LawrenceBerkeleyNationalLaboratoryDecember2021ElectricityMarketsandPolicyDepartmentEnergyAnalysisandEnvironmentalImpactsDivisionLeast-CostPathwayforIndia’sPowerSystemInvestmentsthrough2030NikitAbhyankar,ShrutiDeorah,AmolPhadkeAStudyUndertheFlexibleResourcesInitiativeoftheU.S.-IndiaCleanEnergyFinanceTaskForce2DisclaimerThisdocumentwaspreparedasanaccountofworksponsoredbytheUnitedStatesGovernment.Whilethisdocumentisbelievedtocontaincorrectinformation,neithertheUnitedStatesGovernmentnoranyagencythereof,northeRegentsoftheUniversityofCalifornia,noranyoftheiremployees,makesanywarranty,expressorimplied,orassumesanylegalresponsibilityfortheaccuracy,completeness,orusefulnessofanyinformation,apparatus,product,orprocessdisclosed,orrepresentsthatitsusewouldnotinfringeprivatelyownedrights.Referencehereintoanyspecificcommercialproduct,process,orservicebyitstradename,trademark,manufacturer,orotherwise,doesnotnecessarilyconstituteorimplyitsendorsement,recommendation,orfavoringbytheUnitedStatesGovernmentoranyagencythereof,ortheRegentsoftheUniversityofCalifornia.TheviewsandopinionsofauthorsexpressedhereindonotnecessarilystateorreflectthoseoftheUnitedStatesGovernmentoranyagencythereof,ortheRegentsoftheUniversityofCalifornia.ErnestOrlandoLawrenceBerkeleyNationalLaboratoryisanequalopportunityemployer.CopyrightNoticeThismanuscripthasbeenauthoredbyanauthoratLawrenceBerkeleyNationalLaboratoryunderContractNo.DE-AC02-05CH11231withtheU.S.DepartmentofEnergy.TheU.S.Governmentretains,andthepublisher,byacceptingthearticleforpublication,acknowledgesthattheU.S.Governmentretainsanonexclusive,paid-up,irrevocable,world-widelicensetopublishorreproducethepublishedformofthismanuscript,orallowotherstodoso,forU.S.Governmentpurposes.3Least-CostPathwayforIndia’sPowerSystemInvestmentsthrough2030AStudyundertheFlexibleResourcesInitiativeoftheU.S.-IndiaCleanEnergyFinanceTaskForceNikitAbhyankar,ShrutiDeorah,AmolPhadkeCorrespondingauthor(nabhyankar@lbl.gov)ElectricityMarketsandPolicyDepartmentLawrenceBerkeleyNationalLaboratoryDecember20214TableofContentsAcknowledgments.........................................................................................................................................5SummaryforPolicymakers...........................................................................................................................6ExecutiveSummary......................................................................................................................................71Introduction.........................................................................................................................................122Methods,Data,andAssumptions.......................................................................................................163KeyFindings.......................................................................................................................................214SensitivityAnalysis............................................................................................................................395Conclusions.........................................................................................................................................436PolicyandRegulatoryRecommendations..........................................................................................457CaveatsandFutureWork....................................................................................................................47References...................................................................................................................................................488AppendixI:KeyAssumptionsandData.............................................................................................529AppendixII:AdditionalResults.........................................................................................................5810AppendixIII:ComparativeEconomicsofPumpedHydroandBatteryStorage.................................715AcknowledgmentsWearethankfultotheU.S.DepartmentofStateforfundingthisworkandtotheBureauofEnergyResourcesformanagingit.ThisstudywouldnothavebeenpossiblewithoutthecollaborationwithIndia’sMinistryofPower(MOP),MinistryofNewandRenewableEnergy(MNRE),CentralElectricityAuthority(CEA),PowerSystemOperationCorporation(POSOCO),CentralElectricityRegulatoryCommission(CERC).Inparticular,wesincerelythankthefollowingindividualsforprovidinghelpfulandcriticalreviewsandsuggestionsateachstageofthisstudy:Mr.AlokKumar(Secretary,MOP),Mr.GhanshyamPrasad(JointSecretary,MOP),Mr.AmiteshSinha(JointSecretary,MNRE),Mr.DineshJagdale(JointSecretary,MNRE),Mr.S.R.Narasimhan(Director,POSOCO),Mr.PraveenGupta(ChiefEngineer,CEA),Dr.SushantaChatterjee(Chief,RegulatoryAffairs,CERC),Mr.S.K.Mishra(Director,PowerSystems,SolarEnergyCorporationofIndia),Mr.N.Nallarasan(HeadNorthernRegionalLoadDispatchCenter,POSOCO),Mr.RajivPorwal(POSOCO),Mr.SandeepNaik(Director,MOP),Ms.EstherKamala(CEA),Ms.AmmiToppo(CEA),Mr.ApoorvaAnand(CEA),Mr.SaifRahman(POSOCO),Mr.PriyamJain(POSOCO),Ms.RashmiNair(CERC),andMr.RavindraKadam(CERC).ThestudyalsobenefitedfromthoughtfulinputsfromDr.VarunSivaramoftheU.S.DepartmentofState,andMr.DavidRosner,Mr.ThomasPinkston,Mr.KeithMasill,andMr.EricCiccorettioftheU.S.FederalElectricityRegulatoryCommission(FERC).WethankMr.DavidPalchak,Dr.AmyRose,Mr.MohitJoshi,andMr.IlyaChernyakhovskiyoftheNationalRenewableEnergyLaboratory,Ms.SurbhiGoyal,Ms.ManiKhurana,andDr.PhillipHannamofTheWorldBank,Mr.PabloHevia-KochandMs.ZoeHungerfordoftheInternationalEnergyAgency,andDr.HyungkwanKimandDr.JiangLinofLawrenceBerkeleyNationalLaboratoryfortheirhelpfulreviewsofthereport.WethankMr.AdityaKhandekarofLawrenceBerkeleyNationalLaboratoryforhisassistanceonsolarandwindsiteselectionanalysis,IdamInfrastructureAdvisoryPvtLtdforthebackgroundresearchonsolarpanelsafeguarddutyandvariablecostsofexistingcoalpowerplants,Mr.AlbertoDiaz-GonzalezofLawrenceBerkeleyNationalLaboratoryforprojectmanagementsupport,Mr.JarettZuboyforeditorialsupport,andMs.KristanJohnson,Ms.BanafshehCobisano,andMs.WingLeungofLawrenceBerkeleyNationalLaboratoryfortheadministrativesupport.Finally,wearegratefultoMs.RuthKuandMr.MoisesBeharoftheU.S.DepartmentofState,BureauofEnergyResourcesforhelpfulreviewsaswellasprojectmanagementtokeepthisstudyontrack.Theviewsandopinionsexpressedinthisreportaresolelyoftheauthorsanddonotnecessarilyreflecttheviewsoftheorganizationsandpeoplesupportingthiswork.TheworkdescribedinthisstudywasconductedatLawrenceBerkeleyNationalLaboratoryandsupportedbytheU.S.DepartmentofState.6SummaryforPolicymakersAchievingIndia’sgoalof500GWofnon-fossilcapacity(predominantlyrenewable)istheleast-costandmosteconomicalpathwaytomeetIndia’srisingelectricitydemand,whilesafeguardinggridreliability,aslongasrenewableenergy(RE)canbesupplementedbyflexibleresourcesincluding252GWh(63GW)ofgrid-scalebatterystorageandstoragecostscontinuetodecline.IfadequatestorageandREcapacityarenotdeployedatscale,substantialadditionalthermalcapacitymayberequiredtomeettherisingdemand.Deployingstorageandrenewablesatscalewillrequireaddressingsupplychainchallengesandsecuringadequatefinancing.Usingcomprehensivegridsimulations,thestudyassessesthetechnicalandeconomicimplicationsofREdeploymentatascalesimilartoIndia’sambitioustargetof500GWofnon-fossilcapacityby2030.Keyfindingsfromthestudyareasfollows:1.AchievingIndia’sgoalof500GWofnon-fossilcapacity(predominantlyrenewables)canbetheleast-costandmosteconomicalpathwaytomeettherisingelectricitydemand.Thiswillrequire252GWh(63GW)ofbatterystorageandstepstoensuregridbalancingandstability.Giventherapidincreaseinthepowerdemand,Indiawillstillrequireamodestnetincreaseinthefossilfuelbasedcapacity–albeititsshareintheinstalledcapacityandelectricitygenerationmixwilldeclinesubstantially.Therewillbehigherinfrastructurefinancingrequirementsfortheeconomy,giventhelarge-scaleexpansionoftransmissionandstorage.Theincreaseinthermalcapacityisprimarilyrequiredtomeettheincreaseinnight-timebaseloadnetofthereductionachievedbytheongoingshiftingofnighttimeagriculturalconsumptiontothedaytime.Anetincreaseof23GWofthermalcapacitythrough2030isrequiredunderstudy’smodeledcost-optimalscenarioassuminganincreaseinpowerdemandof70percentby2030.2.Additionalthermalcapacityby2030willberequiredifadequatestorage(63GW/252GWh)cannotbedeployed.Ifthecostofbatterystoragedoesnotdeclineanditbecomesdifficulttodeploystorageatsuchscale,additionalfossilfuelbasedcapacitybeyondthe23GWnetadditionsintheprimarycost-optimalscenariowillbeneededthrough2030tomeeteveningpeakdemand,buttheseplantswilloperateatlowcapacityfactors.3.Adeclineinstoragecostswillberequiredfor500GWofnon-fossilcapacitytobetheleast-costandmosteconomicalpathwayby2030.Suchdeclinesareconsistentwithhistoricaltrendsandcrediblefutureprojectionsbythird-partyexperts,aswellasIndia’sCentralElectricityAuthority.4.Deployingbatterystorageandrenewablesatsuchasignificantscalewilllikelyrequireaddressingsupplychainchallengesandsecuringadequatefinancing.7ExecutiveSummaryIndiahassetambitiousinstalledrenewableenergy(RE)capacitytargetsof175GWby2022and450-500GWby2030.Dramaticcostreductionsoverthelastdecadeforwind,solar,andbatterystoragetechnologiespositionIndiatoleapfrogtoamoreflexible,robust,andsustainablepowersystem—muchofwhichisyettobebuilt—fordeliveringaffordableandreliablepowertoservedemandthatwillnearlydoubleby2030.AsIndia’sgridattainshigherpenetrationsofrenewables,balancinggenerationvariabilitythroughaspectrumofflexibleresourcesbecomesincreasinglyimportantforensuringtheaffordability,stability,andreliabilityofgridpower.Thisstudyassessesaleast-costandoperationallyfeasiblepathwayforIndia’selectricitygridthrough2030thatvalidates—andsurpasses—India’s2030targetof500GWofinstallednon-fossilcapacity.ThestudyusesthelatestREandbatterycostdata,anindustry-standardpowersystemmodelingplatform(PLEXOS),andexhaustiveanalyticalmethods(optimalcapacityexpansionandpowerplant-levelhourlygriddispatchsimulations).Thestudyhighlightsthecriticalroleofenhancingsystemflexibilityandmaintaininggriddependabilitythroughaspectrumofflexibleresources,suchasenergystorage,demandresponse(loadshifting),existingnaturalgaspowerplantsusedmoreflexibly,andelectricitymarkets.Specifically,wefindthattheleast-costresourcemixtomeetIndia’sloadin2030(the“PrimaryLeastCostCase”)consistsprimarilyofacombinationofREandflexibleresourcesasfollows:465GWofRE(307GWDCsolar,142GWwind,and15GWotherRE),63GW(252GWh)ofbatterystorage,60GWofloadshiftingtosolarhours(50GWagricultural+10GWindustrial),andflexibleoperationoftheexistingnaturalgasfleetof25GW.Acoalpowerplantcapacityof229GW(23GWnetadditionover2020)isfoundtobecost-effective(TableES-1).Thestudysignalsinvestmentopportunitiesthatcouldspurcreationofarobustpipelineofflexibleresources,mostnotablybatterystorage.Forexample,thetotalinvestmentrequiredby2030forbatterystoragealoneisRs300,000Cr($40billion)for63GW(252GWh)ofbatteries.Iflow-costenergystorageisnotdeployedatsuchscale,additionalthermalinvestmentsbeyondthe23GWofnetadditionswillbeneededthrough2030tomeetpeakdemand,butsuchassetswilloperateatlowcapacityfactors.Importantly,thestudyshowsthatbetween2020and2030theaveragecostofelectricitygenerationdropsbynearly8-10%owingtotheinflation-proof,low-costrenewablepowerandimprovedcapacityfactorsofexistingcoalpowerplants.Despiteaneardoublingofelectricitydemandbetween2020and2030,theemissionsintensityofelectricitygenerationdropsby43-50%,whiletotalCO2emissionsfromthepowersectorstayalmostthesameas2020levels(TableES-1).Also,India’scoalconsumptioninthepowersectorby2030iscomparabletothe2020level,implyingthatthecleanenergytransitionmaynotleadtolossofcoalmining/supplychainjobsintheneartomediumterm,potentiallygivingIndiasufficienttimetoprepareforalong-termtransition.ForIndiatoachievetheleast-costresourcemixindicatedinthisstudy,amodestdeclineinthecurrentRE(5-10%by2030)andamorepronounceddeclineinthecurrentstoragecosts(30-40%by2030),consistentwithhistoricaltrendsandprojectionsbyotherstudies,willberequired.Also,deployingREandstorageatsuchasignificantscalewilllikelyrequireaddressingsupplychainchallengesandsecuringadequatefinancing.Finally,criticalpolicyandregulatorychangessuchasalong-termresourceadequacyframework8forsystemplanningandprocurement,aregulatoryframeworkforenergystoragethatvaluesitsfullfunctionality,andnaturalgasreformsthatpromoteflexibleandefficientoperationsofthegaspipelinesandpowerplantsshouldbeimplemented.KeyStudyFindings:1.India’sincrementalelectricitydemandthrough2030islargelymetbynewinvestmentsinREandenergystoragealongwithexistingthermalassets.ThePrimaryLeastCostCasecombines465GWofRE(307GWDCsolar,142GWwind,and15GWotherRE),63GW(252GWh)ofbatterystorage,60GWofloadshiftingtosolarhours(50GWagricultural+10GWindustrial),flexibleoperationoftheexistingnaturalgasfleetof25GW,and140GWofadditionalinterstate/interregionalpowertransfercapacity(TableES-1).Acoalpowerplantcapacityof229GW(23GWofnetadditionsover2020)willbeneededby2030.Totalnon-fossilcapacityby2030wouldbe545GW.UnderaLow-RECostCase,whichassumesthatREandbatterycostscontinuetodeclineathistoricalrates(withthesolarlevelizedcostofenergyatthebestsitesdroppingtoRs1.5/kWhby2030),thecapacityofREandbatterystorageintheleast-costmixincreasesto547GWofRE(385GWDCsolar,147GWwind,and15GWotherRE)and84GW(336GWh),respectively(TableES-1).Coalpowerplantcapacityof206GWat2020levelsremainsstablein2030.TableES-1:Installedcapacities,averagecostsofgeneration,andemissionsinIndia(2020and2030)PropertyTechnologyActual(2020)PrimaryLeastCost(2030)Low-RECost(2030)InstalledCapacity(GW)Coal206229206Naturalgas252525Nuclear71919Hydropower436262Wind38142147Solar35307385OtherRE151515Storage06384Total369862943AverageCostofGeneration(Rs/kWh)3.903.593.50Power-SectorCO2Emissions(MT/yr)1,0081,080981EmissionsIntensity(kgCO2/kWh)0.820.470.41Thisnumberisamodelestimateandclosetotheactualnumber.InthePrimaryLeastCostCase,coal’sshareoftotalelectricitygenerationdecreasesfrom73%in2020to48%in2030,whilethesharefromsolarpluswindincreasesto35%.Thetotalshareofelectricitygenerationfromnon-fossilresources,includinghydropowerandnuclear,is50%in2030.IntheLow-9RECostCase,windandsolarresourcesprovide42%oftotalelectricitygenerationby2030,whilethetotalnon-fossilshareincreasesto58%.Inflation-proof,low-costREandbatterystoragearetheprimarydriversoftheseresults.Batterystorageobviatestheneedforbuildingthermalcapacitytomeetmorningandeveningpeakloads,whileagriculturalandindustrialloadshiftingfromeveningtosolarhourssignificantlyreducesthenighttimeloadand,inturn,therequirementfornewbaseloadcoal-firedcapacity.Theaveragegenerationcostin2030inthePrimaryLeastCostCaseis8%lowerthanin2020owingtotheinflation-proof,low-costREandimprovedcoalcapacityfactorsforexistingunits(TableES-1).2.Flexibleresourceshelppreventthestrandingofcoalcapacitywhilemaintaininggriddependabilityandenablingexistingcoalassetstooperatemoreefficiently.Intheabsenceofflexibleresources,particularlybatterystorageandagriculturalloadshifting,Indiamayneedtobuildsignificantnewcoalresourcesprimarilyasafirmcapacityresource,asotherstudiessuggest.Forexample,CEA(2020)showsthat,by2030,Indiawouldneedanetcoalcapacityadditionof60GWbeyondthe2020levels.However,suchacoalbuildout—intandemwiththeREbuildout—wouldlikelycausetheaveragefleet-levelcoalcapacityfactortodropto56%(gross),withover100GWofcoalcapacity(mostlyexistingplantswithhighvariablecost)operatingatcapacityfactorsof15%–40%(gross).Thisresultcouldputsuchassetsatincreasedriskofbeingstrandedandneedingregulatorysupport.Deployingflexibleresourcescanpreventthestrandingofcoalcapacitybyreducingthenewcoalbuildoutwhilemaintaininggriddependabilityandenablingexistingcoalassetstooperatemoreefficiently.InthePrimaryLeastCostCase,theaveragefleet-levelcoalcapacityfactorincreasesto65%(gross)in2030,fromlessthan60%in2020.However,20–36GWofexistingcoalcapacitywithhighvariablecostsmaystilloperateatcapacityfactorsbelow40%.3.WithlargeadditionsofREandbatterystoragecapacity,India’selectricgridremainsdependableExistingandunder-constructionthermalpowerplantscombinedwithhydropower,nuclear,andnewbatterystoragecapacityenableIndiatomeetelectricitydemanddependably—ineveryhouroftheyearineachstate—with465GWofinstalledREcapacityin2030.India’sREgeneration,particularlywindgeneration,ishighlyseasonal.Flexibleresourcesworkintandemtomaintaingriddependabilitythroughouttheyear,includingtimesofhighsystemstresssuchasperiodswithpeakannualload,highREvariability,andhighnetload.DuringhighREgenerationseasons(JunethroughSeptemberforwind,MarchthroughJuneforsolar),energystorageandagriculturalloadshiftingprovidediurnalgridbalancing.Batterieschargeduringtheday(coincidentwithsolargeneration)anddischargeduringmorningandeveningpeakperiods(4–610totalhourseachday).Batteriesalsohelpmeetsteepsystemramps.Shiftingagriculturalloadtosolarhoursincreasesthedaytimeloadby30–60GWdependingontheseason,whilereducingthenighttimeloadandtherebythebaseloadcapacityrequirementby30–50GW.Asaresult,only180GWofcoalcapacityaredispatched,mainlyasabaseloadresource(FigureES-1).FigureES-1:Averagehourlydispatchforkeymonthsin2030inthePrimaryLeastCostCaseDuringthelowREgenerationseason(OctoberthroughFebruary),the25GWofexistingnaturalgascapacity(inlieuofcoal-firedassets)playacrucialroleprovidingseasonalbalancing,withmostofthiscapacitydispatchedduringthesemonths.Iflow-costenergystorageisnotdeployedatsuchscale,additionalthermalinvestmentsbeyondthe23GWofnetadditionswillbeneededthrough2030tomaintaingridreliability,butsuchassetswilloperateatlowcapacityfactors.4.Anadditionalinterstateelectricitytransfercapacitybuildoutof140GWiscost-effective.UnderthePrimaryLeastCostCase,about140GWofnewelectricitytransfercapacitymustbebuiltby2030:40GWoninterregionalcorridorsand100GWoninterstatecorridors.BecauseofananticipateddoublingofIndia’selectricityloadbetween2020and2030,significantadditionaltransmissioncapacityinvestmentswillbeneededirrespectiveofREexpansion.5.Between2020and2030,theemissionsintensityofelectricitygenerationdrops43%–50%.By2030,theaverageCO2emissionsintensityoftheIndianpowersectordropsfrom0.82kg/kWhin2020to0.47kg/kWhinthePrimaryLeastCostCase(43%reduction),andto0.41kg/kWhintheLow-11RECostCase(50%reduction).Totalpower-sectorCO2emissionsfall3%—from1,008MT/yrin2020to981MT/yrin2030—intheLow-RECostCase;emissionsincreasebyonly7%(to1,080MT/yr)inthePrimaryLeastCostCase,despitetheneardoublingofelectricitydemand.Importantly,underthePrimaryLeastCostCase,nearly80%ofthenetincrementalgenerationbetween2020and2030isfromnewcleanenergyassets,includingnewRE,nuclear,andhydropowerassets.UndertheLow-RECostCase,newcleanenergyassetscontributeabout90%ofthenetincrementalgeneration.6.Intheneartomediumterm,India’scleanenergytransitionisunlikelytocausealossofjobsincoalminingandtransportation.By2030,India’stotalcoalconsumptionfromthepowersectoris750MT/yrinthePrimaryLeastCostCaseand667MT/yrintheLow-RECostCase—comparableto2020consumption(647MT/yr).Thus,thecleanenergytransitionmaynotleadtolossofcoalmining/supplychainjobsintheneartomediumterm,potentiallygivingIndiasufficienttimetoprepareforalong-termtransition.Wetestedthesensitivityoftheseresultstootherparametersandpolicies,assummarizedbelow:ImpactofMarket-BasedEconomicDispatch(MBED):IfIndiaimplementsanationalwholesaleelectricitymarketby2030asoutlinedintheMBEDproposalbytheCentralElectricityRegulatoryCommission,theresultingefficientthermaldispatchsavesRs14,000Cr/yr($2billion/yr)or6%inthermalpowerplantvariablecosts,albeitwithasignificantincreaseininterstateelectricitytrade.ImpactofLowDemandGrowth:IftheeconomicrecoveryfromtheCOVID19pandemicisslowanddemandgrowthbetween2020and2030decreasesby20%(2030peakloadis290GW,comparedwith340GWinthePrimaryLeastCostCase),nonewcoalcapacityiscost-effective,whilethecost-effectiveREcapacitydecreasesto355GW,from465GWinthePrimaryLeastCostCase.ImpactofLowLiquefiedNaturalGas(LNG)Price:IftheLNGpricedropsto$4.5/MMBTU(landed)by2030,LNGstartscompetingwithexpensivecoastalcoalpowerplants.Althoughbuildingnewgas-firedassetsstillisnotcost-effective,generationfromgaspowerplantsfueledprimarilybyLNGincreasestoabout120TWh/yrby2030(comparedwithabout50TWh/yrinthePrimaryLeastCostCase).LNGconsumptionincreasestoabout14bcm(10milliontons)peryearby2030.ImpactofPostponedCoalRetirements:ThePrimaryLeastCostCaseassumestheretirementofabout25GWofexistingcoalassetsby2027pertheNationalElectricityPlan.Ifthiscoalcapacitydoesnotretireasplanned,totalcost-effectivecoalcapacityby2030wouldbe238GW.AninstalledREcapacityof453GW(301GWDCsolar,137GWwind,and15GWotherRE),alongwithflexibleresources,isstillmorecost-effectivethanoperatingsomeoftheinefficientcoalcapacitywithhighvariablecosts.However,theriskofpotentiallystrandingsomeoftheoldercoalcapacityincreasessignificantly.Thisstudyindicatesseveralkeypolicyandregulatorystrategiesinthepowerandgassectors,whichweassessinaseparatereportandsummarizeinSection6ofthisreport.121Introduction1.1BackgroundandObjectivesIndiahassetanambitiouscleanenergytargetforthepowersector,namely175GWofrenewableenergy(RE)installedcapacityby2022.In2021,PrimeMinisterModiincreasedthisambitionbyannouncingatargetof500GWofinstallednon-fossilcapacityby2030.Indiahasmaderapidprogresstowardsachievingthesegoals.Between2015and2021,India’srenewableenergycapacitymorethandoubledfrom40GWto100GW,supplyingnearly10%ofthetotalelectricitygeneratedinthefiscalyear2021(CEA,2021).Overthelastdecade,IndiahasbeensuccessfulinachievingsomeofthelowestREcostsintheworld.Between2010and2020,itsawthelargestreductionincountry-levelsolarlevelizedcostofenergy(LCOE),85%,whiletheaveragesolartariffin2020was34%lowerthantheglobalweightedaverage.Indiaalsohadthelowestcountry-levelinstalledcostforsolarandwindin2020(BNEF,2020a)(Figure1).Figure1:SolarandWindenergypricesinkeycountries,includingIndia.Source:BNEF(2020a)Itiswellacceptedthatrenewableelectricitycostshavedroppedbelowcoalcostsonalevelizedbasis.Nonetheless,manycountriesaroundtheworld,includingIndia,continuetoinvestinnewcoalpowerplantsprimarilybecause:(a)REgenerationisintermittentandmayneedsignificantsystemflexibilityforgridintegration,(b)REgenerationdoesnotcoincidewithpeakelectricitydemandperiodswhichisintheeveningforIndia,and(c)legacyplanningandregulatoryframeworksthatmaynotfullycapturethevalueandcapabilitiesofREandenergystoragetechnologies.Inthiscontext,thedramaticdeclineinbatterystoragecosts—90%costreductionatthebatterypacklevelsince2010—couldserveasaturningpoint,becauseitenablesthecost-effectivesupplyoflow-costrenewableelectricityduringpeaktimes(Figure2).Notably,severallargeutilityscaleRE+storageprojectsareunderwaygloballyand,inseveralcases,offerelectricitygenerationpriceswellbelowthatfromfossilpowerplants.Forexample,arecentsolar+storageauctionbyLosAngelesDepartmentofWaterandPower(LADWP)resultedinacombinedPPApriceof$39/MWh(Rs3/kWh)forstoringover50%ofthesolarenergyinbatteries(effectivecapacityfactorofover40%)(Figure2).Similarly,inIndia,therecentRE+storagepeakingpowerauctionresultedinalevelizedtariffofRs3.5/kWh,whichiscost-competitivewiththehighervariablecostof40GWofIndia’sexistingcoalunits.13Figure2:Globalaveragebatterypackpriceoveryears(left)andSolar+BatteryStoragePPApricesintheUnitedStates(right)DataSource:BNEF(2020b)andDeorah,etal(2020).Indianutilitiesarealsousingseveralotherflexibleresourcessuchasdemandresponseforintegratingrenewableenergy.Severalstates(e.g.,Karnataka,Maharashtra,andGujarat)havealreadyshiftedamajorpartoftheiragriculturalloadfromnighttimetosolarhours(over6GWtotalin2020).ElectricitymarketreformsinIndia,fallingglobalnaturalgasprices,anddemandresponsealsooffersomeimportantflexibilityoptionstothegrid.GiventhatalargepartofIndia’selectricitygridinfrastructureisyettobebuilt,suchcostreductionsofferIndiaauniqueopportunitytoleapfrogtoamoreflexible,robust,andsustainablepowersystem.Theobjectiveofthisstudyistoassesstheleast-costresourcemixforIndiatomeetitsloadreliablythrough2030,withaparticularfocusonkeyflexibleresourcessuchasenergystorage,loadshifting,gas,andelectricitymarkets,tosupportIndia’slow-carbonenergytransitionoverthenextdecade.Severalrecentstudieshaveassessedasimilarquestion(e.g.,CEA(2020),NREL(2020&2021b),TERI(2020),BNEF(2020a),andIEA(2021)).However,mostoftherecentstudieseitherdonotconsidertherecentdramaticdeclineinthecleanenergyandstoragecosts,oraccountforsignificantchangestothedailydemandpatternofagriculturalload,orprovidespatialandtemporalgranularitytoassessindetailthetechnicalandeconomicimpactsonthepowersystem.Ourstudyattemptstobuildontheexistingliteratureandaddresssomeofthesegapsby(a)developingaspatiallyandtemporallyresolvedcapacityexpansionandeconomicdispatchmodelusinganindustrystandardplatform,PLEXOS,thatassessestheleastcostresourcemixatthestatelevel,interstatetransmissionrequirement,andpowerplantlevelhourlyeconomicdispatch,(b)usingthelatestrenewableenergyandstoragecostestimatesandtrends,informedbypricesobservedinthemarket,and(c)includingdemandsideresources,inparticular,shiftingoftheagriculturalloadfromnight-timetosolarhours,whichmanyIndianutilitiesarepracticing.1.2SummaryofRecentStudiesOurstudydrawsfromandexpandsonagrowingbodyofliteratureandmethodologiesthatassessanoptimalresourcemixforIndiatomeetitsloadinthemediumtolongrun.Allthesestudiesgrapplewithanumberofkeyissues,suchasrapidlychangingcostsandcapabilitiesofnewenergytechnologiessuchaswind,solarandbatterystorageandoperationsofastateornationalpowersystem.14Forexample,TERI(2020)focusesontheoperationalstrategiesforintegratingthe450GWofrenewablecapacityby2030.CEA(2020)assessesaleast-costresourcemixfor2030andvalidatesthetechnicalfeasibilityofthisresourcemixbysimulatinghourlydispatch.However,theynotconductaspatiallyresolvedanalysis(state/regional/othersub-nationallevel)andassesstheimpactofagriculturalloadshifting.NREL(2020)conductslong-termcapacityexpansionmodeling(2047timeframe)atthestatelevelandNREL(2021b)quantifiestheenergystorageopportunitiesinSouthAsia,includingIndia,byassessingthestoragerequirementandoperationalstrategies.WhileBNEF(2020a)usesrecentrenewableenergyauctionprices,itsresultsarenotspatiallyresolved,anditisunclearwhethertheymodelhourlygriddispatch.Moreover,noneofthestudiesmodeltheroleofdemandsideresources,suchasshiftingagriculturalloadsfromeveningtosolarhours.IEA(2021)focusesonrenewablegridintegrationissuesanddonotconductanoptimalcapacityexpansionanalysis.Theyuseproductioncostmodels–thefive-regionIndiaRegionalPowerSystemModelandtheGujaratStatePowerSystemModel,toassesstheflexibilitychallengesandsolutionsspecifictotheIndiacontext.Table1showsprojectedgeneratingcapacitybyresourcetypefromstudiesofIndia’sresourcemixinthetimeframesindicatedbelow,ascomparedwiththePrimaryLeastCostCase.Table1:Installedcapacity(GW)–2020(actual)and2030projectionsfromrecentstudies(Indiatotal)TechnologyActual(2020)CEA(2030)NREL(2030)BNEF(2030)TERI(2030)Coal206267170234238Naturalgas2525492525Nuclear719113317Hydro5461548184Wind38140200109169Solar36280250204229Batterystorage027(4-hour)16(2-hour)68(4-hour)#N/A60(2-hour)NewpumpedstorageN/A101.500LoadShifting0000OtherRE1515Total381844824734822ACcapacitywithanInverterLoadingRatioof1.30,implyingtheDCcapacitytobeapproximately265GW.Note:Eachstudyusesadifferentsetofassumptionsontechnologycosts,baselineyear,andoperationalparameters.Therefore,thecomparisonacrossstudiesisshownforillustrativepurposesonlyandshouldbeinterpretedcarefully.DataSources:CEA(2020a),CEA(2020b),NREL(2021b),BNEF(2020a),TERI(2020)TERI(2020)examinesvariousscenariosofREpenetrationontheIndiangridin2030,andtheyrunanhourlyproductioncostmodeltodetermineunitcommitment,systemcosts,andtransmissionflows.Theyconcludethatthesystemcostofahigh-REpathway(32%ofgenerationfromsolarandwind)iscomparabletothecostofthebaselinecase(26%ofgenerationfromsolarandwind).Theyexploresystemflexibility15requirementsbyexaminingoptionsforcoalflexibilityandincorporatingbatterystorageintotheresourcemix.However,theydonotconductleast-cost,capacity-expansionmodeling.NREL(2021b)usestheRegionalEnergyDeploymentSystem(ReEDS)modeltoassesscost-effectiveopportunitiesforgrid-scaleenergystoragedeploymentinSouthAsiabothintheneartermandthelongterm,includingadetailedanalysisofenergystoragedrivers,potentialbarriers,andtheroleofenergystorageinsystemoperations.Theyconductscenarios-basedcapacityexpansionmodelingandfindthatIndia’smarketforgridscalestoragewillbeintherangeof50GWto120GWby2030,mostlyfromlithium-ionbatterystorage.BNEF(2020a)estimatesthattheleast-costresourcemixin2030includes313GWacofsolarandwind(26%ofgeneration)and234GWofcoal(55%ofgeneration)outofatotalestimatedinstalledcapacityof734GW.By2034,REcapacitywouldgrowrapidlytoreach450GW,whilecoalcapacitywouldbe252GW;flexibleresources,includingpeakergas,pumpedhydro,batteriesetc.,wouldcontributeatotalofover90GW.Buildingontheexistingliterature,ourstudyattemptstoaddresssomeofthesegapsby(a)developingaspatiallyandtemporallyresolvedcapacityexpansionandeconomicdispatchmodelthatassessestheleastcostresourcemixatthestatelevel,interstatetransmissionrequirement,andpowerplantlevelhourlyeconomicdispatch,(b)usingthelatestrenewableenergyandstoragecostestimatesandtrends,informedbypricesobservedintheglobalandIndianmarkets,and(c)includingdemandsideresources,inparticular,shiftingagriculturalloadfromnight-timetosolarhours,whichmanyIndianutilitiesarepracticing.Theremainderofthisreportisorganizedasfollows.Insection2,wesummarizeourkeyassumptions,scenarios,anddata.Insection3,wepresentthekeyfindingsfollowedbysensitivityanalysisinsection4andkeyconclusionsandpolicyimplicationsinsection5.Insection6,wesummarizethepolicyandregulatoryrecommendations(assessedindetailseparately)thatwouldenableIndia’stransitiontoaflexible,robust,andcleanerpowersystem.Section7provideskeycaveatsinusingthisanalysisandidentifiesthefuturework.AppendixIprovidesdetailedassumptionsanddatasources.AppendixIIshowsadditionalresultsonsystemoperationsandtransmissioninvestments,includingsomestatelevelfindings.InappendixIII,weofferahighlevelcomparisonoftheeconomicsofbatterystorageandpumpedhydrosystemsinIndia.162Methods,Data,andAssumptionsWeusePLEXOS1tobuildacapacity-expansionmodeltoassesstheleast-cost(“optimal”)generationmixatthestatelevelandinterstate/inter-regionaltransmissioninvestmentsforeachyearbetweenFY2020andFY2030.2Themodelminimizestotalgenerationcost(fixedplusvariablecosts)fortheentiresystem,includingexistingandnewgenerationcapacityandtransmissionnetworks.Weassesstheoptimalresourcemixunderarangeofscenariosexaminingtechnologycosts,naturalgasprices,coalplantretirements,demandgrowth,electricitymarketdesign,demandresponse,andsupplychainchallenges.ForFY2030,wealsomodeleconomicdispatchatthepowerplantleveltoensurethatthegridcanrunreliablyforall8,760hoursintheyear,includingthehourswhenthesystemismostconstrained.WemodeltheIndianelectricitygridusing36nodes:onenodeforeachstate/UnionTerritory(Figure3).Figure4depictsouroverallmethodandthevariousdatacomponents.Figure3:RepresentationofIndia’stransmissionnetworkwithasimplifiedinterstatenetwork(36nodes)alongwiththelocationofexistingpowergenerationplants1FormoreinformationonPLEXOS,seewww.energyexemplar.com.PLEXOSusesdeterministicorstochastic,mixed-integeroptimizationtominimizethecostofmeetingloadgivenphysical(e.g.,generatorcapacities,ramprates,transmissionlimits)andeconomic(e.g.,fuelprices,start-upcosts,import/exportlimits)gridparameters.Moreover,PLEXOSsimulatesunitcommitmentandactualenergydispatchforeachhour(orat1-minuteintervals)ofagivenperiod.Asatransparentmodel,PLEXOSmakesavailabletotheusertheentiremathematicalproblemformulation.2ThefiscalyearinIndiarunsfromApril1throughMarch31.Forexample,FY2030runsfromApril1,2029toMarch31,2030.Inthisreport,weusethetermsfiscalyearandyearinterchangeably.Anyreferencetoayearimpliesfiscalyear,unlessspecifiedotherwise.17Figure4:OverviewofthemodelingframeworkWeassessthefollowingtwoprimaryscenarios:1.PrimaryLeastCost:Thisscenarioassumesamid-costtrajectoryforcleantechnologies,60GWofloadshiftingtosolarhours,state-levelbalancing,andabasecostforliquefiednaturalgas(LNG).2.Low-RECost:ThisscenarioassumesREandstoragecostreductionthrough2030inlinewiththehistoricaltrends.Assumptionsaboutloadshifting,systembalancing,andLNGcostsarethesameasthePrimaryLeastCost.Importantassumptionsanddatasourcesareasfollows(seeAppendixIfordetails):●Cleantechnologycosts:Wemodelthreecases(low,mid,andhigh)ofsolar,wind,andbatterycosttrajectories(Figure3).SeeAppendixIfordetails.○Thebaseormid-cost(orbase-cost)caseinthePrimaryLeastCostCaseassumesthecostreductionsforsolarandwindtechnologiesoverthenextdecadearehalftheobservedhistoricalrate.AveragesolarLCOEdropsfromRs.2.8/kWhin2020toRs.2.0/kWhin2030whilewindLCOEgoesfromRs.3.2/kWhin2020toRs.3.0/kWhin2030.NotethattheseprojectionsaresomewhatmoreconservativecomparedtootherglobalprojectionssuchasBNEForNRELATB(moderatecostcase)(BNEF,2021;NREL,2021a).OurassumptionsforLi-ionbatterylevelizedcostofstorage(LCOS),basedonourpreviousbottom-upcostanalysis,areRs.6.0/kWhin2020andRs.3.7/kWhin2030for4-hourstorage(Deorahetal,2020).3○Thelow-costcaseintheLow-RECostCaseassumescostreductionsareinlinewithhistoricaltrends,withtheaverageLCOEin2030droppingtoRs.1.5/kWhforsolar,Rs.2.5/kWhforwind.TheseprojectionsaremoreinlinewithotherglobalforecastssuchasBNEFandNRELATB(moderatecostcase)(BNEF,2021;NREL,2021a).Forbatteries,weassumethattheLCOSofa4-hourstorageprojectdropstoRs.3.0/kWhby2030.3Inparticular,batterypacklifeisassumedtobe3,000cyclesor10years,whiletheprojectlifeisassumedtobe20yearsmeaningthattherewillbeonebatterypackreplacementinyear11.ThermaloperationalconstraintsFY2030loadprojectionsCapital,O&M,andtransmissioncostsCapacityundercon-struction+announcedannannouncedtargetsHourlydispatch&interstatetransmissionflowOptimalResourceMix(Generation+Transmission)Wind/solarprofiles+hydroconstraintsTotalelectricitysupplycostPLEXOSCapacityExpansionandEconomicDispatchmodelwithhourlyresolutionatplantlevelSystemoperationsandmarketsFuelpricesandavailabilityconstraintsHourlyloadprofiles18○Thehigh-costcaseassumesthecosttrajectoryofcleantechnologiesishigherthaninthebasecase(solarandwindLCOEofRs2.3/kWhandRs3.1/kWhby2030,respectivelyand4-hourbatteryLCOSofRs4.9/kWhby2030),whichcouldoccurforvariousreasons,suchasslowerreductionsinglobalprices,restrictionsonimports,orsolarandbatterysupplychaindisruptionsthatlimitthecapacitythatcouldbeinstalledinthefirstfewyearsofthedecade(10GW/yr).Weassumedomesticmanufacturingcatchesupbymiddleofthedecade,andnewinstallationsarenotconstrainedbeyond2025.Underthisscenario,Indiadoesnotachieveits175-GWREtargetby2022.●Demandforecast:Weusestate-leveldemandprojectionsfromCEA’s19thElectricPowerSurvey(EPS)(CEA,2017b).India’speakloadisexpectedtogrowfrom180GWin2020to340GWin2030,whilethetotalenergydemand(bus-bar)increasesfrom1,357TWhto2,363TWhperyearoverthesameperiod(CEA,2017b).Usingthestatelevelhourlyloaddatain2018,weprojectthehourlyloadpatternforfutureyears.WealsorunaLowDemandGrowthcase,whichassumesa25%lowerdemandgrowth,implyinga2030peakloadof290GWandatotalenergydemandof2,000TWhperyear.●REgenerationprofiles:Weassesshourlywindgenerationprofilesandhydrodispatchconstraintsusinghistoricalgenerationdata(fortheloadsynchronized2018weatheryear).Forsolar,wecreatehourlygenerationprofilesusingGlobalHorizontalIrradiance(GHI)orDirectNormalIrradiance(DNI)dataforkeysiteswithineachstate(Deshmukhetal2019;Abhyankaretal,2016).●Agricultural(Ag)andindustrialdemandresponse:Severalstateshaveseparateddistributionfeedersforagriculturalconsumersfromotherfeeders,andsomestates(e.g.,Karnataka,Maharashtra,andGujarat)havealreadyshiftedamajorpartoftheagriculturalloadtosolarhours(over6GWtotalin2020)(KPTCL,2020;MSLDC,2020).Weassumethesametrendtocontinueinthefuture,andby2030,about50GWofagriculturalloadand10GWofindustrialloadcouldbeshiftedfromnight-timetosolarhours.●Coalcapacity:WeincorporatethecoalcapacitythatisalreadyunderconstructionperCEAprogressreports(about38GWbetween2021and2025-23GWuntil2022and15GWbetween2023and2025)(CEA,2021).TheNEPstipulatesthatabout8GWofexistingcoalcapacitywouldretireby2022,andabout25GWretiresby2027—thisincludesplantsthathavesurpassedtheirusefullifeorplantsthatare/willbeunabletomeetrequiredemissionstandards(CEA,2018b).ThePrimaryLeastCostCaseaccountsforalltheseadditionsandretirements.Wealsosimulateacaseinwhichthe25GWofcapacity(plannedtoberetiredbetween2022and2027)donotretireasanticipated.●Regulatoryframeworkforbalancing:Wetreateachstateasanindependentbalancingareawithacertainimport-exportcapacity.Thisisthecurrentpracticeunderwhichdistributionutilities,oftentermedasDiscoms,“self-schedule”thegenerationtheyhaveundercontracts(includingthecentralsectorplants).Currently,limitedelectricitytradeoccursbetweenstates,exceptforoccasionalbilateralcontracts,orintheday-aheadwholesaleelectricitymarket.Wesimulatethe19limitedinterstateelectricitytrade(exceptforthecentralsectorplants)byapplyinganeconomichurdleofRs1.5/kWhtoallelectricitythatastatewouldimportfromotherstates.Central-sectorgeneratingstationsundercontractwouldnotfacesuchhurdleratebecausemostofthemalreadyhavecontractswithmultiplestates.Wealsomodelasaseparatescenarioanationalorcentralizedpool-baseddispatchpursuanttoCERCandMOP’sMarketBasedEconomicDispatch(MBED)proposal(CERC,2018;MOP,2021).●Conventionalfixedcostsandcostofcapital:WeuseCERCgenerationtariffregulations,CEAassumptions,andindustryconsultationsforestimatingthecapitalcostandfixedoperationsandmaintenance(O&M)costsforeachconventionaltechnology(coal,naturalgas,hydro,biomass,anddiesel)(CERC,2019;CEA,2020).Regulatorynormsforcoalcapitalcostsexcludeadditionalinvestmentsrequiredtomeetnewpollutionstandardsforparticulatematter,SOx,andNOxemissions.●WeightedAverageCostofCapital:Weassumethereal(inflation-adjusted)weightedaveragecostofcapital(WACC)of8%,whichisequivalenttoanominalWACCof11.6%(nominalinterestrateof11%andreturnonequityof14%,assumingadebt-to-equityratioof80:20).●Variablecostsofexistingpowerplants:Wetakethevariablecostsofexistinginterstategeneratingstations(ISGS)fromreportsavailableundertheReservesRegulationAncillaryServices(RRAS)mechanism(POSOCO,2021).VariablecostsforstategeneratorsandIPPsarefromtariffordersbytherespectivestateelectricityregulatorycommissionsforFY2020,whereavailable.Variablecostsforpowerplantswithnorecentregulatorydata(nearly5%or10GWoftheexistingthermalpowerplants)aretakeneitherfromtheMERITwebsiteorassumedtoberegional/statelevelaverage.●Coalpricesfornewpowerplants:WeassumeapitheadcoalpriceofRs2000-2500/ton(incltaxes),whichisequivalenttoavariablecostofRs1.59/kWh,increasingat1%peryear(halfthehistoricalgrowthrateofCoalIndiaLimited’sactualcoalpricesperCIL(2020))between2020and2030.ImportedcoalpricesaretakenfromglobalmarketreportsattheIndonesianhub.Althoughimportedcoalpricesarehigherthandomesticcoalprices,theyimproveplantheatrates.●Naturalgasprices:Weassumethatdomesticgasavailabilityforpowersectorwillremainthesameas2020(8.4bcm/yror23mmscmd)(MOSPI,2020).TotalLNGimportcapabilityincreasesfrom15milliontonsperannum(MTPA)in2020to50MTPAin2030.Domesticgaspricein2030isassumedtoremainalmostthesameas2020($4.2/mmbtu).WeexaminetwoLNGpricescenarios:1)alandedpriceof$5.5/MMBTU(plusregasificationcostof$0.6/mmbtuandpipelinecharges,asapplicable)whichisinlinewithcurrentprices,and2)alowerpriceof$4.5/MMBTU(plusregasificationcostof$0.6/mmbtuandpipelinecharges,asapplicable).●Operationalparameters:Wetakekeyoperationalparametersforthermalpowerplants(ramprates,technicalminimumgenerationlevels,auxiliaryconsumption,forcedoutageandplannedmaintenancerates,warm/coldstarttimes,secondaryfueluse,startcost,etc.)fromtheprevalentregulations,performancedata,andnormativevaluesusedbysystemoperatorsandprovidedinthe20CEAThermalPerformanceReview(CEA,2020;CEA2018;CERC,2019).Coalpowerplanttechnicalminimumgenerationlevelisassumedtobe55%forcentralsectorgeneratingstationsandIPPsperCERCregulations,whileitisassumedtobe70%forstate-levelownedgeneratingstations(CERC,2019).●Heatrates:Weuseactualheatratedataforeverypowerplantbasedonseveralsources,suchasregulatoryfilings,CEAThermalPerformanceReview,CEACO2EmissionsBaseline,MERIT4dataonvariablecosts,etc(CEA,2018;CEA2021;MERIT,2021).Wemodeltheheatrateasafunctionofgeneratorloading,asperCERCregulationsoncompensatingforpartialloadoperations.●Reserves:Forcapacity-expansionmodeling,weassumea5%planningreservemargin.Fordispatchmodeling,weinclude5%spinningreserves.Wealsomodelforcedandplannedmaintenanceoutagesforallpowerplants,usingtheactualvaluesforexistingcapacityperCEAThermalPerformanceReviewandwheresuchvaluesarenotavailable,weusenormativevaluesperCERCtariffnorms(CEA,2018;CERC,2019;CEA,2020).●Hydropowerplants:Wedividethehydroplantsineachstateintothreecategories:(i)reservoir,(ii)run-of-river,and(iii)pumpedhydro.Reservoirhydroplantsaremodeledusingamonthlyenergybudgetapproachusingtheactualmonthlygeneration/capacityfactorsinweatheryear2018(CEA,2020;CEA,2019).Forrun-of-riverplants,hourlyoutputisassumedtobeconstantthroughouttheweek/monthsubjecttotheenergybudgetconstraint.Pumpedstoragecapacityisoptimallydispatchedsubjecttohead/tailreservoirstoragecapacities.