拟合迭代学习数据的工业过程控制器

Fitting industrial controller to iterative learning control

为进一步提高工业过程控制系统的跟踪能力,实现稳、准、快等性能,本文利用迭代学习算法产生的各类信息,在控制器函数拟合的基础上,设计了一种高精度跟踪的鲁棒控制器.首先在频域对闭环迭代学习算法进行分析,得出迭代学习控制器等同于级联控制器的结论,进而采用一个低阶结构的控制器去拟合误差序列与控制序列,避免了难以物理实现的高阶控制器,最后通过对一般的工业过程对象进行实验设计,结果表明这种控制器在快速性、无超调及控制精度上具有很好的优势,并且具有良好的抑制干扰能力.

In order to improve the tracking performance of the industrial process system, we fit a robust controller to the information derived from the iterative learning controller. First, the iterative learning algorithm is analyzed in frequency domain by considering the iterative learning controller a cascade controller, and then, a controller is designed to fit the control error and the controlled output, resulting in a realizable controller which is lower in order than the cascade controller. An example is given to illustrate the approach, which shows the advantages of the fitting controller in speed, of none overshoot, with high accuracy and good robustness against disturbances.