85032

EJECTORS AND JET PUMPS - DESIGN AND PERFORMANCE FOR

INCOMPRESSIBLE LIQUID FLOW

1. NOTATION AND UNITS

Units

SI British

A cross-sectional area m2 ft2

a, b, c coefficients in quadratic expression for M 0

C fluid density ratio, secondary fluid density/primary fluid

density, C = ρ s ⁄ ρp

C pr diffuser static-pressure recovery coefficient

(see Equation (B2.5))

d diameter of primary nozzle exit m ft

D diameter of mixing chamber m ft

f friction factor, defined from head loss

= 4f ( L ⁄ D ) ( V 2 ⁄ 2g )

g acceleration due to gravity m/s2 ft/s2

H total head m ft

K component loss coefficient

L length of mixing chamber m ft

l characteristic length m ft

M volume flow ratio, secondary flow rate/primary flow rate,

M = Qs ⁄ Q p

M0 value of M when N = 0

N pressure ratio, secondary flow pressure rise/primary

flow pressure drop, N = ( P 5 – P 2 ) ⁄ ( P 1 – P 5 )

N0 value of N when M = 0

N′ alternative pressure ratio, secondary flow pressure

rise/primary-secondary pressure difference,
N′ = ( P 5 – P 2 ) ⁄ ( P 1 – P 2 )

P total pressure Pa or N/m2 lbf/ft2

For footnotes see end of Notation Section.

Issued December 1985 - 101 pages

This page Amendment B
1
 85032
p static pressure Pa or N/m2 lbf/ft2

Q volume flow rate m3/s ft3/s

R area ratio, A n ⁄ A m

Re Reynolds number, Re = Vl ⁄ ν

s distance from primary nozzle exit to mixing chamber m ft

entrance

V velocity m/s ft/s

ε effective roughness height of mixing chamber surface m ft

η jet pump efficiency, η = M × N

φ diffuser half angle deg deg

ν kinematic viscosity m2/s ft2/s

ρ density kg/m3 *slug/ft3

2
σ cavitation number, σ = ( P 2 – p v ) ⁄ ( ½ρs V 3c )
For footnotes see end of Notation Section.

*
1 = slug = 32.174 lb (pound mass).

Subscripts

1, 2, 3, 4, 5 planes within the ejector, as defined in Sketch 1.1

d diffuser

c cavitation-limited flow value

m mixing chamber

n primary nozzle exit

p primary or driving flow

ref reference or datum value

s secondary or induced flow

v vapour

2
 85032

Sketch 1.1 Jet pump configuration and reference planes

3
 85032
2. INTRODUCTION

This Item is concerned with the design and performance of ejectors and jet pumps. Such devices are
characterised by the use of the kinetic energy of one fluid stream (the primary flow) to drive a second fluid
stream (the secondary flow) by direct mixing. The fluids may be gases or liquids and the secondary stream
is not necessarily the same fluid as the primary. For some applications the secondary stream, and the primary
stream also, may contain solid particles. The design parameters, requirements and methods vary
considerably, depending on whether the working fluids are gases, liquids, solids-in-fluids or combinations
of these, and each type is therefore considered in a separate Data Item.

The terms “ejector” and “jet pump” are alternative names for the same device and the term “injector” is
also used.

2.1 Purpose and Scope of this Item

This Data Item provides information for design and performance evaluation of ejectors and jet pumps in
which both working fluids are liquids. A technique for the optimum design of a jet pump or ejector is
presented; the method is based on the equations of continuity and momentum conservation and includes
empirical coefficients that allow for losses in the different components. The performance of an ejector or
jet pump of given dimensions may be determined from an appropriate performance characteristic curve,
which fully describes the operation of the device. A comprehensive series of such curves is presented, based
on a large number of empirical and analytical data.

2.2 Layout of this Item

Section 3 of this Item discusses briefly some applications of ejectors and jet pumps.

Section 4 describes the principles of ejector operation and defines the different components. These may
vary considerably with application and Section 4 considers some of the possible configurations.

Section 5 describes a design method which will determine the on-design operating conditions and optimum
dimensions for an ejector or jet pump when the primary and secondary working fluids are both liquids. The
method is suitable for primary and secondary fluids of equal or differing densities. Mechanical design
aspects are considered in Section 5.5.

Section 6 describes a method by which the performance of an existing ejector or jet pump may be evaluated.
Performance charts are presented for the case when the working fluids are of equal density together with
corrections allowing for differing densities.

Section 7 describes, briefly, more complicated ejector designs, such as multi-nozzle or annular ejectors or
multi-stage configurations, and considers their advantages in certain situations.

Section 8 discusses the problem of cavitation and includes a definition for cavitation index. The index
characterises the conditions under which cavitation occurs and may be used to predict the onset of this
phenomenon. Methods for improving cavitation performance are also described.

Section 9 presents worked examples showing the application of both the design and the performance
prediction methods.

Section 10 lists all the sources of information used in compiling this Data Item and includes further sources
of reference which may be helpful to the engineer.

4
 85032
Appendix A contains a glossary of terms used in describing jet pump components, design parameters and
performance.

Appendix B presents a detailed theoretical analysis of the flow through an ejector, based on one-dimensional
flow representations of mass and momentum conservation, and includes the development of equations used
to describe efficiency and the onset of cavitation.
Appendix C presents flow charts outlining the design and performance prediction procedures.

5
 85032
3. APPLICATIONS OF LIQUID EJECTORS AND JET PUMPS

Ejectors and jet pumps are used in a wide range of engineering fields, as is illustrated by the following
examples. Their particular advantage lies in their flexibility and simplicity of operation and in the low level
of maintenance required. The ejector is especially useful where a source of high energy fluid is available
as the by-product of some other process, although this is by no means an essential requirement.

Aeration

A jet pump may be connected across a centrifugal pump to aerate the fluid. Examples of this type of
application can be found in an abattoir, or in a coal mine where the jet pump acts on unoxygenated mine
water to oxidise dissolved iron compounds and secure their precipitation before discharge.

Booster Pumping

A liquid jet pump may be used to boost the pressure and supply in a pipe network system. Water supplied
to a reservoir under pressure can be utilised to act as the driving flow of a jet pump which is installed at
the outlet from the reservoir. Similarly, the flow rate of a high pressure water supply can be boosted with
ground water by incorporating a jet pump in the pipeline.

Cleaning and Reducing Turbidity

Tests have been performed on the viability of a jet pump used to pump water to or from an estuary. This is
intended to reduce or prevent growth of algae and improve the water quality. If successful, the pump could
obviate the need to construct a canal opening to the sea.

Deep Well Pumping

When a liquid such as oil or water has to be raised from deep wells, the suction lift may exceed the absolute
vapour pressure of the liquid. This makes it necessary to lower the pumping unit into the well shaft if a
centrifugal pump alone is used. An alternative procedure is to use a jet pump in combination with a
centrifugal pump; the centrifugal pump is mounted at ground level and provides the driving head and
primary flow for the jet pump placed in the well below. Mechanical advantages are that there are no moving
parts in the well, very little maintenance of the jet pump is necessary, and the centrifugal pump and motor
may be placed at any convenient point. Hydraulic advantages are that the jet pump-centrifugal pump
combination gives a steep head-mass flow operating characteristic with a higher operating head than that
of a centrifugal pump alone, and a non-overloading brake horsepower curve (Derivation 2).

A similar type of application is the extraction of slops, grit, oil, water, etc. from cargo holds in ships, or
similar situations where space can be highly restricted. Vertical suction lifts of 30 m have been achieved
with deck-mounted units, where there was no room for the pump to be lowered into the hold, by allowing
air to enter the suction orifice with the liquid.

Domestic Water Supply

The simplicity of construction and maintenance of the jet pump has led to its wide-spread use, in conjunction
with a centrifugal pump, in shallow-well pumping for farm and domestic water supplies (particularly in
the USA and Canada). The centrifugal pump and the jet pump are located together at ground level. The
system exhibits an almost constant power characteristic, which makes it extremely versatile.

6
 85032
Dredging

In this application an inclined or vertical dredging ladder comprising at least driving and discharge pipes
is supported from a floating pontoon or hull. The jet pump is at the lower end and may be immersed in the
deposit being dredged. It is often convenient to use water jets for disintegration or fluidisation of the bottom
material, the supply being derived from a separate pump or, if the optimum pressure is compatible, from
the jet pump primary water supply.

The use of a jet pump avoids having solid materials pass through the driving centrifugal pump so the
sediment cannot clog or erode the pump impeller and, although the solids may pass through the jet pump,
the wear rate is low. The motive pump set is able to run at constant speed and constant load. (See Reference
33.)

The system is designed so that the jet pump characteristic and the pipeline head characteristic intersect so
as to give a stable flow; in fact, jet pump dredges can be so stable in operation that they do not require
supervision.

Hydro-electric Schemes

Ejectors may be used to suppress tail-water back-pressure when flood conditions prevail or to increase the
effective head available for power production when relatively high water levels are required downstream,
for navigational or irrigation purposes. The ejector may be incorporated with reaction or impulse turbines,
as well as with tidal power units.

Lubrication Systems

The simplicity and light weight of the jet pump make it attractive for use in lubrication systems, particularly
for aircraft engines. A major advantage is the elimination of the mechanical drive train necessary for
conventional pumps. However, in this application cavitation limitations may cause problems at high
operating altitudes of the aircraft.

Mixing

During the momentum transfer process in the mixing chamber, almost complete mixing takes place. Thus
the jet pump has application in mixing chemical solutions, making gels and suspensions or diluting noxious
liquids.

Multi-pressure Systems

In facilities such as dry docks and laboratories where both high and low pressure systems are required, with
low and high discharges respectively, it can be cheaper to operate a single set of pumps. In this application
the jet pump may be used in combination with a centrifugal pump, as a slightly changing head can vary
demand on the jet pump unit while the centrifugal pump can operate against a fixed discharge at its best
efficiency point. A jet pump may also be used as a pressure limiting device in a multi-pressure system.

7
 85032
Nuclear Industry

Jet pumps are frequently used in the nuclear industry, both for the transport of radioactive materials, slurries,
etc. and to produce a vacuum. Jet pumps have also been used as recirculation devices to circulate coolant
in a reactor core. Their use can reduce the number of recirculation loops, saving pipe, valve, pump and
motor costs and also eliminating equipment susceptible to breakdown. A major advantage of jet pumps in
all nuclear applications is their reliability and the low level of maintenance required.

Priming Devices

Siphons can be swiftly primed by means of ejectors; there is no necessity to use foot-valves and the
continuous operation of the pump throughout the siphoning process can alleviate problems of vapour
accumulation or air leaks in sub-atmospheric conditions. Alternatively, a jet pump may be used by itself as
a priming device.

Pumping Sludges

Jet pumps are used for the removal of oil sludge on board ships and tank de-sludging in the petrochemical
industries and for the pumping of mud on oil rigs, where mud is also used as the primary fluid. Ejectors
may also be used for the pumping of primary-digested sewage to tanks and ponds, having the advantages
that the rate of disposal of the fibrous material is high and that the sewage tends to lose its non-Newtonian
properties after passing through the jet pump.

Solids Transport

Jet pumps are widely used as boosters in pipelines through which solids as widely varied as coal, sand, ash,
radioactive materials and foodstuffs are transported. The solids may pass through the jet pump or may be
intercepted first; this latter configuration is rather more complex but avoids the friction losses and abrasion
problems incurred when the solids do pass through. The turbulence to which the solids are subjected in the
mixing chamber may be used to advantage in the scrubbing of minerals or sand. Conversely, in the transport
of foodstuffs it is important that little damage is incurred. An annular jet pump is suitable for handling
delicate products, since the solids mainly pass up the centre of the jet pump and are kept away from the
mixing chamber walls.

Throttling Devices

A jet pump may be used as a throttling device between two systems or for surge damping.

Thrust Augmenters

The thrust of a jet may be augmented by the use of a mixing chamber with a driving jet; this is comparable
to other methods of jet propulsion, though less common with liquid than with gas ejectors. Such a device
has been used for controlling ships and can avoid the need for tugs in dynamic positioning.

Trench Cutting

Jet pumps are used in offshore applications for cutting trenches required for burying pipelines and cables.

8
 85032
Well Point Systems

Well points, for de-watering building sites, quarries, etc., may be supplied with high pressure water via jet
pump attachments and can discharge ground water against substantial pressures. The advantages of this
system over the usual layout are that the whole system is under pressure thus making leaks clearly visible,
the lift of a single stage is much greater than the normal 6 m and if air is drawn into the well point, say by
an unforeseen drop in the water table, it is educted without depriming.

9
 85032
4. THE BASIC LIQUID EJECTOR OR JET PUMP

4.1 Principle of Operation

An ejector or jet pump is a device in which the kinetic energy of one fluid stream (the primary or driving
fluid) is used to drive another fluid stream (the secondary or induced fluid). The primary fluid is usually
accelerated through a nozzle into a region containing the secondary fluid. The fall in pressure at the nozzle
exit induces the flow of the secondary fluid in the direction of the primary stream. The secondary fluid is
entrained due to viscous friction at the primary jet periphery and carried forward into a mixing chamber.
The jet expands and there is a transfer of energy and momentum from the primary to the secondary fluid
through a process of turbulent mixing. During this process there is some recovery of static pressure, but
mixing two streams of unequal velocity leads to a loss of kinetic energy. At the end of the mixing chamber,
momentum exchange should be complete and the mixed fluid is discharged, often through a diffuser which
is used for further static pressure recovery.

The basic theory of the flow through an ejector or jet pump, derived by considering continuity and
momentum conservation, is described in Appendix B.

A typical axial static pressure distribution, related to the stations within an ejector, is illustrated in Sketch 4.1.

10
 85032

Sketch 4.1 Axial pressure distribution within an ejector

4.2 Component Parts

Sketch 1.1 shows a typical configuration featuring all the usual components of an ejector. However, the
only component common to jet pumps or ejectors in all applications, and hence the one that effectively
characterises the device, is the mixing chamber. This is where the primary-secondary turbulent mixing,
which is the essential principle of the operation of an ejector, takes place. In liquid jet pumps the mixing
chamber and driving nozzle are nearly always of circular cross-section and must be concentric otherwise
efficiency is lost, and in an abrasive application the life of the mixing chamber is greatly reduced.

All the other components are in some way dependent upon the particular application and the following
points may have a bearing upon the final ejector design.

(a) Primary nozzle

In almost all cases a primary nozzle will be present since it is more economical to transport the
primary fluid to the ejector at high pressure, rather than at high velocity with its inherent
flow/pressure losses.

11
 85032
(b) Position of primary nozzle

A single central nozzle may be in the plane of entry to the mixing chamber, although in liquid
ejectors it is more often moved slightly upstream as discussed in Section 5.5.

(c) Annular primary nozzle

As an alternative, the primary or driving nozzle may be annular, either producing a discrete annular
jet separate from the wall of the mixing chamber, or located around the periphery of the mixing
chamber to give a wall jet. The surface area of an annular jet is greater than that from a central
nozzle so that mixing is more rapid, and an annular nozzle jet pump requires a shorter mixing length
than a central nozzle type. The wall-jet type will, however, have higher friction losses. Annular jet
pumps are discussed in greater detail in Section 7.1.

(d) Multi-nozzle ejectors

Multi-nozzle units may be used in conjunction with a single mixing chamber and diffuser. These
have the advantage of reducing considerably the mixing chamber length required and improving
cavitation performance. More details may be found in Section 7.2.

(e) Multi-stage ejectors

Multi-stage units are possible, in which the ejectors are arranged in series, as described more fully
in Section 7.3. This configuration extends the range of flow ratios over which the fluid may be
pumped. In addition, single-stage jet pumps are limited by the occurrence of cavitation caused by
high velocity secondary inlet flows, and this problem is reduced by splitting the secondary flow
between two or more pumps.

(f) Multi-barrel ejectors

Multi-barrel units are also used. These are similar to multi-nozzle ejectors with a common
secondary flow reservoir, but each nozzle has its own mixing chamber and diffuser, i.e. the ejectors
are arranged in parallel. The exit diffusers discharge into a common delivery tube.

(g) Diffuser

There need be no diffuser if the velocity of the fluid discharged from the mixing chamber is already
suitable for existing pipework and the mixing chamber diameter is appropriate, such as in very low
head jet pumps or where a large and dense solid product is being handled.

(h) Input of secondary flow

The secondary fluid may be input at various angles to the primary stream, depending on the
application, but this introduces extra losses, particularly if the secondary fluid is input at right angles
to the primary jet, although they may be minimal if the inlet velocity is low or diffused.

There need be no secondary or induced flow under steady state conditions since the device may be
used to maintain a fixed degree of vacuum or discharge pressure.

