儲能未來研究-未來幾十年的主要學(xué)識

StorageFuturesStudy

KeyLearningsfortheComingDecades

NateBlair,ChadAugustine,WesleyCole,PaulDenholm,WillFrazier,MadelineGeocaris,JennieJorgenson,KevinMcCabe,KaraPodkaminer,AshreetaPrasanna,BenSigrin

StorageFuturesStudy

KeyLearningsfortheComingDecades

NateBlair,ChadAugustine,WesleyCole,PaulDenholm,WillFrazier,MadelineGeocaris,JennieJorgenson,KevinMcCabe,KaraPodkaminer,AshreetaPrasanna,BenSigrin

SUGGESTEDCITATION

Blair,Nate,ChadAugustine,WesleyCole,etal.2022.StorageFuturesStudy:KeyLearningsfortheComingDecades.Golden,CO:NationalRenewableEnergyLaboratory.NREL/TP-7A40-81779.

/docs/fy22osti/81779

.pdf

StorageFuturesStudy:KeyLearningsfortheComingDecades|iii

NOTICE

ThisworkwasauthoredinpartbytheNationalRenewableEnergyLaboratory,operatedbyAllianceforSustainableEnergy,LLC.fortheU.S.DepartmentofEnergy(DOE)underContractNo.DE-AC36-08GO28308.SupportfortheworkwasalsoprovidedbytheInterstateRenewableEnergyCouncil,Inc.underAgreementSUB-2021-10440.TheviewsexpressedinthearticledonotnecessarilyrepresenttheviewsoftheDOEortheU.S.Government.TheU.S.Governmentretainsandthepublisher,byacceptingthearticleforpublication,acknowledgesthattheU.S.Governmentretainsanonexclusive,paid-up,irrevocable,worldwidelicensetopublishorreproducethepublishedformofthiswork,orallowotherstodoso,forU.S.Governmentpurposes.

ThisreportisavailableatnocostfromtheNationalRenewableEnergyLaboratory(NREL)at

/publications.

U.S.DepartmentforEnergy(DOE)reportsproducedafter1991andagrowingnumberofpre-1991documentsareavailablefreevia

www.OSTI.gov

.

StorageFuturesStudy:KeyLearningsfortheComingDecades|iii

PREFACE

ThisreportistheseventhandfinalpublicationfromtheNationalRenewableEnergyLaboratory’s(NREL’s)StorageFuturesStudy(SFS).TheSFSisamultiyearresearchprojectthatexploreshowenergystoragecouldimpacttheevolutionandoperationoftheU.S.powersector.

Thestudyexaminedtheimpactofenergystoragetechnologyadvancementonthedeploymentofutility-scalestorageandtheadoptionofdistributedstorage,aswellasfuturepowersysteminfrastructureinvestmentandoperations.SomeofthequestionsNRELsoughttoanswerthroughoutthisstudyincluded:

?Howmightstoragecostandperformancechangeovertime?

?Whatistheroleofdiurnalenergystorageinthepowersector,evenabsentdriversorpoliciesthatincreaserenewableenergyshares?

?HowmuchdiurnalgridstoragemightbeeconomicallydeployedintheUnitedStates,bothattheutility-scaleanddistribution-scale?

?Whatfactorsmightdrivethatdeployment?

?Howmightincreasedlevelsofdiurnalstorageimpactgridoperations?

Researchfindingsandsupportingdatafromthestudyhavebeenpublishedinaseriesofsevenpublications,whicharelistedinthetableonthenextpage.Keylearningsfromthroughoutthestudyhaveculminatedinthisfinalreportthathelpsshapethevisionofenergystoragemovingforward.

TheSFSseriesprovidesdataandanalysisinsupportoftheU.S.DepartmentofEnergy’s(DOE’s)

EnergyStorageGrand

Challenge

,acomprehensiveprogramtoacceleratethedevelopment,commercialization,andutilizationofnext-generationenergystoragetechnologiesandsustainAmericangloballeadershipinenergystorage.TheEnergyStorageGrandChallengeemploysause-caseframeworktoensurestoragetechnologiescancost-effectivelymeetspecificneeds,anditincorporatesabroadrangeoftechnologiesinseveralcategories:electrochemical,electromechanical,thermal,flexiblegeneration,flexiblebuildings,andpowerelectronics.

