In 1989, all eyes were on the SPAD question, the timing of the fatal accidents at Bellgrove and Purley ensuring their consideration in Anthony Hidden’s analysis of the multi-train collision at Clapham the previous year. Though it was right for British Rail to focus on the issue and move towards adopting an automatic train control system, there seemed to be another problem looming – lower down the risk ranking, but there all the same: permissive working. And in the February of that year, two staff members lost their lives because of errors around it.

Permissive working is a useful operational tool, which (where authorised) lets trains proceed (with caution) into a block section, signal section or dead-end platform road already occupied by another train. While ‘breaking’ the golden ‘one train, one section’ rule of railway safety, this can clearly increase capacity and aid efficiency. It’s just that when things go wrong, they do go wrong – usually on the freight side. Indeed, the permissive working of freight trains was in the causal chain of at least nine staff fatalities between 1968 and 1989. Many occurred on freight-only lines, but on 3 February 1984 a driver and his guard were killed when a Speedlink service was ‘called on’ into an occupied section at Wigan North Western and struck the rear of a container train whose tail lamp had gone out. Readers can find out more in a forthcoming issue of Rail, but in brief this led to the replacement of the traditional oil lamp with (essentially) the battery powered design we have today.

Not that it was all about tail lamps, as an accident on 27 February 1989 would demonstrate. It happened in the darkness outside Warrington Bank Quay station. Just before 21:30, the guard of Doncaster-bound 6E26 checked his consist in nearby Walton Old Yard. All being well, and the position light signal (PLS) being clear, the driver eased the train over the points and onto the Down Helsby line, bringing it to a stand at WN216 signal. Soon after, red became green, but before the driver could open up, he felt a severe bump. He quickly climbed down from the cab and went back to investigate. What he found filled him with horror: the Dover–Mossend service had struck the rear of his train, sending two wagons laden with lorries down the embankment, and badly damaging the cab of the locomotive. Inside, a 32-year-old rolling stock technician was already dead; the 56-year-old driver, a veteran railwayman of some 42 years’ service, died later in hospital.

The investigation recognised that the ‘Mossend’ had been signalled by a ground PLS (rather than one mounted beneath a main aspect), noted that there were ‘a number of lights and other features’ that could have distracted the driver, and raised concerns about tail lamp reliability. However, it also questioned whether permissive working was needed on the Down Helsby line at all, and called for a review of all lines where it was authorised. It wasn’t the first time this recommendation had been made, yet it would be an accident at Stafford in August 1990 that brought actual change. Here, an ECS formation was ‘called on’ to a platform occupied by a Manchester Piccadilly–Penzance and failed to stop in time. The collision killed the driver of the former and injured 36 people on the latter. Though the investigation suggested that the dead man might not have been in a fit state to drive, having worked 25 shifts in a row without a day off and attended a presentation ‘do’ the day before, it added that the passenger train had been given a proceed aspect, which might have been taken by the ECS driver as applying to his train.

The internal inquiry led to an enhancement of the Track Circuit Block regulations, specifically the addition of an instruction that, before signalling a train into an occupied platform line at a station where it’s not booked to call, the signaller must advise the driver of the circumstances verbally. The HSE report of 1994 recommended a different addition, similar to what later became Regulation 3.4.4.5: ‘Once you have signalled a second train into an occupied platform, you must wait until the second train has stopped in the platform before you can allow the first train to leave.’ However, when debated by industry the following year, this was felt to be too restrictive for general adoption.

But Stafford – and a number of other accidents during the 1990’s – also led to a 60 per cent reduction in places where permissive working was practiced, and to the introduction of ‘Huddersfield controls’, which prohibits two trains moving in a section simultaneously (that is, one departing while a second enters the platform). Where the signalling doesn’t enforce this, the signalling regulations require the signaller to do so him/herself.

All these changes helped reduce the risk associated with permissive working to a point where incidents have fallen since the 1990’s. Looking ahead, the greater approach speed control possible with ERTMS may allow more permissive moves to be undertaken, which may increase flexibility further by allowing more attachments to be made in stations, and so on. But while collisions may be less common, the incident at Norwich on 21 July 2013 shows that the inherent risk from putting two trains into one section remains. This means that, while drivers must remain vigilant, so must we as an industry.

Greg Morse is RSSB’s Operational Feedback Specialist. The views expressed in this article are his own. Greg’s book on railway accidents will be published by Shire later this year.
Follow Greg on Twitter: @GregMorseAuthor