SeawaySizedBulkCarrierModelforHydrodynamicOptimization ofBallastFreeShipDesign

Prof.MichaelG.Parsons DepartmentofNavalArchitectureandMarineEngineering 2600DraperRoad UniversityofMichigan AnnArbor,MI48109 Prof.MiltiadisKotinis EngineeringDepartment,Office239 SUNYMaritimeCollege 6PennyfieldAvenue ThroggsNeck,NY10465

October2, 2006

DepartmentofNavalArchitectureandMarineEngineering UniversityofMichigan AnnArbor,MI481092145

Thisreportrepresentstheresultsofresearchconductedbytheauthorsanddoesnotnecessarily representtheviewsorpoliciesoftheGreatLakesMaritimeResearchInstitute.Thisreportdoes not contain a standard or specified technique. The authors and the Great Lakes Maritime Research Institute do not endorse products or manufacturers. Trade or manufacturers names appearhereinsolelybecausetheyareconsideredessentialtothisreport.

ResearchfundedinpartbytheGreatLakesMaritimeResearchInstitute

Thispageintentionallyleftblank.

TableofContents
page 1.Introduction 2. Background 3.GeneralDescriptionofDesignProcedure 4.ModelConstructionConstraints 5.ModelPropellerSelectionandGeometricScaleFactor 6.ComputationalFluidDynamics(CFD)AnalysisoftheExternalFlow 7.ComputationalFluidDynamics(CFD)Analysisofthe InternalBallastTrunk Flow 8.ModelConstruction 9.PotentialEconomicImpactsoftheResearchResults 10.References 6 7 10 13 14 15

18 19 21 22

ListofTables
page Table3.1:BallastFreeBulkCarrierMainParticulars Table3.2:BallastFreeBulkCarrierBallastCondition Table4.1:MHLTowingTankDimensions Table5.1:MHLNo.23StockPropellerCharacteristics Table5.2:BallastFreeBulkCarrierModelCharacteristics Table6.1:BallastFreeBulkCarrierModelResistanceCalculations 11 11 13 14 14 15

Thispageintentionallyleftblank.

ListofFigures
page Figure2.1: TypicalSeawaysizeBulkCarrier(left)and BallastFreeConceptBulkCarrier(right) Figure2.2: TypicalForwardPlenumandCollisionBulkheadArrangement Figure2.3: TypicalAftPlenumArrangement Figure3.1: PolsteamSeawaysizedBulkCarrierIsa Figure3.2: BallastFreeBulkCarrierBodyPlan Figure3.3: BallastFreeBulkCarrierProfileView Figure3.4:BallastFreeBulkCarrierPlanView Figure6.1:PressureContoursVelocityVectorsattheBowofthe BallastFreeBulkCarrierModel Figure6.2:PressureContoursattheSternoftheBallastFreeBulkCarrierModel Figure6.3:VelocityVectorsattheBowoftheBallastFreeBulkCarrierModel Figure6.4:VelocityVectorsattheSternoftheBallastFreeBulkCarrierModel Figure7.1:ComputationalModelGeometryofDoubleBottomStructure Figure8.1: AftPortionoftheScaleModelSeawaySizedBulkCarrier InvertedSternQuarterView Figure8.2: AftPortionoftheScaleModelSeawaysizedBulkCarrier InvertedSideView