Moreinformation,supportingdataassociatedwiththisreport,linkstootherreportsintheseries,andotherinformationaboutthebroaderstudyareavailableat

/analysis/storage-futures.html.

iv|StorageFuturesStudy:KeyLearningsfortheComingDecades

Table1

StorageFutureStudySeriesReports

Title

Description

RelationtoThisReport

TheFourPhasesof

Explorestherolesandopportunitiesfornew,

Providesbroadercontexton

StorageDeployment:

cost-competitivestationaryenergystorage

theimplicationsofthecostand

AFrameworkforthe

withaconceptualframeworkbasedonfour

performancecharacteristicsdiscussed

ExpandingRoleof

phasesofcurrentandpotentialfuturestorage

inthisreport,includingspecificgrid

StorageintheU.S.Power

deploymentandpresentsavaluepropositionfor

servicestheymayenableinvarious

System(Denholmetal.

energystoragethatcouldresultincost-effective

phasesofstoragedeployment.This

2020)

deploymentsreachinghundredsofgigawattsofinstalledcapacity.

frameworkissupportedbytheresultsofscenariosinthisproject.

EnergyStorage

Reviewsthecurrentcharacteristicsofa

Providesdetailedbackgroundabout

TechnologyModeling

broadrangeofmechanical,thermal,and

thebatteryandpumpedstorage

InputDataReport

electrochemicalstoragetechnologieswith

hydropowercostandperformance

(Augustineetal.2021)

applicationtothepowersector.Providescurrentandfutureprojectionsofcost,performancecharacteristics,andlocationalavailabilityofspecificcommercialtechnologiesalreadydeployed,includinglithium-ionbatterysystemsandpumpedstoragehydropower.

valuesusedasinputstothemodelingperformedinthisproject.

EconomicPotentialof

Assessestheeconomicpotentialforutility-scale

Thisreportfeaturesaseriesofcost-

DiurnalStorageinthe

diurnalstorageandtheeffectsthatstorage

drivengrid-scalecapacityexpansion

U.S.PowerSector(Frazier

capacityadditionscouldhaveonpowersystem

scenariosfortheU.S.gridthrough2050

etal.2021)

evolutionandoperations.

andexaminesthedriversforstoragedeployment.

DistributedStorage

Assessesthecustomeradoptionofdistributed

Analyzesdistributedstorageadoption

CustomerAdoption

diurnalstorageforseveralfuturescenariosand

scenariostotestthevariouscost

Scenarios(Prasannaet

theimplicationsforthedeploymentofdistributed

trajectoriesandassumptionsinparallel

al.2021)

generationandpowersystemevolution.

tothegridstoragedeployments

modeledinthisreport.

TheChallengesof

Describesthechallengeofasingleuniform

Advancesdialoguearoundthemeaning

DefiningLong-Duration

definitionforlong-durationenergystorageto

oflong-durationenergystorageand

EnergyStorage

reflectbothdurationandapplicationofthe

howitfitsintofuturepowersystems.

(Denholmetal.2021)

storedenergy.

GridOperational

Assessestheoperationandassociatedvalue

Considerstheoperationalimplicationsof

Implicationsof

streamsofenergystorageforseveralpower

storagedeploymentandgridevolution

WidespreadStorage

systemevolutionscenariosandexplores

scenariostoexamineandexpandonthe

Deployment(Jorgenson

theimplicationsofseasonalstorageongrid

grid-scalescenarioresultsfoundwith

etal.2022)

operations.

NREL’sRegionalEnergyDeploymentSystemmodelinthisreport.

StorageFuturesStudy:

Synthesizesandsummarizesfindingsfromthe

Thisreport.

KeyLearningsForthe

entireseriesandrelatedanalysesandreports

ComingDecades

andidentifiestopicsforfurtherresearch.

StorageFuturesStudy:KeyLearningsfortheComingDecades|v

ACKNOWLEDGMENTS

WewouldliketoacknowledgethecontributionsoftheentireStorageFuturesStudyteam(listedascoauthors)forthisreport,aswellasourDOEOfficeofStrategicAnalysiscolleaguesascorecontributors,specificallyKaraPodkaminer,PaulSpitsen,andSarahGarman.FeedbackandcontributionsalsocamefromotherNRELstaff,includingGianPorro,DougArent,KarlynnCory,AdamWarren,ChadHunter,EvanReznicek,MichaelPenev,GregStark,VigneshRamasamy,DavidFeldman,GregBrinkman,andTrieuMai.Wealsowouldliketothankourtechnicalreviewcommittee(seeTable2)fortheirinput.

