For public convenience, basement excavations for shopping malls and/or car parks are constructed very close to existing tunnels. Construction of basement may induce unsymmetrical and highly skewed loadings and/or stress changes in an existing tunnel, not only in the transverse but also in the longitudinal direction of the tunnel. Although the basement-tunnel interaction has attracted considerable attention recently, it is often simply treated as a plane strain problem.

This research aims at investigating the fundamental mechanisms of the basement-tunnel interaction three-dimensionally. Two major research methodologies, i.e., centrifuge modelling and finite element analysis, were adopted. Based on a dimensional analysis of the governing parameters, four centrifuge tests were con...[ Read more ]

For public convenience, basement excavations for shopping malls and/or car parks are constructed very close to existing tunnels. Construction of basement may induce unsymmetrical and highly skewed loadings and/or stress changes in an existing tunnel, not only in the transverse but also in the longitudinal direction of the tunnel. Although the basement-tunnel interaction has attracted considerable attention recently, it is often simply treated as a plane strain problem.

This research aims at investigating the fundamental mechanisms of the basement-tunnel interaction three-dimensionally. Two major research methodologies, i.e., centrifuge modelling and finite element analysis, were adopted. Based on a dimensional analysis of the governing parameters, four centrifuge tests were conducted in dry sand to investigate the influence of tunnel location, sand density and retaining wall stiffness on the three-dimensional tunnel responses due to basement excavation. Moreover, two tests were carried out in saturated clay to explore long-term tunnel responses due to overlying basement excavation. A continuous aluminium tube (i.e., empty) was used to model existing tunnel. Effects of excavation were simulated by draining heavy fluid away in-flight. To enhance the fundamental understanding of the basement-tunnel interaction, three-dimensional numerical back-analysis of centrifuge tests and systematic parametric study were conducted. The parameters considered on the interaction in sand included sand density, retaining wall stiffness, excavation geometry, aspect ratio, unloading ratio, tunnel stiffness and joint stiffness ratio.

For the tunnel located directly beneath the basement, heave was induced due to vertical stress relief. Because of symmetrical stress relief around the tunnel, it was vertically elongated. As the relative sand density decreased from 90% to 30%, the maximum heave and tensile strain induced in the tunnel increased by 90% and 80%, respectively. This is because a looser sand has smaller stiffness. Moreover, it is found that the tensile strain induced along the longitudinal direction was insensitive to retaining wall stiffness, but that induced along the transverse direction was significantly reduced by a stiff wall. In addition, the basement-tunnel interaction at the basement centre reached a plane strain condition when excavation length along the longitudinal tunnel direction was longer than 9 H_e (final excavation depth). Both heave and transverse tensile strain of the tunnel exceeded the allowable movement limit and cracking strain when excavation length was longer than 5 H_e (final excavation depth) and excavation width was wider than 2 H_e. For the tunnel located outside the basement, it demonstrated settlement resulting from inward soil movements behind the wall. Due to unsymmetrical stress relief and shearing, the tunnel was distorted (i.e., elongate toward basement). The induced tunnel response in this case was less than 35% of the corresponding value for the tunnel located directly beneath basement. Thus, it is suggested to construct a basement at a side of tunnel rather than above it. The use of a diaphragm wall reduced the maximum settlements and tensile strains induced in the tunnel by up to 22% and 58%, respectively, compared with the use of a sheet pile wall. This is because a stiffer diaphragm wall can significantly reduce the ground movements behind it.

Because of a smaller initial void ratio in a stiffer clay, induced tunnel heave and tensile strain upon completion of excavation in heavily overconsolidated clay (overconsolidation ratio (OCR) = 6.0) were 25% and 16% smaller than that in lightly overconsolidated clay (OCR = 1.7). Due to the dissipation of excess negative pore water pressure, the maximum tunnel heave and tensile strain increased by up to 210% and 50%, respectively, in heavily and lightly overconsolidated clays. Observed larger long-term tunnel response in the stiffer clay was due to shearing induced larger excess negative pore water pressure.