12
 85032
5. DESIGN OF A LIQUID EJECTOR OR JET PUMP

5.1 Introduction

The aim of the design procedure is to produce an ejector or jet pump that will operate at maximum efficiency
within the specified constraints of primary and secondary flow rates and pressures.

The design parameters of an ejector or jet pump are usually expressed in terms of dimensionless ratios that
describe the geometry and performance of the device. The five principal design ratios used are defined as
follows.

(1) Pressure ratio, secondary flow pressure rise/primary flow pressure drop,

N = ( P 5 – P2 ) ⁄ ( P 1 – P5 ) . (5.1)

The pressure values refer to points just upstream or downstream of the pump components, so pipe
or fitting losses upstream of the point of measurement are not included.

When the primary and secondary fluids are of equal density, N can be expressed as a head ratio,

N = ( H 5 – H2 ) ⁄ ( H 1 – H 5 ) . (5.2)

(2) Volume flow ratio, secondary flow to primary flow,

M = Qs ⁄ Q p . (5.3)

When the primary and secondary fluids are of equal density, the volume flow ratio is the same as
the mass flow ratio, and the same symbol is often used.

(3) Jet pump efficiency (always less than 1),

Q s ( P 5 – P2 )
η = --------------------------------- = M × N . (5.4)
Qp( P1 – P5)

(4) Ratio of nozzle exit cross-sectional area to mixing chamber cross-sectional area,

R = An ⁄ A m . (5.5)

The area ratio is the most important non-dimensional design parameter for jet pumps and serves
as a type criterion in the same way as specific speed for a centrifugal pump. A small area ratio jet
pump ( R ≈ 0.1 ) is characteristically low-pressure, high flow ratio. Conversely, a large area ratio jet
pump ( R ≈ 0.6 ) is high-pressure, low flow ratio in nature. The practical range of R values can be
considered to be 0.02 to 0.9.

(5) Length to diameter ratio of mixing chamber, L ⁄ D .

13
 85032
It is, however, possible to define the efficiency of a jet pump in several different ways, leading to a different
definition of pressure ratio. This is discussed in greater detail in Appendix B, where each expression is
derived from a fundamental definition of efficiency. The particular expression used often depends on the
jet pump application.

The definition of jet pump efficiency most commonly quoted, and hence the one used in this Data Item, is
that corresponding to Equation (5.4). The pressure ratio is then correspondingly defined according to
Equation (5.1).

The efficiency defined by Equation (5.4) is reduced by friction, shock losses, throttling at the secondary
inlet, loss of kinetic energy due to mixing, and because velocities are non-uniform across the duct. Maximum
practical efficiencies are of the order of 35 to 40%.

The best efficiency is usually independent of ejector size, except where the ejector is very small or the
working fluids have a very high viscosity. Under those conditions the primary and secondary flows have
low Reynolds number, i.e. viscous forces predominate over inertial forces and the turbulent mixing process
in the mixing chamber is impaired; incomplete mixing leads to a reduced performance. The primary nozzle
loss coefficient, K p , becomes greater for primary jet Reynolds numbers less than about 10 5 and operation
in the laminar regime should be avoided (see Derivation 4 for more details).

5.2 The Characteristic Curve

The characteristic curve of a jet pump or ejector, which describes its performance over the complete
operating range, is expressed non-dimensionally in terms of the flow and pressure ratios. The M versus N
curve of a jet pump corresponds to the Q versus H curve of a centrifugal pump.

For a jet pump with a specific area ratio, R , the M versus N characteristic is a plot with slight curvature,
whose general slope is dictated by R .

A typical characteristic curve is illustrated in Sketch 5.1.

Sketch 5.1 Typical jet pump characteristic and efficiency curves

As the discharge pressure, P 5 , of a jet pump is increased, the flow will decrease until the secondary flow
ceases and only the primary flow passes into the mixing chamber and on through the ejector; this is known
as the “first shut-off”, where M = 0 and N = N 0 , and is the point at which the secondary flow reverses
direction.

14
 85032
Continuing restriction of the discharge eventually reduces the outlet flow to zero, the “second shut-off”
(see Sketch 5.1). At this point the secondary flow is reversed with the primary flow leaving via the secondary
inlet and so the flow ratio, M , is negative. The point of second shut-off is required so that the maximum
pressure possible in the jet pump in the event of line blockage or valve closure may be assessed and
components rated accordingly.

Conversely, as the discharge pressure, P 5 , and hence the pressure ratio, N , is reduced, the flow ratio is
increased until the characteristic crosses the axis ( M = M 0 ) . It is common for cavitation to commence
before this point is reached, although it is possible for a jet pump to operate under conditions where the
discharge pressure is lower than the secondary inlet pressure so that the pressure ratio is negative.

The normal operating region for a jet pump lies near the mid-point of its characteristic curve.

As stated in the previous section, when the working fluids have a high viscosity the jet pump performance
is impaired. Figure 1, which illustrates some experimental results from Derivation 2, shows the influence
of reducing Re p (increasing primary flow viscosity) on jet pump performance.

A typical jet pump efficiency is also illustrated in Sketch 5.1. This is easily plotted using Equation (5.4)
once the M versus N curve is established. The efficiency curve exhibits a well-defined maximum, with
the position of the peak dependent on area ratio, and is approximately symmetric about the value of M
corresponding to the maximum. Values of η must always be less than 1.

5.3 Design Method

The characteristic performance of a jet pump can be expressed in terms of an equation relating the pressure
ratio, N , to the area ratio, R , and the flow ratio, M . The equation, derived as Equation (B5.1) in Appendix
B, is,

2C M2R2 C M2R2
2 R + ----------------------- – R 2 ( 1 + C M ) ( 1 + M ) ( 1 + K m + K d ) – ---------------------- ( 1 + K s )
1–R ( 1 – R )2
- ,(5.6)
N = ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
2C M 2 R 2
( 1 + K p ) – 2R – ----------------------- + R ( 1 + C M ) ( 1 + M ) ( 1 + K m + K d )
2
1–R

where C is the density ratio. The loss coefficients K p , K s , K m and K d account for losses in the primary
nozzle, secondary flow inlet, mixing chamber and diffuser respectively, and are defined by Equations (B2.1)
to (B2.4) in Appendix B.

Equation (5.6) is based on a one-dimensional analysis of flow through an ideal jet pump, but, by use of
empirical values for the loss coefficients, it can be applied to practical jet pumps in which the flow, especially
in the mixing process, is essentially three-dimensional.

Equation (5.6) provides the basis for the design method but is relatively complex and, if speed of design is
important, the M versus N curve represented can be approximated by a straight line. This approximation
is fair for values of R less than about 0.4 and greater than about 0.75 but for other R values results are
considerably poorer. Formulae for the extreme coordinates ( 0, N 0 ) and ( M 0, 0 ) are given in Section 5.4
(Equations (5.15) and (5.16)). If the M versus N curve may be approximated by a straight line, the
efficiency curve ( η = M × N ) may be approximated by a parabola.

15
 85032
For the straight-line characteristic, the following relationships hold.

N = N 0 – ( N 0 ⁄ M 0 )M , (5.7)
M corresponding to best efficiency = M 0 ⁄ 2 ; (5.8)
the efficiency, η , for a particular flow ratio, M , is given by
η = M N 0 – ( N 0 ⁄ M 0 )M 2 (5.9)
and the maximum efficiency for the jet pump is given by
ma x η = ( M 0 N 0 ) ⁄ 4 . (5.10)

Given a particular value of M , jet pumps having a range of R values will function, but only one will yield
an optimum performance. To follow through a design procedure it is necessary in some cases to estimate
an 'optimum ' R value. Figures 3a and 3b present curves of optimum area ratios, as functions of M , N and
the density ratio, C .

A theoretical relationship between M and optimum R can be found if the basic continuity equation is
simplified by assuming identical fluids. This produces an expression for the mixing chamber Reynolds
number,

Re m = ( 1 + M )Re p R . (5.11)

It has been proposed (Derivation 2) that primary/secondary mixing, and hence jet pump performance, is
most efficient when the primary nozzle and mixing chamber Reynolds numbers are equal, for which
Equation (5.11) rearranges to

M = (1 ⁄ R) – 1 . (5.12)

This curve is also presented in Figure 3a and can be seen to agree very well with the empirical data.

Essentially, if only a minimum of constraint is imposed, the design of a jet pump or ejector is an iterative
process of estimation and re-evaluation until a characteristic curve can be drawn to satisfy the pressure or
flow requirements and yield an operating point at the position of maximum efficiency. It is often convenient
in the initial stages to use the straight line approximate characteristic rather than the curve given by Equation
(5.6). The following Section details a Design Procedure in a step-by-step manner.

5.4 Design Procedure (see Flow Chart in Appendix C Section C1)

Listed in Table 11.1 are the parameters completely specifying a jet pump. These parameters are not listed
in any particular order as jet pumps are subject to a vast range of widely varying design criteria. This Table
may be copied for use as a convenient summary of the known parameters at any stage in a calculation
procedure.

Stage 1

Determine the design criteria imposed by the conditions in which the jet pump is required to operate.

The criteria may be operational, specifying the pressures at inlet or outlet to the jet pump or the fluid mass
or volume flow rates, or dimensional, restricting the size of certain of the jet pump components.

16
 85032
From the design criteria it may be possible to calculate a value for flow ratio, M , and/or pressure ratio, N .
For most jet pump applications the design requirements fall into one of three categories:

(a) the jet pump must operate at optimum efficiency;

(b) the main requirement is a particular flow rate or flow ratio;

(c) the main requirement is a particular pressure or pressure ratio.

It is unusual for both flow ratio, M , and pressure ratio, N , to be pre-defined simultaneously. However, if
this is the case, it should be noted that it is quite possible to specify mutually exclusive values of M and
N . If the design procedure was followed with such erroneous values, an impossible value for efficiency
would result. It is advisable to check for this possibility by calculating η at the outset from Equation (5.4);
a value for efficiency less than about 20% indicates that the design would be poor, while a value greater
than about 50% is impossible to achieve and shows that there is inconsistency in the design requirements.
If M and N are mutually exclusive, one or other must be changed. In such circumstances it is recommended
that the design procedure be carried out several times, changing M and N in turn until the best design
compromise is reached.

Calculate the density ratio, C = ρ s ⁄ ρ p , for the primary and secondary fluids with which the jet pump
will be operated.

Determine the vapour pressure, p v , of the mixed fluid.

Select a suitable value for the cavitation index, σ . A value of 1.35 is recommended as providing a
conservative estimate of the onset of cavitation.

Stage 2

Obtain values for the loss coefficients.

Values of the loss coefficients K p , K s and K m , based on measurements on a wide range of research and
full-scale jet pumps, are given in Figures 2a to 2c.

Choose a value for K p , the primary nozzle loss coefficient, using Figure 2a and Table 11.2, taking the
primary nozzle Reynolds number as 10 5 for a first approximation. Figure 2a has been plotted on a large
scale for clarity, but for most nozzles there is only a small variation in K p .

Choose a value for K s , the secondary inlet loss coefficient. A good first approximation is a value of 0.1;
this can be refined later when the ejector geometry has been determined so that the secondary inlet Reynolds
number can be calculated.

Using a roughness value for the mixing chamber material* from the Table on Figure 2c and an estimate of
the mixing chamber diameter, calculate a value for the relative roughness, ε ⁄ D , of the mixing chamber.
Using Figure 2c, estimate a value for K m , the mixing chamber loss coefficient. The flow through the mixing
chamber is fully turbulent and Figure 2c has been derived from the Moody diagram (Reference 37). The
K m value may be read directly for mixing chamber length to diameter ratios of 7.0 and 5.0 (recommended
for central jet and annular jet pumps respectively), but otherwise the engineer must interpolate.

*
Reference 40 gives roughness values for a wide range of surfaces.

17
 85032
Obtain a value of K d , the diffuser loss coefficient, either directly from Reference 42 or from the pressure
recovery coefficient, C pr , (from References 41 and 42, for example), using the relationship

K d = 1 – ( A 4 ⁄ A 5 ) 2 – C pr . (5.13)

In the absence of other constraints, a diffuser area ratio of about 5 and an included angle of 6 to 8 degrees
is recommended.

The coefficients K m and K d have the most significant effect on the characteristic curve of the jet pump,
while performance is much less affected by variations in the values of K p and K s .

Stage 3

Obtain a value for the area ratio, R .

R may be defined directly by nozzle and mixing chamber diameters specified as part of the design criteria.
If this is not so, it is important to obtain the particular value of R which will give optimum performance.

Values of M and N with corresponding optimum values of R are presented in Figure 3a for a central nozzle
jet pump at a density ratio of 1.0. The M versus R curve is derived mainly from experimental results, some
of which are also plotted, while the N versus R curve is derived from computed results, using Equation
(5.6) and loss coefficients typical of a well-designed ejector. Similar curves, derived from computed results,
are presented in Figure 3b for several different density ratios. It should be noted that the optimum area ratio
will be different for geometries other than a central nozzle type jet pump but no data are available for those
cases.

Proceed as follows, according to the design criteria.

(a) If flow ratio, M , is defined as one of the design requirements, find the corresponding
optimum area ratio, R , and pressure ratio, N , using Figures 3a and 3b.

(b) If pressure ratio, N , is defined as one of the design requirements, find the corresponding
optimum area ratio and flow ratio using Figures 3a and 3b.

(c) If neither the flow ratio nor the pressure ratio is restricted, take R ≈ 0.27 , as this has been
shown to give close to maximum efficiency. The corresponding values for M and N may
be found from Figures 3a and 3b. These values may be refined by iteration later.

(d) If pressure ratio, N , is fixed as well as flow ratio, M , R is found by substituting their values
in the following quartic equation in R (derived by rearranging Equation (5.6));

R4a1( N + 1) + 2R3[ ( N + 1 )( C M2 – a1 – 1) ]

– R 2 [ ( N + 1 ) ( 2CM 2 – a 1 – 4 ) – CM 2 ( 1 + K s ) – N ( 1 + K p ) ]
– 2 R [ ( N + 1 ) + N ( 1 + Kp ) ] + N ( 1 + Kp ) = 0 , (5.14)

where

a1 = ( 1 + C M ) ( 1 + M ) ( 1 + Km + Kd ) , (5.14a)

18
 85032
and solving for R . Although this is a quartic in R , there is only one value of R for which the efficiency
( η = N × M ) is less than 1 and that satisfies the design requirements. Use Figures 4a to 4t to see whether
the values chosen for N and M are irreconcilable, and must be revised, and to obtain a first approximation
of R .

Stage 4

If the jet pump is to function at a single operating point, then calculate the pressure ratio, N , from Equation
(5.6).

If the jet pump is to operate over a range of conditions, the N values may be calculated from Equation (5.6)
for the range of flow ratios expected. The jet pump should be designed for the middle of the range. Rather
more simply, if slightly less accurately, the characteristic curve may be plotted using a straight-line
approximation; this method is best used for R values less than 0.4 and greater than 0.75. The curve is
approximated as follows.

Calculate M 0 and N 0 , the values at which the characteristic curve crosses the axes.

When M = 0 ,

2R – R 2 ( 1 + K m + K d )
N 0 = ------------------------------------------------------------------------------------------- . (5.15)
( 1 + K p ) – 2 R + R2 ( 1 + Km + K d )

When N = 0 , a quadratic equation for M 0 results which may be solved in the form

– b – b 2 – 4a c
M 0 = ------------------------------------------- , (5.16)
2a

2C R 2 R 2
where a = -------------- – C R 2 ( 1 + K m + K d ) – C  -------------- ( 1 + K s ) , (5.17)
1–R  1 – R
b = –R 2 ( 1 + C ) ( 1 + Km + Kd ) (5.18)

and c = 2R – R 2 ( 1 + K m + K d ) . (5.19)

It should be noted that only the negative root is required and that coefficient a is always negative so that
M 0 is positive.

Plot the two end-points ( 0, N 0 ) and ( M 0, 0 ) and draw the straight line characteristic.

Stage 5

Calculate the efficiency of the jet pump from Equation (5.4).

A jet pump, designed for a particular operating point, should be able to achieve an efficiency in the range
30% to 40%. It is preferable to know at this stage whether or not the current design parameters will produce
a device which will operate at 'peak' efficiency. It is possible, using Equation (5.4), to pick points from the
M versus N curve and plot the η versus M curve. It is then simple to see where the operating flow ratio
lies in relation to the maximum efficiency point.

19
 85032
If the area ratio, R , has been selected using Figures 3a and 3b, a peak efficiency operating point should
result. If this is not the case, and the operating point lies to the left of maximum efficiency (i.e. the optimum
flow ratio is greater than the required flow ratio), R should be increased for subsequent design iterations,
whereas, if the operating point lies to the right of maximum efficiency (i.e. the optimum flow ratio is too
low), R should be reduced. The logic behind these observations can be seen by examining the trends
illustrated by Figures 3a and 3b.

