Crossrail: Progressing but slipping


News on progress of existing TfL schemes has been in short supply during most of 2019. As if to make up for this, the 23rd October meeting of the Programmes & Investment Committee provided not only comprehensive background papers on a range of subjects but also a very informative discussion of the issues raised.

We hope to try and cover most of the major areas where we have updates of what is happening and, to that end, we kick off with Crossrail using information gained from the above committee and other sources to establish where we are today. The main source of information was Mark Wild’s (chief executive at Crossrail) presentation to the board members of the above committee.

Accepting the reality of today

Various threads weave their way throughout the current Crossrail planning and project assurance process. The first notable thread is that there is very much a ‘cold light of day’, ‘this is where we are now’ approach which is a welcome contrast to the optimistic ‘only give the good news’ approach of two to three years ago. It is clear that the TfL board members on the committee have confidence in the current Crossrail board – not only to complete the job but also report back the truth and not try to hide anything. Obviously there are some things that, because of commercial sensitivity or other reasons, cannot be put in the public domain but nowadays a lot of effort all round appears to be made to ensure that this is kept to the minimum necessary.

Another really noticeable change in attitude is that focus has moved away from opening as much of Crossrail as possible as quickly as possible to an approach that would maximise revenue and minimise costs. Consequently, various stages are considered more important than others and the order of importance is in marked contrast to the ranking that would have been given just two years ago.

The income-based plan

Currently, in terms of enhanced journey opportunities, very little has been achieved to date compared to what was present before Crossrail. The Liverpool St – Shenfield service now has nice new shiny trains (sometimes) but the timetable is basically unaltered. The Heathrow service continues much as before using the same rolling stock. Improvements amount to a 2tph service on Sundays (it was just 1tph) and a new half-hourly service between Paddington and Hayes & Harlington. Even that latter service was, strictly speaking, just taking over a recently-introduced electric service provided by GWR.

In terms of revenue it is hard to see how any such progress as there has been in implementing TfL Rail services, prior to the Elizabeth line opening, has provided any significant new sources of revenue (as opposed to reallocating existing sources).

Getting to Reading is really important

One of the surprises to outsiders when the delay to Crossrail was announced must have been the determination to press forward with introducing Crossrail services to Reading in December 2019. This was originally to be done as part of the stage that was due to be implemented last. To insiders this was no surprise. The line to Reading, which is outside the London fares zone, is the only part of the future Elizabeth line expected to generate significant amounts of revenue from each passenger journey.

Much effort has been expended during 2019 to ensure that TfL runs a Paddington – Reading service on commencement of the December 2019 timetable. With the tunnel section of Crossrail not yet available, this means continuing to utilise platforms at Paddington main line station that were originally expected to be freed up by the full opening of Crossrail.

Providing a Reading to Paddington service (now designated phase 5a) on time but out of sequence appears to provide a win all round. It brings in revenue to TfL, it enables existing drivers to be used productively and gain experience, it shows progress is being made and it enables the operating team to gain more real world experience.

You might have thought that TfL’s gain in going to Reading would be GWR’s loss but this appears not to be the case. It turns out that it is actually quite difficult for GWR to efficiently roster their stopping services to London and it isn’t the moneymaker one might expect it to be. One suspects this is at least partly because of where the stock is stabled and where GWR drivers’ depots are located. A further factor could be that TfL has more opportunity to interwork the stock with other services (initially Crossrail to Heathrow) which should lead to productivity gains. Apart from any financial benefit, one suspects GWR would rather concentrate their management resources on the potentially-far-more-profitable long- distance services to Wales and the West Country.

What is quite concerning is just how tight the schedule has been to get the railway ready for Class 345 (Crossrail) trains to Reading. Network Rail have so far kept their promise in getting all the necessary platform CCTV monitors installed and working. However, as is often the case, Network Rail have cut their works schedule rather fine. Indeed Twyford and Maidenhead (not lightly-used stations) have yet to have CCTV works completed although they are not behind schedule. This uncompleted state so close to the timetable implementation date has raised concerns by at least one outside review body.

Getting 9-car trains in use

Another concern in going to Reading is that the ideal would be to use 9-car Crossrail trains rather than the 7-car ones that are generally currently used. The difference between a 7-car and 9-car train would appear to be two cars but in the case of Crossrail the real difference is found in the train-based software. 7-car trains use software which is old, of limited functionality and not intended to be enhanced. 9-car trains contain the next generation of train-based software which has yet to be extensively used – at least on a Class 345.

