How likely is an ATO-failure and what is the procedure?

[RWMarcus]

Good evening!

I'm from Germany, 30 years old and a little bit interested in the London Underground.

I have a question about disruption procedures on the London Underground.
Some lines are fitted with ATO or similar systems which allow the driver to lean back and let the train drive from station to station

I think ATO is a great system to increase capacity of a line. My last visit was about 5 years ago and I still remember the overcrowding of some trains what I don't know from Germany. I mean I know overcrowded trains in Germany too but not with that high frequency of 2 minutes.

I have to say that I am a supporter of the traditional train driving. I think ATO destroys the job of the train operator. I think it will also be possible to run trains every 100 seconds without ATO but with correct signalling systems (low signal distance, short blocks). As long as you don't have moving block (ATO) there is no difference in capacity between trains driven manually or by ATO with fixed blocks.

What I would like to know is how often the ATO system has failures. And what would happen if the system doesn't work? Can operators switch to Full Speed Manual and drive the train? Or will the line be suspended until the problem has been solved? I think the most drivers on the Jubilee Line aren't trained to stop excatly at the stopping marker without crawling. And as I know from openBVE the "Acurate Stop"-lamp isn't gracious. But maybe there are still many operators who were also on duty when the Jubilee-Line-ATO-System (I know it isn't called ATO on the JL) wasn't installed on the Jubilee Line (Extension). As I read about the Jubilee Line history I was very suprised that ATO wasn't installed in the first operation years of the Jubilee Line Extension. That must have been a challenge to stop at the correct position.

I'm very interested in that topic and I hope that someone can answer my questions.

 

[AM9]

AFAIK, ATOs are just a part of the signalling system, so if there is a fault that prevents it to operate, it is likely to affect all of the trains movements. Reverting to a limited driver operated mode would seem like the next step. No doubt there will be a signalling expert along here soon.

 

[Domh245]

Like @AM9 says, the ATO is part of the signalling system, so if it goes down then the train isn't going anywhere at normal speed. They can be put into a "restricted manual" mode that allows the T/Op to drive the train at a significantly reduced speed (10mph at least on S stock) but generally that's only so that the train can be moved forward to a more convenient place to offload passengers

 

[Dstock7080]

(CBTC sub-surface; TBTC Northern/Jubilee) If the failure is on an individual train, all train movements must be stopped and the train driven in Restricted Manual (9mph) until it can communicate again. There are no signals, only occasional lineside markers.
If a station based SCS Station Control Subsystem or Control Room based VCC Vehicle Control Centre failed then all train movements in the area must stop and each train would need to be moved, one at a time only, in restricted manual mode until it can communicate again in another area

 

[Mojo]

It really depends what line, and what type of failure. For instance on the Central line it is known for trains to come to a stop in between stations (or more common halfway into a platform). In most cases then ATP and codes are still available and therefore the train can be driven at normal speed in Coded Manual to the next stopping mark where ATO should become available again.

 

[Jona26]

Sort of related...I've been on trains before (Central I think) where either brakes have automatically been applied or traction current lost (I'm not sure which) due to door interlock being momentarily lost as passengers have been leaning against the doors.

I know the cause as the train op announced it!

What happens in this situation? Is Auto mode regained when interlock is re-established or do the trains have to be manually driven to the next station?

 

[Mojo]

Sort of related...I've been on trains before (Central I think) where either brakes have automatically been applied or traction current lost (I'm not sure which) due to door interlock being momentarily lost as passengers have been leaning against the doors.

I know the cause as the train op announced it!

What happens in this situation? Is Auto mode regained when interlock is re-established or do the trains have to be manually driven to the next station?

The interlock being lost doesn’t cause emergency braking or loss of traction current, the train motors just cut out. The train can continue in ATO if that is the mode the train left the station in. If in manual then the handle must be put back into off and release (ie. not motoring, but doesn’t have to be braking either) before a motoring position is reapplied.

 

[Bald Rick]

As long as you don't have moving block (ATO) there is no difference in capacity between trains driven manually or by ATO with fixed blocks.

