最大相关峭度解卷积结合1.5维谱的滚动轴承早期故障特征提取方法

Feature extraction for rolling bearing incipient fault based on maximum correlated kurtosis deconvolution and 1. 5 dimension spectrum

滚动轴承处于早期故障阶段时,特征信号微弱,并且受环境噪声影响严重,因此故障特征提取困难。针对这一问题,尝试将最大相关峭度解卷积方法引入到滚动轴承故障诊断领域,并与1.5维谱结合,提出了最大相关峭度解卷积结合1.5维谱的轴承早期故障特征提取方法。首先对故障信号做最大相关峭度解卷积预处理,然后计算解卷积信号的包络信号,最后对包络信号做1.5维谱分析,从而得到解卷积信号的1.5维包络谱,通过分析谱图中幅值突出的频率成分来判断故障类型。滚动轴承故障模拟及实测信号分析结果表明,该方法可有效提取早期故障特征频率信息,具有一定可靠性。

Early fault feature of rolling bearing is very weak and affected by environment noise seriously,so it is difficult to be drawn.Aiming at solving this problem,maximum correlated kurtosis deconvolution （MCKD） was introduced to the field of fault diagnosis for rolling bearing and combining with the 1.5 dimension spectrum,a feature extraction method for rolling bearing incipient fault was proposed.The fault signal was processed by MCKD method and the envelope of its deconvolution signal was calculated,then the envelope signal was analysed using 1.5 dimension spectrum method.The bearing fault was judged by analyzing the frequency components of 1.5 dimension envelope spectrum.The analysis results of simulated and measured fault signals of rolling bearings show that the method can effectively extract the feature frequency information of incipient fault and has a certain reliability.