The contract C310 is part of the current biggest infrastructure project of Europe "Crossrail" which is a major new cross-London rail link project and comprises the construction of the North Woolwich Portal, Plumstead Portal and the twin tube Thames Tunnels with a length of approximately 2.6km. Two Mixshield TBMs (diameter 7.12m) are driving through differing challenging ground conditions (Thanet Sand, Gravels, Flint and Chalk) below the ground water table. The two tunnels are underpassing several listed buildings, sensitive structures, operational railway tracks with low overburden and close to existing subway tunnels in an urban environment. Several additional measures as compensation grouting, micro piles and an intensive monitoring have been carried out to ensure a safe tunnelling process. During the drive underneath the River Thames, the tunnels will only have an overburden of approximately 12m.
The effect of pressure variation due to the tidal River Thames has to be taken into account for the control of the face support pressure. The excavated and the dissolved soil were pumped to a treatment plant where a dewatering process with filter presses was used with great success. After finishing the two main tunnels four cross passages were opened using a sprayed concrete lining. To reduce the water inflow in the chalk aquifer fissure grouting has been executed prior to construction of the cross passages.
Project overview Crossrail
Project Crossrail
Crossrail is currently the biggest infrastructure project in Europe which has been developed to serve London and the southeast of England.
The project includes the construction of twin-bore tunnels on a west-east alignment under central London and the upgrading of existing National Rail lines to the east and west of central London.
The new rail track, as shown in Figure 1, starting in Shenfield/ Abbey Wood in the east of London and ending in Reading / Heathrow Airport in the west of London, includes the construction of nine central area stations, providing interchange with London Underground, National Rail and London Bus services, and the upgrading or renewal of existing stations outside central London. Altogether there will be 118km of new rail tracks, including 42km of tunnels and 40 train stations. After completion in 2018, Crossrail expects 200 million passengers per year (Tauriainen, Rädle, Ingram 2013).
Contract C310 Thames Tunnels (bored tunnel drive H)
Hochtief Murphy Joint Venture (HMJV) is responsible for the construction of two tunnels underneath the River Thames with the contract name C310 Thames Tunnel. The contract C310 comprises the construction of the Plumstead and the North Woolwich Portal, as well as the twin bored tunnel with a length of approximately 2.6km. Contract commencement date was early March 2011 and in November 2011 the construction of the tunnel portal in Plumstead started. The main tunnelling works started in January 2013 and the anticipated completion of the tunnelling work including cross passages and pump sumps is September 2014. Completion of the whole works is expected by mid June 2015. C310 is a challenging project, both technically and operationally, due to its location and difficult geotechnical conditions.
The two TBMs will drive through varying ground conditions (Thanet Sand, Alluvium, and Chalk) below the water table (see Figure 2). During the drive underneath the River Thames, the tunnels will have an overburden of approximately 12m. The effect of pressure variation due to the tidal River Thames has to be accounted for in the control of the tunnelling operations. The tunnel will pass underneath several listed buildings, operational railway tracks, utilities, close to existing subway tunnels and will pass through Woolwich Box Station (Tauriainen, Rädle, Ingram 2013).
Geotechnical and Hydro Geological conditions
Geotechnical situation
The site geology consists of the Chalk group with intermediate flint layers overlain in parts by Thanet Sands and in parts by floodplain Terrace Gravels, Alluvial layers (Clay, Silt and Peat) and Made Ground.
Tunnelling started from Plumstead Portal south of the River Thames in the river terrace gravels which are typically described as medium dense to very dense, grey brown or dark grey, slightly silty, slightly clayey, fine to coarse sand and fine to coarse sub-angular to rounded flint gravel. The sand and gravel proportions varied from very gravelly sand to sandy gravel.
The underlying Thanet Sands were met shortly after TBM launch. These are predominately sequences of dense to very dense, fine-grained sands, with higher proportions of clay and silts in the lower part. The unweathered formation is grey to brownish grey, and at the surface it weathers to a pale yellowish grey. The genuine flint belongs to the basal Bullhead Bed which is a conglomerate comprised of rounded coarse flint gravels and nodular flints in a matrix of dark greenish grey, clayey fine to coarse grained sand.
