Ground freezing and tunnel jacking have been used to protect a key railway route during construction of an environmentally sympathetic flood alleviation scheme in Berkshire. Max Soudain reports.
Maidenhead, Windsor and Eton in Berkshire are prone to flooding. The Environment Agency predicts that if an event on the scale of the last major Thames flood in 1947 occurred today - and on average such events happen twice a century - 5,500 homes would suffer, as well as businesses, transport routes and utilities, at a cost of £40M.
Lesser floods tend to occur every five to seven years. In 1990 500 homes were affected, at an estimated cost of £0.4M.
The £6.7M Dorney Bridge project on the outskirts of Maidenhead is part of the EA's flood alleviation scheme which will divert water from the Thames at Maidenhead along an 11.6km artificial but natural looking channel which will rejoin the river just east of Eton. Work on the overall project began in 1996 and is due to finish in winter 2001.
At Dorney Bridge, where site work began in September 1997, a huge box culvert had to be jacked through an embankment under the Great Western Railway which runs from London to south west England.
The project was designed by consultant Scott Wilson Rail for main contractor Edmund Nuttall, which decided to use ground freezing to stabilise the embankment while it jacked the 50m long, 23m wide and 9.5m high reinforced concrete box culvert through.
Two mainline railway tracks and two local lines run along the 10m high embankment, which was built in two parts. The southern half was built by Brunel in 1840 and consists of a mixture of clay and gravel, while the northern half, built when the embankment was widened in 1893, comprises clay fill.
This is underlain by up to 1.5m of topsoil and fill overlying Thames Terrace Gravels up to 7m thick and varying from fine sand up to boulder size and containing lenses and bands of sand and layers of very coarse material. It is highly permeable and underlain by weak fissured chalk. The water table is very high, a minimum of 1.5m below ground level.
The culvert was jacked from a casting pit built on the southern side of the embankment. This 50m wide and 80m long excavation goes 6m into the chalk and was formed using a combination of slurry and diaphragm walling installed by Taylor Woodrow Foundation Engineering and steel sheet piles. To the north a 30m wide and 10m long steel sheet pile cofferdam was formed. Both excavations were dewatered using a total of 14 pumps installed by WJ Groundwater, pumping out some 4,000m3 a day.
'Because a significant part of the culvert is below the water table we had to deal with the groundwater and anticipated being able to grout the gravels,' explains Nuttall project manager Steve Brackenbury.
Initially grouting was also to be used to stabilise the embankment so that there was minimal disturbance of the railway lines during jacking. Unfortunately, trials showed the method could not be used.
Instead, Nuttall decided to use ground freezing to stabilise the embankment and to control groundwater. While the firm had not used the method in the UK before, its sister company HBG was using the method on the Boston Central Artery project (GE July). 'We called on that expertise,' says Brackenbury.
The ground freezing scheme at Dorney was designed by Alan Auld Associates.
Some 200, 90mm diameter horizontal freeze pipes, up to 60m long, were installed in eight rows through the embankment, covering an area 29m wide and 13.5m high. This produced a 2m thick frozen zone above and below the box and 3m wide zones either side. Most of the boreholes were drilled from the north side (to allow construction of the culvert) by Thrustbore Contracting, using up to three Vermeer directional drilling rigs with bentonite flush.
Additional drilling was carried out by contractor Dosco using a down the hole hammer on part of the lowest level of pipes where ground had already been tightened up by earlier grouting.
Steel freeze pipes were used in the perimeter zone around the box but plastic pipes were used inside the box area to ease excavation. Each hole was drilled through the embankment before pulling the MDPE freeze pipe back through, so that there was 'as little interference as possible to the ground', says Brackenbury.
Drilling was not easy because of the highly variable nature of the ground and the large boulders present but only a few holes had to be redrilled or pipes replaced. All the pipes were pressure tested to ensure that there were no leaks.
