New train fleets are being introduced into service, but they are almost all late. Malcolm Dobell considers why
With an unprecedented number of new vehicles ordered since 2010 – over 8,000 – and with more orders to come, getting them safely, reliably and efficiently into service is a priority. However, challenges with testing, acceptance, software, stabling, depot facilities and long fixed formation trains seem to result in almost all of them being delivered into service well after they should have been.
A recent conference organised by the Institution of Mechanical Engineers tackled the reasons behind this and the possible solutions. The speakers were in an unusually candid mood and the overall conclusion was “could, should and must do better”. No one actually used the word “crisis”, but “late”, “more costly”, “more risk” and “not performing as well as hoped” were all terms that featured in the presentations.
Bill Reeve, director of rail at Transport Scotland, said he was embarrassed by the record of new train introduction over the last few years and it was an inconvenient truth that the last three rolling stock projects sponsored by his department, and those for Northern and TPE, where he is independent chair of the Rail North Partnership, had all been late and had suffered teething troubles. This had led to customer and other benefits being delayed.
As an engineer himself, Bill said that he understood that things can go wrong, but he emphasised that government ministers do not understand why rolling stock suppliers appear to promise what they cannot deliver. Moreover, they remember these problems when they determine the next round of investment, and the rail industry often fails to recognise that it is in competition with other transport modes.
He did admit, however, that government is at the heart of the challenge, as incentives between the infrastructure manager and the train operators have often been misaligned. He speculated that the output of current reviews might well recommend simplified processes, aligned incentives and joined up railway undertakings. Such a move would lead to problems and issues falling away as people tackle problems in teams.
Dave Hickson from GTR and Hans Benker from Siemens reviewed lessons from the Class 700 fleet introduction onto the Thameslink routes. This was the second biggest individual order in UK history, a total of 1,140 vehicles made up into 115 trains, in a programme from start of procurement to final delivery lasting over ten years.
The sheer length of the programme brought its own problems. For example, it was only six years after procurement started, and one year after contract placement, that GTR was appointed, and the opportunity to influence important aspects of the train design had passed. Details, such as aspects of the cab design, led to issues with signal sighting that caused additional infrastructure cost.
With such a large fleet, two new depots, many new stabling points and dependency on the Thameslink infrastructure programme, there were a huge number of risks to manage. Offsite testing at Siemens’ test track was highlighted as a key benefit, something noted by others, but it was emphasised that offsite testing was not a substitute for the rigours of passenger operation.
As was seen during the aftermath of the May 2018 timetable introduction, driver training was a major challenge, both for stock and route learning. Stabling was also a major challenge during the transition process, when there were more trains on the network than usual.
The Class 700 is a software-driven train. This means that Siemens can monitor service trains from the depot and improve predictive maintenance by collecting additional data. However, the focus shifting from hardware to software does mean that adjustments are needed by both operator and maintainer since every train has a collection of sub-systems, each with its own operating system, application and communications interface. Keeping software up to date is significantly affecting how change control is managed.
DRS Classes 68 and 88
Andy Martlew from Direct Rail Services talked about the procurement and authorisation of Class 68 and 88 locomotives. His approach was refreshing. He believed firmly in engagement and relationship building, with the supplier, with the operating staff and with Network Rail, that assurance is something that needs to be planned as part of the design, development and testing process, and that the customer – the railway undertaking (RU) – should be helping the manufacturer to succeed, especially where the manufacturer has limited UK experience.
Andy talked particularly about the network compatibility process, emphasising that the RU must deal with this and, whilst elements might need to be done by others, the overall responsibility cannot be sub-contracted.
Using the example of gauging, Andy outlined how he worked with Network Rail to resolve an initial list of nearly 5,500 “tight spots” by prioritising assessment and, if absolutely necessary, by modifying the locomotive gauge or by fixing the tight spot.
Andy stressed the importance and value of off-network testing. This allows a much shorter testing programme on UK’s crowded railway – a recurring theme of the seminar.
Things can and do go wrong, leading to what Andy euphemistically described as “in-field product development”, especially with brand new designs. Which is when the relationship built with the supplier pays off!
A view from over the water
Peter Smyth, Irish Rail’s chief mechanical engineer, presented his experience of buying new trains. Irish Rail is state owned but organised into two accounting groups – Railway Undertaking and Infrastructure Manager. Irish Rail is roughly a “medium size TOC”, carrying nearly 50 million passengers per annum with a fleet of 900 vehicles.
It has purchased over one billion Euros-worth of rolling stock since 2000 from Europe, Asia and the USA. Two tenders are currently in progress – for electric/battery electric vehicles for an expansion of the Dublin Area Rapid Transit network and for additional vehicles for diesel units bought from Hyundai/Rotem.
Peter outlined the process used in Ireland which will be familiar to anyone who works in an organisation that owns and operates its own vehicles: specify, procure, design, develop manufacture, take delivery and get into service, all over a period of approximately five years. He emphasised the importance of the specification: “If it’s important to you, include it in the specification.”
He highlighted particular issues he faces in a country with no local rolling stock builder, often having to work with suppliers building for Ireland for the first time, and the difficulty of getting any significant off-site testing due to Ireland’s unique 1,600mm gauge.
Steve Mitchell from Abellio Greater Anglia discussed the complications surrounding driver training. The company is introducing three classes of new train and its original plans would have delivered 10 to 16 trained drivers per week, enough for the May 2019 timetable while keeping driver training off the critical path and avoiding the “three-month refresh”.
In reality, there was a delay in having a train ready for driver training, leading to a large number of drivers who had been Part A trained, but who could not do Part B, causing “re-training due to three months”. Even when trains were available for part B, the automatic selective door-opening system was still not available and a catch-up plan was needed. Finally, training was sometimes cancelled due to no units being available.
As a result, driver training became near critical path, emphasising the need for very close liaison between the project and operations teams on train status for driver training.
Since then, many of the routes operated by the new bi-mode trains have suffered delays caused apparently by wrong-side track circuit issues and a serious near miss at a level crossing, currently under investigation by the RAIB. Beware of emergent properties!
Mark Molyneux from the Rail Delivery Group (RDG) explored the reliability impact of so many new trains. He presented this against a background of a) generally improving fleet reliability, b) about 15 per cent of delays being down to fleet issues and c) new trains taking a while to grow their reliability, although they are, generally, more reliable than the trains they replace in the long term.
With about half the national train fleet being replaced, the industry is rightly concerned that overall fleet performance will take a hit before recovering. Typical issues highlighted include software function and validation (it is apparent that software developers and train engineers do not understand each other!), the need for clear pass/fail criteria for the demonstration of compatibility, and improvements to specifications and contracts so that trains are specified in terms of reliability, against realistic timescales and using industry guidance such as Key Train Requirements and Key Interface Requirements. All this is necessary as it is unacceptable to subject the railway’s customers to debugging in public!
Several further speakers followed this theme. Some comments seemed so blindingly obvious that it was almost a surprise that they had to be mentioned – for example, when buying faster and more powerful electric trains, it is important to calculate whether the power supply has enough juice to run them!
A number of speakers questioned whether standards were out of date, appropriate or flexible enough, and whether designers could react quickly to discovered problems. These ranged from train surfers discovering they could use the inter-car connecting cables as a ladder to climb onto the train’s roof, curved windscreens showing ghost reflections of signalling and trains with automatic doors having to be risk-assessed at every platform they could visit.
This very frank and open conference illustrated the scale of the challenge that the UK rail industry has over the next two to three years. It is one that will have to be solved.