"I hope no one's been killed because they'll be on my conscience"
July 25, 2013 3:36 PM   Subscribe

Spain has declared three days of mourning following a deadly train crash that killed at least 80 people and injured many more.

The train was travelling at 190 km/h (118 mph), more than double the posted speed limit, when it derailed near the city of Santiago de Compostela. The drivers are now under investigation. One driver, Francisco Jose Garzon, has been a train operator for 30 years, and helped rescue victims after the crash.

Video footage shows the train derailing as it comes around a curve (may be disturbing to watch; no injured people are visible in the video but graphic photos are shown in the news articles, particularly the National Post story).
posted by randomnity (88 comments total) 7 users marked this as a favorite
 
While trapped in the cab, the driver was reported to have given an account over the radio to officials at Santiago station. He was quoted saying, "I hope there are no dead because they would fall on my conscience" and having repeated over and over: "We're human. We're human."

Too err is human and this seems like a horrible human error.
posted by three blind mice at 3:49 PM on July 25, 2013 [1 favorite]


News tonight said one of the drivers was quoted as "If I drove any faster they would fine me." He bragged about hugging the speed limits. It was a bit odd to me, since if he stays within the limits, then who cares if he pushes right up against them? So a weird thing to comment on. In this case the limits weren't being obeyed.

This is basically a story of what happens when you skimp. Spain wanted high speed rail, but didn't want to pay for the correct infrastructure, so they just trust people to slow down. Not a good plan.
posted by cjorgensen at 3:49 PM on July 25, 2013 [1 favorite]


So other high speed rail systems have automatic limiters?
posted by mecran01 at 3:51 PM on July 25, 2013


.
posted by Joey Michaels at 3:53 PM on July 25, 2013


cjorgensen: "This is basically a story of what happens when you skimp. Spain wanted high speed rail, but didn't want to pay for the correct infrastructure, so they just trust people to slow down. Not a good plan."

One of the linked articles mentions that failsafes that were in place should have limited the speed of the train, and that despite the Spanish government's austerity measures, the track maintenance budget had not been reduced.

Do you have information that contradicts this?
posted by danny the boy at 4:00 PM on July 25, 2013 [4 favorites]


From what I've read, the track had a spanish system that only enters in action above 200 km/h, so it didn't activate. The more expensive euro system might have limited the speed of the train.
posted by Monday, stony Monday at 4:07 PM on July 25, 2013


So other high speed rail systems have automatic limiters?

They make the tracks more able to handle the speeds, so more gentle curves. To retrofit the tracks would have required laying new track. Instead they put fast trains on tracks designed for slower trains and expected the drivers to back off the speeds.
posted by cjorgensen at 4:12 PM on July 25, 2013


. (×80)

I'm amazed at the huge death toll in this crash compared to the only very slightly slower Grayrigg Derailment in Britain. It seems very odd that colliding with the concrete wall parallel to the track could be as disastrous as the Japanese train crash direct into an apartment building. I wonder if the unusual bogie design is as crashworthy as a conventional axled bogie on an articulated trainset.

This is the kind of thing that driver error alone shouldn't be able to cause and a huge. That said, there's no evidence that RENFE or ADIF bear any direct responsibility, and overspeed crashes on curves happen on all speeds of rail network.
posted by ambrosen at 4:12 PM on July 25, 2013 [3 favorites]


.
posted by longdaysjourney at 4:14 PM on July 25, 2013


I am by no means a train expert. I was just regurgitating what was on the news.
posted by cjorgensen at 4:14 PM on July 25, 2013


why do train drivers even exist anymore. this is so easy to automate.
posted by bhnyc at 4:17 PM on July 25, 2013 [2 favorites]


Early reports were that the train was doing 120 MPH in a 50 MPH zone. When you watch the train (fail to) make that turn, it sure does look like it just fell off the track because it was going too fast.

I imagine that engineer is going to spend the rest of his life behind bars.
posted by Chocolate Pickle at 4:20 PM on July 25, 2013


cjorgensen, it was running on track that was layed for high speed trains, as it was running at standard gauge, not Iberian gauge (it uses variable gauge bogies, which is one of my other queries about crashworthiness), but the alignment had curves that are much sharper than usual on high speed track. It's definitely not unusual to use existing infrastructure with speed limits in built up areas rather than building out all curves, especially in a small historic city like Santiago.
posted by ambrosen at 4:20 PM on July 25, 2013 [2 favorites]


Was this train going fast because the engineers were hotdogging or was it going fast because management gave them a schedule that was impossible to comply with otherwise? From what I know about how the trucking system operates in the US the latter seems at least plausible.
posted by ishrinkmajeans at 4:24 PM on July 25, 2013


When Gallegos cancel the festivities surrounding the primary engine of tourism in their region, you know it's a bad deal. Santiago, Apóstol de España, ruega por nosotros.
posted by resurrexit at 4:25 PM on July 25, 2013 [1 favorite]


"We're human. We're human."

