Intelligent Diagnostics of Radial Internal Clearance in Ball Bearings with Machine Learning Methods

Area of Science:

• Mechanical Engineering
• Vibration Analysis
• Machine Learning

Background:

• Radial internal clearance in rolling bearings is challenging to detect via dynamic analysis.
• Existing methods for bearing clearance classification lack sufficient accuracy.
• Intelligent detection of bearing clearance is crucial for predictive maintenance.

Purpose of the Study:

• To develop an intelligent method for classifying rolling bearing radial internal clearance.
• To improve the accuracy of bearing clearance detection using advanced signal processing.
• To evaluate the effectiveness of Variational Mode Decomposition (VMD) in bearing diagnostics.

Main Methods:

• Analysis of short-time vibration data from rolling bearings.
• Calculation of statistical indicators for feature extraction.
• Application of machine learning models for classification.
• Utilizing Variational Mode Decomposition (VMD) for enhanced time series processing.

Main Results:

• Initial classification accuracy using statistical indicators was unsatisfactory.
• VMD successfully extracted dominant frequencies into experimental modes.
• Applying statistical indicators to VMD modes increased classification accuracy to over 90%.
• The study demonstrated the effectiveness of VMD in classifying bearing internal clearance.

Conclusions:

• VMD combined with statistical indicators offers a highly accurate method for classifying rolling bearing radial internal clearance.
• This approach enhances the diagnostic capabilities for rolling element bearings.
• The findings contribute to more reliable condition monitoring and predictive maintenance strategies.