TBTC Testing

[Deleted]

Transmission-based train control.

[jardine01]

Does anybody know if the Jubilee line will get high prefomance upgrade once the line is fully ATO. like the central line did With higher accelration rates ect. Thanks

[21146]

Feb 18, 2011 16:43:29 GMT @jardine01 said:

Does anybody know if the Jubilee line will get high prefomance upgrade once the line is fully ATO. like the central line did With higher accelration rates ect. Thanks

You mean until the first motor falls off after which it'll be turned down again?

Will the Central Line ever go back to the original line speeds that the CLP was supposed to deliver?

Returning to topic I hope there is eventually some gain for all the pain.

[Deleted]

When a new timetable comes into effect, yes; the SRS then will not be telling all the trains to slow down!

[tubeprune]

Feb 18, 2011 16:43:29 GMT @jardine01 said:

Does anybody know if the Jubilee line will get high prefomance upgrade once the line is fully ATO. like the central line did With higher accelration rates ect. Thanks

I am informed by a very reliable source that one of the reasons for the variable speeds on the Jubilee is that the control system tries to adjust train speeds to the present timetable. So, if a train is running late, or there is a gap ahead of it, it goes full speed. If it's running to time or getting to be early, the system slows it down a bit to match. Once the whole line is turned over to ATO and de-bugged, there will be a new WTT with better runing times.

[Deleted]

Yep, that is correct. The performance upgrade on the trains is in place already, and it is the TBTC control centre which is regulating the trains to lesser speeds if on time / timetable slack allows.
I've been travelling on the Jub every day since Xmas and depending how many delays depends how fast the trains go. If they are late, they fly like the Central, if they are on time its more like being on milk float.
Once the north end of the line is resignalled and 3rd platform at Stanmore brought into use, the timetable is to be stepped up towards 30tph from the current 24tph. To do this, more trains will be out and they will all need to move through much more quickly.

[tubeprune]

Feb 18, 2011 22:46:31 GMT @sw206 said:

....
Once the north end of the line is resignalled and 3rd platform at Stanmore brought into use, the timetable is to be stepped up towards 30tph from the current 24tph. To do this, more trains will be out and they will all need to move through much more quickly.

Actually, part of the increase in the number of trains per hour (tph) is, in the case of the Jubilee Line, because the new TBTC allows the removal of full speed fixed overlaps. Instead, overlaps match train speed (more or less). Hence, with overlaps matching speeds, trains can run closer together, particularly at lower speeds. Thus, you can get a more frequent service for the existing line speeds.

The other factor is ATO. If you take the judgment for driving and braking away from the driver, you can usually get a faster run in to stations and up to signals because the ATO element will take the braking rate up to the maximum possible. This alone will account for almost half the 6 more trains per hour on the new Jubilee service.

Experience will show what can be done, once the new system has settled down but the things which have to be skillfully managed are the station stop times, which must not go over 40s, crew changeovers and the tip out times for short workings like Willesden Green. This is where the 30 tph could be lost.

[DrOne]

Yeah I used Wembley Park - Canary wharf last week. Journey took the usual time and I did notice the intermittent start-stop throughout. There was, however one interlude where we had a blistering run from Baker St - Bond St. It was incredible.

[norbitonflyer]

Feb 19, 2011 7:02:25 GMT tubeprune said:

If you take the judgment for driving and braking away from the driver, you can usually get a faster run in to stations and up to signals because the ATO element will take the braking rate up to the maximum possible. ............

There are pros and cons to this style of driving though - high acceleration and maximum braking leads to more wear and tear, and power consumption, and leaves less margin for error if something goes wrong - poor adhesion, object or person seen on the track, that a human operator may be able to allow for.

Feb 19, 2011 7:02:25 GMT tubeprune said:

the things which have to be skillfully managed are the station stop times, which must not go over 40s, ............

This seems to be a misplaced priority: the object of the exercise is to move the maximum number of people (NOT trains) from A to B. Dwell times need to be long enough to allow anyone who can board to do so. I've been at stations where an attempt has been made to close the doors even whilst people were still getting off - let alone anyone having a chance to get on!
If dwell times are a Bad Thing, why stop at all? - you could get far more trains through the system that way!

[North End]

Feb 19, 2011 14:42:34 GMT norbitonflyer said:

Feb 19, 2011 7:02:25 GMT tubeprune said:

If you take the judgment for driving and braking away from the driver, you can usually get a faster run in to stations and up to signals because the ATO element will take the braking rate up to the maximum possible. ............

There are pros and cons to this style of driving though - high acceleration and maximum braking leads to more wear and tear, and power consumption, and leaves less margin for error if something goes wrong - poor adhesion, object or person seen on the track, that a human operator may be able to allow for.

