The cost of upgrading the three of the London Underground lines to driverless trains would cost at least £20 billion, for just three of the oldest lines.
The figure was revealed in a recent Mayor of London answer, and more details about how the number was obtained are available in a report by Transport for London (TfL) on the costs and implications of driverless trains.
Although previous studies have examined how driverless trains could be introduced on the London Underground, the latest report was a condition of TfL’s funding settlement with the government to keep transport services running during the pandemic.
As it happens, much of the London Underground is already semi-automated — what’s known in the trade as “Grade of Automation 2” (GoA2), which is relatively automated driving but with a driver in the cab. The “driverless” Docklands Light Railway is GoA3, which is driverless except when it isn’t, and still has an attendant on board.
The government-mandated report looked at the cost and benefits of what is known as GoA4, which means no staff would be on board the trains.
A heavily redacted copy obtained through a Freedom of Information request filed by ianVisits examines the many options available, how some other cities have managed the migration, and the benefits of upgrading the London Underground to driverless trains.
Although it might seem sensible to upgrade in phases from DLR-style driverless to eventually fully driverless, the report noted that although it’s cheaper to upgrade to a DLR-type service than going totally staffless on trains, the higher wage bill on a DLR-type service effectively wipes out the savings. In effect, either offers the same long-term costs and savings.
So, if you’re going to upgrade, you might as well do it to the full specification.
The report’s early indications suggest that the London Underground could be upgraded to fully driverless operation, as has been done elsewhere in the world, although the London Underground’s age and design would prove more of a challenge.
TfL and the DfT identified 88 requirements to upgrade an existing London Underground line to full GoA4 standard, and while most of the changes were within scope to provide, 11 of those requirements were considered particularly challenging due to the London Underground’s design and age.
Apart from the signalling that controls the trains, there are also considerations about dealing with the public. Modern standards would require all stations to have platform edge doors fitted, and a lot of work would need to be done to segregate TfL and Network Rail services where they overlap.
There are a number of modern detection systems that can reduce the investment cost of physical barriers usually needed to segregate driverless from driven trains, such as those in use in Australia. However, these would need to pass UK regulatory safety checks, and that’s not guaranteed for the much older London Underground network.
One of the biggest challenges is when different train types share a station, particularly the Piccadilly, Bakerloo, and District lines and the London Overground. Platform edge doors are not an option there, so either services need to be reduced to just one line and train type, or a radar or laser-based detection system needs to be found to stop a train if there’s an obstruction on the tracks.
Nuremberg uses a radar system, although, as the report notes, it will likely be scrapped by 2035 in favour of platform edge doors.
One interesting fact that emerges from the debate is that Platform Edge Doors require trains to dwell for longer at platforms, and on highly intensive services such as the Victoria line, that could mean fewer trains per hour on the line.
And of course, there’s the issue of what to do in a breakdown, although the report acknowledges that these are manageable issues, and not a deal breaker for going fully driverless.
The report candidly states that upgrading to GoA3 or GoA4 on some lines would be value for money. It’s just that it costs an awful lot of money to deliver the value for money.
The Mayor’s answer last December broke the numbers down:
Bakerloo line – £4.4 billion
Central line – £10 billion
Piccadilly line – £4.9 billion (on top of the new trains already on order)
If they were to go ahead with upgrading to driverless trains, it would likely be phased in over time as part of the routine upgrades of worn-out equipment—but such an upgrade typically takes a decade or more.
For example, the ongoing sub-surface line trains and signalling upgrades—the Four Lines Modernisation project—started in 2007 and is still not fully completed. The new Piccadilly line trainshave the option to be retrofitted for driverless operation, but even the basic signalling upgrade to run more trains under normal service is on hold because of the cost.
Although a lot of people will want to focus on cost savings — if there are still staff on board, then the cost savings will be sharply reduced. In fact, the main revenue uplift is expected to be from the improvement in the service that automation offers in encouraging more passengers to use public transport.
The report also warns that the revenue uplift is mainly concentrated in the very old lines that are still largely driven manually — such as the Piccadilly and Bakerloo lines — and that the uplift on, for example, the Victoria line would be “negligible.”
That means if you had a magic pot of cash to spend on upgrading the London Underground, improvements to the older lines to automatic operation would pay back that investment far quicker than upgrading more modern lines to fully driverless trains.
However, the revenue uplift was still only around 32p in every £1 worth of overall benefit from going driverless.
There are also some important and useful non-financial benefits from the upgrade, such as an improved customer experience from adding platform edge doors, and the associated improvements in accessibility that would come with the changes to station platforms needed to add the doors.
However, the analysis also found some disadvantages, which might put people off travelling by public transport. The “journey ambience” effect, where people like or dislike travelling, could be reduced if train drivers are removed from trains entirely, compared to the DLR-style service with a train attendant in the carriage.
If, as was widely reported at the time, the government’s interest in funding the billions it would cost to deliver driverless trains was to reduce strikes, then it has to be remembered that having train attendants in the carriage won’t solve the strike issue.
However, the report’s comparison of the costs and benefits of DLR-style service and a totally unstaffed service was so similar that it would be very difficult to argue in favour of the fully unstaffed train option, especially as it seems to be offputting to the public.
The report concluded that the majority of the costs and the benefits come from upgrading the older lines to GoA2 grade of automatic operation (as on the Victoria line), and while there’s still a benefit in going to GoA3+ or GoA4, the benefits are much lower. There’s still a benefit, but not much of one, and most of the benefit is intangible, such as passenger preferences for stations with platform edge doors rather than any practical improvement in train service.
Over time, yes, the services will get ever more automated, and it’s possible that at some point in the future, incremental upgrades will result in a D-Day moment when a line can go fully driverless. However, to spend billions upfront for fairly negligible savings and revenue gains is a very hard argument to make.
Considering the report’s findings that the main benefits come from upgrading the older lines to Victoria line-style automatic operation rather than upgrading modern lines to DLR-style operation, it’s increasingly understandable why everyone involved quietly dropped the idea of upgrading the London Underground to driverless trains.
Comments