TUNNELLING EXPERTS last month questioned contractors' decision to use the New Austrian Tunnelling Method (NATM) on the section of the São Paulo Metro which collapsed in January.
They pointed out that ground conditions made NATM a risky tunnelling option especially as the city has a history of similar failures.
The collapse on 15 January at the site of Pinheiros station, near São Paulo University killed seven people.
The tunnel failure occurred close to a junction with a 40m diameter, 40m deep access shaft, whose sprayed concrete wall also collapsed.
The victims were passers by who were sucked into the crater that opened up as the tunnel caved in. Four were in a mini-bus.
It was the fourth NATM failure on São Paulo Metro projects in 25 years.
ATM is an 'open face' excavation technique. Small sections of the tunnel are worked at a time, with thin layers of sprayed concrete and steel girders providing support. Tunnel walls are usually monitored for movement, with additional concrete applied to halt excessive deformation.
The 18.5m diameter 45m long section of tunnel that collapsed was designed to house two platforms, either side of the metro's twin tracks. It was being excavated in three stages. A 4m high heading had already been dug and work was underway on the rst of two 4m bench excavations.
'It seems there's been a failure to learn from past experience, ' said UK-based NATM expert John Anderson. Anderson played a central role in the Health and Safety Executive's investigation into the October 1994 Heathrow Express NATM collapse.
He added: 'After NATM collapses in Sydney, Lausanne and Barcelona within the last two years, there are enough messages around for people to know that this can be a dodgy method in certain conditions.'
spokesman for turnkey contractor Consórcio Via Amarela, consisting of Odebrecht, OAS, Queiroz Galvao, Andrade Gutierrez and Camargo Correa, conrmed that the 18.5m diameter tunnel was being built using NATM to designs by São Paulo based consultant Engecorps.
ATM was being used to excavate the 18.5m diameter station tunnel as part of the metro's £735M Line 4, or Yellow Line. The dig was being carried out in gneiss overlain by tertiary sediments, consisting of stiff clays and compact sand.
An engineer on the project said that a bored tunnel was not considered because 'there is no equipment for such dimensions'.
'A tunnel of 18.5m diameter is very large, ' said Anderson. 'If you have an open face in dodgy ground and it starts giving way because it's wet or it has low shear strength there's not a lot you can do other than run away, ' he said.
Another tunnelling specialist, who asked not to be named, questioned the choice of NATM in soft or variable ground.
Open face tunnelling is cheap. NATM is much less expensive than a soft ground TBM, but much less safe, ' he said.
A combination of explosives and mechanical excavation techniques were being used.
'The top heading of the failed tunnel was fully excavated in rock. There was approximately 6m of rock above the tunnel, ' the consortium spokesman said.
Support consisted of steel arch girders at 800mm centres, and 350mm thick steel fibre reinforced sprayed concrete.
'The vertical displacements measured at the tunnel crown were in the range of 3mm when the top heading excavation was completed.
Two to three days before the accident, the rate of those displacements increased, and they reached 15mm to 20mm.
The decision was made to install additional support consisting of rock bolts.'
But the spokesman said the collapse happened before any anchors could be installed.
Anderson expressed surprise that NATM was being used to create such a largetunnel in potentially poor ground.
'Generally a single track metro tunnel is 7m across. The running tunnel [for the Yellow Line] is designed for twin tracks and is 9m across, which is big.
The sections adjoining the station shaft were even larger at 18.5m.' The 6m of rock cover over this large tunnel was 'thinnish', Anderson said.
The contractor also described the gneiss as heavily ssured and fractured. It blamed heavy rain for the collapse. The Pinheiros River is also only a few hundred metres from the tunnel.These fractures could have provided routes for water into the tunnel, and could have resulted in sudden changes in rock strength and integrity.
Tunnelling expert Sir Alan MuirWood said: 'You have to assume the ground and its behaviour.
Groundwater is a feature you don't want to meet, and if you have perched ground water it'll tend to cause a wash out [when punctured]. 'The of water in ground that is otherwise stable can cause problems.'
He also warned: 'If you have blocky ground you may have chunks displaced by ground water pressure, which may have built up over time.'
The water table across São Paulo state is typically only a couple of metres below ground.
Anderson added that creating interfaces between tunnels and shafts was notoriously difcult, whatever tunnelling method is used. He said it was particularly problematic using NATM.
Because NATM excavation involves nibbling away at the ground the tunnel profile is asymetrical during construction, with forces acting on it out of equilibrium.
Andrew Mylius NATM failures: the facts
There have been 67 NATM collapses since 1973, research published by the Health & Safety Executive last year revealed.
The risk to third parties from bored tunnelling in soft ground lists three NATM failures on São Paulo Metro alone:
December 1981. Tunnel collapses during excavation in soft ground, leading to local instability. Buildings have to be demolished.
November 1991. Tunnel in soft ground collapses and oods.
November 1993. A sink hole opens up following a collapse in soft ground.
The report also highlights that:
0% of NATM tunnel incidents occur close to the face, before the structure is complete.
Over half of collapses result in a surface crater.
The primary cause of failures is unpredicted ground conditions.
The report emphasises 'the necessity for a good prior understanding of the ground and for thorough contingency and emergency pre-planning.
'It is also possible for ground conditions to change rapidly over short tunnel advances and this can result in sudden unstable open tunnel faces.'
UK NATM expert and coauthor of the guidance, John Anderson, warned: 'You can get pockets of soft material, perched water, boulders - things that don't allow you any mistakes.
'You have to design your tunnelling system to maintain safety in all circumstances.
'The Heathrow Express tunnel was excavated in London Clay, which is very well understood and is generally a good, consistent material.
'If you can get NATM wrong there you can get it wrong anywhere, ' he said.
'You have to have a very exact knowledge of the kind of ground you are operating in.'
The tunnel collapsed in October 1994 but was only fully explained in February 1999.