Crossrail (finally) reaches Stage 2 of opening

You may wonder how you missed the news of Crossrail achieving its second milestone (out of five) in its opening strategy to fully open the Elizabeth line. The reason is simple. It hasn’t been publicised. It is a contender for the title of biggest soft launch of a transport upgrade, certainly in recent times.

We exaggerate. At London Reconnections we are used to the odd ominous sentence appearing buried in the middle of an upbeat press release assuring us that all is right with the world. We live and thrive on such things. What we are not used to is a single sentence in a Letter to the Assembly confirming that Crossrail Class 345 trains are now running in public service from Paddington to Heathrow.

Let us be clear – this means that a major milestone has been achieved. What is more, the milestone has been achieved with 9-car trains with the latest problematic software crucial for full running in central London, not the 7-car trains with simplified software only really suited for above-ground services in legacy-signalled areas.

The full-length Class 345 trains (9 car units) are now operating on services out of Paddington and Heathrow. This is a significant milestone as it uses the ETCS signalling system and means that we continue to build mileage on the Class 345 fleet that will be used on the full Elizabeth line.

From the latest Crossrail press release (near the bottom)

An unusual lack of publicity

It is typical of TfL to quietly implement not-obviously-visible upgrades (such as improved frequency or changeover to automatic train operation) so that, if there are problems, they are not left with red faces. However, replacing 5-car not-very-good Class 360 trains originally used for Heathrow Connect with 9-car high capacity Crossrail trains is far from a minor improvement. It is also in great contrast to the Stage 1 opening (the first Class 345 train) on the Liverpool St – Shenfield route) which took place with much publicity. Certainly, in a Covid-19 world you don’t want journalists and rail fans descending onto the line in a mass but to have virtually no official publicity at all is a bit of a shame.

It could just be that the delay to Stage 2 is rather embarrassing and Crossrail and TfL would rather not remind people of this. After all, the perception of the average person who shows any interest in this is that TfL Rail is already running a service to Heathrow from Paddington and all that has happened is that there has been a change of rolling stock. Indeed, when Crossrail had to report to the Mayor back in 2018 about the delays in opening, it rather disingenuously suggested that Stage 2 had been achieved because TfL trains were running between Paddington and Heathrow on the planned date.

2 becomes 2A and 2B

Subsequently, Crossrail became somewhat more honest and retrospectively split Stage 2 further into 2 stages – Stage 2A and Stage 2B.

Stage 2A was the takeover of Heathrow Connect service using Class 360 trains. This, to the user, involved little more than replacement of the decal on the trains to reflect its new operator (TfL Rail). In fact, the actual service available to the customer was a retrograde step because the toilets on the trains were locked out of use as it was not TfL’s policy to provide toilets on trains. In fairness to TfL, their policy is to provide toilet facilities on stations. Nevertheless, given that new toilets did not suddenly appear when TfL Rail took over the service, this had to be a retrograde step at the time as far as the travelling public was concerned.

It is true that TfL’s takeover of the Paddington Heathrow service led to a reduction in fares. Better still for Londoners with an Over 60 or Freedom Pass, travel between Heathrow and Paddington was now free, so long as you didn’t use the Heathrow Express service.

Stage 2B, which is now at least partially implemented, is the running of Class 345 Crossrail trains between Paddington and Heathrow. There are a couple of caveats. The first is that the Class 360 trains haven’t quite been withdrawn and could potentially still appear as a substitute if a Class 345 was not available. A last-minute substitution of a Class 345 by a Class 360 happened on the first day of planned operation (Thursday 30th July 2020).

The second caveat is a technicality. The planned implementation was to be between Paddington and Heathrow Terminal 4. In fact, Terminal 4 station is not served and trains are running instead to Terminal 5. The impact of Covid-19 on air travel has led to Heathrow operating everything from either Terminal 2 or Terminal 5. This is unlikely to change soon as one runway is closed to take advantage of the opportunity for major refurbishment and repair and Terminal 4 is neither needed nor conveniently sited for the other runway.

A short primer on Crossrail signalling

For those not familiar with the complexities of Crossrail signalling, it is extremely fortuitous that Crossrail have recently released an excellent video describing the challenges of signalling on Crossrail and it is well worth watching in full:

In addition, this screenshot taken from the video shows the intended final areas of different signalling on Crossrail.

The problems of ATP

This unheralded milestone may not appear significant, but it is. Regular readers may remember that there was considerable doubt as to whether it would ever be possible for Crossrail trains to use the Heathrow tunnels without switching off the obsolescent ATP (Automatic Train Protection) signalling system operating in the tunnels. ATP is used by both the Class 360 trains which deputised for the Crossrail trains, and the Heathrow Express Class 332. Unfortunately, ATP is what the electrical engineers would call ‘noisy’ with plenty of stray radio emissions that interfere with the modern ETCS (European Train Control System) signalling now installed in the tunnels.

