频带多尺度复合模糊熵及其在轴承故障诊断中的应用

Spectral band refined composite multiscale fuzzy entropy and its application in fault diagnosis of rolling bearings

旋转机械设备发生滚动轴承故障的早期,受环境噪声影响,故障特征轻微.为了有效提取滚动轴承的故障信号冲击特征,以时频分析为基础,结合信息熵理论,提出一种频带多尺度复合模糊熵的故障诊断方法.与模糊熵相比,基于方差的频带多尺度复合模糊熵可以定量地表征非平稳信号的数据信息,抗干扰性强,更好地反映出不同频带分量在时间轴上的变化特性.引入自适应带通滤波器,成功实现对微弱故障的特征提取和故障识别.仿真分析和实验结果表明,提出的方法较传统滤波方法在降噪抑制方面效果更好,能够快速识别滚动轴承的冲击特征.

The fault characteristic frequency is mostly weak and submerged in background noise when the early fault occurs in the rotating machinery. A new method based on time-frequency analysis and spectral band refined composite multiscale fuzzy entropy was proposed for fault diagnosis. The proposed refined composite multiscale fuzzy entropy based on variance showed better performance than the conditional fuzzy entropy in quantifying the complexity of non-stationary signal and effectively reflected the variation of each frequency band component with time. The adaptive bandpass filter was introduced to extract weak fault feature through strong background noise. Compared with other filtered methods, the proposed method has advantage in recognizing weak shock features of rolling bearings, which is verified by the simulation and experimental results.