Rolling Bearing Fault Diagnosis Based on Optimal Notch Filter and Enhanced Singular Value Decomposition

Area of Science:

• Mechanical Engineering
• Vibration Analysis
• Condition Monitoring

Background:

• Early-stage rolling bearing fault signals are often obscured by fundamental frequency signals and background noise.
• Accurate detection of these weak fault signatures is crucial for preventing catastrophic failures and ensuring machinery reliability.

Purpose of the Study:

• To develop a robust method for diagnosing weak faults in rolling bearings.
• To overcome the limitations of existing methods in isolating fault signals from interference.

Main Methods:

• Utilized an optimal notch filter to eliminate the fundamental frequency signal from raw vibration data.
• Employed Teager energy entropy to adaptively determine the optimal filter bandwidth.
• Applied an enhanced singular value decomposition (SVD) technique for signal denoising.
• Conducted envelope spectrum analysis on the denoised signal to extract fault characteristic frequencies.

Main Results:

• The proposed method successfully isolated weak fault feature signals from noise and fundamental frequency interference.
• Simulation and experimental results validated the effectiveness of the combined optimal notch filter and enhanced SVD approach.
• Comparative analysis demonstrated superior performance over Minimum Entropy Deconvolution, Kurtogram, and Infogram methods.

Conclusions:

• The presented rolling bearing weak fault diagnosis method offers enhanced accuracy and reliability.
• The combination of adaptive filtering and advanced SVD denoising is effective for early fault detection in rotating machinery.