Although digitalisation’s impact on society is a topical subject at present, the 'digital railway' concept has been around for some time now. In fact, urban metro railways have been at the forefront of utilising digital technologies since the 1970s and the rapid progress of broadband radio communications in recent years has offered even more opportunities for automation, writes Chris Cox.
This is evidenced by the fact that the majority of new metro lines opened in the past few years are driverless and, increasingly, owners of existing lines are contemplating resignalling projects to enable the substitution of drivers with mobile stewards.
If physical distancing becomes a normal feature of using public transport well into the future, greater capacity needs to be delivered to allow sufficient vacant space around individual passengers to be maintained. This will place more of an emphasis on automation.
Automation classifications for train operation are defined in the international standard, IEC-62290. Starting at GoA1 (Grade of Automation) which represents manually driven trains according to signal aspects, the standard goes up to GoA4 for trains that can operate entirely without on-board staff under typical conditions.
Automating the normal driver functions of controlling traction and brakes according to signal aspects is relatively straightforward and was achieved even before microprocessors became commonplace.
The challenge is to design in resilience so that service operations can be maintained under failure conditions, without a driver to troubleshoot and recover from breakdowns.
The introduction of broadband wireless communication made an enormous contribution to solving this problem. Despite the challenges, there is growing evidence that automated metro railways will become the default choice in the near future.
Indeed, even mainline railways are moving towards supervised automatic train operation, with examples such as the Thameslink core section using an evolution of the European standard ETCS (European Train Control System) to maximise capacity as a GoA2 system.
While unattended train operation (UTO) began in the 1980s, the earliest implementations were small scale and utilised proprietary technologies that locked the owners into a sole supplier arrangement for trains, track and signalling.
More recently, conventional rolling stock designs using standard gauge track with 54E1 profile rails have become the norm. Track-circuit based signalling systems have been replaced by communication-based train control (CBTC) systems.
These operate on fundamentally different principles in that the position of the train is declared by the train itself, rather than the train’s location being detected by wayside track circuit equipment.
With CBTC, a continuous communication channel between the central computer and the computer on-board the train ensures that the central computer knows where each and every train is and provides instructions to each train about where it can drive to – its limit of movement authority – based on the state of the route ahead.
Advances in wifi technology have enabled refinement of the level of control. Improvements to the reliability of the transmission path, increased data rates and reduced latency have made it possible to not only continuously control the exact position of each train and the subsidiary on-board functions remotely, but also to have extensive real-time monitoring of the status of on-board systems and CCTV feeds.
Three or four CBTC system products from major European railway signalling manufacturers now dominate the mass transit sector market, delivering essentially similar capabilities.
The Cityring line which opened in September 2019 (with a more recent extension to the north harbour) is a prime example of the latest greenfield driverless projects in Europe using the most modern digitalisation solutions.
The developer, Metroselskabet, had a vision to expand the ‘always on’ high intensity service of the first metro line which opened in Copenhagen in 2002 to provide a ‘turn up and go’ quality of service virtually round the clock for the entire inner-city area.
With high frequency services, the economic advantages of automation are more obvious. Headways of less than two minutes between trains demand them to be driven automatically to achieve the consistency of operation expected.
Like virtually all automated mass transit projects in recent years, CBTC was specified by Metroselskabet so that the highest levels of service flexibility would be achievable.
Developments in digital technology also allowed standardisation and increased use of commercial off-the-shelf components. These were chosen to achieve a balance between the reduced cost and the potential effects of obsolescence on the life span.
A comprehensive fibre optic based IP network supports the wifi network for the train data communication system. The network links access points with routers and switches back to the zone controllers and interlocks central computers that control the movement authorities and routing functions. Redundant designs are used so that operation is maintained in the event of failure or other catastrophe, e.g. a fire in the tunnel.
When the Copenhagen Cityring service builds up to its maximum anticipated capacity over the next year, a team of approximately 14 roving stewards will support the six-hour peak service of about 28 trains.
The average will be about 0.72 stewards per train in service over the entire day. Automation has released human resources from operating machinery to a more rewarding role of interacting with passengers and problem solving when failures do occur.