The chalk horizon will be met after 100m from the launch site and will lead to mixed face conditions. Chalk is generally considered to be a soft/weak rock, very pure white limestone formed from the skeletal remains of submicroscopic algae.
However, unlike many limestones, chalk is very widespread as a consequence of its entirely planktonic origin. Two features commonly found in Chalk are flints and marl seams. Flint is a microcrystalline silicarock that occurs as dispersed, usually black nodules or as tabular bands or sheets. Flints represent very strong, brittle inclusions in contrast to the comparatively weak host Chalk matrix. Marl seams are horizons with increased concentrations of clay.
Two different chalk layers were hit by the C310 tunnel, the Chalk of the Haven Brow Beds and the Cuckmere Beds. The Chalk is very weak at the top of the layer, highly weathered and described as low density white Chalk improving in both strength and density with depth. Generally the Chalk is described as medium density chalk.
Approximately 80 per cent of the tunnel length is located within the chalk horizon. Only 150m to 300m before reaching and after leaving the portals and Woolwich Box the tunnel face is located in the Quaternary and Palaeogene soils. (Crossrail, 2007)
Groundwater conditions
In the area of C310 the hydrogeology is controlled by two aquifer systems made up of the Upper Chalk and the overlying floodplain Terrace Gravels. Since the gravels or the Thanet Sands lie directly on the pervious chalk, the two aquifer systems are in hydrostatic contact and have achieved equilibrium.
On the south and north side of the River Thames the aquifers of Chalk, Thanet Sands and Terrace Gravels are sometimes confined by overlying Alluvium (clay, silt and peat). In parts of the river, the Alluvium is non-existent and hence the chalk-gravel aquifer is connected directly to the River Thames.
A response to the tidal motion of the River Thames was observed, depending on the distance to the river. The maximum groundwater level is set at 104.5m above tunnel datum (mATD) whereas the minimum is set at 96.5 mATD.
The variation of ± 4m to the average groundwater level of 100.5 mATD can be found in close proximity to the River Thames. The rest of the tunnel alignment has a preset variation of ± 1 m. Based on this information Hochtief developed the water level diagram in Figure 3 and defined three design water levels which were used for the support pressure calculations. These water levels are not the water levels of the Thames but the expected pore pressure in the chalk/ gravel/sand aquifer.
An additional item which needs to be considered in the calculations is the time dependence of the damping of the pore pressure. In the boreholes adjacent to the river the measured delay was approximately 0.5 hour.
TBM Tunnelling Underneath Sensitive Urban Structures
Launching in close proximity of North Kent Railway Line
Directly after cutting through the diaphragm walls at Plumstead Portal for duration of approximately 36 calendar days the TBMs were excavating in Network Rail’s zone of influence close/adjacent to operational North Kent Line, underneath White Hard Road Bridge and Cathedral Substation. In this area the tunnel cross section is located within Mixed- Face Conditions (Gravel, Thanet Sand and Chalk) with low overburden of approximately 7 to 10m for the first 150m of tunnelling.
Real-Time Monitoring devices consisting of ATS and prisms on the railway and hydrostatic levelling cells on the bridge and substation were installed. The HMJV monitoring system, Advanced Tunnel Drive Steering (ATDS), was collecting any movement and/or potential settlement on an ongoing 24/7 basis and showing the data in real time.
Underpinning White Hart Road Bridge
White Hart Road Bridge is located approximately 40m from the launching point of the TBMs. To mitigate possible settlements caused by the tunnelling operations of the closer passing TBM 1, the foundations of the bridge required supporting by means of underpinning.
To strengthen/support the foundation of White Hart Road Bridge, a micro-pile and soil anchor scheme was the preferred method for the underpinning of the bridge. The scheme comprised of a set of three piles at approximately 500mm c/c along the length of the abutment.
