Abstract

The number of larger tunnels in seismic regions has grown significantly over the last decades. The behaviour of tunnels under seismic actions may be assessed using simplified or more complex approaches. Plane–strain centrifuge tests with dynamic loading on a model tunnel are used as experimental benchmark on the seismic behaviour of tunnels, with the ultimate aim of calibrating numerical and analytical design methods. Two models with dry uniform fine sand were prepared at two different densities, in which an aluminium-alloy tube was installed. This paper describes the numerical simulation of these tests with an elastoplastic model. The tunnel response recorded in the centrifuge tests is compared with the numerical prediction, showing the evolution of accelerations and internal forces along the tunnel lining during the model earthquakes. In general, the numerical simulation diverges from the recorded at the centrifuge tests. The numerical simulation largely amplifies the motion at the fundamental frequency of the soil deposit, while this effect is not significant in the centrifuge tests. It is shown that the peak increments in lining forces during dynamic loading measured in the centrifuge test disagree with the values from the numerical simulation and from the Wang’s elastic solution. The divergence observed between simulation and centrifuge tests may result mainly from the real initial stiffness of the sand in the centrifuge tests which are lower than those measured in laboratory tests and to the insufficient knowledge of all relevant stress paths to be imposed to soil for the calibration of model parameters.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call