●Inter-annualvariabilityinwind,solar,andhydropowergeneration:Choosingaspecificweatheryearformodelingwind,solar,andhydropowergenerationcouldmisscapturingthelow-probabilityrisksduetoverylargeinter-annualvariabilityingeneration.Inordertoassesstheimpactofsuchvariationsduringtheperiodsofhighestsystemstress,wesimulatethehourlydispatchduringthepeakloadandthenetloadpeakweeksbyassumingsolarandwindoutputtobe20%and50%lower,respectively,basedonthesimilarassessmentsintheUSbyShaneretal(2018)andPhadkeetal(2020).●Transmissionnetwork:Wemodeltheinterstate(400kVandabove)transmissionnetworkusingareducedform36-nodemodel(onenodeforeachstate/UnionTerritory),whichallowsustoassessthetransmissionflowsandrequirementsattheinterstatelevel.4MeritOrderDispatchofElectricityforRejuvenationofIncomeandTransparency(MERIT),anapplicationofMinistryofPower:http://meritindia.in213KeyFindings3.1Incrementaldemandthrough2030couldlargelybemetbynewinvestmentsinREplusstorageandexistingthermalassets.Acoalpowerplantcapacityof229GW(23GWnetadditionover2020)willbeneededby2030.(a)Theleast-costresourcemixin2030includes465GWofRE,63GWofenergystorage,60GWofloadshifting,and229GWofcoalcapacityDespitedemandnearlydoublingbetween2020and2030,meetingmostincrementaldemandbybuildingnewsolar,wind,andflexibleresourcesiscostoptimal.UnderthePrimaryLeastCostCase(mid-REcost),about23GWofnewcoalcapacity,mostlyatpit-headlocationsintheeasternandwesternregions—beyondthecoalcapacitycurrentlyunderconstructioniscost-effective.Iflow-costenergystorageisnotdeployedatsuchscale,additionalthermalinvestmentsbeyondthe23GWofnetadditionswillbeneededthrough2030tomeetpeakdemand.IntheLowRECostCase,theleast-costmixofresourcesin2030includesover530GWofsolarandwindcapacity,coupledwith84GWofenergystorage.Table2andFigure5showtheinstalledpowermixovertimeinthePrimaryLeastCostandLowRECostCases.Table2:Installedcapacityin2020vs.2030inthePrimaryLeast-CostandLow-RECostscenariosTechnologyInstalledCapacity(GW)Actual(2020)PrimaryLeastCost(2030)Low-RECost(2030)Coal206229206Naturalgas252525Nuclear71919Hydro436262Wind38142147Solar35307385OtherRE(SmallHydro+Biomass)151515BatteryStorage063(252GWh)84(336GWh)Total369862943AverageGenerationCostRs/kWh3.903.593.50Modeledasstandalonebatteryenergystoragesystems.Thisisestimatedbythemodelandisveryclosetotheactualnumber.ThenationalaveragecostofpowerpurchaseestimatedbyCERCinApril2021isRs3.85/kWh(CERC,2021).22Figure5:InstalledcapacitybyresourcetypeinPrimaryLeast-Cost(left)andLow-REcost(right),2020–2030Withthismix,theshareofnon-fossilresourcesintotalinstalledcapacityis545GWor63%inthePrimaryLeastCostcase(or50%shareinannualgeneration)and628GWor67%inthelow-REcostcase(or56%shareinannualgeneration).Figure6showsthewindandsolarsitesinthePrimaryLeast-Costscenarioin2030.Thesitesarechosenusingmultiplecriteria,suchasresourcequalityandproximitytotheexistingroadandtransmissioninfrastructure,whileexcludingagriculturallands,sensitiveareas,waterbodies,urbanbodies,forests,etc.5Thetotallandfootprintofsolarplantsin2030isabout1.7millionacres,or0.2%oftotallandarea,comparedwith0.16millionacresofdirectfootprint(0.02%oftotallandarea)forwindcapacity.65Allunderlyingdataandotherassumptionsareavailableathttp://mapre.lbl.gov.6Forwindenergy,0.16millionacresisthedirectfootprintontheground.Totallandarearequirementwouldbehigher,butmostofthatlandcouldbeutilizedforotherapplicationsincludingagriculture.23Figure6:Sitesforinstallationofsolar(left)andwind(right)plantsinPrimaryLeastCostCase,2030Theresourcemixinthebasecaseisleastcostduetothreemainreasons:1.Plummetingcostsofsolar,wind,andbatteriesdrivethesystemaveragecostdown.Averagecostofgeneration(includinginterstatetransmission)islowerinFY2030underthePrimaryLeastCostCasethan2020levels.TheaveragecostofgenerationinFY2030underthePrimaryLeastCostcaseisestimatedtobeRs.3.59/kWh,whichis8%lowerthantheestimated2020averagecostofRs.3.90/kWh.Lowergenerationcostswouldtranslateintolowerretailelectricityprices,assumingelectricitydistributioncostsdonotchangesignificantlyinthePrimaryLeastCostCase.TheprimaryreasonforthecostdeclinesislowREandstoragecostsmakethemcompetitivewiththermalpowergenerationthroughoutthecountry,eveninregionspreviouslyconsideredresource-poorforrenewableenergygeneration.Forexample,over150GWofexistingcoalcapacityinIndiahasavariablecostofmorethanRs2/kWh,thelowestsolarPPApricein2020.EvenafteraddingastoragecostofRs1/kWh(for20-25%solarPVenergystoredinbatteries),thecostofevening-peakingsolarpowerwouldbeRs3/kWh,whichislowerthanthevariablecostofover80GWoftheexistingcoalcapacity.7Thisimpliesthatutilitieswillbebetteroffbuildingnewsolar+storageprojectsfordiurnalbalancingandnotdispatchingexpensivecoalplants,whilestillpayingtheirfixedcosts.7BasedonglobalmarkettrendsandbottomcostsinIndia,Levelizedcostofco-locatedstorageisRs6/kWhin2020,droppingtoRs3.7/kWhby2030.By2024-25,thecostsdroptoRs4.8-5/kWh.For20-25%ofsolarenergytobestoredinbatteries,thenetstorageadderwhenspreadoverthesolargenerationfromtheprojectwould20-25%ofRs4.8-5/kWhoraboutRs1-1.2/kWh.242.60GWofdemandresponsereducesthenight-timebaseloadrequirement.Shiftingofagriculturalload,whichisprimarilysuppliedduringnighthours(10PMto6AM),tosolarhourswouldreducesignificantlythenight-timebaseloadpowerrequirementtypicallymetbycoalpowerplants.Shifting60GWofloadawayfromnighthoursreducestheneedforbaseloadcoalcapacitybyover30GW.8Suchloadshifttosolarhoursalsofacilitatescost-effectivegridintegrationof30GWofnewsolarcapacity.3.Cheapgrid-scalebatterystorageenhancesthecapacityvalueofsolar.India’sloadiseveningpeakingusuallyaround7or8pm,implyinglow-capacityvalueforsolarintheIndiangrid.Batterieschangethisdynamicbystoringexcesssolarenergyproducedduringtheafternoonanddischargingitduringtheeveningpeakhours.Thisinterplaybetweensolarandbatteriesalsoenablescleansourcestoprovidefirmcapacityandmeetthereservemarginrequirements.Thestoragecapacityrequiredonthegridisabout10%oftheaveragedailyREgenerationby2030,equivalentto63GWx4hours=252GWh.9(b)About23GWofnetadditiontothecoalcapacityiseconomicalBetween2020and2030,beyondthecoalcapacitycurrentlyunderconstruction(38GW)andplannedretirements(33GW),wefind23GWofnewcoalcapacitytobecost-effectiveunderthePrimaryLeastCostCase.Thismeansabout229GWofcoalcapacitywillberequiredin2030,comparedwith206GWin2020.IntheLowRECostCase,wefindnonewcoalcapacitybeyondthecapacityalreadyunderconstructioniscost-effective.(c)Non-fossilresourcescontributetooverhalfofelectricitygenerationby2030,withsolarandwindconstitutingathirdoftotal2030electricitygenerationThePrimaryLeastCostmixsuggestsafive-foldincreaseintotalREcapacity,from90GWin2020to465GWin2030.ThisincreaseisaidedbytheplummetingcostofLi-ionbatteries,becausestorageenhancesthevalueofsolarenergytothegrid.By2030,solarandwindresourcesprovide36%(about850TWh)oftotalelectricitygenerationinthisscenario(Figure7).Theshareofgenerationfromnon-fossilresources,includinghydroandnuclear,increasesfrom24%in2020to50%in2030—demonstratingthatIndiacanmakemajorgainsinreducingemissionsandlocalairpollutionfromitspowersector,whilereducingcostsforutilitiesandendconsumers.Between2020and2030,coal’sshareoftotalelectricitygenerationdropsfrom73%to48%,whiletotalcoalgenerationincreasesfrom988TWhto1,145TWh,implyingcoalpowerplantoperationsatbettercapacityfactors.Wealsoinferthatcoalconsumptionfromthepowersectorisunlikelytodecreaseoverthenextdecade;infact,inthePrimaryLeastCostCase,itincreasesby15%,from647MTinFY2020to750MTinFY2030.EvenintheLowRECostCase,wheretheREandnon-fossilshareintotalelectricitygenerationincreasesto42%and56%respectively,thetotalcoalconsumptionremainsashighas659MT8Coalcapacityreductionislowerthan60GWbecauseoftheseasonalvariationinagriculturalconsumption.9Notethatwehavemodeledbatterystorageasastandaloneresourceandnotco-locatedwitharenewableenergygenerator.25byFY2030.Therefore,thecleanenergytransitionmaynotleadtolossofcoalmining/supplychainjobsintheneartomediumterm,givingIndiasufficienttimetoprepareforalong-termtransition.Figure7.AnnualgenerationbyresourcetypeinPrimaryLeast-Cost(left)andLow-RECostscenarios(right),2020–2030(d)Coalpowerplantsoperateatimprovedplantloadfactors(PLFs),reducingtheirfinancialandoperationalstressIntheabsenceofflexibleresources,particularlybatterystorageandagriculturalloadshifting,Indiamaycontinuetobuildsignificantnewcoalresourcesprimarilyasafirmcapacityresource,asotherstudiessuggest.However,suchcoalbuildout,intandemwithREbuildout,maycausetheaveragefleet-levelcoalcapacityfactortodropto56%,withover100GWofexistingcoalcapacityhighvariablecostoperatingat15-40%capacityfactors,potentiallyplacingsuchassetsatanincreasedriskoftechnicalandfinancialstressandstranding.At40%PLF,theaveragecostofgenerationincreasestoaboutRs.6.0/kWh,whilesuchcostrisestoaboutRs.10.0/kWhat15%PLF.InthePrimaryLeast-CostCase,theaveragePLFofthecoalfleetincreasesto63%(gross),whichishigherthanthegrossPLFof61%in2020(Figure8).PLFsacrossthecoalfleetstillvary,butthequantumofcapacityoperatingatverylowPLFsreducessignificantly.In2030,about36GWofcapacity,consistingmostlyofexistingplantswithhighvariablecost,operatesatlessthan40%PLF.Theseplantsmayneedcertainregulatorysupportbecauseofthestrandingrisk.26Figure8:CoalgenerationandgrossPLFinthePrimaryLeast-Costscenario(e)Theaveragecostofelectricitygenerationislowerthantoday'scostofgenerationTheaveragecostofelectricityincludesthefixedcosts(annualizedcapitalserviceandO&M)ofallexistingandnewpowerplants,batteryassets(includingbatterypackreplacementcosts),andthetransmissionnetwork,fuelcostsofthermal,biomass,andnucleargenerators,andanystartup/shutdowncosts.WemodeltheheatrateofthermalpowerplantsusingCERC’sheatratecurve,whichimpliesthat,ifthermalpowerplantsoperateatlowloads,thethermalefficiencydecreases.AsFigure9shows,theaveragecostofgenerationdecreasesfromRs.3.90/kWh(5.20cents/kWh)in2020toanestimatedRs.3.59/kWh(4.78cents/kWh)in2030inthePrimaryLeast-CostCase,adropofabout8%.IntheLowRECostCase,theaveragecostofgenerationdropstoRs3.50/kWh(4.67cents/kWh),or10%by2030.10Thesedropsresultfromtwofactors.First,theLCOEofmostnewcapacityadditions(solarandwind)ismuchlowerandfollowsadecreasingtrendoverthedecade.Second,thePPAsarefixedinnominaltermsfor25years,makingahugeportionofpowerprocurementinflationproof.Withadecreasingshareofcoalpowerinthemix,theinflation-relatedincreaseinaveragecost(fuelandtransportationcosts,laborcosts,O&Mcosts)ismitigated.10Thesearethecostsimpliedfromthecapacityexpansionmodel.Ifoneusestheproductioncostsfromhourlydispatchmodel,thefuelcostsmaybeslightlydifferent.Theseaveragecostnumbersshouldnotbetakenfortheirabsolutevalue,butmoreforthetrendbetween2020and2030.27Note:Allcostnumbersexpressedas2020real.Figure9:Nationalaveragegenerationcost,2020through2030,inthePrimaryLeast-CostandLowRECostscenariosFigure10showsthedriversoftotalsystemcostin2020andinthe2030PrimaryLeast-CostCase.In2030,thefixedandvariablecostsofcoal-firedgenerationstillaccountfor52%ofthetotalsystemcost.Ontheotherhand,wind,solar,andstoragecostsjointlyaccountfor29%ofthetotal,whileproviding36%oftheannualenergygeneration.ThecostofnewtransmissionbuildoutwillbeRs0.09/kWh(0.11cents/kWh),ifspreadovertheentiregenerationbase.Note:Allcostnumbersexpressedas2020real.Figure10:Totalsystemcostandcostdriversin2020and2030inthePrimaryLeast-Costscenario283.2ThegridisdependableineveryhouroftheyearOurdispatchmodelingvalidatesthattheoptimalresourcemixcanmeetdemandineveryhouroftheyearin2030.Thereisnolossofload,evenduringdayswhenthesystemisstressed,suchasdaysofpeakload,highestnetload,highestREvariability,etc.Figure11showsaveragehourlysystemdispatchinFY2030forkeymonthsinthePrimaryLeastCostCase.Theflexibleresourcesworkintandemtomaintaingriddependability.Agriculturalloadshiftingandenergystoragetogetherarecriticalfordiurnalbalancingofthegrid,whilenaturalgasplantsarecriticalforseasonalbalancing.Agriculturalloadshiftingreducesthenighttimebaseloadrequirementbyabout30-50GW,suchthatduringtheevening(7-9PM)andmorning(6-8AM)peaks,baseloadcoal’scontributionisminimal.Energystorage,includingbatteriesandpumpedhydro,chargesduringthedayanddischargesduringeveningandmorningpeakhours,whilealsoprovidingtherampingsupportduringthemostcriticalrampevents.NaturalgasplantsoperatemostlyduringthelowREseason(OctoberthroughFebruary)andarecriticalforseasonalbalancingofthegrid.Thereissmallamountofrenewableenergycurtailment(0.2%annually),mostlyoccurringduringhighwindseason(June-September).Thecurtailmentisfoundtobesmallbecauseofthefollowingtworeasons:(a)significantquantumofflexibleresourcessuchasenergystorage,agriculturalloadshifting,andflexibleoperationofgaspowerplantsenablecost-effectivegridintegration,and(b)wedonotmodelanyintra-statetransmissionconstraints,whichismainlyresponsibleforREcurtailmentthatiscurrentlyobservedinIndia.AdditionaldispatchresultsareshowninAppendixII.Figure11:Averagehourlydispatchforkeymonthsin2030inthePrimaryLeast-Costscenario29(a)AgriculturalloadshifthelpsreducethenighttimeloadFigure12showstheprojectedloadcurvesforMayandOctoberinFY2030.Thedottedcurveshowstheoptimallyshiftedloadcurve,if60GWofload(50GWofagriculturalloadand10GWofheavyindustrialload)shiftstosolarhoursbyFY2030.AsmentionedearlieranddemonstratedinAppendixII,statessuchasKarnataka,Gujarat,andMaharashtrahavealreadyshiftedsignificantagriculturalloadtodaytime(6GWintotal),andseveralotherstatesarefollowingsuit.Giventheseasonalandregionalvariationsinagriculturalload,onaverage,suchashiftwouldreducethenighttimeloadby30GWnationally.Figure12:Averageload(solidline)andshiftedload(dottedline)curves,FY2030,inthePrimaryLeast-Costscenario.Annually,122TWhofagriculturalandheavyindustrialloadgetsshiftedfromnighthourstosolarhours.