Is the European Train Control System (ETCS) ever going to be implemented?

[tomuk]

Disagree, we're in the current mess because there's a lack of competition. Basically the mainline has a choice between Hitachi or Siemens, with other companies being largely confined to mass transit/light rail or the margins of the network. They have no interest in taking each other's jobs, nor of allowing anybody else to muscle in so they are incentivised to capture the installation.

Are Bombardier Electrostars, Aventras and Bombardier Turbostars not the most numerous trains on the network ?

 

[eldomtom2]

Decades of a vertically integrated railway has lead us to the steam-ERA AWS, primitive TPWS, and "ATP would have been be nice but it's too complicated and expensive". The strategy is a proven failure.

AWS and TPWS are doing their job very well. The case for ETCS rests more on it enabling service improvements than safety.

 

[AM9]

AWS and TPWS are doing their job very well. The case for ETCS rests more on it enabling service improvements than safety.

ETCS does also improve safety on routes with speed limits greater than 75mph, (which affects most services except on a few rural and metro lines) without capacity sapping double blocking needed to provide protection at slightly higher speeds.

 

[zwk500]

Correct, there's nothing special in the UK network that makes ERTMS unsuitable in it's current and developing forms and not suitable for integration into both existing and new rolling stock. In fact the only 'special' thing about the UK rail system is that it is a hotch potch of systems that have been added piecemeal through lack of investment. An international standard is both cheaper and more accessible than home brew bodges because it has been developed by numerous supplier to common standards.

Technical point, ERTMS is the Traffic management system, of which ETCS is only one part. ERTMS has a lot of stuff that only really affects Train Running Control on top of the signalling system.

The difficulty with implementing ETCS in the UK has several factors, core among them are the restricted ability to gain operational experience with the system, how badly a lot of the technical data is stored, the mixed imperial/metric legacy of our system, and the need to be able to tie ETCS into a number of different systems of differing ages and design.

However a big thing is simply experience (and this is not unique to the UK at all) - because we haven't installed much ETCS, there isn't the institutional knowledge among designers, engineers, drivers and signallers to enable a swift rollout. Each new section now needs long lead times to train staff in ETCS standards, practices and usage and there isn't somebody in the office/depot who's done it for 40 years to ask for help. The issues when ETCS was rolled out on the Thameslink Core speak to this - they were quickly bedded down as the line is intensively used but they were important lessons. It's also why the Hertford loop was used for testing for ages and why the Moorgate line is doing overnight testing now. As more lines open, familiarity will allow the speed of installation to be picked up. Usage of Eurobalises for TASS, APCOs, some ASDO systems will also contribute to experience, as well as continental experience but the big gains will be when we have a major mainline in full signals-away operation.

AWS and TPWS are doing their job very well. The case for ETCS rests more on it enabling service improvements than safety.

ETCS does turnout speeds which TPWS doesn't do. The recent Lumo Overspeed wouldn't have happened with ETCS, and (in theory) ESRs and TSRs will also be supervised by ETCS (I'm aware of incidents on the Cambrian when the download didn't happen) which is a step above a temporary AWS magnet and better than TPWS which doesn't do them at all.

TPWS has done the job it was meant to do very well, but it was always a step below an ATP system (because it was designed to be cheaper) and should really have been accompanied by a programme to fit ATP to the major lines.

 

[D365]

ETCS is being installed. If Network Rail wanted to install it faster they could build their own installation team and buy the hardware off the shelf. It has nothing to do with developing a completely new system at a probably 9 figure cost, specifically for UK use and nowhere else. That would be utterly bonkers.

In effect, that’s what Network Rail is doing right now, by means of various design/install contracts that fall under the East Coast Digital umbrella.

 

[43066]

TPWS has done the job it was meant to do very well, but it was always a step below an ATP system (because it was designed to be cheaper) and should really have been accompanied by a programme to fit ATP to the major lines.

It’s probably hard to convincingly make that argument on a cost-benefit basis. Since TPWS was introduced in the early 2000s, how many lives have been lost that would have been saved had ATP also have been rolled out?

 

[Tester]

It’s probably hard to convincingly make that argument on a cost-benefit basis. Since TPWS was introduced in the early 2000s, how many lives have been lost that would have been saved had ATP also have been rolled out?

<ironyfont>...ahh but if it saves a single life...</ironyfont>

 

[zwk500]

It’s probably hard to convincingly make that argument on a cost-benefit basis. Since TPWS was introduced in the early 2000s, how many lives have been lost that would have been saved had ATP also have been rolled out?

And yet, Belgium, France, Spain, Italy, Netherlands, Germany and Switzerland all thought it worth implementing ATP-equivalent systems. However the cost-effectiveness is exactly why TPWS was chosen.

