resourcewinds

0,04,08,012,016,056789101112veer(°)windspeed(m/s)site1site2site3max(p95)Practicalmeasurementsofwindconditionsattallhubheightsindifferentterrain:ModellingWindConditionsforTallHubHeightsDavidHilbert(1),CarolinSchmitt(2),CatherineMeissner(3)(1)REpowerSystemsSE,Germany(2)juwiWindGmbH,Germany(3)WindSimAS,NorwayTallhubheightsreachintotheso-calledEkmanlayer,wherethecommonscalingassumptionsofthesurfacelayerarenotvalidanymore.
Withincreasingheightcoriolisforce,stabilityeffectsandmesoscaleeffectsgetmoreandmoreimportant,makingaunifieddescriptionofthewindprofileintheABLdifficult.
TothisdatenocommercialmicroscalemodellingsoftwarecontainsthefullphysicaldescriptionoftheentireABL.
Attemptstodescribethewindprofileabovethesurfacelayerrequiremeteorologicalparameterswhicharenotavailablefromstandardmastmeasurements([2],[3],[4]).
Aflowmodelwouldfurthermoreneedtomaintainthecappinginversion(Fig.
1)todescribethewholeABLsuccessfully.
Thisinvolvesassumptionswhicharehardtoverifyforpracticalcases.
Therefore,itmayalsobereasonabletoapplysurfacelayerflowmodelsandstudytheiruncertainties.
Dependingonthesitespecificconditions,anyofthoseattemptsmayleadtofalseestimatesofthewindconditionsattallhubheights.
Tallhubheightswhichreach100mandmoreposeachallengeforwindresourceassessment:Conventionalmeasurementsseldomcovertheseheightsandcommonindustryflowmodelsaretechnicallynotvalidinthoselevelsoftheatmosphericboundarylayer(ABL).
Theoryandpracticalmeasurementsshowthatflowfeaturesfortallhubheightsdifferfromthewell-knownrelationsofthesurfacelayer.
Thedemonstrationofconsequentmodellingerrorsofcommonindustryflowmodelsandsomeimmediateremediesconcludethisanalysis.
Resultsfromsixdifferentflowmodelscomparedtoremotesensingmeasurementsatasiteinsimpleterraindemonstratemodellingerrors(Fig.
8).
Inthisexamplemeasureddatafrom100mservedasareference.
Allmodelsunderestimatethewindresourceabove100mwhilstmostofthemachievegoodprofilematchesforlowerheights.
CFDresultsandanalyticalprofilesaccordingtoGryning[3]usingdifferentassumptionsfortheABLheightarecomparedtomeasurements(Fig.
9).
AgaintheCFDmodelhasdifficultiespredictingtheflowabove100m.
TheanalyticalprofilesdependontheABLheight,aparameterwhichisnotavailablefromstandardmastmeasurements.
Thus,thecorrectorincorrectassumptionoftheABLheighthasadecisiveinfluenceontheirperformance.
Highmeasurementscanserveasanimmediateremedy:InFig.
10itcanbeobservedthatmeasurementswithinthesurfacelayerdonotsufficeforareliableextrapolationtohighhubheights.
Ifconverselymeasurementsuptoorevenabovehubheightareavailable,theuncertaintyofmodelbasedverticalextrapolationfortallturbinescanbereducedsignificantly.
Furthermorehighmeasurementscanhelpinapropermodelsetup.
Fig.
11showsastandardCFDmodelwhichroughlymatchesmeasurementsbelow80m,butperformspoorlyabove.
Usingavailablehighmeasurementsthemodelsettingscanbeadjustedsotoachievesatisfyingresultsevenabove100m.
AbstractPO.
ID195ModellingErrorsTheoryRemediesMeasurementsConclusionEWEA2013,Vienna,Austria:Europe'sPremierWindEnergyEvent[1]Adaptedfromhttp://commons.
wikimedia.
org/wiki/File:Atmospheric_boundary_layer.
svg[2]Wilson,J.
D.
,Flesch,T.
