增量式函数观测器的概念及其状态反馈控制系统的应用  

CONCEPT OF INCREMENT FUNCTION OBSERVER AND ITS APPLICATION IN STATE FEEDBACK CONREOL SYSTEM

作　　者：韩忠旭[1] 

机构地区：[1]中国电力科学研究院,北京100085

出　　处：《中国电机工程学报》2004年第4期210-216,共7页Proceedings of the CSEE

摘　　要：Luenberger 提出的状态观测器理论可以解决在确定性条件下受控线性系统的状态重构问题,然而,物理世界中一切实际的系统都具有不同程度的非线性。该文采用Hammerstein 模型描述一类非线性系统,明确提出增量式函数观测器的概念,从基本定义出发,通过数学分析,证明 Luenberger 函数观测器、全维状态观测器和增量式状态观测器均为增量式函数观测器的特例。由于增量式函数观测器的数学定义是建立在非线性受控系统的基础上,因此,观测器的应用便可扩展到非线性控制领域。理论分析表明,在选择同样的观测器参数的前提下,增量式观测器的鲁棒性要优于鲁棒 Luenberger 观测器。在 3 座火力发电厂的 6台 300MW 以上容量单元机组工程应用的实践已证明,基于增量式观测器的状态反馈控制与 PID 控制相结合,能实现对大滞后复杂受控对象的高质量控制,具有很强的稳定鲁棒性,其控制算法简单实用,且便于推广。The state observer theory proffered by Luenberger can resolve the state reconstruct of linear control system in case of the conclusive condition. But the entire actual world has characteristic of nonlinear at some extent. This paper describes a type of nonlinear system by means of model of Hammerstein. Based on which, the concept of increment function observer was presented clearly. Starting from the basic definition, it was certified by mathematical analyses that all of Luenberger function observer, full dimension state observer and increment state observer is special case of increment function observer. With mathematical definition of increment function observer is based on nonlinear controlled system, so application of observer can be able to extend nonlinear control field. The theory analysis shows that the robustness of increment function observer precede that of robust Luenberger observer, premising parameters of the two type of observer are same. It has been certificated by practices of engineering application of 6×300MW unit generators in 3 power plants that state feedback control based on increment function observer combined with PID control can be able to achieve high quality control index to large lag complexity controlled object, which has strong steady robustness and its control algorithm is briefness, practicality and easy to spread.

关 键 词：状态反馈控制系统 增量式函数观测器 HAMMERSTEIN模型 PID控制 分散控制系统 控制算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]