LU WORLD SPECIAL
Driverless trains on London Underground have become something of a political football recently, but are they even a realistic possibility in the near future?
HAT work has been done by London Underground since 2016 on driverless trains?” asked Greater London Authority member Tony Devenish to Mayor of London Sadiq Khan during a mayor's question time on January 17, 2019.
Sadiq Khan's reply was: “The London Underground system is an extremely complex environment and, for reasons of safety, they are not suitable for driverless trains and there are no plans for their introduction.” This is despite the current rollout of signalling technology across the Sub-Surface Railway that will allow trains to operate semi-automatically.
Tony Devenish's question may have been motivated by the Future of Mobility document published by the Government Office for Science that same month. The preface, written by the Government's chief scientific adviser Sir Patrick Vallance, states: “One attraction of self-driving trains is potential cost savings. Self-driving trains break the link between the frequency of train services and staff costs. If staff costs are lowered, fully autonomous trains can run more frequently with lower additional costs.”
Then in July this year, while on a visit to the future Siemens factory in Goole, Prime Minister (and Sadiq Khan's predecessor as Mayor of London) Boris Johnson said: “Let's take advantage of this technological leap forward, let's not be the prisoners of the unions any more, let's go to driverless trains, and let's make that a condition of the funding settlement for Transport for London this autumn.”
LEVELS OF AUTOMATION
So is automation really possible or likely on London Underground? To answer this question, it helps to define what is meant by ‘automation' as, in railway terms, there are four main Grades of Automation (GoA).
A District Line ‘R' Stock unit waits at South Ealing on March 21, 1963, prior to testing experimental Automatic Train Operation equipment.
GoA1: apart from Automatic Train Protection
(ATP), all operations, including driving, braking and doors, are controlled by the train operator.
GoA2: trains are fitted with ATP and Automatic Train Operation (ATO), such that traction and braking are controlled automatically. Doors and emergency operations are controlled by a train operator.
GoA3: driverless trains have full automation, but an on-board attendant controls the doors and operates the unit in emergencies.
GoA4: no staff on board and the train uses full automation technology, including starting and stopping the service, operating doors and dealing with emergencies such as obstacle detection.
Studies have shown that the efficiency of a conventional railway directly relates to the availability of train operators/drivers, so automation enables more services to operate, generating more passengers and increasing income. Even upgrading from full manual GoA1 to GoA2 has been shown to reduce delays by up to 33%, while full
GoA4 automation allows additional services to be planned and run with as little as an hour's notice as demand changes.
Traditionally, staff are present at all stations and on all trains. However, automation can reduce staff coverage at stations or on trains in different combinations, potentially using multi-skilled staff instead who roam the network. While this could theoretically reduce staffing by up to 70%, most networks that migrate to GoA4 report a need for 30% fewer staff, while European lines that have migrated to GoA2 have reported savings up to 15%.
LONDON'S EXPERIENCE
Globally, there are numerous metro systems that operate GoA4 ‘driverless trains' in one form or another. Many take the form of relatively short, custom-built lines at airports or other limited sites, and most have been custom-built during or since the 1990s, with only a few converted from manual operation. There is not one entire metro system in the world that has been converted from GoA1 manual operation to full GoA4 automation.
The most well-known driverless trains in the UK are on London's Docklands Light Railway, which was designed from the outset to operate GoA3 driverless trains, carrying its first passengers on August 31, 1987. Then, as now, the DLR requires passenger service agents on all services to act as dispatchers and check tickets.
However, the first rail transport system to introduce automated technology was London Underground's Victoria Line. Launched to the travelling public on September 1, 1968, it featured Automatic Train Operation developed from tests using ‘R' Stock on the District Line (in 1962/63) and then Central
Line 1960 prototype Tube Stock on the Woodford and Hainault shuttle services (from April 1964).
The successful introduction of ATO resulted in the technology being developed further. The Fully Automatically Controlled Trains (FACT) project resulted in a very premature headline in London newspaper the Evening Standard on October 10, 1977, claiming
London Transport Museum
the entire Underground would be driverless by 1990. That same article claimed the Central Line would be the first to introduce the technology, saying: “It will be smoother, faster, safer and more reliable than the present service”.
However, FACT was overtaken by One Person Operation,(OPO) which would see the end of guards, and required LU train operators to control and operate trains alone. OPO was first introduced on the Circle and Hammersmith & City Lines in 1984, and the final LU services with guards were on the Northern Line in January 2000.
Since then, other deep tube lines have migrated to ATO: the Central in 1999–2001; Jubilee in 2011; and Northern 2013/14. Yet drivers remain in the cabs of all trains, controlling doors and dealing with emergencies, meaning progress has not been made beyond GoA2.
David Henderson collection.
Following the withdrawal of ‘A', ‘C' and ‘D' Stocks on the Metropolitan, Circle, Hammersmith & City and District Lines in the late-2010s, their replacement by seven- and eight-car ‘S' Stock trains has allowed the Sub-Surface Railway to move towards GoA2 operation.
However, budget cuts caused by the coronavirus pandemic mean the SSR's Four Lines Modernisation (4LM) programme has now been paused beyond the Signal Migration Area 07 (which is that between Dagenham Heathway and Upminster). This means automation beyond Finchley Road on the Metropolitan Line, and Barons Court (to Ealing Broadway/Richmond) and Fulham Broadway (to Wimbledon) on the District Line, will not now take place for the foreseeable future.
