A Dual-Optimization Fault Diagnosis Method for Rolling Bearings Based on Hierarchical Slope Entropy and SVM...

Area of Science:

• Engineering
• Signal Processing
• Machine Learning

Background:

• Slope entropy (SlopEn) is effective for fault diagnosis but struggles with threshold selection.
• Existing methods require robust threshold selection for accurate fault identification.

Purpose of the Study:

• Introduce hierarchical slope entropy (HSlopEn) to enhance fault diagnosis capabilities.
• Develop a dual-optimization fault diagnosis method using white shark optimizer (WSO) for HSlopEn and Support Vector Machine (SVM).

Main Methods:

• Proposed hierarchical slope entropy (HSlopEn) by incorporating hierarchy into SlopEn.
• Applied white shark optimizer (WSO) to optimize both HSlopEn and SVM parameters, creating WSO-HSlopEn and WSO-SVM.
• Tested the dual-optimization method on rolling bearing fault diagnosis using single- and multi-feature scenarios.

Main Results:

• The WSO-HSlopEn and WSO-SVM method demonstrated superior recognition rates compared to other hierarchical entropies.
• Under multi-feature conditions, recognition rates exceeded 97.5%, improving with more selected features.
• Achieved a maximum recognition rate of 100% when five nodes were selected.

Conclusions:

• The proposed WSO-HSlopEn and WSO-SVM method significantly enhances rolling bearing fault diagnosis accuracy.
• Feature selection positively impacts diagnostic performance, with optimal results achieved using five nodes.
• This approach offers a robust and high-performing solution for complex fault diagnosis tasks.