文档名：基于复合神经网络重构对象的永磁同步直线电机变参数型位移速度并行控制
摘要：针对永磁同步直线伺服电机(PMSLM)传统位移控制算法中控制器系数固定带来的控制精度不足等问题,提出一种基于复合神经网络重构对象的PMSLM变参数型位移速度并行控制策略.首先,利用动子位移、线速度的误差信息设计变参数并行控制器;其次,建立含有控制对象多维信息的复合径向基神经网络观测动子位移,并得到控制对象的偏导信息;再次,基于闭环稳定条件,以周期检索的误差与控制目标的比较结果为基础,构建完整的位移速度并行控制器参数更新策略;最后,实验结果表明,该文所提控制策略能实现不同给定位移的高精度控制,且具有控制不同对象参数的泛用性.
Abstract：Apermanentmagnetsynchronouslinearservomotorhasthecharacteristicsofhighthrustdensity,highefficiency,andfastcontrolresponse.Itiswidelyusedinrailtransit,intelligentmanufacturing,andaerospacefields.Theelectricalormechanicalparametersofthemotorandnonlinearresistanceinterferencewillaffectthehighprecisioncontrolofthelinearservo.Therefore,thestableandefficientdisplacementcontrolalgorithmhasimportantengineeringapplicationvaluetoimprovethesystemperformance.Servocontrolcanbedividedintoserialandparallelsystemsaccordingtotherelativepositionofdisplacementandvelocityloop.However,whetherserialorparallelcontrol,mostcontrollerparametersneedtobesetinadvanceandaregenerallyfixed,andadjustingtheactualcontroleffectisimpossible.Thispaperproposesaparallelcontrolstrategyofvariableparameterdisplacementvelocityofpermanentmagnetsynchronouslinearmotorbasedoncompositeneuralnetworkreconstructionobject,whichrealizesthehigh-performancecontroloflinearservomotor.Firstly,aparallelcontrollerwithvariableparametersisdesignedusingtheerrorinformationofthemovingdisplacementandlinearvelocity.Then,thedisplacementoutputofthepermanentmagnetsynchronouslinearmotorisreconstructedbyacompositeradialbasisneuralnetworkcontainingmulti-dimensionalinformationofthecontrolobject,andthepartialderivativeofthedisplacementtothecontrolsignalisobtained.Finally,basedontheclosed-loopstabilityconditionofthesystem,acompleteparameterupdatestrategyofparallelcontrollerwithdisplacementvelocityisprovidedaccordingtothecomparisonbetweenperiodicretrievalerrorsandcontroltargets.Intheexperiment,apermanentmagnetsynchronouslinearservosystemisdesigned,anddifferentnetworkobservations,positioncontrols,andobjectparametercontrolsarecompared.Thenetworkcomparisonshowsthatthecompositeradial-basedneuralnetworkhasfastobservationspeedandhighobservationaccuracy,andtheobservationtimeandrootmeansquareerrorare18.63μsand0.063mm,respectively.Thepositioncontrolcomparisonshowsthatthevariableparameterparallelalgorithmhashighercontrolprecisionthanthefixedparameteralgorithm,suchasPIDcontrol,witha30%～50%improvementeffectinthedisplacementandvelocitycontrolerrorindexes.Theparametercontrolcomparisonshowsthatinthevariableparameterparallelcontrolstrategy,whenthemovingmasschanges,therootmeansquareofdisplacementerrorislessthan0.015mm,andtheabsolutemaximumerrorislessthan0.060mm,whichmeansthatthealgorithmhasstabilityanduniversality.Thefollowingconclusionscanbedrawn:(1)ThecompositeRBFneuralnetworkstructurehasbetterobservationperformancethanthetraditionalRBFneuralnetwork.Comparedwiththetraditionalimprovementmethod,ithasabettereffectwithlesscomputingpressureontheprocessor.(2)Comparedwiththefixedcoefficientalgorithm,suchasPID,theproposedvariableparameterparallelcontrolstrategycaneffectivelyimprovethecontrolperformanceoftheactuatordisplacement.(3)Thevariableparameterparallelcontrolstrategycanguaranteethehighprecisioncontroleffectunderdifferentdisplacementsettingsandobjectparameters.
作者：鲍明堃 周扬忠Author：BaoMingkun ZhouYangzhong
作者单位：福州大学福建省新能源发电与电能变换重点实验室福州350116
刊名：电工技术学报
Journal：TransactionsofChinaElectrotechnicalSociety
年，卷(期)：2024, 39(8)
分类号：TM359.4
关键词：永磁同步直线电机 并行控制 复合径向基神经网络 变参数 更新机制
Keywords：Permanentmagnetsynchronouslinearmotor parallelcontrol compositeradialbasisfunctionneuralnetwork variableparameters updatemechanism
机标分类号：TM359.4TP273.3TP391.9
在线出版日期：2024年4月29日
基金项目：福建省自然科学基金资助项目基于复合神经网络重构对象的永磁同步直线电机变参数型位移速度并行控制[
期刊论文] 电工技术学报--2024, 39(8)鲍明堃 周扬忠针对永磁同步直线伺服电机(PMSLM)传统位移控制算法中控制器系数固定带来的控制精度不足等问题,提出一种基于复合神经网络重构对象的PMSLM变参数型位移速度并行控制策略.首先,利用动子位移、线速度的误差信息设计变参数...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
基于复合神经网络重构对象的永磁同步直线电机变参数型位移速度并行控制 Parallel Displacement Velocity Control of Permanent Magnet Synchronous Linear Motor with Variable Parameters Based on Composite Neural Network Reconstruction Object
基于复合神经网络重构对象的永磁同步直线电机变参数型位移速度并行控制.pdf
- 文件大小:
- 1.72 MB
- 下载次数:
- 60
-
高速下载
|
