自适应MCKD和VMD在行星齿轮箱早期故障诊断中的应用

Early Fault Diagnosis of Adaptive MCKD and VMD in Planetary Gearbox

针对行星齿轮箱早期故障信号微弱且受强背景噪声影响,致使故障信号特征频率难提取,通过自适应最大峭度解卷积(MCKD)和变分模态分解(VMD)进行早期故障特征提取。首先,利用变步长搜索,以峭度值为评判标准,搜索最优滤波器长度L;然后,将信号通过优化后的自适应MCKD算法降噪;最后,利用VMD分解降噪信号,通过包络谱进行分析,寻找故障特征频率。经仿真信号和实验信号验证,这里所提方法能够有效地提取出强噪声背景下的行星齿轮箱故障特征。

The early fault signal of the planetary gearbox is weak and strongly influenced by the background noise,which makes the characteristic frequency of the fault signal difficult to extract. The early fault feature extraction is performed by adaptive maximum kurtosis deconvolution(MCKD)and variational mode decomposition(VMD). Firstly,using the variable step size search,the kurtosis value is used as the criterion to search for the optimal filter length L. Then,the signal is denoised by the optimized adaptive MCKD algorithm. Finally,the VSD is used to decompose the noise reduction signal and pass the packet. The spectrum is analyzed to find the frequency of the fault feature. The simulated signal and experimental signal verify that the proposed method can effectively extract the planetary gearbox fault characteristics under strong noise background.