Optimal weighted envelope spectrum with informative multi-band selection for bearing fault diagnosis

Area of Science:

• Mechanical Engineering
• Vibration Analysis
• Signal Processing

Background:

• Spectral coherence theory is vital for bearing fault diagnosis.
• Existing methods struggle to integrate fault information across multiple spectral bands, limiting diagnostic accuracy.

Purpose of the Study:

• To propose an optimal weighted envelope spectrum (OWES) method for enhanced bearing fault diagnosis.
• To address limitations in identifying and integrating fault information distributed across multiple spectral bands.

Main Methods:

• Constructed expected signal to expected noise (ESEN) with local background correction and adaptive harmonic identification.
• Employed combinatorial optimization to select informative spectral bands for feature integration.
• Developed OWES by weighting and integrating selected informative spectral bands.

Main Results:

• The proposed OWES method demonstrated clearer identification of fault characteristic frequencies and harmonics.
• ESEN enhanced robustness against frequency deviation, amplitude variation, and background noise.
• Quantitative comparisons confirmed OWES superiority over advanced methods using ESEN and kurtosis.

Conclusions:

• OWES effectively extracts and integrates fault features from multi-band spectral data.
• The method significantly improves bearing fault diagnosis performance, especially for distributed fault information.
• OWES offers a robust and accurate solution for bearing condition monitoring.