Written evidence submitted by Roger Ford FCILT Companion IRSE (RTC0002)
Industry & Technology Editor – MODERN RAILWAYS
Founding Editor – RAIL BUSINESS INTELLIGENCE.
Executive Summary
Contrary to the impression given by Network Rail the main lines of the rail network are signalled with modern technology with extensive use of digital signalling and control.
Development of the European Train Control System (ETCS) remains a work in progress. Network Rail is collaborating with other European Railways, notably Germany and Holland, in an attempt to accelerate the availability of a ‘standard’ specification. However, other than Thameslink, it is unlikely that ETCS roll out will start before Control Period 6 (2019-2024).
Network Rail’s ‘Digital Railway’ is a concept based on unrealistic claims, such as the potential to increase capacity, which are not supported by experienced signal engineers and railway operators. Instead of being a separate directorate it should be brought within the Signalling and telecommunications function and its resources allocated to the incremental up-grading of existing digital signalling and control systems where current technology is already offering considerable scope for progressive enhancements.
Introduction
This submission draws on my extensive experience in reporting on the application of digital technology to signalling and telecommunications on Britain’s railways. Signalling has been one of my specialist subjects as a technical writer.
In 2007 I was elected a Companion of the IRSE for my contribution to informing lay readers on signalling and telecommunications issues.
Most recently, the Chartered Institution of Logistics & Transport’s award of Transport Journalist of the Year in 2015 was based on my coverage of London Underground’s aborted procurement of new signalling for the Sub-Surface Lines.
For both my publications I have reported on Network Rail’s programmes to implement the European Train Control System (ETCS), the Traffic Management System (TMS), the Rail Operating Centre programme and, latterly, the development of the Digital Railway concept.
Structure
This submission is in two parts.
Part 1 seeks to put the so-called ‘Digital Railway’ in its historic context and describes the current state of digital signalling technology in the UK.
Part 2 addresses some of the key areas of interest listed by the Committee.
Part 1
1.0 Digital signalling & control
1.1.0 Interlockings
The advent of affordable computing power at the start of the 1980s initiated a step-change in railway safety signalling technology.
1.1.1 Safety signalling uses logic to prevent a route being set which could allow a conflicting movement which could cause trains to collide. In large power signal-boxes this logic, known as ‘interlocking’, involved thousands of electro-magnetic relays.
This structure meant that interlockings were a natural application for digital computer logic. However, the software had to maintain the same levels of vital safety, integrity and reliability as the proven relays.
British Rail Research solved this problem and in 1985 the first Solid State (computer based) Interlocking (SSI) was commissioned for trial operation at Leamington Spa. Two British signalling companies commercialised SSI which became the BR standard and subsequently has been commissioned in 10 countries in Europe, Africa and the Pacific rim. Current new signalling schemes in the UK use the latest interlockings from Siemens and Alstom running software based on SSI principles.
1.2.0 Control Centres
BR Research then used SSI as the basis of a computerised replacement for the power signal box. This innovation was known as the Integrated Electronic Control Centre (IECC).
1.2.1 A 20th Century Power Signal Box incorporates large panels with a track diagram showing points and signals, each of which has a switch or button which is used to set a route manually.
1.2.2 In an IECC the signaller sits at a workstation with the track diagram displayed on one or more visual display units. Routes are set using a mouse to click on points and signals. The first major IECC was commissioned at Liverpool Street station, London in 1989.
1.3.0 Automatic Repute Setting
A key element of the IECC concept is Automatic Route Setting (ARS) also developed by BR Research. ARS is a computer based system which compares real-time train running with a timetable database and sets routes automatically. The aim is to reduce the workload on the signaller at peak times and allow control of a wider area.
1.3.1 ARS is a versatile system. For example, the software includes rules defining how to deal with service perturbations, such as establishing train priority at junctions. As a result it can accommodate late running trains and other forms of service disruption.
1.3.2 The latest version, ‘Enhanced ARS’, can be pre-programmed with operating strategies for different timetable scenarios which are stored in memory for future use. Typical scenarios can cover running special trains for sports events, partial closure of a junction for renewals or enhancements, or emergency operation in the case of a failed train. These scenarios can be developed and tested off-line before being stored for future use.
