Summaries of Papers read at Presidential Programme Technical Meetings

Transcription

1 IRSE PROCEEDINGS Summaries of Papers read at Presidential Programme Technical Meetings Each year the President of the IRSE invites six keynote speakers to produce papers and presentations on selected topics. These are traditionally presented in London, although in more recent years other locations have been used as well. The second paper listed below was presented in Birmingham, UK. In the papers and speakers were as follows (a summary of each appears on the following pages, together with a summary of the discussion that followed): 1. Banedanmark Resignalling with ECTS by M Søndergaard (Banedanmark), presented on 6th October 2015 in London (IRSE NEWS 215). 2. Next Generation Railway Information Systems and their Impact on Railway Signalling and Control by C. Roberts and J. East (University of Birmingham), presented on 11th November 2015 in Birmingham (IRSE NEWS 216). 3. Mathematical proof in an automated environment for railway interlockings by P. Duggan (Siemens) and A. Borälv (Prover Technology AB, Sweden), presented on 8th December 2015 in London (IRSE NEWS 217). 4. Telecommunications: The Heart of the Digital Railway by P. Jenkins (Paul Jenkins Associates), presented on 14th January 2016 in London (IRSE NEWS 220). 5. Data Quality Management systems: Network Rail by T. King (LSC Group) and D. Crowley-Sweet (Network Rail) presented on 10th February 2016 in London (IRSE NEWS 220). 6. Towards the Digital Railway by Patrick Bossert (Network Rail) presented on 15 March 2016 in London (IRSE NEWS 221). Page 1

2 Banedanmark Resignalling with ECTS by M Søndergaard (Banedanmark), presented on 6th October 2015 in London (IRSE NEWS 215). In 2009 the Danish parliament authorised the Danish Signalling Programme, with the objective of renewing the signalling throughout Denmark s national rail network over a twelve-year period, at a total cost of 2.5 billion (approximately 2 billion). The decision was supported by a simple fact, namely that the majority of Denmark s railway signalling systems were at, or in a number of cases, some way past, the end of their service life. The Signalling Programme will deliver two distinctly different signalling systems: in the Copenhagen urban area the existing lineside signalling and ATP on the 170km S-bane suburban rail network are being replaced by a wireless-based system of Communication-Based Train Control (CBTC) delivering improvements in line capacity by the use of moving block techniques and Automatic Train Operation, supervised by means of a new Traffic Management System (TMS). On the remaining 2,000km of the national network (known as the Fjernbane ), all existing signalling will be replaced by ERTMS Level 2, again with a new TMS, in this case, centred in different locations for the East and West sections of the network. Equipping contracts have been awarded to correspond with this split: ERTMS Infrastructure, Fjernbane West; ERTMS Infrastructure, Fjernbane East; ERTMS/STM Onboard equipment (for the whole Fjernbane network); CBTC infrastructure and onboard equipment (for S-bane); GSM-R radio for voice and data (for Fjernbane and S-bane). A number of in-house projects support the Signalling Programme with the development of new Operational Rules, training of operational staff, civil works required to accommodate the new systems, and upgrading and extension of existing data networks. The paper explains the procurement and implementation strategies adopted by Banedanmark for the Signalling Programme (Banedanmark is the state-owned organisation responsible for operation and maintenance of infrastructure on most of the Danish rail network) which have sought to take advantage of the opportunities offered by network resignalling on a national scale, something never before attempted in Europe. In particular, the need to overhaul the operational and organisational structures that support the signalling systems at all its levels of management, maintenance and development has led to the emergence of a holistic approach that will continue to evolve over the remaining years of the Signalling Programme. Discussion: The discussion was opened by P. Woodbridge (Siemens) who thanked the speaker for his paper and believed that he noted four areas of lessons learnt these being don t fit old trains but if you do get the Train Operators to do the fitting; have a fully funded programme and undertake the integration testing yourself and he wondered if there were any other lessons that the UK could learn. M Søndergaard agreed with these but advised that it is not always possible to undertake the integration testing in-house unless you have all the relevant resources and skills available at the same time. Page 2

