This article covers the work carried out last year on an embankment at Harlesden on the southern part of the West Coast Main Line (WCML). Rail Engineer met up with Luke Swain, principal route engineer (geotech) at Network Rail to learn more about the project.
The WCML route is extremely busy and site access at this location is also very constrained. While essentially a geotechnical ‘earthworks’ project with a deep-seated rotational failure trapped between two railway lines, the work required involved every engineering discipline.
Connecting London and Glasgow, with junctions and lines to the major cities of Birmingham, Liverpool, and Manchester, the WCML is the most significant railway corridor in Great Britain. It is also one of the busiest railway routes in Europe, carrying 40% of all UK rail traffic and made up of a mixture of inter-city, regional, commuter, and rail freight traffic.
The southern part of the line was constructed by the London and Birmingham Railway (L&BR) and opened in 1838. The line was engineered by no less than Robert Stephenson. His assistant engineer was Peter Lecount, who produced a number of comparisons to demonstrate the size of the line’s construction. He suggested that the Great Pyramid of Giza required the lifting of 445,500,000 m3 of stone and 100,000 men working for 20 years, and the railway 710,000,000 m3 of material – similar to the weight of stone used in the pyramid – and 20,000 men for five years. While some machinery would have been used to construct the railway, the vast majority of the work would have been undertaken by manual labour.
Rail Engineer has made the point many times that railway earthworks constructed in the 19th century do not comply with modern geotechnic standards. Earthwork construction when the railways were built was undertaken on huge scale in a relatively short time frame, which has left a legacy of over-steep embankments and cuttings, with many made using questionable construction techniques. This makes the asset stewardship of geotechnical earthworks very challenging, especially on a railway which is busier than ever. This was the case with Harlesden.
The problem
Luke explained that the first indication of a problem was when the track maintenance team raised concerns with the deterioration rate of the mainline track quality, at a location 6.5 miles north of London Euston. The line speed of the WCML at this point is 125mph. A desk study quickly established a basic ground model. In simple terms it was determined that the geotechnical asset showed all the characteristics of a 150-year-old reworked London Clay embankment. This type of embankment is problematic for a modern railway, due to its susceptibility to seasonal shrink swell.
Shrink swell is the volume change that occurs as a result of changes in the moisture content of clay-rich soils. Swelling pressures can cause heave, while shrinkage can cause settlement or subsidence. This shrink swell behaviour is the most damaging geohazard in Britain today, costing the economy millions and made worse by the drought and wet spells which are now common.
Prior to a site inspection it was considered that the type of embankment would be a significant contributory factor to the track issues reported. What awaited on site was much more complex, said Luke. The track quality of the line, which is supported by the embankment before it crosses over an intersection bridge, showed deterioration and loss of ballast. Ballast being critical for maintaining track geometry. When a track quality issue arises repeatedly, geotechnical engineering is likely to be a cause and during the site inspection the initial desk top ground model was verified and refined.
The presence of inclinometers on site suggested ground problems had occurred in the area before. The reworked London Clay hypothesis was most likely correct. At the top of the embankment, a loss of ballast shoulder was observed in conjunction with a reduction in the crest level of the embankment. A brick-lined under track cable crossing a manhole also indicated problems. Beneath this was a wedge of shallow angled slope, with inclinometers before the coping stones at the top of a wingwall of the intersection bridge.
It was clear that the crest of the embankment had been reducing in height over time. At the base of the wall, which is between the WCML and a highly used goods line, were some interesting features from a ground engineering perspective. An overhead line structure was leaning significantly towards the wingwall. A signal structure, which appeared to have recently been replaced, was also leaning. The wingwall had been stitched in multiple locations and the drainage system below ground level had become sheared. Finally there was a telecoms cabinet which had rotated on its foundation. In a short space of time the ground model and hazard log was updated, and the risk profile changed from a serviceability to a safety of the railway issue.
There is an asset stewardship system issue here worthy of noting. It is possible that the overhead line, signalling, and telecoms asset stewards and designers when faced with the ground issues such as Harlesden, may individually go to the trouble and expense of providing enhanced foundations for their assets so that they do not move. But having multiple indications of a problem to the geotechnical asset steward enables the root cause of the problem to be identified.
Plan of action
It was evident that a deep-seated landslip was present which was rotating beneath the wingwall and subjecting many of the infrastructure assets to foundation failure. A rapidly moving landslip would require speed restrictions on the affected railway lines or, in the worst case, a line closure. A slower and more gradual movement would permit normal line speed to continue, but with an increased monitoring regime.
An increased monitoring regime was quickly established. Track monitoring frequency was also increased and a remote theodolite monitoring system provided. An emergency action plan was developed, which would be used in the event of a trigger breach. To obtain information on the failure surface and rate of movement, a new borehole with an inclinometer was established at the toe of the wall. The findings were conclusive. This borehole became BH01 and provided a useful dataset, with the data being very important for the next phase of works.
With many engineering disciplines at risk, stakeholder engagement was critical, Luke explained. It was important to ensure that the risk posed by the geotechnical feature was escalated appropriately, as the next few years of this asset’s life would be critical for railway safety and performance. The site constraints were obvious and most construction options were not practical. It would be a challenge to complete the design and construction, to balance the programme urgency with suitable railway access, and minimise the disruption to railway customers.