 

[MarkyT]

It’s probably hard to convincingly make that argument on a cost-benefit basis. Since TPWS was introduced in the early 2000s, how many lives have been lost that would have been saved had ATP also have been rolled out?

TPWS was the only system ready to go when it was needed in a hurry post-Ladbrook Grove and Southall. It was simple to interface to both train and infrastructure, and the programme's key cost and time-saving decision as far as national rollout was concerned was to only equip signals on the network that posed the most significant risk if passed at danger, mainly those protecting against collisions at junctions that weren't already mitigated by flank protection. This policy was informed by a deep understanding of signal overrun risk, about which the industry had developed a sophisticated risk assessment methodology. TPWS has drawbacks clearly beyond its limited fitment. It is a dumb system that can only check speed instantaneously at an overspeed trap and has no state memory or functionality to prevent reacceleration after passing without intervention. The equipment is not intrinsically failsafe either. A local power failure at the lineside equipment cabinet means no signal from the grid and hence no protection; that's partly mitigated by proving the equipment energised with oscillators functioning in the aspect of the previous signal on approach, as well as providing failure alarms to signallers.

Strategically, TPWS has bought the industry time to understand and become proficient at implementing ETCS in line with planned signal system renewals. Twenty years ago the decision to go for an ETCS-based system might have been made, but the tech was fairly cutting-edge at the time and has come a long way since in terms of techniques and supplier capability. The specifications have also been developed with each 'baseline'. Sometimes it is worth waiting for a new development to become more mainstream and mature rather than diving in immediately as a risky early adopter. Another factor is that the long wait has given time for older fleets of trains to retire. The non-computer-based controls of older trains are more difficult to interface to the onboard ETCS equipment, as, like other forms of full supervision ATP, to get full performance a tight precision control loop is required to finely monitor and adjust speed. Conversely, the delay in implementation has allowed time for companies to come up with innovative solutions for fitting older traction, track machines etc. with limited functionality yet compliant onboard systems that are effective and affordable, like those fitments being made to some mainline certified steam locos currently for continued ability to run excursions on the ECML, UKs first major mainline application of the technology.

And yet, Belgium, France, Spain, Italy, Netherlands, Germany and Switzerland all thought it worth implementing ATP-equivalent systems. However the cost-effectiveness is exactly why TPWS was chosen.

UK had an opportunity to go for a eurobalise-based limited supervision system from Siemens when the later phase of national fitment, TPWS+, was being developed. It had no chance of being cost-effective though, as the TPWS equipment was already fitted to the trains so a solution that simply provided more compatible track grids was bound to win. A trial of the active balise-based kit took place near Slough on the GWML however in the early noughties with a HST power car having the new onboard equipment. Although still a 'limited supervision' concept, the more failsafe processor-based box of tricks onboard could provide additional protection such as preventing reacceleration etc. I still think TPWS would have been better with digital balises like that, just as all those continental railways have implemented as a stepping stone to ETCS, but the analogue TPWS tech we have today was a solution fully tested and ready to go when it was needed so that is what we are stuck with. The main takeaway is that TPWS has provided a great improvement in safety while buying the UK time to plan rationally for future ERTMS.

 

[43066]

And yet, Belgium, France, Spain, Italy, Netherlands, Germany and Switzerland all thought it worth implementing ATP-equivalent systems. However the cost-effectiveness is exactly why TPWS was chosen.

Was that decision based on safety, or retirement of legacy systems/performance, though? I have always understood the UK railway in its current format is safer than most others in Europe (I’m not sure whether there are statistics to back that up, though).

TPWS was the only system ready to go when it was needed in a hurry post-Ladbrook Grove and Southall. It was simple to interface to both train and infrastructure, and the programme's key cost and time-saving decision as far as national rollout was concerned was to only equip signals on the network that posed the most significant risk if passed at danger, mainly those protecting against collisions at junctions that weren't already mitigated by flank protection. This policy was informed by a deep understanding of signal overrun risk, about which the industry had developed a sophisticated risk assessment methodology. TPWS has drawbacks clearly beyond its limited fitment. It is a dumb system that can only check speed instantaneously at an overspeed trap and has no state memory or functionality to prevent reacceleration after passing without intervention. The equipment is not intrinsically failsafe either. A local power failure at the lineside equipment cabinet means no signal from the grid and hence no protection; that's partly mitigated by proving the equipment energised with oscillators functioning in the aspect of the previous signal on approach, as well as providing failure alarms to signallers.

