At the Gerede water Transmission Tunnel in Central Turkey, a 31.6km-long water supply line has been designated a national priority due to severe and chronic droughts in the capital city Ankara. Drawing water from the Gerede River, it will be the longest water tunnel in Turkey once complete.

But completing the tunnel has been an obstacle in itself. The project has been called the most challenging tunnel currently under construction in Turkey, and with good reason. Out of three standard Double Shield TBMs originally supplied by a European manufacturer to bore the tunnel, 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 tunnel, but to do so it would need to cross dozens of fault zones and withstand intense water pressures up to 20 bars.

The Gerede Water Tunnel is just one tunnel at the top of a long list of challenging projects in Turkey, however. Like so much of what is considered difficult in tunnelling, the story of why Turkey’s projects are so tough begins and ends with geology.

WHAT MAKES A PROJECT CHALLENGING?

Turkey is in a tectonically active region controlled on a grand scale by the collision of the Arabian Plate and the Eurasian Plate. At a more detailed level, a large piece of continental crust almost the size of Turkey, called the Anatolian block, is being squeezed to the west. The block is bounded to the north by the North Anatolian Fault and to the southeast by the East Anatolian fault. Geology in the fault zones tends to be highly variable and unstable. Several current and recent projects are affected by these faults, including the Kargi HEPP, the Bahce-Nurdag High Speed Rail Tunnels, and the Gerede water tunnel.

At Gerede for example, geologic testing and borehole samples showed a mix of volcanic rock including tuff, basalt, and breccia, giving way to sedimentary formations like sandstone, shale, and limestone, all punctuated by fault zones that contained clay and alluvium. The ground conditions are of immense interest to Nuh Bilgin, professor of mine and tunnel mechanisation at Istanbul Technical University and Chairman of the Turkish Tunnelling Society. “I believe that [Gerede] is the most challenging project of all current tunnelling projects. I believe that after the completion of the project, it will have a special place in the tunnelling history of Turkey.”

It’s a very different geology from that of Istanbul, where many TBMs have been used for metro and sewer tunnels. “The geology in Istanbul necessitated the use of EPB TBMs in the area,” explains Bilgin. Over the years it was discovered that EPB TBMs could outperform open-type and shielded rock machines as well as slurry TBMs in the city’s mixed geology. That’s not true in the rocky fault zones of Central Turkey. “I believe that in the North Anatolian Fault and East Anatolian fault zones or in high overburden, mountainous areas with high tectonic stresses, Crossover/Dual Mode type machines will have more chance to succeed.”

Bilgin has had four decades of experience in geological research and tunnelling projects, and his decision is informed by field experiences. He also believes that a shielded rock machine with certain capabilities could get through the fault zones. “We had an experience in the North Anatolian fault zone at the Kargi HEPP where the squeezing of the TBM was a big problem. The TBM’s emergency thrust system, shield lubrication facilities and multi-speed gearboxes played a big role in the success of the project. However, I believe that an experienced tunnelling crew, which is capable of understanding these necessities, is a key factor for the success of the tunnelling activities. In such cases I believe that the close cooperation between the machine manufacturer and the contractor will increase the effectiveness and the credibility of the project.”

The 10m double shield machine was modified in the tunnel, effectively allowing it to operate like an EPB in fault zones, with high torque and low RPM. These are the same principles used in Crossover machine designs.

Also on the Kargi project, the machine utilised a canopy drill and umbrella arch to consolidate ground directly above and in front of the machine, and also operated with continuous probe drilling. “When probe drilling and umbrella arches are used together in a tunnelling project, machine utilisation time and mean daily advance rates may decrease considerably. However, the contractor must be aware of the fact that if these techniques are not used, the machine will most likely become jammed, necessitating a bypass tunnel that will require much more downtime,” says Bilgin. As far as the selection of TBM type, Bilgin says there are several trade-offs. “A Crossover /Dual Mode type TBM may be a little more expensive than a classical single shield TBM, but this type of TBM may overcome many difficulties arising from the complex geology.”

