Grayling announces Network Rail to use digital signalling

[HSTEd]

How much does fitting ETCS equipment to rolling stock cost?

Perhaps it would be cheaper to simply require all rolling stock be so fitted, and then forget overlays in favour of pure ETCS implementations?

Less trackside work means less cost in most cases in Britain now.

 

[Joseph_Locke]

The DfT plan is ETCS level 2*. This is fixed block (so on its own no additional capacity) but without lineside signals, allowing a mite more flexibility in placement of the ends of block sections due to the removal of signal sighting constraints. It does not improve junction speeds, overlap lengths and platform dwells. It also cannot magically allow a fast train to teleport in front of a stopping train.

There are a few classes where fitting all the gubbins, antennae and DMI is going to be a major project, but given that the bulk of the DR plan isn't until post 2027 even the 142s should have gone to a better place by then.

In The North, on a typical two-track line punctuated by flat junctions and carrying a mix of fast, semi-fast and freight services, a small reduction in journey time might result from ETCS2, but that's about it.

On the other hand, the system will know instantly and to the nearest second exactly how late your train is ...

* and still is, 10 years later - this is still the current plan

 

[The Planner]

This. Level 2 is not the panacea to all our problems.

 

[Joseph_Locke]

This. Level 2 is not the panacea to all our problems.

The problem is that someone has told a couple of economists and a History graduate (aka the D of T) that this is the Philosopher's Stone, and that this (cheap) magick allows Newtonian mechanics to be ignored.

I predict that every rail capacity scheme (north of Watford Gap at least) will be required to justify why it cannot just use this snake oil to slave its issues ...

 

[MarkyT]

The point of the level 2 overlay concept is you can add extra blocks easily where they are required for capacity, usually approaching and through station platforms. This is nothing new as it is similar to an old signalling concept known as 'closing up' which adds extra signals on approach to a station. It was routinely employed at nearly every busy station on London Underground under the old style colour light signalling for example. With lights on sticks however, closing up is inherently difficult to provide on main lines where (unlike London Underground) the through speed is high and the required signal spacing for (stopping) capacity becomes significantly less than braking distance or even half braking distance for the through fasts so simple 3 or 4 aspect sequences are no longer appropriate. The technique requires extra expensive signals clearly, often right in the middle of complex throat junctions (so needing expensive multitrack gantries), and all kinds of funny special aspect sequences, with extra yellows and timed approach releases. This is possible within the rules (subject to risk assessment), but funny aspect sequences can also introduce a heightened risk of 'pre-emption' leading to SPADs, so they're difficult to justify. ETCS level 2 overlay allows these additional closing up blocks to be provided without the risks and at much lower cost, once most of the fleet is equipped on board. Out on higher speed plain line there is usually no need for the extra blocks so they can be applied selectively only where necessary. Retaining conventional signals as well allows trains to run that are not fitted with ETCS. While if many unfitted trains ran during traffic peaks that would affect performance adversely as they will not be able to close up around stations, an occasional such train makes little difference, especially if they are scheduled outside the peaks, and it also provides a fallback means of keeping the show rolling if a particular fitted train suffers an ETCS or radio failure.

 

[Joseph_Locke]

You can't have a block end in the middle of a junction; No amount of Digital Railway is going to fix the issue that many of the existing flat junctions are full to capacity , a function of the time each one takes to cross the path of another train - if you can't get through the junctions at the ends, what use is cmore apacity on the bits in between?

Most of the (TCB bits) of the NW already have a planning headway of 5minutes or better - on paper that's 12 tph. Castlefield Corridor has a planning headway of 3 - that's 20 tph. What bit of Digital Railway is going to magick up the missing four, which are currently "lost" at Castlefield Junction during crossing moves?

pPs. the DfT is talking about no lineside signals, just fixed block ETCS. You can't run non-ETCS train without taking a possession.

 

[Senex]

The problem is that someone has told a couple of economists and a History graduate (aka the D of T) that this is the Philosopher's Stone, and that this (cheap) magick allows Newtonian mechanics to be ignored.

