Alpine baseline railway tribulations May 1997

Shani Wallis, TunnelTalk

There is great satisfaction in conquering the impregnability of the Alps with road, rail and hydroelectricity tunnels but the achievements come at a cost. At a time when tunnellers are celebrating the breakthrough of the Vereina rail tunnel in Switzerland, some five months ahead of programme, others are recovering from the flooding of the Semmering pilot tunnel in Austria. Closer inspection of these two tunnels and their tales illustrate how it is not only the geology but also politics and project planning which influence successful tunnelling. The technical and practical experiences from these two projects will carry forward to the proposed excavation of the long base-line railway tunnels planned to cut through the Alps in Switzerland and Austria and to high overburden mountain tunnels elsewhere.

Semmering pilot tunnel

At the end of March 1997, engineers working on the Semmering pilot tunnel in Austria had nearly recovered the face following an inrush of water which brought tunnelling to a standstill in October 1996. Over the past six months, Porr-Technobau, the contractor, has been pumping about 200 litre/sec non-stop day and night, 7 days a week to pump out the flooded downhill heading. More than 2 million m3 of water have been pumped from the tunnel.

Semmering baseline project within the new rail links in Austria
Semmering baseline project within the new rail links in Austria

Drill+blast of the 16 m2 pilot tunnel had progressed some 2.3km when water ingress of more than 350 litre/sec overpowered the tunnel’s pumping system causing the heading to be abandoned. Left behind was the Atlas Copco two-boom drilling jumbo, the Schaeff-ITC excavator and the 300m conveyor system which transferred muck back to the muck trains. This conveyor was anchored to the crown of the tunnel and prevented withdrawal of the face equipment without first being dismantled. A control bulkhead was hastily erected but this was unable to hold the inundation pressure.

Ingress eventually slowed to a steady 200 litre/sec but inundation filled the slight 1% downhill tunnel to within 25m of the portal - a volume of 30,000m3.

Water ingress was no surprise according to Geoconsult, one of a group of design engineers involved on the proposed 22km base main tunnel through Semmering massif and its 10 km-long pilot. Water was well predicted in this and in a second aquifer some 8 km into the pilot tunnel alignment and probing with a minimum 40m long probe and a minimum 10m overlap was a contract specification. The surprise was the quantity. The pumping system installed by Porr was designed to deal with 140 litre/sec, double the maximum 70 litre/sec expected at the face. Inundation of 350 litre/sec illustrates the difficulty of quantifying expected water ingress.

Water ingress would not have been a probIem on an uphill tunnel but because of political differences this was impossible at Semmering. The portals of the 110m2 x 22km single-tube, double-track Semmering railway tunnel lie in two different local government constituencies, one of which supports the project and the other of which is adamantly opposed and supported in its opposition by environmentalists and other lobby groups. This imposed a downhill operation from the south portal rather than an uphill heading from the north.

The Semmering Pass is one of the busiest north-south international road and rail routes across the Austrian Alps and is being upgraded to meet higher rail traffic demand. The 22 km-long baseline tunnel is one of many projects adopted by the Austrian Government to improve rail links across the country.

One of the most ambitious proposals is a 350 km-long baseline rail route beneath the Brenner Pass in the Tyrol, some 83% of which would be in single-track, twin tube tunnels.

A pumping train with an installed capacity of 400 litres/sec recovered about 45m of flooded tunnel/day, pumping water against a maximum head of 20-25m
A pumping train with an installed capacity of 400 litres/sec recovered about 45m of flooded tunnel/day, pumping water against a maximum head of 20-25m


Recovery of the tunnel was based on a pumping train designed and assembled by Porr. The train carried eight 110kW pumps providing a total pumping capacity of 400 litre/sec. It was, however, very difficult to keep all the pumps and electrical systems working at full capacity round the clock, according to Martin Diewald, site manager for Porr. The pumping train was launched in mid-January and had since maintained a non-stop average pumping operation of 220-250 litre/sec to successfully recover about 45m of the tunnel/day using 6m lengths of pipe to extend the pumping pipeline behind the advancing train. On reaching the face, the pumps were lifting water against a head of about 20-25m along the 2.3km downhill tunnel.

At the end of March, the tunnel face had been regained and the flooded equipment had been retrieved. At the time of writing, fortunately there has been no collapse, deterioration or failure of the tunnel. The dry-mix shotcrete lining remains intact and work was expected to resume in April.

Since October 1996, discussions between client, engineer and contractor have centred on exactly how to proceed. The client, HL-AG, Eisenbahn-Hochleistungsstrecken AG for the Austrian Government remains committed to the project although the cost of continuing is a concern. For the engineer, the event has quantified the water potential, and methods of coping with this known situation can be designed. Renegotiating the contract in the light of extra works completed and now envisaged is the priority for the contractor. Porr was awarded the five-year pilot tunnel contract in 1994 for a tender value of ATS450 million (about US$37.5 million).

