E.R.T.M.S.

[phil-b259]

24 minutes ago, david.hill64 said:

Reading back through the last pages it seems that the ETCS applications are being done as an overlay. I can understand why that is (cheaper and migration easier) but that eliminates one of the benefits of ETCS: improved capacity. (You can use shorter block sections where you have continuous ATP as some of the safety margins are not required and where you have slow lines next to fast you can optimise the block sections on the slow for that traffic rather than being constrained to use longer sections than necessary so that signals on adjacent lines are aligned).

 

Is the plan generally to retain existing blocks or will they be optimized? Or will higher speeds be allowed?

 

Firstly techniques do exist to allow co located signals even with different line speeds - on the ECML they slow lines (max 100mph) generally use three aspect signalling and the fast (125mph) lines use 4 aspect signals.

 

Secondly higher speeds come later because that will also need other non ECTS things addressing like eliminating level crossings and making sure the track / OLE is up to it.

 

Thats the beauty of ECTS - it can be bought in bit by bit and as an overlay initially rather than requiring a 'big bang' approach - which is quite important when you factor in things like driver training* with lineside signals being removed later on.

 

The ECML is due to become the first UK ECTS installation which sees lineside signals eliminated, but even that will be a gradual process with sections being done one at a time - I had to pay a visit to Welwyn last week and noticed lots of 'CAB' signage appearing at what will presumably the boundary of one of the phases.

 

* I am told that even on Thameslink because of the backlog of driver training that last year there were still occasional days when 70% of trains were being driven manually rather than using ECTS and its rare for every train on a particular day to be in ATO mode.

 

 

[DY444]

5 hours ago, david.hill64 said:

Reading back through the last pages it seems that the ETCS applications are being done as an overlay. I can understand why that is (cheaper and migration easier) but that eliminates one of the benefits of ETCS: improved capacity. (You can use shorter block sections where you have continuous ATP as some of the safety margins are not required and where you have slow lines next to fast you can optimise the block sections on the slow for that traffic rather than being constrained to use longer sections than necessary so that signals on adjacent lines are aligned).

 

Is the plan generally to retain existing blocks or will they be optimized? Or will higher speeds be allowed?

 

I think the increased capacity argument is going to turn out to be something of a fallacy unless there is a major change in philosophy wrt defensive driving.  I don't know what happens on the Cambrian but it's not difficult to imagine training on ETCS going something along the lines of if there's a speed reduction incoming then always stay 20 mph or whatever under the curve . 

 

This kind of thing already happens on TVM430 lines where drivers typically want to have one block in hand in front of an incoming speed reduction which I suspect came originally from SNCF's long standing doctrine of being very cautious when running under adverse signals.  There's nothing wrong with this from a safety perspective but it does chew up capacity so I expect the real world usable capacity will be nothing like it is on paper. 

 

Now full supervision ATP should give the confidence that you can drive closer to the braking curve in safety however after years of defensive driving being hammered in and possibly fear that an ATP intervention will be a hanging offence, I just don't see it.

Edited May 8 by DY444

[david.hill64]

15 minutes ago, phil-b259 said:

 

Firstly techniques do exist to allow co located signals even with different line speeds - on the ECML they slow lines (max 100mph) generally use three aspect signalling and the fast (125mph) lines use 4 aspect signals.

Yes: and that just proves that the slow lines are not optimised for capacity. Slow lines with 100mph line speed  and 4 aspect optimised signalling would have higher capacity but would not be done with conventional signalling because of the probability of misreads.

15 minutes ago, phil-b259 said:

Secondly higher speeds come later because that will also need other non ECTS things addressing like eliminating level crossings and making sure the track / OLE is up to it.

Absolutely agree, but it was the only way I could think of getting extra capacity with existing block lengths.

15 minutes ago, phil-b259 said:

 

Thats the beauty of ECTS - it can be bought in bit by bit and as an overlay initially rather than requiring a 'big bang' approach - which is quite important when you factor in things like driver training* with lineside signals being removed later on.

