Waterjet Propulsion 5 1112 December 2008 Royal

Institution Of Naval Architects download

https://ebookbell.com/product/waterjet-
propulsion-5-1112-december-2008-royal-institution-of-naval-
architects-4729326

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Versatility Of Waterjet Technology Multimode Machining Of Most

Materials From Macro To Microscale Peter Ht Liu

https://ebookbell.com/product/versatility-of-waterjet-technology-
multimode-machining-of-most-materials-from-macro-to-microscale-peter-
ht-liu-57677670

Abrasive Water Jet Machining Of Engineering Materials 1st Ed 2020

Jagadish

https://ebookbell.com/product/abrasive-water-jet-machining-of-
engineering-materials-1st-ed-2020-jagadish-10801660

Advances In Water Jetting Selected Papers From The International

Conference On Water Jet 2019 Research Development Applications
November 2022 2019 Eladn Czech Republic 1st Ed Dagmar Klichov

https://ebookbell.com/product/advances-in-water-jetting-selected-
papers-from-the-international-conference-on-water-jet-2019-research-
development-applications-november-2022-2019-eladn-czech-republic-1st-
ed-dagmar-klichov-22501782

Geomechanics 93 Strata Mechanics Numerical Methodswater Jet Cutting

Rakowski Zikmund Rakowski

https://ebookbell.com/product/geomechanics-93-strata-mechanics-
numerical-methodswater-jet-cutting-rakowski-zikmund-rakowski-7160160
Vibrations In The Production System Measurement And Analysis With
Water Jet Technology 1st Ed Tefnia Olejrov

https://ebookbell.com/product/vibrations-in-the-production-system-
measurement-and-analysis-with-water-jet-technology-1st-ed-tefnia-
olejrov-7320636

Watermen Of Reedville And The Chesapeake Bay Shawn Hall

https://ebookbell.com/product/watermen-of-reedville-and-the-
chesapeake-bay-shawn-hall-51848074

Fields Watered With Blood Critical Essays On Margaret Walker Maryemma

Graham

https://ebookbell.com/product/fields-watered-with-blood-critical-
essays-on-margaret-walker-maryemma-graham-5047586

The Watermen Selections From Chesapeake James A Michener

https://ebookbell.com/product/the-watermen-selections-from-chesapeake-
james-a-michener-50549452

The Watermen The Birth Of American Swimming And One Young Mans Fight
To Capture Olympic Gold Loynd

https://ebookbell.com/product/the-watermen-the-birth-of-american-
swimming-and-one-young-mans-fight-to-capture-olympic-gold-
loynd-55341228
 RINA
WATERJET PROPULSION 5
11 - 12 December 2008

© 2008: The Royal Institution of Naval Architects

The Institution is not, as a body, responsible for the opinions expressed by the individual authors or
speakers

THE ROYAL INSTITUTION OF NAVAL ARCHITECTS

10 Upper Belgrave Street
London SW1X 8BQ

Telephone: 020 7235 4622

Fax: 020 7259 5912

ISBN No: 978-1-905040-50-6

Waterjet Propulsion 5, London, UK

CONTENTS
Waterjet Applications in Vessels that Operate in Multiple Modes
D Borrett and P Rae, Hamilton Jet, New Zealand
Off-Design Behaviour of Waterjets
N Bulten and R Verbeek, Wärtsilä Propulsion Netherlands BV, The Netherlands
Research on the Optimum Blades Number of Mixed Flow Pump Based on CFD
Qiongfang Yang, Yongsheng Wang, Zhihong Zhang, and Mingmin Zhang, Naval
University of Engineering, China
A Multi-Objective Automatic Optomisation Stratergy for Design of Waterjet
Pumps
M Zangeneh, University College London, UK
Waterjet Pump Development for High Performance and Higher Power Density
R Aartojärvi and M Heder, Rolls-Royce AB, Sweden
Research on Hydrodynamic Performance of Hybrid Propulsion System
C Sun, Y Wang and Z Li, University of Naval Engineering, China
Testing Air-Augmented Waterjet Propulsion
A Gany, A Shemer, A Gofer and D Har-Lev, Faculty of Aerospace Engineering,
Technion – Israel Institute of Technology, Haifa, Israel
Simulation of Dynamic Characteristics of Waterjet and its Application on
Troubleshooting
Yongsheng Wang and Jiangming Ding, University of Naval Engineering, China;
Zhirong Liu and Wenshan Xu, Guangzhou Marine Engineering Corp., China
Numerical Simulation and Analysis of Cavitation Performance of Waterjet
Chengjiang Liu, Yongsheng Wang and Zhihong Zhang, Naval University of
Engineering, China
Overview of Recent Developments in Testing of Waterjets at NSWCCD
M Donnelly and S Gowing, Naval Surface Warfare Center, Carderock Division,
W.Bethesda, MD, USA
Active Rudder Control
Nils Morén, MJP Marine Jet Power AB, Sweden
Stanislav Pavlov, MTD Company, Ltd., Russia
Scaling of Waterjet Propulsor Inlet Wakes
M B Wilson, Naval Surface Warfare Center, Carderock Division (NSWCCD), USA
Calculation and Analysis for Vortex-Induced Vibration of Waterjet Grid
Shuping Chang, Yongsheng Wang and Zhiyang Pang, Naval University of
Engineering, China
Wenshan Xu, Guangzhou Marine Engineering Corporation
Authors’ Contact Details

© 2008: The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

WATERJET APPLICATIONS IN VESSELS THAT OPERATE IN MULTIPLE MODES

D Borrett and P Rae, Hamilton Jet, New Zealand

SUMMARY

Commercial waterjet installations emerged predominantly in high-speed vessels such as passenger ferries that operate
over a relatively narrow displacement and speed range. The design goals have therefore focused mainly on high-speed
propulsive efficiency. However, waterjets are increasingly employed in vessels that have significant duty-cycles over a
wide range of speeds and loading conditions, together with demanding manoeuvring requirements, Understanding the
design characteristics that enable waterjets to achieve both high-speed efficiency and high manoeuvrability is critical to
meeting the requirements of emerging waterjet markets.

A waterjet's ability to absorb high power without breakaway cavitation is governed by the nozzle-to-inlet diameter ratio,
pump configuration and intake geometry. Steering nozzles and reverse deflectors can be designed to allow the waterjet
to effectively emulate an azimuth thruster. With minimal dead band, low actuation loads and dedicated electro-hydraulic
controls, rapid, accurate and efficient thrust vectoring is possible.

An offshore crew boat equipped with a dynamic positioning system is used to illustrate the requirements and capabilities
of waterjets in “multi-mode” applications. Waterjets for this type of vessel must perform over the wide range of transit
speeds resulting from variations in cargo loading, and deliver the high manoeuvrability needed for accurate station
keeping. In addition to the waterjet propulsive efficiency at speed, key factors in achieving these goals are the steering
efficiency, cavitation margins under high loading conditions, high available static thrust, and the speed and accuracy of
thrust vectoring.

With careful design optimisation, waterjets can offer efficient performance across a wide range of vessel operating
modes, providing an effective and versatile propulsion solution.

1. INTRODUCTION 2.1 PILOT BOATS

Commercial waterjet installations emerged Pilot boats typically have twin or triple waterjets giving a
predominantly in high-speed vessels such as passenger transit speed in the 30 to 35 knot range. They have
ferries that operate over a relatively narrow displacement special manoeuvring requirements for transferring pilots
and speed range. The design goals have therefore focused to and from large vessels that are making way at around
mainly on high-speed propulsive efficiency. However, 8-10 knots. The transfer process is a potentially
waterjets are increasingly employed in vessels that have dangerous operation due to interaction between the two
to perform well in several different operating modes, moving vessels, and requires a high degree of skill on the
covering a wide range of speeds and loading conditions, boat operator’s part. High thrust levels with very rapid
together with demanding manoeuvring and station response are needed. A common technique while
keeping in offshore conditions. manoeuvring alongside a moving vessel is to run the
engines at a fixed high RPM and 'trim' the waterjet
This paper introduces some typical multi-mode vessel reverse buckets to control the forward speed. The jet
applications, examines the design characteristics that thrust response is therefore dependent only on the reverse
enable waterjet propulsion systems to achieve high and steering hydraulics and not the engine response.
performance levels over these wide ranging types of Combined with small travel on the helm wheel, this
vessel operations, and looks in more detail at the allows very precise control during the critical pilot
propulsion solution for an offshore crew boat. transfer stage with the ability to quickly pull away from
the moving vessel when required.

2. MULTIPLE MODE VESSEL For pilot boats the key waterjet requirements are to
APPLICATIONS provide high thrust without cavitation at relatively low
forward speeds, together with rapid response of the
The following section outlines the special requirements reverse bucket and steering nozzle. These characteristics
of three different vessel types that operate for significant are simply not available with other types of propulsor.
periods in distinct, multiple modes.
2.2 PATROL BOATS

Patrol boats typically have three distinct modes - a low

speed 'loitering' mode, a patrolling/cruise mode (at

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

perhaps 70-80% power) and a high-speed chase/attack propulsive coefficient or QPC) is determined mainly by
mode. They have to manoeuvre close to other vessels for the uniformity of the flow field entering the pump, the
boarding and may also perform rescue duties. Their efficiency of the pump and discharge nozzle, and the
particular type of high-speed operation requires the minimisation of losses from other necessary features
ability to manoeuvre aggressively and to accelerate and such as intake screens. Other variables affecting the
decelerate rapidly. overall propulsive coefficient (OPC) such as the wake
fraction and thrust deduction values are hull and speed
Many different propulsion configurations are used dependent and are not included in this discussion.
including propellers, waterjets, surface drives and
combinations of these systems. Waterjets are well suited 2.4 (a) Intake Flow Uniformity
for this application due to their good manoeuvring
characteristics for both low and high speed operations, An ideal intake would present a uniform flow field to the
rapid acceleration for chase/attack, 'crash stop' capability pump rotor under all operating conditions. Achieving this
and high efficiency at transit speeds. Triple waterjet would necessitate a variable geometry intake which is
installations are common, with the centre unit often being generally deemed too complex and costly. The challenge
of a smaller size to provide efficient loitering capability. for the waterjet designer therefore is to achieve good
The centre jet may also be employed to provide high- performance across the required range of operating
speed boost as well as loitering and therefore must conditions within the fixed-geometry constraints. In the
operate over a very wide speed range. type of vessels being discussed, the duty cycle can
involve significant periods of operation anywhere from
The key waterjet requirements are the ability to provide zero up to 50 knots..
efficient loitering and transiting, rapid acceleration
without cavitating, and responsive manoeuvring at both A long shallow intake is the obvious solution to
high and low speeds. achieving good flow uniformity at the impeller plane
under high-speed conditions. Such long intakes generally
2.3 CREW BOATS have a less than desirable shape for good low speed and
static thrust performance, where a steeper ramp and
Crew boats are used to transfer personnel and cargo to intake floor plus larger radii around the sides and rear of
and from offshore platforms. In recent years these vessels the intake are desirable. Long intakes also increase the
have increased in size to the point where they now carry entrained water mass, lengthen the drive shaft and
out many of the tasks previously done by offshore supply increase the inboard structural intrusion. Waterjet
vessels, such as transferring cargo, fuel and other liquids. performance tests at Hamilton Jet have shown that a
number of intake designs intended for optimal high speed
The cargo-carrying role of these vessels means that they performance could, in fact be reduced in length by 20%
operate at greatly different displacements between with no reduction in the high-speed performance of the
lightship and full load. One reason for the success of waterjet.
waterjets in these vessels is that, compared to an
equivalent propeller driven boat, waterjets make use of An important feature of the waterjet intake is the
full engine power irrespective of the state of loading and cutwater, or ’lip' at the rear of the intake opening. For
transit speed. Propellers have to be 'pitched' to meet the high-speed it is generally best to employ a small lip
required speed at the maximum loading condition and radius, no ramp underneath the lip and a shallow intake
cannot propel the boat significantly faster when it is floor. However, this is virtually the opposite of the
light. Waterjets, however, allow the full engine power to geometry required for good low speed performance,
be used at all conditions: this translates into significantly which profoundly affects bollard pull and manoeuvring
higher speeds when a vessel is light or partially loaded. thrust. A larger lip radius can be employed on an intake
otherwise optimised for higher speeds, in order to
The second requirement of these vessels is for accurate improve the static thrust performance.
station keeping while at an offshore platform.
Increasingly, and predominately for safety reasons, this is For a given nozzle size, power input, vessel speed and
being done under automatic control via a dedicated therefore flow rate, a waterjet design with a smaller inlet
dynamic positioning (DP) system. The waterjets are diameter (larger Nozzle-to-Inlet Ratio or NIR) has higher
required to deliver high thrust without cavitation at inlet velocities, an improved Inlet Velocity Ratio (IVR)
'bollard pull' conditions for this mode of operation, as and greater impeller plane flow uniformity at high speed.
well as fast and accurate response of the control surfaces. Increasing the waterjet NIR has a much greater influence
over the QPC at higher speeds than changes to the intake
2.4 HYDRODYNAMIC DESIGN ASPECTS OF geometry. The downside of a higher NIR waterjet design
WATERJETS FOR MULTIPLE MODES is reduced cavitation performance and static thrust for a
given pump configuration, unless a large impeller blade
For a given waterjet nozzle size, power input and vessel area is employed - with consequential loss of pump
speed, the efficiency of the waterjet system (quasi- efficiency.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

2.4 (b) Pump Performance Comparing two commercially available waterjets from
different manufacturers, one axial flow and the other
The waterjet NIR is the principal design parameter that mixed flow, both of the same nozzle size and operating at
determines the pump configuration to be employed. A the same power input and rpm, the axial flow waterjet
larger NIR pump will generally be selected for high has a 5-6 knot lower speed limit for continuous
speed as noted above. However, larger NIR pumps can cavitation-free operation, and over 20% higher maximum
exhibit poor low speed or static thrust performance and static thrust, which greatly increases its versatility in
significant pump taper (radius change) is required to multi-mode applications.
achieve even a minimum acceptable level of cavitation
performance. 2.4 (c) Steering Efficiency

Figure 1 shows the relationship between NIR and A further design consideration that has a particular
maximum static thrust (bollard pull) for a family of bearing on the vessel OPC is the performance of the
similar waterjet pumps operating on the same intake, waterjet steering system.
showing the reduction in static thrust with increased NIR.
When a vessel is operating in a seaway with significant
The ability of a particular waterjet design to be wind and wave conditions, the steering demands are
successful in a wide range of vessel applications is often high in order to maintain a set course. Experience
critically dependent on the choice of NIR. It is fair to say with well-designed nozzle type steering systems has
that in hydrodynamic design, reducing the trade-offs shown that transit times can be reduced by up to 5% on
between good low and high speed performance is a longer runs, when compared to external deflector type
significant area of research for most major waterjet systems, translating directly into an improvement in
manufacturers. overall propulsive efficiency of the same order. This
’course-keeping’ efficiency has less benefit on shorter
runs in calm waters, but these conditions are generally
much less common.

On the passenger ferry ’Ocean Flyte’, a 31m, 30 knot

Maximum Static Thrust

monohull operating in Singapore waters, a 1.5 knot

higher average speed over a one-hour run was measured
following an upgrade from an external spherical deflector
type steering to a high efficiency ‘JT’ type steering
nozzle system. The latter also provides greater
manoeuvring thrust as the design is not affected by the
’flooding’ of the steering system when the vessel is
stationary.

00 2.4 (d) Design Parameters

0.5 0.52 0.54 0.56 0.58 0.6 0.62 0.64
Nozzle to Inlet Ratio (NIR) Table 1 summarises the relevant design parameters for
various waterjet operating modes. Achieving the widest
Figure 1 - Reduction in Static Thrust with NIR effective operational range possible is of course a
significant development goal for waterjet designers.
Another focus of recent waterjet development has been
the reduction in waterjet envelope size. It is a relatively Operati High Laden Low Static
straightforward task to design a highly efficient waterjet ng Speed Cruise Speed, Thrust
pump within a large envelope (higher radial flow Mode High
component) but somewhat more challenging to design a Power
highly efficient, compact pump with very good cavitation Speed
resistance. Range 35 - 50 20 - 30 8 - 12 -2 - +2
(kts)
Under uniform flow conditions the achievable efficiency Exampl Passeng Crew Pilot Station
of axial flow waterjet pump designs can however, be e er ferry boat - transfers, keeping
very close to that of mixed flow types. laden towing
Intake Long/lo Medium Medium Short/
Most mixed flow pumps employed in waterjets today w ramp steep
have a maximum pump diameter around 40% larger than angle
the inlet, resulting in a less than compact installation Intake Small Medium Medium Large
package – typically 20% larger in diameter than the Lip
equivalent nozzle size axial flow pump. Radius

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

NIR High Medium Medium/ Low power only results in increased cavitation and reduced
low thrust. With the engine matching shown, high levels of
Pump Efficien Efficien Breakawa Breaka cavitation are avoided, with maximum thrust being
optimis cy cy with y way achieved at the full engine power. For station keeping,
ed cavitati cavitation cavitati thrust at this maximum power level would normally only
for on resistance on occur for short periods, and practical experience with
margins resistan crew boats over many years has shown that cavitation
for ce related damage is therefore avoided.
unlimite
d
operatio
n
Steerin High High Fast Fast
g course course response, respons
Nozzle keeping keeping high e most
efficien efficien accuracy importa
cy cy nt
Reverse N/A N/A Fast Fast
Duct response, respons
high e
accuracy

Table 1 - Summary of Design Parameters for Different

Operating Modes

2.5 STATIC THRUST CHARACTERISTICS Figure 2 – HM811 Static Thrust vs Engine

Power
The power level that can be used with a particular
waterjet under bollard-pull or static thrust conditions is
determined by the waterjets’ resistance to cavitation; a If more powerful engines are used with the same
potential problem in all types of propulsor. It occurs waterjets, measures need to be taken to ensure that the
when the propulsor blades become very highly loaded, power is limited, to avoid sustained operation in the
creating sufficiently low pressure in the water to cause a cavitation region. In crew boats, this limit is controlled
rupture and form a cavity. The subsequent collapse of the by the dynamic positioning system. When the vessel
cavity due to the pressure of the surrounding fluid, changes to transit mode, the cavitation margin increases
releases a significant amount of energy in the form of an with forward speed, allowing higher power to be applied.
acoustic shock-wave. This energy can be sufficient to
cause erosion of the metal parts in the vicinity of the 2.6 CONTROL OF THRUST FOR
cavitation, resulting in serious damage if allowed to MANOEUVRING
continue. Thrust levels from a waterjet can reduce
dramatically during cavitation, and permanent damage to Apart from the availability of high static thrust, the key
components causes gradual degradation of propulsive to good manoeuvrability lies in the design of the control
efficiency. surfaces that direct the thrust. Manoeuvring thrust is
produced in a waterjet by splitting the jet stream into
For vessels that must achieve the best possible static different components and varying the ratio of flow
thrust and manoeuvring performance, waterjets are often between these components. There are two different
selected in a larger sizing than for other vessel types. design approaches in common use for achieving this.
This is one way of improving cavitation margins, as well
as contributing to better propulsive efficiency. The first method, seen in many larger waterjets, uses a
box shaped steering deflector with a reverse deflector
Figure 2 illustrates the relationship between static thrust 'flap' mounted on it so that the whole assembly swivels
and engine power for a Hamilton HM811 waterjet when the steering is actuated. When the reverse deflector
matched with a 1340kW engine – a common is deployed the flow is split into two components that are
combination on many crew boats. At a power level opposed by 180 degrees at all steering angles. When the
around 1000 kW, the pump efficiency is starting to reverse deflector is set so as to balance the ahead and
reduce - indicating the initial onset of cavitation. Further astern components of flow (zero speed), no side thrust
gains in static thrust can be achieved beyond this point as vector can be generated since the flow components
more power is applied, but at the expense of increasing remain opposed by 180 degrees at all steering angles.
cavitation levels. At a power level of approximately 1350 Figure 3(a) shows the approximate thrust envelope for
kW the thrust level peaks, and attempting to apply more this type of steering/reverse mechanism. Waterjets using

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

this design are more limited in manoeuvring capability ƒ The 'Starboard' component which goes into
and are not well suited to DP operations. the starboard side of the reverse duct.

The 'Ahead' component is only acted on by the steering

nozzle. The volume of flow in this component is
dependent on how far the reverse duct is lowered. The
remaining flow goes into the reverse duct where it is
further split into port and starboard components with the
ratio determined by the steering nozzle position.

When the reverse duct is fully raised all the flow goes
astern, providing maximum ahead thrust. This is the
normal under-way condition. By moving the steering
nozzle, this flow (shown right in plan view) can be
diverted to provide steering forces.

Figure 3 - Thrust Envelopes for (a) Steering-

Mounted and (b) Split Duct Reverse

In the second design approach, the reverse duct is

attached to the waterjet housing. It can be raised and
lowered independently and does not rotate with the
Steering Steering
steering nozzle. The jet stream first passes through the to Port Centred
steering nozzle before being intercepted by the reverse
duct which then turns part or all of the flow back under When the reverse duct is partially lowered, the jet stream
the hull to produce reverse thrust. This type of reverse is split into two or three components. The flow below the
duct has split passages, which allow either two or three duct goes straight astern as before and can be directed to
flow components to be generated. It produces a thrust port or starboard by the steering nozzle. The upper flow
envelope as shown in Figure 3(b), allowing a thrust component that enters the reverse duct is further split into
vector to be obtained at any azimuth. The method of port and starboard components. The reverse duct turns
achieving this is described below. these flow components round and ejects them from the
port and/or starboard outlets in a forward direction
underneath the hull, producing astern thrust. By altering
the steering nozzle angle, the ratio of flow entering the
Stbd Reverse Port Reverse port and starboard sides of the reverse duct can be varied.

Ahead
Steering Steering
to Port Centred
Figure 4 - View into Reverse Duct along Jet Stream
When the thrust from the ahead and astern components is
Figure 4 shows a view looking aft along the jet axis into in equilibrium, the jet is said to be at the 'zero speed'
the reverse duct, which is partly lowered into the jet position. The key feature of the split passage reverse duct
stream. In this view the steering nozzle is angled slightly is that in this condition, even though there is no net
to starboard. The jet stream is split into three components ahead/astern thrust, by moving the steering nozzle, side
as follows: thrust is generated, allowing the vessel to 'steer' even
when stationary. In the case of DP operations, this side
ƒ The 'Ahead' component which goes thrust can be directly used with thrust from a bow
underneath the reverse duct. thruster to hold the vessel against cross winds and
ƒ The 'Port' component which goes into the currents.
port side of the reverse duct.

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

When the reverse duct is fully lowered, it captures all the above a value of IJr = 0.25, position errors increase
flow and maximum astern thrust is produced. By moving exponentially.
the steering nozzle, the ratio of flow entering the port and
starboard passages can be varied, thus providing steering In addition, with a slower hydraulic response, since the
forces in the astern direction. vessel moves further off position, the correcting thrust
has to be applied for longer, possibly involving a higher
engine speed and resulting in more energy being used.
Applying a correcting thrust rapidly and early is more
fuel efficient.

Steering Steering
to Port Centred

IJr = 1
When the jet stream is deflected by the steering nozzle
and reverse duct, some energy is lost. With an efficient
steering nozzle design this loss may be no more than 5% IJr = 0
at maximum steering deflection.

