Ground movement is probably what most geotechnical engineers would think of first when it comes to the risk related to tunnelling but there are other more insidious issues that should also be considered. Deoxygenated air can present a potential danger during construction and for maintenance during the operational phase. Research is currently underway at Imperial College to reveal the cause of the problem.
The research centres on deoxygenation in excavations within the Upnor Formation of the Lambeth Group, which has been linked to a number of construction fatalities.
“During construction of boreholes through the Upnor Formation on the route in 2008 we had a gas blow out with high levels of pressurised deoxygenated air venting from the borehole”
Tim Newman, Thames Water
The best known of these involved the deaths of an engineer and assistant on the London Water Ring Main in 1988 from asphyxiation after entering a confined space within the finished tunnel which was located within the Upnor formation. Investigations into this and other cases have linked the deaths to glauconite but the current research suggests that the cause is something different from common perception.
“The inquest into the two fatalities on Thames Water’s London Water Ring Main concluded that the cause was the presence of glauconite, and the construction industry has never questioned this,” says Thames Water Tideway Tunnel senior engineering geologist and Imperial College PhD research student Tim Newman.
However, Newman says that glauconite is a common mineral and also fairly inert. So how could its presence have led to such a catastrophic outcome and why didn’t the problem occur more frequently given that glauconite is not rare? With 3km and seven shafts on Thames Water’s Tideway supersewer’s route expected to be constructed through the Upnor Formation, and given the company’s previous experience of the problem, Thames was keen to fully understand the issue. As a result, the company is funding Newman’s research which he is combining with his work on the design side of the Tideway scheme.
“Even if a lining is watertight, it could still allow the nitrogen to pass into the tunnel”
Tim Newman, Thames Water
Newman started his research in November 2010 and hopes to complete his report by the end of this year.
Newman’s first experience of the problem came during his work on the ground investigations for the Tideway project five years ago.
“During construction of boreholes through the Upnor Formation on the route in 2008 we had a gas blow out with high levels of pressurised deoxygenated air venting from the borehole,” he says. “The Health & Safety Executive’s senior engineer John Greenwood asked Thames Water to look into the cause, which ultimately led to me undertaking a PhD on the topic.”
The research is being overseen by Imperial College engineering geology lecturer Richard Ghail and GCG senior engineering geologist Jackie Skipper, who has also long believed that glauconite was not the cause of deoxygenation.
“We still don’t understand why the deoxygenation occurs in some places and not in others”
Tim Newman, Thames Water
During a desk study of the problem, it became clear that the issues on the London Water Ring Main and the Tideway borehole were not isolated incidents but the problem was not widely reported, making it difficult for links to be made until now.
Normal dry air consists of 78% nitrogen, 20.9% oxygen and 0.9% argon, by volume, with other gases, principally carbon dioxide, making up the remaining 0.1%. At one site where deoxygenation in the Upnor Formation was identified, oxygen levels were as low as 11%.
The initial focus of Newman’s research was on finding the cause. “We looked at pyrite and organic matter because both have been linked to the production of acid on oxidation. But we found no evidence to support this,” says Newman.
The next cause that was considered was green rust which is an oxidation of iron that creates a mixed Fe(II) and Fe(III) layered double hydroxide.
However, the oxidation happens quickly, usually within a matter of seconds, so the compound is rarely seen. “The oxidation leaves behind no trace element other than goethite, but that is quite common,” says Newman.
According to Newman, the oxidation in these deposits would have occurred when they were brought into contact with air during the regional-scale dewatering and drawdown of the Lower Aquifer during the industrial growth of London. He believes that the postindustrial recharge and resaturation resulted in accumulations of often compressed deoxygenated air, trapped beneath overlying impermeable clay strata.
But the effects can also be induced in a short period of time and Newman points to the deoxygenation effects that were seen in a shaft constructed at Honour Oak by Thames Water.
“There, the site was only dewatered for the project and the deoxygenation effect was seen within a month of the groundwater lowering,” he says.
Although the concept is plausible, Newman has yet to be able to physically observe green rust in the field or laboratory. “Until we find a technique to sample material before the green rust forms, it is difficult to conclusively prove the cause. Capturing is a real issue,” he says.
Techniques trialled so far include encapsulating a sample in wax, cling film and vacuum bags. But none has yet preserved the samples well enough. “Nitrogen bags could be used but that would have health and safety implications on site,” says Newburn.
Analysing the samples is also a challenge and Newman has constructed a special deoxygenated glove box to reduce the oxygen levels to 2% to handle samples in the search for the elusive green rust. “I think we need to find a way of lowering the oxygen levels further still,” says Newman.
While the material itself has yet to be seen first-hand, its effects still can be and this is where Newman’s research will help to safeguard future construction work in the Upnor Formation.
This issue is important - site investigation contractors need to be aware of the potential for the problem to cause blow outs while drilling but it also has implication for the construction phase itself.
“There is a need to consider permanent blast ventilation during the construction period to avoid deoxygenation happening when fans are switched off at the end of a shift,” he says.
“Thames Water’s primary concern is not the main shafts so much, but the connection shafts, which can be smaller than 2.5m internal diameter and hand driven before they are spray concrete lined.”
Newman admits that there are a lot of unanswered questions and there is potential for more research into the issue.
“For example, we still don’t understand why the deoxygenation occurs in some places and not in others,” he says. “Is this due to faulting or maybe local variations in composition?”
To illustrate the problem, Newman points to the Tideway borehole where the initial problem was encountered. “Three years on and the gas is still venting from it at a rate of 20l/h but others nearby have no gas present,” he says. “Some boreholes we have investigated that were drilled in the early 1990s still have depleted oxygen levels.”
The longevity of the issue gives rise to issues for the operational use of tunnels in the Upnor Formation.
“After construction, even if a lining is watertight, it could still allow the nitrogen to pass into the tunnel,” says Newman. “This presents a problem for maintenance teams accessing sections outside of normally ventilated areas like station platforms.”
According to Newman, Crossrail has been keeping an eye on his research and given him use of information from its ground investigations at Farringdon and in the Connaught Tunnel which passes through the Upnor Formation.
While the research has yet to prove the presence of green rust, it is unlikely that the infrastructure developers will rely on the presence of glauconite to determine the risk posed by deoxygenated air during and after construction.
Newman hopes to be able to perfect the sampling method to conclusively prove the green rust risk but, until then, he has helped to develop a better understanding of the conditions under which green rust, and deoxygenation, might occur.