基于最优解耦的关联大系统的故障检测  

The Fault Diagnosis of the Large-scale System Based on Optimal Decoupling

作　　者：徐启程[1] 靖新[1] 孙常春[1] 

机构地区：[1]沈阳建筑大学理学院,辽宁沈阳110168

出　　处：《计算技术与自动化》2007年第4期28-29,共2页Computing Technology and Automation

基　　金：辽宁省自然科学基金项目(20062003)

摘　　要：对于关联大系统的故障检测,通常的做法是通过进行观测器的特征结构配置使残差信号与动态关联项完全解耦,这种方法具有很大的局限性。本文将对这一问题作进一步扩展,当系统模型不能满足残差与关联项解耦时,运用最优解耦理论,通过矩阵的奇异值分解实现最优解耦,从而保证故障的检测效果。For the large- scale system, the residual and interconnected terms can be decoupled by using eigenstucure assignment, this paper deals with the situation which the system can not completely decouple with the interconnected terms, and gives a method to obtain a residual which is sensitive to the sensor faults, but insensitive to the interconnected terms, Based on the theory of optimal decoupling, we employ the theory of singular value decomposition on solving optimal decoupling.

关 键 词：关联大系统 故障检测 残差 观测器 最优解耦 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]