The Greatest Challenges in TBM Tunneling:

Experiences from the Field

Jim Clark and Steve Chorley

The Robbins Company

ABSTRACT: TBM tunneling is an ever-increasing prospect for underground construction, and with each
new tunnel bored there are unknown elements. When boring through the earth, even extensive Geotechnical
Baseline Reports can miss fault lines, water inflows, squeezing ground, rock bursting, and other types of
extreme conditions. This paper will draw on the considerable field service experience within Robbins to
analyze successful methods of dealing with the most challenging conditions encountered.

INTRODUCTION Issues Encountered (Trapped Cutterhead)

Many tunnel projects are located in areas with rela- The machine was launched in the spring of 2012 and
tively poor access along the tunnel alignment and almost immediately encountered geology that was
bored under extremely high overburden. These two substantially more problematic than was described
factors often result in limited geological information. in the geological reports. The geology consisted
It would be reasonable to state that the deeper the of blocky rock, sand and clays. As a countermea-
tunnel, the greater the level of uncertainties. When sure that was immediately put into place to avoid
faced with these uncertainties everyone involved the cutterhead becoming stuck in the blocky mate-
with project including the owner, the contractor, and rial, crews began boring half strokes and half resets.
the machine supplier must be prepared to tackle geo- This ensured that there was always the option of
logical surprises. This paper describes the problem- rapidly retracting the cutterhead in the event that
atic geological conditions and associated difficulties torque reached critical levels. After boring through
faced on three separate projects and the measures 80 meters of these difficult ground conditions, the
that were taken to overcome these difficulties. machine encountered a section of extremely loose
running ground with high clay content. A collapse
KARGI KIZILIRMAK HYDROELECTRIC occurred in front of the cutterhead and the cathedral
PROJECT effect resulted in a cavity forming that extended
more than 10 m above the crown of the tunnel. The
Background
weight of the collapsed material trapped the cutter-
The Kargi Kizilirmak Hydroelectric Project is head. After several unsuccessful attempts to clean
located on the Sakarya River, near the Beypazarı out and restart the cutterhead, consolidation of the
district of Ankara province in Turkey. The Robbins ground above and in front of the machine was car-
Company supplied a 9.84 meter diameter Double ried out. Injection of polyurethane resins via lances
Shield TBM and continuous conveyor system to inserted through the cutter housings and muck buck-
Gülermak for excavation of the 11.8 kilometer head ets was the method utilized for consolidation opera-
race tunnel (see Figure 1). The tunnel is being driven tions; however, injection locations were restricted to
through a mountainside with up to 600 m of over- the available openings and subsequent attempts to
burden. The geology consists of volcanic rock and restart the cutterhead proved to be unsuccessful.
softer limestone for the first 3.0 kilometers, followed
by harder rock including marble and basalt for the Bypass Tunnel
remainder of the tunnel alignment. Due to the varia-
After assessing all the available options it was
tion in geology the ground support regimes range
decided that a bypass tunnel would be required.
from pre-cast segmental lining for the first 3.0 kilo-
Robbins Field Service assisted Gülermak with
meters transitioning into ring beams, rock bolts and
bypass tunnel design and work procedures to free
shotcrete as the tunnel moves into more competent
the cutterhead and stabilize the disturbed ground.
geology. Several unique features were incorporated
Blasting techniques were ruled out due to concern
into the TBM design to facilitate installation of the
over further collapses caused by blast induced
various ground support regimes.

101
 North American Tunneling Conference

Figure 1. Double shield TBM for Kargi

Figure 2. Bypass tunnel

vibration; hence, the excavation was undertaken At this point in time it was believed that the
using pneumatic hand held breakers. Details of the collapse was an isolated event and that the geol-
bypass tunnel can be seen in Figures 2 and 3. ogy would improve as the overburden increased;
Upon completion of the bypass tunnel, further however, material for a second bypass tunnel was
stabilization of the collapsed material above the stored at site. Unfortunately this measure proved to
machine and the ground ahead of the machine was be prudent planning. Although the machine passed
carried out. The injection process this time was far through several weak zones successfully, a fur-
more comprehensive due to the vastly improved ther five bypass tunnels were required to free the
access provided by the bypass tunnel. The area cutterhead during the first 2 kilometers of boring.
around the cutterhead was able to be cleared of mate- Robbins and Gülermak analyzed the bypass tunnel
rial and the cutterhead was freed, allowing boring to excavation procedures and implemented improve-
recommence. ments that resulted in a reduction in the time taken

