Will advances in rail technology reach a dead end at some point?

[778]

Will there come a time when rail technology will reach a dead end with little or no scope of improvement?

The maximum speed of trains on conventional tracks is about 350 mph, so eventually any new trains built will not see any advances in speed. Trains built in the last 10 years are not that much faster than trains built in the 80s/90s, apart from better acceleration and braking. Acceleration and braking also has limits, so at some point journey times will not get any faster.

There are lots of things that could improve the railways in the short to medium term like automation (driverless trains), alternative power to replace diesels (batteries, hydrogen etc) and digital signalling, but even if all of those things happen will rail technology eventually reach its limits, where no more advancement is possible?

 

[AM9]

Will there come a time when rail technology will reach a dead end with little or no scope of improvement?

The maximum speed of trains on conventional tracks is about 350 mph, so eventually any new trains built will not see any advances in speed. Trains built in the last 10 years are not that much faster than trains built in the 80s/90s, apart from better acceleration and braking. Acceleration and braking also has limits, so at some point journey times will not get any faster.

There are lots of things that could improve the railways in the short to medium term like automation (driverless trains), alternative power to replace diesels (batteries, hydrogen etc) and digital signalling, but even if all of those things happen will rail technology eventually reach its limits, where no more advancement is possible?

== Doublepost prevention - post automatically merged: 6 May 2023 ==

Sorry, I should have posted this in "Speculative Discussion". My mistake.

Well ask the same question of road transport. Minor improvements in power, torque, comfort and economy. All the effort is directed to those things that are either mandated in law, or designed to perpetuate the demand for replacement, i.e. planned obsolescence. Journey times on most roads will however increase as congestion increases.

 

[TurbineEnginee]

Timetables set every six months or whatever seems old.

Maybe a more digital process which matches supply with demand in a more dynamic way.

Might combine with digital signalling and ATO.

Trains will run when closer to when people currently need them, rather than use a 12 month old estimate, based on dissatisfaction with the prior estimate from 12 months earlier.

 

[zwk500]

I suspect 400kph will be about the limit of economics for conventional railways purely because of the practicalities of building alignments suitable.
However efficiency and comfort.asvamces will continue, as will traffic management and signalling advances. I wouldn't be surprised to see Guard-Only operation at some point.

 

[Fleetmaster]

Obviously far into the future something will have happened to drastically change our ridiculous approach to massive engineering projects in this country, and assuming something even better hasn't arrived by then, we too can enjoy the benefits of super fast trains where the limits of what is possible genuinely are more about the fundamental issues of the laws of physics.

A hard limit will still be reached well before what is possible and indeed practical in simple engineering terms has been fully exhausted. It's a simple matter of the law of diminishing returns. I am pretty sure this is already evident in bullet trains, where for each step change in performance, there is an almost exponential rise in cost to get there.

Although clearly even there we are still at the stage where one new idea or discovery can unlock major advances. But the hard limit is out there. Perhaps closer than we think.

 

[Irascible]

It's generally improvements in materials driving things ( this applies to all transport anyway ) - solid state power control has lead to simplification elsewhere in the powertrain, new materials & welding techniques result in improvements being possible in body design, better cfd means better aerodynamics & also parallel improvements in trackwork. There are areas like IT where there's a vast amount of possibilities. I think there's bern more revolutionary changes in rail than commercial aviation in the past 40 years, we're a long way from running out of ideas.

 

[zwk500]

Timetables set every six months or whatever seems old.

Maybe a more digital process which matches supply with demand in a more dynamic way.

Might combine with digital signalling and ATO.

Trains will run when closer to when people currently need them, rather than use a 12 month old estimate, based on dissatisfaction with the prior estimate from 12 months earlier.

Timetables are already moving to 3 times a year, but that's an organisational thing not a technology issue.

Advances in passenger counting and ticketing technology might hopefully lead to more accurate data about which journeys are actually being made though.

 

[D365]

... and digital signalling...

Technically, we've had digital signalling for centuries

But I understand your point; communications-based onboard signalling is going to be a key advancement in GB railway signalling.

 

[zwk500]

Technically, we've had digital signalling for centuries

Not quite true: Absolute block is binary but not reprogrammable - it is very much a 'mechanical computer', not a digital one. If you wanted to change the interlocking you had to effectively rebuild the interlocking every time.

But I understand your point; communications-based onboard signalling is going to be a key advancement in GB railway signalling.

It's going to be a noticeable improvement but in general it won't be the game changer politicians and senior management keep saying it is. It's not like In-cab signalling or ATP functionality is new - the GWR had in-cab indications and HS1 has been running TVM for years.