REEXAMINING GEREDE

At Gerede, the Anatolian Fault Zone has certainly presented many obstacles. During the original excavation, the joint venture of Kolin and Limak purchased three 5.56m diameter double shield TBMs from a European manufacturer to deal with the challenging geology. Each machine was to bore a roughly 10km section of tunnel. The TBMs arrived at the site in 2011—the first machine (TBM-1) was launched from the north portal in a relatively homogenous section of rock with low cover of 13m. The TBM completed its 9,588m of tunnel while achieving good average advance rates. The machine encountered some ground water inflows and squeezing that caused delays but it was still able to complete its tunnel.

TBM-2 was launched from an intermediate shaft under higher cover, starting at 60m and reaching over 400m as it bored toward the south. The rock was more transitional in this section, and the TBM had bored a significant section of its 10,339m tunnel when it encountered a massive inrush of water that flooded the TBM and tunnel.

The TBM was boring downhill and the water had to be pumped out, which took some time. The TBM was deemed a loss, and removed from the tunnel.

TBM-3 began boring from the south portal under increasingly high cover that would reach a maximum of more than 500m. The TBM was several kilometres into its 11,653m downhill drive, struggling in karstic aquifer conditions that required polyurethane injection and slowed tunnelling, when its problem became worse. A high water inrush of 1,500 litres/second flowed into the tunnel, causing the machine to become stuck. This inflow resulted in enough pressure to crush the TBM shields and send cylinders catapulting into the back-up. Dye tests showed that the water had come from a river flowing overhead and entered into the tunnel through a cave system. As quickly as it had started, the Gerede Water Transmission Tunnel ground to a halt with two TBMs stuck 9km apart.

A new strategy

The Kolin/Limak JV had to develop a new strategy given the incredibly difficult ground conditions. They contacted The Robbins Company, which suggested a Crossover (dual-mode type) TBM for the remaining 9km section of tunnel. The 5.56m diameter XRE (standing for a Crossover between Rock and EPB) could effectively bore conditions in both rock and mixed ground under water pressure by converting between modes. The revised geology was now understood to contain more significant fault zones and an aquifer system that could cause high-pressure water inrushes of up to 20 bar.

However the ground was expected to improve as the TBM advanced and consist mostly of sandstone, limestone and tuff with a maximum UCS in the range of 100MPa.

Kolin/Limak needed a machine that could effectively bore in those wideranging conditions, but also statically hold water pressure up to 20 bar in the event of an emergency flow—a failsafe with which none of the standard double shield TBMs were equipped. Cutterhead design

Due to the geology, the Gerede machine required a convertible cutterhead optimised for hard rock. Multiple abrasion-resistant deflector plates were included to handle abrasive rock chips as they entered the cutterhead chamber and into a screw conveyor. 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. The problem of regrind occurs in bidirectional heads when already-excavated muck enters through the cutterhead and back out of the next opening, wearing the back portion of the cutterhead. The phenomenon can be very severe in bi-directional cutterheads depending on the ground conditions.

Drive system

To cope with difficult ground, the Gerede machine was also equipped with special gearing, allowing the machine to function as either an EPB or a hard rock TBM. This function is done by adding another gear reduction – heavy duty pinions and bull gears accommodate high torque at low speed, allowing the machine to bore through fault zones and soft ground without becoming stuck. By shifting gears the machine is able to output high speed in order to cut hard rock using the same installed horsepower.

High-water design

The new TBM was designed to statically hold up to 20 bar pressure in the event of a massive water inflow. In order to protect the machine from such 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. In the event that 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 also the gripper and stabiliser shoes are sealed off in the same manner. All of these locations have two rows of seals with a grease-filled cavity between to hold constant pressure.

Specialised screw conveyor

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. In order to account for potential wear, the screw has been designed with replaceable wear plates along its entire length. The screw itself is also made up of short sections that can be removed and replaced if needed. Multiple access hatches are included for maintenance of the wear plates, while two large, removable outer casings can accommodate the change-out of entire screw sections.

A special feature of the conveyor is the ability to seal itself off so the TBM can continue boring. If a fault zone is encountered with large amounts of water, the machine will still be able to continue excavation. In this case, the screw can be used in a sequential operation. The rear screw conveyor gate is closed, sealing off the interior of the machine from the incoming water. The screw extension cylinders will then push the rear of the screw back, thus pulling the screw out of the cutting chamber and into the screw casing. Next, the bulkhead gates will close. The rear gates can be reopened and the screw conveyor can run, emptying the casings of water and muck.

