Sydney alliance holding up under pressure 01 May 2000

By early March 2000, some 70% of the 20km of 6m- and 3.8m-diameter tunnelling associated with the Northside storm water retention project in Sydney was complete, but this is well behind the original project programme, which is based on 760m of tunnel/week. The objective of the four partners in the project's procurement alliance - project owner Sydney Water, project consultants Connell Wagner and Montgomery Watson, and contractor Transfield - was to have tunnel excavation finished by early 2000 and the facility in operation before the Sydney Olympic Games start in September.

The 1.6m-diameter ventilation ducts that carry fresh air down to the advancing TBM tunnels

"We set ourselves a very tight project programme," said Allan Henderson, Deputy Project Director for Sydney Water and member of the Northside Storage Tunnel project alliance management committee. "Tunnelling is currently (early March 2000) running about four months behind the original combined four-TBM tunnelling schedule, but we have experienced some serious delaying factors.

"These centre mainly on a longer than expected schedule to complete the main Tunks Park working adit and underground construction site and difficult geological conditions at launch of the two 6m-diameter TBMs at Tunks Park. However, through continued party cooperation, made possible by the project's contractual alliance, we believe we can recoup this lost time and still bring the project in close to schedule and budget."

The TBMs were then averaging 200m to 300m/week for a combined total advance of more than 3,600m/month, in advance of the original 760m/month/machine programme average. If this advance rate were adhered to, the goal was still in sight.

The four TBMs, one Robbins and three Wirth machines, are working through the Hawkesbury Sandstone that underlies the Sydney area. The sedimentary formation is a competent, relatively soft rock of 25MPa to 70MPa in unconfined compressive strength, with few major faults or discontinuities. A distinctive quality of the rock is that, after millennia of erosion, the formation has a higher in-situ ratio of horizontal to vertical stress. In tunnelling, this can cause either 'dog-ear' spalling at tunnel spring-line or instability and spalling in the crown and invert. The sandstone, however, is well known by engineers who have worked in the Sydney area and is considered to be a favourable tunnelling medium on the whole.

Tunnel access set-ups

The working chambers and access adits at Tunks Park and the North Head working sites are extensive and complex. These underground chambers and adits were excavated through the sandstone using roadheaders - two at Tunks Park and three at the deeper, more extensive North Head site, where chambers for the pumping station to lift storm water into the treatment plant were also excavated by roadheader. These works were major projects in themselves. Work at North Head and Tunks Park started in May and August 1998, respectively, and the roadheaders - two Mitsui 300s, two Voest Alpine 105s, and a Mitsui 200 machine - averaged about 20m³/h.

"We are very pleased with the final profiles of the large adits and huge cavern cross sections," said Henderson, "but excavation did take longer than anticipated, which seriously delayed start of the TBMs. Support in the adits and chambers was much as expected, but work on the 14% declines was particularly tough. In addition, the sandstone was very hard and abrasive in places and was tough to excavate."

At Tunks Park, the underground working areas include two 8,000m³ muck storage caverns to hold muck during times when the barges are not permitted to operate (i.e., during the night or in bad weather). They also accommodate the extension and tension loop section of the horizontal conveyor systems, the construction water treatment facilities, and the Korfmann ventilation fans.

At North Head, muck is transferred from the horizontal conveyor behind the TBM to the base of the project's 170m-deep lift shaft, where a vertical conveyor lifts muck to the surface. From here, it is transferred to the barges via the 1.4km-long conveyor tunnel. The project's conveyor systems have a capacity of 675 tonne/hr at North Head and 1,300 tonne/hr at Tunks Park, one of the largest capacity conveyors in the world. As the TBMs near their respective destinations, the conveyor system will total more than 25km of belt in operation. Once the project is complete, more than 800,000m³ of material will have been hauled. The 16 barges on the project are 55m to 60m long x 17m wide, with a cargo capacity of 2,000 tonne.

Ventilation on the project is supplied by nine sets of 100kW rated Korfmann ventilation fans capable of delivering 16m³/sec of fresh air to each tunnel heading via 1.2- to 1.6m-diameter Protan Ventiflex ducting from Norway. More than 23km of ducting will have been installed by the end of the project.

Tunnel statistics

The 20km-long Northside Storage Tunnel is designed to be largely unlined. As an overflow retention facility, the tunnel is expected to operate some 20 times/year during heavy rain storms to store diluted sewage that would otherwise flow into Sydney Harbour. In smaller storm events, flow remains in the mains of the existing sewerage network. This allows entry into the storage tunnel for inspection and maintenance if necessary.

