The latest Aspect conference of the Institution of Railway Signal Engineers (IRSE) was held in Singapore, and a report appeared in last month’s Rail Engineer. However, due to its location, the first time that Aspect has been held outside the UK, it was inevitable that much of the proceedings would be taken up with metro technology and operations.
Today, most large cities have a metro system, but the types and variety of systems in terms of sophistication, capacity, safety provision and engineering are almost bewildering. Few, if any, are identical, standardisation being virtually non-existent. Proprietary systems from many different suppliers dominate the market but, in order to keep abreast of the latest technology, these same systems are updated on a regular basis such that lines equipped with one technology may be different to another technology inside the same metro network.
This has both advantages and drawbacks. It is good that the latest technology is deployed, so as to maximise operational requirements, but the interworking of rolling stock between lines, spares holdings and staff familiarity can be a real problem.
Many systems exist under the CBTC (Communications Based Train Control) banner but, even here, the means of achieving the vital transmission link can vary: coded track circuits, track loops and radio.
Several of the papers presented at the conference explored the different metro engineering practices and operation. Whilst this gave an opportunity to ‘showcase’ the systems in particular cities or countries, it did highlight the variety of systems deployed and the challenge when planning new investment.
Singapore itself has a superb metro system, so commented Chua Chong Kheng, the deputy chief executive of the LTA (Land Transport Authority). The first line opened in 2003 with many more since and a further four extensions being built.
The lines have a mixture of different proprietary CBTC systems, with much of the equipment being duplicated. Single points of failures do, however, occur. The ongoing objective is to have everything duplicated, including point machines. Track circuits provide secondary train detection should the radio-based primary systems fail. Problem areas are radio spectrum availability and the lack of interoperability and interchangeability between the different systems.
These issues were to be a common theme at Aspect.
Metro trends and challenges
Whilst metro technology has advanced massively in the last two decades, the decision-making process for new or upgraded provision has become ever more complex, with an increasing number of factors needing to be considered, according to Andrew Love from SNC-Lavalin.
Many are obvious:
- Frequency of service;
- Complexity of service patterns;
- Criticality of service (‘must run’ or ‘good to run’);
- Timetable dependability;
- Number of passengers and impact of station dwell time;
- Type of signalling and sensitivity to response time;
- Level of automation and whether it is a requirement or a solution;
- Variety of rolling stock, often with different performances;
- Physical environment – tunnels, fire safety, space and access, ventilation;
- Need or otherwise for platform edge doors (PEDs);
- Interfaces to other lines.
Other new challenges are also emerging with the increasing trend to mix metro and main line operation in city suburban areas. Interoperability then becomes the buzzword, with the overall goal being to provide ‘safe and uninterrupted movement of trains to achieve the desired levels of performance’. Easily said, but not so easy to achieve when the mindsets of metro and main line engineers can be very different. Increasing automation in both camps should be leading to a convergence of technologies, but this is a long way from becoming a reality.
As indicated, metro signalling and control tends to be based on proprietary technology, developed by different supply companies with small metros just buying ‘off the shelf’ and only larger metro networks being able to influence design changes to achieve an element of customisation. Such systems yield a capacity of between 12 and 36tph (trains per hour), where every second counts and a timetable becomes almost irrelevant. If service-affecting problems occur, degraded mode operation can be critical, often by introducing speed control to bring trains closer together while avoiding the risk of stopping trains in tunnels for long periods.
Main line operations have evolved from historic national principles with generic technologies. However, supplier competition has led to the software associated with modern electronic interlockings being essentially secret, ATP systems becoming largely proprietary (except ETCS) and control systems always having subtle differences. Capacity is around 18tph with clock-face timetables and a large variety of rolling stock to cover both passenger and freight needs. Prevention of SPADs (signals passed at danger) is still to the forefront.
If metro and main line rail systems are to be integrated in urban and suburban areas, then converging the technologies will need to happen. This has to be industry led, since the technical expertise lies there. Most of the big supply companies have both metro and main line knowledge, but often in separate divisions. Merging the two would be a useful first step and should lead to products having the same vital hardware, with large portions of code being re-used. Software dominance over hardware is changing the way the market works, but being mindful of disruptive technologies will be important.
Instances of metro type main lines are there to be seen – London’s Thameslink, Gautrain in South Africa, even some high-speed lines, whilst mainline-type metros are also appearing – the London Underground Metropolitan line and Crossrail. Adapting the available technology to suit both needs currently requires innovative one-off solutions, often with multiple installations on the train. Getting an integrated solution will not be easy, but industry alliances must surely recognise the business advantages that this would bring.
