On the working principles of energy saving devices

Ship H y d r o m e c h a n i c s and S t r u c t u r e s L a b o r a t o r y M e k e l w e g 2, 2 6 2 8 CD D e l f t

Delft University of Technology

TUDelft

On the working principles of energy saving d e v i c e s .

by

Tom van T e r w i s g a

Report No. 1 8 8 9 - P 2013

Proceedings of the Third I n t e r n a t i o n a l Symposium on Marine Propulsors, s m p ' 1 3 , L a u n c e s t o n , T a s m a n i a , A u s t r a l i a , May 2 0 1 3 .

International Symposiums on Marine

Propulsors

Homepage of the series of symposiums on marine propulsors

Home Topics Previous Symposiums Committees Contact

Proceedings of the Third International Symposium

on Marine Propulsors - smp13

5 - 8 May 2013, Launceston, Tasmania, Australia

Edited by: Jonathan Binns, Renee Brown & Neil Bose

ISBN (printed proceedings): 978-0-646-90334-7

© smp chair committee (Kourosh Koushan and Sverre Steen)

Publisher: Australian Maritime College, University of Tasmania

www.amc.edu.au/

Organised by:

•

Australian Maritime College

Sponsored by:

•

MARINTEK

•

Norwegian University of Science and Technology

•

Australian Maritime College

•

Cussons

Table of contents

Session I A - Ventilation, Off Design, Unsteady

Propulsion 1

1A.1

Numerical and Physical Investigation of a Surface- 1

Piercing Hydrofoil

Yin Lu Young, Stefano Brizzolara

1 A.2 Propeller and Rudder in Off-Design Conditions 9

Thomas LCicke

1 A.3 On t h e Propulsive Efficiency of Unsteady Propulsors 18

Michael Krieg, Kamran Mohseni

Session I B - Propulsor-Ice Interaction

1 B.I Experimental Study o n Ice Management t h r o u g h the 26

use of Podded Propeller Wash

Jenna M. Ferrieri, Brian Veitch, Ayhan Akinturl<

1 B.2 Podded Propeller Ice Interaction in a Cavitation Tunnel 34

R. Sampson, M. Atlar, J. W. St John, N. Sasaki

1 B.3 Propeller-Ice Impacts Measurements w i t h a Six- 47

Component Blade Load Sensor

Joris Brouwer, Gerco Hagesteijn, René Bosman

Session 2A - T-Pods And Thrusters 1

2A.1 Energy Saving Possibilities in Twin or Triple Propeller 55

Cruise Liners

Raimo Hamalainen,jaap van Heerd

2A.2 Numerical Investigation of Ducted Propeller Added 69

Mass

Suzanne Hutchison, Sverre Steen, AbhishekSanghani

2A.3

An Estimation Method of Full Scale Performance f o r 78

Pulling Type Podded Propellers

Session 2B - Impulse and Steady State Waterjets

2B.1

Waterjet System Performance and Cavitation Test 87

Procedures

Jie Dang, Runwen Liu, Ctiristiaan Pouw

2B.2

Design and Commissioning Tests f o r Waterjet Self- 97

Propulsion Testing of a M e d i u m - S p e e d Catamaran

Ferry using a Single Demihull

Konrad ZCirctier, Neil Bose, Jonattian R. Binns, Giles

Thomas, Gary Davidson

2B.3

Converging Radial Velocity and Thrust Enhancement in 104

Nature's Jetting Swimmers

Michael Krieg, Kamran Mohseni

Session 3A - Complex Flows

3A.1

Hub Effect in Propeller Design and Analysis 110

5. Brizzolara, S. Gaggero, D. Grassi

3>t\2.

