Welcome to European Tribune. It's gone a bit quiet around here these days, but it's still going.

Monday Train Blogging: Trainwreck

by DoDo Wed Nov 30th, 2005 at 04:08:24 AM EST

From the front page ~ whataboutbob

Trainwrecks elicit some dark fascination in most people. Today I will talk about the horrific 3 June 1998 accident near the German town Eschede (click image for larger image on the German Hochgeschwindigkeitszuege.com site):

Photo from Die schnellsten Züge der Welt

100 passengers of InterCityExpress (ICE) train „Wilhelm Conrad Röntgen“ died, about the same number of people was hurt, financial damage was in the hundreds of millions of Euros.

The train was serviced by an ICE-1 (series 401) high-speed trainset. However, what I want to show - following the train from construction to when the dust settled – is that contrary to how the disaster was framed in the media, this was not a failure of high-tech, but a series of failures due to lack of high-tech. I will also address questions whether the back tractor head* pushed up the train, or whether the train would have survived if it were built with bogies like the French TGV.

  • ICE-1 trains consist of 12 cars and two streamlined locomotives at either ends, called 'tractor heads'. The back tractor head in question is the only intact car on the lower right of the above photo.


The faulty wheel design
The jury identified faulty wheel design as the root cause of the disaster: the wheel has outer and inner parts, with an elastic rubber-like material in-between – and stresses in the outer steel ring led to cracking from the inside.

Was this wheel high-tech? Not at all. It was a design taken from tramways – without testing at high speeds – as an easy solution to bad ride quality (i.e. vibrations passengers feel). But that bad ride quality was a result of saving on the use of air springs in the first-generation ICE trains – then a proven but still new technology.

Politics vs. development
The background of this saving (and some I'll tell about below) was that ICE trains were kind of a stepchild of German transport policy in the eighties, developed by the German Railways without much government support – while the toy of the big boys of the Kohl government was the Transrapid, Germany's magnetic levitation train.

The Maglev has its technological advances alright, but has some severe economic and operational problems: its track is very expensive, and while most European high-speed trains reach destinations beyond high-speed lines, you'd have to build dedicated track for the Transrapid everywhere for comparable service. But said big boys only had that glimmer in the eye, and made sure that the ICE won't steal the Transrapid's limelight. So, even tough the ICE was intended to be a much more modern train than the TGV, many key technologies were delayed until the ICE-2 or even ICE-3.

Pre-emptive detection fails
The faulty wheel could have been identified before breaking during the frequent regular checks this train has. In fact, it was! The problem is: in another cost-saving measure, the maintenance facility was only equipped with some simple hand-held devices, which were also unreliable – the workers thought the extreme value they measured was an instrument error like ones they had before.

No emergency detection
The ICE trains were originally to be fitted with on-board detection systems, which could have warned the train driver. But this was also 'saved'.

Stability after derailment
ICE trains have bogies like normal carriages – so it has widely been claimed that had the ICE had Jacobs bogies between the cars like the TGV (picture below), it wouldn't have derailed "so easily". But, in truth, the train was very stable: with a broken and derailed wheel, it travelled another 5.5 km at full speed before the disaster. In fact passengers would have fared better if the train had left the tracks after the wheel broke...

Photo from RailFanEurope.net

The train breaks apart
The onset of the disaster was when the broken wheel (the third on the first carriage behind the front tractor head) got struck in a switch. The impact tore apart the connection between the locomotive and the first car, and switched the switch to the deviating route. The train broke apart the second time between the third and fourth cars – the front tractor head rolled out (stopping 2–3 km away), the first three cars just derailed and came to halt on the tracks with minor damage, but the switch directed the rest of the train sharply to the right at 198 km/h.

A high-tech accident? Not at all: high-speed switches are designed so that nothing gets struck and its moving parts are locked – but the disaster happened on an upgraded old line, on which (cost-cutting again) not all switches were replaced...

The bridge collapses
The really big disaster came because the fourth car kicked out the pillar of a road overpass, which fell on the fifth and sixth cars and "stopped" the rest dead. How could this have happened? You guessed it, high-tech cost-cutting again: along high-speed and fully upgraded old lines, no switches are supposed to be near overpasses, and overpasses have no central pillars.