Feb 19, 2011 7:02:25 GMT tubeprune said:

the things which have to be skillfully managed are the station stop times, which must not go over 40s, ............

This seems to be a misplaced priority: the object of the exercise is to move the maximum number of people (NOT trains) from A to B. Dwell times need to be long enough to allow anyone who can board to do so. I've been at stations where an attempt has been made to close the doors even whilst people were still getting off - let alone anyone having a chance to get on!
If dwell times are a Bad Thing, why stop at all? - you could get far more trains through the system that way!

Excessive dwell times are *seriously* bad. Not only do they contribute to longer journey times, but as Tubeprune rightly says, excess dwell time can badly kill the throughput of trains. An extra few seconds for a (relatively) small number of people to get on means the following train will approach on restrictive signals, which in turn slows following trains down, and can end up costing trains per hour, each of which can carry 1,000 people.

The most important thing is for people to get on & off quickly, which is achieved through letting people alight first, moving down inside the car, and using all the doors on the train to board/alight.

It's partly because the importance of dwell times has been forgotten since the 1980s, that we are now spending billions on systems like TBTC - which will be wasted if we don't maximise the benefits.

[rincew1nd]

I took a jolly on the Jubilee today from Finchley Road. Unfortunately the on-off nature of the speed control had the effect of gently rocking me to sleep and I ended up at Canning Town :S

[Deleted]

Feb 19, 2011 17:17:55 GMT North End said:

Excessive dwell times are *seriously* bad. Not only do they contribute to longer journey times, but as Tubeprune rightly says, excess dwell time can badly kill the throughput of trains. An extra few seconds for a (relatively) small number of people to get on means the following train will approach on restrictive signals, which in turn slows following trains down, and can end up costing trains per hour, each of which can carry 1,000 people.

The most important thing is for people to get on & off quickly, which is achieved through letting people alight first, moving down inside the car, and using all the doors on the train to board/alight.

It's partly because the importance of dwell times has been forgotten since the 1980s, that we are now spending billions on systems like TBTC - which will be wasted if we don't maximise the benefits.

I honestly think that's slightly overstating the case. At certain locations at certain times it's absolutely counter-productive to adhere too rigidly to dwell time targets. If I'm at Canary Wharf or London Bridge, say, in rush hour and there's a three-plus-minute gap to the train behind and the platform is chokka it's not sensible to close the doors too hastily. The timetable currently has enough slack that you shouldn't lose time on the road. The important thing is to be aware of your surroundings.

I see little evidence of excessively long dwell times. I do, or rather did, see a lot of evidence of slow driving, sluggish detrainments of reversing trains, bad signal positioning causing bottlenecks, wrong signals being given at Wembley Park, and disorganised depot manouvres. These were more responsible for delays than excessive dwell-times, in my experience.

I'd also point out that the way trains are currently running in TBTC (when it's running to time) doesn't exactly promote the cause of getting the maximum flow of people on and off the train. As sw206 memorably said, it can be like being on a milk float between stations - then the dwell-time, or rather the time before you get your target point, can be something like 10-15 seconds.

With a more ambitious timetable to come under TBTC dwell times will no doubt be more crucial. In theory trains should be that little bit less crowded which will help.

[tubeprune]

Feb 19, 2011 14:42:34 GMT norbitonflyer said:

Feb 19, 2011 7:02:25 GMT tubeprune said:

If you take the judgment for driving and braking away from the driver, you can usually get a faster run in to stations and up to signals because the ATO element will take the braking rate up to the maximum possible. ............

There are pros and cons to this style of driving though - high acceleration and maximum braking leads to more wear and tear, and power consumption, and leaves less margin for error if something goes wrong - poor adhesion, object or person seen on the track, that a human operator may be able to allow for.

You are right, but it depends on your priorities. LU's main one now is to manage the overcrowding. That means using the system to its maximum capability. If you buy an expensive piece of kit like a train (£6m each), you have to sweat the asset. It's built to take it. If you buy even more expensive signalling kit, you must sweat that too!

Variations for adhesion are taken into account. That's one of the reasons why the signalling is designed for trains to run at <95s headways but the timetable allows 120s headways. As for allowing for objects or persons on the track, any of the drivers on this forum will tell you that stopping before hitting anything which shouldn't be there is almost always impossible, regardless of the speed you are going.

Feb 19, 2011 7:02:25 GMT tubeprune said:

the things which have to be skillfully managed are the station stop times, which must not go over 40s, ............

This seems to be a misplaced priority: the object of the exercise is to move the maximum number of people (NOT trains) from A to B. Dwell times need to be long enough to allow anyone who can board to do so. I've been at stations where an attempt has been made to close the doors even whilst people were still getting off - let alone anyone having a chance to get on!
If dwell times are a Bad Thing, why stop at all? - you could get far more trains through the system that way!