3 An unidentified speaker asked what safety management system had been put in place. M Søndergaard thought that safety approval wasn t difficult but simply time consuming but confirmed that a process had been put into place and agreed with the relevant bodies. B. Kent (Bechtel) questioned if any performance modelling had been undertaken. M Søndergaard explained that they had but the calculations had proven difficult to undertake. T. Foulkes (Network Rail) wondered if the interlockings were to be installed within the Control Centre or geographically distributed around the country. M Søndergaard explained that this depended upon the architecture of the suppliers equipment. P. Grey (Network Rail) asked if there had been many failures shortly after commissioning. M Søndergaard acknowledged that there had been some failures both during the testing phase and post-commissioning but these were all thoroughly investigated. S. Eastmond (Arup) asked how far the integration testing had included the various suppliers. M Søndergaard explained this this was carefully considered and all interfaces reviewed. D. Fry (SSL) wanted to know if there were standardised interfaces with the existing equipment or were each treated as a bespoke solution. M Søndergaard confirmed that there were some unique interfaces when the new system abutted an existing system. P. Thomas (LU) questioned if some of the early software failings could have been prevented by off-site testing. M Søndergaard didn t believe this was the case the failings would only ever have been revealed once installed on-site under normal running conditions. C. Maron (Interfleet) asked what additional testing was undertaken on the test tracks to obtain authorisation. M Søndergaard explained that they are not quite there yet fitment of the on-track equipment and the trains is currently ongoing. P. Bartlett (Leigh Fisher) asked what the strategy for keeping the passengers was and if the vertical integration extended to the Passenger Information Systems. M Søndergaard confirmed that it was difficult to keep passengers advised of on-going works but he confirmed that all of the available data would be filtered down into the Passenger Information Systems. P. McFadden (Network Rail) noted that there was no overlaying of ETCS on the existing lines and questioned what were the pros and cons of this approach. M Søndergaard explained that the conclusion of the studies undertaken was that it was easier to take this approach as most of the equipment needed renewing. A. Simmons (President) thanked the speaker for his paper and the subsequent question and answer session. Page 3

4 Next Generation Railway Information Systems and their Impact on Railway Signalling and Control by C. Roberts and J. East (University of Birmingham), presented on 11th November 2015 in Birmingham (IRSE NEWS 216). This paper takes as its starting point a finding from the UK Department for Transport s (DfT) Value for Money study on UK railways, that effectiveness of the rail industry s Information Systems (IS) is inhibited by legacy systems that are expensive to run, outdated and hence encourage one-off workarounds to overcome their limitations. The 2012 Network Rail Technical Strategy (RTS) however presented a future where businesses in the industry are information-rich and able to use the information to drive decision-making, with decision support tools in place and in daily use. The paper introduces the concept of the Semantic Web, a web of data that can be processed by machines, and explains the use of data models to represent uses and structures of data in a given context so that the British rail industry may realise the vision presented in the RTS, The paper begins by introducing the key technologies and standards of ontology (a system of formalised naming and definitions of data) and RDF (Resource Description Framework), and explains their relationship to each other and to the rest of the Semantic Web. It references common concerns about the technologies (particularly security) and presents a series of brief case studies of applications of Semantic Web technology to other domains (oil and gas, retail and the public sector). The ontology model for the rail domain described in the paper was developed at the University of Birmingham as part of a study jointly funded by the Engineering and Physical Sciences Research Council (EPSRC) and Siemens and used as an example data relating to vehicle positioning to demonstrate how the functionality of a structured information system could be preserved through upgrades to current (and future) data sources from legacy systems and technologies. The paper describes how the concept is exemplified by means of a technology demonstrator to show evolution of an IS application from a rail network based on low resolution train positioning (track circuits or axle counters) to the higher resolution provided by communication-based systems such as ERTMS and illustrates this by means of a storyboard explaining the above scenario. The authors also describe how such an intelligent IS system can increase data availability during periods of degraded system reliability by the use of ontology reasoning to resolve live train locations based on the signalling system s degraded mode train location data (track circuits). The authors conclude by considering whether or not the promising findings of research projects into the topics of ontology and linked open data could be translated into tangible benefits for the industry. Discussion: The discussion was opened by D. Robson (SSL) who asked if there was any form of co-ordination between this and other projects of a similar nature that are currently being developed. C. Roberts recognised that there were other ongoing projects but this particular project was specifically looking to build a model of the railway system without a particular application in mind becoming application focus restricts the ability to communicate with other systems and he thought that the industry should have a defined requirement otherwise the way forward may be dictated by external factors. Page 4