I predict that every rail capacity scheme (north of Watford Gap at least) will be required to justify why it cannot just use this snake oil to slave its issues ...

Surely the Philosopher's Stone is supposed to be Level 3, but as far as I know this doesn't exist on any main line anywhere yet. Maybe someone in Dft knows it's "only just round the corner" ...

 

[MarkyT]

The issue with overlaps.

For the uninitiated, an overlap is a section of track usually around 200m that is proved and locked clear as an 'overrun' ahead of a signal when at red and being approached by a train. Combined with a trainstop system such as TPWS it can ensure that a train passing the red at no greater than an assumed speed will not collide with conflicting traffic passing through junctions beyond the overlap. The interlocking disengages the overlap locking when the train is assumed to be at a stand which is determined when the trains has occupied the 'berth' track circuit for a calculated time based on its length. Ideal layouts ensure there are no junctions within the overlap for any signal so they can all be approached regardless of what movements are taking place beyond, but sometimes this cannot be achieved in the space available so junctions are sometimes locked by overlaps. There are techniques for minimising the adverse effect of this locking on flexibility however: Alternative overlaps can be provided to direct an overrun path to an different track away from the conflicting movement and these can even be moved while the train is approaching using 'swinging overlap' controls. Overlaps can also be shortened where required as long as the maximum approach speed is controlled appropriately, either by a permanent speed restriction for all trains or a 'warning' approach where the clearance of the previous signal is delayed by the interlocking to control speed. Despite all these safegaurds, drivers today on the national network approach reds at dead slow in accordance with their TOCs' defensive driving policies. This is a relatively new phenomenon to avoid even minor SPADs that present no real danger whatsoever and is markedly different to driving behaviour seen on traditionally signalled lines of London Underground, where trains routinely approach reds at fairly high speed and brake heavily at the right moment to stop accurately. LU has always been protected by a trainstop system and the driving behaviour reflects that, so it is perhaps time for a new approach to this on National Rail, given the provision of TPWS trainstops today and ETCS in the future.

Combined with the closing up techniques more widely provided as described in my previous post, more aggressive driving towards platform starters at red could also help to improve capacity under Level 2 and even reduce journey times to a small extent.

 

[Ianno87]

Combined with the closing up techniques more widely provided as described in my previous post, more aggressive driving towards platform starters at red could also help to improve capacity under Level 2 and even reduce journey times to a small extent.

That is true where the binding constraint on a railway is the ability to re-occupy the same platforms to maximise train throughput unconstrained by any other issues. Namely a metro raikway, like the SWML Slow Lines.

The TransPennine route is however considerably more conplex than this.

The ruling constraint is the fact that a train stopping at Mossley, Greenfield, Marsden and Slaithwaite takes about 8-10 minutes longer between Stalybridge and Huddersfield than a non-stop train. Ditto stopping trains between Huddersfield and Leeds. Or indeed freight paths between Stalybridge and Huddersfield.

No manner of whizzy signalling does anything to sort this fundamental fact.

The next constraint is that Diggle Jn to Marsden is one block sectioms. This could be solved with conventional lights on sticks (or rather mounted to tunnels) if one wanted to.

Solve these, then you've still got flat junctions (Guide Bridge, Stalybridge), Bradley Wood, Thornhill) to solve and lastly city centre capacity at Manchester and Leeds. It is only the very latter where ETCS *mighrlt* make the slightest difference.

 

[MarkyT]

You can't have a block end in the middle of a junction; No amount of Digital Railway is going to fix the issue that many of the existing flat junctions are full to capacity, a function of the time each one takes to cross the path of another train - if you can't get through the junctions at the ends, what use is more apacity on the bits in between?

There are cases of closing up signals 'in the middle' of junctions. Ashford International has some examples. It's not ideal as you say because if a trains stops there it is by definition completely blocking other movements through the junction. They're not usually used like that though as they are there to allow a following train to keep moving through the junction while the previous one is about to clear the platform ahead.

