粗集理论及其在旋转机械故障诊断中的应用(四)决策表属性的约简  

Rough sets and fault diagnosis of rotating machinery on rough sets theory (IV) Reducts of decision attributes

作　　者：许琦[1] 李永生[1] 詹月林[2] 

机构地区：[1]南京工业大学机械与动力工程学院,江苏南京210009 [2]盐城工学院化学工程系,江苏盐城224002

出　　处：《南京工业大学学报（自然科学版）》2004年第1期24-27,共4页Journal of Nanjing Tech University(Natural Science Edition)

基　　金：江苏省自然科学基金资助项目(BK2001095)

摘　　要：在旋转机械典型故障实验的基础上,按照粗糙集理论的要求,对采集的各种信息由数据库向符合粗集理论要求的知识库的转换,建立故障振动特征决策表。在决策功能不变的前提下,利用区分矩阵和区分方程,对决策属性进行简化,并寻找属性集的核和属性集合的选优。通过对4种典型旋转机械故障的频域特性表的约简,将6个条件属性简化为3个条件属性。On the basis of typical experiment of the rotating machinery, according to the requests of rough sets theory, all kinds of information that collected by system transformed form the database to repository which accord with the requests of rough sets theory, establish the character decision table of the fault vibration,at the premise of invariable decision attribute, taking advantage of discernibility matrix and discernibility function, simplifying the properties of the decision-making table, and seeking the core of attribute sets of properties and the minimum reducts of attribute value. The decision table of four faults type spectrum is reduced by rough sets theory, the three conditions are taken out form six conditions.

关 键 词：旋转机械 故障诊断 粗糙集 数据库 知识库 频域特性表 转子 决策表 

分 类 号：TH133[机械工程—机械制造及自动化] O159[理学—数学]