基于声发射信号的滚动轴承外圈疲劳剥落故障双冲击特征提取

摘要

Spalling is a major type of fatigue failures of rolling element bearings (REB).As the double impulses phenomenon can be observed in the vibration when a rolling element is entering or exiting the spall in the outer race or the inner race of the faulty REB,the double impulses phenomenon can be also observed in the acoustic emission (AE) signals generated by the faulty REB.If the space between the double impulses can be measured accurately,the width of the spall may be estimated.It will be helpful on the evaluation of remaining useful life (RUL)of REBs.The AE signal is more sensitive to the incipient fault and better to realize interferences isolation.Then,the separate treatment is used to extract the double impulses from the AE signal in this paper.In the proposed scheme,the AE signal was separate into two parts (the entering part and the exiting part) at first.Then the autoregressive model and the minimum entropy deconvolution method were used to enhance the two parts,respectively.After that,the complex Morlet wavelet based kurtogram were utilized to extract the optimal envelope,respectively.At last,the extracted envelopes were added together for the space measurement of the two impulses.Experiment analysis indicates that,the double impulse phenomenon can be effectively extracted in the AE signals of outer race spalled rolling element bearings.%疲劳剥落是引起滚动轴承失效的主要原因。跟振动信号一样，当滚道出现疲劳剥落故障时滚动体在进入和退出剥落区时的声发射信号也存在对应的两类不同冲击特征，称为双冲击现象。对双冲击特征的提取可实现双冲击间隔的有效测量。声发射信号具有对早期故障敏感、不易受噪声干扰等优点。采用将两类特征分离处理的方法，将声发射信号中两类冲击特征分为两部分，通过 AR 模型和最小熵解卷积滤波增强故障特征信号，和基于复 Morlet 小波的谱峭度图算法提取优化解调频带对应的包络信号，对包络信号相加并进行双冲击间隔测量。实验研究表明，该方法能够有效地分离出滚动轴承外圈疲劳剥落故障声发射信号中的双冲击特征。