2.1m
60
Outlet angle f3 2 = 14.5
= 18.33 mls
At the hub, r = 1 m From the inlet velocity triangle
Wx1 = 17.41-(4.8 x 2.1) U m = Ca[cota i + cot(180 - f3i)]
= 7.33 m/s Axial velocity
Whence 18.33
C =------
a a cot 42 + cot 32
(180 - f31) = tan -1 ( C )
W x1 18.33
1 9.51) 1.11+1.6
= tan - ( 7.33
= 6.76m/s
At r = 1 m
Flow rate
Inlet angle fJ 1 = 127.6 Q = C.n(D
2
- d
2
)
since C a is constant
4
At r ~ 1m
W x2 =U 6.76 x n(2 2 - 0.8 2 )
= 17.41 n1/s 4
Whence
IJ 2 = tan - 1( 9.51
- -)
17.41 (b) From the outlet velocity triangle
At r = 1 m ta n f3 2 = Cal U m
Outlet angle fJ2 = 28.6
6.76
, I' J. 18.33
The point to note in this problcIn is that we havc assuIl1ed the energy Outlet blade angle
transfer across the annulus to remain constant and that the axial flow
fJ2 = 20.24
velocity ren1ains constant.
 HYDRAULIC TURBINES 129
128 HYDRAULIC AND COMPRESSlnLE FLOW TURBOMACHINES

(c) The theoretical power is calculated by determining the energy transfer

0.5 ----------1
through an annulus of thickness dr and integrating between the hub and tip
radii for the total theoretical power.
From Eq. (1.24) tJ 0.4
b Safe zone
...
~
~

E
But C x2 is zero and C xl is constant and writing m and U as functions of the e 0.3
~
0..
radius we get for an annulus of thickness dr: c::
.~
~
21tNr)
dW= pCi2nrdr) ( 60 C x1
.~ 0.2
~
()

Integrating ca
:~
W=PC a 1t NC x1
15
2
I1

0.4
r
2d
r
U 0.1
o
This point is in
~ the cavitating zone

= 10 3 X 6.76 x n x 250 x 6.76 x cot 42 [r

2 3
J1
0.25
0.05 0.10 0.15 0.20
15 3 0.4
Dimensionless specific speed, Ns (rev)
= 8348(0.333 - 0.0213)10 3
Figure 3.21 Critical cavitation parameter
Theoretical power = 2602 kW

If the power is calculated at the mean diameter

W=mU m C x1 . . d' F' 3 21 and it is seen that cavitation will
-y
The characterIstIcs are platte In Ig. .
3
= 10 x 17.84 x 18.33 x 6.76 x cot 42
occur. .. . above the tailrace for no cavitation is when
= 2455 kW (b) The limIting heIght
(J = 0.1. Solving for 2 2
The difference is 6 per cen 1. c
3
4 X 10 2
50xO.l==10.3-9.81 X 103- 2
Exercise 3.11 (a) Dimensionless power specific speed is given by Eg. (1.20)
Np1/2 when limiting height
N =---- 2 2 == 10.3 - 0.408 - 5
sp pl/2(gH)5/4
== 4.89 m
300(2 X10 6)1/2 1
= 60 l()3 (9.81 X 50)5/4 . . f Fig 3 11 between sections 2 and 3:
(c) USIng the notatIon a .,
= 0.0969 rev 2 V + Z + Losses
2

From Eq. (3.25)

P2
--+ ---+ V2 Z _ P3
2-
+~ 3
P{j 2{j pg 2g

a= [ (Palm ~ Pvap ) - Z 2 JI H But

P3 = Patm
V3 == 0 (negligible)
=[(10.3- 104 109.81 )-6J;'50 3
X
x
3
2 3 == 0 (datum)
P2 == Pvap
= 0.0778
130 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES

Therefore the head lost in the draft tube is

CHAPTER
hd = (Pva Palm) + Z2 + V;
p -
FOUR
pg 2g
At the limit of cavitation 2 2 = 4.89 m and substituting into the above equation
CEN,TRIFUGAL COMPRESSORS AND FANS
4 x 10 3 10.4 2
hd = 3 -10.3+4.89+---
10 x 9.81 2 x 9.81
= 0.408 - 10.3 + 4.89 + 5.51
Head loss lId = 0.511 In

4.1 INTRODUCTION

During the Second World War, great progress was made in the development
of gas turbines using the centrifugal compressor. This followed from the need
for gas turbines to be supplied with large amounts of high-pressure air, and the
centrifugal compressor became a natural choice as it had previously been
researched for use in small high-speed internal combustion engines. Although
the centrifugal compressor has been superseded by the axial flow compressor
injet aircraft engines, it is useful where a short overall engine length is required
and where it is likely that deposits will be formed in the air passages, since,
because of the relatively short passage length, loss of performance due to build-
up of deposits will not be as great as in the axial compressor. The centrifugal
compressor is mainly found in turbochargers, where it is placed on the same
shaft as an inward now radial gas turbine, which is driven by engine exhaust
gases (Fig. 4.1). Pressure ratios of 4: 1 are typical in a single stage, and ratios of
7: 1 are possible if exotic Inaterials are used for impeller manufacture. The best
efficiencies are 3-4 per ccnt below those obtainable from an axial flow
compressor designed for the saine duty. However, at very low mass flow rates
the axial flow cOlnprcssor efficiency drops ofT rapidly. It is also difficult to hold
the tolerances required for small axial flow blading, and manufacture of the
axial compressor bladcs becomes more expensive. If the density ratio across
the compressor is less than about 1.05, the term 4fan' is used to describe the
machine. In that case the nuid is treated as being incompressible; otherwise
compressible flow eq uations must be used. The term 4blower' is often used in
place of 4fan'.
131
132 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES

Actual W 2
4.1.1 Elements of the Centrifugal Compressor
The elements of a centrifugal compressor are similar to those of a hydraulic
pump, with some important difTerences, and these are illustrated in Fig. 4.1. - - - No slip
The gas enters the compressor at the eye in an axial direction with - - - With slip
absolute velocity C l' and moves into the inducer section, which can be
separate from or be a part of the blades. The inducer section transfers the gas
onto the blades, and enables it to move smoothly into the radial direction.
Energy is imparted to the gas by the rotating blades, thereby increasing its
static pressure as it moves from radius r 1 to r 2 (Fig. 4.2), and the gas moves ofT
the blades with absolute velocity C 2 It should be noted that the blades are
radial, and since it is conventional to measure blade angles from the radial
direction in centrifugal compressors, the blade angle /32 is zero, while the
relative velocity vector W 2 is at angle f3~ because of slip. Ideally, the
component C x2 equals U 2' but it is reduced due to slip. The relative velocity Hub radius, r
vector WI at the inlet is obtained by subtracting U 1 from C 1. Pre-whirl may be
given to the gas at inlet, but this will be discussed in a later section. The Stanitz

Eye tip or shroud radius, R

Vaneless
/diffuser

Impeller

Collector
(volute)

Figure 4.1 Elements of a centrifugal compressor

Blade width, b
Figure 4.2 Velocity diagrams for a centrifugal compressor
134 HYDRAULIC AND COMPRESSIULE FLO\V TURUOr...1AClIlt\:ES CENTRIFUGAL COMPRESSORS AND FANS 135

slip factor (Eq. (2.9)) is best applied for radial vanes and with {12 = 0, Thus E becolnes
as = 1 - 0.63n/Z (4.3)
By Euler's pump equation (Eq. (1.25)) without slip, Upon leaving the impeller the gas enters a vaneless space where it moves
in a spiral path before entering the diffuser, in which the static pressure is
E = U 2Cx2/g = U~/y (4.1)
further increased. The clearance between the impeller blades and inner walls of
and with slip,
the casing must be kept as small as possible to reduce leakage and in some
(4.2)
cases the blades themselves are shrouded. Since we are dealing with a gas and
Although Eg. (4.1) has been n10dified by the slip factor to give Eq. (4.2), since the rise in temperature and pressure causes the density to change, it is
(JsU~/g is still the theoretical work done on the air, since slip will be present convenient to examine the performance of the machine in terms of the
even if the fluid is frictionless. In a real fluid, some of the power supplied by the thermodynamic properties of the gas, and this is done through the Mollier
impeller is used in overcoming losses that have a braking effect on the air chart of Fig. 4.3.
conveyed by the vanes, and these include windage, disc friction and casing
friction. The total power per unit weight of flow is therefore modified by a
4.1.2 Inlet Casing
power input factor ljJ, which typically takes values between 1.035 and 1.040.
The energy equation along a streamline may be written as
Total enthalpy ho = h + C 2/2 = Constant
Therefore, for the fluid that is being drawn from the atmosphere into the
PO'l
P0 3 P3 inducer section, the total enthalpy is
02 .-_0_3__----~--__r_
hoo = ho + C512
02s
~
l /1
/03s
C :1'l
2 Total enthalpy at section 1 is
!
03S5 I

I and since no shaft work has been done and assuming adiabatic steady
I
!
flow,

Thus
(4.4)

4.1.3 Impeller

hUll = hUI
From section 1 to 2 the fluid moves through the impeller where work is done
OO'P'----:w---+------r~~
C'l hU'l = h03 on it to increase its static pressure from PI to P2' Writing the work done per
o -:f unit mass on the fluid in terms of enthalpy we get
Wlnz = h02 - hOl
=U2Cx2-UICxl (4.5)

Entropy, S or after substituting for ho,

Figure 4.3 Mollier chart for a centrifugal compressor I = hI + C f12 - U 1 ex 1 = h 2 + C ~/2 - U 2 ex 2
 136 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES
CENTRIFUGAL COMPRESSORS AND FANS 137
where I is a constant throughout the impeller. In general
The slip factor should be as high as possible since it limits the energy
I=h+C 2 /2-VC x transfer to the fluid even under isentropic conditions and it is seen from the
= h + (C; + C;)/2 - VC x velocity diagrams that C x2 approaches U 2 as the slip factor is increased. The
= h + (W
2
+ C;)/2 - VC x
- W; slip factor may be increased by increasing the number of vanes but this
2 increases the solidity at the impeller eye, resulting in a decrease in the flow area
= h + [W - (V - C x )2 + C;]/2 - VC
x at inlet. For the same mass flow rate, the flow velocity Cn at inlet must therefore
2 2
= h + W /2 - V /2 - C;/2 + VC x - VC x + C;/2 be increased and this increases the loss due to friction. A compromise is usually
= h + W /2 - V 2 /2
2
reached, slip factors of about 0.9 being typical for a compressor with 19-21
2 vanes.
= hOrel - V /2 (4.6)
While it may seem that a high value of power input factor t/J is desirable, it
where ho.reJ is the total enthalpy based on the relative velocity of the fluid. Thus is found that the rate of decrease of isentropic efficiency with increase in t/J
negates any apparent advantage, so the ideal should be to have a power input
h2 - hI = (V~ - Vi)/2 + (Wi - W~)/2 (4.7)
factor of unity.
since I I = 12 , In Eq. (4.7) the main contribution to the static enthalpy rise is The pressure ratio increases with the impeller tip speed but material
from the term (V~ - Vi)/2. strength considerations preclude this being increased indefinitely. Centrifugal
h' .In preliminary design calculations it is usual to assume Cxl = 0, alth oug h stresses are proportional to the square of the tip speed and, for a light alloy
t I~ IS not always the case, whence from Eq. (4.3) the work done on the fluid per impeller, tip speeds are limited to about 460 m/s. This gives a pressure ratio of
unIt mass becomes about 4: 1. Pressure ratios of 7: 1 are possible if materials such as titanium are
used.
h02 - hal = t/J(JsV~ (J/kg)
Equation (4.11) can be written in terms offluid properties and flow angles
After writing Cp To in place of ho' we get that the work input is given by
T 02 - Tal = t/J(JsV~/Cp (4.8)
where.C is th~ mean specific heat over this temperature range. Also, since no
p

work IS done In the diffuser, h02 = h03 , and Eq. (4.8) becomes

T 03 - T01 = t/J(JsV~/Cp (4.9) Got = 340 m/s

y = 1.4 0.90
. Wit~ refer.ence to Figs 1.4 and 4.3, a compressor overall total-to-total (j!. = 0.92

IsentropIc efficiency 1]c may be defined as Vt = 1.04 0.85

.8 I

1] = Total isentropic enthalpy rise between inlet and outlet

0.75
C Actual enthalpy rise between same total pressure limits
=(h 03ss - hod/(h o3 - hod (4.10)
where the sUbscri~t '~s' represents the end state on the total pressure line p
when the process IS Isentropic. Thus 03

'7c = (T03ss - T OI )/(T03 - Tod

= To 1 (T03ss/To I - 1)/(T03 - To 1)
But 2

P03/PO 1 = (T03ss/To dY/(Y - 1)

= [1 + 1]c(T03 - T od/T01 ]Y/(Y-l) 380 400 420 440 460 500

Blade tip speed (m/s)
= (1 + 1]ct/J(JsV~/CpTOl)Y/(Y-l) (4.11 ) Figure 4.4 Overall pressure ratio versus impeller tip speed (from Eq. (4.12))
138 HYDRAULIC AND COMPRESSIULE FLOW TURUOMACHINES CENTRIFUGAL COMPRESSORS AND FANS 139

as follows: since C1

then
and
~UI
(4.12) 1

The change of pressure ratio with blade tip speed is shown in Fig. 4.4 for
various isentropic efficiencies.

4.1.4 Diffuser (a) (b)

The stagnation temperature of the gas at outlet fron1 the diffuser should have Figure 4.5 Velocity triangles for (a) large and (b) small inlet areas
as small a kinetic energy term as possible, as this eases the problem of
combustion chamber design. Typical compressor outlet velocities are of the
order of90 m/s. The diffusion process is carried out in a diffuser as described in 2. If the eye tip diameter is small, the blade speed is small but the axial velocity
Secs 2.5.2 and 2.5.3, some diffusion also taking place in the vaneless space C I is large, and the velocity diagram of Fig. 4.5b may be drawn.
between the impeller tip and diffuser vanes. The flow theory described in those
sections is applicable here. The maxin1um included angle of the vaned diffuser For both of these extreme cases, the relative velocity vector WI is high but it
passage is about 11, any increase in this angle leading to a loss of efficiency must reach a minimum value when moving from one extreme to another. If
through boundary-layer separation on the passage walls. It should also be this relative velocity can be determined, Mach-number effects can be avoided
remembered that any change from the design mass flow rate and pressure ratio by proper design.
will change the sInooth flow direction into the difTuser passage and will Flow into the eye takes place through the annulus formed by the shroud
therefore also result in a loss of efficiency. This may be rectified by utilizing radius R and the hub radius r. For uniform axial flow into the eye
variable-angle diffuser vanes.
m=P1A1C I (4.14)
For adiabatic deceleration of the fluid fron1 absolute velocity C 2 to C 3
with a corresponding increase of static pressure froln P2 to P3' but from the inlet velocity triangle (Fig. 4.5),
"02 == hOJ and
or
The flow area is
(4.13)
Al == nR 2 (I - r 2 jR 2 )
== nR 2 k (4.15)
4.2 INLET "ELOCrrV LIMITATIONS where k == (1 - r /R 2 2
). Substitution for A 1 into Eq. (4.14) gives

Mach-number considerations at the eye of a centrifugal cOlnpressor Illake the 111 == P l rrR 2 kC l
relative velocity WI a very sensitive value as far as cOlllpressor perfornlance is == P l nVikC I /w 2
concerned. Should the Mach nlllnber at entry to the illlpeller be greater than
where V 1 is the inlet tangential velocity of the impeller at the shroud radius
unity, then shock waves will fonn, \vith all their attendant losses. Assulne that
and w is the angular velocity. Rearranging gives
we have a uniform absolute velocity C I with zero \vhirl (C.d == 0) at entry to a
centrifugal compressor. Two cases Inay be cXi.llnined for the salnc Inass flow Itlw
2
/P l nk == vic l
rate, both cases being ex treInes. == Wi(sin 2 fJ 1)(COS Ii tl (4.16)

1. If the eye tip diameter is large, then from continuity considerations the axial From isentropic flow relationships the ratio of stagnation to static pressure at
velocity C I is low and the blade speed is high, resulting in the velocity inlet may be written as
diagram of Fig. 4.5a. Pal/PI == [I + (y - I)Mi/2JY/(Y-l) (4.17)
 140 HYDRAULIC AND COMPRESSIRLE FLOW TURROMACHINES
CENTRIFUGAL COMPRESSORS AND FANS 141
and
atmosphere), the right-hand side of Eq. (4.22) may be plotted and the
TOI/TI = 1 + (y - 1)Mi/2 maximum value determined along with the corresponding blade angle at
Then (4.18)
which the maximum occurs. This maximum value may then be equated to the
PI/TI = (Pol/ToI)[l + (y - I)Mi/2] -Y!(Y-I)[1 + (y - I)Mi/2] left-hand side of Eq. (4.22) and the maximum mass flow rate determined.
Figure 4.6 shows Eq. (4.22) plotted for air at Po 1 = 101.3 kPa and To 1 = 288 K
= (Pol/ToI ) [1 + (y - I)Mi/2] -l/(y-l) where it is seen that the blade angle is almost constant at 60 for maximum
Now mass flow. Therefore, by specifying the relative Mach number M l,rel' the
PI = PI/RT I maximum value of mass flow may be calculated. Relative Mach numbers are
= (POI/ RTod [1 + (y - I)Mi/2] -l/(y-l) (4.19) usually restricted to about 0.8 to ensure there is no shock-wave formation.

Therefore substituting for PI from Eq. (4.19) into Eq. (4.16),

2
mw RTotlrrkpol = W~(sin2 /31)(COS /31)/[1 + (y - 1)Mi!2]l/(Y-1) (4.20) 4.3 PRE-WHIRL AND INLET GUIDE VANES

Writing the relative Mach number based on the relative velocity WI' then Should it not be possible to restrict the Mach number at inlet to an acceptable
= MLela~(sin2 /31)(COS /31)/[1 + (y -
2 value as described in Sec. 4.2, it may be achieved by placing guide vanes at the
mw RTotlrrkpol 1)Mi!2]l/(Y-l) (4.21)
inlet. Figure 4.7 clearly shows that the inlet guide vanes impart a whirl
Fro~ E~. (4.18), a~l/al = [1 + (y - 1)Mi/2]1!2 since a = (yRT)I/2 and after component exl to the fluid, thus reducing W l to an acceptable value.
substItutIng for a l In Eq. (4.21) and putting M = M P I I, reI COS I' However, the work capacity is reduced since ext is no longer zero. It is not
(mw 2)/[nkYPOI (yRToI )I!2] necessary to impart pre-whirl down to the hub, as, in this region, the fluid is
2 nowhere near sonic conditions due to the lower blade speed. The pre-whirl is
= MLel(sin /31)(COS /3 d/[ I + (y - I )MLeI(cos 2 f1 d/2] l/(y - 1) + 3/2 (4.22)
therefore gradually reduced to zero by twisting the inlet guide vanes.
It should be rem:mbered that Eq. (4.22) is applied at the shroud radius R
and consequ~ntly fJ 1 IS also the blade angle at radius R since it is at this radius
that the maXImum value of relative velocity onto the blade will occur. 4.4 MACH NUMBER IN THE DIFFUSER
Therefore, for a gas of known inlet stagnation conditions (e.g. the
The absolute Mach nunlber of the fluid leaving the impeller may well be in
excess of 1. However, it has been found that as long as the radial flow velocity
Cr2 is subsonic, then no loss in efficiency is caused by the formation of shock
0.25,-------- _

Crl = 0
0.20 y = 1.4
..-
N CI
N
~ 0.15
CI
a. VI
& VI
l.1J Cn Cn
MI,rt'1
l-.

~ 0.10 0.90
:t
~
0.85
0.80 fl.
\
\
/1 1 """ Wi
0.75
\
\
~I
I ""
0.05 \ I ""
\
\ ,1 " "
I
I
- - - - - Without guide vanes
\
\
I
I
'"
o 20 30 40 50 60 70 80 - - - With guide vanes
\
\ I
~
Relative flow angle, PI' at the shroud (deg) (a) (b)
Figure 4.6 Optimization of the mass now function
Figure 4.7 The effect of inlet guide vanes on the inlet relative velocity: (a) at shroud; (b) at hub
142 HYDRAULIC AND COMPRESSIBLE FLOW TURUOMACHINES
waves. In addition, provided constant angular m0l11entUI11 with vortex motion
is maintained in the vaneless space between ilnpeller tip and difTuser, then
supersonic diffusion can take place in the vaneless space. This reduces the
Mach nUInber at inlet to the difTuser vanes to about 0.8. High Mach numbers
at inlet to the diffuser vanes will also cause high pressures at the stagnation
points on the diffuser vane tips, which leads to a variation of static pressure
around the circumference of the difTuser. This pressure variation is transmitted
radially across the vaneless space and can cause cyclic loading of the impeller,
which may lead to early fatigue failure.
4.5 CENTRIFUGAL COMPI{ESSOR CHAI{ACTEI~ISTIC
The characteristics of compressible flow machines arc usually described in
terms of the groups of variables derived in Eg. (1.14). The characteristics are
given as a series of curves of POJ/fJo 1 plotted against the Inass flow parameter
In Tb /1 /PO 1 for fixed speed intervals of N/Tb l/. This nlight be done by
2
controlling the flow through a valve situated downstrealn of the compressor
outlet flange.
An idealized fixed-speed characteristic is shown in Fig. 4.8. In discussing
the shape of this curve, much of what will be said will also apply to the axial
flow compressor characteristics of Sec. 5.10.
Consider a centrifugal cOlnprcssor delivering through a flow control valve
situated after the difTuser. If the valve is fully closed, a certain pressure ratio
from inlet across to the difTuser ou tlet will be developed, and this is shown at
point 1 in Fig. 4.8. This pressure ratio is due solely to the vanes Inoving the air
about in the inlpeller. As the flow control valve is opened, and air flow begins,
the diffuser contributes to the pressure ratio, and at point 2 the n1aximum
6 vanes. In this way a pair of diffuser passages will be supplied with air from an
Figure 4.8 The theoretical centrifugal
compressor characlerislic
CENTRIFUGAL COMPRESSORS AND FANS 143
pressure ratio is reached, although the cOlnpressor efficiency at this pressure
ratio will bejust below the Inaximum efficiency. A further increase in mass flow
brings conditions to point 3, where the pressure has dropped slightly from the
maximum value but the efficiency is now a maximum. This is likely to be the
design mass flow rate pressure ratio. A further increase in mass flow sees the
slope of the curve increasing until it is almost vertical at point 4 where the
pressure rise is zero. Theoretically, point 4 would be reached when all the input
power is absorbed in overcoming internal friction. However, the curve just
described is not obtainable practically and some of the reasons for this are now
discussed.
4.5.1 Surging
Suppose we have a compressor operating at point 3 on the negative slope of
the characteristic (Fig. 4.8). A reduction in mass flow due, say, to a momentary
blockage will cause an increase in the delivery pressure PO), which will tend to
return the mass flow to point 3, and this negative slope represents a region of
stable operation. It is self-correcting. If the flow rate drops to a point
corresponding to 6 on the positive slope of the characteristics, the delivery
pressure PO) will continue to decrease, causing a further drop in mass flow and
a further drop in P03' and so on until point 1 is reached, where the mass flow is
zero. The mass flow may even become negative through the compressor. When
the back-pressure PO) has reduced itself sufficiently due to the reduced flow
rate, the positive flow beconles established once again and the compressor
picks up until the restricted Inass flow is again reached, when pressure
reduction once again takes place. The pressure therefore surges back and forth
in an unstable fashion, which, if severe enough, could lead to failure of parts of
the compressor. Because of the reduction of Inass flow, the axial velocity Ca
into the eye is reduced and therefore the relative flow angle onto the blade P1 is
increased. I'he air now onto the blade will no longer be tangential. Surging
tends to originate in dilTuser passages where frictional efTects of the fluid next
to the vane surfaces retard the flow. Indeed, flow nlay well be in reverse from
one blade passage to the next. The likelihood of surging can be reduced by
making the nUJnber of difTuser vanes an odd-number Inultiple of the impeller
odd number of vanes and pressure fluctuations are nlore likely to be evened
out around the circulnference than if exact nlultiples of difTuser vanes are
employed.
4.5.2 Rotating Stall
Rotating stall is a separate phenolnenon, which may lead to surging but can
exist on its own in a stable operating condition. Figure 4.9 illustrates the air
flow directions in a nUlnber of blade passages.
 CENTRIFUGAL COMPRESSORS AND FANS 145
144 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIlINES

Increased angle of attack Decreased angle Now

hal = hI + Cf/2 (4.24)

~/I
of attack

/ /8 = h 1 + (Wi - Ui)/2 (4.25)

Ifchoking occurs when the relative velocity equals the acoustic velocity a 1, Eq.
(4.25) becomes

- and
(4.26)

Direction of stall propagation For isentropic now,

Figure 4.9 Rotating stall propagation

PI/Pal = (TI/Todl/{Y- 1)
- and therefore from 111 = paA
~f the air angle of incidence onto blade A is excessive, due perhaps to a HJ/A = POIaol(TI/Tol)(Y+ 1)/2(y-1)
partIal blockage or uneven flow in the diffuser, the blade may stall, and because
since a oc TIll. Substituting from Eq. (4.26) and rearranging gives
of the partial mass flow decrease in the blade passage, the deflected air causes
~he. angle of inci?ence to the left of blade A to increase while the angle of m/A = POI(lOl[2(l + Ui/ 2C pTol )/(Y + 1)J(y+I)/2(y-l)
InCIdence to the rIght of blade A will tend to decrease. Thus blade B will be the = {YPOIPOI[2(1 + ui/ 2C pT01 )/(Y+ 1)](y+I)/(y-1)}1/2 (4.27)
next to stall while blade A will be unstalled and the process is repeated about
the periphery of the disc. Prolonged cyclic loading and unloading of the rotor Equation (4.27) is simply a modified version of Eq. (4.23) and shows that the
blades can lead to fatigue failure or even immediate catastrophic failure. The choking mass now rate increases with impeller speed.
stall propagates in the opposite direction to blade motion at a frequency In the diffuser passages Eq. (4.23) also applies with the subscripts changed
related to shaft speed. In compressor tests, rotating stall may be audibly to the impeller outlet conditions
recognized as a high-frequency 'screech'. trt/A = {YP02P02[2/(y + l)](Y+ 1)/(y-1)} 1/2 (4.28)

The areas in Eqs (4.23), (4.27) and (4.28) refer to the now areas at the respective
4.5.3 Choking
locations.
If the mass flow is increased to the right of point 3 on the negative slope of the
characteristic (Fig. 4.8), a point 5 is reached where no further increase in mass 4.5.4 Characteristic Curve
flow is possible no matter how wide open the flow control valve is. Thisindicates
Figure 4.10 shows the overall pressure ratio and efficiency plotted against
that at some point within the compressor sonic conditions have been reached
causing the limiting maximum mass flow rate to be set as in the case of
mTb// Ipo I at fixed speed intervals of N /Tb// It is usual to transfer constant-
efficiency points onto the corresponding constant-speed curves of the pressure
compressible flow through a converging-diverging nozzle. Indeed should this
ratio characteristics and then join those points together to form constant-
condition .arise, shock wa ves may well be formed wi thin certain' passages.
ChokIng may take place at the inlet, within the impeller, or in the diffuser efficiency contours.
It is evident that at all speeds the range of mass flow over which the
section. It will occur in the inlet if stationary guide vanes are fitted the
centrifugal compressor will operate before surging or choking occurs is quite
~aximu~ mass flow being governed by the following standard equatio~ for wide, although the margin decreases as the speed increases. The onset of surge
IsentropIc flow at the throat of a converging nozzle:
occurs at increasingly high mass flows as the speed increases, while the locus of
m/A = {YPooPoo[2/(y + l)J{Y+ 1)/(Y-l)}1/2 (4.23) the limit of stability is called the surge line. Maximum efficiency is seen to occur
well within the surge point, making this type of compressor relatively
where stagnation conditions at inlet are known and A is the flow area.
insensitive to mass flow changes as far as instability is concerned. The limit of
In stationary passages, the velocity that is choked is the absolute velocity.
maximum flow is usually set by choking in the impeller, while the surge limit of
In the rotating impeller it is the relative velocity W that is the choked velocity.
 CENTRIFUGAL COMPRESSORS AND FANS 147
146 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIiINES

conditions of 100 kPa and 15C. The hub diameter is 0.13 m and the eye tip diameter 0.3 m. If the
mass now into the eye of the compressor is 8 kg/s and the speed is 16200 rpm, calculate the blade
80

