A Novel Method of Bearing Fault Diagnosis Based on FDS-MOMEDA and IMCKDA
Gangjin Huang, Jing Yang, Yuhao Zhang, Jiayu Ou, Zicheng Wei · 2024
Rolling bearings play an important role in ensuring the normal operation of machines, so the importance of bearing fault diagnosis is increasingly being emphasized. However, due to severe noise interference in the measurement signal, traditional methods are difficult to extract bearings features. Therefore, this study proposes a new method that combines the adjusted multipoint optimal minimum entropy deconvolution (FDS-MOMEDA) and the improved maximum correlation kurtosis deconvolution adjusted (IMCKDA) to distinguish the fault type of the bearings. Firstly, the improved MOMEDA method, FDS-MOMEDA, is used to denoise the original signal. Secondly, IMCKDA is applied to the denoised signal to enhance the pulse pattern of bearing faults, thereby extracting effective fault feature information. Finally, the fault type of the bearing is determined through the Hilbert envelope spectrum. By processing vibration signals from real experimental datasets, the results show that this method can effectively extract fault features of rolling bearings.