Operating a rail line through an active landslide complex presents a high risk but Network Rail hopes to gain EU funding to assist with installation of an innovative new monitoring system
Dealing with geotechnical failures is challenging in terms of budget and closures, as well as being vital to risk management, but predicting when they are going to happen is tricky. Nonetheless, Network Rail is working on forecasting ground movement on a Kent rail line in a bid to manage slope stability issues that have affected the route throughout its history.
The Folkestone to Dover rail line opened in 1844 but soon started to suffer from the effects of the landslide it passed through with one slope failure resulting in the line being closed for four years. Maintaining the rail route through Folkestone Warren over the last 170 years has resulted in numerous remedial projects and monitoring schemes, but EU-funding could now help add more certainty to managing the risks.
Network Rail is waiting to hear whether it has gained funding from the EU’s Life+Project match-funding initiative that could secure £1.7M for a new instrumentation system that will give real-time monitoring information.
“Network Rail is primarily a track operator and it needs to know whether it is safe to send a train onto a section of track or not,” says leading slope stability expert Eddie Bromhead, who is working as a special adviser to Network Rail. “Once the train has passed the last signal before the Warren it is too late to stop it if ground movement starts, so a system is needed to monitor movements and check trends in real time so that Network Rail has confidence that the line is safe for use.”
The line carries around 50 trains a day, including some of the high speed services, through the 1.7km long landslide complex.
While operating the trains, or not to operate the trains based on the risk of a landslide, is a challenge for Network Rail at Folkestone Warren, this is not the only issue at the site – walkers, homeowners and road users also come into the equation.
“The Warren is a Site of Special Scientific Interest and there are lots of footpaths through it so we have a duty of care and we use signage to warn people of the dangers and gates to ensure people stop and see the warnings,” says Network Rail route asset manager for civil engineering (Kent) Derek Butcher.
“We have lots of monitoring systems at the Warren already with extensometers in the drainage adits and cables attached to the
signal system with suction cups, so if there is a landslide or rocks hit the catch fence beside the track the signals turn to red. However, the current system would not prevent a train from hitting a landslide if it has already passed the last signal,” he adds.
“A real-time system would allow us to learn from the readings and plan our operation.”
Bromhead adds: “There has been a long history of slope problems at Folkestone Warren because the whole outcrop of the Gault at the Warren is a historic post-glacial landslide and the rail line runs straight through the middle of it.
“The railway line was built in the 1840s to connect the ports of Folkestone and Dover as although both towns had rail connections to London, they were operated by different companies and ships often docked at the wrong port leaving passengers with pre-booked tickets with no choice but to buy another ticket with the other company.
“Just before the rail line was built, the harbour at Folkestone was extended and this had a significant impact on the coastline at the foot of Folkestone Warren because it prevented longshore drift and depleted the beach levels.”
It is believed that the change in beach material may have triggered a new phase of instability in the landslide complex. Nonetheless, Bromhead believes that the landslide complex was triggered due to sea level changes. “There is a cliff type feature in the chalk that is known as the Horse’s Head, but it is not actually a natural block of chalk – it is a palaeosol formed from redeposited chalk resulting from a major slope failure that has been dated as having been formed 8,500 years ago. This is around the time when sea levels reached their current level which may have been the trigger for the landslide.”
From west to east the rail line through Folkestone Warren passes through three tunnels – the Martello Tunnel at the western end, the Abbots Cliff Tunnel in the middle and the Shakespeare Tunnel at the eastern end near Dover.
After the rail line was built there were some small slope failures, but larger movements occurred in 1876 and 1896 that both caused disruption to the tracks. The 1896 movement is recorded to have cracked the Martello Tunnel and part of it was demolished and the portal rebuilt further west to prevent a future landslide shearing the tunnel.
Then, in 1915, a major failure occurred that resulted in the line being closed for four years. “The movement occurred over the whole length of the Warren and several cliffs collapsed resulting in the chalk fluidising and burying the rail lines with up to 20m of debris and creating a flow 70m out to sea,” says Bromhead.
“Soldiers stationed in the signal box managed to halt a train at the mouth of the Martello Tunnel and, although the train came to a halt partly on the landslide, the passengers were able to walk back to Folkestone along the beach and no one was injured.”
During the four-year closure drainage adits were bored from beach level up through the complex, but further small failures continued to occurred periodically.