At one stage it was believed that Crossrail trains might have to wait until the Heathrow Express trains were replaced by Class 387 trains operated on Heathrow Express’ behalf by GWR. The fact that the original Heathrow Express stock is still operating and that Class 360 can still be used shows that ATP is still functioning in the tunnels, so the engineers working on Crossrail must have finally managed to get Class 345 trains to work in an ATP environment. It is important to give credit as well as criticise and here credit is most definitely due.

ETCS – the new baddie

In fact, as it turns out, it was not the interference of the legacy ATP signalling that ultimately delayed the introduction of Class 345 trains to Heathrow. Once that got fixed, it was the software on the trains themselves that had to work with the new signalling, ETCS (European Train Control System), that was giving problems in the train tunnels and – as it turned out – elsewhere.

ETCS generally relies on continuous communication and that is done by GSM-R, which is a dedicated radio band on the mobile phone spectrum allocated exclusively for railways (hence the -R suffix). Not surprisingly, this tends to work less well in tunnels. It also turns out that the longer the train, the more opportunity for that train to block a signal while in them. Unfortunately, but vital for safety, if the train no longer receives the expected GSM-R signal the brakes come on.

As it subsequently turned out, there were other ETCS problems too that gave issues above ground. This meant that the 9-car ETCS-enabled Class 345 trains recently introduced between Paddington and Reading had to be replaced by 7-car trains that previously ran the service.

The software on the 7-car trains is merely an update of software that has been around for about twenty years. As such it is quite crude by modern standards. It was written when ETCS was in its infancy so does not have any support for that signalling protocol and, as such, is suitable only for the older and more established TPWS/AWS (Train Protection and Warning System/Automatic Warning System). In complete contrast, the 9-car software is new and expressly written with modern signalling systems and Crossrail in mind.

Although safe, one problem with depending on the software on the 7-car trains is that it is not considered as safe as the existing system in the Heathrow tunnels, since TPWS and AWS are not considered as offering as much protection as the legacy ATP system. This means it falls foul of the Office of Rail and Road (ORR) rule that replacement systems must be at least as safe as what they are replacing.

The significance of it all

Crossrail has overcome two major hurdles. It has demonstrated that 9-car Class 345 trains can run in passenger service on ETCS-equipped lines. This means that 9-car Class 345 trains can run from Paddington to both Reading and Heathrow. With platform lengthening work at Liverpool St expected to commence by Christmas 2020, this paves the way for TfL Rail to convert all Crossrail trains to their final planned 9-car configuration. This in turn eliminates various logistical problems in keeping the two configurations in the right place for current operations. It also eliminates the problem of storing the extra two carriages at various locations in the country.

Possibly more important still, Crossrail has finally, properly, achieved its objective of running Class 345 trains to Heathrow, which was one of the five major stages of implementation of Crossrail.

In fact, no hurry

TfL Rail nearly managed to introduce 9-car trains much earlier (around May 2020), but a whole series of last-minute problems started appearing. This delayed official approval, and various technical issues followed. In normal circumstances the spotlight would have been on this issue, with the TfL Board asking questions and bad publicity in general. One effect of Covid is to take the pressure off the implementers whilst they sort out the various problems.

One reason for a lack of concern is that it was hoped the delay would mean that trains could run under ETCS from Heathrow all the way to Acton Main Line station. This would have meant that TfL Heathrow trains would only have to rely on TPWS/AWS between Acton Main Line and Paddington. Full ETCS into Paddington is unlikely to be operational in the short term due to the complexity of the layout.

As it turned out, ETCS still isn’t live between Airport Junction (where Heathrow trains leave and join the Great Western Main Line) and Acton Main Line station so TfL Rail have had to accept the additional complexity of changing from one signalling system to another whilst on the move.

Benefits for TfL Rail

Once TfL Rail are committed to the Class 345, the Class 360 trains can finally be withdrawn. One presumes that TfL Rail will be pleased to see these go. As well as having to pay to lease them (even while their own trains are sitting in the sidings), they were problematic to maintain and were becoming unreliable. Indeed, if it wasn’t for Covid-19 and the reduced timetable along with introduction of Class 345 from Paddington as far as Hayes & Harlington, there probably would have been a lot of bad publicity about the unreliability of the service.

A Class 345 at Hayes & Harlington

Benefits for HS2

Less obvious is the potential benefit to HS2. The Class 360 trains are maintained at Old Oak Common in a depot that it shares with the current fleet of Heathrow Express trains.

The depot was supposed to be handed over to HS2 a couple of years ago but had to continue to function until the fleets of TfL-leased Class 360s and Class 332s (Heathrow Express) based there were withdrawn. It is expected that the Heathrow Express trains will start to be withdrawn in the next few weeks, when they will be replaced by GWR-maintained Class 387s dedicated to Heathrow Express, based at Reading for maintenance purposes. However, even with the current reduced requirements, the lack of Class 387 trains ready and tested for use on Heathrow Express services means that it will be some time before the entire fleet is withdrawn.

