A fault feature extraction algorithm based on CEEMD-TVD-MOMEDA

Abstract

Rolling bearings are indispensable key components in mechanical equipment, and they are also one of the most easily damaged components. To solve the problem of bearing fault feature extraction under strong noise interference, a combination of complementary ensemble average empirical mode decomposition (CEEMD), total variation denoising (TVD), and multipoint optimal minimum entropy deconvolution adjusted(MOMEDA) is proposed. Firstly, decompose the vibration signal into several signal components. Secondly, the qualified IMF signal components are selected by combining with the cross-correlation analysis criteria for reconstruction, and TVD is used to reduce the noise of the signal. Thirdly, MOMEDA is used to filter the denoised signal, so as to enhance the periodic impact component. Finally, the envelope spectrum of the filtered signal is analyzed. The effectiveness of the proposed method is verified by the simulation signals and the bearing fault data set of Case Western Reserve University. The experimental results show that the proposed method can not only reduce the noise interference but also effectively extract and identify the rolling bearing fault features. Compared with the results obtained by traditional LMD and ITD methods, it has a better recognition effect.

Keywords

• complementary ensemble average empirical mode decomposition (ceemd)
• total variation denoising (tvd)
• multipoint optimal minimum entropy deconvolution adjusted
• fault feature extraction