Shani Wallis, TunnelTalk

Construction of the new highway tunnel in Dublin is a complex project to undertake and manage as TunnelTalk discovered on a visit to the site in September 2002.

It is difficult not to ask the obvious on a visit to the Dublin highway tunnel project in Ireland. Questions like: Why two tunnelling machines for such a relatively short distance? Why two machines of such different design? Why is the shaft here and not over there? Why not keep the machines moving on a 24h basis to avoid the possible detrimental effects of standing time? All becomes clear, however, with reference back to decisions taken during the pre-tender period.

The two tunnelling machines on the project will complete a 2.6km long twin-tube bored tunnel section of the new 5.6km section of four-lane highway designed to take heavy truck off the city streets and channel them directly from the port to the national freeway feeding ring road on the outskirts of the city.

• 
• 

Tunnelled route (left) beneath the suburbs of Dublin (right)

The two machines, supplied by Herrenknecht, are each 11.77m diameter and of completely different design. The first is a full-face hard rock shielded TBM needed to work primarily through medium strength limestone with interbedded layers of softer shales. The other is an open-faced shield fitted with three large backhoes working on three levels - the invert and two intermediate working decks. It is also equipped with comprehensive sets of breasting plates for working through boulder clay described as having a good stand-up time but with predicted lenses of water bearing sands and gravels and a high content of cobbles and boulders up to 750mm in dimension. Both machines are refurbished. The rock machine was first used for the Bözberg project in Switzerland, followed by a second project in Italy, and the backhoe shield was used on the Thalwill highspeed railway tunnel project near Zurich in Switzerland.

During a visit to Dublin in early September 2002, both machines were in the early stages of being launched from the one intermediate 56m diameter x 30m deep access/working shaft. Going south, the full-face shielded rock machine was about 85m into its first 2,250m drive, principally through limestone. On the other side of the shaft and heading in the opposite direction, the backhoe machine was just preparing to start its first of two 350m drives through boulder clay.

Ground water was identified as present in both geological formations. The water table in the boulder clay rises to within 1m of the surface while water in the fractures and fissures in the limestone is predicted to be under some artesian pressure. In neither geological deposit was ground water considered a problem that would require EPB excavation or other pressure controlling techniques. The large access/working shaft was excavated from within the support of full depth diaphragm walls, and a dewatering system is being maintained through the construction period to lower hydrostatic pressure on the concrete base slab and partially dewater the sand and gravel deposits.

The full-face rock TBM

Mucking out behind each of the non-pressurised tunnelling machines is via continuous conveyor with vertical conveyors lifting material to the surface for disposal. At the end of their respective first drives both machines will break through diaphragm walls into the highway cut-and-cover transition zones and will be turned into the alignment of the second parallel drives, breaking through finally into the temporary working shaft from which they started. The continuous conveyors will be rearranged to muck out to alternative sites at the portals of the second drives.

The tunnels are lined with 1.7m wide rings of bolted primary precast concrete segments with six segments and a key in each ring. The 350mm thick segments are fitted with a single hydrophilic gasket placed some 40mm from the outside edge. Vacuum segment erectors lift the 8.1 tonne maximum segments into place. Although each machine is different in design, they are both 11.77m in diameter and erect the same design of segments.

Planning and logistics
With no major surface route to follow, the total 2.6km section of bored tunnel on the new highway to its junction with the M1 lies beneath a long established residential suburb of the city. This restricted the location of access points and imposed strict limits on various working conditions. All of these were investigated and discussed with several local resident groups, and are specified in detail in the EIS (environmental impact statement) completed early in the development of the project by the client's tender stage project engineers.

The open faced backhoe shield

One of the EIS stipulations locates the single working shaft for the bored tunnels in open grounds owned by a convent and adjacent to Whitehall Church. This places the 30m deep shaft almost entirely in boulder clay and about 200m, at tunnel horizon, from the limestone interface. From the fixed position of the shaft, the hard rock TBM has an initial 200m of boulder clay to excavate. In addition, both TBMs start their drives from the shaft on the maximum 4% highway gradient with the south-bound rock TBM starting on a 4% downward slope.

The rock machine was the first assembled and launched from its 11m long starter adit in late June 2002. Initially, all involved were confident that the rock TBM would be able to handle the boulder clay but early problems were encountered. After leaving the support and control of the launch cradle, the shield, on its 4% downhill alignment, began to lose level. When TunnelTalk visited the site in early September, the 11.77m o.d. TBM was about 400mm below level after completing 85m or 50 rings.

The causes were said to be several fold, but primarily, the cutterhead and front end of the 1,100 tonne hard rock TBM is heavy. With the harder limestone up to 3m below the base of the access shaft and running tunnel, it was felt that the clay was unable to support fully the weight or allow the shield to respond to corrective steering.

