Three-dimensional interaction of multiple crossing tunnels : centrifuge and numerical modeling

Abstract

Due to the growing demand for underground space in densely populated cities worldwide, an increasing number of closely spaced multiple tunnels are being constructed. The effects of new tunnel excavation cause ground movements and stress changes which in turn result in an adverse impact on the adjacent tunnels. Adverse effects of tunnel driving on adjacent existing tunnels, such as excessive settlement, large angular distortion and cracking of tunnel linings, have been reported in a number of case histories. However, the crossing-tunnel interaction is complex and is influenced by many factors. The main objectives of this research are to improve the fundamental understanding of the interaction of multiple crossing tunnels and to provide high quality physical data for numerical modelers and engineers for checking their designs. A total of eight centrifuge tests were carried out in dry Toyoura sand. Factors influencing the interaction of crossing tunnels, namely the effects of modeling technique, construction sequence, cover depth, pillar depth, shielding and twin new tunnel excavation in side-by-side and vertically stacked arrangements, were investigated. Tunnel excavation was simulated three-dimensionally in-flight using a novel device called a “Donut”, which modeled the effects of tunnel volume loss equaling 2% and weight loss. The measured results were back-analyzed using the finite element method to enhance understanding of the stress transfer mechanism, strain induced and mobilization of stiffness in crossing-tunnel interaction. An advanced hypoplasticity constitutive model with small strain stiffness was adopted. In addition, numerical parametric studies were also performed to examine the influence of volume loss, relative density, tunneling in saturated sand and tunnel diameter in regards to the interaction of crossing tunnels. One of factors that strongly influenced the crossing-tunnel interaction is the pillar depth-to-diameter ratio (P/D). Note that the pillar depth is the vertical clear distance between tunnels. In the excavation of a new tunnel underneath at P/D of 0.5, the maximum settlement, tensile strain and shear stress induced in the existing tunnel exceeded the permissible limits given by LTA (2000), ACI (2001) and ACI (2011), respectively. By increasing P/D from 0.5 to 2, the tunnel settlement was reduced by 50%. This is attributed to a larger shear modulus and a smaller reduction in confining stress of soil in the case for P/D of 2 along the invert of the existing tunnel than for P/D of 0.5. The existing tunnel was elongated horizontally when P/D equaled to 0.5. This is because the stress reduction in the horizontal direction was greater than that in the vertical direction. The stress relief caused by the new tunnel not only led to a reduction in the vertical stress at the invert but also resulted in substantial stress reduction at the springline of the existing tunnel. On the contrary, the existing tunnel was elongated vertically as the new tunnel was excavated at P/D of 2.0 since the reduction in stress in the vertical direction dominated. For multiple crossing-tunnel interaction, the settlement of the existing tunnel caused by the vertically stacked tunnel arrangement was smaller than that due to the side-by-side tunnel case. This is attributed to larger P/D of the lower new tunnel in the vertically stacked tunnel case than the new side-by-side tunnels. In addition, the shielding effects provided by the upper new tunnel minimized the effects of the new tunnel excavation on the existing tunnel.

Date of Award	2014
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology