HomeInfrastructureCrossrail Thames Tunnels

Crossrail Thames Tunnels

Listen to this article

When Paul Baker of Hochtief Murphy Joint Venture (HMJV) was kind enough to invite me to visit the Plumstead portal of the Crossrail Thames Tunnels I was delighted.

I spent most of my career maintaining Victorian rail tunnels, and have visited the interior of some more modern tunnels such as the Channel Tunnel, but I had never been close up to a tunnel boring machine (TBM) in action. This was not to be missed. My colleagues Colin Carr and Graeme Bickerdike felt the same when I let them know that Paul could take several of us along.

The joint venture of Hochtief AG of Germany and Murphy Group Ltd. had previously delivered the HS1 Thames Tunnel further downstream and from this success they were awarded this contract for Crossrail.

View from the sewer

Paul introduced us to the site by making us visit a Victorian structure, an old Metropolitan sewer! There was logic to this though. At Plumstead, the sewer runs above ground under an earth bank that gave a good vantage point to allow us to look each way along the alignment of the Crossrail tunnels. First we looked west towards the cranes that indicated the site of the construction of Woolwich Crossrail station, then the eastward view showed us the spoil disposal site where slurry is delivered from the TBM cutter faces by pipeline, dried and loaded away for disposal. The Plumstead Network Rail DC traction sub- station “Cathedral” building of 1920’s vintage was clearly evident in the latter view, and this became of interest again later in the visit.

As we approached we saw that adjacent to the site offices were stacked piles of tunnel ring segments ready to be taken down to the TBMs. Each pile contained the eight segments required to form a complete, 1.6 metre long, 6.2 metre diameter ring. The segments are being fabricated in Ireland by Shay Murtagh and brought to Plumstead by road and ferry. Each set for a ring comprises a full lorry load.

Crossrail Contract C310, for the construction of the Thames tunnels, was awarded to the Hochtief Murphy Joint Venture. Paul’s involvement was in the railway interface liaison between the client, HMJV and the railway authorities.

This meant managing the interfaces with Network Rail and Docklands Light Railway (DLR). At Plumstead the North Kent Line, carrying an intensive service from London Bridge to Woolwich Arsenal, Plumstead, Abbey Wood and beyond, runs immediately adjacent to the portal box over some 600 metres. It also involved liaison with the DLR since the Crossrail alignment passes directly over its Thames tunnels as they approach Woolwich Arsenal.

The two TBMs that have started their journeys from Plumstead are named Sophia and Mary. Sophia began her first drive to Woolwich, about 1,200 metres away, in January 2013 and this phase of the work clearly shows the joint venture in its true sense as Hochtief AG’s Stephen Assenmacher is tunnel construction manager while Murphy’s Chris Ashton supports him as assistant.

Close up of TBM Mary entering the portal [online]

Sophia had completed this a few days before our visit and attention had shifted to Mary. This second TBM had been erected and tested in the portal box and had begun her drive towards Woolwich on 19 May, although her first complete tunnel ring was not installed until the following day as we were able to see on the progress whiteboard in the site office.

Both Sophia and Mary are slurry machines, meaning that their 7.1 metre diameter cutting shields are filled with a pressurised slurry of bentonite. This prevents the ingress of water from the surrounding ground and is mixed with the cut material at the TBM face to enable it to be pumped back to the slurry treatment plant at the Plumstead site. At the treatment plant the slurry is separated with the spoil extracted being dried and then taken away for disposal.

The bentonite from the slurry is cleaned and refreshed before being pumped back to the TBM face to repeat the cycle. At the early stage of the drive that we saw, the material being produced was gravelly as it was driving through gravel beds, but as the TBM gets down to the levels required for the main drive it will enter chalk and the spoil will become brilliant white.

Heavy machinery

Like the other TBMs being used on the Crossrail project, Sophia and Mary were supplied by the German firm Herrenknecht. This company has grown since its inception in 1975 and is now the world’s leading supplier of tunnelling machinery. Whilst the Crossrail TBMs are not as big as the largest yet, which was over 19 metres in diameter and drove a tunnel in St Petersburg, Russia, they are massive. Each is over 100 metres long and weighs about 1,000 tonnes.

