One traditional view of water in tunnels was that it was merely a nuisance until it came over the top of the foreman’s wellingtons! Today we take a very different view, and it is recognised that even small quantities of water can cause surface settlement in certain situations, as has been observed in Hong Kong and Jersey, for example.
The issues are much broader than just sizing pumps to get rid of water in the tunnel, which was and still is the only issue in many rock tunnels. The impact of pumping on water resources (Hallandsas in Sweden is a typical example, see news) and potential surface settlements are now major considerations in rock tunnels, particularly in urban environments, as will be seen from the article on the Copenhagen mini metro elsewhere.
In the UK there is current interest in the London tunnels that will form part of the high speed Channel Tunnel Rail Link where the Engineer has developed a design for a regional dewatering scheme to permit tunnel construction in the dry below the London Clay. The material below the London Clay contains extensive areas of sandy materials, including the Thanet Sands which are otoriously unstable in the presence of water. These are all currently water charged with pressures up to 3 bars. Many readers will remember the flooding of the London water ring main tunnel in 1988 in just such conditions.
This article considers the precedents for regional dewatering schemes and other dewatering technologies available. As so often in life, nothing is quite as new as it seems; over 100 years ago Robert Stephenson used deep wells to dewater and assist in construction of the Kilsby railway tunnel north of London.
In the modern era one of the first novel uses of dewatering was to install wells from a jack up barge in the Fleet behind Chesil beach in Dorset, UK, to reduce water pressure in the Sandsfoot grits to enable tunnelling of the sewerage outfall in 1980. This seems also to be the first occasion when the dewatering allowed a reduction in the pressure in compressed air working. Subsequently the subsea dewatering technique was used on a grander scale for the Storebælt eastern railway tunnels in Denmark.
This offshore dewatering was possibly the largest and most spectacular dewatering scheme for tunnel construction so far. It became known as project Moses, standing for Method of Safety by Emptying the Storebælt! The statistics are impressive: a total of 41 offshore wells were drilled into the marl, which was protected from recharge from the sea by the glacial tills. A total of 45M.m³ of water was pumped at a cost of DKr200M ($30M) to reduce water pressures, particularly to aid interventions to maintain the TBM’s.
In recent times it is not widely appreciated that the dewatering for stations and shafts on the London Jubilee Line Extension had a major impact on the pressures to be dealt with by the tunnel boring operations. This benefit is difficult to quantify but recognised as significant by the designers of the London tunnels of the Channel Tunnel Rail Link.
London enjoys one major advantage in this type of work in that the Chalk aquifer and the overlying Thanet Sands have been historically dewatered for water supplies in London. These water levels are now recovering and risk of settlement associated with dewatering is much reduced.
While modern tunnelling machines can deal with high water pressures and unstable ground, interventions and cross passage construction remain a problem. Cross passages are a typical feature of modern tunnels and are usually excavated by hand, being relatively short and having complex connections to the main tunnels. Stabilising the ground so that construction can take place outside the safety of the tunnel lining presents challenges in unstable and pressurised ground. Typical solutions involve ground freezing and grouting/jet grouting where access is available from surface. Over the past eight years a number of such projects have been undertaken by dewatering using subhorizontal drains drilled from the tunnel. These case histories are described below and practical constraints outlined.
The stability of soils in vertical wells, except those being drilled into truly artesian conditions, is relatively easy to maintain by balancing the pore pressure in the ground with the column of water or drilling fluid in the drillhole. In a horizontal drain where the pressure is not maintained, the fluid will be at atmospheric pressure and a very high hydraulic gradient from the pore pressure in the soil can rapidly lead to instability of the drillhole wall. This problem is overcome by sealing the hole at the wellhead with a gland or blow out preventer which usually takes the form of a leather or rubber cup that is squeezed around the drillstring in a special housing. The sealing of the hole prevents the introduction of a conventional wellscreen or filter.
STOREBAELT RAILWAY TUNNELS, DENMARK, 1991 – 1995
The Storebælt Eastern Railway tunnels in Denmark form a major communications route joining the two parts of Denmark for the first time by a fixed link that includes two 8km long undersea tunnels. For safety reasons the two 8.5m od railway tunnels are connected, at 250m intervals, by cross passages. At Storebælt, 29 cross passages were required to be constructed under the sea, of which 13 are in glacial tills with permeable zones and pore pressures up to 4.5 bar (Doran et al 1995). A typical layout – from London’s Docklands Light Railway project.
The intention at tender was to stabilise the soils around the cross passages by freezing. However, a review suggested that the soils were rather stiff and that if the flow of water towards the excavation could be controlled, then the negative pore pressures that develop, due to relaxation immediately after excavation, could be maintained to give a stable excavation. The soils are too impermeable for conventional dewatering with a relatively high clay content and permeability in the range 10-5 to 10-7 m/sec. A drilling and installation system was devised that involved drilling through preventers and valves that allowed the injection of a polymer sand suspension around the wellscreen.
The first cross passage was treated with the use of 24 vacuum pumped wellpoints from each side of the cross passage. Pressure criteria were established for 12 piezometers within the excavation such that the head was less than 0.5m and the hydraulic gradient of less than 0.1. The cross passage was successfully excavated and the ground conditions were found to be stiff glacial till with no water inflow. Subsequent cross passages were treated in a similar manner except for four where extreme conditions necessitated the use of partial freezing solutions in combination with the dewatering.