With HS2 having taken over the Old Oak Common site at the beginning of August 2020, the need to demolish the maintenance depot is becoming critical, as this is necessary before the station box can be built. It may be with the withdrawal of the Class 360 trains and the need for only a reduced fleet of Heathrow Express Class 332 trains that enough of the site can be released to HS2 so as not to delay construction.

The current situation

So far, despite the introduction of the Class 345 trains to Heathrow, the timetable remains the same. This is the temporary timetable (not published) that sees 4tph from Paddington to Heathrow Terminal 5. Two of the trains are Heathrow Express trains (leaving Paddington at xx25 and xx55) and two are TfL Rail trains (leaving at xx32 and xx02) – so there is approximately a 15 minute headway through the timetable-critical Heathrow tunnels.

Current advertised service Heathrow – Paddington
Current advertised service Paddington – Heathrow

No TfL Rail trains are currently advertised as running between Paddington and Hayes & Harlington (and terminating there), with the National Journey Planner not showing any indication that this service currently exists. However, sites like OpenTrainTimes.com reflect the true situation, which is that they are in fact running. TfL Rail is starting to use 9-car trains on this service. It is not known when this service will be publicised but, presumably, TfL Rail want to be confident is it reliably running with 9-car trains before doing so.

The service to Reading is intended to be run using 9-car trains but when new signalling problems with the ETCS signalling manifest themselves (regardless of whether the cause is track or train based), the 7-car trains are temporarily substituted. It remains to be seen for how long TfL Rail feel they need to keep a 7-car contingency fleet for such purposes, but this is largely beyond their control as Network Rail iron out various snagging issues.

Transition at Westbourne Park

Those who have been following the Crossrail signalling scenario will know that the original plan was to have ETCS signalling out of Paddington ready prior to Crossrail preparing for service. This would mean that when going west, Crossrail trains would transition from Communications Based Train Control (CBTC), the signalling system in the tunnels, to ETCS. As both of these are advanced ‘intelligent’ systems this transition should be possible without any reduction of speed.

Subsequently, Network Rail’s plan to introduce ETCS on the Great Western Main Line was delayed, so Crossrail had to prepare for transitioning from CBTC to TPWS when travelling west. This would mean going from a data-rich environment to a data-poor environment and would lead to a reduction of speed or capacity or both. It now seems that Crossrail can (and should) go back to transitioning to ETCS.

Eliminate the risk

As we are all too aware, the Crossrail Project still has many risk factors associated with it. By operating Stage 2 some of that risk is removed. The signalling west of Paddington becomes known and proven. The 9-car Class 345 trains can accumulate more mileage. With plans to extend the platforms at Liverpool St over Christmas and many surrounding weekends, the need for 7-car trains on Crossrail should be eliminated. This means the risk involved in converting from 7-car to 9-car will disappear by the time the core service (Stage 3) is up and running. Testing will be able to take place in the Central London tunnels with trains that are known to work with ETCS in tunnels elsewhere once Network Rail sort out various infrastructure issues.

In a way, the good news about Stage 2 does nothing to aid opening Crossrail between Paddington (Low Level) and Abbey Wood (Stage 3), but it should give the team a boost of confidence and give them slightly less to worry about. We are now seeing low-key tangible progress with Crossrail.

As usual, many thanks to ngh who has provided a lot of additional information and made many corrections to technical points. Thanks also to Long Branch Mike and Jonathan Roberts.

London reconnections is a community-run website. If you like what we do, then consider supporting us on Patreon. Every little helps pay our bills.

39 comments

  1. Does this mean that the Class 345 FLU AWS issues, which I an writing about this very minute, have been resolved?

  2. Informative, thank you.

    “As it turned out, ETCS still isn’t live between Airport Junction (where Heathrow trains leave and join the Great Western Main Line) ”

    between Airport Junction and where please?

    [Acton Main Line. Text now corrected. PoP]

  3. Roger Ford,

    I believe not. As I understand it the primary problem with AWS is at West Drayton and so does not affect trains to Heathrow – or 7-car trains to Reading.

    I expect ngh can comment better on this.

  4. Re Roger,

    Yes, resolving that was required to introduce the Heathrow Services last week.
    The AWS issue became prominent shortly after the ORR ETCS authorisation in mid May, which forced the delay in starting Heathrow services till last Thursday.
    First testing on the new software with the fix was early July.

  5. Re PoP,

    The West Drayton issues are separate signalling ones but that only restricts services to the use of 7 car units on the Reading / Maidenhead services

  6. “This means it falls foul of the Office of Rail and Road (ORR) rule that replacement systems must be at least as safe as what they are replacing.”

