Gearbox Fault Detection and Diagnosis Based on EEMD De-noising and Power Spectrum*

The vibration signal of fault gear is related to non-stationary property, and the gear fault diagnosis will be seriously interfered by bearing signal and other noise signal. EEMD can adaptive decomposition of the vibration signal into different frequency components, but the precise extraction of useful signal is still a question. In this paper, a double determination criterion is presented to the EEMD method to reduce noise, and combined with the power spectrum of time-frequency analysis to diagnosis the fault of gearbox. The experimental results show that EEMD combined with Power spectrum can effectively extract the characteristic frequency from strong noise signals and diagnose the fault of gear.