Blue Mountains record breaker Mar 1995

Shani Wallis, Editor

In the mid-1990s, McConnell Dowell worked on a long sewer tunnel project in the Blue Mountains north-east of Sydney, Australia. The 13.4km long tunnel was excavated using a 3.4m diameter Robbins TBM through sandstone. Some 17 years later, progress rates achieved by the machine, its support systems and its working crews, retain the industry’s world records for the best day, best week and best average monthly advance rates. Excavation of the tunnel was finished months earlier than programme and set a solid foundation for the company's subsequent tunnelling successes. Reprint of a 1995 site visit article about the Blue Mountains project is published in conjunction with the last in our series of articles about the company's four current projects - the new Pike River coal mine development on the west coast of New Zealand’s South Island; the Rosedale outfall and Hobson Bay sewer tunnel contracts in Auckland, New Zealand; and the Bogong hydropower headrace tunnel in the Snowy Mountains of northern Victoria, Australia.

Reports of high tunnelling advance rates can engender skepticism but claims of almost 1,200m bored in a single production month on the Blue Mountains sewerage transfer scheme near Sydney in Australia are real. In combination, a 3.4m diameter Robbins Mk12 TBM and a continuous conveyor muck haulage system, has recorded a best 8h shift of 72.5m, a best day of 172.4m and a best week of 703.3m through the soft to medium strength sandstone of the area.

During a visit to the project in November 1994, the TBM and its continuous conveyor system was 10.8km into a 13.4km long drive. A trip to the face with the 3pm shift change took 1h in a man rider with the return trip taking another hour on a supply train. At 13.4km long and 3.4m in diameter, this tunnel is believed to be one of the longest of its size constructed from a single access and one of the smallest diameter tunnels to use a continuous conveyor mucking system. The best advance rates achieved on the drive to date set various benchmarks against which similar projects in the future will gauge their performance.

Sharp end of the Robbins TBM

Private enterprise
The tunnel is being built as one of five privately financed build-own-operate-and-transfer (BOOT) public infrastructure schemes promoted by the New South Wales state government along with two water treatment plants and two toll roads. In July 1990, three groups were invited to bid for the 35-year concession with its extra 15-year option. Eventually, in June 1993, after an evaluation that focused as much on the financial structure as the design concept, the scheme was let to a joint venture between McConnell Dowell Constructors of Australia and the Obayashi Corporation of Japan.

On completion, the civil works will be sold to a pension fund investment trust that will operate the facility and charge the Sydney Water Board a fee for its use. In today's economic climate, it was said that infrastructure projects provide a more attractive private investment return than commercial buildngs or other real estate options.

Close up of the 17in cutters

"As civil contractors, the joint venture was interested in the civil contract rather than in owning or operating the investment project", said David Logan, Project Manager for the McConnell Dowell Obayashi JV. "Sale of the facility to the superannuation fund was part of the original tender package."

The civil contract, which is being financed principally by Japanese banks, together with taxes and financing charges over the full term, represents a total project cost of about $80 million. This is also the selling price to the fund. "The facility will be sold and will go into operation immediately on completion", said Logan. "It is therefore in our interest to finish as quickly as possible. The sooner it is sold, the sooner we terminate our capital financing costs. All construction risks are carried by the joint venture and although we are not finished yet, we are at present about 60% through the construction phase and we are running about 15 months ahead of our original contract programme."

The current project is actually the final section of a larger sewerage strategy started some years ago by the Sydney Water Board to provide upgraded sewerage facilities for urban expansion in the Blue Mountains about 90km west of Sydney. It involves a total 18km of tunnelling, the longest section being the present 13.4km-long TBM drive following which the TBM will excavate a further 3.5km-long tunnel. There are two major branch lines 1.3km and 150m long, the first being excavated by a small Mitsui roadheader from Japan and the latter by drill+blast using hand held equipment

The contract also includes nine ventilation and drop shafts of either 3m or 1.5m diameter and between 60m and 115m deep. These were sunk as blind holes prior to tunnel excavation using a cranemounted kelly bar rig.

