摘要
提出了一种基于特征空间矢量的故障诊断方法。该方法物理意义明确，可以通过对故障诊断误差的学习，实时修正故障模型，实现对不确定性和慢时变性对象的鲁棒故障诊断。在对象故障先验知识不完备的情况下，能够通过学习逐步建立对象完善的故障诊断模型。还可以对诊断模型的故障可分离度进行评判，为优化征兆信号选取提供条件。示例表明，该方法能适用于具有不确定性和慢时变性的复杂对象的故障诊断。
This paper presents a fault diagnosis method based on characteristic space vector. The method which has definite physical meaning, can change fault diagnosis model to adapt to the variability and uncertainty of the plant by learning the fault diagnosis err. It can build perfect fault diagnosis model through learning if the fault information is not complete and has the capability to assess the diagnosability of the fault diagnosis model which can be used for optimal symptom selection. Fault diagnosis examples demonstrated that this method can be used as robust fault diagnosis of complicated system with variability and uncertainty.
出处
《中国电机工程学报》
EI
CSCD
北大核心
1999年第9期53-56,共4页
Proceedings of the CSEE
关键词
故障诊断
特征空间矢量
专家系统
fault diagnosis
characteristic space vector
symptom
learning