Finally,weacknowledgevarioustechnicalexpertsatDOE,includingEricHsieh,AlejandroMoreno,andmanyothers,fortheiradditionalthoughtsandsuggestionsthroughouttheStorageFuturesStudy,asnotedintheindividualreports.

Table2

TechnicalReviewCommitteeMembers

DougArent

(NREL)–TRCChair

PaulAlbertus

(UniversityofMaryland)

InezAzevedo

(StanfordUniversity)

RyanWiser

(LawrenceBerkeley

NationalLaboratory)

SueBabinec(ArgonneNationalLaboratory)

AaronBloom

(NextEra)

ChrisNamovicz

(U.S.EnergyInformation

Administration)

HowardGruenspecht

(MassachusettsInstitute

ofTechnology)

ArvindJaggi

(NYIndependentSystemOperator)

KeithParks

(XcelEnergy)

KiranKumaraswamy

(Fluence)

GrangerMorgan(CarnegieMellonUniversity)

CaraMarcy

(U.S.Environmental

ProtectionAgency)

MaheshMorjaria

(TerabaseEnergy)

OliverSchmidt(ImperialCollege-London)

VincentSprenkle

(PacificNorthwest

NationalLaboratory)

JohnGavan(ColoradoPUCCommissioner)

vi|StorageFuturesStudy:KeyLearningsfortheComingDecades

LISTOF

BESS

DOE

DR

FC

GW

GWh

H2

H2Elec-saltcavern-CT

H2Elec-saltcavern-FC

kW

kWh

LIB

NG

NREL

PV

RE

SFS

VRE

ACRONYMS

—batteryenergystoragesystem(s)

—U.S.DepartmentofEnergy

—distributedresource

—fuelcell

—gigawatts

—gigawatt-hour

—hydrogen(asastoragefluid)

—hydrogenstorageusingelectrolyzers,saltcaverns,andcombustionturbines

—hydrogenstorageusingelectrolyzers,saltcaverns,andstationaryfuelcells

—kilowatt

—kilowatt-hour(eitheraunitofenergyoraunitofstoragecapacity)

—lithium-ionbattery

—naturalgas

—NationalRenewableEnergyLaboratory

—photovoltaics

—renewableenergy

—StorageFuturesStudy

—variablerenewableenergy

StorageFuturesStudy:KeyLearningsfortheComingDecades|vii

TABLEOFCONTENTS

TheComingDecadesofEnergyStorageDeployment 1

KEYLEARNING1:StorageIsPoisedforRapidGrowth 3

KEYLEARNING2:RecentStorageCostReductionsAreProjectedToContinue,withLithium-IonBattery

ContinuingToLeadinMarketShareforSomeTime 4

KEYLEARNING3:TheAbilityofStorageToProvideFirmCapacityIsaPrimaryDriver

forCost-CompetitiveDeployment 7

KEYLEARNING4:StorageIsNottheOnlyFlexibilityOption,butItsDecliningCostsHaveChanged

WhenItIsDeployedVersusOtherOptions 8

KEYLEARNING5:StorageandPVComplementEachOther 10

KEYLEARNING6:CostReductionsandtheValueofBackupPowerIncreasetheAdoptionof

Building-levelStorage 12

KEYLEARNING7:StorageDurationsWillLikelyIncreaseasDeploymentsIncrease 13

KEYLEARNING8:SeasonalStorageTechnologiesBecomeEspeciallyImportantfor100%

CleanEnergySystems 14

ConclusionsandRemainingUncertainties 16

References 18

viii|StorageFuturesStudy:KeyLearningsfortheComingDecades

LISTOFFIGURES

Figure1.Nationalstoragecapacityinthereferencecasegrowstoabout200GWby2050,

deployingarangeofdurations(left)

