Could digital signalling systems be used as a workaround for power supply limitations?

[Crithylum]

Signalling systems already prevent too many trains from occupying the same section of track. Could they also be used to prevent too many electric trains sucking up too much power in a section fed from a feeder station? Trains could be provided ’power authority’, where a computer guesses* how much power a train is likely to need, and if the driver demands more power, either power authority can be transferred from other trains, or excess/spare power can be allocated. Would this be realistic to implement as a feature of ETCS (or some other system)?

*Either a hard coded algorithm or machine learning that improves over time

 

[Harpo]

Signalling systems already prevent too many trains from occupying the same section of track.

Don’t get taken in by the ‘digital’ or ‘smart’ hype. Signalling is still overwhelmingly binary (even ETCS) with the correct answer being one train or zero in section.

Could they also be used to prevent too many electric trains sucking up too much power in a section fed from a feeder station?

Signalling doesn’t know if a train is using the OHLE or not. The UK’s myriad electrification-avoiding traction fudges mean that signallers might not know for certain either

 

[D365]

Signalling doesn’t know if a train is using the OHLE or not. The UK’s myriad electrification-avoiding traction fudges mean that signallers might not know for certain either

In theory ETCS will have the capability to open and close VCBs (vacuum circuit breaker) for neutral sections within OHLE. In reality, though, that is some way off from happening.

 

[SynthD]

This is highly plausible, but hard to do to a safety/performance critical standard. Yes, the computer controller for a wider area can know about the sections powered by each substation and the current acceleration/draw of each train in that section. Adding on safe predictions, like what a train due to leave in one minute will be drawing in one minute, it could tell a BEMU to pause using OHL for 5 minutes. With how timid the railway rightly is about dead trains on active rail, I wonder how this would ever be tested. At night, or on the Derby test track?

 

[BerkshireRails]

Maybe the solution is to add a 5th digit to the headcode to tell the signalling system what sort of traction is on the service, or bodge a system with TOPS.

 

[Harpo]

Maybe the solution is to add a 5th digit to the headcode to tell the signalling system what sort of traction is on the service, or bodge a system with TOPS.

Doesn't help with tri-mode/bi-mode services that might/might not be panned-up.

 

[The Planner]

Maybe the solution is to add a 5th digit to the headcode to tell the signalling system what sort of traction is on the service, or bodge a system with TOPS.

The cost of that would truly be astronomical. You would be changing all the NX panels that show a headcode and VDU systems.

 

[Snow1964]

What it could possibly do, is stagger release of stop (red equivalent) signals, to avoid all trains drawing power together after disruption and coming to a stand in adjacent sections. In normal operations shouldn't be needed as system should cope with normal service levels.

But it's different question if you should have weak electric supply sections that can't handle multiple trains, that can be signalled legitimately.

It seems fairly basic idea that if you install signalling that allows trains every 3-4 minutes, then rest of infrastructure should allow same frequency, otherwise why not save money and only install longer block sections.

 

[MarkyT]

I think a better strategy is for all electric traction to carry a moderately sized battery on board. In addition to service resilience benefits in planned or emergency isolations, batteries could also allow stored power to be blended with line power for short periods, maintaining performance while reducing current draw during extreme peaks of demand that affect line voltage.

 

[BerkshireRails]

The cost of that would truly be astronomical. You would be changing all the NX panels that show a headcode and VDU systems.

It might be cheaper to resignal NX and VDU boxes into ROCs. I was thinking it would be a digit which would only show when there is space on the panel.

I think a better strategy is for all electric traction to carry a moderately sized battery on board. In addition to service resilience benefits in planned or emergency isolations, batteries could also allow stored power to be blended with line power for short periods, maintaining performance while reducing current draw during extreme peaks of demand that affect line voltage.

That seems to be the way forward with Southeastern wanting that to enable there metro trains to keep running during trespassing events.

 

[HSTEd]

The cost of that would truly be astronomical. You would be changing all the NX panels that show a headcode and VDU systems.

