The Lee Tunnel, part of the Thames Tideway Scheme is currently under construction in east London. A joint venture comprising Morgan Sindall, Vinci Construction Grands Projets and Bachy Soletanche (MVB) is undertaking the works. Thames Water awarded the Lee Tunnel contract, which is the largest project awarded in the UK water industry since privatisation in 1989, to MVB in January 2010. The contract incorporates four deep shafts, the largest of which is 38m in diameter, and, in addition, a 6.9km long tunnel.

The deepest tunnel in London, it is intended to capture the discharge from the combined sewer overflow at Abbey Mills preventing it from flowing into the River Lee and ultimately the Thames. The scheme transfers the flow through the new tunnel for treatment at Beckton Sewage Treatment Works. The TBM being used to drive the tunnel was manufactured by Herrenknecht. It is a Mixshield slurry machine, 9m in diameter, and designed to build an 8.5m external diameter fibre reinforced segmental concrete primary lining.

The Beckton site comprises three shafts; the overflow shaft from which the TBM was launched; the connection shaft, an intermediate online shaft and the pump shaft which is offline of the main tunnel, but connected to the connection shaft via a smaller diameter sprayed concrete lined tunnel. The TBM reception shaft located at the upstream end of the tunnel is located at the Abbey Mills Pumping Station. All four shafts are in excess of 70m deep and have been constructed using diaphragm walling techniques for ground support during excavation and a double sided concrete slip-formed chimney to produce the internal shaft finish.

The focus of this paper is the crossing of the connection shaft by the TBM. Due to ground contamination at Beckton, dewatering of the ground at the shaft for tunnel entry was precluded. Furthermore Thames Water and MVB wanted to ensure that high pressure annulus tailskin grouting was possible on both entry and exit of the shaft. The TBM was required to break into the shaft which had to be partly filled and flooded, at full face pressure in order to balance the groundwater pressure for grouting. This would maintain positive pressure within the shaft and ensure no contamination could migrate through the tunnel eye. In addition to the challenges this presented, a method of changing the TBM tooling and nose cone was required which would not affect the critical tunnelling programme

Thesis
This paper considers two key aspects of the TBM crossing of the connection shaft; firstly the part filling and flooding of the shaft to enable the crossing of the TBM without drawing in any contamination and enable annulus tailskin grouting at full pressure; secondly the design of the buried sacrificial chamber which enabled a fast tool change with minimum impact on the tunnelling programme within the shaft fill.

As the connection shaft and portals were programmed for construction before the tunnel arrived it was always known that the TBM would have to cross the connection shaft with the axis of the machine 13m above the shaft base slab. Various systems were discussed including, a partial concrete bath to support the TBM up to axis level, a steel crossing can or a foamed concrete fill, both of which would fully encapsulate the machine. Unfortunately due to the presence of the contamination and the need to maintain positive pressure in the shaft therefore excluding any potential contamination ingress an alternative method of shaft entry was required. This necessitated a full face pressure entry (6bar) from the slurry machine, rendering none of these preferred options possible.

In order to exclude the contamination as the TBM broke through the shaft lining, a partial shaft fill with fluid or solids was required. The fill placed in the connection shaft had to be capable of resisting 6bar of slurry pressure generated by the TBM entering the shaft at full pressure, and also physically supporting the weight of the TBM until the machine had transited 13m into the shaft and subsequently built and grouted tunnelling rings behind itself through the shaft wall break in zone, to close the contamination pathway. Once this point was reached the TBM could then transit across the shaft and stop to perform a tool change on the TBM front face, before exiting the shaft again at full-face slurry pressure.

The shaft was filled with a designed cement bound granular material (CBGM); this consisted of Thanet Sand which was a beneficial use of recycled spoil from the shaft excavations and added cement. The fill was compaction specified at point of placement, batched on site and delivered to the pit bottom in skips, water being added as it was spread out in the shaft by excavators. This CBGM layer was made up of 14.5m depth of material (2m beneath the TBM and 4m above). It was underlain by a bed of single sized stone placed around the shaft bottom permanent works that were completed before shaft filling commenced. Above the CBGM layer was placed a 3.5m layer of Thanet Sand then 50m of water. The CBGM layer above the crown of the machine plus the layer of un-bound Thanet sand was designed incorporating a factor of safety of 1.2 to withstand the minimum operating slurry face pressure of the TBM of 1.2 bar. The remaining 4.8 bar was balanced by the 50m of added water on entry and exit of the machine. Therefore, the shaft had to be flooded before the machine broke in and again prior to its exit.

