Due to the inevitability of soil movements associated with tunnel construction, tunnelling in urban areas can have a significant impact on nearby structures. The effects can be even more pronounced in the case of piled structures, as tunnelling could potentially be carried out very close to piles. In addition, stresses around piles are generally higher than near shallow foundations, so that stress relief associated with tunnelling might have more pronounced effects.
The effects of tunnelling on piled foundations were investigated by means of a model study in the geotechnical centrifuge. A model tunnel, surrounded by dense dry fine sand and capable of inducing plane strain deformation, was used. Model piles, loaded to approximately 50% of their penetration resistance were used. The study initially focussed on the tunnelling induced settlement of single piles, from which a zone of influence around the tunnel was defined in which a potential for large tunnelling related settlement exists. Piles were installed to the depth of the tunnel crown and shallower, as such piles are likely to be more prone to settlement.
Instrumented model piles were used to investigate changes in the load distribution on piles in response to tunnelling related volume loss. Piles inside the zone of influence (mentioned above) suffered a reduction in base load, so that the shaft capacity was mobilised to maintain equilibrium. The mobilisation of shaft capacity was associated with a significant increase in the normal stress on the pile shaft. A lower bound method for the estimation of this stress increase is presented. Once the maximum shaft capacity had been mobilised, continued volume loss resulted in rapid pile settlement. Piles outside the zone of influence did not suffer a reduction in base load during volume loss. The base loads remained nearly constant or increased slightly due to negative shaft friction during volume loss. These piles underwent only small settlements.
Groups of two and three piles were tested to investigate load transfer within a group during volume loss. The presence of more than one loaded pile resulted in larger volume losses being required to induce a given pile settlement than in the case of single piles. In the case of group piles within the zone of influence, volume loss first caused a base load reduction on the individual piles, resulting in the mobilisation of shaft capacity. Once the maximum shaft capacity of a pile had been mobilised, its settlement accelerated, causing load to be transferred to neighbouring piles. Except for piles with their bases very close to the tunnel, large volume losses (which should not occur in practice) were required to induce load transfer between piles.
In addition to pile behaviour, stress changes and soil movements around the deforming tunnel were investigated. Scaled surface and subsurface settlement troughs compare well with results observed in practice. As volume loss was imposed, significant stress redistribution occurred, with the radial stress near the tunnel reducing, while the circumferential stress increased. The increase in circumferential stress (i.e. horizontal stress) above the tunnel resulted in slightly higher shear stresses being mobilised on the pile shaft during volume loss than during pile driving. Cavity contraction and radial equilibrium models were investigated to predict tunnelling related stress changes on the tunnel centreline in the soil.