 

[stuu]

I suppose if you look far enough into the future with automation there is the possibility for "trains" to split and join automatically en-route, avoiding the need to change. Or AI controlled networks regulating all the trains in an area so nothing stops at a red signal and no connections are missed?

But really the advances are going to be at the margin. Slightly more efficient, slightly quieter, slightly higher capacity etc. There is unlikely to be a big leap forward in the way that electrification was.

 

[rf_ioliver]

Will there come a time when rail technology will reach a dead end with little or no scope of improvement?

The maximum speed of trains on conventional tracks is about 350 mph, so eventually any new trains built will not see any advances in speed. Trains built in the last 10 years are not that much faster than trains built in the 80s/90s, apart from better acceleration and braking. Acceleration and braking also has limits, so at some point journey times will not get any faster.

There are lots of things that could improve the railways in the short to medium term like automation (driverless trains), alternative power to replace diesels (batteries, hydrogen etc) and digital signalling, but even if all of those things happen will rail technology eventually reach its limits, where no more advancement is possible?

350kmh seems to the limit based on journey time, capacity, comfort etc. The 560+kmh tests showed that the TGV could have been pushed further, but cost, need and other physical constraints, eg: length of track , come into play; overall however the speed record there was a bonus to the scientific part.

At the moment, from my experience, the advances are coming from technological upgrades and digitalisation - for example, the move from GSM-R to 5G brings a number of communications and control improvements; this coupled with more advanced and dynamic planning and management means more capacity, speed etc.

Are the upper limits, other than the laws of physics, yes; are we going to reach these? Probably not, economics has more of an effect first.

 

[Magdalia]

My impression is that the railway has barely scratched the surface when it comes to using big data and AI to improve the way it operates.

On timetabling, the railway ought to have models, a bit like those used by weather forecasters, to simulate timetable disturbances to identify timetabling scenarios that are wonderful when everything runs on time but fall apart when they don't. There's possibly an analogy with gaming here too.

And the railway ought to be able to manage supply not just demand with a capability to switch paths on long distance routes between destinations. This example is a bit of a moot point because of industrial action, but the railway ought to be able to reallocate supply onto Euston-Liverpool for the Eurovision Song contest but onto Manchester-Euston for the FA Cup Final.

On the demand side big data has huge potential for analysis of ticket sales. But the railway would be doing this with one hand tied behind its back while it still has such a complicated ticketing system that encourages split ticketing, thus rendering the data much less useful for analysis.

Some of this may be happening, but looking from the outside there's not much evidence of it.

 

[zwk500]

My impression is that the railway has barely scratched the surface when it comes to using big data and AI to improve the way it operates.

Agree

On timetabling, the railway ought to have models, a bit like those used by weather forecasters, to simulate timetable disturbances to identify timetabling scenarios that are wonderful when everything runs on time but fall apart when they don't. There's possibly an analogy with gaming here too.

They do. The models are hampered though by the information about the physical network that is made available to them, and the natural limitations of timetable modelling.

And the railway ought to be able to manage supply not just demand with a capability to switch paths on long distance routes between destinations. This example is a bit of a moot point because of industrial action, but the railway ought to be able to reallocate supply onto Euston-Liverpool for the Eurovision Song contest but onto Manchester-Euston for the FA Cup Final.

This can be done if wanted already, but you either end up wasting capacity at the end you're not using or having to bugger everybody else about to fit the extra path in.

On the demand side big data has huge potential for analysis of ticket sales. But the railway would be doing this with one hand tied behind its back while it still has such a complicated ticketing system that encourages split ticketing, thus rendering the data much less useful for analysis.

I'm sure this is done already, but certainly not at the scale it should be doing.

 

[Irascible]

Intelligence in electric power delivery - lets say all trains have battery packs and some substations too - could quite possibly save a chunk of wasted electricity & improve performance.. Even just granular power switching might, if systems know exactly where a train is... where could one get that data I wonder

Freight could be improved a lot, but not with current margins I guess. Plenty of room for technology to advance though.

I believe there was a report on European maglev that listed some effective limits for HSR, I dont remember if 300 or 350kph was the peak. Changes in airline economics might change that considerably, but just going faster isn't really a development at this point.

Superconductors would change so much if there's a practical version even if it needs some refrigeration still.

Maybe there'll be a breakthrough in seat design...

 

[zwk500]

Intelligence in electric power delivery - lets say all trains have battery packs and some substations too - could quite possibly save a chunk of wasted electricity & improve performance.. Even just granular power switching might, if systems know exactly where a train is... where could one get that data I wonder

Freight could be improved a lot, but not with current margins I guess. Plenty of room for technology to advance though.

