global offices of cd-adapco · pressure and frictional loads on body surface, before updating the...
TRANSCRIPT
SUBMARINEManeuvering Simulations
FEATURES
AZIMUT-BENNETISuper Yacht Design
SAFER UAV LANDINGSOn Battle Shipswww.cd-adapco.com
OPTIMIZING CANOESFor the Olympicswww.cd-adapco.com
CDaEs Tokyo [email protected]
CD-adapco KoreaSeoul [email protected]
CD-adapco IndiaBangalore [email protected]
CD-adapco SEAsia Singapore [email protected]
HeadquartersCD-adapco • New York office60 Broadhollow RoadMelville, NY 11747, USATel.: (+1) 631 549 [email protected]
Atlanta GA Austin TX Cincinnati OH Detroit MI Houston TXLebanon NH Los Angeles CA Seattle WA State College PA Tulsa OK [email protected]
For S. America - please contactMelville Office
HeadquartersCD-adapco • London office200 Shepherds Bush RoadLondon, W6 7NL, UKTel.: (+44) 20 7471 6200 [email protected] www.cd-adapco.com
FranceParis officeLyon [email protected]
GermanyNürnberg [email protected]
ItalyTurin officeRome [email protected]
Global offices of CD-adapco
New ZealandMatrix Applied [email protected]
South AfricaAerotherm [email protected]
TurkeyA-Ztech [email protected]
CDAJ JapanYokohama • [email protected]
CDAJ ChinaBeijing • [email protected]
Resellers
AustraliaVeta [email protected]
Israel [email protected]
Europe Asia-PacificAmericas
40 42
2622
14 18
34
EDITORIALWe welcome editorial from all users of CD-adapco software or services.To submit an article:Email [email protected] Telephone: +44 (0)20 7471 6200
Editor Stephen Ferguson [email protected] Joel Davison, Dejan Matic, Lucy FarringtonArt Direction Brandon BothaAdvertising Geri Jackman [email protected] Director Mark Adlington
SubscriptionsSpecial Reports and Dynamics Magazine are available on subscription.To subscribe or unsubscribe, please email [email protected] +44 (0)207 471 6200
Media Kit available online at: http://www.cd-adapco.com/press_room
04 06
28
20
12
Printed on ‘9 Lives’ recycled paper using vegetable inks. We’re doing our bit, are you?Reduce your Carbon Footprint today http://www.carbonfootprint.com/
Contents marine SPECIAL REPORT
Introduction
03 Marine Issues
Introduction by Milovan Perić
DFBI
04 Just When You Thought it was Safe to go Back in the Water:
STAR-CCM+ V3.06
06 Simulation of Lifeboat Launching
Under Storm Conditions
Classification Societies
10 Partner News
DNV
Germanischer Lloyd
12 State-of-the-Art Simulation
for the Marine Industry
Defense
14 Submarine
Maneuvering Simulations
16 CFD Modeling of Underwater Explosions
Lockheed Martin
18 Simulation of Wave Slap
on Submarines
20 Safer UAV Landings
on Battle Ships
22 Tackling Fluid Structure Interaction Problems
DCN
Ship Design
24 Numerical Modeling
of the Flow around a Tanker Hull
26 Lloyd’s Register of Shipping:
The value of CFD in Ship Design & Analysis
28 Validation of STAR-CCM+:
Prediction of Tanker Resistance
30 Propelling Marine Applications
ROLLA Propellers
32 Predicting Drag
of Floating Vessels
34 Aker Yards Marine
Case Studies
38 Simulation of Scavenging
& Exhaust Valve Rotation
Yacht Design
40 CFD for America's Cup Challengers
HSVA
42 R&D Centre & the CFD Design of Luxury Yachts
Azimut-Benetti
44 Yacht Vibrations
CFD Analysis Helps Solve Problems
46 Volvo Ocean Racing:
Using CFD for Optimal Boat Design
Facts
Olympics
48 Breaking Waves at the Olympics
Fillipi Boats
50 CTO & Plastex Paddle to Olympic Glory
Using CFD Simulation
Offshore
52 Hurricane-Resistant
Offshore Platform Design
54 Offshore Mooring
Delmar
55 Improving Traffic Safety on Waterways
University of Hannover
4644 48
..::INTRODUCTION Marine Issues
MARINE REPORT
The CFD-methods based on RANSE (Reynolds-averaged Navier-Stokesequations) have not been as readily accepted by marine industry as in otherindustries, for several reasons:• Computational methods based on potential theory are much faster and givein many cases acceptable results (e.g. for traditional designs where a lot ofpast experience and experimental data exists);
• Complex geometry, huge dimensions and interaction between fluid andstructure make the computation based on RANSE complicated and theresults were - until recently - often not worth the effort.
However, in recent years, substantial improvements in CFD-technology havemade its application in the maritime sector an indispensable supplement totraditional experimental and computational design and optimization methods.This is especially true for new, innovative designs of marine structures, forwhich no past experience exists and extrapolation is not possible. CFD hasalready been used to design and optimize new hull shapes, propulsion andsteering systems, and in particular to study the interaction between marinestructures and their environment (wind, current and waves). The mostimportant features of CFD-software that influenced this development are theability to easily generate computational grids for complex structures based onCAD-data, to compute free-surface deformation including violent wavebreaking and cavitation, and to account for motion of marine structures whencomputing the flow around them.
While simpler methods may be suitable to optimize a single component on itsown, the strength of CFD lies in its ability to account for interactions betweenall components of a system (e.g. hull, propulsion and steering system, and
other appendages on a ship) in full scale and under realistic initial andboundary conditions, including fluid-structure interaction. CD-adapco has beenat the forefront of this development and it is no surprise that its software ismost widely used in marine industry.
This report presents some CFD-applications to solving real world problems byour customers, ranging from classification societies to small consultingcompanies. It gives an overview of where CFD is being used today and thefeeling for where this development will lead us in the future. There is no doubtthat CFD will play an ever increasing role in the optimization of marinestructures to increase the fuel efficiency, improve safety for passengers, crewand payload, and reduce pollution of environment.
Milovan PerićDirector of TechnologyCD-adapco
Computational methods have been used in marine engineering for a long time to designand optimize marine vessels, from yachts to offshore platforms. It is just that thosetraditional computational methods are often not considered as “proper” CFD(Computational Fluid Dynamics) - because they are based on potential flow theory andempirical correlations.
Marine IssuesIntroduction by Milovan Perić
�� EMAIL�[email protected]
CFD has already beenused to design andoptimize new hullshapes, propulsionand steeringsystems, and inparticular to studythe interactionbetween marinestructures and theirenvironment.
“
”
3
marine SPECIAL�REPORT
Dynamic fluid body interaction“The�significant�advantage�of�our�new�DFBI�model�is�that�it�allowsthe� fluid-induced� motion� of� an� object� to� be� predicted� in� sixdegrees� of� freedom,� with� no� additional� effort� from� the� user,
compared�to�performing�a�simple�steady�state�simulation�of�flow�past�thesame� body,” says� Jean–Claude� Ercolanelli,� CD-adapco’s� VP� ProductManagement.
“The�meshing�process�for�a�DFBI�simulation�is�identical�to�that�for�a�simplenon-moving� calculation� and� the� DFBI� model� accomplishes� its� task� withoutinvoking�the�complex�and�time�consuming�re-meshing�schemes�used�by�otherCFD�tools.”
STAR-CCM+�uses�a�fully-implicit�iterative�method�for�coupling�the�computationof�fluid�flow�and�body�motion.�In�each�iteration�it�calculates�the�flow-inducedforces� and� moments� acting� on� the� body� by� automatically� integrating� thepressure� and� frictional� loads� on� body� surface,� before� updating� theacceleration,� velocity� and� body� motion� in� six� degrees� of� freedom.� Severaloptions�for�an�automatic�mesh�adaptation�to�the�changing�body�position�areavailable,� from� translating�and� rotating� the�entire�mesh�with� the�body,�overusing�sliding�interfaces�for�rotational�motion,�to�mesh�morphing.�The�iterationsare�repeated�in�each�time�step�until�convergence,�ensuring�a�robust�solutionprocedure�without�stability�limitations�to�time-step�size.
A complete tool for marine applicationsDFBI� is�useful� for�any�simulation� in�which�an�unrestrained�object�moves� inresponse�to�fluid�excitation.�However,�it�is�particularly�useful�when�combinedwith�the�powerful�free�surface�modeling�capability�in�STAR-CCM+,�to�simulatethe� behavior� of� floating� vessels.� One� can� thus� study� highly� non-linearphenomena�that�cannot�be�analyzed�using�simple�methods�based�on�potentialflow� theory,�such�as�wave�slamming,�wave-piercing,�water-entry,�green�watereffects,�wind�effects�etc.�There�are�no�limitations�to�body�motion,�so�that�sea-keeping�and�maneuvering�simulations�including�capsizing�can�be�performed.
STAR-CCM+� also� includes� an� automatic� wave� generator,� which� is� fullyintegrated�into�the�GUI,�allowing�users�to�generate�both�linear�and�Stokes�5th-order�nonlinear�waves�using�just�a�few�simple�parameters.�The�DFBI�solver�canalso�be�used�to�calculate�the�vessel’s�actual�steady�state�position�in�the�waterfor�a�given�speed,�automatically�calculating�sinkage,�heel�and�trim.
In�June�2008�we�released�STAR-CCM+�V3.04,�followed�by�V3.06�in�October.�Central�to�bothreleases�was�the�introduction�of�a�new�capability�for�“Dynamic�Fluid-Body�Interaction”.Unlike�traditional�CFD�software,�which�is�typically�limited�to�simulating�the�motion�of�a�bodyalong�some�prescribed�path,�the�STAR-CCM+�DFBI�model�enables�engineers�to�easilysimulate�the�fluid-induced�movement�of�a�body,�for�everything�from�a�tanker�listing�underheavy�seas�to�an�aerodynamic�body�tumbling�in�free�fall.
Just When You Thought it was Safe to go Back in the Water:
STAR-CCM+ DFBIStephen Ferguson, CD-adapco.
..::FEATURE ARTICLE DFBI
MARINE�REPORT4
STAR-CCM+ V4.02Due for release early 2009, will include a DFBImaneuvering capability, allowing the motion of avessel that results from a rudder deflection tobe simulated.
Beyond boatsEach� new� release� of� STAR-CCM+� contains� a� raft� of� new� features� andenhancements� that� help� to� increase� productivity� while� reducing� analysisturnaround�times.
In�STAR-CCM+�V3.04�we� improved�the�automatic�hexahedral�mesher,�whichnow�has�the�ability�to�refine�cells�in�specific�directions�allowing�greater�controlof� cell� count� without� compromising� accuracy,� which� is� extremely� useful� foraerodynamics�and�free�surface�flow�simulations.�A�new�surface�offset�tool�wasadded�to�the�Surface�Preparation�GUI�to�ease�and�speed-up�the�handling�ofcomplex� and� poor� geometry.� Indirect� topological� interfaces� for� multi-stageturbines�may�now�be�modeled�where�each�stage�has�a�different�pitch.�Multi-band�thermal�radiation�in�participating�media�has�also�been�implemented�inSTAR-CCM+� for� the� first� time,�allowing� the� radiation�properties� to� vary�withwavelength.
STAR-CCM+�V3.06�introduces�a�new�capability�that�will�allow�users�to�performstructural�analysis�calculations�from�within�the�STAR-CCM+�GUI,�the�first�timethat�an�integrated�process�for�flow,�thermal�and�stress�simulation�has�beenavailable� in� a� single� general-purpose� commercial� finite-volume� code.� Theintroduction� of� a� mesh� morphing� capability� simplifies� moving� meshsimulations,� fluid-structure-interaction� calculations�and�optimization�studies.The�process�from�CAD�to�CAE�solution�is�further�improved�with�the�introductionof�a�new�selection�of�native�CAD�Importers�that�allow�you�to�robustly�importnative�CAD�without�translations,�and�through�the�new�surface�mesh�projectioncapability�that�maintains�the�integrity�of�the�mesh�by�projecting�the�grid�pointsback�to�the�imported�analytical�geometry.�Finally�a�new�transition�model�helpsto�predict�the�transition�from�laminar�to�turbulent�boundary�layer.
New�versions�of�STAR-CCM+�are�released�every�four�months,�for�the�mostup�to�date�information�please�visit�www.cd-adapco.com���
..::FEATURE ARTICLE DFBI
MARINE�REPORT 5
STAR-CCM V4.02
AVAILABLE
FEBRUARY 2009
!
www.cd-adapco
.com
The meshing process for aDFBI simulation isidentical to that for asimple non-movingcalculation and the DFBImodel accomplishes itstask without invoking thecomplex and timeconsuming re-meshingschemes used by otherCFD tools.
“
”
� Major STAR-CCM+ releases in the last 18 months. Literature is available upon request: [email protected]
� MORE�INFORMATION� [email protected]
..::FEATURE ARTICLE DFBI
MARINE�REPORT6
Lifeboats�are�an�important�component�of�safety�measures�for�the�passengers�and�crew�offloating�vessels�and�offshore�platforms.�They�need�to�be�designed�so�that�people�onboard�can�be�evacuated�safely.��This�requires�that:�the�lifeboat�is�not�damaged�duringwater�entry;�the�lifeboat�moves�sufficiently�far�away�from�the�launching�point�before�itsown�propulsion�system�is�started;�and�the�accelerations�experienced�by�occupants�do�notexceed�a�certain�level�over�a�certain�period�of�time.
Simulation ofLifeboat LaunchingUnder StormConditions Hans Jørgen Mørch, CFD Marin and Agder University, Norway.Sven Enger, Milovan Perić and Eberhard Schreck, CD-adapco.
In� the� past,� lifeboat� designs� have� been� tested� solely� byexperimental� means.� In� these� experiments,� pressure� ismeasured� at� a� certain� number� of� locations� on� the� hull,� andmotion�and�acceleration� is� recorded.�However,�due� to�a� large
number�of�different�lifeboat�sizes�and�a�large�variety�of�conditions�underwhich�they�may�have�to�be�used,�the�number�of�necessary�tests�becomesunmanageable.�
Experiments�may�also�be�deficient�due�to�the�following�reasons:• In�model�tests�the�scale�factor�should�not�be�less�than�1:10�in�order�toobtain� a� sufficiently� large�model� to� fit� sensors� and� to� avoid� adversescale� effects.� The� maximum� wave� height� in� model� test� facilities� isnormally�less�than�1�meter.�In�order�to�obtain�results�corresponding�todesign�waves,�one�has�to�rely�on�extrapolation�of�experimental�resultsat�lower�wave�heights�with�the�same�wave�steepness.�This�introducesadditional�uncertainty�in�the�results.
• Full� scale� experiments� can� only� be� performed� at� good� weatherconditions� with� little� wind� and� small� wave� heights,� whereas� designconditions�can�have�wave�heights�of�15m�and�very�strong�wind.� Themismatch�between�testing�conditions�and�reality�makes�the�evaluationof�test�data�difficult.
• With�respect�to�drop�height,�model�test�facilities�have�limitations�givenby�the�height�of�the�ceiling�inside�the�laboratory.�In�full�scale,�launchinga� lifeboat� from� a� ramp� with� heights� greater� than� those� of� existinginstallations�is�both�impracticable�and�expensive.
• Model�and�full�scale�tests�are�normally�documented�by�video�recording.Minimum�instrumentation�is�accelerometers�fore�and�aft�and�pressuremeasurements�are�limited�to�a�few�locations.�Relating�time�history�ofaccelerations�and�pressures� to� the� trajectory�of� the� lifeboat� requiressynchronization�with�a�high-speed�camera.
• Experiments�are�suitable�to�determine�the�actual�loads�on�structure�andpeople� inside� lifeboat,� but� they� do� not� provide� enough� informationnecessary� to� improve� the�design.�For� this�purpose,� it� is� important� toknow�the�pressure�and�velocity�distribution�around�the�hull�during�waterentry�and�the�subsequent�diving�and�re-surfacing�of�the�lifeboat.
It�was�after�the�sinking�of�the�RMS�Titanic�on�April�15,�1912,�that�a�movement�began�torequire�a�sufficient�number�of�lifeboats�on�passenger�ships�for�all�people�on�board.�TheTitanic,�with�a�gross�tonnage�of�46,000�tons�and�carrying�20�lifeboats,�met�and�exceededthe� regulations� laid� down� by� the� Board� of� Trade,� which� required� a� ship� of� her� size�(i.e.�over�10,000�tons)�to�carry�boats�capable�of�carrying�a�total�of�1,060�people.�TheTitanic's� boats� had� a� capacity� of� 1,178� people� on� a� ship� capable� of� carrying� 3,330people.�The�need�for�so�many�more�lifeboats�on�the�decks�of�passenger�ships�after�1912led�to�the�use�of�most�of�the�deck�space�available�even�on�the�large�ships,�creating�theproblem�of�restricted�passageways.�This�was�resolved�by�the�introduction�of�collapsiblelifeboats,�a�number�of�which�(Berthon�Boats)�had�been�carried�on�the�Titanic.
� FACTS�
Titanic�Survivor.Lifeboat�"Collapsible�B,"�washed�off�Titanic�upside�down.
..::FEATURE ARTICLE DFBI
MARINE�REPORT 7
� Fig:01The original hull (top), modified aft body (middle) and modified aft body and bow (bottom).
� Fig:02Time record of vertical acceleration expressed in multiples of gravity acceleration forthe experiment and simulations at the front (left) and rear (right) of the lifeboat.
(contd.)
� Fig:03 Comparison of normalized maximum CAR-values for full load plus 10t of ballast for thethree hull shapes obtained in experiment (blue) and in simulation (red) at the front (left)and rear (right) of the lifeboat.
..::FEATURE ARTICLE DFBI
MARINE�REPORT
Recent� advances� in� computational� fluid� dynamics� (CFD)� have� made� itpossible�to�perform�simulations�of�lifeboat�launching�at�full�scale�and�underrealistic� initial� and� boundary� conditions.� Simulations� also� allow� aninvestigation�of�the�effects�of�changes�in�hull�shape�without�having�to�makea�physical�model.�This�makes�it�possible�to�investigate�a�larger�range�of�hullshapes� at� a� variety� of� launch� conditions� and� finding� a� design� that� isacceptable�for�the�expected�use.
The�method�described�here�uses�state-of-the-art�CFD�software�coupled�to�aCAD�design�tool�and�a�solver�for�6�degrees-of-freedom�motion�of�rigid�bodiesto� efficiently� evaluate� the� effects� of� both� design� changes� and� launchingconditions�on�performance�of�lifeboats.�The�launching�conditions�that�can�beanalyzed� include� wind,� current,� water� depth� and� wave� profiles� in� anycombination.�The�flow�and�flow-induced�motion�of�lifeboats�(which�are�hereconsidered�to�be�rigid)�is�computed�in�a�coupled�manner.�Since�the�numberof�necessary�simulations�in�an�optimization�study�is�large,�it�is�important�thatthe�method�is�computationally�efficient.�This�requires�local�mesh�refinementand�an�efficient�handling�of�mesh�adaptation�to�the�position�of�lifeboat�as�itmoves.
After� trying� several� alternative� approaches,� the� authors� settled� upon� amethod�that�employs�overlapping�grids,�in�which�a�background�grid�is�adaptedto�the�free�surface�and�outside�boundaries�(such�as�the�sea�bed,�oil�platformor�marine�vessel),�while�the�overlapping�grid�is�attached�to�the�lifeboat�andmoves�with�it�without�deformation.�This�overlapping�grid�method�is�applicableto�unlimited�motions�(including�overturning)�and�the�boundary�conditions�(likewave�generation)�are�easier�to�implement�than�in�other�approaches.
A� lifeboat� launched�from�a�ramp�undergoes�three�stages�before� it�hits�thewater.� First,� it� accelerates� forward� as� it� slides� down� the� ramp� under� theinfluence�of�gravitational�and�frictional�forces.�As�the�center�of�gravity�passesthe�edge�of�the�ramp,�the�lifeboat�will�experience�an�angular�acceleration�dueto� the�bow-down�pitch�moment� from� the� reaction� force�at� the�edge�of� theramp,�acting�behind�the�center�of�gravity.�This�pitch�moment�ceases�when�therail� of� the� lifeboat� leaves� the� ramp.� The� lifeboat� will� then� acceleratedownwards�with�nearly�constant�horizontal�and�angular�velocities�before�theimpact�with� the�water.�During� this� third�stage� it� is�appropriate� to�start� thecoupled�simulation�of�fluid�flow�and�flow-induced�motions.�Initial�conditionsfor� simulations� include� position� of� the� center� of� gravity� relative� to� watersurface�and�trim�angle.�Initial�horizontal,�vertical�and�angular�velocities�needalso�to�be�prescribed.
To�gain�confidence�with�this�technique�of�simulation,�a�validation�study�wasperformed� in� collaboration�with�Norsafe� AS� to� examine� the� effects� of� hullshape�on�forward�motion�of�a�lifeboat�and�accelerations�experienced�by�theoccupants.� In� this� study,� three� different� hull� forms� were� considered:� theoriginal�base�form,�modified�aft�body,�and�modified�aft�body�and�bow�section.
The�results�of�experiments�and�simulations�for�all�three�forms�are�presentedin� Figures�02�and�03.� The� results�are�both�qualitatively� and�quantitativelysensible,� reflecting� the�effects�of�design�changes� in� the�same�way�as� theexperiment.
The�usual�quantitative�measure�of�acceleration�is�expressed�as�the�so�calledCAR-value.�For�all�three�forms,�the�accelerations�are�higher�in�the�rear�than�in
� Fig:04Detail of the mesh structure in symmetry plane near hull and free surface, alsoshowing position of the lifeboat 2s after launching.
� Fig:06Distribution of pressure on hull surface 1.8s (upper),1.9s (middle) and 2s (lower) after launching (withwind); also shown is the shape of free surface andthe wetted hull area.
� Fig:05Distribution of pressure (upper) and velocity magnitude (lower) in symmetry planearound lifeboat 2s after launching (with wind).
8
..::FEATURE ARTICLE DFBI
MARINE�REPORT
� MORE�INFORMATION���[email protected] http://www.uia.no/en
the� front� part,� but� the� difference� reduces� as� more� changes� are� carried� out� to� the� original� design.Furthermore,�the�absolute�level�of�the�CAR-values�is�reduced�with�each�modification.�The�hull�with�modifiedaft�body�and�bow�leads�to�a�reduction�of�CAR-values�at�the�rear�by�20%�relative�to�the�original�design.�Thereduction�of�CAR-values�in�the�front�is�significantly�lower�-�of�the�order�of�5%.
Calculations�performed�so�far�show�that�a�simulation�of�lifeboat�motion�with�three�degrees�of�freedom�(twolinear�and�one�rotation�motion)�in�full�size,�from�the�time�it�is�launched�until�it�re-surfaces�from�water�andmoves�about�40�m�away�from�the�launching�point,�can�be�performed�on�a�single�processor�(i.e.�a�singlecore�of�a�multicore�processor)�in�less�than�a�day�using�a�grid�made�of�about�300,000�control�volumes.�
Validations� against� tests� performed� in� calm� water� indicate� that� this� mesh� resolution� is� sufficient� todistinguish�the�effects�of�design�changes.�Final�validations�for�optimal�design�may�require�finer�meshesand�the�use�of�more�processors�to�keep�the�run�time�of�the�same�order�(1�to�2�days).�This�makes�the�useof�CFD�for�the�investigation�of�loads�onto�lifeboat�structure�and�accelerations�to�which�people�at�differentseats�are�exposed�during�water�entry,�an�ideal�supplement�to�experimental�investigations,�which�can�belimited�to�final�validations�of�an�optimized�design.�With�a�cluster�of�hundred�processor�units,�it�is�possibleto�perform�thousands�of�simulations�and�evaluate�the�results�within�just�a�few�weeks.��
� The University of Agder is the fifthlargest higher education institution inNorway. The university was establishedon September 1, 2007 when AgderUniversity College officially became theUniversity of Agder. Agder refers to theregion, consisting of the two counties ofVest-Agder and Aust-Agder.
9
Fig:07 �Predicted angular position (top),angular velocity (middle) and angularacceleration (lower) of the lifeboatduring the first 6s after launching,for the no wind and wind condition.
The method described hereuses state-of-the-art CFDsoftware coupled to a CADdesign tool and a solver for6 degrees-of-freedommotion of rigid bodies toefficiently evaluate theeffects of both designchanges and launchingconditions on performanceof lifeboats.
“
”
Drawing� on� more� than� 140� years� of� experience,DNV�deliver�services�that�predict�and�assess�themotions,� loads� and� other� dynamic� responses� ofships�and�offshore�structures�in�waves�or�related
fluid�flow�problems.�
Says�Dr�Bo�Cerup-Simonsen,�head�of�DNV�Maritime�TechnicalConsulting�and�DNV�Fellow� in�computational�mechanics:�“Theshipping�and�energy�industries�are�faced�with�a�number�of�newchallenges,�driving�the�need�for�novel�designs�and�technologies.For�shipping� this�concerns,�among�others,�container�and�LNGvessels� as� well� as� more� specialized� ships.� The� lack� ofexperience�for�a�novel�design�demands�accurate�prediction�ofloads,� motions,� resistance� and� propulsion� efficiency.� Forexample,�slamming�pressures�on� the�bow�and�aft�part�of� theship�and�sloshing�effects�in�LNG�tanks�are�some�of�the�areasthat� are� critical� and� challenging.� This� new� CFD� solutioncombined� with� our� world-class� competence� will� extend� ourcapabilities�to�better�meet�this�demand.”�
CD-adapco�has�a� long�history�of�successful�partnerships�withleading� companies� in� both� the� maritime� and� petrochemicalindustries.� The� company� has� invested� heavily� in� providingcapabilities�within�its�software�that�meet�the�most�challengingproblems�within�these�industries.�
“We�are�delighted�that�DNV,�with�its�global�presence�and�as�oneof� the� ‘big-three’� classification� societies,� has� justified� thisinvestment� by� choosing� our� software.�By�working� closely�withDNV,� we� intend� to� further� refine� our� technology� to� meet� theindustry� demands,”� says� Dr� Dennis� Nagy,� CD-adapco’s� VicePresident�of�Marketing�and�Business�Development�and�Directorfor�the�Energy�Sector.