3

Figure2.Lithium-ionbatterypackcostshavedroppedbymorethan80%overthepastdecadeand

areexpectedtocontinuetofallbasedoncontinuedscaleofproduction,drivenlargelybyelectric

vehicledemand

4

Figure3.Theutility-scaleBESSReferenceScenarioprojectscontinuedcostreductions 5

Figure4.Capitalcostforenergy($/kWh)versuscapitalcostforcapacity($/kW)

forvarioustechnologies

6

Figure5.Restrictingservicesthatstoragecanprovideshowscapacityservicesaremoreimportant

thantime-shiftingoroperatingreservestoachievestorage’smaximumpotential 7

Figure6.Theflexibilitysupplycurve

8

Figure7.Increasingloadflexibilityandresponsivedemandreducestheneedforstorage

capacityin2050forthelowREcostandlowRE/batterycostscenarioswithandwithout

highdemandresponsecontribution

9

Figure8.IncreaseddeploymentofPVdemonstratesthereduceddurationofnetloadpeaks 10

Figure9.IncreaseddeploymentofPVdemonstratesthereduceddurationrequiredfor

energystoragetoprovidefirmcapacity

10

Figure10.Nationalpeakingcapacitypotentialfordiurnalstorage(upto12hours)asafunction

ofPVcontribution(left)andnationaldiurnalenergytime-shiftingpotentialasafunction

ofPVcontribution(right)

11

Figure11.Storagecapacityasafunctionofrenewableenergycontribution(%) 11

Figure12.Projectedadoptionofdistributedstorage(GWof2-hourdurationstoragesystemscoupledwith

PV)increasesovertimeascostsdecrease,withasignificantjumpiftherearebreakthroughPVcosts 12

Figure13.Asstoragedeploymentincreases,thenetloadpeakwidens,requiringlonger-duration

storagetoprovidefirmcapacity

13

Figure14.Theaveragedurationofnewstoragedeploymentsincreasesasthetotalamountof

storagecapacitygrows,uptoapproximately200GW(usingreferencestoragecosts) 13

Figure15.Seasonalmismatchofrenewableenergysupplyandelectricitydemand

demonstratesthepotentialopportunityforseasonalstorage

14

Figure16.Capacityandgenerationin2050forthescenariosthatreachthe100%requirement 15

LISTOFTABLES

Table1.StorageFutureStudySeriesReports

v

Table2.TechnicalReviewCommitteeMembers

vi

StorageFuturesStudy:KeyLearningsfortheComingDecades|ix

THECOMINGDECADESOFENERGYSTORAGEDEPLOYMENT

Energystorageisverylikelytobecomeacriticalelementofalow-carbon,flexible,resilientfutureelectricgrid.

Inthepastseveralyears,therehasbeenadramaticincreaseofvariablerenewablegenerationintheU.S.powersector,andsignificantgrowthisanticipatedinthefuture.Inaddition,therehasbeenincreasedfocusintheUnitedStatesandgloballyonaddressingnumerousinstancesofpowersystemdisruptionsandincreasedfocusonresearchandanalysisonpowersystemreliabilityandresiliencywithincreasingamountsofvariablerenewablepower—emphasizingtheimportanceofcleanenergydeploymentwhilemaintainingareliablepowersystem.

Atthesametime,therehavebeensignificantcostdeclinesinenergystoragetechnologies(particularlybatteries)overthepastfewyears,andmanymorestoragetechnologiesareunderdevelopment.Theseconvergingfactorshaveincreasedattentiononthepotentialroleofenergystorageasacriticalassetfordecarbonizationandtoensurereliableelectricityfortheevolvinggrid.

Energystorageoffersmanypotentialbenefitstothegrid.ItcouldprovidegenerationtocomplementthedeploymentofwindandsolarPV,providingcapacitywhentheseresourceshavereducedavailability.Whenusedinconjunctionwithrenewableenergy(RE)orothercleanenergyresources,energystoragehastheabilitytoreducegreenhousegasemissions.

Energystoragecanalsoincreaseutilizationofnewandexistingtransmissionlines,whileoffsettingtheneedtobuildnewpowerplantstoprovidepeakingcapacityoroperatingreserves.Finally,distributedenergystoragecanreducestressonthedistributiongridduringpeakdemandtimes.Thisflexibilitywillbeimportantwiththeanticipatedproliferationofelectricvehiclesandpotentialincreasedloadfromotherend-useelectrification.

Asthecostofenergystoragetechnologiescontinuestodeclineandthegridintegratesmorevariablerenewablegeneration,ourmodelingindicatessignificantincreaseddeploymentofenergystoragedeploymentintheelectricsysteminthecomingdecades.Questionsarise,suchashowcouldthisimpacthowthegridoperatesandevolvesoverthecomingdecades?

