If ever a race or comparison of TBMs were to be made you would want to equalise, as much as possible, the conditions facing both machines. Ground conditions should be equal, with limited variation and few localised faults or intrusions that might affect progress in one but not the other. The tunnel lengths would have to be identical and preferably sufficiently great to average out any temporary quirks in boring rates. A long tunnel would also test stamina and staying power. You could even make sure the machine manufacturers were from similar economic and manufacturing backgrounds so that the technology alone was up for assessment. Tunnel design would need to be from the same source and the methodology strictly controlled to ensure that, apart from minor stylistic issues, the contractors’ operations were the same. Finally, you could double the experiment, carrying out the same test from either end of the tunnel.
It is a highly improbable combination. But even so, all of these conditions apply at the Guadarrama high-speed rail tunnel project, just north of Madrid in central Spain. Here two 9.5m diameter Herrenknecht machines and two 9.5m diameter Wirth/NFM hard rock TBMs (German and German-Franco) are driving a passage underneath the Guadarrama national park, a medium height mountain range north of Madrid (Figure 1).
Both contractors have one machine of each kind. The tunnel is very long, with twin parallel bores of 28.3km, on axes just 30m apart. Because of environmental restrictions, with the route passing under highly protected natural countryside (Figure 2), no construction adits or shafts are allowed. Access is only through the portals, which means a continuous drive of over 14km is required from each machine.
Of course this is not a race as such, though the tunnelling industry will undoubtedly be monitoring the comparison. And so far, almost one year into the project, it is much too early on to make any meaningful assessments anyway. Even so, the Guadarrama tunnel is already of great interest as one of the world’s longest hard rock tunnel drives; and also as a precursor perhaps to a number of future long tunnels in Spain as the high speed rail network is extended on a vast scale.
Guadarrama is part of a north west high-speed rail connection, a line that will run from Madrid to the once Roman city of Segovia and onwards to Vallalodid and then the north west provinces. The line is not due to be completed until 2007 and much of its construction, across the relatively flat Spanish plains, will begin later. The tunnelling work needs to begin earlier however, to take the route under the only major natural obstacle on the line, a mountain range between Madrid and the city of Segovia 90km northwards.
Not only must the line pass the 2,000m height of the mountains but it must also avoid disrupting what is a highly valued natural area. Once the cool summer retreat for the Spanish aristocracy, a mainstay for modern Madrid recreation and also, now and historically, the source of most of its water supplies. So important is the area that the routing caused some controversy earlier on. Though client GIF, the government owned infrastructure company for the high-speed works, does not comment extensively on the protests, they had enough impact to change the originally planned design for the tunnels. “The alignment was increased by about 3km,” says Manuel Herrera Alvarez, work director for the supervising and co-ordinating engineer Ineco, “which put the portals outside the protected areas.” Even then he says there are restrictions on operations at the sites.
For both contracting groups the establishment of segment factories and muck disposal has been a key part of getting sites up and running. Both have reasonable space on the north and south edges of the national park. Both have created two steam-cured segment lines using French-made equipment. They reach a 10N/mm2 strength after six hours and 40N after 28 days “though the early strength for handling is the most important factor,” says Ignacio Botella Rodriguèz, project manager on the northern drives for the FCC/Ferrovial led consortium. The 320mm thick, 1.6m long segments are variable in size with a larger invert element, supporting a construction railway for deliveries and access, and a small key. Configuration is six segments plus a key. Tunnel spoil is used as aggregate for concrete, after washing and sorting. “That helps with spoil disposal as well,” says Teador Comeaga, one of the 20-strong team of engineers with Ineco, “about 10% of the rock goes into the segments.”
Strict controls apply to the disposal of the remainder of the approximately 1Mm3 of rock excavated from each of the four drives. Spoil is removed from the tunnels by conveyor, using Continental Conveyor units on the northern contract and HDI equipment on the southern contract. Though it is good quality rock, there is a ban on its sale in the north, where it must be used for infilling old quarry workings and other environmental improvements. In the south the contract includes a requirement to connect the site to a conveniently situated nearby railway line, from which trains can haul the spoil for disposal. Environmental restrictions also include keeping the washing and concrete mix water in a closed circuit; the area is a significant catchment for drinking water reservoirs.
There are also restrictions on excavations within the national park. Essentially there are to be none, which has ruled out an initially planned central shaft for ventilation and escape, and any chance of additional working faces. That said, the tunnel alignment remains very straight-forwards, quite literally. For high-speed rail use, the reduction of curves and gradients is important and the line of the tunnel therefore includes no curve greater than 7,000m radius. The gradient is also limited to a maximum 1.25% as the tunnel climbs from a southern portal, at about 1,000m altitude, to a high point in the centre of 1,200m. A slightly gentler slope of 0.95% then takes the line back down towards Segovia on the high plains. At its highest, overburden will be nearly 1,000m, though most of the tunnel has less.
