Third World Class Capacity: The Austerity Upgrades


In our previous article on London Underground’s World Class Capacity programme we looked at the reasons for its cancellation. Now it is time to look at exactly what has been cancelled and what can be salvaged from the aborted programme.

The World Class Capacity programme…

The four main parts of the World Class Capacity (WCC) programme were:

  • VLU2 – upgrade the Victoria line to 36 trains per hour (tph) between Brixton and Walthamstow Central
  • NLU2 – upgrade the Northern line north of Kennington to 30tph
  • JLE2 – upgrade the Jubilee line with an aim of providing 36tph between West Hampstead and North Greenwich
  • JNAT – Purchase additional trains to support upgrades on the Jubilee and Northern lines

The was also a small, but important, project that ended up being part of the World Class Capacity Programme – which we’ll cover shortly.

…Comes to a halt

It appears that once London Underground found out that the additional trains were more expensive than they were prepared to pay (and that the TfL Board had authorised them to spend), the whole World Class Capacity Programme was cancelled. Or as TfL put it, ‘paused’. There is a certain amount of inference here because there appears to be no definitive official statement from TfL that has been put in the public domain.

On the Northern line, the implication is that we will be left with a maximum of around 24tph north of Kennington. It is a little bit more complicated than that because the Northern line timetable contains numerous tweaks to eke a little more out of it in critical places at critical times, but 24tph is a good working value.

On the Jubilee line, it seems the current maximum of 30tph between North Greenwich and Willesden Green will remain.

Or maybe not

In fact, things are not as bad as they appear. There are enhancements already in the pipeline that are far too far gone to be stopped and are not dependent on new trains.

Furthermore, it is hard to believe that work on enhancing the capacity of the Jubilee and Northern lines has come to an abrupt end. There are plenty of precedents for TfL investment programmes being shut down and some of the proposals being carried over to other projects. We are not expecting to see a revival of any part of the programme that was only included to support the new trains. Typically, this is work that would have been necessary to increase the capacity of the depots. That still leaves plenty of items that may continue to be worthwhile to implement so as to make the most of the existing train fleet.

To look at where we are and possibilities for the future on the Northern line, we first need to look at that small, often forgotten part of the World Class Capacity programme we mentioned earlier. One which doesn’t even appear to have a consistent formal name.

The Northern line Interim Improvement package

In the most recent WCC paper presented to the Programmes and Investment committee, in June 2017, there is a part of the programme rather clumsily referred to as “Interim improvement to the Northern Line train service (Morden Branch)”. This is something of a rare mention for this work. The sum of money involved is relatively small, so it tends to get left out of reports into investment programme updates. Its importance means it really deserves more emphasis than it gets, although perhaps it isn’t mentioned much so as not to highlight the one major failing of the original Northern line upgrade (now generally referred to as NLU1).

Modified diagram showing corrected version of TfL Northern line upgrade plan as at September 2013. It shows 32tph to Morden.

In September 2013 London Underground gave their first public briefing of proposals for Northern line frequencies. This included 32tph in the peak period between Morden and Kennington and was due to be implemented by December 2014. What actually happened was that, at some point, it was realised that this 32tph was not easily achievable but that 30tph was. Amongst other things, the 32tph plan relied on track improvements at Morden, braking improvements to the trains, software improvements to the signalling system and some beefing up of power supplies between Morden and Kennington.

Of all the issues preventing 32tph on the Morden – Kennington stretch, it seems it was the software upgrade that was causing the greatest delay. It wasn’t that the software upgrade was so difficult, just that it had been deemed that the software upgrades for the Night Tube had to take priority. This left a practically-complete programme about to slide into the red area on TfL’s project list, indicating that the programme was behind schedule.

What followed was one of those little administrative tweaks that makes it so hard to follow what is going on. Rather than have NLU1 upgrade be seen to fall behind schedule, the 32tph element was moved into its own sub-programme of the subsequent World Class Capacity programme, taking an appropriate amount of the money originally allocated to NLU1 with it. And thus the NLU1 programme was finished early and within budget.