Now, now! ;D ;D You have to remember that when a train is crowded, the rate of loading and unloading slows down and the dwell time lengthens. If the train stands at the platform too long, the following train will "catch up" and stop outside the station. When the train in front eventually leaves, the waiting train has to restart, accelerate run in and brake to a stop. This wastes more valuable seconds so that the third train will now get a slow down command etc. etc. A few seconds makes a huge difference.

If trains are allowed to wait too long at stations, the whole line suffers and passengers will have to wait longer and will travel more slowly. That's why I make such a fuss about dwell times.

Or, think of it this way: The Jubilee has 29 stations. If you waste 2 seconds at each station, you will have lost almost a minute by the end of the trip. If every train arrives a minute late at the terminus, you would quickly end up with a 20 tph service.

[tubeprune]

Feb 19, 2011 19:14:57 GMT rincew1nd said:

I took a jolly on the Jubilee today from Finchley Road. Unfortunately the on-off nature of the speed control had the effect of gently rocking me to sleep and I ended up at Canning Town :S

This is an interesting problem. Most ATO systems are designed to allow the train to run up to the permitted full speed between stations. When the train reaches the permitted speed, power is shut off and a touch of brake applied to bring it back under the permitted speed level. The train then tries to accelerate back up to the permitted speed and the whole process is repeated. This happens on the Central and Victoria Lines as well as the Jubilee.

The problem is that you get a rocking sensation several times between stations as the power/brake sequence repeats at the permitted speed. It's sometimes referred to a speed hunting. I think there is an argument for looking at ways of eliminating this.

[Ben]

Stupid question, is the touch of brake necessary? In the sense that instead of a maximum speed could not a maximum speed range be developed, say -1%+3% or something? The maximum speeds on conventional lines above 5mph increase so it seems in 5mph increments, can that resolution be increased with ATO/TBTC? Or is it already done like that? Sure I read somewhere that something similar to having a range is implimented anyway in terms of repowering speed being slightly below that of maximum speed.

[tubeprune]

Feb 20, 2011 11:15:21 GMT Ben said:

Stupid question, is the touch of brake necessary?

Not always, so you can get power-coast-power-coast.

[Deleted]

Finally the outright untruths about the management statistics for Jubilee Line "reliability" are starting to get noticed and revealed in the media.

www.bbc.co.uk/news/uk-england-london-12518589

Transport for London (TfL) has claimed 95% of scheduled Jubilee Line services ran in 2009-10.

The Politics Show London calculated that when weekend shutdowns were added to the figures, however, only 78% of trains ran in that period.

Professor Tony Travers of the London School of Economics said "the closures were removed from the statistics as if somehow they don't happen".

It's not just weekends, of course. Look at all the days after Christmas when there was no Jubilee service. Doubtless those weren't counted as well.

For the information of the Jubilee Line management, the "schedule" is actually posted up on every platform. This shows that trains start early in the morning 7 days a week, run the full length of the route from Stratford to Stanmore at all times, and finish late evening. This is the definition of "scheduled services" you will find from the dictionary.

Auntie Diana has noticed on several occasions three adjacent posters on the platform at Canning Town. The first shows this generous schedule, the second gives details of yet more continuing closures and shutdowns on weekends way into the future, and the final self-congratulatory one says how nearly 100% of trains are being run .......

[Deleted]

Feb 20, 2011 9:44:14 GMT tubeprune said:

This is an interesting problem. Most ATO systems are designed to allow the train to run up to the permitted full speed between stations. When the train reaches the permitted speed, power is shut off and a touch of brake applied to bring it back under the permitted speed level. The train then tries to accelerate back up to the permitted speed and the whole process is repeated. This happens on the Central and Victoria Lines as well as the Jubilee.

The problem is that you get a rocking sensation several times between stations as the power/brake sequence repeats at the permitted speed. It's sometimes referred to a speed hunting. I think there is an argument for looking at ways of eliminating this.

Presumably there's some characteristic of electric traction that doesn't allow power to be kept on at whatever lower level is necessary to maintain a cruising speed? When driving a road vehicle you don't power/brake; you hold the accelerator at a reduced opening to suit. Why are ATO Underground trains apparently all-or-nothing?

[Deleted]

Being that the Jubilee line's stock (1996) has effectively stepless control, you would have thought that TBTC could use this to set the speed, as there is going to be a position somewhere that holds the speed steady.

A skilled human driver could do this. I think TBTC can do things similar but it's nowhere near as crude as the go/anchors method used on the Central..

[Deleted]

Surely with the load measuring devices on the cars and the route knowledge of the TBTC system, the ATO can come up with a relatively accurate value of the power required.