    Tell that to Chiltern who have got permission to replace their ATP with TPWS+…

  7. Meanwhile …..
    It seems that C-19 was both a benefit ( More work could be done, because fewer trains were running ) & a drag ( because work could not proceed so rapidly, because of distancing requirements on personnel.)
    I think we’ve all seen the discreet announcement that stage 3 is now likely to be “later” (unspecified) in 2021 than previously thought.
    How much later?
    “Autumn” or “Winter” i.e. before 31/12/2012?

    Most informative, though – the NR video was timely.

  8. Muzer,

    You can make TPWS+ extremely safe providing you make it ‘+’ enough. Ultimately though there comes a point where the risk of signalling workers experiencing injury or death on the track maintaining the additional equipment negates the extra safety to train staff and passengers.

    I would imagine that, above ground, you could present a case to ORR where a well-specced TPWS+ system is as good as ATP. Chiltern are, arguably, in a better position because the track equipment is becoming absolutely life-expired (so the ATP isn’t better if it cannot be maintained). In the Heathrow tunnels you could argue that TfL were moving away from ATP by choice. Remember the GWR Inter-City Express Trains were fitted with ATP and (as far as I am aware) are still using it.

  9. Greg T,

    That seems to be exactly the case. So some major project advancements where it has become possible to have a blockade for a couple of weeks where it would normally be unthinkable (e.g. Kilsby tunnel). But also social distancing can slow a project down especially signalling. So resignalling in the Hither Green area due to be carried out over four days at Easter had to be rescheduled for 9 days in July/August due to the need for signal engineers to socially distance.

  10. Heathrow airport returned to two runway operation from 2 August 2020.

  11. Another great job by the team, this helps enormously in understanding where the project has got to.
    Re Terminal 4. Although the airport is now using both runways (only at some times, there is still some overnight work on the southern runway), the volume of flights is still so far below normal that there is no immediate prospect of terminal 4 reopening.

  12. Re Island Dweller,

    Indeed Heathrow Airport is only operating at around 20-25% of capacity flight wise so far this month with big variations in how busy it is during the day.

  13. Excellent as always PoP – thank you. Interesting about the knock-on /spin-off ‘re HS2. Is there a forthcoming commentary on HS2??

    Back to Crossrail, if we were to assign a factor of difficulty to the integration of the various signalling systems,how far out was the plan? Did the plan underestimate by a factor of 2,3,4? I ask because this seems to be the single area where the project has stalled/faltered/derailed (choose your own adjective)?? Best, Mike

  14. Mike Wilks,

    Back to Crossrail, if we were to assign a factor of difficulty to the integration of the various signalling systems,how far out was the plan? Did the plan underestimate by a factor of 2,3,4?

    A question I find very interesting. In all honesty I would say a minimum of 6 and possibly as high as 12. I base this on coming to the conclusion that the testing phase allowed as many months for signalling and software testing as it should have allowed years. Modern software is extremely complex – but very powerful once it works.

    I came to the conclusion months ago that the original Crossrail Board just had no concept of how software works in the modern world and how integrated it has become. I have some sympathy with them as it changed dramatically during the period of Crosrail development. But they should have recognised this and demanded they had an extra year (which wouldn’t have been enough but at least would be a step in the right direction).

    I have only just understood that getting Class 345s running between Liverpool St and Shenfield didn’t really advance the project that much because all the critical software aboard the trains was dead-end stuff. Nothing of value got tested in any way that helped implement the full Crossrail service. I think this stage 1 opening gave the project sponsors and others false comfort.

    To use other analogies, it is like the difference between a Boeing plane (an aeroplane with computers aboard e.g. 737max) and an Airbus plane (a computer with wings). Never publicised about the Miracle on the Hudson is a belief amongst many that if Captain Sullenburger had been flying a Boeing he wouldn’t have stood a chance of getting the plane down safely. The Airbus computer took over a lot of the critical stuff (how far to push he plane) which meant he and the co-pilot could fly (or glide) the plane to its limits.

    Or its like having a Tesla and an electric car that isn’t a Tesla.

    If really interested, I wrote an article for Railway Gazette last February that touched on this subject. The article is an update of one that appeared on this website but with much more emphasis on the background behind software problems. I have uploaded it and you can read it here.

  15. Thanks for the article, very interesting as always.

    @Mike Wilks: this seems to be the single area where the project has stalled/faltered/derailed

    I would say that it is one area in which the project has stalled. But Mark Wild’s update to the London Assembly makes it clear that even if all the software worked perfectly today, trial running couldn’t start because the route way isn’t complete (hence the 6 week blockade about to start), nor is Stepney Green shaft (expected handover mid-August), nor is Bond Street station ready for trial running (expected “in the coming months”).

