一种预测辨识周期性干扰校正的控制算法  

An Online Identification Control Algorithm Based on Prediction for Periodic Disturbance Correction

作　　者：陈良[1] 任正云[1] 石建平 

机构地区：[1]东华大学自动化系,上海201620 [2]丹阳市农业资源开发局,江苏丹阳212300

出　　处：《控制工程》2010年第S3期6-9,共4页Control Engineering of China

基　　金：国家自然基金资助项目(60674088);中央高校基本科研业务费专项基金资助

摘　　要：针对过程控制中广泛存在的周期性干扰问题,提出了一种抑制周期性干扰的在线频率辨识控制方法。在被控对象的输入端施加阶跃输入信号,首先通过预测控制,提高了系统的动态性能。在加入周期性干扰的情况下,根据频率辨识环节在线地辨识周期性干扰的频率,将数据传入重复控制器,动态地修正重复控制器参数,确保了对周期性干扰的抑制作用。此过程随着周期性干扰频率的改变而循环往复,不断地修改控制器相关参数,以随时抵消不断改变的干扰,使系统永远处于稳定状态。此算法克服了传统的控制器难以对周期性干扰进行有效的实时跟踪控制,最终使系统的品质逐渐变坏的弱点。仿真结果说明,经过早期振荡,系统能够准确地辨识出周期性干扰的频率大小,整个过程的系统输出经过一段时间的振荡后最终能够达到稳定状态。A frequency identification online control methods' for periodic disturbance is proposed,because the problem of periodic disturbance is widespread in the process control.A step input is imposed to charged object.By the prediction control,it can improve the dynamic performance of the system.The controller's parameters online can be corrected according to the link of frequency identification in joining the periodic disturbance.It ensures the elimination of the periodic disturbance.So it overcomes the weaknesses of traditional controller that is difficult to control periodic disturbance,ultimately leading to deterioration in the quality of the system.Simulation results show that after the early oscillations,the system can identify the size of the frequency of periodic disturbance,the output of the whole process eventually reaches a steady state over a period of oscillation.

关 键 词：在线辨识 周期性干扰 预测控制 重复控制器 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]