When placing the CBGM fill around the GRP chamber, to be discussed in further detail in due course, particular care had to be taken to avoid cracking the structure as vibrocompaction was being used. An exclusion zone was put in place around and above the chamber to limit the vibration impact. Furthermore a layer of lean mix concrete was placed above the chamber and lightly compacted with wacker plates. Plywood cladding was used to protect the shaft internal slipform lining from any damage from the plant or fill as this formed a part of the permanent works, the plywood effectively acted as a separation membrane between the shaft lining and the fill. In the two zones where the nose cone (at the centre of the TBM cutterhead) would intersect the shaft lining several cores were drilled in a circular pattern in order to weaken the lining and minimise the damage to the nose cone as it broke into the shaft. This damage limitation detail was of greater importance on the exit of the machine from the shaft as the nose cone would have only just been replaced.

It would have been preferable to keep the shaft flooded whilst the machine transited across the shaft, but in order for repairs to be undertaken on the cutting head, nose cone and brush seals, the shaft had to be emptied of water to prevent the risk of sudden inundation of the tunnel.

At that point in the design and planning, the permeability of the CBGM material for dewatering purposes was as then still unknown and could only be estimated, although due to the coarse nature it was expected to be relatively permeable .

Transit sequence across connection shaft

1. Flood the shaft using the site water supply at the connection shaft site.

2. Approach the outside of shaft with the TBM and breakthrough the diaphragm wall, annulus infill and slipform lining.

3. Transit 13m across the CBGM filled shaft, until tunnel ring building and grouting closes the contamination pathway behind the TBM.

4. Reduce the 6 bar slurry face pressure to 1.2 bar and drain down the shaft through the pipe work positioned within the fill using the TBM slurry handling system.

5. Transit across the CBGM shaft fill and intersect the head intervention chamber stopping 200mm short to avoid damaging it, perform the nose cone and cutterhead changes.

6. Transit through the sacrificial chamber.

7. Build two tunnel rings and then take them down, in order to create enough space to change the brush seals within the tailcan.

8. Transit to within 1m of the internal slipform lining and stop then fill the shaft through the pipework positioned in the fill using the TBM slurry handling system.

9. Exit the shaft at 6bar face pressure

10. Transit 13m out of shaft until ring building and grouting closes the contamination pathway.

11. Drain down the shaft using submersible pumps from the pit top at the connection shaft site.

Flooding & draining of the shaft
Initial flooding of the shaft was completed off-line of the main tunnelling programme from the connection and pumping shaft site water supply. This took in excess of a week to achieve but could be accommodated within the programme as the TBM was still several weeks away from breaking into the connection shaft.

Draining down and re-flooding the shaft to safely facilitate the tool changes was achieved using the TBM’s own pumps and the large capacity storage of the slurry treatment plant, usually used to separate out the tunnel spoil from the slurry. The water used to fill the slurry treatment plant tanks and then the shaft was taken from the final effluent channel near to the overflow shaft site and fed through the slurry handling system to the face of the TBM. Once there it was transferred into the shaft by a specially designed and modelled piping system. This arrangement consisted of two side by side pipes running along the tunnel alignment (in the path of the advancing TBM). This horizontal pipework ran across the full width of the shaft and then turned vertically upwards emerging at the top of the fill. Where it emerged through the fill a remote controlled actuator was fitted, this allowed flow into and out of the pipe to be controlled remotely and acted as a cut-off against sudden inundation during the shaft crossing.

To resist the internal pressure it would experience, the uPVC pipework had to be encased in concrete on both the horizontal and vertical sections. Further design was required to calculate the time it would take to fill/empty the shaft using the TBM slurry handling system. This developed into a detailed staged hydraulic calculation involving; high friction losses, local losses, varying delivery flow rates due to capacity issues and a requirement for a constant discharge pressure utilising a vertical standpipe and the bubble pressure of the TBM.

Unfortunately, the horizontal section of the pipework become blocked as the TBM transited across the first section of the shaft, just after breaking through the shaft lining approaching the intervention chamber. This was due in part to the alignment of the pipework which had been diverted below the intervention chamber effectively creating a trap. It was thought that the high pressure of delivery of the water would clear any blockages but draining the shaft before the head intervention took much longer than expected, due for the most part to the semi-blocked pipework. In hindsight this horizontal section should have been detailed to include a slight rise all the way to the opposite side of the shaft to the point at which it turns vertical.

