Great Western Coffee Shop

From the Local (https://www.thelocal.fr/20200305/french-tgv-driver-hurt-after-derails-on-strasbourg-to-paris-line):

Reported as 30 km out from Strasbourg towards Paris, and running at 300 km/hr with no chance to brake. There are deep cuttings, and the question will I think be why such earthworks have failed when they are new (but not so new as not to have settled).

Embankment failure, land slipped down slope derailing train, see photos from Twitter

https://twitter.com/infosfrancaises/status/1235496682944712704?s=21

Wow - that's a pretty enormous rotational landslip. We have though had an exceptionally wet winter, so I guess the original designers and engineers might have some excuse. I do not though recall anything similar on any (relatively) newly built lines in Western Europe during the even wetter winter of 2013/4.

...and probably a relatively modern and supposedly properly engineered cutting slope?   

I dread to think what skeletons the Victorians left buried under our original earthworks, and I suspect Network Rail will often have no idea either...

Paul

Quite so. Construction of the LGV began around 2012, I think.

I do understand that soil mechanics is not the same kind of exact applied science as some. When looking at something built long ago the question of what materials are you dealing with underneath what you can see (or probe) is always there. If you build it yourselves, there will still be a question about what's underneath the deepest level you made or got to look at. And then there's groundwater, which wasn't there when you looked in any case. But I would hope that a modern programme like this one would be better able to predict such things, or to monitor what's going on if you still don't know, on a slope of this size and obvious hazard potential.

The science is well developed - though I must admit I always found the maths a bit of a challenge - it is the variability of the materials and the difficulty of investigating them that is always the problem.  Clay is a particularity complex material.

I our part of the UK we are used to the need for embankment repairs along the GWR route.  In the early 1840's for about 18 months the western terminus of the line from Paddington was a station then called Faringdon Road (later called Challow), This was because of the engineering challenge of building the clay embankments from there, through Swindon to Chippenham.  The line to Kemble was similarly found to be a challenge that nearly delayed its opening. 

One famous tunneling engineer was once asked to say what a good soil investigation would be for a tunneling project.  His reply was that the only good investigation would involve digging an exploratory tunnel on the line of the proposed tunnel - and of twice the size of the proposed tunnel!

I can't see any new news today, but there's a longer piece in French from last night on Franceinfo (https://france3-regions.francetvinfo.fr/grand-est/bas-rhin/strasbourg-0/temoignage-deraillement-tgv-strasbourg-paris-vitres-se-sont-fissurees-ont-explose-1795127.html), and a gallery of pictures on Dernières Nouvelles d'Alasace (https://www.dna.fr/techniques/2020/03/05/diaporama-une-coule-de-boue-provoque-le-deraillement-d-un-tgv-colmar-paris).

Looking at those pictures, it strikes me that this was a lucky escape from a much more serious crash. What you can see (with hints from other sources) is that the front of the power car was thrown sideways onto the other track, and its rear end pushed off the track but not by far. Presumably the vehicle itself acted as a plough to push the soil back off the track, and the first half of the carriage rake derailled but also cleared the track so the second half did not.

The leading power car kept pointing, and thus running, straight ahead almost to the end - why? How? It's true that both bogies were straddling tracks, which provides some guidance, but that is limited. With a large retarding force at each bogie, and the train pushing at only one side, the moment turning the power car further out of line could easily exceed those guiding forces. Turning only a little further would have seen it dive off among the OLE supports; inevitably that's a much more serious kind of accident.

The torsional rigidity of TGV carriage rakes is well attested by previous derailments, and is attributed to the use of Jacobs bogies. However, that's misleading - only one of the two chassis is solidly pivoted on the bogie, the other one is supported by an eye on a hook on the upper shell of the bogie mechanism. (That, at least, was true with older TGVs; I think it still is with this Euroduplex.) I assume there are mechanical links that restrain it hidden in the inter-carriage gap. (The outer body skins extend to cover most of that gap).

However, the joint of the power car to the leading carriage is more conventional, and if you look at the pictures there's no sign of extra linkages at that point. You can see, because that coupling is about 1 m out of line. So this may have been closer to major disaster than everyone thinks.

One other point I noted is that the train stopped very quickly - 1600 m from 292 km/hr. That comes to over 2.6 g, and on wet rails too, so I can't believe it was down to rail brakes. I think these train have eddy current brakes, and those would likely be used in an emergency brake application. That only works if the bogie is on the track, but then it has the side effect of pulling down quite hard on the bogie - which can only help. However, even that doesn't sound like enough, so the extra drag of the derailed bogies must have been substantial, and the forces turning the power car likewise.

Here are a few points from yesterday's reports I held back from the last (ever-lengthening) post:

There was a large group of railway staff in the front carriage, presumably operations management as they "put on their hi-viz and took charge".

There was some SNCF comment about the embankment at yesterday's news conference: that is was closely monitored (yes, but..) and probably the collapse started with a split at the top.

There had been five earlier trains that day, including two (one each way, I presume) empty ones to prove the line. That was said as if it happens every day, though it could also be only after works on the track overnight. That should perhaps be seen in the context of high-speed services (in France, at least) having a rather long overnight shutdown.

How long does it take to remove a bent TGV? Four days, in this case. The first two days were spent bringing a rail crane, taking down all the OLE wires so it could work, and bringing all the other bits and bobs re-railing calls for. In this case one of those bits (or more likely two) is a new bogie, as the leading one at least was, unsurprisingly, too wrecked to even drag. Then the power car was put on its (new) bogies, and hauled off to the nearest siding at a literal walking pace. The three derailed carriages were then put back on their track, I'm pretty sure without bogie changes - hard to do in an articulated train. Remember that removing that train from within Marseille-St-Charles took two days, a carriage at a time. Only then does the track inspection start ... so it'll be a long job still. At least it was plain line!