基于核极限学习机风光容量配置研究

《基于核极限学习机风光容量配置研究》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、基于核极限学习机的风光容量配置研究*徐永海，涂菁菁，尹忠东（华北电力大学，北京102206）摘要：风光发电的大量接入，电动汽车充电需求的持续增长，将引起配电网规划与运行特征的根本性改变，因而，研究风光发电以及电动汽车充电站容量配置问题对配电网的稳定与经济运行具有重要意义。本文通过利用网络节点电压、风光电源出力等数据对核极限学习机进行训练学习，构建了基于核极限学习机的容量选择模型，并利用均方根误差对模型精度进行评估。采用IEEE33节点系统作为算例进行仿真，给出满足总投资成本和网络损耗最小的容量配置结果，引入电压稳定性评价指标对结果进行评估，并与支持向量机、遗传算法和粒子群算法获得的结果进行

2、对比分析，验证了所提模型和方法的可行性和有效性。关键词：分布式电源；核极限学习机；均方根误差；容量；总投资成本；有功损耗中图分类号：TM933文献标识码：B文章编号：1001-1390（2019）09-0000-00ThecapacityselectionofwindphotovoltaicpowergenerationsBasedbasedonaKELMmethodforthecapacityselectionofwindphotovoltaicpowergenerationsXuYonghai，,TuJingjing，,YinZhongdong(NorthChinaElectricPo

3、werUniversity,Beijing102206,China)Abstract：Thelarge-scaleaccesstowindpowergenerationandthecontinuousincreaseinthedemandforelectricvehiclechargingwillcausefundamentalchangesintheplanningandoperationcharacteristicsofthedistributionnetwork.Therefore,studyingthewindandphotovoltaicpowergenerationandthe

4、capacityallocationofelectricvehiclechargingstationsisofgreatsignificancetothestabilityandeconomicoperationofthedistributionnetwork.Inthispaper,thekernelextremelearningmachine(KELM)istrainingbyusingdatasuchasnetworknodevoltageandwindandphotovoltaicpoweroutput,andacapacityselectionmodelbasedonkernel

5、extremelearningmachineisconstructed,andthemodelaccuracyisevaluatedbytherootmeansquareerror(RMSE).AcapacityallocationschemeisGiveoutobtainedwhichacapacityallocationschemethatsatisfiesthetotalinvestmentcostandnetworkloss;themodelaccuracyisevaluatedbytherootmeansquareerror.TheIEEE33IEEE33-nodesystemi

6、susedasanexampletosimulate,andthevoltagestabilityevaluationindexisintroducedtoevaluatetheresult.Theresultsobtainedbysupportvectormachine,geneticalgorithmandparticleswarmoptimizationalgorithmarecomparedandanalyzed,andthefeasibilityandeffectivenessoftheproposedmodelandmethodareverified.Keywords：Dist

7、ributeddistributedGenerationgeneration(DG),kernelextremelearningmachine(KELM),rootmeansquareerror(rmseRMSE),Capacitycapacity,totalinvestmentcost,activepowerloss-1-0　引言*基金项目：2016年国家重点研发计划项目（2016YFB0101900）近年来，分布式电