Strategically, TPWS has bought the industry time to understand and become proficient at implementing ETCS in line with planned signal system renewals. Twenty years ago the decision to go for an ETCS-based system might have been made, but the tech was fairly cutting-edge at the time and has come a long way since in terms of techniques and supplier capability. The specifications have also been developed with each 'baseline'. Sometimes it is worth waiting for a new development to become more mainstream and mature rather than diving in immediately as a risky early adopter. Another factor is that the long wait has given time for older fleets of trains to retire. The non-computer-based controls of older trains are more difficult to interface to the onboard ETCS equipment, as, like other forms of full supervision ATP, to get full performance a tight precision control loop is required to finely monitor and adjust speed. Conversely, the delay in implementation has allowed time for companies to come up with innovative solutions for fitting older traction, track machines etc. with limited functionality yet compliant onboard systems that are effective and affordable, like those fitments being made to some mainline certified steam locos currently for continued ability to run excursions on the ECML, UKs first major mainline application of the technology.

Interesting points. TPWS plus defensive driving has certainly achieved a lot of “bang for the buck”, in terms of safety. I understand the main drivers of ETCS installation are capacity rather than safety.

 

[D365]

I understand the main drivers of ETCS installation are capacity rather than safety.

plus moving a lot of the maintenance liability off the lineside and onto the vehicles

 

[zwk500]

Was that decision based on safety, or retirement of legacy systems/performance, though? I had understand the UK railway in its current format is safer than most others in Europe.

I'm fairly sure the Dutch ATB (which is nearly universal on the NS network) and French KVB (which is on lines with the overwhelming majority of traffic) were in response to significant collisions. Don't know about the others. Not sure if their networks had other systems that would have prevented TPWS, or if the technology wasn't available when ATB was first installed.

Interesting points. TPWS plus defensive driving has certainly achieved a lot of “bang for the buck”, in terms of safety. I understand the main drivers of ETCS installation are capacity rather than safety.

The safety is an important part of it because it does a number of things TPWS can't do, and part of the capacity benefits are related to the safety applications. E.g., the supervision of the braking curve allows safety sections (overlaps) to be shortened or done away with, so that standage at platforms can be maximised and junction occupation times are reduced. Also the abolition of approach control because of the supervision of the speed for the diverging route means trains don't have to slow down more than necessary and then accelerate again. Another capacity benefit released by the safety controls is that block lengths are not linked to braking distance.

 

[43066]

I'm fairly sure the Dutch ATB (which is nearly universal on the NS network) and French KVB (which is on lines with the overwhelming majority of traffic) were in response to significant collisions. Don't know about the others. Not sure if their networks had other systems that would have prevented TPWS, or if the technology wasn't available when ATB was first installed.

Interesting, thanks. No doubt higher levels of subsidy in Europe would also have been a factor.

The safety is an important part of it because it does a number of things TPWS can't do, and part of the capacity benefits are related to the safety applications. E.g., the supervision of the braking curve allows safety sections (overlaps) to be shortened or done away with, so that standage at platforms can be maximised and junction occupation times are reduced. Also the abolition of approach control because of the supervision of the speed for the diverging route means trains don't have to slow down more than necessary and then accelerate again. Another capacity benefit released by the safety controls is that block lengths are not linked to braking distance.

ETCS is undoubtedly safer, I’m just not sure it (or equivalent ATP systems) offer enough of a safety improvement over TPWS to really support the suggestion that the UK should have gone with an older system around the turn of the century.

@MarkyT’s points above make a lot of sense in this regard. The benefits will come in time but it’s telling that, even with ETCS now a well proven system, the main driver behind installing it in the UK remains efficiency/capacity rather than safety per se.

 

[zwk500]

Interesting, thanks. No doubt higher levels of subsidy in Europe would also have been a factor.

Also the different operating rules and traffic patterns in Europe. AIUI Beligum made do with the Crocodile (functoinally similar to AWS) for a very long time and had only started to install TBL on a few lines when ETCS began to mature.

ETCS is undoubtedly safer, I’m just not sure it (or equivalent ATP systems) offer enough of a safety improvement over TPWS to really support the suggestion that the UK should have gone with an older system around the turn of the century.

I think TPWS was the right call for something that could be rolled out nationally in short order to fill an important safety hole. The safety benefits alone of ETCS/ATP clearly didn't justify replacement of TPWS, but although I can't offhand think of any incidents that would have been prevented by ATP, I do think wider operational experience of such a system would have been useful and the additional safety would have been justified on the ECML and WCML core routes and maybe on certain branches, like Leeds and Manchester, as well as possible edge cases for the MML, GEML and SWML.

As I mentioned, the UK rail industry is having to build up a completely new institutional capability for ECTS from scratch. The Netherlands, France and Germany are in a similar situation with limited installations whilst training and familiarisation are ramped up. Compare it to Belgium, which started aligning TBL+ towards ETCS standards and framework a while ago and now has a significant proportion of it's network covered by L1 and all but one of it's HS Lines covered by L2 (HSL-1 to Lille is still TVM). A good map is here: https://www.openrailwaymap.org/.

@MarkyT’s points above make a lot of sense in this regard. The benefits will come in time but it’s telling that, even with ETCS now a well proven system, the main driver behind installing it in the UK remains efficiency/capacity rather than safety per se.

