Numerical modelling of tunnel face stability in homogeneous and layered soft ground

Soft ground tunnelling in urban areas is more frequently being performed using the shield method. Due to its great influence on both ground settlement and construction safety, face stability is one of the most critical problems in shield tunnelling. In this research, a series of 3 - D finite element simulations were conducted using the Midas - GTS NX software in order to determine the required collapse pressure of a tunnel face during tunnelling in homogeneous or layered soils. In terms of homogeneous ground, the effects of different soil strength parameters, cover-to-diameter ratios and tunnel diameters were investigated. Based on the numerical results of 140 analyses, a new design equation has been derived to calculate the required face collapse pressure during tunnelling in a purely frictional soil or a c - phi soil above the groundwater table. The results of this equation are in close agreement with the results from available experimental tests and theoretical approaches, and hence the equation provides a very useful method for estimating face collapse pressure. Furthermore, the arching effect has been explicitly investigated, and the failure mechanisms ahead of the tunnel face have been presented for various cases. For layered ground, two stratification scenarios were considered. Each scenario comprised two strata; an upper and a lower stratum. The first scenario (Case 1) involved the upper stratum intersecting with the lower stratum at the tunnel crown, while in the second scenario (Case 2), both strata intersect at the tunnel axis. In each case, the shear strength parameters of the upper and lower strata were changed to study the influence of these variations on face collapse pressure. In Case 1, it was found that the face collapse pressure is far more sensitive to parameter variations in the lower stratum than those of the upper stratum. In Case 2, however, almost the same values of face collapse pressure were obtained for both sets of parameter variation. Furthermore, if the lower stratum is stronger than the upper stratum, the required face collapse pressure in Case 2 is greater than that in Case 1.