中文摘要：

针对目前化工过程复杂,在分布式模型预测控制（MPC）的实施中会面临强耦合以及慢速收敛的问题,提出了改进全局性能指标的快速分布式MPC算法。首先在每个采样时刻分别求解子系统自身的局部优化问题,同时考虑关联子系统之间的相互作用,然后在协调过程中对全局最优性能指标进行改进以减少迭代次数。该方法降低了控制问题的复杂度,减少了迭代时间,有效地改善了收敛速度。最后分别对二元精馏塔过程和苯乙烯聚合反应过程进行了仿真,验证了所提算法的有效性。

英文摘要：

Combined with the characteristics of current complex chemical process,there will be some problems in the process of DMPC such as the strong interaction effects and the slow convergence. To solve these problems,the fast distributed MPC algorithm for improving global performance index is proposed. Firstly,the optimal control of each subsystem is obtained at each sampling time and the interaction between interconnected subsystems should also be considered at the same time. Then,the global optimal performance index is improved to reduce the number of iterations in the process of coordination. This method reduces the complexity and iterations time of the control problem. It also improves the convergence speed effectively. In the end,the simulation in a distillation column process and styrene polymerization reaction process shows the validity of the proposed algorithm.