Of various efforts being tried to correct the situation, one was to use only the invert thrust rams during each stroke to try and increase the look-up orientation of the shield. Steel straps across the circle joints in the crown counteracted the asymmetric loading that was trying to spring the circle joints in the crown apart but the uneven thrust had caused cracking and spalling of the precast concrete segments that will require repair.

Side-by-side TBM launch in the 53m diameter x 30m deep access/reception working shaft

The method achieved only limited success. The corrective margin gained during the stroke was being lost during the ring-build standing time. As a result, an additional technique was being implemented when TunnelTalk was on site.

This entailed a system of restraining cables anchored to the thrust wall in the shaft and attached to the upper part of the TBM's forward shield bulkhead to pull on the top of the shield and allow the thrust rams at the bottom to push the cutterhead forward and upward. These 24 cables, passing along the soffit of the tunnel and through the ring build area of the shield, have a maximum load capacity of 240 tonne and are gradually loaded to achieve the millimetres of correction needed with each shove. To make way for erecting the segments, the tension on the cables is released during ring build.

Reports from Dublin in late September confirmed that the method was having the required effect. The TBM was making upward corrective progress and with enough cable on the payout reels, the system was available to assist steering through the remaining 115m or 68 rings before the TBM meets and passes into the limestone.

Dublin Port twin tube four-lane highway tunnel project Client: Dublin City Council Employer's Representative: Tim Brick, Deputy City Engineer Deputy Project Engineer: John Flanagan Tender Phase Engineers: Ove Arup, UK with GeoConsult, Austria Design-build Consortium: Nishimatsu, Mowlem, Irishenco Project Manager: Chris Sedman Design-build designers: Haswell Consulting Engineers with Carl Bro and GCG On-site Representative: Bill Newns Surface works contractor: Mowlem Irishenco Project Manager: Oliver Miret Bored tunnel contractor: Nishimatsu Construction Company Project Manager: Jiro Goto Construction Manager: John Wallis Herrenknecht on-site project manager: Oliver Ferretti Construction Supervisor: Brown & Root Project Director: David Lawrence Chief Resident Engineer: Geoff Featherstone Senior Resident Engineer (Tunnels): Martin Banham Construction start: June 2001 Construction period: 43 months Highway inauguration: Early 2005 Project cost: €448 million Government financing agency: The National Roads Authority as part of the government's National Development Plan for transportation links.

To remain within highway tolerances and avoid ring-binding the segmental lining in the tail shield, overall corrective measures will include incremental adjustments along the remaining alignment of the 2.2km long tunnel drive as well as further correction if necessary during casting of the tunnel's internal in-situ concrete lining. This nominal 275mm thick lining over the tunnel‘s arch above the road deck is a non-load bearing sacrificial finish of a concrete mix with polypropylene fibres designed to degrade and spall in the event of a fire and protect the load-bearing segmental lining behind. As well as the hydrophilic gasket in the primary lining, the tunnel has a drained waterproofing membrane system between the segmental and in-situ linings to prevent water seepage causing a build up of hydrostatic pressure on the internal lining. Clogging clay Another problem slowing progress for the rock TBM, when TunnelTalk visited, was caused by the damp clay.

 

Ground water seeping into the face was mixing with the excavated clay making it sticky and sloppy. This caused blockages and caking on the cutterhead as well as in the excavation chamber, on the TBM transfer conveyor as it passes centrally into the excavation chamber through the middle of the machine's main bearing, and along the continuous conveyor muck hauling system. The problem was exacerbated during extended periods of downtime. Working hours require that the machines stop overnight to avoid perceptible disturbance to residents from the hours of 11pm to 7am each night and there is no production work on Sundays. Manual clean up after the first turn of the cutterhead following these extended stops was losing precious time in the 16h production cycle in the two 12h shifts/day, 6 days/week schedule.

Cut-and-cover tunnelling

The hard rock TBM must pass through another section of boulder clay at the far end of this first drive and negotiate the same two clay sections on its second parallel drive working uphill into the final working shaft breakthrough. As such methods of mitigating these sticky clay problems were being considered. These might include extending production hours to limit the periods of downtime as well as perhaps introducing conditioning agents to improve the behaviour of the excavated clay.

Conforming bid
When first discussed in the early 1990s, the four-lane highway tunnel was promoted as suited to the application of NATM - or SCL as many in the UK and elsewhere prefer.

At the time NATM had been used successfully in the UK and had proven itself over several decades on many hundreds to thousands of kilometres of large diameter road, rail, and metro tunnels in Europe and around the world. The method had also been adopted as the specified design or as value engineered alternatives on several current and important projects in the UK including the London Bridge and Waterloo station caverns on London's Jubilee Line Extension of the Underground. There were few voices at the time expressing serious concern about using NATM in the boulder clay and limestone strata on the Dublin project.

