Class 345 progress

[JonathanH]

The Heathrow branch is shut all weekend Saturday 21 / Sunday 22 November with the TfL Rail service shown working 4tph into the Hayes & Harlington bay on Saturday and, with the relief line blocked, 2tph Hayes shuttles running empty to Maidenhead (!) to reverse on the Sunday.

Is this just track maintenance or something to help the ETCS failures on 345s? I can't immediately see any paths for engineering trains.

As an aside, running empty to Maidenhead means that the units will operate further empty than in service on each journey. I assume that 7-cars will be used on the Sunday - it will be interesting to see what happens on the Saturday.

 

[matt_world2004]

The Heathrow branch is shut all weekend Saturday 21 / Sunday 22 November with the TfL Rail service shown working 4tph into the Hayes & Harlington bay on Saturday and, with the relief line blocked, Hayes shuttles running empty to Maidenhead (!) to reverse on the Sunday.

Is this just track maintenance or something to help the ETCS failures on 345s? I can't immediately see any paths for engineering trains.

As an aside, running empty to Maidenhead means that the units will operate further empty than in service on each journey. I assume that 7-cars will be used on the Sunday - it will be interesting to see what happens on the Saturday.

Running 4tph to maidenhead means theres going to be more trains running empty than in service!

 

[Peter Sarf]

As an old software engineer I feel it is all getting hopelessly complicated. One job I had ended when it became obvious that the new system was too expensive to bother with. Some of the requirements were really everything possible on a wish list. They sold the customer base instead. Now are we really needing all this fancy software ?. Is the king wearing clothes ?.

Is it time to keep it simple ?. You know - lay some more tracks etc to gain capacity. Costly etc I know but not "pushing the boundaries" of science and technology.

 

[samuelmorris]

As an old software engineer I feel it is all getting hopelessly complicated. One job I had ended when it became obvious that the new system was too expensive to bother with. Some of the requirements were really everything possible on a wish list. They sold the customer base instead. Now are we really needing all this fancy software ?. Is the king wearing clothes ?.

Is it time to keep it simple ?. You know - lay some more tracks etc to gain capacity. Costly etc I know but not "pushing the boundaries" of science and technology.

While I in principle agree with the sentiment, if you want an example of how it goes when you lay new tracks for extra capacity, take a look at HS2

 

[Peter Sarf]

While I in principle agree with the sentiment, if you want an example of how it goes when you lay new tracks for extra capacity, take a look at HS2

Yes, gosh that is a good point. Lets wait and see. I agree it is certainly not an environmental push over. I suppose the comparison awaiting HS2 is with the West Coast Mainline upgrade that partly was.

 

[Energy]

As an old software engineer I feel it is all getting hopelessly complicated. One job I had ended when it became obvious that the new system was too expensive to bother with. Some of the requirements were really everything possible on a wish list. They sold the customer base instead. Now are we really needing all this fancy software ?. Is the king wearing clothes ?.

Is it time to keep it simple ?. You know - lay some more tracks etc to gain capacity. Costly etc I know but not "pushing the boundaries" of science and technology.

Going ETCS L2 with ATO would have been a better choice, like they did on Thameslink. Crossrail has been treated as a metro with rail bits tagged onto the ends instead of a railway which goes underground which is why they choose CBTC for the metro bit (This is what is being used on LU and other metros around the world) and TPWS and ETCS L2 (Heathrow) for the rail bits and they switch between the systems which makes it very complicated.

 

[Peter Sarf]

Going ETCS L2 with ATO would have been a better choice, like they did on Thameslink. Crossrail has been treated as a metro with rail bits tagged onto the ends instead of a railway which goes underground which is why they choose CBTC for the metro bit (This is what is being used on LU and other metros around the world) and TPWS and ETCS L2 (Heathrow) for the rail bits and they switch between the systems which makes it very complicated.

That is true. Does remind me that it is London transports baby.

 

[swt_passenger]

Going ETCS L2 with ATO would have been a better choice, like they did on Thameslink. Crossrail has been treated as a metro with rail bits tagged onto the ends instead of a railway which goes underground which is why they choose CBTC for the metro bit (This is what is being used on LU and other metros around the world) and TPWS and ETCS L2 (Heathrow) for the rail bits and they switch between the systems which makes it very complicated.