Sounds like a confession. To my reading, the engineers may have been joy-riding, egging each other on perhaps, led along by a senior member with a penchant for pushing the boundaries. I mean, 120 in a 50 MPH zone? There must be culpability on grounds of recklessness here. To say "we're human" is nothing less than to say "our decisions led to this."
posted by troll at 4:27 PM on July 25, 2013 [1 favorite]


Here's the thing. This is a trivial signal problem. Shinkansen, TGV, Acela -- they all deal with situations where the allowable track speed is much less than what the train can do.

And how do they handle it? With in-cab signaling that, if not followed, automatically stops the train. This isn't new. The Chicago L has this - since 1974.

The very fact that he got within a kilometer of an 80 kph zone doing 200kph shows that there was a massive human failure. Some human desinged a completly inadequate signal system -- for 1980 -- in 2008.

The fact that some other human blew through that in 2013 pales in comparison. If that track was limited to 80kph, why the fuck wasn't the train emergency braking half a kilometer before the curve if it was doing 200?
posted by eriko at 4:27 PM on July 25, 2013 [42 favorites]


It's a national tragedy. I wish I could get that stupid SNL sketch out of my head though.
posted by BrotherCaine at 4:27 PM on July 25, 2013


.
posted by Lynsey at 4:32 PM on July 25, 2013


I'm similarly baffled by the technically trivial safety systems not in place, both here and in the Lac Megantic negligent. (Accident is a bullshit word for something that could have been prevented.) it looks like the companies involved in the Quebec disaster are going to be sued into the ground, though.
posted by seanmpuckett at 4:33 PM on July 25, 2013


Renfe's online credit card processing only works when the bulls shit on a certian corner of a field in Aragorn. Maybe their web team configured the limiters for that segment of track?
posted by jeffburdges at 4:43 PM on July 25, 2013 [5 favorites]


This is an analysis of the automatic train controls available on the Spanish high speed and legacy networks. There was an older cab signalling and automatic braking system active on the track where the crash happened, and much of the AVE network does have higher levels of automation. Ideally, of course, all railways would be fully protected, but such systems are complicated and integrating them fully into a rail network may well not pass a cost-benefit analysis. Until the worst happens, of course.
posted by ambrosen at 4:50 PM on July 25, 2013 [3 favorites]


bhnyc: "why do train drivers even exist anymore. this is so easy to automate."

Train drivers generally exist to push the big red button when they see something that has gone horribly, horribly wrong. The 2009 collision on the DC Metro would have been much, much worse, had the driver not slammed on the brakes a few seconds before the crash. Also, it's still prudent to have a human visually confirm that a train may safely arrive or depart from a station platform. The DC Metro, the NYC L Line, and portions of the London Underground can all operate fully-autonomously, as long as there's a human there to press "Go." I'm sure there are many other systems with similar levels of automation.

mecran01: "So other high speed rail systems have automatic limiters?"

Many "low-speed" rail networks also now have similar systems (under the guise of "Automatic Train Control (ATC)", "Automatic Train Protection (ATP)," or "Positive Train Control." This is especially common on urban metros and congested commuter systems. Even the freight railroads in the US are beginning to implement these technologies on a wide scale.

*For all the ways that the US is backwards on rail transportation, we take safety very seriously. A few of these guidelines (such as the weight minimums and disregard for articulated railcars) are misguided, but we have a very impressive track record to show for it. A few months ago, two (brand-new) Metro North trains collided head-on outside of New Haven, CT at ~70mph. This should have been a horrific, deadly accident. Instead, the trains were equipped with crumple zones (similar to cars, but quite new to the railroad world), and there were only a handful of injuries. Yes, the accident should have been preventable (again, the US doesn't put quite enough focus on crash prevention), but it was a pretty solid affirmation of our safety standards. This technology is very, very effective.

Every HSR system that I know of uses "cab signaling," which essentially relays any speed limits or stop signals onto a computer in the driver's cab. This is done, because human train drivers literally have no time to observe or react to "traditional" signals (ie. lights or signs on posts on the wayside), because they zoom by too quickly.

Typically, this computer won't allow the operator to disobey the signals, unless a specific override command is given (in the US, doing this will also force the railroad to fill out a huge pile of paperwork, and trains deliberately disobeying a signal usually still have their speeds automatically restricted by the computer).

According to The Guardian, the train in question should have been equipped with cab signaling and an electronic speed limiter. No clue why they weren't working. I'm sure we'll learn more in the coming weeks...

ishrinkmajeans: "Was this train going fast because the engineers were hotdogging or was it going fast because management gave them a schedule that was impossible to comply with otherwise? From what I know about how the trucking system operates in the US the latter seems at least plausible."

If the reports are correct, 120mph on 50mph track is suicide. Nobody in their right mind would obey this kind of order. This accident was particularly bad, because the train was rounding an embankment. This kind of speed doesn't even make sense, because the curve was apparently just outside of a station.

