Cambrian Line Radio Signalling failure - RAIB investigating

[iands]

Nicely filmed. Funniest thing I've seen for a while.

 

Anyone know where it is? The YouTube page suggests it's in the Wirral, which it clearly isn't. I'm pretty sure I've seen maintenance vehicles in Slovakia looking like that, and the level crossing seems the right style (can't read the writing under the sign, but the word lengths look right for "Pozor Vlak" (beware of trains).

 

Curious there seems to be a railway signal on the far side of the crossing for trains approaching it.

 

And to get a bit more on topic...

 

Regarding the comment previously about ERTMS (or ECTS if you like) assuming the maximum train length, if the speed restriction is to slow the approach to a level crossing, why does the train length matter anyway? Clearly it does if you're slowing for a bad bit of track, but if the idea is to give sufficient warning at a level crossing, surely it's the front of the train that matters not the back? Once it's on the level crossing from a sighting point of view it shouldn't matter what speed it's doing.

 

With a 'manual' speed restriction for a level crossing, is the driver expected to wait until the back of the train has cleared the sign before accelerating?

All TSRs (and ESRs) will have a "commencement" and "termination" - i.e. a defined length. By inputting the train length into ERTMS at the start of a journey, the system can calculate when it is okay to accelerate after a TSR (by distance travelled). However, from Big Jim's earlier post it would seem a bit more work is required on this as it would seem to be a bit more restrictive than perhaps anticipated.

 

Regards, Ian.

[Grovenor]

Regarding the comment previously about ERTMS (or ECTS if you like) assuming the maximum train length, if the speed restriction is to slow the approach to a level crossing, why does the train length matter anyway? Clearly it does if you're slowing for a bad bit of track, but if the idea is to give sufficient warning at a level crossing, surely it's the front of the train that matters not the back? Once it's on the level crossing from a sighting point of view it shouldn't matter what speed it's doing.

You are correct, however, I suspect that the system does not have a TSR definition for specifically for level crossings so the definition meant for bad track is being used, hence havng a minimum length equal to the train length.New feature needed!

regards

[big jim]

All TSRs (and ESRs) will have a "commencement" and "termination" - i.e. a defined length. By inputting the train length into ERTMS at the start of a journey, the system can calculate when it is okay to accelerate after a TSR (by distance travelled). However, from Big Jim's earlier post it would seem a bit more work is required on this as it would seem to be a bit more restrictive than perhaps anticipated.

 

Regards, Ian.

Yeah, currently (on the 97s at least) you can’t input train length, ‘standard lengths’ were pre inputted years ago and have never been changed

 

I’ll have to ask regards 158 as I’m sure you can put 2,4 or 6 car lengths in

[Coryton]

You are correct, however, I suspect that the system does not have a TSR definition for specifically for level crossings so the definition meant for bad track is being used, hence havng a minimum length equal to the train length.New feature needed!

regards

 

If that's right, it looks like the sort of thing that shouldn't need a trial to realise is missing.

 

As with Big Jim's problems getting into a siding, shouldn't some kind of tabletop exercise have brought these issues out long before implementation?

 

Maybe somebody should have built a big model raliway and implemented ERTMS to find these things out.

 

All TSRs (and ESRs) will have a "commencement" and "termination" - i.e. a defined length. By inputting the train length into ERTMS at the start of a journey, the system can calculate when it is okay to accelerate after a TSR (by distance travelled). However, from Big Jim's earlier post it would seem a bit more work is required on this as it would seem to be a bit more restrictive than perhaps anticipated.

 

Regards, Ian.

 

Yes I appreciate that.

 

But why does a speed restriction for level crossing sighting purposes have to work that way?

 

(Edited to remove duplicate text)

Edited February 25, 2018 by Coryton

[Reorte]

...ERTMS (or ECTS if you like)...

I've never got this straight - are they two different acronyms for the same thing?

[Coryton]

I've never got this straight - are they two different acronyms for the same thing?

 

As I understand it, in technical terms the Cambrian has ETCS which is the signalling component of ERTMS and can stand alone or be part of ERTMS. ERTMS includes train management and we don't have it yet in the UK.

