HomeInfrastructureRolling Boulders

Rolling Boulders

Listen to this article

There have been several instances over the years where trains have hit debris that has fallen onto a track and caused a derailment.

In high risk areas, it is not uncommon for avalanche shelters to be built which are strong enough to withstand the force of anything that can roll down a hillside.

These can frequently be seen in the Alpine areas of Switzerland and Austria and some exist in the UK, the Cambrian line just south of Fairbourne being an example where rock falls from the cliffs of Friog have, in the past, caused trains to derail and plummet into the sea.

The Pass of Brander Accident

Scotland has its fair share of high risk spots and, on the 6 June 2010, a train from Glasgow to Oban hit a boulder in the Pass of Brander just west of Falls of Cruachan station (see the rail engineer issue 69 July 2010).

The front coach of the 2 car DMU derailed to the left and could have rolled down the embankment onto the A85 road and into Loch Awe if trees and vegetation had not prevented its passage.

The section was equipped with an automatic stone guard system consisting of 10 single strand ‘piano’ wires which, if any were broken, would replace to danger one or more of 17 semaphore ‘stone’ signals either side of the break.

Such was the perceived risk that this detection system had been installed in stages by the Caledonian Railway between 1893 and 1913. In 2010, however, the displaced boulder had tumbled from below this wire screen and thus was not detected.

The subsequent RAIB investigation produced recommendations for a better inspection, maintenance and recording regime including the production of new standards. No comment was made on the effectiveness of the detection system, nor whether an improved system should be investigated.

A New Detection Concept

To protect against any immediate repeat of the incident, major geological protection work was authorised and QTS was awarded a contract to clear vegetation, improve the drainage and fettle up the piano wire detection system.

QTS are a private company specialising in lineside management, the initials originally meaning quality tree surgeons, but now more realistically standing for Quality Technical Services.

All this represents good precautionary measures but Network Rail in Scotland believed that a more effective means of detection was possible. Could lessons be learnt from elsewhere?

It came to light that BT had used fibre optic cables to detect rock falls on to roads, so this concept was investigated.

The principle is based upon a fibre having a marginal change in its refractive index if a vibration occurs.

This can be detected by the injected light source being partially reflected at the vibration point, which in turn can be calculated by distance from the light source. Fibre optic cable faults are located in a similar manner, using an optical time domain reflectometer. In theory, the bigger the ‘thump’, the bigger the refraction change.

More ferreting yielded the information that the principle had been tried in Yorkshire to detect cable thieves since the removal of trough lids and any disturbance of cables would cause a fibre to ‘blip’. This trial was mostly successful and may well be taken further.

Accordingly, a contract has been let via BT to Fotech Solutions, a company based in Fleet, Hampshire, to undertake a ‘Proof of Concept’.

This involves setting up a trial site, establishing a testing methodology, rolling different sizes of boulders down a hill, measuring the disturbance on a controlled length of fibre optic cable and producing a set of test results with accompanying analysis.

The Trials Underway

Getting a suitable test site has not been easy but QTS came to the rescue by creating a small artificial length of railway at their Rench Farm depot site near Drumclog, half way between Hamilton and Kilmarnock.

A hillside has been constructed replicating the Pass of Brander and 30 boulders of varying sizes have been made from drums filled with concrete and coated with plastic.

The fibre optic cable layout is a laser light source and sensor located in an adjacent test hut connected firstly to a 1km reel of mono mode fibre, then secondly to a fibre in a 400 m length of fibre cable as used on the FTN (Fixed Telecommunications Network) project, thirdly to a 5km reel of fibre and lastly to a second 400m length of fibre cable. The fibre cable is run outside at the foot of the ballast shoulder on both sides of the test track.

The light source uses the standard 1540 nm wavelength common on optical transmission systems.

Coupled to this test length are the various measuring equipments with the resultant fibre performance being shown graphically by distance on a laptop computer. Fotech claim that a vibration can be detected to a distance of 2½ metres in a 10 km length.

An important element of the trial is to distinguish, within a reasonable degree of accuracy, the difference between normal vibrations as would be caused by a passing train, or even someone walking in close proximity, and a boulder rolling down the hillside and fouling the track, hence the importance of testing different size and weight boulders.