102
 2014 Proceedings

for bypass operations from 28 days to 14 days. One of drilling closer to the tunnel face more than com-
of the main aspects of the improved procedures was pensates for the time spent adding extensions to the
the implementation of breaking out for the bypass tube length. The location of the canopy drill reduces
tunnel through the telescopic shield area of the TBM the length of each canopy tube by more than 3 meters
rather than the accepted norm of breaking out from when compared to installation using the main TBM
the tail shield. This modification resulted in reducing probe drills. Apart from the obvious savings in drill-
the length of each bypass tunnel by over 4 meters. ing time, the extra 3 meters of drilling length can
result in a significant increase in hole deviation. The
Pipe Roof Canopy diameter of the canopy tubes is 90 mm, each canopy
typically extends up to 10 m from the tunnel face and
The possibility of installing ground support such as
the drill positioner, carriage and slew ring provide
fore-poles or a pipe roof canopy ahead of the tun-
130 degrees of coverage.
nel face was investigated and after consultation with
Gülermak a custom design canopy drill was installed
Squeezing Ground
in the forward shield for installation of a tube canopy
(See Figures 4 and 5). The space in the forward shield The time dependency of ground behavior is due to
area is limited; hence, the extension section of each the creep and consolidation processes taking place
tube is only 1.0 m in length. However the advantages around the tunnel (Anagnostou & Kovári 2005). In
many cases the convergence can be a gradual process
taking place over a period of days, weeks or even
months. On several stretches of the Kargi tunnel,
rapid convergences occurred in the space of a few
hours. The geology at the time of these rapid con-
vergences consisted of Serpentine with high content
of swelling clay. The convergence was of a radial
nature, and distributed relatively evenly around the
profile of the TBM.
Probe drilling ahead of the tunnel face iden-
tified the majority of the areas considered to be at
risk from squeezing conditions. As it is generally
accepted that there is a direct relationship between
TBM advance rates and problems caused by squeez-
ing ground it was essential that TBM downtime was
minimized while boring through these stretches.
On the occasions that squeezing ground had been
Figure 3. Bypass tunnel excavation

Figure 4. Custom canopy drill

103
 North American Tunneling Conference

lubrication reduced the frictional forces between the

shields and converging ground.

Solution (Gear Reduction)