 

However...in the UK ETCS is normally called ERTMS (including in official documentation) so I think we have one of those awkward situations where you can either be "correct" and confuse people, or "incorrect" and be understood.

 

Or maybe it should just be seen as having a different meaning in technical signalling circles and elsewhere.

 

Bit unfortunate really.

[St. Simon]

I've never got this straight - are they two different acronyms for the same thing?

 

No.

 

ETCS - European Train Control System - Is the 'In-Cab' Signalling portion of the system, it is what is fitted to trains and sends out the Movement Authorities from the Interlocking.

 

ERTMS - European Rail Traffic Management System - This is the combination of the GSM-R and ETCS as well as a Traffic Management element to it that is fitted to the control centre.

 

As Coryton says, we don't have full ERTMS in the UK, we have ETCS on the Cambrian  and we have a Rail Traffic Management System installed at Romford and Didcot, until these two are combined, it is not ERTMS.

 

However, ETCS is referred officially as ERTMS, which isn't correct, I have seen it used incorrectly by some in the IRSE even. I don't know the reason behind this, but it is annoying.

 

Simon

[Grovenor]

But why does a speed restriction for level crossing sighting purposes have to work that way?

It doesn't have to work that way, but as indicated above, I suspect the speed restriction for level crossing sighting was not included in the system design, so a standard TSR is being used as a bodge.

Regards

[iands]

 

But why does a speed restriction for level crossing sighting purposes have to work that way?

 

(Edited to remove duplicate text)

 

Hi Coryton,

 

Afraid I can't answer that one for you (yet), I'm not sufficiently high enough up the 'food chain' as it were. I suspect as it is still a "trial" on the Cambrian, such things are currently managed on something akin to "best endeavours" at the moment, but there will be an "issues log" detailing both operational and technical short-comings that will be (or aimed to be) addressed in future software releases.

 

However, I'm acquainted with a couple of people who have been directly involved the Cambrian and I'll see if I can glean a bit more info (but it might not be for a week or so).

 

Regards, Ian.

[caradoc]

But why does a speed restriction for level crossing sighting purposes have to work that way?

 

(Edited to remove duplicate text)

 

Hi Coryton,

 

Afraid I can't answer that one for you (yet), I'm not sufficiently high enough up the 'food chain' as it were. I suspect as it is still a "trial" on the Cambrian, such things are currently managed on something akin to "best endeavours" at the moment, but there will be an "issues log" detailing both operational and technical short-comings that will be (or aimed to be) addressed in future software releases.

 

However, I'm acquainted with a couple of people who have been directly involved the Cambrian and I'll see if I can glean a bit more info (but it might not be for a week or so).

 

Regards, Ian.

 

A Permanent Speed Restriction (PSR) could be imposed instead of a TSR at each location affected, however converting a TSR to a PSR (and incidentally an Emergency Speed Restriction, ESR, to a TSR) was always, in my experience, a protracted process. In any case I am not aware of any restriction, whether ESR, TSR or PSR, which only applies to part of a train; Any such restriction must be observed until the whole train has passed clear of the affected section. I'm not sure that introducing a two-tier system, with some restrictions only applying to the front of a train and others to the whole train, would be feasible.

[big jim]

off the top of my head regarding the above there are bridges on the cambrian which are restricted for locos, bridge 22 near hanwood for example, technically once the loco is over the bridge then there should be no reason why the restriction couldnt be cancelled down allowing the coaches/wagons/whatever to accelerate over the bridge but in reality you have to wait for the 12 "invisible coaches" plus the 100m beyond to pass the restriction before being allowed to speed up

 

saying that even under RETB (or conventional signalling for that matter) i dont know of any restrictions that raise once the affected vehicle is past it, other than when approaching open crossings where you power up once on the crossing

Edited February 25, 2018 by big jim

[Coryton]

off the top of my head regarding the above there are bridges on the cambrian which are restricted for locos, bridge 22 near hanwood for example, technically once the loco is over the bridge then there should be no reason why the restriction couldnt be cancelled down allowing the coaches/wagons/whatever to accelerate over the bridge but in reality you have to wait for the 12 "invisible coaches" plus the 100m beyond to pass the restriction before being allowed to speed up

 

Perhaps getting ETCS to understand that is a step too far.

