Vibration signal analysis methods for planetary gearbox fault diagnostics
Abstract
Planetary gearboxes are widely used in many kinds of machinery for their large power transmission capacity in a compact structure. Their vibration signals are more complicated than those of fixed shaft gear systems, due to the complex structure and gear kinematics, and their vibration signals are totally different from those of fixed shaft gear systems in terms of spectral characteristics. We proposed a signal model to describe planetary gearbox vibrations, and summarized the spectral symptoms of both local and distributed faults of sun, planet and ring gears, as well as planet bearings. We also derived the explicit equations for calculating the characteristic frequency of all the three types of gears and all the components in planet bearings. To mitigate the complexity problem with the traditional spectral analysis caused by the complicated sideband structure, we proposed a joint amplitude and frequency demodulation analysis method. To address the vibration nonstationarity issue in real world applications, we proposed to analyze vibration signals in joint time-frequency domain, and extended the amplitude and frequency demodulation to time-varying demodulation analysis. Furthermore, we proposed envelope and frequency order spectrum analysis, to overcome the difficulty due to the time variability and the complexity of vibration frequency contents under nonstationary running conditions. We illustrated and validated the theoretical derivations and the proposed methods through lab experiments of industrial planetary gearboxes and in-situ wind farm measurements.