Simulating sloshing behavior DNV’s� engineers� will� use� CD-adapco’s� STAR-CCM+� to� tacklesloshing�problems.�Resonance�conditions�can�lead�to�sloshing�impacts� at� corners� and� knuckles� inside� tanks� in� excess� ofdesign� loads,� with� a� potential� risk� of� structural� damage.� Thesoftware�will�allow�DNV�to�simulate�sloshing�behavior�driven�bya� wide� range� of� sea-conditions,� allowing� engineers� both� tovisualize�the�liquid�motion�and�to�identify�critical�events�that�maycause�high�sloshing-induced� impact� forces.�DNV�will�also�useSTAR-CCM+�for�the�analysis�of�vortex-induced�vibration�and�forsimulations� of� general� free-surface� flows� and� six-degrees-of-freedom�motion�of�floating�bodies.��
�� MORE�INFORMATION� [email protected] http://www.dnv.com/
..::PARTNER NEWS Classification�Societies
MARINE�REPORT
DNV�(Det�Norske�Veritas)�has�selected�a�software�solution�for�computational�fluid�dynamics(CFD)�from�CD-adapco�to�extend�its�leading-edge�technical�consulting�service�line.�The�newsoftware�will�be�a�valuable�addition�to�the�DNV�toolbox�to�provide�accurate�and�reliableestimation�of�slamming�and�sloshing�loads,�which�are�critical�for�the�design�and�operationof�both�ships�and�offshore�structures.�
� DNV is a global provider ofservices for managing risk.Established in 1864, DNV isan independent foundationwith the objective ofsafeguarding life, propertyand the environment. DNV comprises 300 officesin 100 countries, with over7000 employees.
DNVNew consulting services from DNV with fluiddynamics software from CD-adapco
10
We are delighted that DNV, with its global presence and as one of the ‘big-three’ classification societies,has justified thisinvestment by choosingour software.
“”
..::PARTNER NEWS Classification�Societies
MARINE�REPORT
DNV
�� MORE�INFORMATION� http://www.gl-group.com
Germanischer Lloyd & CD-adapco join forces
Dr�Ould�el�Moctar,�GL’s�Head�of�Department�Fluid�Dynamics,�explains:“Germanischer� Lloyd� and� CD-adapco� have� already� been� workingtogether�for�many�years.�Both�parties�have�considerable�experiencein� the� field� of� CFD� analyses.” Former� joint� publications� by�
CD-adapco’s�Director� of� Technology,� Prof�Dr�Milovan� Perić,� and�Dr� el�Moctarinclude�papers�on�“Wave�Loads”�or�the�“Simulation�of�Sloshing�in�LNG-Tanks”.
The� importance� of� computer-aided� engineering� in� shipping� cannot� beunderestimated,� the� GL-expert� stresses:� “The� technological� progress� ofsimulation�in�this�sector�is�rapid,�both�for�software�and�hardware.�Simulationtechnology�is�employed�in�a�wide�range�of�marine�applications�and�often�playsa�critical�role�in�the�decision�making�process.” Raimund�Schipp,�CD-adapco’sEuropean�Director�of�Sales,�adds:�“Computer�simulation� technology� is� finallyfulfilling� its� enormous� potential.� The� availability� of� accurate� and� reliablepredictions� at� the� earliest� stage� of� the� design� process� leads� to� significantsavings�in�effort�and�money,�compared�to�performing�tests�and�amending�thedesign�later�in�the�process.”
Germanischer�Lloyd�has�always�been� in� the�vanguard�of�CFD�analysis� in� themaritime�field,�today�offering�the�full�scope�of�maritime�engineering�services�-such�as�the�computation�of�slamming,�sloshing,�wave�loads,�cavitation�or�fluid-structure� interaction.� In� all� cases,� advanced� simulation� software� is� needed.“However,�the�true�value�offered�by�advanced�engineering�providers�lies�in�thesymbiosis�of�software�or�hardware�and�highly�skilled�staff,”�Ould�el�Moctar�says.“Modeling�therefore�also�requires�considerable�collective�experience.”�This�iswhy�the�cooperation�is�so�very�interesting,�Raimund�Schipp�points�out:�“We�willcombine� both� the� theoretical� as� well� as� practical� experience� of� our� twocompanies�to�come�up�with�beneficial�solutions.”��
CD-adapco,�a�global�enterprise�offering�computationally-based�engineering�solutions,�andthe�classification�society�Germanischer�Lloyd�(GL)�have�joined�forces.�The�two�companiessigned�a�cooperation�agreement�whereby�they�will�closely�work�together�in�the�sector�ofComputational�Fluid�Dynamics�(CFD)�for�marine�applications.�The�future�collaboration�will�comprise�of�mutual�publications,�development,�validation�aswell�as�workshops.
Container ShipsContainerized�shipping�is�a�rational�way�of�transporting�most�manufactured�and�semi-
manufactured� goods.� This� rational� way� of� handling� the� goods� is� one� of� the
fundamental� reasons� for� the� globalisation� of� production.� Containerisation� has
therefore� led� to� an� increased� demand� for� transportation� and,� thus,� for� further
containerisation.�A�traditional�conventional�vessel�required�between�8�to�10�days�to
load�or�unload�10,000�tons�of�general�cargo.�A�containership�can�handle�the�same
volume�in�2�days�within�Europe�and�in�3�or�4�days�on�other�continents.
The�use�of�containers�started�during�the�Second�World�War,�but�it�wasn’t�until�1956
that� the� first� container� service� was� opened� between� the� USA� and� Puerto� Rico.
Malcom�McLean,�a�trucking�entrepreneur�from�North�Carolina,�acquired�a�steamship
company� in� 1955� with� the� idea� of� using� its� ships� to� transport� cargo-laden� truck
trailers.� McLean's� experiment� resulted� in� the� world's� first� container� ship,� the�
‘Ideal-X’,�a�converted�oil�tanker�whose�deck�had�been�strengthened�to�accommodate
containers.�It�made�its�inaugural�voyage�from�New�Jersey�to�Texas�on�26�April�1956
with�58�trailers�(containers)�on�its�deck.�The�pioneering�container�ships�could�carry
only�59�containers�having�a�length�of�35�feet�and�stacked�two-high�on�deck.�
Emma�Mærsk�-�the�longest�container�ship�ever�built.
� FACTS�
11
� Fig:03Analysis of cavitation on propeller and rudder
..::FEATURE ARTICLE Classification�Societies
MARINE�REPORT
State-of-the-ArtSimulation for the MarineIndustry
� Fig:01Simulation of free surface flow and motionof the Earthrace boat.
“The Earthrace Boat” (www.earthrace.net)is designed to pierce through wavesinstead of riding over them.
Ould el Moctar, Germanischer Lloyd.
Earthrace boat smashes world recordEarthrace�started�her�attempt�to�set�a�newworld�speed�record�for�a�powerboat�tocircle�the�globe�from�Vulkan�Shipyard,Sagunto,�Spain�on�Sunday�27�April�2008,13.35�GMT/14.35�CET.��She�crossed�thefinish�line�on�27�June�2008�at�12.24.00GMT/14.24.00�CET.
The�previous�record�was�74�days�20�hours58�minutes�30�seconds.�This�record�wasset�by�UK�boat�‘Cable�&�WirelessAdventurer’�in�1998.�Earthrace�smashedthe�record�by�almost�14�days,�completingalmost�24,000�nautical�miles�in�just�60days�23�hours�and�49�minutes!��
� LATEST�NEWS
� Germanischer Lloyd is themarket leader in the classificationof containerships. Every secondcontainership is built according toGL Rules. Devoted to promotinginnovation, sustainability andenvironmental protection,Germanischer Lloyd also coversall other ship types such as tanker,bulker, multi purpose vessels,high speed ferries and cruiseships. The current new buildingorder-book contains more than1,400 vessels with 28 Mio GT.
� Fig:04Exaggerated deformation of ship structure at an instant of time during its motion in waves.
� Fig:02In the examples shown here, courtesy ofGermanischer Lloyd, a commercial super-lineris subjected to waves of 7.3 meters height,while traveling at 26 knots.
12
Bow flare slammingThe�extreme�pitch-and-heave�motion�of�a�ship�operating� in� roughseas� causes� Bow� flare� slamming.� Modern� ultra-large� containerships,�which�typically�rely�on�the�additional�cargo�capacity�of�a�large
bow�flare,�are�often�exposed�to�a�high�risk�of�slamming�due�to�their�relativelyhigh�speed�and�operational�requirements�that�they�be�driven�through�adverseweather�conditions.
CD-adapco’s�simulation�technology�has�been�used�extensively�throughout�theindustry� to�understand� the�mechanisms�behind�bow� flare�slamming�and� tohelp� mitigate� the� risk� of� damage.� A� significant� advantage� of� numericalsimulation�is�that�tests�can�be�carried�out�at�full�scale�and�that�pressure�loadscan�be�predicted�at�every�point�on�the�hull.
In� the� example� shown� in� Figure� 2,� courtesy� of� Germanischer� Lloyd,� acommercial� super-liner� is� subjected� to� waves� of� 7.3� meters� height,� whiletraveling�at�26�knots.�A�graphical�comparison�of�the�pressure�experienced�atvarious� points� on� the� hull� surface� shows� outstanding� agreement� betweennumerical� and� experimental� studies.� The� overall� quality� of� the� simulationresults�gave�Germanischer�Lloyd’s�engineers�confidence�in�the�methodology,enabling�them�to�assess�the�safety�of�a�wide�range�of�ship�designs,�operatingunder�the�most�adverse�conditions.
Pitch and roll simulationsThe� motion� of� a� vessel� under� the� influence� of� a� rough� sea� is� a� complexcombination�of�translation,�pitching,�rolling�and�yawing.�CD-adapco’s�simulationtechnology�allows�the�motion�of�a�vessel�to�be�predicted�in�all�six-degrees�offreedom,�using�a�fully�coupled�simulation�technique�that�accounts�for�both�theinfluence�of�the�flow�on�the�boat,�and�the�influence�of�the�boat�motion�on�theflow.
Germanischer�Lloyd�has�used�this�technology�to�good�effect�in�their�analysisof�the�unconventional�Earthrace�vessel,�which�has�recently�set�the�new�worldrecord�for�a�powered�boat�circling�around�the�world.�In�rough�conditions,�theEarthrace� boat� is� designed� to� pierce� through� waves� instead� of� riding� overthem.�Figure�1�illustrates�how�the�bow�of�the�boat�enters�a�simulated�wave.�
Simulation�results�such�as�these�allowed�the�boat�designers�to�understand�itsperformance�in�very�rough�conditions,�from�the�comfort�of�their�design-studio,before�even�the�first�prototype�had�been�built.
Propeller and rudder cavitationCavitation� is�a�significant�cause�of�damage� to�ship�propellers�and� rudders,often� causing� surface� pitting� and� fatigue-inducing� vibration.� CD-adapco’ssimulation� technology� accurately� predicts� the� onset� of� cavitation� and� theunsteady� phenomena� associated� with� the� build-up� and� break-up� of� largecavitation�regions.�An�example�of�a�result�of�such�a�simulation� is�shown� inFigure�3.
Detailed�analysis�of�both�steady�and�unsteady�cavitation�has�been�performedwith� considerable� success� across� the� industry.� For� a� given� design,� thistechnology�allows�designers�to�identify�under�which�operating�conditions�theworst�cavitation�problems�are�likely�to�occur,�or�alternatively�which�design�isleast�prone�to�cavitation�under�a�given�operating�condition.�
Structural deformations from wave impactSlamming� loads�can�cause�deformation�of� local�structural�components�andinduce�high�stresses.�The�accurate�assessment�of�such�loads�is�essential�forthe�design�of�a�ship’s�structure.
Classification� society� rules� contain� formulas� for� slamming� loads.�Generally,these� formulas�are�adequate� for� conventional� ships,� as� they�are�based�onoperational� experience.� However,� for� many� modern� ships� it� becomesnecessary�to�resort�to�direct�computations�of�slamming�loads.
As�an�extension�of�the�described�bow�flare�slamming�work,�deformation�of�theship’s� structure� resulting� from� the� impact� of� a� series� of� large� waves� waspredicted.� In� a� coupled� fluid-structure-interaction� simulation,� the� forcescalculated� from� the� simulation� of� the� flow� were� used� (via� an� interfacedeveloped� by� Germanischer� Lloyd)� to� provide� boundary� conditions� for� astructural-analysis�simulation,�which�predicted�the�structural�deformation�(seeFigure�4).��
..::FEATURE ARTICLE Classification�Societies
MARINE�REPORT
CD-adapco�has�a�27-year�history�of�providing�state-of-the-art�flow,�thermal�and�stresssimulation�to�the�marine�industry.�From�large�shipyards�to�suppliers�of�small�components,the�use�of�our�technology�has�become�a�standard�feature�in�marine�design�and�safetyassurance�process.�Using�our�cutting-edge�solver�technology,�our�customers�have�been�ableto�tackle�some�of�the�most�demanding�problems�that�the�industry�has�to�offer.�Recent�successes�include:
�� MORE�INFORMATION� [email protected] http://www.gl-group.com
13
The� physics-based� simulation� of� a� full-scalesubmarine�performing�maneuvers�is�an�expensiveproposition�relative�to�many�CFD�applications.�Thisis�principally�due�to�the�wide�range�of�length�and
timescales� that� must� be� resolved� in� order� to� predictaccurately�the�flow�around�the�submarine�hull.�An�additionalchallenge�involves�representing�the�full�geometric�complexityof�an�appended�submarine�and�propulsion�unit.�The� lengthscales� range� from� the� very� thin� boundary� layer� to� the� fulllength� of� the� submarine.� The� time� scales� range� from� afraction� of� the� propeller� blade� passing� period� to� the� totalduration� of� a� maneuver� -� more� if� several� maneuvers� arecombined� in�a�single�simulation.�These�disparities� in�scalelead� to� very� large� computational� meshes� and� simulationtimes�that,�until�recently,�have�challenged�the�state�of�the�artin�computational�resources.�
The� submarine� in� question� is� propelled� by� a� three-bladedrotating� propeller.� Maneuvers� were� executed� through� theapplication� of� rudder� and� stern� planes,� and� controlled� byvarying�the�position�of�these�control�surfaces�in�response�tothe�submarine�motion�predicted�by�the�simulation.�
Numerical method During�the�course�of�a�maneuver,�the�submarine�changes�itsposition�and�orientation�continuously� in� time� in� response� tothe� pressure� field� generated� by� application� of� the� controlsurfaces.�The�simulation�of�a�maneuver�requires�the�coupledsolution�of�equations�of�motion�of�the�rigid�body�(in�six�degreesof� freedom)� with� unsteady� Reynolds-averaged� Navier-Stokesequations� (URANS).� The� URANS� solver� uses� a� fully-implicititerative� time-integration� scheme.� It� computes� the� flow� fieldaround� the� body� first� and� integrates� the� computed� shear
The�numerical�simulation�of�submarine�maneuvering�is�a�challenging�problem�that�has�only�recently�been�addressed�by�technological�advances�in�commercial�Computational�FluidDynamics�(CFD)�software.�In�this�article,�we�demonstrate�how�CD-adapco’s�simulationtechnology�can�be�applied�to�accurately�predict�how�a�submarine’s�motion�is�driven�byhydrodynamic�forces,�and�compare�numerical�results�with�experimental�data.�
Office of Naval Research, USA.Dejan Matic, Bill Clark, Ganesh Venkatesan, CD-adapco.
� Fig:01Mesh resolution on propeller andcontrol surfaces.
..::FEATURE ARTICLE Defense
MARINE�REPORT14
Submarine ManeuveringSimulations
The�‘Crazy�Ivan’�is�a�naval�term�for�a�submarine�maneuver,�characterized�by�anynumber�of�sudden�and�sharp�turns,�used�by�submarine�crews�to�"look�behind"�theirboat�using�sonar.�The�"Crazy"�part�of�the�name�comes�from�the�fact�that�thesemaneuvers�were�very�sudden�and�"Ivan"�was�a�common�nickname�used�to�refer�tothe�Russians�during�the�Cold�War.�A�standard�tactic�of�pursuing�submarines�wouldbe�to�closely�follow�the�Soviet�submarine�hidden�right�in�the�sonar�gap,�causing�theU.S.�submarine�to�go�undetected.
� FACTS�
..::FEATURE ARTICLE Defense
MARINE�REPORT 15
� MORE�INFORMATION�VISIT� http://www.onr.navy.mil/
stresses�and�pressure�distribution�on�the�surface�of�the�body,�providing�thehydrodynamic�forces�and�moments�acting�on�it.�The�equations�of�motion�arethen� solved� in� order� to� obtain� instantaneous� displacements� and� rotations.This�information�is�used�to�update�the�computational�mesh�which�is�rotatedand�translated�as�a�rigid�body�with�respect�to�an�inertial�frame�of�reference.�
The�integration�and�rigid�body�mesh�movement�are�performed�automaticallyusing� CD-adapco's� Dynamic� Fluid-Body� Interaction� (DFBI)� model� at� eachiteration.� By� converging� this� iteration� process� at� each� time� step,� thetrajectory� of� the� body� is� obtained.� The� implicit� nature� of� the�method� (inwhich�equations�of�motion�are�calculated�simultaneously�with�the�flow�field)is�important�to�ensure�the�overall�stability�of�the�simulation�without�usingan�impractically�small�time�step.
Computational mesh The�discretized�domain�consisted�of�3�million�computational�cells,� includinglayers�of�prismatic�cells�next� to� the�walls,�which�was�prescribed� in�order� tocapture�the�near�wall�boundary�layer.�The�mesh�was�automatically�constructedusing� CD-adapco's� automatic� hexahedral� meshing� methodology:� a� simplebackground� hexahedral� mesh� was� created� within� the� boundaries� of� thecomputational� domain,� overlapping� the� geometry� of� the� submarine.� Anyhexahedral�cells�that�were�located�completely�inside�the�body�or�the�extrudedlayer�were�deleted,�while�those�that�intersect�this�layer�were�trimmed�so�thatany�overlaps�were� removed.�Finally,� the�mesh�was� locally� refined� in� regionswhere�large�flow�variations�were�expected.�
The� propeller� was� enclosed� inside� the� cylindrical� mesh� block� that� rotatesabout�the�propeller�axis,�with�a�sliding�interface�between�the�cylindrical�meshblock�and�the�surrounding�fluid�domain.�Rudder�control�surface�motions�wereaccounted�for�by�using�mesh�distortion.�As�the�rudder�is�deflected�to�a�newposition� at� each� time� step,� the� mesh� in� this� structured� block� is� locallydeformed� and� smoothed.� By� employing� this� procedure� only� a� singlecomputational�mesh�had�to�be�generated�for�the�entire�simulation�-�rather�thancreating� several� meshes� for� various� rudder� positions� and� interpolatingbetween� them.� Because� the� rudder� mesh� motion� was� integrated� into� thesolution�process,�less�user�input�was�required.�
Maneuvering simulations For�the�case�of�constant�heading�and�large�depth,�the�submarine�is�assumedto�be�traveling�through�an�infinite�domain�of�stagnant�water.�The�motion�of�thesubmarine�is�controlled�by�a�3-bladed�propeller,�rudder�and�stern�planes.�Theentire�computational�mesh� including� the�submarine�body� is�assumed�to�bemoving�with� the�body�without�any�deformation.� The� flow� field� computationswere� performed� in� the� inertial� frame� of� reference,� which� makes� thespecification� of� boundary� conditions� easier.� Since� the� body�moves� throughinfinite� volume� of� stagnant� water,� the� velocity� specified� at� the� far� fieldboundaries�of�the�computational�domain�is�zero.
For�the�case�of�horizontal�overshoot�maneuvering,�the�top�and�bottom�ruddersurfaces�were�actuated�to�initiate�the�maneuver.�In�the�experiment,�the�rudderwas� first� deflected� to� 10� degrees� and� held� in� this� position� until� the� bodyreached�a�yaw�angle�of�30�degrees.�The�rudder�was�then�reversed.�Figure�2shows� predicted� pressure� � distribution� on� walls� and� streamlines� behindpropeller.� Predicted� time� history� of� roll,� pitch� and� yaw� angles� show� goodqualitative�agreement��with�measurements,�see�Figures�3�and�4.
Conclusions Good� qualitative� agreement� has� been� shown� between� predictions� andmeasurements�for�the�studied�maneuvers.�The�results�obtained�demonstratethe� suitability� of� the� present�methodology� for� the� simulation� of� submarinemaneuvers� and� motion� of� similar� underwater� autonomous� vehicles.� CFDsimulation� tools� will� help� engineers� to� optimize� the� design� and� analysisprocess�and� improve� the�maneuvering�capabilities,�survivability�and�cost�ofsubmarines.��
� Fig:02Surface pressure distribution with streamlines.
� The Office of Naval Research (ONR)coordinates, executes, and promotes thescience and technology programs of theUnited States Navy and Marine Corpsthrough schools, universities, governmentlaboratories, and nonprofit and for-profitorganizations. It provides technical adviceto the Chief of Naval Operations and theSecretary of the Navy and works withindustry to improve technologymanufacturing processes.
� Fig:03Comparison of predicted yaw angle withmeasurements for horizontal overshoot maneuver.
� Fig:04Comparison of predicted in-plane trajectory of body center-of-gravity with measurements for horizontal overshoot maneuver.
16
UNDEX CFD MODELING
SPECIFIC HEAT RATIO
BUBBLE PER
IOD
� Fig:02Specific heat ratio vs. bubble period computed using STAR-CD.
In�ensuring�that�marine�equipment�exceeds�the�requirements�ofthis�specification,�Lockheed�Martin�employs�numerical�simulation,including�Fluid-Structure-Interaction�simulations�of�the�secondarywave�impact�on�a�vessel.�This�article�describes�how�CD-adapco’s
CFD�software�is�used�to�generate�those�waves.
IntroductionUnderwater�Explosions�(UNDEX)�are�complex�phenomena�that�consist�of�ashock�wave� followed�by�a�secondary�pressure�pulse�wave;� this�secondarywave� is� created� by� the� contraction� of� the� explosive� product� gases.� Thesecondary� wave� pressure� pulse� profile� was� successfully� created� by� usingSTAR-CD�V4.0� and� its� Volume-of-Fluid� (VOF)�model� for� compressible� flowswith�free�surface.
CFD modelingSTAR-CD�was�used�to�define�a�secondary�wave�pressure�profile�which�waslater�used�for�a�Fluid-Structure-Interaction�(FSI)�analysis�involving�a�MaritimeVessel� (MV)� used� for�MIL-S-901D� testing.� � The� secondary�wave,� althoughsmaller�in�overall�induced�force�levels�that�are�incident�upon�the�MV,�has�asimilar�amount�of�impulse�compared�to�that�of�the�initial�shock�wave.��Theseimpulses�can�cause�significant�damage�to�slower�rebounding�systems�-�suchas� shock� isolation� equipment� and� installed� payloads� that� are� tested� forimpact�resistance�on�the�MV.
The�whole�purpose�of�MIL-S-901D�testing�is�to�qualify�shipboard�equipment-�many�of�which�are�shock�isolated�-�for�naval�usage�under�battle�conditions.��This� test�consists�of�a�series�of�explosions�set�off�at�specified�distances
� Fig:01MIL-S-901D testing, MV shown.
CFD Modelingof UnderwaterExplosions
..::FEATURE ARTICLE Defense
The�US�Navy’s�MIL-S-901D�specification�covers�shock�testing�requirements�for�ship�boardmachinery,�equipment,�systems,�and�structures�(excluding�submarine�pressure�hullpenetrations).�The�purpose�of�these�requirements�is�to�verify�the�ability�of�shipboardinstallations�to�withstand�shock�loadings,�which�may�be�incurred�during�wartime�service�dueto�the�effects�of�nuclear�or�conventional�weapons.
Brian Van Lear, Lockheed Martin Company, USA.
MARINE�REPORT
from�the�MV,�see�Figure�1.�These�explosions�can�cause�considerable�damage,even�though�they�do�not�directly�hit�the�vessel.
The�secondary�wave�pressure�pulse�occurs�after�the�initial�shock�wave�that�isinduced� by� the� explosion� and� is� caused� by� the� behavior� of� the� gaseousproducts�of�the�explosion.�The�gaseous�bubble�initially�overextends,�with�theresult�that�its�internal�pressure�falls�below�ambient�conditions;�this�unstableimbalance� in� pressure� causes� the� bubble� to� compress.� The� compressionends�when�the�bubble�reaches�a�minimum�radius,�at�which�time�the�bubble’sinternal� pressure� reaches� a� maximum� and� one� entire� bubble� oscillationperiod�is�completed.�The�creation�of�the�pressure�maximum�is�what�causesthe�secondary�pressure�wave.