The Hydrodynamic Performance of Propellers with 120

Trans-Velocity Sections in Inclined Shaft Conditions

Ching- Yeh Hsin, Shang-Sheng Chin, Kuan-Kai Chang,

Ya-LIn Tsai, Jeng-Lih Hwang

3A.3

Development of a Shape-Adaptive Composite 128

Propeller Using Bend-Twist Coupling Characteristics

of Composites

Manudha T. Herath, B. Gangadhara Prusty, GH. Yeoh,

M. Chowdhury, Nigel St John

3A.4

An Advanced Scaling Procedure f o r Marine Propellers 136

Session 3B - Unconventional Propulsors 1

38.1

A Development of a Propeller w i t h Backward Tip 143

Raked Fin

Yasuhiko Inukai

3B.2

Simulation of Unsteady Interaction Forces on a Ducted 149

Propeller with Pre-swirl Stators

Zhi-Qiang Rao, Wei Li, Chen-Jun Yang

3B.3

Proposal and Fundamental Experiments on 156

Unconventional Ducted Propulsor w i t h o u t Rotating

Blades

Kazuo Suzuki, Takeshi Yokota

3B.4

Development and Performance Estimates of a Ducted 161

Tandem CRP

iVIasaJu Igeta, Hongbin Yuan

Session 4A - Cavitation, Flow Measurement &

Visualization 1

4A.1

Axiom Propeller tests in t h e Emerson Cavitation 168

Tunnel

iVl. Atiar, R. Sampson, K.C. Seo, A. Watts, D. Watts

AfK2

A Modern Approach t o t h e Representation and use of 176

the KCA Systematic Propeller Series

G.I-I.G iVlitcheil, R. Sampson, IVI. Attar

4A.3

Propeller-Hull Interaction in a Single-Screw Vessel 185

Andrea Pecoraro, Fabio Di Felice, Mario Felli,

Francesco Salvatore, Michele Viviani

4A.4

A Numerical Method f o r t h e Analysis of Unsteady 193

Cavitating Rotor and Stator Interaction

Ye Tian, SpyrosA. Kin nas

Session 4B - OceanNoise & Noise & Vibration

48.1

Hydro-Acoustic Noise f r o m Merchant Ships - Impacts 201

and Practical Mitigation Techniques

IVIartin Renilson, Russell Leaper, Oliver Boisseau

48.2

Investigation of Scale Effects on Propeller-Induced 209

Pressure Fluctuations by a Viscous/lnviscid Coupling

Approach

S. Berger, M. Bauer, M. Druckenbrod, M.

Abdel-Maksoud

48.3

Detached Eddy Simulation of the Flow Behind an 218

Isolated Propeller

Roberto Muscari, Andrea Di Mascio

48.4

Design, Analysis and Experimental Characterization of 227

a Propeller in Decelerating Duct

S. Gaggero, CM. Rizzo, G. Tani, M. Viviani

Session 5A - Ocean Renewable Energy

5A.1

Performance Predictions of a Horizontal Axis Tidal 236

Stream Turbine Considering t h e Effects of Blade

Deformation

Se Wan Park, Sunho Park, Shin Hyung Rhee

5A.2

ParametricAnalysisof Horizontal Axis Tidal Turbine 242

Hydrodynamics f o r O p t i m u m Energy Generation

Session 5B - Powering Prediction

5B.1

A Study on the Powering Performance of Drillship in 257

Transit Mode w i t h Azimuth Thrusters

Seokcheon Go, Youwon Ahn,Jeonghu Kim, IHeungwon

Seo

5B.2

A Holistic Design Approach f o r Propulsion Pacl<ages 263

Lars Greitsch, Markus Druckenbrod, Sven Bednarek,

Hans-Jürgen i-ieinke

Session 6A - Unconventional Propulsors 2

6A.1

Prediction of t h e Unsteady Cavitating Performance of 269

Ducted Propellers Subject to an Inclined Inflow

SpyrosA. Kin nas, Chan-l-ioojeon. Jay Purohit, Ye Tian

6A.2

Round and Round with Paddlewheel Propulsion 279

Robert Ciifford, Tim Roberts

6A.3

Contemporary Bulk Carrier Design t o Meet IMO EEDI 283

Requirements

Anton Minchev, Michael Schmidt, Soeren Schnack

Session 6B - Ventilation, Off Design, Unsteady

Propulsion 2

6A.1

Numerical Analysis of Unsteady Open Water 292

Characteristics of Surface Piercing Propeller

Kohei Himei

CFD Analysis of Propeller Performance in Oblique Flow 298

6A.3

Experimental Investigation of the Effect of Waves and 306

Ventilation in Depressurised Conditions on a

POD-Propeller of a Cruise Liner Model

Joris Brouwer, Gerco l-iagesteijn

Session 7A - Energy Saving Devices 1

7A.1

An Automatic Optimization Process f o r Optimal 314

Ducted Propeller Design and Its Application Based on

CFD Techniques

Long Yu, Marl<us Druclcenbrod, Martin Greve,

Moustafa Abdel-Malcsoud

7A.2

An Integrative Design Method of Propeller and PBCF 320

(Propeller Boss Cap Fins)

CAIHao-peng MA Cheng CHEN Ke,

QiANZheng-fang, YANG Chen-jun

Session 7B - Renewable & Low Environmental Impact

Propulsion 1

7B.1

ZEUS and NOAH Projects of NMRI 324

Noriyul<i Sasal<i

7B.2

Energy Audits on Australian Fishing Vessels 331

John Walceford, Neil Bose

Session 8A - T-Pods And Thrusters 2

8A.1

A Comparison of Panel Method and RANS Calculations 338

f o r a Ducted Propeller System in O p e n - W a t e r

J. Baltazar, D. Rijpkema, J.A.C. Falcao de Campos,J.

Bosschers

8A.2

Combination of Pod, CLT and CRP Propulsion f o r 347

Improving Ship Efficiency: the TRIPOD Project

A. Sanchez-Caja, M. Pérez-Sobrino, R. Quereda, M.

Nijiand, T. VeikonheimoJ. Gonzalez-Adalid, I. Saisto,

A. Auriarte

8A.3

Study on Hydrodynamic Performance of Podded 358

Propulsor at Steering Conditions

Shen Xingrong, Cai Yuejin

Session 8B - Flow Measurement & Visualization 2

88.1

The Modeling of Hub Vortex f o r Numerical Analysis of 365

Marine Propeller Using a Simple Surface Panel Method

"SQCM"

