Standardisation of signalling and train control?

[mrmartin]

Reading through the crossrail disaster it seems the big sticking point has been integration between the Siemens CBTC signalling system and the Bombardier on train management system. I'm not even convinced this is at all working by 'itself', as core testing hasn't started yet, nevermind the other two signalling/train protection systems that it will need to integrate to work on NR lines to Shenfield and Heathrow.

It seems to be pretty crazy to have to integrate two proprietary products like this. You've theoretically got x signalling suppliers and y train manufacturers, which means that you are looking at x * y giant integration projects to do in theory.

I mistakenly thought that ERTMS was the 'standard' for this and everyone would work towards it, a bit like LTE or WiFi. What am I missing here? Is Siemens technology in the core based on ERTMS or is it a totally different system?

Edited to add: it may well be that it is, but it seems to me that there is an enormous amount of bespoke integration work being done, much more than 'tweaking' you'd expect to put two systems together which use the same standard...

 

[swt_passenger]

IIRC TfL didn’t think that ERTMS was mature enough technology, so deliberately chose a different system for their core route, (a more LU style of CBTC) in the hope that at least their infrastructure would be ready in time, even if the NR sections might not. I think they needed some sort of derogation or dispensation to not use ERTMS.

This thread a couple of years ago is one of a few to mention it:
https://www.railforums.co.uk/threads/tfl-consultation-crossrail-central-operating-section.134191/

I linked to a railengineer magazine article in that thread that explains the reasoning for a CBTC system.

 

[hwl]

IIRC TfL didn’t think that ERTMS was mature enough technology, so deliberately chose a different system for their core route, (a more LU style of CBTC) in the hope that at least their infrastructure would be ready in time, even if the NR sections might not. I think they needed some sort of derogation or dispensation to not use ERTMS.

This thread a couple of years ago is one of a few to mention it:
https://www.railforums.co.uk/threads/tfl-consultation-crossrail-central-operating-section.134191/

I linked to a railengineer magazine article in that thread that explains the reasoning for a CBTC system.

Correct and they got a derogation in 2012 not to use ETCS. Possibly best described as an "Oops" moment in hindsight given how well ETCS has worked for Thameslink.(Ok not used daily in passenger service yet as intended yet but works)
Also ETCS issues in the Heathrow Tunnels.

The Siemens CBTC system is an evolution of a Matra system developed for Paris Metro Line 14 automation and nothing to do with ETCS.

 

[chriskeene]

ETCS has worked for Thameslink.(Ok not used daily in passenger service yet as intended yet but works)
.

Is there a reason Thameslink is not using ETCS (and ATM?) in the core yet? Obviously, they have had timetable issues and are not yet up to full capacity but is this stopping the role out of ETCS?

 

[hwl]

Is there a reason Thameslink is not using ETCS (and ATM?) in the core yet? Obviously, they have had timetable issues and are not yet up to full capacity but is this stopping the role out of ETCS?

No need to for the current service level (will need to for the next step up) and they need to train the drivers - Route knowledge probably far more useful to run decent GTR service a the moment than ATO training.

 

[Randomer]

Wasn't the original reason for not using ERTMS / ETCS on Crossrail that they were not sure it had the accuracy to work with platform edge screens/doors which Thameslink does not have to worry about?

 

[Surreytraveller]

Is there a reason Thameslink is not using ETCS (and ATM?) in the core yet? Obviously, they have had timetable issues and are not yet up to full capacity but is this stopping the role out of ETCS?

One more thing to go wrong on top of everything else. One thing at a time gives a better chance of things working. Otherwise you've got drivers who not only don't sign the route, are unfamiliar with stock, but trying to use a new signalling system too

 

[hwl]

Wasn't the original reason for not using ERTMS / ETCS on Crossrail that they were not sure it had the accuracy to work with platform edge screens/doors which Thameslink does not have to worry about?

Yep but the stopping accuracy is apparently good enough on TL

 

[Randomer]

Isn't there a big difference between stopping close enough to the PRM hump (which is somewhat bigger than the door) for that door to work and with all other doors on the platform for Thameslink vs. within a 20 or so CM for the platform screen doors to all be aligned properly.

 

[MarkyT]

Signalling specialists on Thameslink eventually stated that they had achieved a level of stopping accuracy sufficient for PEDs using their ETCS based ATO system, but that was a long time after the Crossrail system was specified and was an unknown at that time.

 

[Bald Rick]

When the decision was taken - a long time ago - there were a number of things that ETCS couldn’t do which Crossrail needed. In addition to the PED interface, there was also auto-reverse (turning a train round in ATO with no driver input) and few other things I can’t remember. AIUI, all these things are now available in ETCS, or in the late stages of development. Indeed I have heard talk of retro-fitting Crossrail to ETCS, albeit that is a long way in the future.

 

[hooverboy]

I mistakenly thought that ERTMS was the 'standard' for this and everyone would work towards it, a bit like LTE or WiFi. What am I missing here?

you aren't missing anything apart from certain organisations being a pain in the backside.
you're quite correct in your assessment that transport signalling and control systems should be working towards a standard like 3GPP is for mobile telephony.

in fact LTE-R is being standardised globally as next gen transport comms.Should have happened A LOT sooner, these guys are 20-30 years behind the curve.