The�VOF�method�was�employed�along�with�user-supplied�(extended)�data�toquantify� the� secondary� wave� pressure� pulse.� The� user-supplied� data� wereused� to� provide� the� pressure-density� relationships� for� the� bubble� gasbehavior.
Figure�2�shows�a�graph�which�relates�the�specific�heat�ratio�of� the�bubblegases�to�the�bubble�period�predicted�using�STAR-CD.��The�bubble�period�wascompared�to�available�Similarity�Relations�from�which�the�specific�heat�ratioof�these�gases�can�be�determined.�The�Similarity�Relations�are�equations�thatwere�derived�from�test�data,�see�Cole�[1]�for�more�details.��The�specific�heatratio�is�one�of�the�missing�pieces�of�data�needed�for�defining�the�secondarywave.
Other�important�pieces�of�information�needed�for�the�analysis�are�the�initialbubble�pressure�and�bubble�radius.��These�values�were�taken�from�Geers�andHunter�[2]�along�with�the�specification�of�the�initial�explosive�charge�weight.
Figures�3�and�4�respectively�show�the�secondary�wave�bubble�at�its�maximumand�minimum�radius.��At�its�minimum�radius,�the�gas�sphere�forms�a�toroid.This� toroidal�shape� is�a� result�of�an� imbalance�of�buoyancy� forces�(due�todifferences�in�depth),�which�cause�greater�acceleration�at�the�bottom�of�thebubble�relative�to�that�at�the�top.��This�in�turn�causes�the�gases�and�waternear�the�bottom�of�the�bubble�to�punch�through�the�top�surface.��
Figure�5� shows�a� secondary�wave� pressure� profile� generated� by�STAR-CD.This�pressure�profile�was�successfully�used�for�a�MIL-S-901D�simulation�ofthe�UNDEX�secondary�wave,�after�the�specific�heat�ratio�was�determined�forthe�product�gases.
ConclusionSTAR-CD�made� it� possible� to� capture� the�secondary�wave�pressure�profileusing�VOF�method�and�treating�both�gas�and�liquid�as�compressible�fluids.The�specific�heat�ratio�for�the�bubble�gas�was�successfully�quantified�by�usingSTAR-CD�and�its�VOF�method.��
� For more than 75 years, Lockheed Martin have helped revolutionize theaerospace industry with a passion for invention that is unparalleled. The Lockheed Martin corporation, headquartered in Bethesda, Md., is organizedaround its core business areas - Aeronautics, Electronic Systems, InformationSystems & Global Services, and Space Systems.Lockheed Martin Aeronautics Company is known for building the finestmilitary aircraft in the world. This recognition was earned through relentlessresearch and development of high-performance aircraft and by continuouslyseeking innovative and low-cost design and manufacturing strategies.
REFERENCES1.�Cole�R.�H.,�Underwater�Explosions,�Princeton,�NJ,�Princeton�University�Press,�1948.2.�Geers,�T.�L.�and�Hunter,�K.�S.,�April�2002,�“An�integrated�wave-effects�model�for�an�underwater�explosion�bubble”,�J.�Acoust.�Soc.�Am.�111�(4)�Pages�1584-1601.
..::FEATURE ARTICLE Defense
MARINE�REPORT 17
An example of a deep underwater explosion isthe�WAHOO�test,�which�was�carried�out�in1958�as�part�of�Operation�Hardtack*.�The�nuclear�device�was�detonated�at�a�depth�of500�feet�(150�m)�in�deep�water.�There�waslittle�evidence�of�a�fireball.�
The�spray�dome�rose�to�a�height�of�900�feet(270�m).�Gas�from�the�bubble�broke�throughthe�spray�dome�to�form�jets�which�shot�out�inall�directions�and�reached�heights�of�up�to1700�feet�(520�m).�The�base�surge�at�itsmaximum�size�was�2�square�miles�(4�km)�indiameter�and�1000�feet�(300�m)�high.
*�A�series�of�72�nuclear�tests�conducted�by�the�UnitedStates�in�1958.
� MORE�INFORMATION����http://www.lockheedmartin.com/
� Fig:05Typical pressure profile generated by STAR-CD
TIME
P
� Fig:06Payload on the MV used in subsequent FSI-analysis.
� FACTS�
Operation�Hardtack
� Fig:03Bubble at maximum radius.
� Fig:04Bubble at minimum radius.
..::FEATURE ARTICLE Defense
MARINE�REPORT18
Prior� to� the� end� of� the� Cold� War� conflict� between� the� twosuperpowers,�the�majority�of�submarine�operations�were�carried�outwhile� completely� submerged� in� the� ocean,� so� the� opportunity� toencounter�waves�was� very� rare.�Now,�while� dealing�with�multiple
small�conflicts�throughout�the�world,�the�submarine’s�operations�no�longer�justtake�place�in�the�silence�of�the�deep�sea.�In�the�wake�of�the�Cold�War�conflict,the� submarine’s� purpose� is� changing� and� its� function� has� become� moreversatile.��Modern�submarines�are�now�expected�to�accept�new�componentsand� systems� such� as� the� Dry� Deck� Shelter� (DDS)� or� the� Advanced� SealDelivery� System� (ASDS),� which� is� designed� to� assist� with� the� efficientdeployment� of� specialized� personnel� and� equipment.� With� new� operatingmodes�come�new�design�demands�and�the�submarine�equipment�has�to�be
able�to�withstand�dynamic�loadings�provided�by�the�waves,�in�addition�to�highhydrostatic�pressures.�
The� rapid� technological� progress� both� in� software� and� hardware� providesnumerous� benefits� in� the� application� of� computer-aided� engineering� to� themaritime� industry.� Accurate� and� reliable� predictions� in� the� early� stages� ofdesign� process� can� reduce� the� need� for� physical� prototypes� which� in� turngreatly� accelerates� development� process� and� cuts� costs.� Furthermore,� thenew�features�of�CFD�software�such�as�Dynamic�Fluid-Body�Interaction�(DFBI)available�since�the�release�of�CD-adapco’s�STAR-CCM+�3.04,�widen�the�scopeof�software�application.�Whether� it� is� the�case�of�a�blade,�a�ship�hull� or�acomplete�offshore�platform�design,�naval�engineers�recognize�these�benefits
Minyee Jiang, Naval Surface Warfare Center, Carderock.Dejan Matic, CD-adapco.
Submarines�become�more�multi-mission�oriented�and�their�mode�of�operation�is�more�oftenat�the�sea’s�surface.�As�such,�waves�become�an�important�parameter�of�a�submarine’sdesign�and�could�have�a�critical�impact�on�its�structural�integrity.�In�an�effort�to�address�these�new�challenges,�the�Naval�Surface�Warfare�Center�atCarderock,�recently�chose�to�use�CD-adapco’s�cutting�edge�CFD�solution�to�mitigate�the�riskof�damage�from�surface�wave�impact.
Simulation of Wave Slapon Submarines
The Carderock Division has�a�very�comprehensive�set�of�technical�capabilities�to�support�its�mission.These�capabilities�cover�all�aspects�of�surface�ship�and�submarine�hull�mechanical�and�electricalsystems� (HM&E)� and� cross� all� life� cycles.� In� essence,� the� Division� supports� ships� from� keel� tomasthead�and�from�cradle�to�grave.
The� Division's� technical� capabilities� were� developed� to� support� Navy� and� military� requirements.However,� because� the� mission� of� the� Division� is� oriented� to� platform� systems� and� not� weaponsystems,�many�of�these�capabilities�are�equally�applicable�to�commercial�ships.�In�fact,�the�legislationestablishing� the� Division� specifically� requires� the� Division� to� "provide� support� to� the� MaritimeAdministration�and�maritime�industry."
� FACTS
US�Navy�submarine
and�use�CFD�to�tackle�some�of�the�most�demanding�projectsmaritime�industry�has�to�offer.
One� of� the� centers� using� CD-adapco� solution� technology� isThe�Naval� Surface�Warfare� Center� at� Carderock.� For� over� acentury,� the� Naval� Surface�Warfare� Center� has� been� at� theforefront�of�technologies�vital�to�the�success�of�the�U.S.�Navyand� Maritime� Industry.� It� has� earned� a� distinguishedreputation� as� the� birthplace� of� superior� naval� technologyworking� in� more� than� forty� disciplines� ranging� fromfundamental� science� to� applied� engineering.� The� center’smission� is� to� provide� research,� development,� test� andevaluation,�logistics�and�integration�of�surface�and�underseavehicles� and� associated� systems.� The� ComputationalHydromechanics�Division�of�The�Naval�Surface�Warfare�Centeris�the�home�of�The�Navy’s�only�major�computational�capabilityfor�surface�and�undersea�vehicle�hull� forms�and�propulsors.By� using� computational� simulation� to� develop� thetechnologies� and� procedures� for� systems� which� define� theexternal�shape�of�the�vehicle,�the�division�supports�all�navalvehicles� from� surface� ships� and� submarines� to� unmannedvehicles�and�craft.�
Led�by�Dr.�Minyee�Jiang,�research�was�recently�conducted�withthe�objective�of�simulating�forces�on�the�ASDS�under�variouswave� height,� wave� length� and� wave� heading� to� identifyconditions� under� which� the� ASDS� systems�may� experiencesignificant� forces.�CD-adapco’s�CFD�software�was�applied� todetermine�pressure�loading�by�simulation�of�the�interaction�ofthe�wave�field�with�the�ASDS�mated�to�the�host�submarine.�
The�interface-capturing�method�(known�also�as�the�Volume-of-Fluid� or� VOF� method)� and� CD-adapco's� high-resolutioninterface-capturing� (HRIC)� scheme� were� used� to� accurately
compute� the� free� surface� flows� and� breaking� waves.Hexahedral� cells� were� applied� across� the� computationaldomain�with�local�grid�refinement�applied�to�the�regions�wherehigh� pressure� and� velocity� gradients� are� expected.�Two� wave� headings� were� investigated� for� the� averagedwavelength�and�height.
The�results�showed�surface�pressure�distribution�over�ASDSand�time�history�of�the�wave�induced�loads.�By�revealing�thelocations�with�high�loading�where�possible�structural�damagecan� occur,� the� pressure� distribution� provided� valuablefeedback� to� design� engineers� about� how� to�make� a� vesselmore�structurally�reliable.�Due�to�the�transient�nature�of�theCFD�simulation,�it�was�also�possible�to�analyze�how�surfacepressure� and� loads� are� changing� with� respect� to� time� andlocation.� This� in� turn� provides� valuable� information� on� theimpact� load� history� of� various� elements� of� the� ASDS.� Theoverall�results�of�this�project�allowed�a�broader�understandingof�the�wave�slap�mechanisms�and�a�good�reference�for�futureinvestigations.�Figures�1�to�3�show�samples�of�these�results.
Following�an�evaluation�of�the�latest�release�of�the�release�ofSTAR-CCM+ 3.04,�Dr.�Jiang�commends�its�ease-of-use.�In�hiswords,�“STAR-CCM+�is�very�easy�to�work�with.�It�is�easy�to�setthe�mesh�and�solver�properties,�perform�mesh�generation�andremedy�using�automated�tools�and�post�process�the�numericalresults”.�Dr.�Jiang�also�notices�that�with�STAR-CCM+ there�isno�need�to�compromise�ease-of-use�over�accuracy.��“We�willcontinue� with� this� project,” says� Dr.� Jiang,� “by� conductingmore�wave� slap� and� slam� validations� and� simulations.� Thewaves� and� ship� interaction� will� be� simulated� with� largercomputer�resources�and�future�applications�will�benefit�fromthe�DFBI�feature�in�STAR-CCM+”.��
� MORE�INFORMATION�EMAIL�[email protected]
� Naval Surface Warfare Center,Carderock Division: The CarderockDivision addresses the full spectrumof applied maritime science andtechnology, from the theoretical andconceptual beginnings, through designand acquisition, to implementationand follow-on engineering.The scope of Division productsencompasses individual components,subsystems, the ship/submarine as awhole, and the Fleet. This includes alltechnical aspects of improving theperformance of ships, submarines,military water craft, and unmannedvehicles, as well as research formilitary logistics systems.In addition, the Division is uniquelychartered by Congress to supportAmerica's maritime industry.
..::FEATURE ARTICLE Defense
MARINE�REPORT 19
� Fig:02Sample of the surface pressure: time = 45.50s (upper), time = 67.00s (lower).
� Fig:01Geometries for the numerical model of ASDS on a submarine
� Fig:03History of Impact load on ASDS appendages.
..::FEATURE ARTICLE Defense
MARINE�REPORT
Dr. Ulf Specht, IABG mbH, Germany.
Safer UAV Landings
UAVs�have�been�used�in�areconnaissance�and�intelligence-gatheringrole�since�the�1950s,�and�morechallenging�roles�are�envisioned,�includingcombat�missions.�Since�1964�the�USDefense�Department�has�developed�11different�UAVs,�though�due�to�acquisitionand�development�problems�only�3�enteredproduction.�The�US�Navy�has�studied�thefeasibility�of�operating�Vertical�Take-Offand�Landing�(VTOL)�UAV’s�since�the�early1960s,�the�QH-50�Gyrodyne�torpedo-delivery�drone�being�an�early�example.However,�high�cost�and�technologicalimmaturity�have�precluded�acquiring�andfielding�operational�VTOL�UAV�systems.
� FACTS�
QH-50
20
Performing�landing�operations�of�a�helicopter�or�an�UAV�(unmanned�air�vehicle)�on�ahelideck�of�a�battle�ship�is�a�critical�situation�for�both�the�helicopter�pilot�and�the�crew�ofthe�ship.�A�detailed�understanding�of�the�flow�structures�and�the�magnitude�of�theturbulent�fluctuations�is�therefore�necessary.�Taking�into�consideration�the�very�detailedsuperstructure�of�the�complete�ship,�STAR-CD�can�be�used,�as�it�is�in�this�case,�to�predictturbulence�fields.
� IABG was founded as a centralanalysis and testing organizationfor the aeronautics industry andthe Ministry of Defense in 1961as part of an initiative by theGerman government. Today, it isa leading European technologyand science service provider.The service offering of IABGincludes analytical, technical andoperational solutions in thefollowing areas: Automotive,InfoCom, Transport &Environment, Aeronautics, Space and Defense & Security.
..::FEATURE ARTICLE Defense
BackgroundReconnaissance� is�one�of� the�main� tasks�of� theGerman�Navy.�An�UAV�will�ensure�identification�ofobjects�beyond�the�horizon�of�the�ship�even�under
insufficient� optical� conditions.� Since� the� landing� operationshould�be�a�completely�automated�task,�the�software�of�thecontrol�system�has�to�take� into�account�anything�that�mightaffect� the� smooth� landing� of� the� UAV,� such� as� theenvironmental�conditions.�Therefore�it�is�necessary�to�predictthe� flow� field�around� the�ship,�by� considering�different�windspeeds�and�wind�directions�relative�to�the�ship.
CFD simulationsStarting� from� the� water� surface� level,� a� narrow� box� wasdiscretized� using� ICEM/Tetra� containing� the� complete� ship.This�included�the�complicated�structure�with�a�refined�regionnear�the�helideck.�The�mesh�was�completed�by�adding�blocksof� hexahedral� cells� in� front� of� the� bow,� behind� the� stern,portside,� starboard,� and� using�STAR-CD’s� ‘arbitrary� couples’methodology� above� the� tetrahedral� cells.� The� final� meshconsisted� of� approximately� 2� million� cells.� The� flow� wasassumed� to� be� steady,� incompressible� and� turbulent.Turbulence� is� modeled� by� the� standard� high� Reynolds� k-εmodel.
ResultsA�general�view�of�the�flow�field�containing�isosurfaces�of�theturbulent� kinetic� energy� is� shown� in� Figure� 1.� There� is� aremarkable�production�of�the�turbulent�kinetic�energy�due�tothe�superstructure.�Since�most�of�the�turbulent�kinetic�energyhas� been� dissipated� before� reaching� the� helideck� theseturbulent�fields�hardly�affect�the�situation�there.
A� large�vortex�generated�by�the�main�flow�at�the�end�of�thehangar� (similar� to� a� backward-facing� step)� however,� has� asignificant� impact� on� the� landing� procedure� (Figure� 2).� Thisvortex� interacts�with� the�main� flow�generating�a�shear� layerand�producing�a�region�of�high�turbulent�kinetic�energy.
Simulating�many�different�flow�directions,�these�computationscan�provide�a�“Best�Practice�Guide”�on�how�to�maneuver�theship�prior�to�a�landing�operation.�Furthermore�the�data�sets�ofthe�velocities�and�the�turbulent�kinetic�energy�can�be�preparedas� an� input� for� a� real-time� UAV� approach� and� touch-downsimulation.�
ConclusionsThe�detailed�information�of�complex�flow�patterns�obtained�bySTAR-CD�simulations�improves�significantly�the�understandingof� how� the� turbulent� fields� are� generated.� Furthermore� thecomplete� flow� field,� which� is� impossible� to� achieve� inexperiments,�can�be�used�as�an�input�for�an�UAV�simulationenvironment.��
�� MORE�INFORMATION�ON�IABG�VISIT http://www.iabg.de
� Fig:01Streamlines and isosurfaces of the turbulent kinetic energy (general view).
� Fig:02Streamlines and isosurfaces of the turbulent kinetic energy (detailed view of the helideck).
MARINE�REPORT 21
The� numerical� design� of� naval� propulsionsystems� must� take� into� account� fluid-structure�interaction�phenomena.�We�haveimplemented� a� weak� coupling� procedure
within� the�STAR-CD�code,�based�on�a� finite�elementdiscretization� of� the� structure� problem� and� a� finitevolume�resolution�of�the�fluid�problem.�This�procedureis�applied�to�the�generic�case�described�by�Figure�1,an�elastic�beam�coupled�with�an�incompressible�fluidin� a� confined� vessel� and� subjected� to� imposedtransverse�motion.
The�dynamic�problem�is�solved�using�STAR-CD�for�thenumerical�calculation�of� fluid� forces�and� free�surfacemotion�using�a�VOF�method,�together�with�user�Fortransubroutines�implemented�to�solve�the�linear�equationof�motion�for�the�structure�problem.
� Fig:01Elastic beam coupled with an incompressible fluid with free surface.
� Fig:02Explicit fluid-structure coupling procedure: numerical principle.
� The French Navyawarded DCN the contracts to construct the La Fayette (F710), Surcoef (F711) and Courbet (F712)frigates in 1988, and Aconit (F713)and Guepratte (F714) in 1992. The propulsion system in the LaFayette Class vessels is a CombinedDiesel and Diesel (CODAD)arrangement. The system is based on four SEMT Pielstick 12 PA6 V 280 STCdiesel engines, rated at 21,000hp. Two shafts drive controllable pitchpropellers. The ship is fitted with a bow thruster. The propulsion system provides a maximum speed of 25kt and, at an economical speed of 12kt, the range is 9,000nm.
The�numerical�simulation�of�coupled�fluid-structure�problems�has�made�tremendousprogress�over�the�past�few�years�and�various�numerical�techniques�have�been�implementedin�order�to�describe�fluid-structure�coupling�effects�(many�numerical�examples�are�given�inReferences�3�and�4).
DCN Tackles Fluid-StructureInteraction ProblemJean-François Sigrist, R&D Engineer, DCN Propulsion, France.
..::FEATURE ARTICLE Defense
MARINE�REPORT22
The� numerical� principle� of� the� explicit� coupling� procedure� isillustrated�in�Figure�2.�The�coupling�procedure�uses�the�followingSTAR-CD�subroutines:�the�uparm subroutine�is�used�to�handle�thefluid-structure� interface� motion,� whereas� the� posdat subroutineperforms� the� numerical� calculation� of� fluid� forces� on� the� fluid-structure�interface.
The�equation�of�motion�for�the�elastic�structure�can�be�solved�intwo� different� ways,� using� a� finite� element� discretization� of� thestructure� problem� or� a� modal� decomposition� of� the� structuremotion.�In�the�latter,�a�few�scalar�equations�of�motions�have�to�betime-integrated,�whereas�in�the�former,�a�large�scaled�differentialmatrix� system� has� to� be� solved.� The� time� integration� of� thedynamic�structure�problem�is�performed�with�standard�numericaltechniques,� for� instance,� implemented� in� Fortran� subroutineswithin�STAR-CD.
The�fluid�and�structure�problem�exchange�structure�displacementand� fluid� forces�at� the� fluid-structure� interface,�as�presented� inFigure�3.�Figure�4�gives�a�representation�of�the�CFD�mesh�for�thefluid� part� of� the� coupled� problem,� using� the� standard� VOFapproach� of� STAR-CD� for� the� numerical� treatment� of� fluid� free-surface�problem.
Figure�5�plots�the�time�history�of�the�relative�displacement�of�thebeam’s�free�end�in�response�to�a�sine�wave�shock.�The�numericalcalculation�is�performed�using�the�coupled�procedure�with�a�modalapproach� for� the� structure� problem.� This� figure� shows� theinfluence�of�the�number�of�structure�modes�to�take�into�accountthe�beam’s�dynamic�response.�The�imposed�motion�is�such�thatthe�free�surface�does�not�undergo�large-amplitude�motion;�in�thiscase,� the� calculated� beam� response� is� almost� identical� to� theanalytical�solution�for�the�generic�problem.
Figure� 6� shows� the� fluid� free� surface� and� the� beam� lineardisplacement� at� various� times� during� the� simulation,� for� small-amplitude� free-surface� motions.� This� example� validates� ourapproach� for� the� numerical� simulation� of� the� coupled� genericproblem� with� standard� tools� provided� by� the� STAR-CD� code.�The� coupling� procedure� is� currently� applied� to� the� numericalsimulation� of� coupled� problems� with� various� dynamic� Froudenumbers,�as�well�as�problems�with�large�motion�of�the�fluid�freesurface.� Current� developments� tend� to� describe� non-linearstructural�effects�(see�for�instance�Reference�5).�Future�work�willbe�carried�out�to�describe�the�dynamic�problem�using�CD-adapco’sSTAR-CCM+� (computational� continuum� mechanics)� solvertechnology,�in�order�to�perform�strongly�coupled�analysis�with�finitevolume�discretization�of�the�fluid-structure�interaction�problem.��
�� MORE�INFORMATION:���[email protected]
DCN Tackles Fluid-StructureInteraction Problem
..::FEATURE ARTICLE Defense
MARINE�REPORT 23
FSI for beginners:This� occurs� when� a� fluid� interacts� with� asolid� structure,� exerting� pressure� on� itwhich� may� cause� deformation� in� thestructure�and�thus�alter�the�flow�of�the�fluiditself.�Such� interactions�may�be�stable�oroscillatory,�and�are�a�crucial�considerationin�the�design�of�many�engineering�systems,especially� aircraft.� Failing� to� consider� theeffects� of� FSI� can� be� catastrophic,especially� in� large� scale� structures� andthose�comprising�materials�susceptible�tofatigue.� Galloping�Gertie� (pictured� above),the� first� Tacoma� Narrows� Bridge,� isprobably� one� of� the� most� infamousexamples�of�large-scale�failure.
� FACTS�
� Fig:03Finite element/finite volume coupling procedure: structure displacement and fluid forces exchanges at the fluid/structure interaction.
� Fig:04CFD mesh for the fluid problem of the generic coupled system.
� Fig:06Inner beam motion and fluid free surface at different times.
� Fig:05Relative displacement of the beam free end for various modes taken into account.
REFERENCES1.�Ferziger,�J.H.,�Perić,�M.:Computational�Methods�for�FluidDynamics.�Springer-Verlag,�19992.�Hughes,�T.J.R.,�Belytschko,�T.:�APrécis�of�Developments�inComputational�Methods�forTransient�Analysis.�Journal�ofApplied�Mechanics,�50,�1033-1041,19833.�Makerle,�J.:�Fluid-StructureInteraction�Problems,�Finite�ElementApproach�and�Boundary�ElementsApproaches.�A�Bibliography.�FiniteElements�in�Analysis�and�Design,31,�231-240,�19994.�Schäfer,�M.:�Coupled�Fluid-SolidProblems:�Survey�on�NumericalApproaches�and�Applications.Pressure�Vessel�and�Piping,Cleveland,�20-24�July�20035.�Sigrist,�J.F.,�Melot,�V.,�Lainé,�C.,Peseux,�B.:�Numerical�Simulation�ofFluid�Structure�Problems�by�CouplingFluid�Finite�Volume�and�StructureFinite�Element�or�Modal�Approach.Flow�Induced�Vibrations,�Paris,�5-9July�2004�
� �
� �
� �
� Fig:01Calculation domain: extents (upper) and boundary conditions (lower).
� Fig:02Finite volume mesh of the tanker model: design draught (lower) and ballast draught (upper).
With�the�rapid�growth�of�low-cost�desktop�computational�power�andaccess� to� high� performance� computing� platforms,� numericalanalysis�has�become�a�reliable�design�tool,�which�is�fully�integratedinto� the� overall� product� development� process.� Ongoing
development� in� continuum� mechanics� methods� and� their� numericalimplementation� is� enabling� computer� processors� to� crunch� numbers� evenfaster.� In� this� light,� BI� has� established� a� new� Numerical� HydrodynamicsDepartment�with�the�main�objective�of�increasing�its�level�of�competitivenessin�the�global�market.�
For� the� purpose� of� a� quantitative� evaluation� of� STAR-CCM+’s� simulationtechnology,�a�hull�form�of�a�tanker�model�extensively�tested�in�BI’s�towing�tankwas� chosen� and� provided� a� set� of� reference� experimental� data� for� acomparison.�A�quantitative�margin�of�±10%�to�measured�data�was�consideredreasonable� in� early� simulations,� leaving� considerable� space� for� furtherimprovements�of�the�numerical�model�through�means�of�higher�accuracy�andreduction�in�computation�time.�Two�simulations�were�performed�on�the�baretanker�model�at�design�and�ballast�loading�conditions.�
A�3-D�rectangular�computation�domain�representing�a�part�of�the�towing�tankwas�simulated.�A�domain� length�of� four�model� lengths�was�prescribed,�onelength�in�front�and�two�lengths�behind�the�hull.�Furthermore,�a�domain�size�ofone�model�length�wide�and�one�and�a�half�model�lengths�high�was�chosen,with�an�expected�free-surface�at�one�model�length�from�the�bottom,�definingthe�water�depth.
Finite�volume�meshes�for�two�loading�conditions�were�generated�entirely�withhexahedral� cells.� Parameters� used� in� the� generation� of� the� meshes� wereidentical�for�both�loading�conditions,�with�the�exception�of�the�total�number�ofgenerated�cells�due�to�the�difference�in�tanker�model�position�relative�to�thefree-surface.� Cell� sizes� in� the� domain� were� gradually� refined� towards� theregions� where� high� gradients� of� velocity� and� pressure� were� expected.� Aprismatic�layer�was�generated�around�the�hull�model,�defining�the�appropriatecell�thicknesses�in�order�to�resolve�the�boundary�layer.
STAR-CCM+� RANSE� solver� was� used� to� solve� the� governing� equations� ofcontinuity� and� momentum� for� the� viscous� flow� around� the� hull.� The� freesurface�was�modeled�using� the�Volume-of-Fluid� approach� (VOF)�with� a�HighResolution�Interface�Capturing�(HRIC)�scheme.�The�solution�convergence�wasjudged�using�residuals,�and�by�monitoring�variation�in�the�hydrodynamic�forcesexerted�upon�the�hull.�Mean�values�of�the�drag,�the�shear�and�the�pressurecomponent�of� the�drag,�and� the� lift� forces�over�a�specific� time�period�werecalculated�for�comparison�with�experimental�data.
The�results�of�the�free-surface�elevation�computation�around�the�tanker�modelat�the�design�and�the�ballast� loading�condition�were�presented�qualitatively,and�compared�with�photographs�taken�during�the�testing�of�the�model�in�thetowing� tank.� The� development� of� the� bow� and� stern� wave� systems� wascaptured�realistically�at�both�loading�conditions,�forming�wave�patterns�aroundthe�hull�model�similar�to�the�ones�observed�in�the�experiment.
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT24
Numerical Modeling ofthe Flow Around aTanker Hull Boris Bućan, Marta Pedišić Buča, Stanislav Ruzić, Brodarski Institut, Croatia.
Dejan Matic, CD-adapco.
Building�on�fifty�years�of�experience�in�shipbuilding�science�and�practice,�BrodarskiInstitut�(BI)�has�achieved�a�reputation�as�a�leading�research�and�development�organizationintegrated�into�a�pan-European�network�of�similar�institutes.�After�evaluating�severalcommercially�available�CFD�codes,�BI�chose�CD-adapco’s�STAR-CCM+�to�deliver�on-time,on-cost,�best�in�class�technology�products�and�services.��
� Fig:07Wake field at propeller position:design draught (lower) and ballastdraught (upper). Color representssimulation result and lines arefrom experimental data.
5d
5c
5b
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT
Streamlines� plotted� on� the� hull� model� indicate� a� possible� region� of� flowseparation�at�the�stern�bulb,�near�to�the�propeller,�at�both�loading�conditions.�Afurther�analysis�of�the�shear�stress�on�the�hull�model�reveals�the�presence�ofreverse�flow�region�at�the�same�position�-�a�result�also�observed�in�the�towingtank.�Wake�fields�computed�at�the�propeller�disk�for�both�loading�conditions�arecompared�to�the�experiment�by�overlapping�the�measured�black-and-white�iso-wake�curves�over�computed�color-filled�results.�Qualitatively,�shapes�of�the�wakefields�are�very�well�captured�at�both�loading�conditions.�
The�analysis�of�the�results�of�the�numerical�prediction�of�flow�around�the�tankermodel�and�the�presented�comparison�with�experimental�data�show�that�valuesof� computed� drag� and� lift� coefficients� are� satisfactory� for� both� loadingconditions,�with�a�margin�of�±3%�to�the�experiment.�The�elevations�of�the�freesurface� calculated� around� the� tanker� model� at� both� loading� conditions� arecaptured� realistically,� according� to� photographs� taken� during� the� experiment.Finally,�some�flow�separation�detected�in�the�stern�region�of�the�tanker�modelnear�the�propeller�is�at�the�same�position�as�observed�in�the�experiment.
In�conclusion,�it�can�be�said�that�the�obtained�computational�results�met�thecriteria� of� this� quantitative� evaluation� and� that� the� use� of� STAR-CCM+considerably�helped�to�provide�an�understanding�of� the�flow�patterns�aroundthe� tanker� model.� Nevertheless,� Brodarski� Institut� plans� to� continueinvestigations�on�further�improvements�of�the�presented�numerical�model.�Thesubject�of�subsequent� research�and�development�will�be� the�optimization�oftotal�cell�number,�sizes�and�distribution,�and�the�application�of�more�accurateturbulence�models. �
�� MORE�INFORMATION� http://www.hrbi.hr/
� Fig:03Comparison of the free-surface elevation around the tanker model at design draught; 3a) & 3c) numerical computation; 3b) & 3d) photograph taken during the experiment.
� Fig:04Comparison of the free-surface elevation around the tanker model at ballast draught; 4a) & 4c) numerical computation; 4b) & 4d) photograph taken during the experiment.
25
� Fig:06Flow separation region at the stern of the tanker model; 6b) numerical computation; 6a) photograph taken during the experiment.
The first tankers, built�from�the�1870s�onwards,�generally�had�single�hulls�divided�into�aseries�of�tanks.�Due�to�environmental�and�security�concerns,�modern�tankers�now�have
double-hulls,� so� that� if� the� outer� hull� is� damaged� the� cargo� in� the� inner� hull� will� be
protected.�The�1990�Oil�Pollution�Act�made� it�mandatory� for�all� tankers�calling�at�U.S.
ports�to�have�double�hulls.�It�was�passed�in�reaction�to�the�worst�oil�spill�in�U.S.�waters,
the�1989�Exxon�Valdez�tanker�spill.�In�1992�the�International�Maritime�Organization�(IMO)
recommended�making�double�hulls�mandatory�for�all�tankers�carrying�heavy�crudes�and
fuel.�It�opted�for�a�staggered�phase-out,�in�recognition�of�the�move's�cost�to�the�industry.
The�massive�Erika�oil�tanker�spill,�off�the�coast�of�France�in�1999,�led�to�the�timetable�for
the�global�phasing�out�of�single-hulls�being�accelerated�to�2010.
� FACTS�
Boris Bućan, Marta Pedišić Buča, Stanislav Ruzić, Brodarski Institut, Croatia.Dejan Matic, CD-adapco.
3d
3c
3b
3a
Exxon�Valdez�spilled�an�estimated�10.8�million�US�gallons�(40.9�million�liters)�of�crude�oil.The�oil�spill�fouled�1,500�miles�of�Alaskan�shoreline�and�killed�more�wildlife�than�any�prior�environmental�disaster.
� Fig:05Streamlines on the tanker model hull; 5a) & 5b) design draught; 5c) & 5d) ballast draught.
� The Brodarski Institutwas founded in 1948 as aninstitution of special interest tothe Republic of Croatia in thefield of marine industry, scienceand technical systems.The Brodarski Institut is anadvanced technologydevelopment organization thatcreates, transfers and appliesknowledge to innovative productsand services of high added valueand quality on the domestic andinternational market.
6b6a
4d
4c
4b
4a 5a
This� article� reviews� a� small� selection� of� cases� on� which� Lloyd’sRegister�has�worked�over� the�past�15�years�and,� in� the�process,illustrates�the�rapid�pace�at�which�CFD�technology�has�developedand�demonstrates�the�extent�to�which�CFD�has�become�a�practical
tool�for�marine�design�and�safety�assessment.
Lloyd’s�Register�were�early�adopters�of�CFD�technology�and�amongst�the�firstin� the�marine� industry� to�deploy� flow�simulation�as�a�standard�part�of� theirassessment� process,� expending� significant� effort� on� the� validation� andverification�of�CFD�methodologies�along�the�way:
“We�were�among�the�first�to�evaluate�CFD�technology�in�the�early�1990s�formarine�applications�and�subsequently�adopt�it�for�advanced�modeling,” saysJohn�Carlton,�Global�Head�-�Marine�Technology�and�Investigations.�“Over�thepast�15�years,�the�CFD�technology�itself�has�evolved�and�matured�significantly,
with�advances� in�mesh�generation�and� turbulence�modeling,” he�continues.“During�this�time�Lloyd’s�Register�has�gained�extensive�experience�in�applyingCFD�to�the�solution�of�a�wide�range�of�technical�problems�and�we�have�helpedto� validate�and� improve� the�accuracy�of� the�predictions�obtained� to�a�pointwhere� shipyards� are� preparing� to� submit� designs� based� largely� on� CFDanalyses.”
In�one�of� the�earliest�projects�Lloyd’s�Register�applied�CFD� to�help�a� clientmitigate� the� effect� of� severe� erosion� observed� during� the� operation� of� awaterjet.�A�series�of� two-dimensional�CFD�simulations�helped� to� identify� theregions�of�low-pressure�that�were�responsible�for�the�erosion,�and�guided�thedesign�of� the�shape�modification� to� the� inlet� throat� that�ultimately� led� to�asignificant�improvement�in�the�flow�through�the�intake�and�the�mitigation�of�theprincipal�mechanism�of� erosion.� In� 1994� this�was� no�mean� feat,� as�DejanRadosavljevic,�Lloyd’s�Register’s�CFD�Manager�recalls:
Lloyd’s Register of Shipping:
The value of CFD in Ship Design & AnalysisComputational�Fluid�Dynamics�(CFD)�analysis�is�fast�becoming�an�essential�part�of�designand�performance�assessment�procedures�within�the�marine�industry.�With�its�strategic�rolein�evaluating�and�applying�leading�edge�technologies�for�the�benefit�of�its�clients�and�theindustry�as�a�whole,�Lloyd’s�Register�has�expended�considerable�effort�in�verifying�andvalidating�CFD�technology.
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT26
� Our marine business principally involves classification of ships, whichsets standards of quality and reliability during their design, constructionandoperation. As well as carrying out statutory inspections for nationaladministrations, required by international shipping conventions and codes,we also help ship owners and operators understand risks and improvebusiness performance. And for a century, our expertise has extended further. The Lloyd's Register Group also provides independent risk managementsolutions and inspection services that optimize asset performance in oil andgas, rail, utilities, general engineering and manufacturing.Lloyd’s Register are experienced users of both STAR-CD and STAR-CCM+,and have an extensive track record of co-operation with CD-adapco.
� Photograph of empty LNG tank.
�� MORE�INFORMATION�ABOUT�LLOYDS�REGISTER PLEASE�VISIT: http://www.lr.org
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT 27
“Despite�modeling�the�problem�in�just�two-dimensions,�each�simulation�tookseveral�weeks�to�complete�with�the�computing�resources�available�to�us�at�thetime.”
A�year� later,�Lloyds’�Register�was� involved,�among�many�other�projects,� in�astudy� that� examined� the� shielding� effect� of� the� funnel� guard� for� protectingpassengers�at�the�aft�sundecks�of�a�large�passenger�ship.�“Using�CFD�it�waspossible�to�evaluate�various�wind�directions�and�cruising�speeds� in�order� todetermine�conditions�that�could� lead�to�smoke�being�sucked�down�onto�thedeck�and�potentially�threatening�the�health�and�wellbeing�of�passengers,”�saysRadosavljevic.�“The�atmospheric�boundary�layer�was�fully�modeled�to�providerealistic�flow�conditions�around�the�ship�and�an�optimal�solution�was�found.”
Ensuring�passenger�safety�is�a�principal�concern.�In�another�project�CFD�wasutilized� to�model� the� spread�of� smoke� in� the�event� of� a� sudden� fire� in� therestaurant�area�of�a�large�passenger�ship.�The�restaurant�was�spread�over�twofloors�with�a�large�atrium�in�the�middle.�Several�scenarios�were�simulated,�eachof�which�examined�different�ventilation�strategies�and�the�influence�of�varyingthe�size�of�fire.�“The�CFD�simulation�showed�the�spread�of�smoke�with�timeallowing�us�to�ascertain�how�long�it�would�take�the�smoke�to�prevent�passengerevacuation,”�says�Dejan.
Another�significant,�and�constant,�challenge�for�ship�designers�is�the�control�ofcavitation.�One�of�the�first�projects�to�study�cavitation�erosion�was�carried�outfor�an�A-bracket�in�1998.�The�main�source�of�cavitation�was�identified�as�a�localzone�of�very�low�pressures�that�led�to�flow�separation.�Predictions�from�the�CFDsimulations�were�corroborated�by�in-service�measurements�and�observationsand� the� findings� used� to� propose� modifications� to� the� bracket� shape� toeliminate�cavitation.�“In�the�early�days�the�available�CFD�technology�restrictedus� to� identifying� the� low� pressure� regions� in� which� cavitation� was� likely� tooccur,” says�Dejan.�“However,�with�the�help�of�advancements�in�CFD�we�have,since�2005,�been�able� to�perform� full-two�phase�cavitation�simulations� thatallow� us,� for� example,� to�minimize� the� likely� occurrence� of� cavitation� on� arudder�surface�very�early�in�the�design�process.”
As�part�of�a�commitment�to�monitor,�validate�and�contribute�to�development�inmarine� CFD,� Lloyd’s� Register� recently� participated� in� the� three-year� EU-sponsored� project� EFFORT.� The� main� focus� of� EFFORT� was� to� increaseconfidence�in�CFD�predictions�of�aft�end�flows�through�extensive�validation�andfurther�development�of�CFD�software.�Due�to�the�uncertainties�associated�withempirical�scaling�procedures�for�spatial�wake�distributions,�wake�prediction�atboth� scales� and� the� effects� of� turbulence� modeling� were� prioritized.� SinceEFFORT’s�conclusion,�Lloyd’s�Register�has�carried�out� its�own�validation�of�awake�prediction�on�a�selection�of�cases�as�a�step�on�the�road�to�modeling�flowaround�a�fully�appended�hull�that�includes�a�working�propeller.�The�figure�showsa�full-scale�measurement�of�a�propeller�wake�field�behind�a�container�ship�asprovided� by� the� Hamburgische� Schiffbau-Versuchsanstalt� GmbH� (HSVA)database�as�an�example.
Most� recently,� CFD� has� been� applied� to� the� development� of� technologicalsolutions�for�liquefied�natural�gas�(LNG)�shipping.�“As�LNG�ships�increase�insize�and�ship�owners�and�operators�need�to�be�able�to�operate�them�at�moreflexible� fill� ranges,� sloshing� become� an� important� issue� demanding� carefulconsideration,” says� Radosavljevic.� “CFD� predictions� have� become� anessential�part�of�the�design�phase�for�LNG�containment�systems�on�large�shipsdue�to�the�lack�of�in-service�data.”
Dejan� Radosavljevic� is� confident� that� CFD� will� play� an� even� greater� role� inmarine�engineering�in�the�future:�“With�a�constant�stream�of�new�challenges,the�need�to�stay�at�the�forefront�of�technological�developments�has�never�beengreater,” he�says.�“While�operational�experience�of�the�new,�larger�and�morecomplex� ships� coming� into� service� is� limited,� the� marine� industry� needsvalidated�CFD� tools� to� be�able� to�maintain� high� levels� of� confidence� in� thesafety�of�their�design.�As�computer�hardware�gets�faster,�more�powerful�andmore� affordable,� increasingly� fine� computational� meshes� can� be� achieved,allowing� CFD� to� handle� ever� more� sophisticated� physical� models� to� anincreasingly�high�level�of�accuracy.�It�is�conceivable�that�phenomena�such�asturbulence,�cavitation�and�multi-phase�effects�that�are�still�difficult�to�calculatewill�be�modeled�routinely�and�accurately�within�the�next�10�years.”
By�coupling�finite�element�analysis�(FEA)�with�CFD,�users�will�be�able�to�applyaccurate�structural�loadings�which�could�minimize�the�need�for�safety�margins.“This�coupled�analysis�would�also�allow�for�much�more�accurate�assessmentof� sloshing� loads� in� LNG� tanks� at� full� scale� which� can� currently� only� beestimated� by� applying� various� ‘correction’� factors� to� small-scale�model� testresults,” says�Radosavljevic.�“FEA�used�together�with�CFD�could�allow�the�fullreplication� of� the� flexibility� of� LNG� tank� walls� including� the� effect� of� theadditional�loading�on�the�underlying�flexible�hull�structure�caused�by�the�seaenvironment.”
However,�as�Radosavljevic�recognizes�that�CFD�is�a�tool�that�is�only�useful�inthe�hands�of�an�experienced�engineer:�“To�fully�utilize�the�benefits�of� theseadvances�the� industry�also�depends�on�the�experience�and�expertise�of� thepeople�using�these�technologies,” he�says.�“In�order�to�get�meaningful�results,a�practitioner�must�have�knowledge�and�understanding�of�what�he�or�she�islooking� at.� Lloyd’s� Register’s� consultancy� team� has� vast� experience� invalidating�CFD� results�and�verifying� internal�procedures,�placing� itself�at� theforefront�of�CFD�developments.”�
� Comparison of propeller wakesimulation with experimental results
� Photograph of empty LNG tank.
� LNG sloshing simulation using STAR-CCM+.
� STAR-CD simulation of rudder cavitation.
..::FEATURE ARTICLE Ship Design
MARINE�REPORT28
The� design� of� a� typical� tanker� vessel� involvesreaching�a�compromise�between� the�maximizationof�cargo�volume�and�minimiz-ation�of�required�powerfor� given� ship� size.� The� largest� determinant� of
resistance� is� the� shape� of� bow� and� stern;� the� latter� isparticularly�important,�since�it�needs�to�provide�an�adequateinflow�to�propeller.�Large�variation�of�velocity�magnitude�anddirection�ahead�of�propeller�-�especially�flow�separation�-�canbe�detrimental�to�propulsion�efficiency.
In�the�early�design�stage,�one�usually�deals�with�a�bare�hull.Several� international�workshops�have�been�organized� in� thepast�with�the�aim�of�comparing�results�from�CFD-codes�for�flowaround�ship�hulls�with�experimental�data.�As�part�of�one�such
effort,� Korean� Research� Institute� of� Ships� and� OceanEngineering�(KRISO)�provided�detailed�measurements�for�theKVLCC�(KRISO�Very�Large�Crude�Carrier)�hull,�which�has�servedas�benchmark�for�many�CFD-studies.�Here�the�Version�2�of�thehull�has�been�used�to�validate�STAR-CCM+�and�its�Reynolds-stress�turbulence�model�for�such�applications.
Experiments�were�performed�in�a�wind�tunnel�using�a�doublebody;� therefore,� in� the�simulation�no�free�surface� is�presentbut� instead�a�symmetry�boundary�condition� is�prescribed�atthe�top�boundary.�The�ship�model�was�5.5172�m�long�and�itsnominal� speed� in� water� would� be� 1.047� m/s� (Reynoldsnumber� of� 4600000).� Simulations� were� performed� on� atrimmed�mesh�with�4�million�cells;�prism�layers�were�applied
The�accurate�prediction�of�ship�resistance�is�important�for�determiningthe�required�propulsion�power.�Due�to�long�lead�times�in�the�enginemanufacturing�process,�a�suitable�power-plant�must�be�selected�veryearly�in�the�ship�design�process.�Consequently,�being�able�to�determinethe�resistance�of�the�vessel�and�the�velocity�distribution�at�the�propellerplane�is�of�very�great�importance�in�ship�design.
Ji-Hun Jeong & Won-Dae Jeon, CD-adapco Korea
Validation of STAR-CCM+:Predictionof TankerResistance
� Fig:01Details of mesh structure at bow (top), at stern (middle) and in asection through propeller plane (bottom).
S-bottom�(A)Hard-chined�(B)Soft-chined�(C)
Hulls�come�in�many�varieties�and�can�have�composite�shape,�(e.g.,�a�fineentry�forward�and�inverted�bell�shape�aft),�but�are�grouped�primarily�as�follows:Moulded, Round Bilged or Soft-chined: Examples�are�the�round�bilge,�semi-round�bilge�and�s-bottom�hull.�DEFINED AS SMOOTH CURVES
Chined and Hard-chined: Examples�are�the�flat-bottom�(chined),�v-bottom�andmulti-bottom�hull�(hard�chined).HAVE AT LEAST ONE PRONOUNCED KNUCKLE THROUGHOUT ALL OR MOST OF THEIR LENGTH
� FACTS�
� MORE�INFORMATION�VISIT� http://www.kr.cd-adapco.com/
along�the�ship�hull,�and�the�mesh�was�successively�locally�refined�towards�thehull.�Figure�1�shows�segments�of�the�mesh�with�refinement�structure.
Inlet� boundary� was� prescribed� about� one� hull� length� upstream� of� the� bow;outlet�boundary�was�located�one�hull�length�behind�stern.�The�side�and�bottomboundaries�were�also�located�one�hull�length�away�from�body.�Only�starboardhalf�of�the�model�was�considered,�with�symmetry�boundary�condition�applied.The�sim-ulation�took�about�20�hours�on�8�CPUs.
Figure�2�shows�predicted�streamlines�around� the�hull.� The�complex� vorticalstructure�of�the�flow�in�the�stern�region�poses�a�great�challenge�for�the�CFD-method�and� turbulence�model� in�particular;�most� turbulence�models�do�notcorrectly�predict�the�velocity�field�in�this�region.�Previous�experience�suggeststhat�Reynolds-stress�turbulence�models�typically�produce�the�produce�the�bestresults,�so�a�Reynolds-stress�model�was�selected�here.
The�streamlines�along�wall,�colored�by�the�wall�shear�stress�magnitude,�areshown� in� Figure�3.� They� show� the� complex� near-wall� flow� structure� in� thestern�region,�which�complements�the�information�given�by�the�streamlines�inFigure�2.
Figure�4�shows�lines�of�constant�velocity�at�several�cross-sections,�indicatingthickening�of�boundary�layer�along�the�hull�and�especially�in�the�stern�region.The�hook-shaped�contours�in�propeller�plane�are�typical�for�this�kind�of�hull�andare� difficult� to� predict� accurately.�With� the� automatically� generated� trimmedmesh�(utilizing�appropriate�local�refinements)�and�the�Reynolds-stress�model�ofturbulence,�STAR-CCM+�predicts�this�flow�feature�accurately,�as�can�be�seen�inFigure�5.�Both�predicted�contours�of�the�stream-wise�velocity�component�and
the�vectors�in�propeller�plane�agree�very�well�with�those�from�experiment.�Thecomputed�total� resistance�was�compared�with�experimental�data� for� the� fullrange� of� operating.� The� comparison� shows� very� good� agreement,� with� thelargest�discrepancy�at�the�highest�speed�being�of�the�order�of�1%,�while�at�thelowest�speed�the�discrepancy�was�just�0.1%.
This� study� has� shown� that� STAR-CCM+� offers� not� only� the� means� ofautomatically�generating�a�high-quality�mesh�but�also�provides�a�high-qualitynumerics�and�turbulence�models� for�accurate�prediction�of� flow�around�shiphulls.�Similar�results�have�been�obtained�in�several�other�studies,�where�fully-appended� hulls� including� free-surface� effects� have� been� analyzed.� Thoseresults�cannot�be�shown�due�to�confidentiality�of�geometry�and�measurementdata,�which�was�the�reason�for�choosing�a�public-domain�data�for�this�study.
..::FEATURE ARTICLE Ship Design
MARINE�REPORT 29
� Fig:03Wall shear stress on the front side (upper) and rear side of the hull (lower).
� Fig:05Contours of the normal velocitycomponent (left) and projectionof velocity vector in the cross-section (right), from experiment(top) and simulation (bottom).
� Fig:04Velocity iso-lines in several cross-sections.� Fig:02
EXPERIMENT
SIMULATION
A� properly� designed� propeller� is� a� compromizebetween� structural� and� hydrodynamic� consid-erations;�moreover� it� should�match� the� boat� andengine� characteristics.� An� incorrect� design� may
easily� lead� to�performance�and�speed� losses,�as�well� as� tocavitation,� noise,� vibrations� and� blade� erosion.� A� clearunderstanding�of�propeller�flow�is�therefore�required.�Firstly,�itis�necessary�to�know�how�the�drag,�trim�and�flow�field�aroundthe�boat�change�with�speed;� this�should�be�done,� in� theory,with�model�testing�in�the�so-called�"towing�tank".�Second,�oncethe� propeller� has� been� designed,� it� should� be� analyzed� inmodel� scale� in� a� "cavitation� tunnel".� Both� tests� are� timeconsuming�and�expensive�(much�more�that�the�propeller�itselfin�most�cases)�and�for�this�reason�they�are�seldom�carried�outfor�custom�applications�on�small�and�medium�size�boats.
