Assessment of fatigue safety factors for deep-water risers i

涡激振动

NTNU,

DeptMarineStructures,N-7491Trondheim,Norway

BerntJ.Leira

TrondStokkaMeling

Statoil,

N-4035Stavanger,Norway

NTNU,

DeptMarineStructures,N-7491Trondheim,Norway

http://www.77cn.com.cnrsen

AssessmentofFatigueSafetyFactorsforDeep-WaterRisersinRelationtoVIV

Safetyfactorsrequiredtocontrolfatiguedamageofdeepwatermetallicriserscausedbyvortex-inducedvibration(VIV)areconsidered.Fourdifferentrisercon gurationsarestudied:CasesIandII:Verticaltensioned12in.riserssuspendedfromasparbuoyatwaterdepthsof500and1500m.CasesIIIandIV:Steelcatenaryriserssuspendedfromasparbuoy,bothat1000m.ForCaseIII,theriserdiameteris12in.,whileforCaseIVitis33in.Foreachrisercon guration,relevantdesignandanalysisparameterswhicharesubjecttouncertaintyareidenti ed.Forthesequantities,randomvariablesareestablishedalsorepresentingmodeluncertainties.Subsequently,repeatedanalysesoffatiguedamageareperformedbyvaryingtheinputparameterswithinrepresentativeintervals.Theresultsareappliedto tanalyticalexpressions(i.e.,so-calledresponsesurfaces)utilizedtodescribethelimitstatefunctionandtodeveloptheprobabilisticmodelforreliabilityanalysisoftherisers.Bycombiningtherandomvariablesfortheinputparameterswiththeresultsfromtheparametervariations,arelationshipbetweenthefatiguesafetyfactorandthefailureprobabilityisestablishedforeachrisercon guration. DOI:10.1115/1.2087548

SINTEF,Marintek,Div.Struct.Eng.,

N-7491Trondheim,Norway

VidarBerntsen

ConsultingEngineer

Houston,TX77059-5544,USA

BernieStahl

BP-Exploration,Sunbury-on-Thames,TW167LN,UnitedKingdom

AndrewTrim

Introduction

Inthefollowing,reliabilitymodelsforfatigueassessmentofrisersareestablished.Therisercasesanalyzedaredescribed.Theresponsesurfaceapproachusedtoperformthereliabilityanalysisisoutlinedandtherandomvariablesusedintheanalysesareprovided.Reliabilityresultsareestimatedforthreesafetylevels.Sensitivitiesofthereliabilityresultstochangesintheinputvari-ablesarealsoindicated.

DescriptionofRiserCon gurationCaseStudies.Inordertocoveracertainrangeofrelevantrisercon gurations,bothverticalandcatenaryrisersareanalyzed.Thefollowingexamplerisercon- gurationsareconsidered:

Sparbuoywitha12in.verticaltensionedriseratwaterdepthsof500and1500m

Sparbuoywitha12in.anda30in.SCRatawaterdepthof1000m,correspondingtoa“stretched”versionoftheOr-menLangecase atawaterdepthof850m

i.ii.

Suf cientlyaccurateexpressionswhichrelatethecom-putedfatiguelifetimetoallthebasicinputparametersRandomvariablesfortheinputparametersmustbeestablished

Thereliabilityanalysisissubsequentlyperformedbycombiningtheresultsobtainedfromthesetwobasicstepsinaprobabilisticmodelfortheriserreliabilityanalysis.

DescriptionofProcedure.Thefailurefunction orlimitstatefunction correspondingtoVIV-inducedfatiguecanbeexpressedas

g X =XFail D Top,X2,X3,...,XN

=XFail DTref X2,X3,...,XN · Top/Tref

1

whereX1=XFailisarandomvariablewhichrepresentstheMiner-Palmgrensumatfailure,X2,X3,...XNarerandomvariablesre-latedtotherisersystem,Topisthetimeinoperationfortherisersystem i.e.,thetimeduringwhichfatiguedamageaccumulationtakesplace .Trefisagivenreferencedurationwhichcanbedif-ferentfromthetimeinoperation e.g.,1year .

TheparticularexpressionforthefatiguedamageinEq. 1 isdifferentforone-andtwo-slopeS-Ncurves.Thesimplestexpres-sioncorrespondstoone-slopecurves whichareexpressedintheformN· Sm=C,where Sisthestressrange andreads

DTref X2,X3,...,XN = Nref/C S m 1+ m/

2

TheenvironmentaldataaretakenfromtheOrmenLangearea.Forthepresentpurpose,thecurrentismodeledbya xedpro lewhichisscaledbythesurfacevelocity.

Forallthecases,a singleslope S-Nfatiguecurve,ClassEwithparameterm=3,isapplied.ReliabilityAnalysis.

General.Basedontheparameterstudies,relationshipsbetweenthebasicinputparametersandthefatiguelifetimeareobtained.Inordertoperformareliabilityanalysis,twobuildingblocksarerequired:

ContributedbytheOceanOffshoreandArcticEngineeringDivisionofASMEforpublicationintheJOURNALOFOFFSHOREMECHANICSANDARCTICENGINEERING.Manu-scriptreceivedNovember26,2003;revisionreceivedMarch11,2005.Reviewcon-ductedby:StephenLiu.

whereithasbeenassumedthatthestressrangeisgivenbyaWeibulldistributionwithscaleparameterSandshapeparameter .ThequantitieswhicharefunctionsofthebasicvariablesX2,X3,...,XNareS, andNref.Thelatterquantity i.e.,Nref istheexpectednumberofstresscycleswhichcorrespondtothetimeperiodTref.

Subsequently,thevariationinfatiguedamageasafunctionoftherandomvariables X2,X3,...,XN isfactoredoutasaseparate

NOVEMBER2005,Vol.127/353

JournalofOffshoreMechanicsandArcticEngineering

Copyright©2005byASME

涡激振动

function.Thisisasmoothanalyticalfunctionwhichisobtainedby ttingaresponsesurfacethroughanumberofdiscretepoints.These“points”areobtainedbyperforminganumberof niteelementresponseanalysesincludingestimationoffatiguedamage 1–7 .Theresponsesurfaceisexpressedasaproductofpolyno-mials,witheachpolynomialrepresentingarandomvariable.Theresultingmodi edformofEq. 2 thenbecomesDTref X2,X3,...,XN

= Nref/C SBasecase m· 1+ m/ Basecase ·f X1,X2,...,XN =DTref,Basecase Nref ·f X1,X2,...,XN

3

DTref,Basecase Nref · Top 1 /Tref = 1/SF DTref,Basecase Nref /Tref = 1/SF 1/ SF·Top,permissible = 1/SF · Top,permissible 1 /Top,permissible

7

wherethequantityTop,permissibleisrelatedtothebasecasedamagethroughEq. 5 .Thisexpressionissubsequentlyappliedinthenewfailurefunction.

Byevaluatingthefailureprobabilitybasedonthefailurefunc-tioninEq. 6 minusthefailureprobabilitybasedonthenewfailurefunction,thefailureprobabilityduringthe nalyearofoperationisobtained.

ComputationofthereliabilityindicesandtheprobabilityoffailureisbasedonFORM rstorderreliabilitymethod approxi-mations.Thesearchforthedesignpointisperformedinno …… 此处隐藏：20174字，全部文档内容请下载后查看。喜欢就下载吧 ……