5 A. Simmons (President) confirmed that within the Digital Railway Programme there are a number of phases looking at how best to increase overall capacity by use of an overall modelled railway system. M. Thomas (retired) thought that with all the work on data descriptions that had previously been undertaken, it would be a relatively simple case to get a business case for this work. C. Roberts explained that he believed to take the big bang approach would be difficult to justify and therefore this process was being undertaken in smaller stages using existing databases to derive the data models. A. Simmons (President) asked about the capability of the system from a manpower viewpoint. C. Roberts advised that as part of the National High Speed Rail College he thought that the required skills could form part of the syllabus to ensure that the staff were educated with the necessary knowledge. M. Slade (CBC Systems) asked if any work had previously been undertaken to develop a standard railway ontology. C. Roberts explained that Birmingham had a core ontology that had been developed with various companies and other work is ongoing in Europe but none of these have a vision of an overall railway model. I. Mitchell (DeltaRail) believed that for the project to be successful, it would be necessary for all stakeholders to buy into the concept and he asked how this would be achieved taking into account all of the commercial and contractual issues. C. Roberts was unsure but he thought that these sort of initiatives would eventually bring about closer cooperation between the various companies although it would need to be centrally driven. N. Barnett (Digital Railway) asked how the concepts would be kept current and up to date as the technology evolves. C. Roberts believed that both governance and a strategy for maintaining the data is required. D. Robson (SSL) asked if any specific framework was being followed. C. Roberts confirmed that this lesson had already been learnt and standard models are now built into the ontology although the governance type issues were still being developed. M. Thomas (retired) believed that the evolution of the concept must be an integral part of the concept but any future evolution results in abandoning what has previously been undertaken. C. Roberts agreed that this was the case and that it was essential to control any system changes. I. Mitchell (DeltaRail) referred to the incorporation of concepts from various organisations and wondered how the original assumptions had been filtered out. C. Roberts explained validation of the functional level was key to this. S. Singh (Atkins) noted that there was a lot of dynamic data available but asked who was responsible for its accuracy. Page 5

6 C. Roberts was unsure who would be liable. M. Slade (CBC Systems) thought that the data backbone would be the most expensive element. C. Roberts agreed that the initial costs would be the most expensive but would only need to be undertaken once rather than multiple times for each element. K. Peters (Thales) was concerned that multiple repositories were being created. C. Roberts confirmed that the core data would be held in one source. A. Simmons (President) thanked the speaker for this paper highlighting the work being undertaken to make the Digital Railway a reality. Page 6