Losses in the reverse duct are higher due to the greater

angles through which the water is deflected and the fact
that the flow astern is directed down and to the sides to
avoid hitting the vessel's transom, resulting in a non-ideal
vector. The astern thrust achieved with the split duct
reverse bucket is thus approximately 60% of the ahead Figure 5 - Position Errors for Different
thrust. Reverse Hydraulic Response Times

2.7 THRUST RESPONSE Obviously, as vessel size increases, so its response

becomes slower, and longer time constants in the
Vessels that use waterjet propulsion are comparatively hydraulic system can be tolerated. However, as the above
light in weight and therefore respond quickly when acted analysis demonstrates, it is important to match the
on by wind, wave and current forces. For manoeuvring waterjet control response to the expected vessel response;
and dynamic positioning the waterjet thrust response otherwise position keeping accuracy will be
must be fast enough to counter these disturbances before compromised, particularly where this is automated.
the vessel moves too far from the desired position and
heading. The thrust can be directed rapidly and precisely The JT type steering nozzle and the split duct reverse
on a waterjet by moving the reverse duct and steering bucket are both designed for minimum actuation loads.
nozzle to new positions and adjusting the engine RPM. This helps to reduce the size, weight and power
requirements of the hydraulic system, while achieving
A simple analysis demonstrates the relationship between rapid response rates.
thrust response and position keeping ability for a 7m test
vessel. Figure 5 shows simulation results of the surge As well as the characteristics of the hydraulics, thrust
position error following an input disturbance of 300 N. response is also dependent on how quickly the engine
The reverse hydraulic system is modelled by a first order can change speed. This varies with engine size and type,
lag with the time constant IJr varying between 0 and 1 but is invariably slower than the steering and reverse
second. It can be seen that the position error following actuators. To reduce dependency on the engine response
the disturbance is only 0.15m when IJr = 0, increasing to while manoeuvring, the RPM can be held at a higher
0.64m when IJr = 1. setting so that greater thrust levels are produced
immediately the jet hydraulics move.
In each case the vessel controller has been optimised to
match the different values of IJr so that the position In Figure 6 the innermost area shows a thrust envelope
control loop does not become unstable. As IJr increases, with an engine at low idle. Thrust direction and
the position control loop gain has to be reduced to magnitude are controlled within this area by positioning
maintain stability and it is this gain reduction that results the reverse duct and steering nozzle. Raising the engine
in a greater position error. speed above low idle increases the size of the inner
envelope as indicated. This allows a greater proportion of
It can be seen that for this example, a value of IJr = 0.25 the thrust to be controlled only by the reverse and
seconds would be acceptable. Not much improvement is steering positions, giving a faster response that is not
possible if this lag is eliminated completely (IJr = 0), but

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

limited by the engine response. This mode of operation is four waterjets, an excess of thrust is available at the stern
commonly used on pilot boats and it is also applicable to of the vessel for manoeuvring and the maximum
station-keeping. athwartships thrust is limited by the bow thruster
capability not by the waterjets. Selecting waterjets as the
2.8 THRUST RESOLUTION main propulsion provides powerful manoeuvring
thrusters ‘for free’.
Waterjets provide virtually infinite thrust resolution
during manoeuvring. At lower thrust levels this is
dependent on the accuracy with which the reverse duct 3. CASE STUDY – OFFSHORE CREW BOAT
and steering nozzle can be positioned. Using electronic
closed loop control, positioning accuracy of a hydraulic This section examines in more detail the application of
cylinder to better than 1% is readily achievable and this waterjet propulsion to a recent offshore crew boat, the
equates to a resolution of about 0.2% of the maximum ‘Joyce McCall’.
thrust. At higher thrust levels, the resolution is dependent
on how accurately the engine speed can be controlled. 3.1 MARKET DRIVERS
Given engine speed control to within 10 RPM for
example, about 0.7% thrust resolution would be obtained Over the last 8 years waterjets have taken significant
for an HM811 waterjet. market share from conventional propeller systems in
crew boats built in the USA. This is quite surprising
since most crew boats are designed for speeds somewhat
below the typical waterjet domain, and therefore cannot
compete with propellers on grounds of propulsive
efficiency alone. The primary drivers cited for selecting
Maximum
waterjets for these vessels are as follows:-

x The ability to achieve higher speeds in light to

mid range cargo load conditions.
Thrust with
increased
x Improved manoeuvrability under manual and
automatic control with high bollard pull and
Low Idle 'azimuth thruster' characteristics.
x Reduced vessel draft which allows shallow
water operations.
x Reduced vessel dry-dockings, waterjets being
largely immune to damage from debris and
underwater obstacles.
x Less maintenance and improved life for
Figure 6 - Jet Thrust Envelopes with propulsion machinery since waterjets do not
Varying Engine Speed overload engines and gearboxes remain
permanently engaged when manoeuvring.
Conventional main drive propeller systems cannot x Jet driven steering and reverse hydraulics do not
achieve such accuracy over the direction and magnitude require AC power - no loss of steering control in
of thrust for manoeuvring. It is the combination of high a 'deadship' situation.
thrust levels, rapid response and fine resolution that make x Integrated engine, steering and reverse control
waterjets excel in high performance manoeuvring system provides ease of installation for the
applications. builder.
x Simple interface with DP systems and rapid set
2.9 MANOUEVRING THRUST EFFICIENCY up.

Using main drive propulsion (propellers or waterjets) for 3.2 PROPULSION REQUIREMENTS
manoeuvring of high speed craft represents a
compromise that cannot match the efficiency of At 54 metres in length with a loaded displacement of 500
dedicated manoeuvring thrusters. However, to fit vessels tonnes, this vessel is near the top of the current size range
such as crew boats with a full set of separate thrusters for for crew boats. Waterjet selection was based on the
station keeping is not considered a viable option due to required speeds at light and loaded displacements and the
the added weight, cost and complexity. need to maintain adequate cavitation margins under full
Typical thrust/power figures for a propeller azimuth load cruise and while manoeuvring. Four waterjets were
thruster are quoted as 150N/kW1. By comparison, main specified, this being the most common arrangement,
drive waterjets achieve approximately 72N/kW. This although some vessels using six jets are now under
reduced efficiency is compensated by the fact that high construction in the USA.
power levels are available from the main engines. With

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

The selected HM811 waterjets have the following In contrast, the waterjet absorbs power more consistently
characteristics:- across the whole operating speed range of the vessel. It
cannot overload the engine, even when the vessel is fully
Max. Continuous Power (cruise) 2800 skW laden. When the vessel is running light, the waterjet can
Max. Continuous Power (DP) 1000 skW convert the available engine power into a much higher
Max. Thrust (DP) 70 kN transit speed. Figure 7 shows the jet thrust and resistance
curves for the 'Joyce McCall' at the different loading
In order to meet the requirements for dynamic conditions.
positioning, this vessel is fitted with a retractable azimuth
thruster at the bow. A separate tunnel thruster uses the
same power source for docking purposes. The particulars
for the 'Joyce McCall', are summarised in Jet Thrust 450 LT
Table 2.
330 LT

210 LT

Length overall 53.8 m

Figure 7 - Waterjet Crew Boat Speed at
Length BP 48.01 m
Different Loading Conditions
Beam 9.14 m
Depth 4.11 m
Light Displacement 205 MT Under fully loaded operating conditions, a propeller
driven vessel will be somewhat more efficient than an
Loaded Displacement 508 MT
equivalent waterjet vessel. However, when considering
Main engines 4 x Cummins KTA50 M2
the overall duty cycle, much of which is with light or
Power 4 x 1340 kW @ 1900 RPM
medium loads, waterjets achieve significantly faster
Waterjets 4 x Hamilton HM811 round trips and/or a greater operating range.
Bow thrusters 1 x 150 kW Tunnel
1 x 150 kW azimuth 3.4 STATION KEEPING PERFORMANCE
Speed at 450 LT 18.5 kts
displacement Dynamic positioning systems operate by actively
Speed at 210 LT 31 kts controlling the vessel in three degrees of freedom –
displacement surge, sway and yaw, based on one or more high
accuracy positioning systems and heading sensors. The
Table 2 - Joyce McCall Particulars ability of a vessel to hold position under the influence of
current, wind and wave forces is defined by the ‘DP
3.3 TRANSIT PERFORMANCE capability plot’. This plot provides a measure of the
combined forces that the vessels thrusters are capable of.
In a vessel of this type, the hull resistance increases by a A crew boat uses the main drive waterjets in combination
factor of approximately three times between the light and with one or more bow thrusters to create the desired
fully loaded conditions. If this vessel was to be driven by forces and moments for station keeping.
fixed-pitch propellers (as the majority of crew boats are),
these would be selected for the required speed of the The DP capability plot for the 'Joyce McCall' is shown in
vessel at the fully laden condition, otherwise engine Figure 8. The plot clearly illustrates that high thrust
overloading would occur. When the same (prop driven) levels are available in the surge direction, but limited
vessel is at light displacement, its top speed is only thrust is available in sway. In the latter case, this vessel is
slightly higher than the fully laden speed because this able to hold station when side on to a 1.5 knot current
becomes limited by how fast the propeller can rotate. In and approximately 10 knots of wind, but beyond that,
the light condition the engine RPM becomes limited by insufficient thrust is available. In fact this limitation is
the governor but the engine is only lightly loaded, unable purely due to the bow thruster which has to
to make use of the full power available2. This is rather approximately match the side thrust from the waterjets,
like being stuck in too low a gear. of which there are four with significantly higher power

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

available. This is not really a problem, since these vessels Assuming that the current and wave drift forces act at the
would normally hold station bow or stern-on to the wind vessel mid point (centre of rotation) while the wind
and waves forces act 3 metres ahead of this point in the same
direction, each waterjet need only produce about 15% of
it's available side thrust in order to balance the bow
thruster and 10 knot wind moments. Thus the main
propulsion would operate at quite low power levels to
hold station against these side forces, resulting in good
fuel efficiency.

The jet thrust envelope is limited to a circular area

having a radius equal to the minimum thrust level, as
shown in Figure 10, to more closely emulate an azimuth
thruster and simplify the DP calculations. Compensation
for the different thrust efficiencies that vary with the
azimuth demand, is automatically carried out in the
DPIM by adjustments to the engine RPM. The imposed
restriction in ahead/astern thrust using the circular
envelope is not a limitation in practice but does restrict
Figure 8 - DP Capability Plot
the derived DP capability plot in the ahead/astern
directions.
3.5 CONTROL SYSTEM DESIGN
Apart from low power usage, a further advantage of this
Larger Hamilton waterjets currently use the Modular method is that all waterjets can thrust in the same
Electronic Control System (MECS) which provides direction and the thruster re-allocation in the event of a
steering, reverse, engine speed and gearbox control. An failure becomes straightforward. The loss of one or two
additional Dynamic Positioning Interface Module waterjets out of four would be unlikely to have a great
(DPIM) provides the link between the jet controls and a effect on sway controllability under DP as excess side
DP system. thrust is available from the remaining jets.

The DP system computes desired thrust vectors for each

waterjet and the bow thrusters(s), as required to hold the
vessel in position. To simplify the interface with the
waterjet controls, the DPIM incorporates a real-time jet
thrust model that computes the waterjet steering and
reverse and engine speed demands, in response to the DP
demanded thrust vectors. From the DP systems
perspective, the waterjet therefore has similar
characteristics to an azimuth thrusters. This greatly
simplifies system installation and commissioning.

3.6 THRUST ALLOCATION FOR DYNAMIC

POSITIONING

With four waterjets plus a bow thruster available for

station keeping, there are many ways in which to
combine the thrust forces for maintaining position. The
methods, referred to as ‘thruster allocation’ are
developed against various optimisation measures, such as
minimising thruster power or minimising the effects on
the vessel of thruster failure. Different allocation
schemes may be employed depending on the weather
conditions or the type of operation being carried out. The
method for generating sway thrust for the Joyce McCall Figure 9 - Allocation Methods for Side Thrust
is illustrated in Figure 9(a).

As shown, the jet and bow thruster vectors are directed Figure Figure 9(b) and (c) illustrate two alternative
athwartships together to generate sway thrust while methods for producing side thrust using four waterjets.
balancing the rotational moment; the jet side thrust being Both methods make use of the excess jet thrust that is
allocated equally to all waterjets. available in the longitudinal direction to produce
additional side thrust by a ‘push-pull’ technique (used in

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

propeller crew boats3), achieving an increase of 5. REFERENCES

approximately 35% over method (a). However,
considerably higher power levels are required from the 1. 2. Bekker, Joe, R. Is a Drop-Down Better
waterjets in order to obtain this increase. Than a Tunnel Thruster?
http://www.thrustmastertexas.com/applications/
dropDownBetterThanTunnelThruster.html
3. Alexander, K. (1995). Waterjet versus Propeller
Matching Characteristics. Naval Engineers Journal, May
1995, 129-139.
Moen, Lars, K. (2004). Crew Boat Station Keeping –
Challenges and Solutions. Dynamic Positioning
Conference Workboats/OSVs, September 2004.
http://seacormarine.com/crewzer/index html

Our thanks to Gulf Craft,Inc. and Kongsberg Maritime

for the information that they contributed towards this
paper.

6. AUTHORS BIOGRAPHIES

Dick Borrett is a Senior Development Engineer at

Hamilton Jet, New Zealand. He worked in the field of
hydrodynamics and control of underwater vehicles for
British Aerospace and Dowty Maritime in the UK before
Figure 10 - Circular Thrust Envelope for moving to New Zealand in 1992. There he joined
Dynamic Positioning Hamilton Jet to establish an in-house capability in
electronic control system design and manufacturing.

Since that time he has been involved in new product

development for subsequent generations of control
4. CONCLUSIONS
system. As well as leading the design of the current
Modular Electronic Control System (MECS), he was also
Waterjet propulsion provides an excellent solution for
responsible for analysis and software development on the
vessels that operating in multiple modes. Through
DP Interface Module that provides Hamilton waterjets
appropriate design optimisation it is possible to achieve
with 'azimuth thruster' capabilities.
efficient high-speed propulsion in different operational
regimes, as well as high levels of static thrust. Together
Dick is currently researching new ideas for waterjet
with fast and accurate control of thrust through optimised
manoeuvring and control and has a number of New
steering nozzle and reverse duct designs, the demanding
Zealand and USA patents granted and pending.
needs for manoeuvring and automated station keeping
can be readily met. The ability of the waterjet to emulate
Philip Rae has managed the research, product
an azimuth thrusters also provides ease of interfacing
development and product support functions at Hamilton
with dynamic positioning systems.
Jet for nearly 10 years. A mechanical engineer by
training, he worked principally in application engineering
The problem of cavitation under high loading can be
and business development prior to completing a business
managed through careful hydrodynamic design,
degree and moving into technical management.
appropriate selection of jet size, correct matching to
engine power, and where necessary, controlled limiting
He has extensive experience with a wide range of
of the applied power levels.
waterjet applications and is directing new product
development programs for both waterjets and electronic
The demand for multiple mode vessels continues to
control systems.
provide challenges, a recent example being a catamaran
crew boat design having a top speed of over 40 knots,
together with more stringent, higher redundancy station
keeping capabilities4. Waterjets are ideally suited to
fulfilling these new propulsion requirements.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

OFF-DESIGN BEHAVIOUR OF WATERJETS

N Bulten and R Verbeek, Wärtsilä Propulsion Netherlands BV, The Netherlands

SUMMARY

Waterjets and waterjet inlets are usually designed for a limited number of operational conditions. During a turn and in
manoeuvring conditions the flow in the inlet will deviate quite considerably from the flow in normal sailing conditions.
This will have an effect on the required shaft torque and the available pressure just upstream of the pump, expressed in
available NPSH. A third typical off-design condition, besides low speed manoeuvring and high speed steering is three jet
operation for a vessel with four waterjets installed. Simulations of the effects in off-design conditions are carried out
with aid of a commercial CFD code. The paper gives an overview of the different phenomena occurring at the various
off-design conditions and how this will affect overall performance. The effects on the performance are also reviewed in
light of the improved cavitation behaviour of the new axial flow pump types LJX and WLD.

NOMENCLATURE conditions will be presented in section 4. In section 5, the

benefits of the new axial flow LJX/WLD pumps with
D Diameter [m] improved cavitations margins will be evaluated. The
NPSH Net positive suction head [m] conclusions of the research will be presented in section 6.
n Shaft revolution rate [s-1]
p pressure [Pa]
vship ship speed [m s-1] 2. BACKGROUND OF NUMERICAL
w Wake fraction [-] ANALYSES

D Drift angle [degrees] The numerical analyses are based on the Reynolds-
H Inlet loss coefficient [-] Averaged Navier Stokes (RANS) equations. For the
N Pump specific NPSH [-] calculations a mesh has been built, which includes the
U Density [kg m-3] waterjet inlet duct, the E-type impeller, stator bowl and
the nozzle. The CFD method and the generation of the
mesh are described in detail in the PhD-thesis of Bulten
1. INTRODUCTION [1].

Evaluation of the performance of a waterjet installation is Figure 1 shows a picture of the complete mesh as used in
mainly based on the design operating condition. When the CFD analyses. The numerical domain is meshed with
the performance of various installations at the design hexahedral cells, based on a multi-block approach. This
point is more or less identical, additional operating method ensures good control over the quality of the cells
conditions can be included in the evaluation. In this near the walls, in which the effects of the boundary layer
paper the performance at some typical off-design development are modelled. The complete mesh of the
conditions will be addressed. Examples of these pump unit consists of about 1.35 M cells.
conditions are: (i) low speed manoeuvring, (ii) high
speed steering and (iii) 3 jet operation at a vessel with 4 Effects of turbulent flow are captured with the standard
installations. Differences in performance can be related k-H turbulence model. This model is utilised at the
to the propulsive efficiency, power absorption of the authors’ company for many years. Implementation of the
pump or the available margins against cavitation. Large body forces due to rotation of the impeller is based on the
variations in power absorption can lead either to quasi-steady Multiple-Frame-of-Reference method. In
overloading of the engine, resulting in black smoke in the this way the impeller is frozen at a certain fixed angular
exhaust gases or in over-speeding of the impeller. position.
Reduction of the cavitation margins will result in more
cavitation and in extreme cases even to thrust break- The inflow is prescribed at the front plane and depending
down due to severe cavitation. on the drift angle on one of the side planes. The
remainder of the sides of the numerical domain are
The research described in this paper is based on treated as constant-pressure planes. The flow rate
numerical analyses of the flow through a complete through the pump is governed by the prescribed RPM of
waterjet installation of the Lips-Jets E-type pump. The the impeller.
background of the Computational Fluid Dynamics (CFD)
method is described in the following section. It has been found that the results of calculations of the
waterjet installation operating at free sailing operating
Section 3 deals with the descriptions of the analysed conditions were in good agreement with the full scale
conditions. The calculated performance for the various performance of the installation [2].

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

and +30 degrees. Under normal circumstances the drift

angles will be much lower than 30 degrees.

3.3 THREE-JET OPERATION

In case of failure of an installation, the active number of

waterjet installations on a vessel will be reduced. So in
case of a vessel with four waterjets installed and one in
failure mode, only three jets are left for propulsion. This
can have a significant influence on the margins against
cavitation, as shown in the two figures below. These
figures show that the ship speed will drop from 44 knots
in four-jet operation to about 33 knots in three-jet
operation. Moreover, the margin against cavitation of
about 11 knots at four-jet operation vanishes almost
completely at three-jet operation.

120%
CAVITATION
LIMIT
100%

100%
Figure 1: mesh of complete waterjet installation as used Normalised thrust [%]
90%
80%
in CFD analyses 80%

60%

3. ANALYSED CONDITIONS
40%

Three typical conditions have been identified as typical

off-design conditions: low speed manoeuvring, high 20% Thrust - 4 jets
30%

speed steering and 3-jet operation. It should be clear that Resistance

off-design does not mean not-used-in-operation for these 0%
conditions. Evaluation of the performance in these 0 10 20 30 40 50
Ship speed [knots]
conditions can thus be important, when the overall
performance is reviewed.
120%
3.1 LOW SPEED MANOEUVRING
100%
The CFD calculations for the low speed manoeuvring
CAVITATION
conditions are carried out for 2 ship speeds, i.e. 3 and 6 LIMIT
Normalised thrust [%]

80%
knots. The revolutions of the pump are set to a realistic 100%
value corresponding to about 30% power. This condition 90%
80%
is marked in figure 2 (6 knots, 31% power). The drift 60%

angles have been varied between -60 and +60 degrees.

Both negative and positive angles have been analysed 40%

because of the non-symmetrical nature of the problem.

This asymmetry is created by the rotation of the impeller. 20% Thrust - 3 jets
30%
It is therefore expected that the torque of the impeller Resistance
might show differences in behaviour when operating in 0%
positive and negative drift angles. 0 10 20 30 40 50
Ship speed [knots]

3.2 HIGH SPEED STEERING Figure 2: E-type thrust diagram for 4 jets (top) and 3 jets
for given resistance line
The calculations for the high speed steering are
performed for 30 and 35 knots. Though this is a few
knots below the design speed, it is expected to represent 4. PERFORMANCE EVALUATION AT OFF-
a realistic condition, since the vessel will always lose DESIGN CONDITIONS
some speed during a steering manoeuvre [3]. The
revolutions of the pump are selected to simulate 100% The performance analysis of the waterjet installation is
power. The drift angles have been varied between -30 based on the impeller torque and the available total
pressure just upstream of the impeller among others.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

Evaluation of the torque of the impeller can give an Torque variation

indication of the risk of overloading the engine. In case
110%
of increased shaft torque the engine temperatures might
3 knots
increase too much and black smoke can be produced. 6 knots
This should be avoided as much as possible.
105%

Normalised torque [-]

The total pressure can be expressed as the available Net
Positive Suction Head (NPSH):
100%

ptot
NPSH A [1]
Ug 95%

with the density of the water.

For normal free sailing conditions the available NPSH 90%
-60 -40 -20 0 20 40 60
can be estimated based on: Drift angle [degrees]

2 Figure 3: normalised torque variation at low speed

p f  pv v ship 2 manoeuvring for 3 and 6 knots
NPSH A  1 H 1w [2]
Ug 2g
It can be seen that the effect of the drift angle is
where pv is the vapour pressure, vship the ship speed, a asymmetrical on the impeller torque. For the largest drift
coefficient representing losses in the inlet ducting and w angle analysed, an increase of torque of about 5% is
the wake fraction. observed, moreover the effect of the ship speed seems to
be linear. Extrapolation of the results towards larger drift
In order to evaluate the performance of the pump the angles and larger ship speeds may give torque increases
available NPSH is compared to the required NPSH of the in order of 10%.
pump. This required NPSH is related to each pump
design, and it is related to the operational point of the Analysis of the flow field just upstream of the impeller
pump according to: revealed the presence of a swirl-component in the flow,
which is related to the drift angle. This swirl of the flow
N can be regarded as a small variation of the impeller
NPSH R ˜ n 2D 2 [3] rotational speed. Thus with a swirl in the direction of the
g
impeller rotation a small reduction of the RPM is felt,
whereas a counter rotating swirl results in a small
The specific NPSH is represented by . The pump
increase of RPM.
operation is given by the pump speed n and diameter D.
As long as the available NPSH exceeds the required
The effect of manoeuvring on the available NPSH is
NPSH of the pump, the margins against cavitation are
shown in figure 4. The values are normalised with the
sufficient. A significant drop of the available NPSH, for
NPSH at zero drift angle for both ship speeds. It can be
example during manoeuvring, might lead to increased
seen that the variation in NPSH is more or less
cavitation of the pump. In such condition, the benefits of
symmetrical, which is in accordance with the
the improved cavitation margins of the LJX/WLD pumps
expectations. Since the reduction in NPSH can be
can be exploited.
regarded as additional frictional loss in the inlet ducting
(thus upstream of the impeller), it is expected that the
4.1 LOW SPEED MANOEUVRING
behaviour would be similar for positive and negative
drift angles.
The evaluation of the performance during manoeuvring
is based on calculations for two different ship speeds. All
For the largest drift angles a reduction in NPSH of about
calculations are carried out at a fixed rate of revolutions
8% is observed. This reduction will decrease the margins
of the pump. The impeller torque at zero drift angle has
against cavitation. Since it was shown in figure 2 that the
been used to normalise the results of all other drift
manoeuvring point is located near the cavitation limit, it
angles. Figure 3 shows the normalised torque for the two
is likely that the reduction in NPSH due to manoeuvring
ship speeds for various drift angles between -60 and + 60
will result in increased cavitation.
degrees.
For identical cavitation behaviour during manoeuvring
the RPM of the pump should to be reduced with about
4%, based on eqn. [3], which is equivalent with about
12% power.

