A Brief Review of Acoustic and Vibration Signal-Based Fault Detection for Belt Conveyor Idlers Using Machine Learning...

Area of Science:

• Engineering
• Industrial Monitoring
• Machine Learning Applications

Background:

• Belt conveyors are crucial for bulk material transport, with idlers being essential but challenging components to monitor.
• Idler defects can lead to system failures, necessitating effective fault detection (FD) methods.
• Existing monitoring often overlooks the large number of distributed idlers, making them prone to undetected issues.

Purpose of the Study:

• To provide a comprehensive review of acoustic and vibration signal-based fault detection for belt conveyor idlers.
• To consolidate current research on using machine learning (ML) models for idler defect detection.
• To identify key steps and challenges in developing and implementing ML-based FD for idlers.

Main Methods:

• Review of existing literature on fault detection for belt conveyor idlers.
• Analysis of approaches involving acoustic and vibration signal processing.
• Examination of machine learning model construction for defect identification.

Main Results:

• Identified ML models as a promising approach for detecting idler defects using vibration and acoustic data.
• Detailed the common stages in ML-based FD: data collection, signal processing, feature extraction, and model building.
• Highlighted the lack of comprehensive reviews in this specific domain.

Conclusions:

• Machine learning offers a viable solution for enhancing the reliability of belt conveyor systems through effective idler fault detection.
• Further research is needed to address open challenges and refine ML-based monitoring techniques.
• This review serves as a foundation for future advancements in industrial fault detection.