(b)About60GW(250GWh)ofenergystoragehelpsmeetmorningandeveningpeakloadsEnergystorageiscrucialfordiurnalbalancingofvariableREgeneration(i.e.,shiftingtheREgenerationtomorningandeveningpeakdemandhours(6–8amand7–10pm))andavoidingbuild-outofnewthermalcapacitythatwouldberequiredprimarilyformeetingthepeakload.11Figure13showscharginganddischarginghoursforbatteriesforanaveragedayduring4monthsoftheyearin2030.Thebatteriestypicallychargeduringthedayanddischargeover6-8hoursduringthemorningandeveningpeakhours.DuringwintermonthsbetweenOctoberandJanuary,batteriesalsochargeatnight.Thisismainlybecausebatteriesareunabletofullychargeduringthedayduetothesteepreductioninwindgenerationduringwinter.Fastrespondingbatteriesarecrucialformeetingthegrid’smorningandeveningramprequirementsin2030.Theserampsaretimedwiththelargeon-rampofsolargenerationbetween7and8am,andoff-ramparound6pm.Batteriesalsohelpreducetherampingstressonthermalplants.Storagewouldbeacriticalsourceofflexibilitystartingasearlyas2023,especiallyinstateswithhighsolardeploymentandlowhydroresourcessuchasRajasthanandGujarat.Table4showstheoptimalbatterystoragerequirementintheintermediateyearsbetween2020and2030.11Insomestates/Unionterritories,suchasDelhi,thepeakloadisshiftingtoevenlaterinthenight(about11pmtomidnight)inthesummer,drivenbyresidentialspacecoolingdemand.30Table4:BatteryStorageRequirementinthePrimaryLeastCostCase202520272030BatteryStorageRequirement(All-India)12GW/48GWh31GW/125GWh63GW/252GWhFigure13:AveragehourlynetgenerationfrombatteriesinFY2030inthePrimaryLeast-CostCaseNote:PositivevaluesimplybatterydischargewhilenegativevaluesimplybatterieschargingBatteriesandagriculturalloadshiftprovidediurnalbalancingbutcannotaddressRE’sseasonalmismatchwithelectricitydemand.Nationally,theelectricitydemandpeaksinSeptember-October,butREoutputpeaksbetweenJuneandSeptember(monsoon),mainlyduetohighwindgeneration.BetweenOctoberandFebruary,windgenerationdropssignificantly(reachinganaveragecapacityfactoraslowas10%–15%,vs.50%–60%duringmonsoonseason),whilesolargenerationalsodrops.Naturalgaspowerplantsplayacrucialroleinprovidingsuchseasonalenergybalancing.Iflow-costenergystoragecouldnotbedeployedatsuchascale,additionalthermalinvestmentsbeyondthe23GWnetadditionswillbeneededthrough2030tomaintaingridreliability,butsuchassetswilloperateatlowcapacityfactors.(c)Existingnaturalgaspowerplantsassistinseasonalbalancing.Evenifthedomesticnaturalgasallocationtopowersectorremainsthesame(8bcm/yr),flexibleoperationofIndia’sexistingnaturalgasassets(25GW)couldbecrucialformaintaininggriddependability,especiallyduringlow-REmonths.FromOctobertotheendofJanuary,asthemonsoonwindwanesandsolaroutputdrops,thegridneedssignificantadditionalenergywhichthebatteriesareunabletosupplyduetolowRE31generation.Wefindthattheexistingnaturalgasplantscanfillthisgapcost-effectivelyrelativetobuildingnewcoalpowerplantsoperatingatlowPLFstomeetsuchseasonalloads.However,seasonal,flexibleoperationofgaspowerplants,andbycorollarygaspipelines,wouldrequirecoordinatedregulatoryinterventionsinthepoweraswellasgassector,whicharediscussedinaseparatereport.Optimalutilizationofhydropowerresources,particularlythe20-25GWofreservoir-basedcapacity,wouldalsobeimportantforseasonalbalancing,especiallyformeetingtheeveningpeakdemandduringthelowREseason.However,severalrestrictionsonhydropowerdispatchsuchasmaintainingthewaterflowsinkeyriverbasins,thecascadedandmultipurposenatureofhydropowerprojectsetc.,limittheirvaluetothegrid.(d)Flexibleresourceshelpmanagethermalramps.Figure14showscoalrampsinthePrimaryLeast-CostCaseandtheCEAcase.12Givenhigherflexibilityinthesystem,coalrampsdecreasesignificantlyinthePrimaryLeast-CostCase,eventhoughbothcaseshaveasimilaramountofREcapacityonthegrid.Batteries,hydropowerresources,andnaturalgasplants(ifoperatedinopen-cyclemode)arebestsuitedtotacklethesteeprampupanddownfromthemiddaysolargeneration.Themaximumsystemramprequirementin2030is60GW/hour.Whilebatteriesperformmostofthisramping,coalpowerplantsarerequiredtorampatlessthan25GW/hour(Figure14).Withoutsufficientbatterycapacity,coalplantsmayhavetomeetover50GW/hroftheramprequirementasseenintheCEAcase,leadingtolowerheatrates,additionalwearandtearandincreasingvariableandO&Mcosts.Figure14:HourlycoalrampsinthePrimaryLeast-CostCaseandCEACasein2030.Eachpointshowstotalcoalfleetrampingperhour.12NotethattheCEACaseshowstheresultsofhourlydispatchsimulationsassumingthecapacityadditionenvisagedinCEA’sOptimalCapacityExpansionreport.Ourassumptionsonoperationalparameters,generationcapacitysiting,transmissionbuildoutetcwouldbesomewhatdifferentfromthoseusedbyCEAintheirstudy.Therefore,theCEAcaseresultsshouldbeinterpretedasindicativeonly.32(e)Thegridhassufficientcapacitytorundependablyduring“high-stress”periods.Herewelookathowgridreliabilityismetduringdaysofexcessivestress,examiningdispatchresultsforeachofthepeakloadweek,thehighestnetloadweek,andthehighestsystemrampsweek.1.PeakLoad:Nationally,thesystemloadpeakof340GWinFY2030wouldoccuronSeptember15at7pm(Figure15).Atthistime,windgenerationisashighas50GW(~33%capacityfactor).Coalplants(ex-busgenerationof170GW)andnuclearplants(ex-busgenerationof13GW)willoperateatnearfullcapacity,providingmostlythebaseloadsupport.Peaksupportisprovidedby60GWofbatteriescombinedwith40GWofhydropower.Naturalgasplantsalsogenerateduringtheeveningpeakandcontinuetooperatethroughthenight.Figure15:Hourlydispatchinthepeakloadweek(2030)inthePrimaryLeast-Costscenario2.NetLoadPeak:Netload(orresidualload)isdefinedasloadminustheoutputfromvariableREsources(solarandwind).Netloadiscriticalfromthesystemplanningandoperationsperspectivebecauseitistheeffectiveloadthattherestofthesystemresources,suchasthermal,nuclear,andhydropower,havetomeet.Thehighestnetload(national)occursinOctober,whenthesystemloadisstillhighbutthewindgenerationhasreduceddrastically(Figure16).InFY2030,netloadpeakof312GWoccursonOctober13.Similartothepeakloadweek,allresourcesoperateatnearfullcapacityduringnetloadpeak.Forexample,coal(ex-busgenerationof175GW)andNuclear33(ex-bus13GW)providethebaseloadsupportwhilebatteries(58GW)andhydropower(40GWincludingsmallhydro)providetheeveningpeakingsupport.Naturalgaspowerplantsrunatacapacityfactorofover60%providingtheseasonalbalancing.Thehighestnetloadpeakdaywouldstillhaveabufferof20GWofundispatchedcoal(ex-buscapacityof15GW,afteraccountingforforcedoutagesandauxiliaryconsumption)andabout5GWofbatterycapacity,whichwouldprovidetheoperatingreservesneededtocoveranycontingencies,suchaserrorsinday-aheadREorloadforecasts.Figure16:Hourlydispatchinthehighestnetloadweek(2030)inthePrimaryLeast-Costscenario3.HighestSystemRamps:Giventhesteepreductioninsolargenerationcoupledwithloadincreaserightaftersunset,thesystemrampingrequirementsincreasesignificantlyby2030owingtohighsolarinstalledcapacity.Thehighestsystemrampofabout61GW/houroccursonMarch26at7pm.Thesystemcanmeettheserampswiththelargebatterycapacity,rampingat45-50GW/hranddispatchablehydropowercapacity,rampingat10-15GW/hr,withsupportof2-5GW/hrprovidedbygaspowerplantsand5-10GW/hrprovidedbycoalpowerplants(Figure17).34Figure17:HourlydispatchinthehighestREvariabilityweek(2030)inthePrimaryLeast-CostCaseAdditionaldispatchresultsareshowninAppendixII.4.Inter-annualvariationinrenewableenergygeneration:Ifsolarandwindenergyresourcesexperiencesignificantlyreducedoutputoverseveralconsecutivedays,thereisasignificantriskthatREgenerationandstoredelectricitymaynotbeabletomeetdemand,especiallyduringpeakornetloadpeakperiods.Suchasituationdoesnotariseinoursimulationyear(2018weatheryear),and,foreveryhouroftheyear,theIndianpowersystemoperatesdependablyevenwhenabout80%ofIndia’sinstantaneousloadismetbyrenewableenergy.However,simulatingonespecificweatheryeardeterministicallymaynotcapturethelow-probability,high-costeventofextremelylowREgenerationduringhighdemandperiodsthatmayoccuronceeveryseveralyears.Assessingtheperiod1981–2015,Shaneretal(2018)findthelowestsolaroutputaggregatedoverthecontinentalUnitedStatesonanydayisatmost20%lowerthanthemeansolaroutputonthatdayoftheyear,whereasthewindoutputcanbeasmuchas50%lower.Theyalsofindthetemporalcorrelationinsolargenerationdecreasesrelativelyweaklywithdistance,implyinglimiteddiversitybenefitsofaggregationoveralargerarea.Unfortunately,nosuchassessmentexistsinIndia.Asafirstapproximation,webelievethatthesolaroutputaggregatedoverIndiaonanydaylikelywillbeatmost20%lowerthanthemeansolar35outputonthatday,similartowhatisobservedintheUnitedStates(muchbiggerlandmasscomparedwithIndia).Inordertoassesstheimpactduringtheperiodsofhighestsystemstress,wesimulatethehourlydispatchduringthepeakloadandthenetloadpeakweeksbyassumingsolarandwindoutputtobe20%and50%lower,respectively.Wefindthatduringbothweeks,thesystemwasabletomeetsuchdeepdropsintherenewablegeneration.Duringthepeakloadhour(September)thesystemstillhassignificantslackincoalcapacity(25GWofundispatchedavailablecapacitythatisnotgenerating)andgascapacity(5GWofundispatchedavailablecapacity).Sincesolargenerationisalmostzeroduringeveningpeakhours,ifwindgenerationdropsby50%(from33GWto16GW),thereisenoughslackfirmcapacityinthesystemthatcouldcompensateforsuchdropandmeetthedemand.Duringthenetloadpeakweek(October),thewindenergygenerationhasalreadydroppedsignificantly(generatingabout15GWduringthenetloadpeakhour),systemisabletohandleanadditional7GW(50%)dropinwindgeneration.Asshownpreviously,coalandbatteriesstillhaveabout20-25GWundispatchedavailablecapacitythathelpsfillinthedropinwindgeneration.Fromtheenergybalanceperspective,thesystemisstillabletohandlesuchdropsbecausebatteriescouldbechargedusingcoalorgasbasedgenerationduringsustainedperiodsofreducedgenerationfromrenewableenergyresources.ItisimportanttounderstandthatthestudyhassimulatedhourlygridoperationsusingaDCOptimalPowerFlowformulation.ThisimpliesthatsomeoftheoperationalissuesthatmayariseinanACpowersystemsuchasreactivepowercompensation,impactonlinevoltagesandgridfrequencyetc.couldnotbeassessedinthisstudy.Deeperanalysesusingappropriatesimulationtools(suchasPowerSystemSimulatorforEngineering(PSSE)etc.)wouldbeneededtofullyunderstandsuchimpacts.Also,whileweinclude5%spinningreservesinourproductioncostmodel,wehavenotexplicitlymodeledREorloadforecasterrorsandsomeadditionalworkwouldbeneededforamorenuancedassessment.However,withstateoftheartforecastingtechniques3.3.By2030,about140GWofadditionaltransfercapacitybuildoutisfoundeconomicalBetween2020and2030,about140GWofadditionaltransfercapacityneedstobebuiltinthePrimaryLeast-CostCase,ofwhichabout33GWarerequiredoninterregionalcorridorsand107GWoninterstatecorridors(Figure18).Therequiredtransmissioncapacitybuildoutcouldbeapproximatelytwiceashighi.e.280GW.Thetransmissioncorridorsthatneedthemostadditionsareasfollows:MaharashtratoChhattisgarh,UttarPradeshtoDelhi,MaharashtratoKarnataka,RajasthantoHaryana,RajasthantoPunjab,GujarattoMadhyaPradesh,MadhyaPradeshtoUttarPradeshetc.AppendixIIprovidesadditionalinsightsonwhichtransmissioncorridorswouldbestrengthened.TheaveragetransmissionbuildoutcostbyFY2030wouldbearoundRs.0.08–0.10/kWh(0.11-0.13cents/kWh).In2020,thetotalinterstateandinter-regionaltransmissioncapacitywasabout200GW.36ER=EasternRegion,SR=SouthernRegion,NR=NorthernRegion,WR=WesternRegionFigure18:TransmissionbuildoutinthePrimaryLeast-CostCase(2020–2030)Importantly,theadditionalinterstate/inter-regionaltransmissioninvestmentsarenotprimarilydrivenbytheREcapacityaddition.BecauseIndia’sloadisexpectedtoalmostdoubleoverthedecade,significantadditionaltransmissionsinvestmentswouldberequiredirrespectiveoftheresourcemix.Infact,wefindthatbecauseofthedeepreductioninsolarandbatteryprices,andgenerallygoodsolarresourcequalityinmoststates,itisfeasibletospreadoutREandstorageinvestmentsthroughoutthecountry,whichmaynotbefeasiblewithcoalpowerplantsthatneedtobesitednearthecoalminestoachievelowcosts.Weranahypothetical“NoNewRE”scenario,wherenonewrenewableenergyorstoragecapacityisbuiltandallincrementalload(2021-2030)ismetonlybybuildingadditionalcoalpowerplants,andfoundthatthetotalinterstate/inter-regionaltransmissionbuildouttobe10%higherthanthePrimaryLeastCostcase.Wehavenotassessedindetailtheinvestmentsinspur-linesthatconnectREpoolingsubstationswiththemaintransmissionnetwork.Nonetheless,initialestimatessuggestthatRs.10,500crore13ofinvestmentinthespur-lineinfrastructuremayberequiredby2030,whichisequivalenttoacostadderofRs.0.01/kWh(0.01cents/kWh)ofREgeneration.Table5showsinterregionalinterchangeforthePrimaryLeast-CostCase.Giventhelowerrelianceoncoal-firedpowerplantsconcentratedinafewregions,thenetinterchangedecreasesinthePrimaryLeast-CostCase,especiallybetweenwesternandnorthernregions,andeasternandsouthernregions.Figure19plotshourlylineloadingforkeyinterstatelinesduring2030.Ingeneral,becauseofthesignificantbatterystoragecapacity,wedonotfindanysignificant/consistentcongestiononthekeyinterstate/inter-regionalcorridorsexceptfortheChhattisgarh-Haryanacorridor.However,morenuancedassessmentusingmoredetailedtransmissionnetwork,sub-hourlyresolution,andACpowerflowanalysiswouldbeneededtoassessthetrueimpactonthetransmissionsystemoperations.13Assumingtheaveragelengthofspur-linestobe20km,theaveragecostofspur-linestobeRs.15,000/MW-km,andthelifeoflinestobe40years.37Table5:InterregionalinterchangeinPrimaryLeast-CostCaseTWh/yr(2030)Inter-regionalcorridorNetAnnualInterchange(TWh/yr)ER-NR27.6WR-NR56.2WR-SR-0.4ER-WR1.9ER-SR23ER-NER7.1Figure19:Hourlylineloadingforkeyinterfaces,PrimaryLeast-CostCase(2030)3.4Emissionsintensityofpowergenerationdropsby43-50%By2030,theaverageCO2emissionsintensityoftheIndianpowersectordropsfrom0.82kg/kWhin2020to0.47kg/kWhinthePrimaryLeastCostcase(43%reduction),andto0.41kg/kWhintheLowRECostcase(50%reduction).Relativetothe2020levels(1008MT/yr),totalCO2emissionsfrompowersectorin2030dropby3%intheLow-RECostcase(981MT/yr)andincreaseonlyby7%(1081MT/yr)inthePrimaryLeastCostcase,despiteneardoublingoftheelectricitydemand(Figure20).Importantly,underthePrimaryLeastCostcase,nearly80%ofthenetincrementalgenerationbetween2020and2030willbecontributedfromnewcleanenergyassets,includingnewRE,nuclear,hydropowerassets.UnderaLow-RECostcase,newcleanenergyassetscontributeabout90%ofthenetincrementalgeneration.Theavoidedcoalgenerationduetosolarandwindgenerationhasimmensehealthbenefitsin38theformofavoidedairpollution.Theresultantbenefitsduetoreducedmortalityandmorbidityaresignificantandneedtobeassessedinmoredetail.Figure20:Carbon-dioxideemissionsandintensityin2020and2030394SensitivityAnalysisWeassessthesensitivityofourresultsonthekeyassumptions:(1)cleantechnologycostsanddisruptionstothesolar/batteriessupplychainoradditionalsafeguarddutiestoreduceimports,(2)LNGprices,(3)utilities’unwillingnesstoretiretheirexistingcoalassets,(4)demandgrowth,and(5)implementationofanationalelectricitymarketliketheproposedMarkedBasedEconomicDispatch.Table6summarizesthesealternatepathways,andkeyinsightssetforthbelow.Table6:Summaryofresultsindifferentsensitivitycases(2030)ScenarioScenarioDescriptionCoal(GW)Gas(GW)Solar(GWDC)Wind(GW)Storage(GW)AvgCost(Rs/kWh)1LowDemandGrowthLoadgrowthis25%lower(2030Peak=290GW)20625230109323.642NoCoalRetirement25GWcoal(NEP)doesnotretireperplan23825301137573.623HighRECost(withsupplychainconstraints)Significantlyhighercostanddeploymentconstraintsforsolarandbatteries24245220142323.634LowLNGPriceLNGprice=$4.5/MMBTU(landed)22428308132663.585MBEDNationalleveleconomicdispatch23225302151593.565.1Nonewcoalcapacitybeyondwhatisunderconstructioniscost-effectiveifdemandgrowthislowerthanexpectedIfthepost-COVIDeconomicrecoverytakeslongerthanexpectedandtherateofloadgrowththrough2030dropsto4.7%peryear(insteadof6.3%perCEA’s19thEPSassumedinthePrimaryLeast-CostCasei.e.,ifpeakloadin2030wouldbe290GW),nonewcoalcapacityadditionisfoundtobeeconomical.Thesystemwouldonlyneedatotalof355GWoftotalREcapacitytomeetthelowerdemand:Solar:230GWWind:109GWOtherRE:15GWBatterystorage:32GW/128GWhTheaveragecostofgenerationinthisLowDemandGrowthcaseisRs.4.08/kWh,slightlyhigherthaninthePrimaryLeast-Costscenario,mainlybecauseofthelowerassetutilizationoftheexistingassets.5.2If25GWofcoalplantsdonotretireasplanned,438GWofsolarandwindcapacitywouldstillbecost-effective,thoughtheaveragecostofgenerationwouldincreasetoRs3.62/kWhascomparedtoRs3.59/kWhunderthePrimaryLeastCostCase40InthelatestNEP,CEAidentifiedabout25GWofagingcoalcapacityforretirementbetween2022and2027.Weanalyzeascenarioinwhichstateutilitiesdonotretireanyofthiscapacityby2030.Themodelinthiscasebuilds8GWofadditionalcoalcapacitybeyondthecapacityunderconstructionresultingin238GWofinstalledcoalcapacity,301GWofsolar,137GWofwind,and57GWofbatterystoragetobecost-effective(Table6).Generationfromcoalplantsincreasesmarginallyto1,168TWh/yr(49%ofgeneration),whilesolarandwindsourcesprovide35%oftotalgeneration.5.3WithhighREcostsduetohigherdutiesand/orsupplychainconstraints,additionalcoalandnaturalgascapacityisneeded,increasingtheaverageelectricitycostbyRs.0.047/kWhInthehighREcostscenario,inadditiontothehighcapitalcost/importtaxesonsolarandbatteries,wealsoassumesignificantsupplychainchallengesleadingtoonlyalimiteddeploymentoftheseresources.Asaresult,coalandnaturalgasexpandsignificantlytomeettheincreasingdemand.Theresourcemixis220GWDCofsolar,142GWofwind,35GWofbatteries,242GWofcoal,and45GWofnaturalgas(Table6).Coalprovidesoverhalfoftheannualgenerationin2030,andsolarandwindaccountforabout30%.Theaverageelectricitycostby2030wouldbeRs3.63/kWh,whichis1.4%higherthaninthePrimaryLeast-CostCase.Thisboundarycaseshowsthepotentialconsequencesofserioustradeorsupplydisruptions.5.4IftheLNGpricedropsto$4.5/MMBTU,flexibleoperationofgas-firedgenerationstartscompetingwithexpensivecoalpowerplantsTheLowLNGPricescenarioassumestheLNGprice(deliveredontheIndianshore)dropsto$4.5/MMBTU,whilethedomesticnaturalgaspricestaysthesame.Whilethispricedropisstillnotenoughtojustifybuildoutofnewnaturalgaspowerplants,gasgenerationincreasestoabout120TWh/yrby2030,comparedwith50TWh/yrinthePrimaryLeast-CostCase.Thereisnochangeincoalcapacity,butabout42TWhofexpensivecoalgenerationisreplacedbynaturalgasgeneration.Thefollowingistheresourcemixforthisscenario:Solar:308GWWind:132GWBatterystorage:66GW/264GWhDomesticgasavailabilityforthepowersectorisabout8.5bcm/yr(23mmscmd)by2030,andLNGconsumptionisabout14bcm/yr(10MTPA).5.5EffectsofMarket-BasedEconomicDispatch(MBED)ForFY2030,wesimulateanationalwholesaleelectricitymarketorMarketBasedEconomicDispatch(MBED)proposedbyCERCtoassesstheimpactonsystemoperationsandcostsinaREheavygrid.TosimulateMBED,weusetheresourcemixinourPrimaryLeast-CostCaseandrunthegriddispatchsimulationbyremovingtheeconomichurdlesthatstatesfaceforimporting/exportingelectricitytoother41states.Thephysicallimitsofthetransmissionsystemandstandardtransmissionwheelingchargeswouldstillapply.1.Moreefficientdispatch:Figure21showstheannualPLFvs.variablecostofeachindividualcoalpowerplantsdispatchedunderthePrimaryLeast-CostCaseandMBEDscenario.IntheMBEDscenario,thesystemdispatchestheplantswithlowestvariablecostatthehighestPLFs,whilethePLFcontinuestodropasthevariablecostincreases.WithMBED,althoughnationalcoalgenerationremainsalmostthesameasinthePrimaryLeast-CostCase(withstate-levelbalancing),thedistributionamongstateschangessignificantly.Becausetheutilizationofcheaperthermalassetsincreasessignificantly,thetotalvariablecostofgeneration(coalplusnaturalgas)decreasesby5%orRs14,000Cr/yr(~$2billion/yr)byFY2030(Table7)andthetotalCO2emissionsfromthepowersectorreduceby5%relativetothePrimaryLeastCostcase.Figure21:PLFvs.variablecost(VC)ofcoalplantsinthePrimaryLeast-CostCase(left)andMBED(right)scenario.Colorofthepointshowsthevariablecostofeachpowerplant.Table7:Totalsystemvariablecost(RsThousandCr/yr)in2030forthePrimaryLeast-Cost(statebalancing)andMBEDscenariosPrimaryLeastCost(StateBalancing)MBEDCoalVariableCost285276GasVariableCost1813Total3032892.Changeinpowerflowsandtransmissioncongestion:Figure22showsdurationcurvesforlineloadingoncertainkeyinterfacesintheMBEDscenario.MBEDincreasesthepossibilityofcongestiononcertaininterfaces,sotransmissionrequirementsshouldbeassessedindetail.42Figure22:HourlyloadingonkeylineinterfacesintheMBEDscenarioAsshowninTable7,inthePrimaryLeast-CostCasewithstate-levelbalancing,thereissignificantinterregionalexchange,mainlybecausemostREcapacityisconcentratedinthewestandsouth.IntheMBEDscenario,theinterregionalexchangesincreasesignificantly(Table8).Thisisexpectedasthesystemtriestodispatchthecheapestresourcesavailableacrossthecountry.Thetransmissioncapacityisenoughtohandletheseflows,butcongestionbecomesmorepronounced.Table8:Interregionaltransmissionflows(TWh/yr)inFY2030NetInterregionalExchange(TWh/yr)PrimaryLeast-CostMBEDER-NR27.667.2WR-NR56.2106.2WR-SR-0.45.9ER-WR1.97.8ER-SR2346.9ER-NER7.19.7435ConclusionsDramaticcostreductionsoverthelastdecadeforwind,solarandbatterystoragepositionIndiatoleapfrogtoamoreflexible,robust,andsustainablepowersystem—muchofwhichisyettobebuilt—fordeliveringaffordableandreliablepowertoserveanearlydoublingpowerdemandby2030.Inthisstudy,weassessacost-effectiveandoperationallyfeasibleinvestmentpathwayforIndia’selectricitygridbyenhancingsystemflexibilityandrobustnessthroughrenewableenergy(RE)andaspectrumofflexibleresources,suchasenergystorage,demandresponse(loadshifting),naturalgas,andelectricitymarkets.Thestudyachievesthisobjectivebyusinganindustrystandardpowersystemmodelingplatform(PLEXOS)andcomprehensiveelectricitygriddataattheindividualpowerplantlevel.WefindthattheleastcostresourcemixtomeetIndia’sloadin2030(the“PrimaryLeastCostCase”)consistsprimarilyofacombinationofREandflexibleresourcesasfollows:465GWofRE(307GWDCsolar,142GWwind,and15GWotherRE),63GW(252GWh)ofbatterystorage,60GWofloadshiftingtosolarhours(50GWagricultural+10GWindustrial),flexibleoperationoftheexistingnaturalgasfleetof25GW.About23GWofnetnewthermalcapacityinvestmentsarefoundtobecosteffectivebetween2020and2030–makingthetotalcoalcapacityrequirementin2030tobe229GW.IfREandstoragepricesdropperhistoricaltrends(LowRECostCase),theleastcostmixincludes547GWofRE(385GWDCsolar,147GWwind,and15GWotherRE),84GW/336GWhofbatterystorage,and206GWofcoalcapacity.Notethatiflow-costenergystoragecouldnotbedeployedatsuchascale,additionalthermalinvestmentsbeyondthe23GWnetadditionswillbeneededthrough2030tomeetthepeakdemand,butsuchassetswilloperateatlowcapacityfactors.ThisimpliesthatIndia’sincrementalelectricitydemandthrough2030willlargelybemetbynewinvestmentsinrenewableenergy(RE)andenergystorageandexistingthermalassets.Asaresult,between2020and2030,despiteneardoublingofIndia’selectricitydemand,thetotalCO2emissionsfromthepowersectorremainalmostthesamewhileemissionsintensityofelectricitygenerationdropsby43%-50%.Also,totalcoalconsumptioninthepowersectorremainsalmostthesameasthatin2020,implyingcleanenergytransitionisunlikelytoleadtolossofcoalminingandtransportationjobsinthenear-tomedium-term.Continuedbuild-outofnewcoal-firedassetswillcausesignificantfinancialandtechnicalstressontheexistingcoalpowerplants.About23GWofnewthermalinvestmentsbeyondthecapacitythatisalreadyunderconstructionisfoundtobeeconomical.WealsofindthatIndia’selectricgridwithmassiveREandbatterystoragecapacitywillbedependableineveryhouroftheyear.Between2020and2030,averagecostofgenerationisfoundtoreduceby8-10%becauseofthefallingrenewableenergyandstorageprices.Theinterstatetransmissioninvestmentisfoundtobemodest-140GWofadditionalinterstatetransfercapacitybuildoutthrough2030.Overall,asIndia’sgridattainshigherpenetrationsofrenewables,balancingitsvariabilitythroughaspectrumofflexibleresources–suchasenergystorage,demandresponse(agriculturalloadshifting),flexibleoperationofgaspowerplants,andbecomesincreasinglyimportantforensuringtheaffordability,stability,andreliabilityofgridpower.Theflexibleresourcesworkintandemtomaintainthehourlysupply-demandbalance.DuringthehighREgenerationseason(JunethroughSeptemberforwindandMarchthroughJuneforsolar),energystorageandagriculturalloadshiftingprovidediurnalgridbalancing.44Batterieschargeduringthedaytime(coincidentwithsolargeneration)anddischargeduringthemorningandeveningpeakperiods(4-6hourstotaleachday).Theyalsohelptomeetsteepsystemramps.Shiftingagriculturalloadtosolarhoursincreasestheday-timeloadby40-60GWwhilereducingthenight-timeloadandtherebythebaseloadcapacityrequirementby30-50GW.Asaresult,only180GWofcoalcapacityisdispatchedmainlyasabaseloadresource.DuringthelowREgenerationseason(OctoberthroughFebruary),the25GWofexistingnaturalgascapacity(inlieuofcoal-firedassets)playsacrucialroleinprovidingtheseasonalbalancingwithmostoftheirdispatchoccurringduringthesemonths.Wideninganddeepeningoftheelectricitymarkets,suchasimplementationoftheMBED,canprovideadditionalflexibilityinsystemdispatchwhilereducingthethermalvariablecostsandCO2emissionsby5%in2030.ForIndiatoachievetheleast-costresourcemixindicatedinthisstudy,amodestdeclineinthecurrentRE(5-10%by2030)andamorepronounceddeclineinthecurrentstoragecosts(30-40%by2030),consistentwithhistoricaltrendsandprojectionsbyotherstudies,willberequired.Also,deployingREandstorageatsuchasignificantscalewilllikelyrequireaddressingsupplychainchallengesandsecuringadequatefinancing.Finally,criticalpolicyandregulatorychangesmustbeexpeditiouslyimplementedinorderforIndiatomoveontotheleast-costpathway.Thesechangesinclude,amongotherthings,anuancedlong-termresourceadequacyframeworkforsystemplanningandprocurement,aregulatoryframeworkforenergystoragethatvaluesandcompensatesthisresourceforitsfullfunctionality,andgasreformsthatpromoteflexibleoperationsandincreasedutilizationofIndia’sgaspipelinesystemtoenablecost-effective,flexibleoperationofIndia’sexisting25GWfleetofgas-firedpowerplantsforseasonalanddiurnalgridbalancing.Weconductedaseparateanalysisofspecificpolicyandregulatorychangesthatmaybeneeded,whicharesummarizedinthenextsection.456PolicyandRegulatoryRecommendationsThisstudypointstoaninvestmentpathwayforIndia’stransitiontoaflexible,robust,andcleanerpowersystemwhilemeetingitsgrowingloadreliablyandatleastcost.Strongpoliciesandregulatoryinterventions–particularly,aroundthreemainareas:resourceadequacy(RA),stateresourceplanningandprocurement,andshort-termmarketsandsystemoperationswouldberequiredforachievingtheprojected2030resourcemixandchangesinsystemoperations,aresummarizedinthissection.Aseparatereportprovidesanin-depthdiscussiononeachofthefollowingpolicyandregulatoryrecommendations.ResourceAdequacy(RA)●Nearer-term(1–3years).DevelopanationalRAmechanismthatrequiresareservemarginstudy,mandatesstandardsforstateloadforecasting,createsRArequirementsforstates,definestransparentmethodsforcapacitycrediting,developsmarketsforRAcapacity,andimplementsdeficiencypenaltiesfornon-complianceandincentivesforgeneratoravailability.●Longer-term(4–10years).DeveloptransparentpricingformechanismstoprovideRAandprobabilisticmethodsforcapacitycrediting(demandresponse,solar,storage,wind).ResourcePlanningandProcurement●Nearer-term(1–3years).IntegrateRArequirementsintostateandDiscomresourceplanningandprocurement,pilotall-sourcecompetitiveprocurementinafewstates,andbuildthecapacityofstates/Discomstoconducteconomicmodelingneededtosupportall-sourcecompetitivesolicitations.●Longer-term(4–10years).Developnationalguidelinesforcompetitiveprocurementandexpandall-sourcecompetitiveprocurementtoallstates.MarketsandSystemOperations●Nearer-term(1–3years).Completeimplementationofcurrentmarketreformsandreviewmarketparticipationrulesforenergystoragetoensurethatitsfullflexibilityandfunctionalitycanbeutilizedandcompensatedthroughmarkets.