Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

Out of sight

RAIL ALL THE WAY - Tunnelling risk posed perhaps the greatest threat to the smooth delivery of the high speed line. Andrew Mylius finds out how it was controlled.

On their way from Dagenham to St Pancras the highspeed line's twin-bored tunnels pass beneath 2,600 properties, numerous industrial facilities, 67 bridges, 50 masonry retaining walls, 12km of surface railway, 12 existing tunnels (including four Tube stations with lifts and escalators) and 600 gas, water and sewer pipelines. With one dramatic exception, the public did not notice a thing.

CTRL's tunnelling incident occurred at Stratford, east London, when an uncharted void - perhaps an old well or unconsolidated bomb crater - caved in, opening a vast hole in gardens behind a terrace of houses.

The tunnel itself was not affected by the collapse - it happened to one side of the tunnel, and the boring machine (TBM) had advanced roughly 30m beyond the site when it happened.

Tunnelling stopped while the ground was stabilised and the Health & Safety Executive checked the project's risk assessment and working methodology, but as former RLE settlement monitoring and engineering manager Keith Bowers points out: 'There was no prosecution - no fault was found.' However, the collapse did cost the project three months against programme.

Bowers says that the project's ability to virtually recover time lost, emerging at St Pancras on schedule, was a testament to the way the tunnelling contracts were set up, equipped and run.

The London tunnels were executed under three separate contracts. A joint venture of Nishimatsu/Cementation/ Skanska bored the 7.5km from Stratford to St Pancras, while a Costain/Skanska/Bachy joint venture worked in the opposite direction to deliver the 4.7km from Stratford to the Barrington Road vent shaft, and a Nuttall/Wayss & Freytag/Kier joint venture took the tunnel the 5.3km from Barrington Road to Dagenham.

Unusually for projects of huge individual value, a common approach to tunnelling was adopted from the project's outset. All of the contractors were tasked to meet the same high minimum standards, set by RLE. And all were run from the same office, blurring boundaries between them and encouraging exceptional openness about technical and commercial issues.

Tunnelling presented a major risk to the project. With a combined length of 17.5km, the high-speed line's London tunnels were three times longer than the Jubilee Line Extension, an epic undertaking that was delivered in 1999 late and over budget. The CTRL tunnels were 60% greater in diameter.

'Ground conditions along the alignment were relatively untested because there aren't many tunnels that far east, ' says Eddie Woods, project engineer and tunnel engineering manager at the time of construction.

However, site investigation showed that the tunnels would be 'through some of the worst potential ground conditions in London'.

Over the length of the London tunnels, self-supporting stiff clays at either end gave way to Lambeth Group sands and gravels, Thanet Sand, Thames river terrace gravels, Upper Chalk, alluvium and peat. Some 60% of the tunnels were driven through Upnor and Thanet Sands.

And over much of the route the water table stood close to ground level, meaning tunnelling would have to be done at water pressures of up to 3.3 bar.

'We were very focused on the outturn cost - we couldn't afford any uncertainty - and specied the TBMs accordingly, ' says former senior eld engineer Gordon Battye.

To cope with the high water pressure and soft ground conditions earth pressure balance machines were needed.

RLE wrote the specication for the TBMs and demanded guarantees of mechanical reliability and performance. 'For example, we specied that the machines' main bearings should have a working life of 10,000 hours; a safety factor of three, ' says Battye.

'We had them designed so you could replace the bearings from the back - normally you have to do that from the front.

We specied the cutter head, the gear boxes, the pumps and thrust rams. We insisted on long-haul conveyors for spoil removal instead of muck away wagons to ensure there would be no constraints on output. We had bentonite injection around the shield annulus and continuous grouting through the tail skin.

'Most important of all, we got the manufacturers to sign up to the requirement that the machines would be capable of achieving 1% face loss or less. The ability to cause no more than 1% greatly reduced settlement risk.' Woods continues: 'This allowed us to use the TBMs as a prime risk mitigation tool. We equipped ourselves with machines that gave us an unprecedented degree of control.' Sensors on the TBM face provided information about earth pressure and wear to the picks. Grouting around the TBM tail skin enabled immediate compensation for overcut, lling the annulus between tunnel lining ring and excavated diameter.