There absolutely is, as drivers tend to have slightly varied driving styles, and the capacity has to be set at the ‘average’ level. ATO can work at or close to the best.

 

[Jona26]

The interlock being lost doesn’t cause emergency braking or loss of traction current, the train motors just cut out. The train can continue in ATO if that is the mode the train left the station in. If in manual then the handle must be put back into off and release (ie. not motoring, but doesn’t have to be braking either) before a motoring position is reapplied.

Thank you - well explained to a non-expert.

 

[Nym]

Small point.
ATO isn't part of the signalling system, it's part of the rolling stock system. The ATP is part of the Signalling System. There is a very distinct gap between the two systems in terms of design, in that the ATP has to be much more highly safety rated than the ATO. Going back to the first such systems, the ATP on 1967TS was designed by signalling engineers and in a completely separate box to the "Auto Drive" system, which was designed by rolling stock engineers.
Hence,
If the ATO fails, and only the ATO fails, one will continue driving manually with ATP.

ATP Failure depends on the type of ATP being applied, if it is the more "Modern" CBTC or TBTC systems such as Thales S40 and S60 used on Northern, Jubilee, DLR, and Sub Surface Lines, the ATP On Board Unit (OBU) includes the reporting of train position for the purpouses of infrastructure interlocking and protection scanners, hence, the loss of communication to any one of these will cause all trains in the section or control area to be cautioned to move in Line of Sight until the lost train passes the next entry point and is again detected by the system.

However, on the Victoria and Central Lines (and the 1st proposed system for the Sub Surface Lines) the train detection is carried out by the infrastructure, not the train. Hence, a single unit failure in ATP will affect only that unit directly, others will be cautioned in terms of the line ahead not being cleared, but will not directly be required to be driven in a different mode, and can indeed continue to function in automatic driving modes.

There absolutely is, as drivers tend to have slightly varied driving styles, and the capacity has to be set at the ‘average’ level. ATO can work at or close to the best.

Particularly when (like the latest two stocks on LUL) manual driving mode does not have access to as much power or retardardation rate as the automatic mode.

 

[RWMarcus]

Thanks for the explanations. It seems to be a really difficult and complex system.

Has there ever been such a failure or disruption? Or is this a typical "Signal failure" which you can see on the tfl-website under status update? Will there be less signal failures when District, H & C, Met and Circle Lines become automated? Because on the deep level lines signal failures seem to be rare.

 

[Dstock7080]

Currently one of the problems on the sub-surface Lines is the migration from legacy tripcock mode to CBTC and vice-versa, a train can suddenly not communicate with the signalling system.
Once many of these boundary points are removed the system should be smoother. However West Harrow, Fulham Broadway, Barons Court are likely to be around for sometime.

For a legacy signal to indicate a clear route requires bulbs, tripcock proving etc to be functioning, something removed in CBTC

 

[Nym]

But one introduces many many more, way many more "black box" failures with CBTC than the highly serviceable equipment used to implement trainstops on the Sub Surface.
Bulb failures for example, you can replace the bulb, use dual filerment bulbs, lamp failure detection loops, dual holders, multi element LED lamps etc.
There isn't really any way to improve the "little red wire" issue, or the base station communication issues on S60. Or in spite of nearly 15 years of operation, units failing to communicate on entry and come out of "sleep mode" properly.

 

[mark-h]

If the ATO fails, and only the ATO fails, one will continue driving manually with ATP.

Not all ATO systems give the driver enough information to enable them to drive at full speed. The original Victoria Line ATO could not be driven above 25mph in coded manual mode as the driver is not informed of an upcoming change in speed from 50mph to 25mph until they enter the reduced speed section, triggering the emergency brake.

There is some discussion on manual driving on ATO lines in this thread.

 

[Nym]

Not all ATO systems give the driver enough information to enable them to drive at full speed. The original Victoria Line ATO could not be driven above 25mph in coded manual mode as the driver is not informed of an upcoming change in speed from 50mph to 25mph until they enter the reduced speed section, triggering the emergency brake.