Fig 1(above).The current tunnel links into the existing Hazelbrook sewer to complete a major sewerage upgrading scheme for the rapidly expanding, naturally beautiful, and highly environmentally sensitive area of the Blue Mountains west of Sydney. The curved alignment follows the contours of the topography to maintain sufficient lever over the tunnel and prevent breaking the length into several shorter tunnels. A straight tunnel would have required a deeper alignment with deeper portal cuttings or a deep working shaft. Roadheader excavation of the intermediate spur tunnel started once the TBM had provided access. Muck from this operation is also loaded onto and transported by the continuous conveyor. (Below) A special bridge section provides the transition from the advancing TBM's transfer conveyor and the continuous conveyor

At the down stream end, the project joins the existing Hazelbrook tunnel to conduct flow to the Winmalee treatment plant and allow several smaller and now inefficient treatment plants to be decommissioned

TBM ready for its 13.4km long drive

The 9.5km long Hazelbrook tunnel was built under a conventional contract for the Water Board in 1991-93 by Costain Mining, now owned by Peabody Resources of the USA. It was excavated using a 3.35m diameter Robbins TBM with a loco and skip muck haulage system. Rock conditions were similar to the current project, with the TBM boring easily through competent, dry and largely homogeneous sandstone of between 50-80MPa unconfined compressive strength. At best, the Hazelbrook operation achieved an advance rate of about 1,000m/month and an average of about 580m/month with its rail bound mucking system.

"Given the length of our longest drive and the limitations of a loco and skip system in a small diameter tunnel, we decided from the start to use a continuous conveyor mucking system", said Logan. "As it is, the 3.4m diameter tunnel is larger than necessary, but anything smaller would make rapid, man-entry excavation impossible." Under current predictions, the tunnel will run at about 20% of the available capacity when commissioned.

In the competent sandstone, the tunnels will remain largely unlined with 50-150mm of shotcrete applied to seal and protect areas of poorer, jointed or softer rock. There will also be a concrete invert to protect against erosion. At the start, McConnell Dowell/Obayashi decided to use preassembled sets of timber sleepers and rail track during excavation, and to cast the invert as a second-pass operation. This programme has since been changed to installing 1.5m long precast concrete invert segments.

Loop take up of the continuous conveyor system and its steep rise to the discharge

At tender stage, the geological report prepared by Connell Wagner, the joint venture's consulting engineer, predicted that immediate support would be minimal with spot bolting and wire mesh required in areas where interfaces with potential seams of claystone and mudstone could cause instability in the crown. "With this information in hand, we based our tunnelling programme on the Hazelbrook average with a 15-20% increase for a continuous conveyor mucking system", said Logan. "This gave us a conservative 700m/month base rate and an overall TBM excavation programme of 24 months. With work starting on site in June 1993, we planned to have the facility in operation by October 1997. On current predictions, we now anticipate completion in the second quarter of 1996."

The TBM being used on the project is the first sold after the merger between Atlas Copco and The Robbins Company and the first to use technology from both companies. The 3.4m diameter machine has an installed cutterhead capacity of 1,000kW, a cutterhead speed of 12.5rev/min, and a maximum thrust of 6350kN. With a maximum torque of 763kNm, the cutterhead is dressed with 25 extended tip 17in Robbins cutters. four in the centre, 16 face cutters and five gauge cutters, each with a rating of 22 tonne/cutter. The TBM has a 1.5m stroke and is driven by four AC water cooled ABB electric motors. It has a hydraulic cutterhead jog motor and a torque-limiting coupling and adjustable maximum torque transfer between the connection motor and the planetary gear. The 15m long x 600mm wide TBM conveyor has a maximum speed of 2.4m/sec.