It is still the case that the Northern line Interim programme (for want of a better name) is a long way off from completion, timewise, but most of the work is believed to have been done, or only awaits that vital software upgrade. As things stand, the 32tph service is due to be implemented in April 2019.

An improvement of 2tph might not sound like much, nor may it initially seem worth the effort. When you think of it as a 6.7% increase in capacity without a need for more trains though then the benefits are clearer. Looking further to the future, this upgrade reinforces the lesson learnt from the Victoria line that, if you look hard enough, you can often do more with the same number of trains.

Another reason to not dismiss this 2tph increase lightly is that the Clapham-Kennington section is the busiest stretch of the Northern line. Now NLU1 is complete, the Northern line Interim programme is the only scheme that will increase capacity on the Northern line where it is most needed.

Improvements to the Jubilee line

Unlike NLU2, JLU2 has already had improvement work take place. By far the most important of these improvements is the replacement turnback siding at West Hampstead. Prior to this, the siding at West Hampstead was not really ideal for daily use and was only used for unscheduled reversing of trains during periods of disruption.

In March 2018 it is planned that the relaid West Hampstead turnback will come into use on a regular basis, at least during peak hours. This will make it possible to extend the peak service from approximately one hour to three – something that is starting to become the standard London Underground objective on all its lines.

Inspiration from the Victoria line

It is important to give credit where credit is due, and it must not be forgotten that one part of the World Class Capacity programme has been very successful indeed. This is the increase in service on the Victoria line from around 33tph Brixton – Seven Sisters to 36tph Brixton – Walthamstow Central. What is more, the combined capacity upgrade programme (VLU1 and VLU2) required very few extra trains. Admittedly if you start from 1967 tube stock and replace it with 2009 tube stock you have a lot of potential for improvement but, that aside, VLU2 has really highlighted how much more you can achieve with your rolling stock if you have a completely holistic look at how to increase capacity. We have written about the efforts to improve the Victoria line previously.

Given that we are where we are, it would seem to us that what is needed now is to accept that there won’t be more trains and to look at how London Underground can make the most of the ones they have. Our belief (backed up by some of our commentators who had very similar ideas) is that here there are still things that can be done.

Smart ways to increase capacity

The consensus is that there are three main ways to increase capacity when you only have a fixed number of trains. These are to:

  • increase the availability of trains
  • reduce journey time
  • terminate more trains short of the line terminus

Train availability

Train availability – in terms of the number in service – has generally risen over the past few years as newer, more reliable, trains appear or older trains are refitted with more modern, more reliable parts. Historically, a figure of around 10% of the number of trains needed in service was added to cover for maintenance and repairs. Figures, even for older rolling stock, have improved significantly in recent years. Southern Railway has managed to achieve availability closing in on 96% and the Waterloo & City line (admittedly a special case) has a timetable that relies on 100% availability of stock in both the morning and evening peak. Perhaps more representative of what is possible is the DLR, which relies on 146 out of 149 units being available in order to be able to run the peak timetable.

In contrast, the Northern line only requires 96 out of 106 trains to be in service to run the timetable (so 10% ‘spare’). For the Jubilee it is 58 out of 63 trains (about 8% spare).

Reduced Journey Time

To reduce journey time on the London Underground you generally need to make the trains go faster. This is because reducing the dwell time at stations is not realistic and closing lesser-used station in peak hours is probably politically unacceptable. Tweaking the Automatic Train Operation (ATO) to try to ensure trains go as fast as they are permitted to go can produce benefits. Sometimes it is possible to re-lay the track to produce potential time savings. In general though, the usual restriction is a lack of electrical power at the rail. Upgrading the power supply may only save seconds, but seconds count. Even changing from a working timetable with a 15-second granulation to one with a 5-second granulation may produce benefits in speeding up services.

Terminating trains short of the end terminus

In the old days, it was usual to thin services out on the Underground as lines went further out and the number of passengers on board was reduced. Modern practice, for a number of reasons, is to try to run trains all the way to the final terminus if possible. Such a policy only works well if you establish how many trains you need to achieve this and then acquire them. If there is a shortfall you either need to accept that you will run a less frequent service in the busiest section than desired or that not all trains will go all the way to end of the line.