8 9 9 10 11 12 12

16 16 17 17 18

19

20

1.Introduction TheBallastFreeShipConcepthasbeenrecentlyinvented[USPatent#66949082004]and investigated [Kotinis et al. 2004, Kotinis 2005, Ballast Water News 2004] atthe University of Michigan. Even though the feasibility of the concept was demonstrated, the aforementioned analysiswaslimitedbyitsrequiredcomprehensiveresearchscopeandassociatedbudget.Thus, itwasonlyfeasibletosupportmodeltestingthatutilizedanexistingmodel.Althoughthevessel typeofgreatestinterestfortheGreatLakesnonindigenousaquaticspeciesintroductionproblem istheSeawaysizedbulkcarrier,thebestavailablemodelwasofarelativelyfiner,higherspeed bargecarryingLighterAboardShip(LASH)vessel.Thisexistingmodelwasmodifiedtoutilize amoreconventionalstern,butthemodeltestresultswerenotdirectlyapplicabletotheSeaway sizedbulkcarriersstudiedindetailintherestoftheresearcheffort. Thecurrentresearch focusesonthedesignofatypicalSeawaysizedbulkcarrierandthe constructionofascaledmodeltobeutilizedinsubsequenttowingtankexperiments.Thismodel hasalready beenprocuredandisexpectedtobedeliveredtotheUniversityof Michiganatthe endofOctober,2006. A major part of the initial hydrodynamic analysis that was performed on the modified LASH vessel was based on the shipspecific Computational Fluid Dynamics (CFD) code SHIPFLOW [Flowtech Int. 1998], which has been shown to have significant limitations regardingthepredictionoftheimportantviscousflownearthestern. In the current research, the external flow around the BallastFree bulk carrier has been investigatednumericallyusingFLUENT [Fluent2005].Theseexternalflowresultshavebeen utilized to obtain the necessary boundary conditions for numerical simulation (also in FLUENT )ofthewaterflowinthelongitudinalballasttrunks.Theinternalflowwassimulated inordertoestimatetheflowratescalingfactorneededinpreparationforthesubsequenttowing tankexperiments. Inthenextstageofthisresearch,themodelwillbeutilizedfortowingtankexperimentsin ordertostudy,inmoredetail,thehydrodynamicimpactoftheBallastFreeShipConcept.The experimental and numerical hydrodynamic investigation, combined with an optimization procedure, is expected to lead to a design solution that could offer a net savings in Required Freight Rate (RFR) relative to alternate ballast watertreatment methods and approaches. The numericalsolutionwillberefinedandvalidatedthroughthescalemodeltesting.

2.Background TheinitialSeaGrantsupporteddevelopmentoftheBallastFreeShipConceptwasreported inapaperbeforetheAnnualMeetingoftheSocietyofNavalArchitectsandMarineEngineersin Washington, DC, on October 2004 [Kotinis et al. 2004]. Overall, the investigation of the BallastFreeShipConcepthasshownthatitprovidesaviablealternativetotheadditionofcostly ballast water treatment systems in order to meet the evolving performance requirements for ballastwatertreatment.Theconceptessentiallyeliminatesthetransportofforeignballastwater. This should be more effective than currenttreatment methods in reducing the potential for the further introduction of nonindigenous aquatic species into the Great Lakes and coastal waters. Furthermore, it should be equally effective as international requirements extend below the 50 micronrange[IMO2004]. Thetraditionalapproachtoballastoperations,sincethe introductionofsteam machinery, hasbeentheuseofwaterballasttoincreasetheweightofthevesselinthelightcargocondition. A paradigm shift in thinking would be to view the ballast condition as a change of buoyancy ratherthananadditionofweightinordertogetthevesseltoitssafeballastdrafts.Suchashift inthinkingledtotheinventionoftheBallastFreeShipConcept[USPatent#66949082004]. Inthisconcept,thetraditional ballasttanksarereplaced by longitudinal,structural ballast trunksthatextendbeneaththecargoregionoftheshipbelowtheballastdraft.Thearrangement ofanequalcapacityconventionalSeawaysize bulkcarrierisshownontheleftinFig.2.1the arrangementofaBallastFreeShipConceptSeawaysize bulk carrier isshown forcomparison ontheright.Inthisexample,thethreeballasttrunkspersideareconnectedtotheseathrougha plenumatthebowandasecondplenumatthestern.Typicaltrunkandplenumarrangementsat thebowandsternofthevesselareillustratedinFig.2.2and2.3,respectively.Thesetrunksare floodedwithseawatertoreducethebuoyancyofthevesselintheballastconditioninordertoget thevesseldowntoitsballastdrafts.Sincethereisanaturalhydrodynamicpressuredifferential created between the bow region and the stern region of a ship due to its motion through the water,aslowflowisinducedintheseopenballasttrunks.Thisensuresthattheballasttrunksare alwaysfilledwithslowlymovinglocalseawater.Thiswillessentiallyensurethatthereisno transportnonindigenousaquaticspeciesacrosstheglobe.Therefore,thevesselbecomesforeign ballastfreefromthetraditionalviewpoint. When the ballast voyage is completed,the ballasttrunks can be isolated from the sea by valvesandthenpumpeddryusingconventionalballastpumps.Theneedforcostlyballastwater treatment equipment or ballast water treatment chemicals would, thus, be eliminated. This approachwouldalsobeequallyeffectiveforbiotasmallerthan50microns.Duringthefullload conditionoranyconditionwhereballastisnotnecessary,thedoublebottomballasttrunkscanbe segregated utilizing sluice gate valves. This is needed to provide the vessel adequate damage survivability. In order to provide adequate intact stability, equivalent damage survivability, equivalent cargo capacity, etc., the entire vessel design needs to be developed to support this concept of ballastoperationsasillustratedinFig.2.1.Theshiprequiresahighertanktopinordertolocate enoughballasttrunkvolumebelowtheballastdraftandrequiresagreaterhulldepthinorderto

