The Rio de Janeiro Metro system consists of two lines, totaling 35km, that transport some 400,000 passengers daily. Work is currently progressing at Line 1’s south end, to extend it further south into the Copacabana District, one of the most densely populated areas in Brazil.
The extension’s first stage, planned in the early 80’s, was awarded in 1987 on a unit price basis to local contractor Construtora Andrade Gutierrez. At that time the Contractor was only authorised to build Arcoverde Station.
The estimated US$80M new phase, being built by the same contractor and discussed here, was authorised in December 1999 by the contract client, Rio Trilhos Metrô Do Rio De Janeiro, and includes the following main structures;
- The 150m long Siqueira Campos Station at the heart of the Copacabana district. The station will be one of the three busiest stations in Rio, and is expected to provide access to a further 150,000 passenger a day to the metro system. It is scheduled to start operation in December 2002;
- A 180m long, double section (each 6m diameter) tunnel, linking the operating Arcoverde Station to the new Siqueira Campos Station;
- A 250m long tunnel of the same dimensions, beyond Siqueira Campos Station, which will serve as a provisional train parking tunnel, until Phase 2, which will extend Line 1 into the Ipanema District. Construction access for this tunnel is through a temporary shaft, located at Garibaldi Street, at the middle length of this tunnel.
Geology
The existing Arcoverde Station is located in a sound Gneiss rock hill formation and was excavated by drill and blast. The 180m tunnel starts in this rock and progresses from a freshly weathered rock to a completely decomposed residual soil. This is a silty and clayey sand, which loses strength when exposed at an excavation face and when subjected to groundwater percolation. As it approaches Siqueira Campos, the tunnel traverses soft alluvium soils (mixed sand and marine organic clay layers). The groundwater level is always above the tunnel.
Some high rise buildings are found above the tunnel, two of them on shallow foundations in the alluvium, about 6m above the tunnel crown, with damage due to settlement from their own weight.
From Siqueira Campos through the 250m parking tunnel, the soil consists exclusively of alluvium, made of cohesionless marine sands and soft organic clay layers, again, below groundwater level.
At the southern end of Siqueira Campos Station part of the tunnel section is directly below a shallow foundation of a 12 storey building, with 5m cover to tunnel crown.
The surface topography changes from a hilly profile near Arcoverde Station, to a flat terrain near Siqueira Campos, and along the 250m parking tunnel. Ground cover above the tunnels varies from some 30m, near Arcoverde, in rock, to 6m, in alluvium, over the parking tunnel.
Construction methods
Siqueira Campos is 16m deep and is being excavated as a top down sequence (cover and cut), with 80cm thick diaphragm walls acting as retaining walls, braced by the top slab, mezzanine slab and bottom slab.
The Station inner columns are temporarily formed by steel profiles founded on diaphragm wall panels, below the bottom slab. These steel profiles are lined with concrete, after casting of the bottom slab.
Two 15m diameter shafts were excavated at both Station ends to allow tunnelling to start from two fronts without having to wait for Station excavation to reach the bottom slab, after casting of the top and mezzanine slabs. In fact, the Station excavation reached the bottom slab some 12 months after tunnelling began.
A short 20m stretch near Arcoverde Station was excavated entirely by drill and blast with support provided by rock anchors and sprayed concrete. This condition changed to a mixed face, where the rock and soil interface gradually dips towards the tunnel invert.
The design concept first adopted for the tunnels in soil was based on a closed face EPB or slurry shield. The difficulty in stabilising cohesionless sands and soft clays below groundwater level, such as in the alluvium, the lack of available space at surface level, and the need to excavate very close to building foundations reinforced this selection.
However, the client, Rio Trilhos Metrô Do Rio De Janeiro, abandoned the shield technique after a significant devaluation of the local currency in the first half of 1999 resulted in unacceptable cost increases as the equipment needed to be imported from abroad. Other factors also contributed to this decision: the time period between the equipment order and actual excavation start was too long to fit the schedule. Also the shield would have to excavate a mixed face with a high strength; very abrasive rock at the bottom and cohesionless soil at the top. That would not rule out a shield but would imply specifically designed equipment, adding cost and time compared to a conventional EPB Shield.
The alternate method chosen was a sprayed concrete lined tunnel in a previously treated soil by jet grouting. The tunnel was basically hand excavated, with the aid of portable percussion tools. Soil removal was via a backhoe excavator and small size dumper trucks.