Takashi Kanemaru, Tomohiro Ryu, Akira Yoshitake,Jun

Ando, Kuniharu Nakatake

88.2

Hydroacoustic and Hydrodynamic Analysis of a 373

Propeller- Rudder Configuration by Pressure Signal

Wavelet Decomposition and Optical Techniques

Mario Felli, Silvano Grizzi, Massimo Falchi

88.3

Measurements and Computations f o r 8lade Spindle 381

Torque of Controllable Pitch Propellers in Open Water

Isao Funeno, Christiaan Pouw, René Bosman

Session 9A - Cavitation

9A.1

URANS Simulations of Cavitation and Hull Pressure 389

Fluctuation f o r Marine Propeller w i t h Hull Interaction

Kwang-Jun Paik, Hyung-Gil Park, Jongsoo Seo

9A.2

Parameter Influence Analysis in Propeller Optimisation 397

9A.3

Study on Marine Propeller Running in Bubbly Flow 405

Chiharu Kawakita

9A.4

Full Scale and Model Scale Propeller Ventilation 413

Behind Ship

Luca Savio, Silas Spence, Kourosh Koushan, Sverre

Steen

Session 9B - Power Prediction

9B.1

Numerical Simulation of Propeller-Hull Interaction 421

and Determination of The Effective Wake Field Using a

Hybrid RANS-BEM Approach

Douwe Rijpkema, Bram Starke, Johan Bosschers

9B.2

Self-Propulsion RANS Computations w i t h a Single- 430

Screw Container Ship

Vladimir I. Krasilnikov

9B.3

Measurement of Speed Loss Due t o Waves 439

Sverre Steen, Zhenju Chuang

9B.4

Uncertainty in Bollard Pull Predictions 447

Arthur Vrijdag, Jochem de Jong, Main van Nuland

9B.5

Reliability Assessment of Ship Powering Performance 454

Extrapolations Using Monte Carlo Methods

Iwan M. Kamal, Jonathan Binns, Neil Bose, Giles

Thomas

Session 10A - Biomimetic Propulsion

10A.2

Flexible Elliptic Oscillating Duct. Taking t h e FOD One 472

Step Further

Gerasimos Poiitis, Theodoros loannou, Vasileios

Tsarsitalidis

10A.3

Marine Propulsor based on a T h r e e - D e g r e e - o f - 481

Freedom Actuated Spherical Joint

Bassem Sudki, Michel Lauria, Flavio Noca

Session 10B - Viscous Flow

10B.1

Model and Full Scale CFD Analysis of Propeller Boss 486

Cap Fins (PBCF)

Takafumi Kawamura, Kazuyuki Ouchi, Susumu

Takeuchi

10B.2

Influence of Tip Roughness and Application Area on 494

Tip Vortex Pressure

Christian KrCiger, Nikolai Kornev, Mathias Paschen,

Christian Sem low

10B.3

Full Scale Thruster Performance and Load 501

Determination Based on Numerical Simulations

Norbert Bulten, Rik Suijkerbuijk

Session 11A - Energy Saving Devices 2

11A.1

On t h e Working Principles of Energy Saving Devices 510

Tom van Terwisga

11 A.2

The Becker Mewis Duct (r) - Challenges in Full-Scale 519

Design and new Developments f o r Fast Ships

11 A.3 Numerical Study o f Energy-saving Mechanism of Duct 528

on a VLCC with Real-geometry Propeller

Yuhai, Kong Weiping, Cai Rongquan, Wangjinbao,

Zhang Yuefeng

I I A . 4 Study on Performance of a Ship Propeller Using a 536

Composite Material

Tadashi Tal<etani, Koyu Kimura, Satol<o Ando, Koutal<u

Yamamoto

Session 11B - Renewable & Low Environmental Impact

Propulsion 2

11 B.I Experimental Characterization of Collective and Cyclic 542

Pitch Propulsion f o r Underwater Vehicle

Poowadol Niyoml<a, Neil Bose, Jonathan Binns, Hung

Nguyen

11 B.2 The Effect of a Fixed Foil on Ship Propulsion and 553

Motions

Eirik B0ckmann, Sverre Steen

11 B.3 "Wind Challenger" t h e Next Generation Hybrid Sailing 562

Vessel

Kazuyuki Ouchi, Kiyoshi Uzawa, Akihiro Kanal,

Masanobu Katori

I I B . 4 LNG The New Fuel f o r Fast Ferries 568

Gary Davidson, Tim Roberts

« smp'11

On the working principles of Energy Saving Devices

Tom van Tenvisga

'iVIaritime Research Institute Netherlands (MARIN), Wageningen, The Netherlands

^Delft University of Teclinology, Faculty of 3ME, Maritime and Transport Technology, Delff, The Netherlands

A B S T R A C T

Tlie aim o f this paper is to explain tlie principal mechanisms f o r the reduction in power demand by so called "Energy Saving Devices" (ESDs). This paper follows a similar approach as taken by Wald (1965) and Dyne (1995), where the authors use considerations o f the momentum and energy equations.

Different energy saving concepts are evaluated on their contribution to the abatement o f energy losses. The discussion on hydrodynamic mechanisms is then focused on Pre- or Upstream Ducts where four different mechanisms are hypothesized. The effect o f these mechanisms is then investigated f o r a systematically varied series o f Pre-Ducts on an axisymmetrlc body i n deeply submerged conditions. It is concluded that the benefits o f this Pre-Duct configuration do not come f r o m the same principles as used by ducts around propellers, nor that they come f r o m an improved propeller-hull interaction. I t is recommended to further study the remaining hypothesis that a Pre-Duct conditions the f l o w into the propeller i n a favorable way so that the propeller operates at a higher efficiency.

Keywords

Energy Saving Devices, Energy considerations, CFD

1. INTRODUCTION

This paper aims to review working mechanisms o f frequently used Energy Saving Devices. Before discussing the working principles, the energy losses produced by an open propeller, and a method to assess the effect o f propeller-hull interaction on delivered power is reviewed. A f t e r a review o f w o r k i n g mechanisms, the discussion is focused on the principles o f pre- or upstream ducts. Four different mechanisms are hypothesized, which are subsequently investigated by C F D computations on a systematically varied series o f Pre-Ducts on an axisymmetrlc deeply submerged body. The paper concludes with an evaluation o f these mechanisms f o r the Pre-Duct and draws conclusions on the computational modeling o f the propeller i n CFD computations.

2. E N E R G Y B A L A N C E CONSIDERATIONS

This section aims at introducing the energy loss terms i n the wake o f the propelled ship, which f u l l y accounts for the power required to propel the ship.

