Fault Diagnose Of Gearbox Under Varied Working Condition Based On Morphological Component Analysis

Gearbox is a key component with greater chance of failure in mechanical drive system.it is difficult for traditional methods of gearbox failure diagnose to measure correctly and distinguish effectively.researching the vibration signal features of gearbox under varied condition through acquisition signal resampled in experiment platform,virified the features of gearbox vibration signal under varied condition,inchuding varied rotating speed,varied load,nolinear and multicomponent.The nolinear and multicomponent signal lead the difficulty of extracting failure features;the phenomenon of frequency smearing result in hard to confirm characteristic frequency;varied load cause the trouble of recognizing between normal gear and failure gear.The above problems were improved:Firstly,Based on the fixed dictionary of morphological component analysis(MCA),a new method for fault diagnosis of rolling bearings based on learning dictionary of MCA was proposed.The local failure of inner ring and outer ring present different forms in the time domain waveform when the bearing rings fault.Take the faults data of inner ring and outer ring as training sample respectively,searching for optimal dictionary space through K-singular value decomposition(K-SVD);Replacement of fixed dictionary in the method of MCA by learning dictionary,Separation of the fault characteristics component of bearing rings and noise according to the morphological differences of each component;Analysis the faults location of bearing via Hilbert envelope spectral of each component.Secondly,extracting infromationg via method of computing rotation rate based on numerical differentiation or evaluating of instaneous frequency based on Chirplet Path Pursuit;gained time-angle function by integral of speed curve.According to the theory of order tracking to ensure the resample order and resample angle,then,computing resample time by interpolation of time-angle function;The original vibration signal interpolated by uniform angle re-sampling time through order tracking based on numerical differentiation,which transforms non-stationary vibration signals into Angle domain stationary signal.Avoid frequency smearing phenomenon result from the change of speed.The original vibration signal interpolated by uniform angle re-sampling time through order tracking Thirdly,Fault Diagnose of Gearbox under varied working condition Based on Morphological Component Analysis was proposed.The original vibration signal interpolated by uniform angle re-sampling time through order tracking,which transforms non-stationary vibration signals into Angle domain stationary signal;subsequently,shock component,harmonic component and noise component is separated from angular domain stationary signal through MCA,which can extract the fault feature from non-linear and multi-component signal in gearbox;then,The angle domain averaging of shock component highlight the fault characteristic;finally,using Instaneous power spectrum(IPS)to identify failure of gearbox.The paper apply the methoud of dictionary learning when choosing dictionary.Simulation and experiment show that dictionary through learning is better matching structure characteristics of the complex signal than fixed dictionary,In view of variable speed and variable load of vibration signal with the characteristic of nonlinear,multi-component,The combination of Morphological component analysis,order tracking and instaneous power spectrum achieve gearbox fault diagnose in varied working condition.