One of the Family: Crossrail’s Transition to Being a Tube

386 comments

In June 2015 we reported that Crossrail was well on the path to becoming a functioning railway in the making rather than a construction project nearing completion. The man whose job it is to create this operational railway is Howard Smith, Operations Director at Crossrail. He was brought in early to ensure that the high standards achieved in construction were reflected in the railway operation. As one of the latest TfL Investment and Programmes Committee papers shows, Smith has clearly been busy.

Back to the basics

Anyone who has had a railway-based conversation with Smith, or who has been to any of his talks, will know that he strives to get a few basics right. He is a strong advocate of ‘turn up and go’ rather than expecting people to remember timetables, as befits a graduate of the London Overground leadership team – the Prussian Military Academy of modern metro services.

Smith has spoken before about the need to keep things simple for the public. He is also not a fan of off-peak services. At talks, he has indicated that in his ideal world operators would provide a fixed interval service (eg quarter-hourly) from the start of the day until the last train. In his eyes, the marginal cost of running a frequent service out of peak hours is not that great and can be justified by the social benefits and increased revenue brought about by a more frequent service.

Of course, anyone can have ideas, but often they are not practical. In Smith’s case, however, evidence of the practicality of his beliefs can be found by looking at London Overground and the East London Line in particular. As it is on the ELL, so it is here and over time it has become clear that he has been working hard behind the scenes to achieve many of the same ideas with the Elizabeth line (without doing anything to jeopardise service reliability).

The timetable is the key

Having the right timetable can make a huge difference to a service in terms of reliability, attractiveness and cost-effectiveness. There are probably many railwaymen who believe you cannot spend too much time getting the timetable right, and the evidence so far seems to be that Smith may well agree with that sentiment.

To use the well-worn adage, sorting out the timetable is a marathon, not a sprint. It is a highly iterative process. To begin with, an initial timetable is produced. This may not be actually be used in anger, but just to show proof of concept. It may be to prove that a particular track layout is operationally workable, for example. As time goes on conditions change, knowledge increases and objectives are modified. In line with this, the timetable also changes. Each time the aim is generally to produce something a bit better and, in particular, to try and eliminate some of the weaknesses of the previous iteration.

The key that unlocks the service to Reading

The first variation on the Elizabeth line operational timetable was necessitated by the agreement to extend it to Reading. Be in no doubt: this did not come about because key decision-makers thought it would be a good idea that would benefit passengers. It happened because there was an issue with the original pre-Reading-electrification proposal for a Great Western diesel shuttle between Reading and Slough. By extending both the overhead lines and the Elizabeth line to Reading this costly shuttle could be eliminated.

The Elizabeth line service beyond Maidenhead to Reading, introduced to obviate the need for the Reading-Slough shuttle, would be paltry – 2tph (trains per hour) throughout the day. No doubt the policy at the time was to get this proposal through by producing an unarguable if unadventurous case. Any improvements could then come later without jeopardising this.

The off-peak weakness

At the outset, it was highly publicised that the Elizabeth line would have a peak service of 24tph with only 10tph going west of Paddington. The off-peak service was hardly mentioned but, slightly implausibly, it looked like there would be 6tph on the Shenfield branch, 8tph on the Abbey Wood branch and only 8tph west of Paddington.

It did not escape the notice of LR Towers that as long as the off-peak service to Shenfield remained at 6tph then any off-peak service provided by the Elizabeth line in central London was either going to be an embarrassingly low 12tph or a rather messy service to run. Clearly, Smith and his team were well ahead of us and had already set the wheels in motion to improve the frequency on the Shenfield branch in the off-peak. This is not so much because the Shenfield branch actually needs a more frequent service, but simply because this determined the maximum frequency that could be run in the central London core. Also highly pertinent was the fact that the Abbey Wood branch would have the same frequency as the Shenfield branch and hence the service provided at the important station at Canary Wharf was dependent on it.

Increasing the off-peak frequency on the Shenfield branch was not simply a matter of just doing it. Crossrail does not ‘own’ the line and there are occasional freight services that cross those tracks via a flat junction. Part of the line is sometimes used by c2c – either for diverted trains or their regular weekend service via Stratford. As such, a case has to be made to the rail regulator if you wish for a guaranteed long-term right to run an enhanced service. This was quietly done and it was established that the Elizabeth line could run 8tph off-peak on the Shenfield line.

What a relief (line)

At some point last year a table appeared on the Crossrail website from which one could deduce the proposed use of the relief lines out of Paddington now that Crossrail was going to Reading. For the uninitiated, most sections of railway, if four-track, have fast and slow lines. The Great Western Railway (GWR), with its history of being different, decided to call their fast lines the “main” and the slow lines the “relief”. To some extent, the terminology is more appropriate as is seems strange to talk about a 100mph speed restriction on the slow lines.


