The many articles on ERTMS (European Rail Traffic Management System), in both this magazine and other publications covering the rail industry, have generally reported that the system with its component parts of ETCS (European Train Control system), GSM-R and the ETML (the still to come traffic management level) will yield significant benefits in both reliability and capacity with a considerable cost saving in the longer term. No-one however with even a modicum of insider knowledge will claim that the route to implementation has been easy. Railways across Europe have struggled with the technology, the operation and the funding.

The objective is simple: to have a single command and control (i.e. signalling) system from a multi-supplier base that is fully interoperable such that trains can seamlessly cross borders between countries without having to change locomotives and with a common man machine interface for both signallers and drivers. Achieving this goal has, however, been fraught for many reasons, some of which are listed:

  • A reluctance to change long standing operating rules and signalling principles inside individual countries;
  • Initially getting the signalling design and supply industry to work together in the production of compatible products although this is no longer such a major issue;
  • The logistics of fitting and retro-fitting rolling stock with the necessary train and cab equipment;
  • The challenge of producing reliable and robust software that can be backward compatible with earlier versions;
  • The protracted approval process within the regime that has emerged for safety assessment and independent safety verification although this is now better defined and understood than previously;
  • The high cost of implementation and the difficulty in producing a convincing business case.

All these are significant issues and some well known UK commentators have recently questioned whether the whole ERTMS programme is capable of delivering the stated benefits and in the time frame that many railways would like to see. These reports are contributing to a growing nervousness as to the wisdom of committing significant funds to continue the rollout.In the UK, the ORR (Office of Rail and Road) recently invited Rail Engineer, with its base of informed knowledge, to discuss the ongoing situation. With the declared intention of government to cut back on all aspects of public expenditure, rail must be expected to take its share of this. When assessing the priorities, where will ERTMS/ETCS sit in the fight for funds?

The current situation

The UK has been slow to adopt ERTMS with only the Cambrian Line in service as an early deployment scheme (i.e. a trial) and from which many lessons have been learned. Two dominant ones emerged: firstly that the signalling rules imposed by ETCS with the version of software used for the project meant the operation of the railway was more constrained when compared to RETB (Radio Electronic Token Block) which it replaced and, secondly, that the retro-fitting of rolling stock is something to be avoided if possible.

Subsequently, a section of the Hertford Loop has been used for interoperability tests between different suppliers’ equipment, with good results within the scope of the trials. These are now expanded to test the ATO (Automatic Train Operation) overlay intended for use on the core section of Thameslink.

Elsewhere, plans exist for the deployment on the Great Western main line (GWML) to Bristol and South Wales – but only as a Level 2 overlay to the existing signalling system primarily because of problems in getting sufficient rolling stock (particularly freight) fitted. The project is still only at GRIP 1-3 stage, which makes it highly likely commissioning will be delayed beyond the 2019 target date.

The more radical project is to equip the East Coast main line (ECML) with ERTMS such that lineside signals can be removed – initially between Kings Cross and Peterborough and subsequently northwards. This is a real challenge for both Network Rail and the train operators and is likely to include more significant changes to the technical and operational rules, widespread fitting of trains, plus staff training and maintenance issues. The projected date for achieving this is 2020 but there is a growing lack of confidence as to whether it will be met.

In Europe, deployment is more widespread with many of the later high-speed lines being equipped and operating successfully with ERTMS Level 2. The earlier EC Directive, mandating that all TEN-T (Trans European Networks) must be equipped with ERTMS technology when renewal is due, has been expanded to cover all lines. Only a few countries so far, namely Denmark, Belgium, Austria, Switzerland (standard gauge lines) and Norway have declared an intention to fit their entire network with ETCS technology, primarily as a renewal-driven strategy. Denmark is leading the way with this but is encountering problems with supplier integration, the lack of technical resources, rewriting the rulebook and the ever- present funding issues.

Some countries (e.g. the Netherlands) will expand ETCS to most lines but will stop short of equipping the more rural routes. No European country has found deployment easy and many have had to make significant changes to the application design in the course of installation because of technical and software shortcomings. Despite all this, there remains a positive view that ERTMS is the way forward and, even in Germany where acceptance has been slow because of the installed base of their own train control design, there are now plans to fit ERTMS on routes that carry through traffic.

The goal of interoperability

To some, interoperability is a political rather than a technical objective. It does require co- operation between countries. A major issue has been the insistence by some member states that ERTMS must allow them to perpetuate longstanding country-specific operating practices, thus requiring options to be incorporated into what should be an otherwise- standard system.

The programme has also required collaboration between suppliers, which has had its challenges since it involves the makers of both ETCS and GSM-R equipment for both infrastructure and train-borne systems. When considering the number of suppliers in all these categories, it is a considerable matrix of firms and no wonder that the interworking required has proved to be difficult. Add to this both hardware and software factors and the progressive implementation of updates and changes, one can see that it needs a strong co-ordinating body to keep this under control and to ensure that all suppliers follow the same course.