    Meanwhile the independent Project Representative say bluntly that “CRL has failed to achieve a programme that has achievable milestones and dependable forecast dates but built around overly optimistic production rates and schedule durations.” They note that “The Critical Path to Trial Running start is through the completion of D25 noise mitigations on Shafts, the Tunnel Ventilation System; Bond Street Station mechanical and electrical works, and Bond Street Station SC1 completion”.

    So assuming that the blockade resolves the first things on this list, it looks like Bond Street will ultimately determine when Trial Running can start (I think SC1 is when the station is complete enough that you can run a trial service through it). “The coming months” is a very vague timetable.

    @Greg T: How much later?

    Well if Trial Running takes about 12 months and can’t start until Bond Street is finished in “the coming months” (ie. more than 2) that gives an earliest possible date of about October 2021… But see the Project Representative comments above.

  16. @PoP: the original Crossrail Board just had no concept of how software works in the modern world

    I think in particular, no concept that delays to software projects aren’t like delays to construction projects in that you can’t just add personnel or extra shifts etc to recoup delays. Brooks’s law applies: “adding manpower to a late software project makes it later”.

    And software applies not just to the signalling, but to the stations as well (all those fire alarm systems etc, which have to be integrated with the existing Tube station systems), and to tunnel ventilation etc.

    There’s an interesting page here about the control system for the ventilation system, for example. It is a safety critical system so has to be certified to a certain level, and it interfaces with the signalling system – the ventilation system needs to know where trains are to provide emergency ventilation, and if a fan shaft fails the signalling has to restrict the number of trains in a given section.

  17. A few observations on 9-car operation.

    9-car 345s have run one diagram daily to Heathrow since 30 July. The other two diagrams remain 360s for the time being. From 1 August, 9-car 345s started to run the two Hayes diagrams.

    This introduction hasn’t been without issue, with a number of services having 30 minutes at Heathrow Tunnel Junction on the transition between the two signalling systems although the amount of downtime when these problems occur does seem to have reduced in the light of experience. There are certainly reports of Bombardier staff travelling on these services, which may be why only one Heathrow diagram has been converted.

    Yesterday, it appears that a 7-car unit was moved from Old Oak Depot to Ilford EMU Depot (leaving thirteen 7-car units at Old Oak Depot for nine passenger diagrams to Reading and a further ‘hot spare’ for the peaks which sits at Maidenhead in the morning peak and on the Paddington carriage line in the evening peak) which is a welcome sign that they are confident that 9-car operation can be sustained. A further movement to Ilford today appears to be of a 9-car unit.

    In terms of back story, from observations, I think this is how it has gone.

    9-car units started testing on the western route in February 2018. As noted in a previous London Reconnections article, a contingency plan was needed for the May 2018 introduction of TfL Rail services to Hayes & Harlington which saw 7-car units move over from the east side of London.

    The first passenger operation of 9-car units was during July 2019. A unit would run empty to Maidenhead during the afternoon peak and then, run empty to Hayes & Harlington to come into service at 1948 replacing the 7-car unit which had been in use during the day. There also appear to have been some weekends when 9-car units worked all Hayes & Harlington shuttles. This operation continued until the December 2019 timetable change but then stopped.

    Past reports suggest that it was originally envisaged that 9-car operation would start to Reading in December 2019 but by July it must have been clear that this would not happen and the operators started a program of converting 9-car units to 7-car units to rely on the software solution that was running reliably on the east side of London.

    Seventeen units were converted from 9-car to 7-car by the end of December 2019 to add to the fifteen units already delivered in this formation.

    In early March 2020, just before the lockdown, 9-car 345s started to be reintroduced on Hayes & Harlington shuttles. By early April 2020, once TfL had reduced its service to half-hourly Reading and Heathrow service, 9-car 345s were running some services to Reading alongside 7-car units. By the end of April and in early May, Reading services were almost exclusively 9-car units although there were only six daily diagrams at the time.

    Come the May timetable change, the Reading service required nine units. From 18 May, TfL Rail were running seven diagrams with 9-car units and the other two with 7-car units, swapping the 7-car units mid morning. (This appears to have been reported as full 9-car operation.) The Hayes service wasn’t running and 360s were running the Heathrow services.

    There was a period of poor reliabilty in the 360 fleet at around this time, 7-car 345s were used to fill in for failed 360s but, of course, could only run to the bay at Hayes & Harlington.

    On 4 June 2020, TfL Rail ran the full Reading service with 9-car units for the first time since the nine diagram service had started. Hayes shuttles restarted at peak times the following week. Heathrow operation was rumoured to start on 11 June but didn’t happen.

    On the night of 11 June 2020, the 9-car 345 were taken out of service. Three Reading diagrams end at Old Oak Depot, two were replaced by the expedient of sending empty 7-car units to Maidenhead to swap out 9-cars that finish after the evening peak and three were replaced at Paddington. One other unit worked a service to Reading and returned empty the following morning.

