THREE GREAT CHALLENGES: Experience from TBM Tunnelling in Difficult Ground’, was the subject of the BTS online presentation in January 2021 given by Robbins president Lok Home. He presented three demanding mechanised tunnelling projects in rock with unique challenges and each providing a great learning experience. The projects discussed were:

1. AMR Water Tunnel, India

2. Gerede Water Tunnel, Turkey

3. Yin Han Ji Wei Water Transfer Tunnel, China

WHAT IS MEANT BY DIFFICULT GROUND?

Difficult ground in rock can include fault zones, rock bursting, high water inflows, squeezing ground, fractured rock, mixed-face conditions, high water pressure and more.

As one of the challenges in hard rock tunnelling, a rock burst is a sudden and explosive failure of the rock which can cause damage to the TBM and personnel. High stress concentration in the rock can lead to rock bursting close to the excavation face or in the excavated tunnel. Compared to conventional tunnelling, rock bursts require a different approach to rock support, with the focus being to handle dynamic loads and rock deformation.

Fault zones are the other common challenges in tunnelling projects which often contain unstable and collapsing ground that can cause a machine to be delayed or stuck. Fault zones are often points of high-pressure water inflows that in extreme cases can inundate the machine and tunnel.

Very hard rock is the other challenge in tunnelling. In extremely hard rock, wear on the cutterhead, cutters and main bearing is typically high. This is compounded when the rock is also abrasive, causing additional wear. A failure of the cutterhead on the machine in this ground condition can quickly result in a sequence of disc-cutter failures if not detected and managed quickly. Though automatic cutter-failure systems can detect remotely when cutters are about to fail, they do not have 100% accuracy and so visual inspection is usually required. But visual inspection is very difficult where a slurry TBM is used in rock. Therefore, the combination of rock tunnels and slurry machines is not an ideal option and is not recommended.

Long tunnels are more likely to run into challenging geology, such as fault zones, water inflows and others. They have a higher rate of machine wear and tear. Logistics and ventilation of long tunnels are more complex and the longer the tunnel, the more difficult the conditions will be.

It is very common for tunnelling projects to be given a geological report with hundreds or even thousands of pages. But the unfortunate reality is that the chance of encountering different geological conditions to those in the report is very high.

Regardless of geologists’ knowledge and experience, providing the required degree of accuracy in the geological information prior to the construction of a project is very difficult – even with an excessive number of boreholes – as the information at this stage relies heavily on interpretation. The advice here would be: ‘’Buy some insurance when you buy a tunnel boring machine.’’

AMR WATER TUNNEL IN INDIA:

India’s Alimineti Madhava Reddy (AMR) water tunnel is a 43.5km irrigation tunnel passing beneath Rajiv Gandhi Tiger Reserve; when complete, it will be the longest tunnel in the world without intermediate access. It will transfer floodwater from the Krishna River to arid regions of India’s Andhra Pradesh state, providing irrigation to 1,200km2 (400,000 acres) of farmland and clean drinking water to 516 villages.

The local government forbade diamond drilling within the park and the geological data was limited to satellite imagery, surface mapping and historic data from the adjacent Srisailam hydropower project. A 43.5km-long tunnel through granite with high quartz content and UCS exceeding 300MPa with very limited geological information was an obvious indication of a very challenging project on the horizon and this proved to be correct.

Two identical 10m-diameter double-shield tunnel boring machines (TBMs) were used for this project. Both were initially assembled successfully onsite using the Onsite First Time Assembly (OFTA) process. Rather than assembling the machine in a manufacturing facility, OFTA saves contractors both time and money in terms of personnel and shipping costs.

Assembly of the first machine took place in a large launch pit at the outlet portal site, using gantry cranes to hoist components into place. Machine parts including the cutterhead, gripper system, forward shield and telescopic shield were then assembled in a concrete ‘cradle’. The assembled TBM and back-up then crawled forward by reacting against invert segments installed progressively up to the tunnel entrance.

The first machine was launched in March 2008. The second Robbins machine was assembled onsite at the opposite end (the inlet portal). While the assembly of the TBM at the inlet took approximately two months, the dam downstream of the inlet portal flooded the area and it delayed the start of the project by eight months.

Available data at the start of the project indicated that the tunnel would be driven through 60% quartzite and 40% granite, but in practice it was the other way round and that made a big difference in progress rates.

It was assumed that cutter consumption in the quartzites would be considerably higher than that in the granites. This has proved not to be the case and the average cutter consumption in quartzite has been 295m3 per cutter, while the average cutter consumption in granite has been 205m3 per cutter. The TBM penetration rate in granite was 1.42m/hour while the penetration rate in quartzite was 2.07m/hour. A summary of the TBM cutter consumption and penetration rates in granite and quartzite on the AMR Water Tunnel project makes interesting reading.

With around 70% of the tunnel length completed, the project has been on hold for about two years due to financial issues. While the major structural components of the TBMs are robust enough to finish the tunnel, some parts of the machine require refurbishment prior to project restart because of the long stoppage period.

