The Silvertown Tunnel Project will be a 1.4km-long road link under the River Thames linking Silvertown with the Greenwich Peninsula in East London. The twin-tube’s tunnels are 10.8m i.d. and are under construction. The scheme’s purpose is to relieve the traffic congestion near the Blackwall Tunnel and allow better public transport links, including more cross river bus journeys.
Transport for London (TfL) contracted with the Riverlinx Construction Joint Venture (CJV) – a joint venture of Ferrovial Construction UK, BAM Nuttall and SK Ecoplant – to build the tunnels and approaches to the existing roads.
The Riverlinx CJV instructed to tunnel boring machine (TBM) manufacturer Herrenknecht to design and fabricate the largest TBM built for the UK, with an excavation diameter of 11.91m, to perform the tunnel construction. The single TBM is of earth pressure balance machine (EPB) type and is driving both of the link’s parallel tunnels.
In September 2022, the machine was launched from the Silvertown site, on the north side of the river, to excavate the Southbound drive towards the Greenwich peninsula. Upon arrival at Greenwich, the TBM would be rotated using nitrogen cushions, sliding it around inside the shaft into re-launch position. The machine would then excavate the northbound tunnel back to Silvertown, where it will be dismantled.
WHY USE THE PRESSURE RING FOR TBM LAUNCH?
The Pressure Ring System of launching allows the large TBM to be launched, twice, in compact shafts on either side of the river. The shafts were sized to the minimum because of neighbouring land use and the proximity of a disused lock entrance in Silvertown.
A more common rectangular shaft was dismissed at Silvertown site because of the complexity of design and execution. An alternative layout of a four intersecting shaft cells that would create the shape of a ‘peanut’ was chosen.
The excavation of a pre-tunnel or cavern was rejected due to the size of the adit and the shallow and difficult ground conditions.
Also, the system adopted would allow the reuse of the same launching structures and, in Silvertown, achieve the learning curve for the re-launch of the machine at the Greenwich rotation chamber.
One of the main challenges in the project was the assembly and launch of the TBM in such a confined shaft. Not only was the TBM to be fitted in but also all the logistics and muck conveying. This limitation with the space drove the study and design of a launching structure that was literally constructed around the shield thus using all the available space within the chamber.
Although this type of launch was developed and used in previous projects, our challenge was complicated by the size of the tunnel, its gradient at 4.12% downhill and the constraints of the assembly area.
I was tasked with the detailed planning and execution of the launch in the Silvertown shaft. This paper presents the solution implemented by the Riverlinx CJV to launch the UK’s largest TBM.
The decision taken for the method to launch the machine is driven by different factors that converged into one special approach for the TBM to start the excavation.
COORDINATION AND INTERFACE WITH CONSTRUCTION OF THE LAUNCH CHAMBER
The interface between the tunnel and the civil departments was crucial. A detailed sequence to assemble the machine had to be studied in depth to minimise any unforeseen collision or interference that might occur during the assembly of the shield, the backup gantries as well as the launching structures and the other auxiliary equipment.
With such a small shaft it was essential that the shaft base and walls were designed and constructed to allow the sequence of the TBM launch.
The base slab also had to be designed and constructed to support the weight and size of the TBM and gantries. This was complicated by the gradient and need for all temporary upright structures and concrete headwall to be inclined so as to be square to the gradient.
LAUNCHING STRUCTURES
Shutter Can
This circular structure was embedded into the concrete headwall and acted as a support for the subsequent structures to be bolted to Due to the size of the portal, the headwall had to be constructed in several lifts, leaving a void inside of the structure to place the cutterhead.
It was very important to survey the installation of this first structure to understand the clearance left between this ring and the TBM Shield. The scanner showed that an average of 180mm gap around the shield was achieved after the installation of the ‘shutter can’ with the TBM shield sitting on the cradle.
Seal Can
Another circular structure is the ‘seal can’, which was assembled on surface and lifted down the shaft as one piece to be bolted onto the exposed flanged connection of the embedded ‘shutter can’.
The special feature of this ring is the installation of two internal rows of rubber that would act as a seal to isolate the TBM shield and the tunnel from any potential ingress of soil and water that could happen during the excavation through the headwall and into the ground. Thanks to the rubber, we could also maintain the earth pressure in front of the cutterhead as required by the designer during the first stages of the excavation.
Thrust Frame
The ‘thrust frame’ is the structure that receives the loads imposed during the TBM launch and transfers them to the ground. The structure was composed of a main frame where the last item of the launching structures, the ‘steel ring’ (described below), was tied to, and also several props that would strengthen the structure and help transmit the loads to the slab.