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

NPSH variation
due to the applied method of normalising the results,
both available NPSH values are set to 100% at zero drift
102%
angle. It is now clearly shown that the effect of the drift
3 knots
6 knots angle is identical for both ship speeds. Comparison of
100%
figures 4 and 6 learns that the amount of NPSH reduction
at high speed is obtained at lower drift angles. A
Normalised NPSH [-]

98%
reduction of 8% at 6 knots is found at a drift angle of 60
degrees, whereas 30 degrees drift angle is found at high
96%
speed. In general, the critical margins against cavitation
94%
are found below 30 knots. It is therefore concluded that
high speed steering is regarded to be a critical off-design
92%
condition.

90% NPSH variation

-60 -40 -20 0 20 40 60
Drift angle [degrees] 102%

30 knots
Figure 4: normalised NPSH variation at low speed 35 knots
100%
manoeuvring for 3 and 6 knots

Normalised NPSH [-]

98%
4.2 HIGH SPEED STEERING
96%
The CFD calculations for the high speed steering
conditions are carried for 30 and 35 knots. The drift 94%
angles are varied between -30 and +30 degrees. It is
expected that normal steering will not result in these 92%
large drift angles, due to the limitations in the steering
angles of the waterjet installations. 90%
-60 -40 -20 0 20 40 60
Drift angle [degrees]
Figure 5 shows the normalised torque for both ship
speeds. Similar to the results presented in the previous Figure 6: normalised NPSH variation at high speed
subsection, all results have been normalised with the steering
values found for the zero drift angle.
4.3 Three-jet operation
The asymmetrical behaviour of the torque is in line with
the results found for the low speed manoeuvring. For the analysis of the effects of three jet operation a
Differences between the two ship speeds are negligible series of calculations is made at constant power with
for all drift angles. varying ship speed. Due to variations in flow rate
through the waterjet, also a variation of impeller torque is
Torque variation
found. The RPM of the pump has been varied for each
ship speed to meet the full power condition.
110%

30 knots Figure 7 shows the effect of the ship speed on the

35 knots
impeller torque, for full power operation. The torque has
105%
been normalised with the torque at 44 knots, which
Normalised torque [-]

represents the operating condition as shown before in

figure 2. The observed increase of torque for operation at
100%
lower ship speeds is only limited. On the other hand, the
effect of ship speed on the available NPSH is significant,
95%
as shown in figure 8. The quadratic relation of the NPSH
with the ship speed as shown in eqn. [2] can be
recognised in this chart.
90%
-60 -40 -20 0 20 40 60 Figure 8 shows that in case of a reduction of the ship
Drift angle [degrees] speed from 44 to 34 knots, the available NPSH decreases
Figure 5: normalised torque variation at high speed with 27%. This significant reduction causes the large
steering effect on the cavitation margins as shown in figure 2.

Figure 6 shows the normalised NPSH for the high speed

steering conditions. Based on eqn. [2], there should be a
difference in NPSH for the two ship speeds. However,

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

Torque variation at full power

The LJX/WLD pump does not only have a different
cross-sectional shape, but also the pump specific NPSH N
104% is better. As a result, the cavitation margins are larger
and therefore the cavitation behaviour is improved
compared to the E-type.
102%
Normalised torque [%]

100%

98%

96%
20 25 30 35 40 45 50
Ship speed [knots]

Figure 7: normalised torque at full power for varying

ship speed

NPSH variation at full power

110%

100%
Normalised NPSH [%]

90%
Figure 9: Cross-sectional views of mixed-flow E-type
and axial flow LJX/WLD-types
80%
5.1 EFFECT OF IMPROVED CAVITATION
70% MARGINS FOR MANOEUVRING

60% As shown in figure 2, the cavitation limit at 6 knots is

found at 32% of the full power. This represents an
50%
20 25 30 35 40 45 50
available thrust of 69% of the design thrust. Figure 10
Ship speed [knots]
shows the comparable thrust diagram for 4 waterjets
equipped with the axial flow type pumps. This figure
Figure 8: normalised NPSH variation at full power for shows that the cavitation limit at 6 knots is now found at
varying ship speed almost 41% power, which is equivalent with 83% thrust.

120%
5. EFFECT OF IMPROVED CAVITATION CAVITATION
LIMIT
MARGINS ON PERFORMANCE
100%

100%
5.1 DESCRIPTION OF NEW PUMP TYPES
Normalised thrust [%]

80% 90%
80%
The CFD analyses presented in the section before are
60%
based on the Lips-Jets E-type waterjet installation, which
has been introduced about 10 years ago. Since some
years, Wartsila has two new pump types in its portfolio, 40%
40%
namely the LJX and the WLD type. One of the main
differences between the two new pumps and the 20% Thrust - 4 jets
conventional E-type is the building type of the pump. Resistance

The original pump has the typical geometrical shape of a 0%

mixed-flow pump, whereas the new pump types are 0 10 20 30 40 50
Ship speed [knots]
shaped according to axial flow pumps. This is illustrated
in figure 9. This figure also shows the effect of the
transom flange diameter for identical inlet diameter. The Figure 10: LJX/WLD-type thrust diagram for 4 jets
outer diameter of the mixed-flow pump increases from operation for given resistance line
the inlet towards the transom, whereas the axial-flow
pump has a constant outer diameter over this part. The axial pump can absorb about 28% (41/32) more
power at manoeuvring speed of 6 knots before the

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

cavitation margins of the mixed-flow E-type are met. The effects of drift angles on torque and NPSH are
This compensates the effects of the inflow under drift significant at low speed manoeuvring conditions.
angles by far. Available margins for engine overloading and cavitation
might not be sufficient for all conditions encountered.
When the benefits of the improved cavitation margins are
expressed in the increase of available thrust, an increase High speed steering seems to be less sensitive to the
of 20% (83/69) is found. Increase in thrust increases the variations in torque and NPSH under normal conditions.
operational envelope of the installation, which will result Torque increase at realistic drift angles is only a few
in better acceleration amongst others. percent. In general the cavitation margins are substantial
above 30 knots (assumed all jets in operation).
5.2 EFFECT OF IMPROVED CAVITATION Consequently, the decrease of NPSH should not lead to
MARGINS FOR THREE JET OPERATION increased cavitation.

The evaluation of three jet operation for a given When a vessel with four waterjet installations is operated
resistance curve, as shown in figure 2, revealed that the with only three jets, the available cavitation margins at
available cavitation margins vanished for the E-type full power decrease significantly and they might even
pump. The same evaluation is shown in figure 11 for the vanish completely. A reduction in NPSH of almost 30%
axial flow pump types. At full power the speed margin is is found when the ship speed drops from 44 knots (with
about 6 knots. This is sufficient for good operation of the four jets in operation) to 34 knots with three jets.
vessel even at three jets.
Two new axial flow pump types, denoted LJX and WLD
120%
have been developed with improved cavitation margins.
6 knots
The axial pump types can absorb almost 30% more
100% CAVITATION
LIMIT power at manoeuvring speeds before the cavitation limits
are reached, which results in about 20% more available
Normalised thrust [%]

80%
thrust.
100%
60% 90%
80% The increased cavitation performance of the axial pump
enables operation of a vessel with three jets at full power,
40% whilst keeping sufficient margins against cavitation.
40%
20% Thrust - 3 jets
Resistance 7. REFERENCES
0%
0 10 20 30 40 50 1. Bulten, N.W.H., ‘Numerical Analysis of Waterjet
Ship speed [knots]
Propulsion System’, PhD thesis, Technical University of
Eindhoven, 2006
Figure 11: LJX/WLD-type thrust diagram for 3 jets
operation for given resistance line 2. Bulten, N.W.H., Verbeek, R., Van Esch, B.P.M.,
‘CFD simulations of the flow through a waterjet
installation’, International Journal of Maritime
6. CONCLUSIONS Engineering, Vol 148 part A3, pp 23-34, 2006

Both at low speed manoeuvring and high speed steering 3. Verbeek, R. & Bulten, N., 'Recent developments in
shaft torque of the impeller is influenced by the inflow waterjet design', RINA Waterjet Propulsion II
field under a drift angle. Negative drift results in a conference, Amsterdam, 1998
decrease of torque and positive drift lead to a higher
torque. The variation in torque is related to the presence
of pre-swirl in the inlet ducting. If the pump has to 8. AUTHORS BIOGRAPHY
operate at large positive drift angles an increase of torque
of about 5% is observed. This might lead to overloading Norbert Bulten is team leader of the CFD department of
of the engine. Wärtsilä Propulsion Netherlands BV. He has written a
PhD thesis at the Eindhoven Technical University (TU/e)
The available NPSH is influenced both at low speed about the numerical analysis of a waterjet installation.
manoeuvring and high speed steering. The decrease of
NPSH is related to the magnitude of the drift angle. Rob Verbeek is responsible for waterjet hydrodynamics
However, the behaviour is more or less identical for and product development within Wärtsilä Propulsion
negative and positive drift angles. Netherlands BV, formerly known as Lips Jets.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

RESEARCH ON THE OPTIMUM BLADES NUMBER OF MIXED FLOW PUMP BASED

ON CFD

Qiongfang Yang, Yongsheng Wang, Zhihong Zhang, and Mingmin Zhang, Naval University of Engineering, China

SUMMARY

Effects of changing the number of pump blades of both rotor and stator on the characteristics of flow pattern and
waterjet propulsion performances are principally investigated in this paper, which includes the changes of rotor's blades
from 5 to 7 to the effect of Stodola slip factor, of stator's vanes from 9 to 13 to the effect of straightening the outflow of
impeller, and of both rotor and stator's blade numbers simultaneously. It shows that, given the condition of fixed pump
revolution and vessel speed, the Stodola slip factor, the head, axial-thrust and brake power are all incline as the number
of rotor's blades increase, but the efficiencies decline; the circumferential velocity of the nozzle outflow weakens with
the increase of number of stator's vanes, but the improvement is approximately unchangeable when it comes to 11. The
waterjet performances are close when the stator's vanes are 10 and 11, both of their head and axial-thrust are bigger
than that of with 12 or 13 stator's vanes; The waterjet optimum propulsion performances can obtained when 5 blades of
rotor match 9 vanes of stator, while the rotor blades are 6 or 7, the best number of stator's vanes both are 11, but the 7
blades and 11 vanes combination presents a better cavitaion performances.

NOMENCLATURE ȕb2 Mean blade angle at the rotor discharge ( e)

ȕ2 Flow angle at the rotor discharge ( e)
b2 Width of rotor outlet ( m) U Density of water ( kg m-3)
D1m Average diameter of the rotor at inlet( m ) ȝ Dynamic viscosity ( N s m-2)
D2m Average diameter of the rotor at outlet( m ) ȝt Turbulence viscosity ( N s m-2 )
D2e Minimum diameter of the rotor at outlet( m ) UgQH
D2a Maximum diameter of the rotor at outlet( m ) Ș Pump efficiency ( - ), K
f Vector of inertia force ( N) N
g acceleration due to gravity (m s2)
H Head of pump ( m )
l Blade chord ( m ) 1. INTRODUCTION
N Brake power ( kW )
n Rotating speed of pump ( r min-1 ) The number of pump blades has significant effects on its
head, efficiency, and cavitation performances.
3.65n Q Generally, the effects present non-linear characteristics,
ns Specific speed ( m-3/4 s-3/2), ns
H 3/ 4 and a given device exist an optimum number of rotating
NPSH Net positive suction head ( m ) blades in the framework of optimum integrated
NPSHi Incipient net positive suction head ( m ) performances in practical engineering application.
p Pressure ( Pa ) Stepanoff (1957) pointed out empirical formulas to
p1 Upstream (suction) static pressure ( Pa ) analyze the optimum blades number for centrifugal and
p01 Upstream (suction) total pressure ( Pa ) axial flow pumps. It is important to recognize that the
pv Vapor pressure ( Pa ) extrusion action of blades and its surface friction loss
PV' Introduced parameter to determine local p<pv should be weakened as much as possible when the
region ( Pa ) number of blades is calculated, additionally, the blade
pmin Minimum pressure in-house the pump ( Pa) passage should be sufficiently extruded to ensure stable
Q Volume flow rate ( m3 s-1 ) flow field and sufficient actions blades act on the flow
r Vector of space ( m ) [2].
SS Stodola slip factor ( - ), S S 1 S sin E b 2
Z The key to design pumps that operate more efficiently,
s Solidity ( - ), s l / t m quietly, and reliably at lower cost is a better
t Time ( s ) understanding of, and ability to, predict their
tm Cascade interval ( m ) hydrodynamics accurately, which requires a detailed
U Vector of fluid velocity ( m s-1) scene of the flow fields within the blade passages.
Vs Ship speed ( m s-1) Nowadays computational fluid dynamics (CFD) is
V Pump suction speed ( m s-1) seeing more and more use successfully in predicting the
Ȧ Vector of angular velocity ( rad s-1) flow fields in both stationary and rotating blade
Z Number of blades ( - ) passages. Miner[3] analyzed the flow field within the
Į1 Blade angle at the rotor discharge ( e) first-stage rotor and stator of a two-stage mixed flow
Į2 Flow angle on shroud side at rotor outlet ( e) pump and made a comparison with the measured data of
the rotor's velocity and pressure profiles. Hu et al.[4]
Į3 Flow angle on hub side at rotor outlet ( e)

©2008 The Royal Institution on Naval Architects

 Waterjet Propulsion 5, London, UK

used several models, including single blade to blade wU

channel model and the whole impeller model with or  ’ ˜ ( UU) 0 (1)
wt
without the shaft and the stator, to compute the flow
field in the impeller and nozzle of a waterjet pump and D
compare the torque with the experimental data. Bulten ( UU)  2 UȦ u U  UȦ u Ȧ u r (2)
Dt
(2006) made a detailed analysis of a mixed-flow Uf ’p  ( P  P t )’ U 2
waterjet propulsion system in his Ph.D. dissertation,
especially to quantify the effects of the non-uniform
inflow and the resulting non-stationary flow on the where f is inertia force, water is sucked from the bottom
system performances, and to quantify the forces on the of the ship with the change of potential energy, so the
complete waterjet installation in both axial and vertical gravity is taken into account. p is modified to account
direction. These cases demonstrate the ability of the for effects due to rotation, U density, and P the
CFD method to predict the waterjet pump performances, dynamic viscosity, Pt the turbulence viscosity. U is the
including the thrust distribution accurately. velocity vector, and Z angular velocity vector, r the
space vector. As the stator lies in the stationary
This paper with an emphasis on the effects of various
reference frame, the continuity equation remains the
number of blades on the waterjet thrust, torque,
same but the momentum equation changes to
efficiency and incipient cavitation performances. The
governing equations, numerical methods and the D
suitable computational domain for the waterjet ( UU ) Uf ’p  ( P  P t )’ 2 U ˄3˅
Dt
numerical simulation will be presented in section 2.
Section 3 will present the validation of numerical These equations along with the inlet velocities and outlet
method, based on the mesh resolution to obtain pressure boundary condition are solved for the waterjet
reasonable results for such devices flows, by comparing in-house flow field.
numerically calculated thrust and torque of the waterjet
under both design and off-design conditions with the The waterjet consists of inlet duct, a rotor-stator, and
waterjet-ship load-drive characteristics curve. The nozzle. Considering the effects on inflow caused by the
effects of number of blades to the propulsion ship speed and hull boundary layer, a sizable region
performances lie in section 4, which will address 3 under hull, named flow control volume (FCV), should
cases, ν) only change the rotor's blades form 5 to 7, be chosen with a care. So the stator(including the
ξ) only change the stator's vanes from 9 to 13, ο) nozzle), rotor, inlet duct(including the fairing) along
change the rotor and stator's blades simultaneously to with the FCV make up of the whole computational
reconstruct different devices. Section 5 will present the domain of the waterjet, see Fig.1. Fig.2 shows the mixed
incipient cavitation performances of the two optimum flow pump geometry.
combination pump above-argumentation. Section 6 will
summarize the results that have been obtained in this
study.

2. GEOMETRY AND COMPUTATIONAL

INVESTIGATIONS

The solver is a cell-centered finite-volume-based code

which solves the hydrodynamic equations as a single
system, and uses a fully implicit discretization of the Fig.1 Computational domain of the waterjet
equations at any given timestep. Turbulence is modeled
Nozzle
using SST turbulence model. It combines the N  H
and N  Z models. For the free stream region the
N  H model is used and for the near wall flow region
(y+<5) the N  Z model is applied. It has been shown
to eliminate the free stream sensitivity problem without
sacrificing the N  Z near wall performance[6]. It is
reckoned to perform very well close to walls in
Sta

boundary layer flows, particularly under strong adverse

ort

pressure gradients[6,7].
Fa
Rotor irin
Equations governing the turbulent incompressible flow gr
oo
within the rotor are formulated in a rotating reference m
frame. The continuity and momentum equations are:

Fig.2 Mixed flow pump geometry

©2008 The Royal Institution on Naval Architects

Waterjet Propulsion 5, London, UK

3. VALIDATION OF FLOW FIELD The mixed plane interfaces are used to handle with the
COMPUTATION rotating and stationary domains. Throughout the
computations presented in subsections, such schemes are
The grid used in the computations has about one million. used unless otherwise stated.
The inlet duct and the FCV are discretized into
unstructured cells as a whole. Special emphasis was put Results for the waterjet show comparison between
on modeling the inlet lip turbulent eddy motion and computed and manufacturer global data on design speed
computing streamwise acceleration of the boundary and revolution and off-design conditions, see Fig.4,
layer near the hull, so fine grid spacing are used in such where all wall forces due to pressure and wall shear
local region. The rest solution domain are filled with stress are integrated to get the thrust. All variables are
hexahedral cells. Considering the periodicity of the normalized by the design value. The power absorption
blade passages, Fig.3 shows the single blade to blade prediction differs by less than 2%, while the difference
channel surface mesh for both the rotor and stator. for the thrust is a litter higher, but still less than 4%.
1.2
manufacturer data
Interface 1.1
CFD prediction
1.0

Nondimensional thrust T
0.9
Outlet Inlet
0.8
0.7

0.6
0.5
0.4
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
Nozzle Stator Rotor Nondimensional speed Vs
(a)
Fig.3 Surface mesh of the waterjet single passage
1.2
An appropriate mesh density gives a compromise manufacturer data
between the accuracy and the computational cost. Miner 1.1
CFD prediction
[8] made a comparison of the velocity profiles with an
Nondimensional power N

1.0
axial flow impeller from two meshes, one with 22176
0.9
nodes and the other with 40131 nodes, and showed no
significant differences. A continuation of work referring 0.8
to a mixed flow pump (specific speed is 388.54) was 0.7
performed by him later[3]. Based on the experience
0.6
gained in the analysis of the axial flow impeller, the
rotor model has 26299 nodes and the stator with 20519 0.5
nodes, the shapes and magnitudes of the velocity and 0.4
0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
static pressure profiles were correctly predicted again. Nondimensional speed Vs
As the mixed flow waterjet pump for this study is close
to that of Miner both on the specific speed and design (b)
parameters, the mesh resolution of the rotor and stator
gets 20000 nodes as basic orders of magnitude Fig.4 Thrust and power versus speed: (a) thrust and (b)
throughout the computations presented in subsections. power
The height of the first cell adjacent to the blade surface
is approximately 0.00001 D1m , which is 3 to 50 in terms
4. EFFECTS OF VARIABLE NUMBLE OF
of y  for all surfaces, where D1m is the average BLADES ON PROPULSION
diameter at the inlet of the rotor. Ten layers of PERFORMANCES
hexahedral cells have been attached to the surface with a
grid-stretching ratio equal to 1.1. The number of the This section will describe different number of blades for
computational elements of the inlet duct and FCV rotor-stator interaction on the waterjet performances
domain is approximately 190000 after a comparison. with three steps. Firstly, just changing only the rotor's
Velocity components of uniform stream with the given blades from 5 to 7, the Stodola slip factor will change
inflow speed are imposed. On the nozzle exit boundary, subsequently. Secondly, increase of the stator's vanes
the static pressure is set to background pressure while resulting in different commutating action is of our
other variables are extrapolated, which assumes the jet interest. Finally, several configuration by changing both
diameter at vena contracta equals to nozzle diameter. of the rotor and stator's blades simultaneously, the

©2008 The Royal Institution on Naval Architects

 Waterjet Propulsion 5, London, UK

global quantities are compared with each other in order Z2

to select the optimum match of the two components. D 2 a2 D1 b 2
E 2 Eb 2

4.1 CHANGE ROTOR'S BLADES ONLY m2

a1 e2
The mixed flow waterjet pump consists of a six blades l
rotor and an eleven vanes stator. The design parameters

D2 m
D2 a
m1

D2 e
of the rotor on design ship speed and rotational speed
are specific speed ns=465.7, average diameter of the
rotor at inlet D1m=0.136m and at outlet D2m=0.137m, e1
width of outlet b2=0.073m, minimum and maximum

D1m
D3

D1a
diameter at outlet D2e=0.887m and D2a=0.950m

D1e
respectively, the mean blade angel at the discharge
Eb2 =21.6°, and the Stodola slip factor SS=0.803, which
are calculated based on literatures[2,9]. Changing only
Fig.5 Meridional surface and geometrical parameters
the rotor's blades, the solidity s l / t m will vary
significantly, where l is the chord of the blade 1.02

Normalized Head
89.2

Efficiency Eta/%
measured in the developed meridional plane (Fig.5), t m is 0.99 88.8
the cascade interval, while the response of the other 88.4
0.96
geometrical parameters are entirely negligible[10], see 88
Table1. Fig.5 shows a significant fraction of the 0.94 87.6
geometrical parameters on the meridional plane. On this 5 6 7 5 6 7
Number of blades Number of blades
surface, the subscripts 1 and 2 denote particular values
at inlet and discharge, and the corresponding subscripts (a) (b)
a , e and m relate to the tip, root and mean span 1.01 1.01

Normalized Power
Normalized Thrust

location respectively. The angles, D1 ǃ D 2 and D 3 are 1

blade angle at the discharge, flow angle on shroud side 0.97
0.99
and on hub side, which can be directly obtained form the 0.97 0.92
schematic in literature[9]. The condition
0.96
0.995 Vsdesign and 1.001 ndesign is used to analyze in the 5 6 7 5 6 7
Number of blades Number of blades
subsequent computations. The corresponding rotor
geometrical parameters with different rotor's blades are (c) (d)
compared in Table1. Fig.6 The waterjet propulsion performance curves
As shown in Fig.5, the flow angle E 2 is not identical to versus the number of rotor's blades: (a) head, (b)
the discharge blade angle E b 2 , and, therefore, implies an propulsion efficiency, (c) thrust (d) brake power.
effective slip S , which is due to the non-uniformity in
the discharge flow caused by the rotational flow within As shown in Fig.6, on the off-design condition, the
an individual blade passage. Stodola (1927) was among propulsion efficiency of the waterjet is highest in the
the first to recognize the importance of the rotational case of the rotor with five blades, while the lowest with
component flow. Busemann (1928) first calculated its seven blades; but both of the corresponding axial thrust
effect upon the head/flow characteristic for the case of and the brake power show an opposite way. According
infinitely thin blades. Stodola also gave the estimated to the definition of pump efficiency:
slip factor, S S , was S S 1 S sin E b 2 , where Z was the
Z UgQH
K (4)
number of blades[11]. As the number of blades gets N
larger, S S tends to unity as the rotational flow where,
increasingly weakens, which minimizes the decrease in Q =flow through the waterjet pump;
the head[11]. As shown in Table1, all the three rotor's H =head of the pump;
blade angle at the discharge is approximate 21.6e. N =brake power required from an engine
Fig.6 shows the three waterjet's head, propulsion
efficiency, axial thrust and brake power corresponding Although the head is maximum for seven blades, the
to three rotors with different blades. The head-blades power absorption increase more rapidly at increased
relationship is approximately linear. In practice, number of blades, hence a decline for the efficiency.
however, the frictional losses will increase with the According to the operational characteristics of the
number of blades. One popular engineering criterion waterjet, the power required with the seven blades rotor
(Stepanoff 1948) is that Z should be one third of the is larger than the delivered engine power, which will
discharge blade angle (in degrees). result in a heavy condition. On the other side, when the

©2008 The Royal Institution on Naval Architects

Waterjet Propulsion 5, London, UK

number of rotor's blades turns to five, a light condition nine to thirteen vanes, combines the origin six blades
will present, the engine is not able to deliver the rotor to be analyzed. And both of the shroud and hub for
maximum power again in such a case. Integrating the the rotor- stator interaction is the same as the origin. The
flow and propulsion predictions, the waterjet device computational condition is the same as section 4.1.
reaches the optimum propulsion performances with the
six blades impeller. The same conclusion can also be Fig.8 shows the iso-velocity plot distribution at the exit
obtained under the other four off-design points, see of the nozzle for five waterjet devices with different
Fig.7. number of vanes stator, the corresponding predicted
thrust and power can be found in Table2. As the number
0.97
Five blades rotor of vanes gets larger, the rotating component of the
Six blades rotor outflow increasingly weakens, however, it shows no
0.90 significant difference when it comes to 11. Uniting
Seven blades rotor
Nondimensional thrust T

Table 2, the head, propulsion efficiency, thrust and

0.83
brake power of the waterjet are close to each other for
the ten vanes stator and eleven vanes. It shows the
maximum thrust for the eleven vanes, while the
0.76 efficiency for the nine vanes stator is the highest, due to
its minimum frictional losses. Compared to the waterjet
with a eleven-vane stator, the thrust and brake power of
0.69
the waterjet with twelve or thirteen vanes are much
smaller, and the hydrodynamic characteristics for the
0.63 twelve and thirteen vanes are similar at the same time.
0.7 0.75 0.8 0.85 0.9 0.95 Integrating the flow and propulsion predictions, the
Nondimensional speed Vs waterjet device reaches the optimum propulsion
(a) performances with the eleven-vane stator.

