LargeScaleMetrologyinAerospaceAssemblyDETv3.docx

LargeScaleMetrologyinAerospaceAssemblyDETv3.docx

《LargeScaleMetrologyinAerospaceAssemblyDETv3.docx》由会员分享，可在线阅读，更多相关《LargeScaleMetrologyinAerospaceAssemblyDETv3.docx（21页珍藏版）》请在冰点文库上搜索。

LargeScaleMetrologyinAerospaceAssemblyDETv3

LargeScaleMetrologyinAerospaceAssembly

JodyMuelaner

TheUniversityofBath

J.E.Muelaner@bath.ac.uk

PaulGMaropoulos

TheUniversityofBath

P.G.Maropoulos@bath.ac.uk

Abstract

Thepaperpresentsareviewoftheprinciplesandstateoftheartininstrumentationusedtomakelargescalemeasurementswithinaerospaceassembly.Theabilitytomeasurelargeartefactsaccuratelyisakeyenablingtechnologytoimprovequalityandfacilitateautomation.Particularemphasisisplacedonissuesofuncertaintywiththeimportanceofacceptancecriteriaexplainedandverificationstandardscomparedanddiscussed.Thefundamentaltechnologiesdeployedareexplainedincludinglasertrackers,indoorGPSandphotogrammetry.Commerciallyavailablesystemsarecomparedintermsofuncertainty,rangeanddeploymentrelatedissues.

keywords

Metrology,LargeVolume,Uncertainty,LaserTracker,iGPS,Photogrammetry

1.Introduction

Theassemblyoflargeaerospacestructuresischaracterizedbyarelianceonmonolithicjigsandhighlevelsofmanuallyintensivereworking,fettlinganddrillingoperations.Insimpletermstheprocessistobringtogetherlargeflexiblecomponentsandsecurethemtoarigidjigwhichcontrolstheshapeoftheemergingstructure.Anymismatchbetweencomponentsisdetectedthroughtheuseofslipgaugesandothermanualinspectiontechniques.Componentsareshimmedorfettledtoensurethatinterfacetolerancesaremaintained.Holesarethendrilledthroughthecomponentsandtheyarefastenedtogether.Thishasbeensummarizedas,“Place,clamp,fastenandrelease”（Pickettetal,1999）.AgenericaerospaceassemblyisshowninmoredetailinFigure1.

Assemblymayaccountforasmuchas40%ofthetotalcostofmanufacturinganairframeduelargelytothelabourandqualityissuesinherenttodrillingthousandsofholesperaircraft（Bullen1997）.Approximately5%ofthetotalmanufacturingcostofanaircraft（Rooks2005）or10%oftheairframe（Burleyetal,1999）isrelatedtotheuseoffixedtoolingwhilereworkingalsorepresentsasignificantproportionofthetotalcostofaircraft（Curranetal,2002）.

Figure1–GenericAssemblyProcess

Largescaleframelessmetrologysystemssuchaslaserandvisionbasedtechnologieshavethepotentialtoovercomemanyproblemsinaerospaceassemblybyenablingflexibleautomationsystems.Largescalereconfigurabletoolinghasbeensuccessfullydemonstratedusingtheinherentaccuracyofmachinetoolstoplacefixtureswhicharethenlockedinplace（Stone2004）.Theuseofnewmetrologytechnologiesmakesthereconfiguringoftoolinginotherapplicationsapracticalandaffordableproposition（Burley,Odietal.1999）eliminatingtherequirementforfixedjigs.Thereisalsothepotentialtofacilitatetheautomationofinspection（Buckinghametal,2007）,fettling（WebbandEastwood2004）anddrilling（Rooks2001）.

Amajorfactorimpedingtheintroductionofautomationisthedifficultyinmakingaccuratemeasurementsandtoolplacementsatthescalerequiredforcommercialaircraftproduction.Largescalemetrologysystemsaddresstheseissues.

2.UncertaintyinRelationtoPartACCEPTANCE

Acommonmistakemadebythosenotfamiliarwiththeprinciplesofmetrologyistoassumethattheresolutionofaninstrumentisthesameasitsaccuracy.Thiscanbeeasilyunderstoodwiththeexampleofatapemeasure.Theresolutionofthetapeis1mmandausermightassumethattheaccuracyistherefore±0.5mm.Whenmeasuringhorizontallywiththetapeunsupportedtheaccuracywillbeconsiderablyworsethanthissincesagandstretchwillbehighlydependentonthetensioninthetape.

InFigure2ameasurementisbeingmadebetweentwobrackets.Thedimensionis1,500mmwithatoleranceof±5mm.Thegraduationsonthetapeshowthedistancetobe1,497mmandsoiscouldbeassumedthatthepartiswithintolerance.Inactualfactthesaginthetapehastakenup3mmandsotheactualdistancebetweenthebracketsis1,494mm–outoftolerance!

Ifthetoleranceforapartgivesaminimumandamaximumvaluethenwhenthepartismeasuredusingagiveninstrument,allowancemustbemadeforthatinstrument’suncertainty.Theexpandeduncertainty,atagivenconfidencelevel,fortheinstrumentisaddedtotheminimumvaluetogiveaminimumacceptancevalue.Similarlytheexpandeduncertaintyissubtractedfromthemaximumvaluetogiveamaximumacceptancevalue.Whenthepartismeasuredthereadingmustbewithintherangeoftheacceptancevaluesinordertosaythatthepartiswithinthetoleranceatthegivenconfidencelevel（BSI1999）.

