There has been an increasing demand for new tunnels in respond to rapid growth in cities. Earth pressure balanced or slurry shield tunneling method is commonly used to construct tunnels in soft ground to improve stability and safety. Over the decades, many studies have been carried out to investigate active failure of tunnel face in sand and clay. However, investigation on passive failure of tunnel face is relatively rare. Failures of a compressed air section during construction of Dockland Light Railway extension and the 2^nd Heinenroord Tunnel during drilling imply that a systematic investigation on passive failure at tunnel face is required. The objectives of this study are to investigate passive failure and deformation mechanisms due to tunnelling in sand and clay by carrying out cen...[ Read more ]

There has been an increasing demand for new tunnels in respond to rapid growth in cities. Earth pressure balanced or slurry shield tunneling method is commonly used to construct tunnels in soft ground to improve stability and safety. Over the decades, many studies have been carried out to investigate active failure of tunnel face in sand and clay. However, investigation on passive failure of tunnel face is relatively rare. Failures of a compressed air section during construction of Dockland Light Railway extension and the 2^nd Heinenroord Tunnel during drilling imply that a systematic investigation on passive failure at tunnel face is required. The objectives of this study are to investigate passive failure and deformation mechanisms due to tunnelling in sand and clay by carrying out centrifuge model tests and numerical simulations together with plasticity and cavity expansion solutions.

Centrifuge model tests were carried out for tunnels located at cover to diameter (C/D) ratios ranged from 2.0 to 4.0 in medium dense sand and soft clay. Three-dimensional finite element analyses were carried out to back-analyse the measured results. After calibrating the numerical tool, numerical parametric study was performed to extend the investigation to consider C/D in a wider range of 1.0 to 6.0. In addition, tunnels located at C/D ratio of 2.0 in loose and dense sand were also included in the numerical parametric study to investigate effect of relative density. The measured and computed results are used to evaluate some existing upper bound and cavity expansion solutions in estimating passive failure pressure. Besides, Gaussion distributions are adopted to describe the induced surface heave.

A localised failure mechanism associated with surface settlement is observed for shallow tunnel located at C/D of 2.2 in loose sand due to soil compression. As initial relative density is increased, on the contrary, a funnel-type failure mechanism associated with surface heave is induced attributed to an increase in soil dilation. With an increase in C/D ratio resulting in an increase in confining stress around the tunnel, soil surrounding the tunnel becomes less dilative and hence a localised failure mechanism associated with surface settlement is induced in medium dense sand. The surface settlement trough become wider and shallower as C/D ratio is increased. In soft clay, funnel-type and localised failure mechanisms are observed for shallow and deep tunnel, respectively. It is found that surface heave due to passive failure is less significant for deep tunnel located in clay at C/D equal to or larger than 3.0. It is observed that surface heave in clay may be described using a two-dimensional Gaussian distribution. An average of 90% degree of consolidation is reached about 1.5 years after passive failure. For shallow tunnel, initial heave due to tunnel face displacement is beneficial in reducing subsequent consolidation settlement but induced extra consolidation settlement for deep tunnel. It is found that existing upper bound solution adopting a five-block failure mechanism may be used in estimating passive failure pressures for tunnels with a funnel-type failure mechanism. The upper bound solutions overestimate passive failure pressures for tunnels with localised failure mechanism but cavity expansion solutions is fairly consistent with measured and computed values for deep tunnel. This is attributed to large discrepancy between localised and five-block failure mechanisms. For tunnelling in clay, passive failure pressure is about two times of the initial vertical total stress at tunnel centreline.