Comparison of Smart Rotor Blade Concepts for Large Offshore Wind Turbines.docx

Comparison of Smart Rotor Blade Concepts for Large Offshore Wind Turbines.docx

《Comparison of Smart Rotor Blade Concepts for Large Offshore Wind Turbines.docx》由会员分享，可在线阅读，更多相关《Comparison of Smart Rotor Blade Concepts for Large Offshore Wind Turbines.docx（16页珍藏版）》请在冰豆网上搜索。

ComparisonofSmartRotorBladeConceptsforLargeOffshoreWindTurbines

ComparisonofSmartRotorBladeConceptsforLargeOffshoreWindTurbines

B.A.H.MarrantandTh.vanHolten

DelftUniversityofTechnology,FacultyofAerospaceEngineering

Kluyverweg1,2629HSDelft,TheNetherlands

Tel.:

+31（0）152785171

Fax.:

+31（0）152783444

E-mail:

B.Marrant@lr.tudelft.nl

Keywords:

offshore,windturbines,fatigueloads,smartmaterials,activerotorcontrol

Abstract

Thispaperprovidestheresultsofacomparisonoffourdifferentsmartstructureconceptstoobtainactiverotorcontrolonlargeoffshorewindturbines.Theconceptsareactivetrailing-edgeflapcontrol,micro-electro-mechanicaltabcontrol,cambercontrolwithinflatablestructuresandactivebladetwistcontrol.Thedifferentsmartrotorbladeconceptsarecomparedwitheachotherbasedontheirpotentialtoreducefatigueloadsforparticulardimensions,theiraerodynamicefficiency,bandwidthandcomplexity.

Introduction

Theaimoftheresearchprojectonsmartdynamicrotorcontrolforlargeoffshorewindturbinesistodevelopnewtechnologycapableofconsiderablyreducingtheextremeandfatigueloadsonwindturbines–inparticularonverylargewindturbinesforoffshoreapplication–andtherebytoreducethecostsofwindturbines.Asecondaimistoreducemaintenancerequirementsandimprovereliabilitybyapplyingcondition-monitoringtechniques.

Thewaytoachievethisistoimplementrecentadvancesincontroltheory,sensor-andactuatortechnology,smartstructures,etc.takingintoaccountthespecialrequirementsandconditionsofoffshorewindturbines.TheFLEXHAT-programperformedintheNetherlandssomeyearsagohasshownthat“smart”controlmethodsmayhaveasignificanteffectonthevariousloads[1].Thepurposeoftheformerprojectwastodecreaseloadsinthedrivetrainandtherotorbladesbyusingpassivetipcontrolandflexibilitiesintherotorsystem.Unfortunately,thetechniquesusedwithintheFLEXHATconfigurationwerenotsuitableforincorporationintoverylargewindturbines.Therefore,differentsolutionsneedtobedeveloped.

Inapreliminarystudy[2,3]onsmartdynamicrotorcontrolforlargeoffshorewindturbinesdifferentconceptshavealreadybeeninvestigatedinwindenergyandhelicopterliterature.Helicopterliteraturehasalsobeenstudiedbecauseofitscloserelationtothefieldofwindturbines.Moreover,smartdynamicrotorcontrolforthepurposeofe.g.vibrationreductionisarelativelynewconceptinwindenergywhereasthistopichasbeenthesubjectofstudyformanyyearsinthefieldofhelicopters[4],[10].

Thetwoapproacheswhichwillbeusedtoreducethestructuralloadsonwindturbinesarethereductionofthefluctuationsofaerodynamicloadsandtheactive/passivedampingofstructuralmodes[5].Inthepreliminarystudyalistofcontroldevicesforrotorbladeswasmadewhichcouldbeusedtocontroltheextremeandfatigueloads.Thislistincludedtrailing-edgeflapcontrol（possiblycombinedwithleading-edgeflapcontrol）,Micro-Electro-Mechanicaltabcontrol,activetwistcontrol,part-spanandfull-spanpitchcontrolandcambercontrol.Thisstudyfocusedmainlyonthefeasibilityofsmartmaterialsasameanstoactuatethedevices.Theseconceptshavealreadybeenpresentedinreference[6].

Thispaperwillcontinuethepreviousworkwiththepurposetomakearankingofdevicesforthepurposeofsmartdynamicrotorcontrol.Themostpromisingdevicesareconsideredtobetrailing-edgeflapsandMicro-Electro-Mechanicaltabsbecauseoftheirrelativesimplicityandtheirpotential.Ontopofthesetwodevicesactivetwistandvariablecamberwiththeuseofinflatablestructuresarealsoincluded.Theactuatorswhichareconsideredinthispaperaremainlysmartmaterialactuatorsbasedonpiezoelectricmaterials.Thedifferentconceptsthatresultfromthedevicesandactuatorsarecomparedwitheachotherbasedontheirpotentialtoreduceaerodynamicloads,theiraerodynamicefficiency,bandwidthandcomplexity.

Approachtocomparethesmartbladeconcepts

Thefoursmartbladeconceptswhichwerementionedbeforewillbecomparedwitheachotherbasedontheirabilitytoreducefatigueloadsduringnormaloperationofthewindturbine.Thefatigueloadsareusedasabasistocomparetheconceptsbecausewindturbinedesignsareoftengovernedbyfatigue[12]andbecausenoneofthesmartbladeconceptsmentionedpreviouslywillhavethepowertoreducetheextremeloadscompletely.Thefatigueloadcaseduringnormalpowerproduction（DLC1.2）,asdescribedintheIECstandard[7],hasbeenusedasabasisforthecomparisonbecausethiswillbethephaseduringwhichthesmartbladewillbeoperatingmostofitstime.Thismeansthatnormalpowerproductionwiththeoccurrenceofanemergency,startup,normalshutdownandstandstillloadcaseshavenotbeenconsidered.

ThefatigueloadcalculationsfortheconventionalbladeandthesmartbladeconceptshavebeenperformedforwindturbineclassIBbecausethisinvolvesahighreferencewindspeedaverageover10minutes（Vref=50m/s）andamediumturbulenceintensityat15m/s（Iref=0.14）whichisconsideredtoberepresentativeforoffshorewindconditions.

Theturbulencemodelwhichhasbeenusedisathree-dimensional,onecomponentmodel,whichmeansthatthewindspeedvariesovertherotordiscareaintime.ThemethodtosimulateturbulencemakesuseofFourierseriestocreateanumberofcorrelatedtimeseriesfromthelongitudinalvelocitycomponentspectrum（S1（f））andacoherencefunction.TheturbulencespectrumusedfortheanalysisistheKaimalspectrum:

（1）

where

1=Iref（0.75Vhub+5.6）:

isthelongitudinalturbulencestandarddeviationwithIref=0.14

L1:

isthelongitudinalvelocityintegralscale,whereL1=8.11

Vhub:

isthewindspeedathubheight

f:

isthefrequencyinHertz

isthelongitudinalturbulencescaleparameter

Thefollowingexponentialcoherencemodel（Coh（r,f））isusedinconjunctionwiththeKaimalautospectrumtoaccountforthespatialcorrelationofthelongitudinalwindspeedcomponent:

（2）

whereristhemagnitudeoftheprojectionoftheseparationvectorbetweenthetwopointsontoaplanenormaltotheaveragewinddirection.

Figure1:

Turbulenceforawindspeedof13m/sovertheheight（z）andthewidth（y）

Anexampleoftheturbulenceatawindspeedof13m/sattimet=0canbeseeninfigure1wherethetotalgridspans60mx60m,thegridinterspacingis4mandthehubheightis91.4m.Theturbulencewasrotationallysampledbyselectingwindspeedsoutofthecompletewindfieldatpointsinspaceandtimecorrespondingtopositionsoftherotatingbladeofahorizontalaxiswindturbine.Correlatedtimeseriesweregeneratedin60equallyspacedpointsontherotorbladeandlinearinterpolationhasbeenusedwhenthepointonthebladewasinbetweenthegridpoints.

Thedeterministicpartofthewindfieldonlyconsistedofwindshearwherethelongitudinalwindspeedisgivenbythefollowingpowerlaw:

（3）

Thefatigueloadsforthedifferentsmartrotorbladeconceptsarecomparedmakinguseofabenchmarkwindturbine.ForthispurposeanoffshorewindturbinederivedfromtheDOWEC（DutchOffshoreWindEnergyConverter）windturbinestudy[9]isusedsincethisinvolvesahorizontalaxis（HAWT）,upwind,pitch-regulated,variablespeedturbinewhichhasthesizeandcharacteristicsofawindturbinethisresearchisintendedfor.AdrawingoftheoriginalDOWECdesigncanbeseeninfigure2andsomecharacteristicdataofthisbenchmarkturbinearepresentedintable1.

Cut-inwindspeedVin

3.0m/s

Cut-outwindspeedVout

25m/s

RatedwindspeedVr

12m/s

Tipspeedratior

7.4

Hubheightzhub（abovethewater）

91.4m

RotordiameterD

120m

RatedpowerPr

6.0MW

Figure2:

DOWECwindturbineTable1:

Characteristicdataofthebenchmarkwindturbine

ForthecalculationsoftheloadstheWindsim[18]packagehasbeenusedwithafewalterationsinordertobeabletocalculatetheaerodynamicloadingduetoadistributedwindfield.ThispackageusesBEMtheoryforthecalculationsoftheaerodynamicloadsinwhichthewakeisapproximatedasconstantovertherotordiscarea.Thesmartrotorbladeconceptswerecomparedbycalculatingtheirfatiguedamagerelativetotheconventionalblade.Asafirstapproximationthedynamicsofthebladeareneglectedandonlythevariableaerodynamicloadingduetothevariablewindfieldisconsidered.Themaximumloadalleviationcapacityofthesmartstructureshasbeenusedintheanalysiswhereithasbeenassumedthatthesmartrotorbladeknowsexactlywhatthewindfieldlookslikeateverytimestep,moreoverasafirstapproximationthesmartbladeisassumedtoreactinstantaneouslytotheloadchange.Makingtheseassumptionshastheinherentadvantagethatthecontrollercanbeleftoutoftheanalysiswhichleadstoamorestraightforwardcomparisonofthesmartrotorbladeconcepts.Thiswayafirstestimatecanbemadeoftheminimumrequireddimensions,deflectionsanddeflectionratesforthedifferentsmartstructureconcepts.

Thesmartrotorbladeconceptswhichareabletoaltertheaerodynamicbladeloadscanbedividedintotwocategories:

bladeswhichareabletoactivelychangetheairfoil’scambertherebyshiftingthecl-curveup-ordownwardandbladeswhichareabletochangethelocalangle-of-attackinordertoobtainchangesinliftcoefficient.Trailing-edgeflaps,MEM-tabsandvariablecambercontrolbelongtothefirstcategorywhereasactivetwistcontrolbelongstothesecond.

Figure3:

Variationofaerodynamicbladerootbending