    By this time, there were only eight 7-car 345s at Old Oak Depot, one of those due for conversion to 9-car with coaches delivered from Worksop and some of the others due to be taken to Ilford. It appears that it was only good fortune that meant they could run a service when the 9-cars were withdrawn.

    The eight available 7-car 345s at Old Oak Depot were put into service on 12 June leaving one diagram uncovered and five 360s worked the five Heathrow and Hayes & Harlington diagrams. The fact that there weren’t enough units for the fourteen diagrams suggests that this was not planned.

    Over the next few days, a unit under conversion to 9-car was converted back to 7-car and a unit came across from Ilford to provide sufficient cover. Three further 9-car units were subsequently converted to 7-car with a further unit transferred from Ilford to give fourteen units for twelve diagrams.

    On the nights of 13/14 and 14/15 July, there were overnight test runs of a 9-car unit to Reading. Test runs to Heathrow then followed for the rest of the next two weeks and, of course, the 9-car operation has now started on one diagram. A 9-car hasn’t been to Reading since those two runs. Nothing appears to suggest that a return to 9-car operation to Reading is imminent.

    There is also 9-car testing to be done on the eastern side. A 9-car unit has been unused at Ilford since February, presumably for this purpose with a further one taken over today.

  18. @GregT – How much later? Only Mark Wild & Tony Meggs still had 6/20 as a performance target. TfL Budget now has 2023 but budgeting for revenue is just a prudence exercise.

    @PoP – Stage 1 Class 345s running between Liverpool St and Shenfield didn’t really advance the project that much.
    Officially it helped Bombardier with mileage accumulation, driver training, and systems checking, and a little marketing to keep public support.

    @IanJ – Trial Running takes about 12 months and can’t start until Bond Street is finished in “the coming months” that gives an earliest possible date of about October 2021.
    All schedules under Covid are likely to take longer so 12+ but can it be “completed” without crush loads and evacuation testing? Maybe mitigation could allow a trial service introduction using load restrictions with staff on-board and on platforms.

    @JONATHANH – helpful notes on whereabouts of 7 car 345s but the riders were not informed.

    Liverpool St lengthening in December 2020 is a significant announcement if that is a definite commitment and not an extract from a shelf plan. The Eversholt leases on the 315 fleet expires then.
    LR article “crossrail-progressing-but-slipping” – If a summer period were the favoured slot for Liverpool St platform works, as they were before, then an autumn or winter stage 4 opening in, say, December 2021, could see peak-time Crossrail services terminating at Liverpool Street continuing to be formed with 7-car trains until summer 2022. They would then be withdrawn and, after the extension works are completed, they would be replaced by 9-car trains.
    Is this scheduling an advancement during the absence of passengers?
    Now that 345/7 reliability is as good as 315s will the fleet be swapped?
    The 345/9 east capability commissioning was paused while the focus moved west, do we anticipate that a restart will wait until a full Reading service is operational to then transfer the experience gained?

  19. Given that the 7 car and 9 car versions are quite fundamentally different, I’m a bit surprised that there wasn’t some attempt to make them easier to describe such as giving one batch a different class no. or calling them 345/1 and 345/2. I’d previously been puzzled why they couldn’t just add a couple of cars to the shorter trains!

  20. @Andrew S

    That is exactly what is being done, and will be done to all of the 345s once Crossrail finally opens fully.

    The difference, as I understand it, is purely in software – the version used on the 7 car versions is stable, but does not support ETCS or CBTC. There were a number of updates to the 9 car software version to enable this, but these changes were never carried over to the 7 car version.

    Therefore the 7 car versions have older, working software, but with limited functionality, whereas the 9 car software version is more complete, but less stable as it is much newer and is still being worked on. This newer software version cannot however be used on the 7 car trains (I imagine there was just little point testing it on the 7 car trains, since they will all eventually be lengthened anyway).

    As far as nuts and bolts go, the 7 car and 9 car 345s are exactly the same train.

  21. @DM1 – the class 755 units for Greater Anglia are classed as 755/3 and 755/4 for 3-car and 4-car units. There are other cases where sub-classes are very similar, so as Andrew asks, it seems odd that the 345’s weren’t split into 345/7 and 345/9 or similar. One difference is that the unit and coach numbers don’t change when the trains move between 7-car and 9-car, but even so, seems a bit odd not to renumber the unit.

  22. Re sub-classes, my understanding is that a temporary change of formation doesn’t trigger a class number change. Within TfL and the project their short hand titles are RLU (Reduced Length Unit) and FLU (Full Length Unit)

  23. In terms of the issues with over-running software projects, a very relevant adage is Hofstadter’s Law, which says:

    It always takes longer than you expect, even when you take into account Hofstadter’s Law.