 

I wonder where the figure of 100 m comes from - because they think people can't work out accurately how long their train is, or because ETCS might be out by 100 m as to where the train is? Or a bit of both?

 

saying that even under RETB (or conventional signalling for that matter) i dont know of any restrictions that raise once the affected vehicle is past it, other than when approaching open crossings where you power up once on the crossing

 

Thanks - that's what I was trying to find out. So under conventional signalling you don't have to wait for the rest of the train at a speed restriction for an open crossing, but ETCS does make you wait (either for real or invisible coaches).

 

That doesn't sound great - though it's not likely to be an issue on main lines.

 

Of course one way round both problems would be for the limits to display in the cab but not to override the driver - after all we cope with the driver's judgement everywhere else in the UK.

 

 

 

 

 

 

 

 

off the top of my head regarding the above there are bridges on the cambrian which are restricted for locos, bridge 22 near hanwood for example, technically once the loco is over the bridge then there should be no reason why the restriction couldnt be cancelled down allowing the coaches/wagons/whatever to accelerate over the bridge but in reality you have to wait for the 12 "invisible coaches" plus the 100m beyond to pass the restriction before being allowed to speed up

 

saying that even under RETB (or conventional signalling for that matter) i dont know of any restrictions that raise once the affected vehicle is past it, other than when approaching open crossings where you power up once on the crossing

 

 

 

 

[phil-b259]

No.

 

ETCS - European Train Control System - Is the 'In-Cab' Signalling portion of the system, it is what is fitted to trains and sends out the Movement Authorities from the Interlocking.

 

ERTMS - European Rail Traffic Management System - This is the combination of the GSM-R and ETCS as well as a Traffic Management element to it that is fitted to the control centre.

 

As Coryton says, we don't have full ERTMS in the UK, we have ETCS on the Cambrian  and we have a Rail Traffic Management System installed at Romford and Didcot, until these two are combined, it is not ERTMS.

 

However, ETCS is referred officially as ERTMS, which isn't correct, I have seen it used incorrectly by some in the IRSE even. I don't know the reason behind this, but it is annoying.

 

Simon

 

Presumably because the aim of it all is to eventually have full ERTMS at some stage.

 

Railways are not short term things - and Acronyms have a habit of lingering on for decades and being absorbed by staff such that changing things later on is a pain.

 

For example there is a type of track circuit known  officially as an EBI track - for over two decades this was officially described as a TI21 track circuit - and guess what? thats what a lot of people (including me) still know it as!

 

OK so the Cambrian scheme may not be ERTMS - but does it really matter that much? ETCS replacing lineside signals is by far the most obvious and significant bit of ERTMS - Traffic management systems (despite the fancy name and the hype) are at the end of the day basically little more than a advanced form of ARS (automatic route setting) that is able to factor other things into the equation much like human regulators do at present.

 

So, while as a designer, the correct use of terminology obviously matters to you (personally and professionally) - for the rest of us its all rather irreverent at this stage in proceedings.

[The Stationmaster]

 

 

 

 

With a 'manual' speed restriction for a level crossing, is the driver expected to wait until the back of the train has cleared the sign before accelerating?

 

The Rule Book instruction is that Drivers should accelerate their train as soon as it is on the level crossing.  Thus if a restriction of speed imposed by ERTMS does not allow that the system is not compliant with the Rule Book - are we coming across another instance here of people getting involved in operational safety matters who are not properly competent to do so?

 

Overall I remain puzzled why a temporary restriction of speed (TROS) has been imposed at these level crossing and what has changed in any way to require any sort of restriction of speed as these crossings have no doubt been around for a considerable time and train speeds have surely not suddenly increased?  In any case the restrictions should be permanent and not temporary judging by the impact they are reportedly having on train running times as making them permanent would require their time cost to be taken into account in the timetable - but taht would be subject t agreement through the Network Change process.  Maybe somebody in NR thinks he/she is being rather clever by simply imposing a TROS without understanding the full; implications of what it means?  Or is it all down to basic level crossing information and speeds not being in the data for ERTMS on the Cambrian?