Let us start w i t h the relatively simple case o f a deeply submerged axisymmetrlc hull in an axial f l o w as depicted in Figure 1.

A n interesting propeily o f the velocity distribution in this plane f o l l o w s f r o m Newton's second law. The resistance force that is experienced by the hull due to the water f l o w , is completely represented by the change i n momentum f l u x between the incoming, undisturbed momentum f l u x , and the outgoing momentum f l u x at the considered transverse exit plane in the wake.

Limiting streamline boundary layer / ...-^ / /

/

- /

Figure 1 Velocitj' profile in the wake of the bare hull. The hatched area in the velocity profile indicates the hull induced

velocities.

We now introduce a propeller, to counteract the resistance force on the hull (see Figure 2), and obsei-ve that the wake and the resistance o f the hull have changed due to the propeller suction. We can now rewrite the momentum equation as follows (see e.g. Wald (1965)):

(1)

Where Rp = resistance force on the hull w i t h active propeller and subscript P refers to the self propelled condition. Limiting streamline boundary layer

/

.

- V

Figure 2 Velocity profile in the wake of the self propelled hull. The hatched area in the velocity profile indicates the

propeller induced velocities.

likely to be different f r o m the resistance o f the bare hull, the difference being accounted f o r by a thrust deduction fraction t:

T { l - t ) = R„ (2) A n important conclusion f r o m Newton's second law and

the momentum balance consideration is that whatever BSD we mount to the hull, it does not produce a net change in momentum f l u x between i n and outflow plane, as long as the ship moves at a steady speed. This implies that a change in propeller thnist, cannot directly be interpreted as a gain or loss i n power requirement. Consequently, all effects o f BSDs must appear through an effect in the energy losses i n the f l o w .

These energy losses are obtained f r o m a consideration o f the energy balance o f a control volume that contains the ship-propulsor system. This energy balance in words states that the power needed to propel the ship must be equal to the hydraulic losses that can be traced back i n the wake o f the ship. The energy balance f o r the self propelled ship i n steady state, can thus be written as:

Pp, = AXL„ + PRESL„ + TRANSL,yp (3) The three loss terms on the right hand side can be expressed in local velocities and pressures by applying the conservation laws to a control surface in a f l u i d . A n additional term is present in the pressure loss term, that should be expressed i n terms o f heat. This energy f l u x results f r o m a conversion o f kinetic energy into heat caused by viscous dissipation.

Let us now take a closer look at the different loss terms. The rationale f o r this analysis is that i f we know the magnitude and the origin o f the energy loss, we may be able to recover part o f these energy losses with an Energy Saving Device.

To this end Figure 3 defines the different (time averaged) axial velocity profiles that can be distinguished in the wake o f the ship and the constituting components.

V

\

•••"A

1

Wakt from t^t prapiliir ii Wikeffcmthe bare

Figure 3 Axial velocity profiles in the wake of the ship, the propeller and the hull

propulsor.

The propulsor produces thrust by accelerating the flow through the propeller disk, thereby increasing both the axial momentum and the energy flux i n axial direction (See Figure 3). For a given thrust, the required axial acceleration o f the flow follows f r o m the momentum equation and f r o m that, the axial component o f the energy flux.

The term "axial losses" is also used i n the actuator disk model f o r a propeller, where the propeller is represented by a disk in a non-viscous flow, which exerts an axial force on the flow. This model yields the f o l l o w i n g relation f o r the axial losses i n the far field :

(4)

These propeller axial losses are accounted f o r by the well known ideal efficiency r]^, which solely depends on the propeller loading Cj.:

AXL, l + yJÏTc'p

(5)

Where AXL. refers to the axial losses i n the actuator disk model

'jpülAp

and Ap = propeller disk area.

For a real propeller in open water conditions, the non-uniform velocity distribution and viscous losses contribute to additional axial losses.

A x i a l losses cannot easily be separated in a real flow as the origin o f the change in axial velocity cannot be traced. These combined losses in the propeller-hull system w i l l be referred to as axial losses AXL,yp i n the f o l l o w i n g , where the subscript WP refers to the Wake o f the self Propelled condition.

The non-uniform velocity distribution does have an effect on the axial losses, as was demonstrated f o r the wake o f a waterjet by Scherer et al. [2001]. The effect o f the non-uniformity o f this wake was computed to give 2 to 3% more axial kinetic energy losses than a u n i f o r m o u t f l o w w o u l d give at the same thrust coefficient.

2.1.2 Pressure losses

The pressure losses and the viscous dissipation term, w i l l be referred to as Pressure Losses i n the f o l l o w i n g :

PRESL,yp = l j { p - P o ) ii.uTdA + yy,,,.

(6)

losses for a propeller in open water conditions. I n self propelled condition, the transverse velocities are likely to increase due to the slope o f the buttocks and waterlines in the aftbody o f the hull. The transverse losses are defined by the energy flux containing the tangential and radial velocity components:

(7)

Analogous to the axial losses, the non-uniformity in the wake also causes an increase in tranverse energy losses. This is caused by the finite number o f blades and the radial load distribution over the blades o f a propeller in open water. I n addition, a non-uniform i n f l o w field f r o m the hull causes a variation in radial loading distribution over the propeller blade, thereby causing extra transverse energy losses.

3. A S S E S S M E N T O F E N E R G Y L O S S E S 3.1 Propeller Open Water Analysis

To get an appreciation o f the magnitude o f the energy losses by a propeller, we w i l l consider the propeller open water case here. A extensive study on the energy losses typical f o r a propeller in open water has been published by Olsen [2004].

