On 22 July 2005, the second 7.2m diameter TBM broke through at Chatswood, virtually completing the excavation for the 12.5km long Epping to Chatswood Rail Line. It followed the first TBM which broke through on 8 June. The TBM’s breakthrough represents the near completion of excavation of some 2.53M tonnes of spoil from the four underground stations and twin running tunnels. The only excavation remaining consists of a crossover, fan niches and several cross-passages.

The breakthrough marked a significant milestone for the Thiess Hochtief JV, which was awarded the US$676M contract for the civil and systems works, including tunnelling and station excavation, and its lead design firms, GHD and Parsons Brinckerhoff. During the tunnelling, TBM No. 2 set a world record in April 2005 by excavating 92m in one day. In early 2005, TBM No.1 narrowly missed the world record for weekly excavation when it drove 368m, 4m short of the current record.

Project scope

The US$679M Epping to Chatswood Rail Line (ECRL), the largest current publicly funded infrastructure project in New South Wales, is on schedule for 2008 rail operations. Connecting Sydney’s second business centre at Chatswood with suburban Epping, the new rail line is the first addition to Sydney’s Cityrail network in 80 years and forms an integral part of the government’s US$26.9bn dollar infrastructure plan and investment in public transport. The line includes four new underground rail stations: Epping, Macquarie University, Macquarie Park and Delhi Road. It will deliver faster trips to the city from Epping, and help relieve congestion between the main rail line and the CBD (figure 1).

The main tunnel drives were excavated by two refurbished Robbins TBMs, each 210m long, worth US$38.4M, and weighing 1060 tonnes. Four roadheaders, rock hammers and rock saws were used to excavate the four underground stations, service building shaft excavations, crossovers, cross-passages and associated ancillary excavations.

The two TBM drives, including stations, were 12.5km long separated by a centre pillar, typically 5.6m wide but as narrow as 3m at Chatswood and as wide as 15m at Epping. The machines began their drive westward from the project’s mid-point at the M2 service shaft on 21 September, and 4 November 2003. They typically excavated 120m-180m per week, with rates on occasion exceeding 8m per hour.

Over half of the 6.7km long westerly drives were in curves with radii ranging from 500m-870m. The large volumes of muck and the non-linear nature of the alignment made muck handling and removal an important part of the project (T&TI, February 2004).

The TBMs broke through at Epping on 27 and 29 July 2004. The cutterheads were removed and lifted to the surface, then refurbished and refitted. The trailing gear was backed down the tunnel to the M2 site, lifted to surface by crane via the M2 shaft and turned around after maintenance. Both bores began again in November 2004, this time in an eastward direction, on a 5.15km alignment with almost 90% in curve.

Running tunnels

The running tunnels intersected a sequence of Ashfield Shale, Mittagong Formation and upper Hawkesbury Sandstone, typical to the Sydney Basin. Design of the tunnel support system considered the risks of potential variations in geology in the development of the range of tunnel support systems. One of the most significant parameters providing an impact on design was the pre-existing in situ stress fields that influence tunnelling conditions, and post excavation induced ground movements. Rock mass and rock defect properties, including shear strength, cohesion and friction values, were developed on the basis of geotechnical information, data and testing. Published information and geological experience in the Sydney Basin environment were also used to develop rational design parameters.

The running tunnel support consisted of initial or temporary primary support, and final support. An array of six 2.7m long, 310kN capacity, temporary full-resin-encapsulated rock bolts generally comprised the initial, or temporary support. They were installed approximately 6m back from the face directly behind the TBM cutterhead. About 55,000 bolts were used in the TBM No. 1 excavation alone, and a similar number in the TBM No. 2 excavation. Mesh was also installed where required.

When very poor ground was encountered, steel sets were an option, but they were only used where the excavation crossed the North Ryde fault zone. A rock bolt/anchor and shotcrete support design was developed as an alternative to steel sets in all the other areas where the original design envisioned steel sets. This significantly speeded construction in areas of poor geological conditions.

Distinct cross-beds that daylighted in the crown frequently caused instantaneous break-back up to 20m. Altering the design by installing a staggered rock bolt pattern secured the tunnel obvert when these cross-beds were present, arresting and generally preventing the break-back.

Ground support did not change from TBM No. 1 to TBM No. 2. However, from the detailed projection of the geological structures encountered in TBM No. 1, the geotechnical team was able to provide TBM No. 2 crew with accurate forecasts of impending ground conditions. This provided the TBM No. 2 management the opportunity for safe support with no surprises, and with associated efficiency.