0.92 Five blades rotor

Six blades rotor
Nondimensional power N

0.83 Seven blades rotor

0.74

0.64

0.55 (a) (b)

0.46

0.7 0.75 0.8 0.85 0.9 0.95 1

Nondimensional speed Vs

(b)
Fig.7 Waterjet propulsion predictions under off-design
conditions: (a) thrust versus speed and (b) power versus
speed (c) (d)

4.2 CHANGE STATOR'S VANES ONLY

The stator is just downstream of the rotor. It's blade

angle is designed to straighten the flow so that the fluid
leaves the stator nearly in axial direction. For reasons of
continuity, the decrease of the cross-section area of the
stator results in an increase of the kinetic energy and a
corresponding decrease of pressure. It recovers (e)
non-tangential component in idealism. According to the
momentum principle, the developed thrust of the Fig.8 Iso-velocity plot distribution at the exit of the
waterjet is closely relative to the velocity distribution at discharge nozzle with different number of vanes: (a) 9;
the exit of the discharge nozzle, which will be used to and (b) 10; and (c) 11; and (d) 12; and (e) 13
reflect the uniformity of the outflow. Considering
changing the origin eleven vanes stator, over a range of

©2008 The Royal Institution on Naval Architects

 Waterjet Propulsion 5, London, UK

1.02 90.0
4.3 CHANGE ROTOR'S BLADES AND STATOR'S

Efficiency Eta/%
Normalized Head
1.01 89.2
VANES SIMULTANEOUSLY 1.00
0.99 88.4
The stator of the mixed flow pump is 3-dimention with a 0.98 87.6
0.97
long axial distance but short spanwise. The rotor-stator 0.96 86.8
interaction should be taken into account when the 9 10 11 12 13 8 9 10 11 12 13
Number of stator's vanes Number of stator's vanes
stator's vanes are designed, so the vanes should be
matched the discharge of the rotor. Fig.9 shows the (a) (b)
pressure distribution of the origin waterjet pump's 1.01 1.04

Normalized Power
Normalized Thrust
meridional surface, including both of the rotor and 1.00 1.02
stator. The leading edge of the vane is approximately 0.99 1.00
parallel to the trailing edge of the rotor's blade, and the 0.98
0.98
0.96
axial distance between them are very small. Such stator's 0.97 0.94
vanes are generally called with big distortion. The 0.96 0.92
variety of rotor's blades and stator's vanes simultaneous 8 9 10 11 12 13 14 8 9 10 11 12 14
Number of stator's vanes Number of stator's vanes
will result in the change of the meridional surface of the
pump, hence of the streamline within the blade passage, (c) (d)
which will affect the waterjet propulsion performances.
Five blades rotor
The operational point is also the same as that in section
4.1. Six blades rotor
Seven blades rotor
For different combination between rotor's blades and
stator's vanes, Fig.10 shows the comparison of the head, Fig.10 Waterjet propulsion predictions for the
propulsion efficiency, thrust and the absorbed power of different combination between rotor's blades and stator's
the different waterjet devices. For the waterjet attached vanes: (a) head, (b) efficiency, (c) thrust and (d) power
by the six blades rotor, all the global quantities lie
intermediately compared to that of five-blade rotor and
seven-blade rotor matched by five different stator's 5. COMPARISON OF PUMP INCIPIENT
vanes. Under the same operational point, the head of the CAVITATION PERFORMANCES WITH 6
waterjet with seven blades rotor gets the maximum BLADES ROTOR TO 7
value, but its efficiency is the lowest, however, as for
five blades rotor, it is just on the contrary, which is the Since we have concluded that the optimum propulsion
same as just changing the rotor's blades only. All the performances of the waterjet can be presented by
global quantities of the six blade rotor integrating with combination of 6 or 7 blades rotor and 11 vanes stator,
the series of stator's vanes vary similarly with that of now form the pump incipient cavitation point of view
seven blades rotor. When the rotor's blades is five, the there are some basic introduce. Cavitation is defined as
optimum propulsion performances can be obtained by the process of formation and disappearance of the vapor
attaching the nine vanes stator. Another optimum match phase of a liquid when it is subjected to reduced and
is seven blades rotor and eleven vanes stator, which is subsequently increased pressure at constant ambient
the same as the combination for six blades rotor. temperatures. The potential for cavitation is typically
evaluated in terms of cavitaion parameters: net positive
Exit of rotor suction head, NPSH, which is regarded as a measure for
the margin against vaporization of the fluid entering the
pump. The formula to compute it reads:

p01  pv
NPSH (5)
Ug

Inlet of stator In which p01 is the upstream (suction) total pressure,

pv is vapor pressure, and g is acceleration due to
gravity. The total pressure equals

1
Fig.9 Pressure distribution of the origin waterjet p01 p1  UV 2 (6)
pump's meridional surface 2

Where p1 is upstream (suction) static pressure, and

V is suction velocity. The incipient cavitation
characteristic plays a key role when designing and

©2008 The Royal Institution on Naval Architects

Waterjet Propulsion 5, London, UK

evaluating rotors with regards to suction performance. in which p 1' is the pipe inlet static pressure, see
By definition (5) one has
Fig.11. Then, substituting equation (5) and (6) in
equation (10) it gets:
p01, i  pv
NPSH i (7)
Ug pv' p '01  UgNPSH (11)

where p01,i is the total upstream (suction) pressure,

where p '01 is the total pressure of the pipe inlet. So the
associated with the situation that cavitation starts
somewhere downstream. At a given condition it exists a region p  pv can be visualized during
minimum pressure pmin at a particular location post-processing after the CFD run through the
downstream in-house the pump. Considering the two iso-surface of pv' .
streamline through point of pmin and pv respectively,
corresponding to the condition of p1 and p01,i . It Fig.12 gives an comparison of the local low pressure
follows that: region above-analyzed between the pumps with 6 blades
rotor and 7 blades. For both the pumps the region starts
p1  p1,i p min  pv (8) a little distance after the blade leading edge, and has
longest streamwise length near the rotor shroud. While
the 6 blades rotor pump has the larger region for the
Substituting equation (8) in equation (7) and (5) it gets: bigger single blade load. Fig.13 shows the comparison
of blade load at 0.7 times span location of the two pump.
p01  pmin The pressure side of the rotor with the number of 6 is
NPSH i (9)
Ug higher around the leading edge. The CFD calculation
states that the NPSH i of the 6 blades rotor pump is
Equation (9) stated that incipient NPSH can be obtained 114.0m, which is smaller than the 7 blades rotor pump
with total upstream pressure and the minimum pressure. of 126.5m under the same design condition
Here the fluid is assumed to be pure, without dissolved above-argumentation, indicating a easier susceptibility
gas. to incipient cavitaion.

Next it turns to the waterjet pump. The real inlet duct

and ship hull geometry are simplified as a straight
suction pipe, see Fig.11, so the inflow velocity
distributions are uniform. Bulten(2006) analyzed the
influence of non-uniform axial inflow in detail in his
Ph.D dissertation. For the origin 6 blades rotor and 11
vanes stator combination, by CFD calculation it states
that under the designed rotating speed and capacity
(a) (b)
condition the axial force of the pump enlarges 20% with
the shaft power decreases 0.7% at the same time, so the
efficiency of the pump increases with the uniform Fig.12 Comparison of the local p  pv region
inflow. between the two pump: (a) 6 blades rotor-11 vanes
stator; (b) 7 blades rotor-11 vanes stator
8
6 blades rotor pump
6 7 blades rotor pump

0
Cp

-2

Fig.11 Sketch of waterjet pump with simplified inlet -4

-6
To determine the region where the local static pressure
-8
drops below the vapor pressure,(i.e. p  pv ), another
-10
parameter pv' is introduced, with reference to equation -0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
streamwise(0 1)
(8), the formula reads:
Fig.13 Comparison of the pressure coefficient at
pv' pv  ( p 1'  p1 ) (10) 0.7span location between the 6 blades and 7 blades

©2008 The Royal Institution on Naval Architects

 Waterjet Propulsion 5, London, UK

6. CONCLUSIONS 8. REFERENCES

Based on CFD, the flow pattern and propulsion 1. STEPANOFF A J. Centrifugal and Axial Flow
performances of the waterjet are analyzed on both Pumps[M]. New York: John Wiley & Sons.INC.,1957.
design and off-design conditions. Effects of the rotor's
blades and stator's vanes to the propulsion performances 2. GUANG Xingfang. The Master Handbook of Modern
are mainly investigated, and the results obtained are as Pumps[M]. Beijing: Press of Astronautics, 1995.(in
follows: Chinese)

(1) Given the operational points with dynamic 3. MINER S M. CFD Analysis of The First-Stage Rotor
parameters, the propulsion efficiency reaches the highest and Stator in A Two-Stage Mixed Flow Pump[J].
with the five blades rotor configuration, the lowest with International Journal of Rotating Machinery, 2005(1):
the seven blades rotor; while both the thrust and brake 23-29.
power are just on the contrast, the lowest with five
blades rotor and the highest with seven blades rotor. As 4. HU P, ZANGENEH M. CFD Calculation of The
the increase of rotor's blades number, the Stodola slip Flow Through A Waterjet Pump[C]. International
factor gets larger, which weakens the rotational flow Conference on Waterjet Propulsion III, Sweden: RINA,
within the blade passage, hence results in the increase of 2001: 1-10.
the head.
5. BULTEN N. Numerical Analysis of a Waterjet
(2) Just changing the stator's vanes only, the rotating Propulsion System[D]. PhD dissertation, Library
discharge flow at the exit of the nozzle tends to Eindhoven University of Technology, Netherlands, 2006.
uniformity as the vanes' number get larger, but the
improvement is approximately unchangeable when the 6. LLOYD G, ASTILLEROS E. Best Practice
vanes comes to 11, and makes the efficiency decline due Guidelines for Marine Applications of Computational
to frictional losses at the same time. Under the certain Fluid Dynamics[R]. WS Atkins Consultants and
operational point, compared to the waterjet with a members of the NSC, MARNET-CFD Thernatic
eleven-vane stator, the thrust and brake power of the Network, 2002: 37-42.
waterjet with twelve or thirteen vanes are much smaller,
and the hydrodynamic characteristics for the twelve and 7. MENTER F R. Ten Years of Industrial Experience
thirteen vanes are similarity as well. with SST Turbulence Model[C]. Fourth International
Symposium on Turbulence, Heat and Mass Transfer
(3) Increasing the rotor's blades from five to seven, and (ICHMT), Turkey: Antalya, 2003: 625–632.
changing the stator's vanes from nine to thirteen at the
same time, it can reconstruct 15 different waterjet 8. MINER S M. 3D Viscous Flow Analysis of An Axial
pumps. In which the characteristics for the series of six Flow Pump Impeller[C]. International Symposium on
blades rotor lie intermediately compared to that of Transport Phenomena and Dynamics of Rotating
five-blade rotor and seven-blade rotor matching five Machinery (ISROMAC’96), Hawaii: Honolulu, 1996, 2:
different stator's vanes respectively. The waterjet series 336-344.
with five-blade rotor get the maximum efficiency and
minimum head, while the waterjet series with 9. HE Xijie, LAO Xuesu. Design Method of
seven-blade rotor are on the contrast. A five blades rotor Mixed-Flow Pump Impeller's Blade[J]. Journal of Hebei
combining a nine vanes stator can obtained the optimum Engineering and Technical College, 2000(1): 1-5.(in
propulsion performances, when the number of rotor's Chinese)
blades is six or seven, the optimum match is both the
eleven-vane stator, while the 6 blades rotor pump's local 10. YAN Bipeng, TANG Fangping. Study on The
low pressure region is bigger and so is easier to incipient Performance of Axial-Flow Pump by Changing The
cavitation. Number of Its Blades[J]. Journal of Yangzhou
University, 1998, 1(3): 53-55.(in Chinese)

7. ACKNOWLEDGEMENTS 11. BRENNER C E. Hydrodynamics of Pumps[M].

Vermont: Concepts ETI, Inc., 1994.
This research was sponsored by one of the Office of
Naval Research Grant under the administration of Dr. 9. AUTHORS BIOGRAPHY
Yongsheng Wang and partially supported to prepare this
paper by the basic research project in E450. Qiongfang Yang received a B.S. degree (Ship Power
Engineering, 2006) and an M.S. degree (Marine
Engineering, 2008) from Naval University of
Engineering, and now is a Ph.D candidate within the
marine engineering realm. He is a member of the Office

©2008 The Royal Institution on Naval Architects

Waterjet Propulsion 5, London, UK

of New Technology of Marine Propulsion, has received Zhihong Zhang holds the current position of Director at
the first scalar award in Hubei province for his B.S. the Department of Fluid Mechanics, professor.
thesis, and has written 8 papers in the area of marine
propulsor up to now. Mingmin Zhang is a member of the Institute of
Underwater Acoustics, professor, mainly focuses on the
Yongsheng Wang holds the current position of Director underwater radiated noise of the navy machinery.
at the Office of New Technology of Marine Propulsion.
He is Professor of Marine Engineering in the University
of Naval Engineering.

TABLE 1 Geometrical parameters of the rotor's blade

Number Specific Width Average Average Velocity Maximum Minimum Blade Stodoal
of speed of diameter diameter meridional diameter diameter angle slip
blades outlet at inlet at outlet at outlet at outlet at outlet e) factor
(m) (m) (m) (m/s) (m) (m)
         
         
         

TABLE 2 Waterjet propulsion prediction with 6 blades rotor attached by different stator's vanes

Number of vanes Thrust (kN) Brake power (kW) Propulsion efficiency (%) Head (m)
    
    
    
    
    

©2008 The Royal Institution on Naval Architects

Waterjet Propulsion 5, London, UK

A MULTI-OBJECTIVE AUTOMATIC OPTIMIZATION STRATEGY FOR DESIGN OF

WATERJET PUMPS

M. Zangeneh, University College London, UK

K. Daneshkhah and B. DaCosta, Advanced Design Technology Ltd, UK

SUMMARY

A methodology is presented for designing waterjet pumps to meet multi-objective design criteria. The method combines
a 3D inviscid inverse design method with multi-objective genetic algorithm to design pumps which meet various
aerodynamic and geometrical requirements. The parameterization of the blade shape through the blade loading enables
3D optimization with very few design parameters. A generic pump stage is used to demonstrate the proposed
methodology. The main design objectives are improving cavitation performance and reducing leading edge sweep. The
optimization is performed subject to certain constraints on Euler head, throat area, thickness and meridional shape so
that the resulting pump can meet both design and off-design conditions. A Pareto Front is generated for the two
objective functions and 3 different configurations on the Pareto front are selected for detailed study by 3D RANS code.
The CFD results confirm the main outcomes of the optimization process.

1. INTRODUCTION Since the viscous losses and cavitation behaviour in the

pump is to large extent controlled by the 3D pressure
Waterjets are increasingly being used as the propulsive distribution, by using this approach one can obtain a
method of choice for high speed marine vehicles. A more direct control over the design process.
pump designed for waterjet application should have high
propulsive efficiency, compact size and low entrained A 3D inverse design approach, TURBOdesign-1 [4],
water. Furthermore it needs to withstand non-uniform that has been extensively used in pump design is that
inflow conditions, which have an adverse effect on its proposed in [5], in which a 3D method is used to design
hydrodynamic and suction performance. In order to meet the blade geometry subject to specified blade loading
these contrasting requirements, a design strategy is distribution and blade thickness. This method has been
required that can consider a large part of the design space used to improve exit flow non-uniformity from
and provide information on trade-offs between centrifugal and mixed flow impellers [6] and eliminate
contrasting design objectives. corner separation in diffusers [7]. Furthermore it has
proved to be very effective in developing a very compact
Traditionally waterjet pumps are designed based on mixed flow pump [8]. Development of a compact pump
empiricism and trial and error approach, in which the is particularly important in marine waterjet applications
impeller and diffuser geometry are modified manually by as it reduces the weight of the waterjet system and the
changes to blade angle distribution. The flow through the amount of entrained water. However, as shown in [8],
resulting impeller (or diffuser) is then analysed by some reduction of volume of the mixed flow pump by 60% is
form of quasi-3D (Q3D) or 3D numerical method. For only possible by very careful control of cavitation
example, in [1] and [2] a lifting surface method is phenomena. However, a design that improves cavitation
proposed for this purpose. While in [3] a combination of may have an adverse effect on performance or on Static
Q3D inviscid method together with a 3D RANS code is or dynamic stresses. Hence the importance of a multi-
proposed. These methods are then used to evaluate the objective design which takes account of the different
flow in the pump. However, these methods do not contrasting requirements.
provide any guidance on how the blade geometry should
be modified in order to improve the flow field and hence TURBOdesign-1 has already been coupled to automatic
the designer has to rely on trial and error. Such a trial and optimizers and 3D RANS code for minimizing losses [9]
error process, however, by its nature restricts the and cavitation performance [10]. Parametrizing the blade
designer to relatively small part of the design space, in terms of blade loading parameters enable one to
limited to blade angle distributions that have worked in represent large part of the design space with a few design
the past, and does not allow the easy exploitation of a parameters and hence provide distinct advantages in 3D
wide part of the design space. optimization of turbomachinery blades, see [11]. In this
paper we propose a new strategy in which
An alternative approach to the design of the pump stage TURBOdesign-1 is coupled with Multi-objective Genetic
is to use an inverse method. In such an approach the Algorithm in order to create a Pareto Front for the
impeller or diffuser geometry is designed for a specified contrasting design requirements.
distribution of pressure distribution or blade loading.

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

2. METHODOLOGY

2.1 INVERSE DESIGN METHOD

The commercial software TURBOdesign-1 [ 4] is used to

parametrically describe the blade geometry.
Turbodesign-1 is a three-dimensional inviscid inverse
design method, where the distribution of the
circumferentially averaged swirl velocity ( rVT ) is
prescribed on the meridional channel of the blade and the
corresponding blade shape is computed iteratively.

The circulation distribution is specified by imposing the

spanwise rVT distribution at leading and trailing edge
and the meridional derivative of the circulation wrVT /wm
(blade loading) inside the blade channel. The input Fig. 1 The blade loading parameters used in the
design parameters required by the program are the optimization
following:
2.2 MULTI-OBJECTIVE OPTIMIZATION
x Meridional channel shape in terms of hub, METHODOLOGY
shroud, leading and trailing edge contours.
x Normal/tangential thickness distribution. Optimization Technique Description:
x Fluid properties and design specifications.
x Inlet flow conditions in terms of spanwise A Non-dominated Sorting Genetic Algorithm (NSGA-II)
distributions of total temperature and velocity [12] is used for the multi-objective optimization. This
components. technique is well-suited for highly non-linear and
x Exit rVș spanwise distribution. By controlling its discontinuous design spaces. Each objective is treated
value, the work coefficient (or Euler head) are separately and a pareto front is constructed by selecting
fixed. feasible non-dominated designs. Standard genetic
x Blade loading distribution ( wrVT /wm ). It is operation of mutation and crossover are performed on the
imposed at two or more span locations between designs. Selection process is based on two main
hub and shroud. The code then automatically mechanisms, "non-dominated sorting" and "crowding
interpolates in the spanwise direction to obtain distance sorting". By the end of the optimization run a
the two-dimensional distribution over the pareto set is constructed where each design has the "best"
meridional channel. combination of objective values and improving one
x Stacking condition. The stacking condition must objective is impossible without sacrificing one or more
be imposed at a chord-wise location between of the other objectives.
leading and trailing edge. Everywhere else the
blade is free to adjust itself according to the population size 80
loading specifications. number of generations 200
crossover probablity 0.9
In this optimization, the meridional shape, blade crossover distribution 10
thickness and Euler head were fixed and the blade index
loading and stacking conditions were modified. The crossover distribution 20
blade loading parameters are shown in Fig. 1. The values index
of blade loading can be specified at a number of initialization mode random
streamlines. In Fig. 1 the values are shown at the hub and Table 1 – The setting used for optimization using NSGA-
shroud streamlines. The loading on each streamline is II
defined by a parabolic distribution from leading edge to a
user defined point (NCHUB or NCSHROUD) , followed by a
straight line section that the user can specify the slope 3. DESCRIPTION OF TEST CASE
(SLOPEShroud or SLOPEHub) and then another parabolic
section starting from ND that brings the loading down to The pump stage used for this study is a generic mixed
zero at the trailing edge in order to satisfy the Kutta flow stage with specific speed of 946 (based on rpm, m
conditions. The value of loading at the leading edge is an and m3/min).
important parameter that affects the blade incidence and
hence can be used to adjust the peak efficiency point of
the design.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

