After nearly three years of underground excavations it is a testament to the veracity of the extensive geotechnical studies carried out in the years before the Crossrail project finally got under way, that, bar “a couple of exceptions”, there have been few surprises. This is some achievement for a project with scope that runs to 42km of running tunnels, hundreds of cross passages, some of the largest soft ground SCL caverns to be found anywhere in the world, and eight new station excavations.
“In the construction phase, the information pretty much panned out as we thought it would,” said Crossrail Geotechnical Manager Mike Black. Where the pre-excavation data did not quite match actual conditions were at Farringdon Station and Stepney Green junction.
Farringdon Station is in the middle of congested London, located on a fault and where a number of TBMs were due to finish their drives. The site therefore represented a point of contractual boundaries between major tunnelling packages: east meets west.
The final pair of 7.08m diameter Herrenknecht TBMs to hole through at Farringdon have been driving from the east; the two TBMs from the west arrived at the station in the winter of 2013-14. Those earlier EPBMs made their mark in another way, too – helping to address local excavation difficulties for the proposed SCL works, arising out of the site’s geological complexities within the Lambeth Group.
“One of the fascinating rewards of this project, in terms of geology, was to have seen a full, open and stable excavated face of the Lambeth Group, excavated at Farringdon Station,” said Black.
Other tunnelling works that demanded detailed ground investigation data were at other, larger-scale SCL excavations – caverns that are among the largest in Europe, such as those carved out for the branch junction at Stepney Green in east London.
The near twin caverns were cut out at the bottom of a slender box structure at Stepney Green, where groundwater level is locally high and the structure is deep. The construction solution was to employ a two-step approach to dewatering – major drawdown, initially, and then locally ahead of the advancing excavations. For the ground investigation works, expensive inclined drilling played an important role.
More broadly, in considering what the project has learned about geology below London and how it may help future tunnelling schemes, Black said: “It’s still very difficult for future projects, but what we’ve done is to increase the knowledge of the problems that may be met. That knowledge informs a toolbox of options to help prepare for the works to be undertaken.”
What has been learned is that sand is scarce west of central London, but “is massive when found”; and, that sand in the east side of the capital lies in smaller, more sinuous layers, even becoming completely like sheets of sand.
Geology in the Farringdon Station area includes London Clay, Lambeth Group and Thanet Sands – and faults. The ground conditions have been among the most challenging anywhere on Crossrail.
Farringdon Station comprises running tunnels and large, deep structures at each end for the eastern and western ticket halls. The other tunnels of various sizes and lengths include cross-passages, concourse tunnels, escalator inclines and ventilation adits.
At Farringdon, the station’s twin running tunnels are about 25m below the surface, about where London Clay meets the Lambeth Group – which mix at the east section of the station; at the west, on the other side of the principal fault zone, it is Lambeth group and Thanet Sands. The water table is generally deeper than the tunnels – roughly where the Thanet Sands meets the still lower Chalk.
“We expected sand units at the Lambeth Group to be full of groundwater, but in the event it was one of the best stations of all to tunnel through,” said Black. The faults here were relatively “clean, dry and tight” with “no groundwater of significance.”
As a result, the material stood up very well at the face, and even with the SCL hand-tunnel work there were no significant problems – “although it turned out the sand was more extensive.”
At the main faces there were more sand channels and some water lower in the face, though no particular problems were presented by this. The groundwater here was tapped with passive wells, and there was no recharge.
Black said: “There were bigger sand units in the crown area, specifically Mottled units of the Lambeth Group. Due to their location, they were of potential concern for groundwater flow and stability.”
Initially, the plan was for each pair of approaching EPBMs to terminate at the ticket halls, with smaller TBMs being utilised to drive pilot bores between them for the following SCL platform tunnel excavations. However, as the EPBMs would be coming to the site anyway, it was decided to use the first pair – those boring from the west – to simply keep driving right through the station. The eastern ticket hall, therefore, became the common termination point for all TBMs.
A key benefit from this approach was enabling the SCL team to use the TBM bores to treat the ground for various tunnels throughout the station from inside the running tunnel bores – “like working from inside a pilot tunnel.” This was possible because there was dry space from which to treat the ground.
The key activities were probing, grouting and passive dewatering. The grout mix system was “very important” since the sand in the area was “mostly quite fine grained” according to Black. Due to the fine sand grades, “we needed a grout of low viscosity – sodium silicate grouts.”
“Only a small amount of grout work was needed at Farringdon,” said Black, but this location “was strategic to the entire construction plan.” Overall, one of the key aspects of the tunnelling experience at Farringdon was the rare chance to see so much of the Lambeth Group, large and up-close. The excavated faces gave “excellent exposure,” reported Black.
It was vital to establish the geometry of sand channels at Farringdon, and elsewhere, yet while much was known in advance, “we didn’t know the shape,” said Black.
Crossrail utilised the BGS predictive model with which to make 3D views. The model and information was then given to the contractors. “They adapted and updated the information as they undertook excavations. Their information confirmed that the original predictive model was good, but it was really good to have the updated data.”
Once the project is completed Crossrail will be sending all its geological data to the British Geological Society (BGS). Having received valuable input during the early site investigation phases, Crossrail is now giving back.
The project information is stored in the Asset Register, and held in the standard format of the Association of Geotechnical and Geo-environmental Specialists (AGS), allowing the data to be shared with other projects across London in future. The information details ground conditions around the running tunnels and the stations – the geology immediately around the underground structures.
