Composite Multiscale Fractional Fuzzy Diversity Entropy and Its Application in Bearing Fault Identification

This paper presents an intelligent fault identification approach integrating composite multiscale fractional fuzzy diversity entropy (CMFFDE) for feature extraction, joint mutual information (JMI) for feature selection, and an extreme learning machine (ELM) for classification. First, the CMFFDE method is developed by incorporating composite multiscale analysis into the proposed fractional fuzzy diversity entropy to extract multiscale fault characteristics. JMI feature selection is then applied to identify sensitive features, which are subsequently used as input to the ELM classifier for fault identification. The effectiveness and superiority of the proposed approach are verified using bearing experimental data. Analysis results demonstrate that the proposed approach achieves better identification performance in bearing vibration signal analysis than alternative methods.