As a result, support in the headings is required principally for immediate stability and safety during construction. Fibreglass and galvanised steel rockbolts, wire mesh, and wet mix fibre reinforced shotcrete provide immediate support where necessary and an in-situ concrete lining will be cast into the invert as required.

Of the four machines working on the project, all are refurbished. The three Wirth machines are open gripper machines and the Robbins is a double­-shielded machine. Three machines are working from the main Tunks Park construction site and access adit, while the fourth is advancing towards Tunks Park from the North Head treatment plant access adit for a mid-tunnel junction with its partner from Tunks Park. Mucking out behind all four TBMs is achieved via continuous conveyor with a sequence of tug boat powered barges carrying muck via the harbour waterways to a central off-loading point. From here, muck continues on a 50km-long haulage route by train to its final licensed dumping site. The spoil is being used for a state rail project.

Given the project's urban and residential location, strict noise, dust, and construction activity restrictions were imposed at all sites. At the main Tunks Park site, most of the construction activity and support equipment is located underground in large working chambers off the main access decline to the deep tunnel level. Surface operations at the site, including the barge loading facility, are housed in large acoustic sheds to limit noise and painted green to blend in with the park site surroundings.

TBM tunnelling starts

The first machine on the project was the 3.8m Wirth TBM that excavated the 1.4km-long Little Manly Point conveyor tunnel to transfer muck from the North Head works to the barge loading jetty at the waterfront. In the competent sandstone, the TBM recorded an average advance of 150m/120hr production week and completed the 1.4km-long drive in 12 working weeks, finishing in December 1998. The same machine was then the first away of the three Tunks Park TBMs. After its launch in August 1999, it has again made good progress of 230m average/140hr production week and is expected to break through at the Lane Cove River reception shaft in April 2000.

"Our most disappointing advance is on the two larger diameter TBMs from Tunks Park," explained Henderson when TunnelTalk visited the site in December 1999. "The higher horizontal to vertical stresses in the sandstone are causing severe spalling and instability in the crown of the 6m-diameter tunnels, and this is increasing substantially the amount of support required and the time taken to install these elements. There are bolting rigs on the TBM, but in these conditions it is imperative to install the support as soon as possible and as close as possible to the face, and this is causing significant TBM downtime and delay."

The first machine to encounter the spalling problem was the 6.3m-diameter Wirth machine launched in September 1999 on the Middle Harbour heading from Tunks Park. Spalling occurring within 60 minutes of excavation. In December 1999, TunnelTalkwatched crews on the open gripper Wirth TBM use Tamrock drill rigs mounted on the TBM outer kelly forward of the grippers to install dense groups of rockbolts directly behind the cutterhead about 5m from the face. The secondary drilling and shotcreting station on the TBM is about 50m from the cutterhead on the backup trailers.

In early November 1999, on the 6m-diameter Robbins machine in the Scotts Creek heading, the situation was more serious. Because it was a shielded machine, the first opportunity to install rockbolts was at the rear of the gripper shield about 15m from the face. By that time, the stress relieving forces had initiated severe spalling, with accumulated debris loading the outside of the shield and fingers and waiting to fall out as the machine took its next stroke.

"We were aware of the spalling phenomenon before we launched the Robbins machine," said Henderson, "but the TBM was already in assembly at the base of the Tunks Park adit - quite an operation in itself to get the large-sized components into the assembly chamber. In addition, launch of the TBMs was already running behind schedule. As a result, it was agreed within the alliance that the machine would be launched in its shielded configuration without removal or modification of the gripper shield."

During the visit in December 1999, the situation was impacting severely on the project schedule and cost, and was the central topic of discussion at meetings at all levels of the project alliance. "My feeling, and that of engineers who also have experience of the Hawkesbury Sandstone, is that the stress relief at the Tunks Park adit is more severe because the area is beneath a basin of lower overburden," explained Henderson. "We also had spalling problems at the start of the 3.8m-diameter Wirth machine to Lane Cove, albeit less severe with the smaller diameter, and conditions improved as the heading advanced under increasing overburden. We believe the same will happen on the two larger diameter headings to Scotts Creek and toward Middle Harbour." By early December, the Wirth machine was 1,000m into the Middle Harbour drive and, although conditions were improving, spalling was still an issue requiring more support than expected and impacting adversely on TBM utilisation.

The Robbins machine, meanwhile, was 500m into the drive, having started in early November, and was running seriously behind schedule. "We can expect conditions to improve as the overburden increases," said Russell Cuttler, Project Manager for the alliance, "but, in the meantime, we have decided that modification of the shield will be necessary over the end-of-year holiday break to allow installation of rockbolts closer to the face."