Even though signal problems are the cause of most accidents like this, the circumstances don't add up in my head. If I were investigating this accident, I'd be strongly considering the possibility of foul play.

I should also note that 120MPH isn't quite HSR. Qualifying this as a "high speed rail" accident is something of a misnomer, given that the train was traveling at speeds less than HSR speeds, on low-speed track. Almost all of Amtrak's rolling stock is qualified to run at 125mph, because that's the speed limit on the Northeast Corridor.

I'm not a rail expert and am not intimately familiar with the details of this accident. Take that for what it's worth.
posted by schmod at 4:51 PM on July 25, 2013 [33 favorites]


If the train was going 2.5x the max allowed speed why was the train still operating.
If the train was going 2.5x the max allowed speed why was the train still operating.
If the train was going 2.5x the max allowed speed why was the train still operating.
If the train was going 2.5x the max allowed speed why was the train still operating.
If the train was going 2.5x the max allowed speed why was the train still operating.
....75 more times...
posted by Cosine at 5:25 PM on July 25, 2013 [9 favorites]


Every time there's a train crash, I know my uncle's packing his suitcase. He spent his career running signalling in Scotland, and after retirement, is part of a team that investigates the root cause of rail incidents. As one of the signalling investigators, he usually has the toughest time, since it's nearly always signalling: either if failed, or was ignored, or over-ridden, or ... he's seen it all.

He takes long vacations, and although impossibly happy to talk at length about everything to do with rail, he doesn't like to talk about the accidents.

.
posted by scruss at 5:33 PM on July 25, 2013 [16 favorites]


The fireman screams and the engine just gleams.

It's a little early to say if this was human error, mechanical failure, or a bit of both, and certainly the design of the track seems to be implicated (I think I read last evening that the train's speed was normal for most of the route, just not that particular bend -- why didn't they redo that bend? I'm sure the Spanish people and government are looking into everything right now). Anyway, as an American, I don't feel qualified to criticize other countries' infrastructure. This is just very sad.
posted by uosuaq at 5:44 PM on July 25, 2013


On the driver's Facebook page he had publicly posted an image of his speedometer at 200km/h, joking about how he wished he could set off a police speed camera.

'One of his friends commented: “You are going way too fast – Braaaaaaake!!”'

This is evidence in favor of manslaughter. It's also evidence that high-speed rail is very safe: this dangerous imbecile had been operating for years without incident.
posted by East Manitoba Regional Junior Kabaddi Champion '94 at 5:48 PM on July 25, 2013 [10 favorites]


This is evidence in favor of manslaughter. It's also evidence that high-speed rail is very safe: this dangerous imbecile had been operating for years without incident.

But it's also unbelievable that the system would allow that to happen. The signalling system for France's TGV, for instance, automatically stops the train if it's going more than 10 km/h over the speed limit for the block.
posted by junco at 5:58 PM on July 25, 2013 [1 favorite]


This all looks very bad, but I would offer two cautionary comments: 1) reading too much into what someone says in the immediate aftermath of surviving a horrific accident like this would be foolish. 2) It if amazingly common for the early reporting on disasters like this to all point towards some particular cause which turns out to be entirely irrelevant on further investigation.
posted by yoink at 6:10 PM on July 25, 2013 [15 favorites]


The recent Asiana plane crash looks like pilot error.
The Southwest landing gear collapse looks like pilot error.
This train derailment looks like driver error.

In none of these cases am I going to go anywhere near claiming we know what caused them before I read the investigative authority's report, much less invoke words like manslaughter or start making claims about the pilot/driver's nationality or their national infrastructure. Completely uncalled for.

(on preview: seconding yoink)
posted by kiltedtaco at 6:13 PM on July 25, 2013 [5 favorites]


Such a tragedy. Apparently the train was going too fast as it was exiting from a stretch that allowed higher speed limits. Now if only my friends would reach me; the government has not released the list yet.

This partial listing of bios is heartbreaking (in Spanish).
posted by ersatz at 6:14 PM on July 25, 2013


Spanish friends talking about this last night commenting they were wondering if it'd make it into the media. Similar incidents allegedly having gotten very little coverage in the past.
posted by yoHighness at 6:23 PM on July 25, 2013


It wouldn't cost much for every train cab to be equipped with DriveCam and drivers informed that DriveCam footage is reviewed on random scheduling.

I'm not talking about a high tech black box. I'm talking about the camera they put in US truck fleets.

A policy like that would keep the drivers following mandatory procedures and keep them from texting, speeding, talking on the phone, surfing the web, etc.
posted by surplus at 6:27 PM on July 25, 2013


You're never going to be able to impose sufficient technological safety measures to counteract determined idiot drivers. I have every sympathy for the lowly employee who follows all the rules and guidelines only to discover that they were insufficient to prevent an accident and then gets blamed for everything. But this guy (these guys? which one was actually driving?) was being reckless, careless, negligent, and ended up killing most of his passengers as a result.
posted by ceribus peribus at 6:30 PM on July 25, 2013


was being reckless, careless, negligent, and ended up killing most of his passengers as a result.

