Detection and Localization of Excavation-Disturbance-Related Near-Field Microseismic Events During TBM Tunneling

Area of Science:

• Geophysics
• Civil Engineering
• Signal Processing

Background:

• Tunnel boring machine (TBM) excavation generates continuous vibrations that mask subtle seismic signals.
• Detecting structural plane disturbances during TBM operations is crucial for safety and efficiency.

Purpose of the Study:

• To develop and validate a novel framework for detecting and localizing structural plane activation events during TBM excavation.
• To overcome challenges posed by TBM-induced mechanical vibrations in seismic monitoring.

Main Methods:

• Developed a Signal-Constrained Activation Detection and Localization (SCADL) framework using geophone-array data.
• Implemented adaptive multi-station consistency triggers, coherence, and polarization constraints to filter noise.
• Utilized fused evidence for first arrival picking and an ahead-of-face velocity model for event localization.
• Integrated multiple factors into a multi-evidence activation index for high-confidence event identification.

Main Results:

• SCADL successfully identified 263 valid propagating events from a 16-hour monitoring record.
• Achieved a high event-level F1-score of 0.88, with reduced arrival-time picking error (2.4 ms) and localization residual (2.9 ms).
• Identified 86 high-confidence events clustered in zones consistent with post-excavation geological validation.

Conclusions:

• The SCADL framework demonstrates feasibility for effective single-section TBM monitoring.
• SCADL significantly enhances the detection and localization accuracy of structural plane activation events amidst TBM vibrations.
• This method provides a reliable tool for real-time assessment of geological conditions during tunneling.