WILMSLOW, JACKED BOX UNDER RAILWAY, UK, 1994
In 1994 a large jacked box for a four lane highway was to be pushed under an active railway. Face stability required that dewatering was required under the railway but, due to the low permeability, the cone of influence was insufficient to reach the centre of the railway. A scheme was devised using horizontal and raking ejector wells. In the event raking wells were installed at angles as shallow as 22° to the horizontal in cased holes. There is a general belief that filters can only be placed in vertical holes and the use of such highly raked filter wells in soils is further development of recent work.
DRUID STREET, JUBILEE LINE EXTENSION, LONDON, 1994 – 1996
As part of the Jubilee Line Extension various ventilation and escape tunnels were to be constructed under the brick arches of the multiple rail tracks near London Bridge station. The ground conditions were predominantly clays but thick layers of sand and silt were encountered at the horizons to be excavated. Due to limited surface access and the relatively thin layers of permeable material (where holes could be drilled) it was decided to draw on the Storebælt experience and install horizontal drains pumped under vacuum.
Drains were drilled from shafts and the running tunnels, using a blow out preventer system, into the sandy layer and with 0.5mm slotted wellscreens, a natural filter was developed. Again the learning curve for the drilling techniques was significant but the end result was successful. As on previous occasions drilling into pressures above one bar proved to be significantly more difficult than at lower pressures. It was concluded that conventional dewatering or short sacrificial drains to lower the pressure initially are of great value.
KELVIN VALLEY SEWER, GLASGOW, 1997
At the beginning of 1997 a 1.8m diameter drill and blast tunnel was being driven through Coal Measures sandstones and mudstones for a new sewer. After one particular blast the tunnel collapsed and started to fill with gravels, boulders and heavy water flows. The tunnel had encountered a buried glacial valley with water pressures above 3 bar. Immediate investigations confirmed the variable nature of the infill material.
A solution was developed whereby an open shield was used from the opposite direction using drill and blast techniques until the valley was approached; grouting of the infill via holes drilled from the tunnel was then undertaken. Some surface dewatering was used to reduce the pressures in the area but problems inevitably occurred with grout flowing to wells and causing them to fail (despite planned efforts to turn off those closest to the grouting operation).
Holes drilled in the centre and low down in the grout pattern were not grouted immediately. These holes usually had strong flows of clear water that sometimes stopped as the grouting in other holes took effect. When flows continued the holes were not grouted, despite some resistance from the mining crews and foremen, but were left to operate as horizontal drains. Due to the coarse nature of the ground these holes were stable with only a short starter pipe. The effect of the holes was to reduce the pressure and improve stability in the critical area ahead of the tunnel during the next blast round.
DLR TUNNEL CROSS PASSAGE AND PUMPING STATION, 1998
As part of the light rail extension (DLR) in London, being built by Nishimatsu Construction, a cross passage and pumping station had to be constructed between the two running tunnels under the River Thames. The excavation was to be wholly within the Thanet Sands, a silty fine sand, with water pressure up to 2.9 bars. The ground conditions were Thanet Sands with the Woolwich & Reading Beds (WRB), which are dominantly clay materials, forming a roof just above the tunnel crown. The Chalk is about five metres below the cross passage sump.
A review led to the detailed consideration of two options: either ground freezing or a combination of compressed air and dewatering. The use of freezing to recover the flooded London water ring main tunnel at Tooting Bec in the same strata and the Storebælt tunnel experience suggested this option could work. A design was developed which consisted of 45 freeze pipes from each tunnel. The two options were compared in terms of risk, cost and programme. Risk analysis and cost did not lead to an obvious choice between the options. The freezing option was not pursued due to complex programme issues, particularly access to the second tunnel and drilling time that together indicated a potential programme benefit from the dewatering/compressed air scheme.
The chosen scheme consisted of multiple dewatering schemes with low pressure compressed air. Dewatering was designed so that compressed air pressure less than one bar was needed, permitting eight hour shifts and limiting health risks to compressed air workers. In order to provide complete stability and safety in the fine sands a high level of redundancy within the dewatering elements was incorporated. Dewatering comprised three independent elements: an array of deep vertical wells drilled from the river and abstracting from the chalk, resulting in downward drainage from the Thanet Sands; a second, independent array, similar to the first, and an array of sub horizontal wellpoints installed from the northbound tunnel around the cross passage.
The in-tunnel scheme consisted of sub horizontal drains into the Thanet Sands. The in-tunnel drain element could run without the use of any external power, by draining from the positive pressure in the ground to free air in the tunnel. This would provide the residual dewatering effort required with the worst case scenario in the event of complete non functioning of the two river well systems. The problem of drilling into unstable high pressure sand and installing an efficient wellscreen and filter was recognised as extremely difficult.
On the basis of a trial the combined dewatering and compressed air scheme was implemented for cross passage construction. Wellpoint flows were measured and showed initial flows totalling 8 litres/s with all wellpoints running. Water pressures were monitored by pressure gauges in the area of the excavation and a substantial pressure reduction was apparent (leaving a pore pressure at axis of between 1.0 and 1.2 bar before the river wells were pumping), and a smaller tidal range than before drainage.
The cross passage was successfully excavated in July 1998 using less than one bar of compressed air pressure and the railway system opened two months ahead of schedule in November 1999.
Related Files
Cross passage dewatering installations on the Docklands Light Railway in London
Soil and filter gradings for Storebaelt drains