There is some discussion on manual driving on ATO lines in this thread.

I'm aware of this, but it was already a monster post so didn't want to drown everything in information.

The majority of traction control systems on "modern" units doesn't allow access to the full power of traction or braking either when being driven in a coded manual, protected manual or manned manual (someone used that term once on a Victoria Line report and it stuck for far too long). Victoria Line allows 0.05m/s/s less in terms of acceleration and braking when in Protected Manual (enforced by the traction and brake control systems), I believe that Northern and Jubilee Lines are the same. This can be implimented on older units given how the traction and brake control PWM is implemented, in that the traction brake controller and auto drive boxes have separate access to the PWM train wires, changing the bounds of access is therefore not too challenging.

This wasn't available on 1967TS with the much, much simpler control system, but the <25mph in manual did function on the mechanical governors (provided they where working(!)), it may have (and I've not worked on 67TS for a long time) cut out access to weak fielding or locked the unit in Rate 1 without access to Rate 2.

Weak fielding effectively increases the top speed. Rate 2 effectively increases the rate of acceleration on a camshaft controlled DC motor, by allowing a larger current to flow (excluding Notch S1).

 

[bramling]

I'm aware of this, but it was already a monster post so didn't want to drown everything in information.

The majority of traction control systems on "modern" units doesn't allow access to the full power of traction or braking either when being driven in a coded manual, protected manual or manned manual (someone used that term once on a Victoria Line report and it stuck for far too long). Victoria Line allows 0.05m/s/s less in terms of acceleration and braking when in Protected Manual (enforced by the traction and brake control systems), I believe that Northern and Jubilee Lines are the same. This can be implimented on older units given how the traction and brake control PWM is implemented, in that the traction brake controller and auto drive boxes have separate access to the PWM train wires, changing the bounds of access is therefore not too challenging.

This wasn't available on 1967TS with the much, much simpler control system, but the <25mph in manual did function on the mechanical governors (provided they where working(!)), it may have (and I've not worked on 67TS for a long time) cut out access to weak fielding or locked the unit in Rate 1 without access to Rate 2.

Weak fielding effectively increases the top speed. Rate 2 effectively increases the rate of acceleration on a camshaft controlled DC motor, by allowing a larger current to flow (excluding Notch S1).

I don’t believe the Jubilee and Northern lines enforce any restrictions with manual driving. Certainly it’s absolutely possible for the best drivers to meet or even slightly surpass ATO performance, which supports this. The Seltrac ATO is generally not as refined as the Victoria Line.

 

[Nym]

You can beat ATO timings driving manually on the Victoria Line as well... It takes quite a bit of skill but it is possible.

 

[RWMarcus]

I think 1992/1995/1996 TS have restrictions when driving in any manual mode. On openBVE are some lines simulated. If you let the ATP drive from station to station it will normally run with a very fast speed into the station. If you drive in any manual mode you also can run with fast speed in the station but you will overrun the stop marker and the plattform because you don't get the brake force which the automatic mode has available. So I don't know if the 1992 TS is simulated correctly but the devs says it is. Idk

 

[Dstock7080]

I think 1992/1995/1996 TS have restrictions when driving in any manual mode.

This is the same with sub-surface CBTC.
In PM Protected Manual the acceleration, braking and top speed rates are reduced from that available in ATO.

 

[bramling]

I think 1992/1995/1996 TS have restrictions when driving in any manual mode. On openBVE are some lines simulated. If you let the ATP drive from station to station it will normally run with a very fast speed into the station. If you drive in any manual mode you also can run with fast speed in the station but you will overrun the stop marker and the plattform because you don't get the brake force which the automatic mode has available. So I don't know if the 1992 TS is simulated correctly but the devs says it is. Idk

I’m not sure these simulations are authentic in that respect. It is entirely possible to hit platforms on all three lines in the same way that ATO does, or even a touch later, and still stop in the correct place without issue. Likewise the target speeds displayed to the driver (and braking curves in the case of the Jubilee and Northern) are the same.

