CFD and Wind Tunnel

ScalingTechniquesUsingCFDandWindTunnelMeasurementsforuseinAircraftDesignKarlPetterssonSeptember2006iiKarlPetterssonLicentiateThesisDepartmentofAeronauticalandVehicleEngineeringDivisionofAerodynamicsSE-10044Stockholm,SwedenTRITA-AVE2006:64ISSN1651-7660AkademiskavhandlingsommedtillståndavKungligaTekniskaHögskolaniStockholmframläggestilloffentliggranskningföravläggandeavteknologielicentiatexamenfredagenden13:eOktoberkl10:00iSalS40,KungligaTekniskaHögskolan,Teknikringen8,Stockholm.©KarlPetterssonUniversitetsserviceUSAB,Stockholm2006LicentiatethesisiiiPrefaceThisworkwascarriedoutattheDepartmentofAeronauticalandVehi-cleEngineering.ThefinancialsupporttothisworkwasprovidedbytheREMFIproject(contractAST3-CT-2004-502895),6thEUframeworkpro-grammeandisgratefullyacknowledged.SNICandPDCarealsoacknowl-edgedfortheirhelpandsupportwiththeparallelcomputeronwhichthesecalculationshavebeendone.TheauthorisverygratefultoProf.ArthurRizziforthesupportanden-couragementandforalwaysbeinggenerouswithbothtimeandmoney.Myfellowco-workersatthedepartmentandmyfellowgraduatestudentsintheresearchgroupSimoneandHanyoandthealreadygraduatedTomasaretobeespeciallythanked.Lastlyandmostimportantly,Iwouldliketothankmyfamilyandmygirl-friendHelenaforalwaysbeingsupportiveandcaring.Thankyou!ivKarlPetterssonLicentiatethesisvAbstractThisthesisdealswiththeproblemsofscalingaerodynamicdatafromwindtunnelconditionstofreeflight.Themainchallengeswhenthisscalingshouldbeperformedishowthemodelsupport,wallinterferenceandthepotentiallylowerReynoldsnumberinthewindtunnelshouldbecorrected.ComputationalFluidDynamics(CFD)simulationshavebeenperformedonamoderntransonictransportaircraftinordertorevealReynoldsnumbereffectsandhowtheseshouldbescaledaccurately.Thisinvestigationalsoex-aminedhowtheEuropeanTransonicWindtunnel(ETW)twinstingmodelsupportinfluencestheflowovertheaircraft.InordertofurtherexamineReynoldsnumbereffectsaMATLABbasedcodecapableofextractinglocalboundarylayerpropertiesfromstructuredandunstructuredCFDcalcula-tionshavebeendevelopedandvalidatedagainstwindtunnelmeasurements.Ageneralscalingmethodologyispresented.viKarlPetterssonLicentiatethesisviiLicentiateThesisThethesisconsistsofanintroductionandthefollowingthreepapers:PaperAK.PetterssonandA.Rizzi,“EstimatingReynoldsnumberscalingandwind-tunnelboomeffectswiththehelpofCFDmethods”,AIAA2006,SanFran-ciscoUSA,June2006.PaperBK.PetterssonandA.Rizzi,“ReynoldsnumbereffectsidentifiedwithCFDmethodscomparedtosemi-empiricalmethods”,ICAS2006,HamburgGer-many,September2006.PaperCK.PetterssonandA.Rizzi,“EstimatinglocalboundarylayerpropertiesusingCFDandwindtunnelmeasurements”,KTHInternalreport2006,StockholmSweden,September2006.viiiKarlPetterssonLicentiatethesisixDivisionofworkbetweenauthorsPaperAPetterssonperformedthecomputationsandwrotethepaper.Rizzisuper-visedtheworkandcontributedwithvaluablecommentsfortheanalysisoftheresults.PaperBPetterssonperformedthecomputationsandwrotethepaper.Rizzisuper-visedtheworkandcontributedwithvaluablecommentsfortheanalysisoftheresults.PaperCPetterssonperformedthecomputationsandwrotethepaper.Rizzisuper-visedtheworkandcontributedwithvaluablecommentsfortheanalysisoftheresults.xKarlPetterssonLicentiatethesisxiContentsPrefaceiiiAbstractvLicentiateThesisviiDivisionofworkbetweenauthorsix1Overviewandsummary11.1Background............................11.2Scalingwindtunnelresults...................21.2.1Windtunnelwallandmodelsupportinterference...81.2.2ScalingDragwithReynoldsnumber..........91.3Newgeneralscalingmethodology................112Summaryofpapers133Conclusionsandfuturework15References17PaperA21PaperB29PaperC45xiiKarlPetterssonLicentiatethesis1Chapter1Overviewandsummary1.1BackgroundTodayarewindtunneltestingandCFDcalculationsnaturalandnecessarypartsinthedevelopmentofanaircraft.ThetrendsarethatmoreandmoretimearespentbothinthewindtunnelandperformingCFDcalcula-tions.[19]Thisinordertoavoidacostlystepbackwardstoapreviousdesignphaseinordertocorrectpotentialmistakes.ThemaingoaloftheCFDcal-culationsandthewindtunneltestingaretopredictthefreeflightconditionsoftheaircraft.Accurateflightperformancepredictionisachallengingtaskbecausemostofthetestinghasbeendoneatsub-scaleconditions.Someofthephenomenawhichhastobeaccountedforwhenscalingwindtunneldatatofreeflightconditionarethewallandmodelsupportinterferenceef-fectsandpotentiallyalowerReynoldsnumber.SomerecentresearchonthistopicisgivenbyEckert[9],wheretheinfluenceofwallandstinginterferenceandtheimpactofthepropelleroftheAirbusA-400Misinvestigated.Thehighcostassociatedwithacquiringfreeflighttestdatamakestheamountofinformationaboutscalinggroundtoflightmethodologyrareintheopenliteratureandareoftencompanyproprietaryandpartoftheircompetitiveedge.[1]Someofthedriversofanincreasedaccuracyinscalingmethodologyistheeconomicalbenefitfromhavinganoptimalchoiceofengineforagivenaircraftconfigurationandtheincreasedneedsofreductionofemissions.[20]Anerroneouslypredictedscaleeffectwhichwouldimplyanincreaseindragof1%foranultrahighcapacityaircraftwouldequatetoaround3tonnesofextrafuelatconstantrangeorareductioninrangeo