摘要
基于预测控制应用在多模型系统时的切换扰动，在切换时间ｔ０ 已知条件下，通过重新构造系统的步限矩阵，提出了一种控制系统平滑切换的广义预测控制算法．预测控制中步限矩阵是预测步限Ｎ 步内Ｄｉｏｐｈａｎｔｉｎｅ 方程的系数．研究了当前计算时刻ｔ在计算步长ｋ：①ｔ＋ ｋ＜ ｔ０＋ １；②ｔ＋ ｋ＝ ｔ０＋ １；③ｔ０＋ １＜ ｔ＋ ｋ≤ｔ＋ Ｎ 三种情况下步限矩阵的元素构成，实现了多模型控制系统的平稳切换．
Hybrid systems for which continuous parts and discrete event parts interact are usual in industrial processes. These systems, such as variable structure processes (with relays, valves, clutches,…) or processes with different operating modes, can often be represented by a linear transfer function with jumping parameters associated to step like perturbations. An effective Generalized Predictive Control (GPC) algorithm that can diminish the effect of commutation in the multi model systems was presented. It is suitable when the switching time is known in advance, which is often the case. It can greatly reduce the effect of the perturbation on the output and on the variations of the control variable. The horizon matrix of GPC algorithm can be turned so as to switch smoothly. The horizon matrix of GPC is composed of the coefficient in Diophantine function with N predictive horizon. If the computing step is k and the present moment is t , the elements of the horizon matrix in three cases: ①t+k<t 0+1; ②t+k=t 0+1; ③t 0+1<t+k≤t+N were studied in this paper. Hence, the control systems can be switched smoothly. The simulation result shows the effect of this method.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
1999年第11期1345-1347,共3页
Journal of Shanghai Jiaotong University