Weighers on the belt conveyors used to remove spoil monitored actual weight of muck away, enabling comparison with anticipated weights. Laser scanners cross-checked the excavated volumes.

Bob Hodgetts, tunnel construction manager, says: 'Monitoring all this information was done in real time, so you could change earth pressure and mitigate any ground movement to prevent settlement.' Data was relayed simultaneously to the TBM operator and a control room on the surface, enabling it to be double-checked for telltale trends or anomalies.

Weeks into construction, Nuttall/Wayss & Freytag/Kier adopted the grouting system being used by Nishimatsu/ Cementation/Skanska. 'They found that tunnel lining rings were floating on the ground water, ' says Hodgetts.

'They were using cement/ sand grout that wasn't curing fast enough. Nishimatsu started off using a thixotropic grout, which sets almost immediately.

It gels in ve seconds and the initial set comes on within 40 seconds.' This fixed lining rings solid before they could float.

In London Clay and Thanet Sand face loss was as low as 0.25%. Face loss for the whole of the London tunnels averaged an amazing 0.5%.

The BMs ere designed to withstand maximum water pressures of 3 bar. Dewatering was used extensively along the tunnel route to make tunnelling easier and it proved especially effective when it came to excavating the 29, 3.5m diameter, cast iron-lined cross passages. 'We found that Thanet Sand, when it's dry, stands up very well, ' says Woods.

Dewatering carried a penalty, however. 'We had to inject water back onto the face at a rate of 30 to 40m3/s to maintain the earth pressure balance, ' says Hodgetts. 'If you put water onto the exposed Thanet Sand face, you could see it being sucked in by the effect of the dewatering pumps.' A careful hydrological balance had to be struck between dewatering for the tunnelling work and maintaining the water table in the wider area. 'To avoid drawing down ground water and affecting other structures we had to recharge at given locations, as well as take water out, ' says Bowers.

Thanet Sand proved to be 'very hard, like a battering ram', says Battye. The pair of Wirth TBMs used by Costain/ Skanska/Bachy were routinely run at 110% of their stated maximum torque, compared to 70% to 80% maximum torque on a normal tunnelling project.

'It created huge heat - sand vitried on the cutter.

'To carry out maintenance you had to fill the head with water, and even then you couldn't get anyone in for three to four hours because of the steam, ' Battye adds.

Bowers says that flints in sandstone encountered by the two Kawasaki TBMs used by Nishimatsu/Cementation/ Skanska meted out massive wear to its cutter head and screw conveyor. 'When we pulled them out of the ground you could shove your arm between the flights of the screw and the conveyor casing.' Cutter heads were refurbished at ventilation shafts - there are ve - as the machines passed through.

Wear and tear notwithstanding, Bowers says that the high specication of the TBMs ensured that they delivered amazing progress of up to 57m a day. And they remained precision instruments right through the project. A handful of particularly ticklish hurdles on the London tunnels tested and justified the team's confidence.

A mere 35m into its drive east from Stratford, Costain/ Skanska/Bachy had to pass under London Underground's (LU) twin cast iron Central Line tunnels. 'The cover between our tunnels and the closest of the Central Line tunnels was 4.3m, ' says Woods.

LU initially talked about suspending services and laying on replacement buses, at vast expense to the project, while the TBMs drove underneath.

After intensive negotiation between RLE, the contractor and LU, the Tube operator was convinced that it should continue to run trains, but with a speed restriction.

The tunnels were fitted with movement instrumentation, monitored by both the project team and LU, and linked to the TBM. To accommodate the small amount of movement that would inevitably occur, every fifth ring of the Central Line's cast iron lining was loosened, enabling the normally rigid structure to flex a little.

In the event, undercrossings of the Central Line were achieved with volume loss of only 0.3% - far short of the maximum 2% that RLE's design had allowed for.

Heading west from Stratford, at Highbury the tunnels pass under four more Tube tunnels and two stations, 'all of different construction, all behaving very differently, all carrying different risks, and all with only 6m to 10m of cover from our tunnel', says design and tunnel settlement engineer Frank Mimnagh.

