Completion will take place late this month of a long running project to tunnel underneath the 19th century central station in the Netherlands capital Amsterdam. The successful outcome will help vindicate the work of civil engineers and tunnellers on the difficult and sometimes fraught north-south line project.
The scheme has taken nearly a decade of planning, design and foundation development and an extension of immersed tube methodology, with major works underway from 2004, in different phases, for this section alone.
But despite the age of the historically protected late 19th-century building, the sensitivity of its old timber pile foundations, and challenging soft ground conditions, a new metro line tunnel space now exists underneath it. There has seemingly been no change to the structure above, which has been kept open and fully operational throughout the project.
As other works are completed the new box section tunnel underneath will be connected to the TBM driven tunnels still being bored to the south, and to a still-to-be- built immersed tube crossing of the Ij river on the north side of the station, which connects the new 9km long metro line to the isolated north Amsterdam districts.
The City client and the contractors on the project are all keen to show off their success here because the north-south line has been a contentious and fraught project from the start of planning in the 1990s, with some public opposition. It follows an even more contentious scheme in the 1970s for the city’s first line, east-west, which caused demonstrations by environmentalists.
Times are not the same as the politically heated 1960s and early 70s but there have still been significant concerns about the line’s route through the historic canal-side city centre, which apart from its cultural significance is also a major world tourist attraction. Virtually all of the buildings here are significant.
Not only that, they are highly sensitive, since the city is built on mostly soft sand and silt of marine and fluvial deposits with tens of thousands of centuries-old timber piles supporting most of it. An accident with a badly placed slurry wall some years ago, which undermined a critical building close to one of the station pits on the southern part of the route, did not help the public image of the scheme.
World economic banking failures in 2008 did not salve the public mood as delays and costs also rose on the project; the scheme is thought now to be costing near to EUR 4bn (USD 5.4bn) compared to an initial EUR 1.5bn (USD 2bn). It will not finish until 2017 instead of an original 2011.
However, the new line of some 9.8km long is an important link to connect northern Amsterdam across the river IJ, through the centre and to the south of the city where there is a new World Trade Centre complex. But to pass the congested and famous old city, it must go underground in a 3.12km long bored tunnel, and making this caused significant fears. It is also below ground in the immersed tube and station area.
In fact, despite two incidents – another smaller failure of sheet piling – tunnelling so far has gone well. An early project on the line was to set up a huge automated total station monitoring system for the buildings along the route, with vast amounts of data being collected hourly. It indicates that settlements for the twin bores have so far been within the very tightly set limits.
Centraal Station
The station project was one of the most sensitive sections of all. It is the central hub of the city at the centre of its concentric rings of canals, and the point of entry for many visitors. But for this reason it was important not just for the line to pass under the station but also for there to be a stop there. It has to connect to the main line and also the east-west which crosses here too.
But how to do it was a major issue. Not only is the main building ‘untouchable’ to quote Bas Obladen, a specialist foundation and jet grouting consultant who worked for contractor Strukton on the scheme, but there were significant problems for the glass and steel arched roof enclosure which covers the tracks and 13 platforms just alongside it. Settlement could unacceptably alter torque forces in the frame shattering glass panels.
On top there was no question of closing the station, which is a key part of the busy Netherlands rail system and now its high speed line to Brussels and Paris. Some 300,000 commuters use it daily.
Ground conditions and the foundations meant conventional methods would prove difficult. The original brickwork station building is founded on a reclaimed sandy island formed on the river Ij shoreline in the 1880s, with wooden piles. There are more than 8,000 in total, ending in stiff clay just before the well-known Amsterdam ‘second sand layer’ about 18m down in the mixed alluvial, clay and silty ground.
“When God created the world he forgot to put any ground here the old story goes,” says Obladen. “So he used the bits and scraps left in his box. It is very mixed and difficult, the layers changing constantly metre by metre, some vanishing, others appearing as you move along.” Usually there are three sand layers, where most piling ends, the deepest at around 60m where bigger modern foundations reach to.
The so called ‘first sand layer’ about 12m down, which is where the old houses are piled to, peters out under the station towards the river, adds Mark Vlaanderen, Strukton’s engineering manager for the immersed tube float in this summer.
All the old piles would have to be removed to tunnel under the station. So engineers were faced with the problem of how to support it.
Station support
The answer was to use a table construction, says Obladen, a giant slab supported on piles at the side to which the building loads could be transferred. Once in, the timber piles could be removed along with much of the ground to make a space underneath.
But even then a giant cavity around 20m for cut and cover construction would probably be too risky, with some movement possible even with struts or top-down methods. It would require very strong and rigid side walls to resist lateral forces.