The jet grouting concept was designed as a continuous application at the whole excavation perimeter (except where the bottom of excavation was in rock, near Arcoverde), and through the excavation face area, forming 2m thick bulkheads at intervals ranging from 7m to 10m, to prevent major face instabilities.
Since crossovers were not required, a twin section with dividing wall between tracks was found to be more economic, for soil treatment, than a conventional double track single section. Also, the double track, single section would be far too wide to permit excavation in one phase, requiring a side drift and enlargement, not dissimilar to a twin section but requiring a larger excavation area and perimeter.
Lattice girders and welded mesh were used with sprayed concrete to form the 25cm thick primary tunnel lining. Cast in situ and sprayed concrete were used to form the 25cm thick secondary lining.
Horizontal and vertical Jets
For tunnel stretches where access from the surface level was either impossible or unacceptable the horizontal single jet grouting system was selected; consisting of near horizontal columns, 30cm to 45cm in diameter, injected from the tunnel face at regular intervals. The total length of injected column, per advance, ranged from 1200m to 2000m (when including the length required to form the bulkhead).
The horizontal jet grouting technique was applied for the 180m tunnel between Arcoverde and Siqueira Campos, and for a 30m stretch at the southern end of Siqueira Campos, to enable crossing under Figueiredo Magalhães street, which is heavily occupied by underground public services, and also below the 12 storey building mentioned earlier.
For most of the 250m parking tunnel, a solution based on vertical and near vertical jet grouting (double jet system; grout and compressed air) was defined. The Column diameter was in the range 120cm – 180cm with the grout pressure before the injecting nozzle being 40MPa, enveloped by a 0.8MPa air pressure. The basic vertical jet grouting array consists of 13 holes, totaling 130m of injected column per metre of double tunnel,
The advantage of the vertical jet grouting system is that it enables uninterrupted tunnel advance. The horizontal system requires a complete stop of tunnel excavation, at 7m to 10m intervals, slowing tunnel advancement rates considerably.
Tunnel starting fronts
Five tunnel excavation fronts were selected. The first was adjacent to Arcoverde Station, towards Siqueira Campos. Since there was no space available for a surface construction site, all the supply of equipment and materials as well as mucking out were done through the Line 1 track to a depot 12km away, during metro non-operating hours.
At Siqueira Campos two shafts were built enabling simultaneous tunnelling fronts towards Arcoverde and the parking tunnel. A shaft, offset from the tunnel axis, was created at Garibaldi Street where the 250m parking tunnel is roughly divided in two. A tunnel transverse to the track tunnels was excavated from this shaft, allowing two portals to be created, one in direction of Siqueira Campos, and the other towards the end of the parking tunnel.
Difficulties and adjustements
Soil loss whilst drilling the horizontal jet grout columns was experienced due to high groundwater pressure at the drill hole. This was aggravated by a lack of cohesion in the sand, resulting in excess soil being flushed out during drilling at the first jet grouting trials, leading to a potentially unacceptable settlement at surface and on buildings.
The horizontal jet grouting sub contractor, Novatecna, implemented a pressure control valve, called a ‘preventer’, at the drill hole mouth, sealed by a sprayed concrete layer at the tunnel face, and used it to control soil flushing during drilling as well as grout return during injection.
The settings of the preventer were adjusted as a function of settlement readings taken by local sub contractor, Technosolo, at deep and surface settlement points, and at building columns. The preventer was gradually adjusted to shut or open positions, depending upon whether the readings indicated settlement or heave during the process.
The readings were based on high precision topographic leveling during drilling and injection of the uppermost columns, with key settlement points leveled at intervals not exceeding 5 minutes.
The double jet system, being carried out by Brasfond, consists of a simultaneous injection of grout and compressed air, through a concentric nozzle. The high potential energy of the grout fluid (at a pump pressure of 40MPa) is converted to kinetic energy after passing through the nozzle, so that this pressure is not transmitted to the ground.
This mechanism does not occur with a compressible fluid like air and in several instances air bubbles were observed to emerge at surface or through basement slabs, at distances as far as 30m from the column being injected. This was usually linked to heave of shallow foundation buildings by as much as 15mm. Deep foundation buildings were hardly affected by jet grouting heave.
At selected buildings, which were already in a precarious state, and whenever heave was recorded or air bubbling was observed, the jet grout column array was redesigned to substitute the double jet system columns for single jet columns (the compressed air was turned off), about 70 to 90cm in diameter (compared to 120cm to 180cm for the double system). This procedure was very effective in preventing the development of heave.