D OAO 0.53 0.67 0.80.0.87 0.93 1.0

\ \~pc:~~n inlet angle at the root and tip of the eye and the Mach number at the tip of the eye. Assume zero
whirl at inlet and no losses in the intake duct.
4.2 A motor rated at 580 kW is available to drive a centrifugal compressor of 480 mm outer
diameter at a speed of20 000 rpm. At the impeller outlet the blade angle is 26.5 measured from the
radial direction and the now velocity (radial component) at exit from the impeller is 122 m/s. If a
N/T~i2
mechanical efficiency of95 per cent is assumed, what air now is to be expected? Assume there is no
slip. What are the eye tip and hub diameters if a radius ratio of 0.3 is chosen for the impeller eye and
if the velocity at inlet is 95 mls with zero whirl? What will be the overall total-to-total isentropic
Design efficiency if an overall total pressure ratio of 5.5 is required? Assume that the now to the inlet is
point incompressible and ambient air conditions are 101.3 kPa and 288 K.
4.3 A centrifugal compressor compresses air at ambient temperature and pressure of 288 K and
101.3 kPa respectively. The impeller runs at a tip speed of 365 mis, the radial velocity at exit from
the impeller is 30 mls and the slip factor is 0.9. Calculate the Mach number of the now at the
impeller tip. If the impeller total-to-total efficiency is 90 per cent and the now area from the
impeller is 0.093 m 2 , calculate the mass now rate of air. Assume zero whirl at inlet, and radial
Surge line blades.
4.4 A centrifugal compressor impeller has 17 radial vanes of tip diametcr 165 mm. 1l rotates at
46000 rpm and the air mass now rate is 0.6 kg/s with no whirl at inlet. Calculate the theoretical
power transferred to the air. At inlet to the impeller, the mean diameter of the eye is 63.5 mm while
the annulus height at the eye is 25 mm. The static pressure and temperature at the impeller inlet are
93 kPa and 293 K respectively. Determine
(a) the blade angle at the mean diameter at impeller inlet,
(b) the stagnation temperature at impeller exit and
(c) the stagnation pressure at impeller exit if the total-to-total efTiciency of the impeller is 90
2.0
per cent.
/
4.5 The relative now Mach numher althe inlet of a centrifugal compressor is to be no greater than
I / I I
I I I / I 0.97. The ratio of hub to lip diameter at the eye is 0.38 and the air cnters the eye without whirl.
1/ I / / Determine
I I / / 0.40
I ,~ / / (a) the optimum speed of rolalion for the maximum mass now rate condition if the mass now
I L':f/ I
1.0 l-..L...:=..L-L.-l----J-L-.-L--L---1--l~O~---:~ is 4 kgls and
o 0.2 0.4 0.6 o.g . Figur~ 4.10 Typical centrifugal (b) the eye tip diameler.
mT~~2 IPol (relative to dcsign value) compressor characteristic At inlet POl = IO!.) kPa and '[~)I = 2HH K.
4.6 A compressor operating at a pressure ratio of 3.8 and a speed of 12000 rpm delivers 8 kg/s of
air. The slip factor is assumcd to be 0.92, the power factor 1.04, and the overall isentropic efficiency
minimum mass flow is set by stalling of the flow onto the diffuser vanes. For. a 0.82. Calculate the impellcr oUllet diameter. ASSUllle zero whirl.
vaneless diffuser the surge line would be further to the left of that shown 1n The Mach number of the air leaving the impeller vanes is to be unity so as to ensure that no
Fig. 4.10, thus increasing the tnass flow range at the design speed. shocks occur. If the losses in the impeller and difTuser are the sanle, what must be the axial depth of
the im peller?
At inlet POl = 101.3 kPa and TOl = 288 K.
EXERCISES 4.7 In exercise 4.6, if the vaneless space is 40 mm wide, the diffuser throat radius 0.4 m and there
are 10 diffuser channels, determine the difTuser vane inlet angle and the throat width of the diffuser
passages.
In all the exercises below, the following properties for air may be assumed:
4.8 A centrifugal compressor has inlet guide vanes fitted at the eye such that free vortex now is
Ratio of specific heats, y = 1.4 achieved at entry to the blades. At the tip radius of the eye the inlet relative Mach number is not to
Specific heat of air at constant pressure, Cl' = 1005 Jjkg K exceed 0.75 and an impeller total-to-total efTiciency of 0.9 is required. The air leaves the tip of the
Gas constant, R = 287 J/kg K inlet guide vanes with a velocity of 90 mis, the impeller tip diameter is OA5 m and the outlet
Assume a power input factor of 1 unless otherwise stated. diameter is 0.76 m. The radial component of velocity at exit from the impeller is 50 mls and the
impeller rotates at 12000 rpm. If a slip factor of 0.9 is assumed, find the guide vane inlet angle at the
tip and the static pressure at impeller outlet.
4.1 A single-sided centrifugal compressor draws air from the atmosphere having ambient
148 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIlINES CENTRIFUGAL COMPRESSORS AND FANS 149

Assume Tot = 28M K and POI = 101.3 kPa. Let the hub and tip radii at the eye be denoted by r h and r t respectively.
4.9 A centrifugal compressor has the following design data: Then the flow area of the impeller inlet annulus is
Mass flow 15 kg/s A 1 = n(rt - r~)
Speed 16000 rpm
Inlet total temperature 288 K = n(0.15 2 - 0.065 2 )
Inlet total pressure 101 kPa
Impeller isentropic efficiency 0.9
= 0.0574 m 2
Width of vaneless space 42mm Assume first a value for density Pl based on ambient conditions:
Axial depth of vaneless space 37 mm
Slip factor 0.9
POl
Power input factor 1.04 Pl = R 1 01T
Impeller outer diameter 0.55m

(a) Assuming no pre-whirl at the inlet, what are the stagnation conditions at the impeller
outlet? 287 x 288
(b) Show that the radial velocity is approximately 100 mls at the impeller outlet and
calculate the Mach number and air angle at the impeller outlet. = 1.21 kg/m 3
(c) Calculate the angle of the diffuser vane leading edges and the Mach number at this radius Then from the continuity equation (Eq. (1.21))
if the diffusion in the vaneless space is isentropic.
m
4.10 The stagnation pressure and temperature at the inlet of a centrifugal compressor are C =--
101 kPa and 290 K respectively. The impeller has 19 radial vanes and no inlet guide vanes. The a PIAl
following data apply.
8
Mass flow rate 2kg/s 1.21 x 0.0574
Impeller tip speed 525 mls
Mechanical efficiency 96 per cent = 115.18 mls
Absolute air velocity at diffuser exit 95 mls
Compressor total-to-total efficiency 81 per cent Since there is no inlet whirl component, Co = C l ' the absolute inlet velocity,
Absolute velocity at impeJler entry 160 mls and the temperature equivalent of this velocity is
Diffuser efficiency 83 per cent
Axial depth of impeller 6mm ci 115.18 2
Power input factor 1.04 2C p 2 x 1005
(a) Determine the shaft power. =6.6K
(b) Calculate the total and static pressures at diffuser outlet. Therefore
(c) If the reaction ratio (h 2 - h d/(h 3 - hd is 0.5, lind the radial velocity, absolute Mach
T l = T01 C 21
number and total and static pressures at the impeller exit. Diffuser efficiency = (h 31 - h2 )/(h 3 - h2 )
(d) Determine the impeller total-to-total efficiency. 2C p
(e) If a vaneless diffuser is litted, what is the ratio of inlet to outlet radius of the diffuser?
(f) Determine the impeller speed. = 288 - 6.6
= 281.4 K
Assuming isentropic flow at the inlet
SOLUTIONS
~ = (li)Y/(Y- 1)

Exercise 4.1 In Fig. 4.2 the blade angle Pl' measured with respect to the axial POl T 01
velocity Co, is required to be found at the hub and tip. First the axial velocity
must be determined from the continuity equation, but since the inlet density PI _ 5(281.4)1.4/(1.4-1)
PI- 10 --
is unknown a trial-and-error procedure must be followed, assuming first a 288
density based on the inlet stagnation conditions. The method is set out below. = 92.2 kPa
150 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES CENTRIFUGAL COMPRESSORS AND FANS 151

Then new At the hub

PI
PI = RT1 2 x n x 16200 x 0.065
VI = - - - - 6 0 - - - -
92.2 X 10 3
= 110.3 mls
287 x 281.4
3
Blade angle
= 1.14 kg/m
and new -1 ( 110.3 )
/31h = tan 122.25
8
C =
a 1.14 X 0.0574 = 42.05

= 122.25 Ill/s The Mach number required is that based on the relative velocity at the eye
Repeat tip. Acoustic velocity
Ci 122.25 2 a 1 = (yRTd 1/ 2
2C p 2 x 1005 = (1.4 x 287 x 280.57)1/2
= 7.43 K = 335.8 mls
T 1 = 288 - 7.43 From the inlet velocity triangle
= 280.57 K Wi = (C; + Vi)1/2
280.57)3.5 = (122.25 2 + 254.5 2 )1 / 2
PI = 10
5
( 288
= 282.3 mls
= 91.25 kPa
R~lativc Mach number = W 11a l
91.25 x 10 3
PI = 287 x 280.56 282.3
335.8
= 1. 13k gl In 3
lV!l.rel = 0.841
Further iterations are unnecessary and the value PI = 1.13 kg/m 3 may be
taken as the inlet density and Ca = C 1 = 122.25 mls as the inlet velocity. Exercise 4.2 Using Fig. 4.2, since there is no slip the blade angle at impeller
At the tip exit equals the relative velocity vector angle (/3~ = f32)'
2n N r t 2 x n x 16 200 x 0.15 nND 2
VI = 6() = 60 V 2 =(;Q

= 254.5 mls n x 20000 x 0.48

Blade angle 60
= 503 mls
- 1( 254.5 )
= tan 122.25 = 503 - 122 tan 26.5
= 64.34 = 442 mls
152 HYDRAULIC AND COMPRESSIHLE FLOW TURBOMACIIINES CENTRIFUGAL COMPRESSORS AND FANS 153

From Eq. (1.25) the power per unit mass flow is rt = 51.7/0.3
W/m=U2Cx2-UtCxl Tip diameter = 172 mm
But Cxl is zero, therefore From Eq. (4.10), compressor total-to-total efficiency is
W/m = 503 x 442 h03ss - "0 t
'1c=----
and 11 03 - hOt

m = W/222.3 x 10 3 m 3 /s CpT01[(T03ss/Tod - IJ
W/nl
Theoretical power = Power available x Mechanical efficiency
But
W = 580 X 10 3 x 0.95
T03 !jS = (P03)(Y-l)!Y
= 551 kW
TOl POl
Hence
Therefore
m = 551/222.3
1005 X288(5.5286 - 1)
Expected mass flow = 2.47 kg/s tlc = 222.3 X 10 3
Using the continuity equation, the mass flow rate through the annulus of
Efficiency = 0.818
the eye is
m=P1A1C l Exercise 4.3 The absolute Mach number of the air leaving the impeller
= Pl n(r; - r~)Cl vanes is
M 2 = C 2 /a 2 = acoustic velocity at impeller tip)
Plnr~G~ -1)C
(a 2
= 1
_ C2
- (yRT 2)1/2
(T2 = static temperature at impeller tip)
Assuming incompressible flow at inlet
Pl = POl The problem therefore requires C 2 and T 2 to be evaluated.
POl Slip factor
RTol
101.3 X 10 3
287 x 288
since C x2 = U 2 for zero slip
= 1.226 kg/m 3
= 0.9 x 365
Thus
= 328.5 m/s
2 2.47
r - - - - - - - - - :2- - - - -
h - n x 1.226 x (3.333 - 1) x 95 From the velocity triangles of Fig. 2.5,

= 6.677 x 10- 4 m c~ = C;2 + (C~2)2

rh = 25.8 mm = 30 + 328.5 2
2

Hub diameter = 51.7 mm = 1.088 x 10 5 m 2 /s 2

154 HYDRAULIC AND COMPRESSIBLE FLOW TURllOMAClllNES CENTRIFUGAL COMPRESSORS AND FANS 155

With zero inlet whirl Eq. (4.2) gives !2 = (353.2)3.5

P02 407.3
W/nl = a s C x2 U 2
= 0.607
= asU~ Therefore

Thus P2= ( -P2 )(P02)

- POl
P02 POl
asU~ = 0.607 x 3.03 x 101.3 x 10 5
T02=Tol+C
p
= 186.3 kPa
0.9 x 365 2
= 288 + 1005 186.3 x 10 3
P2 = 287 x 353.2
= 407.3 K
= 1.84 kg/m 2
Mass flow
Also
c~ /11 = 1.84 x 0.093 x 30
T 2 = T 02 - 2C
p == 5._~ 3 ~~j_s
109 gOO
= 407.3 - 2 x 1005 Exercise 4.4 For radial vanes the slip factor may be found from the Stanitz
equation (Eq. (2.9)) with {12 = 90 0

= 353.2 K Slip factor

Hence a. == 1 _ 0.63~
S Z
1.088 X lOs )1 /2
(
M 2 = 1.4 x 287 x 353.2 = 1 _ 0.63 x n
17
= 0.876
= 0.884
At impeller outlet tn = P2A2Cr2 where P2 = P2/RT2 and note that these are
static properties. The static pressure P2 is found by first solving for the impeller Impeller tip speed
total pressure ratio P02/PO l ' and then relating P2 to P02 by means of the U 2 == nlVD 2
isentropic pressure-temperature relationships. 60
From Eqs (4.10) and (4.11) rearranged for now in the impeller:
n x 46000 x 0.165
P02 [ (T02
- = 1+11 - - 1
)lY/{Y-l) 60
POl C Tal
== 397.4 rnls
407 3
= [ 1 + 0.9 ( 28~ - 1
) lJS W/nl == (JsU 2 C x2 - U 1 Cx !
But e Xl = zero and C x2 equals U 2' the whirl velocity at impeller tip without
= 3.03 slip. So
Now
W == 0.6 x 0.884 x 397.4 2
from isentropic relationships at a point
Theoretical power = 83.76 kW
156 HYDRAULIC AND COMPRESSIRLE FLOW TURIlOMACIIINES CENTRIFUGAL COMPRESSORS AND FANS 157

(a) r t - r h = 25 mm and (r h + r t )/2 = 31.75 mm. Solving for r h gives But from Eq. (4.5),
(2 x 31.75) - 25
rh == 2
83760 5
h02 == - - - + 3.004 x 10
== 19.25 mm 0.6
rt == 44.25 mm = 4.4 x 10~ N m/kg

At the eye Therefore

Pi 4.4 X 10 5
PI == RT T 02 = - - -
1
C"
93 x 10 3 4.4 X 10 5
287 x 293 1005
= 1.106kg/m 3
= 437.8 K
Mass flow
(c) From Fig. 4.3 the impeller efficiency is
m=PIAIC a since Cxl = 0
11 02s - hOI
=P1AIC I tJj==
h02 - hOI
C _ 0.6
I - 1.107 x n(0.044 25 2 - 0.01925 2 ) C p (T02s - Tod == tJjWlm
0.9 x 83.76 x 10 3
= 108.7m/s
0.6
From the velocity triangles of Fig. 4.2
= 125.64 x 10 3
tan PI = U I/C I
At the mean radius T02s =
125.64 X 10 3
+
T
I
Ci
+-
Cp 2C p
U _ nND m 2
l-~ 125.64 X 10 3 29 108.7
----+ 3+---
1005 2 x 1005
= 152.9 m/s
= 423.9 K
f3 I =tan - 1 ( -
152.9)
-
108.7 From isentropic relationships
P T )'1/('1-1)
Inlet blade angle = 54.6 ~= ~
C 2
POI
( T Ol
(b) hOI=hl+~
2
Now
C2
=CpTI+~
2 T Ol = hol/C"
300400
=(1005 x 293)+ I08Y
2 1005
= 3.004 x 10 5 N m/kg == 298.9 K
 CENTRIFUGAL COMPRESSORS AND FANS 159
158 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES

0.3 M r1 = 0.97

29H.9)3.5 ~
r./)

POI = 93 ( 29-3 ~ 0
u 0.2
r./)
0
u
~
=99.7kPa ~
~c
'Vi
=-I
Therefore M ... -b
V)

~ 0 0.1
_ (423.9 )3.5 +
P02 - 99.7 - - .=.
29H.9
= 338.67 k Pa o"-:::::::__ -----1 ----L. --4- --'-- ---J

o 20 40 60 80
Blade inlet angle at shroud. ;/ 1 (deg)
Exercise 4.5 (a) Using Eq. (4.22) the appropriate known data are substituted,
noting that all conditions apply at the eye tip or shroud. Figure 4.11 Mass now parameler versus blade inlet angle at the shroud

2 0.2869
w = ------ . ---
= I -0.38 2 3.08 X 10- 8

== 0.8556 = 9313987

The LHS of Eq. (4.22) beCOlTICS w == 3051 rad/s

2 Optimum speed = 29 143 rpm
LHS= 4w _
n x 0.8556 x 1.4 x 101.3 x 10 3 X (1.4 x 2H7 x 288)1/2 (b) The eye tip diameter is found from the continuity equation for flow into
= 3.08 X IO- 8 w 2 the eye. Once again all cquations apply at the tip radius.

and the RHS of Eq. (4.22) bccorncs 111 == P I A 1C 1 since C 1 = Ca (no whirl)
3 2 Expressing the area in tcnns of the radius ratio and tip radius
RHS = 0.97 sin PI COSPI
[1 + 0.5(1.4 - 1)0.95 2 cos 2 P1J 4 111 == np1kR2Cl
0.9127 sin 2 PI cos Pi It is necessary to determine C I and Pl'
(1 + 0.1805 cos P1)'~ 2 To find C I' from the inlet velocity triangle of Fig. 4.2 at the eye tip radius

Choosing values of PI the RHS of Eq. (4.22) is cvaluatcd: WI == C I/COS fJ I

But
pl(deg) 10 20 30 40 50 55 59 60 61 65 70 80
RHS 0.014 0.055 0.119 0.193 0.258 0.279 0.286 0.287 0.287 0.279 0.253 0.15 and
The results are plotted in Fig. 4.11 fron1 which it is deduced that the RHS Therefore
reaches a maximum at PI == 60. M I == 1\1 r I cos fJ 1
Equating the LHS and RHS with PI == 60,
== 0.97 cos 60
8w 2
0.9127 sin 260 cos 60
3.08 X 10- == - - - - - -2 - - == 0.485
(1 + 0.1805 cos 60)4
160 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES CENTRIFUGAL COMPRESSORS AND FANS 161

Now Therefore

Putting 4
1t x 1.11 x 0.8556 x 161

== 0.008 33
then
R == 0.0913 m
TOl (y - I)Cf
-- = 1 + _._- ------ Eye tip diameter = t H3 mm
Tl 2yRT l
and putting
Exercise 4.6 Froln Eq. (4.1 t) the stagnation ten1pcrature difTcrence is

then T [(') ()J )O.2H6

'['n.\ -'f01 -- --~ - t J
'Ie Po 1
T (y - I)Mi
-Ol= 1 + - - - -
Tl 2 = 288 (3.Ro.2H(l _ t)
2 0.82
= 1 + 0.4 X 0.485
2 == 163.3 mls

= 1.047 Now
Whence
T _ 288 (from Eq. (4.9))
1 - 1.047
Therefore
=275K
2 163.3 x 1005
U 2 =-----
Then 1.04 x 0.92
C 1 = M 1 (yRT 1 )1/2
=171526
= 0.485(1.4 x 287 x 275)1/2
U 2 = 4 t 4.15 mls
= 161 mls Also
From isentropic relationships at a point
~ =(~)l/(Y-I)
POI T 01 414.15 x 60
D2 =
POl
Pl=--
(T )l/(Y-l)
-
l 1t x 12 000

RTo1 TOl ~12.r~_ller ollter diameter == 0.659 m

3
= 101.3 X 10 (_1_)2.5
The overall loss is proportional to (1 - 'Ie) == (1 - 0.82). Half of the overall
287 x 288 1.04 loss is therefore 0.5( t - 0.82) == 0.09 and therefore the efTective efficiency of the
= 1.11 kg/m 3 impeller in compressing from Po 1 to P02 is (t - 0.09) == 0.91.
162 HYDRAULIC AND COMPRESSIBLE FLOW TURUOMACHINES
CENTRIFUGAL COMPRESSORS AND FANS 163

From Eq. (4.11) after rearranging the subscripts To find the flow velocity normal to the periphery of the in1peller
P 1/(T - T ))Y/()'-l)
~= 1+~_ _O_1
(
POI T o,
=0.92x414.15
=(1 + 0.91 x 163.3)3.5
2g8 since T 02 = T 03 =381m/s

= 4.29
T 02 = 163.3 + 2XX = (l)2 X (1.4 x 287 x 376. 1)

=451.3K = 151 117n1 2 /s 2

From the velocity triangles of Fig. 4.2

and
C;2 = C~ - C~22
Therefore substituting for C~ == 151 117 - 145 16 I

T =451.3_(l)2(1.4x2g7T'2) 2
== 5956 m Is
2 2 x 1005
Cr2 = 77.17 mls
T _ 451.3
2 - 1 + 0.2 From the continuity equation the area

= 376.1 K
~ = (12))'!()'-I)
P02 T 02 8
2.127 x 77.17
=(~)3.5
451.3 = 0.0487 m 2
= 0.5282 . A2
Depth of Impeller == -
nD 2
P2 ( P2 )(P02)
Po 1 = P02 p;;; 0.0487
= 0.528 x 4.29 n x 0.659

= 2.266 == 23.5 mm
and
P2 = 2.266 x 101.3 Exercise 4.7 Figure 2.18 shows the absolute velocity vector C at angle (1.' to
= 229.58 kPa the radial direction. When the gas flows into the diffuser passage, it must do so
at this angle and hence a' is the inlet angle of the diffuser vanes:
P2=-
P2
RT2
229.58 X 10 3 In the vaneless space between the impeller outlet and diffuser vanes the flow is
2g7 x 376.1 that of a free vortex which at any radius requires that Cxr = canst.
At the diffuser vane leading edge the radius is (r 2 + 40) mm = (329.5 + 40)
= 2.127 kg/n1 3
= 369.5 mm and
164 HYDRAULIC AND COMPRESSInLE FLOW TURnOMACHINES CENTRIFUGAL COMPRESSORS AND FANS 165

cx = C 2r 2 Throat width Diffuser channel width

r Impeller tip radius 329.5 mm !
(151 117)1/2 X 0.3295 Diffuser tip radius 369.5 mm
Diffuser throat width R.5 mm
0.3695
= 346.7 mls

To find the radial velocity Cr at the diffuser vane entry, start by assuming
the value at the impeller exit, i.e. 77.17 m/s. Then
C2 C; + C;
2C p 2C p
77.17 2 + 346.7 2
2 x J 005
= 125.5 K
Now if we assume that no losses occur across the vaneless space, the other
half of the total losses takes place in the diffuser itself. Then P02 at the impeller
tip equals the stagnation pressure at the diffuser vane inlet Po. Therefore Figure 4.12 Flow from impeller to diffuser

Po = 4.29 Po I (from exercise 4.6)

2 p
C p=-
T= T o - - RT
2C p
139xl0 3
= 451.3 - 125.5
287 x 325.8
= 325.8
= 1.487 kg/m 3
~ =(~)Y/(Y-I)
P02 T 02 With reference to Fig. 4.12, the area of now in the radial direction at
radius 0.369 m is
= (325.8)3.5
451.3 A r = 2n x 0.369 x Depth of vane
= 0.32 = 2n x 0.369 x 0.0235
Also = 0.0545 m 2

P
Pal =
(pP02 )(P02)POI
Therefore
n1
= 0.32 x 4.29 C,= pA
= 1.373 4
p = 1.373 x J01.3 1.4R7 x 0.0545
= J39 kPa = 49.4 mls
166 HYDRAULIC AND COMPRESSIBLE FLO\V TUR130MACHINES
CENTRIFUGAL COMPRESSORS AND FANS 167

Repeat the calculation using this new value of Cr' Neglecting vane thickness
C 2 /(2C p ) == 61 K Flow area == 2n x Radius of diffuser throat x Depth
T== 390.3 K == 2n x 0.4 x 0.0235

P/P02 == 0.601 A r == 0.0591 m 2

p == 261.4 kPa 4
Cr == 2.47 x 0.0591
p == 2.33 kg/Ill J
== 27.4 ITI/S
C, == 31.6 In/s
Recalculate.
Repeat the calculation.
C 2 /(2C p) == 51.38 K
C 2 /(2C p ) == 60.29 K
T== 399.9K
T== 391 K
It may be seen that there is no change in the new values so the radial velocity at
P/P02 == 0.605 the diffuser throat == 27.4 m/s.
p==263kPa The flow direction of the absolute velocity is given by ri at the throat:

p == 2.34 kg/Il1 3 tan a' == Cx/C,

C, == 31.37 rn/s 320.2

27.4
No further iterations are necessary. Thus at the inlet to the vanes C, ==
31.37 m/s. Then (J.' == 85.1

tan rJ.' == 346.7 With reference to Fig. 4.12

31.37 m == pAC == pArCr (r means W.r.t. radial direction)
cl == 84.83 ArC r
A==-
Moving to the radius at the difTuser throat, at the throat radius, 0.4 m, C

C ==(151 117)1/ 2 x 0.3295 But

x 0.4 C == (51.38 X 2c p f/2
== 320.2 In/s == 321.4m/s
Start the iteration at the previous value of C, at the diffuser inlet. and
c 2
/(2C 1') == 51.5 K A, == 0.0235 (the vane depth)
T== 399.8 K Therefore
P/P02 == 0.654 0.0235 x 27.4
A==----
p == 284 kPa 321.4

p == 2.47 kg/In J == 0.002 m 2

168 HYDRAULIC AND COMPRESSIRLE FLOW TURROMACHINES CENTRIFUGAL COMPRESSORS AND FANS 169

As we have 10 diffuser vanes, the width of each throat is 282.7 2 + 90 2 - 253.35 2

C.d == 2 x 282.7
A
Throat width == 0 D h ==42.15m/s
1 x cpt
0.002 But
lOx 0.0235 C.d == C 1 sin (Xl
== 0.0085 Therefore
== 8.5 mm (Xl ==Sln
. -1 (COd)
~

==Sln
. _ 1(42.15)
--
Exercise 4.8 The velocity triangles of Fig. 4.2 may be used.
90
Blade angle at tip == 27.9
c~
T I == Tal - 2C The impeller total-to-total efficiency is given by
P

90 2 rr02s - Tal
== 288 - --------- 'Ii ==
2 x 1005 T 02 - Tal

== 284 K
Therefore
(11 == (yRTd l / 2 Now
Wln1 == C p(T o2 - T Ol )
== (1.4 x 287 x 284)1/2
==asU2Cx2-UICxl
== 337.8 mls
Therefore
At the tip
(since U 2 == Cx2
reND I
Ul==~
Now
12000 x n x 0.45 nND 2
U2==~
60
== 282.7 mls n x 12000 x 0.76
WI == 0.75 x 337.8
60

== 253.35 mls == 477.5 mls

Hence
Now from Fig. 4.7 at the inlet tip
T 02 == 1 + 0.9 X 477.5 2 - 282.7 x 42.15
Wi==(U I -C xl )2+C; (at the tip) T OI 288 x 1005
== U~-2UICXI +C.~l +C~-C;l == 1 + 0.668
== U~-2UICXI +ci == 1.668
 CENTRIFUGAL ('OMPRFSSORS AND FANS 171
170 HYDRAULIC AND COMPRl:SSIBLE i"LO\V TURBOMACIIINLS

From the efficiency equation n x 16000 x 0.55

60
P02/P01 = [(1.668 - 1)0.9 + IJ-'5
= 460.7 mls
= 5.2
From Eq. (4.9)
Now
fJ2 (P02)( P2) T - T - "'(JsU~ and
1005
p;;; = Po 1 P02
03 01 -

= (fJ02)(~)}'i(},-1) Therefore
Po 1 T 02 1.04 x 0.9 x 460.7 2
T 02 - T 01 = 1005
Also
= 197K

T 02 = ~ + 1
Therefore To 1 288
T C~
-2= 1 - - - = 1.68
T 02 2C p T 02
/2 2
Now for the impeller
= 1 _ C x2 + C r2
2C p T 02 P02 = (1 + l1i(To2 - T 01 ))3.5
Po 1 288
But
_ ( 0.9 X 197)3.5
- 1+ 288
= 0.9 x 477.5
= 5.36
= 429.75 m/s
Then
T 02 = 1.668 x 288 p02=541.4kPa
= 480.3 K
and
Substituting T'02 = 197 + 288
2
~ = 1_
2
429.75 + 50 =485 K
T02 2 x 1005 x 480.3
(b) From
= 0.806
C~2==(JsU2
P2/PO 1 = 5.2 X 0.806 3 . 5 '. ,'. ==0.9 x 460

= 2.44 ==414m/s

Impeller outlet static pressure = 265.8 k Pa Let Cr2 == 100 m/s.

