Explaining ERTMS - Europe’s standard signalling system - is the easy bit. Installing it is a different matter entirely.
Britain’s first installation was expected to be working by 2009. In reality, it’s taken until March 2011 to switch the Cambrian Line west of Shrewsbury over to the new system (RAIL 664).
The theory behind ERTMS (European Rail Traffic Management System) is that it provides a signalling system that works across Europe. A locomotive fitted with ERTMS in one country should be compatible with ERTMS tracks in another country, a concept known as interoperability. For a locomotive owner, this should save money because his machine will need just one set of signalling equipment rather than one for each country in which it might work.
For the European Union, ERTMS is one of the keys to unlocking the effects of borders between member states. It should make international freight quicker, easier and therefore cheaper, making European trade more efficient.
For a track owner (known in EU-speak as an infrastructure manager), a signalling standard should help reduce costs. Each signalling supplier offers standard kit, which should work with that of its competitors. If existing signalling is coming to the end of its life, it should be cheaper to resignal with ERTMS. If you’re building a new line it makes sense to install ERTMS from the start.
That’s certainly what China is doing, as it expands its railways. ERTMS might be a European system, but China is its biggest customer. Of the top five world ERTMS countries only two (Spain and Sweden) are European - the other two are Saudi Arabia and Turkey.
Britain has a dense rail network, much of it fitted with modern signalling. It has little international freight to and from Europe, not least because, by European terms, Britain needs special, smaller wagons to fit its lower loading gauge. Our international passenger trains run under modern cab signalling in Britain and in France, and the latter does not need to be in any hurry to update its high-speed line signalling because, other than standardisation, ERTMS offers little advantage over today’s TVM430 kit.
Why, then, resignal the Cambrian Line with ERTMS? In simple terms, Network Rail was facing a 2012 deadline for handing back the frequencies used by the Cambrian’s current RETB signalling system.
NR Principal Programme Sponsor Jim Morgan tells RAIL: “The story of ERTMS goes back quite a few years. The Strategic Rail Authority was keen to develop an early deployment scheme for ERTMS in the UK, and the Cambrian Line was selected. Its discrete fleet was one of the key issues. And if it was a total…” (here Morgan pauses for a long time) “…if it didn’t work properly its effect on the rest of the network could be relatively easily constrained.”
However, the discrete fleet did not last long. Morgan continues: “Since then, Arriva won the franchise and reorganised the ‘158’ fleet, concentrating maintenance in Machynlleth - so all the ‘158s’ need to be fitted and so it has had a bigger impact on the operation than originally planned.”
Fitting train cabs will form a major part of future UK ERTMS schemes, and in many ways has influenced Network Rail’s overall plan. New trains are comparatively simple, according to Morgan, but fitting the kit into existing trains is much harder. For a train to run under ERTMS it needs a new cab display (called the DMI or driver machine interface), GSM-R radio and associated antennae, a computer (known as the European Vital Computer or EVC), a black box recorder called a JRU (juridical recording unit), as well as balise readers, Doppler radar and tachometers.
For Arriva’s Class 158s, the challenge came in finding space for all of this kit, and finding sufficient auxiliary electrical power to drive it all. Even then the challenges were not over. As Morgan recalls: “We have to involve train operators earlier. We have to take more interest in the cab, in the technology for fitting the cab. The infrastructure fitment is fairly routine - there are no big issues there. But the key risk is fitting cabs.
“We fitted some standard European cab driver machine interface (DMI). But on the peculiar circumstances of the Cambrian Line, with the low sun shining off the sea, it meant that the driver couldn’t see the screen properly. So we had to replace - or rather Ansaldo had to replace - all the DMI screens. That took a little bit of time.”
NR fitted the kit to 24 Class 158s and three of the Class 37s it has now classified as Class 97s. Here space was not a problem.
For the tracks, axle counters are fitted to detect trains, balises to provide fixed reference points for the on-train computers, and object controllers that command points. This kit is linked to the control central by the fixed telephone network, while the trains are linked by radio.
In the Department for Transport’s 2007 plan, the East Suffolk Line should have been next in line to receive ERTMS. This route is another controlled by RETB, and faced the same radio frequency deadline. However, the way operator National Express East Anglia (NXEA) organises its diesel fleet meant that any ERTMS conversion would have involved too many units.