Becauseenergystoragecanimpactfeaturesofelectricitygeneration,transmission,anddistribution,quantifyingthevalueofstorageismorecomplicatedthanquantifyingthevalueofotherassetslikesolarPVorwindenergythatarepurelygeneration.ThroughtheStorageFuturesStudy(SFS),theNationalRenewableEnergyLaboratory(NREL)hasaimedtoincreaseunderstandingofhowstorageaddsvalue,andhowmuch,tothepowersystem,howmuchstoragecouldbeeconomicallydeployed,andhowthatdeploymentmightimpactpowersystemevolutionandoperations.

TheStorageFuturesStudystartedwithdefining

aframeworkoffourphasesofincreasingenergy

storagedeploymentanddurationovertime,moved

1|StorageFuturesStudy:KeyLearningsfortheComingDecades

StorageFuturesStudy:KeyLearningsfortheComingDecades|2

ontocreateasetoflong-termprojectionsfordiurnal(<12hours)storagedeploymentintheUnitedStates,andthenapplieddetailedproductioncostandagent-basedmodelingtobetterunderstandtheroleofstorage.Thekeyconclusionoftheresearchisthatdeploymentofenergystoragehasthepotentialtoincreasesignificantly—reachingatleastfivetimestoday’scapacityby2050—anditwillplayanintegral

roleindeterminingthecost-optimalgridmixofthefuture.DrawingontheanalysisacrosstheSFS,previouswork,andadditionalanalysisforthisreport,thestudyidentifiedeightspecifickeylearningsaboutthefutureofenergystorageanditsimpactonthepowersystem.Thesekeylearningscanhelppolicymakers,technologydevelopers,andgridoperatorsprepareforthecomingwaveofstoragedeployment:

KEYLEARNING1:

Storageispoisedforrapidgrowth.

KEYLEARNING2:

Recentstoragecostreductionsareprojectedtocontinue,withlithium-ionbatteries(LIBs)continuingtoleadinmarketshareforsometime.

KEYLEARNING3:

Theabilityofstoragetoprovidefirmcapacityisaprimarydriverofcost-competitivedeployment.

KEYLEARNING4:

Storageisnottheonlyflexibilityoption,butitsdecliningcostshavechangedwhenitisdeployedversusotheroptions.

KEYLEARNING5:

Storageandphotovoltaics(PV)complementeachother.

KEYLEARNING6:

Costreductionsandthevalueofbackuppowerincreasetheadoptionofbuilding-levelstorage.OPTIONOFBUILDING-LEVELSTORAGE.

KEYLEARNING7:

Storagedurationswilllikelyincreaseasdeploymentsincrease.

KEYLEARNING8:

Seasonalstoragetechnologiesbecomeespeciallyimportantfor100%cleanenergysystems.

Eachofthefollowingsectionsprovidesadditionalinsightsintotheeightkeylearnings,andweconcludewithremaininguncertaintiesthatcouldbeexploredtofurtheradvanceunderstandingoftheroleofstorageintheevolvingU.S.powergrid.

KEYLEARNING1

StorageIsPoisedforRapidGrowth

TheSFSreportEconomicPotentialofDiurnalStorageintheU.S.PowerSector(Frazieretal.2021)demonstratesthegrowingcost-competitivenessofenergystorage.Usingastate-of-the-artnational-scalecapacityexpansionmodel,wefindthatdiurnalstorage(<12hoursofduration)iseconomicallycompetitiveacrossavarietyofscenarioswitharangeofcostandperformanceassumptionsforstorage,wind,solarPV,andnaturalgas(NG).

Figure1illustratesthatacrossallscenarios,deploymentsofnewstoragerangesfrom100to650gigawatts(GW)ofnewcapacity.

Thislargerangeisdrivenbyavarietyoffactors,includingstoragecosts(KeyLearning2),naturalgasprices,andrenewableenergycostadvancement,buteventhemostconservativecaserepresentsafivefoldincreasecomparedtotheinstalledstoragecapacityof23GWin2020(themajorityofwhichispumpedstoragehydropower).

Itisimportanttonotethatsignificantdeploymentsofbothrenewableenergyandstoragearedeployedevenwithoutadditionalcarbonpolicies,demonstratingtheirincreasingcost-competitivenessasresourcesforprovisionofenergyandcapacityservices.

