Gearbox Fault Diagnosis Based on Wavelet-AR Model

Abstract

A fault diagnosis technique for gearbox that uses a wavelet-AR model spectrum estimation method is presented in this paper. In the experimental work, the wavelet transform was used for original signal decomposition and de-noising to obtain fault signals, and the fault type was confirmed using AR model spectrum estimation method for gearbox fault signal diagnosis. The experimental results indicated that the proposed wavelet-AR model spectrum estimation method, compared with FFT algorithm, is effective in fault signal diagnosis for gearbox.