To find the relation with the energy losses, the open water efficiency t]^ can be written as:

_ AXL, VISCL, ROIL, ' d

It is noted that the sum o f the pressure losses and transverse losses, as identified i n the preceding section, have been replaced by the sum o f loss terms that refer to the origin o f the loss, i.e. viscosity and rotation. The reason is that this classification better links up w i t h existing literature and w i t h the often non-viscous potential fiow codes f o r propeller analysis.

The axial loss term can be approximated by the ideal efficiency:

( 9 )

It should be noted here that the ideal efficiency accounts for axial kinetic energy losses in the non-viscous flow in w h i c h an actuator disk represents the propeller. This actuator disk consists o f a thin disk w i t h a uniform load distribution, causing a sudden pressure j u m p in the flow. This model does consequently not account f o r the finite number o f blades nor does it account f o r the radial loading distribution, which does cause additional losses (see e.g. Olsen [2004]).

A n estimate o f the different propeller losses is presented in Figure 4. This graph shows a comparison between the ideal efficiency and the open water efficiencies o f a

an L N G earner and a tanker.

Figure 4 shows that a major part o f the losses is caused by axial kinetic energy losses AXL, (approx. 16 to 26 % point), the highest thrust loading Cj producing the highest axial losses. The sum o f the rotational and viscous losses ranges f r o m approx. 17-19% point. The ratio o f these losses can be affected by the choice o f propeller rotation rate and propeller pitch. For a given thrust requirement, a higher pitch angle w i l l reduce the rotation speed and w i l l thus reduce the viscous losses at the cost o f higher rotational losses.

0.0000 1.0000 2.0000 3.0000 4.0000 5.0000 Ttirust loading CT

Figure 4 Relative energy loss terms for B-series propeller

P/D=1.0

3.2 Propeller-hull interaction

A classical treatment o f the energy losses in the propeller-hull interacfion problem is given by Wald [1965]. W a l d derives relations f o r the thrust deduction and ideal efficiency o f an actuator disk that is used as a model in which both potential flow as well as viscous wake effects are important. Wald thereby limits his considerations to the effect o f propeller-hull interaction on the axial energy losses AXL^yp.

I t is found that the thrust deduction fraction t tends to decrease when the propeller is located i n a region o f retardation o f the flow due to viscous effects, but tends to increase when the wake retardation is due to potential flow effects. The efficiency is infiuenced i n opposite directions by these conditions.

Wald derives an expression f o r the ideal efficiency including the effect o f thrust deduction and wake, which can be written i n the f o l l o w i n g f o r m :

2

ViMmd = - . (10)

Where subscript 1 refers to the propeller plane when the propeller is absent, corresponding to the nominal viscous component o f the wake.

I n the case o f a ducted propeller, a similar relation is obtained for the ideal efficiency (actuator disk in a free stream) by e.g. Oostei-veld (1970):

l + ^ l + rCj.

Where T-Tp /Tp the ratio between propeller thrust and total thrust.

This ratio deviates f r o m unity in the case o f a propeller duct f o r example, but can also be used to express the effect o f a thrust deduction fraction on efficiency, thereby assuming a uniform f l o w through the disk. I n the case o f an accelerating duct, the thrust ratio r is smaller than unity and an increase in ideal efficiency occurs. This thrust ratio has a unique relation w i t h the mean velocity through the propeller disk (Oosterveld 1970):

^ = ^ ( i

+ V ö ^ )

(13)

When we now know the change in mass f l o w through the propeller disk, due to e.g. a propeller duct or a Pre-Duct, we can i n principle compute the thrust ratio r (equivalent to the change in thrust deduction factor ( l - ^ ) ) , which then gives us the change i n ideal efficiency. This procedure is based on the assumption that the change o f the effective viscous wake fraction W y w i t h propeller suction, is similar to the change in nominal wake (without propeller). This assumption is likely to be limited to the lighter thrust loadings only (say Cp < 1).

The effect o f a duct resulting i n a change in mass flux through the propeller disk, as w e l l as by a change o f the viscous wake fraction, can now be estimated from eq. (10) once detailed data on velocities and pressures are available f r o m CFD. This w i l l be demonstrated in the worked example on Pre-Ducts.

4. WORKING PRINCIPLES O F E S D s

The previous chapter has classified the energy losses that occur in the wake o f a self propelled ship. W i t h this knowledge, we attempt to better understand the working mechanisms o f three different types o f ESDs.

To this end, it is convenient to study the effect o f ESDs on a simple body (e.g. axisymmetrlc) in deeply submerged condition (no free surface) i n a potential flow. A n d from thereon, gradually add complexity, so as to make a step towards a more 3D ship afterbody in a double body flow (still no free surface), to then add viscosity and flnally add the fiee surface. Although it is realized that ESDs might affect the wavemaking drag, this effect is

It is interesting to see what is happening in a potential flow when we mount a fin to the aftbody, e.g. to improve the fiow into the propulsor. Although a non-lifting body would have a zero contribution to the resistance, a finite fin does contribute to the drag through the shed trailing vortex system, causing the total force vector to act w i t h a component in the downstream x-direction because o f the shed vortex system. The strength o f the shed vortices i n a potential flow f o l l o w f r o m application o f the Kutta Condition at the Trailing Edge o f the fin or f o i l . This vortex system occurs i n the downstream outlet plane as tangential velocities that represent transverse momentum, which is at the cost o f the axial momentum f o r equal power input. The resulting velocity deficit in x-direction is associated w i t h the induced drag o f the f o i l . Only i n the case o f an infinite 2 D f o i l or a circular f o i l w i t h equal loading distribution, no vortices w i l l be shed and the f o i l system may deliver an internal load in x-direction. However, it still w i l l not deliver a positive contribution to thrust.