London Underground is in a situation where there is no realistic opportunity to buy any additional trains. The problem, therefore, needs to be looked at differently. Given that you have a fixed number of trains, the objective is to utilise them in the most beneficial way.

We have already seen how, on the Jubilee line, the plan is to turn a significant portion of trains around at West Hampstead instead of allowing them to continue to Stanmore. This is so that the central core section is beefed up at the expense of the outer stations. Such a solution can be applied to any line (or branch of a line) where passenger numbers significantly tail-off at the outer stations.

If it turns out such a solution of turning trains short of the final terminus in order to beef up the busiest section (normally the service in the central section) is acceptable, then it immediately introduces complications. Proposals to increase the service in the central section will probably lead to a requirement to boost power supplies in that area. In addition, extra cooling may be required to remove the additional heat from the tunnels.

Further complexities for the Northern line

The Northern line has various issues to contend with over the next few years – some beneficial when it comes to running a more frequent service, some not.

The year 2020 is expected to be important for the Northern line because the underground passages at the enhanced Bank station are expected to then be complete. As the station will be able to handle more Northern line passengers, people will expect to see more Northern line trains call there. The desire for more trains will increase in 2022 when the new Cannon St entrance will make the Northern line at Bank station even more attractive to City passengers.

Cannon St entrance of Bank station opening in 2022

The year 2020 is also the year when, if all goes to plan, the Northern line will be extended to Nine Elms and Battersea Power Station. Under current plans, it is the intention to provide an initial service of 16tph to Battersea Power Station. In peak hours this service will feed in to 24tph on the Charing Cross branch of the Northern line and so is unlikely to be an even-interval service.

Northern line to Battersea Power Station

The plan never did quite make sense

What was always strange about the now-cancelled purchase of the extra trains for the Northern line was that they weren’t due to come into service until 2023. Yet it must have been obvious to anyone who followed the fortunes of the line that the trains were ideally needed in 2020.

London Underground have always said that if they don’t have extra trains for the Battersea extension then they will manage with what they have. They have stated that their preferred solution is to increase the availability of trains. This really should not be difficult in this day and age. Initially, we are probably talking about 16tph to Battersea and a maximum of three trains extra in service (so 99 out of 106) in the peak. With a few tweaks elsewhere on the system, it might be possible to get away with just two extra trains.

Camden Town is an issue – as always

In 2023 or 2024 London Underground hope to have Camden Town station upgraded. This will make it feasible to introduce segregation of the branches so that all Bank branch trains go to High Barnet and all Charing Cross branch trains go to Edgware. If they chose to take advantage of this possibility it could lead to further optimisations. In particular, it would make it much easier to run different frequencies on the Bank and Charing Cross branches. This would be beneficial because, in the peak hours, the demand is higher on the Bank branch.

HS2 and Euston

Finally, on the subject of future Northern line issues, from 2026 onward there is the spectre of HS2 and the number of passengers it will discharge at Euston. One might argue that the additional numbers will not be that significant. Nevertheless, the fact is that you will have much longer trains than now depositing a lot of people complete with baggage at Euston, and the fact that the Northern line has both of its central sections (Bank and Charing Cross branch) separately serving Euston means it is highly desirable to have a good a service as possible going to Euston.

Opportunities, but something has to give

The Northern line was due to receive an additional 17 trains. To provide the intended service without them is a tall order and almost certainly cannot be done. That doesn’t mean that one cannot achieve something. After all, if one could get an additional five trains, currently sitting in the depot, into service in the peak period then that would be a good start.

It seems that if TfL are serious about providing an element of capacity upgrade without the extra trains, then they are going to have to accept the principle that one terminates trains in north London before the end of the line. To get the maximum benefit on the Edgware branch this would have to be at Golders Green where there is already an ideal turnback facility with cross-platform interchange both to the north and to the south.

To maximise the potential gain in freeing up trains on the High Barnet branch one really needs to install a turnback siding at East Finchley as originally proposed in NLU2. The WCC proposal shows that London Underground was obviously planning to turn trains here. Originally this was probably because High Barnet could not easily cope with the extra trains, in a future scenario though it could be to claw back some running time for use elsewhere.