Modeledscenariosresultinsignificant,butnotcomplete,decarbonization,wherepowersectoremissionsarereducedby46%–82%comparedto2005,andvariablerenewableenergy(VRE)reachessharesof43%–81%nationallyby2050.Durationswith4–6hoursarethemostcommon,drivenbytheinherentsynergywithPV(KeyLearning5),butlongerdurationsareoftendeployedinthelatermodeledyears(KeyLearning7).TheprimarydriversbehindstoragegrowthandtheevolutionofstoragedevelopmentwereexploredinFrazieretal.(2021)andotherSFSreports—ashighlightedinthefollowingkeylearnings.

Figure1.Nationalstoragecapacityinthereferencecasegrowstoabout200GWby2050,deployingarangeofdurations(left)Thistranslatestoabout1,200gigawatt-hours(GWh)ofstoredenergy(right),withawiderangeofdeployments.

3|StorageFuturesStudy:KeyLearningsfortheComingDecades

StorageFuturesStudy:KeyLearningsfortheComingDecades|4

KEYLEARNING2

RecentStorageCostReductionsAreProjectedToContinue,with

Lithium-IonBatteryContinuingToLeadinMarketShareforSomeTime

TheSFSreportEnergyStorageTechnologyModelingInputDataReportdiscussesthefuturecostprojectionsforutility-scalebatteryenergystoragesystemsandothertechnologiesthatdrivemuchoftheanticipatedgrowthidentifiedinKeyLearning1.

Mostofthestationarystoragedeploymentsthatwilloccurintheneartermareexpectedtobeintheformofbatteries,particularlyLIBs.ThedominanceofLIBs,atleast

inthenearterm,hasbeendrivenbygrowthofthistechnologyacrossmultiplemarkets,includingconsumerelectronics,stationaryapplications,andespeciallyelectricvehicles.

Figure2providesanexampleofhistoricalandprojectedfuturecostsoflithium-ionbatterypacks,illustratingarapiddeclineinrecentyears.Thechartalsoshowsthevastmajorityofbatterydeploymentsarefortransportationapplications,whichwill

likelybethemostimportantdriversofbatterytechnologydevelopmentandbatterycostdeclinesingeneral.

Weusedavarietyoffuturecostprojectionsforutility-scalestationarybatteryenergystoragesystems(BESS)toevaluatetotalsystemcost,includinginverter,balanceofsystem,andinstallation.Anexampleofacostprojectionforbatterieswith2–10hoursofusabledurationthatisusedintheSFSreferencescenarioisshowninFigure3.

Figure2.Lithium-ionbatterypackcostshavedroppedbymorethan80%overthepastdecadeandareexpectedto

continuetofallbasedoncontinuedscaleofproduction,drivenlargelybyelectricvehicledemand.

2021valuesfromBloombergNEF3are$132/kW.DataSource:FrithandGoldie-Scot2019

3“BatteryPackPricesFalltoanAverageof$132/kWh,ButRisingCommodityPricesStarttoBite,”BloombergNEF,November30,2021,

/blog/

battery-pack-prices-fall-to-an-average-of-132-kwh-but-rising-commodity-prices-start-to-bite/

.

Figure3.Theutility-scaleBESSReferenceScenarioprojectscontinuedcostreductions.Theleftpanelmeasurescostona$/kWh(usableenergy)basis,whiletherightpanelmeasurescostsbasedon$/kW(maximumdirectcurrent[DC]outputpower).Projectionsassumea60-megawattDCproject.

Theleftcurveshowsthetotalcostperinstalledkilowatt-hour(kWh)ofusablecapacity,whichisacommonmeasureusedinthebatteryindustry.Thisisthetotalcostofinstallation,whichforstationaryapplicationsincludesboththepower-relatedcosts(associatedwiththeequipmentthatconvertsgridelectricityintostoredelectricityandbackagain)andtheenergy-relatedcosts(thestoragemedium).Thepower-relatedcoststypicallydonotscalewithduration,meaningtheyarethesamefora2-hoursystemanda10-hoursystem,whichiswhythecostsperkWhdecreaseasdurationincreases(powercostsaredividedoveralargernumberofkWh).(ThisbreakdownofcostsforpoweranddurationisillustratedinFigure4.)Therightcurveshowsthecostperkilowatt(kW),whichisamoreconventionalmeasureofpowerplantcostsusedintheutilityindustry.Bythismeasure,costsin