Adaptive VMD-K-SVD-Based Rolling Bearing Fault Signal Enhancement Study

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Rolling bearings are critical components in rotating machinery.
• Signal analysis for fault detection is challenged by noise, non-stationarity, and nonlinearity.
• Effective denoising is crucial for accurate fault feature extraction.

Purpose of the Study:

• To develop a novel combined approach for denoising and enhancing rolling bearing signals.
• To address challenges in fault feature extraction from noisy and complex signals.
• To improve the reliability of condition monitoring systems for rolling bearings.

Main Methods:

• The proposed method integrates Variational Mode Decomposition (VMD) and K-Singular Value Decomposition (K-SVD).
• An Improved Arithmetic Optimization Algorithm (IAOA) optimizes VMD parameters (K, α) for signal decomposition.
• K-SVD is applied to the resulting intrinsic mode functions (IMFs) for further noise reduction.

Main Results:

• The combined VMD-K-SVD approach effectively reduces noise in rolling bearing signals.
• The method enhances the clarity of fault feature signals for better analysis.
• Validation using CWRU and thrust bearing test rig data confirms the approach's efficacy.

Conclusions:

• The novel VMD-K-SVD method offers a robust solution for denoising rolling bearing signals.
• This technique significantly improves the extraction of fault features from complex operational data.
• The proposed approach enhances the performance of condition monitoring and diagnostic systems.