From the beginning it was supposed the most hazardous, contentious and difficult part of the North-South line. The underground tunnelling work through the old city centre of Amsterdam, a national and world heritage site, unique in the world, carried risk and public mistrust. In the event so far the TBM drives have gone exceptionally well.
A total 3.1km of 6.5m outer diameter double bore tunnel plus the length of three deep stations makes up the 3.8km long central section. The tunnel itself is driven close to, and at one point underneath, the brick and timber buildings of the 16th to 18th centuries, all founded on ancient timber piles through soft, difficult and, above all, waterlogged peaty ground. All have suffered movement over the ages and are sensitive. Even outside the historic centre itself there are further pile-founded buildings from the end of the Second World War and equally vulnerable.
Most of the route is parallel twin bores at depth of 18-20m for the tunnel axis in between stations and rising slightly at the stations. For the narrow Ceintuurbaan station it is necessary to bring the tracks in one above the other and the western side tunnel swoops away from the other and goes deeper therefore to about 25m, passing underneath some buildings on its alignment. This is usually forbidden in the Netherlands and required special legislative dispensation.
It was felt that modern TBM methodology could make the drives but it was necessary to develop new levels of tunnelling control and understanding of building foundation behaviour says Frank Kaalberg. He is design team leader for the bored tunnels, from consultant Witteween + Bos which is part of the client’s project engineer, joint venture Tunnel Enginering Consultants. Royal Haskoning forms the other half of the JV.
His team’s work, over nearly two decades has involved both substantial research on the behaviour of the city’s piles, a major survey of the buildings along the alignment, complex 4D finite elements analysis and development of a new level of real-time settlement monitoring above ground. The latter interacts with on-board TBM control systems to give much more precise control of excavation, ring installation and grouting. The whole has been christened the ‘Integrated Boring Control System’ or IBCS.
The first task was to analyse what alignments were possible. "The problem is both to avoid the piles in the ground and to avoid disrupting their bearing mechanism" says Frank Kaalberg. "There are thousands of timber piles usually driven into the dense ‘first sand layer’ about 12 to 14m down. They have some skin friction but also end bearing from a compacted cone formed around the pile tip. You have to avoid straight settlement and stress relief effects on this compacted material."
Conventional advice mostly suggested a machine would have to stay at least two diameters away, based on estimated volume losses on the tunnel line of one per cent to three per cent, causing settlement of 100mm or more above.
"But that would mean staying clear of buildings and driving very deep, which added costs at stations and means deep passenger access is needed," says Kaalberg. "You can route down the streets but they are narrow and anyway even they have pile supports for services and sewer lines."
He believed volume loss could be much less, but there was little published information on that or on timber pile behaviour in the Netherlands. In the 1990s his firm did a series of tests with specially installed piles at the second Heinenoort tunnel, one of the first bored tunnels in this waterlogged country. Using careful control of TBM grout injection it showed losses could be kept to 0.5 per cent. A half diameter was possible, and a distance of half a diameter between the two bores.
The results meant that a shallower, economic alignment could be used. Even so special measures were needed. First was to examine all the buildings in the city and assess their condition. Even though historic building regulations stipulate foundations be kept in good condition, many were unsound, "around 25 per cent," says Kaalberg.
A major survey and finite element analysis of building response was used to classify buildings at risk and notify owners. A 30 per cent subsidy from the city and insurance requirements then stimulated owners to make remedies with 98 per cent of buildings brought to a stable condition.
Monitoring
Meanwhile a complex monitoring system was devised using multiple robotic total stations set at key points along the alignment. Each takes round the clock readings from prism reflectors mounted at three levels on all the building façades, more than 5,500 in all. Readings feed into a massive GIS database with hourly observations, which can track the settlement wave caused by TBM progress.
The original system has been upgraded with the advent of modern laser-based total stations which do not require prisms for accurate readings. These now supplement the building readings with distance reflections direct from the road surface, which means settlement effects can be detected earlier and more directly. Total station points have two instruments in the zone of TBM operations, the laser based ones moved along from point to point with the tunnelling.
The output from this system feeds into the operation of the TBMs below and is one factor in controlling settlement.
TBM design
First however was a special design for the 6.8m diameter Mixshields, devised in conjunction with maker Herrenknecht. The overarching principle was to shorten the length of the TBM head as much as possible explains Joost Joustra, the tunnelling project engineer for the Amsterdam City client.
"The shorter the machine head, the better you can get around curves with minimal ground displacement, which is always a significant factor affecting settlement" says Joustra. Measures include a ‘bubble’ chamber with its entry airlock extended into the chamber, rather than sitting behind it further along the machine. The compressed air bubble used to maintain the slurry pressure in the face chamber therefore surrounds the airlock.
Unusual three section telescoping hydraulic rams were also fitted to the machine for its push forwards from the segment ring, shortening them and their mounting points.
The shortened shield also articulates at the 55 per cent point along its length to further tighten turning capacity, helping it achieve bends in the alignment down to a 140m radius.
"That is encountered at around the point of the Royal Palace," says Joustra, "where the street pattern changes and there are turns nearly as tight."
"The machine is also rather specialized at the front" adds Frank Kaalberg "because it has a very open face with large apertures to cope with the soft ground and clay."
