Condition monitoring and fault diagnosis based on multipoint optimal minimum entropy deconvolution adjusted technique

Planetary gearbox (PG) exhibits unique dynamic behaviour that imposes great challenges in gear fault diagnosis. In particular, multiple and time-varying vibration transmission paths from the gear meshing point to the sensor, usually mounted on the PG housing, cause not only additional spectral components in the signal but also strong noise. Thus, the influence of the transmission paths and multiple vibration sources make fault indications hard to distinguish. This paper presents a new approach for fault diagnosis of PG based on Multipoint Optimal Minimum Entropy Deconvolution Adjusted (MOMEDA). MOMEDA has been demonstrated effective to suppress the path dissertation for linear time-invariant (LTI) system. However, its performance has not been examined with the case of a time-variant system such as PG vibration system. Therefore, an experimental evaluation is carried out to evaluate and optimise MOMEDA analysis for minimising the path influnces and enhancing periodic fault impulses generated by the faulty gear. A set of experimental data acquired from the PG with seeded with common faults on the planet gear and sun gear. The results obtained by the optimised filter length show that the MOMEDA has the expected capability and allows the seeded faults to be diagnostic successfully under different loads, confirming the generality of the approach.