The only ATO system which is audacious is the Victoria Line, though the Central isn’t bad either. By contrast the Jubilee and Northern system is quite lethargic - in the open sections the braking is very early indeed, earlier even than the slowest drivers used to be.

 

[40129]

I seem to remember that when the Central line was converted to ATO, the system appeared to be driving more cautiously than human drivers had been on the above ground sections, whereas in the tunnels the opposite seemed to be the case. I presume this was because the computer had to drive to worst case conditions in terms of slipperyness of the rails on the above ground sections

 

[bramling]

I seem to remember that when the Central line was converted to ATO, the system appeared to be driving more cautiously than human drivers had been on the above ground sections, whereas in the tunnels the opposite seemed to be the case. I presume this was because the computer had to drive to worst case conditions in terms of slipperyness of the rails on the above ground sections

Yes the Central Line started out with a low brake rate in the open, but it soon got ramped up.

The different lines manage adhesion variability in different ways.

The Jubilee and Northern simply go for the worst case scenario and take the hit on journey times that this causes. Unfortunately the design of the system on these two lines is going to make it prove very difficult indeed to improve upon this.

The sub surface lines are similar to the Jubilee and Northern, but the system design is slightly less troublesome, plus they have a software tool which is able to make a prediction on what brake rate should apply at a given time. I believe DLR also select a brake rate based on weather forecasting, though I’m not quite sure on the specifics of exactly what they do (I think it’s simply one brake rate for the whole system - so this doesn’t fully compensate for localised problems).

Victoria Line isn’t an issue as it’s all in tunnels, and this just leaves the Central who deal with it the old-fashioned way - they run adhesion trains in the autumn to improve railhead conditions, and tacitly allow drivers to use their own judgement about when manual driving might be a good idea. They also simply tolerate a certain level of ATO-related operating incidents, for example stations overruns or even SPADs.

None of the ATO systems currently in use on LU are fully able to address the issue of changing railhead adhesion conditions. This is worth remembering when people talk about ATO being inevitable on national rail - so far it’s only happened or been proposed for very specific metro-type applications like Thameslink.

 

[100andthirty]

Northern Line has a number of track based adhesion improvement measures at sites demonstrated to be prone to poor adhesion.
The Metropolitan run adhesion trains, although ATO hasn't reached the affected parts yet, but I don't know whether it has been commissioned yet.
S stock has been equipped with sanders which improve poor adhesion immensely. In trials on the main line, the latest sanders were extremely effective. With the same poor adhesion condition set up for the test, the stopping distance from 55mph was 1200 m without sand and less than 400 m with sand.
The linked article mentions how sanding with ATO for S stock has been simulated as part of the proving trials for the sanding system: WSPER tests

 

[mrredrunner100]

The most common issue seems to be the trains loosing the codes for various reasons, you hear it on the radio frequently. In this case the driver sometimes resets the train or drives in RM to the next station where generally they'll pick the codes back up and carry on as normal. ATP/ATO can fail and lock up an area

Sort of related...I've been on trains before (Central I think) where either brakes have automatically been applied or traction current lost (I'm not sure which) due to door interlock being momentarily lost as passengers have been leaning against the doors.

I know the cause as the train op announced it!

What happens in this situation? Is Auto mode regained when interlock is re-established or do the trains have to be manually driven to the next station?

You won't have lost traction current because of this, you loose the pilot light and the brakes get applied, it can be overridden, but you should check no ones being dragged on the outside first. Most of the time you give it a few seconds and it comes back

 

[Tim M]

Small point.
ATO isn't part of the signalling system, it's part of the rolling stock system. The ATP is part of the Signalling System. There is a very distinct gap between the two systems in terms of design, in that the ATP has to be much more highly safety rated than the ATO. Going back to the first such systems, the ATP on 1967TS was designed by signalling engineers and in a completely separate box to the "Auto Drive”.

I’m sorry but your assertion that the ATO is part of the rolling stock system is misleading.The ATP and ATO systems supplied by Westinghouse for the Central Line and Victoria Line (original and current) are integrated with each other. Indeed the Singapore MRT system as adapted for the Central Line has both functions in the same housing, as was the installation on Beijing Line 1 (I was the Project Engineer for Beijing). Integration includes the trackside elements of both ATP and ATO with the ATP further integrated into the signalling interlocking.