The Brussels based ERTMS Users Group exists to encourage and cajole countries to adopt unified technical specifications and operating procedures. The ERA and a host of other organisations – TEN-T Executive Agency, EIM, CER, UNISIG & UNIFE, GSM-R Industry Group, UIC – all have contributions to make in defining how ERTMS will be provided across Europe.

These bodies represent the suppliers, the train companies, the freight operators and the infrastructure managers, so getting a consensus amongst the many vested interests was always going to be an uphill struggle.

Nonetheless progress has been made and trains do cross borders without having to change locomotives, thus proving that interoperability can be achieved. However, it remains a challenge, the biggest problem being the development of successive versions of the system software. Even when Baseline 3 Release 2 is finally approved, there will never be a situation when the software is stable for all time. Changes will continue to emerge but the vital issue is the need for backward compatibility to give assurance that equipment already in service can continue to perform in a safe and effective manner.

The GSM-R dilemma

The track-to-train radio, known as GSM-R, is a twenty year old design utilising 2G technology that is now largely superseded in the public offering of mobile services. Roll out has been more successful than ETCS with most countries, including the UK, now having nationwide GSM-R networks. Primarily these have been to replace ageing voice radio networks but the concept of using GSM-R for the conveyance of ETCS messages was there from the outset. The advent of packet switching (GPRS), following extensive testing on the UK Hertford Loop, will give much needed extra data capacity to ensure ETCS can operate in busy areas. The UK will, in all likelihood, go it alone with introducing GPRS if the European authorities are slow to sign off the final specifications and approval for its use as part of ERTMS.

ERTMS infrastructure [online]

However, 2G technology is old and will not be supported in the long term future. Guarantees exist to ensure the supply of equipment until 2025 but, after that, commercial considerations will determine how long the supply base will exist. Some thought has been given as to what will replace GSM-R and the UIC has a working party looking at the possible solution. This, in basic terms, is similar to other problems the rail industry often faces: how to cope with rapid technology change in the commercial world when past practice dictates railways need systems to have a 20-30 year life.

The railways were fortunate to get an allocated band of radio frequencies back in the 1990s but a consensus view is that no such allocation will be available for whatever replaces GSM-R. Some people are advocating a 4G solution (sometimes known as LTE) but more enlightened advice suggests they wait for 5G, now in the conceptual design stage. This seems to be more of a functional rather than technical spec and envisages the use of multiple radio applications including public mobile networks, private networks and Wi-Fi combining to provide the connectivity required. Many of the features envisaged will include the requirements of ERTMS and the data capacity will be huge. Safety and security integrity will be built in and anyone who believes this to be a pipedream should go and look at the PTC (Positive Train Control) system being implemented in the USA where the transmission between trains and shore is provided by something akin to this.

Even when a decision is made, the changeover from GSM-R to its replacement will be a logistics nightmare. One can visualise the building of a duplicate infrastructure which itself will require a considerable outlay, but the fitting of trains with duplicated equipment could be difficult as space envelopes are invariably tight. Much of the rolling stock in service now or being built will still be around when the radio change has to happen.

Fitting the trains

As hinted earlier, equipping the trains with ETCS kit can be difficult and expensive and many railways across Europe have been staggered at how much this can cost. Fortunately, in the UK, the situation is looking somewhat better. Since 2012, all new passenger rolling stock has had to be ETCS compatible so that when the time comes to equip any particular fleet, the cabling, power supply, antenna and space provision for the cab display and on-train equipment should make final fitment relatively easy. Particularly challenging are multiple units with end corridor connections that make the driving compartment rather cramped. The Class 158 units on the Cambrian showed how difficult it can be with the price per unit being in the region of £1 million after design and proving costs were appropriated.

For GWML, it is understood the new IEP trains and the Crossrail EMUs to be used on the London suburban services should all arrive from the factory with ETCS fitted. If lineside signals are to be retained, other trains operating over the route need not have ETCS in the short term.

Similarly, on Thameslink, the new Siemens units which will be the only trains operating over the central core section will come equipped with ETCS.

The ECML is another story as removal of signals means every train operating over the London to Peterborough section must have ETCS. For the IEPs, the Thameslink stock and the promised replacement for the Class 313 EMUs, this should not be a problem. However retrofitting some HST Power Cars, the Class 91s and associated DVTs, plus some Class 365s for outer suburban services, will be inevitable but with full width cabs, it may not be too difficult. There are also the Open Access operators – Grand Central and Hull Trains – to be considered, small in number but they must nonetheless be accommodated.