These sets consisted of two raking piles – supporting vertical loadings and one soil anchor – which were designed to control any subsequent horizontal movement. To ensure no load was transferred into the piles during the tunnelling operation, the piles were permanently cased to the invert level of the proposed tunnel, thus transferring the load of the bridge into the chalk strata below the tunnel invert.
The soil anchors were stressed to provide the required reaction force to prevent lateral movement of the bridge foundations. The ground anchors were de-stressed and removed as the tunnelling operations have been ceased; hence they were classed as temporary.
During the passing of the first TBM, no significant movement of the bridge or its foundation were observed. All measured movements where less than 2mm, however the north eastern wing-wall of the bridge, which is not connected to the bridge itself, was noted to have settled 6mm during the time that the TBM was at a standstill at this location for more than 12hrs due to a grouting issue. The bridge was monitored with a number of hydrostatic levelling cells that fed real time data to ATDS so this movement could be observed and monitored in real-time by HMJV. This demonstrated that the employed scheme was effective and the correct chosen method.
Compensation grouting for Cathedral Substation
The existing building Cathedral Substation (Network Rail Substation) is located west of White Hart Road and White Hart Road Bridge and north of the running North Kent Line. For Cathedral Substation the requirement for compensation grouting to prevent settlements caused by tunnelling is driven by the need to protect the highly sensitive equipment within the building. Movement criteria have been agreed with Network Rail and Crossrail after contract award.
The distance between the tunnel crown and the floor was only 6m and compensation grouting was designed for River Terrace Deposits and groundwater with three layers of TAM (tube-amanchette). During installation of the TAM Drillings prior to tunnelling no settlements occurred due to the chosen drilling technique (duplex heading technique with full casing and subsequent immediate grouting of each hole) and the designed pre-heave of approximately 3mm have been achieved.
Both TBMs have successfully underpassed and by use of reactive compensation grouting the movements of the building maintained in the allowable limits (between +5 and – 10mm).
TBM Tunnelling underneath Thames river
Tidal fluctuation
Tunnelling underneath the River Thames was one of the challenges of the TBM drive. The overburden to the river bed was minimum only 10 to 12m.
Due to the proximity to the sea, the Thames Water Level and therefore the adjacent groundwater level, is highly affected by the tides. The duration of one tide cycle is about 12 hours, meaning that two cycles with two minimum and two maximum peaks take place per day. Between low tide and high tide the Thames Water Level alternates over 8m, causing relevant face pressure changes of approx. 0.8 bar. Therefore the support pressure has to be constantly reviewed and adapted.
An additional item is the time dependence of the damping. In the boreholes adjacent to the river the measured delay was approximately 0.5 hour. This time effect will be covered by an increase of the pore water pressure by 2.5 m for low Thames Water levels and the damped value is used to calculate the corresponding theoretical support pressure.
Detailed operational tables for the supporting pressure for the regular tunnel drive were derived from the detailed calculations. For each ring position and respective tunnelmeter the corresponding pressures for the three different water levels have been compiled and interpolated between the calculation cross sections. Based on those tables an automatic calculation and correlation of the support pressure to the tide measurements (illustrated damped water level) within the TBM Data Process Management System TPC takes place and the theoretical support pressure is displayed.
The damped water level for the calculation of the theoretical support pressure is capped at a minimum of 99 mATD although the minimum design river water level is 96.5 mATD. The reason for this is to maintain a conservative assumption. During low Thames Water Level it has to be assumed that through the damping effect some water could remain in the ground, which leads to a higher pore water pressure than the actual water pressure created by the river level.
Interventions before and after Thames Crossing
Before starting the TBM drive there was an intensive discussion on the frequency, the location, the requirements and the procedure for interventions in the working chamber. Based on experiences from the project CTRL 320 the abrasion of the cutting tools was evaluated as low although the flint embedded in the chalk is highly abrasive.
Due to the high pore water pressure below the River Thames supporting pressures of more than 3 bars were required for a full face support. To avoid interventions under such high pressures, interventions were planned before and after the crossing of the River Thames. At these locations the geotechnical ground model predicted a structured medium density chalk with a small amount of thin fissures. This good chalk quality allows an optimisation of the interventions. A procedure of tunnel face inspection and monitoring has been developed to allow a reduction of the required air pressure in the working chamber stepwise to 1 bar.
