Morphological Convolution Undecimated Wavelet: A Novel Frequency Demodulation Analysis Method for Bearing Fault Diagnosis

The morphological undecimated wavelet (MUW) is an efficient feature extraction algorithm for bearing fault diagnosis. Currently, the researched MUW is mainly focused on background noise cancellation and transient impulse extraction but has not been exploited for frequency demodulation. This article proposed a novel frequency demodulation analysis method named morphological convolution undecimated wavelet (MCUW) for bearing fault diagnosis. First, the morphological difference convolution operator (MDCO) is developed, which fully employs the transient impulse enhancement properties of the difference operator and the combined morphological filter-hat transform (CMFHT) operator, as well as the random noise cancellation performance of the convolution operator. Subsequently, the MDCO is utilized into the MUW to produce the MCUW for frequency demodulation to extract the bearing frequency features. Finally, the simulated scenarios and experiment signals are implemented to evaluate the MCUW performance. The analysis results illustrate that the MCUW can efficiently extract fault features and its performance is better than other well-advanced algorithms.