Abstract: To investigate dynamic behaviour of shield tunnels and surrounding soil, a physical model test was conducted. An electromagnetic shaker located at the bottom of the shield tunnel was used to apply sweep excitation and train vibration load. The data of accelerometers are applied to calculate the frequency response function (FRF) and the maximum acceleration of the tunnel and soil. It is found that FRF is insensitive to the excitation amplitude, sweep direction and period, which represents dynamic characteristics of tunnel lining structure and surrounding soil. The results also show that the high-frequency response is greater than the low-frequency response at the tunnel lining. The attenuation of dynamic response along the longitudinal direction of the tunnel is obviously faster at the tunnel invert comparing to at the tunnel apex. For surrounding soil, a variety of dynamic response with depth is observed. A clear degradation of soil response along the longitudinal direction of the tunnel is found at all depths. The soil response increases with the increase of excitation frequency at the first measurement layer above the tunnel lining. However, at the second and third measurement layer, soil response increases linearly at the frequency of 30-90 Hz. At higher frequency range, soil response does not show a clear increasing trend with frequency. The dynamic response under train-vibration load is consistent with sweep excitation load. Both tunnel and soil responses decrease in the longitudinal direction. Tunnel response at the tunnel invert is larger than that at the tunnel apex. With the increase of the train speed, tunnel and soil responses are significantly amplified. It is also found that the soil responses at the free surface are more significant than the soil responses inside the soil layer from train induced vibration.

Key words: shield tunnels, vibration load, dynamic response, frequency response function, tunnel lining, soil