Zafiriou—Morari算法分析及其自适应系统

A Self-Tuning Controller Using Zafiriou-Morari Algorithm

本文简述现有计算机控制算法的不足,介绍了Z-M算法的设计原则并分析其具有稳定性好、适应能力强等优点的原因。仿真证实,该算法在对象模型不准确时可能产生不稳。本文就此提出在线递推辨识、在线求取对象零点的方法,并利用参数可调的Z-M控制器构成自适应控制系统。对该自适应系统进行计算机仿真,取得了满意的效果。

The paper discusses the Zafiriou-Morari' s and other digital control algorithm for linear single-input and single output systems, and finds that the Zafiriou-Morari algorithm possesses better performances than the others. Nevertheless, the Z-M Controller may be unstable when the model is not exact and the modelG(z) and actual plant G(z) do not satisfy G(l) G(l)>0.This paper presents a self-tuning controller which uses.a recurrent square-root algorithm to identify the model of the plant, and a fast recurrent algorithm晅o find the zeroes of the model, then adjusts the Z-M control on-line.Computer simulation shows that the self-tuning controller using Z-M algorithm overcomes the defects caused by the inexactness of model and has better stability.