Identifying seismic disaster precursors and instability early warning signs in tunnels is critical for seismic design and catastrophe warning. Current approaches predominantly rely on static assessments of specific states, either post-event or at peak response. Thus, they fail to capture the continuous evolution and abrupt transitions inherent in nonlinear dynamic systems. For this purpose, the Load/Unload Response Ratio (LURR) theory was introduced to evaluate the seismic stability of tunnels. Definitions were established for loading and unloading parameters, response parameters, and the LURR during the seismic response of tunnels. According to the principles of the LURR theory, shaking table model tests were performed on an unlined tunnel to study how the LURR varies and how the soil’s stability changes with different seismic intensities. Research shows that as seismic amplitude increases, the stability of an unlined tunnel evolves through three distinct phases: stable bearing capacity, localized collapse, and overall collapse. A correlation is observed between the time-history curve of the LURR of the surrounding soil and the progression of stability. During the stable bearing phase, LURR values fluctuate at low magnitudes, while the maximum LURR at the weakest structural location (the arch foot) gradually rises with increasing seismic amplitude. As stability deteriorates, the LURR displays a localized growth pattern in the crown and sidewall regions (monitoring points 1-4). A declining trend in the regional LURR peak value corresponds to the onset of localized collapse in that specific area. Following the initial localized collapse, the maximum LURR shifts from the crown zone to the sidewall foot and invert waist areas. When the regional peak LURR in this secondary zone begins to decrease, the tunnel experiences overall collapse that propagates to the ground surface. Based on shaking table model test results, the decline of the first localized LURR peak, combined with its spatial migration, can serve as a criterion for assessing localized tunnel instability. Furthermore, the decline of two or more localized LURR peaks provides a predictive indicator for the overall instability of the tunnel.
Tan et al. (Sun,) studied this question.