I fully agree that 'digital railway' isn't going to magically squeeze much extra capability out of junctions alone, but combined with some targeted layout
remodelling it could help.

 

[edwin_m]

There are cases of closing up signals 'in the middle' of junctions. Ashford International has some examples. It's not ideal as you say because if a trains stops there it is by definition completely blocking other movements through the junction. They're not usually used like that though as they are there to allow a following train to keep moving through the junction while the previous one is about to clear the platform ahead.

There are also closing up signals at the entrances to each platform at Reading, which would also block the junction if a train stopped there. Presumably they don't allow a second train to approach these until they are pretty sure the previous one is about to move away from the platform.

That is true where the binding constraint on a railway is the ability to re-occupy the same platforms to maximise train throughput unconstrained by any other issues. Namely a metro raikway, like the SWML Slow Lines.

The TransPennine route is however considerably more conplex than this.

The ruling constraint is the fact that a train stopping at Mossley, Greenfield, Marsden and Slaithwaite takes about 8-10 minutes longer between Stalybridge and Huddersfield than a non-stop train. Ditto stopping trains between Huddersfield and Leeds. Or indeed freight paths between Stalybridge and Huddersfield.

No manner of whizzy signalling does anything to sort this fundamental fact.

The next constraint is that Diggle Jn to Marsden is one block sectioms. This could be solved with conventional lights on sticks (or rather mounted to tunnels) if one wanted to.

Solve these, then you've still got flat junctions (Guide Bridge, Stalybridge), Bradley Wood, Thornhill) to solve and lastly city centre capacity at Manchester and Leeds. It is only the very latter where ETCS *mighrlt* make the slightest difference.

I agree with nearly all this - digital signalling won't make much difference to capacity on Transpennine.

However I do suggest the long block section at Standedge shouldn't be too much of a problem with a fast/slow mix, since it is near the middle of the length of track shared by fast and slow trains. The trains are much closer together at either end of the section because the slow sets out just behind the previous fast and is being caught up by the next one as it gets near wherever it is scheduled to get out of the way.

But this logic doesn't apply if the fast/slow mix is replaced by a skip-stopping mix with all the trains having roughly the same average speed ... as is proposed for Transpennine... There remains the safety issue of stopping trains in the middle of a long tunnel though ... especially diesels.

 

[MarkyT]

That is true where the binding constraint on a railway is the ability to re-occupy the same platforms to maximise train throughput unconstrained by any other issues. Namely a metro raikway, like the SWML Slow Lines.

The TransPennine route is however considerably more conplex than this.

The ruling constraint is the fact that a train stopping at Mossley, Greenfield, Marsden and Slaithwaite takes about 8-10 minutes longer between Stalybridge and Huddersfield than a non-stop train. Ditto stopping trains between Huddersfield and Leeds. Or indeed freight paths between Stalybridge and Huddersfield.

No manner of whizzy signalling does anything to sort this fundamental fact.

Platform reoccupation time improvement by closing up facilities (howsoever provided) can also be beneficial where a fast is catching up a slow at the more important stations where both stop. Of course beyond this the slow needs to turn off to get out of the way of the fast onto a slow line, or it could use a differnt platform at the major station to allow the fast to pass, assuming that exists or can be constructed.

I don't disagree that some targetted infrastructure enhancements, extra tracks and platforms will be required at critical locations to allow trains with different characteristics to get out of each others' way.

The next constraint is that Diggle Jn to Marsden is one block sectioms. This could be solved with conventional lights on sticks (or rather mounted to tunnels) if one wanted to.

