摘要
针对连续控制系统,建立了由系统约束集、变量集和边集构成的二部图模型,提出了一种定性描述与定量分析相结合的故障诊断算法。该算法通过分离子系统,求解系统的关联矩阵及最大匹配,定义了描述变量与系统约束之间依赖关系的规则,并设计了关联矩阵分层算法,以此来计算控制系统残差。以一个线性系统为例,探讨了该算法的应用过程,并通过仿真实例验证了该算法的有效性。
After establishing a general fault diagnosis model based on bi-partite graphs,a fault diagnosis method combined qualitative description and quantitative analysis is proposed for continuous control systems.To obtain the reduced associated matrix and its maximum mathing,system composed of known variables is detached;And to calculate residuals for diagnosis,rules about relations between system faults and residuals from the bi-partite graph method are proposed,and the associated matrix＇s layering algorithms are also proposed.Finally,a simulation instance is proposed to discuss the method and validate its result.
出处
《计算机工程与设计》
CSCD
北大核心
2011年第12期4068-4070,4099,共4页
Computer Engineering and Design
基金
中国科学院科技创新基金项目(CXJJ-10-M20)
关键词
控制系统
故障诊断
二部图
匹配
关联矩阵
control system
fault diagnosis
bipartite graph
matching
the associated matrix