Table of frequency of all Heathrow Express trains and all passenger trains using GW relief line

The table, which is well-hidden on the website, is dated March 2016 but was probably not made available until August of that year. It does seem to be slightly misleading as it appears to only reference consistent service patterns. Not apparent are several one-off services that under existing proposals will also use the relief lines, such as the through train from the Henley-On-Thames branch. These do not fit into any pattern.

What is quickly apparent is that, as far as passenger trains are concerned, the plan at the time was for the relief lines out of Paddington to only be used by:

  • Elizabeth line services
  • A 2tph GWR services semi-fast service from Reading to Paddington – which might conceivably originate from further away e.g. Oxford

    and, possibly

  • 4tph Heathrow Express services

In the context of looking at passenger services running on the Great Western relief line, Heathrow Express is a peculiar case. The trains run non-stop from Heathrow (terminals 2 & 3) and Paddington main line terminus. They run in airport-owned tunnels to get out of Heathrow then join the Great Western lines at Airport Junction. Until recently this junction was only a connection with the main line, but now it is also connected to the relief lines as well.

Heathrow Express Ltd would far rather these services ran on the main line – not the relief – as they do today – but as the trains have a maximum speed of 100mph this would seriously reduce future capacity on the main line. They currently “get away with it” because Crossrail is still a project not an operational railway and the lack of top speed is partially offset by the quicker acceleration that the Heathrow Express electric trains provide when compared to the Great Western Railway diesel sets. This will, of course, all change once the dual-mode IEP trains come into service on Great Western territory.

Moving Heathrow Express onto the relief lines, as Network Rail intimated a few years ago that they would like to do, is, unsurprisingly, not ideal either. 100mph non-stop trains and 90mph stopping trains on the same stretch of track are not a good mix and, inevitably, lead to all sorts of timetabling problems.

In the end it seems that the contractual agreement with Heathrow Express just could not be honoured if Heathrow Express was shifted onto the relief lines and so, for the short term future at least, it will continue to use the main lines.

One doesn’t need to be a genius to think that, rather than have a 2tph semi-fast Great Western service between Paddington and Reading and a 2tph (nearly) all stations Elizabeth line service to Reading, it might be a good idea to combine the services in a co-ordinated way and have a 4tph Elizabeth line service to Reading all day. Furthermore, a precedent has already almost been set for this through the decision to abandon the proposed Reading-Slough shuttles and extend Elizabeth services to Reading instead.

The timetabled frequencies that probably won’t be

Alongside all this, TfL released a highly technical consultation about third party rail operators access to the Central Operating Section of Crossrail. It was a legal formality as much as anything else, but as part of the argument to show why sharing the track with other operators would be impractical, a diagram of the latest iteration of the proposed service was shown.


Frequency diagram from page 8 of Crossrail Central Operating Section consultation

Up until then, Crossrail had been very reluctant to provide specifics of the service that would run when the Elizabeth line was fully open. This diagram thus generated a fair amount of interest.

It is clear that the service offered is not sufficient at this stage to satisfy the aspirations of TfL in general and Howard Smith in particular. Specifically:

  • 2tph to Reading is hardly ‘turn up and go’
  • Terminating trains at Maidenhead do not call at Taplow and Burnham in the off-peak period, but do in the peaks (hardly simple)
  • Not all trains call at Hanwell, West Ealing and Acton Main Line

The last point has two unsatisfactory features. The first is that a passenger cannot be sure when they get on the train that it will stop at their station. This could either be to minimise rolling stock needs or simply be down to an operational need to help co-existence of trains with a mixture of stopping patterns. Two trains per hour in the peak period do not stop at any of these three stations. All other trains stop at one or two of the stations in question but not all three.

The second unsatisfactory feature is that this is known as ‘skip stopping’, whereby alternate trains have a different stopping pattern to maximise usage of the line. This leads to a journey between adjacent stations (Hanwell and West Ealing in this case) not being possible without changing trains.

There is an indication that the frequency diagram is a work in progress as the key includes “stations [at which] only off-peak trains call” – yet there are none on the diagram.

Latest proposals

There are less than two years to go until the Elizabeth line is due to open. It is generally reckoned that the simplest timetable change takes about a year to fully go through the necessary processes. It is crucial to know the planned service for the line when fully in operation in the near future as this could possibly impact the day one service. For example, one does not want to offer a service that cannot be sustained upon full opening of the line.