Stage 6

Complete the specification of design parameters as listed in Table 11.1.

If either the primary or the secondary flow rate is unknown, calculate the missing quantity from Equation
(5.3), using the known or estimated values of the flow ratio, M , and the other flow rate.

Find any unknown pressures from the design requirements and the design value of N , by rearranging
Equation (5.1) to make the unknown pressure the subject.

Calculate the primary nozzle exit area, A n from

MR 2 ½
( 1 + K p ) – C ( 1 + K s )  --------------
 1 – R
A n = Q p ----------------------------------------------------------------------------- (5.20)
( P 1 – P 2 ) ⁄ ( ½ρ p )

and hence calculate the nozzle exit diameter, d . Equation (5.20) is derived in Appendix B.

Calculate the mixing chamber cross-sectional area from the area ratio (Equation (5.5)), knowing values for
A n and R , and hence calculate the mixing chamber diameter, D .

Calculate the mixing chamber length, L , taking the optimum length to diameter ratio equal to 7.0 for a
central nozzle jet pump or equal to 5.0 for an annular jet pump, as recommended in Section 5.5.

Stage 7

Although the design is not yet complete, it is important to assess at the earliest stage possible whether or
not cavitation will present a problem. Choose and calculate a suitable criterion to describe the limit set by
cavitation.

This criterion may be either the maximum possible value of flow ratio, defined by

1 – R  P2 – p v  ½
M c = --------------  ----------------------------- , (5.21)
R  ½ρ V 2 C σ
p n

where Vn = Qp ⁄ An ,

20
 85032
or the minimum secondary inlet pressure required to prevent cavitation, defined by

2
 Q p MR 
P2 = ½ ρ p C σ  ------- -------------- + pv . (5.22)
min
 A n 1 – R

The desired criterion is found by taking σ = 1.35 and substituting for the other variables.

Once derived, the cavitation limit on the flow ratio can be plotted onto the design characteristic. The required
operating point should always lie to the left of this limiting ordinate.

The criteria found from Equations (5.21) and (5.22) are worst case values as these equations have been
derived by assuming that the primary nozzle exit is in the same plane as the mixing chamber entrance. It
is possible to improve cavitation performance by retracting the nozzle from the mixing chamber entrance,
although the improvement cannot be evaluated using one-dimensional theory.

If the minimum secondary pressure or the maximum flow ratio are not compatible with design requirements,
it is not possible to produce these requirements with an optimal jet pump. A new value of area ratio, R ,
must be determined and the procedure repeated; the difference will be reflected in the lower value of
efficiency.

The problem of cavitation is discussed more fully in Section 8.

Stage 8

Re-assess the loss coefficient values, based on the jet pump geometry determined, using References 37 to
43, Figures 2a, 2b and 2c and Table 11.3. Figure 2b has been plotted on a large scale for clarity and in
practice there is only small variation in K s . If the secondary inlet Reynolds number is greater than 2 × 10 5 ,
K s ≈ 0.10 .

Repeat the procedure from Stage 2.

The design process is a system of successive refinement and the iteration should proceed until the engineer
is satisfied that the design is, to within the required tolerance, an optimum.

Stage 9

Carry out the detailed design of the primary nozzle, secondary inlet, mixing chamber and diffuser, using
the recommendations for optimum geometry described in Section 5.5.

This will define the following parameters:

secondary inlet geometry (profile specification or included angle if conical),

primary nozzle included angle,
diffuser included angle and length,
details of component junctions.

Decide on the nozzle position with respect to the mixing chamber entrance; the recommended displacement
is one nozzle exit diameter upstream.

21
 85032
5.5 Mechanical Design Considerations

Ejectors and jet pumps can be made of almost any normal constructional material; the choice is usually a
compromise between machinability, strength, resistance to wear, erosion and corrosion, and cost. It is
important that the internal surfaces are smooth to reduce friction losses and useful if the pump is amenable
to part replacement.

To give a complete and definitive description of the optimum geometry of an ejector or jet pump is extremely
difficult, as a large number of the variables concerned show interrelated effects. In addition, space
constraints often limit the size of the device. The following sections present recommendations for the
geometry of each of the ejector components, together with ways of compensating for non-ideal geometry.
The recommendations apply to liquid-liquid jet pumps.

Derivations 12, 17, 25 and 31 are useful sources of information on the optimum geometry of jet pumps.

5.5.1 Primary nozzle

There are two aspects to the design of the primary nozzle, involving specification of both the internal and
external contours and the finish.

The nozzle should have a gently concave external contour, rather than a conical profile (Derivations 15 and
31), the object being to minimise restriction of the secondary flow. Good results have also been found with
nozzles having a quarter ellipse profile. If the nozzle must be conical, the recommended included angle is
16 to 20 degrees (Derivation 17).

There should be no sharp edges in the external profile, as these would tend to cause turbulence in the
secondary flow and increase losses.

The nozzle should have as thin a tip as possible (as near to a feather edge as is practicable), so as to minimise
the size of the wake. The wake increases the turbulence in the mixing layer, leading to an earlier onset of
cavitation. However, if abrasion of the nozzle tip is likely to be severe, it is recommended that the nozzle
walls are not tapered right to the tip, but that the end portions of the internal walls are parallel.

A typical nozzle design is illustrated in Sketch 5.2.

Sketch 5.2 A typical primary nozzle design

22
 85032
5.5.2 Secondary inlet and mixing chamber entry

Good design of the secondary inlet region is important to optimise the jet pump cavitation performance. In
addition, any interference between the primary nozzle exterior and the secondary inlet will increase energy
losses resulting from friction and throttling of the secondary fluid.

When designing the secondary inlet, consideration should therefore be given to the geometry of the primary
nozzle exterior, which may be already determined, so that there is no restriction of the secondary flow. For
example, a nozzle with a convex exterior would be unsuitable in conjunction with a narrow conical inlet.
Flow through the inlet must be constantly converging or parallel up to the primary nozzle exit.

An effort should be made to obtain an even velocity distribution, with the minimum length of passageway
in which high velocities are experienced to reduce friction losses.

A bell-mouth inlet is recommended, although more difficult to machine than a conical inlet, because it
eliminates sharp corners which are sources of high losses, i.e. high K s values. Junctions should be radiused,
as sharp edges in the inlet region are also detrimental to cavitation performance; they generate local vortices,
the low-pressure core regions of which are sources of cavitation.

The quality of the approach to the mixing chamber is less important where the flow ratio is small and
irrelevant where it is zero.

5.5.3 Primary nozzle exit to mixing chamber entry spacing

The optimum spacing between the primary nozzle exit plane and the mixing chamber entry plane actually
depends quite considerably on the geometries of the primary nozzle exterior and the secondary inlet. Where
these are ideal and there is no cavitation, maximum efficiency is found with the nozzle in the plane of the
mixing chamber entrance, i.e. s = 0 (Derivations 18 and 31).

However, the liquid jet pump becomes increasingly prone to cavitation as the nozzle-mixing chamber
spacing approaches zero. This is because the zero spacing causes large pressure drops at the mixing chamber
entry (see Sketch 4.1) and consequently promotes cavitation (Derivations 12 and 25).

To provide a suitable compromise between operating efficiency and cavitation performance, therefore, the
distance from the primary nozzle tip to the mixing chamber entrance is recommended to be in the order of
one nozzle diameter, i.e. s ⁄ d = 1.0 (see Derivation 1, p.176 and Derivation 10, p.110). This also makes
allowance for a non-ideal primary nozzle-secondary inlet geometry. The recommended spacing of one
nozzle diameter is not, however, an absolute optimum for all jet pump geometrical configurations; larger
area ratio jet pumps require longer mixing lengths and may therefore need bigger spacings.

The overall length within which momentum transfer between the primary and secondary fluids takes place
includes the nozzle displacement, s , as well as the mixing chamber length, L , so it is advisable to increase
the displacement distance slightly if space is restricted and the mixing chamber is shorter than recommended
(Derivation 31). It is, however, not practical to define an equivalent mixing length when the nozzle is
retracted because of unknown effects on the turbulent mixing process of the pressure and velocity fields in
the secondary inlet region.

If fine-tuning is necessary, for example if the ejector is to be used for mixing chemicals and an exact flow
ratio is important, it is convenient if the primary nozzle position may be adjusted so that the precise optimum
position for particular conditions may be found.

23
 85032
It should be noted that if the nozzle is retracted from the mixing chamber entrance plane, the practical
performance will differ from the theoretical prediction, as derived in Appendix B, which assumes that the
nozzle exit and the mixing chamber entrance are co-planar. The achieved performance may be better or
worse than predicted, depending on the mixing length required in the particular ejector.

5.5.4 Mixing chamber

The mixing chamber is assumed to be a parallel sided circular cylinder. Its length should be selected so that
energy sharing is complete before entering the diffuser but unless the jet pump is to be used in a mixing
application, the primary and secondary fluids need not necessarily be completely mixed. If the mixing
chamber is too short, energy sharing will continue into the diffuser, resulting in greater losses and possibly
in flow separation; the assumption of complete energy sharing in the mixing chamber is violated and the
use of additional mixing loss factors in the diffuser is required to predict performance. Conversely, if the
mixing chamber is excessively long, the friction losses will more than offset the gain in energy from the
mixing so there will be a net drop in static pressure (see Derivation 1, pp.177-178).

For optimum pressure rise, a mixing chamber length of about seven diameters is recommended for a
central-jet type ejector, and about five diameters for an annular-jet type. Larger area ratio jet pumps require
longer mixing chambers. Multi-nozzle ejectors or even multi-hole nozzles reduce considerably the mixing
chamber length required. For mixing applications, a mixing chamber length of about ten diameters is
recommended, so that mixing as well as energy sharing is complete by the diffuser inlet.

The primary nozzle design has a large effect on mixing chamber length (Derivation 31). If a long tube is
used for the primary nozzle, a thick boundary layer at the nozzle exit results in which case a longer mixing
chamber is required.

Optimum mixing chamber length is also related to diffuser geometry. If the diffuser included angle is larger
than recommended, a longer mixing chamber will produce a more uniform velocity profile at the diffuser
inlet, and hence help prevent flow separation. Similarly, if the diffuser angle is smaller than recommended,
a shorter mixing chamber may be used.

For low values of flow ratio, M , the major loss in the ejector is due to incomplete energy sharing; as M
increases this loss becomes less important, whereas losses due to friction increase sharply. At high flow
ratios the angle of spread of the primary fluid core is very small, i.e. there is gradual diffusion of the primary
kinetic energy to the secondary stream.

Internally the mixing chamber should be as smooth as economically feasible, since high velocities and
turbulence are generated (see Derivation 10). This is particularly important for high values of flow ratio.

5.5.5 Diffuser

The major loss in the diffuser is that caused if flow separation from the boundary occurs; this is dictated
by the divergence angle and the inlet velocity profile. When substantial mixing continues into the diffuser
inlet, there is a danger of separation, as the secondary (low energy) fluid is located near the diffuser wall.

Where the mixing chamber length is insufficient to complete mixing there will be a non-uniform velocity
profile at the diffuser inlet. For two diffusers of the same length, a non-uniform velocity profile will have
a greater adverse effect on the diffuser with the wider angle because of the greater axial static pressure
gradient.

24
 85032
It is recommended that the included angle should be 6 to 8 degrees for optimum efficiency if combined
with a mixing chamber length of 7 diameters. If the diffuser angle must be larger than this, the mixing
chamber length may be increased so as to obtain a more uniform velocity profile, as discussed in Section
5.5.4 although, if the mixing chamber is excessively long, the thicker developed boundary layer can cause
problems with flow separation in the diffuser. The diffuser included angle should not exceed 14 degrees.

The ratio of the diffuser outlet area to its inlet area should not be greater than 5, or the losses become
excessive.

If a short overall length is a practical necessity, a short mixing chamber length may be combined to
advantage with a trumpet-shaped diffuser, which offers a low rate of area change in the diffuser inlet section,
permitting mixing to be completed more efficiently.

In Derivation 6, trumpet-shaped diffusers were found to provide markedly higher efficiencies than conical
diffusers of the same length and outlet-to-inlet area ratio. Though unconventional in shape, and hence more
difficult to design and manufacture, a trumpet-shaped diffuser may be worth considering for special jet
pump applications.

The junction between the mixing chamber exit and the diffuser inlet should be radiused, as a sharp junction
would increase the likelihood of flow separation.

More details on the design of diffusers may be found in References 39 and 42.

25
 85032
6. PERFORMANCE PREDICTION (see Flow Chart in Appendix C Section C2)

Prediction of the performance of an existing ejector or jet pump is very closely related to the initial design
process; the same ratios are used to describe geometry and performance, the same coefficients are used to
allow for losses in the components and the relationship between flow ratio, M , and pressure ratio, N , is
identical.

However, the initial design involves calculating the optimum R value and ensuring that the configuration
is such that the loss coefficients are minimum for pre-defined values of flow ratio and/or pressure ratio. An
existing ejector will, however, have the geometrical ratios pre-defined, and it is unlikely that these will be
optimum if a new pressure or flow ratio is to be specified. In this case it is the predicted performance curves
for the new conditions that are to be found. Alternatively, the engineer may wish to discover the original
design point for the device.

6.1 Performance Prediction Curves

A comprehensive series of jet pump performance curves derived from Equation (5.6) are contained in
Figures 4a to 4t. The curves are presented for a density ratio of unity, several values of ( 1 + K m + K d )
and two values of K s . The coefficient K p , which has much less effect on performance, has been given a
fixed value typical of a well-designed primary nozzle.

6.2 Performance Prediction Procedure (see Flow Chart in Appendix C Section C2)

Stage 1

Obtain values for the loss coefficients.

As described in the Design Procedure (Section 5.4), the loss coefficients must be estimated from empirical
data. However, in this case the geometry of the jet pump is already known, so that the Reynolds number
corresponding to each component may be calculated, on the basis of expected flow rate and component
dimensions, and Figures 2a to 2c, Tables 11.2 and 11.3 and References 37 to 43 may be used directly to
make the first estimates.

Estimate values for the Reynolds numbers of the flow in the primary nozzle and the secondary inlet. Hence
find values of K p and K s using Figures 2a and 2b and References 38, 41 and 43. If the secondary inlet
Reynolds number is greater than 2.0 ×105 , K s may be taken as 0.10.

Estimate the relative roughness, ε ⁄ D , of the mixing chamber, from the diameter and the roughness of the
material of which it is constructed. Consideration must also be given to the length of time the ejector has
been in service, and whether the working fluids have been corrosive or abrasive. Some values of ε are
given in the Table on Figure 2c, and values for a wider range of materials are given in Reference 40. Hence
find a value of K m from Figure 2c or from Reference 37 or 40 using Equation (B2.3).

Measure the diffuser area ratio and length ratio or included angle and estimate an appropriate value for K d
from Reference 42 or calculate K d from the pressure recovery coefficient, C pr (References 41 and 42),
using Equation (5.13).

Calculate the density ratio, C = ρ s ⁄ ρ p , for the primary and secondary fluids with which the jet pump
will be operated.

26
 85032
Stage 2

Measure the primary nozzle exit and mixing chamber diameters and calculate the area ratio of the jet pump,
R = ( d ⁄ D )2 .

Stage 3

Select the performance curve from Figures 4a to 4t appropriate to the values calculated for the loss
coefficients K s , K m and K d . This curve will then define the ejector performance.

If the K p value calculated in Stage 1 is not close to 0.04, the flow ratio will still equal that given by Figures
4a to 4t, but the pressure ratio will vary as illustrated in Figure 5. Values of M and N are much less sensitive
to changes in K s , so the K s value calculated in Stage 1 may usually be approximated by one of the two
values on the Figures.

If the density ratio, C , is other than 1.0, the pressure ratio, N , is still as given by Figures 4a to 4t, but the
flow ratio becomes approximately

M = ( M C = 1.0 ) ⁄ C. (6.1)

This relationship is illustrated in Figures 6a and 6b; the curves have been derived from computed results
using Equation (5.6). The approximation becomes less satisfactory for large values of M and greater
differences in density.

If the loss coefficients are such that no suitable performance curve is included in this Item, or precise
prediction of performance is important, the engineer may construct a curve using the expression for pressure
ratio defined by Equation (5.6), substituting the measured value for area ratio, R , and the calculated loss
coefficients, and taking a suitable range of M values.

For each pair of N and M values, calculate efficiency from Equation (5.4) and plot the η versus M curve
for the ejector. This will indicate the operating point of the device, or show how much loss of efficiency
results for a particular off-design flow ratio.

Note: It is only a slight approximation to assume, on the performance charts, that the locus of the points of
optimum efficiency is, in fact, the envelope of the characteristic curves. This can easily be verified as
indicated above.

