Train protection
Britain has been running automatic trains since the 1960s, but only on closed networks such as London Underground. PHILIP HAIGH discovers how they could be expanded onto Network Rail’s main line tracks
Automatic trains have been running on closed networks since the 1960s, but could now be expanded onto the main line.
One of the joys of growing up in the 1970s was being able to see out of the cab of a British Rail diesel multiple unit. Their windows at the front of the passenger saloon allowed a full view of the cab and the road ahead - unless the driver liked his privacy and dropped the blinds.
To see such views today you need to travel on the Tyne and Wear Metro or on a preserved railway. When BR replaced its first generation DMUs with Pacers and Sprinters, the forward one went, never to return.
There is, however, one other railway that lets you look forward… Docklands Light Railway. DLR doesn’t have driving cabs because it doesn’t have drivers. DLR’s trains are automatic with a crewman provided to take control if needed. Which isn’t often.
Travel on London Underground’s Victoria, Central, Jubilee or Northern lines and you’ll be on an automatic train, albeit with a driver. Indeed, the Victoria Line has always been automatic.
What these four LU lines and DLR have in common is that they’re self-contained railways that operate only one type of train. But out on the national network, it’s a different matter. Expresses running at 125mph mix with 2,000tonne freights. Some stop quickly, others take longer.
Train drivers need to know the line they’re driving. They must know what speed they must run. They must know where signals are placed and what their lights and indicators mean. They must know gradients because they can change brake and acceleration rates. They must know where to brake and stop for stations. They must know where they are. And all at night.
An automatic train also needs to know all this, except for signals because it will receive the information that signals convey electronically rather than visually.
There is growing interest in running automatic trains on the national network, such that the Institution of Mechanical Engineers (IMechE) and Institution of Railway Signal Engineers (IRSE) held a one-day conference in mid-February to discuss the way forward.
Network Rail’s Digital Railway Chief Systems Engineer Andrew Simmonds explained the reasons for the interest in Automatic Train Operation (ATO). He told delegates that train and track reliability in Britain was improving, but with more trains running the delays each incident caused were increasing. “The slightest disruption and it gets fairly messy out there,” he said.
Against a backdrop of a rising population and house prices forcing people to live further from their work (particularly for London), it’s reasonable to assume that demand for rail will continue to increase.
For Simmonds, this presents four options. The railway could do nothing, and so rely on roads to take the increased traffic. It could deliver incremental rail capacity increases through less disruptive conventional enhancements (essentially do a little bit). It could deliver a system upgrade using digital command and control technology. Or it could implement very disruptive conventional enhancements (such as the long and expensive West Coast Route Modernisation). “At Digital Railway, we think the third option is worth looking at,” he commented.
London Underground has delivered system upgrades to its four automatic lines. The Victoria Line has new stock running under a new signalling system and the other lines have something similar, albeit retaining their already reasonably modern stock. The result was an increase in capacity of 20%-40%.
Digital Railway (DR) is already looking to implement European Train Control System (ETCS) signalling. Simmonds’ boss, NR’s Digital Railway Managing Director David Waboso, has told RAIL that he is looking towards ETCS Level 2 with extra track circuits to increase capacity ( RAIL 819). Level 2 brings the prospect of cab signalling, rather than relying on lineside lights on sticks. Trains would receive signalling information via balises on the track.
At the same time, DR is looking to develop
If the driver doesn’t know what the train management system is thinking, there’s little improvement. So go for ATO and a link with train management. Andrew Simmonds, Digital Railway Chief Systems Engineer, Network Rail
train management (TM) systems that can better plan services and recovery from disruption. But it has not found this area easy, with delays to pilot projects at Romford, Cardiff and Three Bridges.
TM is the key to increasing capacity through Thameslink’s central core between St Pancras and Blackfriars. Trains will leave the core for a variety of destinations, much as water sprays from a hose. Feeding the spray into the hose at the other end of the core is much harder, and this is where TM can help by adjusting the position of trains so that they feed seamlessly into the central section at 24 trains every hour.
So far, so good. But as Simmonds told the conference: “If the driver doesn’t know what the train management system is thinking, there’s little improvement. So go for ATO and a link with train management.”
This needs a very reliable communications link, as well as accurate data about both trains and track. It needs the railway to measure punctuality to the second, which will demand a major cultural change.
ATO, argued Simmonds, gives predictable train running and an ability to react to timetable changes. Because it’s more predictable, it should allow closer running, finer regulation of trains, reduced timetable allowances and reduced energy consumption, he said.
Alstom ATO System Manager Benoit Bienfait gave the conference examples of how ATO increases capacity. Every driver drives differently - the differences might be small, but they must be allowed for in timetables.
Bienfait took as his example the six-track line between Brussels North and South. Current capacity is 92 trains per hour (tph), with a spare path inserted every four trains in order to maintain reliability. With consistent driving from ATO, he argued that the spare path could be removed - increasing capacity to 120tph.