Permanent lining and water control

For the TBM running tunnels, the permanent lining is circular-formed un-reinforced concrete, nominally 200mm thick, with a design compressive strength of 32MPa at 28 days. The concrete lining is the same for the drained and the undrained tunnels. The formwork consisted of six shutters, 15m long. Part of the running tunnels is lined with shotcrete for construction reasons.

The tunnel invert consists of a level horizontal slab constructed 3m below the tunnel’s centre. Precast reinforced invert segments, 3200mm long by 450mm wide and shaped to the excavated radius of the tunnel, vary from 215mm at the edge to 650mm at the centre. The segments were spaced 6m apart on a prepared mortar bed. Cast in place, infill slabs of reinforced or un-reinforced concrete of the same depth and profile as the precast segments completed the invert (figure 2). Precast segments were placed behind the bolting deck of the TBM.

Water control for the undrained tunnel was achieved by attaching a geotextile and 2mm HDPE membrane to the rock around the circumference of the tunnel to the invert slab, and a hydrophilic waterstop placed longitudinally along the joints between the invert slab and the final lining. The drained tunnel has a HDPE dimpled sheet membrane and weep holes every 15m.

Groundwater investigation and analysis estimated the long-term steady-state groundwater inflows and inflow control measures to meet the contractual requirements; these requirements included a maximum inflow criterion of 0.1 l/sec/100m of running tunnel, as well as no visible water spotting on the interior of the lining.

Forward probing was required in a number of locations of potentially higher groundwater inflows. The running tunnels are designed and constructed to be either drained or undrained. About 75% of the length of the running tunnels is drained. As the three creeks and Lane Cove River were approached along the tunnel alignment, forward probing ahead of the TBM determined that no grouting from the machine was necessary, so its advance was not significantly impeded. The effect on groundwater and settlement were monitored.


The innovative configuration of the station caverns was defined by the project architects, as the project tender required an asymmetrical arched design, one side of the arch having a flatter arc than the other. Station design drivers included transit functionality, surface constraints, mechanical/electrical plant requirements, fire life safety considerations, and underground construction constraints. Since no Sydney precedents dealt with large, near surface wide-span arched cavern design, the arched design presented unique challenges in relation to shear movements and rock support.

The cavern design was based on the anticipated loads and potential movements the rock bolts and cables would be subjected to during construction and over their 100-year service life, as well as on requirements for watertightness and durability. Surface settlement over the caverns was limited to 40mm, and surface structures were not to be subject to movements likely to cause damage. Maximum permissible anchor extension was governed by Australian Standards, and a maximum shear movement of 15mm was established for the project; after this, the anchor would be deemed to be compromised and would be replaced.

Sequencing of the work

Excavation order for the three station caverns was Macquarie Park Station (Mac Park), Macquarie University Station (Mac Uni), and Delhi Road Station (Delhi Road). At Mac Park, the first station to be designed, excavation began in September 2002.

The 190m long platform cavern’s excavation was 14m high and 20m wide. It was excavated in a series of headings and benches, beginning from the top of the cavern excavation sequentially in two or three headings. Mac Park’s design required installation of permanent support at the earliest opportunity. The bolts were either temporary or permanent bolts/anchors. Temporary rock-bolts are relatively inexpensive and relatively simple to install, but had to be replaced later with corrosion-protected permanent bolts or anchors. Installing permanent anchors near the roof of the cavern was preferable, but risked subsequent excessive movement as the excavation proceeded lower.

The Mac Park two side headings of the top bench were supported by 310kN capacity, 2.7m long, fully resin encapsulated temporary bolts. Next, 6m long Flexibolts (capacity 550kN) were installed before the central pillar was excavated, to support the ground during the pillar’s excavation. Finally, 7m and 7.5m long single and two-stage grouted permanent rock anchors were installed, with a capacity of 580kN and 800kN, respectively.

As reported in February 2004, the ground behaviour was more favourable than early modelling had indicated. Maximum ground settlement above the platform and concourse was 15mm, the crown sag was 12mm, and the maximum shear displacement was 14mm. About 100 temporary bolts were replaced during top-heading excavation, as they experienced more than 12mm of shear. The geotechnical design for the subsequent stations utilised this information to develop a design, in conjunction with the construction sequence, to minimise shear displacement across bolts and anchorages.

