7th Pipeline Technology Conference 2012

A major step for a pipeline: crossing the Serra do Mar State Park in
Brazil
Philipp Elsner, Babendererde Engineers, Germany

Abstract
From 2008 to 2011 the state-owned company Petrobras installed a major gas
pipeline from a newly constructed refinery directly at the southeast coast of Brazil
over roughly 95 km to an inland connection point with the existing major gas
distribution network. The refinery produced natural gas from a recently developed
field 180 km off the coastline.
In order to build the pipeline major logistical and technical challenges had to be
overcome. A literately major step was the crossing of the Serra do Mar State Park, a
5 km wide mountainside quickly rising from sea-level to over 800 m and protecting
the last remains of the Atlantic Rain Forest in that region. In order to cross the park a
6,2 m diameter tunnel had to be drilled under it with a Hard-Rock Tunnel Boring
Machine (TBM). At the end of the tunnel 550 m deep shafts had to be drilled to
connect the tunnel with the surface.
Due to a very strict time schedule and delays in the tunnel excavation the installation
of the pipeline into the tunnel finally had to start parallel to the tunnel excavation.
This was only possible because the pipeline was welded outside of the tunnel portal
and then consecutively pushed into the 3% rising tunnel over the full distance of
5.000 m.
But also the surface pipeline installation had problems to overcome. Heavy rain falls,
related flooding’s and muddy ground sometimes made it impossible to use any heavy
machinery. In order to avoid those negative effects the longest tent in the world was
erected along a critical part of the surface pipeline.
This paper describes the construction works of this pipeline project on land,
especially around the critical section of the Serra do Mar Mountains.

1. Introduction
From 2008 to 2011 the gas field “Mexilhão” 180 km off the southeast coast of Brazil
was connected to the major transport net of the country and significantly cut the gas
imports from other countries to Brazil. The works on land included the construction
of a new refinery close to the coastal town of Caraguatatuba in the state of São
Paulo and a 95 km long 28” pipeline from the refinery to the connection point with the
network.

A special challenge was posed on the construction teams of the pipeline due to
topography and environmental restrains of this region. The coastline between Rio de
Janeiro and São Paulo is often very narrow, in some places less than a kilometre.
Behind that strip the landmass rises within a few kilometres by 800 m and forms a
giant “step” until it continues in a hilly landscape inland. As a consequence the
clouds coming in from the sea often release a lot of rain along the mountainsides
colliding with this large, natural seawall.
Under these conditions a special fauna developed over the past thousands of years,
the so-called Atlantic Rain Forrest. Due to massive clearing the Atlantic Rain Forrest
had been reduced to about 2% of its original size until the remains in that region
finally were turned into a nature protection area, called the “Serra do Mare” State
Park. The strip of the state park just above the new refinery is 5 km wide and
construction works or entry with any kind of machinery was strictly prohibited.

1
 7th Pipeline Technology Conference 2012

Figure 1 – Overview map

Tunnel

As the park extends all along the coast there was no way around. The only way
possible for the pipeline to cross the state park was to tunnel under it.

2. Layout of the pipeline construction

The pipeline installation of the complete project can be divided into several different
sections. The first section extends from the Mexilhão field at the bottom of the 500 m
deep Atlantic over 20 km towards the platform PMXL-1, which had been installed at a
sea depth of 170 m. From there the pipeline continues over 160 km, circling around
the famous island “Ilha Bella”, towards the bay of the town Caraguatatuba, where the
shores approach was made by a conventional trench.
The new refinery was installed on the open grass fields of a huge cattle farm, the
only open area available. From there the pipeline continued towards the foot of the
mountains, where it enters into a 5000 m long tunnel to pass the State Park above it.
Once on the other side of the park and on the upper side of the natural mountain step
the pipeline continues for 90 km over land to the connection point with the main
transport net near the town of Taubaté.

Figure 2 – Aerial photo of the construction side around the Serra do Mar

Shaft

Tunnel

Refinery

Pipeline

2
 7th Pipeline Technology Conference 2012

In the following chapters the article will mainly focus on the tunnel construction and
pipeline installation within the tunnel. In addition to that also some complications of
the pipeline installation on the farm land at the foot of the mountains and of the first
kilometres on top of the mountains will be described.

