诊断汽轮发电机组振动故障的一种模糊聚类分析方法  

A Fuzzy Clustering Method for Turbo-generator Vibration Faulty Diagnosis

作　　者：张彼德[1] 孙才新[1] 欧健 周银春 

机构地区：[1]重庆大学电气工程学院高压系,重庆400044

出　　处：《汽轮机技术》2002年第5期289-291,共3页Turbine Technology

摘　　要：提出的模糊聚类分析方法将标准故障样本与待检样本一起作为分类样本 ,运用基于传递闭包的模糊等价矩阵法和F统计量对原始数据进行初始分类 ,然后求出模糊 c划分的初始叠代矩阵 ,再用ISODATA方法进行验证并得到最优分类矩阵和聚类中心 ,并依此来判断侍检样本的故障类型。通过应用表明 ,提出的这种方法可有效地诊断汽轮发电机组振动故障 ,具有良好的应用前景。This paper introduces a fuzzy clustering method in turbogenerator vibration faulty diagnosis. The classfied samples is composed of standard faulty samples and checked samples. The initial sort is acquired by fuzzy equivalent matrix and F statistic. At the same time, the intial iterative matrix for the fuzzy c means is also acquired. So it gets the most optimize class matrix and the fuzzy clustering center. The experimental results indicate that the method in this paper is available and has a good foreground.

关 键 词：诊断 汽轮发电机组 振动故障 模糊聚类 故障诊断 

分 类 号：TM621.3[电气工程—电力系统及自动化]