CommentaryAnOverviewofIntegrationCostsofVariableRenewablesinthePowerSectorJune2023FarisF.Aljamed,FrankA.Felder,AmroM.ElshurafaIntroductionInrecentyears,thecostofelectricitygeneratedbyrenewableshassignificantlydecreasedtothepointwhereitiscostcompetitivewithconventionalplants(HeptonstallandGross,2021).However,incorporatingvariablerenewableenergy(VRE)technologieslikewindandsolarphotovoltaics(PV)intothepowersystemcreatesintermittency.Byaddingintermittentsourcesofgeneration,demandandsupplywillnotalwaysmatch.Thismeansthatthesystemrequiresmorebackupgeneratorsandmoreflexibilitytobalancethemismatchbetweensupplyanddemand.IntegrationcostsmustbefactoredintodeterminetheoptimalshareofVREandthetotalsystemcost.Traditionally,thelevelizedcostofenergy(LCOE)hasbeenusedtocomparethecostsofdifferentpowergenerators.LCOEcalculatesthetotallifetimediscountedcostsofconstructingandoperatingaplantanddividesitbytheprojectedtotalenergyproducedduringitsassumedlifespan.However,LCOEdoesnotconsiderthecoststhatariseduetoVREintermittencyortheexpensesassociatedwithadaptingthepowersystemtothechangesbroughtbyVRE.Therefore,LCOEalonecannotcapturethetotalsystemcostswhenVREisintroducedtothegrid(Lothetal.,2022).ThiscommentaryprovidesabriefoverviewofhowVREintegrationcostsarecalculatedbyexaminingvariousmethodsavailableintheliterature.IntegrationcostsDefinitionofIntegrationCostsandarecategorizedintotheirComponentsthreecomponents:balancingcosts,gridIntegrationcostsarecategorizedintothreecomponents:balancingcosts,gridcosts,andprofilecosts.Balancingcostsrefertothecostsimposedcosts,andprofilebytheunpredictablenatureofVREgeneration.Supplyuncertaintycausescostsday-aheadforecastingerrors,whichnecessitateoperatingreservesand/orstoragetobalancesupplyanddemand.GridcostsresultfromVREregion-specificrequirements.VREtechnologiesarelessflexiblethanconventionalgenerationintermsofwheretheycanbebuilt.Sometimes,VREgeneratorsarelocatedfarfromloadcenters,requiringadditionaltransmissioninfrastructuretodeliverenergy.ProfilecostsaremostlyduetothevariablenatureofVRE(Ueckerdtetal.2013).Profilecostswerepreviouslyreferredtoas‘adequacycosts.’Adequacycostsaretheexpensesattributabletothelow-capacitycreditofVRE.Conventionalgenerationcapacityisconsidered‘firm’capacity,alwaysreadytomeetdemand,whichisnotthecaseforVRE.Asaresult,capacitycostsincreaseasmoreVREisintegratedintothesystem.ProfilecostsareamorecomprehensiveconceptthatcapturesallcostsimposedbyVREvariability(HeptonstallandGross2021).Profilecostscomprisethreecomponents.Thefirstcomponentisoverproductioncosts,whicharethecostsarisingfromthecurtailmentrequiredforover-generatedpower.Thesecondcomponentisbackupcosts,whicharethecostsofbackupcapacityneededtobalancesupplyAnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector2Figure1.Hierarchyofintegrationcostsandtheircomponents.Source:Ueckerdtetal.