Indeed. It's notable that the majority of installations in Europe have been on new-build HS Lines, with conversion of existing systems or fitment of existing lines wihtout an ATP comparatively rare.

 

[MarkyT]

I'm fairly sure the Dutch ATB (which is nearly universal on the NS network) and French KVB (which is on lines with the overwhelming majority of traffic) were in response to significant collisions. Don't know about the others. Not sure if their networks had other systems that would have prevented TPWS, or if the technology wasn't available when ATB was first installed.

Both these systems and the Swiss Eurosignum/ZUB were an attempt to escape from a custom proprietary physical interface used in their previous legacy protection systems at a time the administrations had decided to go for a much wider rollout, possibly as you suggest in response to incidents. The Swiss system, on which I read up extensively at the time, had a simple onboard system installed on older traction that replicated legacy distant, overspeed and trainstop functionality in the cab although was more advanced 'under the hood'. More modern fully ETCS-equipped traction could emulate this mode using their standard balise readers and the Swiss even got the standards updated some years later to incorporate their limited supervision level 1 (L1 LS) functionality in newer baselines, so foreign traction could easily use the system. They were thus the first nation to declare their standard gauge network fully compliant for European interoperability in 2017 (even though they're not in the EU!). There's a similar Belgian limited protection system, TBL1+, now also fully converted to balises as part of a wider rollout, that in earlier iterations orginally used the same track transponders supplied for the British Rail GWML ATP pilot scheme, still in use of course and now hampered by a supplier long since defunct and being the only implementation of this equipment remaining on the entire planet.

 

[snowball]

Even now the case for ETCS in the UK is sufficiently non-overwhelming that Scotland seems to think it isn't yet worth the money.

 

[The Planner]

Even now the case for ETCS in the UK is sufficiently non-overwhelming that Scotland seems to think it isn't yet worth the money.

Disagree, plan is to redo Warrington, Preston and Carlisle over the next couple of control periods.

 

[snowball]

Disagree, plan is to redo Warrington, Preston and Carlisle over the next couple of control periods.

Warrington, Preston and Carlisle are not in Scotland.

 

[The Planner]

Warrington, Preston and Carlisle are not in Scotland.

You said the UK?

 

[snowball]

You said the UK?

Sorry, maybe I shouldn't have mentioned the UK. I did also mention Scotland. I was thinking of the fact that the recently published Scottish HLOS says

The Scottish Ministers have considered carefully the planned approach to signalling investment elsewhere in Great Britain for CP7, but consider that it does not align with Scotland’s strategic priorities at this time.

In particular, the Scottish Ministers consider that no business case exists for the European Train Control System (ETCS) Level 2 in Scotland at this time, as the railway traffic characteristics and capacity issues are not the same as those for which this system is more effective. Further, that the potential benefits of this system may be secured more cost effectively, more quickly and at lower risk by other investments.

I suppose I assumed, perhaps wrongly, that the case for ETCS level 2 is objectively equally strong or weak in Scotland as in England, and that decision makers were making a different assessment of it in the two places. But it could be instead that the different circumstances in the two countries make the case objectively stronger in England (including Warrington, Preston and Carlisle) than in Scotland.

 

[MarkyT]

Sorry, maybe I shouldn't have mentioned the UK. I did also mention Scotland. I was thinking of the fact that the recently published Scottish HLOS says



I suppose I assumed, perhaps wrongly, that the case for ETCS level 2 is objectively equally strong or weak in Scotland as in England, and that decision makers were making a different assessment of it in the two places. But it could be instead that the different circumstances in the two countries make the case objectively stronger in England (including Warrington, Preston and Carlisle) than in Scotland.

That statement doesn't discount a Level 1, full or limited supervision mode system, nor a more distant future Level 3-based rollout. Perhaps sensibly they are waiting in the light of not having the major pressure of large signalling renewals pending and for manufacturers to innovate a more affordable rural solution within the ETRMS sphere. A policy of making sure any new conventional work in the Central Belt is carried out 'ETCS - ready' should be sufficient in the short and medium term to ensure compatibility, then route-based switchovers can happen when rolling stock is all suitable. RETB-controlled routes in the Highlands have been re-equipped fairly recently and don't need redoing again with a more expensive system (as Cambrian proved).

 

[zwk500]

I suppose I assumed, perhaps wrongly, that the case for ETCS level 2 is objectively equally strong or weak in Scotland as in England, and that decision makers were making a different assessment of it in the two places. But it could be instead that the different circumstances in the two countries make the case objectively stronger in England (including Warrington, Preston and Carlisle) than in Scotland.

The case for ETCS L2 varies from line to line across both countries. I wouldn't think the business case for rural lines in Scotland is intrinsically different from, say, the Settle and Carlisle.