Tunnellers are used to carving out the ground beneath city streets and under buildings, even if such urban work can be fraught at times. But rarely do they need to find their way through the forest of piles just beneath a big building.
At Changi airport in Singapore even more is required; piles lie in the route of the tunnel and must be supported just above it while they are cut away beneath to allow the excavation. Other piles sit within 2m of the tunnel line, which will go deeper than their end points.
All these foundations, just under 50 piles, are ‘live’, continuing to hold up a five storey structure above, the carpark for the airport’s Terminal Two, which abuts the end of the terminal itself. Other piles support a finger pier out to planes and there are foundations too for a bus ramp, and the little elevated electric shuttle, or Skytrain, that runs between terminal Two and Terminal One.
To do the work means forming diaphragm walls from above, working from cramped spaces inside the buildings where it is difficult to manoeuvre big rigs. In places secant piles must go in instead and for a difficult short section only hand excavation will do, mining through and creating the tunnel by a top down method.
“And the consequences if the buildings moved are unthinkable,” says Richard Davies, a ground engineering consultant who has been closely involved in the work.
The airport remains in constant use.
All the buildings date to the end of the 1980s. Now Singapore is adding a mass transit link from the airport into the growing public transport network. A 6km spur connection from the East-West line will bring trains along 2km of elevated route and then through twin 3.5km bored tunnels under the airport and its runways, terminating in a huge, underground station terminus linked to the air terminals.
The TBM drive is spectacular enough, being Singapore’s longest and most precise ever. And the station will be a spectacular architectural feature. But most complex are two short overrun tunnels that must protrude another 126m from the end of the 215m long station box which sits transversely across the main road link into the airport.
“These tunnels pass under buildings to the airside, where they will end as stubs for possible later connections to further terminals,” explains Chelliah Murugamoorthy, project manager for client the Land Transport Authority on the airport contract.
The two overrun tunnels are not all, he says. Underneath one of them must go another short tunnel, to carry conveyors and services for the airport’s famously efficient baggage handling system. This extra tunnel is not being built for the LTA but has been added in to the contract for the civil aviation authority, which will need to bring baggage from a planned third terminal on the other side of the station.
But the older Terminal Two structure has the attention. Singapore’s LTA has designed the spur link tunnel, a track crossover where it enters the airport station and the 25m deep station itself.
For the overrun tunnels it proposed a diaphragm wall system. Compensation grouting would cope with any movements of the building during either wall formation or excavation inside the diaphragms and concreting of the tunnel boxes.
But the station’s main contractor, Japan’s Kumagai Gumi, working with local firm Semcorp, was uncertain about it. Working with advice from ground specialist Richard Davies Associates the JV proposed something different at tender stage. The idea was not to use grouting but make a much more precise assessment of the ground and go ahead with an observational method of construction.
“We suggested an experiment,” says Davies. “I mean, it is possible to calculate things fairly precisely and we did finite element work and other models but even if you do all the sums in the world you don’t know precisely what is going to happen.”
“The problem is the ground – it is not as well known as London Clay, for example” adds Alison Norrish, RDA’s engineer on site. “This is a material called Old Alluvium, a silty cemented sand which some engineers say is a weak rock. But it crumbles in 10 minutes if kept in a bucker of water.” The alluvium underlies some 7m or 8m of fill and weathered material above.
Just how piles would react to nearby excavation was not clear and RDA’s idea was to reschedule work on tunnels. A duplicate bored pile would be made early on in open ground, loaded by reaction against two new tension piles.
The three piles would sit near to the main station box diaphragm wall and, while panels were formed, extensive instrumentation could be used to monitor the behaviour of everything in detail. The three piles were set obliquely about 1-2m from the wall.
The results, produced towards the end of last year, were both surprising and encouraging. Norrish says she was betting the test piles would move as panels were taken out because like those under the carpark they support their loads mainly by skin friction, and the ground would certainly relax as a whole was excavated. But piles did not move even when later overloaded by 50%.