GEREDE WATER TUNNEL IN TURKEY

The Gerede Water Transmission Tunnel is arguably one of the most difficult projects attempted in the world of tunnelling. Numerous fault zones and intense water pressures up to 20bar are just a couple of the challenges that have to be overcome.

The 31.6km-long tunnel is to transfer water from the Gerede River to the capital, Ankara. Once completed it will be the longest water tunnel in Turkey. The geology consists of two formations: volcanics including tuff, basalt, and breccia; and sedimentary including sandstone, shale, and limestone.

Prior to Robbins’ involvement, three standard 5.6m-diameter Double Shield TBMs, originally supplied by another manufacturer, attempted to bore the tunnel. Each TBM was supposed to bore approximately 10km of the tunnel. The geology turned out to be much worse than the geologists and the owner expected. Of the three, two became irretrievably stuck following massive inflows of mud and debris. In 2016, a Robbins Crossover XRE machine was launched to excavate the final 9km of the 31.6km-long water supply tunnel.

TBM FEATURES

The following features were proposed by Robbins for a new TBM to overcome the challenge:

¦ Crossover TBM with single shield and EPB TBM characteristics

¦ Single-direction rock cutterhead

¦ High-strength shield with closed bulkhead to resist water and mud inflows

¦ Bottom screw conveyor

¦ Torque-shift system

¦ High-pressure articulation + tail seals

¦ High trust capacity to overcome high pressure and cutter loading.

¦ Multi-position probe drilling.

One of the key features of Crossover (XRE) machines is that they should run in rock and in EPB mode. Generally, for rock machines the speed is more important, and the torque is hardly needed. In contrast, EPB machines requires a lot of torque because the muck is against the cutterhead, but the speed is hardly needed. Crossover TBMs have both features together with a system called ‘torque shift’ to transfer from one mode to the other. Figure 1 illustrates the range of cutterhead torque and speed in EPB and rock machines.

Due to the geology, the Gerede machine required a convertible cutterhead optimised for hard rock. The cutterhead was designed for ease of conversion between hard rock and EPB modes, and cutter housings can be fitted with either disc cutters or tungsten carbide tooling. In addition, the cutterhead is designed to operate in a single direction. The setup, which is used on all XRE machines, allows for greater efficiency while excavating, with lower power requirements and less chance of regrind. To cope with difficult ground, the Gerede machine was also equipped with special gearing.

In order to protect the machine from the expected high water pressure, an extensive sealing system was put into place. Around the main bearing, there is an outer row of six seals and an inner row of three seals. Between each seal, the cavity is filled with grease to ensure a constant pressure. If the machine is shut down and an inrush of water overtakes the machine, a pressure sensor will detect this presence of water and flush the seal system with grease in order to continually protect the seals. The articulation joint and the gripper and stabiliser shoes are sealed-off in the same manner.

Perhaps one of the most important parts of the Gerede TBM design is the screw conveyor. Because of the potential for massive amounts of water, the machine must have a sealed screw conveyor. However, running rock through a screw conveyor can be highly abrasive. To account for potential wear, the screw was designed with replaceable wear plates along its entire length. The screw itself was also made up of short sections that can be removed and replaced if needed. Multiple access hatches were included for maintenance of the wear plates, while two large, removable outer casings could accommodate the change-out of entire screw sections.

Due to the unpredictable ground conditions, it was necessary to detect and grout-off zones of concern wherever possible in order to protect the machine from loose ground and water pressure. The machine featured a standard array of 12 100mm-diameter ports angled at 7° that were equally spaced around the rear shield. Each port was sealed by a ball valve until it was needed for probing. Ten of the same-sized ports were also located straight through the forward shield for probing and grouting. Six additional hatches were built into the pedestal at the front of the machine. The hatches could be used with a pneumatic percussive drill in the centre section of the cutterhead.

The Robbins XRE TBM was launched in summer 2016 and the machine began boring at a slight angle to the rest of the tunnel, bypassing the stuck machine before gradually meeting up with the original tunnel alignment. Water flow inside the tunnel was still 450-570 lit/sec, straining logistics including rail transport. Probe drilling became routine after advancing 50m past the section that buried the original double-shield TBM. During its bore, the TBM encountered a total of 48 fault zones. Despite the constraints of the difficult geological conditions and the time it took to reach the TBM within the tunnel, crews achieved a best day of 29.4m, best week of 134.6m and a best month of 484m. Excavation of Turkey’s longest water tunnel came to an end in December 2018.

YIN HAN JI WEI WATER TRANSFER TUNNEL

This tunnel is part of a massive scheme to link water supply from two rivers to the towns and agricultural areas of Shaanxi Province in central China, also generating hydroelectricity as part of the water transfer scheme. Much of the total tunnelling is drill and blast, with particularly challenging sections reserved for TBMs with mountainous geology in mainly quartzite and granite without detailed geotechnical report. The project is using an 8m-diameter Robbins Main Beam TBM which is expected to bore two headings of lengths 9.88km and 7.63km.

The TBM was designed for hard-rock tunnelling with high stresses and high water inflows. Some of the TBM features are summarised in table 1.