Some drivers generate better fuel
consumption figures than others. Just as with a car, much depends on the way they drive - harsh acceleration will increase fuel use. Alstom’s analysis suggests potential fuel savings of up to 42% for local trains (which stop frequently) and up to 15% for inter-city trains.
If a driver sets an ETCS train in motion, the ETCS kit will allow it to proceed only as far as the signalling says is safe. If the driver does nothing, it will stop at the next red signal (whether that’s a real signal or a marker board that represents the end of the train’s movement authority). The train stops because it knows where it is and the point beyond which it cannot go.
That’s not really different from instructing a passenger train to stop at a station. Both control where a train should stop, and for this reason they should both use the same principles and communications language.
But if there are similarities between ETCS and ATO, it’s also true that they should be separate. ETCS is the safety net, the limits beyond which a train must not go (in terms of speed or distance). ATO is the control of the train that must remain within the ETCS safety net.
At the European Union Agency for Railways (previously the European Rail Agency), ATO Project Officer Wouter Malfait argues that ATO is to make rail more competitive and only a little bit about making it safer. He says that ATO increases capacity, safety, performance,
efficiency, punctuality and passenger comfort. It decreases costs, energy consumption, rail noise and pollution.
From a European perspective, the agency wants to avoid 25 different types of ATO with no interoperability between them. This is similar to the principles behind ETCS, which include the ability for a train equipped with ETCS in one country to operate on another country’s ETCS. This is particularly important for cross-border freight, and to a lesser extent passenger services. Having one standard for ATO to which several manufacturers’ equipment complies also brings a more open market.
For rolling stock owners this is important, Angel Trains Head of Product Management Euan Smith told the conference. Citing an RSSB project on standard couplers, he said that at privatisation there were three types of coupler while today there are 12. Even Class 220 and 221 Voyagers and Class 222 Meridians could not run together, despite being an outwardly similar train built by the same company. He warned against having different systems on different trains on different tracks, saying this would lead to difficulty interworking fleets and complicate fleet cascades or upgrades.
The lengthy and convoluted development of ETCS’s specification and standards shows that this admirable goal is easier to say than to deliver. Even as Britain leaves the European Union, a standard ATO has advantages in that Network Rail and train operators would not be in hock to a single supplier, as they would be in buying a proprietary system. (Thameslink uses a proprietary system because common ATO standards are still being developed.)
Malfait is determined to see a stable specification for ATO, because he recognised at the February conference that having to upgrade or retrofit kit would ruin the business case for ATO (which is chiefly driven by capacity increase and cost reduction).
Among the challenges is the need to agree standard trackside information that tells the train what’s down the line - in terms of gradients, for example. Unlike a Victoria Line train that can be programmed with the entire line’s characteristics, a go-anywhere European locomotive cannot hold on board full details of every line in Europe. This is one of the differences between metro and heavy rail operation that makes ATO harder for main line railways.
Information is vital to ATO, just as it is for a driver. Bombardier ATC System Engineer Jack Ratcliffe noted: “An ATO is only as good as the information you put in it.”
Britain faces further challenges because its railway is split between many companies and Network Rail is pushing towards route devolution. Delegates at February’s conference were concerned that there was no guiding mind and no central authority. They were also concerned that the costs would fall to train operators and owners while the benefits would go to Network Rail (not least in terms of having more capacity to sell).
Simmonds said that eight route studies would be completed by April, and that outline business cases at route level would incorporate whole-industry costs. It would then be for the Department for Transport to decide the best way to fund improvements over a whole-life basis. This might be via Network Rail or franchises, he suggested.
He called for new trains to be delivered ready for Digital Railway. Whether ETCS or ATO they should have common cabling, for example, to help reduce the costs of fitting equipment later. This should be agreed with the supply industry, he said.
The challenges of European co-operation and standards creation are not the only ones facing ATO. Unlike the subterranean Victoria Line, the national network is subject to a variety of weather that generates a variety of railhead conditions. It’s well known that autumn causes problems with traction and adhesion, as leaf mulch coats the railhead.
Companies tell drivers to brake more slowly, and they extend timetables to maintain punctuality. Despite this, performance figures drop every autumn. Thameslink plans to use an ETCS communications channel (packet 44) to instruct its Class 700s running in automatic mode to brake more slowly on a Monday morning, when engineering work might have closed parts of the track leading to rust on the railhead, according to Engineering Manager Jon Hayes.
Just as drivers are taught to recognise and react to poor braking conditions, so could ATO trains - provided they have a suitable way of telling the train management system in the local control centre. Once again, this is easy to say, but much harder to reliably and safely react to poor railheads.
Consistent driving style is one of ATO’s advantages, but it’s not always good for the track because each type of train will brake or accelerate at the same point. This can lead to defects and rail wear being concentrated on short sections.
