伺服电机模型参数在线辨识研究摘要在伺服电机的研究过程中，由于负载变量的变化，可能会出现经典控制器参数不匹配的情况，导致实验结果出现波动与误差，使控制性能恶化。因此，具备参数在线辨识能力在高性能的伺服电机中是必备的。本次论文通过阐述理论和实验测试对伺服电机模型参数在线辨识进行了深入的研究。首先，对参数在线辨识的方法进行简单的阐述，包括：模型参考自适应法、梯度校正参数估计法、最小二乘参数估计法，本文在此基础上选择其衍生方法：带遗忘因子的递推最小二乘法对参数进行在线辨识。研究表明：模型参考自适应法不能兼顾收敛速度和辨识的精度，梯度校正参数估计法算法简单易懂，实时计算量小，但是收敛速度慢，而最小二乘法原理简单，收敛较快，容易理解，易于编程实现，因此在系统参数估计中被大量使用。其次，在建立数学模型的基础上以自校正PID结构设置控制器，自校正PID控制器以常规PID为基本形式，引入带遗忘因子的递推最小二乘法估计对象参数，并将估计的参数按极点配置法进行设计。自校正PID控制是自校正控制思想和常规PID控制思想结合的产物，具有调整参数少、自适应能力强的优点，能随对象特性变化在线修改参数。最后，搭建基于DSP的直流伺服电机实验平台，进行测速算法和参数辨识的实验研究，采用DSP编程，在实际系统上对自校正控制进行试验测试，观察其实验结果与数据。论文研究成果可以在实际直流伺服电机实验平台中得到较好的应用，以提高系统的性能指标。关键词：“伺服电机”“在线辨识”“自校正PID”“DSP编程”ResearchonOnlineIdentificationofServoMotorModelParametersAbstractDuringtheresearchofservomotor,duetothechangeofloadvariables,themismatchofclassicalcontrollerparametersmayoccur,resultinginfluctuationsanderrorsintheexperimentalresults,whichdeterioratesthecontrolperformance.Therefore,ahigh-performanceservomotormodelrequirestheabilitytoparameterizeonlineidentification.Inthisthesis,throughthetheory,simulationandexperimentaltest,theon-lineidentificationofservomotormodelparametersisdeeplystudied.Firstly,themethodofparameteronlineidentificationisbrieflydescribed,including:modelreferenceadaptivemethod,gradientcorrectionparameterestimationmethod,leastsquaresparameterestimationmethod(includingbatchleastsquaresmethod,recursiveleastsquaremethod,forgettingfactordelivery).Basedontheseriesofderivativealgorithmssuchasleastsquaresmethod,thisthesischoosestherecursiveleastsquaresmethodtoidentifytheparametersonline.Theresearchshowsthatthemodelreferenceadaptivemethodcannotbalancetheconvergencespeedandtheaccuracyoftheidentification.Thegradientcorrectionparameterestimationalgorithmissimpleandeasytounderstand.Thereal-timecalculationissmall,buttheconvergencespeedisslow.Theprincipleofleastsquaresissimple,theconvergenceisfast,anditiseasytounderstand.Itiseasytoprogramandisthereforeusedextensivelyinsystemparameterestimation.Secondly,basedontheestablishmentofthemathematicalmodel,theself-correctingPIDstructureisusedtosetthecontrollerfortheservomotor.Thedesignideaoftheself-tuningPIDcontrolleristousetheregularPIDasthebasicform,introducetherecursivealgorithmtoestimatetheobjectparameters,andpresstheestimationresult.Thepoleconfigurationmethodisusedtodesignthecontrollerparameters.Self-tuningPIDcontrolisacombinationofself-correctingcontrolideaandconventionalPIDcontrolidea.Itcombinestheadvantagesofboth,hastheadvantagesoflessadjustmentparametersandstrongself-adaptiveability,andcanmodifyparametersonlinewithchangesinobjectcharacteristics.Finally,aDSP-basedDCservomotorexperimentalplatformisbuilttocarryoutexperimentalresearchonspeedmeasurementalgorithmandparameteridentification.The...
