一类串联生产过程的分布式解耦预测控制  

Distributed decoupling predictive control of a kind of cascade processes

作　　者：陈庆[1] 李少远[1] 席裕庚[1] 

机构地区：[1]上海交通大学自动化研究所,上海200030

出　　处：《控制与决策》2004年第6期647-650,654,共5页Control and Decision

基　　金：国家自然科学基金重点项目(69934020);高等学校博士点专项科研基金项目(20020248028).

摘　　要：研究一类由多个子过程串联而成的生产过程的预测控制,提出了在分布式通信模式下的预测控制算法.该算法由两部分组成:一部分利用其他子过程在上一时刻的控制策略,对子过程间的弱耦合进行前馈补偿;另一部分通过顺序求解对强耦合进行解耦控制.在线性无约束情况下,得到了分布式解耦动态矩阵控制律.最后通过仿真验证了算法的有效性.The predictive control of a kind of cascade processes, which is made up of several sub-processes, is studied. A decoupling predictive control algorithm in distributed communication mode is proposed. The algorithm is composed of two parts. One is forward compensation for the weak interaction using the control strategy of the previous time. The other is decoupling control for the strong interaction by sequential solving control action for each sub-process. Distributed decoupling dynamic matrix control strategy is obtained for linear unconstrained systems. Simulation results show the efficiency of the algorithm.

关 键 词：多变量预测控制 分布式模型预测控制 解耦控制 串联过程 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]