3. Pipeline installations on the farm land

The pipeline construction on the farm land was conventional weld-in-place installation
in an open trench. But the working conditions were often difficult nonetheless.
The farm land was mostly flat and just a little bit above the sea level. The
groundwater level was just below the surface. Longer or stronger rains rapidly raised
the ground water level and flooded larger parts of the farm land. This happened
repeatedly during the construction phase. The ground was soft and wet and did not
support any heavy machinery. Therefore a working road had to be constructed prior
to the pipeline works.
The working road was basically a mini dam. It was made of two rows of small
wooden piles, in between which several layers of geotextile, logs and the natural
found soil were placed. It was a cheap, environmental friendly and an easy way to
build the working road. Despite many delays through heavy rain and flooding the
pipeline from the coast via the refinery to the tunnel was ready in time.

4. Design and construction along the tunnel

4.1 Tunnel design
The analysis of different designs for the crossing of the state park led the project
engineers to the following solution. A straight dead-end tunnel had to be built starting
at the foot of the mountains. It had a constant inclination of 3,0%. Thus the rock
cover at the end of the tunnel would still be about 550 m and vertical shafts were
required for the final connection of the pipeline from the tunnel to the surface.

Figure 3 – Section of the expected project geology

The knowledge about the geology along the tunnel was rather limited. The rock was
mainly gneiss. Based on a surface inspection several milonitic shear zone and
several diabase dykes could be expected. But a closer inspection of potential critical
points along the tunnel by borings was impossible because it was forbidden to bring
the necessary equipment to the nature protection area.
As consequence of the topography and the environmental restrains the excavation
was only possible from the lower end of the tunnel. Parallel excavations from the top
side or from intermediate access points were impossible. Therefore it was decided
that a Tunnel Boring Machine (TBM), the first ever in hard rock in Brazil, should
excavate the tunnel and due to the limited information about the geology it was

3
 7th Pipeline Technology Conference 2012

specified that the type should be a Double-Shield TBM. This type of TBM protects
the crew inside the machine from the rock completely. When the rock has sufficient
strength this machine fixes its position in the rock by gripping to it by the extension of
very strong hydraulic jacks installed perpendicular to the axis of the machine while
the cutter head at the front is moved forward and excavates the rock. In the same
time a circular lining made of precast concrete elements is installed at the back of the
TBM, but still within the shield. In case the rock mass does not offer enough
resistance to the gripping, the machine can still move by pushing itself forward from
the installed lining in the back. But then the lining installation and the excavation are
two consecutive steps and the overall advance rate reduces significantly.

Figure 4 – Schematic drawing of the double shield tunnel boring machine (TBM)

In the final design it was agreed between the client and the contractor that the lining
of the TBM tunnel would have an inner diameter of 5,4 m and a thickness of 25 cm.
The excavation diameter of the TBM was 6,16 m. The void between the outside of
the lining and the rock was to be filled with pea-gravel which was pumped into the
gap parallel to the construction. The complete support of the TBM for the transport
out of the excavated rock and transport in of the lining segments, the pea-gravel and
all other supplies was planned with trains.

4.2 Shaft design

At the border to the nature preservation area a series of vertical shafts was planned
around the end position of the tunnel for the gas pipeline, permanent ventilation and
other future pipelines. In order to meet the tight time schedule of the project the
drilling of the shafts was planned top-down in three phases. In the first phase
injections drills were done around the positions of the future shafts to close any
fractures in the rock. The second step was the 10” pilot bores for the shafts and the
third step 68” enlargement bores. The pipeline was to be placed into the blind shaft
and fixed in position by concrete before the arrival of the TBM.

4.3 Pipeline design inside the tunnel

The pipeline was to be constructed by a new method just recently developed for a
project called GASDUC III in the state of Rio de Janeiro. In that project the pipeline
was installed by pushing it step-by-step into the tunnel from the entrance. The
pipeline was welded at a welding station at the tunnel entrance. Inside the tunnel the
pipeline was bedded on supports with motorised rollers on the top, which not only

4
 7th Pipeline Technology Conference 2012

provided the bearing for the pipeline but also the ability to move the pipeline through
the tunnel.

Figure 5 – Tunnel cross section (design drawing and as-built)

But because of the dismantling of the TBM at the end of the tunnel excavation all
works regarding the pipeline installation had to wait until the TBM was out of the
tunnel as the size of the dismantled TBM parts to be transported were expected to be
so large that they would require the complete tunnel cross section.

