摘要
在讨论基于自适应神经网络的模糊系统 (ANFIS)建模基础上 ,提出一种基于多模型的在线故障诊断策略。首先应用ANFIS建模方法建立一组模型 ,然后构造了一个基于多模型的故障诊断系统产生偏差信号 ,并首次提出用由偏差绝对值的累加和构成的最小指数来探测故障的发生。综合考虑故障诊断的灵敏度和鲁棒性 ,采用故障指数来确认故障的发生。
On the basis of the discussion of model establishment of fuzzy system (ANFIS), a strategy of on line fault diagnostics based on multimodel is stated. First, a group of models is established by using ANFIS, then a multimodel based fault diagnostics system is structured to create deviation signal. With the accumulation of the absolute deviations and the minimum index created, the fault occurrence is investigated. Comprehensively considering the sensitivity and robustness of fault diagnostics, the fault occurrence is determined by using fault index.
出处
《自动化仪表》
CAS
北大核心
2004年第7期10-16,共7页
Process Automation Instrumentation
