特征尺度谱算法及其在轴承故障诊断中的应用

Feature scale spectrum algorithm and its application in bearing fault diagnosis

针对基于包络谱的方法在轴承故障诊断中存在故障特征谐波衰减过快、并受较强的转频分量干扰,导致对故障程度判断的可靠性降低等问题,提出了一种全新的特征尺度谱轴承故障诊断方法。通过对振动信号进行分帧处理,采用特征算子提取其波动特征,利用二值化阈值处理方式对特征帧信号进行分类,获得信号中的故障区间并对故障区间进行压缩,将故障区间转化成特征脉冲信号,实现故障信号能量增强,并利用仿真信号和轴承故障试验台采集的数据对所提方法进行验证。结果表明:相较基于包络谱的方法,所提方法解决了频谱中故障特征谐波衰减过快的缺点,消除了转频分量的干扰,在低信噪比情况下具有更强的诊断效果。

Here,aiming at problems of the method based on envelope spectrum in bearing fault diagnosis having excessive rapid attenuation of fault feature harmonics and stronger rotating frequency components interference to cause reduction of the reliability of judging fault degree,a new bearing fault diagnosis method based on feature scale spectrum was proposed.Through framing processing of a vibration signal,the feature operator was used to extract their fluctuation characteristics,and the binarization threshold processing method was used to classify feature frame signals,obtain fault intervals in these signals and compress fault intervals,convert fault intervals into feature pulse signals,realize energy enhancement of fault signals,and verify the proposed method by using simulated signals and the data collected with a bearing fault test platform.The results showed that compared with the method based on envelope spectrum,the proposed method can solve the problem of too fast attenuation of fault feature harmonics in spectrum,eliminate interference of rotating frequency components,and have a stronger diagnostic effect under the condition of low signal-to-noise ratio.