Hence my comment about migration issues.

15 minutes ago, phil-b259 said:

 

The ECML is due to become the first UK ECTS installation which sees lineside signals eliminated, but even that will be a gradual process with sections being done one at a time - I had to pay a visit to Welwyn last week and noticed lots of 'CAB' signage appearing at what will presumably the boundary of one of the phases.

I thought that lineside signals have disappeared from the Cambrian?

15 minutes ago, phil-b259 said:

 

* I am told that even on Thameslink because of the backlog of driver training that last year there were still occasional days when 70% of trains were being driven manually rather than using ECTS and its rare for every train on a particular day to be in ATO mode.

 

 

There is a delicious irony in this! Trains are driven manually because the drivers have not been taught how not to drive !!!

[david.hill64]

7 minutes ago, DY444 said:

Now full supervision ATP should give the confidence that you can drive closer to the braking curve in safety however after years of defensive driving being hammered in and possibly fear that an ATP intervention will be a hanging offence, I just don't see it.

 I think you are correct, but then I am sure I read that we don't need HS2 because digital signaling will give us greater capacity..................

[phil-b259]

2 hours ago, david.hill64 said:

 I think you are correct, but then I am sure I read that we don't need HS2 because digital signalling will give us greater capacity..................

 

 

Digital signalling permits very short block sections - the length of which is uneconomic to provide where lineside signals are in use  (On Thameslink there are lots more block sections under ECTS than there are if a driver is using the lineside signals).

 

Automatic Train Operation (ATO) eliminates the need to consider human reaction times and will result in a significantly quicker response to movement authorities or changes in conditions.

 

You need both of the above to truly realise the capacity gains from 'Digital Signaling' (as the marketing gurus call it).

 

Installing ECTS using block sections whose length is designed around colour light signalling and which requires humans to drive the trains won't make any improvement to what currently exists (other than to reduce the amount of lineside infrastructure which requires maintenance (and obviously less maintenance = less staff so you can see why certain people / organisations / HM Treasury still like it) 

Edited May 8 by phil-b259

[Edwin_m]

6 minutes ago, phil-b259 said:

Installing ECTS using block sections whose length is designed around colour light signalling and which requires humans to drive the trains won't make any improvement to what currently exists (other than to reduce the amount of lineside infrastructure which requires maintenance (and obviously less maintenance = less staff so you can see why certain people / organisations / HM Treasury still like it) 

There's a small improvement in the situation where the driver has passed a single yellow so is aiming to stop at a red, but finds it has cleared by the time it is visible.  ETCS can communicate this "aspect" change as soon as it takes place.  It's also possible for trains with better braking to follow more closely, although this is a bit limited unless the blocks are shorter.  

[phil-b259]

1 hour ago, Edwin_m said:

It's also possible for trains with better braking to follow more closely, although this is a bit limited unless the blocks are shorter.  

But blocks can only be shorter because they don’t need to come equipped with fixed signals which add significantly to infrastructure costs.

 

Hence the need for both ATO and ECTS - having one or the other in isolation doesn’t tend to achieve much in terms of train throughput.

 

 

Edited May 8 by phil-b259

[The Stationmaster]

45 minutes ago, phil-b259 said:

But blocks can only be shorter because they don’t need to come equipped with fixed signals!.

 

Hence the need for both ATO and ECTS - having one or the other in isolation doesn’t tend to achieve much in terms of train throughput.

 

 

You can, and NR does, have short blocks with fixed signals. (by short I mean little over a couple of train lengths = c, 2.5 x 3car Class 387).  The Sydney NSW City Loop has even shorter signal sections with trains driven by humans and uses what amounts to 6 aspect signalling.   But the penalty of short signal sections is the impact on line speed, especially for trains which don''t fit the 'high intensity' passenger train model.

