Communications Based Train Control (CBTC) means different things to different people, so says Tom Lee from the Rail Standards and Safety Board who chaired a recent conference in London on the subject organised by Sagacity Media. The engineering / technical aspects are often not seen in the context of operations or passenger service and, whilst CBTC systems are becoming more widespread, they will never be a mass market offering such as road vehicles and public communication networks.
The mix of capacity and energy usage is crucial and far too many examples exist of trains transporting fresh air around for parts of the day. Availability and reliability go alongside safety and using proven technology is often best – being the second application might be preferable but remember someone has to innovate.
So what is CBTC, what does it offer and where are the shortfalls? A number of speakers attempted to provide the answers.
History, standards and broad perspective
David Dimmer, from Thales but working for European railway industry association UNIFE on its NGTC (Next Generation Train Control) project, started the conference off. He commented that, whilst most people associate CBTC with urban metros, it can also embrace ERTMS (European Rail Traffic Management System) technology, particularly where this will be used in high density urban operations.
The technology really began back in the 1960s with the opening of London’s Victoria line, but CBTC only became a recognised terminology in the 1980s. At that time it used track loops as the transmission method although radio became the natural choice in the 1990s.
CBTC consists of three elements – ATP (Automatic Train Protection), ATO (Automatic Train Operation) and ATS (Automatic Train Supervision). Attempts to standardise the functionality and technology have had mixed success. Results so far include:
- IEEE 1474.1 – Functional and Performance Requirements
- IEEE 1474.2 – User Interfaces
- IEEE 1471.3 – Recommended Practice for System Design
- IEEE 1474.4 – Recommended Practice for Functional Testing.
None of these are mandatory and fall way short of getting standardisation in equipment practice. The intended specification IEC 62290 ‘Urban Guided Transport Management – Command and Control’ has too large a scope, thus making little progress. The ModUrban project had an objective to create an interoperability specification but this was not achieved other than a definition of system architecture. Nonetheless the need very much exists and in New York, as an example, the operator has demanded inter-equipment testing from two suppliers.
The NGTC project started in 2012 with a budget of €11 million aiming to produce a European standard that embraces ATO, DTO (Driverless Train Operation) and UTO (Unattended Train Operation) requirements. Its 21 members comprise UNIFE, manufacturers, operators – including London Underground and Paris Metro’s RATP, the ETCS (European Train Control System) Users Group together with various universities and consultants. It sees a growing commonality between CBTC and main line technology, setting out tasks as a series of work packages. These include: technical coherence, common message structure, moving block requirements, IP-based radio communication beyond GSM-R and satellite train positioning (main line applicability only). The focus appears to be on merging ETCS and CBTC technology; perhaps not the right priority at the present time.
It might be concluded that CBTC is more a broad conception rather than a specified technical and operating system. This creates uncertainty amongst both existing and potential users as to what type of system to choose and how much to specify.
The London Underground vision
LU has had a somewhat mixed experience of CBTC over the past few years. The first application (although not known as CBTC in those days) was the use of Automatic Train Operation in the 1960s on the then new Victoria Line. Designed in house, this was a world leader with the system lasting until 2012. Later implementations were not so smooth. The original intention to equip the Jubilee Line Extension prior to the Millennium went horribly wrong and it was to the credit of LU that a conventionally signalled fall back was designed and installed in record time. More recently the
Jubilee Line has been fitted with the Thales Seltrac system similar to that on Docklands. This project proved to be far from an easy upgrade and much press criticism resulted from the prolonged delay and periodic weekend line closures. The system now works well and, from the lessons learned, a similar technology is being installed on the more complex Northern Line. This project is going remarkably smoothly and will be complete by mid 2014 giving a 20% increase in capacity. Disruption to the public has been minimal, most people not even realising the work was progressing.
In parallel, the Victoria line has had its original ATO equipment replaced by the then Invensys Distance-to-Go radio system. This was uncharted waters as it was an upgrade from one CBTC system to another. Hugh Bridge, who works for LU on its Automatic Train Control programme, described the solution which was based around overlaying the new infrastructure upon the old, so allowing new and old trains to run together on the line. The conversion is heralded a great success and a significant increase in capacity – 33 trains per hour – has resulted. Current reliability levels are 4,000 hours for train equipment failure equating to three per week.