7 Mathematical proof in an automated environment for railway interlockings by P. Duggan (Siemens) and A. Borälv (Prover Technology AB, Sweden), presented on 8th December 2015 in London (IRSE NEWS 217). For those working in the field of safety-critical software, the techniques known as Formal Proof of Program or Formal Methods have been recognised for many years. In this paper, the authors describe the evolution of formal methods as applied to one manufacturer s range of railway signalling interlockings over a period of 15 years, application of the techniques to the rail domain having started in the early 1990s with Railway Industry Association Standard No. 23 (RIA 23), the forerunner to IEC subcommittee SC65A and standards IEC and BS EN This evolution started with Invensys Rail s (now Siemens) first exploration in 2001, in conjunction with the National Physical Laboratory (NPL), into formal proof of requirements for an interlocking; the NPL at this time was the distributor in the UK of products from Prover Technology AB in Sweden. Although this exercise did not lead to a commercial application, it did identify a number of issues relating to lack of precision in existing requirements which required a great deal of time to resolve. Several years later, the company s introduction of the Westrace Mk I interlocking, in which the signalling rules and principles were configured by means of ladder logic, presented a new opportunity for formal methods to be applied. Its successor, Westrace Mk. II, with its far greater capacity for configuration data, made the application of automated design methods and formal proof even more attractive and the company partnered with Prover Technology to progress the project, based on the use of Prover s Trident process. The paper describes this process in detail, starting with a library of generic specifications which can be used to define the complete process from design through to testing, together with the specification of the particular installation (i.e. the Scheme Plan) and the mapping of inputs and outputs to the Westrace modules. Following this initial stage, generation of interlocking data is carried out automatically, together with creation of test cases and safety requirements for the specific application. Finally, an automated verification tool, independently assessed, was developed to comply with SIL 4 requirements, provides the safety evidence for the application, based on formal methods. To demonstrate its use on a specific UK application, the process was road-tested by means of an exercise to prepare and verify data for the company s existing Shrewsbury to Crewe signalling scheme. In this case the interlocking code, test cases and safety requirements were created, and the application verified and simulated, in less than an hour. The paper concludes by considering at length the lessons learned to date from the evolution of the above automated processes, in particular that major problems still arise from imprecise, ambiguous or conflicting requirements, although these can be overcome by carefully presenting requirements in natural language that helps to both reveal ambiguities and to drive greater clarity in phrasing requirement statements. Other ongoing developments included the expansion of the scope of application of the tool-assisted processes for development and V&V to include post-commissioning modifications. Looking to the future, the authors note that a number of railway organisations today mandate the use of formal proof for safety verification, prior to commissioning of interlocking and CBTC systems and that the use of automated tools and sign-off verification of their requirements is key to the ability to meet this demand, although they admit that more widespread acceptance of this level of automation will require considerable changes in behaviours within the signalling industry. Page 7

8 Discussion: The discussion was opened by T. Foulkes (retired) who thanked the speakers for their paper and then asked why the modelling didn t start from the track layout and end with the Movement Authorities rather than beginning with the scheme plan. P. Duggan explained that the project had wanted to develop from a known point and the greatest requirement was deemed to be at the interlocking level. He believed that if the concept could be proven, then further developments could ensue. A. Borälv agreed that once proven these methods could be applied to other systems. C. Porter (Past President) pointed out that stabilisation of principles was deemed to be important and questioned how the generic application would be adapted to suit the customer s specific requirements. P. Duggan explained that if change was required it would be made with validation of the generic applications occurring - the economies of scale would still be achieved once proving of those changed principles had taken place although a cost of undertaking them would be incurred. P. Bartlett (Leigh Fischer) was interested in the fact that one company had undertaken the development and verification. P. Duggan clarified that the generation of the data was undertaken using an effective SIL0 system and the certification of the SIL4 data is being safety validated by various organisations. A. Borälv observed that the processes used perform more effectively when verified as a whole system. A. Simmons (President) wondered if the speakers saw a role for the IRSE in undertaking a validation type role of the principles. P. Duggan thought that there would be benefits in this for setting benchmarks within the industry. A. Borälv pointed out that there were already defined safety specifications in other places in the world so this should be feasible. J. Govey (Siemens) noted that signalling principles were constantly evolving and wondered if it would be possible to simply write data that ensured train separation occurred with the system generating its own principles to meet this requirement. P. Duggan agreed that the requirements were constantly changing and believed that the real key to success was to simplify the signalling principles a European initiative to write an interlocking specification was, he believed, still ongoing. F. How (IRSE) wondered if it was the constant change that was causing the issue rather than the need to actually simplify the principles. A. Simmons (President) believed that Canadian Pacific (CP) only IPR the code which could, in theory, be applied to another suppliers hardware. A. Borälv agreed that this was a possibility which suited their (CP) needs. Page 8