"We did consider buying a refurbished TBM, but at the time this new machine was competitively priced and proved a more cost-effective choice for the long drive in hand", said Logan. "The machine is designed to support a cutterhead of between 3.2m and 4.25m diameter and given its performance to date, we would expect to see a reasonable resale value."

Crown fixing of the conveyor

Continuous conveyor
Much of the high performance achieved is attributable, in combination with the TBM, to the continuous conveyor mucking system. While continuous conveyor systems are finding increasing application in the US, this is the first tunnel in Australia to use the technique. The belt itself, unlike the rubber ply belts used in the USA, is a 600mm wide solid woven or PVC belt supplied by local manufacturer BTR Nylex. It is not the longest conveyor system used in tunnels (longer systems were used in the US on the TARP project in Chicago and on the Superconducting Super Collider tunnels in Texas) but it is one of the smallest tunnels in which such a system has been used to date. The conveyor installation was designed by local company Prok. The TBM back-up and its integration with the continuous conveyor was designed and assembled by Terratec Asia Pacific.

Instead of being fixed to the sidewall, as is more usual, the conveyor in this small tunnel is suspended from the crown. At the forward end, it is fixed to the backup. A special bridge section allows non-stop insertion of the sidebars or stringers to transfer the belt from the moving back-up to the fixed crown-mounted structure hung from chains screwed into bolt holes drilled into the rock.

As tunnelling progresses, the conveyor is extended using 400m sections of belt. Each new length is quickly clipped in at the tension control assembly or belt magazine at the portal. The clipped joints are then vulcanised on a Sunday, an operation that takes about half a shift. Like the TBM, the belt system is controlled by a computer-based programmable logic controller (PLC). It runs at 2.5m/sec providing a carrying capacity of about 200 tonne/h.

The belt is driven by 55kW hydraulic drive motors which are installed at about 2km intervals. At the time of the site visit, the conveyor belt was more than 22km long (double the tunnel's 10.5km length) and a sixth drive head was about to be installed. "We were hoping to reach about 11.3km before installing the last drive head", said Logan, "but belt speed is now down to about 1.5m/sec. The next drive head is needed now and will be installed next weekend at about the 11km mark."

Next 400m reel of belt extension

The single continuous belt runs on idlers at 3m intervals and follows four alignment curves, all at a 600m radius. The tunnel is being driven on a 1.5-2% uphill gradient and the belt rises outside the portal on a 15% gradient to discharge onto the spoil heap.

Continuous conveyor systems are relatively expensive compared to conventional loco and skip systems. The system here at the Blue Mountains was said to cost nearly as much as the TBM. As Logan explained: "They can only justify their cost if higher production can be achieved which has certainly been the case on this job. Had we used loco and skips, we would have used 5m³ cars and would have needed large locos of 16 tonne or more. We would have needed much increased ventilation or, alternatively, used battery rather than diesel locos. We would also have needed at least three California switches in the 13.4km drive plus a double track gantry or a towed switch."

Despite high progress rates, current conveyor systems in civil tunnelling projects are a major cause of production downtime. On the Blue Mountains job, downtime caused directly by conveyor problems about equals TBM boring and regripping utilization. Besides extensions and drive head installations, several things can go wrong with conveyors including data cable breaks, computer malfunctions, hydraulic problems, belt breaks, and idler replacements. If conveyor reliability could be improved, utilization would increase by about 15%, according to Logan. This would not necessarily improve the shift or daily averages but it would improve overall weekly and monthly averages. "Non-utilization is a more constant factor than production", said Logan, "and the fact is that as production/h goes up, overall utilization goes down. If the system ran consistently at optimum, we could increase production to about 2,500m/month tops. As it is, the effectiveness of the conveyor in operation together with the reliability of the TBM, has resulted in our best production of 2,197m in a month which included the 27 working days of August 1994 and our best week. This was achieved when we were installing sleepers and rail sets rather than invert segments."