They are operated by 20 people, 12 on board the TBM and eight in the tunnel behind it. Typically a rate of 100 metres per week is being achieved on the Crossrail tunnels, with a total of eight machines being needed to create the 42km of rail tunnels on the whole project. TBMs Phyllis and Ada, working on the western tunnels, have peaked at speeds of around 215 metres per week.

The Thames tunnels are each 2.6 km in length, with the opportunity to pause each drive at Woolwich to refurbish the TBMs in the station box after the drive from Plumstead before each machine starts to tackle the second section beneath the river to North Woolwich. Sophia will start this phase of her drive once Mary is well on her way from Plumstead.

Today’s TBMs are an amazing contrast with the techniques used by Brunel for the first Thames tunnel, not far away from the current site. They are full of sophisticated equipment and monitoring systems, all intended to maximise safety, minimise disturbance to the surroundings and ensure accuracy in the alignment of the final structure.

Sophisticated steering

In a typical day, Sophia and Mary are each likely to progress by fifteen 1.6-metre long tunnel rings. The accuracy of their progress is checked by means of an inertial navigation system (Gyromat) and an automated total station, on the tunnel wall, checking prisms at monitoring points within the TBM and on the tunnel walls. There is a “window” within the TBM, in the upper left hand quadrant of its circular section, which is kept clear at all times throughout the machine’s entire length. This allows the total station to “see” right down the length of the TBM to read a prism at the very front of it.

We were shown, on a computer screen in the site offices, the TBM driver’s view of the outputs from this (pictured right). In a circle on the screen was a symbol shaped a bit like a four finned paper dart. The driver can interpret this to see where the front and rear of the TBM lie in relation to the planned exact centre position of the tunnel.

The aim is to remain within 50mm of the target in both the vertical and horizontal dimensions, although the drivers apparently compete to better this by the greatest margin on each shift. The positional data updates every 11 seconds, and the system can be used to give predicted positions based upon the current position and direction of travel. How I wish that the canal engineers of old had had such systems when our canals were built – it would make steering my narrowboat so much easier in tunnels like Blisworth and Braunston!

Hydrostatic control

One of the most interesting engineering aspects of these particular works is the monitoring system in use. It is usual these days to cover such a site with prisms and automated total stations to establish a network of survey control points. These are monitored frequently and regularly. There will be set thresholds for movements, with alarms automatically triggered should these be exceeded.

The HMJV have elected to use, as an alternative in many areas, a different system. The principles are the same as regards thresholds and alarms, but the measurement technique is totally new and far more accurate. The system is called “HLC”, Hydrostatic Level Control. At each point to be monitored, on a sensitive structure, say, or on the TBM, hydrostatic control cells are fastened at chosen points. These are small boxes containing the hydrostatic sensors which are connected back to the central monitoring system. In certain cases, tiltmeters are also applied. The accuracy of this set-up is claimed to be 10-times better than a system based upon total stations and prisms, being in the hundredths of a millimetre.

The system at Plumstead is set up to take measurements every second, and to record measurements at 15 minute intervals. The difference in periodicity is odd at first sight, but makes sense when one considers the volume of recorded data involved even at the chosen rate. There are a huge number of points being monitored, and so it is sensible to limit the recorded information, which is after all likely only to be required as an audit trail. The greater frequency of real-time monitoring does permit very early detection should there be an alarming movement somewhere, but it is not really necessary to record every measurement if they are all within the agreed acceptable limits.

Monitoring other structures

One structure which is being monitored by HLC is the Network Rail DC traction sub-station, a tall cathedral-like structure built of brick on a concrete raft that sits above the line of the tunnels close to the portal.

This is clearly a sensitive structure and significant movement would not be acceptable to both the structure and equipment therein. Indeed, it was agreed that in order to reduce any settlement as the tunnelling proceeded, it would be lifted 3mm in advance by grouting beneath its foundations.

The effects of this grouting were clearly visible on the HLC records we were shown by Dieter Scherenberg, HMJV’s monitoring engineer. He and his colleagues had set up HLC on the building prior to compensation grouting and have been monitoring it throughout the subsequent passage of the TBMs beneath it.

Ground level view towards portals, showing proximity of passing SE Trains traffic [online]

They expect it to finish up back at its original level after tunnelling is complete, but if it doesn’t, further compensation grouting will be used to re-level it as necessary. That is not expected to be necessary, however.

