Mine gearbox fault diagnosis based on neighboring coefficients of translation-invariant multiwavelets

As the mine gearbox works in a poor working environment,the fault signal is often hidden in background noise when gearbox fault occurs. Thus,the fault feature is very difficult to be extracted through frequency spectrum analysis. Translation-invariant multiwavelets can avoid Gibbs phenomena. Neighboring coefficient denoising considers the relativity of coefficients and overcomes the deficiency of traditional threshold denoising. Translation-invariant multiwavelets denoising using neighboring coefficients is applied to noisy impact simulation signal and extract impact features hidden in the noise. Then the method is applied to mine elevator gearbox. The diagnosis results show that this method can effectively extract the gearbox fault feature frequency,provide an accurate basis for fault diagnosis. Simulation and engineering application verify the effectiveness of translation-invariant multiwavelets denoising using neighboring coefficients in fault diagnosis.