More recently, however, the contract to equip the Sub Surface Routes (Metropolitan, Circle, H&C, District) with CBTC using the Bombardier CityFlo system has had to be abandoned. The reasons for this are unclear but sufficient to say the complexities of the layouts with many flat junctions and joint running with other lines, will stretch the technology of any system that is eventually chosen.
As to the future, George Clark, the engineering director at LU, gave the current view on CBTC system capability to fit in with future London requirements. Ridership is expected to increase by 26% between now and 2024. CBTC introduction will impact on track, train, power, ventilation, signalling, EMC, tunnel cooling, platform management, telecoms, information distribution, ticketing and internet linkage.
Put succinctly, the logical progression of a CBTC is Operating Concept » Functional Requirements » Systems Architecture » Physical Structure. Following from this comes:
- Level of Automation
- Change to Operating Philosophy and Rules
- Migration Strategy (duplicate infrastructure or train equipment?)
- Simulation and Provision of Test Track
- Integration of Different Suppliers Equipment
- Reliability Growth by both Technology and People.
Once the system is decided upon, there come the key decisions relating to standardisation, cost benchmarking, interoperability, interchangeability, maintainability, information from diagnostics
and finally obsolescence. The information opportunities from CBTC are enormous since there is a mass of inherent data. Using this effectively should enable more innovation to
be achieved plus obtaining better security of future infrastructure. The Piccadilly, Bakerloo, Central and Waterloo & City lines will all need CBTC technology within the next few years, so the challenge will be to deploy the best system design without encountering major engineering problems.
The Crossrail Challenges
Choosing the right signalling system for Crossrail is challenging since this railway has to interconnect with existing lines east and west of London, each of which will have its own type of signalling. Equipping the central core section has caused some serious heart searching; should it have a proprietary CBTC system or try to adapt from what will eventually exist on the outer routes?
Duncan Cross from the Crossrail team explained that, once the infrastructure is completed and the trains built, to then not having a reliable control system would be disastrous. Thus the decision is to provide a proven CBTC technology for the central section. This will ensure a 24 tph capability with 30 tph as a future prospect, plus a reliable interface to platform screen doors (PSDs). Two complications arise from this however:
Interfaces and changeovers will need to happen when the trains transit to and from the existing lines.
The trains must be equipped with all the signalling systems of the routes concerned. This will include: CBTC, TPWS, AWS, ETCS Level 2 and possibly the extant GW ATP system put in as a trial back in BR days. In the fullness of time, some of these can be removed once the main line sections are equipped solely with ETCS. Initially, however, the rolling stock will have to host a bizarre collection of signalling kit.
The alternative would be to adopt the Thameslink solution for the central core by using ETCS with an ATO overlay. The fact that this does not yet exist as a proven combination makes people nervous when so much is at stake. Thus the die is cast and at least functions such as the driverless reversal of trains at Westbourne Park will be a formality.
Many countries have deployed CBTC systems. One of the first was Canada with the Vancouver SkyTrain. Since the opening of the 1986 Expo Line, many lessons have been learned. Ian Graham from British Columbia (BC) Rapid Transit described the Seltrac S40 technology with full dependence on the primary system and no axle counter train detection back up.
Being a ‘greenfield’ railway with no regional line interoperability made implementation easier but whilst trains were fully automated, removing a ‘driver’ was thought to be a step too far in terms of public confidence. Later line extensions have moved towards UTO but staffing levels are still considerable at stations since PSDs have not so far been provided. UTO offers many advantages as it minimises human error, eliminates rest time periods at end stations and offers an increased service frequency with less trains.
A system must be in place to handle train failures? Roving attendants are deployed who can get to trains quickly but provision of on-train intercom, alarm buttons and CCTV is part of the safety scenario. The current throughput of 108 seconds between trains (33 tph) is capable of being decreased to 80 seconds.