Way forward

Of the many methods discussed, continued dewatering will be required through the current aquifer which is expected to extend for the next 600m to 1,000m. When encountered in October 1996, water ingress was under about 10bar pressure indicating a head of about 50m above the tunnel. Pumping to date is predicted to have lowered this to about 15m above the tunnel.

Pipe mains lead away from the back of the pumping train used to pump out the flooded downhill pilot tunnel

The immediate plan for progress is to establish an off-line dewatering installation adjacent to the current face. Within this chamber up to 10 deep well boreholes will be sunk to provide a total pumping capacity of 200 litre/sec. The intention is to reduce the water table in the aquifer to about 20-30m below the tunnel invert to present a dry working face through the aquifer. At about 500m advance it is expected that another dewatering installation will be needed to keep the water table lowered.

On the broader scale, other proposals are being discussed. Attempts continue to gain access to the north portal to start excavation of the main tunnel on the uphill gradient. Another plan being discussed is to start excavation of the main tunnel from the downhill south portal. A third idea is to stop the pilot tunnel after passing through this current aquifer and wait for an uphill operation from the north portal to provide access for an uphill north to south operation to complete the last 2km of the 10km pilot. The pilot is needed as an independent offline emergency escape route and will be connected to the main tunnel with cross passages at 1.5km intervals.

Predicting conditions in the second aquifer at pilot tunnel chainage 8,000m will be equally as difficult as it was for the first. The pressure head for the tunnel's pumping system at the 8km mark on a downhill heading will have increased to about 100m. Hydrological models are being studied at Joanneum Research in Graz to help predict the nature of the second aquifer following the experience of the first. In March 1997, while progress was about to resume from where it left off, the entire project - its purpose, its cost, and the alternatives - remained topics of lively political debate.

Vereina breakthrough

0n 26 March 1997, breakthrough of the 19km Vereina tunnel was celebrated when the 7.64m diameter Wirth TBM working from the north cut through into the drill+blast work advanced from the south. Junction at 11.5km along the 19km tunnel comes five months ahead of programme, according to the breakthrough announcement, and within the project's SFr340 million (US$260 million) tunnel excavation budget.

19km Vereina tunnel excavated by TBM from the north and drill+blast from the south
19km Vereina tunnel excavated by TBM from the north and drill+blast from the south

Originally, the meeting point between the north portal TBM consortium headed by Stuag and the south portal drill+blast consortium led by Zschokke was expected in October 1997 and at the tunnel's 12.5km mark. However, while the TBM did reach maximum advance rates of 43m/day in the most favourable rock, it was steady advance through good quality rock by the drill+blast which ensured early completion. Working with an Atlas Copco 3-boom drilling jumbo the tunnelling crews advanced the full 38m2 tunnel at an average 12m/day working three 8h shifts/24h day, 5 days/week to arrive at the appointed 12.5km meeting point in mid-1996, some 14 months ahead of schedule.

In November 1995, with the 46m2 TBM running behind schedule, it was decided to move the meeting point to the 11.5km mark and extend the drill+blast heading by a 1,000m thus amortising the TBM's delay and achieving tunnel breakthrough five months ahead of the original schedule. Having achieved best rates of 14.7m/day the drill+blast reached the revised 11.5km meeting point in January 1997 and has since been building the reception and dismantling chamber for the TBM.

The most difficult section for the TBM was at about chainage 3.5m when it became stuck in a section of crushed and wet mylonitic gneiss. The cutterhead had to be cleared manually, and with each subsequent rotation loose friable material rushed into the face causing voids above and in advance of the cutterhead. It took more than a month to work through 50m of this difficult rock in which deformation was also high.

Rock conditions in the TBM tunnel were classified into six support categories by tunnel designer Amberg Ingenieurbüro which is also project manager and construction supervisor. Precast concrete segments were placed in the invert behind the TBM and the arch was supported with various combinations of fibre-glass rockbolts, layers of wet-mix shotcrete reinforced with wire mesh, and steel arches were necessary on 80cm or 1.6m spacings to match the TBM's gripper stroke and the prevailing rock conditions.

At the start of the drive, significantly more arches were installed than anticipated. These were installed immediately behind the cutteread and limited TBM advance rates to about 25m/day working two 9h production shifts/day with a 6h maintenance shift. Not until the TBM passed into the Amphibolite massif at chainage 4,770m did the need for arches disappear and TBM rates increase significantly to more than 40m/day.