A passive f o i l system ( f i x e d to the hull) might however produce a net thrust to the system, i f this f o i l system is placed in the propeller induced velocity field.

4.1 Pre-Ducts

Before considering Pre-Ducts or upstream mounted ducts, let us first consider the duct o f a ducted propeller. The reason that a nozzle or duct around a propeller is contributing positively to the overall efficiency is because it accelerates the fiow through the propeller plane (pump action), thereby increasing the mass flux through the propeller (eq. (13)) w h i c h gives rise to a higher ideal efficiency (equivalent to lower axial kinetic energy losses eq. (10)). For this purpose, the nozzle is best mounted as close to the propeller as possible, since the actuator induced velocities w i l l disappear gradually w i t h downstream and upstream distance to the actuator.

We thus conclude that the duct f r o m a ducted propeller effectively reduces the axial kinetic energy losses. Furthermore, by increasing the flow rate through the propeller disk, it may also increase the capture o f viscous wake through the propeller disk, thereby increasing the propeller-hull interaction contribution to the efficiency, according to W a l d [1965]. The net contribution to efficiency o f this viscous wake effect depends on the fraction o f the viscous wake through the propeller disk. These two efficiency enhancing mechanisms (increase o f flow rate and viscous wake fraction) can also be assigned to Pre-Ducts. I t should be noted here however, that due to the reduction o f actuator induced velocities w i t h distance f r o m the actuator, the upstream duct must necessarily be less efficient i n a potential flow than a duct mounted i n the propeller (actuator) plane.

resulting i n lower axial losses by the propeller. This is also the a main principle o f a ducted propeller. 2. Use o f propeller hull interaction to reduce axial

losses (through increase o f the viscous wake fraction w v ) or decrease o f thrust deduction t.

3. To increase uniformity o f the far wake by reducing effects o f non-uniformity through:

a. Changing the radial thrust distribution o f the propeller. I f more thrust is generated at lower radii, this could result in a reduction o f the torque and w i t h that o f the efficiency. I t must be noted that the radial thrust distribution can also be altered by changing the radial pitch distribution o f a propeller. I f this gives a positive effect, the radial distribution o f the propeller is probably not optimal.

b. Changing the circumferential velocity distribution into the propeller. I f the propeller i n f l o w is more u n i f o n n , it is expected that the thmst distribution w i l l also be more uniform. This could lead to a more uniform slipstream and with that less kinetic losses in the slipstream. c. Non-uniformity effects o f propeller thrust

distribution on thrust deduction. I f the propeller loading is higher close to the hull, this is likely to increase thrust deduction.

d. The duct can re-direct the f l o w i n axial direction to the propeller. This w i l l have a positive effect on a more evenly distributed blade loading i n circumferential direction, which w i l l give the same benefits as mentioned under the three previous effects

4. I n case o f extensive fiow separation on the aft ship, the Pre-Duct could lead to a decrease o f the extent o f the separated flow over the aft ship. This w o u l d result in a reduction o f the viscous pressure resistance, which w o u l d appear in a reduced thrust deduction. The case o f a Pre-Duct on an axisymmetrlc body is worked out i n the f o l l o w i n g section.

4.2 Pre- and post swirl stators

As discussed in the introduction to this section, passive fin systems can never improve the overall efficiency o f the propulsor-ship system by creating thrust. I t possibly could alleviate the resistance, particularly by interfering favourably with the Free Surface. When looking at the energy losses that are created by the propulsor-hull system, we can however distinguish rotational losses that could be recovered, either by a secondary rotating propeller like system (e.g. in the case o f contra rotating propellers, a Grim's vane wheel or a PBCF), or by static stators either mounted upstream or downstream o f the propeller. These devices are then to decrease the rotational kinetic energy losses at the benefit o f increasing

way that the propeller blades are more heavily and uniformly loaded, resulting i n a reduction o f rotation rate o f the propeller at equal thrust. This in t u m , also reduces the viscous losses incurred by the propeller.

4.3 Rudder bulbs

The effect o f a rudder bulb has been studied in detail by doing a large number o f CFD computations, using different computational models to represent the propeller. Amongst the different models, actuator disks w i t h and without swirl production were used, as w e l l as a R A N S -B E M coupling, where the propeller was modeled through a B E M method.The results o f this study indicate that the beneflt o f a rudder bulb can both be attributed to a smaller thrust requirement, as well as to a relatively lower torque demand.

4.4 Combined systems

I f Energy Saving structures are going to be fitted, it makes sense to investigate whether or not muhiple energy loss reductions can be achieved. A n d although there is often a coupling between the four different types o f energy losses (axial kinetic, transverse (often largely rotational) kinetic, viscous and non-uniformity losses, it is to be investigated whether this coupling is sufficiently weak to successfully abate losses simultaneously.

Successful examples o f combined systems are Contra Rotating propellers (primarily rotational losses, but also viscous losses and axial losses), M e w i s duct (axial, rotational and non-uniformity losses), as w e l l the Grim's Vane Wheel (axial, rotational).

5. NUMERICAL ANALYSIS OF A P R E - D U C T 5.1 Set-up of systematic Pre-Duct series

To v e r i f y , or better falsify, the hypotheses on the working mechanisms o f a Pre-Duct, this secfions summarizes some o f the resuhs o f C F D computations w i t h the M A R I N R A N S solver REFRESCO on the powering performance o f a systematically varied Pre-Duct series, mounted on an axisymmetrlc body. Pre-Ducts are chosen here because their energy saving mechanisms are not free o f debate.

