基于MLD的加权区间预测控制

MLD Based Weighted Zone Predictive Control

针对生产过程中某些被控变量并没有苛刻的设定值要求,只要满足在给定区间内即可的问题,提出了一种基于混合逻辑动态模型的加权区间预测控制算法,在控制器设计中引入逻辑变量,将基于逻辑命题的目标函数以约束条件的形式集成到预测控制器模型框架下。通过逻辑变量的取值检测被控变量预测值所处的偏差区域,设定相应的误差权重值,使预测控制根据输出在不同的偏差区域范围,采取不同的控制强度,从而保证快速控制的同时满足稳定控制的要求。通过对Shell塔的仿真,验证了算法的有效性。

Aiming at the problem of no rigorous set-point request for some controll variables with only satisfaction requirement of a zone control goal for manufacture process, a weighted zone predictive control algorithm based on mixed logical dynamic model is proposed. Logical variables is introduced into controller design with objective function based on logical proposition integrated into framework of predictive controller in form of constraints. Deviation area of controlled variable predictive value is detected through value of logical variable. Corresponding error weight values are set according to different deviation error range. The predict control takes different control strength to meet requirement of guaranteeing fast and stable control. The effectiveness of proposed algorithm is verified with simulation of shell column.