Outlet area normal to impeller periphery

Exercise 4.9 (a) Impeller tip speed A 2 == nD 2 x Impeller depth

== n x 0.55 x 0.037
U _ nND 2
2- == 0.0639 m 2
60
172 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES CENTRIFUGAL COMPRESSORS AND FANS 173

From the outlet velocity triangle = (1.4 x 2R7 x 394.75)1/2

= ~9R.) m/s
C 22 C;2 + C~~
and
2C p 2C p C 2 = (90.25 x 2 x 1005) 1/2
100 2 + 414 2 = 425.9 nl/s
2 x 1005
425.9
M 2 =--
= 90.25 K 398.3
Thus
= 1.07
C~
T 2 = T 02 --
2C p
At the outlet velocity triangle
= 485 - 90.25
cos 0: 2 = Cr2 /C 2
= 394.75 K
100
From isentropic relationships at a point 425.9
_ ( T2 )Y/(Y- 1)
P2 - P02 T02
(c) Assuming free vortex flow in the vaneless space and denoting condi-
= 541.4(394.75)3.5
485 tions at the difTuser vane tip without a subscript,

= 263.4kPa C = C x2 r 2
x
r
Using the equation of state
414 x 0.275
0.317
=359.lnl/s
263.4
Choose as a first try Cr = 100 m/s. Then
287 x 394.75
C2 100 2 + 359. 12
= 2.33 kg/m J
2C p 2 x 1005
The equation of continuity gives
= 69.13 K
Following the same procedure as in part (b)
T=485-69.13 (T = T 02 in vaneless space)
15 = 415.9 K
0.0639 x 2.33
T )3.5
Impeller outlet radial velocity = 100.7 m/s P= P02 ( T
02

415.9
= 541.4 ( - -
)3.5
Impeller outlet Mach number = C2102 485
Now
= 316.3 kPa
174 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHI~ES CENTRIFUGAL COMPRESSORS AND FANS 175

316.3 Power transferred to air

Exercise 4.10 (a) Mechanical efficiency = --I' d h f
P2=287x415.9 Power Supp Ie to sat

= 2.65 W
Shaft power = -
A = 2nr x Depth of vanes '7m

= 2 x n x 0.317 x 0.037 Now for a radial vaned impeller the Stanitz slip factor equation is used.
= 0.0737 m 2
0.63n
(J
s
= 1 -Z
--
C = 15
, 2.65 x 0.0737
0.63n
= 76.8 111/S 19

Next try C, = 76.8 m/s. = 0.8958

Now
C2 76.8 2 + 359.1 2
2CI' 2 x 1005
W/nl = h03 - hOl
= t/J (J s U 2 Cx 2 since C x I = 0
= 67.09 m 2 /s 2
= t/J (J s U ~
si nee U 2 = C x 2
T=417.9K
W = 1.04 x 0.8958 x 525 2 x 2
p=321.7kPa
=513.56kW
P2 = 2.68 kg/In 3
513.56
C, = 75.9 In/s Shaft power = ---
0.96
Therefore there is little change in C, and thus this is the radial velocity = 535 kW
at the diffuser vane tip.
At the vane
(b) From Eq. (4.11)
tan ct.' = C x/Cr PO.3 III
11 elf' (J s
2U 2 )Y/(Y-I)
=
(
1+ --
359.1 POI C p TOI
75.9 = (I + 0.81 ~_L04~ ~8958 X525 2)3.5
1005 x 290
Vane angle
= 6.59

Diffuser outlet stagnation pressure

Mach number at vane
P03 = 6.59 x 101
lvl = Cia
= 665 kPa
= (67.09 x 2 x 1005 )1/2
1.4 x 287 x 417.9

Diffuser vane Mach nUlnber = 0.896

176 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES CENTRIFUGAL COMPRESSORS AND FANS 177

But 3 2 2
= 513.56 x 10 (160 - 95 )
2
C
T03 = T 3 + _ 3 1005 x 2 + 2 x 1005
2C p
= 255.5 + 8.25
T 95 2
-3= 1 - - - - - - = 263.7 K
T03 2 X 1005 X T 03 Substituting
Also T2 - Tl = 0.5 x 263.7
W = 131.87 K
T 03 =--+ Tal
Cpm Now
513.56 x 10
3
290 T 2 = To 1 - -2- + (T2
cf - T 1)
2 x 1005 + Cp
160 2
= 545.5 K = 290 - 2 x 1005 + 131.87
Thus
T 95 2 = 290 - 12.74 + 131.87
- 3= 1 - - - - - - -
T03 2 x 1005 x 545.5
=409.1 K
= 0.9917 Hence
Substituting
C 2 = [2 x 1005(255.5 + 290 - 409.1)J 1/2
P3 = 665(0.9917)3.5 = 523.6 mls
Acoustic velocity
= 646kPa
a 2 = (yRT2 )l/2
Diffuser outlet static pressure = 646 k Pa =(1.4 x 287 x 409.1)1 /2
= 405.4
(c) At the impeller outlet Mach number at inlpellcr outlet
1\1 2 = C 21a 2
523.6
or
405.4
since T 02 = T 03
= 1.29
Rearranging
Radial velocity at impeller outlet
C~ = 2C p [(To3 - Tal) + (T01 - T 2 )]
From the equation for reaction C;2 = C~ - C/2
= 523.6 2 - (as U 2)2
T2 - T1 = 0.5(T3 - T 1)
and = 523.6 2 - (0.8958 X 525)2
= 523.6 2
- 470.3 2
Cr2 = 230.2 mls

Knowing the diffuser efficiency, we can work back from the diffuser ou tlet
178 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES CENTRIFUGAL COMPRESSORS AND FANS 179

to the impeller outlet. Diffuser efficiency (e) Assume that Cx C r and if the gas moves in a free vortex then
11 3s - 11 2 C x3 f 3 == C x2 f 2
liD ==
11 3 - 11 2 Now since C x3 C r3 , then C x3 ~ C 3 and
and this relates the isentropic enthalpy increase to the actual enthalpy
f2 Cx3
increase. Rearranging
f3 C x2

95
0.8958 x 525
whence
P3 _ 1 + T 3 - T 2 )Y/(Y-l)
- ( t'lD
P2 T2
P2
== (1 + 0.83 X 131.87)3.5 (f) P2== RT
409.1 2
282.1 X 10 3
== 2.29
287 x 409.1
646
P2 == 2.29 == 2.402 kg/m 3

== 282.1 kPa Mass flow rate at impeller outlet

111 == P2A2Cr2

Using isentropic relationships at a point == 2np2 f 2C r2 b 2

But
P02 == (T 02 )Y/(Y-l)
where N is in rpm
P2 T2
545.5)3.5 nNnl
P02 == 282.1 ( 409.1

== 772.3 kPa and

525 x 2.402 x 230.2 x 0.006 x 60
N == - - - - - - - - - - - -

(d) /7i=To1 [ ( : : : )
(Y-l)/Y
-1
J!(I T 03 -To1
) == 52253 rpm
2

= 290[

228.9
C:~;3 r 286

- 1 JI (545.5 - 290 )

255.5

== 0.895

CHAPTER
FIVE
AXIAL FLOW COMPRESSORS AND FANS
5.1 INTRODUCTION
Because of a lack of knowledge of the aerodynamic behaviour of axial flow
compressor blades, the initial simple concept of using a reversed turbine as an
axial flow compressor resulted in compressor efficiencies of less than 40 per
cent. Some early gas turbines did use axial flow compressors but these gave
turbine efficiencies of the order of 55 per cent, and it was only with the
development of aeroplanes, and research into the aerodynamic behaviour of
wing sections, that blade design for axial flow compressors became es-
tablished. Some early investigators suggested that efficiencies of 90 per cent
would be reached, and this is now indeed the case, this figure surpassing the
maximum centrifugal compressor efficiency by about 4 per cent. However, a
penalty has to be paid for such high efficiencies and, with the axial flow
compressor, the performance is very sensitive to its mass flow rate at the design
point. Any deviation from the design condition causes the efficiency to drop off
drastically. Thus the axial flow compressor is ideal for constant-load
applications such as in aircraft gas turbine engines. They are also to be found
in fossil fuel power stations where gas turbines are used for topping up the
station output when normal peak loads arc exceeded.
The simple expedient of using a turbine to operate as a compressor ran
into difficulties because of the nature of the air flow in the two cases. In the
turbine, the blades form a converging passage, the area at inlet being greater
than at outlet, with the nuid being accelerated in the passage. In compressors,
the fluid is diffused with a pressure gain taking place through a blade passage
of increasing cross-sectional area. While a fluid can be accelerated over a wide
180
AXIAL FLOW COMPRESSORS AND FANS 181
u
r u
W'y 1
~
w~ ~.
1
;-r --"Wz
I JV 2
I -i
(a) (b)
Figure 5.1 Compressor and turbine blade passages: (a) turbine; (b) compressor
range with high efficiency, the process of diffusion cannot be carried ou t so
rapidly due to the onset of separation on the suction side of the blades, and
consequent stalling. This is similar to the included angle of a diffuser being too
great and separation taking place along the diffuser walls. Typical blade
sections are shown in Fig. 5.1, where it will be noted that the angular turning of
the relative velocity vector is much greater in the turbine than in the
compressor. The maxin1um rate of efficient diffusion within the blade rows is
equivalent to a cone angle of about 7 or 8.
In studying the flow of the nuid through an axial compressor, it is usual to
consider the changes taking place through a compressor stage. A stage consists
ofa row of moving blades attached to the periphery ofa rotor hub followed by
a row of fixed blades attachcd to the walls of the outer casing. The compressor
is made up of a number of such stages to give an overall pressure ratio from
inlet to outlet. Figure 5.2 illustrates a few compressor stages.
It will be seen that at the inlet to the compressor, an extra row of fixed
vanes, called inlet guide vanes, are fitted. These do not form part of the stage
but are solely to guide the air at the correct angle onto the first row of moving
blades. The height of the blades is also seen to decrease as the nuid moves
through the con1pressor. This is so that a constant axial velocity through the
compressor is maintained as the density increases from the low- to high-
pressure regions. i\ constant axial velocity is convenient from the point of view
of design but is by no means a requirement. The analysis for now through the
stage will first be described in terms of two-dimensional flow. The flow through
the stage is assllillcd to takc place at a nlean blade height where the blade
peripheral velocities at inlet and outlet are the same, there being no now in the
radial direction. Whirl cOlnponents of velocity will exist in the direction of
blade motion.
 AXIAL FLOW COMPRESSORS AND FANS 183
182 HYDRAULIC AND COMPRESSIULE FLOW TURBOMACHINES

I PI (Xl
G
R S R V
WI Cl
Ca
W ZI Czl
Ii
ROlorZ
row
U
.1

(xz
--1
i Inlet

'S :sorrow
; area C2
W2 W zz Cz2

_Disc or I. ~I
V
w hub

(X3

Figure 5.2 An axial compressor stage

Figure 5.3 Velocity triangles for an axial now compressor stage

5.2 COMPRESSOR STAGE

The rotor and stator rows of a stage are shown in Fig. 5.3. Note that all angles equation (Eg. (1.25))
are referred to the axial velocity vector Ca' (5.1 )
Air exits from the previous row of stator blades at angle C( 1 with absolute
velocity C l' The rotor row has tangential velocity U, and cOlnbining the two or
velocity vectors gives the relative inlet velocity vector WI at angle {J l' At rotor E=(U2Cx2-U1Cxd/g
row outlet the velocity triangles are silnilar to those drawn for the axial flow But from the velocity triangles, and noting that Ca is constant through the
pump, and the absolute velocity vector C 2 moves into the stator row where the stage, and U 1 = U 2 = U,
flow direction is changed to a 3 with absolute velocity C 3 . The diagrams have
been drawn showing a large gap between the rotor and stator blades; this is for C x2 = U - C el tan f32
clarity. In practice, the clearance between rotor and stator rows is small. and
If the following stage is the saine as the preceding one, the stage is said to be
normal. For a normal stage C 1 = C J and C( 1 = a 3' W 2 is less than WI' showing Thus
that diffusion of the relative velocity has taken place with sOlne static pressure
rise across the rotor blades. The air is turned towards the axial direction by the and
blade camber and the efTective now area is increased froln inlet to outlet, thus E = U Ca(tan f31 - tan #2)/g (W/(N/s)) (5.2)
causing diffusion to take place. Sitnilar diffusion of the absolute velocity takes The energy transfer may also be written in terms of the absolute velocity
place in the stator, where the absolute velocity vector is again turned towards flow angles
the axial direction and a further sta tic pressure rise occurs.
E=UC,,(tana2-tanad/g (W/(N/s)) (5.3)
The energy given to the air per unit tnass flo\v rale is given by Euler's
184 HYDRAULIC AND COMPRESSIBLE FLOW TURHOMACHINES AXIAL FLOW COMPRESSORS AND FANS 185

h
PII2 P(),J or
02
03
Rearranging,
03s P3 ("I

03ss (11 2 - hi) - (C x2 - Cx.)[(U - C x2 ) + (U - Cx1 )]/2 =0

ts)
(1l 2 - hd + (Wx2 - Wxl )(Wx2 + Wx1 )/2 =0
Js (h 2 - hI) + (W.~2 - l.~1)j2 = 0
Jss r I i

')
! "-'1

the enthalpy change in a single-slage axial now cOlnpressor is so lo\\' conlpared

to the centrifugal cOIn pressor. The relative velocities may be of the same order
of magnitude, but the axial now compressor receives no contribution from the
change in tangential velocity.
The isentropic efficiency is \vritten as
Ideal isentropic work input
'Ie = A ctua I wor k .Input
Figure 5.4 Mol1ier chart for an axial now compressor stage

which reduces to
Equation (5.2) or (5.3) may be used depending upon the information
available.
The flow through the stage is shown thernlodynamically on the Mollier Putting
chart in Fig. 5.4 and is similar to that for a centrifugal compressor. The
"OJ/POI = (ToJss/ToIP'/()'-l)
enthalpy change is continuous taking account of irreversibilities in the rotor
and stator. the pressure ratio becolncs
Assuming adiabatic flow through the stage, 11 03 = h02 ' and therefore "OJ/POI = [1 + Yfe(T03 - Tod/To1]y/(y-l) (5.7)
Eq. (5.1) may be written as
The energy input to the nuid will be absorbed usefully in raising the
W/m = h02 - hOI (W/(kg/s)) (5.4) pressure and velocity or the air and sonlC will be wasted in overcoming various
Writing frictional losscs. Ilo\vever. the \vhole or the \vork input \vill appear as a
/, . ('.' ~ Ii: i (. I (
186 HYDRAULIC AND COMPRESSIBLE FLO\V TURllOMACHlj'.;ES AXIAL FLOW COMPRESSORS AND FANS 187

'In practice C a is not constant along the length of the blade and, to account for Figure 5.5 illustrates that it is only at the inlet of the machine that the
this, a work done factor ), is introd uced, defined as velocity profile is fairly constant over the blade length. The solid boundaries of
the rotor and stator exert n10re and more influence on the velocity profile as
Actual work absorbing capacity
Work done factor == .. the air moves through the compressor. The variation in work done factor in
Ideal work absorbIng capacIty Fig. 5.6 shows that ). decreases as the number of compressor stages increases.
Hence
(5.9) 5.3 REACTION RATIO

The reaction ratio is a Il1easure of the static enthalpy rise that occurs in the
tip 1.0 rotor expressed as a percentage of the total static enthalpy rise across the
stage. It is defined as
0.8 Static enthalpy rise in rotor
..c: R == Static enthalpy rise in stage
bD
c:
~
v 0.6 == (il 2 -lzd/(1z 3 - hd (5.10)
"'0
C':l
:0 Stage number Since ho 1 rei == /1 0 2rel' then
4-.
0
v
bO 0.4
~

C 4
v Also if C 1 == C 3' then
~
v
Q...
0.2 (/1 3 - hI) == (h 03 - h01 ) == U(C x2 - Cxd
and substituting for (11 2 - Iz d and (h 3 - h d in Eq. (5.10) the reaction ratio
root 00 becomes
0.2 0.4 0.6 0.8 1.0 1.2
R == (Wi - W~)/[2U(Cx2 - CxdJ
Figure 5.5 Variation of axial velocity along a blade == [(C; + W~l) - (C; + W~2)J/[2U(Cx2 - C x1 )J
== (Wx1 + W x2 )(Wx1 - W x2 )/[2U(C x2 - CxdJ
1.0 But C x2 == U - W x2 and C x1 == U - W x1 ' Therefore
(C x2 - C x1 ) == (Wx1 - Wx2 )
~ 0.95
Hence
o
u
~
R == (Wx1 + Wx2 )/2U
~ 0.9
o
"'0
== Ca(tan /31 + tan /32)/2U
~

~
o0.85 == (Ca/U)(tan /3 cxJ
== <P tan /3 00 (5.11 )
Here, tan /300 == (tan /31 + tan f32)/2, while the ratio of axial velocity to blade
4 8 12 16 20 speed is called the flow coefficient. It may be shown similarly that the reaction
Stage number ratio can be expressed as
Figure 5.6 Variation of work done factor wilh number of slages R == [1 + (tan /32 - tan adJ/2 (5.12)
188 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW COMPRESSORS AND FANS 189

III = WI/nU 2 = (17 03 - ho d/U 2 (5.13)

= }.(C x2 - CxdlU
= ).( CalU)(tan (.(2 - tan a 1)
If = )\, (~(tan a2 - tan ad (5.14)

U
5.5 LIFT AND DRAG COEFFICIENTS
(a) R = 0.5
Consider the rotor blade shown in Fig. 5.8 with relative velocity vectors WI
and W2 at angles PI and P2' This system is similar to flow over an aerofoil so
that lift and drag forces will be set up on the blade while the forces on the air
will act in the opposite direction as shown in Fig. 5.9.
The drag force is defined as acting in the line of the mean velocity vector
Woo at angle fioo to the axial direction as defined by Eq. (5.11), and the lift force
acts perpendicular to this.
The resultant force experienced by the air is therefore given by the vector
~ in Fig. 5.9, so that the force acting in the direction of blade rotation (the x

u
direction) is written as
U
(b) R > 0.5 (c) R < 0.5 Fx = Leos f300 + D sin f300
Figure 5.7 Effect of reaction ratio on the velocity triangles = Leos f300[1 + (CD/CL)tan f300J
But the lift coefficient is defined as
C L = LIO.5pW~A
For a reaction ratio of 50 per cent, (h 2 - hI) = (11 3 - 11 2 ), that is the static
enthalpy and temperature increase in the rotor and stator are equal. Also, from where the blade area is the product of the chord c and the span I, and putting
Eq. (5.12), /32 = aI' and when the outlet and inlet velocity triangles are super-
imposed, the resulting diagram is symmetrical. This is shown in Fig. 5.7a. For
R > 0.5, Fig. 5.7b shows the diagram skewed to the right since f32 > aI' and the
static enthalpy rise in the rotor is greater than in the stator. The static pressure
rise is also greater in the rotor than the stator. If R < 0.5 the diagram is skewed
D
to the left as in Fig. 5.7c, and static enthalpy and pressure rises are greater in
the stator than in the rotor. A reaction ratio of 50 per cent is usually chosen so
that the adverse pressure gradient over the stage is shared equally by the stator
and rotor. This decreases the likelihood of boundary-layer separation in both ---...,u
the stator and rotor blades and is the condition for maximum temperature rise WI /
/
/
and efficiency. L /
/
/
/
/ w2
5.4 STAGE LOADING / /
/
/

If the power input is divided by the term nlU 2 , a dimensionless coefficient ljJ,
called the stage loading factor, results: Figure 5.8 Lift and drag forces on a compressor rotor blade
 AXIAL FLOW COMPRESSORS AND FANS 191
190 HYDRAULIC AND COMPRESSIBLE FLOW TURUOMACllINES

fJ:x
5.6 BLADE CASCADES

~\r----~~~FI I I
The previous sections have concentrated on relating the required energy
transfer or stage work to the blade inlet and outlet angles for both the rotor
and stator. The next requirement is to decide on the blade shape that will give
", I
I : the required stage work at the maximum efficiency along with the minimum of
" I ,I
pressure loss.
',R~ I
I
I
I In Sec. 2.7.1 use was made of blade element theory to relate the blade lift
"
'I
I I
I
coefficient to the energy transfer across the impeller of an axial flow pump
L
',I I (Eq. (2.61)), the blades of axial flow pumps and hydraulic turbines being of
t-...
I' I low solidity. Axial flow C0l11preSSor (and gas turbine) blading is of high
I " I
_-l __ ~1 solidity, with the result that the gas flow around a blade is affected by the
flow around adjacent blades. In order to obtain information on the effect of
D different blade designs on air flow angles, pressure losses and expected energy
tran.sfer across blade rows, one must resort to cascade wind tunnels and
cascade theory.
Figure 5.9 Resolving blaJe forces into the
direction of rotation
A cascade is a row of geometrically similar blades arranged at equal
distflnces from each other and aligned to the flow direction as shown in
Fig. 5.10. The row of blades is installed on a turntable at the end of a wind
tunnel channel such that the angle of incidence of the blades with respect to the
Woo == Ca/cos f3 00 then approaching air ITIay be varied. Vertical traverses between successive blades
may then be made with pitot tubes, and yaw meters to determine pressure
F x == pC;cIC L sec fJoo[1 + (CD/CL)tan {J ocJ/2 losses and air flow angles. Figure 5.10 is known as a linear cascade and can be
The power delivered to the air is given by imagined as a row of compressor blades unwound from the rotor to form the

UF x == nl(h 03 - 11 01 ) (W)

== pC)s(h 03 - ho 1)
where the flow through one blade passage of width s has been considered.
Therefore
l/J == (11 03 - h01 )/U 2
== F:jpClllsU
== (C1JU)(cls) sec fJtr'(C L + CD tan {J,xJ/2
== (c/s) sec {3 (C L + CD tan {J cxJ/2
00 (5.15)
For rnaximun1 efficiency, the Il1Can flow angle fJ ct.' is usually about 45 and
substituting for this into Eq. (5.15), the optin1un1 bladc loading factor t/f
becomes opt

l/JOPI == (<PI J2)(cls)(C L + CD) (5.16)

If CD is much smaller than CL, which is usually the casc for a well-designed r-
ff
blade, then
l/J PI == (4) I
0 J 2) (cis) C L (5.17) Figure 5.10 Blade cascade
192 HYDRAULIC AND COMPRESSIRLE FLOW TURBOMACHINES AXIAL FLOW COMPRESSORS AND FANS 193

cascade. The number of blades comprising the cascade has to be sufficient to

eliminate any wind tunnel wall boundary-layer effects, and suction slots are 40
often let into the tunnel walls to control the boundary layer. '
Mean 35 :s
The data obtained from cascade testing have to be corrected before
application to a prototype compressor can be made. The reasons for the 30 ~r
I
corrections are because of the difTerences between flow in the actual machine 25 -i
and flow through the cascade. These difTerences are as follows: II
20 ~

Trailing
1. In the machine, annulus wall boundary layers exist at the blade hub and tip. 0.3 edge positions
2. In the machine, alternate blade rows interfere with the perceived cascade tJ~
~
da ta flow pattern. V)

0 0.2
3. In the machine, the solidity decreases from hub to tip. ~
4. Blade velocity varies from hub to tip, thereby affecting the blade inlet angle. ~
I
-:; 0.1
From 3 and 4 it is evident that a cascade test only applies for one radius and
inlet angle, and therefore it may be necessary to carry Ollt a number of tests to
obtain a reliable picture of the flow in the blades. 0 '-----L...<--Jr--'''--~_l_---l _"--Joo.....I..-_ _-..L...----'-L--.L-~_-l.....------.J

5.6.1 Cascade Nomenclature and Curves
Before venturing further into cascade testing and blade design, it is necessary
to define various important angles relevant to the design. In Fig. 5.11 a
cambered blade is shown with a curved camber line through the centre.
The tangents to the camber line at inlet and outlet are the camber angles a'l
e
and a~ to the axial direction respectively. The blade camber angle is defined as
a
~ A number of curves such as Fig. 5.12 are obtained for different incidence
Camber line
Figure 5.11 Cascade nomenclature
2 3 4 6
Distance along trailing edge of cascade
Figure 5.12 Cascade denection and pressure loss curves at one angle of incidence
(a'l - (X~). The chord (' is the distance between blade inlet and trailing edges.
The stagger angle ( is that bet\veen the axial direction and the chord and
represents the angle at \vhich the blade is set in the cascade. The blade spacing
is s and represents the blade pitch. If the air enters with velocity C I at angle aI'
the angle of incidence i is (cx l - a't). The air leaves the blade with velocity C 2 at
angle (X2 and the difTerence ((X2 - ex~) is the deviation angle 6. The air deflection
angle is E ( = ex I - ( 2 ), and it is this angle along with the cascade inlet and ou tlet
stagnation pressures Po t and P02 that are nleasured in the traverse along s. The
results of the tra verses a re usually presented as in Fig. 5.12, the stagnation
pressure loss being plotted as a dimensionless number given by
Stagnation pressure loss coefficient = (Pal - P02)/O.5pCi (5.18)
angles and the mean denection and pressure loss coefficient for each curve,
and (Pal - P02)/O.5pcf, arc plotted against incidence angle as in Fig. 5.13. The
deflection increases with angle of incidence up to a maximum Es . This is the
stall point where separation occurs on the suction surface of the blade, and
since this angle may not be well defined in some designs, it is taken as the angle
of incidence where the mean pressure loss is twice the minimum. It is evident
that for a wide range of incidence, the pressure loss is fairly constant and it is
possible to select an angle of denection 1;* that is also compatible with low
pressure loss as representative for the particular design, and by convention
194 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACHINI':S AXIAL FLOW COMPRESSORS AND FANS 195

~I /~~::
50

I
c* = Z c
~0.5
sic

-J 40

l
.2 O.H
*
~ 20 ~
\0.)
1.0
0.100 c: 1.5
.2

~ ~
t) 30

~
iU

~
0.075 I J 10
a::
iU
"'0

+-~1
I-
V)
.~
o
20
--- c;j

I~
c:
0.050 Twice 0 '6
I
:1

M' . 10 minimum 0
InImum pressure \ I z
10
0.025

OL-~_---.l..-_-L-_...l.----lL...-~-~-----'

-20 -15 -10 -5 10 -10 10 30 50 70 Figure 5.14 Cascade nominal deflection

Incidence, i (deg) Air outlet angle, a 2 (deg) versus air outlet angle

Figure 5.13 Cascade mean dcOection and pressure loss curves

e* = O.Se s ' The nominal deflection angle c* is dependent mainly on the
pitch/chord ratio (s/c) and (X2' this being deternlined froln a large number of
cascade tests. I t is th us possible to plot a set of IHaster curvcs of c* versus (X2 for
different values of (s/c). These curvcs are particularly useful to the designer
when any two of the three variables are fixed. For cxanlplc, if the rotor inlet
and outlet angles fJ I and fJ2 are known, c* can be found, and at anglc [32' (s/c)
can be read from Fig. 5.14.
The deviation angle b is caused by the air not relnaining attached to the
blade over its total curvature. () is given by the elnpirical relationship
coefficient Co may be obtained, where CD is the blade profile drag. The method
of approach is to equate all forces acting on the air to the rate of change of
momentum of the air.
In Fig. 5.15 two blades of a cascade having chord c and pitch s are shown.
At sections 1 and 2 the total air pressures are Po 1 and P02 respectively with
corresponding velocities of C 1 and C 2 , while the density change across the
cascade is negligible. The static pressure change across the cascade is therefore
/c,
-~-------:?...,-:-~-

/U'f
where
nl == 0.23(2{//C)2 + 0.1(a 2/50)
(5.19)

(5.20)
1I:""7"r--:-----;- X

/
/
and a is the distance along the chord to the point of 111axillHlIn calnber. For a /
circular arc canlber line, (2a/c) == I, and this blade fOl'll1 is often chosen.

