emanatingfromtheorigin.InFig.7,R=0.1andR=0.5dataareassociatedwithapositive/tensilemeanstress/strainvalue,R=À1correspondstoazeromean-value,whileR=10andR=2pertaintoanegative/compressivemean-value.
ToconstructthecorrespondinglinearizedGoodmandia-gram,constantfatigue-lifedataassociatedwithdifferentR-ratiovaluesareconnectedusingstraightlines.TocompletetheconstructionoftheGoodmandiagram,theconstantfatigue-lifelinesareconnectedtotheultimatetensilestress/strainandtotheultimatecompressivestress/strainpointslocatedonthezero-amplitudehorizontalaxis.ThecompletedGoodmandiagramdisplayedinFig.7thenenables,throughinterpolation,deter-minationofthefatiguelifeforanycombinationofthestress/strainamplitudeandstress/strainmean-value.Hence,athree-dimensionalhistogramsimilartothatoneshowninFig.6(b)canbeconstructedexceptthatthenumberofcycleshererepresentsthetotalnumberofcyclestofailureratherthanthenumberofcyclesintheanalyzedloadsignal.Anexampleofsuchthree-dimensionalhistogramisdisplayedinFig.8.
2.3.3MinerÕsRule.Thecycle-countingproceduredescribedearlierenablescomputationofthenumberofcycles/half-cyclesinthegivenloadsignalwhichfallintobinsofathree-dimensionalhistogram(Fig.6b).TheuseoftheGoodmandiagram,ontheotherhand,enablesthecomputationofasimilartri-dimensionalhistogrambutforthenumberofcyclestofailure(i.e.,thefatiguelife,Fig.8).AccordingtotheMinerÕsrule,aratioofthenumberofcyclesandthecorrespondingtotalnumberofcycles,foragivencombinationofthestress/strainamplitudeandstress/strainmean-value,de nesafractionaldamageassociatedwiththiscomponentoftheloading.Thetotaldamageisthenobtainedbysummingthefractionaldamagesoverallcombinationsofthestress/strainamplitudeandthestress/strainmean-value.
Thetotalfatiguelifeunderthegivennonconstantamplitudetime-varyingloadingisobtainedbydividingtheload-signaldurationbythetotalfractionaldamage.Thisprocedureclearlypostulatesthatfatiguefailurecorrespondstotheconditionwhenthetotaldamageisequaltounity.
2.4Wind-TurbineBladeMaterialSelection
FromsimpleconsiderationofbasicfunctionalandlongevityrequirementsforaHAWT-blade,itcanbereadilyconcluded
Fig.8Anexampleofthethree-dimensionalhistogramshowingtheeffectofstress/strainamplitudeandthestress/strainmean-valueofthematerialfatiguelife
JournalofMaterialsEngineeringandPerformancethatthemainbladematerial-selectionindicesmustbebasedonthefollowingmaterialproperties:
(a)ahighmaterialstiffnesstoensureretentionoftheopti-malaero-dynamicshapebythebladewhilesubjectedtostrong-windloadingconditions;
(b)alowmassdensitytominimizegravity-basedloading;
and
(c)alargehigh-cyclefatiguestrengthtoensuretherequired
20-yearlifecyclewithhighreliability.Asmentionedearlier,theHAWT-bladeisessentiallyacantileverbeam.Ifthematerial-selectionmethodologypro-posedbyAshby(Ref13)isutilized,thenthe rstmaterial-selectionindexcanbede nedbyrequiringthatthebladeattainsaminimalmasswhilemeetingthespeci edbending-stiffnessrequirements(oralternativelythatthebladeattainsmaximumbendingstiffnessatagivenmasslevel).Sincetheblademassscalesdirectlywithitsaveragecross-sectionalareawhileits,stiffnessscalesroughlywiththesquareofits,cross-sectionalarea,followingAshbyÕsmaterial-selectionprocedureonecanreadilyderivethefollowing‘‘light,stiffbeam’’material-selectionindex:M1¼E1=2=q
ðEq2Þ
whereEisthematerialsYoungÕsmodulusandqisitsdensity.
TheuseofM1intheHAWT-bladematerialselectionwouldnormallyidentifyfoam-likematerialsaspotentialcandidates.Inthesematerials,theirlowstiffness(asquanti edbythevalueoftheirYoungÕsmodulus,E)ismorethancompensatedbytheirlowqvalue.Consequently,M1takesonalargevalueinthecaseoffoammaterialssuggestingtheirsuitabilityforuseintheHAWT-bladeapplications.However,foammaterialswouldyieldverybulkybladeswhichcouldpresentseriousdesign,manufacturing,installationandoperationalproblems.Inaddi-tion,potentiallyopen-cellstructureandtheassociatedhighwater-permeability/moisture-absorptioncandisqualifythesematerialsfrombeingusedintheHAWT-bladeapplications.Toovercometheseproblems,asecondmaterial-selectionindex(moreprecisely,alower-boundmaterial-propertylimit)isproposedwhichrequiresthattheHAWT-bladematerialspossessaminimallevelofabsolutestiffness,i.e.,M2!E:
ðEq3Þ
Typically,theminimalleveloftheYoungÕsmodulusrequiredforagiven-bladematerialisina15-20GParange.
Thetwomaterial-selectionindicesde nedaboveutilizetwo(Eandq)outofthethreepreviouslyidenti edmaterialproperties.Inclusionofthethirdmaterialproperty(thefatiguestrength)intoamaterial-selectionindexis,however,quitechallenging.Thereasonisthat,asdiscussedintheprevioussectionwhiletheconstant-amplitudefatiguestrengthassociatedwithagivenloadmean-valueandagivenfatiguelifecanbereadilydetermined,HAWT-bladematerialselectionrequirestheuseofavariable-amplitudefatiguelife.
Asdemonstratedintheprevioussection,thevariable-amplitudefatiguelifecanbe,inprinciple,computedforagivencombinationofthesustainedquasi-staticandtime-varyingloads.However,theprocedurewhichisusedinthiscalculationalsoentailstheknowledgeoftheconstant-amplitudefatiguedataunderdifferentmean-value/R-ratioconditions.Since

the
Volume19(6)August2010—795
热门文档