 

On a mixed traffic railway there will always be a penalty for some of the trains using a line - the decision has to be what the line is optimised for.   Removing fixed signals won't necessarily change that because various trains will still require longer braking distances and;or have to be restricted in speed.  You are always going to come back to the basic definition of headway.  Lines used exclusively by identical trains running to identical or near identical stopping patterns could possibly gain some capacity advantage from ECTS with full ATP.  Lines with mixed traffic and train types with differing brake performance and variations in train length are far less likely to gain any real benefit.

[phil-b259]

 

7 minutes ago, The Stationmaster said:

You can, and NR does, have short blocks with fixed signals. (by short I mean little over a couple of train lengths = c, 2.5 x 3car Class 387).  

 

Naturally in some places this is acceptable in cost terms - but you should note that even on Thameslink not every ECTS block section has a proper signal - over half of them simply have ECTS marker boards.

 

What that means is in laymans terms the capacity is pretty much halved if ECTS is not available.

 

That is the point here - seconds do count and if a 'traditional' block section is subdivided into multiple sections then trains following each other closely are less likely to be bought to a stand (or if they do they are a lot closer to the train in front when that happens)

 

So while obviously mixing trains with different performance characteristics plays a big part in determining headway, having many more block sections can significantly increase throughput - just as it did in the days of semaphore signalling where an Intermediate block signal could be inserted between boxes.

 

 

 

 

 

[ess1uk]

6 hours ago, david.hill64 said:

 

I thought that lineside signals have disappeared from the Cambrian?

 

Me too 

[big jim]

58 minutes ago, ess1uk said:

I thought that lineside signals have disappeared from the Cambrian?

Conventional signals have gone, apart from the bit around Machynlleth depot where there are a couple of ground signals but there are block markers, I wonder if the ECML bit will not have any block markers lineside? 

[corneliuslundie]

"I thought that lineside signals have disappeared from the Cambrian?"

They have but Welwyn is rather a long way from Wales.

The last manual (literally) signal a driver sees going west is at Shrewsbury controlling the junction.

Jonathan

[Edwin_m]

1 hour ago, big jim said:

Conventional signals have gone, apart from the bit around Machynlleth depot where there are a couple of ground signals but there are block markers, I wonder if the ECML bit will not have any block markers lineside? 

I think it would have to, so a signaller can instruct a driver to proceed as far as a particular one when there's some failure with the system.  However, I don't think they have to be alongside each other on "fast" and "slow" lines as signals do, as there's much less risk of reading across.  That means the block lengths can be optimised for the speed and type of trains likely to use them.  

Edited May 8 by Edwin_m

[phil-b259]

2 hours ago, big jim said:

Conventional signals have gone, apart from the bit around Machynlleth depot where there are a couple of ground signals but there are block markers, I wonder if the ECML bit will not have any block markers lineside? 

It will have block markers as far as I know for the simple reason that drivers still need reference marks where they can stop and be cautioned from should the train / ECTS system fail.

 

 

[phil-b259]

1 hour ago, corneliuslundie said:

"I thought that lineside signals have disappeared from the Cambrian?"

They have but Welwyn is rather a long way from Wales.

The last manual (literally) signal a driver sees going west is at Shrewsbury controlling the junction.

Jonathan

They have - but the Cumbrian scheme was classed as a Pilot scheme to test the technology - what is being rolled out on the ECML and has been installed on the Thameslik core / GWML is the  full ‘production system’

 

Thsre is also the fact that ERTMS on the Cambrian replaced the RETB system (which also lacks signals) and if you want to be really picky then you could go back to the first RETB system to find a railway without lineside signals…..

[bécasse]

3 hours ago, phil-b259 said:

Thsre is also the fact that ERTMS on the Cambrian replaced the RETB system (which also lacks signals) and if you want to be really picky then you could go back to the first RETB system to find a railway without lineside signals…..

If one wants to be really, really picky, one would point out that a good few OES single track terminal branch lines were operated without any fixed signals beyond the junction station, especially post the mid-1920s (although the SR, at least, did provide illuminated marker lights). Where fixed signals existed on such lines, they often did no more than protect level crossings.

[corneliuslundie]

Tongue in cheek

The Bishop's Castle Railway had fixed signals but certainly after the first few years none were used. The only signal in use on the line was that protecting the main line at the junction near Craven Arms.