Harmonica effect
The remaining six carriages and the back tractor head curled up like a harmonica. Beyond referring to the TGV's irrelevant supposedly superior stability, newspapers often quoted 'experts' that the back tractor head made things worse by pushing up the train on the bridge's wreck.

Not true: traction stops and braking starts automatically as soon as the train breaks apart! And a basic calculation shows that to stop just the last (12th) car still intact (from 198 km/h to zero in 132 m and 4.8 seconds), something beyond 1G and a force of c. 630 kN would have been needed even for continuous deceleration! That's well beyond what the brakes could give, so even if the back tractor head at least gave to the mass to be stopped, the outcome wouldn't have been much different without it.

Bulletproof windows trap people
There was one high-tech problem that played a role: wounded victims trapped in the wreckage could not break themselves free, as the extremely strong windows weren't designed for that. The ICE's windows are very strong so that passing tunnels, especially passing other trains in tunnels, won't hurt people's ears. However, even this is a result of a cost-cutting measure (which caused other problems too): German high-speed lines were built with single-tube, two-track tunnels, rather than twin tunnels.


Previous Monday Train Bloggings:

  1. (Premiere/ modern Austrian trains & locos)
  2. Adventure
  3. Fast Steam
  4. Heavy Haul
  5. Forgotten Colorado
  6. The Hardest Job
  7. Blowback
  8. Highest Speed
  9. New England Autumn

Display:
Hanover was the last stop of the ICE before it crashed. Everytime I go to Hamburg, the train passes Eschede, and it does feel weird. Some people I know were on the rescue teams at the crash-site. It must have been horribe.

Beside the tragedy: I still believe the ICE system is a great thing. And the ICE II and III are fantastic train. With one serious flaw: the coffee prices are way to high..

by jandsm on Mon Nov 28th, 2005 at 06:59:34 AM EST
Some people I know were on the rescue teams at the crash-site. It must have been horribe.

I once met a Hungarian ex-railroader whose job it was to clean up after accidents. The most horrific sight he told of was one kept secret during communism, when an unbraked sleeper carriage with Russian army officials sleeping in it rolled down the mountain, crashing into a freight train at arond 160 km/h - what your acquintances have seen must have been even worse...

Beside the tragedy: I still believe the ICE system is a great thing. And the ICE II and III are fantastic train.

Yeah, good thing you remind me, I forgot to mention: the ICE-2 (the 'half-set' version: tractor head at one end, driving trailer at the other) already has air springs, so it was already built without the faulty wheels; while the ICE-3 even has distributed traction (every second car is motorised). (The again, more cost-cutting idiocy necessitated the replacement of too weak air conditioners in the ICE-3...)

With one serious flaw: the coffee prices are way to high..

:-(... The Mehdorns of this world either don't know the price of coffee outside five-star hotels, or travel by business jet while directing a railway...

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 07:19:50 AM EST
[ Parent ]
(I added the following sentence to the politics part:

"So, even tough the ICE was intended to be a much more modern train than the TGV, many key technologies were delayed until the ICE-2 or even ICE-3.")

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 01:06:32 PM EST
[ Parent ]
Thanks for the facinating diary, Dodo.

And now for the non-PC rant. Security measures suggested by politicians after train crashes are some of the stupidest, least effective things that can be done. They usually include spending billions on measures that are unlikely to save more than a few lives per year, if that - and would be much more usefully spent on road safety, or fighting certain kinds of diseases (not to mention, if the purpose is to save lives, things like fighting malaria in the third world).

Big train catastrophes, like plane accidents, are extremely rare but seem to generate passionate responses in the public (becuase we all take the train BUT we don't drive it ourselves and are thus not "in control" - extraordinary precautions must thus be taken, obviously, to protect us from all these incompetents).

The real remedies, those suggested by your diary, are never discussed in the public outcry (they may end up being done by the railway company if it has good engineers in management, but it will have little relation to the political noise made around the catastrophe).

