Multiresolution Fourier transform of ripple voltage and current signals for fault detection in a gearbox

This paper deals with detecting defects in a multistage gearbox. The gearbox is driven by an induction motor and the output shaft of the gearbox is connected to a DC generator and a DC tacho-generator. The ripple current and voltage signals supplied by these generators are analyzed for a defect-free gearbox and a defective gearbox with two teeth missing in a particular gear. Multiresolution Fourier transform is applied to these signals in order to improve the signal to noise ratio as well to highlight a specific bandwidth of the signal containing all the gear mesh frequencies. It is found that application of MFT improves the dynamic range of the signal, and hence highlights the frequencies with negligible amplitude. The technique is observed to be effective in distinguishing the defective and the defect-free cases of the gearbox.