I also do not disagree that conventional length blocks would be perfectly adequate for most plain line sections along the route. As to tunnels, provision of signals (or block markers) in tunnels is frowned on generally as evacuation is more difficult if a train becomes disabled or there is an on board emergency, and that is more likely if they are routinely capable of being stopped at a signal for forward traffic or in event of a signalling failure. Where unavoidable (such as protecting a junction immediately beyond the tunnel), signals in tunnels often have 'tunnel controls' applied that prevent the preceding signal from clearing unless the forward section is already clear. Intermediate blocks can be provided in long tunnels (as on metros) if absolutely necessary, but risk would need to be assessed very carefully and mitigations provided as appropriate: Maybe more lights, signs, communications etc.

Solve these, then you've still got flat junctions (Guide Bridge, Stalybridge), Bradley Wood, Thornhill) to solve and lastly city centre capacity at Manchester and Leeds. It is only the very latter where ETCS *mighrlt* make the slightest difference.

One area where ETCS does help is in safety, providing full train protection functionality (ATP) at every signal (or block marker) and speed restriction, something that TPWS currently does not (although could if desired). Expansion of TPWS with current rather dated technology would be a mistake in my opinion. A balise based solution, based on ETCS even if not a full implementation, would be more appropriate. The ETCS specs allow for that now with a technique known as 'limited supervision', a variation of the Level 1 implementation and now being rolled out in Switzerland over the entire network with completion planned for this year.

Please note that the notional level 3, even if it existed, would not in my opinion provide any further benefit over level 2 as the junction and station constraints are not overcome by being able to run trains at closer headways in plain line sections, assuming that was allowed.

 

[DerekC]

All this statement means is that they are going to spend £4M out of the £450M Digital Railway budget investigating whizzy traffic management as a potential improvement to the TPE route. It has nothing to do with ETCS. If you want a vague and woffly list of what Digital Railway is looking at read:

http://digitalrailway.co.uk/2016/12/07/driving-towards-digital-railway-deployment-establishing-investment-route-route/

As I understand it (source close to the horse's mouth) only the LNE project involves ETCS and the timescales for this are still extremely uncertain.

 

[DerekC]

Please note that the notional level 3, even if it existed, would not in my opinion provide any further benefit over level 2 as the junction and station constraints are not overcome by being able to run trains at closer headways in plain line sections, assuming that was allowed.

Level 3 could save lots of money and make the whole system much more flexible (you don't have to resignal if you want to increase plain line capacity to match an improvement made elsewhere). That's why it should be done.

 

[MarkyT]

There are also closing up signals at the entrances to each platform at Reading, which would also block the junction if a train stopped there. Presumably they don't allow a second train to approach these until they are pretty sure the previous one is about to move away from the platform.

One way to do this is to have the overlap for the closing up signal terminate a little way along the platform so that the tail of a train stopped in the platform is still occupying the overlap track circuit, preventing clearance of the home signal heading towards the closing up. The route can be over-set however, so soon after the train in the platform starts to move the overlap becomes clear and the home signal can step up to proceed, allowing the following train to begin moving towards the closing up signal. By the time it reaches the closing up signal, that is clear and a run straight into the platform without blocking the junction is ensured.

However I do suggest the long block section at Standedge shouldn't be too much of a problem with a fast/slow mix, since it is near the middle of the length of track shared by fast and slow trains. The trains are much closer together at either end of the section because the slow sets out just behind the previous fast and is being caught up by the next one as it gets near wherever it is scheduled to get out of the way.

But this logic doesn't apply if the fast/slow mix is replaced by a skip-stopping mix with all the trains having roughly the same average speed ... as is proposed for Transpennine... There remains the safety issue of stopping trains in the middle of a long tunnel though ... especially diesels.

I would have though there will still be some difference in running times, assuming there are some very limited stop services remaining to get the quick headline journey times. I expect it will be the stoppers that become more 'skip stop' in calling pattern, not to entirely equalise all journey times but to reduce the difference compared to the fastest paths.

 

[MarkyT]

Level 3 could save lots of money and make the whole system much more flexible (you don't have to resignal if you want to increase plain line capacity to match an improvement made elsewhere). That's why it should be done.

