Probabilistic seismic capacity models for circular tunnels at different performance levels

This paper presents a systematic study to develop the probabilistic seismic capacity models for circular tunnel linings. The uncertainties of lining parameters, soil properties, and ground motions are all considered. The Bayesian approach is used to estimate the model parameters based on the capacity samples simulated by a large number of refined dynamical numerical analyses. Finally, probabilistic capacity models are separately con-structed for circular tunnels at three performance levels. To demonstrate the properties of the proposed method, the capacity models are used to estimate the capacities and the fragility curves of a typical tunnel. It is concluded that (1) the probabilistic model can quickly estimate the probability distribution of capacity based on existing parameters, with the estimated median value closely approximating the numerical result; (2) the proposed probabilistic capacity models can be used to quickly derive fragility curves for specific tunnels, which facilitates safety evaluation and performance-based seismic design of tunnels.