France - railways, public transport, services and incidents (merged posts)

[whole diary]

[Brucey]

I won't post the whole story here as it is constantly changing, however the BBC^» (British Broadcasting Corporation - home page) are currently reporting 6 people have died in a train crash near Paris.

http://www.bbc.co.uk/news/world-europe-23294630

SNCF^▸ (Societe Nationale des Chemins de fer Francais - French National Railways) have posted many photographs on their Twitter account: https://twitter.com/SNCF_infopresse

[stuving]

Watching French TV, the reports are still garbled - even those commenting on SNCF^▸ (Societe Nationale des Chemins de fer Francais - French National Railways)'s footage.

The best I can make out is that the train consisted of loco plus 6 carriages, the first two followed the loco through the station, and the third split from the second and all the last four derailed. The order of these events, and cause and effect, are not at all clear.

The third car was almost smashed to bits, and the fifth ended up sideways across a platform with the sixth on a different track altogether.

This aerial video gives an impression - the loco and first two cars are out of shot, perhaps moved.
http://www.francetvinfo.fr/faits-divers/accident-de-train-a-bretigny/video-bretigny-images-aeriennes-de-la-catastrophe-ferroviaire_368968.html

[stuving]

I've just seen a better aerial view, which shows two more cars so eight in all:
1,2,3 upright (but 3 may be derailed) and 4 on its side followed the loco
5 severely damaged on the wrong track
6, 7 on top of the platform canopy
8 on the original track.

This clearer video shows cars 3 and 4, though it's hard to see what they are.
http://www.francetvinfo.fr/faits-divers/accident-de-train-a-bretigny/video-bretigny-images-aeriennes-de-la-catastrophe-ferroviaire_368968.html

Clearly an event of great violence.

[TonyK]

Terrible.

The story accompanying that video says it is the worst accident in the past 5 years. It says also that the train left Paris a little early, but does not say whether that may be important. There has been a Red Plan activated, with fire service, ambulances, and helicopters, and the whole of the staff of all hospitals in the Ile-de-France region have been put on standby. That region is bigger than Greater London. There are many tens of people injured, some very badly, some lightly.

The Twitter feed says that three inquiries have been opened, by the police, the transport ministry, and SNCF^▸ (Societe Nationale des Chemins de fer Francais - French National Railways) internally.

[Red Squirrel]

Echoes of Potters Bar 2002..?

[LiskeardRich]

Echoes of Potters Bar 2002..?

This was my thoughts as well, has France been having our heat wave? Hear of speed restrictions because of rail contracting in heat. This may be a result of said effect?

[TonyK]

I thought of Potters Bar too, but decided to keep it to myself because the similarity as things stand is in the effect, rather than the cause. But it is eerily similar.

There was apparently permanent way work done around that area recently. It has been said on Twitter that there are 8 confirmed deaths, but that the fire service have yet to be able to get to the worst damaged carriage. This is a holiday weekend in France, and this is an awful start to it.

[stuving]

Yes ... a lot of things have been mentioned - most are irrelevant.
I was watching French rolling TV news, where they were running the available pictures - basically the same stills and videos you can find on the net - and egging experts (or "experts") on to give them news that had not happened yet (i.e the discovery of what caused it). Almost no attempt was made to relate the words to the very confusing pictures (which was my main motivation in trying to piece that together).

Much of the "what might have caused ..." answers brought up the shortage of cash for "classic" lines, given the cost of LGVs^▸ (Large Goods Vehicle), until recently when RFF started to complain loudly; or the political arguments against separating off RFF, let alone outsourcing maintenance (though major works sometimes are).

From what I heard, the points that were replaced as urgent maintenance were on another line, and most reports said the middle of the train derailed first and then colliding with the station did most of the damage. There were estimated to be 370 passengers, hence the size of the emergency services call-out. Also hence the number of red herrings raised: e.g. in that number there is bound to be one passenger who says the train was going too fast, even if it was not.,

It's actually the worst accident for 25 years excluding level crossing accidents*, since French railways have a good record recently. I suspect that this is partly just random, but has given them a false sense of virtue.

A predictable part of the news reporting has been listing the major post-war rail accidents in France (mostly not familiar to me). Two struck me as worth noting:
A train that derailed on a viaduct due to buckled rails in the heat - very nasty, one carriage fell to the ground below.
The collapse of the roof of a tunnel, with two trains involved in it. this was the worst recent rail accident, with over 100 deaths. It might explain a greater sensitivity to the dangers of tunnels (which I noted elsewhere).

*revision: Allinges (described here http://www.firstgreatwestern.info/coffeeshop/index.php?topic=12589.msg135332#msg135332) was 5 years ago, with 7 deaths. However, it appears to be usual in France to exclude level crossing accidents (or deal with them separately) hence 25 years.

[stuving]

Further reporting this morning:

The official word is that the immediate cause is equipment not human - train speed was 137 km/h where the limit is 150.

The driver is being given credit for raising the alarm very quickly, allowing other trains to be stopped in time to avoid further collisions.