3.1 OPTIMIZATION TARGET AND DESIGN 3.3 CONSTRAINTS

PARAMETERS
In section 2, it was mentioned that the meridional shape
In this study, the blade loading was specified on 3 and blade thickness are fixed during the optimization.
streamlines at hub, midspan and shroud. In order to Another important feature of TURBOdesign-1 is that the
reduce the number of design parameters, the values of Euler head is specified in the program and hence during
NC and ND on all streamlines were fixed at 0.2 and 0.8 the optimization process the Euler head and specific
(i.e. at 20% and 80% of meridional chord) on each work are fixed. In addition to these constraints, which are
streamline and also the value of leading edge loading on implicit to the inverse design process and hence do not
the shroud was fixed at zero so that zero incidence is need to be specified explicitly in the optimizer ( which
maintained on the shroud during the optimization. So the can affect the convergence of the optimizer), 3 additional
design variables modified by the optimizer were as constraints were imposed in the optimization.
shown in Table 2.
The first constraint was on the throat area. Since the
Design Main Blade throat area of the impeller can change by changes to
blade loading distribution the value of throat was set to
Parameter Min Max
vary no more than ±4% of the throat area of the baseline
LEH Ͳ0.3 1.0 impeller in order to ensure that correct peak efficiency
LEM Ͳ0.2 0.5 flow rate is maintained. Furthermore, additional
SlopeH Ͳ2.0 0.5 constraints were imposed on the diffusion ratio on the
shroud, based on the TURBOdesign-1 predicted blade
SlopeM Ͳ2.0 0.5 surface velocity distribution, to limit the possibility of
SlopeS Ͳ2.0 0.5 flow separation. Constraints were also placed on the
Table 2 – Range of Design Parameters used in maximum value of leading edge sweep to be less than
Optimization 106, a value slightly more than that of the baseline
impeller.
So overall 5 design parameters were used to modify the
blade loading. These include the leading edge loading on
the hub (LEH) and midspan (LEM) and slope of loading 4. RESULTS
curve on the hub, midspan and shroud (SlopeH, SlopeM
and SlopeS). In addition the blade was stacked at the The result of the optimization, which correspondes to
trailing edg. The value of wrap angle was fixed as zero at 16,000 different impeller geometries was obtained in
the hub and varied between -10o and 10o at the shroud by about 31 hour of computation on a single core of a P4
the optimizer. So in total 6 design parameters were used. processor PC. The results are summerized in the Pareto
front plot, shown in Fig. 3.
3.2 DESIGN TARGET

The purpose of this multi-objective optimization is to

minimize at the same time criteria for cavitation and
impeller leading edge sweep. High leading edge sweep
can help cavitation performance but can have an adverse
effect on manufacturing or impeller stresses. For
cavitation criteria the value of minimum static pressure
as predicted by TURBOdesign-1 was used.
TURBOdesign-1 provides very accurate prediction of the
surface static pressure as compared to measurement and
CFD predictions. Hence by asking the optimizer to
maximize the minimum static pressure one should arrive
at a design that has improved suction performance. The
criterion used for leading edge sweep can be more easily
defined in terms of the blade shape. In this case we
defined ratio of the difference in Cartesian coordinates of
the hub to shroud at leading edge divided by the Figure 3: The Pareto Front for the Optimization
meridional shape difference multiplied by 100. So in the
case of no sweep a value of 100 will be obtained and the In Fig. 3, the minimum pressure (on vertical axis) and
greater the value of sweep ratio is over 100 the higher is sweep ratio (on the horizontal axis) for every single
the leading edge sweep. configuration obtained by the optimizer are plotted. Each
blue point corresponds to one configuration designed by
TURBOdesign-1 through the modification made to blade
loading by the optimizer. The higher the minimum static
pressure for a configuration, the higher its suction

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

performance. The red dots represent the Pareto Front or pressure and flow angles and the mass flow rate
the “optimum” sets of configurations for the design boundary condition is used at the outflow.
space.
For cavitation analysis, a two phase Rayleigh-Plesset
Three impeller designs along the Pareto front (shown by model is used. The interphase transfer is governed by a
A, B and C) are selected for further detailed study. In mixture model where the interface length scale is 1 mm.
Fig. 4, comparison of the leading edge shape of the A Flow is assumed to be homogeneous and isothermal at
impeller impeller C is presented. One can see clearly that 293.15 K. The saturation pressure is 3619 Pa and the
the A impeller has, as expected, a significantly reduced mean nucleation site diameter is 2e-03 mm.
leading edge sweep as compared to impeller C.
4.2 COMPARISON OF BASIC PERFORMANCE
PARAMETERS

The steady CFD computations of the stage were

performed at various flow rates to be able to make a
comparison of the stage performance with the 3 different
impellers obtained from the optimization versus the
(a) Impeller A (b) Impeller C baseline. In Fig. 5, the normalised pump head versus
flow rate of the 3 different stages are shown. In each
Fig. 4: Comparison of the leading edge shape of Impeller case, the same diffuser was used. The results indicate that
A and C the predicted pump head for impellers A and B is very
similar to the baseline values across all the different flow
In order to make a detailed comparison of the flow field rates. Impeller C, however, seems to have a slightly
in the impellers A to C 3D RANS code was used. lower head.
4.1 DESCRIPTION OF 3D CFD METHOD

CFD computations are performed using ANSYS CFX

which is widely used in industry for turbomachinery flow
simulation. The computational domain consists of a
rotating domain (impeller) and a stationary domain
(diffuser). A structured H-O topology is used to construct
the mesh. A tip clearance of 0.5 mm is used for the
impeller blade. The mesh consists of 300K and 175 K
elements for the impeller and diffuser domains,
respectively. Figure 5 shows the details of the
computational mesh.

Fig. 5 Comparison of normalised predicted head versus

flow rate for different stages.

Figure 5: Computational Mesh

The incompressible RANS equations are solved

simultaneously for the stage configuration. A two-
equation k- turbulence model with scalable wall
functions is used. Flow is assumed to be axisymmetric,
so that only one passage is modelled in each domain and
a non-overlapping mixing plane interface is used
between the impeller and the diffuser domains. The
inflow boundary conditions are total pressure, total
Fig. 6: Comparison of predicted impeller efficiency

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

In Figure 6 the predicted impeller efficiency for the

different cases are compared. Again the results are
normalised with the maximum impeller efficiency of
impeller A.

Finally the stage efficiency normalised by the maximum

efficiency of impeller A are shown in Fig. 7. The results
confirm that all 3 optimization cases are achieving the
same pump stage peak efficiency flow rate as the
specified flow rate.

(b) Impeller B

Fig. 7 Comparison of Predicted Stage Efficiency

4.3 COMPARISON OF CAVITATIONS

PERFORMANCE

In order to compare the performance of the different

impellers in terms of their cavitation performance, two
phase cavitation analysis was performed by reducing the
inlet total pressure at the design flow rate. The results at
a total pressure corresponding to initiation of cavitation
at the design flow rate is shown in Fig. 8. The results
confirm the trend expected from the Pareto front in Fig. c) Impeller C
3, in which impellers A and B have a value of min
surface static pressure which is similar, and as expected Fig. 8. Cavitation Analysis for different impellers
their cavitation performance is similar. However,
impeller C has clearly a better cavitation performance but
higher leading edge sweep values. 5. CONCLUSIONS

A methodology is presented for multi-objective design of

waterjet pumps in which the blade shape is parametrized
in terms of blade loading parameters used as input in a
3D inverse design code. By using the 3D inverse design
code its possible to perform 3D multi-objective
optimization with only 6 design parameters. Computing
the performance objectives and constraints directly from
the output of the inverse method (3D pressure or velocity
field and geometry data) makes the evaluation of each
configuration very rapid. A relatively large population of
16000 design configurations can be computed in about
24-33 hours of CPU time on a single processor PC. The
results shown confirm that by imposing the correct
constraints it is possible to achieve designs that meet
both design and off-design objectives. The proposed
approach can be used to rapidly and automatically
(a) Impeller A explore a large part of the design space to create impeller

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

or diffuser designs that meet contrasting requirements 12. Deb,K., Pratap, A., Agarwal,K. and Meyarivan, T.
relating to efficiency, suction performance, 2002, A Fast and Elitist Multiobjective Genetic
manufacturing limitations, mechanical constraints or Algorithm: NSGA-II, IEEE Transactions On
cost. Evolutionary Computation, Vol. 6, No. 2, pp. 182-189.

6. REFERENCES 7. AUTHORS BIOGRAPHY

1. Taylor, T.E. and J.E. Kerwin (1998), ‘Waterjet pump Mehrdad Zangeneh is Professor of Thermofluids at
design and analysis using a coupled lifting-surface and University College London. His research interest relate
RANS procedure’, International Conference on Waterjet to development of design methods for turbomachinery
Propulsion Latest Developments, 22 & 23 October 1998, based on 3D inverse design approach and automatic
Amsterdam. optimization methods.

2. Taylor, T.E., and R.W. Kimball (1999), ‘Experimental Kasra Daneshkhah is a senior design engineer at
validation of a coupled lifting-surface/RANS procedure’, Advanced design technology and has been involved in
5th International Conference on Fast Sea Transportation, design of various industrial and marine pumps as well as
Seattle, Washington, USA, Sept 1999. low speed fans.

3. Huntsman, I., Hothersall, R., 2001, “Developmet of Bruno Da Costa was an intern student at Advanced
Quasi 3D Methods and 3D Flow Solvers for the Design Technology from University of Centrale Paris,
Hydrodynamic Design France.
of Water Jets”, Proceedings of Waterjet Propulsion III
Conference.

4. Turbodesign-1. Advanced design technology Ltd.

5. Zangeneh, M., 1991, “A 3D Design Method for Radial

and Mixed-Flow Turbomachinery Blades”, Journal of
Numerical Methods in Fluids, 13, pp. 599-624.

6. Zangeneh, M., Goto, A., Harada, H., 1998, “On the

Design Criteria for Suppression of Secondary Flows in
Centrifugal and Mixed Flow Impellers”, Journal of
Turbomachinery, 120, pp. 723-735.

7. Goto, A., Zangeneh, M., 2002, “Hydrodynamic Design

of Pump Diffuser Using Inverse Design Method and
CFD”, Journal of Fluids Engineering, 124, pp. 319-328.

8. Goto, A., Ashihara, K., Sakurai, T., Saito, S., 1999,

“Compact Design of Diffuser Pumps Using Three-
Dimensional Inverse Design Method”, ASME Paper
FEDSM99-6847.

9. C. Yiu and Zangeneh, 2000, "Three Dimensional

Automatic Optimisation method for Turbomachinery
blade design" Journal of Propulsion and Power Vol. 16.
No. 6, Nov-Dec 2000.

10. Ashihara, K., and Goto, A., 2001, “Turbomachinery

Blade Design Using 3D Inverse Design Method, CFD,
and Optimization Algorithm,” ASME Paper No. 2001-
GT-0358.

11. Bonaiuti, D and Zangeneh, M. , 2006, “On the

Coupling of Inverse Design and Optimization
Techniques for Turbomachinery Blade Design”, ASME
Paper GT-2006-90897. To be published in ASME
Journal of Turbomachinery.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

WATERJET PUMP DEVELOPMENT FOR HIGH PERFORMANCE AND HIGHER

POWER DENSITY
R Aartojärvi and M Heder, Rolls-Royce AB, Sweden

SUMMARY

This paper presents the development of a new high performance mixed-flow waterjet pump with a higher power density
than the previous design. The design objective has been to improve the cavitation performance of the pump in order
enable a significant size reduction for a given ship speed and engine power. This will result in lower weight the pump
unit as well as for the whole propulsion system. Size reduction has a positive effect also on the inflow to the pump and
thereby the interaction between inlet duct and pump. An increase of the power density of the waterjet unit presents a
number of challenges. One is to maintain high propulsive efficiency. Another challenge is to cope with the structural
loads. The aim of this paper has been to focus on structural issues in the design process of the new pump.

1. INTRODUCTION The 3D RANS model, illustrated in Fig. 1, is built using

a mixing plane approach together with rotationally
A pump designed for a waterjet application should be periodic boundaries. This approach limits the model to
part of a robust and reliable propulsion unit providing a only one impeller blade and one guide vane. The
high propulsive efficiency. This can be accomplished by computational grid is built using a hexahedral multi
low weight and small dimensions. The new pump needs block topology including the rotor tip gap. Apart from
to cope with various inflow conditions without cavitation the boundary conditions the model is set up with SST k-
erosion and structural problems. These and other Z turbulence model and higher order discretization
requirements should all have an impact on the design. schemes for the solver.

A new high performance waterjet pump has been

developed. This paper describes the design process
where CFD, model tests and structural analysis have
been integrated. The result is a mixed flow pump where
the leading edges of the impeller blades have been swept
forward.

This paper is focused on the development of the impeller

as this component presented the greatest challenge.
During the development process pressure load obtained
by CFD was applied to a FE model of the impeller in
order to compute static stress. The result of the FE Figure. 1 top view of the CFD model including blades
analysis was then verified by static stress obtained in (red), hub (blue) and periodic boundaries (yellow). The
model test. Dynamic stress of the impeller blades was mixing plane (green) is also highlighted in the centre.
obtained from model tests under a number of different
operating conditions. The results are used both to predict pump performance
and to provide a mean load field on the blades later to be
Rolls-Royce AB has developed a modern design used in FE analysis, see Fig. 2. The dynamic load due to
environment based on waterjet system simulations, non-uniform inflow and cavitation are treated by
advanced pump design tools, the extensive use of experimental methods.
analysis tools as well as model scale and full scale
experimental testing. The importance of making best use
of a combination of analytical and experimental methods
is emphasized in this paper.

2. BLADE LOADS BY CFD

The new pump design was developed in three steps using

a 3D inviscid design code, 3D RANS simulations and
experimental tests at Rolls-Royce Hydrodynamic
Research Centre. Figure 2. A contour plot of static pressure acting on the
rotor blade

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

3. MODEL TEST the waterjet system describing limitations for operation

due to cavitation erosion, thrust breakdown etc.
3.1 TEST METHODS
Another important objective for the waterjet system test
In order to determine the final hydrodynamic is to obtain loads acting on different parts of the
performance of the pump a number of model tests need structure, with the focus on the impeller. Strain gauge
to be carried out to supplement the results from CFD measurements on impeller blades are mainly used to
analysis. In this section these methods are described very obtain load variations caused by the non-uniform inflow
briefly. to the pump or the flow interaction with the guide vanes,
see Fig. 5.
A pump loop test set-up, as shown in Fig. 3, gives the
overall performance of the pump in terms of head rise,
power, pump efficiency and cavitation performance as a
function of the flow rate. This test is however done under
ideal conditions since the inflow to the pump is uniform,
which is not true for a waterjet application but it is
valuable as a starting point and verification of the CFD
results.

Figure 5. Strain gauges on impeller pressure side.

3.2 TEST RESULTS FOR THE NEW PUMP

Pump loop tests and waterjet system tests have shown

that the main design objective for the new pump, to
improve the cavitation performance of the pump to
enable a 12% size reduction, was accomplished. In
addition the pump efficiency was increased as well as the
Figure 3. Set-up for pump loop tests at RRHRC. actual flow rate to head rise ratio. This means that the
propulsive efficiency of the new pump is 2-3% higher
The effect on pump performance of the inflow non- compared to the previous design, which is quite a
uniformity needs to be measured in a waterjet system test significant improvement.
set-up as shown in Fig. 4. This test gives the interaction
between the inlet duct and the pump and can be Strain gauge measurements on the pressure side close to
compared to corresponding system analysis using CFD. the leading edge of an impeller blade gave the strain
The difference to the pump loop results increase with the amplitudes used in the structural analysis together with
degree of inflow non-uniformity, which in turn is FE computations of the average strains using pressure
determined by the operating condition of the waterjet loads from CFD analysis in uniform flow.
easiest described by the inlet velocity ratio (IVR).
It is mainly the non-uniformity of the inflow and the
cavitation number that determine the strain amplitude in
the leading edge of the impeller blade. The effect of
cavitation is less straight forward than that of the non-
uniformity, since cavitation can reduce the strain
amplitude close to the leading edge and transfer load
further downstream. Systematic measurements were
done to find the decisive load case. A typical strain signal
from the measurements is shown in Fig. 6.
Figure 4. Set-up for waterjet system tests at RRHRC.

The waterjet system test is also essential to adjust the

preliminary cavitation performance obtained in the pump
loop test. The result is the so-called cavitation zones for

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

Figure 6. Strain signals from the 4 gauges on the pressure Figure 7. FE model of impeller.
side in Fig. 5.

The gauge named PS1 is located close to the leading

edge just outside of the fillet. Gauge PS2 and PS3 are
located aft along the chord. PS4 is located further out on
the blade, see Fig. 5.

4. FE ANALYSIS OF IMPELLER

The mean stress in full scale has been computed

employing a FE model representing the complete
impeller, see Fig. 7. The FE model is composed of 10-
node parabolic tetrahedrons. One of the blades has been
meshed with a finer mesh than the remaining blades in
order to reduce the size of the model. The FE model has
been constrained in all directions on the pump shaft
interface.

Pressure obtained from CFD shown in Fig. 2 has been

scaled to full-scale conditions and applied to the outer
surface of the FE model. The effect of rotational load has
also been included in the analysis.
Figure 8. Maximal principal stress of impeller.
Fig. 8 shows the resulting principal stress corresponding
to a 45 knots operating condition. Due to the forward The strain obtained in the FE analysis is compared to the
swept blade the highest stress will occur in the area close result of the model test in Fig. 9, where the strain has
to the leading edge. In order to optimize the use of been normalized with respect to the highest strain
material the blade thickness variation results in a obtained in the model test. The result of the model test
constant stress over a large area. The stress in the hot has been scaled to correspond to the same size and
spot area is considerably lower than the yield strength of operating condition. The strain obtained in the FE
the material and well below the design limit for static analysis is computed in the same direction as of the strain
stress. gauges. The strain is here compared in four positions, all
located within the red area in Fig. 8. The gauges named
PS1 – PS3 are all located just outside the fillet on the
pressure side, PS1 closest to the leading edge and PS2
and PS3 further aft. The gauge PS4 is located further out
on the blade. The location of the gauges can be seen in
Fig. 5. The comparison shows a good agreement and the
difference between the results is generally within 10%. A
number of operating conditions has been analyzed with
similar results.

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

Figure 9. Non-dimensional strain obtained in model scale

and by FE analysis. Figure 10. Non-dimensional stress amplitude in gauges.

The smean stress of the new impeller is on approximately Fig. 10 shows the non-dimensional dynamic stress
the same level as of the present impeller. The resulting obtained in the positions of the strain gauges obtained
stress level is well below the design criteria. from Fig. 6. All located were on the pressure side of the
Consequently the static stress has not been the main blade in the area close to the leading edge, see Fig. 5.
concern in the design process. Instead the focus has been
on the dynamic stress and fatigue of the impeller blade. The figure shows the stress amplitude under an operating
condition corresponding to a 45 kn ferry. The amplitudes
have been normalized against the maximum value in
5. FATIGUE OF IMPELLER BLADES order to show the relative distribution. The result shows
that the amplitude is approximately the same in gauge
In the previous section it was shown how mean stress of PS1, PS2 and PS4. A tendency of decreasing amplitude
the impeller employing pressure obtained by CFD can be along the chord can be seen. This has been confirmed
computed in a good agreement with to the result of from tests with gauges located close to the trailing edge
model tests. The dynamic stress of the impeller can on where very low amplitudes were found. The amplitude in
the other hand only be obtained by scaling the result of gauge PS4 shows the amplitude is approximately
the model test. Factors like the influence of cavitation constant in radial direction.
can presently not easily and efficiently be represented in
CFD analysis. Also the number of operating conditions Based on the stress amplitude of the impeller blade the
to be studied would render a CFD analysis of the system fatigue life can be estimated employing the Palmgren-
very time consuming. Miner cumulative damage rule. An SN curve has been
established based on a combination of testing and
At an early stage of the design process it was found that extensive full scale experience. In the fatigue test a
the highest stress amplitude will occur in the same area specimen of the actual cast duplex stainless steel was
as the highest mean stress found in the FE analysis. The subjected to an environment of sea water.
dynamic stress in this area is caused by pressure
fluctuations when the blade is passing the wake in the Based on the design life and the operating profile of
intake. Therefore the focus of the design process was to different vessels the cumulative damaged has been
carefully choose a thickness distribution of the blade in computed under different operating conditions. The
order to achieve a constant static and dynamic stress resulting cumulative damage is on the same level as of
spread out over the area close to the leading edge. In this the present Kamewa SII waterjets.
way the material will be utilized in a more efficient way
than if a single hotspot would be present.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

6. CONCLUSIONS

This paper has shown the importance of integrating

structural analysis in the development process of a new
waterjet pump. By carefully choosing the distribution the
blade thickness a good utilizing of the material has been
achieved without influencing the hydrodynamic
performance. For an impeller with forward swept
impeller blades knowledge of the dynamic stress is
essential. In order to obtain the impeller stress variation
with the effects of non-uniform flow in the intake and
cavitation model test are necessary.

7. AUTHORS BIOGRAPHY

Reima Aartojärvi is manager of waterjet

hydrodynamics at Kamewa Waterjets, Rolls-Royce AB,
Kristinehamn, Sweden.

Mats Heder holds the position of senior structural

engineer at Kamewa Waterjets, Rolls-Royce AB,
Kristinehamn, Sweden.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

RESEARCH ON HYDRODYNAMIC PERFORMANCE OF HYBRID PROPULSION

SYSTEM

C Sun, Y Wang and Z Li, University of Naval Engineering, China

SUMMARY

Hybrid propulsion systems of waterjet(s) and propeller(s) have many advantages. An existing ship, South African Navy
3500 ton corvette which was reported in Waterjet Propulsion 4, has greatly proven that. When waterjet(s) and
propeller(s) are combined in a ship, not only the interactions of waterjet-hull and propeller-hull, but also the interaction
between waterjet(s) and propeller(s) should be considered. In this paper, hybrid propulsion system of a waterjet and two
propellers is studied particularly. An isolated waterjet propulsion system and open-water propeller performances are
studied and results are validated by experimental data. Flow region of the hybrid propulsion system is simulated by
solving RANS equations and interactions of the two kinds of propulsors are analyzed. Inflow and outflow of waterjet
and propellers are different from isolated conditions. Performances of propellers are changed more widely than the
waterjet. Different rotating direction of the propellers will affect the velocity distribution at ducting inlet and then the
performance of waterjet. Efficiency of the propulsion system and non-uniformity of pump inflow are better when
propellers rotate outward. Both calculation and theory analysis show that propellers are more sensitive to the change of
flow field on hybrid propulsion system. More attention should be given to propellers when a hybrid propulsion system is
designed.

NOMENCLATURE speeds, both waterjets and diesel engines or gas turbines

have good efficiencies. For naval vessel, propeller's noise
A Propeller disc area (m2) around 20 knots can be optimized without using large
Anozzle Nozzle exit area (m2) propeller diameters which are difficult to mach with the
CTjet Waterjet thrust loading coefficient (1/1) hull geometry.
CTprop Propeller thrust loading coefficient (1/1)
D Propeller diameter (m) For the hybrid propulsion system, propellers are arranged
J Advance ratio (1/1) near the ducting inlet of waterjet. For isolated propeller
kM Torque coefficient (1/1) or waterjet ships, it is enough when the interaction of
kT Thrust coefficient (1/1) propeller-hull or waterjet-hull is considered. When
M Torque (kN m) waterjet and propellers work together, the condition is
n Rotating speed (r.s-1) greatly changed. Flow regions of waterjet and propellers
p Pressure (N m-2) affect each other. Inflow and outflow conditions of two
Q Flow rate (m3 s-1) kind of propulsors are changed and so are the
T Thrust (kN) performances. Such issues should be considered when a
v Ship speed (m.s-1) hybrid propulsion system is designed.
vA Axial velocity entering propeller (m s-1)
vin Average inlet velocity of flow (m s-1) In this paper the interaction of a waterjet and two
vj Average nozzle velocity of flow (m s-1) propellers is researched by using CFD method. Flow
vpump Average axial velocity of pump inflow (m s-1) field of the hybrid propulsion system is simulated by
P Jet velocity ratio (1/1) solving RANS equations. Positions and rotating direction
] Non-uniformity of pump inflow(1/1) of propellers are changed to compare the varying of
U Specific density (kg.m-3) performances. Sensitivity of propellers to the change of
flow field is studied particularly in the end.

