Abstract
As one of the important parts of a mechanical transmission system, a rolling bearing often has multiple faults coexisting, and the mutual coupling between multiple faults poses a challenge for accurate diagnosis of rolling bearings. Aiming at the above problems, this paper proposes a weighted Morlet wavelet-overlapping group sparse (WOGS) algorithm for the multiple fault diagnosis of rolling bearings. On the basis of the overlapping feature of Morlet wavelet transform coefficients, a WOGS optimization model was initially constructed. Thereafter, the weight coefficients in the model were constructed by analyzing the impulse features of the signal. Thus, majorization-minimization was used to solve the optimization problem. A case study on weak multiple fault diagnosis of rolling bearings was performed to validate the effectiveness of the WOGS algorithm. Quantitative indexes are used to further discuss the extraction accuracies of different algorithms, and the results show that the proposed algorithm exhibits better performance than other algorithms.
Highlights
Rolling bearings act as tiny transmission components in a complex mechanical system
The results show that wavelet-overlapping group sparse (WOGS) algorithm has the largest energy ratio, which indicates that it has a stronger ability to extract weak multiple fault features
This paper presents a weak multiple fault detection algorithm for rolling bearings on the basis of WOGS
Summary
Rolling bearings act as tiny transmission components in a complex mechanical system. If a rolling bearing fails, the overall failure rate of a complex system will increase due to its scale effect, which will cause significant economic losses or serious safety accidents [1]. Zhang et al [16] proposed a method on the basis of resonance sparse decomposition and comb filter for gearbox multiple fault diagnosis. Du et al [17] proposed a sparse feature recognition method on the basis of the union of redundant dictionary for multiple fault diagnosis with different morphological waveforms. Obtaining the desired effect by using the difference in fault features to separate the multiple faults of rolling bearings is difficult. On the basis of the overlapping group shrinkage and majorization-minimization (MM), He et al [33,34] extracted the periodic group sparse signal from the vibration signal and realized the compound fault diagnosis of rolling bearings.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