The tender stage engineers for the project, Ove Arup of the UK with specialist NATM subconsultant GeoConsult of Austria, proposed the technical and cost effective advantages of NATM for the Dublin project and the method was subsequently accepted by the client. Ove Arup and GeoConsult went on then to complete the preliminary engineering design for the project, completed the EIS for the works, and prepare contract documents for a fast-track design-build procurement of the highway.

During the visit to Dublin, TunnelTalk was told that the tender phase project engineers had completed extensive site and soil investigation studies which supported the geological suitability of the project to open face sequential excavation and examined thoroughly methods of mitigating potential risks associated with construction, surface settlement, and third party property damage. The results of these studies comprise a full volume of documents within the design-build contract.

Open cut section

It was collapse of the Express Railway station tunnels in London Clay at Heathrow Airport in October 1994 that changed everything. Collapse of those NATM excavations received wide media attention and seriously undermined the confidence clients and many professional engineers had developed in the technique. In addition to the substantial time and cost increases imposed on the Heathrow project, the event had similar time and cost increase impacts on the concurrent Jubilee Line project when the client accepted advice to suspend NATM excavation of the Waterloo and London Bridge station caverns and have them redesigned and completed to more conventional engineering practice.

In Dublin, the Port Tunnel contract was progressing toward the tender process. With an urgent need to sort out chronic heavy freight traffic congestion on the city streets of Dublin, and taking into account the time and money already spent on planning and engineering, the client wished to retain the project's timetable. Objections from the local community and concerns for their property however were intense. To address these concerns, the client and the tender phase design engineers maintained NATM as the conforming bid and included an option for the contractor to used shielded tunnelling machines as an alternative if preferred.

Of the five groups that bid the design-build contract, one tendered on a NATM design. All others proposed a shield alternative and in June 2001 the 43-month design-build contract for the 5.6km of cut-and-cover, surface works and bored tunnel excavations plus design and installation of all permanent M&E services and systems, was awarded for a tender price of €448 million to a consortium of John Mowlem and Company plc of the UK, its Irish subsidiary Irishenco of Dublin, and the European division of Japanese contractor Nishimatsu. Within the consortium Nishimatsu is responsible for excavating the bored tunnels and completing the tunnelled underpass of the main Dublin to Belfast railway tracks, while Mowlem and Irishenco are completing the open-cut, cut-and-cover, and surface works. Mowlem-Irishenco also constructed the 30m deep working shaft for the bored tunnel operation and will install all road decks and M&E services including those in the bored tunnel.

Lead designer to the design-build construction team for the civil and tunnelling works as well as the tunnel ventilation and services systems is Haswell Consulting Engineers of the UK, with UK specialist geotechnical engineers GCG is engaged as subconsultant for the extensive open-cut works, and Carl Bro undertaking the highway design. Mott MacDonald of the UK is the 'Category 3' checker of the design documents presented by the design-build contractor.

NATM cross passage excavation

Construction supervision is by Brown & Root. With a site team of 15 engineers complimented by various support staff, in addition to head office design compliance staff, Brown & Root is the client's representative charged with ensuring that the works are carried out to both design and construction compliance. As a specialist NATM designer, and a subconsultant in the Brown & Root team, the UK office of the Dr Sauer Company (DSC) would have had a more significant supervision role had NATM been the successful tender for the main tunnel bores. As it is DSC is providing design and construction compliance services for proposed NATM excavation of the cross passages that will link the two main highway tubes and of a larger vehicular crossover chamber that will allow traffic to divert from one highway tube to the other should the need arise. There are four such vehicular crossover facility on the highway tunnel. A second will be built into the base of the temporary working/access shaft before it is back filled and another two crossovers are incorporated in the cut-and-cover sections either side of the bored tunnel section.

Preferred engineering
Given the history of the project and its change from the initial NATM design, it is difficult to avoid indulging in a process of preferred engineering, suggesting changes that could have suited better the shielded machine approach. The most obvious would have been a move of the access shaft closer to the limestone interface. This may have been possible within the open space on the surface but would have incurred the risk of delay should there have been a need to revisit the EIS process.

The tender philosophy introduced for the machine alternative was to use only one machine driving south from the shaft with the section to the north constructed as open cut between diaphragm walls. Based on this approach, the shaft was located in the most appropriate position. It was the successful contractor who introduced the idea of a second machine principally to limit traffic disruption at a very congested traffic intersection. Under that approach, the shaft location became less appropriate but by then, the location was fixed. Proposed use of use two machines from the fixed shaft location produced the two short 350m drives for the backhoe shield and a start in clay or the hard rock TBM.

