基于FM^mlet变换与非负矩阵分解的轴承故障诊断

Roller Bearing Fault Diagnosis Method Based on FM^mlet Transform and NMF

针对轴承振动信号的非平稳特征和现实中难以提取故障参数的问题,提出一种基于FM^mlet变换与非负矩阵分解算法(NMF)的轴承故障诊断方法。采用FM^mlet变换对轴承振动信号进行时频分析,能较好匹配信号的线性和非线性时变成分。在此基础上,引入非负矩阵分解计算特征参数,实现了轴承振动谱图像的自动诊断。将该方法应用于轴承4种典型工况的故障诊断实例中,结果证明了方法的有效性。

The vibration signals of bearings are usually non-stationary and it is difficult to extract the fault parameters in reality. Therefore a fault diagnosis method that uses the FM^m let transform and Non- negative Matrix Factorization （NMF） is proposed. FM^m let transform can preferably match the nonlinear time-varying signal components. To let the vibrational spectra diagnosed automatically, NMF is also brought in to calculate the characteristic parameters. The fault diagnosis test of the 4 states of bearing fault proves the effectiveness of method.