Because the dimensions of its main frame, the structure had to be assembled on the ground surface and kept in the right position, level, and dimensions as per the base plates cast previously into the concrete at the bottom of the shaft.
All the castin elements were fixed to the slab using steel bars M47mm OD (Outer Diameter) 950/1050 that had to be meticulously positioned by the surveying team, so that all connections matched when placing the base of the structure.
The total weight of the thrust frame structure was about 90tonne and, therefore, the positioning was a big challenge for the team due to the small clearances within the walls of the shaft.
In fact, the shaft wall at the pit bottom had to be slightly broken out to place the base plates of the main columns.
Once the thrust frame was installed and checked for the correct inclination, the auxiliary props and rear columns were connected to it.
Steel Ring
The ‘steel ring’ is big circular structure of almost 100tonne weight.
The ring is designed to transfer the external loads applied by four groups of cylinders to the TBM thrust cylinders. These external loads were transmitted to the ‘thrust frame’ by means of connecting steel bars in tension.
The steel structure also had to be assembled on the surface due to its size.
The ring was composed of four circular sectors although it was not completely symmetric for two reasons: differences in thrust forces and launching arrangements, respectively.
- The working conditions given to the designer of the structures were based on the weight of the TBM as well as the inclined position of the machine and the excavation parameters. Therefore, the force to be applied was expected to be higher on the bottom sectors rather than the top ones. In addition, the bottom half of the ring had to be more reinforced to transmit this higher thrust to the shield.
- Due to the sequence studied for the launching as well as the design of the TBM backup gantries, the bottom sector had an opening, losing part of the structure to allow the installation of the gantry cradles required for the bogies of the gantries to run over them as the machine advanced.
The steel ring had to be located right behind the rear shield/tail skin and bolted to the first concrete ring of the tunnel, previously erected inside the TBM shield. The load applied by the external cylinders would transfer all the way through the steel and concrete rings and ultimately to the cylinders of the TBM thrust system to propel the machine forward.
Again, because the constraints within the shaft the sequence followed pushing the shield approx 2.5m towards the headwall. This left the cutterhead inside the shutter can and room for the crane to lift down the steel ring, landing it over the same launching rails of the TBM shield so this could slide until the final position.
With the steel ring in position the rest of the TBM continued to be assembled, lifting down the gantries as well as the remaining conveyor structures needed to excavate so the final tests and commissioning of the auxiliary elements and equipment could happen before the TBM was launched.
As previously mentioned, the steel bars used to connect the thrust frame and the steel ring were installed as the rest of the components were being prepared for the start of the excavation.
Steel Tension Bars, Hydraulic Equipment
As the machine was being prepared, the last items of the launching structure were in the process of being installed.
Steel Tension Bars
Four sets of bars were installed across the launching structures to transmit the load from the steel ring to the thrust frame.
They were designed as two sets of 5x47mm OD grade 950/1050 bars installed to each of the top sectors and two sets of larger 6x65mm OD gr 950/1050 installed to each of the bottom sectors. The reason for having a greater number of larger bars is linked to the force that had to be applied at the bottom of the TBM Shield.
To cover the distance during the launching, those bars had to be extended up to 18m that is, two lengths – 6m and 12m per bar – were connected with a coupler.
Hydraulic Equipment
To create the required thrust force, one hydraulic cylinder per bar was bolted to the back of the thrust ring to provide a load of 570tonne on each top quadrant and 1,350tonne on each of the bottom quadrants.
The cylinders were connected to a hydraulic power unit so the jacking forces could be limited to 70% of the bar yields by setting a pressure limiting valve.
TBM LAUNCHING
The solution of the Pressure Ring to put in motion the TBM was nearly completed by the end of August 2022, and so it was then time to brief the crew for familiarisation and operation of the system in place.
Power ON, operating cycle, and rates
The Pressure Ring system works in a similar way to how the TBM advances during the normal operation, where the cycle is limited by the elongation of the thrust cylinders.
In the launch case of the system, the maximum stroke of the external cylinders was 200mm thus, once reached, a new cycle would commence.
The cycles of the Pressure Ring System are explained in the sequence below:
- Rear lock nut tightens against the cylinder rod
- Start-up of the hydraulic power pack. Actuate the cylinders so the force is transmitted to the shield. TBM ready to bore. Commence excavation
- Extension of the hydraulic cylinder as the TBM excavates until the end of the stroke of the cylinder
- Tighten the inner lock nut inside of the stool to maintain the tension. Once all the nuts are locked the cylinder rod is retracted to initiate a new advance
- Repeat from point 1
During the operation, the total advance, movement, and the forces applied were monitored on the four different sectors to get a better control of the machine and the excavation.