Automation also enables the rapid build-up and reduction of train frequencies to match demand by auto-launching the fleet from the control and maintenance centre and auto-routing of trains from revenue service into the maintenance area for cleaning and washing, with minimal input from control room staff.
Since there is a relatively loose relationship between the number of trains and stewards, operation is resilient to minor resource constraints. In fact, what sets automated metros apart from conventional rail services is that the daily delivered service is primarily customer demand driven rather than constrained by the resources available from operator organisations.
Service levels can easily be supplemented with additional trains with less than one hour’s notice. This would not normally be possible if more train drivers were required due to the standby costs.
In an automated metro, trains are functionally integrated into the overall system including the civil work tunnels, stations and MEP. The CBTC and SCADA systems work together to deliver a large part of this functionality.
In Copenhagen, for example, activation of a smoke alarm on a train triggers automatic extension of the platform dwell, i.e. the time the trains spend waiting at the platform, for all trains already in the station and for the train following behind.
Other trains approaching will be non-stop through the station on the opposite track. Operators in the control room have the option to decide which ventilation strategy to implement, which then triggers down escalators to stop, ventilation fans and dampers to start up, emergency direction signs, pressurisation of smoke-free escape routes and automatic notification to the fire brigade.
Such integration requires substantial interface management which can generate significant risk in the project and is typically a key cost driver.
For this reason, contract models for automated metro projects increasingly utilise system packages which include the rolling stock, so that most of the interface management is contained within the contract scope.
Transmitting safety critical information using wifi would appear to offer an easily accessed vulnerability and pose a serious cyber threat.
The industry has recognised this risk and multilayered approaches are incorporated into the design and operation, consistent with IEC 62443.
This makes it as difficult as possible for hackers to detect, decode and otherwise interfere with the data passing between the train and the wayside or elsewhere.
Fortunately, development of control systems is now largely undertaken using simulator rigs that are able to closely model the real world. In Copenhagen, this enabled the CBTC system to be extensively developed and tested before the first real train ran under CBTC supervision on the test track about two years before opening.
As a result, functional and integration testing went much more smoothly than in previous generations of signalling control systems because the majority of tests had already been executed at least once on the various simulators at the supplier’s premises before being attempted in the field.
Undoubtedly, the contribution of skilled and experienced engineers onsite during the commissioning period was vital to the rapid progress from software development through to the start of revenue service.
This process was assisted by the improved diagnostic and data logging systems that are now much more sophisticated than in previous generations, making identifying and correcting problems much faster.
A short trial running period without passengers prior to the opening date also provided reassurance that the overall system was capable of revenue operation.
The impact of the 21 construction sites in the city centre over the seven years of construction was, at times, hard on the local inhabitants and businesses. Expectations were high that the Cityring would adequately compensate for the sacrifice.
With the original metro line performing a service availability of a lot more than 99% (actual versus planned departures), it was publicly assumed that the Cityring would deliver the same.
Complex systems such as these usually have multiple issues at start-up, arising from the high part count and the possibility of undetected systematic failures that can bring the service to a halt.
Largely due to the breadth of testing, comprehensive diagnostics and the use of digital systems, start-up problems were minimised and the Cityring service immediately performed in line with expectations.
Since 1999, Copenhagen has undergone a renaissance that has revitalised the inner city. The construction of the first metro line and now the Cityring line has been an essential ingredient in opening up the city, allowing fast and easy movement for inhabitants.
Digitalisation has been at the heart of the automation of these new railway lines and was a fundamental part of delivering an exciting and innovative transport system to the city of which Copenhageners are justly proud.
Author: Chris Cox is an associate director at Arup. He has more than 30 years’ experience in the railway industry, specialising in systems engineering and technical management roles. He has worked extensively with the design and development of communication based and track-circuit based signalling, particularly in automatic and unattended (GoA4) train operation modes. He has significant experience of bringing railway control systems into revenue operation, maintenance, systems development and managing reliability growth programmes. Prior to joining Arup, he was the Cityring Transportation Systems project director at Metroselskabet. He is a Chartered Member of Engineers Ireland and a Fellow of the Institute of Railway Signal Engineers.