论文部分内容阅读
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。
As the scale of chemical process system increases, centralized predictive control has encountered computational difficulties in modeling and optimization of controllers. Therefore, distributed predictive control based on system decomposition and network control has been widely studied. In view of the former distributed predictive control algorithm considers only some of the inter-subsystem correlation, the complex correlation between subsystems can not be fully expressed. First, considering the correlation between all subsystems, a distributed predictive control algorithm based on related subsystems is proposed. The structure of distributed predictive control algorithm based on related subsystems is given, the predictive output expression of the subsystem is deduced, and the quadratic programming problem of constrained distributed predictive control based on global coordination is constructed. The stability conditions of the algorithm are analyzed. Finally, the simulation experiment is carried out using the TE process. The simulation results of the distributed predictive control and the distributed predictive control verify the effectiveness and feasibility of the proposed algorithm.