Door control may also prove contentious. An on-board attendant might check that all doors are safely closed at each station before pressing a button to tell the train it can move. Or staff on the platform could check and signal via the control centre. The train might use sensors to check doors, but these sensors will need sufficient resolution to ‘see’ a trapped rucksack strap, arm or coat in a door. This should not be impossible, but (as with track conditions) it will take considerable effort to assure passengers and staff that it’s safe.
Further complications could come with platform edge doors, which will need to be linked to train systems to ensure safe departure from stations. Trains must also stop precisely so that train and platform doors align. This is easier if a uniform fleet calls at that platform. If not, then Britain might have to install moveable platform doors (as Japan has).
LU’s Jubilee Line already uses platform doors on its extension east to Stratford. Crossrail, which will use ATO in its central section, also has platform doors in that section. Its trains and drivers will need to switch from ATO to manual driving under standard signalling on the eastern section and manual driving under ETCS Level 2 on the western section.
Sensors, perhaps radar, will play a role in detecting obstacles in the track, whether they’re objects or people who should not be there. In recent years, lorries have fallen from bridges, containers have been blown from freight trains, and (as at Liverpool in February,
RAIL 822) walls have collapsed onto the track. Drivers can spot obstacles but they can’t always stop their trains in time. An ATO system must also be able to spot obstacles and brake, but it too might not be able to stop short.
Preventing trespassers can cut this risk. This demands improved management by Network Rail of boundary fences. NR would also need to improve the way it manages and maintains structures (to prevent collapse) and vegetation (to prevent trees falling onto the line, for example). ATO should trigger changes to the way NR grants access to track workers. encountered many of the problems that would face the national network in bringing ATO into use.
While the Victoria Line might not have adhesion problems caused by weather, the Central Line certainly does. Head of Rolling Stock Graham Neil told the conference that LU uses a combination of information from the London Weather Centre, a network of lineside moisture monitors and information from train wheelslip protection ( WSP) monitors.
From this LU predicts likely rail adhesion and reduces brake rates or speeds, or reverts to manual driving if conditions are beyond ATO’s ability. Accordingly, it follows that a line cannot be used to its full capacity if it must cope with trains taking longer.
Imagine the size of such a monitoring operation if it must cover all Network Rail’s lines, rather than just the parts of LU that extend beyond the central tunnels.
Neil revealed some of the questions that must be answered before ATO can be implemented. They centre on ‘what if’ for a variety of scenarios.
If the driver were to be incapacitated, does the train stop or continue to the next station? What should the train do if a passenger pulls an alarm handle - stop or continue to the next station? Should a train stop automatically if something is caught in a door? Train fires - stop or continue? If platforms have emergency stop buttons, should the train stop? What happens if the traction supply fails?
If a company opts for GoA2 (see panel,
An ATO is only as good as the information you put in it. Jack Ratcliffe, ATC Systems Engineer, Bombardier
not driving the train. This begs the question of how that driver retains his ability and how the driver remains alert to the need to take over. Victoria Line trains work under driver control to and from their depot.
Nevertheless, Neil was enthusiastic: “The benefits are great. They’re absolutely fantastic. We have virtually no SPADs (signals passed at danger) under ATO. When we do they’re usually from a system fault, not the driver.”
Consistent driving style is one of ATO’s advantages, but it’s not always good for the track because each type of train will brake or accelerate at the same point. This can lead to defects and rail wear being concentrated on short sections.
Also tricky will be the decision on how trains should perform. Should they brake or accelerate to the best of their ability, or more gently to save fuel or reduce brake pad wear? Or should trains try to keep as close to the timetable as possible?
Is it better to run at line speed, stop at a signal protecting a junction and then restart after another train has crossed, or run slowly and not need to stop through the junction? Or brake differently and approach the signal slowly, reaching it just as its clears from red? The answer will depend on what’s important to the railway.
Similar questions affect the development of Connected Driver Advisory Systems (CDAS). They take the DAS idea of offering drivers advice on the optimum speed to keep to the timetable by adding a link to the signalling system. In this way, CDAS can advise a driver the best speed that will prevent him having to stop at a blocked junction, which should smooth the flow of trains. ATO could be considered a link between CDAS and the traction and brake controls of a train.
ATO offers the possibility of finer control than a driver can achieve because it can more clearly know what the train management system wants it to do. It could receive information about its ideal speed more frequently than a driver could from signals spaced every mile or so. In this way an ATO train is more accurately a computer-controlled train, even though the computer might be many miles away in a control centre.
Although delegates attending February’s conference were confident that we would see main line ATO in Britain, it’s clear that there is plenty of difficult work to do before it becomes normal.
Europe must agree clear specifications, manufacturers must develop products, Britain must resist the temptation of adding bespoke features to those products, and passengers must be convinced that ATO is safe.
Of course, if there’s no forward view, passengers might not even realise there’s no driver. Personally, I’d trust the technology and enjoy the view!
ATO offers the possibility of finer control than a driver can achieve because it can more clearly know what the train management system wants it to do.