Significant running times

The train running time from East Finchley to High Barnet is 13 minutes. For the return journey, it is 12. Times between Edgware and Golders Green are 11 minutes or 12 minutes depending on direction. So, roughly, on each northern branch, if you turn a train short you will save 24 minutes train running time.

The line south of Kennington is already at capacity so we do not have to consider any extra running time between Kennington and Morden. Running time from Kennington to Golders Green is 27 minutes via Charing Cross and 32 minutes via Bank. Running time to East Finchley from Kennington is 27 minutes and 31 minutes respectively. So, roughly, not allowing for turnround time, you can do a round trip from Kennington to either Golders Green or East Finchley and back in about an hour.

As a very crude guide, for every five existing trains that you turn short at Golders Green or East Finchley you gain 120 minutes, which is equivalent to two extra trains between Kennington and either Golders Green or East Finchley.


Getting fleet availability up to 101 trains out of 106 trains and an aggressive policy of turning back trains early on the northern branches of the Northern line should, theoretically, make it possible to get around 28tph on the Charing Cross and Bank central sections. Just because something may be possible, however, doesn’t necessarily mean it is worth doing. The price to pay to passengers at the outer stations on the Northern branches may be too high. Nevertheless, it does show that some options are there.

Frequencies that may be possible without buying new trains

Longer waits and no direct train

It is not hard to see that there will be a number of problems with such a plan. High Barnet currently has around 19tph in the peak period. If we assume that East Finchley had 28tph and half went on to High Barnet (with the others either terminating at East Finchley or continuing to call at Finchley Central then go on to Mill Hill East) then this would be a reduction of 5tph. Raw passenger numbers suggest that 14tph north of East Finchley Central would be perfectly adequate, but the passengers involved – who would be left with fewer, more crowded trains than they are currently used to, would likely disagree. Worse still for the passengers affected, the trains they still had would probably all go to the same central section (Bank or Charing Cross), so passengers not served by a direct train would have to catch the first train and take it to East Finchley.

High Barnet Branch of the Northern line

Aside from the passenger protests, such a proposal would also produce other problems. Passengers south of East Finchley would possibly not be able to get on the first train, but would instead have to wait for the next train which started at Mill Hill East or East Finchley. They, in turn, might find that the trains they can get on do not go to the central section (Bank or Charing Cross branch) that they want.

Edgware Branch of Northern line

The situation is similar for the Edgware branch but this would probably entail half the peak period trains terminating at Golders Green. So if there were 28tph in the central area, this would mean just 14tph to Edgware instead of almost 24tph. That reduction of 10tph between Golders Green and Edgware would enable around extra four trains per hour to be provided between Kennington and either Golders Green or East Finchley.

We (probably don’t) have the power

Any part of the line which has an increase in the number of trains may well need a suitably upgraded power supply. As outsiders, it is impossible for us to know quite what this involves, although there must be some electrical sub-stations reaching the end of their lives and due for replacement anyway.

Encouragingly, during the seven-week closure of the Bank branch in 2020 when the track and station layout of Bank station is changed, London Underground are promising a 33% increase in the number of trains via Charing Cross (so 32tph instead of 24tph). This does seem to suggest that, by then, at least one of the two central sections could support a 32tph service if required to do so.

More encouraging still, the rebuilding of Bank station includes a new substation suggesting that an upgrade from a notional 24tph now (but in practice it goes up to 26tph for short periods) to 28tph for trains via Bank may well be easily possible. So, if a major power supply upgrade is needed for a substantial stretch of the line, the affected location or locations probably lie north of Camden Town.

Most of the conductor rail on the Northern line is, surprisingly, still steel rather than the higher-conductivity aluminium ones with a steel protective layer on top. This may mean that beefing up power supply where necessary, sufficient for a slight increase in frequency, requires little more than replacing the conductor rail, but that is tricky to know for certain. Replacing the conductor rail was, however, one of the items that was to be carried out under NLU2. It is hard to imagine that it was justified then but cannot be justified now.

Even within the train there may be opportunities to improve the power supply. The efficiency of inverters, now in much more common use with the advent of renewable energy, has improved over the years. If the original ones are still fitted on the trains then it may be cost-effective to replace these for the extra percentage of power made available at the wheels.