MYTH VERSUS REALITY
Taken at face value, automation promises a world where savings can easily be made through reduced staff costs, while enabling effective management of the transport network, with line managers able to identify and resolve problems centrally.
One might therefore be forgiven for thinking investors in transport systems should be highly motivated to dig deep into their pockets in the knowledge they will likely enjoy endless, almost risk-free operation in the future. However, when considering the real-world practicalities of a mature railway serving a capital city, how easy is it to introduce GoA4 automation?
Today's train operators perform a role that is a far cry from drivers and motormen of previous generations. Besides undertaking the role of traditional guards and operating doors, they are responsible for safety and on-train announcements. Train operators are also expected to investigate and overcome minor rolling stock faults, communicate major faults with signallers and line controllers, and manage/control the situation on the ground.
They must do all this following strict, often complicated, rules which need to be learned and accurately recalled at a moment's notice should a significant, or at worst, catastrophic problem occur.
De-staffing a railway is a high-risk strategy: removing train operators means additional security staff will then be needed at stations or on-board services to ensure passenger safety. If staff are to remain on board trains, it makes sense they are trained to undertake technical intervention when necessary, and be able to guide and evacuate passengers in an emergency.
The design of a railway is also important to allow safe conversion to automation. For example, the tunnels for the Channel Tunnel and Crossrail were designed to have regular ventilation shafts to minimise the risk of fire, and are large enough to have passenger walkways along their entire length in case of the need for evacuation. Additionally, they are provided with lighting throughout.
In comparison, most of LU's deep-Tube lines were constructed between the 1890s-1990s and are barely large enough to fit the rolling stock let alone emergency walkways. Therefore, if passengers need to be evacuated in an emergency, they not only face the hazards of walking along the track in a dark tunnel, but also negotiating the potentially electrified third and fourth-rails which can carry in excess of 750V.
Power is usually cut off before a train is evacuated in an emergency, but this then means any train trapped in a tunnel has to rely on its limited batteries to power reduced lighting and other essential features, such as the public address system for relaying instructions. Therefore, if a decision is made to evacuate, train operators need to act quickly to remove passengers, as the specified two-hour battery life is rarely achieved in practice.
It should also be noted that on lines already using GoA2 automation (such as the Central Line), train operators are authorised to operate services manually outside of tunnel sections in times of bad weather or reduced adhesion. Typically, this occurs during heavy rain, sleet, snow and the autumn leaf-fall season.
GOING DRIVERLESS
The language used by Boris Johnson in July is revealing. His soundbite that LU should “take advantage of this technological leap forward” needs to be considered alongside the use of the phrase “let's not be the prisoners of the unions anymore”. These comments seem to indicate his vision is fuelled by politics rather than any understanding of the complexities of railway operation.
He made a similar statement as part of his Mayor of London re-election campaign in 2012, when he said: “TfL (Transport for London) will rapidly establish a timetable for introducing the first driverless trains to become operational on the London Underground network within a decade. I am requesting a mandate from Londoners to push again for changes to national strike law, so that industrial action can no longer be triggered by a small minority of union members.”
If the entire Underground is to be converted for GoA4 automation, technology will need to be developed and applied across the network of variable and interconnected lines, including some that are shared with Network Rail. Each LU line uses different, often ageing, engineering infrastructure which has evolved at different times. Meanwhile, services must be allowed to continue throughout any upgrade work.
Yet London's Crossrail project has revealed how difficult a challenge this can be. Crossrail is already two years late, and it is likely to be another two years before the central section beneath the capital is open, and this jis ust one railway designed from the ground up without having to run any services while it is being built. Most of Crossrail's delays relate to the challenge of integrating the new railway with existing lines east and west of London which use different signalling equipment.
NEW STOCK
Siemens displayed a full-scale mock-up of a potentially cab-less design for the Deep Tube Upgrade Programme in 2013. The wall between the cab and passenger compartment was designed so it could be removed should driverless technology be introduced during the expected 40-year lifespan of rolling stock.
Siemens then won LU's competitive tender for new trains, with an initial order of 94 for the Piccadilly Line being placed in June 2018. These production trains, which should be built at the Goole factory, have also been designed so they can be converted to full GoA4 driverless operation if required. The contract also allows for additional trains to be ordered to replace the fleets of 1972 and 1992 Tube Stock trains on the Bakerloo, Central and Waterloo & City Lines.
Recent conditions imposed upon TfL by the Government during refinancing negotiations in May, following a drop in income because of coronavirus, led to the Department for Transport recruiting accounting firm KPMG to examine TfL's accounts. Part of this review included “exploring the feasibility of extending driverless operation from the DLR to other lines which are already automatic”.
It is understood KPMG's review will recommend GoA4 driverless technologies are re-examined, but only once all lines have introduced full ATO. This is not expected until around 2035, which would push the introduction of driverless technology back for at least another generation.
If a decision is made to introduce full driverless technology, it is likely to be with some staff on board trains. Central Government will need to be willing and sufficiently motivated to invest the very substantial funds necessary to see the project through especially, as in the case of Crossrail, if it starts to run late and over budget.
So can GoA4 driverless train technology be introduced to London Underground? It would seem almost certainly not. But a gradual move towards GoA3 automation is a far more likely prospect.
E