I really don't think we know enough to say this with absolute certainty yet.
posted by yoink at 6:39 PM on July 25, 2013


Totally agree with you; just going off the initial reports. Maybe the complete picture, like ishrinkmajeans suggested, was that the driver was being pressured to rush to meet deadlines. But if he was rushing just for bragging rights then he needs to be permanently banned from operating heavy machinery.
posted by ceribus peribus at 6:46 PM on July 25, 2013


No, we don't know why happened, but the driver has certainly given them more than enough rope to hang him.
posted by double block and bleed at 6:48 PM on July 25, 2013


why do train drivers even exist anymore. this is so easy to automate.

Even if it did require any money to automate all of Spain's rail lines, the Spanish government would surely be able to spare some of its €117 billion budget surplus.
posted by one more dead town's last parade at 6:59 PM on July 25, 2013


Was this train going fast because the engineers were hotdogging or was it going fast because management gave them a schedule that was impossible to comply with otherwise?

Train schedules are an entirely different thing from trucking schedules. They're going to be determined by the speed limits, for one thing, and even more likely from test runs if the trackage is new or they're trying to tighten them up at all. I would be utterly aghast if the schedule were in any way to blame here -- there are just too many people who have to sign off on something like that. It isn't a boss barking "Get to Seville by 3 or you're canned!" Just doesn't work that way.

That said, it is possible that a number of coordinated short-cuts -- in design, in construction of the train or trackage, in training of the engineer -- could be factors. It's likely that more than one will be found to be in play; that's pretty much de rigeur for airplane crashes, for example, or how commonly both drivers in a car crash are found to have violated rules of the road in some way.

It should be noted, though, that however many certified engineers were on board, only one would be driving.

I really don't think we know enough to say this with absolute certainty yet.

Absolute certainty, no, but it seems to have been enough for an arrest. I'll join you with caution as regards his supposedly "bragging" of being a speed demon on Facebook -- just posting a speedometer reading isn't evidence of anything, as it may have been a legal speed where he was moving (and on trains there isn't any good reason to run below the speed limit).

Qualifying this as a "high speed rail" accident is something of a misnomer, given that the train was traveling at speeds less than HSR speeds, on low-speed track.

Nevertheless, it was an Alvia HSR train. It looks very much like a RENFE Class 130 engine. Talgo, the builder, is the company that tried to set up shop in Wisconsin to build HSR for the Amtrak Chicago-Milwaukee-Madison route, and got summarily kicked out by our idiot governor; they already supply trains to Amtrak in the PNW, and hope to win some of the US contracts. Ah: I see now they're saying it was a hybrid version for dual operation.
posted by dhartung at 7:14 PM on July 25, 2013 [1 favorite]


On the driver's Facebook page he had publicly posted an image of his speedometer at 200km/h, joking about how he wished he could set off a police speed camera.

'One of his friends commented: “You are going way too fast – Braaaaaaake!!”'

This is evidence in favor of manslaughter. It's also evidence that high-speed rail is very safe: this dangerous imbecile had been operating for years without incident.


I disagree. It's not as if he posted that photo just as he was rounding the bend into the crash, and it's been stated repeatedly that the speed shown in the photo is actually the speed limit for much of the track. One can absolutely read that message as a sort of "look at how fast I get to go... Imagine the fines if you tried this on the roads!" thing.
posted by polymath at 7:15 PM on July 25, 2013 [19 favorites]


I really don't think we know enough to say this with absolute certainty yet.

If people were only saying things they knew with any degree of certainty or with appropriate background knowledge in this thread, there'd be about three comments.
posted by hoyland at 7:26 PM on July 25, 2013 [5 favorites]


It's a national tragedy. I wish I could get that stupid SNL sketch out of my head though.

Um, did you mean this SNL sketch?
posted by Horace Rumpole at 7:31 PM on July 25, 2013


I disagree. It's not as if he posted that photo just as he was rounding the bend into the crash, and it's been stated repeatedly that the speed shown in the photo is actually the speed limit for much of the track. One can absolutely read that message as a sort of "look at how fast I get to go... Imagine the fines if you tried this on the roads!" thing.

Yeah, I regret posting that and jumping to conclusions.
posted by East Manitoba Regional Junior Kabaddi Champion '94 at 7:33 PM on July 25, 2013 [3 favorites]


i was just on a high-speed train in spain. The one I was on was going 300kph, though, from Barcelona to madrid. It's a lovely travel experience.
posted by empath at 7:58 PM on July 25, 2013


hoyland: "If people were only saying things they knew with any degree of certainty or with appropriate background knowledge in this thread, there'd be about three comments."

You say that like it would be a bad thing.
posted by danny the boy at 8:00 PM on July 25, 2013 [2 favorites]


The comment about being human could have been about their inability to prevent what was happening without their control. I hope that comment doesn't wind up carrying any weight.
posted by bleep at 8:07 PM on July 25, 2013


Train schedules are an entirely different thing from trucking schedules.

