Dr Thomas Staffelbach is Head of Train Protection within the Signalling & Telecommunications unit of Swiss Federal Railways' Infrastructure Division.
Switzerland has consistently led the way in the development and application of ERTMS, with ETCS Level 2 now operational on the 45 km Mattstetten - Rothrist line and in the 34·6 km Lötschberg base tunnel. Including development costs, the Swiss have invested more than €400m on ERTMS over the last 10 years. Has it been worth it?
'If I said "no", it would be a rather flippant answer', says Thomas Staffelbach. 'You have to see it as a long-term investment. And ETCS was in any case needed for operation at 200 km/h on the Mattstetten - Rothrist line to meet the timetabling demands of our Bahn 2000 programme. Apart from that, we will equip the entire SBB network by 2017, mostly using Level 1 with Limited Supervision. The Swiss segment of Corridor A from Basel to Domodossola will be fitted by 2013, with the second route from Basel to Chiasso fitted by 2017 when the Gotthard base tunnel opens. Level 2 will be installed in the Gotthard and Ceneri base tunnels, and we may look at fitting to other lines.'
But will it generate cost savings? 'Level 2 does away with lineside signals, but this will only apply on a couple of hundred km in Switzerland. The most benefit will be derived from reducing the number of train protection systems from four to three. Now we have ETCS, Signum, ZUB and ETM (European Transition Module). In the future we will be able to order vehicles fitted just with ETCS, which means lower capital and operating expenditure.'
Asked about future challenges, Staffelbach emphasises that the current ERTMS specifications permit 'a lot of freedom and flexibility in the design'. Problems arise for suppliers in that onboard components need to be able to provide all functionalities with all combinations of equipment, and all possible ways of implementation. 'That means it is almost impossible for suppliers to guarantee that their equipment will work on all lines, because there is such a degree of freedom in the specification. The railways need to reduce the flexibility in the design for their own benefit.
'This would limit the amount of testing and approval activity, not just for the first time, but also with later upgrades.' There is a safety element too, which came to light before the Lötschberg base tunnel was fully commissioned (RG 1.08 p18). Staffelbach describes this problem as 'an implementation error' which he says was discovered 'during testing'.
He highlights the need for 'intensive testing of the integrated system', drawing attention to the problems that could arise if there were 10 different Radio Block Centre installations from different suppliers in the Rotterdam - Genova corridor, for example. If or when new software needed to be installed, the whole process would become commercially unattractive, but for the moment 'people do not recognise this'.
Another issue is that Europe still has national safety authorities. 'We need to find a way of unifying the approval process - for example an operator running between Rotterdam and Genova would not want to go through different processes to have his onboard equipment approved for the Netherlands, for Germany, for Switzerland and for Italy. We need to unify procedures and regulations as much as possible. Each time you change a bit or byte in your system, you have to run through the whole process again.'
Unifying standards and processes suggests that development must be limited. Does this mean that the System Requirements Specification 3.0.0 should be the final one? Staffelbach points to Microsoft, which offers new software releases at regular intervals. Taking the example of electronic interlockings, he says 'there are upgrades all the time as the technology is continuously evolving. It is no different with ERTMS. As new knowledge emerges, new releases will be needed to deal with changes in functionality and to correct errors. We would be deluding ourselves if we said that SRS 3.0.0 would be the last version - there are already functionalities such as GPS that are not in 3.0.0.'
'I feel there is too much freedom', Staffelbach concludes. 'In the past there were lengthy discussions about when to freeze the content of the specification. I would rather reduce the degree of freedom and limit new releases to every two or four years on the basis of urgent needs. This would allow the market to respond and would avoid the academic questions.'
'Each time you change a bit or byte in your system, you have to run through the whole process again'
Dr Thomas Staffelbach
Head of Train Protection, SBB Infrastructure