Route learning requirements

[TurbineEnginee]

Moderator note: Split from

Transpennine woes and a moan...

I had the (misfortune) to travel over to the North West yesterday for work, and the level of Transpennine North reliability seems to be reaching a low point again. it looks like there was a 3 hour gap in the Liverpool - Newcastle service from 15:40 to 18:31. I realised that the 17:32 ex Man Vic...

railforums.co.uk

The amount of route learning needed to make this happen, from either end, would be astronomical. The timetable is also unit based rather than traincrew based, so there wouldn't be enough time for a break at either end before the return working. So you would have to either have relief en-route, or to take back a different train.

With LIV-GLC only running every 4 hours (even in the theoretical, full timetable) that means you would end up with lots of pass rides to make it all work, possibly only being able to work trains in one direction due to way the timings work out. That's certainly how it used to be when Manchester Airport guards went to Edinburgh.

It sounds simple in theory but as soon as you dig under the surface, it becomes a whole lot more complicated.


The maximum continuous driving at most TOCs is at least 4 hours, and for TPE it's 4.5 hours. With LIV-GLC being just under 3.5 hours on most runs that would likely be OK, although at the upper limit (considering you want to allow some contigency in case of delays).

I hope you don’t mind me asking because you made me so curious.

You say “The amount of route learning needed to make this happen” and I’m sure you are correct 100%.

This is what confuses me.

I have an automotive driving license and within weight and axle limits I can drive any vehicle on any road worldwide.

I have a commercial pilots license and I can fly any qualified type in any (unrestricted) airspace worldwide and with some very limited exceptions I am qualified to use any airfield on the planet.

What is it that is different about operating a train that requires “route learning” when road vehicles and aircraft do not?

 

[Watershed]

I hope you don’t mind me asking because you made me so curious.

You say “The amount of route learning needed to make this happen” and I’m sure you are correct 100%.

This is what confuses me.

I have an automotive driving license and within weight and axle limits I can drive any vehicle on any road worldwide.

I have a commercial pilots license and I can fly any qualified type in any (unrestricted) airspace worldwide and with some very limited exceptions I am qualified to use any airfield on the planet.

What is it that is different about operating a train that requires “route learning” when road vehicles and aircraft do not?

You absolutely make a fair point. I guess possibly the biggest factor is simply that "it's always been that way" and it's (perceived as) very difficult to change established working practices in the industry.

There are also some differences between how the railways operate and how cars and aircraft operate. On the roads and in the air, everything is highly standardised. You learn what to expect on a given type of road/airport/approach etc. when you get your licence.

Any deviation from that norm will be clearly signposted - for instance, if the default speed limit (30mph on roads with lighting at regular intervals, 60mph elsewhere on single carriageways) doesn't apply, there will be a sign to tell you that. And regular repeater signs to remind you. If an ILS approach is non-standard in some way, the approach plate will explain that. Taxiway and runway signage is internationally standardised.

By contrast, on the railways, whilst there's a common Rulebook that all drivers have to learn, it lags behind in terms of the signage and information that drivers get. For instance, not all changes of linespeed have a warning board. There's nothing telling drivers where to start braking for their next station stop. And drivers have to interpret the signals to work out whether they're allowed to proceed (for instance, whether the platform they've been routed into is long enough for their train). The list goes on.

It's a very different mindset and whilst it's clearly resulted in excellent levels of safety, you're right to point out how inflexible it makes the railway. In reality, I think that technology could replace a huge amount of the route learning that goes on. Lineside balises should be placed every few miles so that, together with wheel rotations, trains know exactly where they are at any given point. Drivers should then have a display that tells them all the features of the line that they need to know about, including things such as where the optimal braking point is likely to be. There should be more automation so that drivers don't have to check whether they've been routed correctly - it shouldn't be possible to mis-route trains.

Unfortunately whilst all that technology is perfectly feasible, and has been for at least the last 10 or 20 years, the amount it would cost to move to that sort of operation would be astronomical. And so with the short-termism that's afflicted the industry since privatisation, nobody has deemed it worthwhile to make that investment in order to improve the flexibility of the railway and reduce the cost and time it takes to train new drivers.

