The application of a new deconvolution method to railway bearings fault diagnosis
A new deconvolution method for the fault diagnosis of railway bearings is proposed in this paper. The proposed method aims to deconvolve the impulse-like signal that caused by rolling element bearing fault from the measured vibration signal. As an improved novel deconvolution technique, the new deconvolution method is designed to select an optimal finite impulse response filter to maximize the kurtosis-to-natural logarithm of Shannon entropy ratio of the filtered signal. The optimal FIR filter of the deconvolution technique is solved by an iterative algorithm. The effectiveness of the proposed new deconvolution method is verified by the simulated bearing fault signal as well as the experimental acceleration signal of the high speed trains’ bearings on service.