In a conference of this type, the opportunity is always there for speakers to describe and promote their own country or city systems. Aspect was no exception.
London – the 4LM modernisation taking place on the Metropolitan, District, Circle and Hammersmith & City Lines. Already well documented and described within Rail Engineer articles, the project is, however, a challenge in terms of migration from existing technology, let alone integrating the new technology into other lines with different signalling systems.
Singapore – the system has two operators (SMRT and SBS) and now has to plan for replacement of track, power and signalling as lines are modernised, including provision for an increase in ridership of around 50 per cent by 2030. Digital systems are an obvious choice, but these also need to embrace diagnostics and maintenance using mobile devices, virtual and augmented reality (on tablets) for training, automatic vehicle inspection and depot facilities management.
A 50 per cent increase in the workforce is anticipated, with a rail academy already set up to facilitate the required training with supply companies taking advantage of this facility.
A big challenge will be replacing the signalling system without stopping traffic. The Thales SelTrac system has been selected, with much of the new infrastructure installed and tested. New MMIs (man-machine interfaces) are being built into the existing control consoles. Train equipment is being duplicated, but at the expense of removing two passenger seats to house the kit.
Hong Kong – systems for new lines and replacement on existing lines foresee driverless operation (DTO) as the goal (although still retaining an on board passenger support assistant) with a two-minute headway to achieve 99.9 per cent reliability.
Features include station stopping accuracy of ±350mm with automatic inch-forward or backward, automatic fare collection with self service points at stations, an open cab with an emergency driving console, a remote restricted speed command at 15kph for a maximum distance of 200 metres if communication is lost, continuous CCTV coverage of carriages each with four cameras, automatic ‘wake up’ and ‘sleep’ commands at depots and automatic turn-back short of terminus if line closures have to happen. Also planned is the use of LTE radio communication in 4G, maybe with special facilities for rail usage.
South Africa – the 80km Gautrain from Johannesburg to Pretoria, opened in 2010, is now planning extensions and two further lines. Deciding on the right technology for these is proving difficult, according to Portia Xaba, as it has to be linked to the expected growth in traffic as well as trying to predict the technology evolution.
Gautrain is a metro-like main line, currently equipped with conventional signalling plus ATP. Choosing between CBTC or ERTMS systems is one dilemma, as is the choice of radio bearer – GSM-R, Tetra, LTE, Wi-Fi or satellite. Radio bandwidth allocation is also a concern.
Toronto – a new 10km cross city link known as Eglinton Crosstown Rapid Transit required some novel thinking, according to Colin Williams from SNC-Lavalin. Being half underground and half street running meant providing two signalling systems, with trains needing a very low brake rate to handle the extremes of Canadian weather. The underground sections have ATO with ATP, but street level operation requires manual driving and linkage to the road traffic lights at crossroads.
No priority is given to the trains, as traffic chaos would result, but, to maximise throughput, east- and westbound trains are timetabled to pass each other at stations and road intersections. A maximum capacity of 20tph is achievable, but this concept of mix and match still has its problems.
Malaysia – lessons learned from the Klang Valley MRT SBK line have emphasised the need to have much-improved systems for integration planning and interface management for future lines. Signalling elements including ATP, ATO and ATS (Automatic Train Supervision) need linking to many other rail systems, such as civil infrastructure and stations, using computers and telecoms. Unless a proven interface process is available, future projects are at risk.
12 elements of integration are identified, critical ones being fare management, screen doors and depot access. From there comes the validation of the project using a mechanism known as DOORS (Dynamic Object Orientated Requirements System) that will lead to gap analysis and change control emergence.
Auckland – the re-emergence of rail as an important transport means New Zealand’s largest city has enabled some novel thinking for asset management in the $3.5 billion City Rail Link project. Using the typical asset life cycle flow of Business Case Create Operate and Maintain Dispose, building a database of evidence using BIM techniques (perhaps re-named Better Information Management) is enabling 3D views to be created of the entire system, with all design packages put into a single environment. Condition monitoring sensors feed into the BIM data to show the precise asset situation.
Other metro considerations
Other dilemmas are facing metro operators, according to Robert Cooke from the Singapore LTA. Defining what is wanted from a CBTC railway can prove difficult. Getting the balance right between performance and prescription without stifling innovation but limiting risk is a challenge.
Another difficult issue is whether or not to go for an unmanned railway (UTO = Unattended Train Operation) as this has social, safety and technology factors to consider. Some metros have already used UTO in major cities (Paris has two such lines), so the problems are not insurmountable. Dwell times at stations have to be carefully assessed, as door closure is automatic regardless of platform conditions. Typical would be 28 seconds at non-interchange stations, 45 seconds at interchanges and 60 seconds at terminal or turn-back stations. Terminal station design is critical to ensure efficient reversing. Platform edge (screen) doors (PEDs), and the timing of their operation, can delay passenger movement so these must be integrated into the signalling contract to ensure system ‘ownership’.