maintain the vessels capacity to carry light cargos, such as grain. The BallastFree Ship Conceptalsoincludesfeaturestominimizethebuildupofsedimentwithintheballasttrunksand facilitatetheirrequiredcleaningi.e.,easiertoclean2.4mhighballasttrunks,eliminationofthe lowerpartofthefloorsnexttotheshell,etc.

Figure2.1: TypicalSeawaysizeBulkCarrier(left)and BallastFreeConceptBulkCarrier(right) Asnoted,therecentSeaGrantsponsoredresearchonthedevelopmentoftheBallastFree ShipConceptwas limited by itsrequiredcomprehensiveresearchscopeandassociatedbudget. For budgetary reasons, it was only feasible to support model testing that utilized an existing model.AlthoughthevesseltypeofgreatestinterestfortheGreatLakesnonindigenousaquatic speciesintroductionproblemistheSeawaysizebulkcarrier,thebestavailablemodelwasofa relativelyfiner,higherspeedbargecarryingLighterAboardShip(LASH)vessel.Thisexisting model was modified to utilize a more conventional stern, but the model test results were not directly applicable tothe Seawaysize bulk carriers studied in detail in the restof the research effort.

Figure2.2: TypicalForwardPlenumandCollisionBulkheadArrangement

Figure2.3: TypicalAftPlenumArrangement

Budget restrictions in the initial investigation phase also required that the model tests be limited to a single system design for a single existing model. There was no opportunity to optimizethehydrodynamicdesignofthesystemtominimizetheeconomicimpactoftheBallast FreeShipConceptdesign.ModeltestsandCFDsimulationsusingamodifiedLASHvesselhull showed that the specific ballast intake and discharge locations and method tested in the initial investigation resulted in a modest 2.2% increase in resistance but a more significant 7.4% increaseintherequiredpropulsionpower.Thisspecificresultassumedachangeintheballast waterwithintheballasttrunksonceeverytwohours,whichwouldmeettheenvironmentalintent oftheBallastFreeShipConcept.Thelargepowerincreasecouldresultinan undesirableengine sizeincreaseandwouldresultinfuelcostpenalties.Inthatinvestigation,itwasconcludedthat further hydrodynamic optimization could eliminate most, if not all, of this significant added powerrequirement.

3.GeneralDescriptionofDesignProcedure A typical Seawaysized bulk carrier was designed using data from similar vessels. The prototype vessels were the Isa bulk carrier owned by the Polish Steamship Company (PZM Polsteam) as shown in Figure 3.1 and a recent bulk carrier class designed by the Jiangnan shipyardinChina.

Figure3.1: PolsteamSeawaysizedBulkCarrierIsa The vesselhydrostaticdata matchthedataoftypical bulk carriersofthesamesize.The service speed was assumed to be 14.5 knots, which results in a Froude number of 0.175. The speedofthevesselwheninballastdraftwasassumedtobe15.5knots,whichresultsinaFroude number of 0.185. The stern region was designed to accommodate specific propellerhull clearances.Forthispurpose,theDetNordskeVeritas(DNV)requirementswereutilized.The mainparticularsoftheBallastFreeBulkCarrierarepresentedinTable3.1. The model testing of interest will be performed with the vessel at the ballast draft. A conservativeheavyweatherballastdraftcorrespondstoavalueof40%ofthesummerloadline attheforwardperpendicularand70%attheaftperpendicular.Thesevalueswereutilizedinthe currentstudy.ThevesselparticularsatthisballastconditionarelistedinTable3.2.