The rationale behind the choice f o r an axisymmtric body has a number o f reasons. First, it yields an axisymmetrlc flow into the propeller, avoiding the effects o f any circumferential velocity gradients. Secondly, it is computationally efficient. A n d thirdly, i t allows f o r the energy considerations as used by W a l d (1965).

A systematic series o f Pre-Ducts is designed, keeping both the profile geometry and the leading edge position the same. The loading on the Pre-Duct is increased by moving the trailing edge outward ( f r o m Duct D l ) , w i t h one addifional duct D7, having the lightest loading.

The computational domain consists o f two large cylinders, joined together by means o f a non-conformal

Figure 5 Systematic variation scries A Table 1 # Element s # Surfac e Element s ₊ Averag e y' ^ G R I D l Torpedo 12.3M - - -Axi-symmetric 0.8M 1.0 0.62 body 0.8M 1.0 0.62 Pre-Duct - O . I M 2.2 0.7 G R I D 2 Propeller Propeller Hub 25.6M I . I M O . I M 2.1 0.3 0.3 0.1 The iterative convergence o f the velocities are below 10"^ i n the L „ norm (max residual in whole computational domain) and w e l l below 10"^ i n the L2 norm (RMS value o f all residuals). For the pressure and turbulence quantities the iterative convergence is often to be seen below lO"'' or lower in the L „ norm. Further convergence is possible, but would not lead to other resulting forces or conclusions.

5.2 Results of Pre-Duct computations

The performance o f the different Pre-Duct geometries is presented in Figure 6, where the thrust and torque o f the propeller w i t h the various Pre-Ducts f r o m Figure 5 are plotted relative to the performance o f the open propeller. Most o f the computations were conducted f o r a thi-ust loading Cj.„ = 1 . 6 0 . The performance f o r two ducts has been computed for a higher propeller thrust loading o f

Cj.^ = 2 . 4 2 , which is an increase in thrust loading o f some 50%. The thus obtained thrust loadings are considered representative f o r a large number o f f u l l block ships (see e.g. Meuwis [2011]. The thrust loading Cj.„ used here is based on the advance velocity:

T (14) M _i

/

/

A

Ï

\

REtATIVE DIFFEHENCE IN TORQUE%[1

Figure 6 Performance plot for the systematic variation series A

During these computations, the propeller rotation rate was kept at a constant rpm, and hence the propeller did not operate at the self propulsion point ship, as both the thrust demand and the wake o f the ship is changed slightly for every other Pre-Duct. The required power f o r the new configuration is then obtained f r o m :

: 2 ; r e « (15)

To account f o r this deviation i n rotation rate f r o m the self propulsion point rotation rate, it has been assumed (and numerically verified) that the Thrust-Torque relation in terms o f dT I dQ is unaffected. The required power for any Pre-Duct configuration can thus be assessed f r o m the computed Thrust-Torque relation and a correction in torque f o r the self propulsion point. The correction in torque is thus obtained f r o m the horizontal distance between a calculated point and the IST = line. A further correction is then required f o r the propeller rotation rate, deviating slightly f r o m the rotation rate at this self propulsion point. A first estmiate o f this correction can be made by assuming that both the thrust and the torque coefficient remain constant f o r a small increase or decrease o f the propeller rotation rate. This is strictly speaking not true, but corrections are small and w i l l only amplify the thus obtained power difference with increasing rotation rate correction.

The results f r o m this Power correction procedure are plotted i n Figure 7 as the ratio o f required power P D between the duct configuration and the open propeller.

p C / „ ^ ( l - w . , ) Ap

5 Q. 6 c • . 4 — O

8

I

• I

I

Figure 7 Ratio of required power PD to tlie power required for thc same open propeller for the various duct geometries

One can observe that f o r all Pre-Duct cases, the power required for a given speed is higher than f o r the reference case with open propeller. The best performing duct f o r the lower Cja is approx. 1 % worse in required power than the reference case without duct. The increased propeller loading has a positive effect on the duct effectiveness, but the best duct (D3 H L ) still requires some 1 % increase in power demand relative to the reference open propeller. This finding is somewhat contradictory to many results that were obtained f r o m model tests, where savings up to 7% have been claimed f o r f u l l block ships (e.g. Hollenbach S l V I P ' l l ) . I n many o f these cases any large scale boundary layer f l o w separation could not be detected, suggesting that at least in some cases, the avoidance o f large scale separation (mechanism 4 o f Section 4.1) could not be confirmed.

Let us now evaluate the hypothesized mechanisms 3 f o r energy saving f r o m Section 4.1 w i t h these systematic results. Due to the choice o f an axisymmetrlc hull, the circumferential velocity distribution is not affected. The radial velocity distribution is however significantly affected by the Pre-Duct action, as can be seen in Figure 8, where the radial thrust production o f the propeller is given f o r the different Pre-Duct geometries.

It can be observed that with increasing loading o f the duct, the i n f l o w velocities at the lower propeller radii are increasing and that they are decreasing towards the outer radii, consequently leading to a shift in blade loading towards the blade t i p . I n this case, this axisymmetrlc change i n blade loading can be neutralized by changing the radial pitch and camber distribution o f the blade, so that the optimum blade loading is restored. This remedial action can however not be taken so easily when the velocity field is far f r o m axisymmetrlc, such as i n the case o f a real ship.