1. INTRODUCTION
2. CONCEPT DESIGN INTEGRATION
Hybrid propulsion systems of waterjet(s) and propeller(s)
have been applied on fast yachts, ferries and warships for Parameters of MEKO A-200 Corvette were reported
some time. Higher efficiency and more operation modes detailed in Waterjet propulsion 4[1]. A simple hybrid
are the main characteristics of these applications. From propulsion system similar to MEKO A-200 is selected
the appearance of such systems, the question of how to for this research work. Power ratio, diameter ratio of
tackle the issue of cruising at high overall efficiency with waterjet to propeller are just same as MEKO A-200.
a diesel and/or gas turbine plant is sometimes solved. At However, the selected hybrid propulsion system is small,
low or medium speeds the propellers are driven by waterjet and propellers are different from MEKO A-200.
diesel(s) so that propellers have good efficiency and Furthermore, the main issue concerned is the interaction
diesel engine(s) are loaded in an optimum way. At higher of waterjet and propellers so that only stern part of vessel
is considered and no particular ship model is used. To

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

further simplify the problem, the ship is truncated at the bottom to represent the boundaries of the calculated flow
design waterline, eliminating the need to define the free field. The optimum domain of the control volume from
surface. Table 1 displays the data of selected hybrid my research group is 20D upstream of the inlet, 10D and
propulsion system. 8D of width and depth respectively. D is the rotor inlet
diameter of waterjet.
Table 1 Selected Hybrid Propulsion System Data
Design speed 30knot The generation of the mesh of the complete waterjet is
Waterjet intake duct diameter 71cm split into three separate parts; the inlet mesh and the
Propeller diameter 1.15m pump meshes with rotor and statorbowl. The mesh of
Power of propellers in mode III 2*600kW inlet is done with the commercial mesh code ICEM CFD.
Power of waterjet in mode III 2000kW Near the surface of the ducting, shaft and hull bottom a
Propeller speed 732rpm special procedure is applied to create fine cells at the
Waterjet speed 850rpm walls. This ensures good quality of the y+ values.

Fig.1 CAD model of hybrid propulsion system

Fig.2 Mesh of complete waterjet propulsion system
Fig.1 Shows the CAD model of selected hybrid
propulsion system. Only a waterjet and two propellers
The mesh of pump is done with the code of
are installed at stern part of ship. With a flat bottom
TURBOGRID. The main topology structure of rotor is H
section, the waterjet and the hull join more easily, and the
model and statorbowl is J model. O-grid is generated
properties of waterjet and propellers are not modified by
around the blades. This complete mesh is shown in fig.2.
any built-up surface.
After meshed, the model is brought into CFX Pre to
define the fluid boundaries. The front of the control
After the model of hybrid propulsion system has been
volume is set as an inlet, introducing water to the control
built, CFD method is used to calculate the hydrodynamic
volume at the ship speed. The opposite end of the control
performance of it. In order to compare and validate CFD
volume could have been set as an outlet or an opening,
results, isolated waterjet and open water propeller are
the latter of which is simply a pressure field that allows
analysed using CFD method at first.
fluid to flow in or out. In this paper it is defined as an
outlet. The outboard boundary and the bottom boundary
are set as free slip walls. The hull, ducting and shaft are
3. VALIDATION OF CFD RESULTS
defined as no-slip walls. Rotation of the impeller can be
implemented via the quasi-steady Multiple Frames of
The numerical models of the waterjet and propeller are
Reference method or via the fully transient moving mesh
based on a Reynolds-Averaged-Navier-Stokes (RANS)
option. The latter is significantly more time consuming,
method. Analysis of the flow is made with the
and it is therefore not applied in this analysis. Pump
commercial CFD code CFX. The CFD code uses the
outlet is set as an outlet boundary condition.
finite volume method to solve the discretised set of
equations. Besides the equations for conservation of
The working point of waterjet is determined by the ship
mass and momentum, two additional equations are
speed and the shaft speed of the pump. Many groups of
solved to model the turbulence. For all simulations the
ship speed and shaft speed are set and calculated. Fig.3
sst turbulent model is used.
shows the comparison of the calculated thrust and results
provided by KaMeWa company.
3.1 VALIDATION OF WATERJET RESULTS

There are many reports using CFD method to calculate

the performance of waterjet[2-4]. Norbet W.H.B have done
especially comprehensive and detailed calculation about
this in his doctoral dissertation[5]. In this paper, the
complete waterjet propulsion system is the analysis
object. A control volume is constructed around the ship

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

 1.8
CFD
H/D=I CFD

KaMeWa 1.5 Experiment
 Cavitation region
p1 1.2 II
 p2 10KM
III

kT,10KM
 0.9
Thrust

 p3
0.6 Kt

0.3
 p4
III II I
 0
J
 -0.3 0 3 05 07 09 11 13 15

   Ship speed    

Fig.4 CFD results and experimental data

Fig.3 Thrust prediction for waterjet system using CFD
code and KaMeWa data To analyze the computational results, the thrust and
torque coefficients, kT and kM, are selected as the global
Agreement between the calculated thrust and KaMeWa quantities of interests:
data is very good for most of the calculated conditions.
Near the cavitation zone which is restricted using every
year errors become large. It is because cavitation model
kT T /( Un 2 D 4 ) ; k M M /( Un 2 D 5 ) ; J v /(nD)
is not used in this analysis. Over the complete range of
calculated conditions, the largest deviations of power and The comparison with the experimental data for both
thrust are 3.8% and 8.6%. thrust and torque coefficients is illustrated in Fig.4. The
results agree reasonably well with experiment data. The
3.2 VALIDATION OF OPEN WATER PROPELLER maximum errors of kT and kM are 8%.
RESULTS

The CFD analysis of the open water propeller is 4. CFD ANALYSIS OF HYBRID PROPULSION
described in 2 parts. First the generation of the numerical SYSTEM
model will be discussed. Then the CFD predictions of
thrust and torque will be compared with experimental After the computational results of isolated waterjet and
data. open water propeller are validated, the hybrid propulsion
system is analyzed. Mesh is generated according to the
The present simulations are performed on a 4 blades isolated waterjet and the open water propeller system, the
propeller with the diameter D=0.25m. The computational mesh is shown in fig.5. Most boundary conditions are
domain has been identified with a cylinder surrounding same as the isolated waterjet and propeller. For lack of
the propeller and aligned with the shaft axis. The inlet is better option, the top surface (which would otherwise be
4D upstream, the outflow 6D downstream, the diameter the free surface) was set as a symmetry plane. This
of the lateral cylindrical boundary is 5D. The whole restricts the water available to the inlet by denying the
domain is spilt into 2 parts to generate mesh. One domain ability for the inlet to draw water from above the
is just around the propeller which is filled with waterline.
tetrahedral cells, ten lays of prismatic cells have been
attached to the blades and hub surface. Another domain is
meshed by hexagonal cells. The mesh is shown in fig.3.

Fig.5 Mesh of hybrid propulsion system

4.1 FLOW CHARACTERISTICS OF HYBRID

PROPULSION SYSTEM

Fig. 3 Mesh of the open water propeller Figure 6 shows the streamline of hybrid propulsion
system. There is an important difference with the isolated
Boundary conditions are set to simulate the flow around waterjet and propeller. For hybrid propulsion system,
a rotating propeller in open water: on the inlet boundary, both waterjet and propeller are sucking water from the
velocity components of uniform stream with the given same direction and origin. Space between two propellers
inflow speed are imposed; on the exit boundary, the static is narrower than inflow width of the isolated waterjet
pressure is set; on the outer boundary, the opening plus propellers. The suction flow of propellers are
boundary condition is imposed; on the blade and hub inclined to the side of the ship. After water been
surface, the no-slip condition is imposed. accelerated by propellers and ducting inlet, they flow to

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

the stern part of the ship. However, velocity of water 4.2 INFLUENCE OF THE RELATIVE POSITION
around the ducting inlet is higher than the water OF PROPELLERS AND WATERJET
accelerated by propellers, shapes of propellers' streamline
are seriously distorted. For hybrid propulsion system, the waterjet permits to be
moved as far aft in the ship as desired without the
Furthermore, propellers' blades pressure distribution are problem of unacceptable large inclination of the shaftline
greatly changed. Pressure distribution on four blades is and the engine foundation. So the relative position
not identical. Fig. 7(a) shows the pressure coefficient of between waterjet inlet and propellers can be changed
blade 1 and blade 3 at four different blade spans. more widely. In this paper, performances of five different
Pressure coefficient on blade 1 which is near the ducting positions of propellers are analyzed. The sketch map of
inlet is lower than blade 3. Cavitation should be propellers' positions is shown in figure 8. From position
restricted at blade 1 and 4, especially at the region near 1 to position 5, propellers are close to the ducting inlet
the hub. gradually. When the calculations are done, mesh size is
identical for five positions.

5 4 3 2 1

Fig 8. Sketch map of propeller positions.

a. view from the afterbody of ship
It is expected that the change of propellers' positions will
result in a varying flow conditions of the waterjet and
propellers and thus their performances. Fig. 9(a) shows
the torque and thrust of waterjet at five different
propellers' positions. The results have been normalised
with values of the isolated waterjet. It can be observed
that the effect of propellers' positions on waterjet
performance is small. Torque is nearly constant at five
propellers' positions. Thrust fluctuates at the range of 2%.
Two causes contribute to the fluctuation of thrust. One is
b. view from the bottom of ship the varying of mass flow rate which can be seen from
Fig. 6 Streamline of hybrid propulsion system(inward fig.9(d). From momentum balance we know that the
rotating) thrust is related to the square of the volume flow rate.
2 Cp
Cp
1
r/R=0.3 2

1
r/R=0.5 Consequently, a relatively small change of the volume
0 0 flow rate can lead to a large change in the thrust of the
0 0. 2 0. 0. 6 0. 8 1 0 0. 2 0. 4 0. 6 0. 8 1
-1
streamwise
-1

-2
streamwise installation. The other is the change of wake fraction.
-2
Blade 1 Blade 3
When a waterjet and two propellers work together,
2

Cp r/R=0.7 Cp
1
r/R=0.9
velocity at the region of ducting inlet is higher than the
1 0. 5 isolated condition and wake fraction is larger. For the two
0 0
0 0. 2 0. 4 0. 6 0 8
0. 5
0 0. 2 0. 4 0. 6 0. 8 reasons we can learn that thrust of hybrid propulsion is
-1
streamwise -1 streamwise
smaller than the isolated conditions and some
-2
fluctuations may exist, it can be seen in figure 9(a).
a. pressure values of blade 1 and blade 3.
Furthermore, pressure distribution along the cutwater is
changed. Analysis of the pressure in figure 10 shows that
although the two locations of the stagnation point and
minimum value are almost the same for five propellers'
positions, pressure values are different. Minimum
pressure of position 5 is lowest.
4
4
1
3
1
Easy Cavitation region 3 Compared with the waterjet, thrust and torque of
propellers fluctuate widely. The deviation can reach to
2 2 6%, it is shown more detail in figure 9(b). This is mainly
due to the decrease of advance ratio. Figure 9(d) shows
the velocity at propeller disc plane of five positions.
b. pressure contours of propellers' blades
From position 1 to position 5, velocity decreases
Fig. 7 Pressure distribution of propellers blades
gradually and thus the advance ratio. We can derived the

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

change of thrust and torque from the open water


propeller diagram.
 Inward
:DWHUMHW

(IILFLHQF\
Over the complete propellers' positions of calculated 3URSHOOHU
2YHUDOO Outward

conditions, performances of waterjet and propellers are
all changed. However, there are only a rather limited 
varying of efficiency. This is shown in figure 9(c). 



    
 SRVLWLRQ
7KUXVW 7RUXTH



 Inward
c efficiency of waterjet and propeller
:DWHUMHW 7KUXVW
:DWHUMHW 7RUTXH 
 Outward




Mass flow rate, Velocity


 

     Inward
Outward
3RVLWLRQ 

a waterjet thrust and torque Mass flow rate of waterjet


Velocity at propeller disc plane
 
3URSHOOHU 7KUXVW     
 3URSHOOHU 7RUTXH Position

Inward

7KUXVW7RUTXH

Outward

d mass flow rate and velocity
 Fig. 9 Comparison of calculations of some results for
 five different propellers' positions





  Position   

b propeller thrust and torque

1.5
0 03 Position 1
0 02

1 0 01
Isolated wa erjet
0
Cp [-]

-0 01

-0 02

0.5 -0 03
Position 5
-0 04
-0 49 0.485 0 48 0.475 0.47 0.465
Distances along cutwater [mm]
Cp [-]

0.2 Position 1
-0.5 0.18
0.16
S>0 Isolated waterjet
Cp [-]

0.14
0.12
-1 S=0 0.1
Position 5
0.08
S<0 0.06
0.335 0.34 0.345 0.35 0.355 0.36 0.365 0.37 0.375 0.38
Distance s a ong cutwater [mm]
-1.5
-0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4
Distance s along cutwater [mm]

Fig. 10 Calculated pressure coefficient Cp along the cutwater for different positions of propellers.

4.3 INFLUENCE OF PROPELLER SHAFT influence of propellers' rotation direction on waterjet

ROTATING DIRECTION and propellers' hydrodynamic performance will be
researched particularly in this paper. For the above
For current most vessels the normal propulsion is two calculation, the starboard and port propellers are
propellers driven by diesels and/or gas turbines. The left-handed and right-handed. It is called inward
rotating direction of propeller shaft is mainly rotating propellers, and the reverse is called outward
associated with ship manoeuvring ability. However, rotating propellers. Fig. 11 shows the streamline of

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

hybrid propulsion system at the condition of outward


rotating propellers. Differences between figure 6 and Inward
Outward
figure 11 show that outward rotating propellers' 
streamlines are assembled under the ducting inlet

while dispersed of inward rotating propellers. Pressure

.VL
and velocity distribution under the hull are different at 
two cases. Figure 9 displays the comparison of the

results of two cases. Obviously, thrust and torque of
waterjet is more close to isolated condition when 
propellers rotate outward. This is mainly due to     
3RVLWLRQ
varying of wake fraction for the increase of mass flow
rate when propellers rotate outward cannot change the
values so much.
Fig. 12 Comparision of non-uniformity at two
propeller rotation direction

5. MOMENTUM BALANCE FOR BOTH AN

OPEN PROPELLER AND A WATERJET

Some hydrodynamic performances of hybrid

propulsion system have been discussed above, from
a. view from the afterbody of ship which we know that flow shapes of waterjet and
propellers are different from each isolated condition
and performance of propellers changed more widely.
In fact, thrust of waterjet and propeller can be derived
from the momentum balance for an incompressibility
fluid. In this section, in order to explain propeller is
more sensitive in hybrid propulsion system, thrust of
waterjet and open propeller will be analysed from
theory.
b. view from the bottom of ship
Fig. 11 Streamline of hybrid propulsion 5.1 OPEN PROPELLER THRUST
system(outward rotating)
Assume that the propeller is an actuator disc, i.e. a disc
For further comparison of different propellers rotation with diameter D and area A, causing a sudden increase
directions , the level of non-uniformity is expressed as in pressure 'p . The propeller acts on a circular
a single value ] [6]. column of fluid. Upstream, the flow is undisturbed and
has a velocity of advance v A while passing though
1
] (v v pump ) 2 dA
Q³
area A0 with diameter D0 . Downstream the flow has
contracted; the diameter has decreased to D1 and the

where v is the local axial velocity and v pump the speed has increased to v A  'v . The pressure in the
average axial velocity. Non-uniform pump inflow slipstream is the same as the pressure in the
velocity distribution can cause flow rate fluctuations undisturbed flow. Figure 13 shows a sketch of the
through an impeller channel and the variations of the control volume of an open propeller with the
inflow angle at the leading edge of the blade[5]. nomenclature of the velocities.
Obviously, the non-uniformity should be kept minimal
Undisturbed flow
from a hydrodynamic point of view. Calculation Slipstream
Propeller disc
D0 A 0
results show that the level of non-uniformity of D A
D1 A 1
outward rotating propellers is smaller than inward p0 p  'p p p0

rotating propellers (fig.12). That's because when v A  'v vA

v
propellers rotate outward, water is gathered under the
ducting inlet, hull boundary is destroyed and can't play
a role in the non-uniformity. From this point of view,
the case of outward rotating propellers is better.
Fig 13The propeller as actuator disc: the momentum
theory

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

If Bernoulli’s law is applied, It can be got: T 2(1 P )(1 w) 2

CTjet
Aft of propeller: 1/ 2 U v 2 ship Anozzle P2
1 1 Where w is the wake fraction. The wake fraction
p  'p  U v 2 p0  U (v A  'v) 2 becomes zero, when the inflow velocity is equal to the
2 2
ship speed. This is equivalent with an open water test
Ahead of propeller:
of a propeller with uniform inflow. The resulting
1 1 loading coefficient for a waterjet with undisturbed
p  U v2 p0  U v A 2
2 2 inflow yields:

If the second equation is subtracted from the first, the 2(1 P )

pressure rise over the propeller disc can be solved: CTjet
P2
1 1
'p U (vA  'v) 2 U vA2 Comparison with the open propeller thrust loading
2 2 coefficient reveals a difference between the waterjet
and the open propeller. This is due to the fact that a
Assume that the thrust T exerted by the propeller on waterjet is an internal flow machine. For a waterjet the
the fluid is uniformly distributed so the pressure ratio between the inlet and nozzle area is fixed,
increase at the disc is the same in every position. whereas it is related to the thrust for an open propeller.
For a hybrid propulsion system of waterjet(s) and
1 1 propeller(s), propeller flow condition is greatly
T 'p. A [ U (v A  'v) 2 U vA2 ]. A disturbed while waterjet flow condition only has small
2 2 changes. This can be seen from the streamline of
hybrid propulsion system. So propeller is weaker in
Define the non-dimensional thrust loading coefficient hybrid propulsion system, more attention should be
as: given to it.

Tprop
CTprop 6. CONCLUSION
1/ 2 U vA 2 Aprop
Behavior of the flow pattern though a hybrid
Where Aprop is the cross-sectional area of the propulsion system of a waterjet and two propellers is
simulated and some results are analyzed. Results show
propeller disk, based on the propeller diameter. Jet
that hydrodynamics of hybrid propulsion are obviously
velocity ratio P is defined as: different from the isolated conditions. Flow field of
waterjet and propellers interact on each other.
vin Efficiency and non-uniformity of pump inflow in
P
vout outward rotating condition are better than in inward
So, condition. Both calculation and theory analysis show
that propellers are more sensitive to the change of flow
Tprop (v A  'v) 2 1 P2 field on hybrid propulsion system. More attention
CTprop 1
1/ 2 U v A 2 Aprop vA2 P2 should be given to propellers when a hybrid
propulsion system is designed.
5.2 WATERJET THRUST

For the determination of the thrust of a waterjet in 7. REFERENCES

general the same approach as for the open propeller is
used. The waterjet thrust is defined as: [1] Juergen Wessel, 'Waterjet Propulsion for a 3500
Ton Corvette from BLOHM+VOSS', International
Conference on Waterjet Propulsion, RINA, London,
T U Q(v j vin ) 2004
[2] Seil G.J., 'The Effects of the Shaft Rotation and
The thrust loading coefficient based on nozzle exit Scale on the Flow in Waterjet Inlets', International
area is discussed in [8]. With the nozzle area as Conference on Waterjet Propulsion III, RINA,
reference area, the relation between jet velocity ratio Gothenborg, Sweden 2001
and the thrust loading coefficient becomes: [3] Park, W.G et al., 'Numerical Analysis of Intake
Flow of Waterjet Pump', The 2nd PNU International
Colloquium on Waterjet, Pusan, Korea, pp.171-178
[4] Bulten N.W.H, Verbeek, R., 'CFD simulation of

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

the flow through a waterjet installation', International

Conference on Waterjet Propulsion 4, RINA, London,
2004
[5] Bulten N.W.H., 'Numerical Analysis of a
Waterjet Propulsion System'ˊPhD thesis, Technical
University of Eindhoven ,2006
[6] Roberts J LˈWalkeG J R 'Boundary Layer
Ingestion Effects in Flush Waterjet Intakes',
International Conference on Waterjet Propulsion,
RINA, The Netherlands, 1998
[7] Hans K.Woud, Douwe S., 'Design of Propulsion
and Electric Power Generation Systems'. The Institute
of Marine Engineering, Science and Technology,
London, 2002
[8] MacPherson, D.M., ‘A universal parametric
model for waterjet performance’, Proceedings
FAST’99 conference, pp. 879-882, Seattle, 1999

8. AUTHORS BIOGRAPHY

Cunlou Sun is a Ph.D. candidate of Marine

Engineering at the University of Naval Engineering, P.
R. China. Last year he got his master degree and his
master thesis is about waterjet numerical analysis.
Yongsheng Wang is a Professor at the University of
Naval Engineering, P. R. China. Since 1982 he has
published more than 80 papers and his main research
interests are marine power plant and its characteristics
analysis.
Zhuying Li is a assistant professor at the University of
Naval Engineering, P. R. China. Her main research
interest is material mechanics.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

TESTING AIR-AUGMENTED WATERJET PROPULSION

A Gany, A Shemer, A Gofer and D Har-Lev, Faculty of Aerospace Engineering, Technion – Israel Institute of
Technology, Haifa, Israel

SUMMARY

A unique concept of an air-augmented waterjet propulsion system has been successfully tested, resulting in a remarkable
increase in the system’s thrust. This research may present a conceptual revolution for increasing boost capability and
maximum attainable speed from a given waterjet system, similarly to the role played by an after-burner in a jet engine.
Static tests in a water-tank have been conducted using a Yamaha jet-ski waterjet propulsion unit. Data obtained from the
original motor operation at different rpm's have been used as a reference. Air injection in the air-augmented tests has
been done through an extension section specially designed and installed between the pump and the nozzle exit cone.
Typical thrust increase in the range of 15-30% and more compared to the original thrust has been obtained due to the
injection of air bubbles, without affecting the pump operation and without causing cavitation problems. Such a
capability may have significant implications on improving vessel’s maneuvering, boost and speed.