With this timetable deadline looming, it was a pleasant surprise to see that Crossrail have put forward a paper to the Programmes & Investments committee entitled “Elizabeth Line – Increasing Service Frequency”.


Outline of latest proposed Elizabeth line frequencies

There are three basic improvements proposed.

A better off-peak service

Off-Peak Frequency is to be 20tph between Paddington and Whitechapel and 10tph on each of the Shenfield and Abbey Wood branches.

This seems to be an eminently sensible thing to do and the reasoning behind it is stated in the paper:

5.1 The overall business case for the combined Off Peak 20tph and Peak 24tph recast service pattern demonstrates an exceptionally strong case as it makes better use of the £15bn investment that has been made in the Crossrail project.

The impression given is that Crossrail already have approval from the Rail Regulator for 10tph off-peak on the Shenfield branch. The only problem is that it is not in their specification for the service they have to provide, so Crossrail have to ask the TfL Board (the project sponsor) to mandate that they run 20tph in the off-peak – otherwise they have no authority to do so. This may all seem to be a touch of the ‘Sir Humphreys’ but these things must be done by the book.

To the travelling public, this will be good news. In particular, it could be argued that off-peak frequencies have reached a tipping point where you don’t have to think about them. It will also make the frequency of the Elizabeth line seem typical of a deep-level Tube line, as 24tph peak/20tph off-peak is approximately what the Piccadilly, Bakerloo and Northern line (each central section) operate Mondays to Fridays.

No slow or semi-fast GWR services to Paddington

The second change, which should be no great surprise to anyone who has studied the previously-mentioned Reading-Paddington service interval table, is that TfL are proposing to take over the semi-fast Reading-Paddington 2tph service as well as the 2tph (almost) all stations service which is already part of the plan. This has all sorts of consequences, though. Again, the way the paper reads, this sounds like a done deal – presumably with the DfT. All that is needed, apparently, is for the TfL Board to approve the purchase of four new Crossrail trains to cover the service.

The advertised (almost) all stations running time from Paddington to Reading on the Elizabeth line will be 49 minutes. Any Elizabeth line semi-fast services should do this considerably quicker. Just cutting this down to 45 minutes should mean that only three trains ought to be needed to include this service as part of the Elizabeth line pattern. In fact, the proposal is to purchase four extra trains.

The paper notes that no new stabling area needs to be provided which, presumably, is because provision was made at the outset for more trains than have currently been ordered. The cost of the four trains is not recorded and the need for confidentiality means that it will not be openly discussed at the Programmes & Investment Committee meeting itself. There is also no mention of any payment adjustment as a result of the Great Western franchise losing a revenue stream and that may well also be down to confidentiality.

A more regular service pattern west of Paddington

The paper further notes that a few individual Great Western services that currently use the relief lines in the peak period would need to be withdrawn.

On the Great Western Route, the enhanced Elizabeth line Peak service results in a need to remove five Great Western Franchise services (in both morning and evening Peak) that are specified to operate throughout the Peak on the relief lines between Paddington and Reading with a semi-fast stopping pattern. Great Western Franchise services to Maidenhead, Twyford, Reading and Thames Valley stations will continue to be provided during the Peak period by other trains which operate over the main lines between Paddington and Maidenhead

So Henley-on-Thames and Bourne End (on the Marlow branch) will lose their direct trains and three other semi-fast services that were due to be maintained would now be withdrawn under these proposals. It would appear that this is why the fourth train is needed. The idea is to provide an enhanced stopping pattern to cater for these trains that have been withdrawn from the relief lines. As Crossrail trains will be much longer, it seems that the intention is to realise the dream of providing a better service with fewer trains.

As a consequence, all remaining Great Western services to Paddington will run on the main line – which is probably why Network Rail were so keen to try to move Heathrow Express over to the relief lines.

Subtle changes

If this proposal goes forward then:

  • No Great Western services will run on the Great Western relief lines into Paddington.
  • Great Western services on branch lines into Twyford and Maidenhead will, for the most part at any rate, just be Elizabeth line feeders.
  • At the West Ealing end, the West Ealing – Greenford shuttle will exclusively feed into the Elizabeth line.
  • Under normal (non-engineering work) circumstances, only Elizabeth line trains will call at on the relief line platforms at all stations between Twyford and Acton Main Line.
  • The Elizabeth line will have exclusive use of the track on one of its main branches (Abbey Wood) and almost exclusive use of a second (Shenfield) and third branch (West of Paddington) it will share passenger services with a non-stop airport service that is generally punctual and reliable. The early fears of performance pollution, the odd freight train excepted, seem to have largely been eliminated.
  • The Elizabeth line (née Crossrail) is starting to look a bit like a self-contained railway. In doing so Crossrail 2 is starting to look like a different railway altogether and not really “like Crossrail 1 but bigger”.

