COMPOSITE MATERIALS FATIGUE ISSUES IN WIND TURBINE

ThispaperisdeclaredaworkoftheU.S.GovernmentandisnotsubjecttocopyrightprotectionintheUnitedStates.COMPOSITEMATERIALSFATIGUEISSUESINWINDTURBINEBLADECONSTRUCTIONJohnF.Mandell,DanielD.Samborsky,andPancasatyaAgastraDepartmentofChemicalandBiologicalEngineering,MontanaStateUniversity,Bozeman,MT59717ABSTRACTThispaperprovidesanoverviewoftheresultsofrecentstudiesofcompositelaminatesofinterestforwindturbinebladeconstruction.Inadditiontotheprimaryrequirementsofstiffness,strength,andeaseofprocessing,windbladematerialsmustwithstandseverefatigueloadingunderserviceenvironments.Thelargematerialvolumesandcostconstraintsalsoleadtounusuallythickpliesandfabrics,aswellasthickadhesivebonds,which,combinedwithrelativelybrittle,lowcostresins,canexacerbatedelaminationrelatedstructuralintegrityissuesfoundinmostcompositesstructures.Importantdifferencesinperformanceareshownforthemajorfiberandresintypesofrelevancetoblades.Detailsoffabricconstruction,fibercontentandplydropandgeometrycanproducemajordifferencesinperformance,particularlyunderfatigueloading.Materialsandconditionsareidentifiedwhereparticularlylowstraindamagefailurecanoccurathighcycles.KEYWORDS:Fatigue,WindTurbineBladeTechnology,GlassFiberComposites,CarbonFiberComposites1.INTRODUCTIONWindturbinebladesaredesignedtoseveralmajorstructuralconditions,includingtipdeflection,strengthandbucklingduringsevereloading,aswellasveryhighnumbersoffatiguecyclesduringoperation,varyingbetweentension,compressionandreversedtension-compressionloadsaccordingtotheparticularloadsspectrumfortheturbineandwindconditions.Themajorstaticstrengthandstiffnesspropertiesdependprimarilyonfibertype,content,andorientation,followingcompositemechanicspredictionswidelyavailableintheliterature.Thefatigueofcompositelaminatesappropriateforwindturbinebladeshasbeenthetopicofresearchstudiesformorethantwodecades;ageneraloverviewofthisareacanbefoundinreference(1).Thefindingsofthesestudiesaresummarizedinrecentreports(2-4),andintwocurrentpublicdatabases(5,6).Recentpublications(7-11)aresummarizedhere,withadditionalnewdatainseveralareas.Thedatabasesprovideadequateconstantamplitudefatiguedatafortherangeofloadingconditionsnecessarytodefineconstantlifediagramsandpredictfailureunderspectrumloading(2,3,12).Thisrequirestestingforatleastfiveorsixloadconditions,asdescribedindetailinReference3and12.Preciselaminateconfigurationsforparticularbladesmaynotbeincludedinthedatabases,but,intheabsenceofdataforparticularlaminates,thefatiguetrendsmaybeassumedtoapplyintermsofstrains.Structuraldetailssuchasplydropsusedinthicknesstapering,andspecialfeaturessuchassandwichpanelclose-outsandjointsrequireseparateattention.Thefatigueresponseofstructuraldetailsistypicallydominatedbycrackinitiationandgrowthinthematrixoradhesive(1,2).Recentstudieshavefocusedonthosematerialsissueswhichappearmostlikelytoproducedamageandfailureforotherwisewelldesignedandconstructedblades(7-11):1.tensilefatigueloadingofglassfiberlaminates,2.compressionstaticandfatigueloadingofcarbonfiberlaminates,3.plydelaminationunderarangeoffatigueloadingconditions,and4.matrixcrackingandtransversedirectionfailure.Thesucceedingsectionsdescribethesensitivitytotheseissuesofarangelaminatesofcurrentinterestinblades,intermsoffiberandmatrixdifferences,fibercontentandlaminateconstruction,infusedfabricarchitecture,processing,loadingconditions,andplydropgeometry.2.EXPERIMENTALMETHODSANDDATAREDUCTION2.1MaterialsAbroadrangeofpotentialbladematerialshavebeenincludedinthecourseofthisstudy,includingE-glass,WindStrandTMandcarbonfibers;polyester,vinylesterandepoxyresins;avarietyoflaminateconstructionsandfibercontents,manystitchedfabricsandseveralprepregs.Laminateswereprocessedbyresintransfermolding(RTM),vacuumassistedRTM(VARTM),SCRIMPTMinfusion,andvacuumbagprepregmolding.Thematerialslistcoversmostmaterialsandprocessdetails.Othermaterialswillbedescribedintheforegoing,andfurtherdetailsmaybefoundinthecitedreferences.Mostofthematerialsareintheformofmultidirectionallaminatescontaining0oand±45oplies,withfibervolumefractionsintherangeofcurrentinfusedorprepregblades.Laminatesusedinbladestypicallyvaryfromallunidirectionalinsomesparstoaall±45oinsomeskinsandwebs.Testingexperiencebothinthisprogram(7,8)andEuropeanOPTIMATprogram(3)hasfoundthatitisincreasinglydifficult,oftenimpossible,toobtaingage-sectionfatiguefailuresundermanytestingconditionsforlaminateswithstrongfibers,highfibercontentsandhighfractionsof0oplies.Oneoutcomeofthisproblemisafocusofthedatabasesonlaminateswithsignificant±45oplycontent.Thetestingphilosophyisthentorepresentfatigueresultsintermsofstrainratherthanstress.Sinceallpliesexperiencethesamestrains,otherlaminateconfigurationswithasignificantfractionof0o(mainloaddirection)plies,includingunidirectional,areassumedtofailatconsistentstrain-cycleconditions;thisassumptionissupportedbytestdatainthisstudy.2.1.1MaterialsList1.DDSeries.ThisisanearlyseriesoftestsonE-glass/polyesterlaminateswith0oFabricAand±45oFabricE(2).Lay-upand%0o-material:[0/±45/0]s;72%-0oFibervolumefractionandthickness:bothvariedMatrix:polyester(CoR