An elongation sensor was installed on each sector so the squareness of the steel ring could be controlled and to correct any discrepancy created across them. Also, another elongation sensor was installed on one of the cylinders at each sector to read the extension of the cylinder on each cycle.
A pressure transmitter indicated the force created by the cylinders showing both the TBM operator and the jacking system operators the total force applied by the cylinders on each of the sectors so it could be adjusted trying to avoid eccentricity between the left and right sectors.
As stated, due to safety and design reasons, the load applied by the cylinders was limited by a pressure reducing valve avoiding over-tensioning the bars and the structure, which was designed to work up to a maximum total force of 37,500kN.
Originally, it was estimated that the operation would last approximately nine days. This allowed time for the extension of the gantry rails and the structure of the conveyor as the machine advanced from the Silvertown launch shaft.. The extension of those happened at every 5 cycles of the cylinder strokes. The team achieved the goal in seven days.
The table shows two cases with the estimated duration of a cycle when the cylinders are operating at the maximum and minimum speed.
Of course, the rates shown could not always be achievable for various reasons. For instance, not all the cylinders extended at the same speed so to keep the squareness of the ring, the cycle had to be stopped when the first set of cylinders was about to reach the maximum elongation. On the other hand, there were planned downtimes related to assembly of the gantry rails and the conveyor structure that also affected the advance rate during the operation.
Parameters for the excavation
The team decided to split the advance of the TBM into three different scenarios based on the different geological conditions the machine was going to encounter during the launch:
Stage 1
The TBM had to excavate through a full face 50MPa concrete headwall of 3000mm thickness, then GFRP secant piles of 2000mm-diameter. An earth pressure of 0bar in front of the working chamber was maintained keeping the chamber completely empty so the concrete could not harden inside.
The maximum thrust force expected was between 2000kN and 9000kN. At this phase, the force varied depending upon the machine location, whether just upon the cradle or mining through the concrete headwall. Thrust force was calculated by assessing friction among other factors, such as cutting force or the towing of the back-up gantries. The least favourable combination of individual resistance was decisive thus a safety factor was included within the assessment.w
Stage 2
The TBM had to excavate to 2000mm into soft concrete/ mixed ground at 0.5bar earth pressure in front of the cutterhead.
The maximum thrust force expected was between 8500kN and 18,800kN.
Stage 3
The TBM had to excavate to 4000mm into mixed face London Clay and 10MPa soil mixed piles at 1bar earth pressure in front of the cutterhead.
The maximum thrust force expected was between 14,000kN and 24,300kN.
Incidents during the launching
As the earth pressure increased, the force required to overcome the resistance climbed to a peak of about 24,000kN. At this peak force, a loud noise was heard in the vicinity of the bottom of the steel ring that put on hold the operation.
Investigation followed for possible damages to the structure. It was found that the bottom joint of the steel ring had buckled, several bolts snapped and the inner stiffeners near the joint had slightly deformed.
The team called for an emergency meeting with the designer to propose a quick and feasible solution that would permit the operation to continue. The solution was to reinforce the joint by adding stiffeners to different locations, so the bolted connection was better reinforced to carry on with the excavation.
Further investigations followed, and a better detail for this connection would be designed and prepared for the relaunch at Greenwich.
After the remedial works agreed with the designer were completed at Silvertown, the launch restarted. More attention was paid to any big difference between the forces so that almost no eccentricity could result, and the relative movement was better balanced.
As the TBM continued to the last strokes, the trailing edge of the concrete ring passed through the in-bye rubber seal inside the seal can. It was spotted that some water came inside the TBM. This had found a pathway through the gasket of the key segment that was not fully in contact with the leading edge of the steel ring.
An expansive foam that reacted with humidity had to be injected to stop the ingress of water into the tunnel. With the pathway sealed the TBM could continue the last metre to reach its final position.
Seven days after the TBM started its journey at Silvertown, the launch of the machine using the Pressure Ring System was successfully completed.
Upon completion of the launch, the machine continued excavating using its own thrust system until the gantries were introduced inside the tunnel, leaving enough space behind to assemble the rest of the machine and recommence excavation with the final TBM configuration.
The tension between the thrust frame and the steel ring was kept by tightening the inner lock nuts, so the structure could receive part of the reaction of the TBM during the excavation. Then, the external cylinders were decommissioned. This configuration had to be in place during the installation of the following 45 concrete rings.
DISASSEMBLY OF THE PRESSURE RING LAUNCH SYSTEM
The configuration of the structures left after the launch had to be maintained to provide enough support so the thrust of the TBM could be absorbed.
The theoretical factored friction between the concrete rings installed within the tunnel and the ground was not sufficient to counteract the force from the TBM rams.