Research�and�innovation�have�always�been�the�key�factors�ofour�success.�In�the�past,�a�large�effort�has�been�undertakenby�our�company�to�test�and�optimize�families�of�propellers�in
a� cavitation� tunnel.� Some� years� ago� it� became� clear� that,besides�experimental�investigations,�a�break-through�could�beachieved�using�CFD.�This�is�because,�in�our�opinion,�not�onlythe� raw� numbers� obtainable� by� the� towing� tank� and� thecavitation�tunnel�are�important,�but�also�the�flow�visualizationoffered�by�CFD�programs�allows�better�comprehension�of�theflow� phenomena,� thus� speeding� up� the� "learning� curve"� ofthe�designer.
Our�first�objective�was�to�use�a�tool�capable�of�computing�thehydrodynamic�resistance�of�fast�planing�hulls,�which�representthe�majority�of�our�clients’�applications.�This�is�a�field�wherestate-of-the-art� "Panel� Methods"� fail,� mainly� due� to� thecomplexity� of� the� free� surface� generated� (spray,� breakingwaves,�complex�flow�at�the�transom�stern).�
Among� many� commercial� codes� tested,� we� found� that�CD-adapco’s�CFD�solution�has�best�fulfilled�our�requirements.After�a�set�up�period,�we�obtained�results�of�the�same�order
� Rolla Propellers offers a widevariety of services for virtuallyevery aspect of propeller design,manufacture and application. Rolla stainless steel and NiBrAlsurface-piercing propellers rangefrom 28 inches to 9.8 feet indiameter and include linesspecifically designed for all majorsurface drive systems.Since 1963 the name Rolla has beensynonymous with the highestefficiency, highest quality propellersin the world. This means completehydrodynamic engineering service,unique in the industry. Starting witha CFD analysis of the hull,including sea keeping, to the mostsophisticated CFD propeller designmethods for submerged and surfacepropellers.In 2004, Rolla SP Propellers SA hasbecome a part of Twin Disc, Inc.
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT30
ROLLA�SP�Propellers�SA�was�founded�in�1980�by�Philip�Rolla�inSwitzerland.�Since�then�the�company�has�become�the�worldwide�leader�inthe�design�and�manufacture�of�high-speed�propellers.
PropellingMarine Applications
Fig:01 �Calculated pressure distribution on a surface-piercing propeller.
Propeller 841-B, J=0.8
Experiments�(Olofsun)
Rolla
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT 31
of�accuracy�as�a�towing�tank.�This�enabled�us�to�understand�the�boundaryconditions� and� the� input� data� needed� for� propeller� design.� Moreover,� wecould�look�at�the�pressure�distribution�and�the�flow�field,�detecting�problemsand�suggesting�possible�improvements�to�designers�of�hull�shapes.
Besides�the�computation�of�planing�hulls�in�calm�water,�we�found�it�useful�toexplore� a� similar� field� where,� once� again,� no� other� tool� was� available� forcomputation.�Performances�of�planing�boats�are�strongly�affected�by�naturalsea�waves,�and�the�best�hull�shape�must�be�a�trade-off�between�calm�waterand�sea-keeping�qualities.�Sea-keeping�simulations�at� any� speed�and�wavecharacteristic,�are�now�routinely�performed�using�CFD.�The�most�useful�resultsare�the�level�of�acceleration�onboard,�the�added�resistance�to�waves�and�theinstantaneous�pressure�distribution�over�the�hull�bottom,�this�last�result�beingextremely�useful�for�structure�scantling.
Another�field�where�the�CD-adapco’s�CFD�solver�proved�to�be�a�valuable�toolwas� in� the� computation� of� Surface-Piercing� Propellers� (SPP).� At� very� highspeed,�when�cavitation�becomes�unavoidable�and�strongly�erosive,�it�is�helpfulto� let� the� propeller� work� at� the� interface� between� the� water� and� theatmosphere,�allowing�the�water�vapor�cavities�formed�over�the�blades�to�befilled�with�air.�In�this�case�the�free�surface�becomes�extremely�complex;�a�thinunsteady�pocket�of�air�is�formed�over�the�blade�surface�and�a�large�spray�isdeveloped.
Using� the� VOF� and� "sliding� surface"� capabilities� of� the� code,� it� has� beenpossible�to�obtain�results�that�match�the�data�obtained�in�the�cavitation�tunnelwithin�a�few�percent.�Submerged�propellers�are�designed�using�an�"in-house"developed�Panel�Method;�this�gives�a�reasonable�approximation�of�the�flow�forconventional� propellers� working� at� the� design� point.� Unfortunately,� for� off-design�conditions�or�heavily�loaded�propellers,�unacceptably�large�errors�mayoccur.�For�this�reason�every�propeller�designed�is�then�checked�in�a�wide�rangeof� conditions� using� the� CFD� solver� and� eventually� corrected� beforemanufacture.�
An�effort�of�flexibility�has�been�essential�to�switch�from�the�old�established�wayof�designing�marine�propellers�to�these�new�techniques;�the�ROLLA�Researchbranch�was�created�for�this�purpose.�Its�aim�is�to�link�engineers,�productionand�customers,�and�provide�them�with�the�best�tools�for�answering�the�manyquestions�coming�up�in�boat�or�ship�design.
There�is�no�doubt�that�these�technologies�have�largely�improved�the�qualityof�all�our�products,�reducing�the�sources�of�error�and�uncertainty.��
�� Fig:04Sea keeping simulation for a fast planing hull.
�� MORE�INFORMATION: http://www.rolla-propellers.ch/ http://www.twindisc.com/MarineProducts/MarineRollaProducts.aspx
� Fig:03Wave field generated by a planing hull.
� Fig:02aComparison of simulation results with towing-tank data for high-speed planing boats.
Total Resistance [t]
Boat Speed [knots]
Model�test
Rolla�calculation
� Fig:02bPressure distribution on a surface-piercing propeller (left) and a comparison of simulation results with cavitation tank data.
State-of-the-Art�Panel�Method
A dimensional 1 blade thrust, 100oKx
Predicting Drag of Floating VesselsDirk Jürgens, Michael Palm, Voith Turbo Schneider Propulsion GmbH & Co. KG, Germany.Milovan Perić, Eberhard Schreck, Dejan Matic, CD-adapco.
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT32
� As a leading specialist forpower transmission and asolid, competent partner ofcustomers all over the world,Voith Turbo is synonymouswith fairness, innovative powerand reliability in a wide varietyof applications. We work enthusiastically at thedevelopment of state-of-the artdrive and braking systems forindustry, rail, road and marine:mechanical, hydrodynamic,electrical and electronicsolutions that drive machinesefficiently, move vehiclescomfortably, save energy andreduce emissions.
Over 80 years ago, Voith-Schneider Propellers, the only ship's propulsion system
of its kind in the world, was developed by Voith from an idea by the Austrian
engineer Ernst Schneider. It allows thrust of any magnitude to be generated in
any direction quickly, precisely and in a continuously variable manner and
combines propulsion and steering in a single unit. This solution is as convincing
as it is straightforward: on the Voith-Schneider Propeller, a rotor casing which
ends flush with the ship's bottom is fitted with a number of axially parallel blades
and rotates about a vertical axis. To generate thrust, each of the propeller blades
performs an oscillating motion about its own axis. This is superimposed on the
uniform rotary motion.
� FACTS�
� Fig:01Free surface deformation around brick-like body: simulation (left) and experiment (right) for a fixed body (upper) and with three degrees of freedom (right).
The�aim�of�the�studies�presented�here�was�to�assessthe� reliability� of� the� prediction� of� marine� vesselresistance,� as� well� as� the� possibility� to� predict� theeffects�of�design�changes�on�resistance.�The�strategy
behind�this�project�is�to�initially�validate�the�CFD�simulation�usinga�simple�shape�object,�a�brick-like�body�both�in�fixed�and�floatingposition.�When�the�acceptable�level�of�accuracy�is�reached,�theevaluation�is�continued�using�real�marine�vessels.�The�final�stepin�this�project�is�drag�reduction�analysis�of�a�generic�motor�yacht.The�CFD�simulation�is�performed�to�predict�the�drag�on�severaldesign�configurations.
The� aim� of� these� studies�was� to� assess� the� reliability� of� theprediction�of�marine�vessel�resistance,�as�well�as�the�possibilityto� predict� the� effects� of� design� changes� on� resistance.� Thestrategy� behind� this� experiment� is� to� initially� validate� the�CFDsimulation�using�a�simple�shape�object,�a�brick-like�body�both�infixed�and�floating�position.�When�the�acceptable�level�of�accuracy
is� reached,� the� evaluation� is� continued� using� real� marinevessels.�The�final�step�in�this�project�is�drag�reduction�analysisof� a� generic�motor� yacht.� The�CFD�simulation� is�performed� topredict�the�drag�on�several�design�configurations.�
ComputationAll� computations� described� here� are� performed� using�CD-adapco’s�CFD�software.�The�flow�is�assumed�to�be�governedby� the� Reynolds-averaged� Navier-Stokes� (RANS)� equations,� inwhich� turbulence� effects� are� included� via� k-ε or� k-ω models.Liquid�and�gas�are�considered�as�two�immiscible�componentsof�a�single�effective�fluid,�whose�properties�are�assumed�to�varyaccording� to� the� volume� fraction� of� each� component.Computations� were� performed� on� three� meshes� of� differentfineness�in�order�to�estimate�discretization�errors.�These�werefound�to�be�of�the�order�of�3%�on�the�finest�mesh.�The�modelswith�the�finest�mesh�have�approximately�2�million�cells.
The�optimization�of�marine�vessels�requires�that�the�interaction�betweenvarious�ship�components�and�floating�positions�is�taken�into�account.�These�interactions�make�experimental�optimization�costly�and�timeconsuming,�often�with�poor�repeatability,�or�a�limited�number�of�data�points.With�recent�developments�in�computer�technology�and�numerical�methods,Computational�Fluid�Dynamics�(CFD)�simulation�emerges�as�the�moresuitable�approach.�In�this�article,�we�present�the�results�of�several�validationstudies,�which�include�both�simulations�and�experiments�and�explore�theability�of�CFD�to�predict�not�only�the�trends,�but�also�quantitative�variation�inresistance�due�to�design�changes.
..::FEATURE ARTICLE Ship�Design
Figure�1�shows�the�body�in�a�fixed�and�floating�position,and� the� free� surface� deformation� around� it� whenexposed� to�water� flow�at�1�m/s.� These�pictures�showthat� there� is� a� high� qualitative� similarity� betweenpredicted�free�surface�shape�and�the�observed�interfacedeformation�in�experiments.
A� quantitative� comparison� of� predicted� and� measuredresistance�for�a�fixed�body�position�is�shown�in�Figure�2.The�selected�turbulence�model�makes�little�difference�-�avery� good� agreement� is� obtained� for� all� free-streamvelocities.� The� results� for� a� floating� position� aresomewhat� less�perfect�due� to�high� trim�angle,�which� inturn�makes�the�angle�between�streamlines�and�gridlinesless�favourable�as�the�velocity�increases.�Since�the�brick-like�body�with� its�sharp�edges�and�severe�separation�isnot� highly� representative� of�marine� vessels,� no� deeperanalysis�of�results�was�undertaken.�The�obtained�resultsare�considered�satisfactory�for�the�intended�purpose.�
The� evaluation� is� continued� using� more� representativeshapes�such�as�typical�marine�vessels,�a�tug�boat�and�amotor�yacht.�Here�too,�as�seen�in�Figure�3,�the�numericalresults� show� very� good� agreement� with� experiment.� Inparticular,�the�effect�of�trim�and�sinkage�is�clearly�seen.Only� at� higher� speeds,� the� simulations� predict� slightlylower�resistance�than�is�found�in�the�experiment.
ResultsThe� numerical� results� show� that� CFD� simulation� canreliably�predict�the�variation�of�resistance�as�a�function�ofvessel�speed�and�position�so� it� is�possible�to�use� it� toevaluate� various� ideas� for� resistance� reduction.� Theexperimental� data� indicates� that� the� resistance� of� thevessel�could�be�substantially�lower�if�it�would�not�sink�andtrim.�This� leads� to�hull� form�modification� to� reduce� thevariation�in�vessel�position�at�higher�speeds.�
This�reduction�can�be�achieved�if�the�stern�of�motor�yachtis�modified�by�simple�add-ons.�In�one�case,�a�wedge�hasbeen�added�to�the�hull�bottom�at�stern.�Three�more�testswere�performed�with�a�plate�added�to�the�stern�surface,which� protruded� by� a� different� extent,� below� the� hullbottom.�Plate�1�is�in�size�about�1/3�and�plate�2�about�2/3of�plate�3.
Adding�such�an�obstacle�to�the�smooth�hull�surface�leadsto�an�increase�in�resistance�when�the�vessel�is�held�fixedin�the�zero-speed�position.�This�is�correctly�reproduced�insimulations,�the�dark�bars�in�Figure�4�show�that�both�thewedge�and�plates�result�in�a�higher�resistance.�When�thevessel� is� free� to� sink� and� trim,� the� resistance�with� thelargest� plate� is� the� same� as� for� the� original� geometry,while�all�other�design�changes�provide�reduction�in�drag.For�plate�1,�the�reduction�amounts�to�about�3.5�%.�Thus,a�very�simple�modification�of�stern�geometry�can�providea�significant�fuel�saving.
In� another� series� of� studies� at� Voith,� the� application� ofCFD� to� predict� the� performance� of� a� Voith-Schneiderpropeller�has�been�analyzed.�It�has�been�found�that�CFDsimulations� predict� the� variation� of� torque� on� eachpropeller�blade�during�its�rotation�with�sufficient�accuracy.Through�various�design�changes�that�were�developed�withthe� aid� of� CFD,� the� efficiency� of� the� Voith-SchneiderPropeller�has�been�significantly�improved.�
The�approach�to�use�CD-adapco�simulation�technology�fordesign� optimization� has� been� used� at� Voith� withconsiderable� success� during� the� past� five� years.Nowadays,�for�every�Voith-Schneider�Propeller�delivered�toa� customer,� a� series� of� CFD� system� simulations� isperformed.� In� conclusion,� it� can� be� said� that� the� CFDtechniques�predict�the�resistance�and�its�dependence�onspeed� and� geometrical� variations� with� a� sufficientaccuracy,� that� allows� simulations� to� become�an� integralpart�of�product�design�and�optimization.��
� MORE�INFORMATION�ON�VOITH�TURBO http://www.voithturbo.com
MARINE�REPORT 33
� Fig:02 Comparison of measured and computed resistance for the fixed (upper) and free (lower) brick.
� Fig:04a (TOP) + Fig:04b (BOTTOM)Variation of resistance when the sterngeometry is varied (upper) and the effectson trim and sinkage (lower).
� Fig:03Wave patterns for two typical vessels andcorresponding drag predictions in fixed position and with two degrees of freedom (heave and pitch).
CD-adapco’s�RANS�software�was� used� to� evaluate� the� impact� ofkeel�coolers�on�propeller� inflow.�Two�CFD�models�of� the�new�hullform�with�and�without�keel�coolers�installed�were�developed.�Thesetwo�models�were�compared�against�three�models�of�existing�proven
hulls�which�were� developed� to� serve� as� a� control� group.� By� comparing� theturbulence�seen�on�the�five�different�hulls,�the�impact�of�both�the�new�hull�formand�the�proposed�keel�cooler�location�could�be�evaluated.�The�results�of�theflow�simulation�showed�that�the�location�chosen�for�the�keel�coolers�had�nosignificant�effect�on�the�extent�of�turbulence�generated�by�the�new�hull.�Thedisturbance�caused�by�the�keel�coolers�was�primarily�a�local�effect,�and�did�notextend�beyond�the�hull’s�boundary�layer.�The�level�of�turbulence�predicted�forthe�new�hull�with�keel�coolers� installed�was�similar�to�that�seen�on�existingproven� designs,� and� the� flow� into� the� propeller� is� not� affected� by� theirinstallation.
AYM�also�investigated�the�impact�of�a�change�to�the�shape�of�the�skeg�on�anoffshore� supply� vessel’s� performance.� The� skeg� had� been� modified� toaccommodate�a�large�stern�thruster,�and�the�purpose�of�the�CFD�model�was�todetermine�what�the�impact�of�the�different�skeg�shape�would�be.�To�build�asuitable�mesh� around� the�more� complex� 3D� shape� of� the� skeg,� the� RANSsoftware�package�STAR-CCM+�was�used.�CFD�models�were�made�of�the�newskeg� shape� for� this� hull� and� of� a� more� conventional� skeg� shape� forcomparison.
Flow�visualization�around�the�new�skeg�showed�that�separation�and�vorticityhad�increased�as�a�result�of�the�redesign.�However,�while�the�skeg�shape�willincrease�drag,�the�extent�of�the�wake�was�not�sufficient�to�cause�problems�withpropeller�inflow,�and�it�was�determined�that�the�design�would�be�acceptable.
The�use�of�advanced�tools�such�as�STAR-CCM+�gives�AYM�the�flexibility�to�offerclients� alternatives� to� model� testing� to� determine� flow� characteristics,� andthese�types�of�analyses�can�be�integrated�into�the�early�design�stage.�
� Aker Yards Marine (AYM) is a consulting naval architecture and marine engineering companywith offices in US and Canada established to serve the marine community in North America andselected international markets. It is a wholly owned subsidiary of one of the oldest ship makersin the world - the Aker Yards ASA from Norway whose experience spans over a century ofexceptional ship building history. The services provided cover wide range of marine industrysector including all aspects of ship design and production technology. In addition, AYM hasspecialized teams of analysis engineers whose expertise in Computational Fluid Dynamics(CFD) serves to provide practical and cost-effective approach in marine designs.CD-adapco is proud to announce that its CFD simulation technology has been successfullyvalidated and used at AYM for number of years providing accurate modeling of virtually anyconceivable fluid flow. It played a key role in analysis of variety of projects, from hull formoptimization or propeller design to ferry car deck ventilation. Selected accomplishments aredescribed below.
Dan McGreer, Aker Yards Marine.
LOADS ON SHIP HULLS
The�hulls�of�ships�are�subjected�to�a�number�of�loads.
• Even�when�sitting�at�dockside�or�at�anchor,�the�pressure�of�surrounding�water�
displaced�by�the�ship�presses�in�on�its�hull.
• The�weight� of� the� hull,� and� of� cargo� and� components�within� the� ship� bears� down�
on�the�hull.
• Wind�blows�against�the�hull,�and�waves�run�into�it.
• When�a�ship�moves,�there�is�additional�hull�drag,�the�force�of�propellors,�water�driven
up�against�the�bow.
• When�a�ship�is�loaded�with�cargo,�it�may�have�many�times�its�own�empty�weight�of�
cargo�pushing�down�on�the�structure.�
If�the�ships�structure,�equipment,�and�cargo�are�distributed�unevenly�there�may�be�large
point�loads�into�the�structure,�and�if�they�are�distributed�differently�than�the�distribution
of�buoyancy�from�displaced�water�then�there�are�bending�forces�on�the�hull.
� FACTS�
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT34
� Fig:02Faired skeg with less vorticity along bottom and no flow separation at trailing edge.
� Fig:01New skeg design, showing flow detachment along trailing edge of skeg and strong vorticity along bottom edge.
CASE STUDY 01: Validating Hull Performance using CFD
During�two�recent�projects,�Aker�Yards�Marine�was�requested�to�evaluate�the�impact�ofdifferent�design�parameters�on�vessel�performance.�The�first�analysis�was�to�determinewhether�the�location�of�a�ship’s�keel�coolers�had�a�detrimental�effect�on�propeller�inflow,�andthe�second�analysis�was�to�determine�the�impact�of�skeg�geometry�on�vessel�performance.
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT 35
While� steady� state� heat� conduction� and� natural� convection� inenclosed�simple�shapes�are�problems�for�which�analytic�solutionsexist,� the�complex�geometry�around�the�cryogenic�tanks�requireadvanced� methods� for� accurate� determination� of� the
temperatures�on�steel�surfaces�in�and�around�the�tanks.�To�determine�therequired� steel� grades� for� use� in� these� applications,� and� to� minimize� therequirement�for�cold�temperature�steels�in�the�hull�structure,�AYM�performeda�Computational�Fluid�Dynamics�(CFD)�analysis�of�flow�and�heat�transfer�inship’s� cargo� spaces� capable� of� transporting� low� temperature� liquids.� Theship’s�structure�was�designed�to�meet�US�Coast�Guard�rules�which�requireconsideration�of�the�sea�temperature�at�0oC�and�an�air�temperature�of�-18oC.Using�CFD�methods�to�analyze�this�problem�allowed�the�accurate�predictionof� the� natural� convection,� as� well� as� heat� conduction� through� the� steelbulkheads.
The� CFD� model� was� generated� using� STAR-CCM+,� a� state� of� the� art,unstructured� RANS� solver.� The� bulkheads� were� modeled� as� conductingbaffles,�and�a�high�level�of�refinement�was�used�along�the�steel�surfaces�tocapture� the� boundary� layer� flow,� which� is� critical� to� accurately� modelingnatural�convection�processes.�The�effect�of�the�cargo�was�modeled�by�takingthe�exterior�surface�of�the�tank�at�the�liquid�temperature�of�the�cargo.�Thenatural�convection�process�necessitates�solving�the�model�using�an�implicitunsteady�solver,�and�then�taking�time�averages�of�the�temperatures�on�allsurfaces.�Since�the�CFD�models�were�quick� to�build�and�analyze,� the�CFDsolutions�were�used�in�an�iterative�manner�to�guide�the�layout�of�cofferdamsand�bulkheads�to�minimize�the�use�of�exotic�steels.
02: Improving Cryogenic Analyses with CFDCarrying�cryogenic�fluids�such�as�LNG�or�LPG�aboard�a�ship�poses�some�unique�designchallenges.�The�extreme�low�temperatures�of�these�fluids�require�low�thermal�conductivityinsulation,�and�may�also�necessitate�heating�in�various�ship�locations�to�prevent�exoticsteels�from�being�required�throughout�the�ship’s�hull.�The�accurate�determination�of�theheat�transfer�between�the�sea�water,�outside�air,�and�the�cargo�is�critical�to�ensure�thetank�and�bulkhead�arrangements�are�feasible.
� FACTS�
>>� Fig:02
CFD simulation in tank containing LNG and cooling down.
� Fig:01Cross-section of a tank containing LNG, showing computed tank wall temperature.
CRYOGENICS may� be� defined� as� low� temperature� technology,� or� the� science� of
ultralow� temperatures.� To� distinguish� between� cryogenics� and� refrigeration,� a
commonly�used�measure�is�to�consider�any�temperature�lower�than�-73.3°C�(-1OO°F)
as�cryogenic.
Another�use�of�cryogenics�is�cryogenic�fuels.�Cryogenic�fuels,�mainly�liquid�hydrogen,
have�been�used�as�rocket�fuels.�(Oxygen�is�used�as�an�oxidizer�of�hydrogen,�but�oxygen
is�not,�strictly�speaking,�a�fuel.)�For�example,�NASA's�workhorse�space�shuttle�uses
cryogenic�hydrogen�fuel�as�its�primary�means�of�getting�into�orbit,�as�did�all�of�the
rockets�built�for�the�Soviet�space�program�by�Sergei�Korolev.
�� MORE�INFORMATION�ON�AKER�YARDS�MARINE: http://www.akermarine.com/ http://www.akeryards.com/
Sputnik�Rocket
While�general�rules�of�thumb�exist�to�guide�designers�on�the�effectof� propeller� rotation� direction,� the� forces� generated�while� in�DPmode� will� vary� depending� on� the� hull� form,� the� location� of� theskeg,�and�stern�thruster�location�relative�to�the�propellers.�
Consequently,�it�is�quite�difficult�to�quantify�the�actual�effect�of�the�rotationdirection� without� conducting� model� tests.� To� investigate� this� effect,� AYMcreated�a�CFD�model�of�the�stern�section�of�the�ship�using�roughly�700,000cells�in�the�RANS-solver�STAR-CCM+,�see�Figure�1.
The�propellers�were�modeled�in�both�the�inboard�and�outboard�turning�config-urations�while�the�stern�thruster�was�in�operation.�A�crabbing�maneuver�wassimulated� by� running� the� port� propeller� ahead,� the� starboard� propellerastern,�placing�the�rudders�hard�over�to�port�and�using�the�stern�thruster.�Theyaw�moment�and�crabbing�force�were�calculated�during�the�simulation�as�afunction�of�engine�RPM.�
The� use� of� state-of-the-art� computing� resources� and� software� codes,combined�with�a�wealth�of� in-house�experimental� expertise�allows�AYM� to
offer�CFD�analysis�for�various�types�of�problems,�such�as�the�above�study.�Inthe� past,� AYM� has� performed� CFD� analyses� for� engine� room� ventilationstudies� and� pressure� drop� calculations� for� various� ducting� configurations.Other�studies,�such�as�the�reduction�of�appendage�drag,�can�be�handled�byour�engineering�team,�with�rapid�turnaround�times�from�CAD�concept�directlyto�CFD�solutions.��
Recently,�AYM�investigated�the�effect�of�the�propeller�rotation�direction�on�theDynamic�Positioning�(DP)�characteristics�of�a�supply�vessel.�
�� MORE�INFORMATION�ON�AKER�YARDS�MARINE: http://www.akermarine.com/ http://www.akeryards.com/
CRABBING (i.e.�traversing)�is�the�ability�of�ships�to�move�sideways�while�controlling�the
forward� motion.� This� maneuver� can� be� induced� by� using� a� combination� of� main
propellers,�pods,�rudders,�lateral�thrusters�or�other�dedicated�devices.�A�reduction�of
berthing-unberthing�time,�an�increase�of�safety�and�agility,�less�need�for�tug�assistance
are�only�a�few�examples�of�the�benefits�related�to�good�crabbing�abilities.