1. INTRODUCTION
Diffuser Air Inlet Nozzle
Marine two-phase jet propulsion has been studied
extensively at the Technion - Israel Institute of
Technology for nearly two decades. The main idea of the Water Bubbly
two-phase jet propulsion is that gas or air bubbles
injected into the water flow within the propulsion unit
Flow
convert their expansion work, ³ PdV , into kinetic energy
of the flow, hence increasing the exhaust jet speed and
generating thrust. (inf)

The thrust generated by both aeronautical and marine jet Mixing Chamber
engines, including the marine two-phase jet propulsor,
can be expressed by Figure 1: Illustration of a marine two-phase ramjet
propulsion unit.
F m (ue  u) (1)

where F is the thrust, m the fluid mass flow rate, u is

the cruise velocity, and ue is the exhaust velocity.
Equation (1) assumes that the pressure at the nozzle exit
is equal to the ambient pressure (adapted nozzle). In the
case of marine two-phase jet propulsion one may take
into account the fact that the air mass flow rate is almost
negligible compared to the water flow rate, m a  m w ,
hence the fluid flow rate through the propulsion unit is
practically equal to the water flow rate.
Figure 2: Schematic of marine air-augmented waterjet.
Earlier research was focused on the operating mode of
results. Albagli and Gany [8] and Mor and Gany [9, 10]
marine two-phase ramjet (Figure 1). Water enters to the
analyzed and solved the two-phase flow characteristics
marine ramjet as a result of the vessel motion. Internal
with relation to marine ramjet. The marine ramjet may be
pressure increases due to deceleration of the flow in the
an elegant and efficient propulsion solution for high-
inlet diffuser. Then, air bubbles are injected into a mixing
speed cruise regime. One should note, however, that
chamber, forming a two-phase flow which accelerates
generally ramjet engines cannot start from rest and have
while flowing in the nozzle. The high speed exhaust jet
relatively low boost capability.
generates thrust without any moving parts in contact with
the water. Past works [1-3] discussed the subject of
The objective of the present article is to present concept
marine ramjet. Patents by Varshay and Gany [4, 5]
and testing of another operating mode of two-phase
described the concept in detail. Gany [6, 7] presented the
marine propulsion, namely, a boost mode, comprising
theoretical thermodynamic cycle as well as experimental
air-augmented waterjet (Figure 2). A similar option is
mentioned in [11]. The idea of the air-augmented

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

waterjet is to combine the operation of a standard

waterjet propulsion unit to a thrust augmentation section 90

based on air injection which causes further increase of 80

the exhaust jet speed. The combination acts like an after- Limiting
Severe
Mild
burner in an aeronautical turbojet engine. It can provide 70 cavitation line
cavitation
cavitation

Thrust (Pump O nly) [kN ]

zone
additional thrust from an existing waterjet propulsor zone Engine
60
when necessary, e.g., to increase vessel acceleration, to Power
50 2000 kW
increase maximum speed, or to overcome a resistance
hump without affecting the regular waterjet pump 1600 kW
40
operation. In this way one can upgrade the waterjet Resistance Line
propulsion, extracting more thrust from the same unit and 30 1200 kW
avoiding cavitation problems which would occur if thrust
20
increase were done by operating the existing waterjet 800 kW
Motor Thrust Lines
engine at a higher power. An example revealing how 10
thrust augmentation by air injection can avoid operation 400 kW
0
in cavitation regime for overcoming a resistance hump is 10 20 30 40 50 60
shown in Figure 3. One can see that instead of increasing Vessel-Speed [knots]
waterjet engine power, implying operation under (a)
cavitation in the waterjet pump (Figure 3a), one may
90
operate at a lower engine power (avoiding cavitation),

Thrust (Combined Pump+15% Air Thrust) [kN]

Severe
adding the necessary thrust via air injection (Figure 3b). Mild
80 Limiting cavitation
cavitation line cavitation
In this sample case, the addition of 15% thrust by the air zone
zone
70 Engine
system is sufficient to overcome the resistance hump,
Power
staying away from cavitation problems. Thus, one does 60 2000 kW
not have to install a larger waterjet unit, yet enjoying the
50
augmented thrust by using a small air supply unit. 1600 kW
Resistance Line
40
1200 kW
In the following sections experiments done and results 30
obtained, revealing the feasibility and actual performance
20 800 kW
of thrust augmentation by air injection will be presented.
Motor Thrust Lines
10
400 kW
0
2. TEST FACILITIES 10 20 30 40 50 60
Vessel-Speed [knots]

Tests have been conducted in a water tank with (b)

connection to air supply. A 50 kW Yamaha jet-ski
waterjet propulsion unit has been used for the tests. Fig. 3: Sample of thrust vs. vessel-speed maps: (a)
Thrust at static operation was measured using a load cell Regular waterjet, requiring operation in the cavitation
at the jet-ski front. Pressure downstream of the pump as zone to overcome resistance hump. (b) Air-augmented
well as motor rpm and airflow rate have been recorded waterjet, avoiding operation in cavitation regime.
continuously. Schematic of the test installation is
presented in Figure 4.
Load Cell
Air Two-Phase
3. RESULTS AND DISCUSSION Jet

3.1 BASIC WATERJET OPERATION

A number of test series have been conducted. For

Pump
baseline data the original waterjet was operated at
different rpm's in a regular manner, without air injection. Water
Then, a cylindrical section was installed (to enable Inlet
introduction of air) between the pump section and the
nozzle cone section. The engine with this installation was Nozzle
run again without air to compare the results with the
original arrangement. Figure 5 presents the results of Fig. 4: Schematic of the test installation for the air-
thrust versus motor rpm for the two installations. One augmented waterjet.
can see that the thrust data overlap one another and no
apparent practical difference is detected.

©2008 The Royal Institution of Naval Architects

Waterjet Propulsion 5, London, UK

to be somewhat better at the higher rpm, where the

2500
No Mixing Chamber, No Air
nozzle is adjusted to higher airflow rates.
With Mixing Chamber, No Air
2000 1.9
Theory, 70% Air Expansion Efficiency
1.8 Experimental Results
1500 2000 RPM
Thrust [Nt]

1.7

1.6
1000

Fmix/Fwj
1.5

1.4
500
1.3

1.2
0
1000 2000 3000 4000 5000 6000 7000 1.1
RPM
1
0 20 40 60 80 100 120 140 160 180
mdota [gr/sec]
Figure 5: Thrust vs. rpm for the original waterjet and for
the waterjet with a mixing chamber extension. No air
injection. Figure 6: Thrust ratio of waterjet engine with and without
air injection vs. airflow rate. Engine at 2000 rpm.
3.2 THRUST AUGMENTATION BY AIR 1.7
INJECTION Theory, 70% Air Expansion Efficiency
1.6 Experimental Results

Thrust increase due to air injection has been studied by

3000 RPM
introducing the air from 16 ports at the extension section 1.5

(mixing chamber) casing into the water flow.

1.4
Fmix/Fwj

Ideal isothermal gas expansion in the nozzle is associated

with expansion work wep per unit mass of air as follows, 1.3

1.2
wexp RT ln r (2)
1.1

where R is the specific gas constant, T is the water

1
temperature, and r is the ratio between the pressure after 0 20 40 60 80 100 120 140 160 180
mdota [gr/sec
the pump and the ambient pressure. It is predicted that an
effective fraction of the air expansion work K wexp per
Fig. 7: Thrust ratio of waterjet engine with and without
unit mass of air is converted to additional kinetic energy
air injection vs. airflow rate. Engine at 3000 rpm.
of the exhaust jet (at efficiencyK ). For optimal design,
the nozzle exhaust plane has to be adjusted according to
the air-to-water mass flow rate ratio and the vessel speed. 1.4
Theory, 70% Air Expansion Efficiency
As stated before, all tests have been conducted in static 1.35 Experimental Results
conditions (zero vessel speed). 4000 RPM
1.3

Figure 6 presents the thrust ratio with and without air as a 1.25
function of airflow rate, for waterjet engine operating at
Fmix/Fwj

2000 rpm. A theoretical line with air expansion work 1.2

conversion efficiency K 70% seems to reflect the 1.15

actual performance. Note that for this low rpm, thrust
1.1
increase may be as high as 55%. In this rpm the nozzle
was adjusted to the low airflow rate ( 60 g / sec ), hence, 1.05
somewhat better efficiency can be observed in that range.
1
Figures 7-9 make a similar presentation for 3000, 4000, 0 20 40 60 80 100 120 140 160 180
mdota [gr/sec]
and 5000 motor rpm, respectively. One can see that the
theoretical line with 70% air expansion efficiency gives a
good correlation in all cases. The thrust ratio decreases at Fig. 8: Thrust ratio of waterjet engine with and without
higher rpm's for the same airflow rate, as expected by the air injection vs. airflow rate. Engine at 4000 rpm.
theory. Note that the air expansion work efficiency seems

©2008 The Royal Institution of Naval Architects

 Waterjet Propulsion 5, London, UK

4. VARSHAY, H. and GANY, A., “Underwater

1.4
Theory, 70% Air Expansion Efficiency
Two-Phase Ramjet Engine,” United States Patent
1.35 Experimental Results Number 5,598,700 , Feb. 4, 1997.

1.3 5. VARSHAY, H. and GANY, A., “Underwater

5000 RPM Two-Phase Ramjet Engine,” United States Patent
1.25
Number 5,6
92,3
17 , Dec. 2, 1997.
Fmix/Fw j

1.2
6. GANY, A., "Analysis of a New Thermodynamic Power
1.15 Cycle for a Two-Phase Waterjet," International Conference
Waterjet Propulsion ,4RINA , London, UK, May 26-27, 2004.
1.1

7. GANY, A., "Two-Phase Marin Ramjet Propulsion

1.05
Research," The Australian Naval Architect, Vol. 12, No. 2, pp.
1
36-40, May 2008.
0 20 40 60 80 100 120 140 160 180
mdota [gr/sec]
8. ALBAGLY, D. and GANY, A., “High Speed
Bubbly Nozzle Flow with Heat, Mass, and Momentum
Fig. 9: Thrust ratio of waterjet engine with and without Interactions,” International J. of H eat and Mass
air injection vs. airflow rate. Engine at 5000 rpm. Transfer, Vol. 46, pp. 1993-2003, 2003.

9. MOR, M. and GANY, A., “Analysis of Two

4. CONCLUSIONS Phase Homogeneous Bubbly Flows Including Friction
and Mass Addition,” J. of F luid Engineering
,
Concept and testing of air augmented waterjet propulsion Transactions of ASME , Vol. 126, pp. 102-109, 2004.
have been presented. It is shown that by injecting air
provided by a relatively small air compressor into the 10. MOR, M. and GANY, A., “Performance Mapping
nozzle section of a waterjet unit, one can increase the of Bubbly Water Ramjet,” International Journal of
waterjet thrust without affecting the pump and engine Maritime nEigneering , Vol. 149, Part A1, pp.45-50,
operation. The additional thrust is produced in a way 2007.
parallel to the operation of an after-burner in an
aeronautical turbojet engine. The main advantage is the 11. MUENCH, R.K. and GARRETT, J.H., “A Review
possibility to increase of boost and speed, and of Two-Phase Marine Propulsion,” AIAA Paper No. 72-
overcoming resistance hump without the use of a larger 589, AIAASNAME
U
/ SN Advanced Marine Propulsion
waterjet unit and with avoiding cavitation problems. Meeting, Annapolis, MD, USA, July 17-19, 1972.

The tests conducted using 50 kW Yamaha waterjet unit 6. AUTHORS BIOGRAPHY

reveal static thrust increase of 15%-50%, depending on
air flow rate and motor rpm, demonstrating air expansion Alon Gany is a Professor, holding the Lena and Ben
work efficiency of about 70%. Fohrman Chair in Aeronautical Engineering and heading
the Sylvia and David IA Fine Rocket Propulsion Center
The concept of air-augmented waterjet is very promising at the Faculty of Aerospace Engineering, Technion –
for upgrading waterjet systems. Israel Institute of Technology, Haifa, Israel. Prof. Gany
has published numerous papers in the field of Propulsion
in general and marine two-phase jet propulsion in
5. REFERENCES particular. Prof. Gany is a fellow of RINA.
Email: [email protected]
1. MOTTARD, E.J., and SHOEMAKER, C.J.,
“Preliminary Investigation of an Underwater Ramjet Arik Shemer is currently an M.Sc. student at the Faculty
Powered by Compressed Air,” NASA Report TND-91 , of Aerospace Engineering, Technion – Israel Institute of
December 1961. Technology, Haifa, Israel. His research deals with air-
augmented waterjet. He obtained a B.Sc. degree in
2. WITTE, J.H., “Predicted Performance of Large marine technology from the University of Newcastle, UK
Water Ramjets,” AIAA 2nd Advanced Marine V ehicles (2000). He is the owner of a company Shemer Marine
and Propulsion Meeting
, Seattle, WA, USA , 1969. Engineering in Israel.
Email: [email protected]
3. CHERNYY, I.M., “Optimal Parameters and
Efficiency of Gas-Water Jet Engines Operating on Cold Aviad Gofer is an M.Sc. student at the Faculty of
and Hot Gas,” lFuid Mechanics –oSviet Research , Vol. Aerospace Engineering, Technion – Israel Institute of
6, No. 2, pp. 87-105, 1977. Technology, Haifa, Israel. His B.Sc. degree (2006) was
also obtained from the Technion. His main research

©2008 The Royal Institution of Naval Architects

 Another Random Document on
Scribd Without Any Related Topics
 MONTJOIE.

Je vous prie de vouloir bien m'excuser auprès de monsieur votre

père; mais ma migraine a un peu augmenté: je préfère me retirer.
ÉDITH, à Coralie.

Venez-vous?
CORALIE.

Il est inutile que vous m'accompagniez, ma chère enfant. Restez

avec Daniel qui n'a pas encore été seul avec vous de toute la
journée. C'est par ici, n'est-ce pas?
ÉDITH.

Oui; tout droit.

CORALIE, elle salue de la tête M. de Montjoie.—A Édith.

A tout à l'heure!
Elle sort.

SCÈNE VIII

Les Mêmes, moins CORALIE.

MONTJOIE, prenant son chapeau.

Il ne me reste plus qu'à vous présenter mes hommages,

mademoiselle. (Il salue Édith.) Au revoir, capitaine.
DANIEL.

Au revoir, monsieur.
Montjoie fait quelques pas vers la porte; puis tout à coup, s'arrêtant.
 MONTJOIE.

Au fait, puisque j'ai le plaisir de vous rencontrer, capitaine, voulez-

vous me permettre de vous demander un conseil?
DANIEL, étonné.

A moi?
MONTJOIE.

Pourquoi non? Je vous tiens pour un parfait galant homme, et sur

certaines questions délicates on est toujours heureux d'avoir l'avis
d'un galant homme.
Daniel s'incline.

ÉDITH, souriant.

Mais non celui d'une jeune fille. Je vais vous laisser...

MONTJOIE.

Veuillez demeurer, mademoiselle. En vérité, je serais trop humilié de

vous mettre en fuite. D'autant que l'affaire n'a rien de mystérieux. Je
comptais demander à M. Godefroy, à M. Bonchamp leur opinion.
Mais vous aviez tant de monde ce soir que je n'ai pu les aborder.
Imaginez-vous qu'il m'arrive l'aventure la plus désagréable. L'un de
mes amis, de Marseille, M. Merlin, dont vous m'avez peut-être
entendu parler...
ÉDITH.

Non, je ne crois pas.

MONTJOIE.

Peu importe; eh bien, l'un de mes amis de Marseille, M. Merlin, m'a

écrit ce matin qu'il allait marier sa fille avec un jeune homme dont
vous me permettrez de taire le nom. Il croit que son futur gendre est
de bonne maison. Or, je connais des détails très tristes ignorés de
tous. Le père de ce jeune homme a subi une peine infamante...
ÉDITH.

Oh! mon Dieu!

MONTJOIE.

Et j'avoue que, depuis ce matin, j'hésite sans oser prendre un parti.

Les fiancés s'adorent. Révéler la vérité au père, c'est rompre leur
mariage.
ÉDITH.

Pauvres jeunes gens!

MONTJOIE.

La taire, c'est peut-être indélicat. Oh! ma conscience est très

tatillonne. Dois-je parler, dois-je garder le silence? Que feriez-vous à
ma place, capitaine?
DANIEL.

Je dirais la vérité.
MONTJOIE, avec émotion.

Ah! Cependant les fautes sont personnelles, et parce qu'un individu

est coupable, il ne s'ensuit pas que son fils soit un malhonnête
homme.
DANIEL.

J'estime qu'en toutes circonstances, il faut être très soucieux de

l'honneur de ses amis. N'est-ce pas votre avis, Édith?
MONTJOIE.

Alors, si vous étiez dans ma situation?

 DANIEL.

Je n'hésiterais pas. Je crois qu'il faut toujours remplir son devoir

sans regarder aux conséquences. D'ailleurs, si ce mariage est
rompu, la faute n'en sera pas à vous. Le père de la jeune fille n'a
qu'à pardonner à son futur gendre le crime qu'il n'a pas commis.
MONTJOIE.

Certes; mais je ne puis m'empêcher de voir le résultat de ma

révélation. Ces deux jeunes gens s'aiment: voilà deux cœurs brisés
peut-être, et par ma faute.
ÉDITH, gravement.

Ne craignez rien, monsieur. Si leur amour est sincère et immuable,

leur séparation n'aura qu'un temps: ces amours-là renversent tous
les obstacles.
DANIEL.

Chère Édith!
MONTJOIE.

Je n'ai qu'à m'incliner, mademoiselle. Mais vous, capitaine,

reconnaissez qu'il est pour le moins bien douloureux de rejeter sur
un honnête homme le poids de la faute commise par... par son père.
DANIEL.

C'est douloureux, il est vrai, mais vous aurez obéi à votre

conscience[A]. Remarquez que vous avez bien voulu me demander
d'abord mon avis sur un cas spécial. Je vous ai dit en toute loyauté
ce que je pensais. A présent, la conversation dévie; vous vous en
prenez à la grande question de la responsabilité. Permettez-moi de
garder le silence, car j'ai là-dessus des idées tellement particulières
qu'elles vous sembleraient trop paradoxales.
MONTJOIE, vivement.
Pas du tout! J'attache la plus grande importance à connaître votre
opinion tout entière.
DANIEL.

C'est que j'ai à traiter un sujet un peu... scientifique, et devant

mademoiselle Édith...
ÉDITH.

N'est-ce que cela? Je vais m'asseoir au piano, je ne vous entendrai

pas.
DANIEL.

Vous me pardonnez?
ÉDITH, s'éloignant et allant au piano.

A vous!
DANIEL.

Eh bien, je vous dirai que j'ai étudié le système de Darwin sur

l'origine des espèces, et j'en ai tiré des conclusions cruelles, mais
logiques. Vous savez que le naturaliste anglais a divisé tout ce qui
vit, homme, animal ou plante, en deux grands partis: celui des
vaincus, celui des vainqueurs. Telle espèce sera vaincue, parce
qu'elle est dénuée de moyens de défense; telle autre, victorieuse,
parce qu'elle est constituée pour triompher. Les animaux héritent de
leurs ascendants leur faiblesse ou leur force. Eh bien! je crois, pour
ma part, que ce que la science a reconnu exact au point de vue
physique est vrai au point de vue moral. On hérite non seulement la
beauté ou la laideur des formes, mais encore les vertus et les vices.
Il y a bien des chances pour que le fils d'un honnête homme soit un
honnête homme, pour que le fils d'une coquine soit un coquin, de
même que le petit du lion est fatalement brave et le petit de l'hyène
fatalement lâche. Le principe héréditaire de la noblesse n'a pas
d'autres fondements, et même dans notre bourgeoisie
contemporaine, où l'idée d'honorabilité a remplacé l'idée de
noblesse, vous ne verrez presque jamais une famille sans reproche
s'allier à une famille tarée. C'est injuste, c'est épouvantable,
d'accord. Vous ne nierez pas l'évidence. Je vois que ma théorie vous
surprend beaucoup: n'y attachez, si vous voulez, que l'importance
d'une fantaisie philosophique originale. (Allant à Édith.) J'ai fini ma
tirade, mademoiselle, et de nouveau je vous demande pardon
d'avoir fait le pédant pendant cinq minutes.
Montjoie est resté sur le devant de la scène.

MONTJOIE, très ému, à part.

Et c'est le fils de Coralie qui parle ainsi! Pauvre garçon! (Il regarde un
instant Daniel, qui cause avec Édith.) Cet honnête homme souffrirait trop,
je me tairai. (Haut.) Vous ne sauriez croire, monsieur, combien vos
paroles loyales m'ont ému. (Il lui tend la main.—A Édith.) Je vous ai déjà
fait compliment de votre mariage, mademoiselle. Permettez-moi
d'insister. Il y a certains hommes qu'on estime plus à mesure qu'on
les connaît davantage.
Il salue et sort.

SCÈNE IX

ÉDITH, DANIEL.

DANIEL.

Enfin, nous sommes seuls!

ÉDITH.

Et nous n'avons pas pu nous parler depuis ce matin.

Elle lui tend les deux mains.—Ils restent quelques instants à se regarder, charmés.

DANIEL.
Je suis bien heureux!...
ÉDITH.

Je suis bien heureuse!...

DANIEL.

Je me rappelle la première fois que je vous ai vue. C'était à ce bal.

Vous aviez une robe de satin blanc uni, pas un bijou; une seule fleur
dans les cheveux, ici, à droite. Vous étiez très calme. Plusieurs
personnes sont venues vous parler. Vous répondiez d'une façon
distraite, on voyait bien que votre pensée était ailleurs. Jusque-là,
vous aviez refusé de danser. Je me suis fait présenter à vous et nous
avons causé... De quoi? Je ne sais plus. Je ne faisais pas attention à
ce que je disais, je vous regardais et j'étais heureux. Quand vous
m'avez quitté ce soir-là, il m'a semblé qu'une partie de moi-même
s'en allait. Si quelqu'un m'avait vu, il se serait dit: «Daniel est fou.»
Je n'étais pas fou: je vous aimais.
ÉDITH.

Moi, j'ai demandé qui vous étiez. On m'a répondu: «—C'est lui qui
s'est battu si héroïquement sous les murs de Metz.» Alors j'ai songé
que vous deviez être bon, puisque vous étiez brave. Vous m'avez
quittée pendant dix minutes, mais je sentais que vous me regardiez
de loin. Je vous ai regardé aussi. Ma tante est venue me chercher
pour partir; j'étais toute troublée, mon regard avait croisé le vôtre.
Alors, elle m'a demandé si j'étais souffrante. J'ai hoché la tête en
souriant. Je n'étais pas souffrante: je vous aimais.
DANIEL.

Mon Dieu! Vous pouviez ne pas être à ce bal; moi-même je pouvais

ne pas m'y rendre; et c'en était fait, nous ne nous connaissions pas,
et cela me paraît impossible quand j'y pense.
ÉDITH.
C'était impossible en effet. Je m'étais formé un idéal de noblesse et
de loyauté: nous devions nous rencontrer; quand je l'ai trouvé en
vous, cela ne m'a pas étonnée: je vous attendais.
DANIEL.

Chère Édith! quand je pense que votre père a failli répondre non! Et
je vous aurais perdue, et nous nous serions aimés sans pouvoir nous
le dire!
ÉDITH.

Mais non! De même nous devions nous rencontrer, de même nous

devions nous appartenir. Si mon père vous avait répondu: non, je
vous aurais attendu.
DANIEL.

Combien de temps?
ÉDITH.

Toujours.
DANIEL.

Toujours! Un bien grand mot. L'oubli vient si vite!

ÉDITH.

Quand on oublie celui qu'on aime, c'est qu'on ne l'a jamais aimé.
DANIEL.

Édith!
ÉDITH.

Vous avez entendu M. de Montjoie parler tout à l'heure de ces deux

jeunes gens qui s'aiment et seront peut-être séparés? J'ai songé
qu'un pareil malheur pouvait nous atteindre. Eh bien, si une
impossibilité se dressait entre nous, ce serait pour moi la souffrance,
mais pas le renoncement. Je vous garderais là, bien vivant, dans
mon cœur. Vous m'appartiendriez par la pensée, et ma pensée, nul
n'a le pouvoir de la détruire. Je vivrais loin de vous, mais sans cesser
de me souvenir; car si je comprends la séparation, je n'admets pas
l'oubli. Si je n'avais pas été à vous, je n'aurais été à personne.
DANIEL.

Rien ne nous séparera jamais.

ÉDITH.

Rien.
DANIEL.

Il n'y a que deux mois que je vous aime, et je mourrais de vous

perdre!
ÉDITH.

Rien que deux! Mon amour est plus ancien que le vôtre, puisque moi
je vous connaissais... avant de vous connaître.
DANIEL, couvrant sa main de baisers.

Je suis bien heureux!

ÉDITH.

Je suis bien heureuse!

La toile tombe.
 ACTE TROISIÈME

Même décor.
Une grande table au milieu de la scène, le lendemain, de jour.

SCÈNE PREMIÈRE

GODEFROY, CÉSARINE.
Césarine lisant un journal de romans.—Godefroy furette dans ses vitrines.

GODEFROY, lui montrant un objet.

Sais-tu ce que c'est que ça?

CÉSARINE, prenant l'objet et le tournant avec dédain.

Ça?
GODEFROY.

Oui, un paysan m'a apporté hier soir ce morceau de fer rouillé et

bossué: tout autre n'aurait vu là dedans...
CÉSARINE, l'interrompant.

Qu'une vieille boîte de sardines!

 GODEFROY.

C'est une lampe à crochet du douzième siècle, entends-tu? Tu

ignores sans doute qu'à cette époque les habitants du Montalbanais
étaient sous la dépendance des abbés de Montauriol; ceux-ci,
voulant continuer à exercer leur droit de...
CÉSARINE.

Tiens! laisse-moi tranquille; tu m'agaces.

GODEFROY.

Césarine!
CÉSARINE.