In the long run, we're all dead. John Maynard Keynes

by Jerome a Paris (etg@eurotrib.com) on Mon Nov 28th, 2005 at 10:54:56 AM EST
Which is why, in general, before passing any new legislation in response to a catastrophe there should be some kind of parliamentary inquiry. To cool off and call in the experts and see if the proposed new rules would have prevented the disaster had they been in place.

A society committed to the notion that government is always bad will have bad government. And it doesn't have to be that way. — Paul Krugman
by Carrie (migeru at eurotrib dot com) on Mon Nov 28th, 2005 at 11:02:42 AM EST
[ Parent ]
Yes, agreed to that!

On the other hand, I don't want to sound insensitive, but trainwrecks (and airplane crashes and shipwrecks) don't just have a number of killed effect, but a money effect too - and even if more lives could be saved with that money elsewhere, it is often the case that the disaster costs much more than the safety measure that fell victim to some cost-saving measure.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 01:02:21 PM EST
[ Parent ]
the disaster costs much more than the safety measure that fell victim to some cost-saving measure
So it increases the GDP!

A society committed to the notion that government is always bad will have bad government. And it doesn't have to be that way. — Paul Krugman
by Carrie (migeru at eurotrib dot com) on Mon Nov 28th, 2005 at 01:06:17 PM EST
[ Parent ]
I should start a weekly competition for the most un-PC lampooning of neoliberal buzzwords... (I would be a mere novice, but what you, Colman or afew produce on a daily basis cracks me up :-))

*Lunatic*, n.
One whose delusions are out of fashion.
by DoDo on Mon Nov 28th, 2005 at 01:08:46 PM EST
[ Parent ]
Yes, but the secret which eludes most is "which" safety step will be necessary.  To be perfectly safe, take no risk at all.  Simple solution, but the safe solution.  

And to clarify, one of the reasons for the inner/damper/outer wheel design is to dampen vibration, not for "ride quality" but to keep vibration from causing a structural failure of the wheel.  This wheel design has also grounded the Acela trains here in the US, finding the right damper material is a challenge.  

by btower on Mon Nov 28th, 2005 at 04:08:05 PM EST
[ Parent ]
And to clarify, one of the reasons for the inner/damper/outer wheel design is to dampen vibration, not for "ride quality" but to keep vibration from causing a structural failure of the wheel.

It may have been for the sake of the wheel on other vehicles (tough I never heard of this - and a two-part wheel seems more prone to failure to vibration even on a tramway to me), but not the ICE.

The German Railways in fact first tried to run ICEs without any form of dampening, but the result was the infamous "Dröhnen" (carbody vibrations so strong it can be heard), and a rather unsteady run of the biggest car, the restaurant car.

I don't remember the use of such wheels in the Acela, I have to look it up.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 04:22:06 PM EST
[ Parent ]
btower, I looked, and I found absolutely nothing about the Acela being retrofitted with two-part wheels. You must have conflated the earlier wheel profile problem and the newer yaw damper problem.

*Lunatic*, n.
One whose delusions are out of fashion.
by DoDo on Mon Nov 28th, 2005 at 04:30:30 PM EST
[ Parent ]
the secret which eludes most is "which" safety step will be necessary.

All of those I mentioned were necessary, and at some stage foreseen. The technical people did see their necessity in advance.

A similar issue is that of tunnels and tunnel fires. I predict right now that there will be a catastrophic tunnel fire in some European rail tunnel opened in the last 10 years, in which the investigation will find that escape shafts were placed too infrequently.

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 04:34:39 PM EST
[ Parent ]
Thanks for an interesting review of the failures of cost cutting on rail project developments! The pragmatist in me wonders if the train would have been built without those cuts? Any idea how decisions about how much to retrofit & how much to build new are done & what expertise is involved?

Following up on a comment by Migeru was there any legislative action was taken as a consequence (either hasty or after the initial media frenzy dissipated) of this derailment?

by Alexandra in WMass (alexandra_wmass[a|t]yahoo[d|o|t]fr) on Mon Nov 28th, 2005 at 01:22:49 PM EST
wonders if the train would have been built without those cuts?

It would. If more government R&D had gone its way, then as fast as it was. If not, maybe with some years of delay. As I mentioned to jandsm, the delayed technologies were eventually built into later versions of the train.

Any idea how decisions about how much to retrofit & how much to build new are done & what expertise is involved?

