Research On The Techniques Of Feature Extraction And Noise Canceling To Gearbox Fault

When the fault of gear box occurs,the vibration signals often show non-stationary characteristics and the fault information is often submerged in the strong background noise.When the composite fault of gear box occurs,the vibrations become more complex and difficult to achieve effective diagnosis. This paper,based on Hilbert-Huang Transform,is research on the techniques of the fault feature extraction and noise canceling of gearbox.It mainly includes following aspects:Firstly,the vibration mechanism and fault types of gearbox are studied,the spectrum features of fault signals of gears,bearings and other components and characteristic frequency which is corresponding to different fault are analyzed in detail.Secondly,the basic theory of Hilbert-Huang Transform and its characters are introduced,the simulation result show that noise reduction and feature extraction can be achieved by the Hilbert-Huang Transform method in the low SNR conditions,and the Hilbert-Huang Transform will be useful to the fault signal analysis of gearbox.Thirdly,a diagnosis method to single fault and composite fault of gearbox based on Hilbert-Huang Transform power spectrum analysis was proposed.Hilbert transform was used to intrinsic mode functions to obtain the power spectrum. The comparison of spectral analysis and theoretical result to achieve fault diagnosis of gearbox.Fourthly,the analysis result to the test data of gearbox fault was proved that Hilbert-Huang Transform is useful to the fault feature extraction and noise canceling of gearbox,and shown that the proposed method can achieve diagnosis to single fault and composite fault of gearbox.