Are all the train 'classes' in common use these days?
Could you do something by moving the train code classes around to create separate electric and diesel ones?

 

[The Planner]

It might be cheaper to resignal NX and VDU boxes into ROCs. I was thinking it would be a digit which would only show when there is space on the panel.

That wouldn't make it any cheaper either, the costs of a resignalling of a single major box is well into to 9 figures.

Are all the train 'classes' in common use these days?
Could you do something by moving the train code classes around to create separate electric and diesel ones?

Class 7 and 8 have minimal use.

 

[Brubulus]

There is a secondary workaround, which is to fit batteries to feeder stations, since they are unlikely to be consistently working at full power and a battery there would require less co-ordination and cost than on trains or via signalling.

 

[MarkyT]

There is a secondary workaround, which is to fit batteries to feeder stations, since they are unlikely to be consistently working at full power and a battery there would require less co-ordination and cost than on trains or via signalling.

I'd have some storage onboard AND trackside for maximum resilience. Other storage methods, such as flywheels, are also available, which might be excellent for trackside installations. With fewer weight and size constraints trackside, the near-endless charge cycle life of flywheels and the short storage time typically required by the application seem well-suited. 25 years ago, LU trialled some on the Piccadilly Line and they planned to fit them widely on the Central, but that work never went ahead.

 

[zwk500]

Are all the train 'classes' in common use these days?
Could you do something by moving the train code classes around to create separate electric and diesel ones?

And then trains would need to change headcode on the move when transitioning through an APCO. Or needing to change headcode if there's a set swap.
Headcode 'classes' probably do need a review to enable greater distinction of passenger traffic, but that's a different thread.

It might be cheaper to resignal NX and VDU boxes into ROCs. I was thinking it would be a digit which would only show when there is space on the panel.

It wouldn't be cheaper to resignal anything. And I'm fairly sure that the train describer equipment wouldn't be able to recover the 5th digit once it's dropped, if any of the newer systems can even handle 5 digits at all. To show a 5th digit in many areas would require a quite substantial redesign of the layout on the VDU, even in ROCs, as the berths will be 25% longer.

The answer to the problem is to move the impetus for power supply management to train-borne power draw management systems, ideally supplemented by a on-board battery. Substations could also be provided with a trickle-charged local battery as well to cover for short-term high demand.

 

[Snow1964]

Are all the train 'classes' in common use these days?
Could you do something by moving the train code classes around to create separate electric and diesel ones?

Class 7 and 8 have minimal use.

As 7 and 8 are nowadays fairy rare, would be relatively easy to merge them, and use the other number for electric freight, thus identifying high power trains.

Some numbers have been reclassified before, eg 3 was parcels, livestock and perishables (milk, fish etc) trains, now also used for priority empty trains, 9 has been reclassified into passenger trains from mineral / empty freight train. So another reclassification should be easy

 

[HSTEd]

There is a secondary workaround, which is to fit batteries to feeder stations, since they are unlikely to be consistently working at full power and a battery there would require less co-ordination and cost than on trains or via signalling.

Well if we are talking about static batteries, you could go for some pretty esoteric battery technologies.

Sodium-Sulphur batteries are potentially really cheap per kilowatt hour and have pretty good performance characteristics. They are just rather hard to keep on a moving vehicle due to being at 200 celsius.

In a substation it would scarcely matter.

 

[BerkshireRails]

Or very simple solution you limit total number trains to what the electrical supply can handle.

 

[zwk500]

As 7 and 8 are nowadays fairy rare, would be relatively easy to merge them, and use the other number for electric freight, thus identifying high power trains.

Some numbers have been reclassified before, eg 3 was parcels, livestock and perishables (milk, fish etc) trains, now also used for priority empty trains, 9 has been reclassified into passenger trains from mineral / empty freight train. So another reclassification should be easy

Reclassification would not be "easy", you're changing the headcode of a vast number of trains, not to mention that the classes are embedded into operational practice such as instructions. Every single 'goods line' instruction that refers to class 4, 6, 7, and 8 traffic would need to be rewritten, for example. I don't know if that would have other knock-on impacts to equipment such as level crossings or signalling equipment. There's a vast amount of person-hours in identifying everything that would need changing and then making the changes, let alone deciding on what classifications to use!

