一种抗强干扰自适应PID控制算法.pdf

《一种抗强干扰自适应PID控制算法.pdf》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、中文摘要自适应PID控制算法是一种将经典PID控制与现代自适应控制思想相结合的先进控制算法，具有设计简单、鲁棒性强、快速适应环境变化的特点。然而在现代工业过程复杂工况环境中，经常会存在近似阶跃变化的强干扰信号，从而造成系统的参数辨识结果与实际情况完全背离，使自适应PID控制系统的适应性变差甚至难以运行。为此本文针对一种采用最小二乘辨识对象结构参数的算法，分析辨识失效的原因是由闭环条件下激励的不充分引起的，特别是强干扰产生时辨识参数严重失真的情况，提出了在系统运行初期利用测试信号的充分激励性，检测出系统参数的有效范围；当检测到强干扰信号时，暂停对PID参数进行更新，同时采用协方差复位以及加入激励

2、信号的方式，使系统辨识尽快收敛到真实参数以克服强干扰信号的影响，进而推导出了一种抗强干扰自适应PID控制算法。最后通过对存在强干扰信号条件下的系统仿真，从而验证了这种算法的有效性和优越性。关键字：自适应PID算法，参数辨识，抗强干扰黑龙江大学硕士学位论文AbstractAdaptivePIDControlAlgorithmisalladvancedcontrolalgorithmthatcombiningclassicalPIDcontrolalgorithmwithmodemadaptivecontrolthought．Thealgorithmnotonlyhastheadvantageso

3、fsimpledesignandstrongrobustness，butalsohastheabilityofadaptingtotheenvironmentchanges．Howeverstronginterferencesignalswhichcallbeapproximateasstepchangesoftenappearinthecomplexconditionsofmodemindustrialprocessenvironment．Thestronginterferencemayleadtothecompletedeviationofthesystemparameteridentif

4、icationresultandtheactualsituation，thusmaketheadaptabilityoftheadaptivePIDcontrolsystemworse，orevencollapse．Therefore，accordingtotheresearchonanalgorithmoftheleastsquarestructureparameters,theauthorrealizedthatidentificationinvalidationofthesystemiscausedbytheexcitationinadequacyinclosedloop，especia

5、llytheseriousdistortionsituationofidentificationparameter．Onthebasisoftheprecedingresearch，，thispaperputsforwardsaseriesofmeasurestoovercometheviolentinfluenceofstronginterferenceaboutparameterofadaptivePIDcontrol：Firstly,wewilldetecttheeffectiverangeofsystemparametersbyusingthepersistentexcitationf

6、eatureoftestsignalsattheinitialoperatingstageofthesystem．Secondly,assoonaswedetectthestronginterference，wewillmaketheprocessofrenewalthePIDparameterspaused,meanwhile，resetthecovariancematrixoftheestimatortoalargematrixandaddexcitingsignalsinordertomakethesystemidentificationconvergetotherealparamete

7、rsassoonaspossible．Finally，weobtainanadaptivePIDcontrolalgorithmwithresistingstronginterferenceaftersunluptheabovesteps,andsubsequentlytheeffectivenessandsuperiorityofthealgorithmaleprovedthoughsimula