K.
,2004:Anidealizedmeanwindprofilefortheatmosphericboundarylayer.
Bound.
-LayerMeteorol.
110,281-299.
[3]Gryning,S-E.
,Jrgensen,H.
,Larsen,S.
,Batchvarova,2007:Thewindprofileupto300metersoverflatterrain.
J.
Phys.
:Conf.
Ser.
75.
[4]Kelly,M.
,Gryning,S-E.
,2010:Long-TermMeanWindProfilesBasedonSimilaritytheory.
Bound.
-LayerMeteorol.
136,377-390.
TallhubheightsreachintotheEkmanlayer,wherethecommonscalingassumptionsofthesurfacelayerarenotvalidanymore.
Attemptstodescribethewindprofileaboveapprox.
100mrequiremeteorologicalparameterswhicharenotavailablefromstandardmastmeasurements.
Thus,assumptionshavetobemadewhicharehardtoverifyforpracticalcases.
Dependingonthesitespecificconditions,thesefactsmayleadto(severely)falseestimatesofthewindconditionsathubheight.
Whenusingflowmodelswhichhavebeenvalidatedwellinthesurfacelayertoestimatethewindconditionsfortallhubheights,remotesensingmeasurementsreachingheightsaboveareessentialinreducingextrapolationerrorstoacceptablelevels.
Theycanalsoaidinthemodelsetup,thusachievingabetteroverall"modelfit".
HighmeasurementsarefinallyaprerequisitefordevelopingandvalidatingnewparameterizationsforABLflow,specificallyforareasabovethesurfacelayer.
SitesinsimpleterrainSiteincomplexterrainFig.
2:Commonpowerlawfitstendtounderestimatethewindspeedatheightsabove100meter.
Fig.
3:Orographyinduceslowshear,sowindspeedathighheightsisoverestimatedwithpowerlawfit.
WindspeedWindshearWindveerFig.
6:Variationofwindveerbetweentopandlowestmeasurementheight(averageandmaxvalues).
Fig.
7:WindvectorplottedoverheightrevealsthetypicalEkmanspiralforthis10-minsample.
Fig.
8:Sixdifferentmodelresultscomparedtoremotesensingmeasurements.
Themeasureddatahasbeenfilteredforneutralatmosphericconditions.
Fig.
9:CFDmodelresultsandanalyticalprofilesaccordingto[3]fortwoestimatesoftheABLheight(h).
Themeasureddatahasbeenfilteredforneutralatmosphericconditions.
Fig.
10:Verticalextrapolationsbasedonlowermeasurementheightscomparedtoactualmeasurements.
Fig.
11:Comparisonofflowmodelsetups(standard/adaptedforhighermeasurements)Fig.
4:Measurementsshowhowthewindshearvaries–andincreases–overheight.
Fig.
5:Thesheardecreaseswithheightatthissite,andtheextentdependsstronglyonthewinddirection.
Fig.
1:ABLwithitscharacteristicdiurnalfeatures[1].
0501001502002504681012height(m)windspeed(m/s)simplesite1simplesite2simplesite3Pow.
fits0501001502002504681012height(m)windspeed(m/s)complexsitePow.
fit0501001502000,100,200,300,40height(m)shear(Hellmann-alpha)site1site2site300,10,20,30,4030609012015018021024027030033080-100m100-160m0204060801001201401601802000246810height(m)windspeed(m/s)measurementsCFDresultsGryningh=2000mGryningh=300m0501001502002500,50,70,91,11,31,5height(m)relativewindspeed(-)measurementsCFDstandardCFDadapted4060801001201401601802004,55,56,57,58,5height(m)windspeed(m/s)measurementsmodel1model2model3model4model5model6AcknowledgementsTheauthorsherebythanktheMeteorologicalInstituteattheKarlsruheInstituteofTechnologyforprovidingthemeasurementdataforFig.
9.
0501001502002505791113height(m)windspeed(m/s)simplesitebelow100msimplesiteabove100mcomplexsitebelow100mcomplexsiteabove100m