The other side of this observation is also baffling: Any significant overspeed will be obvious by the time the train arrives at the end destination, and may cause problems by messing up the schedule for other trains using the same section of track or passengers expecting to meet the train at intermediate stations. It is trivially easy to identify and take action against overspeeding drivers - trains just aren't supposed to arrive early. It would be a strange decision for the drivers to intentionally exceed not just the maximum speed the track is rated for, but also their expected speed for that part.
posted by Dr Dracator at 9:04 PM on July 25, 2013 [4 favorites]


It's been a hell of a summer for deadly train crashes so far: Quebec, France and now Spain. The immediate causes are different in all cases and it's way too early to draw any conclusions but good old 'human error' is, as always, a prime suspect.
posted by islander at 9:55 PM on July 25, 2013


I don't believe the driver has not been arrested or charged with anything; he is under police guard in the hospital until the train's black boxes are examined, and I believe they are still questioning him as well. They are being very careful about investigating this, and I think that's the right way to go (though it certainly appears the train was going far too fast for that stretch of the track).

I found this article with the railway safety records charted out by country really informative. Spain's acident rate is twice as high as the other countries in the EU.

When I read the first reports of the train derailment, the average rate of railway accidents in the EU even seemed high to me, but then I realized that the train system is very heavily used there (in Florida, hardly anyone rides a train at all, as we have so few operating here; high-speed rail is in the works to eventually connect the tourist routes between Orlando and the coast, but we lack the infrastructure for intercity transit, which is stupidly short-sighted).
posted by misha at 10:22 PM on July 25, 2013


Just saw the 'not' in the first sentence. Please disregard the double negative. I don't think the driver has been arrested yet.
posted by misha at 10:46 PM on July 25, 2013


misha, that link is borked -- can you repost?
posted by KathrynT at 10:47 PM on July 25, 2013


Hmm, what you said about the accident rate contradicts this article, which says:
Spain's rail safety record is better than the European average, ranking 18th out of 27 countries in terms of railway deaths per kilometers traveled, the European Railway Agency said. There were 218 train accidents in Spain between 2008-2011, well below the EU average of 426 for the same period.
posted by Emanuel at 10:51 PM on July 25, 2013


[Fixed Misha's link ]
posted by taz (staff) at 11:09 PM on July 25, 2013


Even if it did require any money to automate all of Spain's rail lines, the Spanish government would surely be able to spare some of its €117 billion budget surplus.
posted by one more dead town's last parade at 6:59 PM on July 25 [+] [!]


Uh, what part of Spain are you thinking of?

Spain’s Budget Deficit Worst in Euro Zone
posted by chavenet at 12:17 AM on July 26, 2013


Thanks, taz!
posted by misha at 12:20 AM on July 26, 2013


why do train drivers even exist anymore. this is so easy to automate.
posted by bhnyc at 12:17 AM on July 26


As an ex-IT bod who is now a train driver on one of the UK's busiest rail networks, I'm going to tell you that you don't know what the hell you're talking about. I could bang on at great length about why it is very, very far from easy to automate a complex, busy, high-capacity network (as opposed to a typical subway/metro line) but I know that doing that just doesn't stop this sort of lazy, semi-informed-at-best remark so eh, fuck it.

Train drivers generally exist to push the big red button when they see something that has gone horribly, horribly wrong.

Oh, is that what I trained intensively for a full year to do? Man, I must be slow.
posted by Decani at 12:49 AM on July 26, 2013 [41 favorites]


Misha wrote: Spain's acident (sic) rate is twice as high as the other countries in the EU.

Your statement is factually incorrect. If you go back and look at the chart, you will see that of all the European countries mentioned, eight have a higher fatality rate than Spain.

It is also worth noting, as the article does, that Europe has a very low fatality rate.
posted by Mister Bijou at 1:08 AM on July 26, 2013




Your statement is factually incorrect. If you go back and look at the chart, you will see that of all the European countries mentioned, eight have a higher fatality rate than Spain.

Accident rate =/= fatality rate.
posted by empath at 1:10 AM on July 26, 2013 [1 favorite]


I don't believe the driver has not been arrested or charged with anything; he is under police guard in the hospital until the train's black boxes are examined, and I believe they are still questioning him as well.

Yeah, it should be noted that Spain has a civil law system and investigating magistrates have broad powers that would be unusual in a English common law system. The closest analogue in a criminal case would be detention as a material witness. For the record, of course, the NTSB in the US has comparable powers.

There's confirmation, btw, that there was some sort of Positive Train Control or at least limiter technology in place: "Investigators were trying to find out why the train was going so fast and why security devices to keep speed within permitted limits had not slowed the train." So it wasn't that Spain merely left them off the budget to save money (although implementation/maintenance and other factors may yet be part of this), it was that they should have worked and didn't.
posted by dhartung at 1:19 AM on July 26, 2013


Accident rate =/= fatality rate

The chart Misha used for the statement is actually a fatality chart. Misha may have misread/misunderstood/confused "acident' and fatality.