SkyTrain adopted linear induction motors for trains on the first lines but a later line has conventional AC motors. In comparison, the former offers better reliability. Understanding how well the system is performing is vital and Chris Moss, responsible for the systems engineering services, explained how the technology has progressed from the transmitting of simple fault codes that were printed out on paper, to trains that have continuous data logging on to a memory stick that then acts as a ‘black box’ recorder. An instrumented test train is timetabled to traverse all lines and continually check the health of all systems, including the linear induction propulsion.
From France, Dr Pierre Messulam, the director for innovation and research at SNCF, gave a pragmatic account of French ETCS progress, now running some eight years late. Testing software between different suppliers / countries and the braking characteristics of older rolling stock have been major problems, as has the high cost of retrofitting trains with ETCS equipment that already have TVM or KVB – these being train protection systems developed by SNCF.
Bug fixing has led to successive versions of the software. More recently, interference from public GSM networks into GSM-R has meant reliability problems on the radio link, which is a serious concern. Authorisation processes are slow and complex with national authorities having different requirements and procedures.
One positive outcome is that drivers are enthusiastic about ETCS, but degraded mode working needs a lot of attention and close co- operation with signallers. Ideally a simulator is the best way of training staff on how to handle emergencies.
Introducing CBTC is seen as essential to run more trains on the existing infrastructure since civil engineering enhancements are just too expensive, so said Said El Fassi, the technical director in SNCF responsible for system modelling. Optimising the interface design with existing infrastructure is vital, especially where more than one train service operates on the same line.
Station dwell times are crucial to performance and human factor studies are important in understanding this. Modelling can save enormous amounts of time in the future and enable a better understanding of both risk and interfaces for trackside and train interworking, operating rules, environmental considerations and maintenance policies.
A new innovation in Paris is project Nexteo, a joint RFF / RATP exercise to produce a control and communication system based on CBTC principles but capable of operating on main lines where there is high density traffic. This is being designed for a new mass transit line being built but will need to integrate with ETCS as well. Something similar to the Crossrail situation comes to mind.
Safety assessment and human factors
A big factor in introducing new CBTC systems is obtaining safety approval. Paul Cheeseman from Technical Programme Delivery Systems explained the processes. The need to independently assess systems by competent people not associated with the project is insisted upon and must focus on design, development and safety measures.
It should be risk rather than compliance based, as the latter does not necessarily mean being fit for purpose. Caution must be used in using standards to mitigate risk. Two elements prevail:
The Specific Application Safety Case (SASC)
Cross-acceptance from similar systems in use elsewhere can be relevant to the GASC and need to tease out the differences from what has been approved before. Getting a GASC is tantamount to having a ‘go anywhere’ ticket.
Whilst many CBTC applications will aim at DTO or UTO operation, some systems, particularly where main line running is required, will retain a driver. Designing the DMI (Driver Machine Interface) is a science in itself and the gap between technology and adoption needs to be understood. Elaine Thompson from Mott MacDonald explained some of the factors.
Full integration means everything on a single screen including controls and speedometer.
The likelihood of confusion where different types of Train Protection System are required is considerable. The choice between touch screen, soft keys or separate keyboard may be influenced by local preferences. Options are being evaluated on a Class 43 HST, with the results intended as important for when the cab design of Crossrail trains is finalised.
More than just signalling
In summing up, Alan Rumsey from Delcan in Canada pronounced that installing CBTC is much more than a resignalling project, it can be considered as a total line upgrade. It is essential to focus on the real ‘needs’ (capacity, trip times, flexibility, enhanced safety, automation, lower maintenance cost.) while challenging the ‘wants’ (historic practices, need for fall back system).
A migration plan to minimise service impact during implementation is important. Operators should start with what they want to end up with and work backwards, they shouldn’t work out the first stage first.
The train is key and integrated factory testing followed by trials on a test track is the best solution. A CBTC system may be regarded as a distributed computer network so it is essential to ensure there is a stable transmission network and reliable computer hardware.
At the end of a fascinating day, Alan and his fellow speakers had done much to dispel many preconceptions as to what CBTC really entails. It really has a part to play, in fact in many cases it is essential, in keeping high-density metro rains running.