[phil-b259]

Perhaps getting ETCS to understand that is a step too far.

 

I wonder where the figure of 100 m comes from - because they think people can't work out accurately how long their train is, or because ETCS might be out by 100 m as to where the train is? Or a bit of both?

 

I think they key here is the words 'pilot scheme'

 

This scheme has been in use for several years now and I have no doubt that advances in computing technology plus things like this learned from the trial have been used to improve subsequent implementations of ETCS.

 

Its also possible that having got to a certain level of maturity, the latest versions of ECTS are incompatible with the 'pilot scheme' setup and as such it is not cost effective to go back and update it for what amounts to relatively few train movements.

Edited February 25, 2018 by phil-b259

[Coryton]

I think they key here is the word 'pilot scheme'

 

This scheme has been in use for several years now and I have no doubt that advances in computing technology plus things like this learned from the trial have been used to improve subsequent implementations of ETCS.

 

Its also possible that having got to a certain level of maturity, the latest versions of ECTS are incompatible with the 'pilot scheme' setup and as such it is not cost effective to go back and update it for what amounts to relatively few train movements.

 

It's always easy to sit on the outside and criticise, but does it take a full scale trial to realise that some speed restrictions are handled in a different way to normal and that ETCS isn't programmed for that?

 

Maybe.

 

It would be shame if nothing could be done about it, though.

 

It looks as ETCS is designed to make absolutely sure that a driver can't do anything wrong - perhaps there needs to be a bit more balance between that and practicality. How serious is the risk that a driver might accelerate early from a speed restriction - and how does the rest of the network cope with this possibility?

[phil-b259]

It's always easy to sit on the outside and criticise, but does it take a full scale trial to realise that some speed restrictions are handled in a different way to normal and that ETCS isn't programmed for that?

 

Maybe.

 

It would be shame if nothing could be done about it, though.

 

It looks as ETCS is designed to make absolutely sure that a driver can't do anything wrong - perhaps there needs to be a bit more balance between that and practicality. How serious is the risk that a driver might accelerate early from a speed restriction - and how does the rest of the network cope with this possibility?

 

This is actually pretty critical when you consider the end goal of ECTS is to allow the removal of not just signals, but also the whole fixed block system.

 

If the driver was allowed to disregard the speed restriction then they could end up compromising the safe braking distance to the train in front.For the whole 'moving block' thing to work obeying maximum speed limits is essential.

 

OK there is a long way to go between what the Cambrian has at the moment and the Utopian goal of ETRMS as a whole - but it has to be remembered that everything is being designed around reaching said goal. As such it doesn't surprise me that some elements of the system are too restrictive in the Cambrian context.

 

What we need is someone familiar with the specifics of the Pilot scheme that covered its installation to tell us what happens now. Is the scheme being updated as new things are learned or has it served its purpose and has ERMTS development left it behind as it were.

[Coryton]

This is actually pretty critical when you consider the end goal of ECTS is to allow the removal of not just signals, but also the whole fixed block system.

 

If the driver was allowed to disregard the speed restriction then they could end up compromising the safe braking distance to the train in front.For the whole 'moving block' thing to work obeying maximum speed limits is essential.

 

OK there is a long way to go between what the Cambrian has at the moment and the Utopian goal of ETRMS as a whole - but it has to be remembered that everything is being designed around reaching said goal. As such it doesn't surprise me that some elements of the system are too restrictive in the Cambrian context.

 

What we need is someone familiar with the specifics of the Pilot scheme that covered its installation to tell us what happens now. Is the scheme being updated as new things are learned or has it served its purpose and has ERMTS development left it behind as it were.

 

I'm not so sure.

 

The issue here is not whether ECTS enforces speed restrictions per se, but the fact that it enforces them until the last part of the train (real or imaginary) is 100 m beyond the speed restriction - even at an open crossing where it seems the driver would normally accelerate as soon as the train was on the crossing.

