Research On Fault Diagnosis Methods For Gearbox Based On Hilbert-Huang Transform

Gearbox is part of connect and power transmission in many mechanical devices, while gear failures such as wear, cracks, broken teeth of gear are important factors to inducing abnormal operation of machinery, so monitoring and diagnosis gearbox fault accurate and timely is very necessary. Since the gear vibration signals present non-stationary, non-linear characteristics, and it is often associated with a strong background noise, noise will affect accuracy of gearbox failure diagnosis. Therefore, a signal de-noising method of combination EMD decomposition and autocorrelation function is proposed. Hilbert-Huang transform method can decompose signal adaptive based on itself characteristic, effectively extract the fault characteristic frequency. Gearbox fault diagnosis is achieved by analyzing the Hilbert spectrum and Hilbert marginal spectrum, etc.The main contents and results of this thesis include:(1) The common form of gear damage and its causes are studied, and the effectiveness of fault detection parameters is analyzed according to the gear fault symptom. The relationship between typical gear fault and corresponding vibration signal characteristic frequency is obtained based on the meshing frequency modulation phenomenon and sideband distribution which appeared when gear fault occurred.(2)In order to suppress noise interference in gear fault signal, and highlight fault characteristic frequency, signal de-nosing method is proposed based on a combination of correlation analysis and EMD. This de-noising method is better than EMD de-noising and its effectiveness is demonstrated by simulation test stationary and non-stationary signals which are polluted by noise.(3) Using Hilbert-Huang transform analyze vibration signals.Gearbox Fault Diagnosis analysis about gear tooth wear, gear cracks, broken teeth, which conducted at the experimental stage. And analyzing gear tooth wear, gear cracks and broken teeth three failure respectively. Gear fault characteristic frequency and its modulation sidebands characteristic are obtained by comprehensively comparing of the vibration signal’s time-domain waveform, amplitude spectrum, Hilbert spectrum, Hilbert marginal spectrum, etc. And accomplish the gearbox fault diagnosis successfully. It is demonstrated that the Hilbert-Huang transform method is very suitable for gearbox fault diagnosis.