基于MEMD与MMSE的滚动轴承退化特征提取方法
The Method of Degradation Feature Extraction of Rolling Bearing Based on MEMD and Multivariate Multiscale Entropy

针对滚动轴承故障信号的非平稳性特征以及其退化状态难以识别的问题，提出了基于多维经验模态分解(MEMD)与多元多尺度熵（MMSE）的退化特征提取方法。该方法利用多维经验模态算法在多尺度化过程中能够有效的捕获信号不同尺度的成分的特性，更好地区分了不同退化状态的复杂度。首先，利用MEMD算法对滚动轴承不同退化状态对应的多通道信号进行同步自适应分解；然后，对多尺度IMF分量重构的信号进行多元多尺度熵分析。对试验信号进行处理，结果表明，该方法能有效反映滚动轴承退化趋势。
The method of extracting degradation features was proposed based on MEMD and MMSE to solve the problem that non-stationarity of fault signals of roller bearing and degradation condition, which was characteristic of non-ststionarity and hard to recognize. The character of MEMD was adopted to catch different scales of signals effectively during the process of multiscalization, which made complexity of different degradation condition distinguished better than other methods. Firstly, multichannel signals corresponding to various degradation condition of roller bearing were decomposed adaptively using MEMD, then, the reconstructed signals by multiscale IMF was dealt with MSE analysis. The results showed that the proposed method could efficiently evaluate the degradation trend of roller bearing by handing the experimental signals.