Space for work on piles or other structures was also limited with headroom restrictions in the main station hall of around 8m height and less under the roofed platforms. The intention here was to use the existing underpass concourse which gives access to the platforms.
The 30m wide passage from the station itself runs towards the river and could be widened and refitted with a very thick floor slab, which would then act as the top slab for excavations underneath. Together with the station itself this would create a space just under 140m long.
But the under-platform concourse is only 3m high under the concrete troughs running above, which hold the tracks for the trains, and little more under the platforms, creating severe headroom problems for many works.
To cut the risk a solution was devised to create the giant cavity without dewatering. Instead the groundwater, which is high everywhere in the Netherlands, would continue to fill the excavated space balancing the pressures from the ground either side, says Vlaanderen.
To install the necessary tunnel box, including enough space for platforms, in this space an immersed tube method could be used, floating the box underneath the building. It was a method that could take advantage of the river running behind the station, and would be an extension of an immersed tube that was being built for the Ij anyway.
That had its complications since this box is a station section and therefore wider, and it does not connect directly to the Ij crossing; there is an intermediate area of concourses and access stairs and escalators, being built conventionally just behind the main station.
But those issues came later. The first challenge for the main contractor Strukton, and joint venture partner Van Oord, was how to form the side walls to support the top slab of the chamber beneath.
The easy part came first. A diaphragm wall could be made for the end of the box, working on the station forecourt. This would seal the end of the excavation underneath the station and separate it from other construction work outside and the TBM work further south.
Inside the station building and underneath the tracks, however, a diaphragm wall was not possible to create the sides of the box. The rigs are just too big and cause too much vibration.
“Pile driving with vibration machines was also ruled out because we could not cause vibration,” says Obladen, “nor very much noise.”
For the station itself “the only kind of process that would fit all the many constraints was jet grouting,” he says. Rigs could be used inside the entry hall area from which the work was done.
For the track concourse another method was devised, using vertical microtunnelling.
But jet grouting is notoriously fraught with the dangers, both of ground heave and of inadequate column formation from shadowing, ground effects and air effects.
A new form of structure was devised using jet grouting combined with piles in a ‘sandwich wall’ some 2.5m thick. This would be formed with two outer rows of Strukton’s Tubex piles, which could be made with a low headroom machine. Two of these were done 30m apart, each 40m long.
For each, two lines of 457mm diameter piles were drilled at 1m centres. In between, 800mm jet grout columns sealed the spaces. A monojet system was used as the least likely to be problematic in the ground.
“By making two pile and jet column walls with the south end closed by the diaphragm wall and piles along the far end, you could make an enclosure,” says Obladen. Dropping 30m, this ended in a clay layer whose impermeability meant “you had a closed box just under 3m wide to make the sandwich.”
It was safe then to jet grout the interior, he says, because any heave pressure would not escape to the outside ground. Interior columns were of varying sizes overlapping to make a stiff impermeable block.
Obladen, an expert on jet grouting, was closely involved over a long period developing and testing the precise grout mixes and parameters for this work, and various control systems from video monitoring to audio detection of the circular sweep of the grout path. He wrote software to build a three-dimensional representation of the jet grout columns using pressure and volume readings from the grout rigs.
Other developments were needed too including new types of low headroom rigs for the Tubex piles which had enough power to go to 65m depth. Existing machines can manage 30m.
“Most of the wall simply seals the ‘sandwich’ but some of the piles extend further into the third sand layer to carry load,” says Obladen.
In construction, part of the station concourse was sealed off and temporary works used to support the station walls in the sandwich area. Another newly created machine called the Brutus then removed the timber piles without vibration, using powerful jacks to pull them like teeth, sometimes taking three days.
“The timber was in astonishingly good condition,” says Obladen. “Not rotting even two years later in the air.”
Extensive ground monitoring and measurement of the building itself was carried out. “We used all the data to constantly adjust the grouting,” says Obladen. “So you can say it was a true observational method.”
Once made, a concrete framework connected the two walls to take the loads from the main building to the sides, leaving the centre free for excavation of the chamber beneath.
The continuation of the chamber for another 90m under the very low headroom track area was done using another innovation, adapting microtunnel machines to work vertically, to bore very large diameter piles.
Machines were adapted by Strukton’s own design team. The manufacturer IHC made the four machines used.
Piled walls
“There are various issues such as spoil transport which are different. Also as you drill down the pressure increases in a way that is not true horizontally. Bearings need to be different too,” says Obladen.
To avoid blowouts the drill head worked inside a slightly over-diameter cone piece 1.94m in diameter, 20mm more than the cutter. Behind came a 1.82m diameter casing surrounded by an annulus filled with bentonite. “The cone seals the drill hole and is sacrificed at the end,” says Obladen.