The jet grout was used firstly to provide a zone of high strength soil, protecting the tunnel excavation; the resulting unconfined compressive strength of the treated soil cores was in the range 2MPa for organic clay, to 5MPa for pure sand. This equaled or exceeded the minimum specified value of 2MPa.
The second main function of the treatment was to maintain dryness, since a defect in the treated zone could allow uncontrollable sand and water inflows into the tunnel (a phenomenon known as piping).
In order to create an impervious chamber, the jet grouting was designed to fully envelope the excavation and provide a bulkhead sealing across the tunnel section at selected intervals, which in practice varied from 7m to 10m.
This geometry was simple to implement in the vertical system, but for the horizontal system the numbers of columns necessary to create an envelope and a bulkhead was such that the treatment of a single chamber lasted for 3 to 4 weeks. Therefore, in the horizontal system the envelope and bulkhead concept was used strictly in alluvium soil, for about 60m in the Arcoverde to Siqueira Campos tunnel.
About 100m of this tunnel was in residual soil. Horizontal vacuum drains were employed to relieve groundwater pressure and simplify and reduce the horizontal jet grouting quantities to about 50% of a full treatment section. Being a stiff material, the residual soil responded well to the effective stress increase, contrary to a soft alluvium.
Experience in other tunnelling works in Brazil with jet grouting indicated that the requirement of a watertight tunnel excavation cannot statistically be achieved to an acceptable confidence level. Sealing a jet grouting treatment zone with such materials as silicates and resins was analysed but abandoned due to excessive costs and time consumption.
The design concept adopted was to install groundwater lowering wells from surface level, and operate if and when necessary. It was considered preferable to allow a moderate consolidation settlement than to risk a collapse. The wells were 60cm in diameter and sunk to 6m to 8m below the tunnel invert level. The wells were equipped with submersible pumps, being automatically switched on and off by electric sensors placed at 1m and 2m elevations above the pump. Their average pumping flow was in the range 1m/hr – 5m/hr.
Horizontal drains were drilled for dewatering each chamber before excavation. If flow continued, a leakage was thought likely through the treated zone. The groundwater wells were then turned on to avoid risk of piping, and excavation proceeded to that specific chamber. Some 60% of the installed wells were operated in that way and this allowed safe excavation in the partially “defect” chambers.
The resulting maximum consolidation settlement was 20mm or less (the soft clay was slightly pre consolidated due to ageing and possibly dry season water lowering). This consolidation settlement was evenly distributed (low distortion), thus inducing only minor damage on some building’s brick walls.
Excavation under buildings
Of the buildings that were directly above the tunnel, two cases should be highlighted. The buildings at 14 and 20 Tabajaras Street are 10 storeys high, built in the ’50’s, with shallow foundations in alluvium, each 6m above the tunnel crown. Both presented a damage record, due to the weak condition of their foundations.
Underpinning of these buildings was not possible since this would temporarily force the inhabitants out of the building, which was not acceptable.
These buildings were densely instrumented, and settlement monitoring was carried out at 12 hour intervals during excavation. Some selected points were read at 5 minutes intervals during drilling and injection of the upermost jet grout columns.
The experience gained through the advancing tunnel fronts and the control of the horizontal jet grouting operations in alluvium indicated that the passage under these buildings could be safely effected. In fact, by controlled use of the preventer, the jet grouting operations did not induce settlement at all. A slight heave of a few millimetres was observed during injection. The final settlement observed on these buildings was some 25mm, almost entirely due to consolidation settlement induced by a groundwater lowering well nearby.
Progress to date
The Siqueira Campos Station top, mezzanine and bottom slabs are cast and the columns lined. The 180m tunnel linking Arcoverde Station to Siqueira Campos is complete, including secondary lining. Of the 250m parking tunnel, about 120m, from Siqueira Campos to the Garibaldi Shaft are completed.
Some 130m of tunnel remains to be excavated, from the Garibaldi shaft towards the end of the parking tunnel. This can be done independently from Siqueira Campos, which will allow an early operation date for the atation, scheduled for the end of 2002.
The remaining tunnel section is expected to be complete in June 2003.
Related Files
Plan and section of the horizontal jet grouting pattern
The tunnel alignment
The tunnel construction sequence
Settlement figures
Section through a vertical jet grout