1.3.3 It should be noted that other proprietary screen based signalling control systems have been acquired since privatisation. However these did not feature ARS. It was assumed that independently developed ARS programme would be added retrospectively, but is only now being installed. Illustrating the power of ARS, pending fitment these control centres have required additional staffing on busy routes such as the West Main Line.
1.3.4 Some IECCs are currently being upgraded with an enhanced version known as IECC Scalable. This offers many more features. I strongly recommend that the Committee arranges a visit to the IECC Scalable installation at Marylebone station for a demonstration of this state of the art technology. The supplier could also provide details of further enhancements under development.
An article on Marylebone IECC Scalable was published in the July 2015 Modern Railways http://www.modern-railways.com/central/download_attachment.asp?File=030-033_MR_July%202015_Informed%202.pdf
1.4.0 Driver Advisory Systems
In its basic form The Driver Advisory System (DAS) is a computer on a train with the timetable stored in its database, together with route data such as gradients and distances between stations. It is in service with a number of train operators.
1.4.1 DAS compares the train’s actual location and speed against the timetable and the route characteristics and uses this information to display the optimum speed to the Driver to maintain timekeeping while reducing energy consumption. In a diesel multiple unit DAS can also switch off one engine should the remaining engines have sufficient power to keep to time.
1.4.2 With improved communications between track and train it is now possible to connect DAS to the signalling system. This is known as Connected-DAS (C-DAS).
1.4.3 A typical application of C-DAS is regulation at junctions. A trial was carried out for High Speed Trains approaching Heathrow Airport Junction on the Great Western Main Line from the west. Heathrow Express trains join the fast line at this point and in the event of variations from the timetable, a High Speed Train leaving Reading could have to brake to allow a red signal to clear behind a Heathrow Express train.
A proprietary DAS system was linked to the ARS of the IECC Scalable at the Thames Valley Signalling Centre. This system detected emerging situations and a display in the High Speed Train cab showed the speed which would allow the train to arrive at the Junction after the red signal had cleared. This avoids unnecessary fuel consumption and wear on the braking system. South West Trains is in the process of fitting C-DAS to its trains.
Part 2
2.0 Issues raised by the Transport committee
2.1.0 The efficiency of Network Rail's planned roll-out of ERTMS and ETCS across the rail network.
2.1.1 . Since the overall European Rail Traffic Management System (ERTMS) exists only as a concept, this section of the response focuses on its signalling component ETCS. It should be recognised that ETCS is not a signalling system developed to meet evolving railway operating requirement but a response to a political imperative.
ERTMS was conceived by the European Commission as a means of removing the obstacles to cross-border pan-European rail travel for passengers and goods caused by the different national signalling systems. A classic example of the complications for through trains was Eurostar which had to incorporate equipment for five national signalling systems.
2.1.2 This year marks the 20th anniversary of the publication of the European Directive covering ERTMS and ETCS. Developing a pan-European advanced signalling system from first principles has been a long process for various reasons. In particular it has been an iterative process, with successive versions incorporated requirements which have emerged from consultation with the national railways as the specification has become more comprehensive.
The definitive System Requirements Specification (SRS) which Network Rail and other European railways are hoping to use was not published until December 2015. This is known as ‘Baseline 3’.
2.1.3 Appendix A outlines the development process for ETCS. I have included this analysis to explain why it would be unfair to criticise Network Rail for the slow progress in implementing ETCS in the UK. Much the same applies in Europe.
For example On 14 January 2016 the Netherlands Government announced that that the national audit office had not been prepared to sanction the 2015 budget for ETCS implementation because of inadequate project management and cost control. Given that ‘large-scale deployment of ERTMS in the Netherlands is a complex, far-reaching and long-term programme’, the State Secretary warned that it would be essential to minimise inconvenience to passengers and freight customers.
It was argued the slowing down the roll-out would allow the Netherlands to benefit from experience in other railways in Europe that committed to network-wide implementation of ETCS.
2.1.6 As with its European counterparts, it is proving difficult for Network Rail to make a business case for installing ETCS Level 2. Provision of cab signalling requires fitting the cabs of all the trains and locomotives on a route.