It is reported that soil mechanics specialist Karl von Terzaghi visited the site in 1939, but concluded that it wasn’t a landslide and the problems were caused by chalk dissolution.
Despite Terzaghi’s opinion, work focused on stabilising the landslide and more work on the drainage was carried out in the 1950s but there was more movement in the coastal margins of the complex afterward. It has been suggested that the drainage may have transferred the groundwater from the landslip into the coastal area.
Stabilisation work then focused on constructing sea defences including a concrete toe-weighting apron, coastal protection and adit improvement, but the area has continued to experience stability problems.
“In the 1980s the risk was managed by basic surface monitoring,” says Butcher. “We dabbled with electrolevels and tiltmeters but none of them proved to be long-term solutions. We still carry out expert walkovers once a tear, track monitoring and earthworks monitoring, but it is more of a watchingbrief.”
In the last 10 years £10M has been invested at Folkestone Warren with much of the money spent on maintaining and improving the rock revetment and the toe-weighting apron to minimise movement in the coastal margins. Network Rail came to the end of a 10-year management plan for the landslide complex last year and started to look at how the issue could be better managed or monitored and asked Bromhead for advice.
“The last 10-year strategy was all managed in-house but we wanted to move on from the watching brief we are currently using here,” says Butcher.
Network Rail was concerned that a failure could happen again at the site, particularly after movement in one area along the back scar under Old Dover Road and some residential properties was noted.
The land is outside of the Network Rail property boundary, but failure could be as catastrophic for the rail line as the 1915 landslide. One property on the edge of the new movement has already been rebuilt, underlining the scale of the movement.
Bromhead invited the Network Rail team to visit a site in Ancona, Italy, where a differential monitoring system (DMS) from Italy-based Centro Servizi di Geoingegneria (CSG) had been installed to monitor a rail line that passes through an active landslide where there are also a number of residential properties within the unstable area. Ancona’s DMS was installed after a major landslide occurred in the area in 1982 and uses strings of inclinometers, piezometers and robotic total stations linked to a weather station and a control room.
The strings of inclinometers allow the analysis of both horizontal and vertical displacements which enables engineers at Ancona to differentiate between ground movement caused by rapid, slow or deep-seated landslides.
The Ancona system uses three 100m deep boreholes that contain modular dynamic columns that each have 85 biaxial inclinometric modules, two peizometric sensors and 85 temperature sensors over an 85m active length. The biaxial inclinometric modules are linked by special flexible joints that allow the system to move as the borehole twists and bends with ground movement.
Information from each column is sent to the control room and continuously compared to threshold values with a warning text or call sent to the monitoring staff if these thresholds are exceeded. The same type of warning system also operates if ground water levels change rapidly. Within the control centre, software models the displacements visually to assist in analysis of the ground movements.
Installation of the DMS and initial monitoring helped to reassure the local authorities at Ancona that it was safe for residents to live within the landslide area and that there was sufficient data to provide them with an early warning of increased ground movements.
Network Rail was keen to implement a similar DMS from CSG at Folkestone Warren to be able to track trends in ground movement and learn to recognise patterns that could help predict when failures – both minor and major – were more likely to occur. Bromhead recognised that the scheme could offer insight into slope stability monitoring for other infrastructure operators and suggested trying to gain EU funding for the work under the Life+Project initiative.
The general objective of Life+Project is to contribute to the implementation, updating and development of EU environmental policy and legislation by co-financing pilot or demonstration projects with European added value.
Network Rail asset engineer for geotechnics (Kent) Evanthia Batzalexi worked with Bromhead to build links with the University of Florence, Ancona’s local authority, Italian rail operator RFI, the University of Ljubljana and the Geological Survey of Slovenia to make a bid for funding.
The Ground Movements and Railway Safety (G-MARS) team is currently waiting to hear if they have been successful but it is hoped that work could start on the installation phase before the end of this year in order to finish the work by summer 2014. However, implementation of the monitoring system is only the first stage of the work and G-MARS will also call for the findings of the operational side of the monitoring system to be disseminated to the wider industry.
Batzalexi hopes that, whatever the decision by the EU, a DMS will be installed at Folkestone Warren but she says failure to get the EU grant could delay the work. While the system would help Network Rail to live more comfortably within its landslide, gaining of the EU grant would ensure that more people than those passing through Folkestone Warren would benefit from the work.