If the rules of safe working on the railway are ignored, the consequences can be horrific. Sim Harris looks back at the Margam tragedy last July, when two track workers were killed.

WORKING on the railway has always involved risks, and those who are most at risk are the staff who work at track level, particularly when trains are passing by.

In an ideal world railways would be closed while anyone was working on them, but this is plainly not always possible.

Instead, a logical system has evolved over almost 200 years of experience, some of it bitter.

One of the hardest lessons to learn was that communication is the keystone of safe working. If communication fails, lives can be lost.

This had became all too clear by the middle of the nineteenth century, and two incidents will be enough to demonstrate the point.

The first had an element of comedy about it. On 20 October 1852, a special Great Western train was making its way from London to Birmingham to mark the extension of GWR’s broad gauge to the West Midlands. It was meant to be a triumphant occasion: the ten coaches were filled with important guests, while the magnificent locomotive ‘Lord of the Isles’ was at the head of the train, with locomotive superintendent Daniel Gooch on the footplate.

As the special approached Aynho at 50mph (80km/h), it encountered a signal which had been installed to protect a siding. This siding was disused, and the signal had been left showing clear. These days, of course, such a signal would be marked out of use with an X, if not removed.

Its inadvertent message proved unfortunate. The crew did not know the road, and assumed it meant the line through Aynho was clear. Accordingly they put on steam, only to collide moments later with the late-running 09.27 ‘mixed’ train from Didcot to Banbury, which was stationary at the platform.

The stability of Brunel’s broad gauge was demonstrated convincingly that day, because neither train left the rails and six passengers on the ‘mixed’ complained only of slight injuries.

But had the special’s crew known the road, they would not have placed such faith in an irrelevant signal, and it would also have helped if they could have known that the line ahead was occupied.

The second incident had darker results, and a p-way team were at the heart of it. On 9 June 1865, they were renewing the wooden baulks which supported the track at Beult viaduct near Staplehurst. It was a modest structure, only three metres above a stream, and there was enough time between trains to replace each baulk in turn. The job was nearly done, but two lengths of rail still had to be replaced when a train arrived unexpectedly. There was a lookout who had been issued with two detonators, but it was a sunny day and he had been told not to place them unless it was foggy.

So it was that the Folkstone Boat Express became disastrously derailed at Beult, because its timings varied, depending on the tides. The team had been issued with two service timetable books, but one had been destroyed by a train, while the foreman misread his own copy. Ten passengers died and 49 were hurt in the five coaches which plunged into the stream below. One passenger in the leading vehicle, which was not one of the five, was no less a person than Charles Dickens. He was not physically injured, but the longer-term effects may have shortened his life.

Again, as at Aynho, communication had failed and the result was a disaster.

Working on the railway in former times could be dangerous indeed.

On 20 February 1899 MP John Burns told the House of Commons: ‘In 15 years, from 1872 to 1886, out of 1,407,000 troops liable to be engaged in battle, there were only 1,396 killed in action, and in that same period of 15 years there were 8,400 men killed on our railways … we cannot permit this wanton sacrifice of human life to go on.’

Unfortunately, in spite of Mr Burns’ remarks, the improvement over the next 20 years was only partial. In 1921 alone 288 railway staff were killed and 2,881 injured. Of those who died, 65 were p-way workers.

These days, we are shocked (rightly) by a single death on the track, and the tragic accident at Margam on 3 July last year was followed by rigorous inquiries.

Modern p-way teams still depend crucially on good communication. If some of them need to work close to a track, and even on it, they are mainly protected by lookouts. The aim is to give them an effective warning ten seconds before a train arrives, so that they can move clear.

One problem is the noise associated with some forms of p-way work, particularly if pneumatic drills or similar tools are involved, and it is believed that the three men at Margam – two of whom were killed – were all ‘almost certainly’ wearing ear defenders.

Technical solutions are feasible. For example, vibrating wrist bands could theoretically supplant an audible warning. These are often sold as stress-relievers rather than safety-critical equipment, although they may have some industrial applications (Amazon has patented a vibrating wrist band that is reported to be capable of guiding the hands of workers).

There are also various warning systems using sirens and, in some cases, flashing lights. One (TOWS) is operated by an approaching train, while another (LOWS) is triggered by a lookout.

Back at the worksite itself, there are special versions of hearing protection known as communication ear defenders, and these encase the ear and have speakers in the earpieces, using FM radio or Bluetooth to provide the communication link. Some also contain a microphone for two way communication, although whether a conversation should be encouraged while a train is approaching at speed is perhaps another matter.

Another variation of communication ear defenders is known as ‘push-to-speak’ and ‘push-to-listen’, and these require the wearer to push a button in order to hear or be heard. Again, because every second counts when a train is approaching, they are probably not suited to p-way use.

These defenders have one flaw: they depend on the equipment working properly. If it does not (perhaps because the radio or Bluetooth link is being corrupted by electronic interference from a source nearby) there is no fall-back. In other words, such equipment does not ‘fail safe’, as all good railway equipment should do.

A much older method goes back to basics. The distant lookout gives a warning using a horn, whistle or chequered flag, and this is heard or seen by a second lookout who is stationed by the workers on the track. As soon as the warning is given, this ‘local’ lookout quickly taps each worker on the shoulder to alert them.

Of course, if the local lookout is so close to the workers that they are within physical reach, the lookout might also need ear defenders if the sound level at the site is above 85 decibels, but the local lookout must ignore the work in progress and instead keep looking in the direction of the distant lookout to see any visual warning as soon as it is given.

The RAIB has said in its interim report that there were no formally appointed lookouts. All three workers were almost certainly wearing ear defenders, because one of them was using a noisy power tool, and all were concentrating on the job. None of them was aware that a train was approaching, until it was too late to move to a position of safety.

The report continues: ‘Working on an open line without a formally appointed lookout meant that no single individual stood apart from the work at the points with the sole responsibility of providing a warning when trains approached.

‘The absence of a lookout with no involvement in the work activity removed a vital safety barrier.’

It is too late to save two lives, but as on so many previous occasions we can at least hope that the final conclusions lead to a further improvement in railway safety.

This is a revised and updated version of a Hot Topic article first published in the August 2019 print edition of Railnews, RN270. Copies of the current or recent editions can be obtained by calling 01438 281200 from UK numbers or +44 (0)1438 281200 internationally, and selecting Option 2.