 

[LNW-GW Joint]

It should be possible to have a fully standardised driving environment on HS2 with ETCS signalling in the cab.
But then most HS2 trains (except Birmingham shuttles) will leave that modern system after about an hour onto the complexities of the classic WCML.
It will be interesting to know how DfT/HS2 Ltd intends to operate HS2 - considering that Avanti are supposed to be the operator.

Whether you could drive purely to the cab indications on the ETCS-fitted Cambrian, ie without specific route knowledge, would be nice to know.

 

[43066]

I have an automotive driving license and within weight and axle limits I can drive any vehicle on any road worldwide.

I have a commercial pilots license and I can fly any qualified type in any (unrestricted) airspace worldwide and with some very limited exceptions I am qualified to use any airfield on the planet.

Also road vehicles are driven on line of sight, whereas trains aren’t. I think @Watershed has described the differences with aviation very well.

I guess possibly the biggest factor is simply that "it's always been that way" and it's (perceived as) very difficult to change established working practices in the industry.

I agree with most of your post, but not sure about this bit! Route learning remains entirely necessary to drive trains safely, in the absence of the technology you describe. If anything people are pressured to pass out in less time than ideal. As you say it’s difficult to see that changing for a long time - perhaps when ETCS becomes more widespread it will gradually reduce, but we are talking decades.

 

[Greybeard33]

In countries that have speed signalling, are the route learning requirements less demanding than for British route signalling?

 

[XAM2175]

In countries that have speed signalling, are the route learning requirements less demanding than for British route signalling?

It reduces the need to learn and recall all the possible routes and their relevant speed restrictions, yes, but the driver obviously still needs to learn and be able to recall unsignalled speed changes, station locations and braking points, etc etc, unless these are also indicated in some way.

Germany is an example myself and others have used here before where drivers do not need to learn routes to the same extent as those in Britain, and where drivers who have learned a route remain qualified on them for much longer, and this is achieved through the use of speed signalling, a much greater amount of lineside signage (including warning boards for all speed reductions, and 'mileposts' every 200 metres), and an in-cab display system (EBuLa) that combines an electronic timetable for the train and a basic line guide that is synchronised to the movement of the train. They also use a sophisticated train protection system called PZB that is more thorough than TPWS.

Of course, this is something that has evolved in Germany over a very long time, rather than just occurring overnight.

ETCS does deliver several features like this in that the driver is constantly informed of their maximum permitted speed and the distance remaining until they must reduce speed or stop the train entirely for a figurative signal at danger, which will in effect convert ETCS-signalled lines to speed signalling, and provide the opportunity to further reduce the amount of exacting detail they need to learn - but that's many years in the future, and still relies on the culture of British railways being willing to explore those possibilities.

 

[King Lazy]

One very small point I would make regarding driving trains and route knowledge as opposed to the other disciplines that I feel is often overlooked is the following.

On a train you often have to remember something that happened some time ago.

For a simple example: You pass a yellow signal, you stop in a platform. The next signal is out of view. For some reason a delay occurs. Not enough to speak to the signaller but a couple of minutes, maybe loading disabled persons. When you get the ready to start it is your route knowledge that keeps you from approaching that signal too fast as that particular risk at that location will have been covered in the training and the signal, being higher risk should have been committed to your memory including other items such as the gradient, general adhesion etc approaching it.

On a plane generally ATC have a live view and instructions are followed immediately or under visual flight rules it’s becomes very similar to road driving maintaining a lookout and giving way accordingly.

In a car you respond to what you can see and what is a few minutes behind you generally doesn’t matter.

I know an ex-Police driving instructor who would often randomly ask “What was the last road sign”. Outside of complex areas where concentration is high he says most people couldn’t answer.

I am generalising because I know people will be able to think of similar road and air issues.

