Gear fault classification using a health index: a comparison of approaches

摘要

Gear fault classification is gaining importance within the use of gearboxes in industry. In this paper, a health index (HI) has been developed using two approaches, namely frequency domain feature-based and time domain feature-based, which quantitatively describes the gear fault condition.The vibration signals for three gear conditions, namely good, half tooth removed and full tooth removed, are acquired from a Maruti 800 (Indian passenger car)-based gearbox test-rig under variable speed and load conditions and used for health index development. A frequency spectrum-based healthindex has been developed using the wavelet coefficients of denoised vibration signals, which is a frequency domain feature-based health index. Statistical features extracted from wavelet coefficients are used to develop a time domain feature-based health index. A comparison of these two approachesis carried out to propose an effective health index for gear fault classification.