Network Rail’s original proposal was to integrate ETCS installation with the signalling renewals programme – the so-called ‘50 year programme’. While the most economic approach, such a programme would have created ‘islands’ of ETCs in the current colour light signalling. This raised safety concerns with drivers switching between two different systems.
2.1.7 Prudently on the two ETCS installations authorised to date – the Thameslink Central Core and the Great Western Route Modernisation, the decision was taken to install new conventional colour light signalling first. ETCS would then be ‘overlaid’. Eventually, as more trains were ETCS fitted, the colour light signals could be removed.
2.1.8 I believe that, overall, Network Rail has done the best it can with ETCS still in development. It is currently working with ProRail, the Netherlands rail infrastructure operator, and Deutsche Bahn on a joint approach to accelerate the establishment of Version 3.5.0 as the European standard. Germany has been taking the lead in seeking to streamline the process and the proposed Memorandum of Understanding could add weight to this approach.
2.2.0 How changes to Network Rail's Enhancement Delivery Plan following the Hendy Review will impact the rollout of ETCS/ERTMS systems.
2.2.1 It should be noted that in the draft updated Enhancements Delivery Programme, the ETCS Cab Fitment Fund for CP5 has been reduced from £194 million to £133.5 million (at 2012-13 prices). The remainder of the original CP5 fund value has been transferred to CP6.
According to the most recent Network Rail ETCS roll out plan, the ECML from Kings Cross to South of Doncaster was intended to be the first route to go directly from Multiple Aspect Signalling to ETCS Level 2 with cab signalling only from 2020. Bids were submitted last year but the move to what is termed ‘signals away’ been deferred to 2022.
On the GWML, ETCS is still at an early stage in the GRIP programme. My latest information is that ETCS is now seen as a project for CP6.
2.2.2 When this on-going development is combined with the financial constraints following re-classification and the Hendy Review it is likely that after Thameslink, further ETCS roll out will not be funded until Control Period 6 (2019-2024
2.3.0 How the state of current GSM-Railway technology impacts rail infrastructure.
2.3.1 In smartphone terms GSM-R is a ‘2G’ system. While it is working well it uses what is termed ‘circuit switching’ where 3G and 4G systems use ‘package switching’. The practical implication is that although ETCS involves transmission of small amounts (packets) of data, each train occupies one voice channel, or circuit, throughout a journey.
2.3.2 This is both inefficient in terms of use of the radio spectrum and more expensive than ‘packet switching’ communications. At the approaches to a busy station GSM-R would not have sufficient capacity for ETCS.
In February the European Union’s Railway Interoperability & Safety Committee approved a new GSM-R Baseline 1 specification which introduces the GPRS packet switching required for ETCS Level 2 in busy locations. Network Rail intends to introduce ‘Enhanced Data Rates for GSM Evolution’ (EDGE) in parallel with future ETCS implementation.
2.4.0 How realistic the timings proposed in Network Rail's "Digital Railway" programme are and how these will be achieved.
2.4.1 ‘Digital Railway’ is seen by Network Rail as a combination of three technologies – ETCS, Traffic Management and C-DAS - which will create a step change in railway control and communications. Claimed benefits include generating ‘up to 40%’ additional capacity in terms of train paths, thus avoiding the need for infrastructure enhancements.
2.4.2 On timing, ‘Digital Railway’ is a prime example of ‘the aspirational delaying the good’. As part 1 of this submission shows, the digital railway already exists and provides the basis for further enhancements in the near term. Also, many of the claims made for ‘Digital Railway’, particularly capacity, are not supported by experienced railway signal engineers and operators.
2.4.3 Currently ‘Digital Railway’ has begun procurement of an Early Deployment Scheme to demonstrate that ETCS Level 2, TM and C-DAS can be integrated successfully. This £35-50 million pilot system will be installed on the Norwich-Great Yarmouth/Lowestoft’ (NYL) lines, which handle three trains an hour. It is due to be commissioned by the end of 2018.
Apart from the fact that a lightly used rural network seems an unlikely test for a system designed to boost capacity on the busiest routes, by the time NYL is commissioned, a similarly combination of ETCS, Automatic Train Operation and TM should be controlling 24 trains/hour through the Thameslink central core, with a short-term maximum of 30 trains/h during recovery from perturbation.