 

[Senex]

For a simple example: You pass a yellow signal, you stop in a platform. The next signal is out of view. For some reason a delay occurs. Not enough to speak to the signaller but a couple of minutes, maybe loading disabled persons. When you get the ready to start it is your route knowledge that keeps you from approaching that signal too fast as that particular risk at that location will have been covered in the training and the signal, being higher risk should have been committed to your memory including other items such as the gradient, general adhesion etc approaching it.

At the end of the 1980s there was a sign at the south end of the up fast platform at Kettering that read: “NOTICE TO DRIVERS OF STOPPING TRAINS, WHAT WAS THE ASPECT OF THE PREVIOUS SIGNAL LR68?” In this three-aspect area the station was then about half way between two signals and the next to the south, protecting Kettering Jn, well and truly out of sight round a curve. In due course this sign was replaced by a repeated located more than 800 yards ahead of the signal to which it referred. Presumably there had been some problems of the sort you mention.

 

[skyhigh]

What is it that is different about operating a train that requires “route learning” when road vehicles and aircraft do not?

As a practical example, there is a location on one of my routes where the linespeed is 90mph. The station is 'hidden' around a corner and by the time it becomes visible you'd have no chance of stopping from linespeed. However, if you crawled round from the previous stop to be able to stop on sight of the station, you'd be considerably late.

 

[Taunton]

I'm afraid that those writing that a train is somewhat more complex and challenging than flying an aircraft, and holding a pilot's licence, have ... well, never flown an aircraft.

 

[scotraildriver]

You could also never have a fixed braking point anyway. Your braking point on a wet autumn day is totally different to a dry sunny day. My braking points for an HST are completely different to a 158. There are so many variables setting everything in stone is virtually impossible.

 

[the sniper]

I'm afraid that those writing that a train is somewhat more complex and challenging than flying an aircraft, and holding a pilot's licence, have ... well, never flown an aircraft.

Who said that?

 

[LCC106]

In terms of braking points, some instructors will tell you to use a low adhesion braking point all year round. Often if you lose time during certain sections of the journey there is sufficient make up time to still arrive at the destination on point. It often makes for a smoother ride for customers and fellow crew alike.

 

[GB]

I'm afraid that those writing that a train is somewhat more complex and challenging than flying an aircraft, and holding a pilot's licence, have ... well, never flown an aircraft.

No one has said that.

 

[Taunton]

No one has said that.

eg "... in the air, everything is highly standardised ..."

 

[Ashley Hill]

together with wheel rotations, trains know exactly where they are at any given point.

Wheelslip/slide and tyre thickness could lead to misinterpretation by such a system.

 

[Watershed]

eg "... in the air, everything is highly standardised ..."

If you're alluding to my post, I didn't make any judgment as to the relative difficulty - trains and aircraft are simply very different.

As for your final comment, you couldn't be further off the mark!

Wheelslip/slide and tyre thickness could lead to misinterpretation by such a system.

Neither are likely to be fatal flaws in such a system. For instance, you could keep one axle unbraked in order to get an accurate reading. And if you couldn't compensate for tyre thickness by e.g. measuring the distance from the rail to the solebar by laser, at the end of the day it's not going to put you more than a few percent off. With beacons every mile or two, you're still going to have a pretty accurate picture of where you are.

 

[bramling]

Also road vehicles are driven on line of sight, whereas trains aren’t. I think @Watershed has described the differences with aviation very well.



I agree with most of your post, but not sure about this bit! Route learning remains entirely necessary to drive trains safely, in the absence of the technology you describe. If anything people are pressured to pass out in less time than ideal. As you say it’s difficult to see that changing for a long time - perhaps when ETCS becomes more widespread it will gradually reduce, but we are talking decades.

Continuous ATP, particularly where braking curves are supervised, definitely reduces the amount of learning for day-to-day driving. If one takes something like the Northern Line to Battersea, someone could quite happily jump in their chair and take a train there and back without needing any prompting at all (though a few lightweights did get caught out with the Nine Elms platform entry which in one direction is rather quick!).