Final lining

Shotcrete provides both the primary and final permanent lining of the Vereina Tunnel in all but a 300m long section at the start of the north portal. Designed by Amberg, the finish is a monocoque lining which is based on corrosion resistant fibreglass rockbolts and incorporates the properties of the primary lining as contributory to the final design. The lining is designed as an equally durable, yet less expensive, finish to install than conventional in-situ concrete. Rough estimates by Amberg conclude that a final shotcrete lining is 10-20% cheaper to install than in-situ concrete.

The final lining shotcrete train in the north heading
The final lining shotcrete train in the north heading

The design specification for the primary and final shotcrete requires an early strength of 10N/mm2 after 12 hours and a minimum strength of between 30-40N/mm2 after 28 days. It also specifies a water penetration of less than 20mm when exposed to 5bar water pressure. Given the choice, the consortia on both the north and south portal contracts chose wet mix shotcrete over the dry mix alternative.

During excavation, primary shotcrete was applied behind the TBM either by hand directly behind the cutterhead if necessary or from a station on the trailing back-up about 60m behind the face where the nozzle was mounted on a travelling frame custom-designed to ensure a uniform covering of shotcrete over the arch from one side of the precast concrete segmental invert to the other. The shotcrete station also included a device to catch rebound and lift it onto the TBM's transfer conveyor for removal in the muck cars.

As part of the expedient construction programme adopted for the Vereina Tunnel, the final shotcrete lining in both the north and south tunnel headings started before excavation finished. The final shotcrete lining in the TBM tunnel started in January 1996 using a rail-mounted shotcreting train which incorporates a partitioning wall which protected the second rail which continued to support the advancing TBM. The shotcrete train is applying the final lining in 500m lengths, completing one side of the tunnel before switching the nozzle to the other side of the partition wall to complete the other side. Each 500m section takes about 15 days to complete and at the end of March 1997 some 6km of the final lining had been completed.

In the south heading, final shotcrete lining started in Spring 1996 and to late March some 6.3km had been completed. During excavation, a special bridge structure maintaining service traffic and allowed the final in-situ concrete invert to be cast about 70m behind the advancing face.

After testing most of the various additives on the market, the best wet mix shotcrete for both the final lining shotcrete and in both the north and south heading tunnels includes chemical additives supplied by Sika. The best quality mix comprises 400-450kg of cement/m3 with a 20kg/m3 silica fume content; a low water:cement ratio of <0.5; maximum aggregate of 8mm; to which is added at the hatching plant 1-1.5% SikaTard-903 liquid stabiliser/superplasticiser. At the nozzle on the south side, the contractor is adding 4% of Sika's Sigunit-120 liquid accelerator. The north heading contractor is adding at the nozzle 3% Sigunit4 9AF, Sika's alkali-free powder accelerator. This is added by the Aliva 405 powder additive dosing unit and is said to provide the best long-term quality shotcrete with strengths of between 50 and 60N/mm2 after 28 days which continue to develop.

To complete the tunnel, a central 2.1km long double-track cross-over is to be enlarged by drill+blast from the TBM drive of the singletube, single-track rail tunnel. Current schedules indicate that the new SFr538 million, 19km Vereina Tunnel with its 2km Zugwald tunnel extension will come into operation in 1999 to provide a vital through train link in Switzerland's eastern rail network and a car-carrying all-weather train shuttle service to and from the ski resort at Klosters.

Lötschberg and St Gotthard baseline updates

Meanwhile, work progresses on the pilot tunnels ahead of main line excavation of the 42km Lötschberg and 51km St Gotthard base line rail tunnels in Switzerland. As these advanced works continue, an enormous cloud in the form of financial uncertainty hangs over the project as politicians in the Swiss Parliament argue about the justification of building both baseline tunnel projects, the financial preference of one project over the other, and the credibility of the government's suggested methods of raising the enormous funds (even by Swiss standards) needed to cover the cost of the AlpTransit projects even in their revised, much slimmed own versions of original aspirations.

Fig 1. Proposed extent of the European Union rail traffic infrastructure by 2010 with the two baseline projects across Switzerland at the heart of the network
Fig 1. Proposed extent of the European Union rail traffic infrastructure by 2010 with the two baseline projects across Switzerland at the heart of the network

For the St Gotthard, the advanced works centre on two projects - the Sedrun intermediate access adit and the Faido/Polmengo exploratory and access site designed to investigate the Piora basin of potentially extremely difficult soft sugar-like granular dolomite lying between the Gotthard Massif of granites and gneisses and the Penninic Gneiss Zone.

Work on the Piora Basin exploratory adit some 300m above the proposed base tunnel alignment started in September 1993 with the launch of a 5m diameter Wirth TBM. In March 1996 the TBM had advanced some 5.5km when the dolomites were intersected by a routine 42m long x 98mm diameter probe hole and up to 400 litre/sec of water carrying the sugar grained dolomite poured into the tunnel. Initial pressures were more than 90bar. More than 4,200m3 of water carrying some 1,400m3 of dolomite poured into the tunnel, flowed out of the uphill tunnel portal and half buried the TBM.