� FACTS�
� Fig:02Section through propeller plane.
� Fig:01CFD model of stern of ship.
03: Dynamic Positioning Studied with CFD
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT36
>>
Crabbing�Misadventure
CD adapco simulationtechnology is widely usedand thoroughly validated atAYM. It is excellent formodeling of wide spectrumof fluid flows as well as heattransfer and chemicalreactions. In addition, rapidturnaround time fromsurface model to meshcreation is accomplishedusing state of the artunstructured polyhedralmesh and surface wrapping.By providing accuratesolutions in a time efficientmanner it became integralcomponent of our navalarchitecture and marinedesign engineering work.Dan�McGreer�-�Aker�Yards�Marine
“
”
Sulzer� engines� are� of� uniflow� scavengingdesign.�Vessel�propellers�are�generally�directlydriven� with� top� speeds� of� 60� to� 140� rpm,resulting� in� stroke-to-bore� ratios� from� 2.6� to
4.2� for� a� given� mean� piston� speed� slightly� aboveapproximately�8�m/s.
The�resulting�cylinder�proportions�are�depicted�in�Figure1�for�a�piston�position�at�BDC.�During�scavenging,�theburnt� and� unburnt� gases� can� be� kept� more� or� lessseparate� and� the� development� of� "plug-like"� flow� isillustrated�by�three�isosurfaces�(170,�180,�190�deg�CA).
To� improve�the�flow�behavior�even�further,�an�accurateprediction� of� the� flow� through� the� ports� (Figure� 2)� isneeded.�Experiments�are�difficult�because�the�flow�mustbe�transient�and�flow�visualization�techniques�in�enginesof�this�size�are�not�available.�On�the�other�hand�CFD�canalso�be�applied�to�calculate�flow�and�pressure�losses�ininlet�and�exhaust�ports�and�swirl�in�the�cylinder,�as�wellas�the�spin�of�the�exhaust�valve.
Sufficient� valve� rotation� is�desirable� for�achieving� longtimes�between�overhauls.�To�achieve�this,�small�vanesare�added�to�the�stem�that�use�the�exhaust�gas�kineticenergy�for�rotating�the�valve�(Figure�3).
The�rotating�valve�is�a�simple�example�of�fluid-structureinteraction,� which� was� implemented� via� STAR-CD� usercoding�capability.�The�rotation�angle�and�a�comparisonof�measured�and�calculated�pressure�traces�is�depictedin�diagram�1.�Note�that�friction�in�the�valve�train�is�notknown�and�was�not�taken�into�account�in�the�calculation.
In�conclusion,�we� found�that� the� residual�gas�content,the�ratio�of�trapped�to�delivered�charge�and�the�pressuretrace�can�be�calculated�accurately�and�even�the�rotationof�the�valve�is�close�to�measured�results.�Thus�we�havea� tool� for�design�optimization�of� the�whole�scavengingcycle,�helping�to�reduce�design�and�testing�times.��
� MORE�INFORMATION:��http://www.wartsila.com
Sulzer�two�stroke�diesel�engines�are�used�as�"prime�movers"�in�container�ships,bulk�carriers�and�tankers.�Such�ships�need�to�be�in�service�for�as�long�aspossible�without�stopping�for�a�pre-scheduled�overhaul,�placing�a�continuousdemand�on�engine�developers�in�terms�of�reducing�costs�and�at�the�same�timeimproving�reliability.�CFD�is�a�proven�tool�for�minimizing�experimental�effort�duringthe�optimization�of�receivers,�blowers,�in-cylinder�flow�and�lubrication�systems.
Simulation of Scavenging &Exhaust Valve RotationDr.-Ing. Reiner Schulz, Wartsila Switzerland Ltd.
..::FEATURE ARTICLE Ship�Design
MARINE�REPORT 37
� Fig:03Close up view of exhaust valve, showing velocity vectors.
� Fig:02Close-up view of inlet ports, showing velocity vectors.
Fig:01 �Visualization of interface between burnt and unburnt gas at different times during scavenging.
� Fig:04 Comparison of experimental and calculated results.
PRESSURE ROTATION
CRANK ANGLE [deg]
MARINE�REPORT
..::REGULARS Yacht�Design
The�America's�Cup,�dating� from�1851,is�the�oldest�trophy�in�international�sportand� is� considered� yacht� racing's� HolyGrail.�In�2000,�a�German�challenger�startedto� work� towards� a� future� challenge,� calledPinta�Racing.� An� international� team�of� expertsfor� hull� form� design,� sailing� aerodynamics,� andfurther� technical� resources� was� brought� togetherto� support� an� outstanding� crew� of� internationallyrenowned�sailors�to�challenge�the�cup�holder.
Bettar el Moctar, Jochen Marzi, Arndt Schumacher, Hamburg Ship Model Basin (HSVA).
CFD for America's CupChallengers
38
In� the�80s�and�90s,�Willi� Illbruck� left� his�mark�on� the� yachting� scene.�His
yachts�called�‘Pinta’�won�the�AdmiralÅfs�Cup�twice�(1983,�1993).�Among�his
many�successes�were�also�the�two�One-Ton�Cup�victories�in�1993�and�1994
and�winning�the�Sardinia�Cup�in�1984.�Willi�Illbruck�was�the�first�German�ocean
yachtsman�and�owner�who� committed�himself� to� professional� yachting�and
supported�this�sport.�His�son�Michael�has�continued�his�father’s�dedication
over�the�past�years.�In�2002,�the�‘Pinta’�won�the�Volvo�Ocean�Race,�the�most
important�blue�water�regatta.
� FACTS�
� The Volvo Ocean 60-Hightech-Racer "illbruck", which won the Volvo Ocean Race Round the World 2001-2002.
Photo:�Daniel�Forster/illbruck�Challenge
� HSVA is a central point forapplied research in all areasrelated to transport systems andship technologies in open waterand ice. As an international, independentcompany, HSVA offersadvanced technologies andmodern test facilities. Customerorientated services have alwaysbeen the solid foundation of ourbusiness activities.
..::REGULARS Yacht�Design
MARINE�REPORT
The�Hamburg�ship�model�basin�(HSVA)�was�asked�to�join�theexpert�team�and�perform�the�hydrodynamic�assessment�of�thenew� ship� hull.� The� main� activities� of� HSVA� comprised� CFDcomputations� and�naturally� experimental� tests� in� their� large
towing�tank.�The�CFD�analysis�was�based�on�potential�flow�results�usingthe� in-house�n-SHALLO�code�and�the�RANS�code�from�CD-adapco.�Theuse� of� a� parallel� Linux-PC-cluster,� enabled� a� large� number� ofcomputations� to� be� performed� for� the� different� sailing� conditions� thatneed�to�be�considered�to�assess�the�quality�of�a�new�design.
Validation�experiments�have�been�performed�in�the�large�towing�tank�for`the�most�promising�candidates.�Here,�the�sophisticated�planar�motioncarriage� (CPMC)� is� used� to� control� different� operational� conditionsincluding�all�trim,�heel�and�yaw�variations�that�need�to�be�modeled.�Figure1�shows�the�experimental�set-up�during�measurements.
The� numerical� grid� for� the� RANS� computations� consisted� of� 3�millionhexahedral�cells.�Rudder,�keel,�bulb�and�winglets�were�considered�in�thecomputation.�The�k-ε model�was�used�to�account�for�turbulence�effects,while� free-surface�effects�were�modeled�by�an� interface-capturing� (VOF-like)�method�using�the�HRIC�(high-resolution�interface-capturing)�scheme.For�the�momentum�equations,�convective�fluxes�were�discretized�using�acentral� differencing� scheme.� The� unsteady� simulation� time� step� waschosen�such�that�the�Courant�number�of�1.0�was�not�exceeded.
Each�simulation�was�performed�using�six�processors�from�our�Linux-PCcluster.� As� shown� in� Figure�2,� the� difference� between� computed� andmeasured� forces� for� different� designs� (Design� 1� and� Design� 2)� atdifferent�sailing�conditions�(0°�and�27°�heel�angle)�was,�in�most�cases,less�than�five�per�cent.�The�RANS�computations�for�the�different�designsalways� predicted� the� same� trends� as� found� in� experiments.� Thesimulations�have�thus�proven�their�ability�to�account�for�relative�changesin�resistance�when�design�or�sailing�conditions�are�changed,�almost�asreliably� as� experiments.� This� allows� a� reduction� in� the� number� ofexperiments�needed�for�validation�purposes,�since�these�can�be�limitedto�the�optimum�design�found�by�simulation.�Figure�3�shows�the�pressuredistribution�on�the�yacht�body,�with�details�of�keel�and�winglets.�Figure4�presents�cross-flow�velocity�vectors�in�a�cross-section�through�the�keelfor� a� heel� angle� of� 27°,� showing� a� complex� flow� structure.� Figure� 5shows� free-surface�deformation�around� the�yacht� for� the�case�of�27°heel�angle,�with�an�asymmetric�wave�pattern.�Red�areas�indicate�highand�blue�areas�low�water�level.�Finally,�Figure�6�shows�the�yacht’s�wettedsurface�as�seen�from�the�front�for�the�27°�heel�angle�condition.
These� results�demonstrate� the�suitability�of�RANS� in�CD-adapco’s�CFDsolver�for�the�analysis�of�flow�around�yachts�under�sailing�conditions�asa�complement�to�experimental�testing.�While�in�the�early�design�phase,potential�flow�solvers�are�still�the�only�viable�simulation�tool�due�to�theirefficiency�(a�few�minutes�of�computing�time)�and�the�possibility�of�running
hundreds� of� test� cases� for� various� design� options.� RANS� solvers� areneeded�to�accurately�account�for�the�effects�of�turbulence�and�high�free-surface�deformation,�including�breaking�waves.
Coupled�analysis�of�wind�flow�around�sails�and�water�flow�around�hull�andappendages�in�sea�waves�will�be�performed�to�model�the�interaction�ofall�yacht�components�affecting�its�performance.��
�� MORE�INFORMATION� http://www.americascup.com http://www.hsva.de/
39
Fig:04 �Velocity vectors in a cross-section through keel normal to sailing direction.
� Fig:05Free-surface deformation around yacht (Design 1) at 27° heel angle.
� Fig:03Pressure distribution on yacht body including rudder and keel with winglets.
� Fig:02Comparison of computed and measured forces for different designs at different sailing conditions.
� Fig:01Experimental set-up in towing tank.
� Fig:06Wetted hull area (red) for the yacht hull at 27° heel angle, viewed from front, indicating wave breaking and air entrainment on the left-hand side.
Dedicated�to�delivering�the�highest�standards�in�both�performanceand�quality,�Azimut-Benetti�have�adopted�STAR-CCM+�as�a�standardCFD� tool� for� the�aerodynamic,�hydrodynamic,�and� thermodynamicdesign� of� their� yachts.� Using� STAR-CCM+� Azimut-Benetti� have
managed� to�optimize� the�performance�of� their� yachts,�while�simultaneouslyreducing�development�costs�and�time�to�market.�
Azimut-Benetti� conducts� structural� and� hydrodynamic� research� anddevelopment�at�a�dedicated�facility�in�Varazze�(SV),�where�innovative�solutionsare�developed�and�later�deployed�across�the�entire�Azimut-Benetti�group,�which
includes�the�brands�Azimut�Yachts,�Atlantis�and�Benetti�Yachts�-�each�aimedat� a� distinct� segment� of� the� boating�market� -� as� well� as� Fraser� Yachts,� aleading�brand�in�the�service�sector.�
Within� the�Azimut-Benetti�group�Computational�Fluid�Dynamics�simulation� isdeployed�for�many�purposes:�• internal environment conditioning; • aerodynamics; • thermal analysis of engine rooms; • exhaust gas flow analysis; • resistance calculation for fixed trim hulls; • calculating the hydrodynamic characteristics, including flow resistance, ofplaning hulls, with fixed and free trim (2 degrees of freedom).
Of� these� calculations� the�most� complex� and� valuable� is� the� hydrodynamicanalysis� of� planing� hulls.� It� is� necessary� to� consider� the� hull� as� a� three-dimensional� body� that� interacts� with� two� immiscible� fluids� (air� and� water),simultaneously�accounting�for�the�manner�in�which�the�fluids�interact�with�eachother� at� the� free-surface� and� for� the� way� in� which� the� trim� of� the� boat� isinfluenced� through� the� interaction� with� both� mediums.� Simple� staticsimulations� (in� which� the� boat’s� running� attitude� is� prescribed)� are� notgenerally�sufficient�for�our�purpose.�
� The Benetti shipyard, based in Viareggio, had been owned by theBenetti family since its foundation in 1873 and in 1985 was takenover by Azimut. The yard had earned an exceptional reputation forbuilding boats in classical lines, fitted out with luxury interiors.
Unsurpassed in producing top class boats based on traditionalcraftsmanship, Benetti were revolutionary on account of being thefirst to introduce and develop the modern concept of “motoryachts”, quoting one example: the “delfino” line.
Simone Bruckner, Francesco Serra, Azimut-Benetti, R&D Centre in Varazze (SV).
The Azimut-Benetti R&DCentre & the CFD Designof Luxury YachtsAzimut-Benetti�has�been�at�the�top�of�theworld�"mega-yacht"�industry�for�over�tenyears:�owning�over�10%�of�the�global�market,and�nearly�30%�of�the�domestic�market;�the�Italian�company�is�the�world�boatingindustry’s�largest�private�group,�producingover�€800�million�worth�of�yachts�every�year.�
..::FEATURE ARTICLE Yacht�Design
MARINE�REPORT40
Similarly,�although�CFD�simulations�with�a�single-phase�fluidmodel�have�been�used�elsewhere�to�provide�information�onthe�resistance�of�immersed�parts,�this�does�not�represent�anacceptable�approximation�for�solving�our�issues.�
Historically�the�process�of�hydrodynamic�design�relied�heavilyon�the�analytical�data�provided�by�multiple�towing�tank�tests,which�are�both�expensive�and�time�consuming�to�perform:�asingle� tank� test� typically� costs� many� thousands� of� Euros.Nowadays�we�adopt�a�“Virtual�Prototyping”�approach�in�whichwe� try� to� reduce� physical� hydrodynamics� testing� to� just� asingle�model� in�a� tank;� validating� the�design� that�we�havefound�to�be�the�best�following�numerous�"virtual"�tank�testsusing�STAR-CCM+.
The�Azimut-Benetti�Group�develops�and�produces�about�fivenew�yachts�a�year.�Using�STAR-CCM+�we�can�carry�out�verymany� tests� on� different� hulls� configurations� under� a� widevariety�of�operating�conditions.�We�can�compare�numeroustests�even�with�designs�that�are�not�finalized.�Thus�we�areable�to�virtually�optimize�the�hull� in�all� its�details.�The�finalproduction� vessel� is� therefore� fully� optimized,� reducingexperimentation� costs�and� time� to�market�and�providing�abetter�overall�product.
In� 2007� Azimut-Benetti� developed,� in� partnership� with�CD-adapco,�a�proprietary�method�that�uses�STAR-CCM+�as�afluid� dynamic� simulation� "engine".� This�method� calculatesthe�running�attitude�of� the�vessel� in�2�degrees�of� freedom("sink"�and�"trim").� � In�order� to� represent� the� full�complexgeometry�of�the�hull,�and�to�provide�appropriate�resolution�atthe� free-surface� waterline,� Azimut-Benetti� opted� for� the
trimmed� cell� calculation� method� available� in� STAR-CCM+.�The�method� prescribes� a� Cartesian� base� grid�with� perfecthexahedrons� that� are� subsequently� “trimmed”� to� the� CADsurfaces.�A�very�important�detail�is�the�treatment�of�the�layerof�cells�near�the�solid�walls:�layers�of�prismatic�cells�are�usedthat�automatically�fit�the�CAD�geometry.�
Figure� 1� illustrates� the� emerged� part� of� a� yacht� and� therelevant�results�from�aerodynamic�calculation.�In�this�case,the� use� of� the� surface� wrapper� outlined� above� wasfundamental�-�the�task�of�the�wrapper�is�to�produce�a�surfacetotally�free�of�connectivity�defects,�which�can�then�be�used�asan� input� datum� for� the� volume� mesher.� In� the� case� ofexternal� aerodynamics,� given� the� large� quantity� of� objectspresent�in�the�original�CAD,�the�surface�wrapper�has�allowedentire� days� to� be� saved� that� would� otherwise� have� beendedicated�to�an�activity�(CAD�cleaning)�with�no�added�valuefor�an�R&D�Centre.�
Figure�2,�finally,�illustrates�the�calculation�grid�for�the�externalaerodynamics.� In� this� case,� Azimut-Benetti� has� opted� forpolyhedral� grids� which� guarantee� automatic� setting� butperform�better�than�the�tetrahedrons.�
Azimut-Benetti�maintains� its�market� leadership�thanks�to�atechnological� leadership� such� as,� for� example,� the� use� ofaccurate� and� dedicated� CFD� software.� This� is� put� intopractice�by�working�alongside�key�players�on�the�CAE�scenesuch�as�CD-adapco�and�by�making�significant�contributionsto� improving�the�products,�with�a� long�term�spin-off� for� theentire�nautical�sector.��
�� MORE�INFORMATION�ON�AZIMUT-BENNETI�VISIT: http://www.azimutyachts.com
Global demand for private luxury
yachts has been increasing steadily
over the last five years. Marinas are
at saturation point in many parts of
the Mediterranean, as well as the
Caribbean. This growth is due to
the largest demand for leisure
boats ever seen with recreational
boat ownership at $25 billion.
International industry and sales
are growing annually at 5-10%.
According to Showboats
International, a publication that
tracks vessel construction, in 1993
the world had fewer than 700
private owners with boats over 24
metres (79 feet). Today, there are
an estimated 7000 yachts over
24 metres (79 feet) in use.
� FACTS�
..::FEATURE ARTICLE Yacht�Design
MARINE�REPORT 41
� Fig:01Results from aerodynamic calculation: pressure distribution on yacht surface and streamlines.
� Fig:02A view of the polyhedral mesh along the symmetry plane.
Before�purchasing�CFD�software�Azimut-Benetti�undertook�anexhaustive�evaluation�process.��For�hydrodynamic�simulation,�wechose�STAR-CCM+�because�of�its�ability�to:
1. Perform bi-phase calculations in air and water
2. Automatically construct a calculation grid that adapts to the physics of the problem
3. Calculate the equilibrium position of a planing hull.
For�aerodynamic�simulation,�we�needed�to�be�able�to�workeffectively�with�CAD�data�to�extract�a�single�wetted�surface,�oftenstarting�from�a�CAD�that�contains�superfluous�details�thatinterfere�with�the�fluid�volume�meshing�software;�a�finalrequirement�therefore�is:�
4. The ability to automatically clean CAD surfaces (surface wrapper).
Avoiding�vibrations�requires�finding�out�what�causes�them.�In�mostcases�the�power�train�is�responsible,�but�flow�unsteadiness�(usuallyvortex�shedding)�can�also�cause�fluctuating�forces�on�the�structureand� this� may� be� difficult� to� pinpoint.� The� problem� that
Germanischer� Lloyd� was� able� to� solve� belonged� to� the� latter� class:� thevibrations� reported� by� the� yacht� owner� and� confirmed� in� on-boardmeasurements� (a� one-node� torsional� hull� mode� at� 3.6� Hz)� could� not� beattributed� to� any� of� the� usual� sources.� � In� collaboration�with�CD-adapco,� adetailed�analysis�of�unsteady�flow�around�the�yacht�was�performed,�which�wasexpected�to�provide�information�on�where�the�excitation�for�those�vibrations�iscoming�from.
The�problem�with�simulations�is�that�one�needs�to�make�compromises,�andthis�is�where�the�know-how�of�experts�is�desirable.�Anyone�can�get�the�right
answer� if� a� suitable� software� is� used,� exact� physics� model� and� boundaryconditions�are�applied,�and�the�grid�and�time�steps�are�fine�enough�to�resolveall�flow�details.�However,�the�complex�geometry�of�a�modern�yacht�(above�andunder�water�surface,�with�all�appendages)�and�the�complex�physics�(turbulentflow,� free� surface�with� breaking�waves,�wind,� fluid-structure� interaction�etc.)makes�it�impossible�to�take�everything�into�account;�one�has�to�choose�whatto�ignore,�what�to�account�for�via�a�simplified�model,�and�where�to�apply�thebest�possible�analysis�approach.
The�critical�oscillations�were�reported�at�a�cruising�speed�between�16�and18� knots.�We� assumed� that� free-surface� effects�were� not� responsible� forvibrations�and�therefore�decided�to�study�the�flow�around�a�double-hull�withsymmetry�at�undisturbed�free�surface.�The�grid�was�locally�refined�in�regionswhich�were�expected�to�give�rise�to�flow�unsteadiness:�bow-thruster�tunnel,exhaust�gas�outlets,�skeg,�shafts,�V-�and�I-brackets.�The�mesh�had�in�the�endabout�6�million�cells.�Unsteady�RANS-simulations�were�performed�using�k-ωSST� turbulence�model� and�a� time�step� such� that� the�Courant� number� onaverage� was� of� the� order� of� unity.� The� simulation� was� performed� in� thereference�system�attached�to�the�hull;�inlet�velocity�was�set�to�yacht�speed(16�kn),�and�the�initial�velocity�was�also�set�equal�to�this�value.�The�flow�was
Bettar El Moctar and Tobias Zorn, Germanischer Lloyd AG, Hamburg.Milovan Perić, Tobias Zorn, CD-adapco, Nürnberg Office.
CFD Analysis HelpsSolve Yacht VibrationProblemGermanischer�Lloyd�was�recently�asked�tosolve�a�problem�related�to�torsional�vibrationson�board�a�mega�yacht.�
..::FEATURE ARTICLE Yacht�Design
MARINE�REPORT42
Fig:01 �Geometry of the mega-yacht which experienced rare flow-induced vibrations; different colors indicate regions for which flow-inducedforces were recorded as a function of time.
computed�over�a�period�of�about�one�minute�of�real�time,�andthe� forces� exerted� by� the� flow� on� various� parts� of� hull� andappendages�have�been�recorded�as�a�function�of�time.
The� analysis� of� the� results� showed� fluctuations� of� forces� inseveral� regions,� but� only� in� the� propeller� tunnel� were� themagnitudes�of�forces�and�the�frequency�of�their�oscillation�of�theorder� that� could� cause� the� problematic� torsional� vibrations.� Adetailed�analysis�of�flow�features�in�this�region�revealed�vortexshedding�in�the�gap�between�propeller�shaft�and�hull.�Figure�2shows� instantaneous� velocity� magnitude� in� one� horizontalsection� through� the� propeller� tunnel� and� in� hull� vicinity.� Thevortex-structures�can�be�clearly�identified.�The�frequency�of�forcefluctuations�on�various�parts�affected�by�the�flow�in�this�regionturned�out�to�be�around�3.8�Hz,�which�was�close�to�the�vibrationfrequency�found�in�measurements�on�board.�
We� did� not� expect� to� obtain� the� exact� answer� from� thesimulation;�for�a�quantitative�agreement�with�measurements,�aLES-type� simulation� and� a� much� finer� mesh� (including� yachtmotion� and� free-surface� effects)� would� be� needed.� However,there�is�no�need�to�predict�exactly�something�that�one�wants�toavoid:�the�simulation�pinpointed�the�source�of�trouble�and�nowthe�designers�had�a�chance�to�suggest�changes�that�would�avoidvortex� shedding.� One� had� to� consider� that� the� yacht� alreadyexisted,�so�there�were�limitations�to�the�extent�of�possible�hullmodification.� Fortunately,� a� relatively� simple� modification� wasboth� easy� to� realize� and� efficient.� However,� before� anymodification�on�the�yacht�was�done,�the�effect�of�the�proposedchange�was�first�investigated�in�another�simulation.
The� modification� involved� insertion� of� streamlined� fairingsbetween�the�shaft�and�hull�over�an�extended�length,�comparedwith� the�original� design� that� had� gaps� in� this� region� (see� red
parts�in�Figure�4).�Only�one�design�was�tested�numerically,�andit�was�a�full�success:�the�earlier�observed�vortex�shedding�andforce�fluctuations�almost�ceased,�and�the�velocity�field�in�thesame�area� shown� in� Figure�2�was�now�much� smoother,� seeFigure�3.�
With�simulations�suggesting�that�excitation�to�vibration�will�begone,� the� fairings� were� installed� on� the� yacht� andmeasurements�of�pressure�and�accelerations�were�performedat�sea.�The� result�was�positive:� the�problematic�vibrations� inthe� owner's� cabin� that� plagued� him� before� could� not� bedetected!