However I’m still not entirely convinced it would make that much of a saving, except on the most metro-oriented routes routes. It’s still possible to mess up when doing shunts or unusual moves, and we saw on the Cambrian how incorrect data can get into the system. Then on top of that there’s still the knowledge a driver needs to know during incidents.

I’d say the only thing a continuous ATP system negates is the driver needing to know the speed profile in great depth, including things like rarely used turnouts. A small saving in learning, but I’m not sure it’s as much as some wishfully think it might be.

 

[XAM2175]

Wheelslip/slide and tyre thickness could lead to misinterpretation by such a system.

Neither are likely to be fatal flaws in such a system. For instance, you could keep one axle unbraked in order to get an accurate reading. And if you couldn't compensate for tyre thickness by e.g. measuring the distance from the rail to the solebar by laser, at the end of the day it's not going to put you more than a few percent off. With beacons every mile or two, you're still going to have a pretty accurate picture of where you are.

The ETCS odometry subsystem deals with this by measuring the rotation of multiple axles, much as is already done for WSP systems, and combining it with data from other systems such as Doppler radar transceivers (that measure the rate at which the train is moving over the railhead or trackbed), accelerometers (that measure the inertial movement of the train), and even GNSS receivers (GPS, Galileo, etc). The ERTMS Performance Standards require that odometry be correct to within plus or minus the sum of 5 meters plus 5% of the distance from the last positioning balise, but I believe that most systems routinely achieve far greater precision.

It's even possible to have the system automatically recalibrate the tyre thickness by arranging a series of balises at fixed distances on a stretch of straight track, should you doubt that your technicians are keeping on top of it.

 

[Taunton]

trains and aircraft are simply very different.

Indeed. Aircraft have to be controlled in three dimensions, one of which is very susceptible to gravity. Cars have to be managed in two dimensions. Trains just have to be managed in one dimension. It's not going anywhere different to where the points are set.

I wonder whether drivers, driving home from the depot at end of shift, if asked by their colleague to give them a lift home first, reply "No, can't do that, I don't know the way from your house on to mine".

 

[Need2]

Indeed. Aircraft have to be controlled in three dimensions, one of which is very susceptible to gravity. Cars have to be managed in two dimensions. Trains just have to be managed in one dimension. It's not going anywhere different to where the points are set.

I wonder whether drivers, driving home from the depot at end of shift, if asked by their colleague to give them a lift home first, reply "No, can't do that, I don't know the way from your house on to mine".

Very tedious, as usual.

 

[GB]

Indeed. Aircraft have to be controlled in three dimensions, one of which is very susceptible to gravity. Cars have to be managed in two dimensions. Trains just have to be managed in one dimension. It's not going anywhere different to where the points are set.

I wonder whether drivers, driving home from the depot at end of shift, if asked by their colleague to give them a lift home first, reply "No, can't do that, I don't know the way from your house on to mine".

So your of the opinion drivers don’t need route knowledge then? Very odd.

 

[Need2]

So your of the opinion drivers don’t need route knowledge then? Very odd.

Ignore him, he is very anti driver.

 

[Agent_Squash]

Ignore him, he is very anti driver.

Don’t see why someone’s opinion should be ignored because you disagree.

I think route learning should stay - but the requirement to know every inch should really be supplemented by new technology to make driving more accessible. Signage could be used, like on the road network, to support stopping distances.

 

[bramling]

Don’t see why someone’s opinion should be ignored because you disagree.

I think route learning should stay - but the requirement to know every inch should really be supplemented by new technology to make driving more accessible. Signage could be used, like on the road network, to support stopping distances.

Relying on signage isn’t really a great way to go. It can go missing or become obscured from view for any number of reasons.

One needs to be very wary of replacing one problem with a set of new ones. Route knowledge has a history of generally working pretty well, so there needs to be a very good reason to change things.

 

[Taunton]

I was actually hoping for some informed discussion ... oh well.

 

[Mag_seven]

Route knowledge has a history of generally working pretty well, so there needs to be a very good reason to change things.

I think with that point we will bring this discussion to a close.

thanks all