渐进辨识法的改进及在化工过程中的应用

A modified asymptotic identification method and its application

提出了一种基于随机数直接搜索法(NLJ)的渐进辨识方法。该方法不但消除了渐进法对输入信号的条件限制,并且提高了NLJ的辨识精度,获得了工业对象的近似模型,对生产运行影响较小,在实际PID整定中得到了较好效果。

This paper introduces an asymptotic identification method based on the NLJ method. The modified procedure combines the advantages of both methods, and can improve the accuracy of the NLJ method without introducing the characteristic signal of the asymptotic method. An exact model can be derived without disturbance of the process and its utility has been shown by industrial application.