A weighted multi-scale morphological gradient filter for rolling element bearing fault detection.

Area of Science:

• Mechanical Engineering
• Signal Processing
• Condition Monitoring

Background:

• Rolling element bearings are critical components in rotating machinery.
• Early fault detection is essential to prevent catastrophic failures and minimize downtime.
• Traditional methods often struggle with noisy vibration signals.

Purpose of the Study:

• To introduce a novel signal processing scheme, the weighted multi-scale morphological gradient filter (WMMG).
• To evaluate the WMMG's effectiveness in detecting faults in rolling element bearings.
• To compare WMMG performance against established techniques like envelope analysis (EA) and wavelet transform-envelope analysis (WT-EA).

Main Methods:

• Development of the weighted multi-scale morphological gradient filter (WMMG).
• Testing WMMG using simulated and experimental vibration signals from a bearing test rig.
• Comparative analysis with traditional envelope analysis and a WT-EA algorithm.
• Investigation of fault classification performance under varying noise levels.

Main Results:

• WMMG effectively suppresses noise at larger scales while preserving impulsive fault details at smaller scales.
• Experimental results confirm WMMG's capability to extract impulsive components from noisy vibration data.
• WMMG demonstrated equal or superior performance to EA and WT-EA in identifying bearing faults.
• The proposed method achieved high classification accuracy across different noise intensities.

Conclusions:

• The WMMG filter is a highly effective tool for rolling element bearing fault detection.
• Its ability to handle noise makes it suitable for real-world applications.
• Low computational cost makes WMMG ideal for on-line condition monitoring systems.