Increasingly the ATO will be integrated with Control Centre Automatic Train Supervision and Regulation systems (ATS/ATR) in order to manage the service, for example dynamic control of coasting and station dwell and departure times. The importance of integrating all these systems has to be emphasised, separating the ATP/ATO from the ATS/ATR on the original Jubilee Line Extension Contract was seen as very odd by those involved at the time.

 

[100andthirty]

Correcting a number of misconceptions:
1) the original ATP/ATO system on the Victoria line was split up so that the Chief Signalling Engineer was responsible for the infrastructure and the on-train ATP box/pick up coils. The Chief Mechanical Engineer was responsible for the ATO box. That said, all the original equipment came from Westinghouse. The ATO box was replaced twice lad by the CME staff.
2) since then all the ATP/ATO equipment has been sourced as an integrated system from Westinghouse/Invensys/Siemens or from Thales. The latter is so integrated that it is barely possible to separate out ATO from ATP. Trainborne equipment is serviced in depot by depot staff but engineering responsibility remains with the signal department
3) I would be very surprised if a current Victoria line train could be driven in protected manual to achieve a faster run time than a flat out train in ATO. This is because the ATO has access to a brake rate that is 0.15 m/s2 higher than the peak rate in PM. The physics of achieving the best possible run time between two adjacent stations on a metro line are, assuming no PSRs between stations, accelerate as hard as possible up to line speed, run at line speed until the braking point for the next station, and then brake as hard as possible. As the ATO Victoria line train can brake harder then the manual one, it must deliver a better time. In practice, the Victoria line usually doesn't run flat out as coasting is built into the run profiles
4) This does not apply to the Northern and Jubilee lines, nor, I believe (but don't know for a fact), to the Sub-Surface lines. They have the same maximum service brake rate as is available to the driver in PM.

 

[Tim M]

3) The physics of achieving the best possible run time between two adjacent stations on a metro line are, assuming no PSRs between stations, accelerate as hard as possible up to line speed, run at line speed until the braking point for the next station, and then brake as hard as possible. As the ATO Victoria line train can brake harder then the manual one, it must deliver a better time. In practice, the Victoria line usually doesn't run flat out as coasting is built into the run profiles

When Westinghouse moved on from Hong Kong and various contracts via sister company Dimetronic in Spain to Singapore, a need to save energy was identified. Note both Hong Kong MTR and Singapore MRT (and for that matter Beijing Subway) used London Transport International as consultants.

In order to save energy an element of coasting was required and intelligent ATO was born. It’s now over thirty years since I was involved but let’s say 5% and 10% coasting options were required. Acceleration was to a speed set from interstation run simulation (mostly done by my colleague Terry Rees), thereafter coasting to an accurate stop against Platform Screen Doors at the next station. I think Singapore was the first Metro as opposed to a people mover to use PSD’s.

The default was maximum coasting to save maximum energy, but this could be varied before a train departed from a station from the Operations Control Centre. That had the potential to bring a late running train back onto schedule. In practice maximum coasting was always used, greater energy use being decidedly frowned upon.

When the Central Line upgrade to ATP/ATO, LU didn’t want coasting, I wasn’t involved in the later Victoria Line job, but as you say coasting is used, I think it was also used on the Contract for the airport line in Taipei using the same DTG-R system. Mind that job is a long story in itself!

 

[100andthirty]

Coasting wasn't originally included on the Victoria line either. The plan was to run flat out through the central.area and somehow use up timetabled recovery time in the suburbs. There was a counter argument that coasting could be used,controlled on a per train basis to provide some ability to catch up if a station stop was excessively long. This arguement won the day and fortunately the Invensys control system had the capability to do this. The Invensys chaps gave a talk about it at one of the institutions a few years ago; it was very clever.

