The jubilee line first opened in 1979, and a section between Baker Street and Bond Street is lined with a 3,810mm expanded precast concrete lining (EPC), with a 3,850mm bolted cast iron lining section on either side.

An earlier version of this type of lining was used for the Victoria line in 1961. The 1970s version consists of 22 segments, with 5 of them forming the invert. A pair of wedge-shaped segments at knee level is used to expand the ring directly against the ground.

It has been used for a number of construction through London Clay for both London Underground passenger and service tunnels – successfully.

According to records it facilitated faster segment erection doing away with the need for bolting and grouting, and London Underground’s lead tunnel engineer Neel Goorvadoo tells Tunnels and Tunnelling that in all other cases it continues to perform well in self-supporting ground like London Clay.

The design alignment of the tunnel drives between Baker Street and Bond Street is such that the southbound is about 10m lower than the northbound rising to the same level at Bond Street platform.

"The geology of the ground shows that this places the tunnel in the Lambeth Group beds, a comparatively less competent soil than London Clay".

The 1974 as-built records shows that the soil to be highly variable, faulted and fissured and containing sand and voids, not uncommon for Lambeth Group beds, which may have been uplifted in this area.

The presence of the voids being key, considering the nature of an expanded lining that requires the ground to act uniformly upon it, to allow it to safely carry the intended load.

First contact
The ‘problem’ as Goorvadoo describes it, was first observed around 2006 when a tunnel inspection revealed localised spalling of the concrete segments with fragments found at track level. This was reported to the London Underground tunnels department who undertook a more comprehensive inspection.

"At first we thought maybe the concrete quality was poor, says Goorvadoo. "However, desk studies and research revealed the lower strength concrete was not the main factor. We then looked at the circularity of the tunnel using an optical surveying methodology, which revealed a non-elliptical profile particularly on the upper section.

The survey was extended to cover other EPC and cast iron tunnel section.

A detailed comparison, looking particularly at the joint eccentricity, revealed a strong correlation between the extent of the large joint eccentricities and the defects observed at this location. This was when we realized that joint rotation was a major contributory factor.

As this is an operational railway tunnel, there was a need to re-assess both the structural and service risks brought about by this manifestation. A trolley mounted laser scanning, using a similar principle to the original optical method, was deployed on a monthly basis to allow us to observe small changes in circularity.

"Additionally a simple method of drumminess testing was used on a monthly basis to locate potential spallings. These two observational approach allows us to stay ahead of the game.

In parallel, further investigations including finite element modelling, analytical assessment and full-scale testing of spare EPC segments broadened our knowledge of the problem. The latter in particular demonstrated post-failure pattern similar to the observed spalling in the tunnel.

A system of real-time monitoring was installed on rings manifesting significant joint eccentricities which acted as early warning system.

The cause of the problem
Firstly the poor ground conditions encountered affected the build quality resulting in over-dig and misalignment of the tunnel rings. It is highly likely that some of the rings acquired an eccentric shape right from the beginning.

Secondly, the effectiveness of grouting any voids at the construction stage in an EPC type ring is very limited; and it is the case that partially filled voids may well have existed from the beginning.

"Given time, a combination of these factors would have allowed the upper segments to rotate around joints as a response to differences between vertical and horizontal stresses as the pore water pressure changes. "Whilst having little effect on a bolted segmental lining, the implications for the EPC ring with convex-convex joint is significant", says Goorvadoo.

Changing requirements
"Thirdly, over the years, as the line underwent several track, train and signaling enhancements, the "piston" effects caused by faster-moving trains resulted in continuous drying of the ground through the permeable EPC lining.

"As the ground behind the lining became drier and more desiccated, soil-lining interface became more and more compromised.

"So, in the absence of sufficient ground resistance, the segments in the inherently eccentric ring rotate to adapt to the changing stresses. Gradually the location of the contact patches at the longitudinal joints changes accordingly. When these reach sufficiently near to the surface of the segment, the area of concrete carrying the load reduces.

"More load on a smaller area of course creates a higher pressure exerted, which was then leading to the delamination that was detected by the druminess test and subsequently spalling"

London underground engineers take immediate action
A denser real-time monitoring in the form LVDTs (linear variable displacement transducers) was implemented which allowed us to monitor joint movements at an improved scale, as an early warning system. "We went back to first principles trying to determine what would happen if several movements occurred on the same joints or movements occurred on several joints on the same ring.

"We were able to model a sequence of joint movements that could lead to a structural failure. We formulated a set of trigger levels from the results of inspection, druminess testing and realtime monitoring.

When movements reached agreed threshold levels, it was time for physical intervention. Over the years, this took several forms. The first development in 2009, referred to as Heavy Duty strap, consisted of arc steel segments bolted together to form an inner lining that was jacked from the end of a steel beam in the invert to keep the problematic ring in shape and helping it to carry the ground load.

This solution was constraint by the size of the steel members having to be shallow enough to avoid infringing the kinematic envelope, the virtual shape of a moving train.

A total of 19 rings were strapped over 3 years.

"Although original measurements indicated that the strapping halted the movements, earlier strapped rings were starting to show sudden movements.

On the positive side, we had now accumulated a larger population of good monitoring data to improve our analysis.