Check it out.
posted by Mister Bijou at 1:19 AM on July 26, 2013


Ah.. well it is twice as high as the EU average. It's just that there are a bunch of countries with, I assume, less rail traffic who happen to have higher fatality rates.
posted by empath at 1:24 AM on July 26, 2013 [1 favorite]


ambrosen: "I'm amazed at the huge death toll in this crash compared to the only very slightly slower Grayrigg Derailment in Britain. It seems very odd that colliding with the concrete wall parallel to the track could be as disastrous as the Japanese train crash direct into an apartment building."
Looking at the pictures in the National Post article it seems some of the carriages telescoped. There's also a bridge overhead directly at the crash site, which brings to mind the Eschede disaster.
posted by brokkr at 1:32 AM on July 26, 2013


Sorry, i did flub that up, didn't I?

I should have just quoted the article I linked, "Spain’s passenger fatality risk is almost double the E.U. average."
posted by misha at 1:59 AM on July 26, 2013


I have an interesting old book about train wrecks. It shows horrific scenes of telescoped passenger cars, bridge collapses and spectacular crown sheet failures from the age of steam. At the conclusion the author suggested that we now live in a safer era. It is shocking and saddening to learn this is not necessarily the case.
posted by kinnakeet at 2:43 AM on July 26, 2013


ambrosen: "I wonder if the unusual bogie design is as crashworthy as a conventional axled bogie on an articulated trainset."
Jacobs-type bogies are said to be less disaster-prone on derailment than normal bogies. But anyway the different bogies probably doesn't matter when the train simply tips over into a concrete wall. Speed and centripetal force did this train in, not the wheels.

Here's a BBC news report from a 300 km/h TGV derailment in France, stating that "Observers say the TGV's low centre of gravity and construction, which links cars by shared wheel carriages, makes such trains more stable than traditional rolling stock."
kinnakeet: "At the conclusion the author suggested that we now live in a safer era. It is shocking and saddening to learn this is not necessarily the case."
Rail travel (in Europe, at least) is vastly safer than 100 years ago. Trains are routinely traveling at more than double the speed of a steam locomotive, with far more passengers carried over longer distances. Today, automated signalling, automated train routing and safety standards for transporting dangerous goods makes a repeat of e.g. the Quintinshill disaster extremely unlikely.
posted by brokkr at 2:48 AM on July 26, 2013


Decani - I'd actually be fascinated to hear more, if you've the time. I often wonder what's going on, hidden away in the cab.
posted by chill at 3:21 AM on July 26, 2013 [2 favorites]


The world's first, known, widely-reported railway casualty: William Huskisson (15 September 1830)
posted by Mister Bijou at 3:24 AM on July 26, 2013


Talgo bogies basically aren't Jacobs bogies though. There's a passive tilt mechanism attaching the carriages to the bogies above the centre of gravity of the carriage, plus some trickery with the axles that I don't fully understand (and seems not to be fully explained on the web). The point of an articulated train is that even when part of the train flies off the track, the rest of the train holds it upright. The coaches shouldn't have crumpled and stacked up as much as they did, in my amateur reading of this.
posted by ambrosen at 3:25 AM on July 26, 2013 [1 favorite]


ambrosen: "The point of an articulated train is that even when part of the train flies off the track, the rest of the train holds it upright."
Agreed, that's how it generally works if the train is derailed due to e.g. heat expansion or objects on the tracks. The problem here (have you seen the video?) seems to me to be that the train leaves the curve, everything holds together just fine, and then there's unfortunately a concrete wall in the way. Had there been an open field instead, it would probably have looked a lot more like Grayrigg.

I am not a train engineer or anything, I just have an unhealthy interest in armchair analysis of transport accidents.
posted by brokkr at 3:40 AM on July 26, 2013


> At the conclusion the author suggested that we now live in a safer era. It is shocking and saddening to learn this is not necessarily the case.

Horrific accidents like this are not evidence the past was safer. If anything, the fact an accident of this scale made international news is evidence that things have improved; if it was routine, it would not get this level of attention and scrutiny from the rest of the world.
posted by ardgedee at 4:29 AM on July 26, 2013 [7 favorites]


Fatality rate might not be as good as accident rate for comparison. A country's medical services and other outisde factors play into how many survive injuries.
posted by Ironmouth at 7:12 AM on July 26, 2013


My train experience is with north american freight trains and I can tell you one thing: do not make the mistake to over-simplify the running of trains. Rail traffic, passenger and freight, is a very complex operation with very complicated machinery; automation of any part of the system (and there is a lot of automation already in place) is extremely difficult and requires extensive investment and maintenance. Despite the recent horrific incidents in the rail industry, rail is very safe; however when something goes wrong it can often go very wrong.