Waterproofing membrane installation

Under the original proposal NATM was to be used in both the clay and the rock strata. NATM would, or could, have progressed on all four faces out of the working shaft; could have advanced from another four faces from the portal zones if necessary; and would have optimised the two-lane highway tunnel cross-section. The 11.77m o.d. TBMs excavate a circular section with a face area of 108m². Part of this is then infilled to create the road deck. During the visit few would be drawn into a hypothetical discussion about whether NATM or SCL would be as suitable as originally suggested in the ground conditions as now exposed.

Perhaps the most eliminating factor for the NATM option however was the need to use drill+blast techniques to excavate limestone in the strength range of 100-120MPa. As stipulated in the EIS, the last blast allowed during each working day was 8pm which limited progress to one, or at the very best, two blasts/working day. This presented significant risk to meeting the necessary excavation schedule in the overall project construction period and avoid liquidated damages for any delays. For the TBM option, the production cycle was allowed to progress a further three hours per working day, to 11pm at night, adding to the machine option's advantage. Nishimatsu's excavation programme is based on an average of 300m/month for the hard rock TBM and 100m/month for the open-faced shield.

The EIS also required strict control of dust. To comply, the contractor chose to provide for a large enclosure of the on-site muck pile to protect the adjacent homes from dust and noise. From the vertical shaft conveyors, muck travels via over-head conveyors for final discharge into one of several bins in the shed. The enclosure also shields the excavated muck from Dublin's frequent wet weather.

TBM progress
Having struggled through the clay zone, the hard rock TBM passed into a full face of limestone in early October and is said to be responding very well. Reports from Dublin at the end of September also indicated that the shielded backhoe machine had started and was making good progress through the boulder clay.

Micro pre-support beneath the railway

As the hard rock TBM advances the tunnel invert is being filled with the permanent base of the subsequent road deck. Trucks running on this infill from shaft to TBM carry segments and other supplies to the trailing backup where they are lifted and the segments carried forward over the advancing infill working zone to the ring build area. Segments for the project are produced under subcontract by Banagher Concrete at a casting yard equipped with CBE moulds from France.

As the rock TBM progresses into its 2.2km long drive, the internal in-situ secondary lining with the waterproofing/drainage membrane behind will progress concurrently and at a distance behind the TBM backup. Lining of the two short clay tunnel drives will be as a second pass.

Within the 43-month design-build contract, excavation of the 2.6km long twin-tube bored tunnels by Nishimatsu is programmed to be complete by November 2003 with the completed tunnels handed over to Mowlem-Irishenco by February 2004 for installation of the finishing M&E systems. The new highway link is scheduled to be in operation by 2005.

Railway underpass
In addition to the TBM tunnels and closer to the port end of the highway, Nishimatsu is also completing the 24m wide x 15m high x 60m long tunnelled underpass of the main Dublin to Belfast railway lines. With a cover of only 3-4m beneath the tracks, excavation will be undertaken from within a pre-support arch of 39, 1.2m diameter interlocked pipe micro bores.

Finished twin tube tunnel motorway

Subcontractor Portun of the UK is using two of its auger boring units to install the pre-support pipes through the graded fill of the railway embankment and the estuarine sediments and underlying boulder clay. When TunnelTalk visited the site, it was taking about one shift to set, weld, test and install each 12m length of steel pipe and about five days to complete one 60m long drive. The first pipe at the centre of the arch was installed to accurate line and level using hand-mining methods. The clutches of the interconnecting system on the steel pipes then guided each subsequent drive. A dewatering system lowered the local ground water table to reduces hydrostatic pressure on the excavation of the pre-support pipes and on the large launch and reception pits, which are excavated and braced with heavy cross supports as the pre-support pipes in the side walls progress downwards.

To ensure uninterrupted operation of the railway line, Mowlem Rail has a 24h x 7day/week gang on call to re-tamp and re-ballast the tracks and compensate for any surface settlement.

Installation of the pre-support pipes is expected to be completed during October and core excavation by Nishimatsu crews expected to start by November.

Project data
The Dublin Port highway tunnel link is the largest civil engineering transportation project ever undertaken by the Republic of Ireland to date. At an overall cost to the nation of €448 million, the project is being paid for entirely by the Irish government, without European Union contributions, and via the National Road Authority as part of the Government's National Development Plan for improved transportation links. In addition to greatly improving the environment of the city, diverting heavy freight traffic into the new highway and its tunnels will promote development of derelict dockland areas to high standard offices and residential properties and further Ireland’s significant economic advance over recent years.

Search

TunnelTalk Articles TunnelCast Videos Archive Feedback Current Diary Dates World News Briefs Career Moves Accolades & Tributes Stay Connected Procurement Job Opportunities Equipment Sales DIRECTORY Marketplace About Us Advertising Contact Us Legal Notices Home Office     Latest Alert Alert Signup Alert