Lift-typewind-basedloadsdescribedabovenotonlycauserotortospin,butalsoleadtotheso-called ap-wisebendingoftheblades.Itshouldberecognizedthatthelift-inducedloadinghasbothapersistent/static-likeandatime-varyingcomponent(thelatteroneisduetonaturalvariabilityofthewind).Inaddition,therelativefractionofthetwoloadcomponentschangesduringrotationoftherotorduetotheso-calledwind-sheareffects(i.e.,duetoanaturalincreaseinthewindspeedwithanincreaseintheheightabovetheterrain).
Inadditiontothelift-relatedloadsdiscussedabove,wind-turbinebladesarealsosubjectedtogravityloads.Theseloadsarethehighestinmagnitudewhenthebladeisinanearlyhorizontalpositionandtheycauseedge-wisebendingofthewind-turbineblade.Since,thebladesbendonewaywhentheyareontheright-handsideofthetowerwhiletheybendintheotherdirectionwhentheyareontheleft-handsideofthetower;gravityloadingalsocontainsavariablecomponent.
Wind-turbinebladesarealsosubjectedtocentrifugalloadingduetorotationoftherotor.Nevertheless,sincetheupper-boundangularvelocityoftherotoristypicallyina10-20rpmrange,centrifugal-tensileloadsalongthebladelengtharegenerallynotconsideredasdesign-controlling/life-limitingloads(andare,hence,ignoredinthiswork).
Toaccountforthetypicalwind-turbinebladeloadingdiscussedabove,aseriesoftwo-dimensionalaerodynamicanalyseswascarriedoutusingtheJavafoilcomputerprogram(Ref6).Thisprogramsolvesthe owequationsoveranairfoilbyimplementingtheboundaryintegralmethod.Forthegivenairfoilpro leandsize,thewindspeedandtheangleofattack,theprogramgeneratesadistributionofpressuresoverthebladesurface.Anexampleoftheresultspertainingtothespatialdistributionofthecoef cientofpressure(aratioofthepressureminusmean-streampressuredifferenceandthehalfproductofmean-streamair-densityandsquaredwindvelocity)isdis-playedinFig.5.Theseanalysesarerepeatedforupto10equallyspacedwind-turbinebladecrosssections.Theresultsobtainedwerethenusedwithinaninterpolationalgorithmtocomputepressuredistributionovertheentirebladesurface.Twowind-inducedloadingconditionswereconsidered:(a)
Forthestructural-responseanalysis,peakloadswerederivedbyconsideringa50-yearextremegustof70m/s(IECClass1(Ref7).Thebladeisassumedtobeinafullyfeatheredposition(i.e.,pitchofthebladeis
Fig.5Anexampleoftheresultspertainingtothetwo-dimensionaldistributionofthecoef cientofpressureandthestreamlinesintheregionsurroundingtheairfoilforthecaseofa10°angleofattack(theanglebetweenthewinddirectionandtheairfoilchord.)
JournalofMaterialsEngineeringandPerformanceadjustedtoobtainthewindattack-angleassociatedwiththelowestaerodynamicloads)witha±15°variationinwinddirection.Toattainthemostconservativeloadingcase,itwasassumedthatthegust-inducedloadingresultsineachbladesectionsimultaneouslyreachingitslocalmaximum-liftcoef cientcondition;and
(b)Forthefatigue-lifeprediction/assessmentanalysis,load-ingwasdeterminedusingtheaveragewindspeedatthewind-turbinepowerrating.ThisvelocitywascomputedusingtheprocedureoutlinedinRef5.Withinthisproce-dure,thespeci cpowerrating(takentobe0.44kW/m2)isde nedasaproductofrotoref ciencycoef cient(a=0.5),adrive-trainef ciency(b=0.925),airden-sity(=1.225kg/m3),andthethirdpowerofthewindratedspeed(=130%ofthewindmeanspeedattherotorhubelevation).Thisprocedureyieldedawindmeanspeedatthehubelevationof7.67m/sinthedirectionofrotoraxis.Itshouldbealsonotedthatthisprocedureenableddeterminationofthemean-levelwind-inducedloadsintheHAWT-blade.Toaccountforthetime-varyingcomponentofthewind-inducedandgravityloading,theso-calledWISPER(windspectrumrefer-ence)loadinghistory/pro le(Ref8)(areferenceloadspectrumtypicallyusedinthedesignofwind-turbinebladesinEurope)wasused(afterproperscaling).Todeterminethequasi-staticstructuralresponseoftheblade,astatic nite-elementanalysiswascarriedoutinwhichtheroot-edgeofthebladewas xedandthebladeoutersurfacessubjectedtotheaforementionedgust-inducedloading.Theresultsoftheseanalyseswereusedtodeterminetheturbine-bladebendingstiffness(asquanti edbytheaveragedisplacementofitstipsection)andbythebladestrength(asmeasuredbythelargestvalueofthevonMisesequivalentstresswithinitsinterior)aswellastheextentofbending-to-torsioncoupling(asmeasuredbytheloading-inducedtwistatthebladetip).Inadditionduetothefactthatwind-inducedloadingwasfoundtobenearlyproportional(i.e.,theorienta-tionofthein-planeprincipalcoordinatessystemoverthemosthighlystressblade-surfacesectionswasfoundnottochangesigni cantlyduringloading),theresultsofthestructuralanalysiswereusedalsointhefatigue-lifeassessmentanalysis(discussedinnextsection).Inotherwords,localstressesareassumedtoscalelinearlywiththeleveloflocalwind-inducedloadingsothatthegust-basedstressescanbeusedtodirectlycalculatethecorrespondingstressesatanylevelofwind-inducedloading.
Allthecalculationspertainingtothestructuralresponseofthewind-turbinebladeweredoneusingABAQUS/Standard,acommerciallyavailablegeneral-purpose nite-elementprogram(Ref9).
2.3Wind-TurbineBladeFatigue-LifePrediction
Itiswell-establishedthatinmostcases,thelifecycleofawind-turbinebladeiscontrolledbyitsfatiguestrength(inthepresenceoflocalthermalandaggressiveenvironmentalcon-ditions).Whileitisgenerallyfairlystraightforwardtoquantifyfatiguestrengthofthestructuralmaterials(glass-orcarbon- berreinforcedpolymer-matrixcomposites,inthecaseofwind-turbineblades)underconstant-amplitudeloadingcondi-tions,relatingthematerialfatiguestrengthtothecomponent(aturbineblade,inthepresentcase)isaquitechallenging

task.
Volume19(6)August2010—793
热门文档