一种滚动轴承故障知识获取的新方法

A New Knowledge Acquisition Method for Fault Diagnosis of Rolling Bearings

针对滚动轴承诊断中故障样本不足和故障模式复杂且难以辨识的特点,提出了一种基于Weka软件数据挖掘平台的滚动轴承故障知识获取的新方法。该方法综合运用滚动轴承时域参数和小波包络谱特征参数,并选取与其运行状态密切相关的多个振动参数作为原始特征模式,然后借助Weka平台的C4.5决策树提取了滚动轴承故障知识规则,并加以解释。最后将该方法应用于现场采集到的大量轴承数据,结果表明该方法正确有效。

A new knowledge acquisition method is proposed for rolling bearing fault diagnosis based on Weka platform,which is able to overcome effectively the problem such as understanding complex diagnosis process,insufficient fault samples for rolling bearing diagnosis.In this new method,firstly,the time domain parameters and the wavelet envelope spectrum characteristic parameters are combined,according to the experience knowledge,some parameters of them are selected as the diagnosis features;secondly,the C4.5 decision tree algorithm in Weka platform is used to extract the rolling bearing fault diagnosis knowledge rules,and they are explained and analyzed;finally,the method is applied in large amounts of on-the-spot data of bearing.The results fully show the correctness and rationality of the method.