大型旋转机械故障诊断专家系统ETHYLENE的理论研究  

A Theoretical Study on the ETHYLENE System—A Fuzzy Diagnostic Expert System for Large Rotating Machinery

作　　者：赵美德[1] 黄文虎[1] 夏松波 须根法 

机构地区：[1]哈尔滨工业大学计算机科学与工程系,哈尔滨150006

出　　处：《振动工程学报》1992年第2期105-113,共9页Journal of Vibration Engineering

基　　金：国家自然科学基金

摘　　要：本文在模糊产生式规则的基础上引入了加权模糊连接算子的概念,并研究了征兆的类型。在此基础上,为旋转机械模糊诊断知识提出了一个灵活的知识表达方法——广义模糊产生式规则。在征兆类型的基础上引入了征兆认可因子的概念.利用这一概念来描述模糊蕴涵关系的语义,基于征兆认可因子提出了一个模糊推理算法。文中还给出了由基于这一广义模糊产生式规则和模糊推理算法建立的透平压缩机组故障诊断专家系统ETHYLENE得到的两个诊断结果。In this paper a weighting fuzzy connection operator is introduced in fuzzy production systems, and four kinds of symptoms are presented. Based on the two new concepts. a new fuzzy knowledge representation method called the generalized fuzzy production rule is proposed for the formal representation of the diagnostic expert knowledge in rotating machinery. Endorsement factor is introduced for describing the semantics of fuzzy implication, and based on it a fuzzy reasoning algorithm is proposed. The ETHYLENE system-a fuzzy diagnostic expert system for turbocompressors is built based on the new fuzzy knowledge representation method and the new fuzzy reasoning algorithm. Two diagnostic results obtained by running the ETHYLENE sytem show that the system can act as a useful decision aid in the fault diagnosis of large rotating machinery.

关 键 词：故障诊断 专家系统 旋转机械 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]