A Review on Rolling Bearing Fault Signal Detection Methods Based on Different Sensors

Area of Science:

• Mechanical Engineering
• Condition Monitoring
• Fault Diagnosis

Background:

• Rolling bearings are critical precision components essential for reducing friction and ensuring high mechanical efficiency in various applications.
• Their widespread use necessitates effective methods for detecting faults to prevent operational failures and maintain performance.

Purpose of the Study:

• To review and categorize existing research on rolling bearing fault detection.
• To analyze different fault signal types, sensor configurations, and diagnostic methodologies.
• To compare the effectiveness and characteristics of various fault detection approaches.

Main Methods:

• Classification of fault signal types into one-dimensional (40.14%) and two-dimensional images (31.69%).
• Analysis of detection methods based on sensor type (internal/external) and signal dimensionality.
• Comparison of detection rates, ranging from over 70% to 100% across different techniques.

Main Results:

• One-dimensional signal detection with internal sensors is prevalent (36.62%).
• Detection rates vary significantly, with some methods achieving 100% accuracy.
• Emerging artificial intelligence (AI) technologies like artificial neural networks and machine learning are rapidly advancing the field.

Conclusions:

• New research methods, particularly AI-driven techniques, are accelerating rolling bearing fault detection and diagnosis.
• These advanced methods offer improved robustness and superiority in identifying faults and their causes.
• Continued development in AI promises further enhancements in bearing condition monitoring and predictive maintenance.