Freezing, carried out by British Drilling & Freezing, took four and a half months 'to ensure that the ground was totally frozen,' says Brackenbury. Calcium chloride brine at -35degreesC was pumped from four freezing plants to create a homogeneous mass with a strength of between 25N/mm2 and 30N/mm2.
Brackenbury says one of the main concerns was that there would be gaps in the frozen ground, as some of the boreholes had to be curved around obstacles in the ground during drilling. Any free water would have melted some of the ground and caused problems during excavation, so pipes were extensively surveyed and plotted up to locate any potential problem areas. 'A few pipes were found to be unsatisfactory, so new holes were drilled or freezing time increased,' he says.
The integrity of the freezing was checked using thermocouples put down 'dummy' boreholes and by probe holes through the steel sheet piles once ice started forming. While the freezing was satisfactory, six vertical freeze pipes up to 15m long were installed 'to reinforce the freeze in a couple of areas'.
Keeping movement in the embankment and the railway lines to an absolute minimum was vital. There are onerous financial penalties if train services are affected, Brackenbury explains. Monitoring was carried out using electrolevels supplied by CMCS, with track movement monitored using EDM and precise levelling. Track tolerance was set to +/-7mm but embankment heave through freezing was expected to be 130mm on the north side and 90mm on the south side, so throughout the freezing, a team from Amey Rail was on site to keep track movement within acceptable limits by adding or removing ballast. The crew also worked to combat settlement during jacking and will remain on site during thawing operations, due to begin this month.
Actual heave over the freezing period was 150mm on the north side, but virtually zero on the south side. Brackenbury says this was because this section of the embankment was built of clay with trapped water, whereas Brunel used a mixture of gravels and clays. The main challenge was to keep tilt of the tracks to a minimum because of the differential movement.
The six week jacking operation started in July this year and was carried out using a method developed by subcontractor John Ropkins. First, a 1.3m thick reinforced concrete jacking slab was cast in the base of the casting pit. Six trenches run along its length and 36 hydraulic jacks were placed in these, in sets of six. At the back of the jacking slab, a thinner slab was cast to act as the reaction for the push. While this slab was only 300mm thick, excavated material from the tunnel was placed on it during the push to provide more reaction.
Because the jacks have a maximum stroke of 1.2m, they had to be moved up the trenches as the culvert edged forward. To enable this, notches were cast at intervals across the trenches and steel packers placed in these slots as jacking points.
The key feature of John Ropkins' method is the anti-drag system, which uses heavily greased steel ropes to reduce friction between the roof and floor of the box and the surrounding ground. This was particularly important on this contract as lateral movement of the railway lines had to be kept to a minimum.
'Casting ropes' were placed on the jacking slab about 150mm apart before the culvert was built and steel plates were cast in to the underside of the box to reduce friction further between the slab and the box.
Nearly 900, 13mm diameter and 55m long steel ropes were laid inside the culvert before jacking. These were anchored to the jacking slab through slots in the floor of the box, with their free ends inside the box itself. Another 760 of these ropes were also used for the roof, suspended from the ceiling, anchored to a steel reaction beam on top of the box at the front and kept tensioned by jacks at the rear of the casting pit. As the box moved forward, these ADS ropes were payed out, resulting in minimum ground movement above and below the culvert.
Excavation of the frozen ground was carried out using two roadheaders, a Komatsu fitted with a specially designed cutter from Webster Schaeff, and a smaller machine that could not reach the full height of the box. Up to 1m was excavated before each jacking. The high strength of the ground meant that overdigging around the base and the sides of the box was possible to ease the jacking. There was some concern that the box could be frozen into the ground if there were any hold-ups in excavation, so heating elements were fitted around the base of the box. In the event they were used throughout jacking to melt a small amount of the surrounding ground around to help the push.
Jacking finished in the middle of August and final preparations are being made before the ground is thawed. Some grouting is being carried out around the box to fill gaps left by the over- digging, to minimise settlement. Brackenbury says that by the time work is complete at the end of the year, about 75mm of embankment settlement is expected.