2.4.4 This miss-match highlights the lack of an experienced engineering director with signal engineering background at Board level exacerbated by an organisational structure which sees the Digital Railway Directorate as separate from the main-stream signalling function. The appointment of David Waboso, London Underground’s Capital Projects Director as Managing Director Digital Railway brings extensive signalling experience to Network Rail.
2.4.5 This need for Network Rail to be a better informed customer has been highlighted by the aborted procurement of a Traffic Management System.
In June 2009, Network Rail issued the OJEU Notice seeking expressions of interest in supplying a new Traffic Management System TMS. Around 60 responses were reduced to 12 by the first stage Pre-qualification Questionnaire. Further, more-detailed, clarification questions reduced the number to 6.
These companies submitted proposals which resulted in three shortlisted companies, each of which was awarded a contract to produce a demonstration TMS office controlling a simulation of the Leeds area. Total value of the contracts was £20m
Following completion of the demonstration phase the three contractors began bidding for the accelerated roll out of the TMS across the network only to be told in March 2015 that the procurement exercise had been ‘curtailed’.
2.4.6 As part 1 of this submission showed, digital signalling and control technology is an established feature of the railway. Its latest evolutions provides the basis for further development in the near term which could provide much of the benefit claimed for digital railway, sooner.
2.5.0 How the transfer of signalling responsibilities to Rail Operating Centres (ROCs) is proceeding, and what implications this transfer has for the overall rollout of new signalling and traffic management technology.
How changes to Network Rail's Enhancement Delivery Plan following the Hendy Review will impact the rollout of ETCS/ERTMS systems.
2.5.1 These two issues are, inter-related. Transfer of signalling control responsibilities to the Rail Operating Centres, inevitably a slow process, is being further delayed as signalling schemes are pushed back to CP6 following the Hendy Review. However, the ROC consolidates the re-signalling programme, rather than determining the programme’s progress.
Each ROC is progressively taking over from control centres and signal boxes in its area as resignalling proceeds. In some cases, a multi-stage transfer is required. For example, when a rural line with manual signal boxes is resignaled, initially control may be transferred to a new workstation in an existing control centre. In time this centre will subsequently be migrated to the area ROC.
2.5.2 The main feature of transfer of control to a ROC is the opportunity to extend the area or length of route under unified control. Modern signalling already facilitates this.
Extending the area controlled also improves regulation. A characteristic of today’s busy railway is that the knock-on effect of a train failure, can extend over hundreds of miles and several hours. With it’s wider over-view and traffic management capabilities, the ROC network should speed recovery from delays and mitigate the impact.
APPENDIX
ETCS timescale
Development of ETCS Level 2, is defined by Versions of the System Requirements Specification (SRS)
1.0 SRS Version 2.2.2 was the first to be authorised, (in 2002) and was installed by a limited number of operators. It was succeeded by SRS2.3.0d – the ‘d’ denotes ‘de-bugged’.
Network Rail used Version 2.3.0d for the pilot scheme on the Cambrian Lines and it is also being installed on the central core of Thameslink. The Thameslink ETCS application is entirely separate from Digital Railway which is looking to the long term national application of ETCS.
Baseline 3 is intended to be the definitive version of ETCS while being compatible with track and trains fitted with Version 2.3.0d
2.0 As an example of the time consuming nature of ETCS development, between authorisation of the original Version 2.2.2 and the introduction of Version 2.3.0d, just one of the rules had, over the years, grown to 14 pages long and was described by the European Rail Agency (ERA), which is responsible for ETCS, as ‘not easily comprehensible any more’.
In addition sub-clauses had proliferated during the multiple iterations. According to the ERA, some rules could have up to six levels of sub-titles with additional subsections within the different subtitles. The new structure for Baseline 3 has imposed a maximum of three subtitles.
3.0 Baseline 3, Version 3.0.0, emerged in December 2008. It then took seven iterations, to achieve the first version of Baseline 3 released for use (Version 3.3.0) in March 2012.
A further four iterations resulted in Version 3.4.0 in May 2014. After another four iterations, Version 3.5.0, the version NR intends to use, was released in December 2015. A possible Version 3.6.0 may be necessary in 2017.