文档名：基于超局部化时间序列的永磁同步电机无模型预测电流滑模控制策略
摘要：在复杂环境、多变负载工况中,由于时变电感参数、磁场耦合、铁心饱和等影响,电机控制精度及可靠性下降.为解决以上问题,该文充分利用时间序列模型反映电机电压、电流等状态变量之间关系,消除传统参数化建模的影响,提出基于超局部化时间序列的无模型预测电流滑模控制(SMC)方法.该方法将时间序列数据驱动模型超局部化,提升传统超局部模型精度,并采用递归最小二乘法(RLS)在线估计和更新模型中全部待定系数,实时精准响应当前系统工作状态.在此基础上,结合超局部时间序列模型,生成滑模控制函数,并设计Lyapunov方法验证函数趋近条件.实验结果表明,与传统控制方法相比,提出方法具有更强的鲁棒性、更优的电流质量和较低的系统噪声.
Abstract：Thelimitedflexibilityandrobustnessofthetypicalslidingmodecontrol(SMC)failtomeetthedemandsofcomplexenvironmentswithvariableloadsandtheinfluenceoftime-varyinginductanceparameters,magneticfieldcoupling,coresaturation,andotherfactors.Onthecontrary,themodel-freeSMCstrategyismoreeffective.Thispaperproposesamodel-freepredictiveSMCstrategyutilizinganultra-localizedtime-seriesmodelforapermanentmagnetsynchronousmotor(PMSM)drivingsystem.Byrepresentingthemotorasacollectionofdiscrete-timelinearfunctionsandmaintaininghighmodelaccuracythroughanonlineestimationalgorithm,theproposedstrategyisbetteralignedwiththemotioncharacteristicsofthemotorsystem.Firstly,thisapproachestablishesanultra-localizedtime-seriesmodelandupdatestheregressivevector,whichonlysummarizesinputandoutputsignalsbasedonsampleddata.Secondly,allundeterminedcoefficientsinthemodelareestimatedusingtherecursiveleastsquare(RLS)algorithm.Consequently,thecurrentoperatingstateofthemotordrivingsystemisdescribedasacollectionofdiscrete-timelinearfunctionsandconvertedintotheultra-localstructuretogeneratetheslidingmodesignal.Finally,controlfunctionsaredesignedbasedonthepower-reachingrule,andthereachingconditionsareverifiedusingtheLyapunovmethod.Thisultra-localizedtime-seriesmodeliseasilyimplementedwithintheSMCstrategy,offeringgoodaccuracyandaddressingtheissuescausedbytime-varyingphysicalparametersintheplantmodelofthetypicalSMCandinputgainoftheconventionalobserver-basedultra-localmodel.SimulationandexperimentalresultsonaPMSMdrivingsystemdemonstratetheeffectivenessoftheproposedmethodinresistingdisturbancesandsuccessfullytrackingthereference.Thedisturbancesprimarilyincludechangedparametermismatchesandloadtorque.Fourieranalysisandaccumulatederrorcomparisonsbetweentheproposedandconventionalmodel-freeSMCstrategiesshowthattheproposedmethodreducesthetotalharmonicdistortion(THD)by3.14%.Comparedtotheconventionalstrategy,theproposedmethodexhibitstheminimumascendingslopeofaccumulatederrorforcurrentandloweroperatingnoiseamplitudesinvariousspeedreferencesandloadtorques.Experimentalresultswithdifferentparametermismatchesofthestatorinductancefurtherverifyitsrobustness.Thevalidationsprovidethefollowingconclusions:(1)Theproposedmethodadoptsanultra-localizedtime-seriesmodeltorepresentthecurrentoperatingstateofthemotordrivingsystem.Unliketheconventionalstrategythatutilizesanultra-localapproach,thismodelformulatesthesystemasacollectionofdiscrete-timelinearfunctions.(2)Theultra-localizedtime-seriesmodelsignificantlyimprovescurrentquality,accumulatedcurrenterror,andsystemnoise.Thisimprovementisattributedtothehighaccuracyoftheultra-localizedtime-seriesmodelandtheeliminatedinfluencesoftheunsuitableinputgain.(3)Byemployingadesignedestimationalgorithmandsampleddata,theultra-localizedtime-seriesmodelreplacesthephysicalmodel,whichinvolvesmultipletime-varyingphysicalparameters,enablingmoreaccuratemodelingandupdatingprocesses.
作者：魏尧 柯栋梁 黄东晓 汪凤翔 张祯滨 Author：WeiYao KeDongliang HuangDongxiao WangFengxiang ZhangZhenbin
作者单位：电机驱动与功率电子国家地方联合研究中心(中国科学院海西研究院泉州装备制造研究中心)晋江362216山东大学电气工程学院济南250061
刊名：电工技术学报
Journal：TransactionsofChinaElectrotechnicalSociety
年，卷(期)：2024, 39(4)
分类号：TM341
关键词：无模型滑模控制 超局部化时间序列模型 递归最小二乘算法 数据驱动模型
Keywords：Model-freeslidingmodecontrol ultra-localizedtimeseriesmodel recursiveleastsquarealgorithm data-drivenmodel
机标分类号：TM351TP273U469.72
在线出版日期：2024年3月5日
基金项目：国家自然科学基金，中国博士后科学基金，福建省科技计划资助项目基于超局部化时间序列的永磁同步电机无模型预测电流滑模控制策略[
期刊论文] 电工技术学报--2024, 39(4)魏尧 柯栋梁 黄东晓 汪凤翔 张祯滨在复杂环境、多变负载工况中,由于时变电感参数、磁场耦合、铁心饱和等影响,电机控制精度及可靠性下降.为解决以上问题,该文充分利用时间序列模型反映电机电压、电流等状态变量之间关系,消除传统参数化建模的影响,提出基于超...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
基于超局部化时间序列的永磁同步电机无模型预测电流滑模控制策略 Model-Free Predictive Sliding Mode Control Using Ultra-Localized Time-Series for Permanent Magnet Synchronous Motor Drives
基于超局部化时间序列的永磁同步电机无模型预测电流滑模控制策略.pdf
- 文件大小:
- 2.6 MB
- 下载次数:
- 60
-
高速下载
|
|