I believe there was a report on European maglev that listed some effective limits for HSR, I dont remember if 300 or 350kph was the peak. Changes in airline economics might change that considerably, but just going faster isn't really a development at this point.

Superconductors would change so much if there's a practical version even if it needs some refrigeration still.

Maybe there'll be a breakthrough in seat design...

China has lots of 350kph lines, HS2 is being designed for 400kph theoretical max. France seems to have settled on 320kph though.

 

[Fleetmaster]

I think there's bern more revolutionary changes in rail than commercial aviation in the past 40 years, we're a long way from running out of ideas.

That's a big claim.

What have been the transformative changes in railways that compare to the switch to joystick control and fly by wire, or the switch to composite materials for large but critical surfaces like wings and tail planes?

And for more gradual changes, what in rail terms compares to how range, fuel economy and interior specifications have been on an ever upward trajectory, to the point that over a span of forty years, they look like a revolution.

Forty years on, which is well past the time slam doors wooden bodies was the cutting edge, and trains are very much still metal tubes with electrical/onic drive and controls and basic reclining seats with tables. Not so much for the planes. Both have computers now, but the differences in what they're doing is vast.

 

[MarkyT]

TGV-M is designed for 350kph, but that might be for neighbouring foreign networks SNCF intend some to operate through services over, or for sales to other nations.

 

[zwk500]

TGV-M is designed for 350kph, but that might be for neighbouring foreign networks SNCF intend some to operate through services over, or for sales to other nations.

There are a couple of lines in France that are 350kph.

 

[Silenos]

Timetables are already moving to 3 times a year, but that's an organisational thing not a technology issue.

Advances in passenger counting and ticketing technology might hopefully lead to more accurate data about which journeys are actually being made though.

A timetable that is constantly changing is the same as no timetable at all. Passengers need to be able to plan, so I imagine that 2 or 3 changes a year is probably the most that would be acceptable

 

[zwk500]

A timetable that is constantly changing is the same as no timetable at all. Passengers need to be able to plan, so I imagine that 2 or 3 changes a year is probably the most that would be acceptable

Well indeed -as the last few years has shown.

 

[Silenos]

To return to the original point of the thread, it seems likely that railway technology will continue to develop, but probably not in the UK, which can always find reasons why it is a special case and things that work elsewhere won’t work here.

 

[NoRoute]

I think the next major advances to rail technology will come from convergence with road vehicles, from the self-driving technologies being developed for road vehicles, once they are refined and we see the emergence of self-driving vehicles with platooning capabilities, driving in convoys along roads, it isn't then a huge leap to move to a point where the same technology is applied to an articulated train able to operate on a conventional road surface.

At that point with the train now capable of steering itself along a route, there's no need for a rail to guide it, it becomes possible to get rid of the rails, replacing the costly and inflexible rail infrastructure with a simple road surface, navigated by self-driving vehicles.

 

[zwk500]

I think the next major advances to rail technology will come from convergence with road vehicles, from the self-driving technologies being developed for road vehicles, once they are refined and we see the emergence of self-driving vehicles with platooning capabilities, driving in convoys along roads, it isn't then a huge leap to move to a point where the same technology is applied to an articulated train able to operate on a conventional road surface.

There's talk of an ETCS L4 with 'virtual coupling' but it's a very, very long way off.

At that point with the train now capable of steering itself along a route, there's no need for a rail to guide it, it becomes possible to get rid of the rails, replacing the costly and inflexible rail infrastructure with a simple road surface, navigated by self-driving vehicles.

The efficiency of steel rails and steel wheels for bulk loads and high speeds will still outweigh the disadvantages of inflexible road infrastructure. We will see a rebalancing of the modal choice, but railways will continue to have a role. If they couldn't be killed off in the 70s motorway heyday, they won't die now.

 

[MarkyT]

There's talk of an ETCS L4 with 'virtual coupling' but it's a very, very long way off.

I'm not sure that'll ever result in any high-speed virtual coupling/decoupling en route. The high energy collision risks of a malfunction during the process are probably just too high. And then there's the switching issue. If a virtual consists splits in motion, then the two parts must separate sufficiently for the rear one to be far enough behind for junctions to be commanded and established for an alternative route without 'cautionary aspects' or their modern control system equivalents. Alternatively, such a 'slipping' technique might only be used where a stopping portion is dropped from a through express and no junction switching is required. Self-switching techniques where each vehicle chooses its own route are a feature of some pod-based gadgetbahns, which helps to overcome some of these spacing issues but they are probably more difficult to apply reliably and safely to larger vehicle systems and inevitably result in greater complexity.