    Or put differently: Software projects have an inherent tendency to over-run, even when you understand that at the outset.
    To minimise this, strong project discipline and relentless focus on sticking to core scope is essential. The very common “we’re all so clever we can just make it up as we go along” approach always ends in disaster.

  24. 345 naming – the RLU were an initial run of 15 units as a fleet introduction for a temporary phase.
    Their number have now been increased operating different software with unique characteristics on distinct services under a different operating brand for likely some more years yet, a longer period than some defined classes.

  25. The Paddington – Hayes & Harlingtom trains now appear on the National Rail website.

    Further visible improvements to look out for are:
    Removal of all Class 360 trains
    Consistent 9-car service for Hayes & Harlington shuttle
    Consistent 9-car service for Reading service

    Not visible but important is for trains to operate ETCS between Airport Junction and Acton Main Line.

    Maybe the December timetable will extend the Hayes & Harlington shuttle to Heathrow.

  26. @PoP, re Martin Smith’s comment on 7th August, “and Acton Main Line” is still missing from the text.

    [Really don’t know how that happened. I did change it. I have tried a second time. I think the change is visible now. PoP]

  27. What could possibly go wrong?

    “ETCS generally relies on continuous communication and that is done by GSM-R, which is a dedicated radio band on the mobile phone spectrum allocated exclusively for railways”

    The word ‘generally’… How much work is it doing in that sentence?

    GSM jammers exist – they are illegal, and rightly so – but the system needs resilience against this.

    Solar flares and geomagnetic storms exist: and the New York Railway Superstorm is, as the name suggests, en example of a solar storm that affected the railways. I doubt that anyone can plan for a catastrophe like the Carrington Event, but I do worry that the World’s mobile data infrastructure came into existence during a solar minimum, and we will soon be entering a period of increasingly disruptive solar ‘weather’.

    http://www.solarstorms.org/SS1921.html

    Your thoughts? And who’s in a position to say “Yes, some thought went into this, and…”

  28. Nile
    Ah but radio comms are so much bettermore convenient than cables or continuous circuits, because ….
    Cables get stolen, cables get laid in the wrong place or their whereabouts are not noted ( The GWR way! ) & are only “found” the hard way …
    Cables also translate as track circuits, don’t they,? Now being replaced by axle counters which are so much better … or maybe not – because they will not, indeed cannot detect track damage & broken rails
    THIS RAIB report shows one such – if axle counters had been in use, I wonder how much more damage would have been done?
    And radio comms are also cheaper – because the maintenance does not involve physically walking miles of track – which itself carries a safety risk & you have to employ so manypeople to do it, as well.
    P.S. Are you the same “Nile” as on “Charlie’s Blog”?

  29. Oh dear – “Standard” saying opening now delayed to 2022 + £1.1 bn “extra” – C-19 has really not helped, apparently.

  30. Is there chapter and verse on why Bond Street seems to still be in a delayed mess? It’s clearly the issue on the critical path.

  31. Alison
    IIRC, Bond Street is the “Kilsby Tunnel” of this project ( Comparing with the historical parallel of building the London & Birmingham )
    I’m fairly certain that the original contractors handed their keys back, as “too difficult” …. there seems to have been more than one change of immediate management of that site since then, I also think the “usual” problem of untilities being in places where they were not supposed to be turned up, & the site is cramped.
    However, even given all of that, it cannot be as complicated as, say the Liverpool St-& Moorgate -interchange, with multiple existing running tunnels involved.
    I’m not sure that even ploughing through all the various TfL reports would actually uncover a clear picture.

    I would like to echo Alison’s question & ask if anyone has a reasonably clear & concise list of the problems & why this particular site is such a p.i.t.a.

  32. Re Bond Street
    https://metro.co.uk/2019/10/03/fifth-crossrail-worker-dies-sleep-just-six-months-10841276/
    I can’t help wondering if there is a reticence of many trades folk to accept work at this site – which translates into the endless delays on this part of the contract.. I don’t expect any of the parties involved to make any official comment – but I do wonder.
    You can assert that these five deaths are wholly unrelated, not statistically significant, until you’re blue in the face. The “mathematical” side of my brain accepts that argument. But if I was a skilled trade worker, with choice about which job to accept, I know the “emotional non rational” side of my brain would wonder why five (fairly) young and fit workers have died without plausible explanation (not one in the public domain at least), and I wouldn’t take a contract there.

  33. Bond St – ‘Critical Path’ : Xrail board minutes from 2 July state that Crossrail Ltd has implemented a “90-day plan” to get outstanding work at Bond Street station complete that consists of the installation of “residual systems and electrical work”.

    Crossrail Ltd’s plan is to tackle a backlog of work during a five-week blockade of the railway beginning at the start of August until the middle of September, with the option to extend for a further two weeks. Crossrail Ltd revealed that there are “450 hazards that need to be addressed by the construction blockade” before the project can move towards trial running. Crossrail added “The construction blockade in the routeway is progressing well and we are achieving our targeted levels of productivity and planned activities with just under two weeks of the blockade remaining.”