5.6.2 Cascade Lift and Drag Coefficients

The pressure ratio is governed by the efficiency of the stage and this efficicncy
I~ 2
I
depends on the total drag for both the rotor and stator rows. Froll1 the
measured cascade test results of (Pal - P02), the lift coefficient CL and drag Figure 5.15 Forces and velocities in a cascade
196 HYDRAULIC AND COMPRESSIRLE FLOW TUR130MACHINES
AXIAL FLOW COMPRESSORS AND FANS 197

given by then
D/cos 2 (/.tXl = (p Ce; s/2 cos (/..xJ(tan 2 a 1 - tan 2 a 2 )
(5.21)
- (pee; s/2 cos ~;)(tan2 a 1 - tan 2 a 2 )
where it should be noted that Po 1 > P02 because no work is done in the cascade
and the flow is irreversible. Eq uation (5.21) will be written as + (SPa/COS Ct.~)
The first two ternlS on the RHS are equal and therefore disappear to leave
(5.22)
D=sj)ocOS~l-
where ~P = (P2 - PI) and Po = (Po 1 - P02)'
Dividing the drag by 0.5 pC~c gives the drag coefficient
The summation of all forces acting on the air in the control volume in the x
and y directions must equal the rate of change of momentum of the air in these C n = 2(s/c)(po/pC~,) cos aX) (5.25)
directions. Considering first the y direction, since Cll is constant, there is no But Coo = Ca/cos aa, and ClI = C 1 COS CXl~ thus substituting in Eq. (5.25) gives
velocity change from 1 to 2 in the y direction and consequently no momentum
change. Hence for a unit length of blade, C n = 2(s/(')(iJo/pC~)(cosJ ~,x./COS2 ad (5.26)
A similar procedure Inay be followed for C L by substituting for D and ~P
L sin a'/, - D cos a,t, - s~r = 0
in Eq. (5.24) to give
Therefore
L = (p c,; s/ ens ~,., )( tan (/. t - tan C( 2)
D = L tan a ,x.) - s~fJ/cos a,t, (5.23)
- {L tan (/.7' - (s/cos C(oo)[p(Ci - C~)/2 - Po]} tan a oo
In the x direction the velocity changes from C xt to C x2 ' and noting that these and
are in the negative x direction,
L(I + tan 2 Ct.'xJ = (p('(;s/cos Ct.oo)(tan a 1 - tan ( 2 )
Lcosa,t, + Dsina oo = - PC ll S(C x2 - Cxd
+ (spC:'l;/2 cos (/.<XJ(tan 2 at - tan 2 a 2)tanCt.<x\
= pC; s(tan a 1 - tan a 2 )
- (sjJo/cos a oo ) tan (/.00
and rearranging
L=(pC(:scosaoo)[(tana l - tan( 2)
L = (pC; s/cos acrJ(tan a 1 - tan a 2 ) - D tan a oo (5.24)
+ (tan 2at - tan 2 Ct. 2)(tan (Xl + tan a 2 )/4]
Substituting for L and ~p in Eq. (5.23)
- (sjJ o cos (X 00 ) tan (X 00
D = [(pC s/cos ao-J(tan a 1 - tan ( 2 )
: - D tan aooJ tan a oo
= (pC(~scosaoo)(tan(Xl
l
- tan( 2)[1 + (tana l + tana 2)2/4]
- (s/cos aoo)[p(Ci - C~)/2 - Po]
- (siJo cos a oo ) tan a oo
and
= (pC; s COS aoo)(tan a 1 - tan a 2)/cos 2 a oo - (sPo cos aC(J tan a oo

Now
- (s/cos a oo ) [p( ci - C~)/2 - PoJ

Now (1 + tan 2 a oo ) = sec 2 a oo and tan a oo = (tan at + tan ( 2 )/2. Therefore

D/cos 2 a oo = (pC; s/2 cos aoo)(tan a 1 - tan a 2)(tan a 1 + tan a2 ) Thus

- (s/cos a oo ) [p( ci - C~)/2 - Po] C L = 2(s/c)cos aoo(tan (X 1 - tan a 2 )

and noting that - 2(s/c)(po/ p Ci) (cos 3 aoo/cos 2 a 1) tan a oo

= 2(s/c)cos c(X)(tan at - tan ( 2 ) - CD tan a oo (5.27)
198 HYDRAULIC AND COMPRESSIBLE FLOW TURUOMACHINLS AXIAL FLOW COMPRESSORS AND FANS 199

1.5 2.0
c....;-

C
v
1.0 'C 1.5
CL !B ..l
U
~
u
~
C
0
0.5 :J 'u sic
fE0 1.0
0 1.5
u
~ 1.0
J
0.075 0
0.5 0.5

c
~

~ 0.050 o L.----L_.......l-_.!----L_--..l...-_-L-----'-_---'
~
o -10 10 30 50 70 Figure 5.17 Nominal lift coeffi-
u cients for the cascade
OD Air outlet angle. iX t (deg)
:":j

CS 0.025

Annulus drag coefficient

0l------L..-----l..---..L..-----4-----"--_-----J
- 20 - 15 - 10 - 5 0 10 C DA = 0.02(sl/) (5.29)
Incidence angle, i (deg)
where / is the blade heigh t.
Figure 5.16 Lift and drag coefficients for the cascade from Eqs (5.26) and (5.27)
Secondary losses
CDS = 0.018Ct (5.30)
The air inlet velocity C l ' the incidence angle i and the blade inlet angle
ell are known and therefore :x} ( = 0:'1 + i) is also known. The deviation B is read The total drag coefficient is given by
from Fig. 5.13 for the angle of incidence, and ct.'Xl = tan - 1 [(tan ct.} + tan ct. 2 )/2J
where ct. 2 = (cx 1 - B). C DT = CD + C DA + CDS (5.31 )

Knowing (sic), values of fJo/0.5pC7 can be rcad frol11 Fig. 5.13 for various
incidence angles and upon substitution of these variables into Eqs (5.26) and 5.7 BLADE EFFICIENCY AND STAGE EFFICIENCY
(5.27) curves of CL and CD Inay be plotted against the incidence angle as shown
in Fig. 5.16. Finally the lift coefficien t can be plotted against the air outlet angle It has been shown in the previous section that linear cascade data may be
CX 2 for the nominal value of l;* for a \vhole series of different geol11etry cascades
effectively used to determine the lift and drag coefficients for the cascade and
to give the variation of C L with air outlet angle for a particular (sic) ratio then be modified by the addition of annulus drag and secondary losses in order
(Fig. 5.17).
to approximate the drag coefficient for an annular cascade. After determining
The drag coefficient is usually very slnall in cOlnparison with CL and is eDT from Eq. (5.31), the pressure loss coefficient fJol0.5pCi is determined from
therefore often ignored so that Eq. (5.27) becol11es
Eq. (5.26). The blade row efficiency 1Jb is defined as
C L = 2(slc)coscx-x,(tana1 - tana 2 ) (5.28) Actual pressure rise in compressor blade row
(5.32)
To the profile drag as given by Eg. (5.26) two further drags Inust be added '7b = Theoretical pressure rise in blade row
for reasons described in Sec. 5.6. These are the drag effects due to the walls of
Now the theoretical pressure rise will occur when Po is zero, that is when there
the con1pressor, usually called the annulus drag, and the secondary losses
is no total pressure loss across the cascade. Therefore from Eq. (5.22)
caused by trailing vortices at the blade tips. Ell1pirical relationships exist for
these drags as follows. ~Ptheor = pC;(tan 2 CX 1 - tan 2 cx 2 )/2 (5.33)
200 HYDRAULIC AND COMPRESSIBLE FLOW TURIJOMACHINES
AXIAL FLOW COMPRESSORS AND FANS 201

and hlade efficiency is defined as (p+dp) (r+dr)dfJ

'7h = (L\Pthcor - Po)/ L\Pthcor

= 1- Po / L\Ptheor (5.34)
dr
where both actual and theoretical static pressure differences are expressed in
terms of known angles. For the normal stage, C 1 = C 3' and therefore the stage
isentropic efficiency can be approximated to
(p + dp/2) dr
(5.35)
(p + dp/2) dr
It can be shown that for reaction ratios of 50 per cent, when (T2 - T 1 ) in the
rotor will equal (T 3 - T 2 ) in the stator, the blade efficiency and stage isentropic
efficiencies are the same. For other reaction ratios, the stage efficiency is given
by
'Is = Rryb.rotor + (1 - R)ryb.stator (5.36)
If, for 50 per cent reaction, total values arc used instead of static values, then
the total-to-total pressure ratio is approximately equal to the static-to-static
pressure ratio.

5.8 THREE-DIMENSIONAL FLOW
So far we have been considering only those flows that are two-dimensional in
nature in that only whirl and axial flow velocities exist with no radial velocity
component. In axial flow turbomachines with hub/tip ratios greater than 0.8
this is a fairly reasonable assumption concerning the flow in the annulus, but
for hub/tip ratios less than 0.8, the assumption of two-dimensional flow is no
longer valid. This is seen to be so in the case of aircraft compressors, and at the
low-pressure end of gas turbines where, because of the high mass flow
requirements and the need for as small a frontal area as possible, the trend is to
longer blades on a smaller hub. Radial velocities set up in the blade row can
now result in an appreciable redistribution of the Jnass flow, which can
seriously afTect the blade outlet velocity distribution. Consequently any blade
Rotor blade Stator hlade
Figure 5.1 R Streamline deviation in
rotor row for radial equilibrium
: (one unit)
/
Figure 5.19 Pressure forces on a Ouid element
designed on the two-ditllensional principle could be seriously in error as
regards blade angles. .
Radial flow is caused by a temporary imbalance between the centnfugal
and radial pressure forces acting on the flu.id: W~en thes~ forc~s bala~~e ~ach
other there is no radial velocity and the flllld IS said to be In radial equlhbrlum.
The :nethod of analysis is known as the ~radial equilibrium. method' ~nd
assumes that all radial movement of the streamlines takes place I.n the ~otatlng
blade passages, \vhile on either side the strearnlines are two-dlmen.slonal as
illustrated in Fig. 5.18. The analysis examines the pressure forces acting ?n an
element of fluid, shown in Fig. 5.19, and equates these forces to the ~ent.nfugal
forces acting on the nuid. The stagnation enthalpy Ito at any radius IS then
introduced and the following equation is derived:
dho/dr = C;/r + C x dCx/dr + CadCa/dr (5.37)
Equation (5.37) is the radial equilibrium. e~u~tion and may be applied to
problems in different ways. For instance, If It IS assumed that the energy
transfer is constant at all radii, then dho/dr = 0 and
C;/r + CxdCx/dr + CadCa/dr = 0 (5.38)
Ifwe now aSSllJl1e (as we have previollsly) that in the axial flow compressor
 AXIAL FLOW COMPRESSORS AND FANS 203
202 HYDRAULIC AND COMPRESSIBLE FLOW TURBO~1ACHINES

Ca is constant over the annulus, then dCa/dr == 0 and hence

dCx/dr == - C x/r (5.39) PI
which upon integration gives
C."r == const (5.40) Pa

and this is the requirement for flow in a free vortex with the whirl velocity being
inversely proportional to the radius. The outlet blade angles would therefore
be calculated using the free vortex velocity distribution.
It is not necessary to stipulate that dho/dr be zero or indeed that Ca is
constant at all radii. It might be desirable to specify a radial distribution of
energy transfer, choose a radial distribution for Ca and calculate C x as a
function of r which satisfies Eq. (5.37). Having discussed the principle of three-
dimensional analysis, however, it should be stressed that it is not a panacea for
all design problems since loss of perfonnance due to boundary-layer growth Figure 5.20 Multi-stage compression
and secondary losses detract even fronl the inlproved accuracy of radial
equilibrium analysis.
where the numerator consists of a number of isentropic enthalpy increases. But
as the entropy increases through the compression, so the constant pressure
5.9 MULTI-STAGE PERFORMANCE lines diverge and

The total pressure ratio across a single stage is dependent upon the total
temperature rise across the stage, and the unthinking I1light assume that, to and thus
find the total pressure ratio across N stages, it would only be necessary to raise l1s> l1c
the pressure ratio to the power of the number of stages, such that if POr is the That is, the overall single isentropic compression efficiency i,s less th~n the
pressure ratio for one stage, then the total pressure ratio is given by (POr)N. stage efficiency. The difference also increases with pressure ratio and With the
However, this is not correct, since for the same tcrnperaturc rise per stage, as number of stages. .
the entropy increases, the pressure rise decreases, as cxarnina tion of the To account for cOlnpression in stages, a slnall stage ?r pol.ytrop~c
Mollier chart will show. It is here that the sinall stage or polytropic cl1iciency efficiency is defined for an clel11cntal compression process. T~IS efficlenc~ IS
l1 p is employed. ' idered to bc constant throughout the whole COlnprcsslon. Assuming
In Fig. 5.20, if we cornprcss in a single cornpressioll fro III 1 to 5, the co ns . ' F' 5 21
constant specific heat, then for the elemental cOInprcsslon In 19. .
isentropic work done is
'ip == dT'/dT (5.41)
HI/ nl == (hils - hi) For an iscntropic proccss,
and the isentropic efficiency of the conlpression is T/p(Y- 1)/y == canst (5.42)
'Ie == (hils - h\)/(h ll - hi) Thus
dT' == const[p-(l!Y)(y - 1)/yJdp
Ifwe now compress from 1 to 5 in a nurnber of small finite stages, the isentropic
work done is Substituting for d T' froln Eg. (5.41), and for the constant in Eq. (5.42), then

Ws/nl == (h 2s - h l ) + (h Js - 1l 2) + (h 4s - II]) + (h ss - h4 ) 11 p (d7'/T) == [(y - l)/yJ(dp/p)

and if for similarly designed stages the efficiency 'Is is the saine, then Integrating betwecn the linlits of the full compression from I to II,
In( TII/TI ) == [("I - l)/Yl1 p J In (Pilip)) (5.43)
'ls == (Wj,Il)/(h ll - hi)
204 HYDRAULIC AND COMPRESSIRLE FLOW TURROMACIIINES AXIAL FLOW COMPRESSORS AND FANS 205

1.2 . . . - - - - - - - - - - - - - - - - - - - - - - - - - ,
T

1.0
Stalling in last stage
_________- /
5
~
I
/1:
\
I I
I

Design point - ... , I I 1.1

5//; //
I 2
~ 0.8 L
I II / /

dT'
.i .. '~
S
urge
I' /,' I / "

10

~.;2 0.6 conSlanl_emCle~ncY

lines / " /::~:~/<;> 0.9

/ / / I 0.8 ~ / TIn
U Stalling in first stage / / / /1 // / / N 01
Figure 5.21 Polytropic or small stage 04 .
- \
~ II
//
/1
/11/
1/ //
I (design = I)
compression V /,.. 4 / / 07
//"'/1 1/ / // .
/ t.._ ..... / 1/ II
/ / I

and rearranging, ~//'0.6'L

"'l1t~ ' . . . . . -----)/
/Chokin g in rear stages
(5.44) Stall In 0.5
o many front sta~~~L_ _______l -'I. ~__ ___.I

A typical value for polytropic efficiency is 0.88 and in initial design calculations
it is often assumed that fl p = fls'
If it is assumed that we have equal total temperature rises in each stage,
and denoting the inlet conditions by 01 and outlet conditions at the last stage
as 011, then for N stages,
Pall/POI = (TolI/Tod'/pY/(Y - t)
Also
where!1 To is the stage total temperature rise. It is also usual to assume that the
polytropic and total-to-total stage isentropic efficiencies are equal at a value of
about 0.88.
While it is possible to make a very rapid calculation of pressure rise
through the compressor by this method, the step-by-step calculation of
conditions throughout the machine should still be made, particularly if blade
forms change.
5.10 AXIAL FLOW COMPRESSOR CHARACTERISTICS
A multi-stage axial compressor characteristic is shown in Fig. 5.22. In
comparing this with Fig. 4.10, it is seen that the pressure ratio of the centrifugal
compressor is less sensitive to mass flow variations at a given speed than is the
axial compressor.
In Fig. 5.22, the design mass flow and pressure ratio are at point 1, and
o 0.2 0.4 0.6 a.x 1.0 1.2
fraction of design mT(I,;'l. / POI
Figure 5.22 Axial now compressor characteristic
when operating at this condition, all air flo\v angles and velocities throughout
(5.45)
the compressor arc at their correct design values. We will examine what
happens within the cOfnpressor \vhen the mass OO\V at the design speed is
(5.46)
changed, and when the speed itself is changed. In practice the design point is
very close to the surge line (point 5) and if the mass flow is only slightly
reduced, the pressure ratio and density in the rear stages will both increase.
Since C a = rn/ pA, the axial velocity will decrease and hence the incidence
angle a 1 will increase sharply in the rear stages, thereby causing stalling in
those stages.
Conversely, a small increase in mass flo\v will lead to a sharp drop in
pressure ratio to point 2. The density also drops sharply so that Cn increases.
This results in the large decrease of the incidence angle in the rear stages,
thereby causing stalling in the rear stages with negative incidence.
If the speed of the compressor is reduced so that the operating point moves
to 4, the mass flow and therefore Cn fall faster than the blade speed U, resulting
in an increased incidence angle. A further slight reduction in mass flow along
the constant-speed characteristic will have little efTect on the first-stage density
but will cause a further increase in at and possible stalling in the first stage. If
the mass flow rate is increased at low speed, the likelihood offirst-stage stalling
recedes but the density in the rear stages is very low and consequently C a
 AXIAL FLOW COM PRESSORS AND FANS 207
206 HYDRAULIC AND COMPRESSIBLE FLOW TUR130MACI-IINES

increases until sonic conditions and choking of the flow In the rear stages Determine:
(a) the blade and air angles at the mean radius,
occurs.
(b) the mean radius,
When the design speed is exceeded (point 3), the characteristic eventually (c) the blade height,
becomes almost vertical. The increased speed allows Inore air to be passed at a (d) the pitch and chord and
higher density and pressure ratio. But at the inlet, the l11ass flow increases faster (e) the number of blades.
than the density and choking of the inlet is usually the first to occur. All of the 5.6 Using the data of exercise 5.5, if a circular arc camber line for the blade is assumed, and also the
limiting conditions discussed above lead to unstable or inefficient operation data of Fig. 5.13, determine:
(a) the blade camber angle,
and should be a voided at all tin1es.
(b) the deviation,
(c) the blade stagger,
(d) the total drag coefficient of the blade,
EXERCISES (e) blade row efficiency and stage efficiency,
(f) stage static pressure ratio and
(g) the stage total pressure ratio.
Unless otherwise stated, the work done factor is unity and inlet stagnation Assume zero incidence and a normal stage (C t = C 3 )
conditions are 101.3 kPa and 288 K. For air C p == 1005 J/kg K; R == 287 J/kg K, 5.7 An axial flow compressor has constant axial velocity throughout the compressor of 160 mis, a
y == 1.4. mean blade speed of 244 mls and delivers a pressure ratio of 5: 1. Each stage is of 50 per ce~t
reaction and the relative outlet air angles are the same (30) for each stage. If a polytropIc
5.1 Using the notation given in the text, show that in an axial llow compressor stage efficiency of 88 per cent is assumed, determine the number of stages in the compressor.
h 2 - h}.s = (POJrrl - f'ol.rd)11' 5.8 An axial flow compressor delivers a total pressure ratio of 6, the total. head p~essur~ and
temperature at entry being 0.408 M Pa and 300 K respectively and the ov~rallisentroplc effiCiency
5.2 An axial now compressor stage with 50 per cent reaction has the following data: being 82 per cent. The degree of reaction is 50 per cent and all stages contnbute an equal. amount.of
work. At a particular stage the blade speed at the mean height is 203 mls and the aXial veloclt.y
Air inlet stagnation temperature 290K 171 m/s. If the absolute air angle entering the rotor at this stage is 15 and the work done factor IS
Relative now angle at rotor outlet measured from the axial direction 32 0.92, determine:
Flow coefficient 0.55 (a) the rotor air inlet angle,
Relative inlet Mach number onto the rotor 0.75 (b) the number of stages required,
(c) the static temperature of the air at entry to the rotor and
(d) the rotor inlet relative Mach number.
If the stage is normal, what is the stagnation temperature rise in the lirst stagc of the compressor?
5.9 An axial flow compressor has 10 stages and the following data apply to each stage at the mean
5.3 An axial now compressor stage is to be designed for a stagnation temperature rise of20 K. The
diameter:
work done factor is 0.92 and the blade velocities at the root, mean radius and tip arc 157.5,210 and
262.5 01/s respectively. Thc axial velocity is constant from root to tip and is 157.5 m/s. If the
reaction ratio at the mean radius is 0.5, what are the inlet and outlet air and blade angles at the Blade speed 200 m/s
root, mean radius and tip for a free vortex Jesign? Calculate also the reaction at the root and tip. Reaction 0.5
5.4 An alternative design proposal to that in exercise 5.3 is to have 50 per cent reaction along the Polytropic efficiency 0.88
whole blade. What, then, will the air and blade angles be? Stage efficiency 0.X4
Angle of absolute air velocity at rotor inlet 13
5.5 The design of the first stage of an axial now compressor calls for the following design data:
Angle of absolute air velocity at rotor outlet 45
Work done factor 0.86
Stage stagnation temperature risc 22 K Inlet stagnation pressure 99.3 kPa
Mass now of air 25 kg/s Inlet stagnation temperature 15C
Rotational speed 150rev/s
Axial velocity through stage 157 m/s
Work done factor 0.95 Determine the total pressure ratio of the first stage and the overall static pressure ratio.
Mean blade speed 200 m/s 5.10 An axial flow compressor under test in a laboratory exhibits a stage loading of 0.4 for a
Reaction at the mean radius 50 per cent reaction ratio of 0.65 and now coefficient 0.55. It is decided to reduce the mass now by 7 per c~nt
Rotor blade aspect ratio 3 while the blade speed is kept constant, and it is assumed under this new co.ndition that the ~elatlve
Inlet stagnation temperature 288 K flow exit angles for both the rotor and stator remain unchanged. What IS the stage loadIng and
Inlet stagnation pressure 101.3 kPa reaction at the new condition? Assume the work done factor is 0.9.
208 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW COMPRESSORS AND FANS 209

SOLUTIONS Therefore

Exercise 5.1 From Eg. (5.6), in the rotor of a turbine or compressor the
quantity I is constant from inlet and ho I rei = h02rcl' Therefore
The problem resolves itself into finding II I and Ca'
From Eq. (5.11)

or 2R
tanfJl =-;j;-tan j i 2

2 x 0.5
=~-tanII2
Across the stage the fluid may be assumed incompressible. Therefore

pWi pW~ = 1.19

POlrel = PI + -2- and P02rei = P2 + -2-
fil = 50 deg
Hence Now

from which
Now from Fig. 5.4, along an isentrope
W 2l=yRM rl
2 ( TOI -- ci )
1 2C p
Tds = dh - -dp = 0
p But W 1 = Ca/coSfJI and C 1 = Ca/cosr:t. l , also f32 = a l . Thus
Therefore C 1 = Ca/cos 32 = Ca/0.848
1 and
dh= -dp
p WI = Ca/cos 50 = Ca/0.643
Along 1-2s Substituting for WI and C 1

C 2 = YR M 2
a rt
(~o t _-144C
_. C; ) cos 50
p
2

Thus
= 1.4 x 2R7 X 0.5625 (290 _ C; ) 0.643 2
_POlrel-PI-P02rel+P2 1.44 X 1005
h 2 -h 2s- P2-PI
----
P P = 2.71 X 10 4 - 0.064C;
Po Irei - P02rei C; = 25470
p Ca = 159.6 rn/s
Substituting for the calculated values
Exercise 5.2 Since the stage is normal and reaction 50 per cent, the velocity
159.6 2
triangle is symmetrical as in Fig. 5.7. T02 - To 1 = (tan 50 - tan 32)
From Eq. (5.4) W/m = C p (T02 - Tal) across the stage. Now 1005 X 0.55

U(C x2 - C xl ) = U(WxI - Wx2 ) Stagnation temperature rise = 26.1 K

210 HYDRAULIC AND COMPRESSIBLE FLOW TURllOMACIlINES AXIAL FLOW COMPRESSORS AND FANS 211

Exercise 5.3 Solution begins at the Hlean radius r m and from Eg. (5.9) But

whence Then
tan alt = 0.8 tan 18.78
20 x 1005
tan/31 - tan/32 = 0.92 x 210 x 157.5 At tip, inlet air angle

=0.66
Also from Eg. (5.11) 262.5
tan /3 == - - - tan 15 .21 u
It 157.5

At tip, blade air angle

Thus
0.5x2x210
tan[3 l + tan/3 2= 157.5
At outlet
= 1.333 rC x2 == const at all radii
Eliminating tan /32 gives Therefore
2tan/J 1 =1.99 tan a 2t == 0.8 tan a 2m
fi1 == 44.8 == 0.8 tan 44.8
Then == 0.794
tan /32 = 1.333 - 0.993 At tip, outlet air angle
u
fJ2 == 18.78

ct l = f32 and ct 2 = fJ1 since R = 0.5 at the Illcan radius.