9 An unidentified speaker asked what would happen if the validation failed and whether the process could be applied across multiple systems. A. Borälv stated that the verification process didn t generally fail but if it did the cause would be investigated. He saw no reason why the process couldn t be applied to multiple systems but has, to date, been undertaken on a site-by-site basis. P. Duggan explained that certain interfaces are already modelled in the tool and could easily be verified. P. Woodbridge (Siemens) wondered if the verification tool proved both what the interlocking controls should and shouldn t do. A. Borälv explained that simulation checks the positives and verification the negatives. J. Barnes (Siemens) wondered how easy it would be to modularise the signalling principles. P. Duggan believed that there was a relatively small amount of work needed to achieve this objective. S. Eastmond (Arup) asked if formal requirements had been developed for interfaces with equipment outside the scope of the process. P. Duggan thought that there were very few standardised interfaces and whilst they could be modelled he felt that most would be one-offs and would be tested as such. N. Barnett (Digital Railway) questioned how the safety principles were demonstrated to be complete and correct. P. Duggan explained that this had been a human activity to model and review all of the interlocking requirements noting that this can never encompass every item that will be encountered. A. Borälv agreed that this diversity and expert review was a good method of validating the safety principles. A. Simmons (President) thanked the speakers for their fascinating paper. Page 9

10 Telecommunications: The Heart of the Digital Railway by P. Jenkins (Paul Jenkins Associates), presented on 14th January 2016 in London (IRSE NEWS 220). In 2004, the author presented a paper Telecommunications: the heart of the signalling system. This paper takes up the theme and explores current and developing railway telecommunications with an eye to the advent of the UK s Digital Railway. The paper explores three issues: current digital technology on the railway, the future use of telecommunications to support the Digital Railway, and ownership and control of railway telecommunications. It concludes with thoughts for the future. The author notes that digital technology has a long history of railway applications, from time-division multiplex signalling remote control systems in the 1960s, through TOPS (Total Operational Processing System in the 1970s, an early system, networked computerised seat reservation and of course Solid State Interlocking in the 1980s. Following the privatisation of British Rail and the loss of its telecoms network, a new Fixed Telecoms Network (FTN) and GSM-R train radio were put in hand. These are now being deployed to support the concentration of the UK s rail operations into 11 Route Operating Centres and the roll-out of European Train Control System (ETCS), as well as asset condition monitoring, passenger information, CCTV and Wi-Fi connectivity on trains; all of these need reliable and secure data communications. The overall theme of the paper is however not the technology, but the organisational environment in which railway telecommunications systems are developed, implemented and operated. The author quotes a number of European examples but singles out Transport for London (TfL) as a good example of what a Digital Railway should be, albeit applied to a vertically integrated network where the management of technological change is easier than on the main line network. In the second part of the paper, the author describes Network Rail s development of the new generation FTNx IP network, using multiplexing technology to provide additional capacity and security, with bandwidths available up to 16Tbit/s on the existing FTN fibre cables. The network for the recently-completed Borders Railway is explained in detail, with communication between trains and trackside provided by GSM-R, which now provides complete coverage of the national rail network. Although GSM-R technology has been overtaken by both 3G and 4G, a new standard is being developed by the European Railway Agency (ERA) to allow ETCS to use IP over GSM-R which will allow Network Rail to deliver its foreseeable communication requirements over its own networks, using existing or evolving technology. The third part of the paper returns to the organisational theme, summarised as the railway infrastructure manager must remain in full control of his telecommunications. The author recalls the British Rail Telecoms (BRT) network as one of the casualties of privatisation, being sold to an outside party who then disposed of its interest such that Railtrack (by this time custodian of rail network infrastructure) had no contractual basis on which to object and which therefore could neither control cost or performance, nor update the network. Further examples are quoted from Germany and France, both of which attempted similar privatisations, but with similarly unfortunate outcomes, owing to fundamental differences between the incentives of railway operators and commercial telecommunications providers. The paper s conclusions are that, not only are efficient, reliable and secure telecommunications critical to the modern railway, but that ownership and control of the networks should be firmly vested in the railway s infrastructure management so that maximum advantage may be taken of the opportunities they provide. Page 10

