第47卷第4期电网技术Vol.47No.42023年4月PowerSystemTechnologyApr.2023文章编号：1000-3673（2023）04-1331-09中图分类号：TM721文献标志码：A学科代码：470·40计及深度信念网络场景生成的风光储协同优化规划方法史昭娣，黄越辉，李湃，王伟胜（新能源与储能运行控制国家重点实验室(中国电力科学研究院有限公司)，北京市海淀区100192）CollaborativeOptimizationPlanningforWind/PV/storageBasedonScenarioGeneratedbyDeepBeliefNetworkSHIZhaodi,HUANGYuehui,LIPai,WANGWeisheng(StateKeyLaboratoryofOperationandControlofRenewableEnergy&StorageSystems(ChinaElectricPowerResearchInstitute),HaidianDistrict,Beijing100192,China)1ABSTRACT:Asrenewableenergyoutputhasstochasticandfluctuatingcharacteristics,itisofgreatsignificancetohavetherenewableenergyoutputstochasticscenariosinsolvingtheproblemsasrenewablegenerationexpansionandenergystorageplanning.However,theexistingmethodsofscenariogenerationbasedonthestatisticalmodelshardlyconsiderthestochasticandfluctuatingcharacteristicsofrenewableenergyoutput.Basedonthis,thispaperfirstproposesatimesequencescenariogenerationbasedonthedeepbeliefnetworktofullyexplorethecharacteristicsofrenewableenergyoutput.Secondly,comprehensivelyconsideringthesystemeconomyandtheenvironmentalprotectioncosts,aoptimizationplanningmodelforwind/photovoltaic(PV)/storagebasedonthemid-to-long-termtimingpowerbalancesimulationisestablished.Inordertofullytakeintoaccountthestochasticsoutput,thispaperadoptsseveralannualwind/PVoutputsequencesastheinputoftheplanningmodel.Takingtheexpectedresultsofallthescenariosasthefinaloptimizationresult,thereliabilityoftheplanningresultsischecked.Finally,aprovincialpowergridinChinaistakenasanexampletoanalyzethewind/PV/storagecapacityplanningschemesunderdifferentstochasticscenariosandparametersensitivity.Therefore,asolutiontotherenewablegenerationplanningproblemconsideringthestochasticoutputofthewind/PVisprovided.KEYWORDS:renewableenergy;stochasticoutput;generationplanning;deepbeliefnetwork;scenariogeneration基金项目：国家电网有限公司总部管理科技项目(1300-202155460...