10

Table3.1:BallastFreeBulkCarrier MainParticulars Waterlinelength(m) Lengthbetweenperpendiculars(m) Maximumbeam(m) Depthtomaindeck(m) Fullloaddraft(m) Blockcoefficient Prismaticcoefficient Waterplaneareacoefficient Midshipsectioncoefficient Fullloaddisplacement(metrictons) 2 Wettedsurfacearea(m ) 195.5 192.0 23.76 16.0 10.70 0.835 0.837 0.909 0.998 42,546 7673

Table3.2:BallastFreeBulkCarrier BallastCondition Waterlinelength(m) Maximumbeam(m) DraftatF.P.(m) DraftatA.P.(m) Speed(knots) Reynoldsnumber(freshwater) 189.5 23.76 4.28 7.49 15.5 9 1.35*10

The hull of the BallastFree Bulk Carrier has been designed in Maxsurf [Formation Design Systems2003].ThevessellinesareillustratedinFigures3.2through3.4.

Figure3.2: BallastFreeBulkCarrierBodyPlan

11

Figure3.3: BallastFreeBulkCarrierProfileView

Figure3.4:BallastFreeBulkCarrierPlanView

12

4.ModelConstructionConstraints Theconstructionofashipmodeltobetestedinatowingtankneedstocomplywithcertain requirements that stem from the experimental procedure and the available testing equipment. The models tests will be performed at the towing tank of the University of Michigan Marine HydrodynamicLaboratory(MHL).ThetowingtankdimensionsareshowninTable4.1. Table4.1:MHLTowingTankDimensions Length(m) Width(m) Depth(m) 2 Tankcrosssectionalarea(m ) 109.7 6.70 3.20 21.44

Asignificantconstraintintheexperimentalprocedureistheinterferencebetweentheside wallsandbottomofthetankandthepressurefieldandwavesystemgeneratedbythemotionof themodel.Thisisusuallytermedblockageeffectandresultsintotalresistancemeasurement errors(artificialincrease).Ageneralguidelineisthatthemodelcrosssectionalareabeatmost 1/200ofthetankcrosssectionalarea(intermsofthetankbeamandwaterdepth).Nevertheless, if the operating Froude number is relatively low (less than 0.20), the wave resistance is not a large proportion of the total resistance, and thus correction factors can be introduced in the calculationprocedurewhenlargemodelsareutilized. Ifthevesselunderinvestigationnormallyoperatesindeepwater,asimilarsituationneeds to be simulated in the towing tank experiments. Operation in shallow water is generally consideredtoincreasetheresistance.Ageneralguidelineisthattheratioofthesquarerootof themodelcrosssectionalareaoverthetankwaterdepthshouldnotexceedavalueof0.25.

13

5.ModelPropellerSelectionandGeometricScaleFactor Averysignificantparameterindeterminingthegeometricscalefactorofthemodelisthe sizeoftheavailablestockpropellers.Twopropellers fromtheMHLstock(No.10andNo.23) were investigated during the propulsion analysis. The thrust requirements were based on Holtropsresistanceandpropulsivecoefficientsestimates[Holtrop1984].Thetotalresistanceof theBallastFreebulkcarrieratthefullloadconditionwasestimatedat573kN.Thepropulsive coefficients were also estimated using Holtrops empirical formulas. These were utilized as inputtothepropelleranalysis. The corresponding nondimensional thrust and torque versus coefficient of advance (Kt, KqJ)chartsofthepropellerswereutilizedfortheanalysis.Theresultsshowedthatthemodel propeller No.23 provides a higher propulsive efficiency (0.55) assuming a fullscale propeller diameterof6.0m.Thecorrespondingvalueofthegeometricscaleratiois l=37.92,whichis within the acceptable range, based on the design constraints. This will result in a model waterline length of approximately 5 m. The amountof back cavitation is expected to be well below the upper limit for merchant ships. A summary of the stock propeller No. 23 characteristics is presented in Table 5.1. The model characteristics are demonstrated in Table 5.2. Table5.1:MHLNo.23StockPropellerCharacteristics Diameter(in.) Pitch/Diameter(P/D)ratio Numberof blades Expandedarearatio( AE/AO) Hubdiameter(in.) Hublength(in.) Material 6.229 1.08 4 0.55 1.226 1.355 Brass

Table5.2:BallastFreeBulkCarrierModelCharacteristics Geometricscalefactor l Waterlinelength(m) Maximumbeam(m) Averagedraft(m) F.P.draft(40%F.L.) A.P.draft(70%F.L.) 2 Wettedsurfacearea(m ) Speed(m/s) Reynoldsnumber(freshwater) 37.92 5.00 0.627 0.155 0.113 0.198 5.34 1.295 6 5.78*10

The extent of laminar flow along the model length that corresponds to the aforementionedscalefactorvalueandmodelspeedis9%(ofthemodellength).Theattachment oftripwiresorsandstripstothemodelbowwillbeusedtostimulateturbulenceatapointcloser totheforwardstagnationpointasexpectedatfullscale.