Another variable will be the height above rail level that the boulder commences its descent. Facilitated by QTS staff, Fotech are arranging test heights of 2, 5, 7 and 10 metres.

The smallest boulders can be lifted into place by hand but the big ones need a ‘grabber’ JCB perched on the hill top to place them in position. Once the go signal is given, the boulder careers down the hill towards the track; the smaller ones tend to ride over the cess, bounce off the rail and return to the cess; the bigger ones ride the cess and stop with a clang against the rail. Any train would hit these with considerable force.

A boulder was not observed to climb the rail into the 4 foot but this is obviously possible since it happened at the Pass of Brander. Back in the test hut, the rollings are observed by a strong peak on the fibre measurement graph. Since the fibre exists on both sides of the rail, two peaks are observed. These peaks can be captured and used to trigger an alarm.

Records are being kept of the exact position where the various boulders come to rest. Further tests will involve a road-railer vehicle being put on the track and moved up and down to compare this with the vibrations caused by the boulders

Next Steps

It looks like the ‘proof of concept’ trial will be successful. If so, the next stage will be to install 400m of fibre cable in the Falls of Cruachan area, including through the Pass of Brander, and connect this to onsite test equipment.

This will establish how well the system performs on a real railway and will evaluate the profiles of normal vibration likely to be encountered. Assuming this is able to pick out unusual events, then a third stage will be embarked on by extending the fibre cable to a 20km length through this whole section of the Oban route and terminating it in an RETB repeater site.

The associated equipment room will be provided with a line that can support ISDN or IP connections, thus enabling the alarms to be monitored from a central point.

Future Potential

Whilst it is acknowledged that much testing / proving has still to be done, this system could completely change the way the problems of falling boulders and other obstructions are managed. If successful, the ancient and difficult to maintain piano wire system will be abolished. It is frequently triggered by deer and walkers so few will miss it. Beyond that, the Network Rail geotechnical group will consider installing similar equipment at other high risk locations. With the roll out of the FTN, most rail routes now have a fibre optic cable installed trackside. A spare fibre within these cables can be utilised for the detection system obviating the need to run out a dedicated cable. Significant cost advantages will result.

The means by which and to where the alarm signals are sent and interpreted will need careful thought. Too many false alarms will quickly give the system a bad name. The opportunity for using the technology to detect cable theft has already been mentioned. It may however be some time before that full potential is realised.

Thanks are expressed to Ian Findlay, the senior project engineer at Network Rail, Phil Jones from QTS for managing the site visit and facilitating the boulder drops and Lindsay McInnes from Fotech for being the onsite brain and explaining the technology.

Clive Kessell
Clive Kessellhttp://therailengineer.com
SPECIALIST AREAS Signalling and telecommunications, traffic management, digital railway Clive Kessell joined British Rail as an Engineering Student in 1961 and graduated via a thin sandwich course in Electrical Engineering from City University, London. He has been involved in railway telecommunications and signalling for his whole working life. He made telecommunications his primary expertise and became responsible for the roll out of Cab Secure Radio and the National Radio Network during the 1970s. He became Telecommunications Engineer for the Southern Region in 1979 and for all of BR in 1984. Appointed Director, Engineering of BR Telecommunications in 1990, Clive moved to Racal in 1995 with privatisation and became Director, Engineering Services for Racal Fieldforce in 1999. He left mainstream employment in 2001 but still offers consultancy services to the rail industry through Centuria Comrail Ltd. Clive has also been heavily involved with various railway industry bodies. He was President of the Institution of Railway Signal Engineers (IRSE) in 1999/2000 and Chairman of the Railway Engineers Forum (REF) from 2003 to 2007. He continues as a member of the IRSE International Technical Committee and is also a Liveryman of the Worshipful Company of Information Technologists. A chartered engineer, Clive has presented many technical papers over the past 30 years and his wide experience has allowed him to write on a wide range of topics for Rail Engineer since 2007.


  1. It seems odd that the test “boulders” are of regular shape. I can understand this as a starting point but surely real boulders are of irregular shape and therefore travel on an unpredictable path, and possibly bounce if the terrain is quite rocky.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.