11 Discussion: The discussion was opened by T. Foulkes (retired) who thanked the speaker for his interesting paper; he thought that the key was getting reasonable user requests and maintaining the system. A. Simmons (President) thought that this was something that the IRSE could become involved in. D. Fenner (retired) asked if the speaker had read the Shaw Report, which potentially advocated the splitting-up of Network Rail, and wondered if his views should be considered. P. Jenkins was aware of the consultation paper and would be sending a copy of this paper to the author. C. Kessel (retired) agreed with the majority of the paper and he asked the speaker if he thought that the Telecommunications network would eventually be sold off as has happened on previous occasions. P. Jenkins thought that there was a danger that further fragmentation of the network could occur. K. Ford (Network Rail) questioned what would be the key challenges for the telecommunications network with the implementation of Traffic Management Systems. P. Jenkins believed that communications would be vital to the concentration of operations under Traffic Management. N. Frisch (ex Deutsche Bahn) thought that there would be a need to utilise external companies and wondered how quality of systems would be maintained. P. Jenkins agreed that it would be difficult but believed that it would be essential for the railway to retain overall control. C. Porter (Past President) was concerned that the pace of technological change within the railway industry is different to external companies and asked if this would cause problems. P. Jenkins acknowledged that technology was rapidly changing but thought that the rate of change within the railway industry needed to accelerate to harness these changes for everybody s benefits the current fragmentation of the industry is, however, hindering this process. A. Simmons (President) questioned if the use of IP would assist in this change. P. Jenkins thought that specifying the use of IP over open-interfaces in signalling was the way forward. An unidentified speaker thought that the use of open-networks could jeopardise railway operations and asked if anything could be done to mitigate against this. P. Jenkins accepted that there was a potential threat, which is why FTN-X had multiple security layers, and while cyber-security would increase in importance; he believed that control of access to the network was essential. H. Barty (TfL) asked if Engineering Safety Management activates needed to be enhanced if using opennetworks. P. Jenkins was unsure but emphasised the principle that each application should control its own security under those circumstances. Page 11

12 G. Clark (LU) asked if there was a blurring of the skill sets between the different disciplines and asked if this would present a challenge for the railway operators when specifying requirements. P. Jenkins agreed that this was a big issue and different skill sets would be required. P. Halliwell (Network Rail) wanted to understand if outsourcing of the different communication elements had to go hand-in-hand. P. Jenkins believed that the outsourcing should go together. An unidentified speaker pointed out that the ETCS industry had stated that it was going with GPRS and not IP - the step to IP taking place at some point in the future - and he thought that the bearer network shouldn t be critical to this. P. Jenkins agreed. S. Docherty (Atkins) questioned how the rolled-out IP network would be maintained with the current maintenance skills gap. K. Ford (Network Rail) responded by stating that this is an area that needs to be carefully considered. A. Simmons (President) noted that this change was probably the biggest fundamental change that the railway industry had faced for some time but it would be necessary to learn the lessons from other industries. He finally thanked the speaker for his thought-provoking paper. Page 12