Stage 4

Calculate a value for cavitation index, σ , based on the expected efficiency of the primary nozzle-secondary
inlet combination, from the recommendations given in Section 8. Find the cavitation-limited flow ratio, as
defined by Equation (5.21) or the minimum secondary inlet pressure required, as defined by Equation (5.22).

The above procedure will result in a characteristic curve that describes the behaviour of an individual jet
pump over a complete range of non-cavitating operating conditions.

27
 85032
7. ASPECTS OF SOME MORE COMPLEX DESIGNS

7.1 Annular Jet Pumps

An annular jet pump requires only around two-thirds of the mixing length of a central jet type because of
the greater surface area of the jet. There is a small loss in the maximum efficiency achievable with an
annular jet pump, as compared to a central jet type, because the high-velocity primary jet has two boundary
layers as opposed to one and because the jet is at an angle to the secondary flow. The maximum efficiency
will occur at a different design point, as illustrated in Sketch 7.1. Annular jet pumps operate at higher R
values to give the same efficiency as central jet pumps.

A large value of R indicates a large nozzle for which, for a given power input, the driving pressure is less.
An annular jet pump therefore suits a low pressure driving supply.

An annular jet pump allows easy achievement of variable geometry, because the internal nozzle profile and
the nozzle exit area, and hence the jet pump area ratio, can be altered readily. However, the configuration
can lead to premature choking at the mixing chamber entry with restriction of the secondary flow, and there
are additional losses in the supply duct to the annular nozzle.

Sketch 7.1 Efficiency curves for annular and central jet pumps

An annular type jet pump is not normally used for abrasive applications because the nozzle cannot be made
in the hardest materials due to the accurate machining required. It is also unsuitable for dredging because
the secondary inlet is constricted, secondary velocities are high and consequently blockage by the solids is
likely. It is, however, ideal for the transport of delicate materials.

The cavitation behaviour of annular and central type jet pumps is identical but, because annular jet pumps
operate at high R values, cavitation will set in at lower values of flow ratio than for central jet pumps. On
the other hand, because of the higher R value, less pressure is needed to drive against a given head.

7.2 Multi-nozzle Jet Pumps

A multi-nozzle jet pump requires a shorter mixing chamber than a single nozzle jet pump because the greater
surface area of the primary jets improves mixing, and so this type of configuration is useful where space
is limited.

28
 85032
A further major advantage is that a multi-nozzle design improves cavitation performance as the improved
mixing reduces turbulence and local relative velocities and increases the static pressure at the entrance to
and within the mixing chamber, so that the likelihood of vapour formation is reduced. However, it is unlikely
that a multi-nozzle design would be as efficient as a single nozzle jet pump, unless the number of nozzles
was very low, because of the higher friction losses incurred.

The number of nozzles required for a particular length to diameter ratio may be calculated approximately
from

( L ⁄ D ) n = 7.0 . (7.1)

The nozzles should be spaced equally across the mixing chamber entrance, and should not be placed so as
to form a ring close to the wall as this arrangement reduces efficiency.

A multi-nozzle jet pump will produce approximately the same discharge flow and pressure as a single
nozzle jet pump at non-cavitating conditions, but will provide a higher outlet pressure at reduced secondary
inlet pressure conditions. This means that the multi-nozzle jet pump is capable of operating at much lower
secondary inlet pressures than a single nozzle jet pump, and thus exhibits an improved range of operation.

In the study reported in Derivation 13, tests were carried out on three types of multi-hole nozzle. These
were: a parallel nozzle which produced several primary jets parallel to the mixing chamber, a convergent
nozzle which produced jets converging towards the end of the mixing chamber, and a skew nozzle which
produced primary streams swirling around the secondary (low velocity) stream. The tests indicated that
there was little difference between the results obtained from each nozzle type; a parallel nozzle is probably
preferable for greatest ease of fabrication.

7.3 Multi-stage Jet Pumps

The use of series-staging of jet pumps greatly increases the flexibility of the device and provides an effective
way of achieving a wide range of operating flow ratios, by varying the number of stages present. It is also
a useful means of achieving a high secondary mass flow rate, i.e. as high a value of M as possible. However,
the configuration will not achieve higher pressure ratios.

The concentrations and relative proportions of the fluids may be controlled relatively easily using a
series-staged jet pump; one of the most direct applications is in the mixing of several streams of different
fluids under controlled conditions.

Single jet pumps are limited by the occurrence of cavitation caused by high velocity secondary inlet flow.
By splitting the secondary flow between two pumps so that the discharge from the first becomes the
secondary flow to the second, the cavitation problem can be significantly reduced.

Alternatively, the discharge flow from the first jet pump may become the primary flow to the second. This
arrangement is extremely useful if there are problems in siting the jet pumps, so that the first stage can act
as a booster pump to the second larger main stage.

A schematic diagram of the layout of a series-staged jet pump is shown in Sketch 7.2. It is common that
only the final stage contains a diffuser.

29
 85032

Sketch 7.2 Schematic layout of an n-stage ejector

30
 85032
8. CAVITATION

8.1 Description

Whenever the pressure within a flowing liquid is locally reduced to its vapour pressure, bubbles form and
then, as the pressure increases again, collapse. This phenomenon is known as cavitation.

Cavitation reduces flow efficiency, causes noise and vibration and damages local surfaces.

One of the most important problems in the design of jet pump systems is the prediction of the onset of
cavitation. When cavitation occurs the secondary flow ceases to respond to the discharge pressure
(cavitation limited flow) and performance deteriorates rapidly. This causes the jet pump characteristic curve
to drop sharply, as illustrated in Sketch 8.1. Superficially, cavitation-limited flow resembles choked
compressible flow. The consequent power loss may be acceptable in small units where deliberate design
within the cavitation regime allows use of the jet pump to give a metering characteristic, i.e. a constant
flow over a range of discharge heads. In general, however, it is preferable to keep cavitation to a minimum.

Sketch 8.1 Effect of cavitation on a jet pump characteristic curve

Cavitation occurs principally at the entrance to the mixing chamber or at the beginning of the diffuser.
Cavitation at the mixing chamber entrance results from high primary jet velocities, low secondary pressure
or low discharge pressure, P 5 . Cavitation at the diffuser inlet will occur if the velocity head of the combined
primary and secondary flows at the end of the mixing chamber exceeds the downstream absolute pressure.
This is much less common than cavitation at the mixing chamber entrance, which is considered in the
following discussion.

Incipient cavitation first appears in the shear layer at the periphery of the primary jet (Derivation 23). This
is promoted by increasing jet velocity and by reducing absolute pressure in the mixing zone. In the mixing
chamber, increasing vapour release occurs as the stress is increased but this bubble formation along the jet
has little or no effect on pump performance. Cavitation-limited flow occurs whenever cavitation is so severe
that the bubble cloud chokes the mixing chamber, touching the wall.

31
 85032
Two factors are important in the promotion of cavitation: the velocity of the driving jet and the secondary
inlet pressure being applied to the pump. Cavitation becomes a greater problem as the flow ratio, M , is
increased because a higher flow ratio produces lower levels of static pressure in the secondary inlet region
of the pump (due to higher velocities), and a lower axial pressure gradient in the mixing chamber. Both
effects are conducive to cavitation.

A secondary factor that contributes to an increased susceptibility to cavitation at small nozzle spacings is
the wake produced by the nozzle exit wall. This wake increases the turbulence in the mixing layer where
cavitation occurs, and has a greater influence at smaller nozzle spacings because the static pressure in the
mixing chamber is low.

8.2 Cavitation Index

Use of a cavitation index is a method whereby the conditions existing at the onset of cavitation in a particular
ejector can be characterised by a single number. Several definitions for the cavitation index, σ , are possible.
The definition used in this Item is based on considerations of energy and momentum conservation in the
secondary inlet, and the assumption that, at the point of total pressure breakdown, the static pressure in the
plane of the nozzle exit is equal to the fluid vapour pressure. A detailed derivation is given in Appendix B
which results in the expression

2
σ = ( P 2 – pv ) ⁄ ( ½ ρs V3 c ) , (8.1)

where V 3 c is the cavitation-limited velocity of the secondary fluid at the mixing chamber entrance. The
term ( P 2 – p v ) is analogous to the NP S H (net positive suction head) in pump technology.

The value of σ specific to a particular jet pump must be measured experimentally. It can be seen, by
reference to Equation (8.1), that the lower the value of σ the less likely it is that the device will experience
cavitation problems during operation. This is because a low σ requires either a low value of P 2 or a high
value of V 3 , both of which should be avoidable in practice.

Conversely, a high measured value of σ indicates that to avoid cavitation, P 2 must be high or V 3 low,
which may place operational restrictions on the device.

Measured values of σ have been found to range from 1.0 to 1.67 approximately. For a well-designed jet
pump σ is around 1.0 - 1.1.

At the design stage the cavitation index for the jet pump will, perforce, be unknown. In order to assess the
likelihood that a proposed configuration will experience cavitation problems at the design point, or indeed
to estimate the limits of its operating range, a value of σ must be assumed.

With regard to the available data (see Derivations 24 and 26), a conservative value of the cavitation index
can be taken to be σ = 1.35 . Using this figure, a pessimistic estimate of the limits imposed by cavitation
should be available. The value holds for water, oils and mercury, but care should be taken when considering
high gas-solubility fluids, as the high gas content can lead to more severe problems. A more suitable value
of σ for an individual jet pump than 1.35 can be found from brief tests on a prototype.

32
 85032
A cavitation-limited flow ratio, M c , may be defined in terms of the index, σ , thus:

1–R P2 – p v ½
Mc = -------------
- ----------------------------
- . (8.2)
R ½ρ p V n 2 C σ

This expression is derived in Appendix B. If M c is less than or equal to the operating flow ratio, which is
usually set equal to the best efficiency value, the pump will not perform according to theoretical predictions.

An alternative criterion to describe cavitation-limited performance is the minimum secondary inlet pressure
required to prevent cavitation. This criterion is useful if design requirements place a limit on the available
secondary inlet pressure and is defined by

Q p MR 2
P2 = ½ρ p Cσ ------- -------------- + p v . (8.3)
min An 1 – R

Equation (8.3) is also derived in Appendix B.

8.3 Improving Cavitation Performance

The cavitation performance of a device may be improved by reducing the design area ratio, reducing the
primary jet velocity or by increasing the net positive suction head, ( P 2 – p v ) . An improvement may also
be obtained by smoothing pump contours, suppressing dissolved gas effects or by retracting the primary
nozzle from the mixing chamber entrance plane. The static pressure in the secondary inlet and mixing
chamber region increases as the nozzle is retracted, because the resulting increase in annular area of the
secondary inlet produces lower velocities and higher static pressures for a given flow rate. However, if the
nozzle is retracted to a large spacing (about 2 to 3 mixing chamber diameters) the static pressure is decreased
near the mixing chamber exit due to increased frictional losses.

As mentioned in Section 5.5.2, the design of the secondary inlet region, including the external contour of
the primary nozzle, is critical to the jet pump cavitation performance. The secondary inlet flow path should
smoothly converge to the mixing chamber entrance and the primary nozzle wall should be as thin as possible.

Another means of improvement is to use a two-stage ejector. In this configuration the first stage operates
at a relatively low supply head with little or no cavitation, its combined flow being fed into the secondary
inlet of the second stage under such a pressure that cavitation is suppressed there also. Alternatively a
multi-nozzle jet pump may be employed.

The presence of sediment in the secondary fluid should not greatly alter the incidence of cavitation, although
in liquids with a high dissolved gas content, sand grains may serve as nucleii for gas bubbles to form around.
This phenomenon can have an ameliorating effect in that it can result in a reduction in the erosion of the
mixing chamber.

33
 85032
9. WORKED EXAMPLES

9.1 Design Procedure - Example 1

A new product in a food processing plant requires, for one of its stages, a pressurised supply of brine. The
salt concentration in the solution must be controlled at 30 g/l with a flowrate of 20 1/s. The delivery pressure
is set by the process at 200 kPa absolute.

The factory uses brine in a number of areas and a large central reservoir is maintained at a concentration
of 60 g/l to be diluted as required. The water for dilution is supplied from a pump whose maximum delivery
pressure is 450 kPa absolute.

It is proposed to use a jet pump to effect both the delivery and the mixing of the brine.

Stage 1 - Determine the design criteria

– To achieve the desired salt concentration the main reservoir brine must be diluted in equal
parts with water. The flow ratio, M , is therefore fixed at unity.
– At a concentration of 60 g/l, the main reservoir brine has a relative density of 1.1. With
water as the primary fluid and the reservoir brine as the secondary, the density ratio, C , is
set at 1.1.
– At 15°C the density of water, i.e. ρ p , is 998 kg/m3.
– The vapour pressure of the reservoir brine is around 3 kPa.
– The total flowrate, Q 5 , is fixed by the process requirements at 0.02 m3/s.
– The delivery pressure, P 5 , is fixed at 200 kPa absolute.
– The secondary inlet pressure, P 2 , is not determined by the process but for convenience it is
proposed to construct a subsidiary constant head tank adjacent to the new equipment. The
head in this tank, 3 m, fixes the secondary pressure.
–3
P 2 = ( 3 × 1.1 × 998 × 9.81 ×10 ) + 101.3 kPa absolute = 133.6 kPa absolute.
– The pump supplying the process water sets an upper limit on the primary pressure,
i.e. P 1 ≤ 450 kPa.
– At this stage a nominally conservative value for the cavitation index, σ , can be assumed,
σ = 1.35 .

Stage 2 - Obtain values for the loss coefficients

5
– Assuming, as a first guess, a primary nozzle Reynolds number of 1 ×10 , the loss
coefficient, K p , can be read from Figure 2a as K p = 0.05 .
– As a first approximation, the secondary inlet loss coefficient, K s , will be assumed to be
0.10.

34
 85032
– The jet pump will be constructed of galvanised steel, a material in common use in the food
industry. From the table in Figure 2c, the roughness value, ε , of galvanised steel is 0.15
mm. As a first approximation assume that the mixing chamber will be 100 mm diameter.
This gives a relative roughness, ε ⁄ D , of 0.0015. Using this value, with the recommended
L ⁄ D ratio for a central jet ejector of 7.0, an estimate of K m can be obtained from Figure 2c
as K m = 0.150 .
– Assuming a diffuser half angle of 3°, as recommended, Reference 42 suggests a diffuser
pressure recovery coefficient of 0.84. Assuming also a recommended area ratio of 5.0, K d
can be calculated using Equation (5.13):
K d = ( 1 – ( 1 ⁄ 5 ) 2 ) – 0.84 = 0.12 .

It is now possible to complete most of the calculation sheet (see Table 9.1). The master for this sheet is
Table 11.1.

Stage 3 - Obtain a value for the area ratio R.

The only parameter now required to allow the design to proceed is the ratio of the primary nozzle exit area
to the area of the mixing chamber. The design flow ratio, M , is fixed at 1.0 and Figures 3a and 3b present
empirical and theoretical optimum area ratios for given flow ratios. Using M = 1.0 with Figure 3b gives
an estimated value for R of 0.25, for C = 1.1 .

Stage 4

The jet pump is to work at a single operating point and it would be possible now to use Equation (5.6) to
calculate the pressure ratio, N , corresponding to the mass flow ratio, M = 1.0 , noting that as
N × M ≤ 1.0, N ≤ 1.0 .

Since this is the first iteration using several estimated parameters, an approximate straight line characteristic
will be developed, from which further information can be derived. The straight line approximation is fair
for R values ≤ 0.4 which is well above the present estimate of 0.25.

To find the end points of the characteristic, Equations (5.15) and (5.16) must be used.

Note that N 0 does not depend on the density ratio, C , whereas M 0 does.

2R – R 2 ( 1 + K m + K d )
N 0 = -------------------------------------------------------------------------------------- .
1 + K p – 2R + R 2 ( 1 + K m + K d )
( 2 × 0.25 ) – 0.25 2 ( 1 + 0.150 + 0.12 )
So that N 0 = -----------------------------------------------------------------------------------------------------------------------------
1 + 0.05 – ( 2 × 0.25 ) + 0.25 2 ( 1 + 0.150 + 0.12 )

= 0.668 .

The quadratic coefficients in Equation (5.16) must now be calculated:

2CR 2 R 2
a = ------------- – CR 2 ( 1 + K m + K d ) – C  ----------- ( 1 + K s )
1–R  1 – R

35
 85032
2 × 1.1 × 0.25 2 0.25 2
= ---------------------------------- – 1.1 × 0.25 2 ( 1 + 0.15 + 0.12 ) – 1.1  ------------------ ( 1 + 0.1 )
1 – 0.25  1 – 0.25

= – 0.0384 ,

b = –R 2 ( 1 + C ) ( 1 + Km + Kd )

= – 0.25 2 ( 1 + 1.1 ) ( 1 + 0.15 + 0.12 )

= – 0.1667 ,

c = 2R – R 2 ( 1 + K m + K d )

= 2 × 0.25 – 0.25 2 ( 1 + 0.15 + 0.12 )

= 0.4206 .