In�this�study,�experimentators�and�structural�experts�were�alsoinvolved.� The� FE-analysis� of� vibration� sensitivity� by� applyingharmonic�unit�excitation�forces�at�various�locations�proved�thatoscillating� forces� predicted� by� flow� simulation� could� indeedcause� torsional� vibrations� of� the� kind� observed� on� board.Pressure� measurements� on� the� hull� in� operation� (at� severallocations� indicated� by� flow� simulation� to� be� critical)� alsoconfirmed� oscillatory� behavior,� thus� strengthening� theconfidence� that� the� source� of� the� problem� was� found.� Thishelped� to� avoid� repeating� flow� simulations� on� finer� meshes(which� would� otherwise� have� been� necessary,� in� order� toestablish�the�confidence�in�numerical�solutions)�and�to�reducethe�number�of� flow�simulations.�Taking� into�account�the�meshfineness�and�the�fact�that�simulations�have�to�run�over�tens�ofthousands�of�time�steps,�this�greatly�helped�to�achieve�the�goalin�the�shortest�possible�time.�
Engineers� often� have� to� combine� various� interdisciplinaryanalysis�approaches�to�solve�engineering�problems,�and�thisstudy� was� a� good� example� of� how� CFD� can� help� in� thisprocess.��
�� MORE�INFORMATION�ON�GL�VISIT: http://www.gl-group.com
..::FEATURE ARTICLE Yacht�Design
MARINE�REPORT 43
� Germanischer Lloyd is at theforefront whenever it comes totesting, researching and continuallyimproving safety-relevant factors.For about 140 years GermanischerLloyd (GL) has been offering itsservices to the shipping industryand setting standards in technology,safety and quality. Over the years,the spectrum of services providedhas steadily broadened so that GLhas ceased to be a pure ship classification society and is now aglobally operating technicalmonitoring group. The fields itcovers range from classicalmaritime requirements such asoperational soundness analysis and anti-corrosion advice, tooffshore and onshore installationsto facility services.
� Fig:04Modified design of the hull-shaft junction, with prolonged, streamlined fairings (red).
� Fig:03 (AFTER)Same information for the modified design of hull-shaft junction - almost steady flow.
� Fig:02 (BEFORE)Instantaneous distribution of velocity magnitude near hull surface (left) and in a horizontal section above shaft (right) for the original design of hull-shaft junction.
..::FEATURE ARTICLE Yacht�Design
MMaarriinnee rreeppoorrtt44
We�run�all�hydrodynamic�and�aerodynamic�simulations�separately.Hydrodynamic�simulations�are�used�to� research�hull� form,�yachtbehavior�in�waves,�and�appendage�shape�and�position.�Early�in�thedesign�process,�aerodynamic�simulations�are�used�to�determine
the�sail�forces�which�are�in�turn�used�as�input�in�hydrodynamic�simulations.Later�we�return�to�aerodynamic�simulations�in�order�to�optimize�sail�shapesand�investigate�new�sail�concepts.�
Our� hydrodynamic� cases� are� run� using� full� size� hull�models�with� rudders,keels�and�foils.�The�free�surface� is�modeled�using�a�volume�fraction�(VoF)method,� and� simulations� allow� for� dynamic� trim� and� sinkage.� Ourcomputational� models� consist� of� unstructured� hexahedral� meshes� with
extensive� local� refinement� and� roughly� two�million� cells.� Appendages� aremeshed� independently,� giving� us� the� freedom� to� test� many� differentappendage�shapes�and�sizes�quite�easily.�
In�aerodynamic�simulations�we�model�all� geometry�above� the�static�waterplane,�including�the�sails,�mast,�boom,�deck�and�hull.�The�rig�is�tilted�to�thecorrect� attitude� accurately� modeling� heel,� pitch� and� yaw.� A� varying� windprofile� is�used� to�account� for� the�boundary� layer�along� the�surface�of� thewater.�The�aerodynamic�computational�models�use�automatically�generatedpolyhedral� meshes� with� prism� layers.� The� aerodynamic� meshes� have� asimilar�cell�count�but�many�more�nodes�than�the�hydrodynamic�meshes.�
At�Cape�Horn�Engineering�we�have�forgone�traditional�towing�tank�and�wind�tunnel�tests�infavor�of�an�exclusively�CFD-based�design�philosophy.�Using�STAR-CCM+�simulations�arecheaper,�faster�and�more�reliable�than�traditional�tests.�We�run�all�simulations�at�full�scalewhich�eliminates�the�inherent�error�in�scaled�test�results.�Enhanced�flow�visualization�andforce�decomposition�give�designers�much�greater�understanding�of�flow�phenomena.�
Volvo Ocean Racing:Using CFD for OptimalBoat Design Rodrigo Azcueta, Cape Horn Engineering.
Photo:�Oskar�Kihlborg�
� Ericsson Racing Team.
Cape�Horn�Engineering�performed�all�the�CFD�work
for� the�design�of� the� two�Ericsson�Racing�Team
boats� that�will� start� as� favorites� in� the� 2008/9
Volvo�Ocean�Race.
These�boats�were�designed�by�Juan�Kouyoumdjian
(www.juanyachtdesign.com)� based� on� the
simulations�performed�at�Cape�Horn�Engineering,
continuing� a� successful� partnership� that� started
with�the�design�of�the�ABN�Amro�boat�that�won�the
overall� race�victory� in�the�previous�edition�of� the
Volvo�Ocean�Race.
http://www.cape-horn-eng.com/references.html
� FACTS�
� MORE�INFORMATION http://www.volvooceanrace.org http://www.cape-horn-eng.com
..::FEATURE ARTICLE Yacht�Design
MARINE�REPORT 45
For�sail�analysis�we�use�parametric�modeling�and�fluid-structure-interaction(FSI)�codes.�In�parametric�models�we�use�STAR-CCM+�to�find�the�optimalaerodynamic� sail� shape.� Designers� then� develop� a� sail� with� theappropriate� structural� elements� in� order� to� achieve� this� optimal� 'flying'shape.�Parametric�variation�is�done�either�by�using�a�predefined�matrix�ofvariations,�or�by�incorporating�an�optimizing�algorithm�in�an�iterative�loop.In� FSI� models,� STAR-CCM+� code� passes� pressure� forces� into� a� finiteelement�model�which�then�calculates�the�deformed�shape�of�the�sail.�Thenew�deformed�shape�is�trimmed�by�the�sail�designer�and�put�back�into�theCFD�simulation.�This�cycle�is�repeated�four�to�five�times�until�convergenceis�reached.�
All�of�our�simulations�are�processed� in�parallel�on� four�cores;�with�436processing�cores�in�our�cluster�we�can�run�over�100�cases�concurrently.Each�case�is�part�of�a�matrix�of�predefined�sailing�conditions.�In�generalhydrodynamic�cases�are�run�as�unsteady�simulations�to�allow�free-surfacewaves� to� develop.� These� take� between� 12� and� 24� hours� to� complete.Alternately,�most� of� our� aerodynamic� cases� can�be� run�as� steady-statesimulations�which�converge�quickly�because�there�is�very�little�separationon�the�sails�of�high-performance�racing�boats.�These�sail�cases�generallyrun�in�only�a�few�hours.�Combined,�this�means�we�are�capable�of�runningseveral�hundred�simulations�per�day.�
In�hull�shape�studies�we�use�fully�appended�models�with�a�reference�setof�rudders,�keels�and�foils.�The�position�of�each�appendage�shifts�to�matchthe�shape�of�each�candidate�hull.�For�example,� the� rudders�will�changeangle�slightly�to�remain�perpendicular�to�the�hull�surface�on�each�hull.�
For� appendage� studies� we� use� a� set� of� reference� hulls� and� changeappendage�concepts,�shapes,�positions�and�orientations.�Each�variation�isevaluated�by�comparing�the�resulting�forces�(drag,�side�force,�roll�and�yawmoments)�and�by�comparing� the� flow�characteristics�using�stream� linesand�other�visual�techniques.�
The� Volvo� 70�Class� boats� present� a� new,� very� complex,design�problem.�Compared�to�America's�Cup�Yachts�thereare� many� more� design� variables;� the� Volvo� 70's� aredesigned�to�a�box�rule�and�experience�sailing�conditionsfrom� all� over� the� world.� Boat� speed� ranges� from� 6knots�as�a�displacement�hull�to�30�knots�planing�andsurfing� down� waves.� The� boats� also� have� cantingkeels� and�water� ballast�which� drastically� changesthe�displacement�and�center�of�gravity.��All�of�thesevariables�lead�to�very�large�testing�matrices.�Here�CFDbecomes� very� attractive;� each� variable� usually� can� bechanged� simply� by� changing� a� number� and� running� thesimulation�again.�
Our� hull� shape� research� program� is� quite� extensive.� The� designspiral�begins�with�the�required�transverse�stability�and�waterline�beam.We�then�obtain�accurate�sail�force�coefficients�using�our�own�aerodynamicsimulations.�Different�sail�sets�are�tested�for�upwind�and�downwind�sailingin�light�and�heavy�conditions.�This�is�important,�especially�for�very�beamyboats,�because�the�sail�forces�can�have�a�large�effect�on�the�longitudinaltrim�which�changes�the�drag.�Instead�of�attempting�to�match�a�given�sailset�to�specific�sailing�conditions,�which�is�difficult�and�error�prone,�we�findthe�center�of�effort�for�the�sails�and�apply�a�force�vector�at�that�point�inthe�hydrodynamic�model.�Then�the�hydrodynamic�simulation�runs�so�thedrag�on�the�boat�matches�the�force�generated�by�the�sails.�
The�hull�shape�investigation�continues�with�studies�of�volume�distribution,prismatic�coefficient,�transom�width�and�immersion,�bow�fullness,�etc.�Hullshapes�are�organized�with�parent�hull�shapes�and�their�derivatives.�Thisallows� for� easy� analysis� of� trends� and� performance� drivers� and� finalselection�of�a�hull�shape.�After�the�final�hull�has�been�chosen�and�the�lineshave�been�sent�to�the�builder,�research�continues�with�appendages�andsails.� We� investigate� the� size,� shape,� and� position� of� the� keel,� bulb,
rudders� and� dagger� boards.� We� adjust� the� transverse� inclination,longitudinal�inclination�(sweep),�alignment�of�the�keel�cant�axis,�pitch�of�thebulb,�angle�of�attack�of�the�dagger�boards,�etc.�
Different�solutions�and�details�for�the�attachment�of�foils�to�the�hull�areinvestigated,� like� the� recess� or� 'bubble'� in� the� hull� where� the� keelattachment�(a�cylinder)�rotates,�and�the�fairings�between�hull�and�foils.�
Candidate�boats�are�then�run�through�race�simulations�to�determine�whichdesign� is� the� best.� VPP� is� used� to� analyze� trade-offs,� such� as� stabilityversus� drag,� and� to� determine� optimum� balance.� Our� Router� programsimulates�the�best�course�for�each�hull�using�statistical�weather�data�forrelevant� parts� of� the� world,� and� then� compares� the� time� needed� tocomplete�the�course�for�each�hull.�Thus,�the�overall�winner�of�the�race�isfound�in�probabilistic�terms.�Other�design�variations,�such�as�the�positionof� a� dagger� board,� are� more� straightforward� to� analyze� and� usually� itsuffices� to� compare� the� amount� of� drag� at� a� given� side� force� to� drawconclusions.�
In�many�cases� the�design�and�modification�of�appendages�and�sails� isgiven�a�reality�check�when�tested�on�the�trial�boat�in�real�sailing�conditions.During� this� on-the-water� training� period�we� are� given� valuable� feedbackfrom�the�real�world�that�keeps�us�motivated�and�focused.�
Finally,�a�set�of�seakeeping�simulations�are�performed�to�investigate�thedynamic�behavior�of�the�final�candidate�hulls.�These�simulations�ascertainthat�what�is�good�in�calm�water�does�not�become�detrimental�in�waves.Due�to�the�nature�of�these�time�consuming�simulations,�only�regular�wavesare� considered.� To� analyze� a� complete� sea� spectrum� would� require� astatistical�analysis�over�a�large�period�of�time,�i.e.�half�an�hour;�however,the� response� to� regular� waves� can� be� found� in� less� than� a�minute� ofsimulated�time.�Despite�this�restriction,�the�information�gathered�in�regularwaves�is�indicative�of�real�sea�state�performance.�Unlike�a�towing�tank,�noother�simplifications�are�necessary;�in�our�simulations,�any�wave�directionis� allowed,� the� boat� is� at� full� scale,� the� center� of�mass� is� in� the� rightposition�and�the�moments�of�inertia�will�be�the�actual�estimated�values.�
� Fig:01Predicted deformation of A3 sail (colored bydisplacement) as a result of FSI-simulation.
� Fig:02Computed free-surface deformation,streamlines in the mid-plane, and pressuredistribution on hull and sails.
The TeamsSo� far� there� are� 8� teams� entered.� Ericsson� (x2),� Puma,Telefonica� (x2),� Team� Russia,� Green� Team,� and� the� newlyanounced�Delta�Lloyd.
What is it?The� Volvo� Ocean� Race� is� the� ultimate� mix� of� world-class� sportingcompetition� and� on-the-edge� adventure,� a� unique� blend� of� onshoreglamour�with�offshore�drama�and�endurance.�It�takes�over�nine�months,covers�over�37,000�nautical�miles�of�the�globe's�most�treacherous�seas.Raced�over�11�legs,�the�2008-2009�race�will�visit�12�ports�around�theworld�and�incorporate�eight�inshore�races,�and�eight�'pro-am'�races.
Established�in�1973�as�The�Whitbread�Round�the�World�Race,�the�VolvoOcean�Race�is�undeniably�the�world's�premier�global�race�and�arguablythe�most� challenging� sporting� event� in� the�world.� It� is� the�ultimate� inhuman� endeavour,� and� it� will� always� demand� faith,� trust� and� respectamong�team-mates.�It�is�life�at�the�extreme�and�will�always�remain�so.
The�teams�comprise�professional�sportsmen�and�women�at�the�top�oftheir�game.�The�race�requires�their�utmost�skill,�physical�endurance�andcompetitive�spirit�as�they�race�from�continent�to�continent�in�an�easterlydirection�around�the�world.
InnovationInnovation�has�always�been�the�signature�of�the�Volvo�Ocean�Race�and�anew� design� of� boat� was� introduced� for� the� 2005-06� event.� Withdevelopment�in�mind,�a�forum�of�skippers,�boat�designers,�sail�and�rigmanufacturers,� syndicate� heads� and� representatives� from� the� VolvoOcean�Race�collaborated�to�produce�a�state-of-the-art�70'�monohull�raceboat,�the�Volvo�Open�70.
The�Volvo�Ocean�Race�2005-2006�was�the�most�successful�round�theworld� race�to�date,� firmly�establishing�the�event�as�one�of� the�world'spremier�sporting�events.�In�2008-09,�it�will�build�on�its�33-year�foundationby�creating�the�platform�to�find�the�most�consummate,�all-round�racingteam�the�world�has�ever�seen.�For�the�first�time�in�the�event's�history,�therace�will� look�at�visiting�new�ports�along�a�new�route�that�may�includeports�in�India�and�Asia.
Building�on�the�enormous�success�of�the�In-port�racing�in�2005-2006,the� in-port� racing� in� 2008-2009� will� again� bring� the� spectacularperformance�of�the�Volvo�Open�70�to�the�ports�and�with�points�countingtowards�20%�of�the�overall�result,�will�be�great�for�the�public,�sponsorsand� media� alike.� The� new� ultra-high-tech� Volvo� Open� 70� set� newbenchmarks� and� redefined� elite� ocean� high-performance� racing.Modifications�to�the�Volvo�Open�70�rule�will�refine�the�boat�to�ensure�it
Let’s�start�with�the�basics.�The�race�consists�of�11�ocean�Legs,�8�Inshore�Races�and�8Pro-Am�Races.�All�up,�the�boats�have�to�sail�about�39,000�miles�on�their�trip�around�theglobe.�The�race�was�originally�known�as�the�Whitbread�Round�the�World�Race�and�occurredfor�the�first�time�in�1973.�(http://livesaildie.com)
Volvo Ocean Racing: Facts
..::FEATURE ARTICLE Yacht�Design
MARINE�REPORT46
� Ericsson Racing Team.
...they live with verylittle sleep(maximum of six toseven hoursinterrupted sleep)and sometimes withwet clothes for twoweeks at an end.
“”
Photo:�Oskar�Kihlborg�
is�as�quick�and�dynamic�as�any�boat�previously�sailedin�the�race.
Volvo Ocean Race 2008 to 2009During�the�nine�months�of�the�2008-09�Volvo,�whichstarts� in� Alicante,� Spain� in� October� 2008� andconcludes�in�St�Petersburg,�Russia,�during�late�June2009,�the�teams�will�sail�over�37,000�nautical�milesof�the�world’s�most�treacherous�seas�via�Cape�Town,Kochi,� Singapore,� Qingdao,� around� Cape� Horn� toRio� de� Janeiro,� Boston,� Galway,� Goteborg� andStockholm.� Eight� boats� are� confirmed� as� entrants� inthe�2008-2009�race.�This�will�be�the�first�time�that�therace�will�pass�through�Asia,�with�second�in-port�race�ofthe�2008-2009�edition�to�be�held�in�Singapore.�
Volvo�Ocean�Race�has�moved�forward�with�its�plans�tobring�the�2008-09�event�to�India�and�it�now�looks�likelythat�the�major�south-western�port�of�Kochi�will�be�thestopover�on�the�new�race�route�through�the�Middle�Eastand�Asia.�Boston�will�be�the�only�North�American�port�ofthe�race.�The�2008-2009�race�will�be�tracked�using�aRace� Management� System� developed� by� the� UKcompany�Cybit's�maritime�division.
Key Words With Regard to the Design“Using�high-tech� to�go� through�hell”;� “Most� importantyacht�race�around�the�world”;�“Innovation�in�design,�lifeat� the� extreme”;� “Competition� between� designers”;“Boats�that�weigh�14�tons,�but�hardly�have�a�gram�toomuch”;� “Boats� made� to� overcome� huge� wave� crests(sometimes� up� to� 30� metres)� at� a� high� speed”;“Coming�down�a�wave,�the�prow�sometimes�pierces�the
water.� A� wall� of� water,� one� and� a� half� metres� high,almost�sweeps�you�away.�After�that�you�sit�in�ice�coldwater�up� to� the�hip…�(Tim�Kröger,�professional�sailor,Hamburg)”;� “The� boats� are� built� in� such� a� way� thatwaves�can�pass�over�them�with�only�the�superstructuresstopping�them.”�
Key Words on Sailors“Huge�physical�challenge,�sailors�need�up�to�6000�kcala�day”;�“they�live�with�very�little�sleep�(maximum�of�sixto�seven�hours�interrupted�sleep)�and�sometimes�withwet� clothes� for� two� weeks� at� an� end”;� “lonelinesssometimes� 2000� nautical� miles� away� from� land”;“every�ten�minutes�you�have�to�take�a�decision�what�youare�going�to�do�next,�but�you�won’t�know�whether�it�wasthe�right�one�until�much�later…”�
� FOR�UP-TO-DATE�INFORMATION�ON�THE�RACE:� http://www.volvooceanrace.org
� THE�CREW
Each�of�the�eight�entries�has�a�sailing�teamof�11�professional�crew,�and�the�racerequires�their�utmost�skills,�physicalendurance�and�competitive�spirit�as�theyrace�day�and�night�for�more�than�30�days�ata�time�on�some�of�the�legs.�They�will�eachtake�on�different�jobs�onboard�the�boat�andon�top�of�these�sailing�roles,�there�will�betwo�sailors�that�have�had�medical�training,as�well�as�a�sailmaker,�an�engineer�and�amedia�specialist.
During�the�race�the�crews�will�experience�lifeat�the�extreme:�no�fresh�food�is�takenonboard�so�they�live�off�freeze�dried�fare,they�will�experience�temperature�variationsfrom�-5�to�+40�degrees�Celsius�and�will�onlytake�one�change�of�clothes.�They�will�trusttheir�lives�to�the�boat�and�the�skipper�andexperience�hunger�and�sleep�deprivation.
� THE�HARDSHIPS
..::FEATURE ARTICLE Yacht�Design
MARINE�REPORT 47
Huge physicalchallenge, sailorsneed up to 6000 kcal a day.
“”
Mike SandersonABN AMRO ONE
� Ericsson Racing Team.
Photo:�Jon�Nash
� Abn-AMRO Team
� Cape Horn Engineering has experience inindustry-leading design projects, including theAmerica's Cup and Volvo Ocean Race. Results havemade us extremely confident in our numericalsimulations and show that our viscous ComputationalFluid Dynamics (CFD) models are superior alternativesto traditional testing. Our methods have been proventime and again in a wide variety of marine applicationsincluding sailing yachts, power boats, high speedvessels, planing craft, advanced and unconventionalhull platforms, and commercial cargo ships.http://www.cape-horn-eng.com
CFD� is� already� widely� used� as� an� engineering� toolwithin�the�maritime�industry.�Long�used�for�optimizinghull� designs� under� steady� cruising� conditions,� it� isalso�becoming� increasingly� important� for� predicting
the� complex� three-dimensional� phenomena� applicable� tomanoeuvring�conditions.�Used�effectively,�it�reduces�the�relianceon�expensive�towing�tank�tests�and�allows�the�investigation�of�awider�variety�of�more�radical�designs�than�would�otherwise�bepossible.
Simulating�the�flow�around�a�rowing�boat,�however,�presents�acomplex� challenge.� While� most� boats� are� propelled� at� aconstant�rate,�a�rowing�boat�moves�forward�under�the�rhythmicrowing� action� of� the� crew.� As� the� boat� accelerates� anddecelerates�through�each�successive�stroke,�both�the�positionof� the� boat� and� its� attitude� in� the� water� are� dynamicallymodified,�making�this�a�complex�problem�with�multiple-degreesof�freedom.�At�the�forefront�of�the�pioneering�work�in�this�fieldis�Filippi�Boats�(part�of�Filippi�Lido�shipyards)�and�partners.�
To�become�an�Olympic�Rowing�champion�you�need�two�qualities�in�fairmeasure:�grace�and�guts.�More�than�any�other�sport,�rowing�combinessheer�explosive�power�with�fine�technique.�With�winning�margins�measuredin�just�tenths�of�a�second,�Gold�Medals�are�traditionally�won�by�workingharder�and�suffering�more�than�your�opponents.�At�the�last�two�Olympics,however,�those�rules�have�changed;�blood,�sweat�and�tears�alone�are�goodenough.�Nowadays,�the�best�teams�have�another�weapon�in�their�armory.�At�the�Olympics,�Computational�Fluid�Dynamics�could�be�the�differencebetween�Gold�and�Bronze.
� Filippi Boats are an Italianmanufacturer of rowing racingshells. The company was founded in1980 by Filippi Lido. Today, therunning of the boatyard isundertaken by Filippi’s son David,the yard employs 50 technicians andproduces just over 700 boats eachyear which supply Federationsworldwide. In the previous 20 years crews inFilippi boats have achieved over 300medals in World Championships andOlympic Games.
..::FEATURE ARTICLE Olympics
MARINE�REPORT48
BreakingWaves at theOlympicsStephen Ferguson, Consultant Engineer, CD-adapco.
Filippi�WRC�Monaco�2007
Filippi�WRC�Monaco�2007
As�a�leading�manufacturer�of�high-quality,�race-standard,�rowing�boats,�they�have�been�using�CD-adapco’s�CFD�code�to�optimize�their�high-tech�designs�for�the�Summer�Olympics.�Workingtogether�with�a�prestigious�Italian�university,�Politecnico�di�Milano-MOX�(Milan),�their�aim�is�toprovide�enough�advantage�to�propel�their�oarsmen�to�the�top�of�the�Olympic�podium.
Using�CD-adapco’s�technology�they�have,�for�the�first�time,�been�able�simulate�the�influenceof�moving�rowers�on�the�boat�and�the�periodic�accelerations�caused�by�each�stroke�of�the�oarsand�thereby�the�time-dependant�changes�of�resistance�and�propulsion.�The�simulation�takesfull�account�of�both�squat�(aka�‘dynamic�sinkage’),�the�tendency�of�a�moving�boat�to�rise�outof�the�water,�and�trim,�its�tendency�to�pitch�in�the�water.�Through�these�simulations,�both�squatand�trim�were�shown�to�have�major�effects�on�resistance�experienced�by�the�boat.
This�simulation�was�only�possible�due�to� the�open�structure�of� the�CFD�solver� that�allowsusers�to�easily�add�their�own�routines�for�rigid�body�movement,�extending�the�simulation�of�allsix�degrees�of�freedom.
CFD�calculates� the�position�of� the�water�surface�around�the�boat� in�a� rapid,�accurate�andefficient�manner.�The�free�surface�between�the�water�and�the�air�is�captured�without�smearingusing�the�proprietary�High-Resolution�Interface-Capturing�scheme�(HRIC).�The�CFD�results�areused�to�get�an�in-depth�understanding�of�the�flow�field�around�the�race�boats�under�actualrace�conditions�-�something�impossible�in�a�scaled-down�towing�tank�test.�Although�a�racingrowing�boat�may�look,�to�the�layman,� like�a�simple�hull� form,� it� is�actually�quite�a�complexgeometry�and�a�difficult�task�to�model�and�optimize.
The�wrong�amount�of�trim�or�squat�in�adverse�conditions�may�allow�water�to�flow�over�the�sideof� the� boat� with� immediate� and� devastating� consequence.� This� point� was� all� too� clearlyillustrated�at�the�trial�regatta�for�the�Olympic�Rowing�Lake�in�2004,�which�had�to�be�abandonedafter�many�of�the�boats�sunk�in�choppy�conditions.�In�rowing,�‘taking�on�fluids’�is�a�real�risk.
