基于灰色关联-K近邻法的设备故障检测研究  

Research on equipment fault diagnosis based on grey correlation and K nearest neighbour

作　　者：周振阳 李亚慧 温丹丽[1] Zhou Zhenyang;Li Yahui;Wen Danli(Software College,Shenyang Normal University,Shenyang 110034,China)

机构地区：[1]沈阳师范大学软件学院,辽宁沈阳110034

出　　处：《无线互联科技》2021年第13期109-110,共2页Wireless Internet Technology

基　　金：2020年沈阳师范大学“大学生创新创业训练计划”资助项目,项目编号:x0202010166127。

摘　　要：文章采用灰色关联-K近邻法,快速、准确地检测球磨机齿轮磨损构件位置,将实时采集到的齿轮振动频谱中齿轮啮合频率、谐波幅值和垂直振动幅值作为训练集信号的特征向量,并将这些作为标准样本,求出每一个实时的测试样本特征向量与标准样本之间的灰色关联系数,并用K近邻法对训练测试样本做出判定,与异常数据库中数据实时进行分类识别,如果属于异常数据类别,则判定齿轮已发生磨损。该方法具有样本量小、计算量小、识别速度较快、准确性能高等优点。The grey correlation and K nearest neighbor method is used to check the position of damaged components of ball mill gears quickly and accurately.Gear meshing frequency,harmonic amplitude and vertical vibration amplitude collected in the real-time in the gear vibration frequency are taken as the feature vectors of the training set signal,and these are taken as the standard samples.The grey correlation coefficient between the feature vector of each real-time test sample and the standard sample is obtained.The K nearest neighbor method is used to determine the training test sample.The data in the abnormal database are classified and identified in real time.If it belongs to the abnormal data category,it is determined that the gear has been worn.This method has the advantages of small sample size,small calculation,fast recognition speed and high accuracy.

关 键 词：灰色关联 近邻法 故障检测 

分 类 号：TH17[机械工程—机械制造及自动化]