基于加权的区域预测控制算法研究

Research on the algorithm of region based on weighted prediction

本文针对多变量有约束区域预测控制问题,提出一种基于加权的区域预测控制算法。该方法首先利用实际输出与预测输出之间的偏差对模型预测输出进行修正并对约束区域进行软约束处理,在此基础上对约束区域进行分段处理,最后提出基于加权的区域预测控制方法。该方法根据预测输出与给定区域之间的偏离值设定误差加权矩阵,如果预测输出超出给定区域,则重新计算控制变量的变化量,进而构成一种基于加权的区域预测控制算法。将本文方法与传统方法进行对比仿真研究。仿真结果表明:该方法能使违反区域的被控变量能快速返回给定区域,并能保证系统的稳定性,很好的协调了控制的快速性与保持系统稳定性之间的矛盾。

A weight method is developed for the complicated industrial process control with zone constraint requirement for multivariable predictive control problem. First revising the model prediction output and adjusting the soft constraint region by using the error between the actual output and the predictive output, on the basis of the constraint region segmentation processing, at the last propose the predictive control algorithm based on the weighted area. The method reset the error weighting matrix according to deviation between the predicted output and the given output, the change of control variables will be recalculated, if the output exceeds the given contain, thus constitute a predictive control algorithm based on weighted. Compare the method in this paper with the traditional one. The simulation results show that make the variables which break the breach area return to the controlled area quickly and guarantee the stability of the system, coordinate the contradiction well between the stability of the control system and the rapidity of the control system.