This force was increased as the machine advanced due to an increase of the production rate thus, speed of the machine, and the rise of the earth pressure that had to be achieved in the working chamber of the TBM to minimise any ground movement or settlement.
The team had to find a balance between the expected total thrust force and the friction created in the tunnel for the TBM to stop in a safe haven for intervention.
To do this, it was agreed to stop 45 rings built after the launch; at this location it was expected a maximum total thrust force of 45,000kN that had to be opposed by a total friction in the tunnel of about 100,000kN plus 37,500kN that the launching structure could be loaded. This would also unlock the use of the maximum thrust force that the TBM could apply.
This short length of tunnel was sufficient to bury the whole machine in the tunnel and leave the shaft clear to commence the dismantling of the launching structures and prepare the bottom of the shaft to improve the logistics for a rapid construction of the tunnel. The disassembly of the structures was not without challenges.
By this time, the crane capacity was reduced on site therefore the methodology for dismantling the structures had to be restudied. The different structures had to be split in a sequence, so they were not damaged as they were to be reinstated at Greenwich for the ‘relaunch’ after the rotation of the TBM.
The most complicated part was dealing with the tension created between the thrust frame and the steel ring. To relax that tension from the bars, the cylinders had to be loaded back so the inner lock nut of the stool could be untightened. Then, the pressure in the cylinder was released in a controlled manner letting free the bar and ultimately the structure.
During the operation, we understood that a single bar could not be released, one at a time, but it was necessary to release the whole sector of the ring together, otherwise the tension would be transferred to the adjacent bars and risk overload.
With the cylinders and bars dismantled, the removal of the structures followed.
The size of the crane required all the structures had to be divided into smaller pieces. The sequence started by pushing out the steel ring so the top half of the ring could be unbolted from the bottom half and then lifted. For stability purposes, the bottom quarters were temporarily supported to remove them in singular lifts. After the steel ring was cleared out of the shaft, the disassembly of the remaining structures continued with no major issues leaving exposed the trailing edge of the first concrete ring built and the entrance of the tunnel.
LESSONS LEARNT FOR THE TBM RELAUNCH
I have reviewed the various options that considered to improve the different stages of the operation for the relaunch of the TBM in Greenwich on the Northbound drive once the machine is rotated.
One of them is to resolve the redesign of some of steel structures based on the new conditions for the launch of the TBM and the shaft that is being constructed at the Greenwich site.
The main change would be to the steel ring to avoid any failure during the launch. The ring, therefore, has been stiffened in all the four bolted connections. Also, the opening created at the bottom of the ring has been filled with a new steel panel to give complete continuity around the ring which will also reinforce the bottom bolted joint.
With the methodology of the relaunch changed, the installation of the gantry cradles installed at Silvertown would not be required at the rotation chamber. Instead, the same cradle that would be used to transport the gantries around the shaft will also be used to work as a supporting structure for them to run over up to the entrance of the tunnel.
From my experience during the assembly of the launching structures, I understood the potential clashes we might encounter during the installation at Greenwich for the re-launch.
Our enhanced understanding of the lifting operations and how the equipment should be set up to manoeuvre the pieces in a controlled manner would improve the operation to erect the structures.
Despite that the excavation during the launch at Silvertown was well performed, even with the programmed stoppages and unplanned downtimes, the plan includes improvement to communications between the TBM driver and the cylinder operatives. Better control of the operation is implemented by:
- A more intuitive display to show the thrust and elongation parameters on a single screen
- Data collection to be saved every 5 seconds to allow a better understanding of the operation and study any circumstance that can happen
- The elongation sensors being changed to a laser instead to measure the distance advanced and the relative position of the steel ring at each of the sectors. This should minimise any deviation with the mechanical sensor and so enable the squareness of the steel ring to be maintained
CONCLUSION
The methodology to launch the TBM can cause significant delays and costs to the project if the decision taken is not well assessed. It is clearly understood that the right decision was adopted by the joint venture upon the conditions studied at the early stages of the project, not only for the launch at the Silvertown site but the possibility of giving a second opportunity to the structures at Greenwich for the relaunch.
The solution implemented using the Pressure Ring System was successfully achieved for the completion of the Southbound tunnel drive, but the team had to examine in detail the design and very complex sequences to minimise and eliminate risks involved during the operation at all levels.
ACKNOWLEDGEMENT
I would like to thank to the Riverlinx CJV, the tunnel team, the designer OTB Engineering ltd, PHL Hydraulics, and Herrenknecht AG for their effort and determination invested during the design process and for the understanding of the required solution.
A special thanks to the people involved during the execution that ultimately made this launch possible. It has been an honour to be involved in this complicated and exciting part of the project and to have this opportunity to acknowledge the team as well as all the people involved on the launching process from the design concept to the performance of the works.