Once the head intervention had been completed and the TBM transited across the shaft to within 1m of the internal lining, the shaft again had to be re-filled. At this point there was very little of the horizontal section of the pipework remaining, the TBM was effectively in direct contact with the vertical section of the pipework. This meant the shaft filling performed much better and behaved in-line with the hydraulic predictions and modelling. Once flooded, the TBM safely exited the shaft at 6 bar slurry face pressure in order to keep the contamination from moving into the shaft. Finally when the TBM had exited the shaft and transited sufficiently far to build and grout tunnel lining rings into the breakout zone, the shaft was drained down again, but the tunnel was now sealed from the connection shaft. Therefore the shaft had to be drained down from the connection and pumping shaft surface site using submersible pumps, as with the initial shaft filling this was now not on the critical path and therefore not time dependent.

Head Intervention Chamber
The option to fill and flood the shaft solved the contamination problem but gave rise to another. When the TBM was to pass through the connection shaft this presented the final opportunity in a ‘safe’ and dry environment to replace the nose cone and perform any potential welded repairs on the cutter head. This, if not for the contamination and grouting requirement leading to the shaft fill, would have been undertaken in the shaft whilst the TBM was positioned on a steel or concrete cradle arrangement. In our case the depth to the nose cone of the cutter head was over 10m within a saturated CBGM block. Access to undertake these works could be gained through the head intervention doors on the TBM but a large void would have to be created in order to provide enough working space to change the cutter head tools and the nose cone. This could have been completed using a staged timber heading, but due to the programme pressures on the tunnel drive this was ruled out.

Sheet piling and excavation from the surface of the fill was also discussed but dismissed due to the depth, plant requirement and again time constraints.

In a weekly tunnelling temporary works meeting between Morgan Sindall Underground Professional Services (MSUnPS) and the MVB TBM site team, the author put forward the idea of a sacrificial ‘intervention chamber’, constructed from concrete. This chamber would be precast on the surface (off of the construction critical path), and could then be lifted into position at the appropriate level during the shaft CBGM filling and buried to await the TBM. The buried intervention chamber had to be capable of supporting the fill above and around it, especially during the TBM approach, which would, due to the high torque exerted by the TBM cause high lateral loading. As the design of the chamber was undertaken before the shaft fill material was finalised, the fill material parameters and the depth of material above had to be conservatively assumed. Due to this very high torque generated by the machine it was decided that the TBM would stop 200mm short of the chamber to avoid any direct contact between the TBM cutter head and the concrete, to avoid causing damage before miners and engineers entered the chamber to undertake the nose cone inspection and any required work.

The head intervention chamber was placed directly on the alignment of the tunnel in order for work to be undertaken on the front of the TBM. The structure had to be sacrificial, i.e. the machine had to be able to cut through the chamber once the repairs were finished, and this presented a problem. The TBM is capable of breaking through concrete and could cut traditional steel reinforcing rebar but the steel would block the slurry pumps and separation plant on the surface and had the potential to cause significant damage to the system.

The geometry of the chamber had to be such that it could facilitate work on the entire cutting head surface. This resulted in the need for a 5.5m clear span working area 2m in height and 1m in depth. By positioning the chamber such that the nose cone (at the centre of the TBM) was in elevation on the far left of the box, by then rotating the cutting head the entire face could be sequentially accessed. Due to the very large span required and in order to keep the structural sections relatively thin a central wall was added with a localised recess to maintain access to the entire cutter head, effectively creating a twin bore culvert.

Analysis
The loading on the head intervention chamber I modelled using the finite element package Plaxis 2D, the fill geotechnical parameters were not known at the time of analysis and design. The material properties had to be conservatively assumed to have minimal shear strength and a unit weight of 19kN/m³. Importantly only the ground loading was applied to the chamber. The 50m head of water pumped into the shaft was ignored as the permeability of the coarse material and the presence of large openings in the chamber (it had an open front face) meant that the water pressure could equalise as required and therefore not transfer any load to the chamber structure. In addition to this high permeability, water pathways existed throughout the shaft between the internal lining and its plywood protection, and the shaft flooding pipework. Clay had been discussed as a capping layer option as opposed to the CBGM; it was thought that due to its cohesive properties clay would better contain the face slurry pressure from ‘blowing’ the fill layer.

However if a clay layer had been used it would have drastically increased the load on the chamber as it would act as a barrier to the flood water finding its way into the chamber for pressure equalization.