It was publicised somewhere that TfL & LU somehow convinced themselves that ETCS/ATO couldn’t or wouldn’t be ready in time. Then I had an idea they did it primarily to keep other operators out of their tunnels...

 

[Non Multi]

It was publicised somewhere that TfL & LU somehow convinced themselves that ETCS/ATO couldn’t or wouldn’t be ready in time. Then I had an idea they did it primarily to keep other operators out of their tunnels...

As well as themselves, as it has turned out!

 

[swt_passenger]

As well as themselves, as it has turned out!

Nice one...

 

[Energy]

It was publicised somewhere that TfL & LU somehow convinced themselves that ETCS/ATO couldn’t or wouldn’t be ready in time. Then I had an idea they did it primarily to keep other operators out of their tunnels...

Surprising that nobody at the DfT said otherwise as it is fine for Thameslink. The later reason seems more likely, as if it wasn't ready at opening then installing ATO on ETCS afterwards would be comparatively easy.

 

[Peter Sarf]

Surprising that nobody at the DfT said otherwise as it is fine for Thameslink. The later reason seems more likely, as if it wasn't ready at opening then installing ATO on ETCS afterwards would be comparatively easy.

More proof that the DfT did not see the importance of what they should be doing ?. Or would that be interfering. My favourite example is how, after privatisation, it would be down to the DfT to ask the EU for the regular derogations that the BR used to get for the European standards for overhead electrification.

 

[Bald Rick]

It was publicised somewhere that TfL & LU somehow convinced themselves that ETCS/ATO couldn’t or wouldn’t be ready in time. Then I had an idea they did it primarily to keep other operators out of their tunnels...

Indeed, and ETCS / ATO is still not available for the full functionality required for Crossrail. Including autoreverse and Platfirm Edge door functionality.

 

[TT-ONR-NRN]

Have there been any West East swaps yet or are they all still on the same side as they have been for a while

 

[kevin_roche]

Tom Godrey of Bombardier in the IRSE talk seemed to imply that the transitions between the TPWS/AWS and CBTC were not the main problem. See post #3651 for a link.

When you consider the interfaces to other systems that are in place on the Crossrail COS its seems unlikely that anything else would have been much easier to implement.

The IRSE talk in May by Tom Godrey of Bombardier and Rory Mitchell of Crossrail goes into some detail. I took several of Rory Mitchell's slides and put them together to build a complete picture. This highlights Mark Wild's comments to the HS2 developers warning them not to over complicate things.

I'm not sure what all the boxes are but I think some of the acronyms are:

PMS TfL Performance Management System
LWOD Line Wide Overview Display
ITPS TfL Integrated Train Planning System
TRUST Network Rail computer system used for monitoring the progress of trains and tracking delays
DWH Data Warehouse
TD Net Network Rail Open Rail Data
ITPS Integrated Train Planning System
SCADA Supervisory Control and Data Acquisition
PSD Platform Screen Doors
CIS Customer Information System
ATS Automatic Train Supervision

If anyone can translate any others please let me know.

There is also a new IRSE video lecture by Alessandra Scholl-Steinberg of Siemens Mobility about the integration testing facility used to test interfaces on Crossrail https://vimeo.com/467293446

 

[JonathanH]

A pretty bad set of runs on Heathrow services today, with, I think, no less than three 9-car units sent to the depot and six units suffering issues.

Diagram starting 9T01 0442 London Paddington to Heathrow Airport Terminal 5
9T11 0602 London Paddington to Heathrow Airport Terminal 5 lost 14 minutes at Heathrow Central - slight knock on delay on the next trip
9P97 1722 Heathrow Airport Terminal 5 to London Paddington delayed 6 minutes from Heathrow Airport Terminal 5
9T07 1802 London Paddington to Heathrow Airport Terminal 5 delayed 12 minutes at Hayes & Harlington
9P09 1852 Heathrow Airport Terminal 5 to London Paddington cancelled as a result - unit went empty to Old Oak Depot

Fresh unit sent to Paddington for 9T19 1932 London Paddington to Heathrow Airport Terminal 5 but subject to 21 minute delayed departure
9P21 2022 Heathrow Airport Terminal 5 to London Paddington delayed 4 minutes at Heathrow Airport Junction