In particular the water inflow has been monitored continuously.
In addition to the interventions required for the TBM drive additional three interventions were made to investigate the chalk characteristic at the locations of the planned cross passages.
Soil treatment
Separation for the slurry shield
Due to the use of a slurry shield TBM a plant is necessary to filtrate the spoil out of the slurry to recycle it. This task is handled by the slurry treatment plant (STP). To manage the excavated material by a maximum tunnelling advance up to 80mm/min a slurry flow rate of 1600m3/h needs to be treated. This material consists of Terrace gravels, Thanet sands and Chalk with a various amount of flint stones (5 to 25 per cent).
In a first separation step, a scalping unit as a rotating drum with an integrated screen cuts off all the material with a size bigger than 8mm. This material drops straight on a belt conveyor and is transported to the dump. All the slurry and the grains, which passed the openings in the screen, are split in three parallel lines including a desanding and desilting section. In each line two big cyclones with a separating cut of 70µm and 12 small cyclones with a separating cut of 40µm are installed. After passing the desanding and desilting section the slurry will be reused in the slurry circuit.
Chalk separation
Based on experience with a Mixshield TM in chalk the support fluid was replaced with water to achieve stability. The chalk dissolves very quickly after excavation and mixes with the water in the working chamber and the slurry pipes. The particles of the excavated chalk spoil, which are up to a 100 per cent smaller than 40µm, require a special handling.
In former projects the dissolved chalk was handled with centrifuges and afterwards the water content was reduced by adding a high amount of cement.
In this project an alternative method was used to suit the agreed spoil management and disposal: aim was that the moisture content of produced spoil from the treatment plant shall not exceed 35 per cent. The last separation step was the treatment of the waste mud by carrying out a filtration process with the filter presses. Before the slurry was pumped into the chambers of the filter presses, a defined amount of lime milk was added.
The settings for lime milk concentration and lime dosage were calculated utilizing the values of outflow of waste mud tank measured by a flow- and densimeter. The purpose of lime milk for the filtration process was to accelerate the process itself and also to reduce the stickiness of the filter cake surface to improve the discharge of the filter cakes.
The filtration process includes the steps of closing, feeding, inflating (only membrane filter presses), core blast, opening and demoulding. There are six filter presses, where each allows a treatment of 14.5t/h of dry mud. Four of them were normal chamber filter presses with a maximum closing pressure up to 250 bar, supplied by two hydraulic jacks.
Another two were membrane chamber filter presses, which were conducting an additional squeezing/inflating process after the normal feeding process. Due to this additional process, the closing pressure was with 400 bar much higher than with the normal chamber filter presses. The excess mud treatment started with the closing of the 100 chambers of a filter press and was followed by the mud feeding. Thereby a volume of approximately 7,300 litres of limed mud out of the storage silo was pumped into the chambers. The solid particles in the slurry were collecting itself on the surface of the filter cloth and creating a, so called, filter cake. At the same time the water flowed through the cake and the cloth into the plates, where it was led via small channels to a drain into a tank. In consequence of the increasing thickness of filter cake, the pressure in the filter presses went up to 7 bar. At this pressure set-point the feeding pumps were regulated to maintain the pressure at this level for a certain time until the filtration process was finished. Before the core blast sequence started, the inflating mode was carried out by the two membrane filter presses. During this mode, the volume of the chamber was reduced by inflating membranes in the plates utilising 14 bar water pressure.
Due to this compression, residual water contained in the cake was removed. The pressure was held over a certain time until the process was ended by a required criterion.
After finishing this mode the liquid core in the presses needed to be removed. This was necessary to prevent the liquid mud falling down on the dry filter cakes. Therefore compressed air was passed through the centre of the filter press to discharge the material into a small tank, where it is pumped back to the limed mud silo. Before the hydraulic pressure was released, to allow an opening of the plates, the membranes got drained.