To further mitigate the effects of squeezing ground or
collapses, custom-made gear reducers were ordered
and retrofitted to the cutterhead motors as a solution.
They were installed between the drive motor and
the primary two-stage planetary gearboxes. During
standard boring operations the gear reducers oper-
ate at a ratio of 1:1, offering no additional reduction
and allowing the cutterhead to reach design speeds
for hard rock boring. When the machine encounters
loose or squeezing ground the reducers are engaged,
which results in a reduction in cutterhead speed but
Figure 5. Canopy tube drilling
the available torque is increased. Figure 6 shows the
torque curves for both standard and reduced gearing.
identified all outstanding maintenance works, repairs Since the installation of the canopy drill and the
and replacement of worn cutters was completed increase in available cutterhead torque, the TBM has
before boring through the zone of concern com- traversed several sections of adverse geology includ-
menced. Inevitably, even after taking these precau- ing stretches of severe convergence without becom-
tions there were unscheduled stoppages. On many ing trapped. As of November 2013 over 4,250 m of
occasions the only successful means of restarting the boring has been completed.
machine after stoppages in convergence zones was
to utilize single shield mode boring. In this mode LOS OLMOS
the TBM gripper shoes are retracted, the main thrust
cylinders are closed up and the auxiliary thrust cyl- The Los Olmos tunnel is a 12.5 km long water
inders are utilized to propel the machine forward by transfer tunnel that was bored through the Andes
thrusting off the segmental lining. The typical thrust Mountains in Peru. Odebrecht was the main contrac-
force for standard boring operations using the main tor and the tunnel was driven using a 5.3 m diameter
thrust cylinders on the Kargi machine is approxi- Robbins main beam TBM. It is the World’s second
mately 21,000 kN. On several occasions thrust force deepest civil works tunnel after the Gotthard base
up to 136,000 kN was applied through the auxiliary tunnel with overburden of up to 2000 meters. The
thrust system before the machine could be freed from tunnel alignment is through complex geology con-
squeezing ground. Generally after boring one or two sisting of quartz porphyry, andesite, and tuff with
meters in single shield mode the TBM was freed and rock strengths ranging from 60 to 225 MPa. The
it was possible to return to double shield mode. machine crossed over 400 fault lines including two
On several stretches of tunnel the rate of conver- major faults of approximately 50 m wide.
gence coupled with the comparative softness of the The machine was launched in March 2007 and
ground caused the gripper shield to act as a plough by February 2008 it had bored over four kilometers.
and force muck into the telescopic shield area. The The geology over the first 4,000  m of boring was
buildup of material became so severe that a muck- far more challenging than was anticipated. As the
ing system had to be installed in the telescopic shield height of the overburden increased, the geological
area. The system consisted of two electric hoists conditions became gradually more severe and long
mounted on a running beam that allowed muck kib- stretches of extremely loose, blocky ground were
bles to be placed, lifted, and emptied directly onto encountered. The rock stresses caused by the high
the TBM conveyor. overburden also resulted in over 16,000 recorded
Another measure utilized to combat the effects rock bursting events. TBM utilization was as low
of the squeezing ground was the application of a as 18.7% of working time because rock support
polymer based biodegradable lubricant to the extra- installation was requiring a very high 43.5% of the
dos of the TBM shields. Eight injection ports were working time (Roby & Willis 2008). One of the
installed around the perimeter of the forward shield main problems faced was ground deterioration and
and lubrication was injected when boring through the resulting falls of blocky ground. The majority of
convergence zones. It is difficult to quantify the these events occurred during the time taken for the
advantage obtained as there was very little consis- newly excavated bore to pass behind the rear fingers
tency in ground conditions and associated thrust of the roof shield, where ring beams and mesh are
pressures; however, it is clear that the application of installed.

104
 2014 Proceedings

45500

40500 Standard Torque

35500 Exceptional Torque

129% (Ƭorq Low

30500
speed)
Cutter Head Torque, kN-m

194% Ƭorq (Low

25500
spd Exceptional)
20500

15500

10500

5500

500
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00
Cutter Head Speed, RPM

Figure 6. Cutterhead torque curves

Figure 7. McNally support system

McNally Roof Supports System forming rectangular pockets with a length of 1.4 m.
The pockets run from the rear side of the cutterhead
During consultations between Robbins and
to the trailing edge of the roof support. At a later
Odebrecht, a decision was taken to modify the
stage when the ground conditions worsened these
machine to facilitate the installation of the McNally
pockets were extended to cover the profile of the side
roof support system, which allows support to be
supports. Figure 7 shows details of the modifications
installed directly behind the main roof shield. The
that were implemented to enable use of the McNally
main components of the initial modification con-
System.
sisted of removing the shield roof fingers and

105
 North American Tunneling Conference

The procedures for installation of the McNally

Roof Support System were as follows:

1. Two slats, formed from 6 mm rebar are

loaded into each of the pockets.
2. The upper slats in each pocket are drawn
from the pocket and pinned to the tunnel wall
by means of ring beams or rock bolts.
3. As the machine advances the slats are held in
place and extruded from the pockets.
4. When the leading edge of the upper slat is
completely withdrawn it is fixed to the trail-
ing edge of the lower slat, with an overlap
of 200 mm. Additional slats are then loaded.