4.4 Tunnel site preparation

In 2008 the preparation of the tunnel construction site started parallel to all other
works along the future pipeline and since mid-2009 the assembly of the TBM
commenced. Unfortunately the works in that area had been delayed by half a year
due to permission problems with environmental agencies. Thus the bottle neck of
the complete campaign, the tunnel construction, which cannot be accelerated as
simple as a construction site on the surface by increasing manpower and machinery,
immediately got onto the critical path.

Figure 6 – TBM installation and preparation of the tunnel portal

Parallel to the TBM assembly the excavation of the tunnel by the conventional
method started. The idea was to ease the start of the TBM by cutting a starter tunnel
through the top layers of the rock, which was weathered to various degrees, until at
least a class IV rock, according to the rock mass rating system by Bieniawski, had
been found. After 270 m of excavation such rock was finally discovered. A starter

5
 7th Pipeline Technology Conference 2012

box was prepared at the end of the conventional tunnel and the TBM was moved in
after that. In November 2009 the TBM finally started the excavation process.

4.5 Tunnel excavation

The TBM advance developed as planned during the starting months. But already to
the end of the first quarter of 2011 it became obvious that the advance rates did not
rise up to the original expectations. Improvements in the logistics, aiming to enable
the TBM to advance at very high top speeds whenever geology and machine allowed
it, did not bring any significant acceleration. When it became obvious that the
tunnelling delays would impact the starting date of the gas delivery, a strategy was
developed to start the pipeline works already during the TBM tunnelling.

4.6 Pipeline support installation

As explained above it had been planned to place the pipeline onto steel support
installed to the sidewall of the tunnel. The supports had been prefabricated and were
fixed with a set of bolts. The bolt holes had to be drilled in a distinct pattern to match
the connection points of the supports. This was done with a set of core drills which
the contractor had installed onto a steel frame. It could be moved with a hydraulic
arm which again was mounted onto a custom made rail car. The necessary power
units for the core drills were on the rail car as well.

Figure 7 – Installation of the pipe supports

The logistic problem of the support installation was that it could not be executed
parallel to the tunnel excavation because in that time a supply train for the TBM was
passing the tunnel every 30 minutes. The compromise was that the support
installation was done during the maintenance phase of the TBM, which happened
once every 24 hours and normally lasted at least four hours. Of course the
disadvantage was that the TBM was cut off from any heavy supply during the
maintenance which required a thorough planning of the TBM crew to not lose any
time.
On the top of each pipe support a roller bearing with an electric motor was installed.
Those motors were operated from a control centre at the tunnel entrance.
During the installation it had to be taken care of that the roller bearings were installed
on a straight line. Although the tunnel alignment had been planned as a straight line,

6
 7th Pipeline Technology Conference 2012

the excavation in reality always deviates up to several centimetres around the design
axis. Tolerances from the lining installation come on top of that. The installed tunnel
ring, in this case made of six individual segments, can deviate from the theoretical
position as well as the perfect circular shape. Therefore the final position of the roller
bearing had to be verified for each support prior to installation and corrective
measures were necessary whenever the bearing would have ended up outside the
acceptable tolerances.

4.7 Pipeline welding

As soon as enough pipe supports had been installed inside the tunnel, the installation
of the pipeline started. A welding station was assembled in front of the tunnel
entrance in line with the future pipeline inside the tunnel. An 18 m piece of steel pipe
was placed on supports in front of the tunnel and welded to the pipe section already
inside the tunnel. But the platform used by the welders was so wide that it extended
into the structural clearing of the tunnel train line. Because of this the platform had
been designed retractable. Every time a trained passed the welding works were
interrupted and the platform was moved aside. As soon as the train had passed the
platform was moved back and the welding continued. Those interruptions lasted one
or two minutes at a maximum.
When a welding seam was finished it was checked and insulated. Then the whole
pipe string was moved forward by the length of one pipe and the procedure started
again.

Figure 8 – Pipeline welding station at the tunnel entrance

Using this method 1,5 km of pipeline could be installed parallel to the excavation of
the tunnel. After that distance an intermediate train station inside the tunnel, where
outbound and inbound trains could pass each other in this otherwise single tracked
tunnel, stopped the further progress of the pipeline. But the installation of the
supports was continued even beyond the position of the train station.

