非参数模型控制在液位控制系统中的应用研究  

Study on nonparametric model learning adaptive control in liquid-level control system

作　　者：曹荣敏[1] 关静丽[1] 张维[1] 

机构地区：[1]北京机械工业学院计算机及自动化系,北京100085

出　　处：《计算机工程与设计》2008年第15期4093-4096,共4页Computer Engineering and Design

摘　　要：针对工业控制过程中液位系统的时变和明显的滞后特征,研究了非参数模型控制方法在液位控制系统中的设计方案,讨论了控制算法中引入的伪偏导数的在线估计问题,实现了通过液位系统的输入输出信息并利用递归最小二乘法对伪偏导数进行在线估计的过程,仿真实验验证了非参数模型算法对液位控制的鲁棒性、快速性及抗干扰性,通过仿真比较,展示了该算法性能优于PID算法和模糊控制的结果。Basis for time-varying and time delay characters of a liquid-level control system in industry control process, the design of nonparametric model control algorithm in a liquid-level control system is studied, estimation of pseudo-partial-derivatives is discussed, controller is based directly on pseudo-partial-derivatlves derived on-line from the input and output information of a liquld-level model using recursive least squares type of identification algorithms. The simulation control results show that the algorithms exhibit such advantages as good robustness, quick response, against exogenous disturbance and noise for time-varying systems with vaguely known dynamics, the performance outperforms traditional PID controller and fuzzy controller.

关 键 词：非参数模型学习自适应控制 液位控制系统 时滞 稳定性 鲁棒性 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]