Designing and the construction of a tunnel in urban areas has their own specific considerations. Usually, excessive settlement caused by tunneling during construction damages the adjacent infrastructures and utilities, especially if the tunnel is excavated by the new Austrian tunneling method (NATM). Thus, it’s important to make accurate predictions and effective control on tunneling-induced settlement. In this study, the soil’s Young’s modulus was modeled using a three-dimensional random field and coupled with a finite difference method (FDM) analysis to reveal the influence of scale of fluctuation (SOF) on the maximum surface settlement (Smax). To generate the field of soil’s Young’s modulus, the Fourier series method is employed and sensitivity studies are further performed via Monte-Carlo simulations (MCS). The results demonstrate both the mean value of Smax and its coefficient of variation (COV) increase from 28 mm to 31 mm and from 0.02 to 0.35 respectively, with an increasing horizontal SOF but they stabilize at higher values of SOF. Furthermore, the probability of failure increases as COV increases for each allowable limit greater than the verification FDM of Smax. It is observed that ignoring the spatial variability of soil’s properties leads to an underestimate of the risk of excessive surface settlement.

No abstract

Designing and the construction of a tunnel in urban areas has their own specific considerations. Usually, excessive settlement caused by tunneling during construction damages the adjacent infrastructures and utilities, especially if the tunnel is excavated by the new Austrian tunneling method (NATM). Thus, it's important to make accurate predictions and effective control on tunneling-induced settlement. In this study, the soil's Young's modulus was modeled using a three-dimensional random field and coupled with a finite difference method (FDM) analysis to reveal the influence of scale of fluctuation (SOF) on the maximum surface settlement (Smax). To generate the field of soil's Young's modulus, the Fourier series method is employed and sensitivity studies are further performed via Monte-Carlo simulations (MCS). The results demonstrate both the mean value of Smax and its coefficient of variation (COV) increase from 28 mm to 31 mm and from 0.02 to 0.35 respectively, with an increasing horizontal SOF but they stabilize at higher values of SOF. Furthermore, the probability of failure increases as COV increases for each allowable limit greater than the verification FDM of Smax. It is observed that ignoring the spatial variability of soil's properties leads to an underestimate of the risk of excessive surface settlement.