Network Rail designers produced a contingency plan, which included temporary construction options in the event of a rapid deterioration of the earthwork. The designs allow a relatively simple solution to allow railway recovery. These would have been used in the event of an increase in slope movement. They would have caused disruption, but would have also been quick to implement. Fortunately, the action plan was never required but played a significant part in allowing Network Rail to deliver a full engineering solution and keep disruption to a minimum during the works.
With the short-term risk to the operational railway controlled, the Harlesden Embankment renewal project was established, although the project development phase had already commenced in parallel with the risk mitigation works.
Project development
Optioneering began within Network Rail’s own internal design organisation. Many options were considered, however the movement being some three metres below ground level and the heavily constrained site limited the solutions available. The available options report provided a comprehensive desk study of the site, including its many constraints. As with many railway locations in London, the review of historic OS maps confirmed many changes to the railway and surrounding land since the construction of the embankment, adding to the complexity of the ground model.
There were multiple railway assets affected by the landslip, so stakeholder engagement was important throughout the design development. The stakeholders included engineers who specialise in the maintenance and management of a specific asset type, but due to the complexity of the geological failure mechanisms the project was led by geotechnical engineers.
The key areas for the designer to be concerned with were: (i) the deep-seated nature of the slip plane; (ii) the age and condition of the brick wing wall; (iii) the condition of key railway infrastructure; (iv) and the site constraints for construction equipment.
Ground engineering solutions were considered, including an embedded bored pile retaining wall, sheet pile wall, diaphragm wall, jet grouting, deep soil mixing, soil nails, anchors, and electrokinetic geosynthetic stabilisation. The advantages and disadvantages of each were explored and a recommendation identified.
After stakeholder review sessions were held, the option selected for the Approval in Principle (AiP) design was an embedded bored piled wall with associated gravity retaining wall, as this solution would provide the renewal of both the earthwork and structural assets, and provide long term asset resilience. In simple terms, a strong well anchored wall would be provided.
As the project developed, Murphy was appointed as the principal contractor to support the detailed design and construction. The other contractors included BAM Richies, Sheet Piling (UK), Ainscough, MJ Hughes Ltd, and InterRail. All complex projects require great collaboration between client, principal contractor, and designer; and the principal contractor brought in AtkinsRéalis, which had the cross-asset engineering experience to undertake the detailed designs.
Detailed design
The single option design was reviewed to ensure buildability and cost efficiency. The complexity of the construction and constraints of the site meant that shutting the railway for a disruptive period would be unavoidable. It took some time, but after much work involving Network Rail and its customers, the required railway access was agreed.
The construction sequence adopted, after a great deal of consideration and industry engagement, resulted in the WCML remaining open throughout the works. This was important as trains pass every three minutes. However, one of the two tracks on the Lower-Level Goods line would need to be shut for a number of weeks. This solution ensured minimal disruption to the railway and the works were aligned with other disruptive works in the London area.
To support the construction, temporary access to non-railway land was important. Multiple sites across the area from Wembley Central through to Willesden Junction were negotiated. The landowners may not have realised the importance of their involvement in this critical UK railway infrastructure project, but their cooperation with providing land was a significant factor for the project to succeed. The sites were used for material storage, site access, and plant positioning.
With the construction access confirmed, the detailed design work could be completed. The close collaboration of client, principal contractor and designer at this stage allowed the long-term solution to be turned into a constructable design. The window of delivery was small and if missed it would have had a long term negative operational impact to both passenger and freight customers.
The detailed design phase contributed to making cost efficiencies, while still maintaining Eurocode and Network Rail standards compliance. This reduced the budget required from £14 million to £9.9 million. The final design comprised an embedded 10-metre-long sheet pile wall along the toe of the embankment to intercept and stabilise the shear surface of the landslip. The design also provided a cable route and staging for the S&T equipment. The wingwall would be replaced with a pre-cast concrete wall, anchored to ensure long term stability. This solution making the original failing wall redundant.
As part of the project, the adjacent embankment would also be soil nailed to improve the resilience of the embankment and provide support against deteriorating soil parameters. In early April 2025, the project team – which since 2018 had included many companies and individuals – could begin the construction.
Construction phase
After years of development, work could now begin on site. The early works included temporary alterations to the electrification systems to enable changes to the normal timetable and to allow for the line closure beneath the embankment. Compound and material handling areas were established at strategic locations and the physical works commenced against a well-structured and tight programme.
Even with the closure, the working area was extremely tight and instrumentation and monitoring of adjacent assets throughout the area was critical to avoiding any unplanned disruption to railway customers. The construction monitoring was assessed against years of monitoring data. This was a benefit as it provided baseline readings with seasonal variations.
All the works went to programme and now one of the railway’s highest risk geotechnical assets has been stabilised for the future. With asset resilience increased, Luke said, the focus will move to other locations across the railway network, as a Victorian built railway offers a never-ending challenge. He added that the Harlesden project broke through any silo working and was a truly collaborative effort involving client, stakeholders, designer, contractors, customers, and third parties. It was delivered within budget, on time, and did not significantly disrupt customers services.
Luke concluded that geotechnical asset management involves making cost effective decisions to ensure a safe status is maintained. Identifying a risk is important, but managing the risk and correctly prioritising the required interventions is also essential.
Image credit: