Boris demands TfL agree to Driverless Trains

[aslefshrugged]

Jul 27, 2020 8:56:20 GMT spsmiler said:

Jul 26, 2020 13:29:03 GMT aslefshrugged said:

If Boris wants to expand "driverless operation from the DLR to other lines which are already automatic" then he's going to need a great big umbrella over the whole of the Central Line as I lost count of the number of times I had to switch to Coded Manual last night because ATO couldn't cope with the rain!

'only' the Central line (for the umbrella)?

I thought that all lines with automation have 'issues' in wet weather, although maybe on some of them (Northern?) this is resolved by using less powerful braking over a longer distance - with the consequential increased journey time resulting in slightly less service frequency.

I've only worked the Central (and W&C) so don't know if rain is a problem on the other ATO lines (I doubt if its much of a problem on the Victoria) or if its a problem on the DLR.

[North End]

Jul 27, 2020 14:11:15 GMT aslefshrugged said:

Jul 27, 2020 8:56:20 GMT spsmiler said:

'only' the Central line (for the umbrella)?

I thought that all lines with automation have 'issues' in wet weather, although maybe on some of them (Northern?) this is resolved by using less powerful braking over a longer distance - with the consequential increased journey time resulting in slightly less service frequency.

I've only worked the Central (and W&C) so don't know if rain is a problem on the other ATO lines (I doubt if its much of a problem on the Victoria) or if its a problem on the DLR.

Rain on its own isn’t really a problem on the Northern, the system is set to drive slower than even the slowest manual drivers used to, so it just loses time instead which has to be made up by drivers doing quick turnarounds.

Leaf fall is more of an issue, especially a wet and windy day, and in sleet or snow forget it!

One thing though, manual driving doesn’t necessarily help, as because there isn’t always much warning when the target speed drops this means the driver can’t always anticipate in the way a conventional driver would, even the Central setup gives the driver plenty of clues as to where the train ahead is, what’s likely to happen next et cetera. Seltrac doesn’t really, and even a fully on-the-ball driver can miss something and find they have then had a speed violation.

I presume same applies to the Jubilee.

[t697]

Yes, whether there's someone at the front of the train or not, we could do with ATC that isn't fazed by the excellent wheelslip control available in the most modern rolling stock. Much of the benefit of that is lost with the system on NL, JL and coming in on S stock.

[punkman]

To be fair the Jubilee is pretty good in the wet, or at least I've not heard of too many problems. I remember my first experience of ATO was back when involved in testing when it was absolutely chucking it down with rain. The train was probably still doing 35-40 mph as it entered the platform and I was convinced an overshoot was inevitable, but with much noise from the WSP (Wheel slip protection) it stopped bang on the mark.

[t697]

Jul 28, 2020 9:40:10 GMT punkman said:

To be fair the Jubilee is pretty good in the wet, or at least I've not heard of too many problems. I remember my first experience of ATO was back when involved in testing when it was absolutely chucking it down with rain. The train was probably still doing 35-40 mph as it entered the platform and I was convinced an overshoot was inevitable, but with much noise from the WSP (Wheel slip protection) it stopped bang on the mark.

Impressive. Possibly JL has it slightly the easiest. Not too much of the railway gets severe rail contamination (usually!), as with the NL trains the braking is biased to the motor cars to minimise friction braking on the trailer cars (which have the ATC tachometers) and as with S stock the ATC is set up so the rear ATC controller is normally in control so the active tachos are on the second last car where adhesion will have been improved a little by the passage of the leading cars (usually!). So the critical ATC axles are less likely to experience slip and get the ATC confused.

User experience welcome to knock me down .

[100andthirty]

Jul 28, 2020 15:21:37 GMT t697 said:

Jul 28, 2020 9:40:10 GMT punkman said:

To be fair the Jubilee is pretty good in the wet, or at least I've not heard of too many problems. I remember my first experience of ATO was back when involved in testing when it was absolutely chucking it down with rain. The train was probably still doing 35-40 mph as it entered the platform and I was convinced an overshoot was inevitable, but with much noise from the WSP (Wheel slip protection) it stopped bang on the mark.

Impressive. Possibly JL has it slightly the easiest. Not too much of the railway gets severe rail contamination (usually!), as with the NL trains the braking is biased to the motor cars to minimise friction braking on the trailer cars (which have the ATC tachometers) and as with S stock the ATC is set up so the rear ATC controller is normally in control so the active tachos are on the second last car where adhesion will have been improved a little by the passage of the leading cars (usually!). So the critical ATC axles are less likely to experience slip and get the ATC confused.

User experience welcome to knock me down .

Fully wet braking isn't usually a problem. The lowest adhesion occurs on damp leaves and where the rail is initially wetted after a dry spell. It has been shown both scientifically and practically that a small amount of water together with normal contaminants such as iron oxide tiny amounts of lubricant (eg carry over from rail lubricators) can deliver adhesion as poor as with damp leaves. Generally though, this phase doesn't last long. Dew is quickly dispersed and more water (eg as a rain shower really gets going) rapidly improves adhesion, albeit not to the dry values.

Trials have been carried out, sponsored by RSSB on a variety of methods to clean the rail head to improve adhesion and one of these was to add more water. The principle was "if very small amounts of water (ie microlitres per metre) cause adhesion problems, then adding more water (ie a few millilitres per metre) adhesion is improved. This is referred to in August's Modern Railways.

[whistlekiller2000]

I've been through this thread and removed every instance of silly political nickname calling and political opinion I can find, with a note to say so in the edit box. Despite staff warnings about it on several occasions before it's STILL happening. We don't want to see any more of it. Personally, I don't care what side of the political coin a contributor is and I don't want to know either as it's outside the remit of this forum and against the rules. If we see it again there'll be consequences. 

[100andthirty]

Jul 28, 2020 16:32:15 GMT 100andthirty said:

Jul 28, 2020 15:21:37 GMT t697 said:

Impressive. Possibly JL has it slightly the easiest. Not too much of the railway gets severe rail contamination (usually!), as with the NL trains the braking is biased to the motor cars to minimise friction braking on the trailer cars (which have the ATC tachometers) and as with S stock the ATC is set up so the rear ATC controller is normally in control so the active tachos are on the second last car where adhesion will have been improved a little by the passage of the leading cars (usually!). So the critical ATC axles are less likely to experience slip and get the ATC confused.

User experience welcome to knock me down .

Fully wet braking isn't usually a problem. The lowest adhesion occurs on damp leaves and where the rail is initially wetted after a dry spell. It has been shown both scientifically and practically that a small amount of water together with normal contaminants such as iron oxide and tiny amounts of lubricant (eg carry over from rail lubricators) can deliver adhesion as poor as with damp leaves. Generally though, this phase doesn't last long. Dew is quickly dispersed and more water (eg as a rain shower really gets going) rapidly improves adhesion, albeit not to the dry values.

Trials have been carried out, sponsored by RSSB on a variety of methods to clean the rail head to improve adhesion and one of these was to add more water. The principle was "if very small amounts of water (ie microlitres per metre) cause adhesion problems, then adding more water (ie a few millilitres per metre) adhesion is improved. This is referred to in August's Modern Railways.

[spsmiler]

Jul 28, 2020 16:32:15 GMT 100andthirty said:

Jul 28, 2020 15:21:37 GMT t697 said:

Impressive. Possibly JL has it slightly the easiest. Not too much of the railway gets severe rail contamination (usually!), as with the NL trains the braking is biased to the motor cars to minimise friction braking on the trailer cars (which have the ATC tachometers) and as with S stock the ATC is set up so the rear ATC controller is normally in control so the active tachos are on the second last car where adhesion will have been improved a little by the passage of the leading cars (usually!). So the critical ATC axles are less likely to experience slip and get the ATC confused.

User experience welcome to knock me down .

Fully wet braking isn't usually a problem. The lowest adhesion occurs on damp leaves and where the rail is initially wetted after a dry spell. It has been shown both scientifically and practically that a small amount of water together with normal contaminants such as iron oxide tiny amounts of lubricant (eg carry over from rail lubricators) can deliver adhesion as poor as with damp leaves. Generally though, this phase doesn't last long. Dew is quickly dispersed and more water (eg as a rain shower really gets going) rapidly improves adhesion, albeit not to the dry values.

Trials have been carried out, sponsored by RSSB on a variety of methods to clean the rail head to improve adhesion and one of these was to add more water. The principle was "if very small amounts of water (ie microlitres per metre) cause adhesion problems, then adding more water (ie a few millilitres per metre) adhesion is improved. This is referred to in August's Modern Railways.

in many ways this sounds counter-intuitive, but hey, if real-world experience proves that it works then its good news

but how water is needed to be carried on the trains to extend this fleetwide? (assuming that such is possible / practical, even in icy weather)

[aslefshrugged]

Someone has leaked TfL's driverless trains business case from August to ASLEF. Estimated cost £7bn which I will admit isn't as much as I thought it would but even so it doesn't provide "value for money".

www.itv.com/news/london/2020-10-26/driverless-tube-trains-would-cost-7-billion-claims-leaked-document


PDF of business case here on Twitter, it would cost £41.643m to convert the Waterloo & City (I'm starting  a whip round at Leytonstone TCD)

[Chris M]

Tom Edwards has published the document (or at least two slides from it) on Twitter

Two salient points for me are:

• The Net Financial Effect (NFE) of an LU network-wide conversion is negative, showing a net cost of c.£5.6bn after accounting for revenue and OPEX savings.
• All LU lines have a negative NFE; therefore, none of the GOA4 conversions would cover their costs over the stated asset life.

(emphasis in original, OPEX = operational expenditure, I don't know the meaning of the GOA4 acronym)
(click here if tweet fails to load)

[quex]

GOA4 is Grade of Automation level 4 - essentially completely unattended train operation. The different grades of automation are defined by the International Association of Public Transport as a standard system so that everyone knows what each other means - "automatic train operation" and "automatic train control" can mean quite different things internationally, for example.

[aslefshrugged]

GoA 0 On-sight; Similar to a tram running in street traffic

GoA 1 Manual; A train driver controls starting and stopping, operation of doors and handling of emergencies or sudden diversions.

GoA 2 Semi-automatic (STO); Starting and stopping are automated, but a driver operates the doors, drives the train if needed and handles emergencies. As used on Victoria, Central, Northern, Jubilee and parts of Sub Surface Lines

GoA 3 Driverless (DTO); Starting and stopping are automated, but a train attendant operates the doors and drives the train in case of emergencies. As used on DLR

GoA 4 Unattended train operation (UTO); Starting and stopping, operation of doors and handling of emergencies are all fully automated without any on-train staff




I've got the full 26-page document available on a PDF through the ASLEF Central East WhatsApp, I'll have a read at some point.

Page 14, TfL have decided that GoA4 wouldn't work so it would have GoA3 with Train Attendants like the DLR.  So like the DLR those TAs would be able to join a union (probably RMT) and go on strike if they chose to.

[Chris M]

The page of the document that explains the various grades of automation levels has been tweeted now:

(backup link)

[North End]

Oct 26, 2020 12:38:27 GMT aslefshrugged said:

GoA 0 On-sight; Similar to a tram running in street traffic

GoA 1 Manual; A train driver controls starting and stopping, operation of doors and handling of emergencies or sudden diversions.

GoA 2 Semi-automatic (STO); Starting and stopping are automated, but a driver operates the doors, drives the train if needed and handles emergencies. As used on Victoria, Central, Northern, Jubilee and parts of Sub Surface Lines

GoA 3 Driverless (DTO); Starting and stopping are automated, but a train attendant operates the doors and drives the train in case of emergencies. As used on DLR

GoA 4 Unattended train operation (UTO); Starting and stopping, operation of doors and handling of emergencies are all fully automated without any on-train staff




I've got the full 26-page document available on a PDF through the ASLEF Central East WhatsApp, I'll have a read at some point.

Page 14, TfL have decided that GoA4 wouldn't work so it would have GoA3 with Train Attendants like the DLR.  So like the DLR those TAs would be able to join a union (probably RMT) and go on strike if they chose to.

And GOA3 doesn’t allow the abolition of duty schedules, which is one of the potential prizes which can be won from trying to increase automation. It may allow some simplifying, but ultimately you still have to have a name in a numbered box, and a body on every train - just like today.

[jimbo]

Oct 26, 2020 19:59:27 GMT North End said:

Oct 26, 2020 12:38:27 GMT aslefshrugged said:

GoA 0 On-sight; Similar to a tram running in street traffic

GoA 1 Manual; A train driver controls starting and stopping, operation of doors and handling of emergencies or sudden diversions.

GoA 2 Semi-automatic (STO); Starting and stopping are automated, but a driver operates the doors, drives the train if needed and handles emergencies. As used on Victoria, Central, Northern, Jubilee and parts of Sub Surface Lines

GoA 3 Driverless (DTO); Starting and stopping are automated, but a train attendant operates the doors and drives the train in case of emergencies. As used on DLR

GoA 4 Unattended train operation (UTO); Starting and stopping, operation of doors and handling of emergencies are all fully automated without any on-train staff




I've got the full 26-page document available on a PDF through the ASLEF Central East WhatsApp, I'll have a read at some point.

Page 14, TfL have decided that GoA4 wouldn't work so it would have GoA3 with Train Attendants like the DLR.  So like the DLR those TAs would be able to join a union (probably RMT) and go on strike if they chose to.

And GOA3 doesn’t allow the abolition of duty schedules, which is one of the potential prizes which can be won from trying to increase automation. It may allow some simplifying, but ultimately you still have to have a name in a numbered box, and a body on every train - just like today.

But see benefits on page 24 include: Faster schedule recovery, independent of the recovery of train operators to rosters or ‘duty sheets’. This suggests that Train Attendants are envisaged to not have a duty allocated to particular trains, but supervisor still needs to allocate staff to individual trains, and monitor their meal break requirement and end of duty relief. There will still be training, leave cover, etc to manage. The real advantage of full automation is operation like station lifts, rather than the staffed operation they once had. Perhaps that is not possible on LU trains.

[aslefshrugged]

Oct 27, 2020 3:42:56 GMT jimbo said:

Oct 26, 2020 19:59:27 GMT North End said:

And GOA3 doesn’t allow the abolition of duty schedules, which is one of the potential prizes which can be won from trying to increase automation. It may allow some simplifying, but ultimately you still have to have a name in a numbered box, and a body on every train - just like today.

But see benefits on page 24 include: Faster schedule recovery, independent of the recovery of train operators to rosters or ‘duty sheets’. This suggests that Train Attendants are envisaged to not have a duty allocated to particular trains, but supervisor still needs to allocate staff to individual trains, and monitor their meal break requirement and end of duty relief. There will still be training, leave cover, etc to manage. The real advantage of full automation is operation like station lifts, rather than the staffed operation they once had. Perhaps that is not possible on LU trains.

We had that when running the "Covid" service in the weeks after lockdown, booking on at our rostered time without knowing beforehand which train we'd be picking up but that was with a greatly reduced service, trying to run a full service on that basis would be an unholy nightmare for the manager on the desk and probably require far more TAs than the current number of TOps which would negate any savings on staff costs.

N.B. last year DLR PSAs were on around £45k, only £10k less than TOps.

---

Page 14 it says the Piccadilly Line would "retain a pool of Train Operators for continued GOA2 operation between Uxbridge and South Harrow".  Page 16 says that on the SSR "Mode changes will also be necessary on branches where interoperation with National Rail and Overground services occurs and where the Thales automatic signalling system is not yet installed" and on the Bakerloo "GoA2 operation would continue North of Queen’s Park where Network Rail track and platforms are shared with TfL London Overground trains".

So there would still be drivers on driverless trains

[alpinejohn]

5 pages on and it seems people here are still cautiously dancing around the underlying message which The Government are trying to seed. Namely they would love to have a totally automated railway where trains can be left to operate without any onboard staff, with the true message being they want to eliminate the prospect of strikes by train drivers.

Given how brilliantly well D/Transport's franchising solution is working on the mainline railway system, perhaps it really is time for the Mayor to call there bluff and simply say - fine - over to you it is all your problem.

If the Government then want to buy fully automatic trains - it won't be cheap but the cost will fall to D/Transport and in future come from central tax revenue. If they really want to upgrade all lines so trains do not need any onboard staff, then that will cost a bit more, and again it will come out of D/Transports budget. Finally any decision to fund this by increasing fares and congestion charge rates/extent will be clearly down to the Government.

At the moment this looks more like political kite flying intended to go down well with their political supporters, but will fail miserably once the Treasury start adding up all the costs.

Meantime it stirs up totally unnecessary concern over future job security among so many hard working railway staff, even though there is no realistic prospect of a totally automated (staff free) tube service within my life-time or indeed probably within the lifetime of every existing member of parliament.

Yes if UK plc was massively cash rich, then perhaps it might be a long term objective, but against the backdrop of a country running up massive debts to cover the economic impact of COVID, I cannot see the Treasury being keen to spend a penny more than really needs to be spent on Transport projects for many many years to come.

I would go further and say even if The Government insist D/Transport push ahead to fully automated operation of the easiest line - the Waterloo and City, then it would perhaps make them realise this dream is not easy, and is definitely not going to payback the upfront costs, within any realistic timeframe.

Can we please ignore this madness.