To give LU and Network Rail condence that their infrastructure was safe, surface and deep-level monitoring was installed over the 400m ahead of Highbury, 'showing we could go across without harm', Mimnagh explains.

'By doing that we increased the length over which we had to pull out all the stops [to minimise settlement] from 100m or so to half a kilometre, but we proved we could carry out the work without disruption to the railway.' For Bowers, the proof that the TBM specication was right came near the end of the Stratford-St Pancras drive, where it scraped under the Caledonian Road. The tunnels cleared a sewer support beam by just 800mm, a basement by 3m, large diameter water mains by 4m and rail bridge foundations by 4.8m. 'It was incredibly tight, ' he says.

'By the time we got to the Caledonian Road we'd flogged the Kawasaki TBM just about to death. We had anticipated it would be used in open mode over about half of the drive, but because ground conditions were more varied than expected, and because we wanted to keep settlement as low as absolutely possible we drove it all the way in closed mode. The machine took a lot more punishment than it was designed for.

'Even so, we still had so much control over the machine that we went under the Caledonian Road with a maximum movement of 6mm. The public didn't even register that.'

Tunnel lining One of the tunnelling contracts' guiding principles was 'no remedial work'. Finding the right tunnel lining was key to this aim, and in designing it, RLE had to consider performance during construction, in fire and under long-term maintenance.

Lining segments are subject to huge horizontal loading during construction as the TBM thrusts against the most recently installed ring to propel itself forward. About 10% of segments constructed with rebar suffer damage and require repair, says Battye.

As well as costing time and money to repair, there is a longterm danger that patches will, over time, come loose and fall out.

'It's not good if that happens in front of a high-speed train, ' Battye says.

And international experience of tunnel maintenance has shown that lining segments, including rebar, were prone to corrosion and spalling, requiring rebar to be replaced and concrete patched.

Fire in the tunnel could be structurally catastrophic with the wrong lining, Battye adds.

Segment trials on steel and polypropylene bre reinforced concrete (FRC) revealed that steel FRC segments tended to spall and split when heavily loaded. Polypropylene FRC, on the other hand was able to resist crushing up to three times the anticipated maximum load, and still had sufcient residual strength after a two-hour worst-case scenario re to work in the worst ground conditions.

Polypropylene FRC lining segments were manufactured for the London tunnels at two casting factories - one at Stratford for the tunnels running east and west of the station box, and the other at Ripple Lane for the Dagenham to Barrington Road section.

Hodgetts says that less than 1% of segments needed repair after installation.

Leakage was also minor - 1,000 of 226,666 segments installed dripped - just 0.44% of the total. High-spec gaskets that swell on contact with water were used in areas where pressure is especially high. Leaks have been dealt with by injecting hydrophilic high-strength polyurethane foam into hairline cracks and leaking joints.

Alliance The three London tunnelling contracts were characterised by a strange blend of camaraderie and competition. Working out of the same ofce, the London tunnelling contracts worked fairly intimately from the start of construction. The creation of a formal tunnelling alliance cemented the relationship.

All of the contractors, plus RLE and Union Railways, put their contingency funds into a single pot, providing a large insurance fund to deal with risk.

All mitigation works design and assessment of all structures was done jointly by RLE and the contractors. RLE took responsibility for design, and the contractors for any design changes.

'We were looking for ways to minimise change and for ways to economise, ' says Woods.

Savings were made by taking rebar out of the walkways used to contain a train derailment. 'We needed shear links to keep the walls in place, but it didn't need rebar to stop cracking. That was a purely cosmetic consideration - it wasn't an issue in tunnel.' Construction of the tunnel walkway and track bed system was carried out using a slip form paver, more commonly used to install concrete motorway crash barriers, which reduced construction time.

As well as sharing segments the alliance shared the same belt conveyor system, meaning that there was never a problem obtaining spare parts or additional conveyor units.

'The alliance removed duplication and prevented any conict between contractors for resources or materials, ' says Woods.

'If a section of tunnel was getting slow we looked across the construction team to take resource off other activities and put them onto it.

We also got specialist teams working across all three contracts.'

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.