Gearbox fault diagnosis via generalized velocity synchronous Fourier transform and order analysis

Abstract

Gearboxes have essential roles in many types of industrial equipment. Fault detection for gearboxes is important yet extremely difficult because volatile working conditions lead to nonstationary vibration signals. Order tracking is a classic and effective technique for nonstationary vibration analysis and fault diagnosis of rotating machinery. Many order tracking methods that do not require a tachometer have been proposed, such as methods based on re-sampling. However, most are complex and often introduce interpolation errors. To avoid such difficulties, a simple yet effective method is proposed in this paper. This method employs the generalized demodulation approach to extract a component with a frequency proportional to the instantaneous shaft rotational frequency from the vibration signal. Then, demodulating the extracted component recovers the instantaneous shaft rotational phase. With such information the order spectrum can be directly obtained via a velocity synchronous discrete Fourier transform. Finally, the fault can be diagnosed by order spectrum analysis. The effectiveness of this method is validated with both simulated and lab experimental vibration signals of a gearbox under time-varying rotational speed conditions.