[zbang]

A minor point-

Oct 26, 2020 12:38:27 GMT aslefshrugged said:

GoA 3 Driverless (DTO); Starting and stopping are automated, but a train attendant operates the doors and drives the train in case of emergencies. As used on DLR<abbr>.</abbr>

My recollection is that DLR does/did have TAs during the busy times but not early/later in the day. Is this still the case?

[Chris M]

Every DLR train has a staff member on board at all times. In normal times they operate the doors and carry out customer service duties such as checking tickets. At busy times, when there are staff on or about the track, in high winds, first trains of the day and at any other time so instructed (and all the time at the moment) they sit at the front observing the road ahead and operate the doors from that position - effectively GOA2. When required they can also drive the train manually from the front seat, although this is very rare in passenger service it is the only mode of operation in depots AIUI.

[aslefshrugged]

That never was the case as far as I'm aware, DLR trains are GoA3 and cannot operate without a member of staff on board.

(as ever if anyone has more information then I am ready to stand corrected).

[jimbo]

Do the DLR train attendants always remain on the leading car? They cannot walk between cars on a moving train.

[johnlinford]

No - and they can use the doors between cars on a stationary train - I have observed this when, for example a major signal failure resulted in a train I was on reversing back to Poplar. They also walk through like this at stations when they have detrained passengers before moving in to a depot when this requires changing ends (eg at All Saints to reverse in to the sidings on the line up to Stratford).

[Chris W]

Oct 27, 2020 19:04:13 GMT jimbo said:

Do the DLR train attendants always remain on the leading car? They cannot walk between cars on a moving train.

In my experience using the DLR, admittedly not extensive, I've seen attendants/captains in either of the units...

I'm sure that I've seen them move between units... but I'm just thinking/pondering how they could do this is they were controlling the doors at the time...

[rincew1nd]

The train can berth, then the PSA walk along the platform between units before closing the doors from another position.

[Chris M]

There is a position on the the control unit adjacent to the doors marked "I" (which I've always assumed stands for "inhibit") that the PSAs select before moving between units.

[johnlinford]

I always assumed it was "Interrupt", which is based on my early computational electronics training, but "Inhibit" makes much more sense!

[jimbo]

Oct 27, 2020 8:55:36 GMT aslefshrugged said:

Oct 27, 2020 3:42:56 GMT jimbo said:

But see benefits on page 24 include: Faster schedule recovery, independent of the recovery of train operators to rosters or ‘duty sheets’. This suggests that Train Attendants are envisaged to not have a duty allocated to particular trains, but supervisor still needs to allocate staff to individual trains, and monitor their meal break requirement and end of duty relief. There will still be training, leave cover, etc to manage. The real advantage of full automation is operation like station lifts, rather than the staffed operation they once had. Perhaps that is not possible on LU trains.

... N.B. last year DLR PSAs were on around £45k, only £10k less than TOps. ...

I don't think the DLR train attendant is relevant: they operate the doors and drive when required, like former tube train guards. The proposed new auto-train attendant will do none of that. The doors are said to be automatic, and train driving during failures will be remote from line control centre or from a failures supervisor making their way to the train. The future train attendant will be more like a platform attendant, and paid accordingly I should think. It is a fully auto train with a staff member to reassure passengers in the event of failure. Passengers will be assured there is a member of staff on a train somewhere, but unless they are in the same car they are unlikely to see them. So to say they will deal with passenger enquiries and tickets is just to pretend they have a purpose other than reassurance, since most travellers will not come across them. What assistance can they provide to 800 passengers on the train, as mentioned elsewhere in the presentation. It is intended that the failure rate will be far lower than on current trains. Who will want these pointless jobs? Riding all day without a seat or cab.

[zbang]

Oct 27, 2020 18:49:56 GMT Chris M said:

Every DLR train has a staff member on board at all times.

Was that always the case? ISTR from 2014(?) that the later-in-the-night trains didn't- leaving the ExCeL center/Prince Regent station after 2200 our group were often the only ones on the train heading west. (My memory is a bit fuzzy on this.)

[norbitonflyer]

Oct 28, 2020 5:09:29 GMT zbang said:

Oct 27, 2020 18:49:56 GMT Chris M said:

Every DLR train has a staff member on board at all times.

Was that always the case? ISTR from 2014(?) that the later-in-the-night trains didn't- leaving the ExCeL center/Prince Regent station after 2200 our group were often the only ones on the train heading west. (My memory is a bit fuzzy on this.)

Who opened the doors?