Planetary gearbox fault diagnosis via rotary encoder signal analysis

Abstract

Planetary gearbox fault diagnosis via vibration signal analysis is highly challenging, because of vibration spectral complexity due to gear fault modulation and especially time-varying transmission path effects. Rotary encoder signals contain gear fault signature, and are free of extra amplitude modulation caused by time-varying transmission path effects. Hence, encoder signal analysis provides a potentially effective approach to overcome difficulties in planetary gearbox fault diagnosis. Gear faults induce jitters in instantaneous angular speed, thus leading to changes in pulse interval of encoder signals. We proposed to analyze the order spectrum of repetitive temporal interval reciprocal between adjacent encoder pulses in the angle domain. Our proposed method is free from the algorithmic complexity of instantaneous angular speed or acceleration estimation, and more importantly addresses the difficult issue of planetary gearbox fault diagnosis under time-varying speeds. It is validated through lab experiments of a planetary gearbox under both constant and time-varying speed conditions. Localized faults on the sun, planet and ring gears are all successfully diagnosed.