Due� to� their� leading-edge�performance� requirements,�sports�applications�are�an� importantbenchmark�for�CFD.�The�application�by�Filippi�Boats�is�a�key�example�of�the�current�trend�inthe� marine� world� of� how� CD-adapco’s� software� and� services� can� perform� multi-fluid� six-degrees-of-freedom�simulations.��
� Fig:01Below surface pressure contours.
�� MORE�INFORMATION�ON�FILIPPI�BOATS: http://www.filippiboats.com
..::FEATURE ARTICLE Olympics
MARINE�REPORT 49
THE WORLD’S OLDEST RACE
Doggett's�Coat�and�Badge�was� first� contested� in�1715�and� is�held
annually� from� London� Bridge� to� Chelsea� and� is� believed� to� be� the
oldest�sporting�contest�in�existence!
Thomas�Doggett,� an� Irish� comedian�and� joint�manager� of� the�Drury
Lane�Theatre,�provided�in�his�will�dated�10th�September�1721,�for�a
prize� of� a� coat� and� silver� badge� to� be� rowed� for� annually� by� six
watermen�within�a�year�of�completing�their�apprenticeships.
"... Five Pounds for a Badge of Silver weighing about Twelve Ounces andrepresenting Liberty to be given to be rowed for by Six Young Watermenaccording to my Custom, Eighteen Shillings for Cloath for a Liverywhereon the said Badge is to be put, One Pound One Shilling for makingup the said Livery and Buttons and Appurtenances to it...............all which I would have to be continued yearly forever in Commemorationof His Majesty King Georges happy Accession to the Brittish Throne..."The�course�was�originally�four�and�half�miles�long�from�"The�Swan"�at
London�Bridge� to� "The�Swan"�at�Chelsea.� The�Barge�Master�of� the
Fishmongers'� Company� would� start� the� race� and� the� Clerk� of� the
Watermen�and�Lightermen's�Hall�would�receive�a�fee�of�thirty�shillings
from�each�competitor�(indicating�that�at�one�time�the�whole�event�was
arranged�by�that�company).�As�a�real�test�of�stay�and�endurance,�the
race�used�to�be�rowed�in�heavy�old�wherries�which�had�to�be�pulled�up
against�the�ebb�tide�-�sometimes�it�took�contestants�nearly�two�hours
to�row�the�distance!
The� race�soon�became�open� to�abuses�as�contestants� realized� the
advantages� of� using� lighter� undersized� vessels.� In� some� instances
riotous�behavior�was�reported�between�competitors.�In�1723,�one�of
the�leaders�of�the�race�had�his�“scull knocked away and a big boat rowedacross his bows".� As� a� result,� in� 1769,� the� Fishmongers'� Companydecided� to� draw� up� some� regulations� to� prevent� such� abuses.� All
vessels� had� to� be� "common� Scullers� Boats"� and� examined� by� the
company.�Originally,�the�six�watermen�were�drawn�by�lots�which�meant
that�not�all�contestants�had�a�fair�chance�of�winning.�Later�in�the�19th
century�a�trial�heat�system�at�Putney�was�introduced�to�select�the�final
best�six�men�for�the�race.������
Guildhall�Library�Manuscripts�Section
� FACTS�
� Fig:02Velocity defect in wake of boat.
“Finish�of�the�Race�for�Doggett's�Coat�&�Badge”�-�Thomas�Rowlandson�(1756-1827)
For�this�reason,�the�battle�for�Gold�begun�long�agoin� the� offices� and� testing� facilities� of� researchcentres,�where�the�sports�equipment�used�by�thecompetitors� in� the� Beijing� Olympics� has� been
constantly� optimized� and� improved.�Canoeing� is� one�of� thefields� where� extensive� research� has� been� performed� intominimizing�hull�resistance�using�state-of-the-art�measurementand� experimental� techniques,� backed� extensively� byComputational� Fluid� Dynamics� simulation� using� CD-adapcosoftware.
The�Polish�company,�Plastex�Composite,�recognized�worldwideas�one�of�the�leading�producers�of�competition�canoes,�hasprovided� state-of-the-art� equipment� for� Olympic� Games� andWorld�Championships�for�many�years.�In�2005,�during�the�1stWorld�Canoe�Championships�in�Poznań,�56�out�of�81�medals
were� won� by� competitors� using� Plastex� boats.� The� boatsthemselves� are� designed� by� the� company’s� owner,� RyszardSeruga,�in�cooperation�with�Tomasz�Bugalski,�Ph.D,�from�theShip�Design�and�Research�Centre�(CTO)�S.A.�-�Poland.
In�the�spring�of�2007,�Plastex�and�CTO�began�to�work�on�theshape� of� new� canoes� for� the� Olympic� Games� in� Beijing.Optimization� of� the� new� design� started� with� extensiveinvestigation�of�the�existing�hull�shapes�in�CTO’s�model�basin,with�the�principal�aim�of�determining�the�dependency�of�thecanoe’s� performance� on� the� basic� design� parameters� andinitial�trim.�Although�such�experiments�provide�a�large�amountof� reliable� data� in� a� short� timescale,� they� do� not� alwaysilluminate� the� physical� mechanisms� that� affect� theperformance�of�the�hull.�
For�most�of�the�competitors�at�the�2008�Beijing�Olympic�Games,�the�possibility�of�mountingthe�podium�to�claim�an�Olympic�medal�represents�the�very�pinnacle�of�sporting�achievement:usually�the�payoff�of�many�years�of�blood,�sweat�and�tears.�However,�in�2008,�being�the�bestathlete�is�not�longer�necessarily�enough:�in�most�events�the�Gold�Medal�winner�will�also�havehad�the�aid�of�the�very�best�sporting�equipment.
� Plastex Composite is producer andexporter of boats and paddles of thehighest quality for canoeing and rowing.They have a track history that far exceedsany of their competitors as far as sportingmedals go, their athletes have simply wonfar more using their products.
Since 1998 Plastex has begun, in cooperation with the Institute ofHydrodynamics, the research of moreeffective models of kayaks and canoesbased on the application of CFD.
..::FEATURE ARTICLE Olympics
MARINE�REPORT50
CTO & Plastex Paddle to Olympic Glory using CFDSimulation Tomasz Bugalski, Ph.D and Marek Kraskowski - Centrum Techniki Okrêtowej.
For� this� reason,� the�experimental� research�was�widely�supportedwith�extensive�CFD�simulation,�which�is�more�suited�to�a�detailedcomparison� of� the� influence� of� flow� properties� such� as� waveelevation�and�pressure�distribution�on�the�hull�for�different�designs.�
Each�CFD� simulation� considered� a� canoe� hull� towed�at� constantspeed� through� calm� water.� The� computational� analyses� yieldednumerical� data,� such� as� hull� resistance,� as�well� as� allowing� thedesign� team� to� visualize� the� flow� field� around� the� hulls,� therebyhelping�them�identify�the�mechanisms�behind�variations�in�physicalperformance,�e.g.�bow�and�stern�wave�height�or�wave�interaction.After�testing�the�existing�boats,�the�best�design�was�chosen�basedon� analysis� results� and� work� on� new� designs� began.� Byimplementing�CFD�into�the�design�process,�timescales�and�costshave�been�significantly� reduced.�The�viability�of�each�new�designwas� first� tested� numerically,� so� that� only� a� small� number� ofoptimized�designs�were�selected�for�manufacturing�and�testing�inthe�model�basin.�Final� tests�were�carried�out� in� real�conditions� -with�the�professional�competitor�rowing�along�the�basin.
The�simulations�were�carried�out�using�the�Volume�Of�Fluid�(VOF)model�for�multiphase�flows�and�the�RNG�k-ε turbulence�model�andspecially� constructed� 1.5� million� cell� hexahedral� meshes.� Thesurface�models� of� the� existing� canoes�were� obtained�by� digitallyscanning�the�hulls,�carried�out�using�an�ATOS�II�optical�scanner�anda�TRIPOD�photogrammetric�system�provided�by�GOM�GmbH.�
Due� to� the� fact� that� the�Olympic� canoes� travel� at� relatively� highspeed�(of�the�order�of�6m/s),�it�is�absolutely�necessary�to�take�intoaccount�the�dynamic�trim�and�sinkage�of�the�hull�in�the�numericalanalysis�of�the�flow�around�it,�requiring�either�experiment�data,�or�ifnot�available,�adjusting�the�hull�position�during�the�CFD�simulationuntil�force�and�moment�equilibrium�is�reached.�Although�this�can�bedone� iteratively,� based� on� the� hull� hydrostatics,� CTO� uses� an� in-house,�automated�procedure�for�coupling�the�flow�solver�to�the�hullmotion�equations,�allowing�for�accurate�evaluation�of�the�canoe’sposition.�The�computational�mesh�in�this�approach�remains�rigid,�itmoves�together�with�the�hull�without�relative�motion�of�the�nodes,which�proved�to�be�sufficiently�accurate,�robust�and�very�simple,�nore-meshing�is�required�when�the�hull�changes�its�position.
At� present,� the�CFD�simulations� and�model� tests� of� the� canoe’sperformance�are�limited�to�steady-state�analyses�-�the�hull�is�towedwith�constant�speed�and�fixed�centre�of�mass.�Such�a�simplifiedapproach� allowed� for� effective� optimization� of� the� hull� shapesbased�upon�resistance�with�an�identified�1%�reduction,�which�couldeasily� be� the� difference� between� Olympic� Glory� and� ignominiousdefeat.� The� use� of� CFD� methods� allowed� reduction� in� costs� bylimiting�the�number�of�designs�tested�and�so�reducing�the�need�tomanufacture�many�hull�shapes.�Further�to�this,�identification�of�theflow�phenomena�by�CFD�allowed�optimization�to�be�carried�out�farquicker�than�previously�possible.
It�is�very�likely�that�the�present�shapes�of�the�Olympic�canoes�arealready�very�close�to�the�absolute�minimum�resistance�obtainablein�steady�flow.�In�the�future,�significant�further�development�will�onlyoccur�by�optimizing�the�dynamic�behavior�of�the�hull,�which�meanstaking�into�account�all�the�phenomena�encountered�during�a�race,motion�of�the�competitor,�and�unsteady�forces�exerted�on�the�hull.For� that� reason,� the� next� step� for� CTO� is� to� study� 6� degrees� offreedom�(6-DOF)�analyses�to�help�aid�hull�optimization�and�to�try�toadjust�the�shape�so�as�to�minimize�the�loss�of�energy�due�to�thehull�motion� and� interaction�with� other� canoes.�CTO�have�alreadyperformed�a�trial�simulation�of�the�Wigley�hull�in�head�waves.�In�thissimulation,� the�hull�was� free� to�pitch�and�heave� (2-DOF�motion),while�sailing�with�constant�speed�and�fixed�zero�drift�angle�resultsrevealed�good�accuracy�and�robustness�of�the�method.��
..::FEATURE ARTICLE Olympics
� Fig:02The canoe hull in its static (left)) and dynamic (right)position in the flow.
� Fig:01K1 (top-left), K2 (top-right) and K4 (left) canoes:predicted free surface elevation.
MARINE�REPORT 51
� Fig:03Example of the hull dynamics analysis: Wigley hull in waves.
� Fig:05Example of the computational mesh.
� Fig:04Final tests were carried out with a professional rower.
� FACTS�
Olympic Classification:Kayak: K1�-�single�seat�kayak
K2�-�double�seated�kayakK4�-�4�seated�kayak
Canoe: C1�-�single�kneeling�canoeC2�-�double�kneeling�canoeC4�-�4�person�kneeling�canoe
Fig:02 �CD-adapco’s CFD solutionsare routinely used in thedesign of marine applicationsto understand how a unit willreact upon impact by a wave. This technology allowsengineers to optimize thehydrodynamic performance ofthe ship, FPSO or platformand understand the range ofconditions under which safeoperation can be assured.
The�scale�of�these�losses�combined�with�pressure�frominsurers,� has� led� to� a� rapid� re-evaluation� to� thetechniques� used� to� design� offshore� platforms.� Manyoperators� are� turning� towards� Computational� Fluid
Dynamics� in� order� to� provide� additional� insight� into� how� theirplatforms�perform�under�the�most�extreme�operating�conditions.
Computational� Fluid� Dynamics� (or� CFD)� is� a� technique� thatsimulates�fluid�flow�phenomena�using�super-computer�technology.Although�its�origin� is� in�the�aerospace�and�automotive� industries,CFD�is�increasingly�finding�application�in�many�areas�of�the�oil�andgas� industry.� CFD� can� be� used� to� simultaneously� simulate� theaerodynamic� effect� of� strong� winds� on� the� platform� with� thehydrodynamic�influence�of�waves�impacting�upon�it.
Although� CFD� technology� has� been� routinely� applied� in� manyindustries�since�the�early�eighties,�it�has�only�recently�begun�to�beseriously�used� in�offshore�platform�design.�Most�current�offshoreplatforms� were� designed� using� extensive� experimental� modeltesting.� Although� experimental� analysis� provides� considerableinsight� into� the� performance� of� a� particular� design,� physicalprototypes�are�expensive�and�time�consuming�to�construct.
Dr�Dennis�Nagy,�CD-adapco’s�Director� for� the�Oil� and�Gas� sectorexplains:� “It� isn’t� that�CFD� technology�wasn’t� available�when� thecurrent�generation�of�platforms�was�designed;�CFD�technology�hasbeen�routinely�applied�in�many�industries�since�the�early�eighties.�Itis� just� that� the� cost� of� performing� the� analysis� would,� until� veryrecently,�have�been�too�prohibitive.”
Nagy�feels�that�the�biggest�advantage�of�CFD�is�that�its�rapid�turn-around�time�helps�to�break�the�dependence�of�offshore�design�onpre-existing� design� codes.� Although� design� wind� and� waveconditions� are� a� useful� starting� condition� for� offshore� platformanalysis,� CFD� simulation� allows� designers� to�more� easily� pursuemultiple�“what�if?”�scenarios.�Once�a�CFD�model�for�a�platform�isset�up,�it� is�relatively�simple�to�repeat�the�calculation�for�multipleloading�scenarios.�“Instead�of�becoming�stuck�by�the�fact�that�thedesign� codes�don’t� deal�with�wave� heights� above�70� feet,� usingCFD,�designers�are�free�to�consider�the�impact�of�wave�heights�of
Current�design�standards�require�that�platforms�be�built�to�survive�so-called�100-year�storms,which�generate�wave�heights�of�up�to�about�70�feet.�However,�during�Hurricane�Ivan�peakwave�heights�of�over�90�ft�were�measured�(including�one�that�severely�damaged�the�ChevronPetronius�platform)�consistent�with�a�once�in�2500-year�storm.�The�problem�is�compoundedby�the�fact�that�many�of�the�4000�platforms�operating�in�the�Gulf�of�Mexico�were�designedbefore�1988,�when�the�current�100-year�design�standards�came�into�operation�(althoughsome�of�the�destroyed�platforms�were�of�recent�design).
Hurricane-ResistantOffshore Platform Design
MARINE�REPORT52
Stephen Ferguson, CD-adapco.
..::FEATURE ARTICLE Offshore
Courtesy�Germanischer�Lloyd
80,�90,�or�even�100�feet,”�says�Nagy.�“All�they�need�todo� is� input� the�new� condition�and�sit� back�while� thecomputer� does� the� number� crunching.� It� is� a� veryeffective�way�of�assessing�the�limit�of�your�design.”
Unlike�testing�of�physical�prototypes,�CFD�simulationsare� typically� carried� out� at� full� scale� (the� computermodel� has� the� same� dimensions� as� the� actualproduction� platform� rather� than� those� of� a� smallerexperimental� model).� This� has� the� considerableadvantage�that�results�can�be�interpreted�directly�anddo� not� have� to� undergo� scaling,� a� process� that� canintroduce� a� significant� uncertainty,� especially� fortransient�phenomena�such�as�the�impact�of�a�wave.
A�further�advantage�is�that,�instead�of�being�restrictedto� retrieving�data� from�a� few�experimental�monitoringprobes,�data�is�available�at�every�point�on�the�platform,at�every�discrete�time�interval�for�which�the�simulationis�performed.�The�wave� impact�on�a�platform�can�beviewed� from�any�angle,�and� the� instantaneous� forcesacting�on�any�part�of�the�structure�can�be�calculated.
Data�from�CFD�calculations�can�also�be�used�to�assistother�types�of�analysis,�for�example,�the�forces�actingon� a� platform� can� be� exported� to� a� stress-analysissoftware�package.�In�extreme�cases,�where�fluid�forcescause� large� deflections� of� components,� the� CFDsimulation� can� be� coupled� directly� with� the� stressanalysis�tool�and�both�stress�and�fluid�simulations�canbe�performed�simultaneously,�each�simulation�feedingnew�boundary�conditions�to�the�other.
In� Nagy’s� view,� the� adoption� of� CFD� technology� as� aroutine� part� of� offshore� design� is� inevitable.� “In� theautomotive� industry� almost� every� component� isdesigned�with�the�aid�of�CFD�technology,�to�bring�a�newproduct�to�market�without�it�would�be�unthinkable,”�hesays.� “The� financial� and�environmental� impact�of� therecent�hurricanes�means�that�the�oil�and�gas�industryhas�no�choice�but�to�follow�suit.” �
� MORE�INFORMATION�[email protected]
� Fig:03 Study of wave impact on offshore platform: free-surface shape at five instants during wave encounter.
� BENEFITS
Rapid turnaround time.
Easy�investigation�of�‘what-if’�scenarios.
Simulations are carried out at full scale.
Data�available�at�every�part�on�the�platformat�every�interval
Can be exported to a stress analysis package
..::FEATURE ARTICLE Offshore
MARINE�REPORT 53
The�first�oil�platform�in�the�world�was�built�in1947.�The�Oil�Rocks�platform�was�built�25miles�off�the�coast�of�Azerbaijan�in�theCaspian�Sea�and�is�a�functional�city�with�apopulation�of�5000�people.�There�are�over100�miles�of�paved�streets�on�this,�theworld's�largest�oil�platform.�There�are�shops,restaurants,�even�a�library.�With�all�workersliving�and�working�together,�this�engineeringmarvel�has�been�in�existence�for�60�years.
� FACTS�
� Fig:01CD-adapco software has been used to evaluate thewind and wave loading on platforms in stormconditions. These analyses allowed platformdesigners and operators to evaluate many differentplatform loading scenarios without the excessivecost of creating physical prototypes. This figureshows pressure on walls and velocity magnitude inone longitudinal section.
� MORE�INFORMATION�VISIT���http://www.delmarus.com
The�anchoring�of�a�drilling�rig�is�somewhat�more�sophisticated�thananchoring�a�boat.�Installing�these�units�is�equivalent�to�towing�aneight�story�building�out�into�the�middle�of�the�Gulf�of�Mexico,�tippingit�over�the�side�of�a�large�vessel,�and�burying�it�two�miles�down�into
the�sea�floor.
The� structure� is� constructed� from�massive� welded� sections� of� plate� steel.Internal�ribbing�and�mounting�flanges�are�designed�to�provide�the�necessaryrigidity�and�strength.�Not�surprisingly,�the�shear�weight�of�this�structure�makesits�transport�a�major�undertaking.
The� most� critical� of� these� is� the� “overboarding� event”,� when� the� pile� islaunched�overboard�using�a�sled�and�winch�system.�There�is�a�potential�fordamage� to� the� structure� during� deployment,� when� the� massive� weight� iscantilevered�off�the�ship.
Proper�design�of�the�sled�structure�and�cable�mounting�system�are�essential;damage�to�the�pile�during�this�phase�would�be�catastrophic.
Another� “event”� that� requires� investigation� is� the� sinking� of� the� structure.Suction� piles� drive� themselves� part�way� into� the� sea� floor� under� their� ownweight.�A�remotely�operated�vehicle�(ROV)�is�then�used�to�pump�the�seawaterfrom�the�tower,�creating�a�vacuum�that�draws�the�tower�further�into�the�seabed.During�this�process,�adequate�internal�ribbing�is�necessary�to�avoid�bucklingof�the�outer�shell.
All�aspects�of�the�installation�were�verified�with�Finite-Element�Analysis�(FEA).Analytical�models�were�constructed�in�pro-fe,�the�FEA�pre-�and�post-processorsupported�by� the�CD-adapco.�High-quality�hexahedral�mesh�was�used� for�allcomponents�of� the�pile,� to�ensure�good�stress�values�at� the�mount�points.Automatic�trimmed-cell�meshing�capability�of�pro-STAR�was�used�to�easily�fillthe�soil�volume�of�the�seabed�surrounding�the�anchor.
The�structural�analyses�included�G-loadings�during�transport�and�over-boarding.Particular�emphasis�was�placed�on�the�skid�support�structure,�as�well�as�thecable� mounting� system.� During� the� initial� design� phase,� the� analysisdetermined�that�the�winch�attachment�needed�to�be�redesigned�for�adequatemargin.�
Additional�analysis�addressed�pressure�loadings�during�installation,�when�theROV�creates�a�negative�pump�pressure�on�the�inside�of�the�pile.�The�seabedwas�modeled�as�a�bi-phasic�material,�using�soil�properties�developed�for�thatparticular� location.� Eigenvalue� buckling� calculations�were� also� performed� tocheck�susceptibility�during�installation.
The�anchor�was�successfully�installed�and�remains�at�its�installed�location�forapproximately�six�months�to�allow�the�soil�to�settle.��After�this�time�the�anchorbecomes�active�as�a�parking�anchor�for�mobile�drilling�units.��
When�Delmar�Systems�wanted�to�perform�a�design�validation�of�the�world’s�largest�suctionpile�anchor�system,�they�called�upon�the�expertise�of�CD-adapco.�These�anchors�are�usedto�“park”�and�anchor�large�offshore�structures�such�as�mobile�drilling�units.
Offshore Mooring
� Delmar Systems, Inc. is the world leader inoffshore mooring, providing the safest, mostefficient mooring solutions for the oil and gasindustry. Every Delmar employee isempowered with knowledge to use the mostsophisticated technology.Our success is measured only againstuncompromising expectations.www.delmarus.com
..::FEATURE ARTICLE Offshore
MARINE�REPORT54
55
Until�now,�the�common�practice�for�developing�outfall�structures�hasbeen�to�carry�out�very�time-�and�cost-consuming�physical�model�tests,but�in�future�numerical�simulations�could�help�to�reduce�this�overhead.An�outfall�structure�was�developed�by�the�University�of�Hannover�using
physical�model�tests�and�later�put�into�operation�this�year�in�the�Main-Danube-Canal� in� the� city� of� Bamberg� (Germany).� In� order� to� judge� the� capabilities� ofnumerical�simulations� in� this�application,� the�structure�has�also�been�studiedusing�STAR-CD.
Figure�1�shows� the�36m-wide�outfall� structure� including� features�such�as� thepressure� pipe,� overfall� weir� and� submerged� wall,� as� well� as� the� topographicsituation�200�metres�up-�and�downstream,�which�were�all� transposed� into�thenumerical�simulation.�A�mesh�of�about�80,000�cells�was�generated,�with�edgelengths�between�0.05m�and�3m.�The�water/air�free�surface�was�modeled�usingthe�VOF�method�and�the�standard�high�Reynolds-number�k-ε turbulence�modelwas�used.�
Figure�2�shows�the�weir�overfall�in�the�outfall�structure�for�the�initial�numericalsimulation.�It�is�abundantly�clear�that�the�well-balanced�weir�outfall�seen�in�themodel�tests�is�similarly�reproduced,�in�that�the�discrepancy�between�calculatedand� measured� overfall� height� never� exceeds� 6%.� A� similarly� low� divergencebetween�physical�model�test�and�numerical�simulation�can�be�recognized�whencomparing�the�calculated�and�measured�flow�field�within�and�in�the�vicinity�of�theoutfall�structure�at�half-water�depth.�This�is�illustrated�in�Figure�3,�which�showsparticularly�good�agreement�in�both�the�flow�velocities�and�directions�in�the�outfallstructure�itself.
For�similar�investigations�in�the�future,�the�use�of�more�numerical�simulations�isrecommended�in�order�to�reduce�the�amount�of�very�time-consuming�and�costlyphysical�model� tests.� Note,� however,� that� physical�model� tests� cannot� yet� bereplaced�completely,�since�although�the�current�simulation�methodology�makes�itpossible�to�develop�an�overfall�structure�from�a�purely�fluid-mechanical�point�ofview,�it�does�not�take�into�account�the�influence�of�ship�traffic�on�the�waterway.This� is�because� the� interaction�of�a�moving�object�and� its�environment� is�notpresently�implemented.�Research�is�under�way�to�close�this�gap.��
Transverse�flows�from�lateralwater�discharges�intowaterways,�for�example�thosecaused�by�storm�wateroutfalls,�may�cause�passingships�to�drift.�In�order�toprovide�traffic�safety,�sucheffects�need�to�be�restrictedso�that�outfall�structurescreate�well-balanced�flowfields�of�low�intensity.
Tobias Linke, University of Hannover.
Bamberg, Germany
Main-Danube-Canal
Overfall Weir
Submerged Wall
Lock
� Fig:02Calculated weir overfall in the outfall structure at the beginning.
Fig:01 �Outfall structure at the Main-Danube-Canal inthe City of Bamberg (Germany).
CFD Improves Traffic Safety on Waterways
..::FEATURE ARTICLE Traffic�Safety
MARINE�REPORT
� MORE�INFORMATION�VISIT� http://www.uni-hannover.de/en/
� Fig:03Calculated (Left) and measured (Right) flow field within and in the vicinity of the outfall structure at half-water depth