And on busy days it sometimes had two engines in steam at once.

Jonathan

[icn]

On 07/05/2024 at 00:45, david.hill64 said:

I think that is rather harsh. 

 

There is a world of difference between equipping self contained monorail systems or small national railways and complex conventional national systems with a variety of rolling stock, complex layouts and existing signalling systems.

 

I think that the UK approach has been rational. Try it first on a lightly used, relatively unimportant line (Cambrian) where problems are unlikely to affect the national network to any serious degree. More than 10 years experience. Success.

 

Then with that experience under your belt, feel confident enough to introduce it - together with an ATO add on -through the Thameslink core route where performance is critical. You still have the simplicity of a unified fleet but are now getting experience in a high risk operational environment. Success.

 

Then take a deep breath and integrate it with two other signalling systems on another line where performance is critical (Elizabeth line). Success.

 

Next up equip a conventional main line and a wide variety of rolling stock - electric, diesel and steam.

 

Seems reasonable to me.

Switzerland is fully ETCS equipped by now. I wouldn't call that small by kilometers (5000km network, vs UK's 15'000 - and I believe the total number of passenger journeys is quite similar at around 400 million/year for both), extremely busy network, and also has the station with the highest volume of rail traffic worldwide.

 

It's true that most of the network is ETCS-L1/overlay - pure ETCS-L2 is only used on select sections, but some of those sections are critical for network stability (NBS, and arguably the Gotthard and Lotschberg tunnels are too even if not as important for commuters). And there is a lot of ETCS-only rolling stock all over the network by now (on top of all the stock that's been retrofitted to operate on the ETCS-only sections). The system works, and it works at scale.

 

The UK approach is certainly rational, but it's still a laggard relatively speaking.

Edited May 9 by icn

[david.hill64]

14 minutes ago, icn said:

Switzerland is fully ETCS equipped by now. I wouldn't call that small by kilometers (5000km network, vs UK's 15'000 - and I believe the total number of passenger journeys is quite similar at around 400 million/year for both), extremely busy network, and also has the station with the highest volume of rail traffic worldwide.

 

It's true that most of the network is ETCS-L1/overlay - pure ETCS-L2 is only used on select sections, but some of those sections are critical for network stability (NBS, and arguably the Gotthard and Lotschberg tunnels are too even if not as important for commuters). And there is a lot of ETCS-only rolling stock all over the network by now (on top of all the stock that's been retrofitted to operate on the ETCS-only sections). The system works, and it works at scale.

 

The UK approach is certainly rational, but it's still a laggard relatively speaking.

Switzerland got in early: I think a lot of the 'hands-on' development was done there. Good for them.

 

ETCS level 1 overlay is relatively simple to achieve. I'm not disputing the Swiss achievement and I have no idea what they had before (Indusi??) But ETCS level 1 would give marginal safety benefit over TPWS (How many ATP preventable accidents have there been in UK since TPWS roll-out?), so going for ETCS level 2 is, in my opinion, the correct option for UK. Could it be done more quickly? I don't know as I'm not working in UK, but I suspect that availability of competent staff as much as funding will be an issue.

 

Having said that, I can still remember vividly being told by Railtrack in 1994 that as the commercial benefits of ETCS level 3 were overwhelming, Industry would deliver within 5 years and Railtrack would complete the installation within 10. Happy days!

[The Stationmaster]

22 hours ago, phil-b259 said:

 

 

Naturally in some places this is acceptable in cost terms - but you should note that even on Thameslink not every ECTS block section has a proper signal - over half of them simply have ECTS marker boards.

 

What that means is in laymans terms the capacity is pretty much halved if ECTS is not available.

 

That is the point here - seconds do count and if a 'traditional' block section is subdivided into multiple sections then trains following each other closely are less likely to be bought to a stand (or if they do they are a lot closer to the train in front when that happens)

 

So while obviously mixing trains with different performance characteristics plays a big part in determining headway, having many more block sections can significantly increase throughput - just as it did in the days of semaphore signalling where an Intermediate block signal could be inserted between boxes.