0.1667 – [ 0.1667 2 + ( 4 × 0.0384 × 0.4206 ) ]

So that, M 0 = ---------------------------------------------------------------------------------------------------------------------
-
2 × ( – 0.0384 )

= 1.787 .

The straight line characteristic can now be plotted and is shown in Figure 7a. For a flow ratio of M = 1.0 ,
the pressure ratio, N , is read off as 0.30.

Stage 5 - Efficiency

By picking pairs of points from the M versus N line the efficiency can be easily calculated.

The efficiency curve has been plotted on Figure 7a and it can be seen that the operating point is almost
exactly at peak efficiency.

Stage 6 - Remaining parameters

In order to be able to refine the loss coefficient values, and also to check that no design constraints have
been violated, this stage of the design is used to fill in the remaining gaps in the Calculation Sheet (Table 9.1).
_
The primary pressure is found from:

P1 = [ P5( 1 + N ) – P2 ] ⁄ N
= [ 200 ( 1 + 0.30 ) – 133.6 ] ⁄ 0.30
= 421.3 kPa absolute

36
 85032
This is below the maximum available driving pump pressure of 450 kPa.
_
The primary nozzle exit area can now be calculated from
MR 2 ½
( 1 + K p ) – C ( 1 + K s )  --------------
 1 – R
A n = Q p ----------------------------------------------------------------------------------
( P 1 – P2 ) ⁄ ½ ρp

1.0 × 0.25 2 ½
( 1 + 0.05 ) – 1.1 ( 1 + 0.1 )  -------------------------
 1 – 0.25 
= 0.01 ------------------------------------------------------------------------------------------------------
3
( 421.3 – 133.6 ) × 10 ⁄ 0.5 × 998

= 0.000398 m2.

So the nozzle exit diameter is given by

4 ½
d =  0.000398 × ---
 π

= 0.02253 m, say 22.5 mm.

The mixing chamber diameter is now available from

Am = An ⁄ R
= 0.000398 ⁄ 0.25
= 0.001594 m2,

from which D = 0.04505 m, say 45 mm.

_
The optimum mixing chamber length for a single central jet ejector is given by

L = 7×D
= 0.315 m.

Stage 7 - Cavitation

It is now necessary to check the cavitation performance. The flow ratio limit for cavitation is given by
2 ½
1 – R ( P2 – p v ) An
M c = -------------- ----------------------------------
R ½ρ p Cσ Q p
2

3 ½
1 – 0.25 ( 133.6 – 3.0 ) × 10 × 0.000398 2
= --------------------- -----------------------------------------------------------------------------------
0.25 0.5 × 998 × 1.1 × 1.35 × 0.01 2

= 1.59 .

37
 85032
Stage 8 - Loss coefficient re-evaluation

The primary nozzle Reynolds number is given by

Re p = V n d ⁄ ν p = Q p d ⁄ ( A n ν p )

0.01 × 0.0225
= -------------------------------------
–6
0.000398 ×10
5
= 5.7 ×10 ,

so from Figure 2a, K p ≈ 0.04 .

The secondary inlet Reynolds number is given by

Qs ( D – d )
Re s = --------------------------------
( A m – A n )ν s
0.01 ( 0.045 – 0.0225 )
= -------------------------------------------------------------------------------------------
–6
( 0.001594 – 0.000398 ) × 1.18 ×10
5
= 1.59 ×10 ,

so, from Figure 2b, K s ≈ 0.11 .

The mixing chamber diameter has been calculated to be around 45 mm which, with an ε value of 0.15 mm,
gives a relative roughness, ε ⁄ D , of 0.0033.

Using the curve for a central jet ejector on Figure 2c gives K m = 0.186 .

No further information on the diffuser is available so the original estimate of K d = 0.12 will be retained.

Stages 4 to 7 can now be repeated with the new loss coefficient values.

Stage 4 repeated.

Again the straight line approximation is used. The new end points are calculated as

N 0 = 0.673 and M 0 = 1.717 ( a = – 0.0421, b = – 0.1714, c = 0.4184 ) .

The relationship has been plotted in Figure 7b and gives an operating pressure ratio of N = 0.28 .

Stage 5 - Efficiency

Once again the efficiency curve can be plotted and is shown on Figure 7b. The operating point is still close
to peak efficiency; operating η = 0.28 , with peak η = 0.288 .

Stage 6 - Remaining parameters

38
 85032
The remaining parameters can now be re-evaluated using the expressions in Stage 6.

– Primary pressure = 437.1 kPa absolute.

– Primary nozzle exit area = 0.000385 m2,
which gives d = 0.0221 m, or 22 mm.
– The mixing chamber area = 0.001540 m2,
which gives D = 0.0443 m or 44 mm.
– The mixing chamber length, L , from the optimum ratio, L ⁄ D = 7 , is 0.308 m.

Stage 7 repeated

The flow ratio limit set by cavitation is recalculated as

M c = 1.53 .

This limit is marked on Figure 7b and will not cause a problem to the jet pump operation.

For comparison purposes, as an alternative to using the straight-line approximate method for the second
iteration, a more accurate characteristic has been plotted in Figure 7c by using Equation (5.6).

Returning to Stages 4 to 7, but now using Figure 7c gives the following results.

– Primary pressure, P 1 , = 429.8 kPa absolute.

– Primary nozzle exit area, A n = 0.000390 m2,
from which d = 0.0223 m or 22.3 mm.
– Mixing chamber diameter, A m = 0.00156 m2
from which D = 0.0446 m, say 45 mm.
– Mixing chamber length, L = 7D = 0.315 m.

Stage 9

The jet pump is operating extremely close to peak efficiency and none of the loss coefficients will change
significantly on further iterations so the design can be considered fixed as far as working pressures and
dimensions are concerned. The guidelines given in Section 5.5 should be followed before completing the
detailed design of the components.

A sketch of the basic configuration is shown in Figure 7d and the final values of the parameters are shown
in the calculation sheet (Table 9.1).

39
 85032
TABLE 9.1 Calculation Sheet for Example 1 - Design Procedure40
Parameter Relationship Value

First iteration Final design

Primary pressure P 1 = [ P5 ( 1 + N ) –P2 ] ⁄ N <450 kPa 429.8 kPa
Secondary pressure P 2 = P 5 – N ( P1 – P5 ) 133.6 kPa 133.6 kPa
Discharge pressure P 5 = ( NP 1 + P 2 ) ⁄ ( N + 1 ) 200.0 kPa 200.0 kPa
Pressure ratio N = ( P5 – P2 ) ⁄ ( P1 – P5 ) ? 0.289
Primary flowrate Qp = Qs ⁄ M 0.01 m3/s 0.01 m3/s
Secondary flowrate Q s = M × Qp 0.01 m3/s 0.01 m3/s
Flow ratio M = Qs ⁄ Qp 1.0 1.0
Mixing chamber diameter D = L ⁄ 7 or L ⁄ 5 ? 45.0 mm
Mixing chamber length L = 7D or 5D ? 0.315 m
Primary nozzle exit diameter d=D R ? 22.3 mm
Primary nozzle spacing s≈d ? 22 mm
Diffuser included angle 2φ 6 6
Primary fluid density ρp 998.0 kg/m3 998.0 kg/m3
Secondary fluid density ρs 1098.0 kg/m3 1098.0 kg/m3
Density ratio C = ρs ⁄ ρp 1.1 1.1
2
Cavitation index σ = ( P 2 – p v ) ⁄ ( ½ρs V 3c ) 1.35 1.35
Primary nozzle loss coefficient Kp 0.05 0.04
Secondary inlet loss coefficient Ks 0.10 0.11
Mixing chamber loss coefficient Km 0.15 0.186
Diffuser loss coefficient Kd 0.12 0.120

40
 85032
9.2 Design Procedure - Example 2

A jet pump is already in use in a nuclear power station circulating liquid sodium, at a temperature of 920
K, through the core of the reactor. It has a mixing chamber diameter of 24.0 mm and an area ratio of 0.0425.
The primary nozzle system can supply an absolute pressure of 550 kPa, which, in the present configuration,
develops a mass flow rate of 0.5 kg/s. The induced, or secondary, flow is developed from an inlet pressure
of 170 kPa absolute at a rate of 1.5 kg/s.

A performance upgrade of the reactor requires that the sodium coolant ejector outlet pressure must be
increased from 195 kPa to 205 kPa absolute. The current device is an integral part of the plant and, rather
than replace the entire unit, it has been decided simply to replace the primary nozzle by one compatible
with the new operating requirement.

The new primary nozzle exit diameter is to be determined.

Stage 1 - Determine the design criteria

– The primary pressure is fixed by the motive unit at 550 kPa absolute.
– The secondary inlet pressure is fixed by the recirculation pipework at 170 kPa absolute.
– The secondary flowrate is similarly fixed at 1.5 kg/s.
– The mixing chamber diameter is measured at 24.0 mm.
– The primary and the secondary fluids are both sodium at 920 K and have a density of 790.5
kg/m3.
– Vapour pressure for the sodium is 6.62 kPa absolute.
– A conservative value of 1.35 will be assumed for the cavitation index.

Stage 2 - Loss coefficients

In terms of overall flow rates and configuration the upgraded ejector will not be significantly different from
the current design. The loss coefficients, therefore, are available and, it can be assumed, will alter negligibly
with the new nozzle. The loss coefficients are:

K p = 0.03 ,
K s = 0.10 ,
K m = 0.10 ,
K d = 0.10 .

It is finally assumed that the original device was an optimum design for its duty, i.e. the mixing chamber,
diffuser, etc. were constructed according to the optimum dimensions and ratios calculated.

41
 85032
Stage 3 - Area ratio

In fact, in this example, the area ratio, R , is the object of the design. The approach in a case such as this is
to 'design' a number of configurations based on a range of area ratios. The resulting parameters can be
plotted as bespoke design charts and the required operating conditions will be used to interpolate, on the
charts, for the missing data.

The required range of area ratios will in this case be from 0.04 (the current design) to, say, 0.07.

Stage 4 - Characteristic curve

The object of this stage is to calculate, for the given range of R , values of M and N . The design constraint,
the exit pressure, P 5 , will then be available as will the efficiency, η .

In this example the operating point of the ejector is so well constrained that a plot of the pump characteristic
is not necessary, a single operating point for each area ratio being sufficient.

The relevant parameters are first evaluated for R = 0.04 .

The mixing chamber area, A m , is found from

π
A m = --- 0.024 2
4
–4
= 4.524 × 10 m2.

The nozzle exit area is then,

A n = RA m

–4
= 0.04 × 4.524 × 10

–5
= 1.810 × 10 m 2,

giving a nozzle diameter from

4 ½
d =  --- × A n .
π 

4 –5 ½
So d =  --- × 1.810 × 10 
π 
= 4.80 mm.

The primary mass flowrate, Q p , is related to the nozzle area in an expression (see Equation (5.20) or
Equation (B6.4)) involving M , the flow ratio. The determination of M is, therefore, an iterative process.

42
 85032
A first approximation for Q p can be obtained by rearranging Equation (5.20) and omitting the term in M ,
which is relatively small:

( P 1 – P2 ) ½
Q p = An -------------------------------
( 1 + K p )½ρ

3 ½
–5 ( 550 – 170 ) × 10
= 1.810 × 10 -----------------------------------------------------
( 1 + 0.03 ) ½ × 790.5
–4
= 5.53 × 10 m3/s.

This gives M as

M = Qs ⁄ Qp
–4
= ( 1.5 ⁄ 790.5 ) ⁄ ( 5.53 × 10 )
= 3.43 .

Substituting this value of M into Equation (5.20) will provide a new estimate for Q p from
½
( P 1 – P 2 ) ⁄ ½ρ p
Q p = A n ---------------------------------------------------------------------------------
MR 2
( 1 + K p ) – C ( 1 + K s )  -------------- 
 1 – R

½
3
–5 ( 550 – 170 ) × 10 ⁄ ( ½ × 790.5 )
as Q p = 1.810 × 10 -----------------------------------------------------------------------------------------
3.43 × 0.04 2
1.03 – 1.0 ( 1 + 0.1 )  ----------------------------
 1 – 0.04 

–4
= 5.59 ×10 m3/s.

This is sufficiently close to the original guess that no further iteration is required and

M = Qs ⁄ Qp
–4
= ( 1.5 ⁄ 790.5 ) ⁄ ( 5.6 × 10 )
= 3.39 .

The primary mass flow rate is

–4
Q p × ρ = 5.60 × 10 × 790.5
= 0.442 kg/s.

43
 85032
The pressure ratio, N , can be evaluated using Equation (5.6) or, for the present case of equal primary and
secondary fluid densities, the slightly shorter version given in Appendix B (Equation (B5.2)), from which

N = 0.0626 .

Knowing the pressure ratio along with the primary and secondary inlet pressures allows a value for the
discharge pressure, P 5 , to be calculated by rearranging Equation (5.1):

N P1 + P 2
P 5 = -------------------------
1+N
0.0626 × 550 + 170
= --------------------------------------------------
1 + 0.0626
= 192.4 kPa absolute.

The efficiency of the device can be calculated from Equation (5.4):

η = M ×N
= 3.39 × 0.0626
= 0.212 ., i.e. 21.2 %.

This calculation method, effectively Stage 4 of the design procedure, can be repeated in an identical manner
for the full range of area ratios required. The pertinent results can then be plotted against R on bespoke
design charts as has been done in Figure 8.

The exit pressure, P 5 , is the relevant design constraint and so, from Figure 8 with P 5 = 205 kPa, the
required value of R can be read off as R = 0.0615 .

Knowing R , the other three design parameters can also be found:

primary nozzle exit diameter, d = 5.95 mm,

primary mass flow rate, Q p = 0.68 kg/s,

efficiency, η = 22.4 %.

Stage 7 - Cavitation

Although the example has shown a slightly different approach to design which has obviated the need for a
number of the procedural stages, now that a suitable configuration has been devised it should be checked
to ensure that cavitation will not present a problem.

The limiting value of the cavitation index was taken to be 1.35, which characterises an averagely constructed
jet pump. Since the secondary pressure is effectively fixed, the cavitation check must centre around the

44
 85032
critical mass ratio, M c , which can be calculated using Equation (5.21), expressing V n as Q p ⁄ A n :

2 ½
1 – R ( P 2 – p v )A n
M c = -------------- ----------------------------------
R ½ρ p Cσ Q p
2

 3 ½
1 – 0.0615  ( 170 – 6.62 ) × 10 × ( π × 0.00595 2 ⁄ 4 ) 2 
= ---------------------------  ---------------------------------------------------------------------------------------------------------- 
0.0615 0.68 2
 ½ × 790.5 × 1 × 1.35 ×  ------------- 
  790.5  
= 8.63 .

This value for M c of 8.63 is, as would be expected for such a low pressure ratio ejector, well above the
operating mass flow ratio.

The re-design has been successfully effected and shows that, by replacing the existing nozzle with a new
5.95 mm diameter unit, the upgraded delivery pressure of 205 kPa will be achieved with no cavitation
problems.

9.3 Performance Prediction - Example 3

A jet pump has been designed to pump 0.8 kg/s of aircraft fuel at a temperature of 40ºC f rom a pressure
of 40 kPa to 80 kPa absolute. This fuel pump must, however, be able to operate over a range of flows and
pressures to meet varying engine requirements, always with a constant flow ratio of 2.5. The primary
pressure can vary up to a maximum of 1000 kPa absolute to meet the flow range requirement; secondary
flow rate can vary from 0.5 to 1.0 kg/s. The primary and secondary fluid densities can be taken to be equal
–6
at 765 kg/m3 and their kinematic viscosities 1.4 × 10 m2/s. The primary nozzle exit diameter is 3.8 mm,
the mixing chamber has a diameter of 17.0 mm and surface roughness is estimated as 0.025 mm. The
diffuser area ratio is 3.5 and its exit diameter is 31.8 mm.

Stage 1 - Determine loss coefficients and fluid properties

_
The design secondary mass flow is 0.8 kg/s, from which the secondary volume flow rate, Q s , is
0.8/765 = 0.00105 m3/s.
_
The flow ratio, M , is held constant at 2.5. The primary flow rate is given by

Qp = Qs ⁄ M
= 0.00105 ⁄ 2.5
= 0.00042 m3/s.
_
Total volume flow rate is

Q p + Q s = 0.00042 + 0.00105

= 0.00147 m3/s.

45
 85032
_ The primary nozzle Reynolds number is given by

Qpd
Re p = ------------
An νp

0.00042 × 0.0038 × 4
= -----------------------------------------------------------------
–6
π × 0.0038 2 × 1.4 × 10
5
= 1 × 10

and, from Figure 2a, K p ≈ 0.05 .

