Localized fault detection of sun gears based on windowed synchronous averaging in the angular domain

Abstract

The vibration transmission paths between the fixed-mounted accelerometer and the meshing points of the sun gear's local fault gear tooth is time varying in a planetary gearbox. Consequently, the conventional time synchronous averaging used in vibration-based fault detection of fixed-axis gearboxes is invalid for the localized fault detection of sun gears. To address the noise reduction, the windowed synchronous averaging originally introduced by McFadden can be employed to extract the useful components and reduce the effect of transmission paths. However, the speed fluctuation makes the above method inapplicable in real-world applications. In order to improve the robustness of windowed synchronous averaging under the speed fluctuation condition, a non-separation version of the windowed synchronous averaging in the angular domain is proposed in this article to address the speed fluctuation. In this scheme, the windowed time-domain signal is resampled at constant angular increments by the computed order tracking, which converts the non-stationary signal in the time domain into the stationary one in the angular domain. Then, the windowed angular-domain signal is processed by the synchronous averaging. In this way, the frequency blur and waveform distortion caused by rotating speed fluctuations can be eliminated. Experiments are conducted to verify the validity of the proposed method.