A new study published in Engineering presents novel resilience models for assessing and quantifying the recovery of tunnels after earthquakes. The research, conducted by a team from Tongji University, Brunel University of London, and University College London, offers a probabilistic approach to predict tunnel recovery, providing valuable insights for infrastructure operators and city planners.
Tunnels are critical components of urban infrastructure, continuously exposed to various hazards, including earthquakes, fires, floods, and aging-related disturbances. Events such as the magnitude-7.8 earthquake in southeastern Türkiye in 2023 and the Chi-Chi earthquake in Taiwan of China have caused significant damage to tunnels, highlighting the vulnerability of these structures and the need for robust resilience assessment tools. Previous research has extensively explored the vulnerability and fragility of tunnels, but studies focusing on restoration to quantify resilience have been limited. This gap has hindered proactive and reactive adaptation measures to ensure seamless tunnel functionality.
To address this issue, the study introduces a damage-level-dependent probabilistic approach for quantifying tunnel recovery. The research team conducted a global expert survey to gather input on restoration tasks, their duration, sequencing, and cost. The survey focused primarily on damage induced by seismic events, incorporating idle times and traffic capacity gains over time. The results were used to generate deterministic and probabilistic reinstatement and restoration models, with the probabilistic models accounting for epistemic uncertainties.
The study proposes a framework for assessing tunnel resilience, which includes hazard characterization, vulnerability assessment, and the development of restoration models. The framework is based on defining structural damage levels and using fragility functions to determine the probability of damage at a given hazard intensity. The restoration models are tailored to tunnel resilience assessments, incorporating expert knowledge on required restoration tasks and their prioritization. The models help quantify resilience and optimize the repair process for tunnels with various levels of damage.
The research highlights that the most time-consuming restoration tasks typically involve replacing tunnel structural components or reinforcing tunnel structures and soil. The study also finds that idle time and cost ratios increase significantly with greater damage severity. For example, the mean idle time for tunnels with minor, moderate, extensive, and complete damage levels were found to be 6.52, 12.09, 24.02, and 50.57 days, respectively. The cost ratio, which represents the restoration cost relative to the total construction cost, also rises with increasing damage severity, ranging from 7.73% for minor damage to 74.05% for complete damage.
The study's findings are demonstrated through a case study of a typical tunnel, showing how the newly developed restoration models can be applied to assess tunnel resilience. The results indicate that the resilience index of the tunnel decreases as seismic intensity increases, with more severe damage levels corresponding to longer restoration times and higher uncertainty in the resilience index.
The research underscores the importance of incorporating resilience models into post-earthquake restoration workflows, guiding practitioners in optimal decision-making. The models provide a scientific basis for estimating downtime and losses due to tunnel disruptions, facilitating proactive tunnel adaptation and resource allocation. Future research is suggested to enhance the reliability and generalizability of the restoration models by collecting more expert responses and incorporating parameters tailored to specific tunnel designs and conditions.
The paper "Resilience Models for Tunnel Recovery After Earthquakes," is authored by Zhong-Kai Huang, Nian-Chen Zeng, Dong-Mei Zhang, Sotirios Argyroudis, and Stergios-Aristoteles Mitoulis. Full text of the open access paper: https://doi.org/10.1016/j.eng.2025.06.028
