Thameslink’s ETCS level 2 signalling and automatic train operation is an impressive achievement. However, it is misleading for the Digital Railway’s website to state that “digital deployment on the Thameslink will allow 24 trains per hour to run”. Increasing Thameslink’s capacity also required significant infrastructure works, which included increasing London Bridge’s through platforms from six to nine, a new viaduct to double the tracks west of the station, and grade separation at the Bermondsey dive-under.

In Manchester, the recently completed Ordsall Chord removes conflicts at Piccadilly’s station throat by re-routing some trains through the congested Castlefield corridor. For this reason, the Ordsall Chord was to be just one part of the Northern Hub scheme whose benefits were assessed at £2.8 billion over a 60-year period with a benefit to cost ratio of four to one.

The Northern Hub was also to increase trains through the Castlefield corridor from 12 to 16 trains per hour by providing two additional through platforms at Piccadilly and lengthening platforms at Oxford Road. A Transport and Works Order for this was submitted in 2015. However, a decision is still awaited and these projects have disappeared from Network Rail’s website. Further doubts about the future of these platforms were raised when, in July 2017, Chris Grayling advised he wished Network Rail to consider whether digital technology could remove the need for them.

Yet Thameslink shows that significant capacity improvements require both digital signalling and infrastructure work. This is also one of the lessons from our feature on Railtrack’s 1994 proposal to introduce moving block signalling on the West Coast main line. We explain why this case study of corporate self-delusion was instrumental in bringing down the company.

The 1999 report which proposed abandoning moving-block signalling concluded that, although the signalling system is the dominant factor for capacity improvements for metro operations, the capacity constraint is mainly infrastructure on a mixed traffic railway.

Even now, the benefits of the digital railway continue to be oversold. In 2016, the House of Commons Transport Committee considered that Network Rail was “over-heroic” in claiming a 40 per cent capacity improvement and that it should not promise moving block signalling until ETCS level 3 is viable.

Hence, although digital signalling offers undoubted benefits, it is difficult to see how it can remove the need for the Castlefield corridor’s extra platforms if the huge benefits of the Northern Hub scheme are to be realised.

Digital in-cab signalling was essential for the first French high-speed lines, for which the TVM system was developed in the 1970s, well before the ERTMS/ETCS specification was finalised. TVM is also used on the London to Paris and Brussels high-speed routes. However, as Clive Kessell explains, if trains on these routes are to run throughout Europe they will need ERTMS. This is a significant long-term project, especially as, unlike TVM, ETCS requires a radio system.

The importance of a dedicated railway telecommunications network is underscored by it being a fundamental element of ERTMS. In a feature about Network Rail Telecom, we explain its Fixed Telecoms Network and the work of its Network Management Centre, which monitors 200,000 devices connected to its 40,000km of cable and 2,500 radio masts.

Our signalling and telecoms focus this month also includes articles on signalling procurement, electromagnetic compatibility and obstacle detectors at level crossings. It also looks to the potential impact of artificial intelligence (AI).

David Shipman’s feature considers how AI can automate simple and repetitive jobs. With a shortage of skilled talent, the automation of everyday, time-consuming activities enables experts to focus on activities requiring their skills, leaving computational power for repetitive and straightforward tasks.

Paul Darlington’s article considers the difficulties of ensuring that ethical decisions are taken. As Paul explains, AI is programmed to ‘learn’ what it can do and improve its own performance. Hence its behaviour can never be completely foreseen. His article illustrates this point with a nice anecdote about children’s behaviour.

Moving away from signalling, Colin Carr gives us a feature on Brunel’s historic bridge at Steventon, which needs to be demolished and replaced for the GWML electrification. However, the Vale of White Horse district council recently denied planning permission for this despite Historic England accepting the case for the demolition of this listed structure. Yet it seems planning permission was refused due to the villagers’ understandable concerns about the impact of the resultant 10-month closure of the road through their village.

Railway civil engineers routinely demonstrate great ingenuity to ensure that bridge replacements only close the railway for a day or two. It will be interesting to see if the road closure duration at Steventon could similarly be significantly reduced to resolve this impasse.

It is nearly time for the big biennial Berlin rail trade fair that is InnoTrans. In his preview of the event, Nigel Wordsworth can only mention a small number of the 3,000 or so exhibitors from 60 countries that will be there. Rail Engineer will, of course be there and is interested in hearing from you if you are going too, see page 10 for details.


Read more: Rail Engineer September 2018: Signalling and InnoTrans