A retractable outer ring of teeth was used on the drill head so that when it finished it could be brought back through the casing.
At the top meanwhile was a mobile drilling frame running on rails, with a jacking mechanism to push the 1.8m sections of tubular casing down behind the machine in three passes. A rail mounted crane unit tailored to fit over the jacking frame was used to bring in casings or other equipment. The casing sections were complex needing the various spoil transport tubes, measuring instruments, access ladders and other items according to their position. They were made up in a yard outside Amsterdam and delivered just in time.
“For the pile walls the lifting crane ran on the same rails as the frame and was carefully designed to just fit over the frame within the very tight headroom,” says Obladen. To allow it past, the jacking frame had retractable wheels; the frame itself was locked down during each pile operation with four screw-in Leeuw anchors.
“Anchors were needed to prevent rotation too,” says Obladen. The tubular casings had a clutch piece either side which connected to the next pile in the same way as a sheet pile, and therefore no twisting was allowable. Hydraulic clamps could adjust the position.
“We made every third pile first,” he says. “Because these were 65m deep and would be foundation load carrying. Then the other two were only 30m deep to the impermeable clay as they simply hold back the ground.”
Between the two rows of 2m diameter piles, the original concourse floor has now been replaced with a heftier post-tensioned slab to act as the chamber roof.
It was done in two halves later tensioned together, because the concourse had to be kept in use for passenger access and so could only be done in two halves, with a 15m wide section available. The outer row of microtunnel piles and the half-width of slab were done on one side and then the other with false work columns were created in the centre line, using Leeuw screw piles. That allowed the existing slab to be demolished and the new one be built.
When the station and concourse slabs were ready, excavation could begin. The first 6m depth was done using excavators in the dry in 2009 with walings along the pile walls and struts across for support.
After that, the cavity was flooded and the work continued using portal cranes running along the top and spoil barges floated in from the river outside.
Some 18 months of work completed the excavation with a 1m thick base slab poured underwater in March this year to complete the cavity, leaving everything ready for the immersion operation.
Immersion
While the station was being prepared, starting in 2005 and running through to 2008, the immersed tube has also been under construction. “That was done as part of another contract for the immersed tube crossing of the Ij and work on the northern section,” explains Mark Vlaanderen.
The station work is divided into three sections; the forecourt of the station; the under station work; and the riverside quay area just outside De Ruyterkade station. As it happens the JV between Strukton and Van Oord has all of these contracts, but they remain separate. It is known as Combinatie Strukton van Oord, or CSO.
“There is another contract for the immersed tube crossing of the river Ij as well” says Vlaanderen. Strukton happens to be in this grouping too, which is a JV with Heijmans called Zink; Strukton’s operational arm Strukton Afzinktechnieken Mergor, is doing the actual immersion work.
Even without the overlaps the building of the station segment, along with the four units needed to cross the Ij, made good sense. The crossing JV had a space on the north bank, a coffer-dammed drydock perpendicular to the river. The space will eventually become the approach ramp and cutting for the north-south line into the river tunnel but was first used to build the segments one by one.
“They also did the station segment, which is different to the others,” says Vlaanderen. The river units are just big enough for the twin tracks with a 12m width and 8m depth but the station piece is 21m wide, and 136m long.
All the units were floated out and have been stored in one of Amsterdam’s nearby harbour basins, initially for two years but after rescheduling changes on the project, for four. First to come out was the station piece which was floated into position this summer in an unusually complex operation.
“The problem is to get the unit into the station and underneath a number of struts at the end,” says Vlaanderen. The chamber has remained strutted across underneath the station building end, not because the sandwich walls are not strong enough, but to add an extra layer of risk protection he says, “Just in case.”
To bring in the section a three-stage operation was required with the water level needing to be dropped 1.5m and then 3m inside in order to have room for flotation tanks and other equipment on the top of the segment and to manoeuvre the unit in stages. The first drop in water level is needed to get under the station at all, the second to fit the equipment. The final part underneath the struts and the far end is then done by fully submerging the element enough to pass through.
For the first phase a sheet pile cofferdam was piled 80m out into the river in May. This, together with the 60m length of the quay works excavation would provide the space for initial reception of the unit. The sheetpiles were installed from barges using vibrators, pushing the piles down 32m, the last two metres keying into an impermeable clay. “There is no sand to get through here in the river,” says Vlaanderen.
At the end of May the big unit was floated from its storage point in a carefully planned operation using four tugs pulling and one pushing. Though a straightforward operation, care had to be taken not to hit the quays for the famous free pedestrian ferries which link the north of the city to the south by the station.
