摘要
本文基于δ算子构造了Ｈｉｌｂｅｒｔ空间中的信号向量，推导了系统的输入输出关系，得出了前向和后向预测误差向量的表达式．借助格形滤波器理论导出了故障残差序列的递归算法，从而设计出故障检测滤波器．将该滤波器与自适应噪声抵消技术相结合，实现了对被噪声污染信号的故障实时在线检测．最后，仿真实验验证了这种故障检测方法的有效性．
In this paper, based on Delta Operator, signal vectors of Hilbert space are constructed, an input-output model is derived, and both the forward and backward prediction error vectors are described. With the help of Lattice filters, recursive algorithms for generating time-updated as well as order-updated residual series are formed, so that a fault detection filter is designed. Combined with the approach of adaptive noise canceling, real time, on-line fault detection is carried out. Finally.simulating experiments show that this fault detection method is effective.
出处
《自动化学报》
EI
CSCD
北大核心
1994年第4期413-419,共7页
Acta Automatica Sinica
基金
国家"八五"科技攻关资助
关键词
故障检测
Δ算子
格形滤波器
fault detection
Delta Operator
Lattice filter.