Things aren’t going well with the introduction of 9-car sets. They do run, at times, between Hayes & Harlington and Paddington but their mean time to failure is reported to be around 500 miles at present. Generally one expects the figure to be at least 20,000 miles and even more with modern debugged stock with London Underground’s ‘S’ stock believed to be around 60,000 miles. Given the current unreliability of 9-car trains, it is no surprise that it is almost inevitable that the initial TfL service to Reading will be formed using 7-car trains.

Getting to Heathrow – properly

One of the bits of good news was that Mark was confident that TfL could be running Class 345 trains to Heathrow Terminal 4 in “the early part” of 2020 now that issues with the signalling have been identified as solvable – if not already solved. It seems that here, and in some other places too, the biggest problem now is to get sufficient railway possessions (exclusive use of the track) in order to either test the system or install hardware.

What was not mentioned at the committee meeting was whether Heathrow would be initially reached by 7-car trains or 9-car trains. Either way it would mean a 15 minute interval service to Heathrow, being able to finally get rid of the old Heathrow Connect class 360s (which are problematic and only 5-cars long), having a far more marketable product and further opportunities for operational experience.

Opening the core section – who cares?

Taking the theme of being concerned with revenue, it now seems that TfL ( or at least parts of TfL) and Mark Wild, seem rather less concerned with opening stage 3 (Abbey Wood – Paddington) than has been the case in the past.

The driving force behind this reduction in importance of opening stage 3 (Paddington to Abbey Wood) is partly down to revenue. It has long been argued that most Elizabeth line revenue from stage 3 opening would simply be abstraction. That is, it is revenue TfL already gets – typically on the Underground – and the financial effect would be to transfer that revenue to a different column in a financial report.

A note of caution was made by one of the project sponsors on the idea that stage 3 was not important from a revenue perspective. He agreed with the short-term prognosis but argued that in the long term it made sense to open as soon as practical. He pointed out that the TfL staff needed for Crossrail to open were, for the most part, already employed. He then further pointed out that, until assets such as stations were handed over, contractors would still have to have their own staff in the building. The implication was that ultimately it would be TfL paying for these staff. Presumably, once TfL staff are in place, it would make sense to open the line as soon as practical. Although, initially, revenue would largely be abstraction, traffic would gradually build up so the sooner that process was initiated the better.

Not wanting an unreliable railway…

In addition to that, in total contrast to the previous Crossrail regime, Mark Wild does not want to open stage 3 until it is believed to be solid and dependable. Or to put it another way, you don’t open it if you know you have outstanding issues. Part of this philosophy is down to the fact that once open it is more difficult to fix things on a live railway than it is on a not-yet-opened railway. As stated by Mark, “Opening Canary Wharf to Abbey Wood would be a cheap shot.”

A factor that may have influenced this de-emphasising the importance of stage 3 is the on- going delay at Bond Street station where it seems that everything that could go wrong has gone wrong. At present a 12tph service between Paddington and Abbey Wood could be run if the only thing holding it up was Bond Street. But, even then, Bond Street must be available for emergency evacuation. However, it has been deemed that it would not be possible to run 24tph without Bond Street being fully open because it wouldn’t be able to safely handle an emergency evacuation from the tunnels.

The argument goes, if you can’t open Bond Street, why open stage 3 if you won’t be able to follow on with stage 4 a few months later? You could argue opening stage 3 will enable the service to settle down and give confidence one is ready for a stage 4 opening but equally it could be argued that without Bond Street open such reassurance is of limited value.

.. but a ‘metronomic one’

Using a delightful phrase, Mark stated that ‘before going to stages 4 and 5 it was fundamental to have a “strong, stable, metronomic 12 trains per hour”. So clearly the current chief executive does not intend to be rushed if things are not right. On this he appeared to have the backing of board members. Certainly there was no dissent from the committee.

Stage 4 – the first critical stage

Stage 4 see most trains from east of Stratford diverted into the Crossrail core rather than terminating at Liverpool St (high level). This stage was stated as being significant for revenue growth. This is probably largely due to new journeys (especially off-peak) between east London and the centre of London such as the West End. It is also a stage that really has to align with a Network timetable change as other trains (especially freight trains) interwork with the current TfL Rail trains to Shenfield. As well as that, the frequency and service pattern east of Stratford have to align with the service pattern in the Crossrail core. So this means calling at all stations at all times of day (unlike at present) and the trains, for the most part, evenly spaced out (also not the case at present).

Because of the restrictions of when stage 4 can open and the need not to miss its target date, Mark Wild is anxious to have a stable stage 3 before even considering this – hence the importance of not opening stage 3 unless he is confident they will soon be ready for stage 4. It might be notable to point out that at around this time last year Mark expressed concern in the original Crossrail plan as to how quickly stage 4 was expected to open after stage 3 – not enough time for the initial service to settle down in his opinion.