_
The secondary inlet Reynolds number is given by

Qs ( D – d ) 4 Qs
Re s = -------------------------------- = ------------------------------
( A m – A n )ν s π ( D + d ) νs

4 × 0.00105
= ---------------------------------------------------------------------------------
–6
π ( 0.017 + 0.0038 ) × 1.4 × 10
4
= 4.6 ×10

and, from Figure 2b, K s ≈ 0.38 .

_
The mixing chamber relative roughness, ε ⁄ D = 0.025/17 = 0.0015. From Figure 2c, this
corresponds to a value for K m of 0.15.
_
The diffuser area ratio is 3.5, and the static pressure recovery coefficient is estimated from
Reference 42 to be 0.80, making K d ≈ 0.12 .
_
The primary and secondary fluids are both aircraft fuel, so that the density ratio, C , is unity.
_
The vapour pressure, p v , for the fuel is 13.8 kPa absolute.

Stage 2 - Determine the area ratio

Primary nozzle exit diameter, d = 0.0038 m and mixing chamber diameter, D = 0.017 m, so that the
area ratio,

R = (d ⁄ D)2
= ( 0.0038 ⁄ 0.017 ) 2
= 0.05 .

46
 85032
Stage 3

The performance curve most closely corresponding to the required conditions is Figure 4o.

The curve for R = 0.05 is replotted in Figure 9a, from which, for a value of M = 2.5, N = 0.078 .

It may be seen from the η versus M curve also plotted on Figure 9a that the jet pump is working fairly
close to peak efficiency at the design operating point.

Primary pressure, P 1 , varies up to a maximum of 1000 kPa to meet the range of flow requirements. Taking
values of P 1 within the range 0 - 1000 kPa, it is possible to calculate corresponding values of P 5 and Q s
for each operating point.

For P 1 = 200 kPa:

P 5 = ( N P1 + P 2 ) ⁄ ( 1 + N )

( 0.078 × 200 ) + 40
= -------------------------------------------------
1 + 0.078
= 51.6 kPa absolute.

Rearranging Equation (5.20) gives the primary flow rate,

½
( P 1 – P 2 ) ⁄ ( ½ρ p )
Q p = A n ---------------------------------------------------------------------------------
MR 2
( 1 + K p ) – C ( 1 + K s )  -------------- 
 1 – R
½
3
π ( 200 – 40 ) × 10 ⁄ ( ½ × 765 )
= --- ( 0.0038 ) 2 ---------------------------------------------------------------------------------------------------
-
4 2.5 × 0.05 2
( 1 + 0.05 ) – 1.0 ( 1 + 0.38 )  -------------------------
 1 – 0.05 
= 0.00024 m3/s.

The secondary volume flow rate is given by

Qs = M × Q p
= 2.5 × 0.00024
= 0.00058 m3/s

from which the secondary mass flow rate

= 0.46 kg/s.

These steps can be repeated for the range of primary pressure, P 1 , to give the following results:

47
 85032
P 1 kPa abs 200 400 600 800 1000
P 5 kPa abs 51.6 66.0 80.5 95.0 109.5
Q p m3/s 0.00024 0.00034 0.00043 0.00050 0.00056
Q s m3/s 0.00060 0.00087 0.00109 0.00125 0.00140
Secondary mass 0.46 0.67 0.83 0.96 1.07
flow kg/s

The engine operating curve may now be plotted and is shown in Figure 9b.

Stage 4

It is now necessary to check for the occurrence of cavitation. A conservative value of σ = 1.35 will be
taken for the cavitation index.

The cavitation-limited flow ratio, M c , will change as primary pressure, P 1 , varies because the primary
flow rate is changing. Equation (8.3) can be expressed as

2 ½
1 – R ( P 2 – p v ) An
M c = -------------- ---------------------------------- .
R ½ρ p Q p 2 C σ

The primary nozzle exit area is

π π –5
A n = --- D 2 = --- × 0.0038 2 = 1.134 × 10 m2.
4 4

For P 1 = 200 kPa absolute: Q p = 0.00023 m3/s, as calculated previously, and hence

3 –5 2 ½
1 – 0.05 ( 40 – 13.8 ) × 10 × ( 1.134 × 10 )
M c = --------------------- ----------------------------------------------------------------------------------------------
0.05 ½ × 765 × 0.00023 2 × 1 × 1.35
= 6.67 .

Repeating the calculation for the required range of primary pressures gives:

P 1 kPa abs 200 400 600 800 1000

Mc 6.67 4.51 3.57 3.07 2.74

The maximum value of M set by the onset of cavitation is always greater than the operating value of 2.5,
within the operating range of secondary mass flows, so that cavitation should not be a problem.

48
 85032
10. DERIVATION AND REFERENCES

10.1 Derivation

The derivation lists selected sources that have assisted in the preparation of this Item.

1. GOSLINE, J.E. The water jet pump. Univ. of California Publ. Eng., Vol.3, No.3,
O'BRIEN, M.P. pp.167-190, 1934.
2. STEPANOFF, A.J. Centrifugal jet pump systems. Test book on mixed and axial flow
pumps, pp.402-421. McGraw Hill, 1948.
3. SMITH, R.A. Theory and design of simple ejectors. Paper 13, Conf. “Some Aspects
of Fluid Flow”, Inst. Physics, 1950.
4. CUNNINGHAM, R.G. Jet pump theory and performance with fluids of high viscosity. Trans.
am. Soc. mech. Engrs, Vol.79, Pt.8, pp.1807-1820, 1957.
5. BONNINGTON, S.T. The cavitation limits of a liquid-liquid jet pump. Rep. RR 605, Brit.
Hydrodynamics Res. Assoc., Cranfield, UK, 1958.
6. HENDERSON, F.D. Effect of profile and length on the efficiency of pump diffusers. Tech.
Note 181. Rocket Propulsion Est., Westcott, UK, Min. Supply, 1959.
7. KLINE, S.J. Optimum design of straight-walled diffusers. J. basic Engng, Vol.81,
ABBOTT, D.E. No.3, pp.321-331, 1959.
FOX, R.W.
8. MASLIN, A.B. Jet pumps. Assoc. Eng. Shipbuilding Draughtsmen, Vol.12/59,
pp.5-62, 1959.
9. SILVESTER, R. Characteristics and applications of the water jet pump. La Houille
Blanche, No.4, pp.451-460, 1961.
10. MUELLER, N.H.G. Water jet pump. Proc. am. Soc. civil Engrs, J. Hydraulics Div., Vol.90,
No.HY3, pp.83-113, 1964.
11. GLUNTZ, D.M. Jet pumps advance BWR recirculation flow-design. Nucleonics,
MOEN, R.H. Vol.23, No.12, 1965.
WRAY, J.L.
12. HANSEN, A.G. The design of water-jet pumps. I - Experimental determination of
KINNAVY, R. optimum design parameters. Paper 65-WA/FE-31, Am. Soc. mech.
Engrs, 1965.
13. LEWIS, R.A. An experimental analysis of a jet inducer with multiple nozzles
calibrated in liquid mercury and in water. Am. Soc. mech. Engrs
Symp. “Cavitation in Fluid Machinery”, pp.109-119, 1965.
14. SIDHOM, M. A study of the performance of staged-jet pumps. Paper 66-WA/FE-37,
HANSEN, A.G. Am. Soc. mech. Engrs, 1966.
15. WILMAN, J.T. Jet pumps. Rep. EUR 3253.e, European Atomic Energy Community,
1966.

49
 85032
16. HANSEN, A.G. A jet pump cavitation parameter based on NPSH. Paper 68-WA/FE-42,
NA, T.Y. Am. Soc. mech. Engrs, 1968.
17. REDDY, Y.R. Theory and performance of water jet pumps. Proc. am. Soc. civil
KAR, S. Engrs., J. Hydraulic Div., Vol.94, No.HY5, pp-1261-1281, 1968.
18. SANGER, N.L. Non-cavitating performance of two low-area-ratio water jet pumps
having throat lengths of 7.25 diameters. NASA tech. Note
D-4445,1968.
19. SANGER, N.L. Cavitating performance of two low-area-ratio jet pumps having throat
lengths of 7.25 diameters. NASA tech.Note D-4592, 1968.
20. SANGER, N.L. Noncavitating and cavitating performance of two low-area-ratio water
jet pumps with throat lengths of 5.66 diameters. NASA tech. Note
D-4759, 1968.
21. SILVESTER, R. Design data for the liquid-liquid jet pump. J. Hydraulic Research,
MUELLER, N.H.G. Vol.6. No.2, pp.129-162, 1968.
22. CAIRNS, J.R. Optimum design of water jet pumps. Paper 68-WA/FE-13, Am. Soc.
NA, T.Y. mech. Engrs, 1969.
23. CUNNINGHAM, R.G. Jet pump cavitation. Paper 69-WA/FE-29, Am. Soc. mech. Engrs,
HANSEN, A.G. 1969.
NA, T.Y.
24. SANGER, N.L. Noncavitating and cavitating performance of several low-area-ratio
water jet pumps having throat lengths of 3.54 diameters. NASA tech.
Note D-5095, 1969.
25. SANGER, N.L. An experimental investigation of several low-area-ratio water jet
pumps. J. basic Engng, Vol.92, No.1, pp.11-20, 1970.
26. SANGER, N.L. Fortran programs for the design of liquid-to-liquid jet pumps. NASA
tech. Note D-6453, 1971.
27. KENTFIELD, J.A.C. The prediction of the optimum performance of ejectors. Proc. Inst.
BARNES, R.W. mech. Engrs, Thermodynamics Fluid Mech. Group, Vol.186, Pt 54/72,
pp.671-681, 1972.
28. SILVESTER, R. The jet pump using liquids of different density. Paper 11, Symp. "Jet
VONGVISESSOMJAI, S. Pumps and Ejectors", Brit. Hydromechanics Res. Assoc., Cranfield,
UK, 1972.
29. VYAS, B.D. Standardisation of water jet pumps. Paper 10, Symp. “Jet Pumps and
KAR, S. Ejectors”, Brit. Hydromechanics Res. Assoc., Cranfield, UK, 1972.
30. WAKEFIELD, A.W. Practical solids-handling jet pumps. Paper 12, Symp. “Jet Pumps and
Ejectors”, Brit. Hydromechanics Res. Assoc., Cranfield, UK, 1972.
31. CUNNINGHAM, R.G. Liquid jet pump modelling; effects of axial dimensions on
theory-experiment agreement, Paper Fl, 2nd Symp. “Jet Pumps and
Ejectors”, Brit. Hydromechanics Res. Assoc., Cranfield, UK, 1975.

50
 85032
32. WAKEFIELD, A.W. Performance of solids-handling jet pumps at low Reynolds numbers.
Paper A3, 2nd Symp. “Jet Pumps and Ejectors”, Brit. Hydromechanics
Res. Assoc., Cranfield, UK, 1975.
33. WAKEFIELD, A.W. Jet pump set to take off. Coal, Gold and Base Minerals of Southern
Africa, 1980.
34. GOVATOS, G.C. The slurry jet pump. J. Pipelines, Vol.1, pp.145-157, 1981.
35. WAKEFIELD, A.W. Hydraulic jet pumps for solids handling, Coal, Gold and Base Minerals
of Southern Africa, 1981.
36. WAKEFIELD, A.W. Private communication from Mr. A.W. Wakefield, Wakefield and
Imberg, Stamford, UK, 1985.

10.2 References

The references are recommended sources of information supplementary to the information in this Item.

37. MOODY, L.F. Friction factors for pipe flow. Trans. am. Soc. mech. Engrs, Vol.66,
pp.671-684, 1944.
38. GREY, R.E. Performance of conical jet nozzles in terms of flow and velocity
WILSTED, H.D. coefficients. NACA Rep. 933, Nat. adv. Comm. Aeronautics, USA, 1949.
39. ESDU Introduction to design and performance data for diffusers. Item
No.76027, ESDU International Ltd, London, November 1976.
40. ESDU Friction losses for fully-developed flow in straight pipes. Item No.66027,
with Amendments A and B, ESDU International Ltd, London, April
1977.
41. MILLER, D.S. Internal flow systems. Vol.5, Fluid Engineering Series, British
Hydromechanics Res. Assoc., Cranfield, UK, 1978.
42. ESDU Performance of conical diffusers in incompressible flow. Item No.73024,
ESDU International Ltd, London, March 1980.
43. BSI Specification for square-edged orifice plates, nozzles and venturi tubes
inserted in circular cross section conduits running full. BS 1042: Section
1.1, British Standards Inst., London, 1981.
44. ESDU Ejectors and jet pumps. Design and performance for compressible air
flow. Item No.84029, ESDU International Ltd, London, December 1984.

51
 85032
11. TABLES

TABLE 11.1 Jet Pump Design Parameters - Calculation Sheet

Parameter Relationship Value

Primary pressure P 1 = [ P5 ( 1 + N ) –P2 ] ⁄ N

Secondary pressure P 2 = P 5 – N ( P1 – P5 )

Discharge pressure P 5 = ( NP 1 + P 2 ) ⁄ ( N + 1 )

Pressure ratio N = ( P5 – P2 ) ⁄ ( P1 – P5 )

Primary flowrate Qp = Qs ⁄ M

Secondary flowrate Q s = M × Qp

Flow ratio M = Qs ⁄ Qp

Mixing chamber diameter D = L ⁄ 7 or L ⁄ 5

Mixing chamber length L = 7D or 5D

Primary nozzle diameter d=D R

Primary nozzle spacing s≈d

Diffuser included angle 2φ

Primary fluid density ρp

Secondary fluid density ρs

Density ratio C = ρs ⁄ ρp
2
Cavitation index σ = ( P 2 – p v ) ⁄ ( ½ρ s V 3c )

Primary nozzle loss coefficient Kp

Secondary inlet loss coefficient Ks

Mixing chamber loss coefficient Km

Diffuser loss coefficient Kd

52
 85032
TABLE 11.2 Test Primary Nozzle Description (see Figure 2a)

Source R value d (mm) Nozzle profile

Derivation 1 0.25 7.9 Conical inlet to
parallel section
(effect of 180°
bend included in
Kp )

Derivation 4 0.10 2.54 Elliptical

0.133 3.58 Elliptical
0.174 2.54 Conical
0.20 3.58 Elliptical
0.30 4.39 Elliptical
0.40 3.58 Elliptical
0.54 4.50 Conical
0.60 4.39 Elliptical

Derivation 10 – 10.7 Conical

– 10.7 Elliptical
– 10.7 Circular inlet to
parallel section

Derivation 16 – Varying Conical

Derivation 20 0.066 8.8 Circular arc, radius

175 mm
0.197 15.2 Circular arc, radius
203 mm

53
 85032
TABLE 11.2 Test Primary Nozzle Description (see Figure 2a)

0.108 11.3 Circular arc, radius

190 mm
0.141 12.8 Circular arc, radius
190 mm
Derivation 27 – Varying Circular arc inlet,
radius 0.8d , to
parallel section,
length
5d

TABLE 11.3 Secondary Flow Inlet |Loss Coefficients

Source Inlet profile K s value

Derivation 1 Conical 0.05
Derivation 15 Conical 0.05
Derivation 17 Bellmouth 0.108
Derivation 30 Bellmouth 0.04

54
 
0.35

Area ratio R = 0.174

0.30
Results approximate

0.25

0.20
Pressure ratio N

0.15
55

Rep

~ 105
0.10 10340
3420
2480

0.05
2120

Rep
437 1720
0.00

85032
00.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

Flow ratio M

FIGURE 1 EFFECT OF VISCOSITY ON JET PUMP CHARACTERISTIC CURVE

 
1.7
Reference 4
Reference 10
Reference 19
1.6 Reference 22
Reference 31

Approximate recommended value

1.5

1.4
} Poor nozzle design, constricted flow path

1 + Kp
56

1.3

1.2

1.1

85032
1.0 2 5 8 2 5 8 2 5 8 5 8 5 8
2 2
102 103 104 105 106 107

Re

FIGURE 2a VARIATION OF Kp WITH REYNOLDS NUMBER

 
(1+Ks)
1.4

Approximate recommended value

1.3

1.2

1.1
57

1.0

0.9
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

Secondary inlet Reynolds number ( V3 (D - d) / νs) + 105

85032
FIGURE 2b VARIATION OF Ks WITH REYNOLDS NUMBER
 
f
0.002 0.0025 0.003 0.004 0.005 0.0075 0.01
0.30
L/D
0.28 7.0
Material Roughness value ε (mm)
Drawn tubing 0.0015
0.26 Commercial steel 0.045
Wrought iron 0.045
0.24 Galvanised metals 0.15
Cast metals 0.200