"We were confident that influencing joint rotation will improve the situation. So, in 2012, we designed a system of steel members that could change to the sudden rotation to a more "ductile" mode. This consisted of a system of curved steel ladder that were bolted to the EPC ring without pre-stressing. As they were lighter and thus much easier to install, these were installed on all the EPC rings.

"They were also designed to support the temporary works for the permanent relining phase."

There was significant costs involved with the continuous monitoring and interventions which would have become uneconomical over a longer period. There was a need for a "fit and forget" solution.

The development of the ideal "fit and forget" solution
After a comprehensive feasibility study, a decision was made to adopt a bolted segmental lining as a replacement. One of the major constraint that led to this decision was the need for the project not to impact on the operational service. Therefore the solution needed to be designed such that it could be implemented safely in engineering hours.

"This was a major challenge as this approach has not been used in a modern urban metro before.

"The dedicated design team was assembled that consisted of geotechnical, structural, tunnel and mechanical engineers".

Due to the uniqueness of the proposed design and corresponding risk profile, there was a need to gain confidence by setting up a trial.

There is a short section of tunnel connected to the disused Charing Cross Station that was lined in exactly the same way as the problem section, but in good London Clay. It was the perfect place to practice the procedure.

A first trial saw six rings replaced. The test section was in benign ground according to Goorvadoo, and so there were no real worries of geotechnical challenges for the engineers to contend with, but they wanted to see how easy it was to use the tools. A derrick consisting of a lifting frame and platform was the main launch pad for the works.

It made use of simple block and tackles lifting methodology. The works were labour intensive, the designed temporary works were problematic and working area very constraints. Nevertheless very useful lessons were learnt for design mark II. It was unanimously understood that a more mechanized system was required.

"Back to the drawing board, we used a wagon from the London Underground fleet and modified this as a platform for a self-contained unit consisting of a segment handling arm, cutter/breaker, lifting cranes and conveyor belt type trolley system"

"With our new Segment Handling Plant (SHP), things were improved significantly as it could be hauled to and from site by 2 locomotives."

This new plant and revised methodology needed to be tested, necessitating a further trial operation.

Second trial
"We returned to Charing Cross for a second trail which successfully improved the issues faced in the original trials. As with any new plant, we now had to resolve new issues, including the passage of this modified engineering train on the existing signaling system, emergency access planning and sequencing of activities to make efficient use of the short working window.

"The final trial was undertaken in early 2013, which allowed us to resolve all the know issues and provided much needed experience in using the plant.

"We demonstrated with confidence that at least one SGI segment can be safely installed per shift. After a diligent tender process, bearing in mind the uniqueness of the project, Specialist Engineering Services (SES) were contracted to undertake the work."

The job
The key is the removal of the hoop load says Goorvadoo.

"The EPC ring is carrying up to 800kN, so dismantling it required significant control, effort and support.

"The ring is first supported with substantial temporary works which make use of the previously installed ring restraints.

"A disc cutter attached to the manipulator arm is used to make cuts along predefined strategic line.

"It is imperative that the depth of cut is controlled, otherwise the disc blade could get jammed in place from the compressive forces in the EPC segment, which is still carrying significant forces.

The disc cutter is replaced by a breaker to break the concrete in steps along the cut lines until a portion segment is broken into two parts.

"At that point, there is no hoop load in the rings and the remaining upper segment are now supported by the temporary works consisting of 30mm solid steel build bars which is fixed to the ring restraints ahead and the fully bolted cast iron or SGI ring behind.

"Once the new SGI segment is in position, it is bolted to the previous completed rings.

"Normally the annulus would be grouted but this is not possible in an incomplete ring. There is an annulus at least 35mm deep that need to be grouted before the tunnel can be safely returned to service (note the limited time window).

Bullflex Bag
"The solution we came up with was the Bullflex bag used in tunnelling but modified to suit our needs. The woven sack shaped as large pillow was tailored to fit most of the central area of the SGI extrados, with an inlet attached to one of the grout holes, accessible from the intrados.

"The bag was securely fixed to the SGI prior to installation and once the segment was bolted, grout was pumped to pressure into the bag. This provided the ground support required in the partially relined ring.

"A system of screws and base plates is used to connect the SGI to the remaining EPC to reform the structural integrity of the ring. The remaining EPC segments are removed in similar way over a number of shifts.

The Bullflex bag allowed the teams to safely delay the back grouting process until a full ring with necessary sealing is completed and the interface between the new ring & the existing invert EPC segments are fully grouted.

Engineering supervision was provided on every shift to ensure that works went as planned and any issues resolved as quickly as possible through the office-based team.

Conclusion and final Thoughts
Goorvadoo says he is confident of the condition of the newly lined tunnel. An inspection is due to be undertaken at handover, but really all the problems have been removed. There will not be flaking from an SGI lining, and bolting will also mean that the tunnel acts as a cylinder and has its own strength. The grouting behind the lining also takes care of voids.

A permanent wireless monitoring array that can be checked from the platform has been put in place, but any movements seen so far have been negligible.

"This project is a good example of whole life asset management. It has also been rewarding to be able to deploy our skill set as engineers to solve this unique problem.

"The proof of the successful execution strategy is that passengers on the train were oblivious to the significant works being managed every night without any delays to the service"