 

If - as I suggested - you enforce the train initially slowing for the speed restriction but leave it to the driver's judgement as to when to speed up, I don't think it makes any difference whether the train is running with fixed or moving block. 

 

At the moment we trust drivers not to speed up too soon. 

 

Was a goal of ECTS to reduce that risk? Is that risk higher under ECTS because we're training drivers to just stare at a screen and use no judgement?

 

Or did somebody decide that because ECTS could enforce this it should?

 

(And who decided on the 100 m margin? Are drivers under conventional signalling required to act as if they have an extra 4 coaches on the back of the train?)

[phil-b259]

I'm not so sure.

 

The issue here is not whether ECTS enforces speed restrictions per se, but the fact that it enforces them until the last part of the train (real or imaginary) is 100 m beyond the speed restriction - even at an open crossing where it seems the driver would normally accelerate as soon as the train was on the crossing.

 

If - as I suggested - you enforce the train initially slowing for the speed restriction but leave it to the driver's judgement as to when to speed up, I don't think it makes any difference whether the train is running with fixed or moving block. 

 

At the moment we trust drivers not to speed up too soon. 

 

Was a goal of ECTS to reduce that risk? Is that risk higher under ECTS because we're training drivers to just stare at a screen and use no judgement?

 

Or did somebody decide that because ECTS could enforce this it should?

 

(And who decided on the 100 m margin? Are drivers under conventional signalling required to act as if they have an extra 4 coaches on the back of the train?)

 

Ultimately I guess the goal is not to have the driver 'driving' at all. Think about it - the advantages of moving block in congested areas can only come from squeezing trains as close together as possible and that requires for the train to be moving at exactly the calculated speed. This is not new as such - Thameslink trains in the core and the even earlier Victoria line use the principle that the train drives itself - the driver is merely in charge of the doors (and can hit the emergency stop button if they see an issue such as a track defect or people on the line ahead)

 

As such in a true ECTS situation then the driver will not need to 'look down at the screen all the time' - the problem comes when the trains are still manually driven and we have replaced all lineside stuff (signals, speed boards, etc) with a screen in the cab.

 

Granted the Cambrian is never likely to need Thameslink style computer control for capacity reasons, but the overall thrust remains the same. Namely increased automation with humans only there in a 'just in case' scenario. We already have that with aeroplanes (they can take off fly then land without any human intervention subject to the appropriate systems being in place) and signalling is going down that route to with the railway running itself through ARS, and in future the much touted 'Train Management System', so why should driver be immune to change.

Edited February 25, 2018 by phil-b259

[Coryton]

Ultimately I guess the goal is not to have the driver 'driving' at all. Think about it - the advantages of moving block in congested areas can only come from squeezing trains as close together as possible and that requires for the train to be moving at exactly the calculated speed. This is not new as such - Thameslink trains in the core and the even earlier Victoria line use the principle that the train drives itself - the driver is merely in charge of the doors (and can hit the emergency stop button if they see an issue such as a track defect or people on the line ahead)

 

As such in a true ECTS situation then the driver will not need to 'look down at the screen all the time' - the problem comes when the trains are still manually driven and we have replaced all lineside stuff (signals, speed boards, etc) with a screen in the cab.

 

Granted the Cambrian is never likely to need Thameslink style computer control for capacity reasons, but the overall thrust remains the same. Namely increased automation with humans only there in a 'just in case' scenario. We already have that with aeroplanes (they can take off fly then land without any human intervention subject to the appropriate systems being in place) and signalling is going down that route to with the railway running itself through ARS, and in future the much touted 'Train Management System', so why should driver be immune to change.

 

I'm not saying that drivers should be immune to change.

 

But if we are going to automate the railway in general, I think it will need something that takes a rather different form to what works for Metro style services and won't just be an extension of ECTS - indeed it might need something with the same level of "intelligence" that self-driving cars use, and which could make a similar judgement to human drivers as to when to speed up at a level crossing.