“We found there is some sort of arching effect in the ground,” says Richard Davies. This meant that stability of the piles was not affected by cutting into the circle of confining pressure.
On top of that the Old Alluvium has a very low permeability, explains Norrish. That meant the initial effect of excavation was to create a negative pore water pressure “sucking” around the pile. Friction remained. This meant there was an important time effect, adds Davies. As the pore pressure rebalances the ground relaxes over roughly a 24 hour period.
“We set some parameters for the construction as a result,” says Norrish. “Panels should be kept as small as possible, just over 2m long, and be done one at a time, and they should not be made closer than 2m to a pile.” That meant a certain amount of realignment and in the event two were within 1.75m. Panels had to be excavated and concreted in 24 hours.
This was easier said than done. To do the work meant operating inside the car park and even with one floor knocked out there was only 5.5m of headroom, much less than the height of most diaphragm rigs. Korean subcontractor Sambo had to bring in two special low headroom hydrofraises for the job, a Bauer machine and another, both with under 5m clearance. Low headroom also meant reinforcement panels could only be pre-made in 4m sections. Before beginning the work there were rehearsals by the crews for these awkward conditions.
But contingencies were vital. In case something went wrong and a panel got “stuck” a special “Zimmer frame” was created, as Davies called it. This big steel girder could be assembled under the car park floor to transfer load to adjacent piles. “And as first line of defence the panel hole would be immediately backfilled,” says Davies.
The same would obviously happen if the building moved beyond preset trigger levels. “We had nearly a £1M worth of instruments watching it,” says Davies, “including a robot total station measuring a series of prisms at short intervals.”
Part of the wall formation was done with secant piles, says Paul Gasson, the LTA’s deputy project manager. “We knew there were some old ground anchors left behind from the original construction and to cut through these we used a tungsten carbide tipped pile barrel attachment, again with the rig specially modified for low headroom.”
Before and during this work, the first part of the excavation under the car park had been done as part of the formation of the main station excavation. This 260m long box has a wider section at each end which will both become huge rising atria, extended by glass curtain walls above ground. The atrium space cuts under the car park too just before the overrun tunnels begin.
For this, the LTA designed a raft structure to hold up the adjacent lift and service core of the car park building, while piles and the structure below were cut away for the excavation.
“We put down temporary barrette piles on one side by the excavation, and on the other side there is a diaphragm wall. We excavated 4m underneath, enough to be able to insert big steel 14m span needle beams, each weighing about 35t and supported on the wall and temporary barettes,” says Gasson.
A 2.8m thick slab around the needles forms a platform for the car park above and, once that was in and loads transferred, the piles were sliced off.
Excavation is now under way within the parallel diaphragm walls and slabs have been formed 4-5m below ground level, around those piles which sit within the walls. These too are cut off after loads are transferred to the walls; excavation then continues with in situ concrete strut/walers.
Tunnel roof and base slabs formed as part of a 6m deep box below, also act as struts. Some 47 piles are being cut away.
On the second of two sets of parallel diaphragm walls, the excavation will continue deeper because the second baggage tunnel box is needed under the train overrun tunnel. Diaphragm walls are 24m deep on this side, 7m below the level of the pile ends.
But there is more. Though the 800mm thick diaphragm walls have been made in the car park and also on the airside where they underlie the Terminal Two finger pier, a central 20m long section remains to be done, This is where the tunnel must pass through the foundations of a live Skytrain viaduct above and a bus ramp. There is no room for rigs or piling machines.
The tunnels will therefore be formed by hand mining. The area is getting a sheetpiled cofferdam and “there will be some chemical grouting” to hold back groundwater, says Norrish.
“We will use shotcrete and soil nails to support the slopes of a hand excavation as we go under the bus ramp to form big underpinning transfer structures,” says Gasson. Beams will span between diaphragm wall and pile sections either side of the bus ramp and the Skytrain.
A top down hand excavation will then take out 2m increments to allow concreting of the side walls of the box and then its base slab. The LTA, RDA and the main contractor are still discussing the exact method, however.
Related Files
Cross section
Changi airport