Similar measures have been taken for other sensitive structures and buildings along the route. One that is being watched closely is a bridge carrying the North Kent Line over White Hart Road. The structure dates back to the opening of the line, with a later span addition for the sidings, and with little information on the actual foundations was of particular concern.

The DLR Thames tunnels, built as recently as 2009, are also being monitored. The Crossrail tunnels have to pass over the DLR ones with not much more than two metres clearance between them. This restricted dimension is dictated by the need for the Crossrail tunnels to pass under some building foundations a short distance away whilst rising to enter into the Woolwich station box, limiting the vertical room for movement.

DLR required that the project monitored its tunnels and the track within for a whole year before the new tunnelling works approached close enough to influence them. Carried out by specialist surveying teams of TES2000, this included both track geometry measurements using a track trolley and, leading up to and during the period of TBM overpass, laser profiling of the tunnels to check for any structural movement. This monitoring is required to continue for at least six months after the last tunnelling work in the area of influence has passed clear of the DLR tunnels. In practice the area of influence is being taken as a zone 100 metres each side of the DLR tunnels.

Unexpected movement

During the monitoring beforehand, DLR was apparently quite concerned by the movements being reported by the HLC system as it had not previously been aware that its tunnels were moving. However, discussions with Hochtief demonstrated that the movements were quite normal for structures like these, it was just that previous survey methods were not sensitive enough to show them.

In fact the sensitivity of the HLC system is such that the small movements caused by the tidal water level changes in the Thames are readily apparent in the records for the tunnels, showing as a waveform in the level plots at the same frequency as the tides. The effects of sunlight-driven temperature changes are also observable in structures exposed to the sun, and looking back at such records one can readily distinguish sunny days from cloudy ones by noting the differences in the recorded movements.

In the event, when Sophia passed over the DLR tunnels no untoward movements of track or tunnels were detected, and after 14 days the JV had still seen nothing reportable. The effects of the TBM’s passage were actually significantly less than the tidal effects.

The third structure affected by the Crossrail project is Network Rail’s North Kent line and and adjacent Plumstead sidings on the line out of London Bridge between Woolwich and Dartford.

Crossrail was required by Network Rail to install detailed monitoring systems to detect any disturbance of its track and structures. The overall 24/7 track monitoring has been carried out by Crossrail, under contract C701, using conventional automated total stations. This was backed up by HMJV carrying out regular planned track geometry measurement trolley runs delivered by TES 2000 Ltd. Under Contract 310, the JV is responsible for checking this, and for undertaking any further special monitoring necessitated by their activities.

A crucial part of that last responsibility concerns the massive portal box structure within which Sophia and Mary were assembled before starting their drives. Some 300 metres long and 23 metres deep at the portal, the southern side of the box runs almost parallel to the route of the South Eastern and Freight trains passing every few minutes about six metres away.

Disruption of the train services along the line due to track or structure movement could not be tolerated, but it was necessary to install both secant and diaphragm wall piles parallel to the line to form the wall for that side of the box structure. In addition, the piles were to be re-enforced with steel cages and all this presented a challenge both to the engineers, and more importantly, Network Rail standards.

What could be done? The solution is an interesting story of its own, and one which is addressed in a separate article in this magazine. Suffice to say that it was a major challenge for the Network Rail asset protection team, led by Geraldine Quinn and colleagues.

Chris Parker
Chris Parkerhttp://therailengineer.com

Conventional and slab-track, permanent way, earthworks and embankments, road-rail plant

Chris Parker has worked in the rail industry since 1972, beginning with British Rail in the civil engineering department in Birmingham and ending his full-time employment at Network Rail HQ in London in 2004. In between, he worked in various locations including Nottingham, Swindon, Derby and York.

His BR experience covered track and structures, design and maintenance, followed by a move into infrastructure management. During the rail privatisation process he was a project manager setting up the Midlands Zone of Railtrack, becoming Zone Civil Engineer before moving into Railtrack HQ in London.

Under Network Rail, he became Track Maintenance Engineer, representing his company and the UK at the UIC and CEN, dealing with international standards for track and interoperability, making full use of his spoken French skills.

Chris is active in the ICE and PWI. He started writing for Rail Engineer in 2006, and also writes for the PWI Journal and other organisations.


  1. What a great, informative article. Just to add, the TBM cutting head rotates at 1 to 3 rpm and is pushing forward with 22 double hydraulic jacks.
    One question though – how many motors are driving the head, 12 or 16?


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.