You need an awful lot of exceedingly reliable communications capacity for a notional fully moving block system and there are interesting difficulties in train integrity and length measurement that have still to be solved on general purpose railways. It is much simpler on metro type railways where all trains are a common length. If you assume all trains on a main line are the longest that ever run you will lose capacity, as main line trains can vary from half mile long freights to single car 153s. Radio communications and processing capability will have to be scaled around a target capacity which can be defined in terms of a numbers of trains in an area at one time or a headway frequency so it can't be assumed that an increase in required frequency can be had 'for free' and that enhanced service will still run into the same junction and station capacity constraints discussed earlier anyway. No main line rail network or manufacturer anywhere in the World has come close to solving these problems and level 3 will remain vapour-ware until they do. We should not be basing any strategies on such snake oil.

 

[Sunset route]

One way to do this is to have the overlap for the closing up signal terminate a little way along the platform so that the tail of a train stopped in the platform is still occupying the overlap track circuit, preventing clearance of the home signal heading towards the closing up. The route can be over-set however, so soon after the train in the platform starts to move the overlap becomes clear and the home signal can step up to proceed, allowing the following train to begin moving towards the closing up signal. By the time it reaches the closing up signal, that is clear and a run straight into the platform without blocking the junction is ensured.

Unless you have automatic restirictive overlaps on your closing up signals then then following train will eventually be signalled forward after the timing relays on the signal in rear time off, then you end with a 12 car snaking across an approaching junction blocking all lines. Aka East Croydon the other day when the train in the platform didn't depart as intended.

 

[DerekC]

You need an awful lot of exceedingly reliable communications capacity for a notional fully moving block system and there are interesting difficulties in train integrity and length measurement that have still to be solved on general purpose railways. It is much simpler on metro type railways where all trains are a common length. If you assume all trains on a main line are the longest that ever run you will lose capacity, as main line trains can vary from half mile long freights to single car 153s. Radio communications and processing capability will have to be scaled around a target capacity which can be defined in terms of a numbers of trains in an area at one time or a headway frequency so it can't be assumed that an increase in required frequency can be had 'for free' and that enhanced service will still run into the same junction and station capacity constraints discussed earlier anyway. No main line rail network or manufacturer anywhere in the World has come close to solving these problems and level 3 will remain vapour-ware until they do. We should not be basing any strategies on such snake oil.

The data rate needed to control a train is in the Kbits/second so to suggest that comms capacity will be a problem is a bit short sighted. As and when the railway catches up with where the rest of the world has already got to, it won't be a problem. Train integrity is easy to solve for multiple unit trains. It's more difficult for locomotive hauled trains, but there are a number of solutions which look promising. Of course a Level 3 system needs to know how long the trains are and I agree this isn't as easy as it sounds. It has a lot to do with integration between the rolling stock and the control system. A multiple unit train can tell you how long it is and the day when a freight train knows how many cars so that data input by the driver can be cross-checked can't be too far off.

Going on reciting the mantra about junction and platform capacity misses the point, which is that Level 3 will be much cheaper for whatever plain line capacity you want, and will cost much less to upgrade. For sure you will need more processing power and maybe comms capacity (although the data rate needed to control a train is very modest by modern standards - see above) but you won't need new interlockings, track circuits, signals, lineside cables and all the other stuff that goes with resignalling.

I accept that it is going to take a long time because the railway is very conservative and wedded to the infrastructure-based control that it has had for 150 years, but we will get there in the end. It's interesting to reflect that the first and most successful moving block control system was designed by aerospace engineers!

 

[MarkyT]

A multiple unit train can tell you how long it is and the day when a freight train knows how many cars so that data input by the driver can be cross-checked can't be too far off.

