A new approach to adaptive VMD based on SSA for rolling bearing fault feature extraction

Abstract

Due to the structure of rolling bearings, will have various problems. So the early detection of rolling bearing faults is very important. Consequently, a precise method for extracting fault features is required. In this study, an adaptive variational modal decomposition (VMD) fault feature extraction method is proposed, utilizing the sparrow search algorithm (SSA). Firstly, a novel measurement index called impulse diversity entropy (IDE) is introduced, which better represents internal changes within the mode components. Secondly, the SSA is employed to select the optimal VMD decomposition parameters based on the IDE. Finally, a spectrum analysis is conducted on the mode component with the highest IDE to extract fault features. The experimental results show that this method has an accurate feature extraction ability and obvious advantages over other methods in distinguishing fault and interference frequencies because it is a special signal decomposition method.

Maximum Gpq–mean deconvolution for the impulsive fault feature enhancement of rolling bearing

“…The probability of success, denoted by p, ranges from 0 to 1. The probability mass function of the Bernoulli distribution is shown in equation (10), where k represents the trial outcome, with 1 indicating success and 0 indicating failure…”

“…The Bernoulli distribution expression shown in equation (10) generates a set of data consisting solely of 0 and 1. The sparsity of the distribution can be adjusted by controlling the probability of occurrence of 0.…”

“…Jiang et al [9] proposed an adaptive and efficient VMD method, drawing from a comprehensive analysis of its convergence properties. Gao et al [10] introduced a new measure called IDE using SSA to improve VMD in enhancing its feature extraction capability for improved accuracy, showcasing evident advantages.…”