Experimental study on the dynamic performance of tunnel damping measures considering initial support

Tunnels in high-intensity seismic regions are susceptible to severe damage from earthquakes, such as lining cracking, spalling, invert uplift, and progressive plastic deformation, which jeopardize structural safety and serviceability. Although seismic isolation layers have been widely applied to mitigate tunnel seismic responses, most studies have primarily focused on the interaction between the secondary lining and isolation layer, often overlooking the role of primary support, an essential component in tunnel construction. This study aims to elucidate the dynamic response characteristics and damage evolution of tunnel linings when seismic mitigation measures are implemented under realistic primary support conditions. By integrating primary support into the mitigation system, the study experimentally evaluates the effectiveness of a composite configuration comprising the secondary lining, seismic isolation layer, primary support, and surrounding rock, providing guidance for resilience-oriented tunnel seismic design.

The Jiedexiu No. 2 Tunnel on the Lhasa-Nyingchi Railway was selected as a representative case for a series of shaking-table model tests. A gravity-distorted similarity model was developed based on the Buckingham π theorem, featuring a geometric similarity ratio of 1:40, alongside appropriate scaling of elastic modulus, density, displacement, and acceleration. The surrounding rock and overburden were simulated using Grade Ⅴ phyllite and gravelly-breccia soils, respectively. The secondary lining was modeled with gypsum to simulate C30 concrete, whereas sponge rubber material was used as the seismic isolation layer. Basalt fiber-reinforced polymer anchors were employed to simulate rock bolts in the primary support. Horizontal excitation was applied using the El Centro earthquake wave with peak ground accelerations of 0.1 g, 0.2 g, 0.3 g, and 0.4 g. Acceleration sensors and strain gauges were arranged symmetrically to compare a conventional section with a mitigated section that incorporated the isolation layer and primary support. Dynamic responses were analyzed in the time and frequency domains, which included acceleration time histories, Fourier spectra, and acceleration response spectra. To quantify cumulative damage, a plastic deformation index (PDI), defined as the ratio of residual strain to peak dynamic strain, was introduced to classify damage evolution into elastic, elasto-plastic, and plastic stages.

Results show that the seismic isolation layer significantly reduces the peak amplitudes of acceleration response spectra and Fourier spectra without altering their overall shapes. Under low-intensity excitation (0.1 g), the response spectra exhibit multipeak characteristics with a predominant period of approximately 0.06 s. As ground-motion intensity increases, the spectra transition to a single-peak pattern, accompanied by a lengthening of the predominant period to approximately 0.08 s, indicating enhanced system nonlinearity and amplification of low-frequency components. Dynamic strain measurements reveal that the mitigated section consistently experiences lower strain peaks and slower strain accumulation, particularly at the crown and invert, whereas the haunch exhibits the weakest mitigation effect due to strong boundary constraints. PDI analysis indicates that the conventional section enters a plastic-dominated state when excitation exceeds 0.3 g, whereas the mitigated section remains primarily in the elasto-plastic stage with substantially lower PDI values. Post-test observations confirm that damage in the mitigated section is markedly reduced compared to the conventional section.

Experimental results demonstrate that a seismic mitigation configuration that explicitly considers primary support and incorporates a seismic isolation layer can effectively improve tunnel seismic performance under high-intensity earthquake loading. This composite system reduces spectral amplitudes and plastic deformation demand while preserving the fundamental spectral characteristics of the lining response. The mitigation effect is most pronounced at the invert and crown, indicating that the haunch remains a critical area requiring additional design attention. The proposed approach provides a practical experimental basis for energy-dissipation-oriented seismic design and retrofitting of tunnels in earthquake-prone regions.