Rolling bearing fault diagnosis based on RCMDE and extreme learning machine

Aiming at the low recognition rate of rolling bearing fault signals, a fault diagnosis method based on refined composite multiscale dispersion entropy (RCMDE) and extreme learning machine (ELM) is proposed. Firstly, extracting 20 scales of refined composite multi-scale discrete entropy from the original vibration signal and constructing a fault feature set based on this, and then using ELM to identify fault types. The validity of the proposed method was verified by bearing data from Case Western Reserve University. Finally, the proposed method is compared with MPE-ELM. The comparison results show that the proposed fault diagnosis method has higher classification accuracy.