The original plan intended to have a stage 3 opening in December 2018 and stage 4 opening in May 2019 whereas Mark suggested the time between these two events should not be less than six months. It will be interesting to see what pressures there are to open stage 4 once stage 3 is eventually opened.

Stage 5b – fully open

Stage 5b is the full opening of Crossrail incorporating through running from the central section onto the Great Western Main Line for services destined for Heathrow, Maidenhead or Reading. This is the second of the two yet-to-be-finalised critical stages in terms of both revenue and aligning it with a Network Rail timetable change – so basically December or May with a possibility of September if really necessary. Stage 5b would also be expected to provide genuine revenue growth. Through services from Heathrow to Central London and Canary Wharf will generate new rail journeys and through services to London from places such as Slough would also generate new journey patterns. Of less significance, but not to be forgotten, is the reverse situation. This could mean tourists based in London being more inclined to visit Windsor or travellers from London to the West Country or Wales choosing to change at Reading rather than Paddington.

Liverpool St platforms not mentioned

Mark did not elaborate on when Network Rail’s works to re-configure the east-side Liverpool Street terminal platforms for longer trains would be taken forward. The current length of the platforms restrict Crossrail trains serving Liverpool St to a maximum length of seven carriages (160 metres or eight ‘conventional’ carriages). What is desired is the ability to operate nine carriages (205 metres, or slightly longer than ten ‘conventional’ carriages). Without this platform lengthening 9-car Crossrail trains wouldn’t be able to use Liverpool St (high level) in an emergency or during engineering work. Furthermore, the few residual peak-period services still intended to terminate at the existing Liverpool St station would have to remain 7-car and not be interworked with the rest of the fleet.

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.

Project Key Areas

The relative importance of different stages of opening means nothing unless you can actually get the railway working. Mark Wild identified three very different key areas that needed to be tackled.

The first of these was, not surprisingly, the physical infrastructure and to this end we are talking about ‘must have’ rather than ‘nice to have’ so issues such as the rebuilding of Network Rail stations or accessibility, important though they are, do not get factored in at this stage.

The second key area was software. In particular, this is generally signalling software but also refers to train-based software which may have a role in many additional functions such as Selective Door Operation. It is indicative of the modern world that for so many projects (not just railway ones) it is software that becomes a major area of concern when trying to complete a project whereas thirty to forty years ago, a software element would not even exist. Before being too quick to blame software, it is worth remembering that many of the projects would not be so beneficial (once they are working properly) without the software element and in some cases these projects simply could not have been implemented without modern software.

The third key area was assurance and handover. Mark emphasised that this wasn’t just a case of ‘getting the paperwork done’. It was a vital component of the project. Procedures had to be in place and tested. Documentation had to be approved by regulators and approval couldn’t be taken for granted and without approval the line could not open.

We shall look at these three key areas in detail.

Physical Infrastructure delays

It will come as no surprise to anyone who has closely followed the Crossrail saga that the big two problems with stations were at Bond St and Whitechapel. Whitechapel is probably less talked about because it will probably be ready before Bond St so is not critical. This assumes that Whitechapel can at least be used for evacuation though, to date, there hasn’t been any suggestion that Whitechapel would not be fully open in time for a stage 3 opening.

There was some good news. GSM-R (the radio system) is now working in the tunnels. It was probably GSM-R that Mark was thinking of when, a year ago, he referred to the tunnels still not being fully kitted out.

Even Farringdon, a station where progress has generally been good, is only now nearly complete and has reached the status of ‘PPE free’ which, in very simple terms, means that workers did not have to walk around in a hi-viz orange jacket whilst wearing steel-capped boots and a hard hat. Amongst the things still to complete here (but not specifically mentioned) are the hostile vehicle mitigation measures (bollards to you and me) to be installed outside the station. Incredibly this is only now being done and it has turned out to be a surprisingly complex job involving advance diversion of utilities which has taken longer than originally expected. It is hard to see how the station could have opened without these in today’s terrorist-aware climate, yet somehow the original Crossrail team believed in the summer of 2018 that this station was still on track to open in December 2018 despite work on the bollards not even having been started.

Much publicity has been given in the press to the inadequacies of the fire alarms which have had to be replaced as not compliant with railway standards. Mark stated this came as a surprise and could not have been discovered before testing. Here at LR Towers, we can’t quite accept that as it is believed the very same problem was one of the factors that delayed the opening of the Walbrook entrance of Bank station which finally opened at the end of November 2018 when it should have opened that summer. So there were at least some warning signs.