Think the unthinkable?

Despite all the above, it may be necessary to think about things that would not normally be acceptable in order to provide the greatest good for the greatest number. If not many people can get on at Clapham Common and Clapham North Tube stations in the morning, would it not be better just to close them at peak times in order to boost the service elsewhere?

Another way of tackling the problem is to reduce demand. Could TfL abandon one of its lines in the sand and, come 2018, reinstate the Thameslink service on the Tube Map to take pressure off the Bank branch of the Northern line?

The Jubilee line

As we have already covered, the turnback at West Hampstead should come into use in March 2018. This will provide 30tph (a train every 120 seconds) between West Hampstead and North Greenwich.

Jubilee line to Stanmore

Under JLU2, the next stage was to have been a very slight increase in frequency. With an implicit move to a timetable graded to the nearest five seconds (rather than fifteen currently), this would mean a train every 115 seconds (31.3tph). This amounts to 4.3% increase, so is big enough to make a difference. This increase was intended to be “using the additional power and cooling capability”. We can’t know for sure, but it seems that if it didn’t require extra trains and could previously be justified, then it too can probably still be justified now.

31.3tph? Or 32tph?

The 115-second proposal is an update to the original plan, which originally proposed 32tph as the first increase in frequency. 32tph is as an even-interval service which amounts to a train every 112.5 seconds, something which clearly can’t be timetabled, but it does make one wonder if this could be achieved by running trains alternatively at 110- and 115-second intervals. It seems that this shouldn’t be too hard, in which case you have achieved a third of the original objective without requiring any additional trains.

Keep cool

The latest Investment Programme Report refers to a London Underground Infrastructure Renewals project where it states:

To reduce temperatures in tunnels on the Jubilee line between Baker Street and Green Park, we continue major tunnel ventilation system upgrades on site at Park Square Gardens and Hay’s Mews. These works will take approximately one year and support the Jubilee line upgrade (stage 2).

We don’t know where the proposed cooling was planned, but it seems quite possible this is not dependent on the JLU2 upgrade and doesn’t come out of that budget.

More than 32tph?

It is possible that slightly better than 32tph can be achieved by reducing the service from Stanmore in the peak period (from around 18tph to around 16tph) and terminating at Willesden Green or West Hampstead rather than Wembley Park.

As others have pointed out, once you start using the turnback sidings at Willesden Green and West Hampstead, you can get further train utilisation efficiency gains by one of :

  • having drivers ‘step back’ so a new driver is available to take out an incoming train straightaway
  • introducing auto-reverse so that a driver can be changing ends by walking through the train as it is automatically driven into the turnback siding (if permitted in older tube trains)
  • introducing auto-reverse so a driver can leave the train then send it to the turnback siding to reverse – the driver can rejoin the train when it stops automatically at the country end of the London-bound platform

If London Underground could operate a train every 110 seconds then that would be close to the fallback option of their original plans (with extra trains), which was for a train every 105 seconds. The original plan aimed to have 36tph – a train every 100 seconds – but it was always recognised that there was a small risk that this might not be achievable.

It is pretty much inconceivable that the original JLU2 plan can be achieved without extra trains. But, largely using London Underground’s own figures and plans, it seems they can get surprisingly close for a lot less money.

Towards a more achievable plan

To many, the failure to continue with the WCC programme is a huge disappointment. At the same time, there always seemed to be something profoundly unsatisfactory about ordering new Tube trains when it was clear they would only have a limited life. This was coupled with more optimism than could perhaps be justified that the trains could be acquired at a sensible price.

What must be hoped for now is that London Underground revisit the project, take into account current best practices and see what can be achieved with the trains they do have. It is likely that London Underground believe that there are things that can still be done to improve the tube travellers lot on the affected lines. It will be very interesting to see what they come up with.

Thanks to ngh, Moosealot and others whose comments on a previous article were very helpful in the writing of this one.

LR is funded largely by its community. Like what we do? Then help us to do more!

Donate £1Visit our shop

Cover image by HahifuhehoOwn work, CC0, Link

Written by Pedantic of Purley