[Alight]

This YouTube clip demonstrates the announcements being played twice:

The original plays 3 quarters of the way through the tunnel, whilst the new TBTC triggered one plays just as the train is pulling into the stations (similar to the 1995/2009/S stocks)

[North End]

Feb 21, 2011 13:31:29 GMT @cj said:

Being that the Jubilee line's stock (1996) has effectively stepless control, you would have thought that TBTC could use this to set the speed, as there is going to be a position somewhere that holds the speed steady.

A skilled human driver could do this. I think TBTC can do things similar but it's nowhere near as crude as the go/anchors method used on the Central..

The thing you have to remember is that even if the system were clever enough to calculate the exact percentage of traction required (taking into account gradient changes on the line ahead, loading of the train, adhesion, how good the motors are on this particular train, etc.), you also have to consider the effect of rail gaps - something which in various forms LU has many.

Although the Central Line ATO is pretty good, the smoothest ATO has to be the 1967 stock. Apart from slightly jolty stops, though even these are no worse than achieved by some Train Operators, the Victoria Line ATO system is superbly smooth, which is a real credit to those who designed the system back in the 1960s. By contrast I haven't found the 09 stock to be particularly smooth, and the Jubilee Line is horrible.

None of the ATO systems are as good as the best Train Operators though, and - I suspect - never will be.

[Deleted]

Nothing can beat the skilled input of a human, simple as!

I'd say part of the reason the 67 ATO is so smooth is the crudeness of the traction systems taking so long to respond. You do make a fair point about rail gaps, as they are of varying length for say pointwork, or some for isolation gaps. Nothing other than a busline would solve the jerkiness from that, and two problems there; a) not permitted b) unsafe.

[malcolmffc]

Feb 21, 2011 21:30:15 GMT @cj said:

Nothing can beat the skilled input of a human, simple as!

I'd say part of the reason the 67 ATO is so smooth is the crudeness of the traction systems taking so long to respond. You do make a fair point about rail gaps, as they are of varying length for say pointwork, or some for isolation gaps. Nothing other than a busline would solve the jerkiness from that, and two problems there; a) not permitted b) unsafe.

But the whole point of ATO is that it can consider the whole line at once, rather than from individual drivers' perspectives, and adjust speeds accordingly

[Deleted]

Give a driver a DTG display and a time either side of timetable and problem solved.

[Chris M]

Feb 22, 2011 8:02:03 GMT @cj said:

Give a driver a DTG display and a time either side of timetable and problem solved.

Would the relative positions of the trains in-front and behind also be useful? For example if the next train is 1 minute behind them the driver can be more rutheless about leaving people on the platform than if the next train is 10 minute behind.

[North End]

Feb 22, 2011 13:31:52 GMT Chris M said:

Feb 22, 2011 8:02:03 GMT @cj said:

Give a driver a DTG display and a time either side of timetable and problem solved.

Would the relative positions of the trains in-front and behind also be useful? For example if the next train is 1 minute behind them the driver can be more rutheless about leaving people on the platform than if the next train is 10 minute behind.

This is nothing new, on most lines the T/Op can judge how close the train in front is due to the numbers of people on the platform, and the dot-matrix gives an indication of the gap behind.

As I've said before, a decent Train Operator can do everything ATO can do.

[Deleted]

Feb 22, 2011 13:49:44 GMT North End said:

Feb 22, 2011 13:31:52 GMT Chris M said:

Would the relative positions of the trains in-front and behind also be useful? For example if the next train is 1 minute behind them the driver can be more rutheless about leaving people on the platform than if the next train is 10 minute behind.

This is nothing new, on most lines the T/Op can judge how close the train in front is due to the numbers of people on the platform, and the dot-matrix gives an indication of the gap behind.

As I've said before, a decent Train Operator can do everything ATO can do.

...and more. He can drink tea. ;D

Seriously tho', we shouldn't get too pro-T/Op; ATO has its advantages. Don't forget that.

Then again, a driver can die and the train won't know. Happened in the very nasty Waterfall crash in 199x, in Sydney.

Basically, the guy had a heart attack and slumped, dead, on the deadmans _{(oh the irony)}. The deadmans only recognises pressure, so the train charged along to its breakfast date with the Maker.

The ATO should be designed such that it can recognise its own death and slam on the brakes, to wait for the doctor.

[phillw48]

Feb 22, 2011 13:55:45 GMT @axewielder said:

The ATO should be designed such that it can recognise its own death and slam on the brakes, to wait for the doctor.

If its dead its too late for the doctor, the undertaker more like it!

[Deleted]

Andrew, you seem to forget the tripcock system. Or moving block ATP which is the safest as it doesn't rely on fixed length overlaps!