双层结构预测控制研究进展

Research development of two-layered predictive control

现代工业大系统的优化控制采用递阶结构,其中以预测控制为代表的先进过程控制已经成为重要的一级.目前,主流的工业预测控制技术均采用双层结构,即包含稳态优化层和动态控制层.双层结构预测控制技术可以有效解决复杂工业过程常见的多目标优化、多变量控制的难点问题.本文简要总结了双层结构预测控制的算法,并从控制输入与被控输出稳态关系入手分析了多变量预测控制稳态解的相容性和唯一性,说明了稳态优化的重要性.针对双层结构预测控制与区间预测控制的性能比较、稳态模型的奇异性以及闭环系统动态特性等提出了一些见解,并指出了需要重点研究的主题.

Modem large-scale industrial control, systems adopt hierarchical structures, where the advanced process con-trol, with model predictive control （MPC） as its representative, has been an important level. The two-layered structure, i.e.,the steady state optimization plus and the dynamic control layer, is dominant in the industrial MPC technology. The two-layered MPC can effectively handle the multi-objective optimization and multi-variable control problems in the complexprocesses. The algorithm of two-layered MPC is briefly summarized. The compatibility and uniqueness of the steady statesolutions for the multi-variable MPC are analyzed based on the steady-state relationship between the control inputs and thecontrolled outputs, demonstrating the importance of the steady state optimization. Viewpoints are presented on issues suchas performance comparisons between two-layered MPC and zone MPC, the singularity of the steady-state model, and thedynamics of the closed-loop system. Topics requiring further researches are pointed out.