Application of Vehicle Scanning Method to Jointed Shield Tunnels: Damage Identification Using an Enhanced Driving Component Extraction Strategy

The vehicle scanning method (VSM) enables indirect structural damage detection using onboard sensors, but its application to shield tunnels remains challenging due to the high stiffness-to-mass ratio of tunnels and the presence of segmental joints. This study investigates whether segmental joints produce false-positive damage indications and evaluates a modified signal extraction strategy. A refined vehicle–tunnel coupled finite element model incorporating shear and rotational springs at segment interfaces is established. Instead of applying Variational Mode Decomposition (VMD) before filtering, the proposed strategy applies low-pass filtering prior to VMD to extract a purer driving component (DC). The DC is quantitatively compared with the previously identified optimal component (D2) in terms of energy amplitude, background suppression, and noise robustness. Results show that (1) segmental joints produce no spurious wavelet energy peaks above the detection threshold in undamaged tunnels; (2) the DC substantially outperforms D2, achieving a signal-to-background ratio of 83.1 (vs. 18.8 for D2) and a 12% higher detection rate for 5% damage under SNR = 10 dB noise; and (3) liner and support damage produce single sharp wavelet energy peaks, while joint damage produces a broadened peak spanning two adjacent elements, suggesting a potential signature for distinguishing joint-related damage from element-related damage in the cases examined. These findings suggest that the method may be applicable to jointed shield tunnels within the range of conditions considered in this numerical study, without requiring joint-specific calibration. The modified DC extraction offers a simpler and more sensitive damage indicator for practical inspection.