7
 7th Pipeline Technology Conference 2012

4.8 End of TBM excavation

In the middle of January 2011 the TBM reached the position planned as the end of
the tunnel. But at the time it had been decided to start the pipeline installation
parallel to the excavation it had also been accepted that the TBM would never again
leave the tunnel due to the lack of sufficient space for the disassembled TBM parts.
Instead the client bought the TBM from the contractor and the tunnel was excavated
for 150 m more than originally planned. This additional excavation took one and a
half weeks but was significantly less than the three months originally reserved for the
TBM disassembly which had been on the critical path of the time schedule. So just
one day after the TBM had reached its new final position it was abandoned by the
tunnel contractor and full control over the tunnel was given to the pipeline contractor.

4.9 Final installation of the tunnel pipeline

As soon as the tunnel was under the sole use for the pipeline installation the erection
of the supports was continued at maximum speed. Three more welding stations
were installed at the tunnel entrance to increase the pace the pipeline could be
extended and it soon reached the end of the tunnel.

Figure 9 – Pipeline advancing inside the tunnel

5. Connection of the pipeline from tunnel and shaft

Unfortunately the delay of the tunnelling works was not the only problem for the
overall schedule of the project. The drilling of the shafts turned out to be much more
difficult than anticipated.
After the first pilot drill had been brought down successfully, the subsequent reaming
drill failed because the reamer started deviating uncontrollably from the pilot bore.
While the second pilot hole was bored the drilling crews were able to recover the first
pilot bore hole over the complete depth. As there was a potential risk that the second
reaming drill could fail as well the client decided to use the two pilot bores for
temporary pipelines. This way he would be able to start the operation of the gas
refinery which would not start the production at full capacity but with a learning curve
anyway. Following those considerations two cross drifts were started by drill and

8
 7th Pipeline Technology Conference 2012

blast from the TBM tunnel to reach the first two pilot drills. Once found they were
connected with the 28” pipeline inside the tunnel and on the surface.

Figure 10 – Temporary connection of the pipeline to the surface

Meanwhile the third pilot drill was started from the surface and successfully reached
to the required depth. Also the subsequent reaming drill succeeded this time.
Once the 28” pipeline had finally been installed in the bore hole the third cross
passage at the end of the tunnel was built. But because the pipeline was already
under operation blasting was not allowed anymore. Thus the rock had to be
excavated by hand using core drills and afterwards hydraulic splitting to break the
rock. As soon as the foot of the vertical pipe had been found, it was integrated into
the major pipeline at both ends and the refinery could go into full production.

Figure 11 – Final cross drift for the connection to find the vertical 28” pipe

9
 7th Pipeline Technology Conference 2012

6. A tent for a pipeline

Like the underground installation of the pipeline also the surface installation for the
first kilometres behind the shafts got under the risks to develop serious delays. The
problem there was that the alignment of the pipeline was passing through very hilly
woodland for the first 3 km. Once cleared from the protective layer of plants the
surface soil was very deceptive to rain and even an improvised road dam like in the
area around the refinery would not have been of any use. But knowing the risk of the
schedule and rather than betting on a dry season the client decided to take proactive
measures to ensure the timely installation of the surface pipeline. In order to be save
from the frequent and strong rains a large, 3,3 km long tent was built over the
alignment.

Figure 12 – 3,3 km long tent from the air and from the ground

Shaft site

For the ease of installation and to limit the amount of necessary material the tent was
inflated by compressed air from the inside. At both ends air locks allowed the
entrance of machinery and more points allowed the access or exit for the workers in
case of an emergency. The weather in 2010 proved that the tent construction had
been a wise decision (it rained a lot) and the pipeline on the surface was finished in
time.

Figure 13 – Working inside the tent

10
 7th Pipeline Technology Conference 2012

7. Summary
From 2008 until 2011 a new gas field off the southeast cost of Brazil was connected
with the country’s major network. 180 km of pipeline in the sea, a new refinery and
95 km of pipeline on land had to be constructed. The biggest challenge for the
pipeline construction was the crossing of the Serra do Mar State Park, a nature
protection area. A 5 km long tunnel had to be built, using a hard rock tunnel boring
machine for the first time ever in Brazil. At the end of the tunnel over 500 m deep
shafts were necessary for the connection to the surface. In the tunnel an innovative
installation method was chosen by the contractor. He pushed the pipeline into the
tunnel from the portal over the complete length. The pipeline installation even had to
be started while the tunnel was still under construction.

11