摘要
在分析输入环节脉冲型故障和阶跃型故障对动态系统的状态变化和测量系统输出信息的影响的基础上 ,本文采用正向容错 M-型滤波与反向 WLS估计结合的方法 ,建立了输入环节脉冲型故障的检测与故障幅度辨识算法 ;对于输入环节的阶跃型故障 ,文中建立了分别适用于定常系统和时变系统的差分变换法与扩展状态空间方法 .
Under quantitatively analyzing the influence of pulse-type faults and step-type faults from input components on state and output, a series of new fault detection approaches and magnitude identification algorithm are built, which are combination of onward fault-tolerant M-type filter with the reverse weighted least squared estimators. As for step-type faults, the difference transformation approach and the expanded state space approach are suggested for time-invariant and time-variant systems, respectively. These new methods are practical.
出处
《自动化学报》
EI
CSCD
北大核心
2002年第6期881-887,共7页
Acta Automatica Sinica
基金
面向 2 1世纪教育振兴行动计划项目 ( YL1318)
中国科学院支持高水平大学建设项目 ( KY2 70 6 )
中国博士后科研基金资助
关键词
动态系统
突发性故障
检测
辨识
Estimation
Fault tolerant computer systems
Least squares approximations
Mathematical transformations
Signal detection