TBM assembly was undertaken in March 2015 in an underground cavern shown in yellow (Figure 4) at the end of a 3.9km drill and blast adit with 8.18% down gradient to the main tunnel alignment.

It was immediately apparent that the rock was much harder and more abrasive than predicted. Unconfined compressive strength on the drive to date averaged 194MPa, with an average abrasive index of 5.36, quartz content of 71.6% and an average integrity coefficient of 0.8. Geological conditions were underestimated by 50% or more which resulted in higher cutter wear and slower progress rates.

Rock bursting is the other major challenge on this project. High stresses and the potential for rock bursts were anticipated for 95.5% of the original two drives to a total of 18.3km with the following details:

¦ Slight rock bursts were predicted in 545m of the alignment.

¦ Moderate rock bursts for 13,030m of the tunnel

¦ Strong bursting along a 3,880m length of tunnel

To prepare for these conditions, the TBM was equipped with the McNally Support System, consisting of pockets in the roof shield and steel slats.

The encountered rock bursting has been much more severe than predicted. During the first 9.3km of tunnelling, there were 397 sections of rock-burst activity. In the second heading which started in September 2020, there have already been 18,045 rock bursts. Of those events, 444 cases have been strong with 1,736 cases recorded at an energy of more than 100kJ. There have been 88 bursts of more than 800kJ with the highest energy rock burst reaching 4,080kJ. In 2,000kJ the grippers usually fail. In this machine, special measures were in place to use relief valves and capacitors for such explosions to control the gripper failure.

A series of 739 rock bursts across a total 1,865m length of tunnel required boring to be stopped so as to install additional support to stabilise the rock and lower the risk of injury to workers and damage to the machine. Since the start of boring, the machine has had to stop eight times to make repairs due to severe bursting, including damage to the grippers, thrust cylinders, cutters and cutterhead.

Once the rock bursting starts, it can continue for minutes, hours or even days but not normally for weeks or months. It stops when the stress is released, allowing the workers to clean up and continue their work. In zones of medium to heavy rock bursting, progress has slowed to about 90m/month due to the need for repairs and for installing extra support.

Later in the project, a micro-seismic monitoring system was used to predict any potential rock-bursting areas ahead of the TBM. The system senses rock stresses in a borehole 20m in front of the face, and the potential for rock bursting can be predicted following a comparative analysis with similar rock burst data in previous occasions. The accuracy rate of the rock burst prediction is approximately 70%.

Water inflow was the other major challenge on this project. Sudden water ingress has occurred a total of 69 times, six of them extreme. The greatest water inflow occurred at the face on 28 February 2016 and exceeded 20,000m3/day from one point. Total water flow into the down gradient heading exceeded 46,000m3/day, flooding the lower TBM motors and the electrical cabinets on the lower deck of the backup gantries. After dramatic flooding events in 2016, the in-tunnel pumping capacity was increased to 41,000 m3/day. This is a 20% additional capacity, and new capacity is 3.4 times greater than the maximum water ingress prediction included in the original design documents.

Systematic probing is an essential activity carried out constantly in front of the TBM to predict the potential rock bursting and to shut-off water ingress in advance of the TBM. At any time, if it is found that the water ingress ahead of the TBM is 70% of the in-tunnel pumping capacity, TBM boring stops to carry out grout injection with special grouting material.

The construction of this tunnel is expected to be completed by the end of 2021. There has been great teamwork and cooperation in this project and the Chinese team have done a great job along with the client team. However, the difficult and challenging condition of the tunnel has added around three years’ delay to the project with additional costs.

RECOMMENDATIONS

1. Good geological advance investigation within the tunnel is cost effective and is highly recommended.

2. Probe drilling in advance of the face is highly recommended (water-powered down-the-hole hammers can drill faster, farther and straighter)

3. Multiple probe holes in faulted areas are recommended. The image (below left) indicates 360-degree probe drilling in advance of the face.

4. Buy insurance! It should be taken into account that whenever geological information is detailed and comprehensive prior to construction in mountain tunnels, there will still be some differences in geology. This is generally not the case in metro tunnels as they are usually closer to the surface and a lot of information can be investigated prior to construction.

5. New techniques can be very helpful, such as Measurement While Drilling (MWD), Amberg TSP System by small blasting and recording the reflection to the receivers to understand the geological condition in advance of tunnelling, Geotechnical BEAM System and other similar methods.

6. Using a water inflow prevention system, robust sealing design and muck-chute closure doors.

7. Ensuring the correct rock support systems are applied.

8. In rock tunnels, it is not recommended to mine under pressure as it is not a practical solution. 

CONCLUSION AND TAKEAWAY POINTS

¦ Geological surprises are frequently encountered in long and deep tunnels.

¦ Contractors often procure a TBM that is suited to the geological baseline reports rather than opting for additional features that insure against geological surprises.

¦ Downtime for retrofitting components is usually substantial.

¦ The cost of additional features installed during TBM manufacture is insignificant compared to the overall cost of a project.

¦ Technical features on the TBM are not the only insurance required.

¦ An action plan needs to be in place to cover all eventualities.

¦ Properly equipped modern tunnel boring machines will go through difficult geology faster and safer than drill and blast.