With much site investigation already done by 2010, before main tunnelling started, “as client, we didn’t do much more,” said Black. “But the contractors did a fair bit more, focusing on refining the details, especially about the box stations.”
A lot of the extra ground investigation couldn’t be completed before the main works due to blocked access; requiring, for example, demolition of buildings on some contracts. The works had to be under way before vital extra site investigation to refine local geotechnical data and models could be completed.
By mid-2010, CRL had built upon historical and other project data to perform, or was planning to execute, a total of 33 packages of ground investigation work. Eight of the packages were classed as large (£1 million – £1.5 million), and typically comprised up to 40 boreholes.
As Black noted in 2010, a key lesson from previous tunnel projects in London, incorporated from the outset, was to carry out high quality ground investigation with detailed testing and description. In the event the final total of ground investigation packages was slightly higher than noted back in mid-2010. “We ended up with 37 packages of client-procured ground investigation,” said Black.
Additional, ad hoc ground investigation was undertaken by some contractors where required. But the large volume of data generated as investigations proceeded didn’t throw up any significant surprises.
Consequently, no major design changes arose for key structures due to unexpected turns in the geotechnical data. But decisions about some, relatively small underground structures – such as cross passages – depended on vital local geotechnical data. Typically on the project, SGI rings are used for cross passages built in the east section and SCL for some of those linking the running tunnels in the west (where there is London Clay, no water). The challenge comes up as east-meets-west in central London, and therefore the choice of cross passage structure at each location demanded specific local geotechnical data.
As volume loss in the ground and consequent concerns around settlement arise from the disturbance caused by TBMs, the planning phase of Crossrail closely studied this factor in tandem with the capabilities of high-specification shields.
“In planning, we started with an estimated volume loss of 1.7% and ended the plans with a best estimate of 1%,’ said Black. Actual experience, however, showed settlements at half this prediction, “and even less in many parts.”
The tunnel contracts specified limits of 1%, down to 0.5% settlement. This proved more than enough margin, with Black reporting that the TBM bores between Royal Oak portal and Paddington Station – where the first TBMs started boring on the project, in west London – the volume loss was recorded at just 0.04%.
“This was a minute volume loss, and the best, but almost similar figures were repeated elsewhere,” said Black.
The unprecedented low volume loss for such a large-scale TBM operation in a dense urban environment such as London, was achieved as a result of a number of factors: the hi-spec shields; as well as monitoring both the spoil material being produced within the TBM and on the ground surface.
Within the TBM, the monitoring task was performed through a combination of three key measures: spoil weighing on the conveyor belt; laser scanning of the spoil since different ground materials don’t all heap the same way on the belt; and, sampling the recorded CCTV images.
The control and monitoring system was applied to all TBMs as Black highlighted it would be before boring began.
“We never intended any plan of dewatering over massive areas – we expected the TBMs to cope, and they did in most cases,” said Black. “Instead, dewatering as required was carried out locally – for example, for cross passages.”
Crossrail, unlike the Channel Tunnel excavations, opted for a local dewatering regime and only in areas where TBMs were not to be used. Local dewatering sometimes had to draw down the water table significantly, by 20–30m. However, as the terrain is fairly flat the settlement effect was minimal; the hydraulic gradient only steepened closest to the dewatering points, and so was quite localised.
Aside from the running tunnel cross passages – most of which have been excavated – the other key sites for dewatering were the SCL caverns, most notably Stepney Green junction.
TBM slow-downs resulting from wet ground have been only occasional, and again due to localised conditions. “It was simply the case that along some short stretches the wet ground would flow or drop on top of the shield. Consequently, there was some increased friction which reduced production, briefly,” explained Black. In all cases the ground conditions changed back to being more favourable again.
Beyond the successes of the giant TBM bores below London, the Crossrail project can also point to other major tunnelling accomplishments – construction of the large SCL caverns at Stepney Green, Fisher Street and the gardens at Vallance Road on contracts C305, C300 and C510, respectively.
While the Stepney Green structure is the junction for two branches of Crossrail in east London, the other two SCL caverns are for crossover tracks between the running tunnels.
Three large SCL cavern sites were a lot to get right on one project – especially as they were on separate contracts. Given the design and construction jump in SCL capability for the structures, each being excavated in congested and expensive urban environments, the client set up an SCL Working Group to ensure lessons and experiences were shared across contracts.
“This was a regular forum established early in the construction phase to bring the various SCL practitioners together to share lessons learnt, innovations, common design issues, etc.,” said Black.
Also vital to enabling the step-up in the scale of SCL cavern construction has been the advances in measurement control in recent years, as TunnelTalk was told prior to commencement of the ambitious works at Stepney Green.
The SCL caverns at Stepney Green junction were up to 17m high x 16.5m wide – among Europe’s largest soft ground/soft rock voids. Monitoring has been vital to the entire project, perhaps especially so for the SCL caverns – which also required major dewatering. “Groundwater is very high at Stepney Green, and together with the Lambeth Group’s sand we expected that would be a potential problem.”
Dewatering was performed from the surface, taking down the water table most of the way required – but not quite far enough. Black said that dewatering “couldn’t get the groundwater low enough to below all excavation levels.”
The solution was to perform supplemental, active in-tunnel depressurisation. Basically, this involved pumping out the groundwater from within the immediate excavation – just ahead of the face. Effectively, they were “chasing the groundwater” as tunnelling advanced, ensuring the lower parts of the excavation remained dry, says Black.
A presentation on the Stepney Green in-tunnel depressurisation work won an internal Crossrail papers competition, and is contained in Vol. 1 of the ICE-published collection of technical papers.