Figure2–TapeMeasureExample

WecansaythattherearefivepossiblescenarioswhenmakingameasurementasillustratedgraphicallyinFigure3.

Figure3–PossibleInteractionsbetweenToleranceZoneandUncertaintyBand

A.Theuncertaintyoftheinstrumentisgreaterthanthetoleranceofthepartandsoitwillneverbepossibletodeterminewhetherthepartiswithintolerance.

B.Theuncertaintyoftheinstrumentislessthanthetoleranceofthepart.Thereadingshowstheparttobesufficientlyoutoftolerancethatthereisnooverlapbetweenthetolerancezoneandtheuncertaintyband.Wecanthereforestatewithconfidencethatthepartisoutoftolerance.

C.Theuncertaintyoftheinstrumentislessthanthetoleranceofthepart.Thereadingshowstheparttobeoutoftolerancebutthereisoverlapbetweenthetolerancezoneandtheuncertaintyband.Thepartmaybeintolerancebutmustberejected.

D.Theuncertaintyoftheinstrumentislessthanthetoleranceofthepart.Thereadingshowstheparttobeintolerancebutthereisoverlapbetweenthetolerancezoneandtheuncertaintyband.Thepartisprobablyintolerancebutwecannotstatethiswithconfidenceandthereforeitmustberejected.

E.Theuncertaintyoftheinstrumentislessthanthetoleranceofthepart.Thereadingshowstheparttobesufficientlywithinthetolerancethatthereisnooverlapbetweenthetolerancezoneandtheuncertaintyband.Wecanthereforestatewithconfidencethatthepartisintolerance.Thisistheonlycasewherethepartshouldbeaccepted.

Ifwereturntotheexampleofthetapemeasureandsaythattheuncertaintyinthemeasurementis±4mmata95%confidencelevel.Sincethedimension1,500mmhasatoleranceof±5mmtheacceptancecriteriaisthatthemeasuredvalueliesbetween1,499mmand1,501mm.Themeasurementof1,497mmwouldthereforefaileventhoughitinitiallyappearstobewithintolerance（conditionD）.

Theseissuesareimportanttorememberwhenusingadigitalinstrumentwhichmayhavearesolutionof0.1µmbutanuncertaintyof±50µm!

3.VerificationStandards

Anumberofpublicationsaresummarizedwhichhaverelevancetotheverificationoflargescalemetrologyinstruments.Althoughthestandardsandpapersreviewedcoveranumberofdifferentinstrumentsthereisagreatdealofcommongroundbetweenthem.

Themeasurementofcalibratedlengthsisabasicprincipleofmaintainingtraceabilityofthemeasurementsmadewithaninstrumentbacktosomereferencestandard.

Isolationofsub-systemsisanapplicationoftheprincipleofdecomposingsourcesoferror.Inallsystemswhichuseaprobethereisanattempttoisolatetheerrorduetotheprobeandquantifythiserrorindependently.Apossiblesourceofprobeerrorisadeviationfromsphereicity.Similarlymostoftheliteraturereviewedencouragedtheisolationofindividualencodererrorsininitialtests.

Inadditiontotestingsub-systemsinisolationtheliteraturealsoencouragestestingthecombinedeffectofthesystemasawhole.Standardtestsarenotabletoestablishtraceabilityofallmeasurementsthatitispossibleforcomplexequipmenttomake.Itisthereforeimportantthatstandardtestsaresupplementedbytestswhichmorecloselyresemblethemeasurementtaskstobecarriedout.

Testsshouldbecarriedoutinaccordancewithnormaloperationoftheinstrumentasrecommendedbythemanufacturer.

Inthestandardsstudiedsimpledecisionruleswereusedtodetermineconformanceornon-conformancewiththeexpectedperformance.

ISO10360-2:

2002

TheISO10360（BSI2002）acceptanceandreverificationtestsareawellestablishedstandardforcoordinatemeasuringmachines（CMMs）;itisdirectlyapplicableonlytoconventionalgantrybasedCMMsusingcontactprobingandoperatinginthediscrete-pointprobingmode.Errorisdividedintoprobingerroranderrorofindicationofsizemeasurement.

Probeerrorisdeterminedbymaking25pointmeasurementsonthesurfaceofaknownsphereandcomputingthedeviationofmeasuredpointsfromtheGaussianassociatedsphere.

TheerrorofindicationofsizemeasurementistheprimarymeasureoftheaccuracyofaCMM.Fivedifferentcalibratedlengthsareplacedinsevendifferentlocationsand/orpositionsandmeasuredthreetimesineachpositionforatotalof105measurements.Thelongestlengthshouldbeatleast66%ofthelongestdiagonalwithinthemeasuringvolume.Thestandarddoesnotstatetheorientationsinwhichthemeasurementsshouldbetaken,howevertheNPLguidetoCMMverification（Flack2001）doessuggestthatthesevendifferentlocationsmightincludesomeofthefourcrossdiagonals,thethreeinplanediagonalsandthelinesnominallyparalleltoanaxis.

ASMEB89.4.19.

TheASMEB89（ASME2006）standarddetailsverificationproceduresspecificto‘SphericalCoordinateMeasurementSystems’usedasindustrialmeasurementtoolssuchaslasertrackersandlaserradar.Thelowlevelgenerictestsmeasuringcalibratedlengthsthataredetailedinthisstandardshouldbesupplementedbytestswhichcloselymirror