Figure 8 Relative radial thrust distribution for the seven different Pre-Ducts

The hypothesized mechanisms 3c and 3d cannot be evaluated because o f the axisymmetrlc f l o w in this case. Let us now take a closer look at the mechanism similar to that o f the propeller duct, where the propeller is unloaded by the duct, such that the axial losses o f the propeller are diminished. I n this case, the mean velocity through the propeller disc should be increased. The effect o f this pumping power by the Pre-Duct (or thrust unloading) can be assessed by the relations (12) and (13) when the thrust loading o f the parent propeller is known and the mass f l u x through the propeller disk is evaluated f r o m the C F D results. I t can be observed f r o m the change i n massflow in Figure 9 that the mean velocity through the propeller disk increases up to some 4 % (up to 6% after correction f o r the self propulsion point).

] r'

A

/f

/*

é }

_Y

@ @ ®

—#—10'Viscous wake friction

Figure 9 The relative changes in mass flow, forces and viscous wake

The consequence o f this increased velocity on the propeller unloading can be read from Figure 10 and appears to give a propeller unloading o f some 5% f o r a

Cj^ = 1.60 . The effect on the ideal efficiency in behind condition follows f r o m eq. (12) and appears to result in an increase o f less than 1% f o r the duct w i t h the highest flow

I "

c t=1.6C \ ^ \Ct=2.42

I 1 2 1 4 1 6 1 s :

u_P/UO

Figure 10 Relation between mean velocity through propeller disk up/Uo and propeller thrust ratio T

The effect o f duct geometiy on viscous wake is also shown i n Figure 9. The ducts D1-D3 show an almost constant viscous wake fraction o f 0.29-0.30, where no separation on the ducts has occurred. For the highest loaded ducts, separation occurs on the duct itself, causing the increasing viscous wake fraction. This effect cannot give a net positive contribution to the efficiency.

The above results lead to the conclusion that the Pre-Ducts do not contribute in a similar way to a higher efficiency as ducts around a propeller (see e.g. Oosterveld (1970)). A l s o their effect on improving the propeller hull interaction through a possible increase i n viscous wake did not lead to significant gains f o r a representative thrust loading o f Cj-^ = 1 . 6 0 , based on ship speed.

This leads to the conclusion that i f Pre-Ducts are successful, it must apparently be caused to a large extent by an improved propeller efficiency due to a change in the 3 D velocity field. This remaining hypothesis can not be evaluated on this axi-symmetric body and is yet to be verified.

On the modeling o f the propeller for this type o f computations, it is concluded that the effect o f a non-u n i f o r m i n f l o w and/or inclined i n f l o w is to be accnon-urately represented by the method. This leaves some doubt w i t h the typical propeller modeling i n B E M codes, as they might not sufficiently accurately model this effect. R A N S codes should capture this effect in theory.

6. CONCLUSIONS AND RECOMMENDATIONS

The paper starts with a review o f energy losses produced by the propeller-hull system, as any Energy Saving Device needs to improve on at least one o f these losses. The losses produced by a representative B-series propeller and a method to assess the effect o f hull-propeller interaction are then discussed. Perhaps one o f the less w e l l understood Energy Saving Devices is the Pre- or Upstream Duct.

power reductions up to 7% have been claimed i n literature for f l i l l block ships. These conclusions were typically based on model tests. It is concluded that i f such benefits would occur, they must be caused by a favourable conditioning o f the flow by the Pre-Duct, so that the propeller can operate at a higher efficiency. The Pre-Ducts used i n this study were not able to reduce the overall power demand.

It is recommended therefore, that f o r a reliable assessment o f the consequences in power demand, an accurate propeller modeling is necessary. This leaves doubts w i t h the adequacy o f B E M methods for such modeling, as the generally used wake modeling in these methods is not likely to accurately account f o r an inclined flow or strong velocity gradients. Full R A N S methods are expected to offer this resolution in modeling.

7. A C K N O W L E D G E M E N T S

The author is much indebted to Bart Schuiling, w h o performed the computations and performed the post processing with endless energy. Also the various discussions w i t h m y colleagues A r j a n Lampe, Jan Holtrop, Jie Dang, Martin Hoekstra and many other colleagues are gratefully acknowledged.

The 7FP E U project GRIP is acknowledged for f u n d i n g much o f the research work that contributed to this paper.

R E F E R E N C E S

Dyne, G. (1995). 'The principles o f propulsion optimization'. Trans. R I N A 137. London, United Kingdom.

Hollenbach, U . and Reinholz, O. (2011). 'Hydrodynamic trends i n optimizing propulsion'. Second International Symposium on Marine Propulsors SMP' 11. Hamburg, Germany.

Mewis, F. and Guiard, T. (2011). ' M e w i s Duct - New Developments, Solutions and Conclusions'. Second International Symposium on Marine Propulsors S M P ' 1 1 . Hamburg, Germany.

Olsen, A . S . (2004). 'Energy coefficients f o r a propeller series'. Ocean Engineering 31. 2004. pp401-4I6. Oosterveld, M . W . C . (1970). 'Wake adapted ducted

propellers'. PhD thesis. D e l f t University. Delft, The Netherlands.

Scherer, O., Mutnick, I . and Lanni, F. (2001). 'Procedure for conducting a towing tank test o f a waterjet propelled craft using Laser Doppler Velocimetry to determine the momentum and energy flux'. 26"' American Towing Tank Conference. Webb Institute. Wald, Q. (1965). 'Performance o f a propeller i n a wake

and the interaction o f propeller and h u l l ' . Journal o f Ship Research.