文档摘要：暂态稳定实时性判别是响应式稳定控制的核心,新能源和直流输电并网下电网的结构、运行和响应特征复杂程度骤增,现有实时性判别方法的准确性和泛化性面临挑战.该文针对含新能源和直流输电的复杂电网暂态稳定性,提出了一种基于可量测暂态能量特征的自适应逻辑推理判稳方法,实现无电网模型依赖且具有可解释性的实时性稳定性判别.首先,基于发电机转子运动的能量关系,结合特勒根定理构建了面向复杂电网响应信息的能量函数并论证了其守恒性;然后,根据系统动能-势能能量转换特征定义了稳定预判因数,并结合最大功角差构成了判稳关键特征量,提出了基于自适应模糊推理神经网络(ANFIS)的关键特征量与稳定状态之间的映射模型,实现了暂态稳定性实时推理评估;最后,在简单系统中量化分析了关键特征量与系统稳定性间的关系,并在修改后含新能源与直流输电的IEEE10机39节点系统中验证了该文所提方法的有效性与泛化性.
Abstract：Real-timetransientstabilityassessmentisthecoreofresponse-drivenstabilitycontrol.WithrenewableenergyandDCtransmissionsconnectingtopowersystems,thecomplexityofgrid'sstructure,operationandcharacteristicsincreasessignificantly,whichhavechallengedtheeffectivenessandaccuracyofexistingstabilityassessmentmethods.Forreal-timestabilityassessmentinthecomplexpowersystemincludingrenewableenergyandDCtransmissions,anadaptivenetworkbasedfuzzyinferencesystem(ANFIS)basedmodel-freeandinterpretablestabilityassessmentmethodwasproposedbyusingmeasurabletransientenergyfeatures.First,basedontheenergyrelationofgeneratormotionequationandtheTellegen'stheorem,thetransientenergyfunctionusingtheresponseinformationofthecomplexgridwasconstructedanditsenergyconservationwasalsovalidated.Then,accordingtotheconversioncharacteristicbetweenkineticenergyandpotentialenergy,thestabilitypredictionfactorwasdefined,whichwascombinedwithmaximumangledeviationtoobtainthekeyfeaturesofthestabilityjudgment.AmappingmodelbetweenthekeyfeaturesandstabilitystatusbasedonANFISadaptivelogicreasoningwasproposedtorealizereal-timepredictionoftransientstability.Finally,thequantitativeanalysisoftherelationshipbetweenthekeyfeaturesandsystemstabilitystatuswasillustratedinasimplesystem,andtheeffectivenessandgeneralizationoftheproposedmethodwereverifiedinthemodifiedIEEE10-machine39-bussystemwithnewenergiesandDCtransmissions.Firstly,basedontheenergyrelationofthegeneratorrotormovementandtheTellegen'stheorem,arefinedenergyfunctionmodelwasdevelopedbyusingtheresponseinformationofacomplexgrid,andtheenergyconservationpropertyofthemodelwasproven.Theenergyfunctionmodelcandepicttheenergyresponsecharacteristicsofvariouscomponentsinthesystem.Byintroducingthevirtualimpedancebranch,thepotentialenergyofeachcomponentcanbeconvenientlycalculatedusingthepre-faultequilibriumpointofpowersystem,whichdoesnotdependonthecomponentmodelanditsparameters,showinggoodgenerality.Secondly,theconversioncharacteristicsbetweenkineticenergyandpotentialenergywereanalyzedunderstableandunstableconditions,andthenastabilitypredictionfactor(SPF)weredefinedatthemaximumpotentialenergypoint.Togetherwiththemaximumangledeviation(MAD)atthesamepoint,thecriticalfeaturesSPFandMADforstabilityassessmentwereobtained,whichhavestrongrepresentationalabilitytoindicatethestable/unstablestatusofthesystem.Finally,aresponse-drivenstabilitydiscriminationmodelwasconstructedwiththekeyfeaturesSPFandMAD.ThestabilitydiscriminationmodelwasestablishedbasedonlogicalreasoningANFIS,whichcanobtainaninterpretablemappingmodelbetweenthekeyfeaturesandthestable/unstablestatusofthesystem.Thismodelenablesreal-timetransientstabilityassessmentwithinterpretability.Thispaperquantitativelyanalyzedtherelationshipbetweenthekeyfeatures(SPFandMAD)andsystemstabilitystatusinamodifiedsimpleIEEE9-bussystem,andevaluatedtheeffectivenessandgeneralizationoftheproposedstabilitydiscriminationmodelinamodifiedIEEE10-machine39-bussystemwithnewenergysourcesandDCtransmissions.ThesimulationresultsdemonstratedthattheproposedkeyfeaturesSPFandMAD,showingstrongphysicalattributesassociatedwiththesystemstability,caneffectivelyreflectthesystemstabilitystatus.ComparedtootherAIstabilitydiscriminationmodels,theproposeddata-drivenANFISstabilitydiscriminationmodelhadahigheraccuracyofstabilitydiscriminationinacomplexpowergrid.Thus,itcanbeeffectivelyappliedinsubsequentemergencyresponse-drivenstabilitycontrol.Theconclusionsofthispaperaregivenasfollows:(1)ThekeyfeaturesSPFandMAD,extractedfromtheconstructedenergyfunctionmodel,haveastrongcorrelationwiththesystemstable/unstablestatus.Theycandirectlycharacterizethesystemstabilityconditioncomparedtootherresponsefeatures.(2)Thedata-drivenANFISbasedstabilitydiscriminationmodelestablishestheif-thenlogicalinferenceprocessbetweentheinputfeatures(SPFandMAD)andthesystemstable/unstablestatus.Itcandetecttheunstableconditioninreal-timeandshowhigheraccuracyofstabilitydiscriminationinacomplexpowergridcomparedtootherAIstabilitydiscriminationmethods.(3)Theproposedstabilityanalysismethodmaintainshighaccuracyeveninthepresenceofchangesinthepowergridtopology,showinggoodgeneralizationperformance.
作者：杨浩 伍柏臻 刘铖 孙正龙 蔡国伟 刘萌 Author：YangHao WuBaizhen LiuCheng SunZhenglong CaiGuowei LiuMeng
作者单位：现代电力系统仿真与绿色电能新技术教育部重点实验室(东北电力大学)吉林132012国网山东省电力公司电力科学研究院济南250003
刊名：电工技术学报 ISTICEIPKU
Journal：TransactionsofChinaElectrotechnicalSociety
年，卷(期)：2024, 39(13)
分类号：TM712
关键词：暂态稳定性 响应驱动 特勒根定理 能量函数 复杂电网 自适应逻辑推理
Keywords：Transientstability response-driven Tellegen'stheorem transientenergyfunction complexpowersystem adaptivelogicalreasoning
机标分类号：TP273TM721.1TP391.41
在线出版日期：2024年7月22日
基金项目：国家重点研发计划基于暂态关键特征逻辑推理的复杂电网响应驱动暂态稳定性判别[
期刊论文] 电工技术学报--2024, 39(13)杨浩 伍柏臻 刘铖 孙正龙 蔡国伟 刘萌暂态稳定实时性判别是响应式稳定控制的核心,新能源和直流输电并网下电网的结构、运行和响应特征复杂程度骤增,现有实时性判别方法的准确性和泛化性面临挑战.该文针对含新能源和直流输电的复杂电网暂态稳定性,提出了一种...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
关键词：暂态稳定性,响应驱动,特勒根定理,能量函数,复杂电网,自适应逻辑推理,
- 文件大小:
- 17.37 MB
- 下载次数:
- 60
-
高速下载
|
|
1
