预测控制拟等效集结的仿真研究.pdf

第 21 卷第 20 期 系 统 仿 真 学 报© Vol. 21 No. 20 2009 年 10 月 Journal of System Simulation Oct., 2009 预测控制拟等效集结的仿真研究 姜 舒，席裕庚，李德伟 (上海交通大学自动化系，上海 200240) 摘 要：由于实际系统中通常存在约束，如何保证预测控制集结策略在有约束系统中的控制质量， 对于集结策略的实际应用有着重要意义。这里通过对预测控制拟等效集结策略进行仿真分析，提出 了一种确定拟等效集结策略中参数取值的算法。应用该算法，可以使基于拟等效集结的预测控制 器对于有约束系统在降低在线计算量的同时保证系统的控制质量，进而达到近似等效的目的。 关键词：预测控制；集结策略；拟等效集结；控制质量 中图分类号：TP273 文献标识码：A 文章编号：1004-731X (2009) 20-6547-05 Research of Quasi-equivalent Aggregation Strategy in Model Predictive Control Based on Simulation JIANG Shu , XI Yu-geng, LI De-wei (Automation Department, Shanghai Jiaotong University, Shanghai 200240, China) Abstract: Constraints always exist in many practical control systems. How to guarantee the control quality is very important for the aggregation based Model Predictive Control (MPC). Through analyzing the quasi-equivalent aggregation strategy of MPC on the basis of the simulations, an algorithm was proposed to determine the parameters of quasi-equivalent aggregation strategy. By this algorithm, for the quasi-equivalent aggregation based MPC, the on-line computation complexity could be reduced as well as the control quality was guaranteed. Key words: model predictive control; aggregation strategy; quasi-equivalent aggregation strategy; control quality 引 言1 了集结前后控制效果不变。在此基础上文[8]又提出了适用于 有约束系统的拟等效集结策略，但没有给出“自由变量个数” 预测控制是一种有限时域的滚动优化控制方法，它可以