An enhanced rolling bearing fault detection method combining sparse code shrinkage denoising with fast spectral...

Area of Science:

• Mechanical Engineering
• Signal Processing

Background:

• Rolling bearings are critical components in rotating machinery, but are susceptible to failures.
• Early and accurate fault detection in rolling bearings is essential for preventing catastrophic failures and ensuring operational reliability.
• Extracting weak fault impulses from strong background noise in vibration signals poses a significant challenge in bearing fault diagnosis.

Purpose of the Study:

• To develop an enhanced fault detection method for rolling bearings that accurately identifies weak impulses buried in noise.
• To improve the speed and accuracy of rolling bearing fault diagnosis.

Main Methods:

• A novel fault detection approach combining sparse code shrinkage denoising with Fast Spectral Correlation (Fast-SC).
• Utilizing the non-Gaussian statistical properties of defective bearing vibration signals for denoising.
• Applying Fast-SC to the denoised signal to obtain cyclic spectral correlation.
• Introducing the squared enhanced envelope spectrum (SEES) for fault detection and identification.

Main Results:

• The proposed method successfully highlights periodic impulses by effectively denoising the original noisy signals.
• The Fast-SC algorithm processed the denoised signal to extract cyclic spectral correlation.
• The squared enhanced envelope spectrum (SEES) demonstrated superior performance in detecting and identifying rolling bearing faults.
• Experimental results validated the method's effectiveness and superiority over existing techniques like Fast-SC, spectral kurtosis, and Infogram.

Conclusions:

• The integrated approach of sparse code shrinkage denoising and Fast Spectral Correlation provides a robust solution for rolling bearing fault diagnosis.
• The proposed method significantly enhances the ability to detect weak fault impulses in noisy vibration signals.
• This technique offers a superior alternative for accurate and efficient rolling bearing fault detection in rotating machinery.