An answer to landslip detection?
In the wake of the horrific crash in Scotland, PHILIP HAIGH looks at a solution that could act as an early warning system for landslides and other track problems
TO say that Carmont’s terrible derailment was a shock to those running Scotland’s Railway is an understatement.
Staff across ScotRail and Network Rail Scotland were busy - to borrow Scotland’s Railway’s tagline - building the best railway Scotland has ever had. And with recent electrification schemes delivering faster and greener journeys (with more wires planned), new trains, new stations and plans for new lines, they’d been making great progress. August 12 hit them hard.
Smoke curling lazily up from Carron Water’s deep and steep valley provided the first sign that something was wrong and prompted a member of the public to call Police Scotland.
At the same time, Nicola Whyte, a ScotRail conductor travelling on the derailed train to her next duty, was hurrying back to Carmont box to raise the alarm.
I salute her clear thinking. She can only have clambered out of the train’s wreckage in shock at the devastation around her. The train’s leading power car was down an embankment. Another coach nearby was ablaze. Two more were rolled upside down with the fourth balanced on top. Only the rear power car remained in line with the track, albeit derailed all wheels.
The line through Carmont is one of only a handful of Network Rail’s lines still signalled under ‘Absolute Block’ regulations.
Entirely safe from a signalling perspective, they permit only one train to enter a section of track between two signal boxes. A second cannot follow until the signaller sending it has received confirmation that the first has arrived at the next box. Once a train leaves a signaller’s immediate area, the signalling system does not know precisely where it is, only that it has left one box and has not yet arrived at another. It’s just over five miles between Carmont box and the next one north, Stonehaven.
Railway rules require a train’s crew to lay detonator protection 1¼ miles from their train in each direction, to warn and stop other trains approaching. With driver and guard incapacitated, it fell to Nicola to do this. She chose to go in the direction of the nearest signal box, Carmont, from which the train had departed just minutes before.
Fortunately, there was no train heading south. Equally fortunately, the derailed train - the 0638 Aberdeen-Glasgow Queen Street returning north - only had nine people on board, including the crew.
If COVID-19 can be praised for anything, it’s that so few people were on the 0638. Looking at the wreckage, and realising that three of those nine people died, it doesn’t bear thinking about the consequences of a full train or a southbound train colliding with the wreckage.
The cause of this wreck appears to be a landslide. Aerial photos show a small slip just back along the track from the mangled coaches.
The 0638 is not the first ScotRail train to hit a landslip, but it’s the first with such devastating results. Landslips often derail trains, but those trains generally stay upright and in line.
A rail accident report a few years ago showed a ScotRail Class 158 sitting down the track with mud and debris slumped over the track in the foreground. It was running at 45mph and stopped 160 metres beyond the slip.
That same report noted that no one had died in a train hitting a landslide for over 50 years in Britain. To emphasise that the risk still exists, it added that nine people died when an Italian train hit a landslide near Bolzano on April 12 2010.
By my reckoning, had the 0638 not hit the parapet of the bridge over Carron Water, it too might have stayed upright and in line with the track, even if it had reached its 75mph maximum speed for that stretch of line.
Just as with Heck in 2000, a small derailment was fatally magnified by other factors. In Heck’s case it was a set of trailing points that pushed the front of the train into the path of a coal train heading the other way. I visited that site the following day and the images of sheer devastation still haunt me today.
But back to the problems of landslips. They are not a new thing. The central plot of The
Railway Children, written by Edith Nesbit in 1905, involved the children successfully warning a train of a landslip so that it could stop in time.
On the real railway, there is a long screen of trip wires alongside the line at Pass of Brander (on the Oban branch of the West Highland Line). Should a rock fall down the bank, the wires are designed to break and place a semaphore signal arm at danger to alert any approaching train. This mechanical apparatus has sat sentinel since 1882.
What Network Rail needs today is the equivalent of the Pass of Brander’s system
“If fibre optics can hear passing trains, cable thieves or lineside work and detect their location, it seems likely they will be able to hear a landslide and raise the alarm.”
installed more widely along its tracks, to keep watch over likely landslip sites. Sending an alert to the controlling signal box, it could provide a way of stopping trains in emergency following a landslide.
NR already has a way of alerting trains. Indeed, the driver of the 0638 received just such an alert via his GSMR radio, telling him of flooding ahead earlier on his southbound journey.
The flood report came from another driver heading north, but it’s clear that using trains and drivers to find floods and landslides is not the best answer. NR knows this and last year published a ‘challenge statement’ with the rather clunky title Detection of Geotechnical Asset Failure by Means Other than Train Drivers or Lineside Staff.
The statement was a prompt to innovators and research and development teams. It notes: “Geotechnical failures are frequently reported by train drivers as bank slips or rough rides. This is too late for preventative measures to be put in place.”
It adds: “Most data capture requires an examiner to go on site. This is time-consuming and limits the frequency that condition data can be collected. Access to inspect assets is often difficult, especially when accessing through third-party land. Some third-party assets also pose a risk to the railway - for example, slopes outside our boundary and boulders.”
When you consider that NR has 190,000 earthwork sites to inspect, you can see the size of its challenge. It suggests that the answer might come from frequently scanning earthworks using LIDAR and comparing each run with previous ones to detect differences - perhaps where the toe of a cutting is starting to bulge.
There might be another answer, and that’s to use long fibre optic cables to listen for sounds of movement and detect real landslides in time to alert signallers and drivers. It might seem odd that something that transmits light can ‘hear’, but it seems that it can.
The Fibre Optic Sensing for Railways Association is a small group of European railway managers. It reports that one early use of fibre optics was to hear cable thieves at work and be able to determine their location. It can also hear trains passing, so provides another method of tracking where they are, and which DB has successfully tested to trigger passenger announcements at stations.
In another trial, a fibre optic cable could distinguish between tamping and grinding work. The association suggests that by merely listening, they can detect when track needs repairing or replacing.
If fibre optics can hear passing trains, cable thieves or lineside work and detect their location, it seems likely they will be able to hear a landslide and raise the alarm. What is particularly intriguing is that many tracks are already flanked by fibre optic cables. Perhaps they can be used to listen for problems while also doing their primary job?
NR’s challenge statement shows that it’s well aware of the problem. I hope that the terrible events at Carmont provide the boost in interest that researchers need to develop fibre optic sensing.
It wouldn’t be the first accident to trigger another major step forward in safety, but I hope that it’s one that contributes to Britain continuing to be the safest railway in Europe.