Signalling obstacles in the path of NR’s Digital Railway

Railways and technology go hand-in-hand. Switching from saturated to superheated boilers improved the efficiency of steam locomotives. Introducing electro-mechanical Automatic Warning System (AWS) improved safety. Tilting trains have allowed speeds to increase.

Philip Haigh Colas 66850 Watton at Stone ERTMS test section 091213 DSC_0309Colas 66850 hustles an infrastructure train through Watton at Stone on December 9 2013. It’s running along the stretch of line Network Rail uses as its ETCS test track. The string of red signals behind the train mimic the signalling being installed on the central section of Thameslink under London and shows that high-capacity signalling is not only possible with ETCS, although ETCS would not need the signals which helps cut maintenance and installation costs. Copyright: PHILIP HAIGH.

The pairing does not always work. Gas turbines never caught on and some technology was too advanced for its day – British Rail’s APT tilting train of the early 1980s comes to mind.

Signalling is one area in which technology has always played a major role. It linked communications systems with computers and incorporated safety features to minimise mistakes causing accidents. The Victorians linked their telegraph method of transmitting information about trains to the mechanical computers that sat under every signalbox and made sure that signals could only be cleared if points and trains were in particular positions. This application of logic is no different today then it was then, albeit it’s done by a few grams of silicon rather than tons of steel.

Network Rail now describes the future as ‘Digital Railway’. This overlooks British Rail’s work in pioneering solid-state interlocking (SSI), which it introduced in the 1980s. Interlocking is the logic that links points, signals and train locations while solid-state merely means that it’s based on solid semiconductors such as silicon chips, that is, it’s computerised, digital.

That said, NR’s ambitions will take the railway to a higher level. It should be easier to plan and implement timetables and it should be easier to deliver accurate and timely information to passengers when trains are delayed. Between now and that nirvana lies a long and expensive road. It relies on implementing systems that cannot be bought off-the-shelf today. No-one knows the cost and no-one knows the timescales. NR’s plan has seen various timescales – it was 50 years, then Mark Carne arrived as chief executive and pledged 2029, now it seems to be settling on 25 years. The 50-year was figure was based on installing Digital Railway signalling when current equipment reached the end of its life. This would give a patchwork with drivers switching from traditional to cab signalling, with a risk of confusion. Faster options would lead to current signalling being removed part-way through its life, which is more expensive. There is no perfect answer.

NR’s vision of the Digital Railway comprises European Train Control System ETCS) signalling (initially at Level 2 and then Level 3), GSM-R radio communications and a traffic management system (TMS). Put all together and they form the European Rail Traffic Management System (ERTMS). Trains on the Cambrian Coast already run under ETCS signalling, which tells drivers how far they can proceed via a screen in their cab, with information coming from a control centre via GSM-R radios. NR’s history of TMS has been more patchy. It pulled plans for widespread implementation and is instead trying a couple of testsites based around Cardiff and Romford. It’s TMS that provides a better ability to plan timetables in real-time and release accurate information following incidents that delay trains.

There’s another strand to NR’s plan that sits outside ERTMS. It’s another acronym, C-DAS, standing for Connected Driver Advisory System. It build on current DAS technology that advises drivers of the best speed to use to keep to their timetable. This can save fuel by promoting coasting when suitable and can reduce the number of red signals drivers encounter by ensuring they don’t run ahead of timetables. But DAS works on fixed timetables and can’t account for what other trains are doing.

C-DAS provides a link from signalling systems. It’s use is best shown by considering a junction busy with trains approaching from two lines to join one line. C-DAS can advise drivers on the best speed to ensure they arrive at the junction in sequence and can pass through it without stopping. It’s rather like car drivers adjusting their speed on a slip road to join a motorway without coming to a halt.

The prospects and pitfalls of all these changes has netted enough interest from the MPs on the Transport Select Committee for them to hold public hearings to quiz rail leaders. Mark Carne took command of the hearing on May 23, leaving committee chairman Louise Ellman almost a bystander. He pushed a strong case for Digital Railway although he wouldn’t be specific on costs, benefits or timings.

I’ve some sympathy for his reticence. NR was badly stung by revealing early costs for Great Western electrification that it couldn’t match as plans developed. Carne is determined not to fall into this trap again but he must also contend with Treasury funding rules now that demand accurate costs before money is released. Beyond admitting that it would be “a great deal of money” Carne said MPs would have to wait until the end of this year before NR would have a better idea.

He argued: “We spend about £1 billion a year renewing signalling systems. Over the next 25 years, if we don’t do anything we will still spend £25bn just renewing worn-out signalling systems. We believe that £25bn can be better spent transforming the whole signalling system and train control system.”

NR’s written evidence said that the annual figure spent on operating, maintaining and renewing signalling was “in excess of one billion pounds” which suggests that Carne might have been taking advantage of the MPs’ lack of knowledge.

This wasn’t the only time he left himself open to challenge. He later said: “At the moment, a lot of our tracks are one-way streets essentially because that’s the way the signalling system is set up. As soon as we move to digital train control, all of those tracks become two-way streets so that we can really run the network in a much more flexible way and a completely different kind of way.”

In itself, it’s true that ETCS cab signalling makes it easier to use a line in either direction. That’s because it doesn’t need a ‘light on stick’ signal to control movement onto and along that line. But it ignores the fact that if the railway is today as busy as NR claims, and tomorrow will be even busier, there’s very likely to be a train coming the other way along that track you wish to use. More bi-directional lines will help the railway recover from incidents but it does little for normal working and little for improved capacity. The flexibility Carne desires also needs points to switch trains from one track to another and any increase in them will need to be factored into Digital Railway’s case.

Carne did give some ground on one of NR’s more controversial claims. That’s the claim that Digital Railway will bring a 40% increase in capacity. Carne stood by the claim for dense commuter lines but admitted that DR wouldn’t deliver this on long-distance routes.

Squeezing more trains onto a line needs shorter gaps between them. This demands more accurate information about their location. Conventional signalling can do this by erecting more signals and installing more track circuits or axle counters. These circuits and counters determine a train’s position and allow signalling systems to more accurately place trains. At Level 2, ETCS does away with the signals but it still needs the circuits or counters. Simply switching signals for a screen in the cab does not improve capacity.

Level 3 removes the need for circuits or counters because the train itself works out its position and sends this via radio to the control centre. This allows for ‘moving block’ (as opposed to the fixed block created by track circuits). The signalling then computes the best distance between trains depending on their speed (just as car drivers do – nose-to-tail in crawling traffic, longer gaps at higher speeds). Signalling company Thales reckons ETCS L3 is ten years away from widespread deployment.

In any case, signalling experts will point out that capacity is not limited by the distance between trains on plain track. More constrains comes from the mixof fast and slow trains, their stopping patterns and the capacity of termini to receive and dispatch trains.

Termini challenge ERTMS, especially its GSM-R radio system, which is based on ageing technology, akin to 2G in mobile phone terms. This means that it does not have capacity to cope with the number of trains in a busy station. Upgrading it to GPRS will help and this forms part of NR’s plan. Elsewhere in Europe, railways swerve around this problem by retaining conventional signalling at busy termini, which negates any capacity benefit ETCS might deliver elsewhere. It shows that Europe sees ETCS installation simply as a signalling renewals. NR sees it as a much wider project.

There are further problems with GSM-R. GPRS is now old technology and will be obsolete in a decade. Even today, commercial mobile phone networks interfere with it. That’s why there’s a 3G transmitter in Cardiff that’s switched off because it interfered with railway communications.

The railway radio of the future must have sufficient capacity and must not be susceptible to interference from other networks because that would be another source of delays to trains. The UIC has just issued the specification for a future rail radio system. Yet, as NR’s chief digital railway engineer, Andrew Simmons, told MPs, this specification is likely to take two to three years of discussing before plans can be further developed.

Part of NR’s problem is that its tracks are crowded and busy now. In the rest of Continental Europe, there’s less pressure for technology to solve congestion and less impetus to move forward. There are hints that signalling manufacturers are in little hurry to move towards ETCS L3 because they want to recoup their investment in L2. The European Railway Agency would like to see L2 being used successfully before moving to L3, according to the Institute of Railway Signal Engineers. This gives NR and Britain an opportunity to lead L3’s development but also the challenge of dragging European railways along a road they don’t yet wish to travel.

All the while, passenger numbers in Britain keep rising. As Carne admits, DR is not a panacea and major projects such as High Speed 2, Crossrail and Crossrail 2 are needed, in addition to smaller improvements. But he’s in a hurry to deliver his vision of a better railway. “150,000 people a day are standing on commuter trains, we have to do something and we have to do something fast,” he told MPs.

Is it churlish to suggest that if he finds seats for those 150,000, their floorspace will simply be taken by another few hundred thousand standees?

This article first appeared in RAIL 802, published on June 8 2016. For more, see

Published by

Philip Haigh

Freelance railway writer, former deputy editor at RAIL magazine - news, views and analysis of today's railway.

Leave a Reply

Your email address will not be published. Required fields are marked *