Within the common theme of "lack of maintenance/investment for over 30 years", the signalling has been cited, though none is older than 1970 and the main line is newer. There are four signal boxes within the Br^tigny area, with some equipment and trackwork as old as 1910! However, this includes operating freight yard, so again the oldest equipment will not be on or even near the main line.

The points before the station are on an embankment, and at that point there is a small bridge over a road (Av. Jules Marquis). A new, bigger, bridge has been built next to it,  apparently by insertion into the embankment without closing the line. The reports now say that subsidence/settlement at this point has been detected by a track measurement train* and urgent work was carried out on another track. The track relevant to yesterday's accident was inspected and passed on July 4th.


* These use a Mauzin - a dynamometer car with two or four extra axles under it, able to duplicate specific ride characteristics and even derailment itself (i.e. simulate with real wheels and track - none of this computer modelling nonsense). I found one of these at Saint-Gilles-Croix-de-Vie last month, but it sneakily avoided being photographed, and I have not found a decent still picture on-line - the best is half-way through this: http://www.youtube.com/watch?v=AlDvMZChpuM

[stuving]

I have just seen an SNCF^▸ (Societe Nationale des Chemins de fer Francais - French National Railways)/RFF/SNCF-infra press conference. Unfortunately the TV pointed their camera at the speaker not the illustrations he referred to, but they said that a fishplate (or part of one) came loose and lodged within a flangeway of the points. Work proceeds to find out what broke (or came loose, or was loosened?) first.

[Red Squirrel]

I note that some tabloid reports are suggesting that passengers were electrocuted. Is that likely? Wouldn't the power have tripped out pretty quickly as soon as a derailing train shorted the circuit?

[TonyK]

Thought so too, Red Squirrel. I'd have thought there would have been earth leakage circuit breakers, as well as residual current protection. But if one bit can fail... who knows?

Having a background which includes light aviation, I know that most "accidents" have at least three causes. We know from SNCF^▸ (Societe Nationale des Chemins de fer Francais - French National Railways)'s initial disclosure that there was a line fault, which is presumably the primary problem. We do not yet know what the other contributory factors are. This does get more like Potters Bar every day. I hope our railway (and especially our government) learn from the French mistakes as they are revealed, rather than making more of their own. This is not a criticism of SNCF or CoRail, or anybody French, more an expression of my own wish that all transport systems are safe, because they learn from any source available.

[Red Squirrel]

I'm always amused when I hear people mistaking the statement that aviation is the safest way to travel (which it may be) with the idea that it is intrinsically safe - which it plainly isn't. Aviation is safe because a lot of highly professional people work very hard indeed to keep it that way. It is harder to imagine a more intrinsically fail-dangerous form of transport than one that involves travelling 6 miles up in something that will only stay there if it keeps moving very fast.

I worry that the kind of highly-evolved safety systems that make aviation viable (let alone safe) are hard to apply to a railway organised as it is - the essence of aviation safety is no-fault reporting, yet the railways seem to have 'blame' written into every contract.

[stuving]

The big puzzle is still how a fishplate had four bolts holding it when inspected on 4th July, and eight days later was loose and jumping about.
The latest info says one of the bolts was still there, and the fishplate rotated, but was so close to the points it ended in the flangeway. SNCF^▸ (Societe Nationale des Chemins de fer Francais - French National Railways) Twitter photos in this report: http://www.rtl.fr/actualites/info/article/train-paris-limoges-les-circonstances-du-drame-se-precisent-7763126734

If someone undid it, then who? why? Could it have undone itself?

Electrocution was reported right from the start. The train mowed down several OLE^▸ (Overhead Line Equipment, more often "OHLE") supports, so the line may have dropped towards the ground before power was removed. In addition, this line is 1500V DC^▸ (Direct Current), not 25 kV AC, so the fault current to trip the breakers will be much higher.

[Electric train]

Thought so too, Red Squirrel. I'd have thought there would have been earth leakage circuit breakers, as well as residual current protection. But if one bit can fail... who knows?

Earth leakage devices are of no use on an ac rail traction systems as they have earthed return system; in the case of dc electrification systems they have an earthy traction current return system even though there is no direct connection between the rectifier negative and earth the fact the rails and ground are close enough in potential not to case a problem.  The reason for not having a direct earth connection is to manage stray dc return current , also extraneous metal work like signal posts, lighting columns etc are not bonded to traction return if they are earthed to other electrical systems they are often screened by an insulting barrier.

Electrocution was reported right from the start. The train mowed down several OLE^▸ (Overhead Line Equipment, more often "OHLE") supports, so the line may have dropped towards the ground before power was removed. In addition, this line is 1500V DC^▸ (Direct Current), not 25 kV AC, so the fault current to trip the breakers will be much higher.

Not sure what the fault current on the French 1500V dc is, generally on traction systems the protection is there to protect the equipment and not to protect protect persons from harm, wiring regs for buildings (eg BS7671) have a emphasise to protect persons from harm.  It is possible that the French operating procedure is to re-close circuit breakers after a tripping which could mean if the wreckage was clear of the OLE and the OLE remained clear of structures it could have been recharged.