Development of underground transportation systems often involves tunnels, which may have to be constructed in close proximity to existing piled-foundations in urban areas. It is well recognised that pile capacity depends on mobilised stress near shaft and toe of the pile. Since tunnelling is a stress release process which induces inevitably soil movement in the ground. Although extensive studies have been conducted to investigate adverse effects of tunnel construction on existing single piles, influence of twin tunnel on existing pile groups, which are commonly found in urban cities, is not well investigated and understood.

This research aims at investigating the interaction between advancing twin tunnels and an existing 2×2 pile group three-dimensionally and quantifying settle...[ Read more ]

Development of underground transportation systems often involves tunnels, which may have to be constructed in close proximity to existing piled-foundations in urban areas. It is well recognised that pile capacity depends on mobilised stress near shaft and toe of the pile. Since tunnelling is a stress release process which induces inevitably soil movement in the ground. Although extensive studies have been conducted to investigate adverse effects of tunnel construction on existing single piles, influence of twin tunnel on existing pile groups, which are commonly found in urban cities, is not well investigated and understood.

This research aims at investigating the interaction between advancing twin tunnels and an existing 2×2 pile group three-dimensionally and quantifying settlement, tilting and load transfer mechanisms within the pile group. The prototype diameter of each pile and tunnel was 0.8 m and 6 m, respectively. Two series of three-dimensional centrifuge model tests were carried out to investigate the response of the existing pile group under working load subjected to twin (i.e., side-by-side and piggy-back) tunnelling at various locations in dry Toyoura sand. Side-by-side twin tunnels (excavated one after the other on both sides of the pile group) were simulated in-flight by controlling 1% volume loss at three selected locations relative to the pile group, namely near the mid-depth, (Test SS), next to the toe (Test TT), and below the toe of the pile group (Test BB), of the pile group. Similarly, in each twin (piggy-back) tunnelling test, the first tunnel was constructed near the mid-depth, while the second tunnel was subsequently advanced either next to, below or right underneath the toe of pile group (i.e., Tests ST, SB and SU, respectively). To enhance fundamental understanding of the interaction between twin tunnels and pile group, systematic three-dimensional numerical back-analyses of the centrifuge model tests and a parametric study were also carried out.

Among the six configurations simulated, the most significant settlement (i.e., 4.6% of pile diameter) of the pile group is caused by twin tunnelling SU, as the largest reduction (35%) of end bearing resistance was occurred after tunnelling underneath the pile group. This is equivalent to the apparent loss of pile capacity of 40%. Downward soil movements due to tunnelling underneath the pile group induced net tension in each pile. On the other hand, the largest tilting of the pile cap (i.e., 0.2%, reaching the limit suggested by Eurocode 7) and hence the most significant bending moment (exceeding the ultimate bending moment pile capacity (i.e., 800 kNm) by 40%) induced near the pile cap are resulted from twin tunnelling ST. Because tunnelling ST causes the most significant plastic yielding near the toes of front piles, resulting in the largest differential settlements between the front and the rear piles. In spite of the severely induced bending moments in most of the tests (Tests ST, TT, SB and BB), no more than 15% increase in axial force is observed in all the tests.

Two distinct load transfer mechanisms can be identified from the tests. Tunnelling near the mid-depth of the pile group and near the pile toes leads to downward load transfer in piles while upward load transfer is resulted from tunnelling below the pile toe. Apart from the load transfer within each pile, load re-distribution also occurs among the four piles through the rigid pile cap. During twin tunnelling, the axial load at pile head reduces only in the pile closet to the advancing tunnel face and the reduction is re-distributed to the other three piles. Compared to a single pile subjected to twin tunnelling, less change of axial load is induced in a pile group because of load-redistribution among piles. In contrast, significant larger bending moment is induced in a pile group than that in a single pile. This is caused by the tilting of pile cap during tunnelling.