(2013).anddemand.Thethirdcomponentisfull-loadhour(FLH)reductioncosts.VREsreducetheFLHofdispatchableplants,resultinginlowergenerationpercapacityfortheseplants(Ueckerdtetal.2013).Figure1summarizestheintegrationcosts.ReviewofHowIntegrationCostsofVariableRenewableEnergyAreCalculatedLoaddurationcurvesmethodOnemethodtoassessintegrationcostsistheloaddurationcurve(LDC)method.AnLDCdisplaysthehourlyloadofayear,sortedfromthehighestloadhourtotheleastloadhour.When(VRE)isadded,theLDCischangedtoresidualloaddurationcurve(RLDC),whichshowshowmuchelectricitydemandisleftaftersubtractingthesupplyfromrenewableresources.TodeterminetheresidualcostsforthesystemwithVRE,oneneedstointegratealongtheinverseoftheRLDCandmultiplythevaluebytherespectiveminimumscreeningcurvevalue.ForthesystemwithoutVRE,theintegrationisalongtheinverseoftheLDC.Screeningcurvesshowthetotalcostperkilowatt(kW)peryearofdifferentgenerationtechnologies.Figure2displaysanexampleofanLDCandanRLDC.AnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector3Figure2.ConceptualschematicexplainingthedifferencebetweenLDCandRLDC.Source:Authors’illustration.Ueckerdtetal.(2013)introducedthesystemLCOEmetric,whichisthesumoftheplant’smarginalgenerationcostsandmarginalintegrationcosts.TheauthorsdividedthecostsofthesystemintoVREgenerationcostandresidualcosts.TheresidualcostsarethegenerationcostsofconventionalplantsandtheintegrationcostsofVRE.Theauthorscomparedtheresidualcostsoftwosystems:onewithVREandonewithout.SincethesystemwithoutVREhasnointegrationcosts,comparingtheresidualcostsofthetwosystemsisolatestheVREintegrationcosts.Theintegrationcostsaredefinedasthedifferencebetweenthespecificcostsperunitofresidualloadofthetwosystemsmultipliedbytheresidualgeneration.Ueckerdtetal.estimatedbalancingandgridcostsfrompreviousstudiesandcalculatedprofilecosts.Forwindsharesfrom5%to30%,balancingcostsrangefrom2.5to5eurospermegawatthour(€/MWh),andgridcostsarearound13€/MWhfor40%windpenetration.Profilecostsreachabout30€/MWhat30%windpenetration,andoverallintegrationcostscangoupto60€/MWhat40%windpenetration.Integrationcostscanbereducedbyintroducingoptionssuchaslongdistanceinterconnection,storage,anddemandmanagement.Notethatthestudybeingrevieweddoesnotoptimizetheenergymix.Theonlyoptionconsideredwasthecapacitymixofresidualpowergenerationbythermalgenerators.Thus,theprofilecostscalculatedareoverestimated.AnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector4CostproductionmodelmethodAnothermethodtoassessintegrationcostsisthecostproductionmodelmethod.Here,themodelercomparesascenariowithoutrenewablestoanotherwithrenewables.Thedifferenceincostbetweenthetwoscenarioswouldbetheintegrationcosts.Themodelscanbebuiltusingstandardsoftwareorprogramminglanguages.Forinstance,Brouweretal.(2015)usedPLEXOS,acommerciallyavailablesoftwarepackagethatmodelspowersystems,tosimulatethepowersectorforWesternEuropein2050.Fordifferentpenetrationlevels,theauthorsconsideredfivecomplementaryoptionstointegrateVREatthelowestcost:demandresponse(DR),gas-firedpowerplantswithandwithoutcarboncapture,increasedinterconnectioncapacity,curtailment,andelectricitystorage.PLEXOSoptimizesunitcommitmentandeconomicdispatchwhilemeetingfiveconstraints:balancingelectricitysupplyanddemand,flexibilityconstraintsofgenerators,limitedtransmissioncapacityforinterconnections,scheduledandunscheduledoutages,andthebalancingreservesrequirements.Profilecostswerecalculatedforvariablerenewableenergy(VRE)penetrationlevelsbetween22%and59%,withvaluesbetween0%and22%linearlyinterpolated.TheincreaseinprofilecostsduetoVREadditionismainlyattributedtotwofactors:thereductioninthecapacityfactorofthermalgeneratorscausedbyincreasedVRE,andtheneedformorecurtailmentduetooverproductionfromrenewables(Brouweretal.2015).Marginalprofilecostsrangedfrom0€/MWhto100€/MWhforpenetrationlevelsof0%to60%.Upto40%penetration,integrationcostsincreasedlinearly,reachingapproximately30€/MWh.However,afterthe40%mark,integrationcostsstartedtogrowexponentially.Reichenbergetal.(2018)focusedontheintegrationcostsofVREinEuropebydividingitinto10regions.Theyusedanelectricityinvestmentmodelthataccountsforvariabilityandvariationmanagementtooptimizethedispatchandinvestmentingeneration,storage,andtransmissionforallpenetrationlevels.Theauthorscalculatedthesystemlevelizedcostofelectricity(LCOE)usingthesamedefinitionasUeckerdtetal.ThemarginalsystemLCOEincreasedlinearlyasVREpenetrationincreased,witharateof6€/MWhforeach10%increaseuntilreaching80%penetration.Afterthatpoint,themarginalsystemLCOEstartedtoincreaseexponentiallyduetoallocatingVREinregionswithlowercapacityfactorsandtheneedtocurtailorstoreexcessenergy.Xietal.(2022)calculatedtheintegrationcostsforthepowersystemintheJilinprovinceofChina,comparingasystemwithnoVREgenerationtoasystemwithVREgeneration.Duetothecoal-dominatednatureoftheJilinpowersystem,itexperiencedarapidincreaseinintegrationcostsatanearlierpenetrationlevelcomparedtootherpowersystems.Yaoetal.(2020)simulatedthepowersystemofGuangdongprovinceinChinaandfoundthatintegrationcostsforwindandsolarPVrangedfrom-2.18€/MWhto11.47€/MWhand-5.21€/MWhto6.73€/MWh,respectively,forpenetrationlevelsupto30%.AnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector5Overall,theOverall,theintegrationcostsofVREvarydependingonthepenetrationintegrationcostslevel,systemflexibility,andthespecificcharacteristicsofthepowersystembeinganalyzed.WhilemarginalsystemLCOEandincrementalofVREvaryoperatingcostsofthermalplantstendtoincreaselinearlywithhigherdependingontheVREpenetration,curtailmentcosts,idlecosts,andbalancingcostscanpenetrationlevel,decreaseorremainconstantinmoreflexiblepowersystems.systemflexibility,andthespecificThetwostudieswediscussedonChinaonlyconsiderintegrationcostscharacteristicsoftheforpenetrationlevelsupto40%.However,toincreasethedeploymentpowersystembeingofVREintheChinesepowersystem,bettersystemflexibilityisneeded.Ruetal.(2022)proposethattheChinesepowersystemcanachieveanalyzedVREpenetrationlevelsbetween70%and85%byimplementingvariationmanagementoptionssuchasdifferentenergystoragetechnologyandultra-highvoltagedirectcurrent-based(UHVDC-based)transmission.DiscussionInUeckerdtetal.’smodel,theintegrationcostsstartedtoincreaseatahigherrateatlowerpenetrationlevelsthaninotherstudies.Forwind,thejumpoccurredat25%penetration,whileforPV,itoccurredat15%penetration.Reichenbergetal.suggestthatthereasonbehindthisrelativelyearlyjumpinintegrationcostsisduetotheabsenceofvariationmanagementsolutionssuchastrade,storage,demandresponse,orcomplementarityofwindandsolar.Brouweretal.calculatedtheintegrationcostsofVREforupto60%penetration.ThesharpincreaseinprofilecostshappenedlaterthaninUeckerdtetal.’sstudy.Brouweretal.’smodelisimplementedacrossEurope,notjustinGermany.Thisgivesitawiderscopethataccountsfortradebetweenregions.However,onedownsidetothemodelisthatVREcapacityandtransmissionlocationswerenotoptimized,andthesharpincreaseinprofilecostsoccursataround40%penetrationduetoreducedFLHandcurtailmentcosts.IntegrationcostsinReichenbergetal.startincreasingsharplyatamuchhigherpenetrationlevelthanthatofpreviousstudies,whichhappensataround80%penetration.Theauthorsstatethatthevaluesoftheintegrationcostaremostlyattributedtocostassumptions,whilethepointatwhichthecostsstarttoincreasesharplystemsfromsystemdynamics.ThisstudyshowsthebenefitsofaccountingforvariationmanagementoptionsandhowtheyaffectthelinearincreaseofintegrationcostswithrespecttoVREpenetrationatsmallshares.Employingdifferentintegrationoptionscouldalsoprovetobecomplementarytoeachother.Forexample,Auguadraetal.(2023)demonstratethatdemandresponseiscomplementarytoenergystorageandprovidesflexibilityforstoragetechnology.AlimitationofReichenbergetal.’smodelisthatitdoesn’tconsidersometechnicalaspectslikeforecastingerrorsandtheneedforbalancingpowerfromthermalplants.Anotherlimitationisthat,whilethemodelinvestsintransmissionbetweentheregions,transmissionwithineachregionisunaccountedfor.AlimitationofsolarPVsisthatthetimeresolutionisnothourly,whichimpactssolaravailability,asitcanchangedrasticallyAnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector6fromonepointintimetothenext(e.g.,from9a.m.to11a.m.).Finally,aAccordingtothelimitationofwindgenerationistheinterannualvariabilityofwindspeed,literature,balancingwhichwasnotaccountedforinthestudy.costsaregenerallylowwhencomparedAccordingtotheliterature,balancingcostsaregenerallylowwhentoothercomponents,comparedtoothercomponents,withestimatestheirvaluestypicallybelowwithestimatestheir6€/MWh.Whenthetrendlineisfittedtothedata,balancingcostsincreasefrom2€/MWhto4€/MWhforwindpenetrationfrom0%to40%(Hirthetvaluestypicallyal.2014).Hirthetal.findthatgridcostsarealsosmall,andtheyarenotbelow6€/MWhusuallyreportedinmarginalterms.Furthermore,theresultsareusuallynotbasedoncostoptimization.Gridcostsareestimatedtobeintheorderof5€/MWh.Windprofilecostsareestimatedtobenegativeorclosetozeroatlowpenetrationrates.However,athigherpenetrationratesofbetween30%and40%,profilecostsforwindareestimatedtobearound15to25€/MWh(Hirthetal.2014).AnotherstudythatreviewedpastliteratureestimateswasconductedbyHeptonstallandGross(2021).Theyestimatedthatadditionalcostsforoperatingreserves(usedtobalancesupplyanddemand)arebelow5€/MWhforupto35%penetrationandbelow10€/MWhforpenetrationlevelsupto45%,withthesizeofthesecostsdependingontheflexibilityofthesystem.Adequacycosts,whichareaspecifictypeofprofilecost,areestimatedtobearound10€/MWhorlessforallpenetrationlevels.Profilecostsareestimatedtorangefrom15to25€/MWhat25%to35%penetration.Theauthorsalsoestimatedgridcoststobeintherangeof7to28€/MWh.However,theynotedthatestimatesforthesecostsvarywidelyacrosstheliterature,anditischallengingtoattributeallgridandtransmissionupgradestothevariablegenerationofVRE.Table1,below,summarizesthecostsdiscussedinthissection.Table1.Anestimateofthecostscalculatedinthestudiescovered.StudyMethodLocationPenetrationBalancingGridcostsProfilecostsPointofpercentagecosts(€/MWh)(€/MWh)(€/MWh)exponentialUeckertdLDCGermanyincreaseetal.calculations0%-40%(wind)2.5-5(wind)0-13(wind)-5-60(wind)25%(wind)BrouwerEurope0%-25%(Solar)-10-100(solar)15%(solar)etal.CostproductionReichenbergmodelEurope0%-60%0.2-1N/A0-10040%etal.80%CostproductionChina0%-100%N/A0-11020%Xietal.modelN/AChina0%-40%IncludedwithN/A0-16.8Yaoetal.CostproductionUSAandprofileN/AmodelEuropean0%-30%Hirthetal.countries-2.18-11.47(wind)N/ACostproductionEurope,0%-40%-5.21-6.73(solar)HeptonstallmodelUSA,andGrossAsia0%-45%(Balancing)2-450-25Costproduction0%-35%(Profile)modelsN/A(Grid)0-107-280-25CostproductionmodelsAnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector7ConclusionTherearegenerallytwomethodsforcalculatingintegrationcosts.Thefirstmethodwecoveredusestheloaddurationcurveandresidualloaddurationcurvetocomparetheresidualcostsofanon-VREsystemandasystemwithVRE.Thesecondmethodusesproductioncostmodels,whichisthemostcommonlyusedmethodduetoitshigheraccuracyanddescriptivepower.Wealsosawthateachsystemhasitsowncharacteristicsandintegrationchallenges,andthereforerequiresadedicatedstudy.Theproductioncostmethodisthemostcommonlyusedmethodtoestimateintegrationcostsduetoitshigheraccuracy,despiteitbeingmoredataandmodelingheavy.Althoughthecalculationsperformedondifferentsystemswerenotexactlythesame,wegenerallyobservesimilartrends.Balancingcostsareusuallybelow10€/MWhandaretypicallyinthesingledigits.Whileliteratureongridcostsisscarce,thesecostsareusuallygreaterthanbalancingcosts.Profilecostsarealwaysthelargestcomponentofintegrationcosts,andtheytypicallyamounttoaround25€/MWhat35%to40%penetration.Insomecases,theycanevenbehigherifthesystemdoesnotaccountforappropriatevariationmanagementsolutions.AtlowerVREpenetrationlevels,integrationcostscanbeclosetozeroorevennegativeinsomecases.However,asmoreVREgenerationisaddedtothesystem,thesecostsincreaserapidly.Dependingonthesystemcharacteristicsandintegrationoptionsconsidered,thelevelofincreaseinintegrationcostsvaries.Moreover,thepointatwhichthesecostsbegintoincreaseexponentiallyalsodependsontheaforementionedfactors.Forflexiblesystems,integrationcostsstarttoincreaseexponentiallyatpenetrationlevelsabove40%.Ifthesystemalsoimplementsvariationmanagementsolutions,thenthepointofexponentialcostincreasescanbedelayedtoupto80%penetration.TheseresultsmustbeconsideredwhenVREtechnologiesareintegratedintothepowersystem.Factorssuchassystemflexibilityandinterconnectivity,forexample,needtobecarefullyevaluatedtoensureasmoothtransitiontothetargetedVREpenetrationlevel.Furthermore,keepinginmindthepowersystem’scharacteristicandchallengesisoftheutmostimportancetosuccessfullyselecttheappropriateintegrationoptions.AlthoughtheseconsiderationswillnotaffecttheLCOEofVRE,thesystemLCOE(theoverallsystemmarginalcosts)willbedecreased.AnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector8ReferencesAuguadra,Marco,DavidRibó-Pérez,andTomásGómez-Navarro.2023.“Planningthedeploymentofenergystoragesystemstointegratehighsharesofrenewables:TheSpaincasestudy.”Energy264:126275.Brouwer,AnneSjoerd,MachteldvandenBroek,WilliamZappa,WimC.Turkenburg,andAndréFaaij.2016.