However, reclassification is a separate thread (if there hasn't been one before).

== Doublepost prevention - post automatically merged: 27 Aug 2025 ==

Or very simple solution you limit total number trains to what the electrical supply can handle.

And in areas where that number of trains, even at maximum length, can't match demand?

 

[The Planner]

Or very simple solution you limit total number trains to what the electrical supply can handle.

Which happens now.

== Doublepost prevention - post automatically merged: 27 Aug 2025 ==

And in areas where that number of trains, even at maximum length, can't match demand?

You declare congested infrastructure. Its likely to happen on some lines.

 

[zwk500]

You declare congested infrastructure. Its likely to happen on some lines.

Which requires a plan to be developed to resolve the congestion, so what's the plan? Either you upgrade the power supply, or you provide alternative methods of managing the power demands (this assumes that track and signaling capacity is higher than the electrical capacity).

 

[The Planner]

Which requires a plan to be developed to resolve the congestion, so what's the plan? Either you upgrade the power supply, or you provide alternative methods of managing the power demands (this assumes that track and signaling capacity is higher than the electrical capacity).

You plan to upgrade, doesn't mean you need to be exact on a date.

 

[MarkyT]

You plan to upgrade, doesn't mean you need to be exact on a date.

There needs to be a 'predict and provide' strategy, perhaps.

== Doublepost prevention - post automatically merged: 27 Aug 2025 ==

Well if we are talking about static batteries, you could go for some pretty esoteric battery technologies.

Sodium-Sulphur batteries are potentially really cheap per kilowatt hour and have pretty good performance characteristics. They are just rather hard to keep on a moving vehicle due to being at 200 celsius.

In a substation it would scarcely matter.

It's like the mercury arc rectifier period all over again!

 

[Bald Rick]

You plan to upgrade, doesn't mean you need to be exact on a date.

Or exact on where the money is coming from!

 

[Harpo]

Or exact on where the money is coming from!

As someone from the erstwhile ‘Operating Strategy’ said to me recently, ‘we burnt through the final millions descoping everything’.

 

[BerkshireRails]

Which happens now.

== Doublepost prevention - post automatically merged: 27 Aug 2025 ==


You declare congested infrastructure. Its likely to happen on some lines.

The Great Western Mainline between London and Reading is congested infrastructure.

 

[mike57]

Would mixing 'control' with 'safety' be a problem? In my industry you keep your safety systems separated (in so far as is possible...) from the control systems, would this apply to railways?

 

[BerkshireRails]

Would mixing 'control' with 'safety' be a problem? In my industry you keep your safety systems separated (in so far as is possible...) from the control systems, would this apply to railways?

Don't think so as you combine them with ETCS (European Train Control System) which does both control and saftey in one. There is other systems like CBTC (Communications Based Train Control) which used on Elizabeth Line as well as ETCS. Sometimes ETCS is known ERTMS (European Rail Traffic Management System) this includes the GSM-R (Global System Mobile - Railway).

 

[zwk500]

Sometimes ETCS is known ERTMS (European Rail Traffic Management System) this includes the GSM-R (Global System Mobile - Railway).

Incorrectly. ETCS and GSM-R are separate components of ERTMS. The RSSB refer to ERTMS as their position is that the safety system involves both ETCS and GSM-R. However there are non-safety related components of ERTMS as well.

 

[BerkshireRails]

Incorrectly. ETCS and GSM-R are separate components of ERTMS. The RSSB refer to ERTMS as their position is that the safety system involves both ETCS and GSM-R. However there are non-safety related components of ERTMS as well.

Network Rail western sectional appendix refers to ERTMS L2 between Ealing Broadway and Heathrow/Stockley Bridge Junction, when Great Western Railway refers to it as ETCS L2.