Diagram starting 9T03 0502 London Paddington to Heathrow Airport Terminal 5
9P41 1022 Heathrow Airport Terminal 5 to London Paddington lost 11 minutes at Southall - following trip missed stops
9T63 1232 London Paddington to Heathrow Airport Terminal 5 lost 11 minutes between Tunnel Junction and Heathrow Central - slight knock on delay on next trip (9P65) - unit taken out of service at Paddington
9T75 1402 London Paddington to Heathrow Airport Terminal 5 and 9P77 1452 Heathrow Airport Terminal 5 to London Paddington cancelled as a result

Fresh unit sent to Paddington for 9T87 1532 London Paddington to Heathrow Airport Terminal 5
9T11 1832 London Paddington to Heathrow Airport Terminal 5 cancelled at London Paddington - unit went empty to Old Oak Depot

Fresh unit sent to Heathrow Airport Terminal 5 for 9P13 1922 Heathrow Airport Terminal 5 to London Paddington
9T35 2132 London Paddington to Heathrow Airport Terminal 5 lost 11 minutes at Heathrow Central

Diagram starting 9T07 0532 London Paddington to Heathrow Airport Terminal 5
9T91 1602 London Paddington to Heathrow Airport Terminal 5 delayed 10 minutes between Tunnel Junction and Heathrow Central

Everything else ran to time but it does seem a bit much to have the ongoing issues with the 9-car 345s.

In terms of 7-car units, the unit which worked 9P60 1222 Reading to London Paddington was taken out of service on arrival and swapped for another unit for 9R72 1343 London Paddington to Reading but otherwise I can't see any issues. That said, a unit was taken out of service at Iver yesterday evening on a late evening up service.

 

[kevin_roche]

A pretty bad set of runs on Heathrow services today, with, I think, no less than three 9-car units sent to the depot and six units suffering issues.

I am assuming it is a software issue. They must be hoping that the new version PD+12 or TR2 software will fix these issues. It was slated for mid-November on the last (August) Jacobs report.

 

[samuelmorris]

Not that I expect any substantial remedial work to have been carried out during the COVID era, but Aventra reliability numbers are still on the decline -

Graph link below

https://i.imgur.com/zOxuZaa.png

 

[JonathanH]

Not that I expect any substantial remedial work to have been carried out during the COVID era, but Aventra reliability numbers are still on the decline -

Graph link below

https://i.imgur.com/zOxuZaa.png

It is a bit too out of date to be draw live conclusions as it only hoes up to August - 9-345 didn't operate in passenger service between 12 June and 29 July and then had some initial problems on Heathrow services (which seem to be ongoing). The fluctuations in the performance of 7-345 are interesting but you could see how they might correspond to the introduction of services to Reading in December 2019 and some more challenging times since.

7-345 performance isn't that bad on the GWML (and neither presumably on the east side other).

The thing about 9-345s is that the problems don't only happen on the Heathrow branch - a quick look at RTT and you see delays at the other stations as well - eg 9T47 1032 London Paddington to Heathrow Airport Terminal 5 twelve minutes late from Paddington earlier, 9T71 1332 London Paddington to Heathrow Airport Terminal 5 delayed six minutes at Hanwell just now - the same unit.

 

[JN114]

9T47 wasn’t a unit fault, there were OHLE issues at Paddington briefly.

1K11, 1A13 and one of the HEX driver trainers got clobbered as well

 

[Horizon22]

It is a bit too out of date to be draw live conclusions as it only hoes up to August - 9-345 didn't operate in passenger service between 12 June and 29 July and then had some initial problems on Heathrow services (which seem to be ongoing). The fluctuations in the performance of 7-345 are interesting but you could see how they might correspond to the introduction of services to Reading in December 2019 and some more challenging times since.

7-345 performance isn't that bad on the GWML (and neither presumably on the east side other).

The thing about 9-345s is that the problems don't only happen on the Heathrow branch - a quick look at RTT and you see delays at the other stations as well - eg 9T47 1032 London Paddington to Heathrow Airport Terminal 5 twelve minutes late from Paddington earlier, 9T71 1332 London Paddington to Heathrow Airport Terminal 5 delayed six minutes at Hanwell just now - the same unit.

Yes that the 7-cars work broadly fine and the 9 cars don't is evidently an issue.

 

[JonathanH]

9T47 wasn’t a unit fault, there were OHLE issues at Paddington briefly.

1K11, 1A13 and one of the HEX driver trainers got clobbered as well

OK, thanks, look like I picked the wrong example - presumably the nine minute delay to 9T55 1132 London Paddington to Heathrow Airport Terminal 5 was a unit problem? Point is that it isn't only on the Heathrow branch where they seem to have issues.

