Fault Diagnosis Method for Rolling Bearings Based on Composite Multiscale Fluctuation Dispersion Entropy.

Area of Science:

• Mechanical Engineering
• Data Science
• Signal Processing

Background:

• Time series complexity analysis is crucial for diagnosing mechanical faults.
• Existing entropy estimation methods can lack stability.
• Rolling bearing fault diagnosis requires robust feature extraction and classification.

Purpose of the Study:

• To introduce composite multiscale fluctuation dispersion entropy (CMFDE) for stable time series complexity measurement.
• To develop an effective rolling bearing fault diagnosis method using CMFDE, mRMR, and kNN.
• To validate the proposed method's performance on experimental data.

Main Methods:

• Composite Multiscale Fluctuation Dispersion Entropy (CMFDE) was developed to measure time series complexity across multiple scales.
• Minimum Redundancy Maximum Relevancy (mRMR) was employed for sensitive fault feature selection.
• K-Nearest Neighbors (kNN) classifier was utilized for rolling bearing condition recognition.

Main Results:

• CMFDE demonstrated improved stability in entropy estimation through simulations.
• The CMFDE-mRMR-kNN method effectively extracted fault characteristics and selected sensitive features.
• Experimental validation confirmed the high accuracy of the proposed fault diagnosis approach.

Conclusions:

• CMFDE offers a stable and effective measure for time series complexity.
• The CMFDE-mRMR-kNN framework provides a robust solution for rolling bearing fault diagnosis.
• The proposed method shows significant potential for industrial applications in condition monitoring.