Then there is freight and the ‘yellow fleet’ of on-track machinery. The freight companies understandably wish to keep a ‘go anywhere’ strategy for their locomotives but a compromise will need to be reached to produce a restricted fleet for working over sections of the ECML in the short to medium term. Whilst a policy exists for fitting all freight locomotives over the period 2017 – 21, and some early design development work is believed to be underway for each class of traction unit, the funding is not yet agreed and thus no invitations to tender are issued. This is part of the ORR dilemma. The same factors will influence the on-track plant fitment.

In short, the longer the ERTMS roll out takes, so the fitting of rolling stock should get easier – always assuming the policy of new trains being ETCS compatible remains in place. The oldest DMUs and EMUs will probably be replaced in the next decade and thus the number of trains that will need to be retrofitted will be a decreasing number.

Realising the savings

Understanding the technical and operational constraints is one thing, but does ERTMS offer value for money? The ‘business case’ justification has recently been discussed by the IRSE International Technical Committee but the findings were somewhat negative. In generic terms, cab signalling with intelligent traffic management systems should produce a positive investment return on a whole life cost basis but the technical and operational logistics of migration from a conventionally signalled railway are enormous. Only with new build high speed lines and self-contained metro routes that change to CBTC will any generic case be sound.

The cost of implementing ERTMS Level 2 on an existing railway is considerable but, with many signalling systems needing to be renewed, it raises the question as to what other options exist. As indicated, some of the smaller countries have committed to a nationwide implementation but none of these are anywhere near achieving this. Renewal of the asset forms the business case but without really studying any alternative.

The crunch question is whether or not lineside signals can be removed. If the answer is no, then the business case in the classic sense is non-existent. Having, in effect, two parallel signalling systems has no business logic, yet this is the situation on many railways because of the difficulties of retrofitting rolling stock.

Level 2, even without signals, still requires a considerable amount of trackside infrastructure to support the track circuits or axle counters required for train detection and this has always been seen as a shortcoming.

Only with Level 3 is this element removed. In many cases, adopting ERTMS is an act of faith with only the EC Directive and the requirement for interoperability providing the real justification.

The elusive Level 3

As defined in the standards, ERTMS comes in three levels:

» Level 1 – an ATP system using balises (not deployed in the UK);

» Level 2 – a system giving both ATP and movement authorities via radio with balises for position reference and with the option of retaining or removing lineside signals. This is the most commonly used application in Europe, including the UK Cambrian line, and is the current spec for future UK schemes;

» Level 3 – a complete radio based system with balises for position reference but with the removal of track circuits / axle counters. It also facilitates moving block that allows trains to ‘close up’ when running at slower speed.

With the operational and cost advantages that Level 3 should yield, why has it never really been developed? The answer may be more political/commercial rather than technical. With industry having invested considerable sums of money in the development of Level 2, the companies wish to see a return on this outlay before venturing down another expensive development that might take years to come to full fruition.

One factor that has been a longstanding sticking point is the proving of train completeness. Without a track circuit or axle counter, how can it be proved that a train has not divided en route? This is not a problem for passenger trains with modern braking systems and on-board train data systems, but for freight trains, despite having a continuous brake, the driver may not realise the train has divided if the break is towards the rear of the consist.

More significantly, Level 3 will cause a shift in the functional responsibilities and cost between the organisations that make up a typical modern railway. The significant transfer of risk to the TOCs and adapting trains to be much more a part of the signalling control system is a heavy responsibility. Penalties for failure of equipment and delay to other trains will be a minefield and until the ‘rules of operation’ are sorted out, they are likely to be barriers that prevent Level 3 being progressed.

All is not lost, however, and maybe the use of Level 3 will be best suited initially to rural lines. This is already happening in Sweden where a system known as Regional ERTMS has been trialled. Under this system, the amount of signalling infrastructure required is minimised and the train crew becomes responsible for monitoring train integrity. For the UK, this would be the ERTMS equivalent of RETB – recently upgraded on remote lines in Scotland – but with the advantage of not having to have special train fitments and captive stock.

The predictions in the early 1990s that Level 3 could be deployed on the WCML proved to be wildly optimistic and even now, some 20 years later, there seems no prospect of main line introduction.

The role of Traffic Management

Whilst many claims are made that ERTMS can significantly increase train path capacity, it is doubtful that this can be achieved without a parallel traffic management system (TMS) in place. The ETML element of ERTMS is not really off the starting blocks and hence several railways have invested in TMS from suppliers with proprietary products. Much will depend on the mix and service pattern of traffic as to just how much capacity can be gained but TMS may be the vital factor in determining train running optimisation.

The UK carried out a ‘beauty parade’ of three such firms earlier this year with the result that Thales was awarded a contract for pilot schemes at Cardiff and Romford. These have yet to be commissioned and are already running late. With the emergence of the ‘Digital Railway’ (see below), all further deployment has been put on hold with the exception of the central core of Thameslink where a further TMS contract has recently been awarded to Hitachi.