 

[Tim M]

Coasting wasn't originally included on the Victoria line either. The plan was to run flat out through the central.area and somehow use up timetabled recovery time in the suburbs. There was a counter argument that coasting could be used,controlled on a per train basis to provide some ability to catch up if a station stop was excessively long. This arguement won the day and fortunately the Invensys control system had the capability to do this. The Invensys chaps gave a talk about it at one of the institutions a few years ago; it was very clever.

I’m sure my former colleagues at Chippenham (I retired a while ago) would be pleased to hear your last comment.

 

[Gadget63]

Good evening!

I'm from Germany, 30 years old and a little bit interested in the London Underground.

I have a question about disruption procedures on the London Underground.
Some lines are fitted with ATO or similar systems which allow the driver to lean back and let the train drive from station to station

I think ATO is a great system to increase capacity of a line. My last visit was about 5 years ago and I still remember the overcrowding of some trains what I don't know from Germany. I mean I know overcrowded trains in Germany too but not with that high frequency of 2 minutes.

I have to say that I am a supporter of the traditional train driving. I think ATO destroys the job of the train operator. I think it will also be possible to run trains every 100 seconds without ATO but with correct signalling systems (low signal distance, short blocks). As long as you don't have moving block (ATO) there is no difference in capacity between trains driven manually or by ATO with fixed blocks.

What I would like to know is how often the ATO system has failures. And what would happen if the system doesn't work? Can operators switch to Full Speed Manual and drive the train? Or will the line be suspended until the problem has been solved? I think the most drivers on the Jubilee Line aren't trained to stop excatly at the stopping marker without crawling. And as I know from openBVE the "Acurate Stop"-lamp isn't gracious. But maybe there are still many operators who were also on duty when the Jubilee-Line-ATO-System (I know it isn't called ATO on the JL) wasn't installed on the Jubilee Line (Extension). As I read about the Jubilee Line history I was very suprised that ATO wasn't installed in the first operation years of the Jubilee Line Extension. That must have been a challenge to stop at the correct position.

I'm very interested in that topic and I hope that someone can answer my questions.

I remember doing the testing of the 96 stock on the jubilee line extension. You had a little leeway in stopping at the PED doors. It was tedious getting the trains in and out of Stratford Market Depot at the start as we used loco's to dead drag them in and out.
The thinking of not putting ATO on the extension at the start is that at the time they didn't want 2 different systems working together, not like now with other projects.

I'm aware of this, but it was already a monster post so didn't want to drown everything in information.

The majority of traction control systems on "modern" units doesn't allow access to the full power of traction or braking either when being driven in a coded manual, protected manual or manned manual (someone used that term once on a Victoria Line report and it stuck for far too long). Victoria Line allows 0.05m/s/s less in terms of acceleration and braking when in Protected Manual (enforced by the traction and brake control systems), I believe that Northern and Jubilee Lines are the same. This can be implimented on older units given how the traction and brake control PWM is implemented, in that the traction brake controller and auto drive boxes have separate access to the PWM train wires, changing the bounds of access is therefore not too challenging.

This wasn't available on 1967TS with the much, much simpler control system, but the <25mph in manual did function on the mechanical governors (provided they where working(!)), it may have (and I've not worked on 67TS for a long time) cut out access to weak fielding or locked the unit in Rate 1 without access to Rate 2.

Weak fielding effectively increases the top speed. Rate 2 effectively increases the rate of acceleration on a camshaft controlled DC motor, by allowing a larger current to flow (excluding Notch S1).

We can do manual driving on the Victoria line on weekends between Seven Sisters an Walthamstow both roads at certain times, you do have to inform the Service Controller though. The only time we normally do manual driving is in and out of Northumberland Park Depot at 16Km/
I have driven an 09 from Walthamstow to Brixton only once which was an experience not to forget!

We can do manual driving on the Victoria line on weekends between Seven Sisters an Walthamstow both roads at certain times, you do have to inform the Service Controller though. The only time we normally do manual driving is in and out of Northumberland Park Depot at 16Km/
I have driven an 09 from Walthamstow to Brixton only once which was an experience not to forget!