The challenge of monitoring Victorian earthworks
WHEN the Rail Accident Investigation Branch looked in 2014 at a series of landslips on Network Rail lines, it found that they were difficult to predict. Following these, it investigated a derailment and collision at Watford on September 16 2016.
In the Watford accident, London Midland’s 0619 Milton Keynes Central-London Euston ran at 66mph into a landslide on the Up Slow line, by the northern portal of Watford Tunnel. The leading coach derailed, and the train stopped
380 metres into the tunnel, partially obstructing the Down
Slow line.
The driver transmitted an emergency stop message via GSMR, which the northbound 0634 Euston-Birmingham’s driver heard. Despite braking, his train struck the derailed one, tearing off the right-hand cab doors of both trains.
The landslip occurred after prolonged intense rain that had started at around 0300 (50mm of rain fell in four hours, around one month’s worth for a September), following several days of hot, dry weather.
RAIB recommended that NR review and improve its process for identifying sites that have a localised water concentration feature which can channel large amounts of water onto the railway.
Flooding beyond the railway’s boundary on June 13 2019 trapped an East Midlands Trains HST. It did not derail, but after it stopped more material washed onto the tracks near Corby, which meant it could not move ( RAIL 907).
RAIB’s wider and earlier report looked at six landslips that had taken place in 2012 and 2013, at Loch Treig (near Tulloch) on June 28 2012, Falls of Cruachan (on the line to Oban) on July 18 2012, Rosyth (near Edinburgh) on July 18 2012, St Bees (Cumbria) on August 30 2012, Bargoed (South Wales) on January 30 2013, and Hatfield Colliery (South Yorkshire) on February 11 2013.
At Loch Treig, GBRf 66734, which was hauling 24 wagons, derailed. The material that had slipped came from 35 metres up the slope from NR’s boundary. A passenger train had passed by without incident around 43 minutes earlier, but shortly before the derailment the area is thought to have had 6mm of rainfall in a half-hour period.
At Falls of Cruachan, ScotRail 156478 and 156453 ran into landslip debris but did not derail. They were carrying 70 passengers. A 30mph temporary speed restriction slowed trains from the usual 45mph because NR was concerned about the cutting side. A blocked culvert under a road further up the slope led to water diverting from its usual stream, and this water dislodged material onto the track.
In Rosyth’s incident, ScotRail 158739 hit debris and derailed at around 45mph, stopping 160 metres beyond the debris. Around 7mm of rain had fallen in the hour before the accident, with 40mm over the previous 48 hours. RAIB said it had no evidence that NR should have recognised that the gentle slope posed a risk and noted that a train running the other way 20 minutes before the accident had reported nothing amiss.
At St Bees, there were four slips. Northern 153317 led a classmate with the 0600 Maryport-Lancaster through the first but derailed its leading bogie on the second. The Met Office estimated that nearby St Bees had received 52mm of rainfall in the 24-hour period of the accident.
A contractor had started work to stabilise the slope at Bargoed the day before a slip displaced a tree into the path of Arriva Trains Wales 150258. A nearby slip in July 2012 had prompted NR to act. Rainfall reached 125mm in the previous five days. RAIB noted that a pipe under a road higher up the slope was feeding water into the area of the slip.
There are no natural slopes around Hatfield, but a slip in a colliery spoil heap outside NR’s boundary resulted in tracks being pushed up and sideways, forcing NR to shut the line.
In all, RAIB suggested that there are several factors that make landslips difficult to predict. They include:
■ Natural weathering processes
that gradually weaken the ground and make it unstable.
■Water makes slopes unstable, but rainfall and its collection in slopes cannot be accurately predicted in quantity or rate.
■Vegetation can change, and this influences the amount of water accumulating within a slope and the strength of the slope (roots can strengthen a slope).
■Changes in land use around the railway can alter how much water reaches the railway and how quickly it does so.
■Many railway cuttings and embankments were built with steeper slopes (therefore a greater likelihood of failure) than would be done today. This makes them more likely to slip.
■Existing drains are often poorer than modern designs and are not always reliable.
RAIB notes that landslips are almost always triggered by relatively high rainfall because:
■Water pressure tends to push soil particles apart, which reduces the friction between them and their ability to hold shape.
■Water flowing through the ground can wash out soil particles.
■Water can increase the likelihood of a landslip by making parts of a slope heavier than in dry conditions.