If these proposals were to be implemented the Great Western relief line to Reading would have a real Elizabeth line feel to it. The service will be better than to the outer reaches of the Metropolitan line, better even than parts of the Central line. It remains to be seen to what extent TfL will be in control of fares policy over their new domain.

An obvious suggestion is to go one step further and hand over the relief line tracks to Crossrail. This will not happen because they are used by freight and are needed for planned engineering work on the main lines when all traffic (generally on a Sunday) is diverted to the relief lines. Besides, the fact that there are no other users now does not mean that there won’t be in future. It is entirely possible that the Western Rail Access To Heathrow project may lead to another rail company using them.

Add the future of Heathrow Express to the mix

It is also interesting to speculate what will happen in 2023 when the Heathrow Express agreement becomes up for renewal. Heathrow Express will still own the Heathrow rail tunnel, by then used by 4tph on the Elizabeth line, so they will have some bargaining power. Nevertheless, one suspects that there would be a reluctance to let them continue to run onto the main line if it turns out it is possible to accommodate them on the relief lines.

Another possibility is that by 2023 Heathrow Express gets subsumed into the Elizabeth line itself. Perhaps, by introducing intermediate stops, the railway becomes less of a mixed traffic railway and capacity increases as a result. And, maybe, the digital railway will enable an extra 2tph to operate on the relief lines. It is even possible that Kensington & Chelsea get their proposed station at Kensal and that leads to more trains continuing beyond Paddington.

If enhancements to the Elizabeth line service west of Paddington do take place, before long those 14tph originally planned to terminate at Paddington will seem to be very few in number indeed. No wonder Crossrail appeared distinctly non-plussed by the former Transport Secretary of State’s announcement that they were looking to run Crossrail trains from Paddington up the West Coast Main Line.

A better, more regular and more reliable service

The third and final proposed improvement is:

[A] revision of the Peak services operating pattern across the network to provide a regular interval of trains, [presumed to mean from central London] including a train approximately every five minutes proceeding west from Paddington

One of the big advantages of taking over the Great Western 2tph semi-fast service is that this now means that there would be 12tph west of Paddington – exactly half the number east of Paddington. This makes potential operations much neater and also slightly ameliorates the criticism of Crossrail as originally planned – that it is ridiculous to have only 10tph (peak) continuing west from Paddington.

The extra 2tph beyond Paddington now potentially makes it possible to simplify the peak period service and have all trains that travel in from the west of Paddington go down the same eastern branch. Whilst operationally it might be preferable to have these go to Shenfield, it would probably make much more sense to ensure that Heathrow can be reached from Canary Wharf without a change of train. This would suggest that now all trains for destinations west of Paddington will originate from Abbey Wood and all Shenfield trains will terminate at Paddington (Westbourne Park sidings).

By good luck, inclusion of the GWR Reading stopping service means that off-peak the services west of Paddington also amount to the exact capacity of each of the two eastern branches. So, basically, if you are trying to provide a neat service pattern, consistent throughout the day, that the public can easily understand then you are starting from a better place.

As stated in the report:

the more regular service pattern will have a positive effect on operational reliability compared to the Iteration 5 timetable. The precise performance and reliability impact of the proposed timetable is being validated through joint modelling with Network Rail, as part of the timetable development process. Initial results indicate that the performance across the network will improve due to the improvement in the interval of trains

If anything shows that there is more to creating a new railway than providing the infrastructure, then Crossrail’s obsession with the timetable and the results they seem to be achieving as a consequence is surely it.

The glorified Tube line

All of this work may mean that by the time Crossrail transitions to the Elizabeth line, it may well be more like the glorified Tube line some have criticized that name for implying. It will have fewer stations and slightly fewer trains than the Central line, platform edge doors like the Jubilee, some semi-fast services on its most distant branch like the Metropolitan (at least in peak hours), frequencies similar to deep Tube lines and almost exclusive use of track. The main difference is that it is much bigger and extends much further than your average Tube line. All this with the distinctive TfL operating philosophy behind London Overground and the London Underground.

Crossrail may have started off as something different, but since the ‘Elizabeth’ name was announced significant effort has been made to integrate with the Underground’s ‘brand’. It seems that if Smith and his operational team have their way, then by the time it opens that integration may well extend to its service pattern as well.

Cover photo by Alex Nevin-Tylee

Written by Pedantic of Purley