“You have retrofitment of ‘156s’ and ‘170s’, you have a change in service pattern and the cab fitment costs because the way NXEA uses its fleet meant there were quite a lot of cabs, and we were concerned that we might not be able to deliver that in time because the prime thing we have to do is deliver a signalled route,” says Morgan.
This has forced NR to drop East Suffolk. It now plans a conventional resignalling.
Warming to his theme, Morgan continues: “The other problem was the south of the river fleets, which are all new. It’s obviously much cheaper to install ERTMS in new trains than retrofitting an old train. We did an independent review for DfT, and went through it with ORR to get the best financial case.
“What we’re doing now is looking at our signalling workbank for Control Period 5 to see what delivers the best value for money for resignalling. There is less equipment out on the ground, so there is less maintenance and it’s cheaper to install. That’s counteracted by the cost of fitting cabs.
“It’s getting that balance right that’s really quite difficult. If you do get that balance right then you can actually save a lot of money by not fitting unnecessary cabs. We’re working very closely with ATOC on this. ATOC is doing a lot of analysis on ‘What’s the best plan? Could we fit this fleet or not fit that fleet?’ and it needs some management of the train cascade policy with DfT.
“You wouldn’t fit a Pacer with ERTMS, whereas a newer train with 20 years’ life ahead of it you might well. Of course, there is a requirement that new trains as they are delivered are fitted with ERTMS, so that will reduce the costs quite significantly and reduce the risk to the industry.
“It’s a very difficult thing to get the cab fitment and numbers strategy right, but that’s why we decided not to do the Western without signals.”
The Great Western routes running from London Paddington will be the first main line to be equipped with ERTMS. First Great Western’s passenger trains are fairly fixed to the route, and a new long-distance fleet beckons in the Government’s Intercity Express Programme.
“If you think about the Western, then obviously the passenger fleet is fairly captive to the route, apart from the CrossCountry fleet,” says Morgan.
“I think probably 65% of the freight locomotives in the UK use the Western, so we’d be facing a massive programme of retrofitting freight locos. I said already that retrofitment of cabs is a big issue. So the decision was made on the Western that what we’d do is install a system with signals so that around 2026, when we actually have more cabs fitted, we’ll take the signals away. First Great Western is well in the picture about this.
“So that’s going ahead with signals. The next route on the plan is East Coast Main Line (South) - that’s Doncaster south to King’s Cross - which is the first installation without signals. Then Midland Main Line (South), again without signals. The exact southern boundary is still under discussion because that’s affected by the Thameslink programme.”
The DfT’s 2007 plan extends no further than 2044, with most of the network equipped. NR now has responsibility for revising and updating that plan on DfT’s behalf.
Morgan cautions against moving too quickly towards ERTMS. “It would be stupid to move to 100% ERTMS renewals at the moment. We’ve only done one route. We don’t have the skills or expertise to do it.
“But if we can move gradually to increase what we do, year-by-year, that’s the sensible approach. I don’t know whether it’s 2044 or 2050, but I should imagine we will move towards the UK renewal policy being a standard of ERTMS fairly soon. But we have to demonstrate that it works.
“The balance will be modular and ERTMS with very few conventional resignallings.”
Western pushed itself to the front of the queue because it’s currently fitted with an ageing Automatic Train Protection (ATP) system, a legacy from British Rail days when the national rail operator tested two systems with a view to installing one of them more widely.
“Western is a complicated installation because we have the requirement for ATP replacement,” says Morgan. “ATP is obsolescent. You can’t replace it with something that is less safe, so TPWS is not possible. That’s why ERTMS is essential on Western route currently fitted with ATP, so we’ll be fitting ERTMS on those routes. And, if we’re going to fit ERTMS then let’s fit the most advanced solution. We’re talking to suppliers just to see who has the most advanced stuff.”
Fitting Great Western lines with an ERTMS system that includes conventional signals does come with a cost penalty, as Morgan readily admits: “It’s expenditure in the short term because of the amount of work we have. On Western we were very conscious of Crossrail, which at the time had a tight timetable although that’s since slightly relaxed. And there’s electrification, which has an even tighter timescale. So it’s a lot simpler and cost-effective in the short term to actually install signals.