Most railways have far-reaching safety management systems in place and the safe runnings of trains is always priority number one. Despite all this there will always be those who are willing to skirt or ignore the rules; to my eyes this incident in Spain would have never happened if the conductor and engineer had just followed the operating speed limits. Track speed varies greatly on any given stretch of track, for curves, grades, cities, slow orders, etc. It is the responsibility of the conductor and engineer to ensure that the train is operating at the correct speed at any given point in time. The question that I have at this time is what checks and balances are in place to ensure that the crew is following proper operating practices? Here in North America we have track side detectors and on-board wireless information systems that report back to the rail traffic center; I am curious to know what they use in Spain.
posted by Vindaloo at 7:22 AM on July 26, 2013 [1 favorite]


chill: "Decani - I'd actually be fascinated to hear more, if you've the time. I often wonder what's going on, hidden away in the cab."

Seconding this. Please.
posted by zarq at 8:29 AM on July 26, 2013 [2 favorites]


If you ever think people don't understand what you do in your job then I recommend speaking to a train driver. The disconnect between reality and the public perception is hilariously wide. I sometimes moan when people think my job is easy but few think they can just rock up and have a go.

Spoke to an underground driver once who said every time there's industrial action because TFL want to make them wear blindfolds or some shit The Sun make s a big thing about how much they get paid. This cause a recruitment frenzy followed six months later by a load of nubs burning out because they thought driving a train must be be like a big car without the steering wheel.
posted by fullerine at 9:00 AM on July 26, 2013 [1 favorite]


why do train drivers even exist anymore. this is so easy to automate.

The majority of what a driver does *isn't* easy to automate.

Automatically stopping a train in the circumstance of going over the speed limit into a curve (or other similarly dangerous situation) where the driver is negligent or incapacitated: that bit is easy to automate. And all indications are that it really should have been.
posted by chimaera at 9:10 AM on July 26, 2013 [1 favorite]


The majority of what a driver does *isn't* easy to automate.

Compared to self-driving cars (which already exist) it is ridiculously easy. A train doesn't even have a steering wheel. And as people have said above, many trains are already automated.
posted by bhnyc at 2:18 PM on July 26, 2013


Compared to self-driving cars (which already exist) it is ridiculously easy. A train doesn't even have a steering wheel. And as people have said above, many trains are already automated.

If you're not giving specific reasons, or can back up your statement with something better than a phrase like "it is ridiculously easy", then you're just conjecturing off the top of your head. Which is less valuable than utter nonsense.
posted by suedehead at 2:33 PM on July 26, 2013 [3 favorites]


Horrific accidents like this are not evidence the past was safer. If anything, the fact an accident of this scale made international news is evidence that things have improved; if it was routine, it would not get this level of attention and scrutiny from the rest of the world.

For instance, based on the annual rate from Wikipedia, there have been roughly 100 car deaths in Spain so far in July, versus 80 by train. Some of the road deaths probably didn't even make the local news.
posted by Homeboy Trouble at 2:49 PM on July 26, 2013


Compared to self-driving cars (which already exist) it is ridiculously easy.

This accident illustrates very clearly why you need several orders of magnitude more reliability in controlling a train rather than a car.
posted by Dr Dracator at 4:26 PM on July 26, 2013


I accept and believe the statement that trains are surprisingly hard to drive. I'm not challenging that. But I don't understand how this reconciles with the fact that unmanned trains actually exist.
posted by Bugbread at 5:20 PM on July 27, 2013


Here is an article on driverless trains, and this is the Wikipedia list of driverless trains: It looks like this is the way things are going, but most of the existing driverless trains are metro or urban light rail systems. It's probably easier to implement in a limited system, where everything is built and operated by a single company.
posted by Dr Dracator at 11:56 PM on July 27, 2013 [2 favorites]


BBC and CNN report that the driver, Francisco Garzon, was charged yesterday with 79 counts of reckless homicide. Train officials have stated that there were no technical problems with the train and that the driver should have started to slow the train 4km before the curve in the track where the crash occurred, following 80 km of straight track with a much higher speed limit. At 190 km/h, the driver was about one minute late in his attempts to slow the train.

Reuters and others have mentioned that the accident occurred at a section controlled by the Spanish safety system ASFA, which warns drivers of excessive speeds but does not directly alter train speeds. In contrast the newer European system ETCS, used elsewhere on the track, automatically slows speeding trains. A regular driver commented, "There is no safety warning for the speed, it's pure human factor, you have to slow down manually and you have no assistance in the cabin."

It looks like the death count has been revised to 79 people (including 2 Americans). 70 passengers remain in the hospital, 22 of them in serious conditon.
posted by randomnity at 11:49 AM on July 29, 2013 [1 favorite]


"There is no safety warning for the speed, it's pure human factor, you have to slow down manually and you have no assistance in the cabin"

If this is true, this is a huge design error. If you design a section of track that requires you to transition from 190km/h to 80km/h, you should have giant warning signs flashing in the cabin -- if not an emergency slowdown system.
posted by suedehead at 12:25 PM on July 30, 2013 [1 favorite]


It should have started slowing down one minute before the crash? Can someone clarify, according to the information so far: was the train's speed legal (or close to it) prior to that?