One of three open gripper Wirth machines working on the project

"What we will do," explained Cliff Wallman, site manager at Tunks Park, "is cut 'windows' into the telescopic hood between the shields through which we can install an initial array of up to four rockbolts. Bolt holes will have to be drilled using stopers or wombat drills in the limited space in the forward section of the shield, but this will bring the first opportunity to install rockbolts some 6m to 7m closer to the face."

Depending on the wear of the gauge cutters, the annulus between the gripper shield and the tunnel crown is 80mm to 125mm. In these windows, split sets can be installed without being damaged by the gripper shield as it passes. The main support is then installed at the tail end of the shield using twin Atlas Copco bolters. Additional secondary galvanised steel and resin anchored rockbolts can be installed from the backup support station about 40m from the face.

Latest reports from the site in early March 2000 confirm that the modification to the Robbins shield had been undertaken and that, while the spalling situation still persisted, progress had improved significantly. "Crews have been installing four initial split sets through the hood windows with every stroke and still applying much more support than anticipated from the backup support station," said Cuttler. "Downtime for installing support on the Scotts Creek heading is still too high, at about 20% at present (early March). But TBM penetration during a 1m to 2m stroke is as anticipated for all the TBMs, at about 75mm/min, and rates of advance for the Robbins machine have improved to 265m, excavated in one 140hr production week."

Contractual conditions

Given the various delays encountered to date, a contract under traditional terms and conditions would have been under considerable pressure by now. Contractual letters would be flying back and forth and dossiers for substantial claims would be growing thicker by the day. As it is, the alliance arrangement entered into by all parties to the Northside Storage Tunnel project is said to be standing up very well to the test.

"Because all parties are aiming for the same goal, we can pool resources and act unanimously on 'best for project' suggestions rather than allowing the situation to fuel individual grievances," said Henderson. "There have been some tense and frank meetings but, because each alliance partner is sharing in any project profit or loss, there is no room for individual vested interest."

In December 1999, TunnelTalk was told that the current forecast was projecting a time and cost overrun but that, as a fast track project, there was a lot of flexibility built into the design and the construction programme. As such, there was scope for recouping time and cost in the tunnelling schedule and in the post excavation activities.

Groups of fibreglass dowels and wire mesh installed immediately behind the trailing fingers of the 6.3m-diameter Wirth TBM

"Increasing the tunnel production schedule is one opportunity already implemented for recouping lost time," said Cuttler. "The alliance's target excavation programme, as verified by independent assessors, was to excavate 20km of tunnel using four TBMs and averaging 190m/machine/week, working 24hr/day (two 10hr production shifts and a 4hr maintenance shift), 6 days/week. By increasing this to a 7 day/week schedule, the average advance estimate increases to 250m/week/machine, which improves overall project advance considerably."

In October 1999, the 3.8m-diameter Wirth TBM working on the Lane Cove drive from Tunks Park increased to a 7 day/week production schedule to take full advantage of the mucking capacity of the working site, while the two 6m machines struggled to gather momentum and reach optimum progress. Since then, all machines have changed to a continuous 24hr/7 day/week production schedule.

Meanwhile, to compensate for the delayed TBM working from Tunks Park, the Wirth machine, which has been progressing well since its launch in early September 1999 from North Head towards Middle Harbour, will advance well beyond the anticipated 4.7km junction point to complete the Middle Harbour undersea crossing.

At Middle Harbour, extensive geotechnical investigations have indicated a Paleochannel that has experienced severe 'valley bulging.' As a result, an extremely high permeability zone is anticipated (greater than 100 lugeon) over a 100m to 250m length of tunnel. Therefore, the 6.6m Wirth TBM has been fitted with twin Tamrock probe drills, and pre-grouting in advance of the face will be undertaken.

Eighteen 38m holes will be drilled 7° to give full 360° coverage. Grouting will be undertaken until acceptable water inflows are achieved, leaving an 8m bulkhead in front of the face. The TBM will advance in 15m to 20m cycles during the crossing, which was anticipated to take place in April 2000. The junction is now set for 5.9km from North Head and 3.7km from Tunks Park. Breakthrough is anticipated in May 2000. To achieve this, the North Head machine has a further 1,426m to excavate from early March 2000, and the Tunks Park machine is required to complete another 853m.

The Robbins machine drive at Scotts Creek was still on the critical path, with some 2.5km remaining to excavate from early March 2000. Although progress for this machine has improved dramatically in more favourable rock conditions, the most optimistic completion date for the 6m-diameter x 3.8km drive is late May 2000. Plans to bring forward finishing works are in place and the target commissioning date of the project has been revised to August 2000.