At the tip, from the velocity triangle of Fig. 5.3 at inlct tan Ii 2l == 1.66 - tan a Zt
tana l + tanfJ1 == VjC u == 1.66 - 0.794
262.5 At tip, outlet blade angle
157.5
== 1.666
Similarly at outlet Moving now to the rool, the same equations apply. At inlet
tan a 2 + tan fJ2 = 1.66 tan ex 1 + tan /31 = 1
For free vortex design rC xl == canst at all radii and therefore between the and
mean radi us and the tip rC x1 = const

rmC xm == rte xt VOl

taOCl..lr==utanalm
r
r mtan aIm = r t tan a 1t
Since C a is a constant == ~ tan 18.78
rm 157.5
tan a 1t == - tan alm
rl == 0.453
212 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW COMPRESSORS AND FANS 213

At inlet, air angle At the tip using Eq. (5.11)

157.5
0.5 == (tan r> 1 + tan (32)
2 x 262.5
tan f31 r == 1 - tan 24.38 and using Eq. (5.9)
At root, inlet blade angle 20 x 1005
tan {J 1 - tan IJ 2 == 0.92 x 157.5 x 262.5

== 0.528
At root, outlet air angle Solving simultaneously for tan {3 1
210 2 tan fJ 1 == 1.666 + 0.528
tan (12r == 157.5 tan Ct 21ll
tan {i 1 == 1.097
== 1.333 tan 44.8 At the tip, blade angle and air angles are equal:
(12r == 52.94 {J 1 == 47.64 == (X2

At root, outlet blade angle

Blade outlet angle
tan fi 2r == 1 - tan (12r tan {3 2 == tan 47.64 - 0.528
== 1 - 1.324
== 1.097 - 0.528
== - 0.324
and f3 2r is given by f32 == 29.64 == ex 1

At the root using the same equations

At the tip, reaction 0.5x2x157.5

tan IJ 1 + tan fJ 2 == 157.5

==1

157.5 and
- (tan 54.37 + tan 40.9) 20 x 1005
262.5 x 2
tan {/l - tan f32 == 0.92 x 157.5 x 157.5
== 0.68
== O.RR 1

At the root, reaction whence

157.5 2 tan f31 == 1.8R 1
R == - - - - (tan 28.68 + tan ( - 17.95))
r 2 x 157.5
IJ 1 == 43.23 == (12
== 0.112
tan f32 == 1 - 0.94
Exercise 5.4 Equations (5.11) and (5.9) are used at all radii. The angles at the
f32 == 3.43 == (11
mean radii will of course remain unchanged.
 AXIAL FLOW COMPRESSORS AND FANS 215
214 HYDRAULIC AND COMPRESSIBLE FLO\V TUR130MACI-IIl"ES

Exercise 5.5 (a) Using Eq. (5.9) at the rnean radius T 1 == T ol --

Cr
2C p
;~UCa
T 03 - Tal == -C(tan {3 l - tan {3 2 ) == 288 _ 162.4
2
p
2 x 1005
0.95 x 200 x 157
22 == 1005 (tan {3 1 - tan {3 2 ) 274.88 K
==
Assuming isentropic flow at inlet we may use isentropic relationships at a point
tan /31 - tan f3 2== 0.741
~ == (~)Y/(Y-I)
Using Eq. (5.11)
Pal Tal
Static pressure
274.9 )3.5
PI == 101.3 ( 288
and hence
0.5 x 2 x 200 == 86.1 kPa
tan /3 1 + tan {32 == - - - - -
157 Then
== 1.27 PI
Pi == RT
Solving simultaneously for /31 and fJ 2 I

86.1 x 10 3
2 tan {3 l == 2.011
287 x 274.88
{3 1 == 45.16 == a 2
== 1.09 kg/m 3
(since diagram is symmetrical) and From the continuity equation
tan {3 2 == 1.27 - tan 45.16 A == 25
{3 2 == 14.81 == a 1 0
1.09 x 157
== 0.146m 2
(b) Mean radius
U
Blade height
r
III
==--
2nN h==~
2nr m
200
0.146
2n x 150
2n x 0.212
== 0.212 m
== 0.11 m
(c) The blade height is found [raIn the annulus area of flow as used in the
The variation of air angle across the annulus is determined as in exercises
continuity equation
5.3 and 5.4 and will not be repeated here. Suffice it to say that they will be
depedent upon the type of flow chosen, i.e. free vortex, contant reaction, etc.
Now (d) The nominal deflection curve (Fig. 5.14) for a number of blade
cascades may now be used to find the pitch and chord. At the mean radius,
C 1-~
-
cosa and noting that blades, {J, an equivalent to cascades, a,
l

157 t;* == {J 1 - {J2

== 45.16 - 14.81
0
cos 14.81
== 30.35
== 162.4 In/s
216 HYDRAULIC AND COMPRESSIBLE FLOW TURDOMACHINES AXIAL FLOW COMPRESSORS AND FANS 217

and hence from Fig. 5.14 at the air outlet angle /32 of 14.81 0, the solidity is found: (b) Camber angle
sic = 0.8 0= a'l - ('/.'2
= a'l - a2 +~
. Span
BIa d e aspect ratIo = - - = a'l - a 2 + 0.2330
Chord
The cascade air inlet angle (,(t = compressor relative air inlet angle /31 = 45.16,
0.11 and since there is no incidence on the blade, cascade blade inlet angle a'l eq uals
c=-
3 air inlet angle (Xt.
Blade chord = 0.0366 m
() = 45.16 - 14.81 + 0.2330
Blade pitch
Camber angle () = 39.57
s = 0.8 x 0.0366
= 0.0293 m (c) From the geometry of Fig. 5.11 the blade stagger for a circular arc
Circumference at mean radius cascade is gi ven by
e urn ber 0 f bl a d es = - - - - - - - - - - -
()N
Pitch at mean radius ()
(= a'l - -
2n x 0.212 2
0.0293 39.57
=4516---
. 2
= 45.46
Stagger angle = 25.4
Blade numbers are usually chosen to avoid those with common multiples. This
is to avoid the chances of resonant frequencies being set up between the fixed All information is now available for construction ofa chosen profile blade
and moving blades. In this exercise a choice of 45 is divisible by 5 and 9 and on the circular arc camber line at the angles calculated.
therefore a prime under such as 43 or 47 blades would be chosen. This alters (d) The cascade lift coefficient for (sic) = 0.8 is given from Fig. 5.1 7 as
the calculations slightly and so if we choose 47 blades CL = 0.9 at (,(2 = 14.81 From Eq. (5.30) the secondary losses are
0.

s = 0.0283 m, c = 0.0354m and Aspect ratio = 3.1 CDs =0.018Ct

= 0.018 X (0.9)2
Exercise 5.6 (a) The deviation angle b = (,(2 - (,(~ as shown in Fig. 5.11 and is = 0.01458
given by Eq. (5.19). In Eq. (5.20) for m, (2alc) is unity (for circular arc camber)
and from Eq. (5.29) the annulus losses are given by
and at the mean radius, (,(2 for the cascade equals /32 for the moving blade, (X2
= 14.81. Therefore C OA = 0.02(s/l1) (span = blade height)

2a)2 + 0.1 ((,(2) 0.0283

m = 0.23 ( ~ 50 =0.02 x - -
0.11
14.81) = 0.005 145
= 0.23 x (1) + 0.1 ( 5()
Then using Fig. 5.13 at zero incidence, Po/tpCi = 0.012 and substituting this
=0.26 into Eq. (5.26) gives for the straight cascade
and S Po
- ) cos 3 (,(00
b = mO(slc)1/2 Cn = ( ~ ) ( 1pCi cos 2 a l
= 0.26(0.8) 1/2 0
COS3 (X 00 )

Deviation b = 0.2330 = 0.8 x 0.012 ( cos 2 45.16

218 HYDRAULIC AND COMPRESSIBLE FLOW TURllOMAClIlNES AXIAL FLOW COMPRESSORS AND FANS 219

Now
tan a 1 + tan C(2 0.0323
tan(J.CX) ==---2--- rIb == 1 - 0.468

tan 45.16 + tan 14.81 == 0.931

2 Blade row efficiency = 0.931

Then The stator blade row efficiency will be almost identical for 50 per cent
CI)= 0.8 x 0.012 x C~S3 J2AI reaction (at the mean diarnetcr). DifTerences in the height a.nd blad~ pitch
cos 2 45.16 between rotor and stator rows can afTect C OA but, as shown In (d), thIS drag
constitutes only a small part of the total drag and its effect therefore is
== 0.0116
negligible.
Therefore substituting for the losses into Eq. (5.31) the total drag coeffi The stage efficiency for 50 per cent reaction is found from Eg. (5.36) to
cient is be the same as the blade efficiency,
COT == CD + C OA + Cos tIs = 0.931
== 0.0116 + 0.005145 + 0.014 58
Blade drag coefficient == 0.0313 ([) The stage static pressure ratio is found from Eg. (5.35)
T 1 T T) - (-P3 )(Y-l)!Y
(e) l!sing Eq. (5.26) again to find the pressure loss coefficient in the moving
3ss
-==
T1 ( +t1s
3 -

T1
1 _

PI
blades SInce the same equations and arguments will apply to circular cascades
as well as straight cascades Now C 1 == C 3 for a normal stage and therefore the stage static temperature
rise equals the stage stagnation temperature rise

and
22
P3 == 1 + 0.931 ( - -
)]3.5
Pi [ 274.88
Po 0.0313 X cos 2 45.16
1- pCi 0.8 cos 3 32.41 == 1.29

== 0.0323 (g) ~O~ == 1 + 0.931 ( -

22 )]3.5
P02
[ 288
Now the theoretical pressure rise through the row of blades is given by
Eq. (5.33) when Po is zero and this leads to the blade efficiency of Eq. (5.34). == 1.27
Blade efficiency
Exercise 5.7 Since we have 50 per cent reaction the velocity diagrams are
symmetrical as in Fig. 5.7a, and ex 1 == fJ2' a 2 = /1 1 , The number of stages may be
Dividing Eg. (5.33) by pCi/2 we get found from Eg. (5.46)
T ol +Nt1To
t1P-
- heor
lpC2
t
= (C-C )2 (tan
a 2
a -tan 2 a)
1 2
T'oll
T OI T OI
2 1 1
Using Eq. (5.8) the stage stagnation temperature rise is
= cos 2 a 1 (tan 2 a 1 - tan 2 ( 2)

= cos 2 45.16 (tan 2 45.16 - tan 2 14.H 1)

= 0.468
220 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW COMPRESSORS AND FANS 221

Now tan fJI = 0.919

Wx2 = Ca tan P2 P1 = 42.6 = (12 for 50 per cent reaction
= 160 tan 30 Rotor inlet angle fJ 1 = 42.6
= 92.38 m/s
(b) The overall isentropic efficiency is obtained by working between the
Therefore
WI = C 2 = [(U - + C~J1/2
W x2 )2 compressor inlet and outlet conditions.
= [(244 - 92.38)2 + 160 2 J1 / 2 Isentropic work done
Overall isentropic efficiency = A I kd
ctua war one
= 220.4m/s
cos PI = C a/W 1 TOils -T OI
T on - T OI
160
220.4 = (TOilS _ 1 )/I(TolI _
TOI TOI
1)
fil = 43.45
Therefore Substituting for (TOil/TO!) from Eq. (5.45) and (TOIls/TOI ), using isentropic
244 x 160 flow relationships we get
11 To = 1005 (tan 43.45 - tan 30)

= 14.37 K
Equation (5.44) gives the overall temperature ratio as 0.82 = (6286 - 1)/(6 o. 286 /r]p - 1)

~=~
rr (p )<Y - I )!'7PY
6o. 286 /"p = ( 1.669 - 1) + 1
TOI POI
0.82
= 50.4/(0.88 x 1.4) = 1.816
= 53247 0.286
--log 6 = log 1.816
= 1.68 '1 p
Thus Polytropic efficiency
(1.68 x 288) - 288
I]p = 0.859
N = 14.37
Therefore
= 13.3 T
-~!~ =
(p~ )0.286/0.859

Number of full stages = 14 TOI POI

= 1.82 K
Exercise 5.8 (a) The velocity diagrams of Fig. 5.3 are once again symmetrical
Substitute into Eq. (5.46) to find the number of stages
and ell = 15 = P2'
TOil T OI + N~ To
= 171 x tan 15 T OI T OI
= 45.8 m/s where
~ To = Stagnation temperature rise in the stage
U-C and
tan Pl = xl
Ca C p ~ To = t/J Ca U(tan P1 - tan fJ2) from Eq. (5.9)
203 - 45.8 = 0.92 x 171 x 203(tan 42.6 - tan 15)
171 = 20.81 kJ/kg
222 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW COMPRESSORS AND FANS 223

Number of stages Exer~ise 5.9 (a) The axial velocity is first found from the velocity triangles
and SInce they are symmetrical for 50 per cent reaction from Fig. 5.7a, (Xl = /32

lV == [
(1.82 x 300) - 300
20.81 x 10 3
J1005 and (X2 = /31' Therefore

== 11.8
== 12 stages
c == 200
a tan 45 + tan 13
(c) At entry to rotor
= 162.5 m/s
Work done in the first stage is given by Eq. (5.9) as
Now
Cp (T03 - TOI ) = t/JC a U(tan /31 - tan /32)
c-~
1 - cos (Xl
= 0.86 x 162.5 x 200(tan 45 - tan 13)
= 21497 J/kg
171
For the first stage, the pressure ratio is given by Eq. (5.7):
cos 15
= 177
P03
_=
(
1+11c
To - T.
3 01
)y/(y- 1)

Therefore POI T01

177
T l =300- 2 x 1005-
2
= (1 + 0.84 x 21497)3.5
288 x 1005
= 300 - 15.58 = 1.24
= 284.4 K
To find the overall static pressure ratio, we may use Eq. (5.44):
(d) The rotor relative inlet Mach number is given by
M lr == Wl/al
where
(J I == ()' RT 1) 1/2 Now
== (1.4 x 287 x 284.4)1/2 TOil = TOI + lOx ~ To
== 338 n1/s
= 288 + lOx 21 49~
Also 1005
~v~ == (U - C xl )2 + C~
== (203 - 45.8)2 + 171 2 = 502K
But
== 53953
W l == 232.3 nl/s
Therefore substituting
232.3 162.5
A1 rl == 338

== 0.687 = 166.8 m/s

224 HYDRAULIC AND COMPRESSIIlLE FLOW TURBOMACJlINES AXIAL FLO\V COMPRESSORS AND FANS 225

Therefore Equation (5.11) gives the reaction ratio

C 21 q)
T1=TolI --
2C p R = -(tan fJ l + tan 1>2)
2
166.8 2
=288---- Solving simultaneously for 1/1 and 1>2
2 x 1005
= 274.1 K lan/II =(R+tJj
and
TIl
C2
= TOII - -I lanfI2= ( R- ~) I
;./
2C p 2
2 Therefore under the initial now conditions
=502- 166.8
2 x 1005 I
tan 1/ I = ( 0.65 + -0.4) I 0.55
= 488.1 K 1.8 I
Thus 1/1 = 57.76
PII = (488.1 )0.88 x 1.4/0.4 and
PI 274.1
Overall static pressure ratio = 5.91
Ian /1 2 = ( 0.65 - ~::) I0.55
1/2 = 37.87
Exercise 5.10 The velocity diagrams at both conditions are sketched below From the velocity triangles
(Fig. 5.23) and it will be noted that they are no longer symmetrical. When the
mass flow is reduced, axial velocity C a is also reduced, and noting that it is
angles /32 and a l that remain constant, the dotted lines give the new velocity Therefore
triangles. The angle a l is constant as this is the exit angle of the air from the tan a 2 = _1_ - tan 37.87
previous stator. 0.55
Equation (5.14) gives the stage loading
l~ = A(tan a 2 - tan a.) Also
= )~(tan PI - tan fJ 2)
then
tan a l = _1_ - tan 57.76
0.55

At the reduced mass now

= 0.55 x 0.93
= 0.5115
Then
tan fJl = _1_ - tan 13.1
0.5115
u Figure 5.23 fJ 1 = 59.86
226 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACIIINLS

Also
1 CHAPTER
tan Cl. 2 = -- - tan 37.H7-'
0.5115
SIX
New stage loading AXIAL FLOW STEAM AND GAS TURBINES
lj; = 0.9 x 0.5115 (tan 59.86 - tan 37.87)
= 0.435
New reaction
0.5115
R = -2-(tan 59.86 + tan 37.87)

=0.64

6.1 INTRODUCTION

In this section, steam and gas turbines will be considered together, with the
assumption that the same theory applies to both types of machines. This
commonality is valid providing the steam is in the superheated state and
therefore behaves approxin1atcly as a perfect gas. Care should be taken, when
the condition of the steam falls below the saturation line on the Mollier chart,
that the theory and equations developed in later sections are not applied in
that case. This would n10st likely occur at the low-pressure end of the steam
turbine.
Axial flow turbines were first built and used successfully by Sir Charles
Parsons when he installed a reaction steam turbine in a small marine vessel for
propulsion. Since then development of the steam turbine has been rapid and
very successful, the power output at the present time ranging from a few
kilowatts to 660 MW. The development of the high outputs is due in part to
the relatively easy matter of superheating steam in a boiler and superheater,
then expanding it through the turbine to below atmospheric pressure in the
condenser to extract the maximum energy from the steam.
Development of the axial flow gas turbine was hindered by the need to
obtain both a high enough flow rate and compression ratio from a compressor
to maintain the air requirement for the combustion process and subsequent
expansion of the exhaust gases. Initially the air was provided by centrifugal
compressors, and later the axial flow compressor was developed, which, in the
case of turbojet and turbofan engines, uses the power developed by the turbine
to drive it. Steam turbines are universally used in fossil fuel power stations and

227
228 HYDRAULIC ANI> COMPRESSIRLE FLOW TURBOMACIlINES
AXIAL FLOW STEAM AND GAS TURBINES 229

for steam-driven propulsion in ships, although gas turbine propulsion units rixed n07.1.Ie rings Stator casing
are often fitted in the smaller class of naval vessel. Gas turbines are universally
used as the power unit for large jet aircraft propulsion, their advantage being Stator
~"S"r-""-..J-f-.f"-L-"""""""~-+YL;~--'-d-"7"'Icasin g
that they have a high power-to-weight ratio.
The principle of energy extraction from the gas is one of gradually Outlet Rotor
reducing the high-pressure energy by converting it into kinetic energy. This is
accomplished by passing the gas alternately through rows offixed and moving
blades. The kinetic energy of the gas is reduced in the moving blades, which are
attached to the turbine hub, and recovered in the fixed stationary blades
attached to the casing. This necessitates a gradual density decrease as the gas Velocity
moves through the turbine and the blade height therefore increases towards
the low-pressure end, if a constant axial flow velocity is to be maintained
through the turbine. The stator row is often termed the nozzle row and in
certain types of steam turbine the nozzle row consists of a set of converging
nozzles spaced around the drum.
Figure 6. t shows a steam turbine rotor in the bottom half of its casing. It
will be noted that the blade height increases in the direction of gas flow. The
two important types of axial flow turbines are the reaction and impulse types, Pressure
\I Pressure
and combinations of each type may be found in a single turbine. Blade types (a)
are discussed in Sec. 6.6 but Fig. 6.2 describes how the velocity and pressure (b)
vary through impulse and reaction stages. The overall trend is a decreasing
Figure 6.2 Pressure and velocity through impulse and reaction axial flow turbines: (a) impulse;
pressure with velocity recovery in the stator row or nozzle ring. A row of stator
(b) reaction
blades followed by a set of rotor blades is considered to be a stage and

conditions through the stage \vill be examined In detail. The following

assumptions will be made:

1. Flow conditions will be evaluated at the mean radius unless otherwise

stated.
2. Blade height/mean radius is small, allowing two-dimensional flow theory
to be used.
3. Radial velocities arc zero.

6.2 TURlllNE STAGE

Figure 6.1 160 MW axial flow steam turbine
(courtesy of Escher Wyss Ltd)
A single turbine stage and velocity triangles are illustrated in Figs 6.3 and 6.4
respectively. The inlet to the stator blades is designated section 0, inlet to the
rotor section 1 and outlet from the rotor section 2. This numbering system
keeps the subscripts of the variables the same as in the case of the axial flow
compressor. All now angles are measured from the axial direction and care
must be taken when reading steam turbine literature in which it is customary
to measure now angles from the direction of blade motion.
230 HYDRAULIC AND COMPRESSIBLE FLOW TURUOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 231

each other as shown in Fig. 6.4, from which the energy transfer is given by
E = U(C x1 - Cx2 )/g
and since Cx2 is in the negative x direction, the work done per unit mass flow is
Eg = W/m = U(C x1 + C x2 ) (6.1)

= U(W x1 + W x2 ) (6.2)

If Cal =1= Ca2 there will be an axial thrust in the flow direction. However, we
will assume that C a is constant and therefore
u W/,n = UCa(tan (,(1 + tan (,(2) (6.3)
= UCa(tan P1 + tan /J 2 ) (6.4)
Rotor row u
Equation (6.4) is often referred to as the diagram work per unit mass flow and
associated with this is the diagram efficiency defined as
-.1 Figure 6.3 Velocity triangles for an
axial now gas (or steam) turbine Diagram efficiency =
Diagram work done per unit mass flow
. . n
Work avallable per unit mass ow
stage
= U(Wx1 + Wx2 )/-}ci (6.5)
The thermodynamic conditions of the gas through the stage are shown in
Fig. 6.5, where the inlet conditions to the stator are at total pressure POO and
Work done total enthalpy hoo . For adiabatic now through the stator row or nozzle ring,
I~ ~I hoo = hOI' but, owing to irreversibilities, the total pressure drops to POl at
stator outlet (rotor inlet). Ex pansion to P02 and total enthalpy h02 takes place
:I
CJ.'l
I
I
in the rotor rowand, as shown in Eq. (5.6), 11 01 rei = h02rel ' The work done
per unit mass flow by the gas is given by
I I('u W/n, = hoo - h02 = ho 1 - h02 (6.6)
or
~---------i~t------1~
u ~ W/n, = Cp{TOI - T 02 ) (6.7)

Substituting for Win' frOlll Eq. (6.4),

Figure 6.4 Superimposed velocity triangles
Cp(TOI - T o2 ) = UCa(tan fJI + tan fJ2) (6.8)
It should be noted that the work done factor ). is not used in Eq. (6.8). This
is because, in a gas or steanl turbine, flow through the blade passages is
The gas leaves the stator blades with absolute velocity C 1 at angle (,(1 and, accelerating as opposed to decelerating. Flow in the compressor and the effect
by subtracting the blade velocity vector U, the relative velocity vector at entry of boundary-layer growth in the fonner are therefore negligible. For a normal
to the rotor W l is detennined. In 1110ving across the rotor blade, the flow stage in which Co = C 2 , the static tcnlpcrature drop across the stage equals the
direction is changed, and the pressure reduced while the absolute velocity is total temperature drop.
decreased and the relative velocity increases. The gas leaves the blade The turbine stage total-to-total isentropic efficiency rJl(t-t) is defined as
tangentially at angle fJ2 with relative velocity W 2. Vectorially subtracting the
Actual work done by the gas
blade speed results in the absolute velocity C 2' This is now the inlet velocity to
llt(l-I) = Isentropic work done
the next stator row at angle 0: 2 , which for a nonnal stage equals Co at (,(0
respectively. The two velocity triangles are conventionally superilnposed upon = efoo - 'r02 )/('[00 - 'r02s J (6.9)
232 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES
AXIAL FLO\V STEAM AND GAS TURBINES 233

Poo
h to the work output. It is defined as
POI '1 1(I-s) = (h oo - h02 )/(11 00 - h25S) (6.11 )
()()
01
Considering the total-la-total efficiency of a normal stage,
('j
-..........
Po GO '11(1 n = (hOD - h 02 )/(11 00 - h02s sl
0 = (11 0 - h 2 )/(11 0 - h 2s J
For a nornlal stage Co = C 2 and ~o = ~2' and upon rearranging
17 1(1-1) = (h o - 11 2 )/[(h o - 11 2 ) + (11 2 - h2 J + (11 2s - h 2ss )J (6.12)
But consider now the slope of a constant-pressure line on the Mollier chart
(Fig. 6.5). Since <.Ill = T d.\' then the slope along a constant-pressure line is given
by
(DIl/rJs)p = T (6.13)

and therefore for a finite change of enthalpy L1h at constant pressure

~h ~ T~s
Is P02
Therefore
02ss (6.14)
and
2 Figure 6.5 Mallier chart for expansion
2ss through an axial now steam or gas But exanlination of Fig. 6.5 shows
turhine stage
(S2s - S2sJ = (SI - SIs)

Thus Therefore substituting into Eg. (6.14) gives

Too - T 02 = 1]1(1-1) ToO (1 - T02ss/Too) ("2" - "2sJ = (T 2 /Td(11 1 - hlJ (6.15)
Too - 7~)2 = 1]l(I-t) Too [1 - (P02/ Poo )(y- I )!y] (6.10) For both the stator (nozzle) and blade ro\vs, dimensionless loss coeffi-
cients may he defined in two \\'ays. For the nozzle
6.3 STATOR (NOZZLE) AND ROTOR LOSSES (N = (II I - Il I s)/O.5cf or
and for the rotor
Before considering the losses occurring in the rotor and stator rows, it will be
worth while distinguishing between two isentropic efficiencies commonly used or
in axial flow turbine work. The first has already been defined in Eq. (6.9) where
The value of Y or ( in the stator and rotor represents the percentage drop of
the temperature limits are taken between total temperatures, and hence,
energy due to friction in the blades, which results in a total pressure and static
defined in this way, the efficiency is the total-to-total efficiency. This form is
enthalpy d rap across the blades. Typical losses are of the order of 10-15 per
used when the kinetic energy at the outlet of the stage is utilized for producing
cent but can be lower for very low values of flow coefficient.
work. Such a case would be the propelling nozzle of a turbojet exhaust and any
Substituting for the blade loss coefficients into Eg. (6.12), we get
intermediate stage of a multi-stage turbine where the leaving kinetic energy is
used in the following stage. The other efficiency is the total-to-static efficiency, 1]1(1-0 + W~(R/2 + (T2 /T1 )(NCi/2J
= (h o - h2 )/[(h o - 11 2 )
which is used when the leaving kinetic energy is wasted, thus not contributing = { I + [(R W~ + (NCi(T 2 /T 1 )]/[2(h o - h 2)J} -I (6.18)
234 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 235

If the exit velocity is not utilized, then following a similar procedure for the where (1 and (nom are the loss coefficients at the new aspect ratio and at an
analysis, the total-to-static efficiency beconles aspect ratio of 3 respectively.
Also, in Eqs (6.21) and (6.22) I is the blade length, while a Reynolds number
tlt(l-S) = {I + [(R W~ + (NCi(Tz/T1) + C~J/[2(llo - liz)]} -1 (6.19) 5
of 10 5 is still assumed. Should the Reynolds number not be 10 , a further
As.a starting point for a calculation, since T z and T 1 are not very different, the correction may be made as follows. The Reynolds number for the flow is given
ratIo (Tz/T 1 ) is often taken as unity. Typical values of ( are given by by
( = 0.04 + 0.06([,/100)2 (6.20) Re = (2p 1 siC 1 cos c(
1 )/[,u(s cos a 1 + I)J (6.23)
5
where the deflection angle c is l; = (Xo + cx 1 for the stator (nozzle) and and if (1 is the loss coefficient at a Reynolds number of 10 , then at the new
= /31 + /32 for the rotor.
f, Reynolds number, the corrected loss coefficient is given by
Equation (6.20) applies only for a Reynolds nUI11bcr of 10 5 , based on the (z = (10 5/Re)o.z5(1 (6.24)
characteristic velocity C 1 at the stator outlet and the characteristic hydraulic This empirical method of loss prediction is based on the Soderberg
diameter dH defined as correlation 6 and is derived from the analysis of a large number of turbine
4 x Flow area at stator outlet characteristics. The error over a wide range of Reynolds numbers is as low as 3
(I H = -.------------- per cent. Nozzle loss coefficients obtained from a large number of turbine tests
Wetted perimeter at stator outlet
are typically 0.09 and 0.05 for the rotor and stator rows respectively. The total-
With reference to Fig. 6.6, to-total efficiency is in the region of 90 per cent, the variation with blade
dB = (4slcosctd/[2(scosal + I)J loading and flow coefficient being plotted in Fig. 6.7, where it is seen that low t/J
and give high stage efficiencies.
= (2 sl cos ct 1 )/(s cos ex 1 + I)
The ratio of nlaxiITIUm blade thickness to chord (tmax/c) is 0.2 and the blade
aspect ratio (span/chord) based on the axial chord (i.e. c cos cxa:J is 3. For
conditions other than the above, for instance if the aspect ratio is different, 6.3.1 Blade Loading Coefficient
the following empirical equations nlay be used:
The work capacity of the stage is expressed in terms of a temperature drop
For rotors coefficient or blade loading coefficient
1 + (1 = (l + (nonJ[O.975 + (0.075ccosacxJ/IJ (6.21 ) t/J = W/nlU z = Cp (T 01 - Toz)/U z (6.25)