It would have been even more open he says but for the need also to carry disc cutters. These are required because the TBMs have to bore through the up to 1.5m thickness of concrete of the deep station diaphragm walls, which is obviously hard. "And because of the delays on the project it has hardened even more than expected," Kaalberg explains.
Transfer chamber
The shorter machine also aided design of another special feature, a transfer chamber made for the transition through the deep stations. The original sequence for the TBM drives involved two machines driving from the start chamber in a big caisson in the Damrak in front of the Central Station. From here the machines would arrive at the first deep station at the Rokin, pass through and move to the other end before recommencing the drive, passing in turn through the other two deep stations.
Entering the station means coping with the groundwater pressure at machine depth as the TBM’s pressurised slurry-filled cutterhead breaks through. "Normally you would use a concrete block in the ground outside the diaphragm wall to make a seal but that requires a steel retaining wall to excavate inside to make it, which is difficult in such deep ground" says Kaalberg.
"You could use a diaphragm wall but that needs a lot of space and a grouted block is difficult too."
The solution devised is a pressurized steel chamber, held with steel reaction struts and sealed against the inner side of the diaphragm wall. It is filled with water at the same pressure as the tunnel face.
"As it cuts through the machine enters this ‘Shield Transfer System’ (STS) and displaces the water," explains Kaalberg. It continues making segment rings up to and in the diaphragm wall, grouted as usual and then additionally sealed. That done the STS can be de-pressurised and drained.
"It then ‘walks’ forwards on a series of vertical jacks," says Kaalberg. This copes with a second problem which is a deep point in concrete floor at the station ends where the base of future escalators will be accommodated. "Further on the floor is higher because we did not want to excavate any more than necessary, particularly to save on air pressure."
After ‘caterpillaring’ to the end, the chamber is again sealed against the wall for the tunnel exit. The TBM erects temporary steel segments within the chamber as it does so, the first with a bulkhead seal for the annulus which contains the pressure. That leaves the temporary tunnel interior open for the back-up train supplying normal concrete segments as the shield end reaches the diaphragm wall.
The STS has not been used as much as supposed however because the TBM drives were rescheduled to cope with delays at the deep stations, most of all the middle Vijzelgracht station which was not ready for arrival of the machines.
"The first drives began from the start-chamber at central station in spring 2010 with the second in August," says Joustra.
These had a complex start as they passed under the water-filled basin of the wet Damrak, where the tourist boats leave. The alignment began at a relatively shallow depth and headed a bit deeper while curving towards the main street.
After discussions with the contractor is was decided to do this work underneath overlapping ground freeze umbrellas about 2-3m thick over the two bores for the first 85m. "That gave additional security and prevented the possibility of a blowout early on" says Kaalberg. Once through, the section the 750m long drives went well, and the settlements were well within the tolerances. "Up to 25mm is allowable but although it was around 20mm at the beginning, mostly it has been less than 10mm and even approaching zero," says Joustra.
"It took a while to learn the system," says Wolf Friedemann, project manager for the TBM tunnel contractor Saturn, which is a joint venture of Germany’s Zublin and Dutch firm Dura Vermeer. "You have to take great care with the grouting and make sure that you inject 6.5m3 to fill the 6m3 space in the annulus. I was a little nervous about it at the beginning."
The TBM operator has information to monitor from pressure detectors in the tailskin and from the ITBS above ground he says. "The pressures change all the time so he has to be vigilant."
But his work team of mainly German tunnellers has got the procedures down to a fine art he says and on the next drives high speeds have been achieved.
These drives however have been made from the other end of the alignment because the ordering of the drives was changed to cope with delays and difficulties on the deep stations.
"Work at Ceintuurbaan, expected to be the last to finish, went better than the rest and so it was decided to do the rest from the south," explains Joustra. The first TBM shields would be sacrificed, and all the equipment transferred to the southern end of the line.
"This was possible because an initial cut and cover section of the line which starts it from the surface south of the main exhibition centre, had been completed" he says. That gave relatively easy access; earlier a southern start was ruled out because it needed shaft works, disrupting Amsterdam’s important international exhibitions business.
To make the change, new TBM shields were ordered and the important equipment and motors inside the first shields were stripped out to rebuild the TBMs. The bentonite plant, grout plant segment supply system and spoil disposal systems were also moved.
"Fortunately," says Joustra "there was a disused area close to the Amstel river, which local residents suggested. It is close to the city ring motorway and has access to barge loading for spoil." A new pipeline runs there. The material cleaned from the bentonite is carried to a section of the old deep harbour which is being partially filled to create shallow water for development.
The two new machines are now driving north, first making the 950m bore to the Ceintuurbaan, before begin transferred through with the STS and then onwards.
"Because of the extended completion time the drives are made one at a time, with one set of shift workers operating each TBM in turn," says Joustra.
Drives for the western side were begun in May 2011 and made through Ceintuurbaan to the Vijzelgracht station where the TBM has been ‘parked’. Meanwhile the other side began in December last year and completed as far as Ceintuurbaan by Easter. The sixth eastern side drive was underway in early summer and the crews were due to switch sides again for the final western drive after that and the final, eighth drive, to complete early next year.
