基于关联子系统的分布式预测控制算法  

Distributed Predictive Control based on Associated Subsystems

作　　者：董婷婷[1] 李丽娟[1] 熊路[1] 徐欧官[2] 

机构地区：[1]南京工业大学电气工程与控制科学学院,南京211800 [2]浙江工业大学之江学院,杭州310024

出　　处：《控制工程》2015年第6期1201-1206,共6页Control Engineering of China

基　　金：国家自然科学基金(61203072;61203133);江苏省六大人才高峰项目;工业控制技术国家重点实验室开放课题(ICT1234)

摘　　要：随着化工过程系统规模的增大,集中式预测控制在建模和控制器优化求解方面遇到了计算上的困难,因而基于系统分解和网络控制的分布式预测控制得到了广泛研究。鉴于前人分布式预测控制算法只考虑了部分子系统间的相互关联,对子系统间的复杂关联不能全面表达。首先,考虑全部子系统间的相互关联,提出基于关联子系统的分布式预测控制算法。给出基于关联子系统的分布式预测控制算法结构,推导子系统的预测输出表达式,构建基于全局协调的有约束分布式预测控制二次规划问题,并分析算法的稳定性条件。最后,采用TE过程进行仿真实验,分布式预测控制与分散式预测控制的仿真对比结果验证该算法的有效性和可行性。With the increase of chemical process scale, centralized predictive control had difficulty in solving optimal solution in the modeling and controllers. Hence system decomposition and network control based distributed predictive control has been widely studied. Given the distributed predictive control algorithm, predecessors only considered partially interrelation between subsystems, which can not completely express the complex relation between subsystems. We consider all interrelation between subsystems, proposed distributed predictive control algorithm based on interrelated subsystems. It shows the algorithm structure of distributed predictive control based on interrelated subsystems, derives subsystem predicted output expression, constructs quadratic programming problems based on global coordination of constrained distributed predictive control, and analyzes the stability conditions for the algorithm. Finally, TE process simulation, comparison of distributed predictive control and decentralized predictive control simulation results verifies the feasibility and effectiveness of the algorithm.

关 键 词：分布式预测控制 关联子系统 TE过程 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]