Once the screw has been emptied, the rear gate can be sealed. The bulkhead gates can then be reopened and the screw re-extended into the cutting chamber. Boring can then commence until the screw conveyor is once again full. Once the screw is refilled, it can again be retracted and sealed, starting the process over again. This process can be slow, but will get the machine through a fault zone and into better ground.

Probe drilling

Due to the unpredictable ground conditions, it is 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 utilises a standard array of 12 Ø100mm ports angled at 7° that are equally spaced around the rear shield. Each port is sealed by a ball valve until it is needed for probing. Ten of the same-sized ports are also located straight through the forward shield for probing and grouting. Six additional hatches are built into the pedestal at the front of the machine. The hatches can be used with a pneumatic percussive drill in the centre section of the cutterhead.

The probe drill on the Gerede machine also has an extra feature for cases of emergency. The drill is designed to pull back behind the tail shield and at an angle of 16°, so it can drill behind the shields and into the segment lining. If water has filled the cutting chamber and the pressure is high, drilling a hole in the roof of the tunnel will allow the water to spill out, thus relieving the buildup of pressure on the machine.

THE BORE PATH

The Robbins XRE TBM was launched in summer 2016. The machine is using some components from the original double shield TBM back-up, as well as the remaining segments being stored for the project. Crews excavated a bypass tunnel to one side of the stuck double shield (TBM-3), and the Robbins TBM components were walked in through the south portal. The machine was assembled using Onsite First Time Assembly in an underground launch chamber.

“The logistics of getting components through the existing tunnel were the most challenging thing. The assembly chamber was 7km from the portal. The water inflow made it difficult to get the materials to the machine,” says Glen Maynard, Robbins Field Service site manager.

To overcome this, custom flat cars equipped with hydraulic lifts were used to transport the bigger sections of the TBM through the tunnel to the build chamber. Large sections of the TBM shield were positioned high enough to pass through the segment lining using the hydraulic lift and side shift adjustments as the cars passes through the tunnel. The Robbins 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.

“The section of tunnel from the launch chamber up to the point adjacent to the buried double shield was reasonably known due to the existing bored tunnel info,” says Maynard. “We knew to expect large water inflows with flowing materials at any point over the initial 300m of tunnel.”

In fact, the machine was required to be used in EPB mode as it encountered water pressures up to 23 bars, alluvium, flowing materials, and clay. Water pressure was lowered by draining the ground water through the rear shield probe drill ports, which were equipped with normally-closed ball valves. Probe drilling became routine after advancing 50m past the section that buried the original Double Shield TBM. The Crossover machine bored at approximately 30mm/min through the bad sections of ground, with any limitations in speed the result of the existing tunnel belt conveyor, which tended to have spillage issues resulting in significant cleanup. Despite the challenges, Maynard was proud of the accomplishment in the first section of tunnel. “We crossed through material that caused the double shield machine to become stuck. We bored 80m to the side of where that machine is currently buried and we passed it.” To date, the machine has bored more than 25 per cent of the remaining tunnel length.

REDUCING RISK

With projects like Gerede as an example of just how difficult Turkey’s geology can be, how can contractors reduce risk? Bilgin says probe drilling is not something to be taken lightly in geology as challenging as the Anatolian Fault zone. “Most contractors try to avoid probe drilling since they think that this decreases daily advance rates, which is not definitely correct. However with careful analysis of the drilling data, it is possible to detect the problematic zones and take mitigation actions such as using an umbrella arch in front of the cutterhead and applying polymers.” Bilgin additionally discussed a technology in development that could help with identifying major zones of failure. “We collected a huge amount of data from TBM tunnelling projects in difficult ground conditions in Turkey. In collapsing or squeezing zones we have observed a few key indicators like a critical peak of thrust/penetration or a change in thrust/torque ratio that serve as an indicator before the failure occurs. However, these changes are very sudden and sometimes it is difficult to detect. We are working on a computer programme for quick detection of these critical zones. I have to admit that this is not an easy task.”

However he is optimistic that tunnelling in challenging conditions like those in Turkey will become easier over time. “Turkey’s complex geology requires investing more on the side of ground investigation and on a custom, high-technology TBM. Contractors can reduce risk by employing a highly qualified and experienced tunnelling crew,” says Bilgin. He adds that obtaining good service from a TBM manufacturer, and maintaining close cooperation among all parties involved, may be the most important risk reduction tools of all.