Then in a defined order compressed air was pushed in hydraulic jacks on both sides between the plates to provide enough space for dropping the filtrate cakes.
The discharge of the cakes was done by gravity and supported by shaking of the plates. The cakes falling down in a provided box for collection with a wheel loader were mixed up with material from the belt conveyor and loaded on a lorry.
Cross Passages
For optimisation the excavation and support concept for the cross passages was changed from SGI Support Elements to Sprayed Concrete Lining. To reduce the permeability of the Chalk and to reduce the water inflow to a manageable level fissure grouting in the area of the future cross passages was carried out within the Chalk. The fissure grouting was executed for two of the four cross passages from the finalized tunnel and for the remaining two cross passages from the ground surface.
Due to the high asymmetric loading on the segmental lining underneath the River Thames (embedding of segments, water pressure and ground treatment) and the low ratio of the diameter of the TBM Tunnel versus the diameter of the cross passage, it was necessary to install a temporary steel support structure at each cross passage opening location prior to the commencement of the cross passage excavation. The support structure remained in-situ until the cross passage works were completed.
In preparation of the excavation the success of the grouting was proofed with probe drilling and the segments were removed. After excavation and the completion of the sprayed concrete primary lining a waterproof membrane and a cast in situ secondary lining were installed.
For the opening of the segmental lining within the tunnels the alternative use concrete segments with a shear bicone (combined steel and polyamide dowels) in the circumferential joints in lieu of the original designed steel opening sets in the area of the opening was proposed.
To allow the alternative design to work effectively, a higher design concrete strength and combined reinforced segments (steel as well as steel fibre reinforcement) have been developed for five fixed rings in the area of the opening which also integrate a "soft keystone" to simplify the opening and removal works.
Questions from the floor
Roger Bridge, BTS chair: Thank you for the presentation. It’s interesting to see some of the challenges that you have faced. It is a credit to you; it’s a superb job. I think now we will take this opportunity to see if anyone has any questions they would like to ask.
Andy Sindle, Crossrail: Could you tell me a bit more about the grouting for the cross passages when you were doing the Scl work in chalk? i would imagine that would have had to be pretty dry for you to be able to do it. You said you treated the ground first, but could you tell us a little bit more about that please?
Answer: At cross passage 19 we did some treatment from the surface in some Thanet Sands on the top because the overburden was only 2m of chalk above cross passage one. For jet grouting in the chalk you have to use microfine cement. This is quite important because the openings are quite small. Cross passages 18, 17 and 16 we did ground treatment from within the bored tunnels. Ascending and descending stage techniques. Open the hole with a picker and microfine cement. and a pattern of 1.5m between the drillings. Performance criteria were agreed with Crossrail.
Colin Mackenzie, retired: First of all, congratulations on a first class presentation. It’s one of the best I have seen in a long time. in the portal areas you had temporary works in which you had two levels of struts and these were clearly heavy struts, and the lower ones were hydraulic, and as you were inducing hydraulic loads you would know exactly what loads you were putting in. Did you find that the loads, how closely the loads corresponded with your design loads because in London there have been cases where the design loads didn’t appear… notably on the limehouse link where a major level of struts was eliminated and a lot of money was saved. Did you find that the loads matched the design loads?
Answer: The loads were about 20 per cent of what we anticipated so they were a little bit over-designed. But one reason was by Plumstead portal next to the railway of course we wanted to be on the conservative side but there was a requirement that if one of the props were accidentally removed, the structure would still take all of the loads. and so that is the reason for the very heavy railers and props. But you are right, it can be optimised.
Mackenzie: Noting the proximity of the railway, I’m not surprised that you were forced into a conservative approach. Thank you.
Steve Price, Mapei: I apologise if it’s a simple question but I would like to know a bit more about the synthetic membrane and how that seems to waterproof the tunnel, that we saw in one of the slides.
Answer: It is a 1.5mm thick membrane, the placing and testing, and welding is quite important. These are the main requirements and ensure that it is watertight to get the 120-year design life. You need a high quality, and a high standard of inspection.
Chair: Thank you very much for the presentation.
