离散多变量递推自校正PID控制

A Method for Discrete Multivariable Recursive Selftuning PID Control

本文给出了一种新的离散多变量递推自校正PID控制器设计方法。该方法只对被控过程进行近似解耦,并通过极小化一个广义性能指标函数以获得PID控制器参数。广义性能指标函数是由预测误差及控制作用所组成,极小化问题的解是以递推形式给出,便于在每一周期里对PID参数进行新的修正。自校正算法不受被控过程模型阶数的限制,可以很方便地为高阶被控过程设计PID控制器。一个仿真算例表明了本文方法的有效性。

A new design method for the self-tuning of discrete multivariabie recursive PID controller is proposed. In this method only an approximate docoupling is needed and the controller parameters are obtained by the minimization of a generalized performance index that includes prediction errors and control efforts over a period of time. The solution of the minimization problem is given as a recursive algorithm which can easily updata the PID parameters at each sampling time. The method is free of the degree of the controlled process model and convenient to design a PID controller for any high degree process. The method has been successfully tested by simulation study and some results are presented.