基于变分模态分解奇异值熵的滚动轴承微弱故障辨识方法

Weak fault identification of rolling bearings based on VMD singular value entropy

针对滚动轴承微弱故障难以识别的问题,提出一种基于变分模态分解(Variational Mode Decomposition,VMD)与奇异值熵融合的滚动轴承微弱故障辨识方法。该方法对滚动轴承的振动信号进行VMD分解获得4个本征模态函数(Intrinsic Mode Function,IMF),并根据一种均方差-欧氏距离指标选择出含丰富故障信息的IMF分量进行信号重构;对重构信号进行奇异值分解获得奇异值对角阵,进而结合信息熵理论求取对角阵的奇异值熵;利用奇异值熵的大小区分滚动轴承的工作状态和故障类型。用美国西储大学的滚动轴承振动信号对所述方法进行验证的结果表明:相比传统EMD奇异值熵故障诊断方法,该方法能够更清晰地划分出滚动轴承微弱故障的类别区间,有助于实现微弱故障类型的准确辨识,为滚动轴承微弱故障诊断提供了一种可靠的评估依据。

Aiming at problems of rolling bearings’ weak failures being difficult to identify,a rolling bearing weak fault identification method based on fusion of the variational mode decomposition（ VMD） and the singular value entropy was proposed. Firstly,VMD was done for vibration signal of a rolling bearing to obtain 4 intrinsic mode functions（ IMFs）. According to a mean square deviation-Euclidean distance index,IMF components containing rich fault information were chosen to perform signal reconstruction. Then,the singular value decomposition（ SVD） was done for the reconstructed signal to obtain the diagonal matrix of singular values,and the diagonal matrix’s singular value entropy was obtained using the information entropy theory. Finally,the singular value entropy was used to distinguish working state and fault type of the bearing. The new method was verified with rolling bearing vibration signals of American West Storage University. The results showed that compared with the traditional EMD singular value entropy fault diagnosis method,this method can more clearly delineate rolling bearing weak fault classes to correctly identify a bearing’s weak fault. This study provided a reliable basis for weak fault diagnosis of rolling bearings.