粗糙集及主元分析的机械故障诊断研究  

The Research on Mechanical Fault Diagnosis Based on Combining Rough Set with Principal Component Analysis

作　　者：姜万录[1] 刘思远[1] 

机构地区：[1]燕山大学机械电子工程系,河北秦皇岛066004

出　　处：《机床与液压》2009年第12期215-218,共4页Machine Tool & Hydraulics

基　　金：国家自然科学基金项目(50775198);河北省自然科学基金项目(E2008000812)

摘　　要：针对粗糙集理论方法与主成分分析方法的特点,引入了一个启发式函数,提出了一种将两种方法相结合的算法。运用粗糙集理论建立了信息系统的分辨矩阵,通过主元分析方法得到了各主成分的表达式和主成分贡献率,通过启发式算法得到了决策信息表的最佳属性约简和故障诊断规则,并通过算例分析验证了该算法的优越性。研究表明该算法在处理机械故障诊断中数据量大、变量多的数据时,具有很好的效果。By introducing a kind of heuristic function, a new algorithm based on rough set theory and principal component analysis method was presented. A resolving matrix of information system was established by using of rough set theory, the expression and the contribution rate of the principal component were got by using of principal component analysis method, the best attribute reduction and fault diagnosis rules were got by means of heuristic algorithm. The diagnosis result of an example shows that the algorithm has superiority on disposing of the problems about mechanical fault diagnosis full of large quantity of data and more variables.

关 键 词：粗糙集 主成分分析 分辨矩阵 故障诊断 属性约简 

分 类 号：TP306.3[自动化与计算机技术—计算机系统结构]