The Main advantage of the McNally support system Figure 8. Loose rock held in place by McNally
is that is installed closer to the face than other ground system
support methods used on TBMs, which reduces the
required standup time of the excavation. It holds smoothly with productions rates of up to 526 m a
loose rock in place (see Figure 8) which in turn helps month. Tunnel support ranged from spot bolting
to mobilize the strength of the rock mass and main- through to complete ring beams, mesh, shotcrete and
tain the inherent strength of the tunnel arch. When rock bolts.
used correctly the system can significantly reduce
the time taken to provide adequate support and can Rock Bursts
also offer reductions in the level of support required.
Incorporation of the McNally support system By mid-October 2006 with over four kilometers of
and various other modifications to the TBM resulted boring completed and overburden of over 1,100 m,
in a steady increase in production rates in spite of several major rock bursting events occurred. The
continuous rock bursting events. The machine broke rock bursting was accompanied by moderate to
though in December 2011 having achieved produc- severe loss of ground so the support regime was
tion rates in excess of 670 m a month. upgraded to include ring beams, rock bolts, lagging
sheets and concrete backfilling. During the following
PARBATI HYDROELECTRIC PROJECT 50 m of boring the incidences of rock bursting events
STAGE II increased to the point that at times they were almost
continuous.
The Parbati Hydroelectric Project Stage II is located
in the Kullu district of Himachal Pradesh in India. Probe Drilling
A nine kilometer section of the head race tunnel is
being driven by a 6.8 m diameter open gripper type The Parbati project is typical of many hydroelectric
TBM, through a highly stressed mountain range at projects in that it is located in a mountainous area
the foot of the Himalayas. Overburden along the where there is limited access and high overburden
TBM section of the headrace tunnel reaches as high above the alignment of the tunnel. These factors
as 1400 m. The geology consists of granite/gneiss- resulted in limited availability of detailed geological
ose, and quartzite with bands of biotite schist and information. Bearing this in mind geological inves-
talc. Rock strengths are expected to exceed 270 MPa. tigation ahead of the tunnel face was essential and
The contractor, Himachal Joint Venture (HJV), was achieved by maintaining a strict regime of probe
purchased a refurbished Robbins-Atlas Jarva TBM drilling.
from Norwegian Company NCC. The machine was A routine probe hole (P1) was drilled at chain-
launched in May 2004 and after the completion of age 4056 m at the 11 o’clock position on the face.
500 m of boring NCC handed over the machine to The depth of the hole was 27 m and minor ingress
HJV. HJV operated the machine up to Chainage of water and silt was observed from probe chainage
1300  m but due to technical difficulties associ- 4066.5 m up to 4077.3 m. A decision was made to
ated with the machine and relatively slow progress drill a second probe hole (P2) at the 1 o’clock face
they approached Robbins for assistance. Robbins position in order to gain further information on the
provided a field service team to supervise repairs, geology/hydrology ahead of the face. During the
maintenance and operation of the TBM. Repairs night shift of the 18th November 2006 the P2 probe
were carried out, the machine restarted and despite drilling operations were underway when the crew
crossing several minor fault zones operations went heard several cracking sounds emanating from the

