Automated monitoring and verification of structural deformation in existing subway tunnels under large-section underground excavation

To ensure operational safety of existing tunnels during large-section underground excavation, this paper proposes an automated deformation signal monitoring method. The method integrates three-layer wavelet decomposition with threshold-based reconstruction for denoising, applies scale factors to determine deformation intervals, employs maximum likelihood estimation to detect anomalies, and introduces a disturbance coefficient to refine convergence results. Engineering case studies demonstrate that this method outperforms conventional approaches: abnormal intervals were reduced by 10%, denoised RMSE remained stable at ≈0.002, and the monitoring process exhibited consistent performance across time periods with higher signal-to-noise ratios. These results confirm that the proposed method is efficient, accurate, and stable in construction environments, providing a feasible solution for the safety monitoring of subway tunnels.