Application of adaptive multiwavelets via lifting scheme in bevel gear fault diagnosis

Different from parallel-axes gears, the bevel gears can change the direction of power transmission. So they are widely used in transmission systems, and their fault diagnosis is of great significance. However, the feature extraction for the faulty spiral bevel gears is quite difficult because of its large overlap ratio as well as amplitude and frequency modulation (AMFM) nature. Therefore, it is urgent to develop an effective feature extraction method for this task. Multiwavelet with multiple wavelet basis functions and many excellent properties provides an effective tool to rotating machinery fault diagnosis. In this paper, firstly, we construct an adaptive multiwavelet via symmetric lifting scheme, and then decompose the original signal; secondly, the signal is reconstructed with chosen sensitive feature bands; thirdly, the reconstructed signal is demodulated to extract the characteristic frequency based on Hilbert transform. The spiral bevel gear breakage and scrape faults simulated on the test bench are taken as examples to verify the effectiveness and reliability of the proposed method.