    Crossrail’s sponsors are concerned about a “potential bow wave” of activity under the revised delivery plan.

  34. Class 345s working all three TfL Rail diagrams to Heathrow from today but it appears that the Hayes shuttles may have reverted to 7-car operation to allow this to happen. My guess would be that this is to concentrate the available technical support on the Heathrow services since there are more than enough 9-car units in service to cover both but would be interested to hear the actual reason.

  35. @ALISONW 21 August 2020 at 13:50
    Is there chapter and verse on why Bond Street seems to still be in a delayed mess? It’s clearly the issue on the critical path.

    &@GREG T 22 August 2020 at 15:34
    I’m fairly certain that the original contractors handed their keys back, as “too difficult”
    I’m not sure that even ploughing through all the various TfL reports would actually uncover a clear picture.
    I would like to echo Alison’s question & ask if anyone has a reasonably clear & concise list of the problems & why this particular site is such a p.i.t.a.

    A one-off £19M payment was made to the Costain Skanska joint venture (CSJV) on 26 June, the day after Crossrail Ltd announced that the two parties had agreed to conclude the existing contract. Crossrail said: “The details pertaining to this matter are confidential and commercially sensitive, therefore Crossrail Ltd will not be able to provide further comment.”
    The termination of the Bond Street contract has had no bearing on CSJV’s work at Paddington station, where it is also the main contractor.

    At the All-Party Parliamentary Rail Group consisting of MPs and Lords from across parliament, Chairman Tony Meggs admitted that : “There are three issues on Bond Street.
    One is that it was late starting because all of the tunnelling that allowed it to begin was itself late.
    Secondly, for some reason it is a very complex design. For technical reasons there is complexity in the station and the design of the station.
    Thirdly, and to be frank, less than perfect contractor performance on that station.

    The way this railway was designed was that nothing could open until everything was open. We have redesigned that such that we can open the central section without Bond Street if we need to, if it is not ready in time. Indeed, that is also true for other stations as well.”

    In the same session, Crossrail deputy chair Nick Raynsford admitted it was “not feasible” that Bond Street station will open in the first half of 2022.

    Outlining lessons learned on the job Meggs said the original fixed deadline by the project sponsors created systemic problems which led to a failure to finish the project in time. As things slipped a bit the immovable end date meant that there was a huge amount of compression which got built into the schedule. Every year the schedule was revised but the end date was never moved, and as a consequence, two or three years’ worth of work ultimately got squished into a few months and of course that ultimately led to a failure to deliver on the date and, frankly, caused a huge loss of confidence in our ability to execute these large programmes.”

    Meggs said that 35% of Crossrail’s workforce had been demobilised after the 2018 deadline was missed, meaning he and deputy chairman Nick Raynsford spent a “good part” of last year rebuilding parts of the team which managed finance and project control.

    Crossrail, while undeniably a complex project, “didn’t need to be that complicated”. Future infrastructure projects should bear in mind the mantra of simplification, standardisation and minimum viable product. At the very beginning of the project think hard of a very systematic process of standardisation, more use of offsite manufacturing and construction and really attempting to simplify the structure of the project, the technology of the project, the number of interfaces on the project and the number of inter-dependencies.

  36. Full 9-car operation restored on the west side of London this morning with a similar swap of 7-car for 9-car units for Reading services last night to that which took place in the opposite direction on evening of 11 June.

    This follows the introduction of new software on the 9-car units at the start of December, initially on Heathrow services and then on Hayes services from 8 December.

  37. According to Jacobs it had not been able to estimate an opening window for Stage 3 until the commencement of testing. Announced today (Mon 10th May 2021) trial running of services has commenced on the future Elizabeth Line in a move that sees four trains per hour running between Abbey Wood and Paddington. Trial running will initially be undertaken from Monday to Thursday, with frequencies gradually increased as preparations for opening ramp up.
    The following stage will be Trial Operations inviting passengers onto trains to test various systems in operation.

    The 12 month program had expected construction works to have been completed. That and Covid disruption may now not be as high as looked possible earlier this year so June 2022 is still possible.

  38. Next week’s unredacted Board report on situation.

    https://board.tfl.gov.uk/documents/s17050/board-20211208-item06-crossrail-update.pdf

    Canary Wharf could be done by year end.
    Bond Street by mid 2022
    Romford & Ilford by a possible March Elizabeth start.
    W/e up lifts at Brentwood and Goodmayes are deferred. One is waiting for a shop lease expiry and the other will be subject to rebuilding for a Goods Yard redevelopment. The interim will need a journey reversal on the down line,

    The Jacobs view was that a joined operation in 2022 required early completion of Trial Ops and possible reduced demand timetable.

Comments are closed.