The efficiency of steel rails and steel wheels for bulk loads and high speeds will still outweigh the disadvantages of inflexible road infrastructure. We will see a rebalancing of the modal choice, but railways will continue to have a role. If they couldn't be killed off in the 70s motorway heyday, they won't die now.

Especially now and going forward into the foreseeable future with the emphasis on energy efficiency. If a free virtually weightless unlimited and unpolluting power source suddenly becomes available then clearly all bets are off! Rubber tyres, particularly those of heavy vehicles, are probably practically and economically limited to around 80mph commercial top speed. Material shedding from rubber tyres as they wear is also an environmental pollution issue and largely a result of the much higher friction interface. There is a related issue with steel wear on rail, but the quantities are much smaller and the material is more benign, although there may be human health issues, particularly in tunnel environments, from dust, as with brake pad wear. On all kinds of vehicles, brake wear can be minimised by using regenerative techniques as much as possible, which with storage techniques, also assists with energy efficiency and power demand management, hence the likely future dominance of electrical drive-trains, whatever the source of motive power. The efficiency and weight of motor/generator and storage technologies in rail will continue to improve in line with broader transport and industrial trends.

 

[TurbineEnginee]

I suppose if you look far enough into the future with automation there is the possibility for "trains" to split and join automatically en-route, avoiding the need to change. Or AI controlled networks regulating all the trains in an area so nothing stops at a red signal and no connections are missed?

But really the advances are going to be at the margin. Slightly more efficient, slightly quieter, slightly higher capacity etc. There is unlikely to be a big leap forward in the way that electrification was.

Your first part is what I was trying to illustrate.

6 months in advance we organize to have a block of iron and some staff moving from here to there at this specific time. Customer go figure it out.

Manufacturing went Just In Time and achieved substantial benefit, for the most part.

That’s about blocks of iron and staff and customers too, but somehow they did it.

It seems to me that a JIT railway could produce much greater benefit than achieving insanely high line speeds.

 

[edwin_m]

6 months in advance we organize to have a block of iron and some staff moving from here to there at this specific time. Customer go figure it out.

There's a clue in the word "customer". They'd like to know in advance when their service operates, which makes it difficult to change a passenger timetable at the drop of a hat (though we seem to be doing that a lot at present...). The long lead time is to a large degree down to processes, where all parties have to bid and agree on a timetable, rather than technology.

 

[Peregrine 4903]

Timetables are already moving to 3 times a year, but that's an organisational thing not a technology issue.

Advances in passenger counting and ticketing technology might hopefully lead to more accurate data about which journeys are actually being made though.

They aren't yet, its still two for now with an opportunity to make changes at a third point but not a defined timetable change.

 

[Snow1964]

Seems Competition Authorities are not enthusiastic about merger of Hitachi and Thales regarding digital signalling

Merger could mean Network Rail and TfL lose out on digital signalling options

CMA’s investigation provisionally concludes that Hitachi’s takeover of Thales GTS could reduce choice for Network Rail and TfL, impacting cost and service quality for passengers

www.gov.uk

 

[MarkyT]

Seems Competition Authorities are not enthusiastic about merger of Hitachi and Thales regarding digital signalling

Merger could mean Network Rail and TfL lose out on digital signalling options

CMA’s investigation provisionally concludes that Hitachi’s takeover of Thales GTS could reduce choice for Network Rail and TfL, impacting cost and service quality for passengers

www.gov.uk

An interesting dilemma. Hitachi bought Ansaldo signalling some years ago. They haven't carried out significant work on UK mainline applications apart from the Cambrian ETCS pilot scheme and the earlier South Manchester conventional resignalling project for WCML modernisation which wasn't considered a success and hasn't been repeated. They also have UK oriented control centre products from their purchase of TRE.

Thales (and predecessors) on the other hand have not been a UK mainline supplier at all apart from overlaid train protection tech including the SelCab ATP pilot scheme on Chiltern, now obsolete, and of course TPWS, rolled out nationwide. On TfL however, Thales has come to dominate, with all recent signalling schemes using SelTrac on LUL and DLR, apart from the Victoria line.

An alternative view to that expressed by the CMA is that, in the UK at least, the combined forces of Hitachi and Thales might stand up better to the traditional duopoly of Siemens and Alstom (formerly Westinghouse and GEC) in signalling. The EU objected strongly to a merger of those two dominant players quite rightly, and Alstom eventually took over the rail interests of Bombardier instead, when they became available.

 

[D365]

I was wondering why I hadn't heard of any progress regarding this merger.