Bearing Fault Diagnosis Using Refined Composite Generalized Multiscale Dispersion Entropy-Based Skewness and Variance...

Area of Science:

• Mechanical Engineering
• Signal Processing
• Fault Diagnosis

Background:

• Bearing vibration signals exhibit nonlinearities affecting fault diagnosis.
• Traditional multiscale entropy algorithms can lose crucial complexity data.
• Generalized multiscale algorithms offer improved complexity measurement.

Purpose of the Study:

• To evaluate refined composite generalized multiscale dispersion entropy (RCGMDispEn) and refined composite multiscale dispersion entropy (RCMDispEn) for bearing fault diagnosis.
• To develop a multiclass FCM-ANFIS model for enhanced rotating machinery fault classification.
• To investigate the efficacy of generalized multiscale algorithms in conjunction with traditional multiscale approaches.

Main Methods:

• Utilized RCGMDispEn based on variance and skewness.
• Employed RCMDispEn for comparative analysis.
• Developed a multiclass fuzzy c-means adaptive network-based fuzzy inference system (FCM-ANFIS).

Main Results:

• Generalized multiscale algorithms, particularly those using variance and skewness, show promise for bearing fault diagnosis.
• The combined use of multiscale and generalized multiscale algorithms significantly improved fault classification across three real-world datasets.
• The FCM-ANFIS model enhanced the efficiency of rotating machinery fault classification.

Conclusions:

• Generalized multiscale algorithms, especially RCGMDispEn, are effective for bearing fault diagnosis.
• Synergistic application of multiscale and generalized multiscale entropy methods boosts diagnostic accuracy.
• The proposed approach offers a robust solution for rotating machinery condition monitoring.