Fault detection of rolling bearings through vibration analysis via the hybrid CEEMD-EMD approach

This paper presents a composited method named as CEEMD-EMD (complementary EEMD and EMD) to analyze rolling bearings´ defective signals. Through CEEMD-EMD method, some interested intrinsic mode functions (IMFs) were collected, and then calculated envelope spectrum was used to identify the defects in the rolling bearings. The advantages of CEEMD-EMD method over EMD and CEEMD methods were obtained by simulation estimation. The calculation showed CEEMD-EMD method not only overcomes the disadvantage of mode mixing problem, but also removes residual noise and solves mode splitting problem. Then CEEMD-EMD approach was used to analyze vibration signals of rolling bearings. The analysis results demonstrated CEEMD-EMD approach combined with envelope spectrum analysis is an ideal and valid way to extract fault features in the testing signals of rolling bearings.