13 Data Quality Management systems: Network Rail by T. King (LSC Group) and D. Crowley-Sweet (Network Rail) presented on 10th February 2016 in London (IRSE NEWS 220). The President s theme for the year is the data-enabled railway, a strategy which seeks to make better use of data to build, operate and maintain our railways for the benefit of its customers. This paper develops a concern raised by Andrew Simmons in his Presidential address regarding the data-enabled railway where [if] inappropriate or corrupted data is present, the ability of the signalling system to achieve its business objectives is likely to be compromised and explores the application of asset management standards and techniques to data. Recent years have seen advances in the evolution of standards for Asset Management, initially with Publicly Available Specification PAS 55 produced by the British Standards Institute in This has subsequently (2014) been developed into the ISO series of international standards which sets out principles, terminology and methodologies whereby owners of physical assets may implement a systematic approach to effective and efficient management of those assets. In this paper, the authors quote examples of non-railway activities that have been adversely affected by poor, inaccurate or corrupted data, and seek to demonstrate that the quality of data should be approached with as much rigour as that applied to physical assets. Network Rail is currently pursuing a programme towards achieving best practice in data quality through the implementation of international standards. At the organisational level, the ISO 9000 family of quality management systems sets out generic principles and definitions which are as applicable to data as to any other product or process or, as stated in the paper, data quality is nothing more or nothing less than the characteristics of the data meeting requirements. The International Standards Organisation (ISO) is currently developing the ISO 8000 suite of standards to specifically address data quality. These are being developed by an ISO technical committee TC 184 (Automation systems and integration), sub-committee SC 4 (Industrial data), referred to as ISO/TC184/SC4 in the paper, which summarises the history and development of the ISO 8000 standards since their inception in At the particular level at which data is considered as an asset, therefore, the ISO 8000 and ISO families of standards are used in conjunction to establish best practice approaches to asset management. The paper quotes as an example the UK Ministry of Defence (MoD) which is supporting an initiative by the UK Council for Electronic Business (UKCeB) to develop a Data Quality Assessment Tool (DQAT) to identify how data quality impacts the safe, effective and efficient execution of asset management across a complex enterprise, a number of parallels with which exist in Network Rail. The paper then explores further the application of a data quality strategy to Network Rail, detailing the approach to be taken to implement the ISO 8000 standards in a rail domain, with a graphical illustration of the process and consideration of non-technical issues such as attracting, motivating and retaining suitably competent resources, which is itself seen as a major challenge. The authors conclusions relating to data quality are summarised in the phrase if the characteristics of data meet requirements, then the right individual can make the right decision at the right time. Discussion: The discussion was opened by A. Salisbury (retired) who referred to the control loop and asked if system response times had been considered in the process to prevent resonance occurring. Page 13

14 D. Crowley-Sweet confirmed that this issue had been considered during the development process and parameters had been included within the user requirements. A. Simmons (President) wondered where the system boundaries were defined. D. Crowley-Sweet explained that these have been defined in the system process definition and the data information architecture. T, Foulkes (retired) asked for clarification on the use of the data to make real-time decisions. T. King confirmed that the process defines a range of conditions that will allow decisions to be taken safely. D. Crowley-Sweet also explained that it will be necessary to perform a series of safety assurances on various criteria to demonstrate both the accuracy of the data and reasons for the decisions that were made. M. Thomas (retired) observed that there needs to be a concentration on interfaces within the railway industry and the existence of just one database. He also asked if the data would be time-stamped. D. Crowley-Sweet accepted that interfaces were key but was looking towards assistance with this and thought that RSSB had a part to play. He also confirmed that the process was being taken to ERA and thought that eventually there would be a TSI on data; he and recognised the importance of keeping the data up to date not necessarily by use of a common database but linking the (existing) differing databases together. T. King also reiterated that the process will consider the accuracy, completeness and timeliness of the data quality as well as provenance. He was also unsure that a single common database was the correct way forward An unidentified speaker believed that if data quality improved, the requests for further improvements also grew and he asked how far from this Network Rail were. D. Crowley-Sweet believed that there was currently a long way to go to achieve this but it was important to get whole-life costs before setting off on that path. M. McLean (Telent) wondered if the quality of data would relieve contractors for any liabilities if it subsequently proved to be inaccurate. D. Crowley-Sweet was unsure how this would stand contractually but data supplied would come with certain assurances based on confidence levels of the data. T. King agreed that the data should come with statements on the accuracy and completeness of information supplied. J. Govey (Siemens) in referring to the BIM model, asked how signalling projects would be managed using this process. D. Crowley-Sweet acknowledged that the data required to build and commission was different from that needed to operate and maintain the system and this was currently under investigation. Big Page 14