 

[samuelmorris]

It is a bit too out of date to be draw live conclusions as it only hoes up to August - 9-345 didn't operate in passenger service between 12 June and 29 July and then had some initial problems on Heathrow services (which seem to be ongoing). The fluctuations in the performance of 7-345 are interesting but you could see how they might correspond to the introduction of services to Reading in December 2019 and some more challenging times since.

7-345 performance isn't that bad on the GWML (and neither presumably on the east side other).

The thing about 9-345s is that the problems don't only happen on the Heathrow branch - a quick look at RTT and you see delays at the other stations as well - eg 9T47 1032 London Paddington to Heathrow Airport Terminal 5 twelve minutes late from Paddington earlier, 9T71 1332 London Paddington to Heathrow Airport Terminal 5 delayed six minutes at Hanwell just now - the same unit.

7-car performance isn't awful, but apart from a couple of outliers, has yet to reliably beat that of the 40 year old 315s they are supposed to replace, and furthermore, it continues to decrease. I'm not aware of any substantial drop in workload for the 7-car fleet apart from the initial kneejerk reaction with the first lockdown, so that is somewhat concerning. The 710 decay is perhaps explained by the introduction of the units to new areas (the more recently introduced /1 units drag down the average from the /2s). I understand software updates to the 345s are only coming via the 9-cars which does make sense, but as of now, there's nothing solid to suggest that, whether the issues are software or hardware, the Aventras are going to provide good enough performance to run the full through service whenever it actually opens. Presently even the 7-car 345s are failing more than 10 times as often as the S-stock used on the underground, and arguably given their longer gaps between stations they should be pushing higher numbers, not lower.

 

[kevin_roche]

I have just noticed that Crossrail's response to the August Jacob's report contains the following:

Four trains per hour to Heathrow depends on Bombardier Transportation delivery of new software fixes to ameliorate the current level of European Train Control System resets required in passenger service (running at circa 0.09 resets per operating hour). A partial fix is planned in October 2020, followed by an additional fix in December 2020.

 

[JonathanH]

I have just noticed that Crossrail's response to the August Jacob's report contains the following:

Four trains per hour to Heathrow depends on Bombardier Transportation delivery of new software fixes to ameliorate the current level of European Train Control System resets required in passenger service (running at circa 0.09 resets per operating hour). A partial fix is planned in October 2020, followed by an additional fix in December 2020.

Good spot. 0.09 resets per hour is the same as one reset every eleven hours so roughly two a day most days. Today and yesterday look as if they have been a lot better than Monday though.

 

[SeaKing]

Have anymore entered service or testing recently, not been able to pick numbers up on Slough Railcams of late.

 

[Bigbru]

Good spot. 0.09 resets per hour is the same as one reset every eleven hours so roughly two a day most days. Today and yesterday look as if they have been a lot better than Monday though.

Is that roughly 2 per day per train. As they are currently running 3 trains for a 2 tph service that's 6 resets a day

 

[Horizon22]

Good spot. 0.09 resets per hour is the same as one reset every eleven hours so roughly two a day most days. Today and yesterday look as if they have been a lot better than Monday though.

My experience would suggest more like 5-6 a day. Often its the same unit although MTR seem to be applying a "3-strikes-and-out" approach before returning the affected unit to the depot.

 

[Peter Sarf]

I have just noticed that Crossrail's response to the August Jacob's report contains the following:

Well spotted. I notice the report you quote says 0.09 resets per hour (sorry I don't know how to get it quoted). I assume that means on average almost one reset every six minutes ?.Edit- just beaten to it which made me spot my mistake !. I suppose it depends how long a reset takes but that seems quite a regular risk of disruption so avoiding a doubling of frequency to 4 trains per hour seems prudent.

 

[Horizon22]

Well spotted. I notice the report you quote says 0.09 resets per hour (sorry I don't know how to get it quoted). I assume that means on average almost one reset every six minutes ?. I suppose it depends how long a reset takes but that seems quite a regular amount of disruption so avoiding a doubling of frequency to 4 trains per hour seems prudent.

You've done the maths the wrong way around there - 0.09 every 60 minutes. So according to their data that's 1 reset approximately every 600 minutes (10 hours).

Edit: An average ETCS reset takes about 8-10 minutes.