A signal selection scheme through order tracking techniques for planetary gearbox condition monitoring

Planetary gearbox is an important rotating machinery, and it always operates under non-stationary working conditions, thus its vibration signal exhibits complex time-varying modulation nature and therefore yields complicated spectral structures. Traditional signal processing methods, such as Fourier analysis, as a result face challenges for a proper fault diagnostic. To achieve a more effective fault diagnosis, this paper proposes a signal selection scheme which is based on order tracking techniques. The basic procedures of the proposed scheme are as follow: Firstly, Computed Order Tracking is applied to reveal order contents and identify the order(s) of interest. Secondly, Vold-Kalman filter order tracking is used to extract the order(s) of interest -These filtered order(s) constitute the so called selected vibration signal. Experimental studies on planetary gearbox demonstrate the effectiveness of the proposed method for planetary gearbox fault diagnosis.