14

6.ComputationalFluidDynamics(CFD)AnalysisoftheExternalFlow
Thecommercial CFDcode FLUENT wasutilizedtostudytheexternal flowaroundthe modeloftheBallastFreeBulkCarrierintheballastcondition.Themodelscalespeedis1.295 m/s,whichiscalculatedassumingFroudescaling.Theobtainedsolutionhasbeendemonstrated to be gridindependent. The Realizable ke turbulence model was utilized in the calculations. The flow close to the hull was approximated with wall functions. The computational grid consistsofahybridmeshwith1,507,546cells.Theinvestigatedflowcorrespondstoadouble model solution the freesurface is ignored and the waterplane is considered to be a plane of symmetry.

TheconvergedCFDsolutioncanbeutilizedtoobtainthehullresistance,whichconsistsof thefrictiondragandthepressure(form)dragforthisconfiguration.Thefrictiondragarounda ship hull is usually estimated through a semiempirical formula adopted by the International TowingTankConference(ITTC).Thecomputedvalueisextremelyclosetothesemiempirical one.Theformdragcorrespondstoonlyasmallfractionofthetotaldrag.Theutilizationofa doublemodelflowallowsforthecomputationoftheformfactor,whichiscalculatedastheratio of the total drag (excluding the wave drag)tothe ITTC friction drag estimate. The computed valueisreasonablefortheshipintheballastcondition.Asummaryoftheresultsispresentedin Table6.1. Table6.1:BallastFreeBulkCarrierModelResistanceCalculations Frictiondragcoefficient(FLUENT) Frictiondragcoefficient(ITTC) Pressure(form)dragcoefficient(FLUENT) Totaldragcoefficient(FLUENT) Formfactork
3 3.32*10 3 3.31*10 3 0.45*10 3 3.77*10 0.139

An important aspectof the BallastFree Ship Concept is the location of the water inlet anddischargeplena.Thesteadypressuredifferentialbetweenthetwolocationswillgeneratea slowwaterflow inthe longitudinal ballasttrunks.Thevolume flowrateisproportionaltothe squarerootofthepressuredifferential.Itwasdemonstrated[Kotinis2005]thatthehigherthe flow rate the more effective (less time required) the flushing of the trunks. The computed pressurecontoursinthebowandsternregionoftheBallastFreeBulkCarrierareshowninFigs. 6.1and6.2,respectively. Thepressurecontoursinthebowregionrevealthatthepressureishighinasmallregion near the bulbous bow. However, starting at the forward perpendicular, the flow is relatively smoothwithlowpressureandpressuregradientvalues.ThisisalsoobservedinFig.6.3,where the velocity vectors in the bow region are plotted. The downward flow direction reduces the heightofthebowwave.Theonlysignificantdrawbackiswithrespecttothelocationofthebow plenum, as the positive pressure is limited to a small area starting at the stagnation point and extendingafttoabout6%ofthewaterlinelength.

15

Figure6.1:PressureContoursVelocityVectorsattheBowofthe BallastFreeBulkCarrierModel

Figure6.2:PressureContoursattheSternoftheBallastFreeBulkCarrierModel Thesternregionisgovernedbysuctionpressure,whichhasapeaknearthebilgejustahead of the stern bulb. A significant positive pressure region exists above the stern bulb. No separation isobserved inthe velocity vectorplotinFig.6.4.The locationofthesternplenum willbeclosetotheaftendoftheengineroom(aroundx/L=0.92)andabovetheinnerbottom (2.4mfromthekeel).