“I think it costs 5% more to do ERTMS with signals than ERTMS without signals. We looked at lots of options, and the best solution was the overlay solution of ERTMS with signals.”
But Morgan does have his eye on longer-term lower costs, in line with the pressure NR is under from the Office of Rail Regulation and wider interest groups to reduce costs and become more efficient. With planning for the next control period (CP5, 2014 to 2019), Morgan says: “We have to demonstrate value for money. We have to deliver cheaper solutions. We do need to move towards reducing costs quickly but improving safety at the same time, which is something ERTMS will do.”
“We’re looking at the whole signalling work plan and looking at what we can put back, as ERTMS depends on value for money because we’re all under pressure to cut costs. We’re under pressure to deliver value for money, and this is one of the ways we can do it because the capital costs are 60% of the capital costs of conventional signalling.”
“Look at Denmark. That’s what they are doing, because they haven’t done any signalling renewals for quite some time. Rather than resignal they are installing ERTMS through their whole network, because it’s the cheapest thing to do.”
Bringing Europe into the picture puts a different perspective onto Britain’s efforts. ERTMS provides a great opportunity for countries that have spent little on signalling (such as Denmark, see panel), or countries such as Britain that did not pursue automatic train protection systems with much vigour.
As Morgan notes: “I have to say there are some mixed feelings around Europe about ERTMS. They are largely the ones such as France and Germany that have an awful lot of Level 2 safety systems in their trains. You can get ERTMS to interface with all of them, but the more you do the more expensive it becomes.”
Denmark is also re-writing its rulebook to take advantage of ERTMS. For a wider Europe, there are clear advantages to having an ERTMS rule book common to its international freight corridors, although the chances of this seem remote given the profusion of different ERTMS standards being introduced across Europe (see panel).
At NR, Morgan sends a mixed message about rulebooks. On the Cambrian Line he says: “ERTMS gives us fantastic potential for making operations easier. Instead of just trying to modify the existing rulebook a little bit, we should perhaps have a fresh approach. We’re looking to appoint an operations expert to help us with that, but at the moment the Cambrian is operated with something that looks very like the existing rulebook.”
He cautions: “There’s a risk from moving dramatically away from the UK rulebook because you might end up with too dramatic a change. But there is something you can do to modify the rulebook. We haven’t reached any conclusion on that. The Danes have offered to share any information with us.
“The pressure from Europe is about international freight corridors. That’s what’s pushing ERTMS, so you can have one loco from Sweden to Italy. But that’s of no relevance to us. What we’re interested in is the cost savings from ERTMS - if you have ERTMS then you can have lower capital costs.
“Suppose you have ERTMS on the southern end of the ECML. There are some routes off the ECML where every vehicle will have to be fitted because it goes on the ECML - routes such as the GN/GE Joint Line. You could resignal that very cheaply. There’s an awful lot you can do if you’re smart and flexible.
“You need to get a sensible route out of ERTMS. Look at a section of the West Coast Main Line that I know well - Warrington to Preston. If you renewed the interlockings there would be six changes from conventional to ERTMS and from ERTMS to conventional. No driver could be expected to cope with that, so you need to get a sensible route out of that.
“We are going to have to do life-extension works in some places to get that but it’s blooming hard work. We’ve been doing that analysis for a few months, but haven’t quite finished yet.”
Key to Network Rail’s efforts to bring ERTMS to Britain’s rail network will be its test track at Hertford North.
NR plans to convert one track of this lightly used double-track line between London and Stevenage. To do this it will resignal a 51⁄2-mile stretch of one line to allow trains to run in both directions, and then convert the other line into its ERTMS test facility.
Morgan says: “Anybody who wants to install ERTMS in the UK on our network will have to install it on Hertford and prove it will work. Hertford was developed for Thameslink, but we’re going to use it also for testing on the Western programme.
“Thameslink needs automatic train operation (ATO) over the core section to get the number of trains per hour reliably, and that’s on an ERTMS platform.
“The new Thameslink trains are specified with ERTMS - they will be delivered with ATO based on the ERTMS platform that will be tested at Hertford.