0.22

0.20 5.0
Mixing duct loss coefficient, Km

0.18

0.16

0.14
58

0.12

0.10

0.08
L/ D
0.06
7.0

0.04 5.0

0.02

85032
0.00 2 3 4 5 6 7 8 2 3 4 5 6 7 8 2 3 4 5 6 7 8 2 3 4 5 6 7 8 2 3 4 5 6 7 8
10-5 10-4 10-3 10-2 10-1 100

ε/D

FIGURE 2c MIXING DUCT LOSS COEFFICIENT VERSUS DUCT RELATIVE ROUGHNESS

 85032
 1 
M =  −1
 r
Computed results derived using:
1 + Kp = 1.04
1 + Ks = 1.10
1 + Km + Kd = 1.25 η
M N C = 1.0
4.0 1.6 0.40

3.5 1.4 0.35

3.0 1.2 0.30

2.5 1.0 0.25

2.0 0.8 0.20

1.5 0.6 0.15

η
1.0 0.4 0.10

0.5 0.2 0.05

M
0 0.0 0
00.0 0.2 0.4 0.6 0.8 1.0

Optimum area ratio R

FIGURE 3a VARIATION OF OPTIMUM AREA RATIO WITH FLOW AND PRESSURE RATIOS, FOR A
DENSITY RATIO OF 1.0

59
 85032
4.5 0.45

4.0 0.40

3.5 0.35

3.0 0.30

C
0.50
2.5 1.00 0.25
1.25
M
1.50 η

2.00
2.0 0.20

1.5 0.15

1.0 0.10

C η
0.50
0.5 1.00 0.05
1.25
1.50
2.00
M
0.0 0
0 0.0 0.2 0.4 0.6 0.8 1.0

Optimum area ratio R

FIGURE 3b EFFECT OF DENSITY RATIO ON VARIATION OF OPTIMUM AREA RATIO WITH FLOW
RATIO

60
 85032
Pressure ratio, N
1.6

1.4
0.40
0.35
1.2
1 + Km + Kd = 1.1
1 + Ks = 1.1
1.0 0.30 1 + Kp = 1.04
C = 1.0
0.8
0.25

0.6
0.20

0.4 0.15
0.10
0.2
0.05
R
0.0
0 1 2 3 4 5 6 7 8

Pressure ratio, N Flow ratio, M

R
0.45
0.85

0.80

0.75

0.70

0.65

0.60

0.55

0.50

0
0.0 0.2 0.4 0.6 0.8 1.0

Flow ratio, M

FIGURE 4a PERFORMANCE PREDICTION CURVE

61
 85032
Pressure ratio, N
1.6

1.4
0.40
0.35
1.2
1 + Km + Kd = 1.14
1 + Ks = 1.1
1.0 0.30 1 + Kp = 1.04
C = 1.0
0.8
0.25

0.6
0.20

0.4 0.15
0.10
0.2
0.05
R
0.0
0 1 2 3 4 5 6 7 8

Flow ratio, M

Pressure ratio, N

0.45
0.60

0.55
0.65

0.50
0.85
0.80

0.75

0.70

0
0.0 0.2 0.4 0.6 0.8 1.0

Flow ratio, M

FIGURE 4b PERFORMANCE PREDICTION CURVE

62
 85032
Pressure ratio, N
1.2

0.40

0.35
1.0
1 + Km + Kd = 1.18
1 + Ks = 1.1
0.30
1 + Kp = 1.04
C = 1.0
0.8

0.25

0.6

0.20

0.4
0.15

0.10
0.2

0.05
R

0.0
0 1 2 3 4 5 6 7 8

Flow ratio, M
Pressure ratio, N

R
0.45
0.80

0.55
0.85

0.65
0.75

0.70

0.60

0.50

0
0.0 0.2 0.4 0.6 0.8 1.0

Flow ratio, M

FIGURE 4c PERFORMANCE PREDICTION CURVE

63
 85032

Pressure ratio, N

1.2

0.40
1 + Km + Kd = 1.22
1.0
0.35
1 + Ks = 1.1
1 + Kp = 1.04

0.30
C = 1.0

0.8

0.25

0.6

0.20

0.4
0.15

0.10
0.2

0.05

0.0
0 1 2 3 4 5 6 7 8

Flow ratio, M
Pressure ratio, N
4

R
0.45
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50
0
0.0 0.2 0.4 0.6 0.8

Flow ratio, M

FIGURE 4d PERFORMANCE PREDICTION CURVE

64
 85032

Pressure ratio, N

1.2

0.40
1 + Km + Kd = 1.26
1.0
0.35
1 + Ks = 1.1
1 + Kp = 1.04
C = 1.0
0.30

0.8

0.25

0.6

0.20

0.4
0.15
Pressure ratio, N
3.5
0.10
0.2

0.05
3.0
R

0.0
0 1 2 3 4 5 6 7
2.5
Flow ratio, M

2.0

1.5

1.0

0.5

R
0.45
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4e PERFORMANCE PREDICTION CURVE

65
 85032

Pressure ratio, N

1.2

0.40
1 + Km + Kd = 1.3
1.0
0.35
1 + Ks = 1.1
1 + Kp = 1.04
C = 1.0
0.30

0.8

0.25

0.6

0.20

0.4
0.15
Pressure ratio, N
3.5
0.10
0.2

0.05
3.0
R

0.0
0 1 2 3 4 5 6 7
2.5
Flow ratio, M

2.0

1.5

1.0

0.5

R
0.45
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4f PERFORMANCE PREDICTION CURVE

66
 85032

Pressure ratio, N

1.2

1.0
0.40 1 + Km + Kd = 1.34
1 + Ks = 1.1
0.35 1 + Kp = 1.04

0.8 C = 1.0
0.30

0.6

0.25
0.20

0.4
0.15

Pressure ratio, N 0.10

0.2
2.5
0.05
R

0.0
0 1 2 3 4 5 6 7
2.0 Flow ratio, M

1.5

1.0

0.5

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4g PERFORMANCE PREDICTION CURVE

67
 85032
Pressure ratio, N

1.2

1.0
0.40 1 + Km + Kd = 1.38
1 + Ks = 1.1
0.35 1 + Kp = 1.04

0.8 C = 1.0
0.30

0.6

0.25

0.4 0.20
0.15

0.10
0.2
Pressure ratio, N
0.05
R

0.0
0 1 2 3 4 5 6 7
2.0 Flow ratio, M

1.5

1.0

0.5

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4h PERFORMANCE PREDICTION CURVE

68
 85032

Pressure ratio, N

1.2

1.0
0.40 1 + Km + Kd = 1.42
1 + Ks = 1.1
0.35 1 + Kp = 1.04

0.8 C = 1.0
0.30

0.6

0.25

0.4 0.20
0.15

0.10
0.2
Pressure ratio, N
0.05
R

2.0 0.0
0 1 2 3 4 5 6

Flow ratio, M

1.5

1.0

0.5

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4i PERFORMANCE PREDICTION CURVE

69
 85032

Pressure ratio, N
1.2

0.40

1.0 0.35
1 + K m + K d = 1.46
1 + K s = 1.1
0.30 1 + K p = 1.04
0.8 C = 1.0

0.6

0.25

0.4 0.20

0.15

0.2 0.10

0.05
R

0.0
0 1 2 3 4 5 6 7

Pressure ratio, N Flow ratio, M

2.0

1.6

1.2

0.8

0.4

R
0.85

0.65

0.60

0.55
0.80

0.75

0.70

0.50

0.45
0.0
0.0 0.2 0.4 0.6 0.8 1.0

Flow ratio, M

FIGURE 4j PERFORMANCE PREDICTION CURVE

70
 85032

Pressure ratio, N
1.6

1.4
0.4 0

1.2 0.3 5
1 + K m + K d = 1.1
1.0 1 + K s = 1.2
0.3 0
1 + K p = 1.04
0.8 C = 1.0
0.2 5
0.6
0.2 0

0.4 0.1 5

0.1 0
0.2
0.0 5
R
0.0
0 1 2 3 4 5 6 7 8

Pressure ratio, N Flow ratio, M

7

R
0.8 5

0.8 0

0.7 5

0.5 0
0.7 0

0.6 5

0.6 0

0.5 5

0.4 5
0
0.0 0.2 0.4 0.6 0.8 1.0

Flow ratio, M

FIGURE 4k PERFORMANCE PREDICTION CURVE

71
 85032
Pressure ratio, N

1.4

0.40
1.2
1 + Km + Kd = 1.14
1 + Ks = 1.2
0.35 1 + Kp = 1.04

1.0 C = 1.0

0.30

0.8

0.25

0.6

0.20

Pressure ratio, N 0.4

0.15
5.5

0.10
5.0 0.2
0.05
4.5 R

0.0
4.0 0 1 2 3 4 5 6 7

Flow ratio, M
3.5

3.0

2.5

2.0

1.5

1.0

R
0.5
0.45
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4l PERFORMANCE PREDICTION CURVE

72
 85032
Pressure ratio, N
1.4

1.2
0.40
1 + Km + Kd = 1.18
0.35
1 + Ks = 1.2
1.0 1 + Kp = 1.04
C = 1.0
0.30

0.8

0.25

0.6

0.20

0.4
Pressure ratio, N 0.15

4.5
0.10
0.2
4.0
0.05
R

3.5 0.0
0 1 2 3 4 5 6 7

Flow ratio, M
3.0

2.5

2.0

1.5

1.0

0.5
R
0.45
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4m PERFORMANCE PREDICTION CURVE

73
 85032
Pressure ratio, N
1.4

1.2

0.40
1 + Km + Kd = 1.22
1 + Ks = 1.2
0.35
1.0 1 + Kp = 1.04
C = 1.0
0.30

0.8

0.25

0.6

0.20

0.4
0.15

0.10
0.2

0.05

R
0.0
0 1 2 3 4 5 6 7

Flow ratio, M
Pressure ratio, N
4.0

3.5

3.0

2.5

2.0

1.5

1.0

0.5
R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4n PERFORMANCE PREDICTION CURVE

74
 85032
Pressure ratio, N
1.4

1.2

0.40
1 + Km + Kd = 1.26
1 + Ks = 1.2
0.35
1.0 1 + Kp = 1.04
C = 1.0

0.30

0.8

0.25

0.6

0.20

0.4
Pressure ratio, N 0.15

3.5
0.10
0.2

3.0 0.05
R

0.0
0 1 2 3 4 5 6 7
2.5
Flow ratio, M

2.0

1.5

1.0

0.5

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4o PERFORMANCE PREDICTION CURVE

75
 85032

Pressure ratio, N
1.4

1.2

0.40 1 + Km + Kd = 1.3
1 + Ks = 1.2
0.35
1.0 1 + Kp = 1.04
C = 1.0

0.30

0.8

0.25

0.6

0.20

Pressure ratio, N 0.4

0.15

3.0
0.10
0.2

0.05
2.5
R

0.0
0 1 2 3 4 5 6 7
2.0 Flow ratio, M

1.5

1.0

0.5

R
0.45
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4p PERFORMANCE PREDICTION CURVE

76
 85032
Pressure ratio, N

1.2

0.40
1.0 1 + Km + Kd = 1.34
1 + Ks = 1.2
0.35
1 + Kp = 1.04

0.30
C = 1.0
0.8

0.6

0.25

0.20
0.4

0.15
Pressure ratio, N
0.10
0.2

0.05
2.4
R
0.0
0 1 2 3 4 5 6 7
2.0 Flow ratio, M

1.6

1.2

0.8

0.4

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4q PERFORMANCE PREDICTION CURVE

77
 85032

Pressure ratio, N

1.2

0.40
1.0 1 + Km + Kd = 1.38
1 + Ks = 1.2
0.35
1 + Kp = 1.04
C = 1.0
0.8 0.30

0.6
0.25

0.20
0.4

0.15

0.2 0.10
Pressure ratio, N

2.4
0.05
R
0.0
0 1 2 3 4 5 6 7
2.0 Flow ratio, M

1.6

1.2

0.8

0.4

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4r PERFORMANCE PREDICTION CURVE

78
 85032
Pressure ratio, N

1.2

0.40
1.0 1 + Km + Kd = 1.42
1 + Ks = 1.2
0.35
1 + Kp = 1.04
C = 1.0
0.8 0.30

0.6
0.25

0.20
0.4

0.15

0.2 0.10

0.05

Pressure ratio, N R

0.0
0 1 2 3 4 5 6
2.0 Flow ratio, M

1.6

1.2

0.8

0.4

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4s PERFORMANCE PREDICTION CURVE

79
 85032
Pressure ratio, N
1.2

1.0
0.40
1 + K m + Kd = 1.46
0.35
1 + K s = 1.2
0.8 1 + Kp = 1.04
0.30 C = 1.0

0.6

0.25

0.4 0.20
0.15

0.10
0.2

0.05
R

0.0
0 1 2 3 4 5 6 7

Flow ratio, M

Pressure ratio, N
2.0

1.6

1.2

0.8

0.4

R
0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8

Flow ratio, M

FIGURE 4t PERFORMANCE PREDICTION CURVE

80
 
1.6
1 + Kp 1.02
1.04

1.4
1.08

1.10

1.2
1.14
O ptim um pressure ratio

1.0 1.20

1.30
0.8
1 + Kp
81

0.6

0.4
1 + K s = 1.1
1 + K m + K d = 1.25
C = 1.0
0.2

0.0

85032
0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

A rea R atio R

FIGURE 5 EFFECT OF Kp ON VARIATION OF OPTIMUM PRESSURE RATIO WITH AREA RATIO

 
1.0
Calculated for C = 1.25

1
R ( M values for C = 1 ) × , N values unchanged
125
.
0.8 0.3

0.6

N
0.2
82

0.4

0.1
0.2
0.05

0.0
0 0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0

85032
M

FIGURE 6a EFFECT OF DENSITY RATIO

 
3.2

2.8

2.4

C alculated for C = 0.5

2.0 1
( M values for C = 1.0 ) × , N values u nchanged
0 .5

N
1.6
83

1.2

0.8

0.4

R 0.8 0.7 0.6 0.5

0.0

85032
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

FIGURE 6b EFFECT OF DENSITY RATIO

 85032

1.0
a. Results of first iteration using
straight line approximation

Cavitation limit
0.8 40

Efficiency η (%)
0.6 30
N
Pressure ratio N

0.4 20
Operating

0.2 10
Point

0.0 0
00.0 0.4 0.8 1.2 1.6 2.0

Flow ratio M

FIGURE 7a RESULTS FOR EXAMPLE 1

84
 85032

1.0
b. Results of second iteration using
straight line approximation

Cavitation limit
0.8 40

Efficiency η (%)
Pressure ratio N

0.6 30
N

0.4 20
Operating

0.2 10
Point

0.0 0
0 0.0 0.4 0.8 1.2 1.6 2.0

Flow ratio M

FIGURE 7b RESULTS FOR EXAMPLE 1

85
 85032

1.0
c. Results of second iteration using
Equation (5.6)

0.8 40

(%)
Pressure ratio N

0.6 30

Efficiency η
N

0.4 20
Operating

0.2 10
Point

0.0 0
0 0.0 0.4 0.8 1.2 1.6 2.0

Flow ratio M

FIGURE 7c RESULTS FOR EXAMPLE 1

86
 85032

22.3 mm D = 45 mm 101 mm

L = 315 mm 534 mm

FIGURE 7d OUTLINE OF FINAL DESIGN CONFIGURATION

87
 85032

210
P5

206 6.5
Discharge pressure P5 ( kPa )
d

(mm)
202 6.0

d
Primary nozzle exit diameter
198 5.5

194 5.0

190 4.5
0.04 0.05 0.06 0.07

Area Ratio R

0.8

Qp
( kg/s )

0.7

23
Qp

η
(%)
Primary flow rate

0.6
η

22
Efficiency

0.5

21

0.4
0.04 0.05 0.06 0.07

Area Ratio R

FIGURE 8 EXAMPLE 2. BESPOKE DESIGN CURVES

88
 85032

a. Performance prediction curve

0.10

0.08 20
Pressure ratio N

0.06 15
η

0.04 10

Efficiency η (%)
0.02 5

0.00 0
00 1 2 3 4 5 6 7

Flow ratio M

FIGURE 9a RESULTS FOR EXAMPLE 3. PERFORMANCE PREDICTION CURVE

89
 85032
120

100

80
Discharge pressure P5 (k Pa)

60

40

Operating range

20

0
00.0 0.2 0.4 0.6 0.8 1.0 1.2

Secondary mass flow (kg/s)

FIGURE 9b RESULTS FOR EXAMPLE 3. ENGINE OPERATING CURVE

90
 85032
APPENDIX AGLOSSARY OF TERMS

Cavitation-limited flow - an operational condition in which the secondary flow rate cannot increase due
to the occurrence of cavitation within the jet pump, usually at the mixing chamber entrance. The condition
is generally to be avoided.

Cavitation index - a number used to predict the onset of cavitation-limited flow.

Driving fluid or (flow) - see primary flow.

Driving nozzle - see primary nozzle.

Ejector - a device in which the kinetic energy of one fluid is used to drive another fluid.

Head ratio - the ratio of the net output head to the net input head, usually denoted by the symbol N . The
head ratio is the same as the pressure ratio when the primary and secondary fluids are of equal density.

Induced fluid (or flow) - see secondary fluid.

Injector - an alternative name for an ejector.

Jet pump - an alternative name for an ejector. Often used when the working fluids are liquids.

Flow ratio - the ratio of the secondary fluid volume flow rate to that of the primary fluid. It is the same as
the mass flow ratio when the working fluids are of the same density, and is often denoted by the same
symbol, M .

Mixing chamber - the duct, usually cylindrical, in which the turbulent mixing of the primary and secondary
flows takes place.

Motive nozzle and motive flow - see primary nozzle and primary flow.

Pressure ratio - the ratio of secondary flow pressure rise to primary flow pressure drop, usually denoted
by the symbol N .

Primary fluid (or flow) - the liquid that is input into the ejector with high energy, generally through a nozzle.

Primary nozzle - the nozzle through which the primary flow enters the jet pump.

Secondary fluid (or flow) - the fluid that is drawn into the ejector with low energy and is entrained by the
primary fluid.

91
 85032
APPENDIX BBASIC THEORY

B1. Assumptions

1. Both primary and secondary fluids are liquids and therefore incompressible.
2. The flow is one-dimensional except within the mixing chamber.
3. The primary nozzle exit plane coincides with the mixing chamber entrance plane.
4. The nozzle wall thickness at the exit is zero.
5. The mixing chamber has constant cross-sectional area.
6. Mixing (momentum transfer) is completed within the mixing chamber prior to the diffuser
entry.

B2. Losses

Losses in the different parts of the ejector are accounted for by dimensionless loss coefficients. Losses
occurring beyond the points just upstream of the primary nozzle and secondary inlet and just downstream
of the diffuser are considered as part of the installation to which the pump is attached.

Primary nozzle

The loss coefficient can be expressed in terms of the pressure difference across the nozzle;

P 1 – pn
K p = -------------------- – 1 . (B2.1)
2
½ρ p V n

Secondary inlet

The loss may be treated in a similar manner to give

P2 – p 3
K s = -------------------- – 1 . (B2.2)
2
½ ρs V 3

Mixing Chamber

The mixing chamber loss coefficient is defined as

K m = 4f L ⁄ D , (B2.3)

where f is the usual pipe friction factor, ∆p ⁄ [ 4 ( L ⁄ D )½ρ V 2 ] .

The velocity at the mixing chamber exit, V 4 , is used in the expression for friction loss.

92
 85032
Although this is lower than the average velocity in the mixing chamber, the friction loss in the mixing
chamber has been found to be much less than in conventional pipe flow (see Derivation 17). The higher
velocities in the mixing chamber are not found near the wall and the pressure gain due to mixing may also
have an effect.

Diffuser

The loss coefficient in the diffuser is defined as

P4 – P 5
K d = -------------------- . (B2.4)
2
½ρ d V 4

However, the diffuser performance is often quoted in terms of a static pressure recovery coefficient, C pr ,
where

p5 – p 4
C pr = -------------------- . (B2.5)
2
½ρ d V 4

The diffuser loss coefficient may be expressed in terms of C pr as

2
 A4
K d = 1 –  ------  – C pr . (B2.6)
 A5

B3. Total Pressure Equations

By consideration of conservation of energy, the total pressure at the primary nozzle, the secondary inlet
and the diffuser exit may be written as;

2
P 1 = p 3 + ½ ρ p V n + primary nozzle pressure loss, (B3.1)
2
P 2 = p 3 + ½ ρ s V 3 + secondary flow inlet pressure loss, (B3.2)
2
P 5 = p 4 + ½ ρ d V 4 – diffuser pressure loss, (B3.3)
2
where primary nozzle pressure loss = K p ( ½ ρp V n ) , (B3.4)
2
secondary flow inlet pressure loss = K s ( ½ρ s V 3 ) , (B3.5)
2
and diffuser pressure loss = K d ( ½ ρd V 4 ) . (B3.6)

93
 85032
Hence,

P 1 = p3 + ½ ρp V n 2 ( 1 + K p ) , (B3.7)
P 2 = p3 + ½ ρs V 3 3 ( 1 + K s ) , (B3.8)
P 5 = p4 + ½ ρd V 4 2 ( 1 – K d ) . (B3.9)

The equation of continuity for the ejector may be written as

ρ p Qp + ρ s Qs = ρ d Qd . (B3.10)

Using the definition of volume flow ratio, M = Q s ⁄ Q p , the velocity at the diffuser entrance, V 4 , and
the secondary fluid velocity at the entrance to the mixing chamber, V 3 , may be expressed in terms of the
primary nozzle exit velocity as

V 4 = R ( 1 + M ) Vn (B3.11)
MR
and . V 3 = -------------- V n . (B3.12)
1–R

Considering conservation of momentum in the mixing chamber,

p 3 A 3 + ρ p Q p V n + ρ s Q s V 3 = p 4 A 4 + ρ m Q m V 4 + mixing chamber friction loss.(B3.13)

The expressions for V 4 and V 3 may be substituted to give

V 2
MR 2 4fL n
( p – p )A = ρ A V 2 + ρ ( A – A )  --------------  V 2 – ρ A R 2 ( 1 + M ) 2 V 2 – ρ --------- R 2 ( 1 + M ) 2 ------- . (B3.14)
4 3 4 p n n s 3 n  1 – R n d 4 n d D 2

The total pressure difference between the diffuser exit and the secondary inlet may be written, utilising
Equations (B3.8) and (B3.9), as

2
P 5 – P 2 = p 4 – p 3 + ( 1 – K d )½ ρ d V 4 2 – ( 1 + K s )½ ρ s V 3 , (B3.15)

and incorporating Equation (B3.14) and the density ratio, C = ρ s ⁄ ρ p , will give

2 M 2 R2 MR 2
P –P = ------------------------ – R 2 ( 1 + K + K ) ( 1 + CM ) ( 1 + M ) – C ( 1 + K )  -------------- 
½ ρ p V n 2R + 2C . (B3.16)
5 2 1–R m d s  1 – R

Similarly, the total pressure difference, P 1 – P 5 , may be written from Equations (B3.7) and (B3.8) as

P – P = p – p + ( 1 + K ) ½ ρ V 2 – ( 1 – K )R 2 ( 1 + M ) ( 1 + CM ) ½ ρ V 2 . (B3.17)
1 5 3 4 p p n d p n

94
 85032
Again incorporating Equation (B3.14) will give

M 2 R2
P –P =
1 5
½ ρ p V n 2 – 2 R – 2C
------------------------ + R 2 ( 1 + M ) ( 1 + CM ) ( 1 + K m + K d ) + ( 1 + K p ) . (B3.18)
1–R

B4. Efficiency

The efficiency of a jet pump is often relatively unimportant, since in many applications it is used in
conjunction with another device whose efficiency is more critical. Computation of efficiency is still useful,
however, for comparing one pump design with another to accomplish the same results.

The efficiency of any device is usually defined as the ratio of the total energy put in to the total energy
extracted. If this concept is applied to the jet pump, then the basic efficiency can be said to be

( Q p + Q s ) × ( P 5 – P ref )
--------------------------------------------------------------------------------- . (B4.1)
Q p ( P 1 – P ref ) + Q s ( P 2 – P ref )

A jet pump is often used in the type of system illustrated in Sketch B4.1, in which the primary fluid is
recirculated.

Sketch B4.1 Schematic of jet pump system

When this is the case, an expression for efficiency may be derived by taking P 2 as the reference pressure.
Adding and subtracting P 5 to the denominator and expanding the numerator will then give the efficiency as

Qp ( P 5 – P2 ) + Q s ( P5 – P 2 )
-------------------------------------------------------------------------- . (B4.2)
Q p ( P 5 – P2 ) + Q p ( P1 – P 5 )

If the energy change of the recirculated fluid is neglected, i.e. the Q p ( P 5 – P 2 ) terms are omitted, the
resulting expression for efficiency is

Qs ( P 5 – P2 )
η = --------------------------------- . (B4.3)
Qp ( P 1 – P5 )

95
 85032
The definition of efficiency gives rise to a convenient form for jet pump pressure ratio,

P5 – P2
N = -------------------- . (B4.4)
P1 – P5

An alternative efficiency expression results from a slightly different procedure. Taking P 2 as the reference
pressure again, and simply neglecting the energy change of the recirculated fluid, represented by the
Q p ( P 5 – P 2 ) term in the numerator, will give an efficiency defined by

Q s ( P 5 – P2 )
η ′ = --------------------------------- . (B4.5)
Qp( P1 – P2)

Values of η ′ are always lower than values of η at the same flow ratio but the maximum value of η ′ occurs
at virtually the same M value as the maximum value of η . Hence, if the η ′ versus M curve is plotted on
the same axes as the η versus M curve, it will have the same end-points and the same shape but lie below
the η versus M curve.

This definition of efficiency gives rise to an alternative pressure ratio sometimes used, the ratio of the rise
in pressure of the secondary flow to difference between the primary and secondary pressures,

P5 – P 2
N ′ = -------------------- . (B4.6)
P1 – P 2

The relationship between the two pressure ratios is:

N = N′ ⁄ (1 – N′) . (B4.7)

These two efficiency definitions, particularly Equation (B4.3), are often used to describe the efficiency of
any jet pump, even where the primary fluid is not recirculating.

B5. Pressure Ratio

The pressure ratio, N , is defined by Equation (B4.4).

Substituting from Equations (B3.16) and (B3.18) yields;

2CM 2 R 2 C M 2 R2
2R + ------------------------ – R 2 ( 1 + CM ) ( 1 + M ) ( 1 + K + K ) – ---------------------- ( 1 + K )
1–R m d (1 – R)2 s
N = ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
-. (B5.1)
2CM 2 R 2
( 1 + K ) – 2R – ------------------------ + R 2 ( 1 + CM ) ( 1 + M ) ( 1 + K + K )
p 1–R m d

96
 85032
When the primary and secondary fluids are of equal density, i.e. C = 1 , this expression reduces to

2M 2 R 2 M2R2
2R + ------------------- – R 2 ( 1 + M ) 2 ( 1 + K m + K d ) – ---------------------- ( 1 + K s )
1–R ( 1 – R )2
-.
N = --------------------------------------------------------------------------------------------------------------------------------------------------------------------- (B5.2)
2M 2 R 2 2 2
( 1 + K ) – 2R – ------------------- + R ( 1 + M ) ( 1 + K + K )
p 1–R m d

B6. Primary Nozzle Exit Area

The total pressure equations may also be used to derive an expression for the primary nozzle exit area, A n .
From Equations (B3.7) and (B3.8),

P 1 – P2 = ½ ρ p V n 2 ( 1 + K p ) – ½ ρs V 3 2 ( 1 + K s ) . (B6.1)

Expressing the secondary inlet velocity in terms of the primary nozzle exit velocity, from Equation (B3.12),
and introducing the density ratio, C , gives

MR 2
P 1 – P 2 = ½ρ p V n 2 ( 1 + K p ) – C ( 1 + K s )  --------------  . (B6.2)
 1 – R

Introducing the primary nozzle exit area, this becomes

2
 Q p MR 2
P 1 – P 2 = ½ ρ p  -------  ( 1 + K p ) – C ( 1 + K s )  --------------  (B6.3)
 An  1 – R

and rearranging gives

MR 2 ½
( 1 + K p ) – C ( 1 + K s )  -------------- 
 1 – R
A n = Q p -------------------------------------------------------------------------------- . (B6.4)
( P 1 – P2 ) ⁄ ( ½ ρp )

B7. Cavitation

A cavitation prediction parameter, or index, may be derived by applying the principles of conservation of
momentum and energy to the secondary inlet.

Equation (B3.8) can be rearranged as

P2 – p 3
-------------------- = 1 + K s . (B7.1)
2
½ ρsV3

97
 85032
The cavitation index, σ , may be defined as

P 2 – pn
σ = -------------------- . (B7.2)
2
½ρ s V 3

If it is assumed that at the point when total pressure rise begins to break down, due to the onset of cavitation,
the static pressure in the plane of the primary nozzle exit is equal to the fluid vapour pressure, then at that
point the limiting value of σ is

P 2 – pv
σ = ---------------------- = 1 + K s
2
½ρ s V 3c

However, it should be noted that, strictly speaking, the secondary inlet loss coefficient is defined only for
non-cavitating conditions.

Expressing the secondary inlet velocity in terms of the primary nozzle exit velocity, from Equation (B3.12),
and introducing the density ratio, C , gives

1 1–R 2
σ = ( P 2 – p v ) ---------------  ------------------ , (B7.3)
½C ρ p  M c R V n 

where M c is the cavitation-limited value of flow ratio. Rearranging gives

1 – R P2 – p v ½
M c = -------------- ----------------------------- . (B7.4)
R ½ρ p V n 2 C σ

An alternative criterion describing cavitating conditions is the minimum secondary inlet pressure required
to prevent cavitation. This is derived as follows:

P 2 required = ½ρ s V 3 2 σ + p v (B7.5)

MR 2
= ½ρ p C σ  -------------- V n + p v
1 – R 
2
 Qp M R 
= ½ρ p C σ  ------- --------------  + p v . (B7.6)
 A n 1 – R

Comparison of the required P 2 with the available P 2 is necessary to determine cavitation-limited

conditions.

98
 85032
APPENDIX C: FLOWCHARTS

C1. FLOW CHART FOR DESIGN PROCEDURE (see Section 5.4)

Determine design criteria and insert values in Table 11.1. Stage1

Obtain values for the loss coefficients using Figures 2a to 2c and Stage 2
References 37 to 43.

Obtain a value for the optimum area ratio using Figures 3a and Stage 3
3b or Equation (5.6)
Check that N × M is less than 1.0.

Determine the form of the characteristic curve using Equations Stage 4

(5.15) to (5.19) or Equation (5.6) and establish operating point
or range.

Calculate the efficiency from pairs of M and N values and using Stage 5
Equation (5.4) plot the efficiency curve.

Complete the specification of the design parameters in Stage 6

Table 11.1.

Check the limit set by cavitation: either the maximum flow ratio Stage 7
or the minimum secondary pressure.

Assess design. Stage 8

inadequate adequate

Carry out the detailed design of the components. Stage 9

99
 85032
C2. FLOW CHART FOR PERFORMANCE PREDICTION METHOD (see Section 6.2)

Estimate values for the loss coefficients using Figures 2a to 2c Stage 1

and References 37 to 43. Calculate the density ratio.

Measure the primary nozzle exit diameter and the mixing Stage 2
chamber diameter and calculate the area ratio.

Select the appropriate performance prediction curve from Stage 3

Figures 4a to 4t and make adjustment for density ratio, if
necessary

or

plot the performance prediction curve using Equation (5.6).

Calculate the jet pump efficiency for each pair of M and N

Calculate or measure a value for the cavitation index, choose a Stage 4

suitable criterion, and calculate the limit set by cavitation using
Equations (5.21) and (5.22).

100
 85032
THE DEVELOPMENT OF THIS DATA ITEM

The work on this particular Item was monitored and guided by the following Working Party:

Mr J.M. McAvoy – National Engineering Laboratory

Mr J.W.E. Campbell – Babcock Power Ltd
Dr D.J. Cockrell – University of Leicester
Prof. R. Sylvester* – University of Western Australia
Dr R.S. Sylvester – British Hydromechanics Research Association
Mr A.W. Wakefield – Wakefield and Imberg
Mr A.R. Whitaker – British Aerospace (Warton)

on behalf of the Internal Flow Panel which has the following constitution:

Chairman
Mr N.G. Worley – Babcock Power Ltd

Members
Dr T.W. Broyd – Atkins Research and Development
Mr D.A. Campbell – Rolls-Royce Ltd, Derby
Mr J. Campbell – Ove Arup and Partners
Dr D.J. Cockrell – University of Leicester
Dr J.A. Eaton – GEC Power Engineering
Prof. D.H. Freeston* – Auckland University, New Zealand
Dr G. Hobson – GEC Turbine Generators
Prof. J.L. Livesey – University of Salford
Dr A. Moore – British Hydromechanics Research Association.

The work on this Item was carried out as part of the programme for the Internal Flow Group of ESDU. The
initial assessment of the available information and the subsequent development of the Item was undertaken
by:

Ms A.E. Gibson – Atkins Research and Development, Epsom, Surrey, UK

Dr S.J. Murray – Atkins Research and Development, Epsom, Surrey, UK

and sponsored by ESDU. The member of ESDU staff concerned was Mr B.C.Freeman.

*
Corresponding Member

101