The tunnel route has not been unduly affected by the area’s geology, which is consistent and relatively free of difficulties. The mountains are ancient igneous and metamorphic rocks, primarily granites and gneiss. Hardness varies from around 80MPa to 200MPa and in the main there is little water expected. “The bulk of the rock is gneiss, about 60% to 65% with granites comprising 20% to 25% and microgranites another 3%-4%,” says Comeaga. “That is, the rock mass is almost entirely composed of granites, ademellites, granodiorites, augen gneiss, pellite gneiss and micaschist, all of which are amenable to hard rock excavation methods and which are in the main self-supporting.”
All that said, rock is not without potential problems. There is an 8% presence of dykes or intrusions across sedimentary layers he says which might lead to filtration problems. And despite an overall dry tunnel drive expected there could be some water. “The structural evolution of the rock through the two orogenetic cycles that have affected them, the Hercynian and Alpine periods, has led to important anostropies that have an effect on geotechnical characteristics and possible circulation and catchment of water. There could be authentic hydrogeological system, which while local would have a significant impact, considering the scale of the massif,” says Comeaga.
One significant potential difficulty that confronts the tunnellers, is a major fault towards the centre of the tunnel line, some 500m wide and filled with weathered and fractured gneiss mixed with clays. “In the north we also have a shorter difficult point to cross under a river valley, the Valparaisa,” says Herrera. Ground cover here will be about 20m. The big, Umbria, fault and the presence of soft and fractured ground at the beginning of the drives led the client GIF to specify the use of an unusual shielded TBM format for the rock drives, incorporating the kind of ground treatment equipment that might be found on a soft ground machine.
The potential for a fast drive is increased by the use of telescopic double shield designs for all four of the machines, which are similar in overall concept, though different in their detailed implementation. They can operate in two modes; primarily as a hard rock machine but with the capacity to behave partly like an EPBM.
“For hard rock operation the machines use twin grippers on the shield to take the thrust,” says Comeaga. These are larger on the Herrenknecht machines with a surface area of 12m2 each, whereas the Wirths use 7.9m2 and a larger thrust force, 89,000kN compared to 80,000kN on the Herrenknecht. The primary advantage of the double shield layout is to allow simultaneous operation of the cutting head and placement of segments. The annulus is backfilled from the machines with pea-gravel and then grout injection. For any difficult soft ground, and particularly across the Umbria fault, the machines will operate in a single shield mode, thrusting from the segment ring and then stopping to install the next segments. This mode should be needed for only about 10% of the drives. “It will mean slower progress at these points of course,” says Herrera “even more so because the main fault will need forward ground treatment about 50m to 70m ahead”.
There will be additional works in the tunnels. Cross passages have to be built every 250m as the main emergency and escape system. These will be constructed using drill and blast methods. In operation the passages will be closed, to prevent smoke escape but accessible by emergency doors for escape into the other tunnel, which will provide the main route for rescue. “There will also be a central emergency area at the half way point,” explains Herrera. “This will be a 500m long section with additional crossovers – at every 50m and a central longitudinal tunnel in between the other two which will be a refuge point.”
But that is to come. The contractors at present have been climbing through a steep learning curve on the machines since the drives began last year. Hard rock tunnelling on any large scale is relatively unfamiliar in Spain and even more so for such a long tunnel. The work crews need to build up experience.
There are no lack of tunnellers in the Madrid region, with work on the metro extensions recently completed, but they bring a soft ground knowledge with them rather than the different skills of hard rock. “Additionally we have been unable to find any data on a hard rock drive of this length where there are no additional access points,” says Botella. “Perhaps there is something from New Zealand, but even the machine makers told us this is primarily new territory.” He says that with machines now operating since the middle of last year “we have passed through the learning curve for their operation.”
“But we still have to increase speed. We achieved 600m in the last thirty days on the Herrenknecht and 450m on the Wirth but we really want to be doing 650m in normal ground.” The problem he says is not to improve machine operation, which is good, but reduce downtime. At present the machines are inspected every three to eight strokes, in practice this means a one hour inspection daily, with tools changed if necessary. Cutter wear has been higher than expected so far says Botella, possibly because the TBMs have found a high proportion of harder rock. He says that it is far too early to make any comparisons between the machines. “They feel very different in use however” says Botella “despite similar parameters.”
In fact both machines are quite similar in specification adds Herrera, “there are some minor differences – the Wirth has slightly more power, but it is not overly significant.” Different rates seen so far do not really reflect on the machines, which have seen some differences in ground conditions. Additionally, one machine started behind the other, the Wirth not beginning until December last year.
The real test of the machines comes later, as the long distance is covered. A big question then will be the resilience of the machines and the need for large scale maintenance. “So we can only draw conclusions if at all, when we finish,” concludes Herrera. And that will not be until March 2005.
Related Files
Figure 3 – Cross section of the completed Guadarrama tunnels
Figure 2 – Detailed location of the Guadarrama tunnels, just north of Madrid (see Fig 1)
Figure 1 – Location of the project
