基于RBFNN的非线性多变量大时滞系统的PID控制  

A non-linear multivariable big delay system designed with PID control based on RBFNN

作　　者：孟国庆 周建芳 

机构地区：[1]奇瑞汽车有限公司,安徽芜湖241009

出　　处：《工业仪表与自动化装置》2009年第1期7-10,共4页Industrial Instrumentation & Automation

摘　　要：针对非线性多变量大时滞系统,研究了一种基于神经网络的智能控制策略。该文给出了网络的结构和算法,对一组二变量强耦合时变系统进行了仿真。仿真结果表明,控制器能根据系统的运行状态获得对应于某种最优控制律下的PID参数,解耦后的系统具有较好的动态和静态性能。该方法具有控制精度高、响应速度快的优点,并具有较强的自适应性和鲁棒性。In view of the characteristic of a non-linear multivariable time-varying system, the paper makes a study of an intelligent control strategy based on a neural network including its structure and relevant algorithm. The paper also shows the result from the real-time sumulation of a double variable and strong-coupled time-varying system. The PID controller can get access to the optimal control parameters according to the system operation state. The decoupling control system has better dynamic and static performance. It has proved that the method has the advantages of high control accuracy, short response, great adaptability and robustness.

关 键 词：RBF神经网络 预测控制 非线性时变系统 PID控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]