“I think the core section is the only installation of ERTMS on Thameslink, but that’s the most important section. Thameslink will need a train management system to ensure they are delivered to the portal on time, and Hertford is the key to getting that right.”
Morgan reckons Hertford test track will be ready for use in 2014, and hopes to have his test train ready as the test bed for manufacturers’ kit. Having skirted around where this train is coming from, Morgan reveals all: “It’s not a new train. It’s a ‘313’. There is one spare ‘313’ and that’s 313121.”
He had hoped to acquire HSTs from First Group on which to install a pilot version of ERTMS, but “for various reasons this plan did not go ahead”. With First Great Western short of rolling stock, it seems likely that the operator simply had none to spare for Morgan’s ERTMS tests.
The test track invokes memories of Railtrack’s disastrous dalliance with advanced signalling systems in the late 1990s, particularly its plans to install moving block signalling on the West Coast Main Line. Such a system equates to ERTMS Level 3, and it’s worth noting that no rules and specifications now even exist for Level 3 despite a general acceptance of the principle of moving block.
Morgan explains: “There’s no way we would take that forward without a lot more work. But we think that with the right approach we can introduce new technology into the UK.”
He’s determined to move forward, arguing: “The UK is famous for innovation. Why should the rail industry be unable to innovate?
“But we have to think about it. We must demonstrate our competence in this area rather than say ‘this is the magic bullet’ and rush ahead. I was at Railtrack when West Coast was being talked about, and there was great excitement and cries of ‘yes, this is the solution’ without doing the analysis.
“NR is very keen on making sure we get this right. Our new chairman and our new chief executive are very keen on doing that analysis to prove we can do it. I’ve had a lot of support to actually try and develop this new technology.
“We’re looking at moving block signalling, but it’s probably best to test that on a rural route. We’re working with DfT to develop new low-cost signalling initiatives. We do know where we went wrong, and we’re determined not to fall into that trap again.”
Back in Wales, Cambrian ERTMS is set to go live in March, following last autumn’s commissioning of the first section between Pwllheli and Harlech (a section which includes the flat crossing with the Ffestiniog Railway). This section was signed into use at 0730 on October 24 2010.
“It is working very well between Pwllheli and Harlech,” says Morgan. “It’s probably almost the most advanced ERTMS solution anywhere. We’ve tried to get towards version 2.3.0d. We’re not quite there yet, but we do have a migration path to get us there.
“If you talk to people who have implemented 2.3.0d there are some who say they have and some who say ‘yes, but it’s a special Scandinavian version’ or something like that. It’s not been without its problems, but we’ve learned a lot of lessons. As far as the original objectives have been concerned it’s been pretty successful.”
But the installation is much later than had been planned. Morgan explains the delay: “It’s taken so long because it’s taken a long time to understand and put right what the issues are. When I joined NR a year ago we were looking at commissioning in March last year, but we hadn’t got the reliability right. We’ve done a number of upgrades of software since then and the reliability wasn’t good. There’s no point in introducing a signalling system that isn’t as reliable as it possibly can be, because what we’d give Arriva is a system that is less reliable than the system we have now.
“I think we made the right decision to defer commissioning until we had the software right. The DMI screens were wrong, so we’ve had to get those changed. There’s various other small issues, all of which take time to look at because this is entirely new for us.”
He defends NR’s decision to go for the most advanced system possible.
“There’s less advanced kit out there. Look at China, for example. China probably has more ERTMS than anywhere else in the world, but it’s largely Level 1 with signals. Now, you have an RETB railway - do you really want to put signals back? Probably not, so it needs to be Level 2 without signals. So, if you’re going to go for Level 2 why don’t you go for the most advanced system, because we’re trying to learn for the future.”
Cambrian’s advanced system has had its problems. Morgan says: “The railway is quite simple. It’s the software and the relationship with the radio block centre and the equipment on the train, the balise readers and the odometry. There’s an awful lot of very new technology on there which hasn’t been as reliable as we’d want it to be.”
He’s clear that reliability is the key, and cites Arriva’s recent improvement work. “ATW’s main concern is reliability. It has done an awful lot to improve the punctuality of the Cambrian Line since it took over the franchise.”
That’s the key to any new railway equipment - it must provide a safe, punctual and reliable railway.