For stators (nozzles) which from Eq. (6.4) may also be written as

1 + (1 = (1 + (nom)[0.933 + (0.021ccos ctcxJ/IJ (6.22) t/J = Citan /31 + tan /32)/U

= (tan /31 + tan /3z) (6.26)
I~ .. \
where is the flow coefficient.
The implication of a low flow coefficient is that frictional losses are
/ = length of blade reduced in the stage since Ca is low, but for a given mass flow rate the annulus
flow area would be large. Low values of t/J imply a small amount of work done
per stage, and therefore for a required overall power output, a large number of
d = 4s/cos.):1
II 2(scosa}+I)
stages is required.
In stationary industrial power plants where the specific fuel consumption
is of prime importance, a large-diameter, relatively long turbine, of low flow
coefficient and low blade loading, giving a high efficiency, would probably be
accepted. However, gas turbines used in aircraft propulsion have minimum
Figure 6.6 The hydraulic diameter weight and a small frontal area as prime considerations. This means using
236 HYDRAULIC AND COMPRESSIRLE FLOW TURROMACIIINES AXIAL FLO\V STEAM AND GAS TURBINES 237

and substituting for (hi - "2) in Eg. (6.27),

2.5
R = (W~ - Wi)/[2(h ol - ho2 )]
7J I (I-I)
0.90 = (W~ - Wr)/[2U(C xl + C x2 )] (6.28)
..:::> 2.0
If it is assurncd that the axial vclocity C" is constant through the stage then
~
f....~ R = (W.;2 - l1'.;\)/[2U(U + W x \ + W x2 - U)J
1.5
I =(Wx2 - + Wxl )/[2U(Wxl + Wx2 )]
W. d )(VVx2
f....;

C,.)
Q, = C(,(tan I~ 2- tan fi 1)f2U
1.0
/I = 4>(tan I~ 2 - tan 11.)/2 (6.29)
~

Equation (6.29) can hc rcarranged into a sccond form,

0.5
R = (C n tan II 2 )/2U - [(Ca tan (Xl)/2U - U/2U]
= 0.5 + C,,(tan fi2 - tan 1)/2 U
(X (6.30)
o 0.5 1.0 1.5
and a third form is given by substituting for
Ca/U =
Figure 6.7 Blade loading coefficient versus flow coefficient tan IJ 2 = (U + Cx2 )/C a = (tan C(2 + U/Ca)
into Eg. (6.30), giving

higher values of flow coefficient and blade loading factor to give a shorter
R = 0.5 + Ca(tan C(2 + U/C a - tan C( 1)/2U
compact turbine but also, as a consequence, accepting a lower efficiency = 1 + C a (tanC(2 - tana l )/2U (6.31)
(Fig. 6.7). Inspection of Eg. (6.29) reveals that for zero reaction, f32 = f31 and
W1 = W2 The gas conditions through the stage, and the accompanying
velocity triangles, are shown in Fig. 6.8, where the velocity triangles are
6.4 REACTION RATIO skewed to the left. Ideally, for reversible adiabatic flow, the points 1, 2 and
2s on the Mollier chart should coincide, and in that case no pressure drop
The reaction ratio has more meaning in the case of an axial flow turbine than occurs in the rotor. Now consider a pure impulse stage where, by definition,
for an aXi~1 flow c~mpress~r where it is usually set at 50 per cent for the stage. there is no pressure drop in the rotor. Figure 6.9 shows the MolHer chart
The .tur?lne reactI?n ratIo can vary widely from 100 per cent to 0, the for the pure impulse stage where, for reversible adiabatic flow, the points 1,
ImplIcatIons of whIch are described in this section. 2 and 2s will coincide, and therefore with isentropic flow conditions prevailing
The reaction ratio is
R = Static enthalpy drop across rotor h
Po
Static enthalpy drop across stage o
02rel
= (It I - h 2 )/(h o - h 2 )
= (It I - h2)/[(h oo - C5/2)- (h 02 - C~/2)J

But for a normal stage Co = C 2 and since hoo = h oI in the nozzle, then 2

R = (hI - h2)/(h ol - h02 ) (6.27) v

Remembering that h Olrel = h02re" then (a) (b)

hOI reI - h02rel = (h l - h2 ) + (Wi - W~)/2 = 0 Figure 6.8 Zero reaction axial gas turbine
 238 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 239

Po Po

o h
h

Is
2ss 2s
u
Figure 6.11 A 50 per cent reaction stage
Figure 6.9 Pure impulse stage

the zero reaction stage is exactly the same as an irnpulse stage. However, Po
h
when the flow is irreversible, the zero reaction and ilnpulse stages are not o
the same and in fact an increase in enthalpy occurs in the rotor of the impulse
stage, the implication being that the reaction is negative. For a nega~ive
reaction stage, the tendency will be for W2 < WI' thereby causing diffusion
of the relative velocity vector in the rotor and a subsequent rise in pressure.
This should be avoided since adverse pressure gradients causing flow
separation on the blade surfaces can then occur. Figure 6.10 illustrates the
Mollier chart for negative reaction. u
With 50 per cent reaction, Eq. (6.30) requires lhal/J 2 = aI' III = tt 2 Also Figure 6.12 A 100 per cent reaction stage
C t = W 2 and C 2 = Wt, resulting in the sylnn1etrical velocity diagram of
Fig. 6.11 with equal enthalpy drops in the stator and rotor.
For 100 per cent reaction, Eq. (6.31) gives a l = a 2 and C 1 = C 2 , with the
velocity diagram skewed to the right as illustrated in Fig. 6.12. Increasing the h

reaction ratio to greater than 1 gives rise to diffusion in the stator passages or
nozzles with C 1 < Co This situation should also be avoided because of the
likelihood of flow separation on the stator blade surfaces (Fig. 6.13).

h Po
o
Figure. 6 13 Stage expansl'on with reaction more than
I ()() per cent

6.5 EF FECT OF REACTION RATIO ON STAG E EFFICIENCY

Reaction and impulse blading lind use in different typ~s of machines. In stea~
turbines where, at the high-pressure end, pressure ratIos acro~s a stage ~ou
Is b y high impulse blading is almost universally used sInce there IS no
Figure 6.10 Negative reaction stage p~e~:~re dro~ in the stage and therefore no tip leakage of the steam from
240 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 241

one blade row to the next. At the low-pressure end of steam turbines and
in gas turbines where the pressure ratios are more modest, reaction blading
fl. (small angle)
is employed. A choice of 50 per cent reaction means a sharing of the total
expansion between the stator and rotor rows.
It was seen in Fig. 6.7 that, for a high total-to-total efficiency, the blade
loading factor should be as low as possible, which implies as high a blade
speed as possible, consistent with allowable blade stresses. However, the
variation in total-to-total efficiency with slight variation in the blade loading
coefficient is very small, this being true for a wide choice of reaction ratio.
In contrast, the total-to-static efficiency is heavily dependent upon the
reaction ratio and it can be shown that the total-to-static efficiency at a given
~ --=_Almost straight
blade loading may be optimized by choosing a suitable value of reaction. at trailing edge

Figure 6.14 Reaction blading

6.6 BLADE TYPES

It has been noted in the previous section that steam turbines are usually
impulse or a mixture of impulse and reaction stages whereas gas turbines
tend to be always of the reaction type. The pressure ratio of steam turbines
can be of the order of 1000: 1 but for a gas turbine it is usually no more than
10:1. It is obvious that a very long steam turbine with many reaction stages
would be required to reduce the pressure by a ratio of 1000: 1, and even if
the pressure drop per stage were made large to red uce the number of stages,
blade tip leakage loss would still lead to very inefficient operation. Therefore
reaction stages are used where the pressure drop per stage is low and also
where the overall pressure ratio of the turbine is relatively low, as would be
the case in an aero-engine, which may typically have only three or four
Figure 6.15 Shrouded gas turbine reaction
reaction stages of or near 50 per cent reaction at the mean radius. blades
The requirement of different types of stages has led to differing designs of
blades for each type, and this section describes some of the pertinent points
relevant to each design. the blade evolved empirically in steam turbine practice, and comparisons
between blades designed in the late nineteenth century for the first steam
turbine and those designed by the later aerodynamic theory showed only
6.6.1 Reaction Blading about a 4 per cent increase in maximum efficiency. However, efficiency
As described in Fig. 6.2, the pressure reduces through succeeding stator and increases of only 0.5 per cent or less are now very important when fuel costs are
rotor rows, the velocity being recovered as the pressure drops, and this taken into account. Reaction blading is often shrouded at the tips, especially if
necessitates a blade passage that is convergent towards the outlet, as in the blades are long. This aids in preventing excessive vibration by tying the
Fig. 6.14. blades together and thus changing them from cantilevers to blades fixed at
For 50 per cent reaction the stator and rotor blades will be the same, both ends (Fig. 6.15).
whereas zero reaction implies impulse rotor blades with constant
cross-sectional area passages and no change in now velocity. Reaction of 6.6.2 Impulse Blading
100 per cent implies that the stator blades are of the constant-area impulse
This type of blading has developed historically from steam turbine practice.
type. It will be noted that the inlet angle fJ 1 for the reaction blade is almost
It was noted in Sec. 6.1 that velocity triangles were drawn at the mean radius
zero while the profile of the back of the blade is almost linear. The form of
242 HYDRAULIC AND COMPRESSIBLE fLOW TURUOMACIIINES AXIAL FLOW STEAM AND GAS TURBINES 243

of the blades. Because of the tip speed variation along the lcngth of the blade,
incidence angles should change fronl root to tip, resulting in a twisted blade.
However, it has been customary to ignore the t\visted blade requirenlent and
to have an untapered, untwisted blade of either 50 per cent reaction or zero
reaction (impulse). The loss of efficiency of the straight blades is very small
for the short blades that are used at the high-pressure end of the turbine.
But for the much longer blades at the low-pressure end, larger losses can Constant-
area passage
occur, and blades in this region are now designed on the free vortex principle
instead of constant reaction ratio. For reversible flow through the rotor,
impulse and zero reaction blading are the same.
Impulse blading is employed successfully at the high-pressure end of
steam turbines. The velocity of the steam is increased in thc convergent nozzle
row to perhaps 800 m/s before entering the rotor blades and passing through
them at constant pressure as in Fig. 6.2b. Fronl Eq. (6.5), we can rearrange
(Wx1 + Wx2 ) as
Figure 6.16 Impulse turbine blades
WXI + Wx2 = WI sin P1 + W2 sin Il 2 (6.32)
and putting PI = P2 for impulse or zero reaction,
Wxl + W x2 = WI sinPl{l + WR )
= (C 1 sinal - U)(l + vVR ) (6.33) the chances of flow separation on the back convex surface of the blade are
reduced by decreasing the radius of curvature of the convex surfa~e. The
where WR is thc relativc vclocity ratio W 2 IW I Substituting Eq. (6.33) into
slightly convergent passage aids in preserving uniform flow as the flUId flows
Eq. (6.5)
into the bend and although the diverging section causes diffusion of the
Diagranl efficiency = 2U(C 1 sin a l - U)(l + WR)/Ci flow, the net ~fTect is beneficial when the turning angles are large and radii
= 2[(UIC 1 )sina l - (UIC l )2](1 + WR ) (6.34) of blade curvature are small.

For maximum diagram efficiency, differentiate Eq. (6.34) with respect to

(U IC 1) and equate to zero. Then
O=sinCX: I -2U/C I
or
(6.35)
"Equation (6.35) indicates that the nozzle angle a 1 should be as high as
possible, the ideal being 90. Howevcr, a 1 is lilllited by C a , since the larger
a I' the smaller Ca becoInes and therefore the longer is the blade length to
accommodate the required Il1aSS flow rate. Typical nozzle angles are between
65 and 78. The rotor blade passages arc usually of constant-area
symmetrical cross section, with inlet and outlet angles of 45 (IJ I and P2)
being typical (Fig. 6.16). The centres of curvature of the convex and concave
surfaces of adjacent blades are then located at the saine point to f'onn parallel
passages. Another design of ilnpulse blading that has been developed is the
convergent-divergent type. ~rhis design has been found useful in that
6.6.3 Free Vortex and Constant-Nozzle-Angle Design
It was pointed out in Sec. 6.6.2 that free vortex and other design crite.ria
may be employed for the design of long blades, where constant-angle deSIgn
would lead to low efficiencies.
As shown in Eq. (5.40), for constant stagnation enthalpy across the
annulus (dho/dr = 0) and constant axial velocity (de a/dr = 0), then the whirl
component of velocity C x is inversely proportional to the radius an~ radial
equilibrium is achieved. Now if we know conditions at the mean diameter
and i1 subscript m dcnotes \at the nlean dianleter', then at any nozzle blade
radius r l ,
(free vortex condition)
But Cal is constant, and therefore at the nozzle outlet
tana l = (r 1nJr.)tana lm
244 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAr,! AND GAS TURBINES 245

and at the stage outlet when there is a whirl component
tan a 2 = (r 2 m!r 2 ) tan a 2m
Another design criterion is that of constant-nozzle-angle design where the
nozzle angle at is constant as well as dho/dr = 0 and dCaldr = O. This leads
to the axial velocity distribution given by
and since ex 1 is constant, then Ca 1 is proportional to C.d as well as dho/dr = O.
Therefore
In many cases the change in vortex design has only a marginal effect on
the performance of the blade while secondary losses may actually increase.
It is left to the experience of the designer and the results of cascade test data
to settle on a particular type of vortex flow.
6.8 MULTI-STAGE GAS TURBINES
When the multi-stage compressor was discussed in Sec. 5.8, the polytropic or
small-stage efficiency was made equal to the stage isentropic efficiency, and
an initial estimate of the expected overall pressure ratio was obtained. In
multi-stage gas turbines a similar procedure could be followed, but since
there are probably only three or four stages at the most, little extra work is
necessary to carry out a calculation stage by stage to determine the
temperatures and pressures through the turbine, the outlet conditions of the
preceding stage beconling the inlet conditions for the following stage. Once
the overall temperature and pressure drops have been established, the
isentropic efficiency can he calculated.
The performance characteristics are usually drawn in terms of the mass
flow parameter ntT~~2 11'01 and efficiency 1]( versus overall pressure ratio POI/Pon
at differing speeds given by the parameter Iv /Tb~2. All speed curves are seen
to be grouped closely together, merging into a single line at a maximum

6.7 AERODYNAMIC BLADE DESIGN 100 , . . . . - - - - - - - - - - - - - - - - - ,

Cascade data similar to that discussed in Sec. 5.1 are used for modern turbine
80
rotor blade design where the lift and drag coefficients are obtained from the
cascade data curves. Pressure losses can then be determined and an estimation
of the efficiency made. The drag coefficient must again be modified due to
the blades actually being .in annular form. Real boundaries exist at the hub
and tip while the ideal flow pattern is disturbed by preceding and following
blades. The drag coefficient is modified by tip clearance loss CDC and
secondary flow loss Cos given by 7
CDC = nC~(kclls) (6.36) o
and
Cos = C~)~cls (6.37) 1.0
Here k is the clearance between the casing and blade tip while n is taken as 0.5
for unshrouded blades and 0.25 for tip-shrouded blades. The parameter A is
~
~o 0.8
estimated from a functional relationship of the form E
So 0.6
A= f[(Blade outlet area normal to flOW)2 I( 1 + Hub radiUS)] 'Vi
o
"0

Blade inlet area normal to flow I Tip radius

e-
o 0.4
c
.9
Hence for the turbine U
~ 0.2
C DT = CD (from cascade data) + CDC + Cos (6.38)
Further information on blade design may be obtained from specialized 2 4 Figure 6.17 Characteristic curves for
texts, detailed design procedures being outside the scope of this book. an axial now gas turbine
246 HYDRAULIC AND COMPRESSIBLE fLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 247

Single stage choking turbine closely follows the ellipse law

1.0 ~- - - - - - - -
lnTb~2 /POI = k[l - (POII/POI)2] 1/2 (6.39)
Design value
Multi-stage where k is a constant

Three-stage
EXERCISES

The following constants should be used, except where otherwise stated:

0.2
1.5 2.0 2.5 3.0
Overall pressure ratio, PUll POll
Figure 6.18 Multi-stage characteristics of an axial flow gas turbine
mass flow value (Fig. 6.17). This is the choked condition and is caused by
velocities in the nozzle or at exit froIn the rotor becolning sonic.
Since the mass flow characteristics are grouped so closely together, even
towards the lower pressure ratios, a I1lean curve through all the speed curves is
often drawn and taken to be representative of that turbine. The isentropic
efficiency remains within a narrow band for a wide range of pressure ratios
(once choking has occurred) and speed, in1plying that the gas turbine is a
very flexible machine as far as Inatching to a cOIn presser is concerned, the
compressor as we have seen in Sec. 5.9 being very sensitive to oIT-design
conditions. Therefore, if the conlpressor design speed is difTerent froln the
turbine design speed, the turbine efficiency will be little changed from its
maximum value when running at the ofT-design speed. This flexibility of
operation is due to a wide range of incidence angle being possible
without incurring high rotor blade pressure loss coenlcients and can be
observed from cascade data.
Finally for this section, Fig. 6.18 shows the effect of increasing the number
of stages. The lil11iting Blass flow for the single-stage turbine occurs at a
pressure ratio just in excess of 2, and is due to choking in the stator passages.
The choking pressure ratio increases, however, with an increase in the number
of stages, but for a given pressure ratio, the s\vallowing capacity (l11ass flow)
decreases as the number of stages increases. The curve for the 111ulti-stage
Specific heat of turbine gas at constant pressure, C p = 1148 J/kg K
Ratio of specific heats, y = 1.333, gas constant R = 287 J/kg K
6.1 An impulse stage of a steam turbine is designed for a nozzle isentropic enthalpy drop of
58 kJ/kg with a mass flow rate of 0.3 kg/s. The steam leaving the nozzle enters the rotor blade
passage at an angle of 75 measured from the axial flow direction and has its relative velocity
reduced by 5 per cent in the blade passage. If the blade speed is 173 m/s and the velocity
coefficient of the nozzle is 0.98. determine:
(a) the blade inlet angle,
(b) the power developed.
(c) the tangential thrust on the blades and
(d) the stage diagram efficiency.
Assume that the blading is symmetrical.
6.2 A single-wheel impulse steam turbine has equiangular rotor blades that develop 3.75 kW and
produce a torque in the disc of 1.62 Nm at a mean radius of 132.5 mm. The rotor receives 0.014 kg/s
of steam from nozzles inclined at 70 to the axial direction and the steam discharges from the
wheel chamber in an axial direction. Find:
(a) the blade angles,
(b) the diagram efficiency.
(c) the end thrust on the shaft and
(d) the tangential thrust on the blades.
6.3 A 50 per cent reaction steam turbine running at 450 rpm develops 5 MW and has a steam
mass flow rate of 6.5 kg/kWh. At a particular stage in the expansion the absolute pressure is
85 k Pa at a steam dryness fraction of 0.94. If the exit angle of the blades is 70, measured from
the axial flow direction. and the outlet relative velocity of the steam is 1.3 times the mean blade
speed. find the blade height if the ratio of rotor hub diameter to blade height is 14.
6.4 In a zero reaction gas turbine, the blade speed at the mean diameter is 290 m/s. Gas leaves the
nozzle ring at an angle of 65 to the axial direction while the stage inlet stagnation temperature is
1100 K. The following pressures were measured at the various locations:
At nozzle entry, stagnation pressure 400 k Pa
At nozzle exit. stagnation pressure 390kPa
At nozzle exit. static pressure 200 kPa
IHR k Pa
At rotor exit, static pressure
Assuming that the magnitude and direction of velocities at entry and exit of the stage are the same,
determine the stage total-to-total erliciency.
6.5 An axial flow turbine stage has a flow coefficient of 0.65, a constant axial velocity and the gas
248 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 249

leaves the stator blades at an angle of 65 to the axial direction. Calculate:

Massnow,m 30 kg/s
(a) the blade loading coefficient,
Inlet stagnation temperature, Too 1225 K
(b) the rotor blade relative now angles,
Inlet stagnation pressure, Poo 800kPa
(c) the degree of reaction and
Temperature drop, (Too - T o2 ) 160K
(d) the total-to-total and total-to-static efficiencies.
Isentropic efficiency, Yf tll Il 0.9
6.6 A small axial now gas turbine with equal stage inlet and outlet velocities has the following Mean blade speed, U m 330 mls
design data based on the mean diameter: Rotational speed, N 15000 rpm
Outlet velocity, C 2 390 mls

Inlet stagnation temperature, Too 1100K

Inlet stagnation pressure, Poo 350 kPa Calculate:
Axial now velocity (constant through stage), C a 260 mls (a) the blade height and radius ratio of the annulus,
Mass flow, m 18 kg/s (b) the axial velocity at nozzle exit if now in the nozzle is isentropic and
Blade speed, U 350 mls (c) the inlet relative Mach number at the rotor blade root radius.
Nozzle emux angle, a l 60
Gas stage exit angle, a 2 12
SOLUTIONS
Calculate: Exercise 6.1 (a) A nozzle velocity coefficient was defined in Eq. (3.7) and in
(a) the rotor blade gas angles,
(b) the degree of reaction, blade loading coefficient and power output and
terms of enthalpy may be redefined as
(c) the total nozzle throat area required if the throat is situated at the nozzle outlet and the C = Actual nozzle velocity
nozzle loss coefficient is 0.05.
v Ideal nozzle velocity
6.7 A single-stage axial flow gas turbine with convergent nozzles has the following data:
Now across the nozzle, hoo = 11 0 l ' and since no work is done in the nozzle,
Inlet stagnation temperature, Too 1150K
C2 C2
Inlet stagnation pressure, Poo 420 kPa hO+-.2=111+~
Pressure ratio, PoolPo2 1.9 2 2
Stage stagnation temperature drop, (Too - T o2 ) 150K
Massflow,m 25 kgjs But in the steam chest at entry to the nozzle, Co is negligibly small compared
Mean blade speed, U 340 mls to C l ' and therefore
Rotational speed, N 15000 rpm
Flow coefficient, if> 0.75 C~ = 2(h 1 - 11 2 )
Angle of gas lea ving stage, a 2 12
The ideal nozzle velocity will be \vhen the flow through the nozzle is isentropic,
that is when (11 0 - hd == (h o - hls)' Thus ideal Ci = 2(h o - h ls )' Actual nozzle
If the axial velocity remains constant and the gas velocities at inlet and outlet are the same velocity
determine at the mean radius:
C 1 = (58 x 10 3 x 2)1/20.98
(a) the blade loading coefficient and degree of reaction,
(b) the gas angles, = 333.8 m/s
(c) the required nozzle throat area,
'N
(d) the annulus area at nozzle inlet and outlet and at rotor outlet if = 0.05 and From the velocity diagrams of Figs 6.3 and 6.4
(e) the height and radius ratio of the annulus at the above locations.
Ca=ClCOSCX l
6.8 Using the data of exercise 6.7, if the turbine is of free vortex design, find the gas angles a h /31' (X2
and /32 at the root and tip of the rotor blades and the relative Mach number at the inlet to the blade = 333.8cos75
tip and root.
= 86.4 m/s
6.9 Once again using the data of exercises 6.7 and 6.8, if the design is based on a constant nozzle Also
angle, calculate the angle fJ I at the blade root and tip, and compare them with those of exercise 6.8.
U + Wx1 = Cl sinal
6.10 The data below apply to a single-stage axial now gas turbine designed on free vortex = 333.8 sin 75
theory. The outlet velocity is axial and the turbine is designed with a constant annulus area
through the stage. = 322.4 m/s
250 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 251

Whence 16.3 X 10 3 x 2
W.d = 322.4 - 173 333.8 2 x 0.3
= 149.4 n1/s = 0.975
Therefore
tan P1 = W x1 /C a Exercise 6.2 (a) In the velocity diagram of Fig. 6.4, C a = C 2 since discharge
from the wheel chamber is axial. Then Wx2 = U at the mean radius.
149.4
Torque = l11r(W x2 + Wxd at the mean radius
86.4
1.62
P1 = 60
Wx2 + Wxl = 0.1325 x 0.014

(b) Relative inlet velocity = 873.3 mls

From Eq. (6.2)
W1 = Ca/COSf31

86.4
whence
cos 60 3.75 X 10 3
U=-----
= 172.8 mls 0.014 x 873.3
Therefore = 306.7 m/s
W2 = 0.95 x 172.8 Now
= 164.1rn/s C - _W_x_l _+_W_X_ 2
al - tan 70
W x2 = W 2 sin P1 since f31 = fJ2
= 164.1 sin 60 873.3
= 142.1rn/s tan 70

From Eq. (6.2) = 317.9 mls

W xl = 873.29 - 306.7
= 566.6 m/s
Power developed Therefore
W=0.3 x 173(142.1 + 172.8) tanPi = Wx1/C td
= 16.3 kW 566.6
(c) Tangential thrust = 111( ~Vx2 + W x d 317.9
= 0.3( 142.1 + 172.8) ~~ ~O.7"

= 94.47 N From Fig. 6.8

r!~ == 60.7"
From Eq. (6.5)
Work done per unitlllaSS flow (b) Also from the velocity diagrams
(d) Diagram efficiency ----- ---- -- -- -------
Work available per unit rnass flow C~ == (Wx1 + Wx2 )2 + C~i
Wlnl == 873.3 2 + 317.9 2
Ci/ 2 == 8.637 x 10 5 m 2 /s 2
252 HYDRAULIC AND COMPRESSInLE FLOW TURnOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 253

Equation (6.5) gives the diagram efficiency as Substituting

ffi -. U(Wx2 + Wx 1) 0.445 X 2n: x 450( 1411 + 11)
O lagram e IClency = C=---------
Ci/ 2 a 2 X 60
306.7 x 873.3 x 2 = 157.3I1m/s
863700 But the area of the annulus of flow = nh(D h + h), i.e.
=0.62 A=15n:h 2
(c) If there is an axial thrust C a 1 =I- C a2 Now use steam tables, at a pressure of 85 kPa. At saturation the specific
volume of the steam is 1.972 m 3 /kg. At a dryness fraction of 0.94, then
C a2 = Wx2
tan f32 Specific volume of steam = 1.972 x 0.94
= 1.85 m 3 /kg
U
(U = Wx2 since C 2 is axial) Now mass flow rate
tan fJ 2
5 x t 03 X 6.5
306.7 111=
3600
tan 60.7
= 9.03 kg/s
= 172.1 mls
C(lA
Axial thrust = m(C al - Ca2 ) (Eg. (1.23) in axial direction) 111=-- where v is specific volume
P
= 0.014(317.9 - 172.1)
Therefore
= 2.04 N 157.311 x 15nh 2
9.03 = ------
1.85
(d) Tangential thrust on blades = m(Wx2 + Wx1 ) (Eq. (1.23))
whence
= 0.014 x 873.3 h 3 = 2.254 x 10- 3
= 12.23 N h = 0.131 m
Blade height = 0.131 m
Exercise 6.3 From the velocity diagram (Fig. 6.4)
W 2 =1.3U Exercise 6.4 The condition of the gas through the stage is shown in Fig. 6.8 for
Now zero reaction where it is noted that there is no temperature (or enthalpy) drop
C a = W 2 cos /32 through the rotor.
= 1.3 V cos 70 For zero reaction
= 0.445U mls from Eg. (6.27)
Now
U=nND Afso in the rotor
at the mean diameter from Eg. (5.6)
60 11 01 reI = h02rel