One thing I confidently predict is that at least some fixed train detection will survive around junctions, to prove they are definitely clear before points are allowed to move. This offers the possibility that trains could actually be measured by the fixed system at junctions and then that value stored for subsequent use in spacing them out through moving block plain line sections between. The front of a train would know its position from balises and odometry and its relative position to the end of the fixed train detection could be calculated at the moment the junction train detection went clear, giving a good approximation of the trains's length which could be cross checked with other sources such as driver consist input. Ideally a moving block system should contain most of its processing power on the trains themselves so the processing capacity scales with the number of trains and frequency of the traffic using it. Trains would effectively lock onto the train in front and most communications for the purposes of spacing would take place between trains rather than back and forth to radio block centres and the like, which would concentrate on managing block direction locking and continuing movement authority for the whole link and all traffic on it. The radio infrastructure would essentially consist of a series of repeaters along the route that allow the trains to remain in contact with each other or the end of section at the next junction.

 

[HSTEd]

The advantages of ERTMS L3 (characterised by on board train-complete detection) are not absolutely coupled to the idea of moving blocks though.

Hence the idea of Regional ERTMS.
If a train can tell if it is complete or not (passengers trains can feasibly do this now, and freight operators should be made to adopt ECP-type brakes, giving them this capability), then you have no need for plain line axle counters or track circuits at all. Blocks become arbitrary definitions in the memory of the interlocking computer.

Potentially huge amounts of trackside equipment can be eliminated, along with the attendant communication systems and power supplies.
The only remaining infrastructure away from junctions relates to GSM-R coverage, which need not be positioned adjacent to the railway and can be placed wherever is convenient for optimal coverage.

 

[MarkyT]

The advantages of ERTMS L3 (characterised by on board train-complete detection) are not absolutely coupled to the idea of moving blocks though.

Hence the idea of Regional ERTMS.
If a train can tell if it is complete or not (passengers trains can feasibly do this now, and freight operators should be made to adopt ECP-type brakes, giving them this capability), then you have no need for plain line axle counters or track circuits at all. Blocks become arbitrary definitions in the memory of the interlocking computer.

Potentially huge amounts of trackside equipment can be eliminated, along with the attendant communication systems and power supplies.
The only remaining infrastructure away from junctions relates to GSM-R coverage, which need not be positioned adjacent to the railway and can be placed wherever is convenient for optimal coverage.

The key thing about 'Regional' is that it is intermittent, i.e you only need radio coverage at and around the sites of the virtual fixed block boundaries and the train 'remembers' it's movement authority between these sites. In this respect it is very similar in concept to RETB. You don't need continuous radio coverage to refresh and maintain your movement authority for such long fixed blocks as you would conventionally in L2 or moving block L3. My idea for measuring train length could also be used with Regional, as it is likely that fixed train detection around junctions will be retained, as it was for the Swedish pilot project. Note it is fairly easy and economical to provide fixed train detection just around junctions, as by definition control I/O and power must be present in order to drive and detect the point machines. Adding a few more local sensors in the vicinity for an axle counter for dead locking is a small price to pay for the additional security provided to ensure points cannot under any circumstance move under a train, a most undesirable event, and that route locking of the junction is released as quickly as possible after passage.

 

[Olaf]

Traffic Management Systems - TMS for short - are being rolled out across the network, it's been going on for at least 4 years to my certain knowledge. None are actually operation so far as I am aware in the UK, Japan uses them quite a lot on their suburban routes, they are nothing new.

At least two systems are due to go LIVE in 2018.

 

[Olaf]

You can't have a block end in the middle of a junction; No amount of Digital Railway is going to fix the issue that many of the existing flat junctions are full to capacity , a function of the time each one takes to cross the path of another train - if you can't get through the junctions at the ends, what use is cmore apacity on the bits in between?
.

I was under the impression that a system approach was to be used going forward such that infrastructure and signalling are to be ungraded hand-in-hand on a route-section basis so that target capability is delivered by also addressing infrastructure bottle-necks prior to cut-out to new systems. Is this not already known about within NR?

 

[Olaf]

Surely the Philosopher's Stone is supposed to be Level 3, but as far as I know this doesn't exist on any main line anywhere yet. Maybe someone in Dft knows it's "only just round the corner" ...