Veux-tu que je te dise la vérité? Tu me rappelles cet avoué de

Toulouse que Bonchamp a beaucoup connu. Un jour qu'il se
promenait au Pirée, il ramassa une coquille d'huître énorme, et la
mit dans sa vitrine, en écrivant dessous: «Elle servit peut-être à
exiler Aristide!»
GODEFROY.

Et dire qu'elle est ma sœur!

CÉSARINE.

Au lieu de débiter des sornettes, parlons raison. Quand doit venir

madame Dubois?
GODEFROY.

Pour le contrat? d'un moment à l'autre... Bonchamp, étant le seul

occupé, a choisi son heure, hier soir, en nous quittant.
CÉSARINE.

Lui as-tu révélé?...

GODEFROY.
Quoi?
CÉSARINE.

Ce qui concerne Daniel?

GODEFROY.

L'irrégularité de sa naissance? Non, que veux-tu? Bonchamp est un

bien vieil ami. Je n'ai pas osé: je crains qu'il ne s'étonne, qu'il ne me
blâme.
CÉSARINE.

Tu le connais bien peu.

SCÈNE II

Les Mêmes, BONCHAMP, une serviette sous le bras.

BONCHAMP.

Tu vas bien. Bonjour, ma chère amie. Madame Dubois n'est pas là?
Bravo, je vois que j'arrive le premier. Elle est très bien, la tante de
Daniel.
GODEFROY, vivement.

Alors, elle te plaît?

BONCHAMP.

Beaucoup, bien que je l'aie vue hier pour la première fois. Elle adore
son neveu. Elle m'a parlé de lui avec une émotion touchante.
GODEFROY.

Aussi, plus j'y réfléchis, plus je suis satisfait de mon futur gendre. Je
suis bien obligé d'en convenir, jamais je n'aurais osé espérer pour

É
Édith un aussi beau parti.
BONCHAMP.

Je ne sais pourquoi tu me chantes cette antienne. On croirait, ma

parole, que je n'ai pas toujours été de cet avis-là. Le vrai mérite de
Daniel, c'est, à mes yeux, moins sa fortune que sa personne. Il est
intelligent, loyal et droit; je le crois incapable de mensonge; enfin il
adore Édith. Serait-il pauvre et sans position que j'applaudirais à ton
choix des deux mains.

SCÈNE III

Les Mêmes, ÉDITH.

ÉDITH, entrant vivement.

Voici madame Dubois. J'étais à la fenêtre par hasard...

BONCHAMP, souriant.

Par hasard?
ÉDITH, très confuse.

Mon parrain...
Elle s'éloigne et va au fond, près du perron, regardant.

GODEFROY, à Bonchamp.

Tu ris; ça la trouble.
BONCHAMP.

Ah l'amour, l'amour, l'amour!

CÉSARINE.

Ce notaire a des élans de poète!

 BONCHAMP.

Et pourquoi pas?
Madame Dubois et Daniel paraissent au fond.

SCÈNE IV

Les Mêmes, CORALIE, DANIEL.

CORALIE.

Est-ce que nous sommes en retard?

GODEFROY.

Pas du tout, chère madame.

CORALIE.

J'ai apporté les pièces nécessaires pour régler nos affaires d'un seul
coup. (Montrant Édith et Daniel.) Ces enfants vont aller se promener
pendant ce temps-là.
CÉSARINE.

Sous ma surveillance rigide! Le temps de mettre un chapeau... Me

voici prête! en route... (A Édith.) Passez devant, vous deux. (A Coralie.)
Je ressemble à un maître d'études qui mène des collégiens à la
promenade!
Tous les trois sortent.

BONCHAMP.

Et nous, aux affaires.

SCÈNE V
 CORALIE, GODEFROY, BONCHAMP.

BONCHAMP.

Veuillez vous asseoir, chère madame. (Tous les trois prennent place, il
déplie quelques papiers.) Un contrat de mariage est généralement chose
très embrouillée... Avez-vous remarqué les défiances naïves qu'on
s'y témoigne? La crainte d'être volées que les parties contractantes
ne se cachent pas réciproquement? J'ai souvent fait cette
supposition invraisemblable: une bande de galériens se constituant
en société... anonyme et rédigeant un acte. J'imagine qu'ils ne
prendront pas plus de précautions que les honnêtes gens.
GODEFROY, à Coralie.

Heureusement ce n'est pas notre cas! D'ailleurs Bonchamp réglera

tout. Je m'entends peu aux affaires d'argent, et j'ai peur que vous
ne soyez pas plus avancée que moi. Je ne suis guère qu'un pauvre
savant de province.
BONCHAMP, prenant une feuille de papier et un crayon.

Abandonnons Godefroy à ses méditations scientifiques. Madame, je

suis au courant de tout ce qui concerne Édith. Son père lui donne
une somme ronde de six cent mille francs: il est de plus spécifié
qu'elle sera l'unique héritière de sa tante, mademoiselle Césarine
Godefroy. Vous voyez que, de notre côté, il n'y aura pas beaucoup
d'écritures... Du vôtre...
CORALIE.

Il n'y en aura pas davantage. Mon neveu a une fortune personnelle

qui se monte à neuf cent mille francs environ, je vous remettrai le
détail complet, désignant les coupons de rente qu'il possède. Il sera
spécifié de plus que, moi aussi, je m'engage à lui laisser toute ma
fortune.
BONCHAMP.
Parfaitement. Voilà qui supprime bien des difficultés. Nous adoptons,
je suppose, le régime de la communauté? C'est le plus sage et le
plus logique.
CORALIE.

Le régime de la communauté me convient à merveille.

BONCHAMP.

Alors, chère madame, il ne me reste plus qu'à rédiger le contrat

selon la coutume; pour en finir tout de suite, je vous demanderai les
nom, prénoms et lieux de naissance des père et mère de votre
neveu.
CORALIE.

C'est que... je croyais que M. Godefroy vous avait expliqué...

GODEFROY, toussant.

Ah!
BONCHAMP.

On dirait que ma question t'offusque.

GODEFROY, se levant.

Elle ne m'offusque pas. Seulement elle m'oblige à te confier un

secret fort délicat, puisque j'ai cru devoir te le cacher jusqu'à
présent; comme il concerne uniquement Daniel, tu comprendras ma
discrétion.
BONCHAMP.

Qu'est-ce que c'est?

GODEFROY.

Madame Dubois ne pourra pas te donner les nom, prénoms et lieux

de naissance des père et mère de son neveu, attendu que Daniel
est... hum!... est... enfant naturel.
BONCHAMP, froidement.

Ah!
GODEFROY, un peu timidement.

Comprends-tu? Très loyalement Daniel est venu me faire son aveu

avant de demander officiellement la main d'Édith. Au premier
moment, j'ai été désagréablement surpris: j'ai même commencé par
déclarer à Daniel que je ne pouvais consentir au mariage. C'est ma
sœur qui a insisté. Elle m'a révélé l'amour d'Édith; elle a fait valoir
un tas de raisons... Bref, j'ai cédé. Est-ce que tu me blâmes?...
BONCHAMP.

Te blâmer, moi? Tu es fou. Daniel est-il, oui ou non, un honnête

homme? Oui ou non, est-il aimé d'Édith? C'est la seule chose que
j'examine, le reste m'importe fort peu. La bâtardise, en ce temps-ci,
n'est qu'un malheur, ce n'est plus une tache.
GODEFROY, vivement, avec joie.

Que je te serre la main! Ah! tu m'enlèves un poids.

BONCHAMP, haussant les épaules.

Un poids! Seulement cela change mes idées pour le contrat. Loin de

moi la pensée de vouloir être indiscret, chère madame; mais j'ai
besoin d'être fixé. De qui Daniel tient-il sa fortune: de son père ou
de sa mère?
CORALIE, avec beaucoup de calme.

De sa mère, qui était ma sœur. Je vous saurai gré de ne pas

m'interroger trop longuement sur un passé qui m'est toujours
douloureux. Notre famille habitait Paris. Mon père possédait une
belle fortune. Ma pauvre sœur eut le malheur de se laisser séduire
par un de nos cousins qui était marié; elle mourut en mettant son
fils au monde. Quand je devins veuve, je me rattachai à cet enfant,
le seul parent que j'eusse encore. Vous savez le reste, je n'ai plus
rien à vous apprendre.
BONCHAMP.

Je comprends, chère madame, tout ce qu'une pareille conversation

doit avoir de pénible, mais vous me permettrez d'insister sur un
point. Votre neveu n'a pas de nom, car, sans doute, il a été déclaré
de père et mère inconnus. Votre sœur n'étant pas sa mère,
légalement parlant, ne lui a donc pas laissé sa fortune par voie
d'héritage naturel. Il a dû intervenir un testament. Il faut que vous
donniez le nom du notaire chez qui il a été déposé. De plus, votre
neveu a eu un tuteur; ses intérêts ont évidemment été mis en
bonnes mains, mais veuillez remarquer que j'agis ici autant comme
officier ministériel que comme ami. Par conséquent, il ne m'est pas
permis de négliger un seul détail. Vous annoncez l'intention de
léguer toute votre fortune à Daniel, c'est une idée généreuse qui ne
m'étonne nullement de vous: encore faut-il que nous puissions
établir votre droit à faire une pareille libéralité; pour cela, j'ai besoin
de l'acte de décès de monsieur votre mari.
CORALIE.

Mon Dieu, cher monsieur, ma pauvre sœur n'a pas fait de testament;
j'étais son héritière naturelle, la fortune de mon père se partageait
donc entre nous deux; à son lit de mort, elle m'a chargée de veiller
sur les intérêts de son enfant, et de lui remettre purement et
simplement la part qui lui reviendrait. C'est ce que vous appelez, je
crois, un fidéicommis en termes de droit?
BONCHAMP.

Un fidéicommis.
CORALIE.

De même pour la question du tuteur, Daniel n'en a pas eu. Quant à

l'acte de décès de mon mari, je l'ai apporté, pensant en effet qu'il
serait nécessaire.
BONCHAMP.

Il est regrettable, très regrettable qu'il n'y ait pas eu testament.

Certes, en réfléchissant, je comprends tout l'avantage d'un
fidéicommis. Comme vous le dites fort bien, vous étiez l'héritière de
votre sœur, donc sa fortune se transmettait tout naturellement à son
fils. Ce qui me contrarie, c'est la nécessité où nous serons de
remonter plus haut. Monsieur votre père est mort à Paris, chère
madame? Veuillez me donner le nom du notaire de votre famille, je
lui écrirai pour qu'il m'envoie copie du testament par lequel vous
avez hérité.
CORALIE, troublée.

Oh! mon Dieu.

GODEFROY, un peu agacé.

En vérité, mon cher ami, tu me confonds... Pourquoi fais-tu

intervenir le père de madame Dubois à propos du contrat de Daniel?
BONCHAMP.

J'avoue que j'excuse ton étonnement et celui de madame, mais je

crois être dans la sagesse et la raison. Le pire mal des choses
irrégulières, c'est qu'elles prêtent à gloser à tout le monde. Il sera
impossible de cacher la vérité; si tu l'as espéré un moment, tu n'es
qu'un naïf. La publication des bans est une petite machine très
simple, mais très indiscrète. Le premier venu pourra lire: «Promesse
de mariage entre demoiselle Édith-Jeanne Godefroy, fille légitime
de...» et «le sieur Daniel, capitaine d'artillerie, fils de personne».
Sois bien persuadé que les cancans iront bon train. J'entends ces
braves gens d'ici: (Avec l'accent du Midi.) «Té! Il est bien riche pour un
bâtard.» «Eh! pécaïré! qui sait ce qu'aura fait mademoiselle sa
mère?» Le second jour, une âme charitable insinuera que ladite
fortune est d'origine douteuse; le troisième, ce doute deviendra une
certitude; le quatrième, on fournira les détails les plus inattendus.
Allons donc! cela me révolte par avance! Avoue franchement la
situation. La franchise est le sauf-conduit de l'honneur. Tu inviteras
tout Montauban à la signature du contrat: il est nécessaire que
chacun puisse dire que les choses se font honnêtement entre
honnêtes gens que nous sommes. «Le capitaine Daniel apporte neuf
cent mille francs, laissés par mademoiselle X... sa mère, laquelle
avait hérité de ses père et mère, ainsi qu'il ressort de tel testament
déposé chez maître X..., notaire à Paris. Il apporte en outre la
fortune à venir de sa tante madame Dubois, veuve de M. Dubois,
etc... etc...» De cette façon les commérages tomberont d'eux-
mêmes. On saura que la fortune de ton gendre est de source pure,
que la famille de sa mère était riche, que le mari de sa tante était
riche; les plus malveillants seront condamnés au silence, car nul
n'osera dire qu'étant né en dehors du mariage, Daniel est riche en
dehors de l'honneur!
CORALIE, écrasée.

Tout est perdu!

BONCHAMP.

Vous ne dites rien, chère madame.

CORALIE, très troublée, se levant.

Je dis que...
BONCHAMP, à part.

Comme elle est troublée! (Haut.) Enfin, quelle est votre opinion?
CORALIE, toujours très émue.

Mon opinion est... est, au contraire, d'éviter le bruit. Moi, je serais

d'avis de n'inviter personne à la signature du contrat. Je puis vous
remettre l'acte de décès de mon mari ainsi que le détail de... de la
fortune qu'il m'a laissée. Quant au testament de mon père, il n'en a
pas fait.
 BONCHAMP.

Le notaire de votre famille a gardé la minute de la liquidation; vous

voyez que rien n'est plus facile que de se procurer la preuve que je
désire. (L'épouvante a gagné peu à peu Coralie.—A part.) Oh! oh! il y a
quelque chose. (Haut) C'est pourtant bien simple.
CORALIE.

Oui, mais je suis forcée de vous demander quelques jours, car, en

vérité, je ne m'attendais pas à ces difficultés. Le temps d'écrire à
mon notaire...
BONCHAMP, avec intention.

Donnez-moi son nom: ces démarches-là nous regardent, nous

autres. C'est moi qui écrirai.
CORALIE.

Vous voulez?...
BONCHAMP, même jeu.

Quelle raison auriez-vous de refuser?

CORALIE, vivement.

Je ne refuse pas. Je vous communiquerai tous ces renseignements-

là demain.
BONCHAMP, toujours avec insistance.

Je préférerais en finir immédiatement, d'autant que nous désirons

tous que ces enfants soient mariés le plus vite possible..
CORALIE.

Votre conversation de jurisprudence m'a un peu étourdie, et pour

l'instant...
GODEFROY.
Voulez-vous passer dans la chambre de ma sœur? Elle donne sur le
jardin.
CORALIE.

Volontiers. (A part.) Que faire?

BONCHAMP.

Je vous serais obligé de me remettre le plus tôt possible ce que je

vous demande.
CORALIE.

Oui, oui... (A part.) Je trouverai!

SCÈNE VI

BONCHAMP, GODEFROY.

BONCHAMP.

Veux-tu que je te dise mon opinion?

GODEFROY.

Sur quoi?
BONCHAMP.

Sur madame Dubois. Je ne serais pas étonné qu'elle fût, non la

tante, mais la mère de Daniel.
GODEFROY.

La mère de Daniel! Qu'est-ce que tu me chantes là?

BONCHAMP.
C'est le seul moyen d'expliquer son trouble, ses hésitations, lorsque
je lui demandais tout à l'heure certaines pièces que je crois
nécessaires.
GODEFROY.

La mère de Daniel!
BONCHAMP.

Vas-tu prendre les choses au tragique? Tu savais Daniel enfant

naturel, seulement tu le croyais orphelin. Il se trouve que sa mère
existe: voilà toute la différence entre ce qui est et ce qui était. En
somme, la pauvre femme n'a pas osé nous avouer la vérité et je ne
saurais lui en vouloir.
GODEFROY, tombant assis.

Tu me renverses!
BONCHAMP.

C'est que tu es facile à renverser.

GODEFROY, se relevant.

La mère de Daniel!
BONCHAMP.

Quand tu répéteras cela jusqu'à demain? Mme Dubois est, après tout,
une femme intelligente, distinguée; mais sa tendresse pour Daniel,
c'est de la tendresse maternelle.
LE DOMESTIQUE, entrant.

M. de Montjoie, M. Morisseau.
GODEFROY.
Il y a un moyen bien simple de recueillir des renseignements. M. de
Montjoie la connaît, je crois, à ce que m'a dit Édith, ou du moins ils
ont un ami commun.
BONCHAMP.

Ah!
GODEFROY.

Faites entrer.
BONCHAMP, à part et préoccupé.

Un ami commun?

SCÈNE VII

Les Mêmes, MONTJOIE, CLAUDE MORISSEAU.

CLAUDE.

J'allais travailler en forêt, lorsque j'ai rencontré M. de Montjoie qui

venait vous faire visite, je suis entré avec lui. (A Godefroy.) A propos,
j'irai demain à votre ferme de Cos. Oui, vous avez là une très belle
basse-cour: elle m'inspire, je veux écrire une symphonie d'après les
vrais principes de notre école symboliste, car la musique sera
symboliste ou elle ne sera pas.
MONTJOIE.

Moi je viens vous faire mes adieux.

GODEFROY.

Vous partez?
MONTJOIE.
Demain ou après-demain au plus tard.
BONCHAMP.

Est-ce que votre absence se prolongera?

MONTJOIE.

Je ne sais au juste. Peut-être mon goût de voyage me reprendra-t-il.

Quand on est seul dans la vie, comme moi, peu importe qu'on soit
ici ou là, à droite ou à gauche.
CLAUDE.

Décidément, depuis hier vous êtes sous une influence de tristesse.

MONTJOIE.

Moi?
CLAUDE.

Oh! je m'en suis bien aperçu quand nous sommes revenus

ensemble. Est-ce la vue de madame Dubois qui a réveillé les
souvenirs de votre passion d'autrefois?
MONTJOIE.

Monsieur Morisseau!
BONCHAMP.

Hein?
CLAUDE.

Je ne commets pas d'indiscrétion, que je sache? Imaginez-vous que

la tante du capitaine Daniel ressemble à s'y méprendre à Coralie,
cette femme que notre ami M. de Montjoie a tant aimée.
GODEFROY.
Vraiment?
BONCHAMP, se rapprochant.

Ah!
MONTJOIE.

Monsieur Morisseau...
CLAUDE.

Vous ne nierez pas que cette ressemblance vous ait frappé. Moi je
l'ai trouvée sans hésiter. Le coup d'œil de l'artiste! (A Bonchamp.) C'est
vraiment extraordinaire. Les mêmes yeux profonds et étranges, le
même nez, la même bouche. Une seule différence: Coralie était
blonde, et madame Dubois est brune. (A Montjoie.) Vous ne direz pas
le contraire.
MONTJOIE.

Non pas, mais ce sont des souvenirs pour le moins inutiles à

évoquer.
BONCHAMP, avec intention.

Vous les évoquiez vous-même hier encore. Mais, j'y pense, vous
avez causé longuement avec madame Dubois. Vous avez un ami
commun, paraît-il?
MONTJOIE.

En effet, et nous nous sommes trouvés tout de suite en pays de

connaissance.
BONCHAMP, bas, à Godefroy.

Emmène cet imbécile de Morisseau, et laisse-moi seul avec M. de

Montjoie.
GODEFROY.
Mais...
BONCHAMP.

Fais ce que je te dis.

GODEFROY.

Enfin... (A Claude.) Vous allez en forêt? Je vais faire un tour avec vous.
CLAUDE.

Tiens!... c'est une idée! (A Montjoie.) Je vous reverrai avant votre

départ?
MONTJOIE.

Je sors avec vous.

GODEFROY, vivement.

Mais non, attendez donc un peu. Ma fille et ma sœur ne tarderont

pas à revenir. Elles seront charmées de vous voir. A tout à l'heure!
Il sort avec Claude.

SCÈNE VIII

BONCHAMP, MONTJOIE.
Un moment de silence.

BONCHAMP, brusquement à Montjoie.

Monsieur, vous êtes un galant homme; il me répugnerait de vous

mentir. Je vais donc droit au but. Vous connaissez madame Dubois;
vous l'avez sûrement vue déjà avant de vous rencontrer ici avec elle.
Je n'en veux pour preuve que votre long entretien d'hier. Or, je
soupçonne quelque chose de grave. Le capitaine Daniel a avoué à
mon ami Godefroy qu'il était enfant naturel, Godefroy a passé par là-
dessus et il a bien fait. Mais le capitaine se croit orphelin; en cela il
se trompe. Sa mère vit, sa mère c'est madame Dubois.
MONTJOIE.

Vous m'apprenez là, monsieur, des choses que je n'ai pas le droit de
connaître, et...
BONCHAMP.

Certains secrets peuvent toujours être confiés à certains hommes;

j'ajoute qu'il dépend de vous de sauver cette maison d'une
catastrophe peut-être; c'est un service que j'attends de vous: car
vous êtes notre ami, n'est-ce pas?
MONTJOIE.

J'espère que vous me faites l'honneur de n'en pas douter.

BONCHAMP.

Non. Je reprends. Tout à l'heure Godefroy et moi nous avons discuté

avec madame Dubois les clauses du contrat de mariage. Si certains
points sont restés trop obscurs, d'autres sont devenus trop clairs;
elle a parlé plusieurs fois de sa sœur, de la mère de son neveu. Elle
a donc cherché à nous tromper sur un point; elle peut chercher à
nous tromper aussi sur le reste. Et voici que l'on me parle
soudainement d'une ressemblance extraordinaire qui existe entre
madame Dubois et cette fille nommée Coralie; voici que vous,
l'ancien amant de Coralie, vous qui ne connaissiez pas madame
Dubois, vous causez longuement avec elle... Eh bien!... c'est
absurde, soit!... mais je me dis: «Est-ce que par hasard madame
Dubois et Coralie ne seraient point la même et unique personne?»
MONTJOIE, froidement.

Je l'ignore, monsieur; puis, tout cela ne me regarde pas.

BONCHAMP.
Ou j'ai raison, ou j'ai tort dans mon soupçon. Si j'ai tort, dites-le-
moi; si j'ai raison, souvenez-vous de l'accueil qu'on vous a fait en
cette maison, et décidez, dans votre conscience, si vous devez vous
taire ou parler.
MONTJOIE.

Monsieur...
BONCHAMP.

Un mot tranchera la question; donnez-moi votre parole d'honneur

que madame Dubois n'est pas Coralie, et je me tiens pour satisfait.
(Montjoie se tait.) Votre silence répond, prenez garde!

MONTJOIE.

Vous vous trompez, monsieur: je me tais parce que je n'ai rien à

vous apprendre. Si vous n'étiez pas l'ami de M. Godefroy, le parrain
de mademoiselle Édith, je m'étonnerais du droit que vous prenez de
faire cette inquisition dans ma vie passée. Vous semblez
m'encourager à remplir un devoir. Permettez-moi de vous dire que je
n'ai besoin de personne pour connaître le mien. Au surplus, brisons
là, monsieur.
BONCHAMP.

Ainsi vous refusez de répondre?

MONTJOIE.

Mais je n'ai rien à vous dire.

SCÈNE IX

Les Mêmes, CORALIE.

CORALIE, entrant, une lettre à la main.

Voici les renseignements que vous m'avez demandés. (Elle s'arrête en
apercevant Montjoie.—Un silence. Montjoie s'incline profondément devant elle et
sort.—A part.) A-t-il parlé?

SCÈNE X

CORALIE, BONCHAMP.
BONCHAMP, à part.

Il faut pourtant que je sache...

CORALIE, inquiète.

C'est moi qui fais fuir monsieur de Montjoie?

BONCHAMP, à part.

C'est le seul moyen.

CORALIE.

Je vous disais donc que je vous apporte...

BONCHAMP.

Inutile, madame.
CORALIE.

Ah!
BONCHAMP.

M. de Montjoie m'a tout dit.

CORALIE.

Le misérable!
BONCHAMP.
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com