2nN(D h + h) Therefore
60 x 2 W 1= W 2 and fJl = f32
where Dh is the rotor diameter at the hub and h IS the blade height. The total-to-total efficiency is given by Eg. (6.12) for a normal stage, which
254 HYDRAULIC AND COMPRESSIBLE FLOW TURllOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 255

may be rearranged in tenns of tClnpcraturc difTcrcl1ces as stage

c~ == C~ + C;2
To - T 2
1J t (t-l) = --- == 263.25 2 + 15.5 2
(To - T 2 ) + (T2 - T2ss )
= 69540 m 2 /s 2
and so the problem resolves itself into an evaluation of the above temperature Thus
differences. C~
~ =~
Through the nozzle row Too = T Ol and at a point -T 2 00
--
2C
- - T2
p

PI )(Y-I)/Y
T I == T OI ( - =1100-( 69540 )-931 since T 1 =T2
POI 2 x 1148
200) (I .333 - 1)/1 .333 == 138.7K
== 1100 ( -
390 Using isentropic relationships
== 931 K T 2ss = (J2)(Y-l)/Y
Then
Too Poo
Ci == 2CI'(Tol - T 1)
188)0.25
=2 x 1148(1100-931) T 2ss = 1100 ( 400
== 388024
C 1 == 622.9 In/s == 910.8 K
Now Therefore
T2 - T 2ss == 931 - 910.8
Cal == C I COS (Xl
= 20.2 K
== 622.9 cos 65
== 263.25 In/s Substituting for these tcmpcrature differences gives
and 138.7
C xl == C I sin (Xl YJl(t-t) == 138.7 + 20.2
= 622.9 sin 65 == 0.873
== 564.5 mls
Therefore Exercise 6.5 (a) From Eg. (6.26), the blade loading coefficient is
W xl == C xl - U
t/J == qJ(tan f3l + tan f32)
= 564.5 - 290
== (tan (Xl + tan (X2)
= 274.5 n11s
Now
but in this case (1.2 == 0 since at outlet the flow is axial. Therefore
C x2 ==Wx2 -U t/J == 0.65 tan 65
and since we have zero reaction thcn ~d = ~t'x2' So == 1.394
C x2 = 274.5 - 290 (b) From the velocity triangles
= - 15.5 In/s
tan f32 == uIC a (if C 2 is axial)
The negative sign indicates that thcre is a whirl vclocity in the opposite
direction to blade motion. 1
Since the stage is normal, C 2 = CO, and assunling constant C a through the 0.65
 256 HYDRAULIC AND COMPRESSIBLE FLOW TURROMACHINES AXIAL FLOW STEAM AND GAS TURBINES 257

Blade outlet angle fJ2 = 57 and

IjJ
tan PI = - tan f3 2 and from Eg. (6.6)
11 0 - 11 2 = V(C x1 + C x2 ) (since Co = C 2)
1.394
= - - - tan 57 = VC x1 (no whirl at exit)
0.65
=VC a tana l
Blade inlet angle [31 = 31.2
Therefore upon substituting
(c) From Eq. (6.30) _( 0.0867 sec 2 57 + 0.0654 sec 2 65(T2 /T 1 ) ) - 1
1'Jt(t-t) - 1+ 2 tan 65 .65 %
R 0.5 + 0.5(tan fJ2 - tan ad
=

= 0.5 + 0.5 x 0.65(tan 57 - tan 65) Putting T 2 /T 1 = 1,

Degree of reaction = 0.303 1'J1(t-t) = 0.909
= 90.9%
(d) Equation (6.18) gives the total-to-total efficiency in terms of the nozzle
and rotor loss coefficien ts:
Equation (6.19) gives the total-to-static efficiency with Co = Ca

11t(t 1)
= (1 + (R W~ + (NCf(T~~2:!)) - 1
2(h -h )
_ ( ~R W~ + (NCi(T 2 /Td + C5) - 1
o 2 'It(t.,,)- 1+ 2(h -h )
o 2
Using the Soderberg correlation of Sec. 6.3, for the stator (nozzle) row
= 0: 0 + a 1
= (I + f~sec2fi2 + (Nsec 2 1+ 1)-1 0:
I:N 2 tan 0: 1/4>
= 0 + 65
0.OR67 sec 2 57 + 0.0654 sec 265 + 1) - 1
= 65 = ( 1+-------------
2 ta n 65 .65 %

For the rotor row

R = fJ 1 + f32 = 0.799
= 31.2 + 57 = 79.9%
= 88.2
Therefore from Eq. (6.20), rotor loss coefficient Exercise 6.6 (a) From the velocity triangles (Fig. 6.4)

88.2)2 Cx1=CatanO: t
(R = 0.04 + 0.06 ( 100 = 260 tan 60
= 0.0867 = 450.3 m/s
and nozzle loss coefficient C x2 = C a tan a 2
= 260 tan 12
65)2
(N = 0.04 + 0.06 ( 100 = 55.26 m/s
Hence
= 0.0654 W x1 = C xt - V
Also = 450.3 - 350
= 100.3 m/s
258 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 259

Therefore Power output

W = I1,U(C.d + C x2 )
= 18 x 350(450.3 + 55.26)
=3185kW

(c) From velocity triangles

C 1 = Ca sec(X1
Also = 260 sec 60
Wx2 = U + C x2 = 520 mls

= 350 + 55.26 To find the area of the nozzle throat we must find the density Pl' In
= 405.3 mls Eq. (6.16) the nozzle enthalpy loss coefficient is
Therefore ( _ C p (T 1 - TIs)
R (Wc::X2
N - tci
P2 = tan -1 )
or
0.05 x 0.5 x 520 2
T -T - - - - - -
Is- 1148
=tan -1 (405.3)
1
--
260
= 57.3 = 5.89 K
Now in the nozzle To 1 = Too. That is, if the flow in the nozzle is adiabatic,
(b) From Eg. (6.29), reaction then the stagnation temperature is constant. Now

T 1 = T Ol -
ci
R= "2 (tan P2 - tan (31) 2Cp
2
260 = 1100 _ 520
= 2 x 350 (tan 57.3 - tan 21.1) 2 x 1148

= 0.435 =982 K
Therefore
TIs = 982 - 5.89
Blade loading coefficient is given by Eg. (6.26) = 976.1 K
Thus
P T )Yi(Y- 1)
~= ~
PI ( T 1s
[Note: Another definition for blade leading coefficient that is used is
l/J = 2
Cp(Too - T o2 )/tu , which has a nUlnerical value twice that of =(~)4
Eq. (6.26).J 976.1
Thus = 1.61
260 350
ljJ = 350 (tan 21.1 + tan 57.3)
P1 = 1.61
= 1.44 = 217 kPa
260 HYDRAULIC AND COMPRESSIBLE FLOW TURHOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 261

Then Therefore
PI
PI = RT 1
I tan 57.1 = 2 5 (1.49 + 2R)
x 0.7
3
217 X 10
287 x 982 whence reaction
3
= 0.77 kgjm R=0.414
Mass flow rate
(b) Solving the above simultaneous equations for tan fJ 1 gives

A _ 18
1
1 - 0.77 x 520 tan PI = 24J (1jJ - 2R)
Nozzle outlet area = 0.045 m 2
1
= 2 x 0.75 (1.49 - 2 x 0.414)
Exercise 6.7 (a) Substituting the data into Eq. (6.25) for the blade loading
coefficient /31 = 23.8

Then
1
1148 x 150 tana 1 =tanfJ1+-

340 2
1
= 1.49 = tan 23.8 + -
0.75
From the turbine velocity diagrams
a 1 = 60.6
U je a = tan /32 - tan C(2
or
1 (c) The nozzle throat area requires the density at the nozzle throat and the
tan fJ 2 = + tan a 2 absolute velocity at that location. Since we have a simple convergent nozzle,
the nozzle throat is at plane 1 in our notation.
1 From the velocity diagram
= - + tan 12
0.75
Cal = U
fJ2=57.1 = 340 x 0.75
From Eqs (6.26) and (6.29), = 255 m/s

t/J = (tan fJ I + tan fJ2) and

and
C 1-~
-
cos a 1
from which 255
1 cos 60.6
tan fJ2 = -(t/J + 2R)
2 =519.5m/s
262 HYDRAULIC AND COMPRESSIULE FLO\V TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 263

Now 25
2
C 0.9 x 519.5
T01 - T1=_1
2C p = 0.0534m 2
2
519.5
2 X 1148 (d) Annulus area at plane 1 is

= 117.6K
From Eq. (6.16)
25
T - T - NCi
I Is- 2C 0.9 x 255
p

0.05 X 519.5 2 Annulus area A 1 = 0.109 m 2

2 x 1148
= 5.87 K A similar procedure is followed for locations 0 and 2. Now
and 255
C o =C 2 =-12
cos
But
= 260.7 m/s
and
therefore
C5 260.7 2
T 1 = 1150 - 117.6
2C p =2 x 1148
= 1032.4 K
and hence = 29.6 K
'f1s = 1032.4 - 5.H7
= 1026.5 K C5
To = Too-U-
p
From isentropic relations bctwccn two points
= 1150 - 29.6
Poo = (Too)Y/(Y-l) = 1120K
Pi TIs Therefore
P T )Y/( Y - 1)

= 420( 1026.5)4 ~= ~
Po ( To
PI 1150
= 266.6 kPa 1120)4
Thus density Po = 420 ( 1150
PI = 377.9 kPa
Pi = RT
1 Hence
272.8 x 10 3 Po
Po= RT
287 x 1032.4 o
= 0.9 kg/m 3 378.4 x 10 3
287 x 1120
Nozzle throat area = _'_11_ = 1.17 kg/m 3
PIC!
264 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 265

Annulus area 25
0.704 x 255
=0.139m 2
25
1.177x255 (e) If h is denoted as the blade height and r is the hub radius while R is the
2 tip radius, then if the mean radius condition is denoted by subscript m
= 0.0838 m
nlVD m
U m =6()
At the stage outlet
T 02 = Too - (Too - T o2 ) 340 x 60
D =----
= 1150 - 150 m 1t X 15000
= 1000K = 0.433 m
and
Now annulus area
C~ 60U m h
T 2 = T o2 - - A=1tD mh =N--
2C p
= 1000 - 29.6 Then
= 970.4 K A~'"
h=--
Then 60U m
~= (!2)y/(y-t) = 0.735 A
P02 T 02
and so
But
11 0 = 0.061 m; h 1 = 0.080 m; 11 2 = 0.102 m
P02 = Pon ( -P02)
POO Also
420 R rm + (11/2)
1.9 r rm - (11/2)
= 221.1 kPa and so
Then
970.4)4 (R/r)o = 1.328; (R/r}t = 1.45; (R/r}2 = 1.61
P2 = 221.1 ( 1000

= 196.1kPa Exercise 6.8 At the root

Hence
196.1 X 10 3 r = rm - (11/2)
P2 = 287 x 970.4 At the tip
R = rm + (11/2)
= 0.704kg/m 3 Therefore
Therefore annulus area

=0.81
 266 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 267

( rrn~) = 1.185
and
1
m
tan fJ 1 = ( -r ) tan ct. 1 rn - ( - r) Urn
-C
( ~)
rrn 2
= 0.764
At the tip
r 1 rrn 1 a

( ~)
rrn 2
= 1.236
tanf32R= - ( rR) -Urn
rn C
+ (rrn)
2
-
R 2
tanct. 2rn
a
For free vortex design Sec. 6.6.3 shows that at the root
1.236 tan 12
=--+--
0.75 1.236
f32R = 61.2
= 1.23 tan 60.6
tan 60.6 1.181
and tan f3 1 R= 1. 181 0.75
tana 2r = C; )2 tana 2m f31R = - 4.0
= 1.31 tan 12
At the root
(X2r = 15.5
tan 12 0.764
At the tip tan f32r = 0.764 + 0.75
tan a 1 R = 0.847 tan 60.6
fJ2r = 52.3
a 1R = 56.4
and and
tan a 2R = 0.809 tan 12 tan fJ = tan 60.6 _ 0.81
1r 0.81 0.75
ct. 2R = 9.8
= 48
To find the relative gas angles, [rain the velocity triangles, at any radius,
U Summarizing the results:
tan{J 2 =tanr:l. 2 +-
Ca (Xl Pi (X2 P2
and
U At tip 56.4 -4.0 9.8 0
61.2
tan fJ 1 = tan ct. 1 - - At root 65.5 48.0 15.5 52.3
Ca At mean radi us 60.6 23.8 12.0 57.1
But
Urn U

Blade angles at other points can be calculated to give the blade geometry
Therefore substituting for U and tan ct., over the whole span.
From exercise 6.7
rrn)
tan f3 2 = ( - ( r)
tancx 2m + -:- -U m
r 2 m 2 C el
'
268 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMAC'"NES AXIAL FLOW STEAM AND GAS TURBINES 269

At the root Then

WI, == 255 sec {i 1, dCal x1 dC
--==--cota 1
== 255 sec 48 dr dr
==381m/s Equation (5.37) therefore becomes
Also
C 1, == Ca sec ai' C;l x1 dC 2 dC xl
0==-+C xl --+C x1 cot a 1 - -
== 255 sec 65.5 r dr dr
==615m/s 2 dC x1 C xl
==(I+cot ( 1 ) - - + -
Thus dr r
T 1,== T OI --
Ci, But
2C p 1 +cot 2 a l ==sin 2 cx 1
2 dCxl . 2 dr
== 1150 _ 615 --==-slna l -
2 x 1148 Cxl r
Integrating
==985K
Thus Mach number
or
(see Sec. 6.6.3)
Now if a l is constant and tan a l == Cxl/C a then C xl is proportional to Ca;
381
therefore
(1.33 x 287 x 985)1/2
Mach number at root == 0.664
At the mean radius Cam == 255 mls and rIm == 0.2165 ffi, a l == 60.6. Therefore
At the tip const == 255 X 0.2165sin260.6
W 1R ==255.7m/s
== 79.83 mls
C 1R == 460.8 mls At the root
T 1R == 1057.4K 79.83
(Cad, = (rIm _ h\/2)o.759
At tip, Mach number
79.83
(0.2165 - 0.08/2) 759
It will be noted that the higher Mach number occurs at the blade root.
== 297.77 mls
Thus
Exercise 6.9 The radial equilibrium equation is given by Eq. (5.37)
dh o C;
dC x dCll
-==--+C - + C - ==Cal(r m)
dr r x dr a dr Um r 1
In plane 1, dho/dr == 0 as before and also if a 1 is constant then 297.77 1
==--x-
340 0.81
== 1.08 ~ 1
270 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 271

1 From isentropic relationships

tan fJ 1, == tan (Xl' - -
, ~
0255
)'1/('1- 1 )
1
P02 = POO ( ---;;;-
100
== tan 60.6 - - (aI' == aIm == const)
1.08
= 800(1047.2)4
fJ lr == 40 1225

At the root PI' == 40 Now

= 427.3 kPa

At the tip C~
T 2 = T 02 --
C _ 79.83 2C p
( a .) R - ( J 12) o. 7 59 2
rIm + 11
= (1225 - 160) _ 390
79.83 2 x 1148
(0.2165 + 0.08/2)0.759 = 998.7 K
Then
== 224.2 mls T2 )'1/('1- 1)
So P2=P02 ( -T
<P = 224.2 (' m) 02

R 340 R 1
= 427.3 (?_?8.7)4
1065
224.2( 1 )
== 340 ~18 = 330 kPa
and
== 0.559
P2
{) -
2 - R'[2

330 X 10 3
1 287 x 998.7
== tan 60.6 - - -
0.559 = 1.153 kg/m 3
fJ 1 R == - 0.81 From continuity

At the ~l == -0.81

Exercise 6.10 (a) From Eq. (6.9) 30

1.153 x 390
Too - T 02
Too - T 02ss = - - - -
'1 l { l - l ) = 0.0667 m 2
At mean radius
160 Um
0.9 'm=-
OJ

= 177.77 K
330 x 60
Therefore
15000 x 2n
T02ss = 1225 - 177.77
= 1047.2 K =0.21 m
272 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES AXIAL FLOW STEAM AND GAS TURBINES 273

Blade height Then

C -~
1-
cosa l

0.0667 198
2n x 0.21 cos 70.4
= 0.0506m = 590.4 m/s
Radius ratio Now
C2
R rm + hl2 T 1 = Too - 2~
p
r rm - hl2
590.42
0.21 + 0.0253 =12 25 - - - -
2 x 1148
0.21 - 0.0253
= 1.27 = 1073 K

Using isentropic relations

(b) An iterative procedure must be followed to determine Cal' This is set out
below. The continuity equation must be satisfied; therefore at the nozzle outlet
_ (T-T
PI-POO
1 )Y/(Y-1)
since Poo = Po I
00
station 1,
A 1 = const and C a 1 = const = 800( 1073)4
1225
As a first guess let
Poo = 470.9 kPa
PI = RT
oo
Thus
800 x 10 3
287 x 1225
= 2.27 kg/m 3 470.9 X 10 3
Then 287 x 1073
30
Cal = -2.-27-x-0.-0-66-7 = 1.529 kg/m J

= 198 m/s This value is now used in a new iteration until Cal and PI do not change. A
Now remembering that C x2 = 0, table in set up below.

Iteration 2 4 6 7

PI 2.27 1.53 1.44 1.42 1.42

1148 x 160 = 330 x 198 tan 0: I Cal 200.2 294 312.3 316.7
(Xl 70.2 62.1 60.67 60.34
from which
Cl 590.4 62RJ 637.6 640
1148 x t 60 Tl 1073 1053 1047.8 1046.4
tano: = - - - -
1 330 x 198 PI 470.9 436.6 428.2 426
PI 1.53 1.44 1.42 1.42
274 HYDRAULIC AND COMPRESSIBLE FLOW TURilOMAClIlNES

At nozzle outlet axial velocity

CHAPTER
C (1 1 = 316.7 111/s
SEVEN
(c) To find (M lrel)" where subscript r refers to the root radius, RADIAL FLOW GAS TURBINES
U, I'

Um rill

U = 330 x 0.21 - 0.0253

, 0.21
= 290.2 Ill/S
At the root

1148 x 160
=----290.2
290.2
= 342.7 mls
7.1 INTRODUCTION
Then
The inward flow radial gas turbine is used for applications where the flow rate
is very low, for example turbochargers for commercial (diesel) engines and fire
= 316.7 2 + 342.7 2 pumps. They are very compact, the maximum diameter being about 0.2 m.
WI = 466.6 In/s Speeds are high, ranging from 40 000 to 180 0000 rpm. They are usually of the
Acoustic velocity 90 type, the blades being perpendicular to the tangent at the rotor outer inlet
Q
1 = (yRT 1)l/2 periphery, and the gas after entering in the radial direction exits axially at
= (1.333 x 287 x 1046.4)1/2 outlet.
The turbine and its essential parts are shown in Fig. 7.1, where its
= 632.7 similarity to the centrifugal compressor is noted, the difTerence being that the
gas flow is in the opposite direction. Figure 7.1 shows that gas enters the scroll
casing, the cross-sectional area of the scroll decreasing as the gas passes
466.6
through it. This keeps the velocity at entry to the nozzle vanes constant as the
632.7 gas is gradually drawn ofT on its circumferential path. The nozzle vanes are
Relative Mach nUI11ber at root = 0.74 converging to increase the kinetic energy of the gas and they set the gas angle
for entry into the rotor. This angle is usually about 70 (measured froln the
radial direction) but the vanes can be pivoted to allow for adjustment of the
flow angle as the load changes. In some designs, there may be no vanes at all,
but a passage similar to that of the vaneless difTuser of Fig. 2.17 is fitted
(Fig. 7.2). A vaneless space exists between the outlet tip of the vanes and the
rotor, this space being utilized by the gas for further flow adjustment and
aiding in the reduction of vibratory disturbances within the turbine.
The rotor, which is usually manufactured of cast nickel alloy, has blades
that are curved to change the flow from the radial to the axial direction. The
275
 RADIAL FLOW GAS TURBINES 277

shrouding for the hladcs is formed by the casing, and a diffuser can be fitted at
the outlet, to reduce further the high kinetic energy at that point and thereby to
increase the enthalpy drop across the rotor.

7.2 VELOCITY DIAGRAMS AND THERMODYNAMICS

OF FLOW
Diffuser

Flow into
The velocity triangles for this turbine are drawn in Fig. 7.3. The usual
inlet nozzles numbering system of 1 to 2 across the rotor will be used and therefore 0
indicates the point of entry to the nozzle vanes and 3 the diffuser outlet section.
The thermodynamic path followed by the gas is shown on the Mollier chart of
. . Scroll
Fig. 7.4. In the nozzle no work is done, therefore hoo = ho 1 although the total
casing pressure drops from Poo to Po 1 because of irreversibilities. Thus
(7.1)
The work done per unit mass flow in the rotor is given by Euler's turbine
Figure 7.1 Elements of a 90 inward now radial gas turbine with inlet nozzle ring equation (Eq. (1.24))
(7.2)
If the whirl velocity is zero at exit then
W/m=V t C x1

Gas inlet to volute

and for radial relative velocity at entry
W/m = vi (7.3)
In more general terms, substituting for W/m
h01 - h02 = V 1 C Xl - V 2 C x2
But it was shown in Sec 4.1.2 that the quantity I for a centrifugal compressor is

u ~,.-+--+-+--

Rotor inlet . P(.w,

C'~
U'J.
Section at radius r'J.av

r2hllh

Figure 7.2 A 90 inward flow radial gas turbine without nozzle ring
Figure 7.3 Velocity triangles for the 90 inward now radial gas turbine
276
278 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES RADIAL FLOW GAS TURBINES 279

I Table 7.1 Spouting velocities

11--
II
00
Poo
01 PUI Po
Total-to-total Total-to-stalic

I 0 With diffuser c; = (h oo - h03ss )/2 c; = (h oo - h3u )/2

Without diffuser c; = (h oo - h02ss )/2 c; = (h oo - II h~)/2

if a diffuser is fitted. This energy change may be related to a kinetic energy,

where the associated velocity term is known as the spouting velocity Cs ' Thus
Is four spouting velocities may be defined, as in Table 7.1, with and without a
diffuser and for total-to-total or total-to-static conditions.
The appropriate definition would be used depending upon the efficiency
02s being determined. It will be noted that for isentropic flow throughout
02ss
03s 3s! Wlnl == vi == C;/2
or
V tiCs == 0.707 (7.7)
In practice V l/C s lies in the range 0.68 to 0.7.

Figure 7.4 Mollier chart for expansion in a 90 inward now radial gas lurbine
7.4 TURBINE EFI7ICIENCY
given by
The method of deterrnining the efficiency of an inward flow radial turbine is to
I == hOrd - V 2 /2 where I is a constant
determine the total-to-static efficiency in terms of loss coefficients for the
Therefore nozzle and rotor, and then obtain a relationship for the total-to-total efficiency
hI - h 2 == [( V ~ - V ~) - ( vvi - vV ~ )J/2 in terms of known turbine dilnensions and the previously determined total-to-
static efficiency. However, before doing that, a number of relationships that
If C x2 == 0, then (W~ - V~) == C~ and
will be needed in the derivation of tlt(t-s) are discussed. From Fig. 7.3
hI - h2 == (Vi - wi + C~)/2 (7.4)
WI==Vlcota l
In the diffuser h02 == 11 03 , Thus
C 2 == U 2 cot fJ 2
f
h 3 - h2 == (C~ - C~)j2 (7.5)
If the losses in the diffuser are neglected, then T03ss == T02ss and the total-to-
I and substituting for WI and C 2 into Eq. (7.4) yields

total isentropic efficiency is given by .l-

"j
ht - h 2 == vi [I - co t 2 a I + (V 2/ V d 2 CO t 2/3 2 J/2
I and putting V 2/ V 1 == r 2/ r 1
'1t(t-l) == (Too - T 02 )/(Too - T02s J (7.6)
h t -1l 2 Vi[1 -cot 2 a t + (r 2 /rd 2 cot 2 /3 2 J/2 (7.8)
I
==
efficiencies being in the region of 80-90 per cent.
i
Now (hI - h 2) == Cp(T I - T2) == yR(T t - T 2)/(), - I) and Eq. (7.8) becomes
1
7.3 SPOUTING VELOCITY f T 2 /T 1 == 1- U~(}'- 1)[I-cot 2 a t + (r2/rl)2cot2/J2J/2yRTt
1
and putting yRT1 ==ai
If the gas passes through the turbine isentropically, then the isentropic total
enthalpy drop is given by (/zoo - h02s J ifno difTuser is fitted and by (h oo - h03ss ) T 2 /T 1 = 1 - (V 1/a t )2(y - 1)[1 - cot 2a t + (r 2 /rd 2 cot 2 P2J/2 (7.9)
280 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES RADIAL FLOW GAS TURBINES 281

The total-to-static efficiency (without the diffuser) is given by However, substituting typical values of the variables into Eq. (7.9) yields
(T 2 /T 1) approximately equal to 1 and so it is often ignored and the
1]t(t-s) = (h oo - h02 )/(h oo - h 2ss ) approximate equation for total-to-static efficiency is given by
= (h oo - h02 )/[ (h oo - h2 J + (h 2s - h2ss ) ] 1]t(t-s) = {I + [(Ncosec 2 (Xl + (r 2av /r 1)2((R cosec 2 fJ2av + cot 2 fJ2aJJ/2} -1 (7.15)
But The total-to-total efficiency may be found from the following equation
h oo = vi + h 02 (since V~ = W 1m)
(1/'1t(t-o) = (1/1]t(t-s) - [(r 2av lr 1 )2 cot 2 fJ2avJ/2 (7.16)
= vi + h2 + C~/2 Loss coefficients usually lie in the following ranges for 90 inward flow
Therefore turbines 8
0.063 ~ (N ~ 0.235
1]t(t-s) = ui/[Vi + C~/2 + (h 2 - h2s ) + (h 2s - h 2ss )] (7.10)
0.384 ~ (R ~ 0.777
Defining a nozzle and rotor loss coefficient in terms of the enthalpy loss in
each divided by the kinetic energy at outlet from each,
7.5 DIMENSIONLESS SPECIFIC SPEED
(N=(h l -h ls )/(0.5Ci) (7.11 )
But from inspection of the Mollier chart The inward now radial gas turbine can be shown to have its maximum
efficiency in a very narrow range of dimensionless specific speeds. However,
hI - hIs ~ Tl(Sl - SIs) whereas the volume now rate through hydraulic machines remains constant,
and that through the radial now gas turbine changes significantly, and this change
must be taken into account. One suggested volume flow rate to use is that at
the outlet Q2' Therefore referring to Eq. (1.16) and writing (gH) in terms of the
from which enthalpy,
hi - his = (T 1 /T2){h 2s - h2ss )
N 5 = N Q~.5 1(11 00 - 11 025S )0.7 5 (7.17)
and substituting into Eq. (7.11) and rearranging gives
Now N = U l/ rel) 1 and "00 - h02ss = C;/2 and, upon substituting, Eq. (7.17)
h 2s - h2ss = (T2 /T 1 )(Nci/2 (7.12) becomes
Similarly N!j = (V 1/rrDIN)O.5(U 1/7tDl)[Q~5/(C;/2)0.75J

or = 0.302(Q2/NDi)o.5(V 1/Cs)1.5 (7.18)

h2-h2s='RW~/2 (7.13) Equation (7.7) showed that for an ideal turbine the ratio U I/C s was equal
Now substituting Eqs (7.12) and (7.13) into Eq. (7.10) to 0.707. Therefore substituting for this into Eq. (7.18) gives

tlt(t-!J) = vi/ {Vi + [C~ + W~(R + Ci(N(T /T )J/2} 2 I

N s = 0.18(Q2/N Di)0.5 (7.19)

