Piles, by their nature, are generally designed to do anything but move. On the northern bank of the Thames in central London, however, lies a group of large 30m deep concrete bored piles that are intended to do just that - slide up to 50mm towards the river.
The 2.1m diameter piles, surrounded for a third of their depth by an annulus of bentonite, form the front two rows of the cable anchor foundation for the capital's new £15.9M footbridge. When the suspension bridge's cables are stressed later this year, the entire anchor block will move bodily forward.
Critically, only 14m in front of the foundation, stands a 200 year old concrete river wall which could be weakened by the increasing pressure as the ground resists the moving piles. Hence the bentonite annulus.
Such innovation at the construction stage is appropriate on a project bristling with cutting edge design.
Two groups of cables span from bank to bank over two piers placed 150m apart in the river. The dip of the cable over this central span is 2.3m, about six times shallower than a more conventional suspension bridge structure.
A shallow cable profile necessarily results in large cable tensions. Ribbon bridges have similar shallow profiles, but are typically single spans, allowing the cables to be anchored directly to substantial stiff abutments.
Restricted space either side of the river ruled out raking piles and so loads have to be transferred to the vertically piled foundations as a horizontal force. Stiff abutments help limit the live load deflections.
The bridge dead load is about 2t/m, and the resulting cable tension is 22.5MN. Live load adds another 8MN under extreme conditions. The design exploits this tension, which provides sufficient stiffness for the deck so that no additional structure, such as a conventional horizontal truss under the deck, is required to resist lateral wind load.
The lack of space is most evident at the northern abutment, where bridge cables run directly into the narrow flight of riverbank steps beneath St Paul's, flanked by multi storey buildings.
Piles will resist the horizontal loading by lateral loading of the soil. The moment is resisted partly by frame action but mainly by vertical push- pull axial loading of the piles. Design of the foundation system is principally constrained by the requirement to limit horizontal deformations at the cable anchorage point due to lateral movement and rotation of the pile cap.
Then there is the issue of the moving piles. Horizontal loading will cause some soil pressure to be applied to the piles supporting the neighbouring buildings and the river walls, which they should be able to sustain without distress.
In particular, abutment piles will tend to apply a pressure to the back of the river walls as they are pulled forwards by the force from the cables.
The original proposal to avoid loading these walls involved sleeving the front row of piles at each abutment, which left a 75mm clear void between the piles and the surrounding soil behind the river walls. The outer sleeve was to comprise a 2.5mm thick ribbed steel tube, which was designed to be flexible enough to distort under loading from the piles behind. However Bachy Soletanche considered sinking such a flimsy outer casing to be impractical.
The alternative system was developed during discussion between Arup, Bachy Soletanche and MTM. The front two rows of piles are surrounded by an annulus filled with a weak, cement-bentonite slurry. The slurry must be sufficiently dense and strong to support the soil, but should only transfer a limited amount of load to the ground as the piles deform.
As Bachy Soletanche piling operations manager Chris Merridew puts it: 'It's not unusual to surround piles with a bentonite ring to protect them from outside ground forces, but it could be a first to do it the other way round.'
The north abutment site was excavated to just below the pile cap soffit level to allow the Museum of London Archaeological Service to investigate its archaeological potential. This excavation was supported by a temporary king post wall system on three sides and the Swiss Bank Building basement wall on the other.
To allow access for the piling rig the excavation was temporarily backfilled before piling. The substructure is formed from a 3m deep pile cap, which is integral with the two walls connecting to the cable anchorages. Foundations comprise a group of 12, 2.1m diameter, bored cast insitu concrete piles, 28m long. Piles were constructed from C40 concrete and reinforced at the top with 44 T50 bars, reducing in number down the length of the piles. Pile lengths and reinforcement have been sized primarily to provide the desired stiffness response of the foundation.
The south abutment site was also archaeologically excavated. As there was more room here, battered side slopes could be formed and the piles installed from the base of the dig. The bridge cables will be anchored to 'wings' attached the pile cap about 4m above ground level. As with the north abutment, the loading on the foundations is predominately horizontal. The foundations comprise a group of 16, 2.1m diameter, bored cast insitu concrete piles, 28m long with a 3m deep pile cap.
Pile caps are separated from the soil along the face by the river by Cellcore, a compressible grillage formed from expanded polystyrene, again to avoid loading the river wall. The other three sides of the pile caps are separated from the soil by 200mm thick expanded polystyrene board.
Within the river, the bridge is supported by two piers. Foundations for the piers comprise a pair of 6m diameter caissons, dug 18m below river bed level within a sheet pile cofferdam. Pier foundations have to withstand a 32MN head-on ship impact without significant permanent displacements as well as providing adequate stiffness to the cable system along the length of the bridge.
Design of the caissons was carried out using a 3-D finite element analysis which modelled the structure and the soil by 40,000 brick elements.
The non-linear behaviour of the soil was simulated to correctly represent its stiffness at relatively small movements in service as well as the large displacements which would occur during a ship impact.
The caissons were dug using a mini-excavator and lined with bolted, pre- cast concrete segments which were grouted to ensure good contact with the surrounding soil. The caissons were subsequently backfilled with reinforced concrete. During excavation the soil beneath the shafts was drained using predrilled relief wells, with monitoring carried out using arrays of vibrating wire piezometers.
John Seaman, Arup Geotechnics
Additional reporting by David Hayward