●Longer-term(4–10years).Morecloselyalignshort-termmarketsandpowersystemoperationbyimplementinglocationalmarginalprice-based(LMP-based)securityconstrainedeconomicdispatch(SCED)inreal-timemarkets.FlexibleoperationoftheexistinggaspowerplantsTooperategaspowerplantsflexibly,existinggas-firedgeneratorswillneedaccesstofuelandtheabilitytochangetheamountoffuelconsumedtoincreaseordecreaseoutput.Thespecificpolicyandregulatorychangesneededinthegassectortoenablesuchoperationalflexibilityandbettergas-electricsectorcoordinationareprovidedinthetablebelow.46Table9:Policyandregulatoryrecommendationsforgas-electriccoordination#Keygas-electriccoordinationtopicsSpecificPolicy&RegulatoryInnovation1InformationtransparencytoensurephysicalaccessandavailabilityofgastopowerplantsRequireestablishmentofanelectronicbulletinboardthatprovidesinformationonpipelinecapacityavailability,criticalnoticesofsystemoutagesandnaturalgaspipelinetariffs.2Enablingcontractualflexibilityincommoditygas/LNGsupplyagreementsCommoditysupplyagreementsshouldincludeflexiblepricingandvolumeoptionality3EnablingflexibilityinpipelinegastransportationagreementsEnableflexibilityingastransportationbyeliminatingtake-or-pay(ship-or-pay)provisions,therebysupportingcommoditycontractsonflexibletermsAmendpipelineaccesscodetorequirepipelineoperatorstoofferinterruptibletransportationservicesinadditiontofirmserviceRequirepipelineoperatorstooffervalue-addedno-notice,hourly,non-ratableflow,andlinepackderivedstorageservicesthatarerelatedtotransportationflexibilityAllowintra-daygascapacitynominationsandrequirepipelineoperatorstomoveinitialnominationschedulestoonedaybeforeflowsAllowresaleoffirmtransportationservicetothirdpartiesRequirepipelineoperatorstoprovideflexibilityinentry/exit(receipt/delivery)pointsAmendpipelineaccesscodetorequirepipelineoperatorstoofferstandardizedgastransportationagreementtermsandconditions4Wholesalegasmarket,pricingandstructuralreformsFacilitatecommercialandoperationalunbundlingofpipelinetransportation,commoditygassalesandmarketingfunctionsandallowopenaccesstothirdpartyshippers5RationalizationofnaturalgaspipelinetariffstructureAmendtariffregulationstomovetowardsatwo-parttariffforfirmtransportationserviceandinterruptibleservicesonvolumetricbasis477CaveatsandFutureWorkAlthoughweassessanoperationallyfeasibleleastcostpathwayforIndia’spowersystemusingweather-synchronizedloadandgenerationdata,furtherworkisneededtoadvanceourunderstandingofotherfacetsofapowersystemwithhighREpenetration.First,thisreportprimarilyfocusesonrenewable-specifictechnologypathwaysanddoesnotexplorethefullportfolioofcleantechnologiesthatcouldcontributetofutureelectricitysupply.First,issuessuchaslossofloadprobability,systeminertia,andalternating-currenttransmissionflowsneedfurtherassessment.Optionstoaddresstheseissueshavebeenidentifiedelsewhere(forexample,Denholm,2020).Second,ourassessmentdoesnotfullyaddresstheoperationalimpactsofday-ahead/intra-dayforecasterrorsinREandload.However,severalstudieshaveshownthatwithstateoftheartforecastingtechniques,theimpactofsuchforecasterrorsappearstobesmall(forexample,Hodge,2015;Martinez-Anido,2016).Althoughthisanalysisdoesnotattemptafullpower-systemreliabilityassessment,weperformscenarioandsensitivityanalysistoensurethatdemandismetinallperiods,includingduringextremeweathereventsandperiodsoflowrenewableenergygeneration.Thismodelingapproachprovidesconfidencethatintegratingover450GWofrenewableenergyintothegridistechnicallyfeasibleandeconomicallydesirablebyFY2030.Thisiscritical,becausepowersectordecarbonizationcanbethecatalystfordecarbonizationacrossalleconomicsectorsviaelectrificationofvehicles,buildings,andindustry.Owingtotheglobalnatureofrenewableenergyandbatterymarkets,ourstudyindicatesthepossibilitythatcost-effectivedecarbonizationcanbeanear-termreality.Finally,althoughthisreportdescribesthesystemcharacteristicsneededtoaccommodatehighlevelsofrenewablegeneration,itdoesnotaddresstheinstitutional,market,andregulatorychangesthatareneededtofacilitatesuchatransformation.Ourcomplementaryanalysis,presentedinaseparatereportandsummarizedinthepolicyrecommendationssection,identifiesmanyofthesesolutions(Deorahetal,2021).Furtherdetailsonthekeyassumptionsandresultscanbefoundintheappendices.48ReferencesBNEF(2020a).India’sCleanPowerRevolution:Asuccessstorywithglobalimplications.BloombergNewEnergyFinance.Authors:Gadre,Rohit,etal.BNEF.(2020b)BatteryPackPricesCitedBelow$100/kWhfortheFirstTimein2020,WhileMarketAverageSitsat$137/kWh.BloombergNewEnergyFinance.Retrievedfrom:https://about.bnef.com/blog/battery-pack-prices-cited-below-100-kwh-for-the-first-time-in-2020-while-market-average-sits-at-137-kwh/Martinez-Anido,CB,BenjaminBotor,AnthonyR.Florita,CarolineDraxl,SiyuanLu,HendrikF.Hamann,Bri-MathiasHodge.(2016),“Thevalueofday-aheadsolarpowerforecastingimprovement”,SolarEnergy(129)pp192-203.CEA(2010).MonthlyEnergyGenerationReports–JanuarythroughDecember2010.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2011).MonthlyEnergyGenerationReports–JanuarythroughDecember2011.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2012).MonthlyEnergyGenerationReports–JanuarythroughDecember2012.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2013).MonthlyEnergyGenerationReports–JanuarythroughDecember2013.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2014).MonthlyEnergyGenerationReports–JanuarythroughDecember2014.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2015).MonthlyEnergyGenerationReports–JanuarythroughDecember2015.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2016a).MonthlyEnergyGenerationReports–JanuarythroughDecember2016.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2016b).“ReviewofPerformanceofThermalPowerStations2014-15”.CentralElectricityAuthority,MinistryofPower,GovernmentofIndia.CEA(2017a).MonthlyEnergyGenerationReports–JanuarythroughDecember2017.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2017b).“ReportonNineteenthElectricPowerSurveyofIndia”.CentralElectricityAuthority,MinistryofPower(GovernmentofIndia).CEA(2018a).MonthlyEnergyGenerationReports–JanuarythroughDecember2018.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.49CEA(2018b).“NationalElectricityPlan:VolumeI”.CentralElectricityAuthority,MinistryofPower,GovernmentofIndia.CEA(2018c).“ReviewofPerformanceofThermalPowerStations2015-16”.CentralElectricityAuthority,MinistryofPower,GovernmentofIndia.CEA(2019a).MonthlyEnergyGenerationReports–JanuarythroughDecember2019.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2019b).“ReviewofPerformanceofHydroPowerStations2017-18”.CentralElectricityAuthority,MinistryofPower,GovernmentofIndia.CEA(2019c).ReviewofPerformanceofThermalPowerStations2017-18.CentralElectricityAuthority,MinistryofPower,GovernmentofIndia.CEA(2020a).BroadStatusofThermalPowerPlants–JanuarythroughDecember2020.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2020b).MonthlyEnergyGenerationReports–JanuarythroughDecember2020.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2020c).MonthlyEnergyGenerationReports–JanuarythroughDecember2020.CentralElectricityAuthority(OperationPerformanceMonitoringDivision),MinistryofPower,GovernmentofIndia.CEA(2020d).“ReportonOptimalGenerationCapacityMixfor2029-30”.CentralElectricityAuthorityMinistryofPower,GovernmentofIndia.CEA(2021).CO2EmissionsDatabasefromPowerSector.CentralElectricityAuthority,MinistryofPower(GovernmentofIndia).CERC(2015).“CentralElectricityRegulatoryCommission(TermsandConditionsforTariffDeterminationfromRenewableEnergySources)(ThirdAmendment)Regulations,2015.”CentralElectricityRegulatoryCommission(GovernmentofIndia).CERC(2016a).“CentralElectricityRegulatoryCommission(IndianElectricityGridCode)(FourthAmendment)Regulations,2016”,L-1/18/2010-CERC.NewDelhi:CentralElectricityRegulatoryCommission(GovernmentofIndia).http://www.cercind.gov.in/2016/regulation/124_1.pdf.CERC(2016b).“CentralElectricityRegulatoryCommission(DeviationSettlementMechanismandRelatedMatters)(ThirdAmendment)Regulations,2016”.CentralElectricityRegulatoryCommission(GovernmentofIndia).CERC(2019).“CentralElectricityRegulatoryCommission(TermsandConditionsofTariff)Regulations,2019”.CentralElectricityRegulatoryCommission(GovernmentofIndia).50CERC(2021).“CalculationofAveragePowerPurchaseCost(APPC)rateatthenationallevel”.CentralElectricityRegulatoryCommission(GovernmentofIndia).Deorah,Shruti,NikitAbhyankaretal(2020).“EstimatingtheCostofGrid-ScaleLithium-IonBatteryStorageinIndia”.LawrenceBerkeleyNationalLaboratory,Berkeley.Deshmukh,Ranjit,GraceWu,andAmolPhadke.(2017).“RenewableEnergyZonesforBalancingSitingTrade-OffsinIndia.”,LawrenceBerkeleyNationalLaboratory,Berkeley.Denholm,Paul,TrieuMai,RickWallaceKenyon,BenKroposki,andMarkO’Malley(2020).“InertiaandthePowerGrid:AGuideWithouttheSpin”.Golden,CO:NationalRenewableEnergyLaboratory.NREL/TP-6120-73856.Hodge,B.M.,A.Florita,J.Sharp,M.MargulisandD.Mcreavy.(2015).“TheValueofImprovedShortTermWindPowerForecasting”,NationalRenewableEnergyLaboratory,GoldenCO.NREL/TP-5D00-63175IEA(2021).“RenewablesIntegrationinIndia”,InternationalEnergyAgency(IEA)andNationalInstitutionforTransformingIndia(NITI)Aayog,GovernmentofIndia.KPTCL(2020).HourlySystemLoadData2010-2020,KarnatakaPowerTransmissionCorporationLimited.MERIT(2020),DatabaseonMarginalCostsofElectricityGeneration.MeritOrderDispatchofElectricityforRejuvenationofIncomeandTransparency.MinistryofPower.Retrievedfrom:http://meritindia.in/MOSPI(2018),“EnergyStatistics2018”.MinistryofStatisticsandProgramImplementation,GovernmentofIndia.MOSPI(2019),“EnergyStatistics2019”.MinistryofStatisticsandProgramImplementation,GovernmentofIndia.MOSPI(2020),“EnergyStatistics2020”.MinistryofStatisticsandProgramImplementation,GovernmentofIndia.MSLDC(2020).HourlySystemLoadData2010-2020,MaharashtraStateLoadDispatchCenter(MaharashtraStateElectricityTransmissionLimited).NREL(2020).“Least-CostPathwaysforIndia’sElectricPowerSector”.NationalRenewableEnergyLaboratory.Authors:AmyRose,IlyaChernyakhovskiy,DavidPalchak,SamKoebrich,andMohitJoshi.NREL(2021a).“AnnualTechnologiesBaseline2021”,NationalRenewableEnergyLaboratory.NREL(2021b).“EnergyStorageinSouthAsia:UnderstandingtheRoleofGridConnectedEnergyStorageinSouthAsia’sPowerSectorTransformation”.NationalRenewableEnergyLaboratory.Authors:IlyaChernyakhovskiy,MohitJoshi,DavidPalchak,andAmyRose.51Phadke,Amol,NikitAbhyankar,RanjitDeshmukh,JuliaSzinaietal(2020),“Cost-effectivedecarbonizationofCalifornia’sPowerSectorby2030WiththeAidofBatteryStorage”,LawrenceBerkeleyNationalLaboratory,Berkeley.POSOCO(2020)“VariableCostDataforGeneratorsProvidingFastResponseAncillaryServices”,PowerSystemOperationsCorporation(POSOCO),MinistryofPower,GovernmentofIndia.Shaner,M.R,S.J.Davis,N.S.Lewis,andK.Caldeira(2018),“GeophysicalconstraintsonthereliabilityofsolarandwindpowerintheUnitedStates,”EnergyEnviron.Sci.,vol.11,no.4,pp.914–925,Apr.2018.TERI(2020).“RenewablePowerPathways:ModellingtheIntegrationofWindandSolarinIndiaby2030”.TheEnergyandResourcesInstitute.Authors:ThomasSpencer,NeshwinRodrigues,RaghavPachouri,ShubhamThakre,G.Renjith.528AppendixI:KeyAssumptionsandData8.1Wind,solar,andbatterystoragecostsForthefiscalyear2019-20,weuseactualreverseauctionresultsforassessingthesolarandwindlevelizedcostofenergy.Toaccountfortheimpactofsafeguarddutyonimportedcomponents,weincreasethe2019-20solarauctionpricesby10%basedonseveralCentralElectricityRegulatoryCommission(CERC)andMaharashtraElectricityRegulatoryCommission(MERC)tariffordersallowingforthepass-throughofsafeguardduty.Forexample,averagesolarreverseauctionpriceinFY2019-20wasfoundtobeRs2.51/kWh;byapplyingthesafeguardduty,thecorrectedLCOEcomestoaboutRs2.76/kWh,foranaverageresourcequalityregionwithDCcapacityfactorof19%.WeadjustLCOEperresourcequalityinaregion.Forexample,inRajasthanandGujarat,weadjustthe2020LCOEtoRs2.4/kWh,whichreflectsthebetterresourcequalityandanaverageDCcapacityfactorof22%.Batterystoragecapitalcostshavebeentakenfromourpreviousanalysisthatassessesthegrid-scalebatterystoragecostsinIndia(Deorahetal,2020).Forprojectingthefuturecostsofwind,solar,andbatterystoragethrough2030,wecreatethreescenarios–low-,base-,andhigh-.Thelow-costcaseassumescostreductionsbetween2020and2030areinlinewithhistoricaltrendsinIndia.Thebaseormid-costcaseassumescostreductionsbetween2020and2030tobehalftheobservedhistoricalrate.Thehigh-costcaseassumesthecosttrajectoryofcleantechnologiesishigherthaninthebasecase,whichcouldoccurforvariousreasons,suchasslowerreductionsinglobalprices,restrictionsonimports,orsolarandbatterysupplychaindisruptionsthatlimitthecapacitythatcouldbeinstalledinthefirstfewyearsofthedecade(10GW/yr).Weassumedomesticmanufacturingcatchesupbymiddleofthedecade,andnewinstallationsarenotconstrainedbeyond2025.TableA1listsourwind,solar,andbatterystoragecostassumptions.TableA1:AssumptionsonAverageLevelizedCostofEnergyforWindandSolarandLevelizedCostofStorageforStandaloneBatteryEnergyStorageSystems(Rs/kWh)HighCostMid(orBase)CostLowCostWindSolarBattery(4-hour)WindSolarBattery(4-hour)WindSolarBattery(4-hour)20203.22.86.03.22.86.03.22.86.020213.22.75.93.22.75.83.22.65.620223.22.75.83.22.65.53.12.45.220233.22.65.73.22.55.23.02.34.920243.22.65.63.12.45.02.92.14.520253.22.55.43.12.34.72.82.04.220263.12.55.33.12.34.52.81.94.020273.12.45.23.02.24.32.71.83.720283.12.45.13.02.14.12.61.73.420293.12.35.03.02.13.92.61.63.220303.12.34.93.02.03.72.51.53.0Note:WindandsolarLCOEsareshownfortypicalprojectswithaveragecapacityfactorsof26%and19.2%,respectively.538.2ConventionalpowerplantfixedcostsConventionaltechnology(coal,naturalgas,hydro,biomass,anddiesel)capitalandfixedO&McostshavebeentakenfrommultiplesourcesincludingCERCtariffnorms,CEA’soptimalcapacityexpansionreport,andindustryconsultations.TableA2summarizestheassumptions.TableA2:AssumptionsonFixedCostsofConventionalTechnologiesTechnologyCapitalCostofNewCapacity(RsCr/MW)FixedO&MCost(RsCr/MW-yr)Coal(Ultrasuper-critical)7.850.188Gas(CCGT)4.50.113Hydro#N/A0.15Nuclear#N/A0.15Biomass#N/A0.113Note:Hydro,Nuclear,andBiomasscapacitiesarenotoptimizedandthecurrentCEA/DAEplansaretakenasgiven.Weassumethereal(inflation-adjusted)weightedaveragecostofcapital(WACC)of8%,whichisequivalenttoanominalWACCof11.6%(nominalinterestrateof11%andreturnonequityof14%,assumingadebt-to-equityratioof80:20).8.3CoalpricesandvariablecostsForexistingcoalpowerplants,wetakethevariablecostsofexistinginterstategeneratingstations(ISGS)fromreportsavailableundertheReservesRegulationAncillaryServices(RRAS)mechanism.VariablecostsforstategeneratorsandIPPsarefromregulatoryordersbyIndianstatecommissions.Forplantswithnorecentdataavailablefromregulatoryorders,weusepowerthevariablecostdatafromMinistryofPower’sMERITdatabase.Forpowerplantswithnodataavailable(lessthan5GW),weusetheaveragevariablecostsforthattechnologyandsizeintheirstate/region.Between2020and2030,weassumea1%peryearofrealincreaseinthevariablecosts,whichishalfthehistoricalgrowthrateofCoalIndiaLimited’sactualcoalprices.FigureA1showsthesupplycurveofthecoalfleet(atindividualunitlevel)forFY2020.Eachpointonthechartrepresentsathermalpowerplantunitinthecountry;thehorizontalaxisshowscumulativetotalinstalledcapacityofthefleetinMWwhiletheverticalaxisshowsthevariablecostinRs/kWh.54FigureA1:SupplyCurveoftheExistingCoalCapacityinFY2020ItisinterestingtonotethaninFY2020,nearly90GWofthecoalcapacityhadavariablecostofhigherthanRs2.76/kWh,theaveragesolarreverseauctionpriceincludingthesafeguardduty.Fornewcoalpowerplants,weassumeapitheadcoalpriceofRs2000-2500/ton(incltaxes),whichisequivalenttoavariablecostofRs1.59/kWh,increasingat1%peryear(halfthehistoricalgrowthrateofCoalIndiaLimited’sactualcoalprices)between2020and2030.ImportedcoalpricesaretakenfromglobalmarketreportsattheIndonesianhub.Averagedeliveredpriceimportedcoalisassessedtobe$70/toninFY2020increasingat1%peryear,whichisequivalenttoavariablecostofRs3.5/kWhforcoastalpowerplants,afteraccountingfortheimprovementinheatratesduetoimportedcoal.8.4GaspricesandsupplyconstraintsWeassumethatthetotaldomesticgasavailabilityforpowersectorwillremainconstrainedatthe2020levels(8.4bcm/yror23mmscmd).TotalLNGimportcapabilitywouldincreasefrom15milliontonsperannum(MTPA)in2020to50MTPAin2030.Domesticgaspricein2030isassumedtoremainalmostthesameas2020($4.2/mmbtu).LNGpricein2020isassumedtobe$3.5/mmbtu(FOB)or$4.5/mmbtu(landed).For2030,weexaminetwoLNGpricescenarios:1)BaseLNGprice:landedpriceof$5.5/MMBTU(plusregasificationcostof$0.6/mmbtuandpipelinecharges,asapplicable),and2)LowLNGprice:landedpriceof$4.5/MMBTU(plusregasificationcostof$0.6/mmbtuandpipelinecharges,asapplicable).558.5OperationalparametersOperationalparameterssuchasramprates,technicalminimumlevels,auxiliaryconsumption,minimumupanddowntimesetchavebeentakenfromtheactualdatafromCEAthermalperformancereviewreports,regulatorynorms,andexpert/industryconsultations.TheyaresummarizedinTableA3.TableA3:AssumptionsonOperationalParametersofPowerPlantsCoal(new)Coal(existing)GasCCGTGasCTHydroNuclearBiomassWindSolarBatteryPumpedHydroPlannedOutagerate5%8-12%5%5%5%10%10%(Availabilityisseasonal)1%1%1%5%ForcedOutagerate5%7-8%5%5%5%10%10%1%1%1%5%TechnicalMinimumLevel%55%Central&IPP=55%40%20%0%90%70%0%0%0%0%State=70%Cold-starttime(hours)2424121#N/A9624000#N/AMinimumup-time(hours)121261096120000Minimumdown-time(hours)663109660000Cold-startCost($/MW)100100301#N/A#N/A#N/A#N/A#N/A#N/A#N/ARamping(%ofinstalledcapacityperminute)1%0.50%2%10%100%#N/A1%#N/A#N/A100%100%AuxiliaryConsumption7%7-8%5%2%1%10%10%0.5%0.5%0.5%1%RoundtripEfficiency#N/A#N/A#N/A#N/A#N/A#N/A#N/A#N/A#N/A90%80%HeatRatekCal/kWh22622214-2819(actual)18102857#N/A#N/A3000#N/A#N/A#N/A#N/A8.6HeatratesWeuseactualheatratedataforeverypowerplantusingseveralsourcessuchasregulatoryfilings,CEAThermalperformancereview,CEACO2EmissionsBaselineetc.Wemodeltheheatrateasafunctionofgeneratorloading,meaningthatasthepowergenerationfromaunitdrops,theheatratewillincrease.TheheatratefunctionistakenfromtheCERCregulationsoncompensatingthegeneratorsforpartialloadoperations.FigureA2showstheheatratefunctionusedforanew660MWsuper-criticalpowerplant.56FigureA2:Averageheatrateofacoalunit(660MWsuper-critical)asafunctionofunitloadingAttechnicalminimumlevelof55%,theheatrateincreasesbyover4%ofthedesignheatrateatratedcapacity.8.7AgriculturalandindustrialdemandresponseSeveralstateshaveseparateddistributionfeedersforagriculturalconsumersfromotherfeeders,andsomestates(e.g.,Karnataka,Maharashtra,andGujarat)havealreadyshiftedamajorpartoftheagriculturalloadtosolarhours(over6GWtotalin2020).Forexample,Karnataka,whichwastraditionallyaneveningpeakingsystem,hasitspeakloadintheafternoonmainlyduetoshiftingofagriculturalloadtosolarday-timehours,asseeninthechartsinFigureA3.57FigureA3:ActualaveragehourlyloadmetinKarnatakainselectedmonthsfor10yearsDatasource:KPTCL(2020)Weassumethesametrendtocontinueinthefuture,andby2030,about50GWofagriculturalloadand10GWofindustrialloadcouldbeshiftedfromnight-timetosolarhours.589AppendixII:AdditionalResults9.1SeasonalnatureofrenewableenergygenerationinIndiaAsshowninFigureA4,windgenerationishighlyseasonalinIndiawithmajorityofthegenerationoccurringinmonsoon(June–September).Solargenerationmoreuniformlydistributedacrosstheyear,withsomedropinwinter;albeitelectricitydemandalsoreducesinwinter.Thesystemload(nationalaggregate)peaksinSeptemberandOctober.Asaresult,thenetload(loadminusrenewableenergygeneration)peaksinOctoberaswindgenerationstartsdroppingrapidlybeyondSeptember.FigureA4:Dailyloadandnetloadenergy(left)anddailyloadandREgeneration(right)inFY2030inthePrimaryLeastCostcase9.2ImpactofincreasingREpenetrationonsystemrampingrequirementsAsREcapacityincreases,thesystemrampingrequirementincreasessignificantlyasshowninFigureA5.59FigureA5:Hourlyload-onlyrampsandnet-loadrampsforthePrimaryLeastCostcaseinFY20309.3WhereisnewREcapacityandstoragecapacitybuilt?India’swindenergypotentialishighlygeographicallyconcentratedwhilesolarenergypotentialismoreuniformlyspreadout.Top-10states,mostlyinthewesternandsouthernregion(TableA4),wouldhaveover90%oftheinstalledREcapacityby2030;specificsiteswheresolarandwindcapacitiescouldbesitedareshowninFigureA6.Batterystorageinstallationsarefoundtobeoptimalinstateswithsignificantsolarcapacitiesandwithlimitedotherflexibilityoptionssuchasagriculturalloadshifting,hydro,andgas.Becauseofthehighlyseasonalnatureofthewindgeneration.60FigureA6:SitesforinstallationofsolarandwindplantsinPrimaryLeastCostCase,2030TableA4:Statelevelwind,solar,andbatterystorageinstalledcapacities(GW)inthePrimaryLeastCostcaseState2020(Actual)2030SolarWindBatteryStorageSolarWindBatteryStorage(x4-hr)Rajasthan5.24.3-57.46.320.1Gujarat3.17.6-35.822.34.3Maharashtra1.95.0-30.235.20.3Karnataka7.34.8-24.734.94.1TamilNadu3.99.3-17.730.83.9AndhraPradesh3.64.1-37.66.78.3MadhyaPradesh2.32.5-31.74.19.6Telangana3.70.1-20.41.13.8UttarPradesh1.2--16.7-4.8Uttarakhand0.3--10.1-1.9OtherStates2.70.1-24.60.72.6All-India35.137.8-306.9142.263.4619.4Whichtransmissioncorridorsneedstrengthening?TableA5:NewTransferCapacityBuildoutthroughFY2030inthePrimaryLeastCostCaseInterfaceAdditionalTransferCapacityBuildoutGW(2021-2030)Maharashtra_Chattisgarh15.7UttarPradesh_MadhyaPradesh8.7MadhyaPradesh_Gujarat8.3Rajasthan_Haryana8.3Bihar_UttarPradesh7.6Maharashtra_Karnataka7.1Punjab_UttarPradesh6.1Rajasthan_Punjab5.6UttarPradesh_Delhi5.1TamilNadu_Karnataka4.9WestBengal_Bihar4.4WestBengal_Odisha4.2WestBengal_Jharkhand4.1Karnataka_Telangana3.7Kerala_Karnataka3.4Odisha_AndhraPradesh3.3Haryana_HimachalPradesh2.9Bihar_Odisha2.4Uttarakhand_Haryana2.4AndhraPradesh_TamilNadu2.3WestBengal_Bhutan2.1HimachalPradesh_Punjab2.0AndhraPradesh_Telangana1.9Maharashtra_Telangana1.9Haryana_Delhi1.5Punjab_JammuandKashmir1.4Uttarakhand_UttarPradesh1.3AndhraPradesh_Karnataka1.2Uttarakhand_JammuandKashmir1.2Maharashtra_MadhyaPradesh1.1Jharkhand_Bihar1.1Odisha_Bihar1.0Otherinterfaces11.7Total139.9629.5HourlycoalandgasgenerationMaximumcoalgenerationofabout180GW(ex-bus)occursduringthenetloadpeakseasonofOctober(FigureA7).FigureA7:Hourlycoalgeneration(nationalaggregate)inthePrimaryLeastCostCase(FY2030)MostofthegasgenerationoccursinthelowREseason(betweenOctoberandFebruary)providingthemuchneededseasonalbalancingtothegrid(FigureA8).FigureA8:Hourlygasgeneration(nationalaggregate)inthePrimaryLeastCostCase(FY2030)639.6HowdoeshydropowercontributetosystemflexibilitySincethereservoirbasedfullydispatchablehydroelectricprojectsareonly~25GW,thediurnalrampingsupportprovidedbythehydrocapacityislimitedto~20-25GW/hour,asshownintheFigureA9.Dispatchablehydropowerplantsmainlygenerateduringthemorningandeveningpeakdemandperiods.FigureA9:Averagehourlyhydrogeneration(includingsmallhydro)inFY2030inthePrimaryLeastCostcase649.7StatelevelgenerationandannualtransmissionflowsFigureA10:Statelevelgenerationandtransmissionflows(TWh/yr)inFY2030inthePrimaryLeastCostcaseNotes:1.AllnumbersinTWh/yr.2.Theindividualstatepiechartsdonotincludeimports/allocationsfromthecentralsectorgeneratingstations.9.8NationaldispatchduringmaximumREgenerationweekInFY2030,theinstantaneousmaximumcontributionofREinmeetingtheloadisashighas73%oftheinstantaneousload,typicallyoccurringat12or1PMinlate-Junewhenwindgenerationhaspickedupandsolargenerationinthewestandnorthhasnotdroppedmuch(FigureA11).InRErichstatessuchasKarnatakaorTamilNadu,theinstantaneousmaximumREgenerationinFY2030couldbeashighas90%.65FigureA11:InstantaneousREgenerationas%oftheinstantaneousloadinthePrimaryLeastCostcase(FY2030)FigureA12showsnationaldispatchduringtheweekofthemaximumdailyREgenerationinFY2030,whichoccursinearlymonsoonwhensolargenerationhasnotdroppedsignificantly.Thereissmallcurtailmentobservedduringthisperiod,butitismostlyconcentratedinthewindrichstatessuchasKarnatakaorTamilNadu.66FigureA12:NationalhourlydispatchduringthemaximumREintegratedweekinthePrimaryLeastCostcase(FY2030)9.9NationaldispatchduringminimumloadweekInwinter,thewindgenerationdropssignificantlybutelectricitydemandalsodrops.Withthehelpofotherresources,thedemandisreliablymetineachhourbutthemannerinwhichenergystorageoperatesisdifferentfromotherseasons.Duringlowdemandperiods,energystoragechargestwiceineachday–majorityofthechargingstillhappensduringthesolarhoursbutsmallamountofchargingalsohappensduringthedemandtroughsintheearlymorninghours(1-5AM).FigureA13showsthenationaldispatchduringminimumloadweek.67FigureA13:NationalhourlydispatchduringtheminimumloadweekinthePrimaryLeastCostcase(FY2030)689.10DispatchinKeyRERichStatesThefollowingchartsshowaveragehourlydispatchinkeyRErichstatesinFY2030forthePrimaryLeastCostcase.Thedifferencebetweenloadandtotalgenerationaretheimports/exportsfromthestate(includingcentralsectorallocationsandbilateral/markettransactions).FigureA14:Averagehourlydispatchinkeywestern&northernRErichstatesinPrimaryLeastCost(FY2030)69FigureA15:AveragehourlydispatchinkeysouthernRErichstatesinthePrimaryLeastCostcase(FY2030)709.11Whatarethetransmissionflowsonkeycorridors?FigureA16:HourlytransmissionflowsonkeycorridorsinthePrimaryLeastCostScenario(FY2030)7110AppendixIII:ComparativeEconomicsofPumpedHydroandBatteryStoragePumpedhydrostoragesystemandbatterystoragesystemsareconsideredtobethetwomajoralternativesfordiurnalstorageinIndia.Inthisappendix,wepresentcomparativeeconomicsofpumpedhydroandbatterystoragesystemsinIndia.Ourkeyassumptionsaregiveninthefollowingtables:TableA6:Capitalcost,life,constructiontime,andlandrequirementofpumpedhydrosystemsAdditionofpumpsystemtoexistinghydroreservoirsNewpumpedhydroplantCapitalCost(RsCr/MW)7.512EconomicLife25years25yearsTechnicalLife50years50yearsConstructiontime3-4years8-10yearsLandrequirement2-5acres/MW(Assuming300mnethead)Giventhematurityofthepumpedhydrotechnology,itscapitalcostisassumedtostaythesameoveryears.Capitalcostsofthebatterystoragesystems,however,isexpectedtoreducerapidlyinthefuture,asshowninthefollowingtable.TableA7:Capitalcost,life,andconstructiontimesofbatterystoragesystemCo-locatedbatterystoragesystemStandalonebatterystoragesystem202020252030202020252030Batterypackprice($/kWh)14388621438862BOS,EPCandothercosts($/kW)176136124240184164TechnicalandEconomicLifePacklife=10years(3000cycles)BOSlife=25yearsPacklife=10years(3000cycles)BOSlife=25yearsConstructiontime6months6monthsNote:Allcostnumbersareexpressedas2020real,unlessspecifiedotherwise.Datasource:Deorahetal(2020)Asshownintheprevioussections,wefind4-6hoursofdiurnalstoragetobecost-effectiveinIndiathrough2030.Thefollowingtableshowsthelevelizedcostofstorageforpumpedhydroandbatterystoragefora4-hoursofdiurnalstorage(1MW/4MWhsystem).72TableA8:Levelizedcostofstoragefor1MW/4MWhpumpedhydroandbatterystoragesystemsinIndia(Rs/kWh,2020real)202020252030PumpedHydro(Additiontoexistingreservoirs)6.186.186.18PumpedHydro(newproject)8.248.248.24Batterystorage(co-locatedwithRE)4.573.132.49Batterystorage(standalone)4.933.402.71Note:Allcostnumbersareexpressedas2020real,unlessspecifiedotherwise.Othereconomicassumptions:WACC=8%(real);Fullcharge/dischargecycles=300cycles/year.Thelevelizedcostofstoragedependsonthehoursofstorage(i.e.MWhtoMWratio).Thefollowingchartshowsthelevelizedcostofstoragein2025and2030forpumpedhydroandbatterystoragesystems.Note:Allcostnumbersareexpressedas2020real,unlessspecifiedotherwise.FigureA17:Levelizedcostofstorageforpumpedhydroandbatterystoragesystemsin2025(left)and2030(right)asafunctionofhoursofstorageItcanbeseenfromFigureA17thatby2025,co-locatedbatterystoragesystemsarecheaperupto10hoursofdiurnalstorage,whencomparedwithaddingpumpedsystemstoexistinghydroreservoirs.Ashoursofstorageincrease,pumpedhydrobecomesmoreeconomical.However,asbatterycostscontinuetoreduceinthefuture,by2030,batteriesarefoundtobecheaperthanpumpedhydro,irrespectiveofthehoursofstorage.Asmentionedpreviously,wefindthatby2030,4-6hoursofdiurnalenergystorageisfoundtobecost-effectiveinIndia,implyingthatbatteriesareamorecost-effectivestorageoptioninIndia.