Answers are grounded in this paper and cite their sources.
A framework assesses structural robustness and recoverability in cross-fault tunnels, indicating improved safety measures.
The long-term operational safety of cross-fault tunnels is threatened by the combined effects of creep-slip rupture and earthquakes. This study presents a novel machine learning-assisted resilience assessment framework for quantitatively evaluating both the structural robustness and recoverability of cross-fault tunnels under such multi-hazard conditions. First, a numerical model accounting for the interaction among surrounding rock, tunnel structure, and fault zone is developed and validated against existing experimental data to ensure reliability. Based on the analysis of structural responses under varying fault rupture deformations and seismic excitations, a convolutional neural network (CNN) architecture is optimized to predict lining deformations and strain evolution. Utilizing the CNN predictions, fragility analysis of tunnels in terms of fault rupture deformation and seismic excitation is conducted through a planar probabilistic demand model, which correlates tunnel lining performance indicators with hazard intensity. Subsequently, a comprehensive performance function is established by integrating tunnel damage probability, economic loss coefficients, and multiple recovery modes. Finally, the influences of key parameters, including creep-slip rate, seismic intensity, and recovery characteristics, on tunnel resilience are thoroughly investigated. The proposed methodology offers a rapid and practical theoretical foundation for performance-based design, efficient resilience assessment, and post-hazard recovery planning of cross-fault tunnels.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: