Gearbox Fault Detection using Wavelet Transform and ANFIS

This paper studies the signal-based approaches to induction gearbox fault detection. A method based on wavelet transform and ANFIS was used to describe the gearbox faults. System faults such as clearance failures or broken gears are investigated. The identification of a single parameter, modeling the gearbox effect has allowed proper detection within a classical analytical redundancy scheme. An experimental benchmark has allowed a practical outline of the methods.