16

Figure6.3:VelocityVectorsattheBowoftheBallastFreeBulkCarrierModel

Figure6.4:VelocityVectorsattheSternoftheBallastFreeBulkCarrierModel

17

7.ComputationalFluidDynamics(CFD)AnalysisoftheInternalBallastTrunk Flow The internal trunk flow was investigated in both full and model scale. A typical trunk configuration[Kotinis2005]wasgeneratedandthecorrespondingpressureboundaryconditions, as calculated from the external flow simulations, were imposed on the inlet/outlet plena. The forward plenum was located at a point about x/L = 0.04 between the 0.2 and 0.35 design waterlines. The aft plenum was located at a point near x/L = 0.925 between the 0.3 and 0.45 designwaterlines. Thepressuredistributionwasassumedtobethesameinbothfullandmodelscale.This assumptionisjustifiedbythefactthattheReynoldsnumberisrelativelyhighinbothcasesand the body at ballast draft is relatively streamlined (low form drag). The computational grid consistsofahybridmeshwith705,915cells.Thetrunkswereassumedtobefilledwithballast water and the natural trunk flushing was initiated by new ballast water entering the system throughtheinletplenum.Thevolumeflowrateattheinletandoutletplenawasmonitoreduntil convergence to a constant value was observed. The unsteady flow solver in FLUENT was utilizedwithatimestepsizeofonesecond(1s).Thesolutionateachtimestepwasconsidered tohaveconvergedwhentheresidualsdroppedbyfourordersofmagnitude.Thecomputational modelgeometryofthedoublebottomstructureisshowninFig.7.1.

Figure7.1:ComputationalModelGeometryofDoubleBottomStructure
3 Theaveragevolumeflowrateatfullscaleis0.969m /s.Themodelscalesolutionhasan 4 3 averageflowrateof1.02*10 m /s.Thecorrespondingratioisapproximately9,500.Thisresult isinagreementwiththescalingfactorbetweentheshipscaleflowrateandthemodelscaleflow 5/2 rate, Qs = Qm l , theoretically derived in the initial investigation of the BallastFree Ship Concept based on Froude scaling. This scaling assumed that the flow rate scaling should be proportional tothe scale factor l raised tothe 2.5 power. The corresponding exponent of the geometricscalingfactorbasedonthenumericalCFDresultswasapproximately2.52.

18

8.ModelConstruction The model of the SeawaySized Bulk Carrier for use in subsequent hydrodynamic optimization testing has been contracted for construction by F.M. Pattern Works of North Vancouver, BC. This is the primary deliverable of the project and its procurement represents approximately halfoftheproject funding.The model isunderconstruction atthetimeofthis writinganddeliveryofthecompleted5mscalemodelisexpectedneartheendofOctober2006. Progressphotographsoftheglueupoftheaftportionofthehullareshownin Figs.8.1and8.2.

Figure8.1: AftPortionoftheScaleModelSeawaySizedBulkCarrier InvertedSternQuarterView

19

Figure8.2: AftPortionoftheScaleModelSeawaysizedBulkCarrier InvertedSideView

20

9.PotentialEconomicImpactsof theResearchResults The hydrodynamic optimization of the BallastFree Ship Concept is expected to demonstrate that there is only a small, if any propulsion power penalty associated with the concept.Inthatcase,itshouldresultinanetRequiredFreightRatesavingsrelativetoalternate ballastwatertreatmentmethodssuchasfiltrationandUVtreatmentofabout0.20$/tonofcargo. The concept should also provide superior protection from the further introduction of nonindigenousaquaticspeciesintotheGreatLakesandcoastalwaters.

21

10.References BallastWaterNews(2004)"TheBallastFreeShipFactorFancy?"BallastWaterNews,Global BallastWaterManagementProgram,IMO,17,AprilJune. FlowtechInternationalAB(1998)SHIPFLOW2.4UsersManual. Fluent(2005) FLUENT6.2UsersManual,FluentInc.,Lebanon,NH. FormationDesignSystemsPty(2003)MAXSURFWindowsVersion9.6UsersManual. Holtrop,J.(1984)AStatisticalReAnalysisofResistanceandPropulsion Data,International ShipbuildingProgress,31363,November. IMO(2004)InternationalConventionfortheControlandManagementofShipsBallastWater &Sediments,DiplomaticConference,February,London. Kotinis,M.,Parsons,M.G.,Lamb,T.andSirviente,A.(2004) DevelopmentandInvestigation oftheBallastFreeShipConceptTransactionsSNAME,112,pp.206240. Kotinis,M.(2005) DevelopmentandInvestigationoftheBallastFreeShipConceptPh.D. Dissertation,UniversityofMichigan,DepartmentofNavalArchitectureandMarine Engineering. U.S.Patent#6694908(2004)BallastFreeShipSystem,U.S.PatentandTrademarkOffice, Washington,DC.

22