 

In any case, if ECTS is intended to be universal perhaps it should be designed to be flexible enough cope with routes like the Cambrian and not just main lines. As you say, the Cambrian doesn't need automatic operation. It does have a lot of open crossings though.

 

(Do Thameslink trains drive themselves in the core? I thought that was Crossrail)

 

As for aircraft - so far as I know, passenger airliners don't have "auto-takeoff" and auto landing systems are good for poor visibility, but not so great when there's a cross-wind. And listening to ATC for incoming flights at a busy airport makes it clear that currently the crew have quite a lot to do even if the plane can follow down the instrument landing system itself. 

[Grovenor]

Overall I remain puzzled why a temporary restriction of speed (TROS) has been imposed at these level crossing and what has changed in any way to require any sort of restriction of speed as these crossings have no doubt been around for a considerable time and train speeds have surely not suddenly increased?

Going by big jim's statement in post #20 it would seem that these restrictions did not exist when the ETCS was specified for the Cambrian so use of a TSR was probably the only tool available in the system as designed. Provision of a function for a Level Crossing Approach speed restriction allowing immediate speed up as per rule book would be easy enough if it was in the spec. after all it just means that its not neccessary to add on the train length.

Regards

Edited February 25, 2018 by Grovenor

[phil-b259]

I'm not saying that drivers should be immune to change.

 

But if we are going to automate the railway in general, I think it will need something that takes a rather different form to what works for Metro style services and won't just be an extension of ECTS - indeed it might need something with the same level of "intelligence" that self-driving cars use, and which could make a similar judgement to human drivers as to when to speed up at a level crossing.

 

In any case, if ECTS is intended to be universal perhaps it should be designed to be flexible enough cope with routes like the Cambrian and not just main lines. As you say, the Cambrian doesn't need automatic operation. It does have a lot of open crossings though.

 

(Do Thameslink trains drive themselves in the core? I thought that was Crossrail)

 

As for aircraft - so far as I know, passenger airliners don't have "auto-takeoff" and auto landing systems are good for poor visibility, but not so great when there's a cross-wind. And listening to ATC for incoming flights at a busy airport makes it clear that currently the crew have quite a lot to do even if the plane can follow down the instrument landing system itself. 

 

All your points are noted - but to an extent the point about ERTMS - its not a 'fixed system' - its evolving as technology evolves. Moreover as I keep repeating the Cambrian line is a pilot scheme precisely to find the pitfalls from which improvements can be made. Naturally those improvements will come at a cost - for example it may simply be the case that there is not a decent enough BCR to go back and re-write the software to cope with shorter train lengths or sections shorter than 100m for example.

 

I used the aeroplane analogy because while no commercial airline has done it, Airbus / Boeing have proved it can technically be done. Whether the systems are good enough to cope with strong crosswinds is not something I can comment on, but again as technology continues to develop the potential is there. Whether it makes financial sense is a different question - though again one of the reasons for pilot schemes is to see how things stack up financially.

Edited February 25, 2018 by phil-b259

[phil-b259]

 

(Do Thameslink trains drive themselves in the core? I thought that was Crossrail)

 

 

They will do from this year. See here https://www.railengineer.uk/2017/03/22/thameslink-signalling-update/

 

And note this paragraph

 

The ERTMS Early Deployment Scheme (EDS), introduced in 2011 on the Cambrian lines in Wales, was the pilot project for Level 2 deployment to other parts of the UK network. The experience gained and lessons learned have provided a valuable input to the development and delivery of the Thameslink ERTMS installation.

[iands]

The attachment below contains 3 extracts from the ERTMS Operational Concept document. They don't specifically answer the queries above but give a bit of an insight into how ERTMS will/might be implemented.

 

ERTMS OC Extract.docx

 

Regards, Ian.

[big jim]

Being as I only sign to Machynlleth I can’t comment on open crossings as to whether the speed opens up straight away but I have asked the question of one of my drivers who does sign the coast bit

 

Regards the comment a few posts back about ERTMS should prevent the driver making mistakes, it is still possible to have a SPAD under ERTMS but it would be a very slow speed one due to there being a ‘release speed’ for approaching block markers

Edited February 25, 2018 by big jim