Great Western Coffee Shop

From Luxemburger Wort (http://www.wort.lu/en/luxembourg/south-luxembourg-missed-stop-sign-cause-of-fatal-train-crash-58a2c336a5e74263e13aa8e9):

There are a lot of pictures in that report showing the serious damage to both trains. The view in the quote above shows the damage to the fronts of the two trains.

(https://blobsvc.wort.lu/picture/fd303cc2932d1f93a3e1bf81d9d27306/700/463/wortv3/99beb6415443f751c7522b54b7a4911782e7e827)

This view (same credits as above) is a little to the left of the first. Assuming the leading carriage started off around the same length as the rest, it has been very badly telescoped indeed.

It appears that this Alstom 3-car train (no. 2211), as well as one (no. 2207) involved in another collision nearby (Zoufftgen) in 2006, are almost identical to the SNCF TERs of type Z 24500. That case too involved a head-on collision with a French goods train, and destroyed the leading carriage, but there were six fatalaties.

That's pretty appalling crash performance for a modern vehicle - are prototypes not crash tested before full production begins as per with road vehicles? I'm amazed more people weren't killed in that. No indication of the speed involved, but I assume it was relatively low speed as most of the rest of the wagons & carriages appear to upright and aligned with the track.

We'll have to wait for more information about this accident, but the similarities with the one ten years ago are remarkable. The inquiry into that was done jointly by BEA-TT and its Luxembourg equivalent  EEAI (now AET), and the report is on BEA-TT's web site. I've been trying to extract some relevant points from that, with some difficulty - it's 136 pages long, and full of strange railmen's terms, all in French but different for the two countries.

The collision speed then was estimated as 78 + 79 km/hr. It was noted that both trains were modern, and built to survive impact, but only up to 36 km/hr with something static. That front carriage was reduced from 27.35 m long to 7.6 m. Given the very large energy involved, the damage was considered to be less than might have been expected.

The 2006 accident was frankly disgraceful. Most of the blame was attributed to the Luxembourg signalling system and staff, and I can't really think of a word than sums up just how bad its performance was. I'll try, as far as I can, to give you the summary of causal factors.

The background was the closure of one track every day for three weeks, with the other track used for both directions. This stretch is controlled from Bettembourg in Luxembourg, and is signalled bidirectionally, so in theory this use should be quite safe. The accident happened just as the staff were changing shift at midday, which appeared to be rather causally done. Hence the goods train en route from Thionville in France to the goods yard at Bettembourg got "forgotten".

The next train to go - as agreed with the French signallers - needed a route to be set, but the interlocking wouldn't allow this (correctly). After two attempts, the head of the signalling centre concluded this was a fault and issued the driver with an order to pass the red signal. He set off, into what was fairly dense fog.

At least one signaller in another control post saw this and rang the Bettembourg centre to remind them the line was not clear. There then followed a couple of minutes of panic while the signallers tried to contact the drivers by radio - which failed - or turn off the power - done too late to help.

So the main causal factor was the issuing of that mistaken order. But:

Factors leading to that order included: the ineffective handover between shifts (mandatory written records not made), the high frequency of signal faults at that centre, the poor quality of the operating instructions, and the layout of the mimic panel which could have been much clearer.

Factors preventing a warning (or order to stop) reaching the drivers included: the signaller operating the alarm button was doing it wrong (i.e. no alarm was produced), the delay before trying to turn off the power, the lack on any direct communication link with the electrical controllers in France, and the limited number of phone lines from the centre.

Factors relating to the signallers competence: insufficient knowledge of non-routine procedures, notably for issuing "pass at red" orders or for handling emergencies, and the absence of any rehearsals of emergencies.

Factors relating to CFL's organisation and operating rules: the unrealistic division of tasks between the head signaller and his three assistants (who were meant to await his instruction or permission before doing almost anything), the lack of any process to invite feedback from operating staff, the lack of discipline among operating staff, and the rigid hierarchy.

In addition, if the cab radio alarm call had been made, it wouldn't have worked. The Bettembourg centre's radio control unit had a fault, as did the radio in the goods locomotive (which was reported but the driver was told to finish his run anyway).

If any of those factors is still there, and implicated this time, I can see a load of senior job vacancies coming up in CFL - and the government too, perhaps.

People complain that our railway costs a fortune in comparison with those in Europe. But some of those costs appear to be because we have a railway that is the safest in Europe.  We have specified trains that don't collapse on impact, and they very rarely do any more these days. Hence why later this month the National Rail network will have gone 10 years without a passenger fatality caused by a crash. How many other rail networks can boast such a record?

There do seem to have been a number of serious crashes in Europe over the last few years that have been directly attributed to incompetence or gross negligence of staff. And as demonstrated, some of their trains don't appear to hold up too well in collisions, although I don't think I'd like to be in a Pacer in a heavy shunt...

That said, UK railways aren't squeaky clean (although I am not fundamentally disagreeing with you!) - don't forget West Coast Railways - they have come close to causing a major incident or two in recent years, pure luck that a high speed collision between a steamer & HST was avoided at Wooton Bassett. Considering the involvement of a steam pressure vessel and old Mk1/2 rolling stock, I am quite sure multiple fatalities and numerous injuries would've been the result had the timing been a little less fortuitous, so I do believe there is a significant element of luck behind that 10 year record.

I'm sure that the standards are the same here as elsewhere in Europe, though how far back that's true I don't know. The current standard, as called up in the relevant TSI, is EN 15227:2008 +A1:2011 - not public of course.

Hitachi say this about the IEP design: "The crashworthy structure for the Class 800/801 is a further development of the technology used for the Class 395 rolling stock. In addition to being lighter and taking up less space, it complies with the latest TSI, the EN 15227 European standard for collision safety, and the GM/RT2100 UK railway standard for strength."

Hitachi also illustrate that with little pictures, showing a head-on collision with a like train at 18 km/hr each (36 km/hr closing), with 40 mm vertical offset; with an 80 tonne wagon at 36 km/hr; and a 15 tonne lorry at 110 km/hr. Note that a with-like head-on collision has the same effect as hitting a solid object, while running into a stationary like train is the same but at half the closing velocity - so 18 km/hr into a solid wall is as severe as 36 km/hr closing with another train.

As it says in the BEA-TT 2006 report, passive safety is not expected to cope alone with a head-on collision; for that you also need active (procedural and engineered) safety.

[pedant]Well it is public of course - anyone can buy a copy - just not free![/pedant]