Research On The Feature Extraction Method Of The Partial Failure Of Planetary Gearbox Gears

As a key component of the powertrain,planetary gearboxes are widely used in military,industrial,and life fields.At the same time,the loss caused by its fault is countless,so the research on planetary gearbox fault diagnosis is of great significance.Feature extraction is the key technology for fault diagnosis.In order to extract the fault feature information from non-stationary and nonlinear impulse signal,a method combining improved dynamic time warping and improved empirical wavelet transform is proposed to extract the features of planetary gearbox gear faults.The main research contents of this paper are as follows:(1)According to the two characteristics of planetary gearbox(the vibration generated by more gear meshing is superimposed on each other at the same time,the transfer path between each the meshing point and the acceleration sensor presents periodic change),the paper lists the gear local fault vibration signal model of the planetary gearbox,and gets through the model and the experiment data of planetary gearbox to simulate the normal signal and fault signal.(2)First,the test signal is aligned with the reference signal by dynamic time warping,and then the original signal is recovered by the resampling method,thereby obtaining a residual vector signal and highlighting the fault characteristic information.Due to the low efficiency of standard dynamic time warping,the residual vector signal after warping is prone to singularity and cannot be naturally aligned.The paper proposes an improved method:by combining data abstraction and window constraints,the efficiency of the standard dynamic time warping method is improved;the cumulative distance formula is modified by the cross-correlation number and the derivative of the two time-domain signals.(3)In order to further extract the features of the vibration signal,the empirical wavelet transform is used to perform modal decomposition of the residual vector signal,obtain the intrinsic modal component,and filter out the irrelevant signal.Since the spectral segmentation of the empirical wavelet transform is prone to error,the paper proposes an improved method,which performs envelope analysis on the Fourier spectrum and segment the frequency spectrum by dividing the envelope region.In the paper,a new fault feature extraction method based on the improved dynamic time warping method and the improved empirical wavelet transform method is proposed.The effectiveness of the method is verified by simulation analysis and experiments.