A New Strategy for Disc Cutter Wear Status Perception Using Vibration Detection and Machine Learning

Area of Science:

• Engineering
• Artificial Intelligence
• Geotechnical Engineering

Background:

• Real-time monitoring of Tunnel Boring Machine (TBM) cutter health is vital for operational efficiency and minimizing downtime.
• Vibration signals generated during the dynamic contact between disc cutters and rock contain valuable information about cutter wear status.
• Existing methods face challenges in signal acquisition and transmission due to high cutting forces and complex operating conditions during TBM operation.

Purpose of the Study:

• To develop and validate a novel method for diagnosing abnormal wear states in TBM disc cutters.
• To leverage brain-like artificial intelligence for processing and analyzing cutter-rock interaction vibration signals.
• To identify different wear states, including normal, wear failure, and angled wear failure, of disc cutters.

Main Methods:

• Proposed a brain-like artificial intelligence approach for analyzing vibration signals from disc cutter-rock interactions.
• Utilized low-thrust operational conditions to ensure smoother cutter operation and minimize environmental interference for signal acquisition.
• Employed frequency-domain characteristics of periodic vibration waveforms during cutter-granite contact.
• Applied artificial neural networks for the classification of cutter wear states.

Main Results:

• The proposed method successfully diagnosed abnormal wear states in disc cutters.
• The artificial neural network achieved a diagnosis accuracy rate of 90% for identifying normal, wear failure, and angled wear failure states.
• The study demonstrated the feasibility of using vibration signal analysis under specific conditions for cutter wear diagnosis.

Conclusions:

• The brain-like artificial intelligence method offers a promising approach for real-time monitoring and fault diagnosis of TBM disc cutter wear.
• Optimized signal acquisition conditions (low thrust) enhance the reliability of vibration-based wear diagnosis.
• Accurate identification of cutter wear states can significantly improve TBM equipment health management and reduce operational losses.