15 benefits were also to be gained when considering all of the interfaces between signalling, civil engineering and train operation. T. King didn t believe that there were any real technical challenges in delivering this but the levels of complexity would need to be dealt with. D. Fenner (retired) asked about the security of the data both malicious and non-malicious. D. Crowley-Sweet explained that some of this would come under the professional head of Network Rail s Cyber Security Group. T. King agreed that security was important and needed to take into account all of the users. A. Kirk (Arup) wondered how people would be encouraged to take up data as a career. D. Crowley-Sweet thought it was an exciting challenge that would attract the necessary individuals. T. King also believed that it wouldn t be difficult to recruit the right staff. A. Simmons (President) thanked the speakers for their paper. Page 15

16 Towards the Digital Railway by Patrick Bossert (Network Rail) presented on 15 March 2016 in London (IRSE NEWS 221). The term Digital Railway has been adopted to describe the UK rail network s medium to long term strategy to take advantage of a range of digital systems and technologies, both existing and developing, in order to overcome a number of current natural and physical constraints and deliver improved services and experiences to users of the rail network, both passenger and non-passenger, at all levels. The Digital Railway Programme is consequently attracting considerable interest and in February 2016 the UK Parliament s Transport Committee announced an Inquiry into Signalling and Traffic Management Technology, looking at the plans for deploying new digital technologies and how they will impact the use of a rail network already working at capacity level on some of its major routes, with passenger numbers continuing to grow significantly over the next twenty years. This paper therefore starts from the premise that, despite existing and upcoming major projects such as HS2, building new tracks and infrastructure alone will not deliver the overall increase in network capacity that is required. It explores Network Rail s strategy to unlock capacity from existing infrastructure that will allow users throughout the network to benefit from more trains, better connections and greater reliability. The Digital Railway Programme is therefore a vital step towards releasing latent capacity in the UK s rail infrastructure by providing more train paths on routes where extra trains are needed most, better real-time management of those paths to improve connections (different parts of the network need different solutions depending on location, time of day, time of year and other factors), an overall increase in reliability and hence a reduction in both delays and their impact on the network and its users. The author outlines the major system components of the Digital Railway strategy, with the European Rail Traffic Management System (ERTMS) being the fundamental digital enabler through implementation of European Train Control System (ETCS) Level 2 and Automatic Train Operation (ATO), supported by Traffic Management, the Connected Driver Advisory System (CDAS) and Combined Positioning and Alternative Signalling System (COMPASS). Future development would include extending the architecture to add an ETCS Level 3 capability. The Digital Railway strategy is to focus the current proposals for national ETCS deployment (replacing conventional signalling systems only when they are life-expired by a piecemeal process that could extend over 50 years), into an accelerated digital infrastructure programme, modernising route by route to take advantage of improved maintenance and operational cost opportunities and improving capacity and resilience by implementing signalling that will provide bi-directional capability by default. The author summarises a case study carried out on the South West Main Line (London Waterloo to Basingstoke, Southampton and branches) and notes that completed analysis of the line s demands and constraints supports the case for a combination of conventional technology and modern developments such as COMPASS, which would maintain an operational railway during failures of the traditional signalling systems. Most significantly, the acceleration of digital modernisation in this way could avoid the need to build a fifth track between Surbiton and London Waterloo. The paper concludes by noting that Network Rail s Initial Industry Plan submitted in September 2016 includes a 25-year baseline option for a Digital Railway that is deliverable and based on proven technology, with an option to deliver a Digital Railway in a year timeframe with greater benefits using an upgraded version of the technology. Page 16

17 Discussion: No record was made of the discussion at this meeting. Page 17