Although the stations are similar, they have significant differences. Importantly, geological conditions at each of the three stations were different, affecting the required rock support for each station. Delhi Road had the highest measured in-situ stresses of the three stations. However, the geology of Delhi Road was more massive. It did not have the pervasive clay beds that were discovered at Mac Uni, nor the joints and faults encountered at Mac Park and as such, the high stresses were effectively locked in, as they were not as redistributed and subsequently dissipated by rock mass deformation over geological time. The stations also differ in their layouts.

Mac Uni and Delhi Road caverns were designed for an anticipated condition instead of a worst-case scenario. An observational method was used to carefully monitor the behaviour of the ground and identify the amount of movement the rock support encountered. Where movement exceeded the 15mm trigger level, supports were rebolted.

By February 2004, both the Macquarie Park and Macquarie University stations had been excavated. Also excavated in 2004 were Delhi Road, and Epping stations, the latter carrying on into 2005. About 850,000 tonnes of rock were taken from the four stations, including Epping.

As at Mac Park, the excavation of Delhi Road Station was driven in three top headings, two side headings and then the central pillar. The excavation of the top heading of Mac Uni Station was driven in two headings. The initial support comprised of 2.7m long 310kN rock bolts, and 6m long 550kN capacity cable anchors, with 2.7m long 310kN rockbolts used as spotbolts between the cable anchor pattern.

Delhi Road’s two side headings were supported by 6m long Flexibolts while the central pillar was excavated. Permanent rock bolts were then installed comprising of 6m long DCP full column grouted 310kN capacity. Mac Uni’s two top headings were supported by 6m long Flexibolts upon excavation, and the permanent anchors were installed later: (7m and 7.5m long single-stage 580kN and 730kN capacity).

Excavation of the stations was principally performed using four Voest Alpine 300kW roadheaders, which achieved high rates of productivity. The roadheaders excavated the 14m high x 20m wide station platform caverns leaving a 5.6m high bench at its base in Mac Park and Mac Uni. The TBMs then excavated through this lower bench, with 1.7m of the 7.2m cutterhead exposed (T&TI, February 2004). The torque and thrust were reduced by about half, and as the second TBM passed through each station, a special brace was used to protect the 5.6m wide pillar between the two TBM excavations at the base of the stations. Roadheaders excavated Delhi Road station platform cavern down to invert level.

Macquarie University Station had a maximum measured shear of 29.5mm; the largest bolt/anchor shear was 16mm, resulting in rebolting of some 65 permanent anchors. Delhi Road station bolts experienced a maximum shear of 8mm, and 15mm of surface settlement was measured. The total number of rockbolts and rock anchors used in the three asymmetrical arch shaped station platform caverns, concourses and passageways is shown in Table 1.

Waterproofing and final lining

All station excavations were designed as ‘drained’ to avoid the need to resist hydrostatic water pressure. Expected water inflows were calculated as part of the project hydrogeologic assessment. Specific waterproofing and drainage were designed for the cavern roofs, and included a drain strip longitudinally spaced at 3.5m centres (with sheet drains at wet spots) and a 2mm-3mm sprayed membrane applied over a smoothing shotcrete layer. The requirement for anchor penetrations meant a sprayed membrane was selected rather than a sheet membrane. Various membranes were considered. An acrylic membrane was chosen which created high-quality adhesion between the shotcrete layers

The design determined that the permanent ground support could not be accommodated by a passive final concrete layer because of the size, shape and geometry of the cavern as well as the sandwiching of a waterproof membrane. A 120mm thick permanent shotcrete lining (reinforced steel fibre) was installed, spanning between the rock anchors at 1.75m centres. Spider plates screwed onto the rock anchors suspend the shotcrete in 1.75m by 1.75m panels below the sprayed membrane.

The lining has been designed to provide fire protection for the anchors. Modelling and testing demonstrated that the total 180mm of shotcrete (which included the smoothing shotcrete and the final lining) would provide sufficient protection. Polypropylene fibres were included in the shotcrete mix at 0.9kg/m3 to resist spalling in the event of a fire.

Remaining work

Many challenges have been addressed in the design and construction of the Epping to Chatswood Rail Line in excavation, support, waterproofing and lining and many innovative solutions have been devised. The project is on track for opening in 2008.

Related Files
Fig 3 – typical section through the station lining showing the waterproofing arrangement
Fig 2 – rock support regime shown in a cross section of a running tunnel
Fig 1 – Plan map of the project with stations and the central M2 worksite shown