Adopt was to be considered for the SWML upgrade options. I do not know what the outcome was.

 

[Olaf]

Going back a bit, there was this article from April this year:

ERTMS Level 3 - A possible way forward
http://www.railengineer.uk/2017/04/25/ertms-level-3-a-possible-way-forward/
- 2017-04-25

Selected highlights:

- whilst ERTMS Level 2 brings considerable benefit for interoperability and
some capacity gains, it was always envisaged that Level 3 would be the
ultimate goal as this offers significant cost savings for infrastructure
equipment.
- In short, ERTMS Level 3 has two main features over and above Level 2. It
facilitates much closer headways by the opportunity of adopting moving
block, meaning that trains in close succession can close up particularly
at lower speeds, and it allows for the removal of track-based train
detection equipment in the form of track circuits or axle counters, which
should reduce capital and maintenance costs and improve reliability.
- Level 3 ETCS (the signalling element within ERTMS) is based on a total
radio solution. A train¿s position is reported back to the RBC (Radio
Block Centre) at least every five seconds. This information is based upon
the data obtained from a series of track-mounted eurobalises (radio
beacons) provided at intervals dependent on the positional accuracy needed
(for instance, where a precision stop is required). The position reference
obtained is then incremented by accurate train-borne odometry that
calculates the distance travelled from the last balise.
- This constant updating of position and speed allows following trains to
run closer to the one in front by adjustment of the MA (Movement
Authority) information displayed to the driver. The potential gain in
capacity is significant.
- Why has Level 3 not yet been developed into a standard way of working?
There are a number of issues but the two main factors are that the train
has to be proven as complete ¿ that it has not become uncoupled en route
with part of the train left behind ¿ and the radio system has to be
completely reliable.
- Hybrid Level 3, it has been in development since 2013 as a joint effort by
Network Rail and ProRail with Alstom and Bombardier both supplying
equipment that demonstrated the feasibility.
- Hybrid Level 3 system is to retain any existing track-circuit or
axle-counter sections and to then create ¿virtual blocks¿ as sub sections
within these.

Of note is the proposal:

Early 2018 – an early deployment trial on a chosen route;

This may be on part of the NTP route.

 

[swt_passenger]

Adopt was to be considered for the SWML upgrade options. I do not know what the outcome was.

What is "Adopt" in this context please?

 

[MarkyT]

Going back a bit, there was this article from April this yearRTMS Level 3 - A possible way forward
http://www.railengineer.uk/2017/04/25/ertms-level-3-a-possible-way-forward/ - 2017-04-25

This seems very similar to the regional concept with virtual fixed L3 blocks to divide up the longer L2 blocks that are furnished with conventional fixed train detection, and the technique might also be able to be overlaid on traditional signalling as with Thameslink L2, but without the need for more train detection sections. Again you only need to apply the additional L3 virtual blocks where necessary for capacity, mostly for closing up on and dividing busy platforms to minimise the reoccupation time. Conventional length blocks elsewhere, whether real or virtual, will probably be adequate.

 

[MarkyT]

Unless you have automatic restirictive overlaps on your closing up signals then then following train will eventually be signalled forward after the timing relays on the signal in rear time off, then you end with a 12 car snaking across an approaching junction blocking all lines. Aka East Croydon the other day when the train in the platform didn't depart as intended.

Ouch! I guess the only way out of that is not to set the route from home to closing up until the full overlap is clear, but having to do this at a particular critical time is unlikely to be practical on a busy panel.

 

[Olaf]

What is "Adopt" in this context please?

Sorry, jumped ahead of myself while typing.

Adoption of ETCS L3 (with and without ATO) was to be considered as one of the options for capacity enhancements on the inner sections of the SWML. The planning period was for a thirty year period but I am assuming that some effort has been expended on this by now.

 

[Olaf]

This seems very similar to the regional concept with virtual fixed L3 blocks ...

Yes, that is exactly what they are looking at as far as I am aware.