Once the situation was recovered, the TBM was withdrawn and a programme of further exploratory boreholes has been progressing since. The results of the additional investigations will determine the need for civil works to sink a 300m shaft into sound rock and excavate lateral tunnels and chambers to carry out further geological investigations and perhaps advanced ground treatment of the treacherous Piora dolomites.

Drill+blast excavation of the 990m Sedrun access started in April 1996 and at the end of March 1997, excavation had progressed to 860m. Excavation of the inclined ventilation adit is yet to start and although excavation of this and the 800m x 8.75m diameter access shaft are included in the current contract, notice to proceed further than the end of the 990m horizontal adit will have to await the outcome of the current political debate.

Fig 3. Geological section of the planned St Gotthard baseline rail tunnel
Fig 3. Geological section of the planned St Gotthard baseline rail tunnel

On the Lötschberg tunnel, excavation of the 9km Frutigen exploratory tunnel and the 1.5km lateral adit at Mitholz is complete. Excavation of the 5m diameter exploratory/pilot tunnel running parallel with and at the same elevation as the main 42km tunnel was adopted to investigate the quality of Helvetic nappe system at the north end of the tunnel which is dominated by subhorizontal structures which are otherwise difficult to predict by traditional investigation methods. Of greatest concern is the interstratification of hard lime series rocks and thick Flysch deposits together with the potential for natural gas concentrations and high water inflows.

Proposed Faido/Polmengo adit exploratory works
Proposed Faido/Polmengo adit exploratory works
Sedran intermediate access point
Sedran intermediate access point

Excavation of the pilot tunnel using a 5m diameter Atlas Copco Robbins TBM started at the Frutigen portal, about 10km from the north portal of the main line tunnels, in January 1995. In February 1997, the TBM was stopped at the end of its 9.5km dead-end heading and was withdrawn.

In the meantime, drill+blast excavation of a lateral 1.5km long x 60m2 intermediate access adit had started much earlier, in mid-1994. At just 70m into the drive, work was interrupted for four months when the face deteriorated in difficult geological conditions. This adit progressing from the right intersects the pilot at its 7.4km mark. It is needed to provide an intermediate adit of sufficient size for efficient mucking out and servicing of the main railway tunnel and to provide improved ventilation facilities both for the pilot tunnel and the eventual main line tunnel excavation. The 5m diameter pilot is just too small. Excavation of the adit reached the 1.5km length in July 1996. The TBM reached the intersection with the adit in September 1996 and the connection was made. The TBM then carried on for a further 2km before being stopped in February 1997.

Fig 4. Plan of the Lötschberg tunnel with the 9km exploratory tunnel from the Frutigan portal and the side adit at Mitholz
Fig 4. Plan of the Lötschberg tunnel with the 9km exploratory tunnel from the Frutigan portal and the side adit at Mitholz

In efforts to trim costs off the Lötschberg baseline tunnel, promoters have devised a scheme whereby the parallel pilot tunnel has effectively cancelled out the second of the twin tube main line rail tunnels in this section of the project. Two tubes will advance from both main line portals to the pilot tunnel section where crossovers will direct trains into a phased single-tube two-way traffic main line tube for the 9.5km section running parallel with the pilot. Cross passages at regular intervals will provide the obligatory escape route from the single main line tube into and through the pilot tunnel.

This plan has cut the cost of the Lötschberg project from an original estimate of SFr4.25 milliard (million million) (about US$2.89 billion) to about SFr3.1 milliard.

Total cost of the trimmed NEAT project with its combined 60km Liitschberg route and the 131km St Gotthard baseline projects (80% of which will run in the deep baseline tunnels) is estimated at SFr13.5 milliard (US$9.2 billion). Cost benefit studies indicated that this investment will be repaid within 60 years of operation but these projects have come under fierce attack by politicians in Switzerland. In addition, the small, independent and very prosperous Alpine country is coming under renewed pressure by the European Union to lift its ban on heavy road freight traffic. Switzerland's refusal to allow trucks of more than 28tonne to travel its highways, principally for ecological and environmental reasons, initiated the country's alpine baseline railway tunnel projects in the first place.


  • Visit our Archive and use our TunnelTalk library search facility for a wealth of information about other Alpine tunnelling and underground engineering projects.

Add your comment

Thank you for taking the time to share your thoughts and comments. You share in the wider tunnelling community, so please keep your comments smart and civil. Don't attack other readers personally, and keep your language professional.
In case of an error submitting Feedback, copy and send the text to
Name :

Date :

Email :

Phone No :

   Security Image Refresh
Enter the security code :
No spaces, case-sensitive