ERTMS graphic [online]

There is a need to link TMS with ETCS since, to obtain maximum capacity, movement authorities must take account of the position of all trains in a wide area and only traffic management can give this information. A further link with C-DAS (Connected Driver Advisory System) will obtain the optimum train path utilisation, taking into account different stopping patterns, braking characteristics and junction conflicts.

Rail organisations are beginning to realise these vital connections but how to manage and co-ordinate the design and supply of the component elements is not yet fully understood. It would be sensible to roll out TMS in parallel with ETCS deployment, even though the benefits of TMS can also be obtained with modern, conventional signalling. However, this would further compound the logistics for the already- complex decisions surrounding the ERTMS programme.

The Digital Railway

The much-publicised Network Rail banner to create a Digital Railway needs some explanation. Certainly it embraces many different aspects of rail operation covering public interfaces, social media and train operation as well as safety and non-safety applications. One suspects that the heart of it is the requirement to create additional capacity and to improve the operational interfaces for both the travelling public and front line staff. ERTMS has to be part of this big picture and, whilst incorporation is a logical step, it is to be hoped that the Digital Railway team understands the many other technologies that will be critical to ETCS success.

Digital systems have been part of the rail scene for nearly 50 years when the first PCM digital transmission system was commissioned between Euston and Bletchley in 1968.

Since then the whole of the railway telecom network and many of the signalling systems have migrated to digital technology. The latest work to upgrade the FTN (the Fixed Telecoms Network) to IP (Internet Protocol) operation – the FTNx – is all part of the digital revolution. Thus the railway is well placed to take advantage of digital bearers that it owns courtesy of NRT (Network Rail Telecom), for which a recent article described the technology and value of this asset.

The alternatives

If not ERTMS, then what else? As one IRSE ITC member said recently, there will be no alternative because industry is going to make only the ETCS product for main line railways. Whilst this statement may be true for cab based signalling systems, there will always be the continuance of traditional signalling with ‘lights on sticks’. In the UK, the development of modular signalling with standardised components and ‘plug and play’ cabling may well be expanded for the re-signalling of secondary routes although the cable element, with the requirement for totally accurate measurement of distance, may revert to traditional cable methods.

Several lines have either been implemented or are in the course of conversion. It has many advantages since it requires no rule book change and avoids the problem of train fitment. It is easy to visualise that this technology will be around for a long time with an expected service life of around 30 years, potentially undermining the business case for ERTMS introduction.

Capacity is rarely a problem on such lines and, with the right signal spacing, headways of five minutes are perfectly possible. It must be remembered, however, that the adoption of ERTMS is now EC policy for even secondary routes and countries will have to make a case for not doing so when signalling renewals are due.

Only on the main trunk routes (including high-speed lines), where speeds will exceed 125mph, does enhanced protection and cab- based signalling become a necessity, and as such ETCS is a logical component. With the predicted increase in capacity that ETCS will yield, the system is also appropriate for busy commuter routes and, as in the central area of Thameslink, the option for an ATO overlay will further enhance the benefit. Just how much capacity gain can be achieved is open to debate with claims of 40% perhaps being somewhat optimistic, but 20% should be achievable on routes where most trains have the same service pattern. If conventional signalling is already optimised for a train service, then obtaining even 20% is unlikely.

Critical first deployment

The ORR is right to question the timing and scope of the current ERTMS programme. The recently declared intention of Network Rail to speed up the deployment through the Digital Railway initiative is to be applauded since, if the advantages to be gained are as claimed, then the programme should proceed with all possible momentum.

The risk is that Network Rail does not have a good track record for the timely and efficient delivery of conventional signalling systems, especially where new technology is deployed. Partly this is due to a critical shortage of experienced staff, partly because of the time taken to approve new equipment (viz East Sussex and the use of a plc for level crossing control) and partly because of the complex matrix of responsibilities between client, consultant, contractor and sub-contractors.

For ERTMS, and particularly the ETCS component, the crunch test will be both the Thameslink core and the ECML deployments.

If these go well and produce the expected operational and cost benefits, then the dominos can be expected to fall and ERTMS provision will gather pace. If, however, one or both of these schemes fails to deliver to budget, timescale or the anticipated benefits, then the whole ERTMS programme could, at worst, be put on hold until confidence of successful delivery can be established.

Taking into account all the factors described, the ORR would be well advised to closely scrutinise the progress of these projects on a very frequent basis so that no nasty surprises emerge when it is too late to back pedal. Fingers crossed all round, I guess.

Acknowledgements are made to Ian Maxwell from the ORR for initiating the investigation and to Francis How, chief executive of the IRSE, for some industry and personal observations.