Fault diagnosis of gearbox based on EEMD and HMM

Abstract

As a complicated mechanical component, gearbox plays a significant role in industrial field. Its fault diagnosis benefits decision making of maintenance and avoids undesired downtime cost. Empirical mode decomposition (EMD) is a self-adaptive signal processing method, which has been applied in non linear and non stationary signal processing successfully. However, the EMD algorithm has its inherent drawbacks. Aiming at the problem of intrinsic mode function (IMF) criterion in the EMD, this paper introduces the mode mixing problem of EMD in Hilbert-Huang Transform (HHT). In order to overcome the mode mixing problem in EMD, ensemble empirical mode decomposition (EEMD) is used. Therefore, this paper proposes a new method based on EEMD and hidden Markov mode (HMM) for gear fault diagnosis. First, a simulation signal is used to verify the advantages of EEMD comparing to EMD. Second, the new method is applied to the gear fault diagnosis. There are two patterns seeded faults in the experiment. One pattern is broken teeth, the other is cracks. The results show that the method can identify gear fault accurately and effectively.