 

 

 

 

 

And as it still can where short signal sections are used with lineside signals - albeit with a speed penalty when t running on close headways.    and that was my point - you don't need ECTS in order to have short signal sections but it does have certain advantages when coupled with ATP and the removal of all the equipment and labour needed to provide, power, and maintain traditional lineside signals.  

 

But the market boards are of course in any case fixed signals in their own right as they still mark the extent of what would traditionally be called signal sections..  And that remains the case until a system advances to a moving block arrangement.

 

It is also very easy for us talking in this century to forget - with only very few remaining examples - the numerous short block sections that existed in the steam/early diesel age with many being only a few hundred yards (and sometimes less)  long.   Short sections  per se are very definitely not a new idea

[Edwin_m]

42 minutes ago, The Stationmaster said:

And as it still can where short signal sections are used with lineside signals - albeit with a speed penalty when t running on close headways.    and that was my point - you don't need ECTS in order to have short signal sections but it does have certain advantages when coupled with ATP and the removal of all the equipment and labour needed to provide, power, and maintain traditional lineside signals.  

 

But the market boards are of course in any case fixed signals in their own right as they still mark the extent of what would traditionally be called signal sections..  And that remains the case until a system advances to a moving block arrangement.

 

It is also very easy for us talking in this century to forget - with only very few remaining examples - the numerous short block sections that existed in the steam/early diesel age with many being only a few hundred yards (and sometimes less)  long.   Short sections  per se are very definitely not a new idea

I recall the Reading remodelling is an example of this.  There are "closing up" signals at the entry to each platform to allow the next train to approach as the previous one departs, but the resulting short sections and the limitation of only four signal aspects leads to a speed restriction being imposed that wouldn't otherwise be necessary.  ETCS here would allow non-stop trains (not that there are very many!) to pass at higher speed.  

[The Stationmaster]

17 minutes ago, Edwin_m said:

I recall the Reading remodelling is an example of this.  There are "closing up" signals at the entry to each platform to allow the next train to approach as the previous one departs, but the resulting short sections and the limitation of only four signal aspects leads to a speed restriction being imposed that wouldn't otherwise be necessary.  ETCS here would allow non-stop trains (not that there are very many!) to pass at higher speed.  

Only on the Relief Line side I think.  I wonder if there are any Signalbox Special Instructions in place for certain non-stopping trains, especially the heavy freights. (although iI understand that when possible the heaviest stone trains are routed on the level via Westbury Line Jcn). 

[Steadfast]

1 hour ago, The Stationmaster said:

Only on the Relief Line side I think.  I wonder if there are any Signalbox Special Instructions in place for certain non-stopping trains, especially the heavy freights. (although iI understand that when possible the heaviest stone trains are routed on the level via Westbury Line Jcn). 

Yes Relief lines only. We don't do any of the really heavy stuff, but everything I've seen seems to go via the Feeders so as not to disrupt the Main lines side of the station by crossing at the East end. It'd be interesting to have a go on a jumbo over the Feeders to be fair, as the back end being on the downhill from Oxford Road probably helps get the train going, even though the loco is climbing from the Feeders towards Reading.

2 hours ago, Edwin_m said:

I recall the Reading remodelling is an example of this.  There are "closing up" signals at the entry to each platform to allow the next train to approach as the previous one departs, but the resulting short sections and the limitation of only four signal aspects leads to a speed restriction being imposed that wouldn't otherwise be necessary.  ETCS here would allow non-stop trains (not that there are very many!) to pass at higher speed.  

The vast majority of the freight through Reading is coming on or off the 30 mph Feeder lines, so even the 40 mph Line speed isn't seen often!

 

Jo

[ess1uk]

ECML testing plan

[david.hill64]

16 hours ago, ess1uk said:

ECML testing plan

And this says no more sections with overlays, which implies that they will be optimizing blocks when ETCS is designed (or at least will have the ability to do so). Good!