Once in the prepared box the end could be sealed with sheet piles and the water pumped down 1.5m to allow the next stage. “It was not possible to do the whole 3m drop in one go because the cofferdam would have needed to be hugely strong and therefore expensive,” he explains.
But the 1.5m drop was sufficient for the levels to match those under the station. A temporary works wall sealing off the station was removed therefore and the unit gingerly pulled forwards.
“We have a winch wire running over a pulley fixed on the diaphragm wall which separates the chamber from the forecourt area,” says Vlaanderen. The pulley cable used only 2t of force on the 20,000t unit to give it a gentle motion in the water.
The unit was halted 40m from the end wall because of the struts above. But that was enough to bring the far end within the quayside works space which could then be sealed with a cofferdam gate on the riverside which was much heftier than the first. “It is part of temporary works for later when the quayside works are dewatered to 18m depth,” says Vlaanderen. “But it is much shorter than the extension into the river and therefore less cost”.
Now the water level could be further dropped by an additional 1.5m creating the space to fit out the unit with immersion pontoons filled with suspension winches, piping for later sand backfill operations, a primary tower for survey purposes, which is folded down and erected later.
All this was for the immersion operation itself, which would guide the unit underneath the far struts and into its final position. Water-filled tanks were used inside for ballast and the units were winched from the flotation units above, along with a pull force from the end pulley.
“It is essentially a standard immersion with which we are very familiar,” he says. He adds that the Busan-Geoje crossing in Korea with summer typhoons, deep water and open seas was much more difficult. The company was a specialist subcontractor there.
But there were unusual risks of collision with the walls and with the new foundation of the station.
Positioning of the unit was slightly unusual. Taking advantage of the internal walls four wood-faced columns were used as guides at the side. Each was fitted with flat pneumatic jacks, flat air bags used in the oil industry to lift heavy objects. By pumping in air they could adjust the plan position of the unit and even rotate its longitudinal alignment slightly. It had to fit with a plus or minus 20mm tolerance.
The end position was adjusted too using steel shim plates inserted from a gantry on the front of the unit. Until the float-in and lifting of the primary survey tower on the unit the exact position of the 1m thick diaphragm wall could only be measured from the far side on the forecourt works.
A more unusual feature of the immersion involved the unit being hung from the ceiling slab, rather than being positioned onto prepared foundation pads as it would be in a normal open trench. But the underwater concrete slab was already serving a structural function and it was better not to do anything to it.
To hang it, special steel brackets had been anchored into the ceiling at four points. On the floating segment were four substantial steel beams, which were on hinges and lying flat. Once in place they were raised up and attached to the ceiling brackets, an operation requiring some delicacy in the positioning.
That done the ballast could be increased to load each of these hangers to 150t rather than the 25t or so during immersion. That made the whole structure much heavier, in preparation for sandflow underneath and for backfilling, due in October.
That awaits completion of a new bulkhead to seal the under station area and simultaneously create a complete cofferdam around the cut and cover works on the Ruyterkade Quay area.
Inside a sand mixer unit on the deck delivers the backfill through 20 pipes to the underside arranged alternately inside and outside. Under each outlet ‘pancakes’ of sand build up.
As they do so the back-pressure on the concrete unit begins to increase, says Vlaanderen, and this is measured by the reduction in load on the four hanging beams, each of which is fitted with hydraulic jacks for fine adjustments. The reduction in load allows the measurement.
With final position adjustment done the backfilling will be done to about one metre, to fix the box position. Then the hangers can be removed and the load taken by the sand ready for a final backfilling in October.
The finished box will still be isolated and must be connected through from the contract in the station forecourt where the main ticket hall and entrances are being built. “We shall use, or rather the contractor for the forecourt, will use ground freezing for that operation,” says Vlaanderen. After that will come track and signal works.
Meanwhile in the 60m box at the quay end, the other entrances and access points will be created in a dry dewatered space inside the now complete sheet pile area. This will also be connected to the under station box eventually by ground freezing.
And next year comes the immersed tube operation across the Ij river.
Figure 1, Amsterdam’s north-south line alignment showing Centraal Station on the south bank of the river Ij Settlement under Centraal Station could cause damage to the station’s glass roof Figures 2 and 3, showing the geology and location of the existing and new piles The grand welcome of Amsterdam’s Centraal station Figure 5, excavation of the chamber beneath the concourse Figures, 6 and 7, visualisations of the front and rear of Centraal Station as the immersed tube segment is floated in Brackets were installed in the concrete floor slab to hang the immersed station box Figure 4, a visualisation of piles being bored within Centraal Station