106
 2014 Proceedings

surrounding rock mass. Shortly after these events pressure gauge and minor inflows of silt and water
the initial probe hole (P1) was observed to be dis- began to flow through fissures in the rock mass close
charging water and silt under high pressure. It took to the face. Further rock bursting fractured the rock
the crew almost 2½ hours to seal the 51 mm hole mass surrounding the collar of probe hole P1 caus-
using a mechanical packer attached to the probe drill. ing the rock to fall away and expose the hole behind
During these 2½ hours approximately 180 m3 of silt resulting in an inrush of water and silt under massive
and 125,000 liters of water were discharged, and pressure. The crew tried unsuccessfully for several
continuous rock bursting was occurring. hours to insert a packer into P1 to stem the flow of
material but at 7:00 am with silt levels rising rapidly
Inundation and rock bursting continually occurring, the tunnel
was evacuated for safety reasons.
Due to the high pressure and high volume of the dis-
During the 25th November it was deemed
charge it was decided that the best course of action
impractical and unsafe to enter the tunnel. Water
would be to drill drainage holes to relieve the pres-
ingress was measured at the portal throughout the
sure ahead of the tunnel face, before a programme
day and flow rates gradually increased until they
of consolidation grouting could be undertaken. Both
exceeded 7000 liters/min. On the 26th November
drainage holes and grout holes were to be drilled via
flow rates stabilized so a team entered the tunnel to
standpipes. The design of the standpipe arrangement
assess the situation. They observed that the inunda-
consisted of drilling a 75 mm hole 5.0 m deep, insert-
tion had almost completely buried the TBM (see
ing a 6.0 m long, 64 mm steel pipe with a threaded
Figure  9) and that silt and water were still flowing
section on the trailing end, and anchoring the pipe in
from the probe hole. However the pressure of the
place by cement grouting. A ball valve and pressure
discharge had reduced and a crew was mobilized
gauge were attached to the threaded end of the pipe.
and managed to seal the probe hole by inserting a
A third probe hole (P3) was drilled utilizing
mechanical packer. The total amount of silt depos-
the standpipe arrangement, to a depth of 38 meters.
ited during this event was over 14,000 m3 and the
Although the location of the P3 probe hole was adja-
cleanup operation took over 2 months.
cent to the P1 probe hole location at the 10 o’clock
face position, it did not encounter silt or high pres-
TBM Refurbishment & Modification
sure water. The next course of action was to attempt
drilling a fourth hole that would intersect probe hole Robbins was awarded a refurbishment contract for
P1 to facilitate drainage operations. The hole was the TBM as many parts and assemblies had been
drilled though a standpipe which was subsequently damaged due to being submerged for a prolonged
fitted with a valve to enable regulation of flow, a period of time. Once the refurbishment was com-
pressure gauge and a length of 75 mm hose to allow plete, cement grouting with OPC was carried out
drainage of material directly into the tunnel muck to consolidate the ground in front of the TBM. The
cars (see Figure 8). project was then held up due to contractual issues
On the 24th November probe hole P1 was until January 2010 when Robbins was awarded a
successfully intersected and drainage operations contract to modify the TBM. The main components
were underway when several rock bursting events of the modifications included installation of pock-
occurred. The pressure in probe hole P1 gradually ets for the McNally support system, upgrading the
increased until it exceeded the 25 bar capacity of the cutterhead support system, and an improved probe
drilling system. The existing probe drilling system
accommodated drilling from two fixed positions
only. The modified system provides 110 degrees of
coverage.
After the modifications were completed fur-
ther consolidation grouting was carried out before
the machine advanced. A system of boring in incre-
ments of 8.0 m advances interspersed by extensive
consolidation grouting proved to be successful and
the machine successfully crossed the geological fea-
ture that had caused the inundation. 50 m of boring
was completed before the project was again held up
due to contractual issues. The project was retendered
early 2013 and works resumed in November 2013,
although boring will not commence immediately as
remedial works to ground support are required in
several sections of the tunnel.
Figure 8. Drilling through stand pipe at Parbati

107
 North American Tunneling Conference

Figure 9. Parbati TBM buried in silt

CONCLUSIONS Technical features on the TBM are not the

only insurance required when faced with geological
TBMs are often the only viable option for the exca-
uncertainties. The contractor should have an action
vation of long tunnels with high overburden, due
plan in place to cover all eventualities. Ground treat-
to the impracticalities of opening several faces via
ment materials and equipment, as well as bypass tun-
adits to enable the application of traditional tunnel-
nel materials and equipment should be available at
ing methods. As with the three case studies outlined
site. Again the cost of these items is almost insignifi-
in this paper geological surprises are frequently
cant when compared to the cost of the project, and
encountered in long and deep tunnels. Due to cost
their availability will provide substantial reductions
constraints contractors often decide to procure a
in project delays should they be required.
TBM that is suited to the geological baseline reports
rather than opting for additional features that insure
REFERENCES
against geological anomalies. It is more often than
not possible to retrofit additional features but TBM Anagnostou, G. & Kovári, K. 2005. Tunnelling
down time for preparatory works, installation, and through geological fault zones. Proceedings of
component lead times is usually substantial. The International symposium on design, construc-
actual cost of the additional features applied to the tion and operation of long tunnels (Taipei).
machines described in this paper would have been a Roby, J. & Willis D. 2008. Coping with difficult
fraction of the costs involved had they been installed ground and 2000m of cover in Peru. Proceeding
during the manufacturing process. When compared of World Tunnel Congress, Underground
to the overall cost of a project, additional features Facilities for Better Environment and Safety
installed during manufacturing become almost (India).
insignificant.

108