Other UTO requirements include mandatory CCTV coverage of every carriage with live viewing, together with loudspeaker announcements if an alarm incident occurs. Also needed will be automatic train ‘wake up and sleep’ commands, a ‘creep’ command to get the train to the next station if the ATO fails, a passenger-initiated emergency evacuation procedure using train-end doors (critical in any fire alert situation) and continuous health monitoring of the train’s condition and performance. With all of this, a 90-second headway should be possible.
Another hot topic is the need, or otherwise, for secondary train detection on CBTC-equipped lines. A modern CBTC system will continuously track the progress of trains by combinations of balises, track loops, radio signals and odometry. These allow moving block to be achieved and trains to close up when traffic levels are high.
However, if the system fails, then trains come to a halt and restoring movement under degraded conditions can be a challenge. Conservative thinking leads to the provision of track circuits and/or axle counters as a secondary detection facility. This gives the control room positive information on the position of every train and allows fixed block operation to kick in with associated commands for train movement.
The pundits claim that, as well as aiding the recovery of service, it does enable stock that is not fitted for CBTC to operate and (with track circuits) enables some types of broken rails to be detected. However, secondary detection does add more complexity, with more interfaces that result in lower reliability.
So who is right? There is no easy answer, but many seem to adopt secondary detection just to be ‘on the safe side’. Recent information from Madrid Metro indicates that it has recently removed its secondary detection equipment, with a consequential improvement in system availability. The consensus seems to be that it is not worth it.
Testing and commissioning
Because metros are essential to the daily life of a city, being granted an extended line closure period to introduce a new signalling system is a rare occurrence. If the line is totally new, then commissioning is all part of the project, so not a problem. For a replacement system on an existing line, the risks are considerable and, according to Daniel Woodland from Ricardo Rail UK, 67 considerations have been identified by the RSSB as needing to be assured before commissioning can take place.
Just understanding the reasons for testing and what tests are necessary is a good starting point. Modern systems are too complicated to test as a part installation, although equipment elements can be tested in the factory. Agreeing the relevant standards for the different elements, which nowadays have to include human factors, is an important consideration.
Project work has to be co-ordinated with the operational railway, which requires a major planning exercise. Beyond the obvious engineering and technical elements, it is essential to include:
- Testing accessories, including tools and test equipment;
- Test trains;
- Test pass criteria;
- Admissibility of earlier test results;
- Competence requirements;
- Recording and reporting;
- Supplier compliance and quality control;
- Stakeholder involvement;
- Briefing the plan;
- Access arrangements;
- Possession details;
- Traction rail de-energisation;
- Safety rules and assurance;
- Rest periods and welfare.
Much of this might sound obvious but how many times do we witness project dates being missed or, worse still, overrun once the implementation has commenced? It is all too easy to get bad headlines in the media, with exposure of blame only worsening the situation.
Metro performance on main lines
As hinted previously, a growing number of main line suburban routes in and around cities are facing metro-style capacity challenges. What technology should be adopted, when these will frequently be mixed traffic railways?
David Gill from Siemens cited the Thameslink central core in London, where ETCS level 2 will be overlaid with an ATO package to achieve 24tph. The system is actually designed to achieve 30tph but is restricted in order to allow recovery from perturbations.
The situation is helped by having only one class of train (Siemens 700) using the section. Could an ETCS Level 3 solution be better? Considerable modelling has taken place, concluding that train delays would be considerably reduced with much better recovery from any system failures. As well as eliminating track-based signalling equipment, trains are presented to junctions at orderly intervals by holding trains back at a platform if a conflict is predicted. Trains would request routes dynamically only at the last moment, and cross-coupling of delay at flat junctions would be eliminated. Under normal running conditions, however, a Level 3 system would offer only marginal performance advantages.
Signalling and operating a metro railway is exciting. Innovation, and the taking advantage of the latest technology, is there for all to see. Increased capacity in terms of trains per hour is the driving force with many examples worldwide enjoying the benefits.
The downside is the absence of standardisation and the ‘locking in’ to proprietary systems.
The rise of mixed metro and main line operation is an increasing demand, so the latest challenge for the supply industry and the rail operators is to establish an improved co-operative relationship, such that interoperability and interchangeability can be achieved without stifling innovative thinking.
Achieving that will be difficult but the advantages will benefit all.
This article was written by Clive Kessell.
Read more: Nokia – What has it got to do with rail?