= vi /{vi + [U~ cot 2 fJ2 + (V~ cosec 2 fJ2)(R Ifa uniform axial velocity at exit is assumed, then Q2 = A 2 C 2 . Also writing the
+ (Ui cosec 2
a l )(T2 /T 1 )(NJ/2} projected area of the rotor as A r = nDi/4, then

= {I + [(N(T2 /Tdcosec a l + (r 2 /r l )2((R cosec 2 fJ2

2 (7.20)
+ cot 2 fJ2)J/2} - 1 (7.14) Multiplying the numerator and denominator by Ap Eq. (7.20) may be
At the outlet section 2, two radii are possible, viz. at the shroud and at the arranged as
hub. It is usual to take the average radius Q21NDi = (A 2 /A r)(n 2 /4)(C 2/U 1)

r 2av = (r 2hub + r 2tip)/2 = (A 2/ A r)(7t 2/4)( 1.414/Cs)C 2

while the temperature ratio (T 2 /T l ) is obtained from Eq. (7.9). = 3.49(A 2/ A r)( C 2/C s)
 RADIAL FLOW GAS TURBINES 283
282 HYDRAULIC AND COMPRESSIBLE FLO\V TURBOMAClllNES

7.2 A small inward flow radial exhaust gas turbine has the following design data:
Francis turbine
100 Axial Rotor inlet tip diameter 90 mm
flow Rotor outlet tip diameter 62 mm
80 Rot?r outlet hub diameter 25 mm
~ 90 radial flow Rat~o C 2/C s 0.447
>-. 60 gas turbine
u
c:
RatIO U I/C s (ideal) 0.707
v Blade. speed. 30000 rpm
'u
tt: 40 DenSIty at Impeller exit 1.8 kg/m 3
U.J

20 Determine:
(a) the dimensionless specific speed of the turbine
0 (b) the volume flow rate at impeller outlet and '
1.0 10.0 (c) the power developed hy the turbine.
0.01 0.1
Dimensionless specific speed, Ns (rad)
7.3 The design data of a proposed inward radial now exhaust gas turbine are as follows:

Figure 7.5 Variation of efficiency with dimensionless specific speed Stagnation pressure at inlet to nozzles p 700 kPa
Stagnation
S . temperature at inlet to no~'zl~~o' T 00 1075 K
tat~c pressure at exit from nozzles, PI 510 kPa
StatIC temperature at exit from nozzles ' T 1 995 K
S tatic pressure at exit from rotor p
Thus substituting for Q2/NDi in Eg. (7.19) gives Static temperature at exit from T r~t02r 350 kPa
N s = 0.336(C 2/CJo. 5 (A 2 /A r )o.5 rev (7.21 ) Stagnation temperature at exit from 'rot~r ~ 918 K
Ratio r2.vlr I 02 920 K
= 2.11(C 2 /C s)O.5(A 2 /A r)o.5 rad (7.22) Rotational speed, N 0.5
26000 rpm
The flow into the rotor is: purcl y rd. d"la I an d at exit. the flow is axial at all radii C a l l '
In practice (a) the total-to-statlc efficiency of the turbine, . cu ate.
(b) the outer diameter of the rotor
0.04 < (C 2jC s )2 < OJ (c) the enthalpy loss coefficient for' the nozzle and rotor rows
(d) the blade outlet angle at the mean diameter {J d .,
0.1 < (A 2 IA r ) < 0.5
(e) the total-to-total efficiency of the turbine. lav an

Then 7.4 Using the data of exercise 7.3 n

Calculate: .. now 0 f exhaust gas availahle to the turbine is 2.66 kg/s.
Ie mass

0.3<N s <1.1 rad

(a) the volume flow rate at rotor exit
Thus the dimensionless specific speed range is very sll1all and the variation of (b) the hub and tip diameters of the ;otor at exit
efficiency with N s is shown in Fig. 7.5, where it is seen to match the axial flow (c) the power developed by the turbine, '
(d) the rotor exit blade angles at the hub and tip
gas turbine over the litnited range of N s (e)the nozzle exit angle and '
(f) the ratio of rotor width at inlet to its inlet tip diameter.
7;' An inward now radial turbine is fitted with a diffuser, which can be assumed to have an
EXERCISES ;e~~~;sc~ho: ~:e~e~~:~~~~t~~tu~it)e c;okeshwhen the ab~olute velocity at exit from the impeller
2 ,s ow t at the lImItIng total pressure ratIO is given by
In the following exercises, unless otherwise stated, use the following values: 1/(1- R fI
) = (C 2/C s )2{ [1 + 0.5(y - 1)MiJ/[O.5(y - 1)A1 22 J} +'lI(t-t)
Specific heat at constant pressure, C p = 1147 Jjkg K where R = (P02IPoo)l't-ll!y.
fl

Ratio of specific heats, y = 1.333, gas constant R = 287 Jjkg K

7.1 An inward now radial gas turbine operates with a total-to-total efficiency (from nozzle inlet to
diffuser outlet) of 0.9. At entry to the nozzles, the stagnation pressure and temperature of the gas SOLUTIONS
are 300 kPa and 1150 K respectivdy. At outlet from the diffuser the pressure is 100 kPa and the
velocity of now is negligible at that point. Find the impeller lip speed and the now angle at the Exercise 7.1 The overall isentropl'c eill'
lClency 0f t hc turbine from inlet to
nozzle outlet if the gas enters the impeller radially and there is no whirl at the impeller exit. The
diffuser is given by
Mach number at exit from the nozzles is 0.9.
284 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACHINES RADIAL FLOW GAS TURBINES 285

Too - T and substituting for C p we get

111(1-1) = - - - 03
-
Too - T03ss !i= 1- Vi(y-l)
Now from Eq. (7.3) Too 2yRToosin 2 al
Ur(y - 1)
=1- 22 ' 2
a 01 SIn a 1
when Cx2 = 0. Also
But
T OO ( 1 - ;:;s) = TOo[ 1 - G::r-1)!Y]
Therefore and
al V1
a0 1 M 1a Ol sin a l
100)0.2498 ] Therefore
=0.9 x 1147 x 1150 [ 1- ( 300

Impeller tip speed V 1 = 533.7 m/s

The Mach number of the absolute flow velocity at exit from the nozzle is and
given by
1= ( - -Lfl
-
a01 sinal
)2(Y-l 1)
--+-2
2 M1

U
sin 2 a l = ( _1 )2 (y - 1+ 1)
--- -2
aOI 2 A1 1
Across the nozzle the now is adiabatic; therefore
But
Too = T Ol = T 1 + -
Cr
2C p
= 1.333 x 287 x 1150
V2
=Tl + 1
2C p sin 2 a l = 439956 m 2/s 2
or Therefore
2
~ = 1- vi . 2 533.7 (0.333
sIn a = - - - - - + -2
1)
Too 2C p Too sin 2
(Xl I 439956 2 0.9
Now = 0.9071

and Nozzle angle (X 1 = 72.25

1 R
1--=-
y Cp
Exercise 7.2 (a) From Eq. (7.21) the dimensionless specific speed is
or
N =0.336 ( ~
c )1/2(A~ )1/2 rev
S Cs Ar
286 HYDRAULIC AND COMPRESSIBLE FLOW TURBOMACIIINES RADIAL FLOW GAS TURBINES 287

Now Exercise 7.3 (a) The total-to-static efficiency is given by

hoo - h02
I1l(l-s) = h - h
00 2ss
n(0.062 2 - 0.025 2)
1 - (T02 /Too )
4
1 - (T2ss/Too)
== 2.53 x 10- 3 m 2
From isentropic relationships
nD 2
A == ---~
r 4
T 2ss == (~)(y- 1 )/y
Too Poo
0.09 2 n
Therefore
4
1 - (T 02 /Too )
== 6.361 X 10- 3 m 2 I1l(l-s) = 1 _ (P2 / Poo )(Y-1)/Y
Dimensionless specific speed
1 - (920/1075)
=0336(0.447 x 2.53)1/2 1 - (350/700)0.2498
Ns 6.36
== 0.144/0.159
== 0.142 rev
= 0.906
== 0.89 rad

If this value is compared with Fig. 7.5 it is seen that a reasonable overall (b) From Eq. (7.3) the specific work done is
efficiency is achieved. W/m == vi
(b) The flow rate at outlet for the ideal turbine is given by Eq. (7.19)
since C x2 = 0 and C x 1 = V 1 for radial inlet flow. Therefore
Q2 )1/2
N s =0.18 ( - - 3 rev Cp(Too - T 02 ) == ui
ND 1
and
Q2 X 60 )1/2
0.142 == 0.18 ( -
30000 x 0.09
3
VI = 1147(1075 - 920)
whence
== 177785 m 2/s 2
VI = 421.6 m/s
(c) From Eq. (7.3) Then
W = I1lui
60U 1
Vi D 1 =--
==P2Q2 nN

= 1.8 x 0.227 x ( n~~ 1 ) 2 60 x 421.6

n x 26000
n x 30000 x 0.09)2 == 0.31 m
== 1.8 x 0.227 x ( 60
Power developed == 8.166 kW (c) Equation (7.11) gives the nozzle loss coefficient, which we can arrange
288 HYDRAULIC AND COMPRESSIIlLE FLOW TURBOMACHINES RADIAL FLOW GAS TURBINES 289

into an equation involving only pressures and temperatures: Also

h 1 - h 1s c~ = 2C p (T o2 - T 2)
(N = 0.5 q = 2 x 1147(920 - 918)
Now =4588m l /s l
Therefore
(since To 1 = Too)
w~ = 4588 + 210.8 2
Therefore = 49025 m 2 /s 2
and for the numerator

We may write
"2 - "2s = C T2- (~2JTl ]p[

-- C
[T (P2 )().- I)IY
T
]
T 1s T 1s ) l - -
= ( -Too Too P PI 1

_ ( P1 )(Y-l)/Y
= 1147(918 - 905.7)
- - x Too
Poo = 14108J/kg (or m 2j s2)
510)0.2498 Therefore
=(- x 1075
14108
700
(R = 0.5 x 49025
= 993.2 K
Substituting = 0.58
( _ 995 - 993.2 (d) The total-to-total efficiency is found from Eq. (7.16).
N - 1075 - 995
____~_ = _1__ ~ (C_~aY cot fJ2 )2
Nozzle loss coefficient = 0.0225 t'll(l--I) 'll(1-s) 2 r1 av

The rotor loss coefficient is given by Eq. (7.13) Now

cot Play = C 2/ U 2
( _h 2 -h 2s (4588)1/2
R - 0.5 W~
210.8
Now
= 0.321
w~ = C~ + U~
fJ 2ay = 72.2
and using the mean diameter for the calculation of conditions at the impeller
outlet
Therefore
U2 r 2ay - 1 = -1- - 0.5(0.5 x 0.321)
2

U.1 r1 t'll(t-I) 0.906

U 2 = 0.5 x 421.6 = 1.09
'11(1-1) = 0.917
= 210.8mjs
290 HYDRAULIC AND COMPRESSIBLE fLOW TURBOMAClIlNES RADIA L FLOW GAS TURBINES 291

Exercise 7.4 The dimensionless specific speed based on the total-to-total Therefore
enthalpy drop in the turbine is given by Eq. (7.17).
NQ1/ 2 2.66
Ns = (J100 - }102ss )3/4 Q2 = 1.328
'
But =2.0m 3 js

(h oo - h02s J = TooC p ( 1 _ T
02SS
)
This gives
Too
2722.7 X (2.0)1/2
Using isentropic relations
N s = (193.78 X 10 3 )3/4
(hoo - h02ss ) = CpTooL 1 -
I ( ::: )()'-I)/YJ
= 0.416 rad
Now Since it was not stated in the problem which was the efficiency of importance
~ = (~)),/(Y-l) the dimensionless specific speed could equally have been based on the total-to-
static enthalpy drop (h oo - h2ss ) to correspond to the total-to-static efficiency.
P02 T 02
The error is small however:
Therefore
920)4
P02 = 350 ( 9i8

= 353.06 kPa = 196035 Jjkg

and and
(11 00 - h02ss ) = 1147 x 1075[ 1 - C~~,~6 r,2498] N s =0.413rad

(b) Volume flow rate

= 193.78 kJ/kg
Rotational speed Q2 = A 2C2

2nN = n(r~ tip - r~ hub)C 2

W=--
60 = 2nr 2av hC 2
2n x 26000 where h is the height of the blade between hub and tip and '2av =
60 (r 2 hub + '2 tip)/2. Now
= 2722.7 rad/s C~ = 2C p (T02 - T2)
Q2 = Hl/P2 =2 x 1147(920-918)
and
= 4588 m 2 /s 2
P2
P2 = RT2 C2 = 67.7 m/s

350 X 10 3 and
287 x 918 r 2av = 0.5 x 0.155
3
= 1.328 kg/m = 0.0775m
292 HYDRAULIC AND COMPRESSIBLE FLOW TURROMACHINES RADIAL FLOW GAS TURBINES 293

Therefore At the hub

h= 2.0
2n x 0.0775 x 67.7 0.047)
U 2hub = 421.6 ( - -
0.155
= 0.061 m
= 127.8mjs
h
r2hUb=r2av-2 127.8
tan fJ2 = 67.7
0.061
= (0.5 x 0.155) - - - At exit hub f32 = 62
2
= 0.047 m
(e) Across the nozzle hoo = ho 1 and
Hub diameter = 0.094 m
ci = 2C p (T oo - T 1 )

=2 x 1147(1075-995)
= 183520 kJjkg
0.061
= (0.5 x 0.155) +--
2 Nozzle exit velocity
= 0.108 m C 1 = 428.4 mjs
Tip diameter = 0.216 m The inlet velocity triangle shows
(c) From Eq. (7.3)
sina 1 =U 1/C l
W=mui 421.6
= 2.66 x 421.6 2 428.4
Power developed = 472.8 kW Nozzle angle a 1 = 79.7
(d) From the outlet velocity triangle, since C 2 is axial and uniform over
the exit annulus, (f) From the continuity equation at impeller inlet,

nl=P1 A l C rl
At the tip
r2tip
= P1 nD l W 1 b 1
U 2tip= U 1 -
r1 where b 1 is the depth of the impeller blades at inlet

= 421.6(108) b1 m
0.155 D1 nplDi WI
= 293.8 mls But
293.8
tan f3 2 = - -
67.7
= 421.6 cot 79.7
At exit tip f32 = 77
= 76.62 mls
294 HYDRAULIC AND COMPRESSIULE FLO\V TURUOMACIlINES RADIAL FLOW GAS TURBINES 295

and Now
PI
Pi = RT (iii)
I

510 and
287 x 995 R
= 1.786 kg/m 3 Cp

Therefore substituting Substituting for C p in Eg. (iii)

T (y - 1) C~
bi 2.66 -oz= 1 + - - - -
Tz 2 yRT2
D I n x 1. 79 X 0.31 2 x 76.62
But
= 0.0642
Blade depth = 0.02 n1
Therefore

Exercise 7.5 Since the now is isentropic in the diffuser T 02 = T03 (iv)
T02 = T 03 and P02 = P03' With the diffuser the total-to-total 'effici;~cy is sS'
Thus
Too - T 02
'1t(t-l) = T _ 'T
00 102ss

From Table 7.1 the spouting velocity is

c; (v)
U = Too - T 02ss
p
and hence
Substituting
C~ !(y - l)M~
1 2T02 C p (vi)
(i) 2C p T oz 1 + t(y - 1)lv!~
t'lt(l-l) = l-(T IT) - ~
025S 00 5
Substituting from Eg. (vi) into Eg. (ii) and writing R" = (PozlpOO)(y-l)/Y
Using isentropic relationships
T 0255 = ()(Y- l)/Y ( Cz)Z
C
(1
= 1 - R" t
-11 (t-t)
)( t(y -
1 + t(y -
I)M~
l)M~
)
P02 s

Too POO and rearranging

Substituting for the temperature ratio and multiplying both sides by c~ we get ___ 1 (C)Z(1+ 1z(Y-1)M Z) +
~ 2
1 - R" - Cs !(y - l)M~ 11t(t-t)

or

or _ _ _1 = (C 2)2( 1 + !(y - l)M~) +

1 - (P02Ipoo)(Y- 1)/'1 Cs !(y - l)M ~ 11t(t-t)

(ii) For chosen values of ~t(t-t) and lv! 2' (C 2 IC s )Z can be varied and (Poz/Poo)
determined.
 REFERENCES
INDEX

1. Stodola, A., Steam and Gas Turbines, vols I and II, McGraw-Hili, New York, 1927. (Reprinted,
Peter Smith, New York, 1945.) Aerofoil section, 57 reaction, 236
2. Wislicenus, G.F., Fluid Mechanics of Turbomachinery, McGraw-Hili, New York, 1947. Angle, 35 stage, 229
3. Stanitz, J.D., 'Some theoretical aerodynamic investigations of impellers in radial and mixed Angle of incidence, 57 stage number, 246
flow centrifugal compressors', Trans. ASM E, vol. 74, p. 4, J 952. Angular momentum: swallo\ving capacity, 246
4. StepanofT, AJ., Centrifugal and Axial Flow Pumps, Wiley, New York, J 948. equation, 20 velocity triangles, 30
5. Jacobs, E.N., Ward, E. and Pinkerton, R.M., 'The characteristics of 78 related aerofoil sections in diffuser, 51 Axial pump, 56
from tests in the variable density wind tunnel', NACA Report No. 460. Axial com pressor, t 80 characteristic, 61
6. Soderberg, C. R., Unpublished note, Gas Turbine Laboratory, Massachusetts Institute of characteristic, 10, 204 Axial hydraulic turbine, 88, 103
Technology, 1949. blade forces, 189
7. Ainley, D.G. and Mathieson, G.C.R., 'A method of performance estimation for axial flow efficiency, 185 Blade:
turbines', A.R.C., Rand M 2974, J95 J. energy transfer, J X3 back ward facing, 46
8. Benson, R.S., 'A review of methods for assessing loss coefficients in radial gas turbines', Int. J. Mollier chart, 184 cascades, 191
M echo Sci., vol. 12, 1970. multi-stage performance, 204 chord, 61
stage, t 82 design, 240, 243, 244
stage pressure ratio, t 85 difTusion, 181, 182
stall, 205 efficiency, t 99
surge line, 205 element, 59
velocity triangles, 183 forces, 189
Axial fan, 180 solidity, 59
Axial flow steam turbine, 227 stagger, 63
Axial gas turbine, 227 stall, 143
blade loading, 235 Bucket friction, 92
blade types, 240 Buseman, 40
cascades, 244
characteristic, 10 Camber, 192
choking, 246 angle, 192
energy transfer, 228 circular are, 194
impulse stage, 237 Cascade, 191
isentropic efficiency, 231 annulus drag, 199
multi-stage performance, 245 curves, 193
296
297
298 I!\DEX INDEX 299

drag, 194, 195 secondary flow loss, 199, 244 hydraulic, 42 t urhine,
88
lift, 194, 195 tip clearance loss, 244 impeller, 42 turbine efficiency, 101
lift coefficient, 197 Con1pressor: mechanical, 42 turbine losses, 101
nomenclature, 192 axial, 131, 180 nozzle, 95 Hydroelectric installation, 98
pressure across, 196 centrifugal, 131 pipe transmission, 95
pressure loss, 193 definition, 4 polytropic, 202 Impeller, 33
profile drag, 195 eye, 132 radial gas turbine, 279 constant, 136, 277
secondary losses, 199 stage, 182 small stage, 202 definition, 4
traverse, 193 Continuity equation, 16 stage, 185, 199 efficiency, 42
Casing, 34 Corresponding: total-to-static, 232, 279 power loss, 40
efficiency, 42 characteristic, 65 volumetric, 42 Impulsc:
loss in punlp, 36 points, 65 Elevation of turbine, 106 blading, 241
loss in turbine, 101 Critical: Energy transfer: turbine, 88, 228
Cavitation: cavitation, 54 axial gas turbine, 228, 231 Incidcnce, 57, 193
critical, 54 NPSH, 54 radial gas turbine, 277 Inducer, 132
erosion, 53, 54 Entropy, 20 Inlet, 34
in pumps, 53 Deflection, 193, 194 Euler: Isentropic:
in turbines, 106 nonlinal, 193 head, 36 efficiency, 136, 185, 231, 280
limits, 107 Deviation, 193, 194 pump equation, 20 flow, 9
parameter, 53 Diagranl efficiency, 231, 242 turbine equation, 20 Isolated blade, 59
Centrifugal conl pressor, 131 Dialneter variation, 66 Eye of compressor, 132
characteristic, 142, 145 DifTuser, 4 Kaplan turbine, 103
choking, 144 centrifugal conlpressor, 138 Fan definition, 4 Kinematic similarity, 8
diffuser, 138 Mach nUlnber, 141 First law of thermodynamics, 16
diffuser Mach number, 141 radial gas turbine, 277, 278 Flow coefficient, 6, 187, 235 Leakagc, 40, 41, 101
elements of, 132 ring, 31 Flow in discharge casing, 48 Lift:
energy transfer, 134 vaned, 51 Flow rate variation, 45 acrofoil, 61
impeller, 135 vaneless, 49 Francis turbine, 96 coefficient, 61, 189, 197,244
inlet casing, 135 DifTusion in blades, 1g1, 182 Free vortex design, 202, 243 in cascade, 194, 197
inlet guide vanes, 141 Dilnensional analysis, 4 Friction losses, 45 Logarithmic spiral, 51
inlet velocity, 139 cOlnpressible !low 111achines, 8 Fundamental dimensions, 5 Loss coefficient:
mass flow, 140, 141 hydraulic lnachines, 4 nozzle, 233
pressure ratio, 136, 138, 142 Din1ensionless: Gas turbine: radial gas turbine, 280
pre-whirl 141 groups, 5, 6, 10 nozzle, 228 rotor, 233
total-to-total efficiency, 136 specific speed, 12, 14, 281 reaction, 228, 236
velocity triangles, 133 Disc friction, 40 stator, 228 Mach number:
Centrifugal fan, 131 Divergence angle, 52 Geometric similarity, 8 in difTuser, 141
Centrifugal pun1p, 31, 33 Draft tube, 98 Guide vanes, 56 Machines, 1
characteristics, 40 head loss, 100 adjustable, 98 Manometric head, 33
Characteristic: Drag: fixed, 98 Matching:
axial compressor, 204 in cascade, 194 head loss, 97 pump, 63
axial pump, 61 total, 199 system, 63
centrifugal compressor, 145 Drag coefficient, 189 Head coefficient, 6 Mechanical:
centrifugal punlp, 43, 47 Duty point, 64 Head: efficiency, 42
impulse turbine, 95, 96 Dynalnic silnilarity, 8 loss in draft tube, 100 loss, 42
Kaplan turbine 105 net turbine, 100 Mixed flow, 31
reaction turbine, 102 Eddies, 46 water, 97 Model:
Choking, 144, 246 EfTect of outlet angle, 46 Hydraulic: efficiency, 11
Circulation, 59 Efficiency: diameter, 234 testing, 7
Chord, 59, 193 axial COIn pressor, 185 efficiency, 7, 42 Mollier chart:
Coefficient: blade, 199 efficiency of Pelton wheel, 93 axial COlnpressor, 184
annulus drag, 199 diagrall1, 231, 242 machine characteristics, 7 axial turbine, 232
 INDEX 301
300 INDEX

centrifugal compressor, 134 Pumps: Soderberg correlation, 235 casing loss, 101
radial turbine, 278 in parallel, 66 Solidity, 191 cavitation parameter, 106
Momentum equation, 19 in series, 66 blade, 59 definition, 2
Multi-stage performance: Spear valve, 94 elevation, 106
axial compressor, 202 Radial: Specific speed, 14 Francis, 96
axial turbine, 245 blades, 39, 46, 47, 132, 277 conversion, 14 hydraulic, 88
equilibrium, 201 dimensionless, 12 efficiency, 108
Net head, 100 thrust, 49 suction, 55 radial, 96
Net positive suction head, 54, 106 Radial gas turbine, 275 System resistance, 64 i111 pulse, 88
Normal stage, 182, 230, 233, 236 diffuser flow, 279 Spouting velocity, 278 Kaplan, 103
Nozzle loss, 232 efficiency, 278, 279, 281 Stage: leakage loss, 101
coefficient, 233 elements of, 275 axial compressor, 182 overall efficiency, 101
Pelton wheel, 95 energy transfer, 277 axial gas turbine, 229 power loss, 101
radial gas turbine, 280 loss coefficients, 279 efficiency, 185, 199, 231 reaction, 88, 96, 103
NPSH, 54,106 Mollier chart, 278 loading, 188, 235 reversible-pump, 89
Number of diffuser vanes, 52 nozzle loss, 280 normal, 182, 230 total head, 100
rotor loss, 280 Stagger angle, 193 Twisted blades, 56, 105
Off-design, 40 Stall: Types, 1
Optimum blade loading, 190 velocity diagrams, 277
Radial hydraulic turbine, 88, 96 axial con1pressor, 205
Outlet, 34, 56 in centrifugal C0I11prCSSor, 143 Unit:
Overall efficiency: Reaction:
50 per cent, 188 rotating, 143 head, 4
pump, 41 Stanitz slip factor, 132 power, 14
turbine, 101 100 per cen t, 238,
axial compressor, 187 Static gauge pressure, 32 quantity, 14
Parallel pumps, 66 axial gas turbine, 236 Static head, 33 speed, 14
Pelton wheel, 88, 89 blading, 240 Stator:
bucket friction, 92 hydraulic turbine, 88 blades, 182 Vaned diffuser, 51
characteristic, 95 negative, 238 los~es, 232 Vaneless diffuser, 49
efficiencies, 94 turbine characteristic, 102 Steam turbine: Variables:
installation, 89 zero, 237 impulse, 228 compressible flow machines, 9
hydraulic efficiency, 93 Recirculation, 40, 41,101 reaction, 228 hydraulic machines,S
load changes, 93 Relative eddy, 38 Stodola, 38, 3 Velocity coefficient, 95
losses, 94 Relative total enthalpy, 136 Surge, 143 Velocity triangles:
nozzle loss, 94 Reversible: Swallowing capacity, 246 axial compressor, 183
spear valve, 94 process, 21 axial gas turbine, 230
velocity triangles, 90 pump-turbine, 89 Tailrace, 97, 100 centrifugal compressor, 133
Pitch, 59, 193 Reynolds number, 6 Thoma number, 106 hydraulic pump, 36, 57
Polytropic efficiency, 202 Rotor, 4 Three-dimensional flow, 200 hydraulic turbine, 99, 105
Positive displacement, 1, 65 Rotor loss coefficient, 233, 280 Total drag, 199 Pelton wheel, 92
Power coefficient, 6 Runner, 4 Total head, 31, 100 radial gas turbine, 277
Power input factor, 135, 137 turbine, 98, 104 Total-to-sta tic efliciency: Volute, 4,
Pre-whirl, 132 axial gas turhine, 232 casing, 4
Propeller, 56 Second law, 20 radial gas turbine, 2HO
Prototype, 4, 8 Secondary losses, 198, 244 Turbine Whirl components, 37
efficiency, 11 Self-limiting characteristic, 48 axial hydraulic, 103 Work done factor, 186
Pump: Series pumps, 66
axial flow, 56 Shock, 40
characteristic, 43 Shock losses, 44
definition, 2 Similarity laws, 66
diameter variation, 66 Slip factor, 37
elevation, 55 Buseman, 40
losses, 40 centrifugal compressor, 137
positive displacement, 1, 65 Stanitz, 40, 132
speed variation, 65 Stodola, 39