“Least-costoptionsforintegratingintermittentrenewablesinlow-carbonpowersystems.”AppliedEnergy161:48-74.Heptonstall,PhilipJ.,andRobertJKGross.2021.“Asystematicreviewofthecostsandimpactsofintegratingvariablerenewablesintopowergrids.”NatureEnergy6(1):72-83.Hirth,Lion,FalkoUeckerdt,andOttmarEdenhofer.2015.“Integrationcostsrevisited–Aneconomicframeworkforwindandsolarvariability.”RenewableEnergy74:925-939.Joos,Michael,andIainStaffell.2018.“Short-termintegrationcostsofvariablerenewableenergy:WindcurtailmentandbalancinginBritainandGermany.”RenewableandSustainableEnergyReviews86:45-65.Li,Ru,Bao-JunTang,BiyingYu,HuaLiao,ChenZhang,andYi-MingWei.2022.“Cost-optimaloperationstrategyforintegratinglargescaleofrenewableenergyinChina’spowersystem:Fromamulti-regionalperspective.”AppliedEnergy325:119780.Loth,Eric,ChrisQin,JulietG.Simpson,andKatherineDykes.2022.“WhywemustmovebeyondLCOEforrenewableenergydesign.”AdvancesinAppliedEnergy8:100112.Reichenberg,Lina,FredrikHedenus,MikaelOdenberger,andFilipJohnsson.2018.“ThemarginalsystemLCOEofvariablerenewables–EvaluatinghighpenetrationlevelsofwindandsolarinEurope.”Energy152:914-924.Ueckerdt,Falko,LionHirth,GunnarLuderer,andOttmarEdenhofer.2013.“SystemLCOE:Whatarethecostsofvariablerenewables?”Energy63:61-75.Xi,Xingxuan,WeirongZhang,YanleiZhu,JianZhang,andJiahaiYuan.2022.“WindintegrationcostinChina:Aproductionsimulationapproachandcasestudy.”SustainableEnergyTechnologiesandAssessments51:101985.Yao,Xing,BowenYi,YangYu,YingFan,andLeiZhu.2020.“Economicanalysisofgridintegrationofvariablesolarandwindpowerwithconventionalpowersystem.”AppliedEnergy264:114706.AnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector9AbouttheProjectTheKingdomintendstoreduceoreliminateitsuseofliquidfuelsbydeployingconsiderablerenewableenergycapacity.TheAssessmentoftheChangingEconomicsoftheSaudiElectricityIndustryprojectassessestheimpactsofdeployingrenewablesontheSaudipowersectorintermsofcosts,reliability,emissions,andnaturalgasconsumption.AnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector10AboutKAPSARCKAPSARCisanadvisorythinktankwithinglobalenergyeconomicsandsustainabilityprovidingadvisoryservicestoentitiesandauthoritiesintheSaudienergysectortoadvanceSaudiArabiaʼsenergysectorandinformglobalpoliciesthroughevidence-basedadviceandappliedresearch.LegalNotice©Copyright2023KingAbdullahPetroleumStudiesandResearchCenter(“KAPSARC”).ThisDocument(andanyinformation,dataormaterialscontainedtherein)(the“Document”)shallnotbeusedwithouttheproperattributiontoKAPSARC.TheDocumentshallnotbereproduced,inwholeorinpart,withoutthewrittenpermissionofKAPSARC.KAPSARCmakesnowarranty,representationorundertakingwhetherexpressedorimplied,nordoesitassumeanylegalliability,whetherdirectorindirect,orresponsibilityfortheaccuracy,completeness,orusefulnessofanyinformationthatiscontainedintheDocument.NothingintheDocumentconstitutesorshallbeimpliedtoconstituteadvice,recommendationoroption.TheviewsandopinionsexpressedinthispublicationarethoseoftheauthorsanddonotnecessarilyreflecttheofficialviewsorpositionofKAPSARC.AnOverviewofIntegrationCostsofVariableRenewablesinthePowerSector11www.kapsarc.org
VIP