Having something as fundamental as train speed require human intervention within a minute would indicate a broken system, regardless of whether the driver was being reckless, malicious, or simply negligent. I'm hoping I'm misunderstanding the description somehow, and that this isn't a normal occurrence that happened to be missed in this instance.
posted by Riki tiki at 12:16 PM on August 5, 2013


The minute is not something that was reported, just my own calculation from the statement that the train should have started to slow 4km before the accident site, according to a driver: "When you exit the high-speed section you start slowing down ... you have like 4 km to the curve," and the rail company has stated the same distance. BBC reported that trains travel up to 220 km/h in the straight sections preceding the curve, although it wasn't completely clear whether the speed limit changed at that 4 km mark or if that was just where the first warnings were (following an earlier speed limit decrease).

Black box data shows:

1) 2 min before crash: train at 199 km/h
2) 1 min 50s before crash: driver is called by the (on-board) conductor to discuss details of the upcoming stop
3) sometime during the conversation: first and second acoustic warnings of "a sharply reduced speed zone ahead" were played in the cabin
4) 11s before crash: last words of conversation
5) moments before the crash: third warning; train at 192 km/h; brakes are applied
6) train derails at an estimated 153 km/h

At 190 km/h the train would travel 4km in about 1.3 minutes, assuming my math is correct and the news articles are accurately reporting the speeds and distances involved.

The driver stated in court, "I have no explanation, I don’t understand how I didn’t see it." Apparently the court has ruled that the phone call "could not be considered a cause of the accident."
posted by randomnity at 9:16 AM on August 6, 2013


3) sometime during the conversation: first and second acoustic warnings of "a sharply reduced speed zone ahead" were played in the cabin

I'm not that familiar with train control systems. However, as a designer, I'll say that the control systems should take into account the cognitive load of the driver into account -- so "acoustic warnings" are good, but ultimately not enough.

-- In cars/buses, you have a constant necessity for attention, since the driver is constantly slowing down/stopping/speeding up/turning/braking. Driving a vehicle engages the majority of the driver's attention, the vast majority of the time -- even if you're driving straight, you have to maintain control on the steering wheel. There's a medium-intensity, mostly-constant cognitive load placed on the driver.

-- In (large commercial) planes, you have two major moments of (mostly) manual control -- the takeoff and the landing. Much of the trip inbetween is often in autopilot, although the pilot may take more manual control the plane at points, of course. So you have two very clearly defined moments of high-intensity cognitive load, with a resting low cognitive load, with scattered amounts of manual control here and there.

-- In trains, my understanding is that driving the train involves a highly technical control of brake/throttle systems, and signaling/announcements. You have numerous stretches of track that may be flat and consistent, but also stretches of track that might have rapid geometry changes and thus require delicate and complex braking control. You have stations dotted throughout your route, perhaps at irregular intervals. For the driver, there are patches of high-intensity cognitive load (reduce 120km/h within a minute) scattered among low-intensity cognitive load. (drive for 190km/h for 120 minutes)

I can't think of any other setup in which it'd be easy to overlook or make human mistakes -- to have stretches of monotony punctuated by irregular, momentary points of high-intensity focus. It's the same cognitive 'rhythm' required of security guards -- of course you have people falling asleep. And from a design perspective, requiring humans to do this is just asking for trouble.

An ideal acoustic warning system would be one that stepped up a series of warnings - that alerted you to a speed change 10 minutes ahead, then 5 minutes ahead, then a minute ahead. Think of your GPS in the car - and how it continually lets you know -- "in 1 mile, take exit 32A", "in 300 ft, take exit 32A", and so on. At 60mi/hr, that's about a 1 minute advance notice ---- which is appropriate for the constant medium-intensity cognitive load of a car driver. Scaling that GPS warning system to the cognitive rhythm of a train driver would be something that notifies the driver beforehand -- "In 25 km, slow down to 90km/hr", "In 5km, slow down to 90km/hr", etrc. -- which is about 7 minutes' notice and 1.5 minutes' notice, respectively.

So A) having only an acoustic signal was a bad idea, and B) having an acoustic signal that alerts the driver 1 minute before a possible catastrophic scenario, taking into consideration a train driver's cognitive state, was a horrible and fatal design decision.
posted by suedehead at 10:52 AM on August 6, 2013


The fact that this kind of accident does not happen every day should tell us that the interface design is adequate for the vast majority of situations. I don't think we will be able to draw useful conclusions until the details are known: every other driver handled the particular section of track without accident in the past as far as I know, the crucial point is to figure out what made this particular run different.
posted by Dr Dracator at 12:17 PM on August 6, 2013


« Older Mouseunculus   |   Class Dublin Taxi Driver Newer »


This thread has been archived and is closed to new comments