Tension control unit keeps belt tension at about 2 tonne

Production cycle
Work on the project is run with three crews on three 5h shifts/day, five days/week plus two production shifts on Saturday. Sundays are off days except for necessary and time consuming maintenance, and each crew gets every third Saturday off. A bonus system is in operation and shift changes take place at the face.

In general, the rock was said to be softer than expected. Where the sandstone was expected to average about 50MPa and range from 20-150MPa, it has actually averaged between 40-50MPa and ranged from about 10-100MPa. It has also contained a higher clay content than expected and there have been zones of jointed rock through which water ingress has been more than anticipated. In one jointed zone, weak material running at the face caused over excavation. Spilling and erecting steel sets held up production for about a week.

"Although there have been long runs where no support has been required, we have installed an average of more than two 1.5m or 2m long resin anchored rockbolts/m", said Logan. "At most we have installed an array of eight bolts in a meter. Where these can be installed from the drill rig station on the back-up, about 15 doesn’t interfere with production. Where they are required immediately behind the cutter head rockbolting in this confined area using hand-held drills does hold up TBM advance. The high clay content has also slowed progress. When mixed with the cutterhead's dust suppressing water-spraying system, the clay clogs the excavation chamber and need to muck buckets and be cleared manually.

“Hard rock has never hindered progress", said Logan. "In fact, the harder the rock the faster we go. In harder rock the TBM cuts bigger chips, there are no blockages and no hold for support. In this soft to medium rock, only about 50% of the TBM's installed 1,000kW (260 amps) power is being used."

Assembled TBM backup

On average, it takes about 10 minutes to complete a 1.5m stroke cycle. Higher rates than this, although possible by the TBM, cause problems such as overloaded cutterhead muck buckets and spillages off the conveyor particularly at the drive head stations. These areas are now monitored by sensors to warn of excessive spills off the overhead conveyor which require manual cleaning and can, if unnoticed, build up quickly, blocking off rail access. Production, on the whole, is limited by the speed and capacity of the conveyor.

With a 50% silica content, most of which is quartz, the sandstone is highly abrasive. Cutter wear is high at about one cutter for about 400m³ (40 linear meters) but this is within expectations. Cutter changes and other service tasks including water supply, ventilation and cable extensions and routine mechanical maintenance, are undertaken during the first four hours of every day shift. Cutter changes take about half this time. Worn discs are reworked on site, and, thanks to the softer rock conditions, there have been no disc bearing failures, no cutter breakages, and no saddle replacements to date. There have been no problems either with the TBM's main bearing, its rubber lip seals or its lubrication and positive grease feed systems. The kelly's slide bearing is also standing up to the task although it has been adjusted twice. TBM availability was said to be very high at about 95%, excluding cutter changes.

As expected, dust was recognized as a potential problem which is in fact exacerbated by the continuous conveyor. The TBM is equipped with normal dust suppressing systems and the excavated rock is sprayed again before starting its long trip on the conveyor. In addition, all tunnel workers and visitors are supplied with dust masks. That the ventilation shafts were sunk prior to excavation also assists ventilation. They provide a natural ventilating effect and, rather than forcing ventilation from the portal, allow the main ventilation fans to be kept within about 3km of the face.

Breakthrough into the 60m deep reception shaft was achieved in January 1995, some nine months earlier than the original programme. The TBM and its back-up will now be lifted out of the 3m diameter shaft and returned to the initial working site for its second 3.5km long drive. The conveyor will be wound back out of the tunnel and cut at each splice for rolling once again into easily-handled 400m rolls.

With the learning curve behind them and the teams in top form, perhaps the second drive will see better advance rates than already achieved. Regardless, the Blue Mountains experience will remain an outstanding achievement in the tunnelling hall of fame.

References

Award winner at Pike River - TunnelTalk, Feb 2009
Design-build responsibility at Rosedale - TunnelTalk, Feb 2009
Underwater works for Hobson Bay - TunnelTalk, Dec 2008
Programme challenges at Bogong - TunnelTalk, Mar 2009