The fire alarm issue was presented as just something else on the list that had to be sorted out and Mark state that ‘fire work’ should be complete by November 2019. However it was not made clear what ‘fire work’ consists of. We are aware that as a consequence of Grenfell Tower a lot of panels have had to be removed and ‘fire work’ done behind them before they are refitted. It probably won’t surprise the reader to know that this was apparently a particular issue at Bond Street and is probably what has led to past stories of contractors calling Crossrail ‘the Hokey Cokey line’ because you put the panels in, then you take the panels out.

Mark stated that construction output was “broadly in line” with the plan. This was reassuring but this reassurance was somewhat shattered when the words “at a P50 level” were added to the sentence. This means there is a 50% chance of being broadly in line with the plan. Having at least an 80% confidence level (P80) would be much more reassuring. Given the history of Crossrail, one is inclined to get nervous when a programme of works is reported to be on schedule but the confidence level in the reporting is effectively as low as it you can get and still report it in a positive way. Any lower and you are saying it is more likely than not that the target will not be met.

Software Delays

There were many failings of the original Crossrail team but probably their biggest one was not comprehending at all just how difficult and time-consuming software validation and integration is. It would be not too much of an exaggeration to suggest that they were allowing a number of months for some things that they should have been allowing a number of years. To a very limited extent this was understandable in the early days of the project but, as the software complexity became apparent during the lifetime of the project ,the timescales should have been adjusted accordingly.

One of Terry Morgan’s (previous Crossrail Chairman) defences to Crossrail’s failure to deliver was that TfL procured the trains so the critical software on the trains that was causing a major problem was TfL’s problem not one for Crossrail Ltd. Thankfully this doesn’t wash with Mark Wild who has previously stated that as systems integrator it is his job to integrate the systems and that includes getting individual systems working regardless of where they came from or where they are located.

A critical point of the project is to get to the point where trial running can be tested in the tunnels. This means running trains to the proposed timetable. Unfortunately, whilst a lot of multiple running has taken place successfully it has identified software problems that mean moving up to trial running, which is the next phase, is not yet possible.

At present Crossrail is on version 10 of the critical software and it is really hoped that version 11 will allow trial running as version 10 is so nearly there. At one stage it was expected that version 8 would be sufficient for trial running so there is quite a bit of delay there. The elapsed time between software versions is relatively long as the software has to be amended and thoroughly retested.

Assurance and Handover

When referencing assurance and handover, Mark emphasised the scale of the job. “Undoubtedly the largest handover outside of China”. Of 30 elements to the process, according to the planned timeline six or seven should have been handed over by now but “only two were handed over so far”.

What’s the timescale ?

Of course, what everyone wants to know is what the latest timescale is and is it being met. Here the news was slightly disappointing but not unexpected. Because of the delay to trial running, it is becoming more challenging to meet the planned opening window of October 2020 – March 2021. The software delays moved it towards February – March 2021. What Mark actually said was “February to May 2021” but in the context of what was being said we presumed he meant to say March not May. He admitted that, if version 11 of the software had a problem that meant trial running was still not possible, then obviously that would make the situation even more challenging.

It doesn’t really matter which phase of the project (other than the first almost trivial one) you look at. It seems that the expected delay is around two years and, in most cases probably slightly longer.

Possible future dates

If we assume that Crossrail succeeds in opening stage 3 in March 2021 then, because of the need to align with Network Rail’s timetable, it would not really be possible to open stage 4 before September 2021. Quite possibly December 2021 is more realistic due to the need to be absolutely confident, in advance of a major national timetable change, that the proposed changes can be implemented. Aside from the repetition of an embarrassing scenario, such as happened in the May 2018 timetable change, Network Rail’s procedures for approval of timetable changes are now far more rigorous than they used to be and rail operators need to be able to demonstrate that implementation of changes will not cause problems.

Even if stage 4 opens in September 2021 then through running to both Heathrow and Reading by December 2021, as currently expected, is going to be pretty challenging and would break Mark Wild’s self-imposed rule about having a minimum of six months between each opening phase. So, it is starting to appear as if a full opening of Crossrail in December 2021 is unlikely with May 2022 as a more realistic earliest possible date and later in the year being more likely.

Where are we now?

One left the meeting not feeling confident that targets and dates would be met but that the project is in a much better organised state than it was before. Although unpleasant unexpected surprises cannot be totally eliminated, there no longer seem to be fundamental problems that may hold the project up a further serious extent. To put it another way, it now all seems doable. The only real questions about dates are how close to March 2021 for stage 3 and December 2021 for stage 5 can the project really get? It might be better for planning purposes if full through running was not envisaged until 2022. Somehow we can’t yet see the Queen’s diary secretary getting ready to put a date in Her Majesty’s diary for the ceremonial opening of stage 3.

Many thanks to Jonathan Roberts. This article was based on his original notes and he has followed up with advice, suggestions and corrections.

Written by Pedantic of Purley