This train was all new, so no retrofit here. The new technologies used weren't entirely novel either - in fact most were developed by German firms. But to build any new train in Europe, epecially for conditions that didn't exist before (like - going faster), you have to design all parts to work together, and prove that they do in tests. This costs money (and so does building the new equipment, so to make that cheap is part of R&D) - so cost-cutters may decide to not use something, or use some old technology instead that is known well enough to be adapted more easily.

Now your question was about the decision process - that's usually some high boss sitting down with an R&D boss, looking through what costs how much, and the high boss making the decision in the end. Which is often silly, and some in the R&D department already see that. I don't know it closer than that.

was there any legislative action was taken as a consequence (either hasty or after the initial media frenzy dissipated) of this derailment?

None of the kind Jérôme mentioned that I know of. The trains were first taken off the rails, then allowed back with standard wheels (speak: a lousy ride quality; I travelled in one), while a jury investigation was left to run its course. Unfortunately, and as all too usual, there was no verdict coming down hard on the responsible higher-ups. Meanwhile, there have been programs to retrofit the trains with window-breaking points and on-board sensors, and to better equip maintenance centers, but AFAIK without 100% coverage so far. Same for track-related issues (e.g. switches, bridges).

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 02:43:22 PM EST
[ Parent ]
Thanks for the reply. If nothing else it seems the temptation to jump to short term legislative solutions to the crash was averted.
by Alexandra in WMass (alexandra_wmass[a|t]yahoo[d|o|t]fr) on Mon Nov 28th, 2005 at 02:54:46 PM EST
[ Parent ]
The pragmatist in me wonders if the train would have been built without those cuts?

Definitely!
As DoDo mentioned the alternative Maglev is very expensive so some sort of high speed train would definitely have been built.

Following up on a comment by Migeru was there any legislative action was taken as a consequence (either hasty or after the initial media frenzy dissipated) of this derailment?

As far as I know, no. The government pretty much stayed out of it.

What happened in the immediate aftermath was a frenzy of checks by the railway company.
June 4: Speed limit of 160 km/h for all ICE 1 trains
        and ultrasound checks of all wheels
June 6: All 59 ICE 1 trains recalled for ultrasound
        checks and not to be used until checked (even
        the 15 already checked since June 4)
June 13:Once again all ICE 1 trains recalled. This time
        because of the supervising Federal Railway
        Agency.
        They found similar problems in some urban rail
        trains using the same wheels.
June 15:The Federal railway Agency forbids the use of
        this kind of wheels. All have to be exchanged
        for standard wheels.

(According to the websites I googled the accident happened on June 3, not July 3?)

Additionally, checks all across Germany if there were similar switches before possible obstacles. Any new built tracks since then don´t have switches in potentially dangerous zones and new bridges - if possible - don´t have a "pillar" near the tracks.
Of course I don´t know how much retrofit was done.

by Detlef (Detlef1961_at_yahoo_dot_de) on Mon Nov 28th, 2005 at 03:06:54 PM EST
[ Parent ]
Excellent additions!

the accident happened on June 3, not July 3?

I always confuse the two... corrected!

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 04:17:49 PM EST
[ Parent ]
BTW, in Germany, Eschede is the worst post-war train disaster, and only two or three during WWII exceeded it. All three were collisions, and two of them happened on the very same day (22.12.1939). The third, a collision of a fast train with a fuel train, may have been the worst (c. 300 dead), but we only have the ridiculously low number of the Nazis' official propaganda (41).

The worst on record is very recent: when the big Southeast Asian Tsunami hit a train on Sri Lanka, 1800 people may have perished.

Before that, the three worst I heard of: a train falling onto a river in India (c. 800 dead), a train in the eighties in the Soviet Union rushing into a cloud of gas spewing forth from broken pipelines and bringing it to explosion (c. 600 dead); and a French military holiday train during WWI, passing the Italian-French border, on which the military officer threatened the engineer with court martial if he doesn't proceed downhill with an overwheight train, on which first the hot and failing brakes caused fire, then all fell into a ravine (c. 5-800 dead).

*Lunatic*, n.
One whose delusions are out of fashion.

by DoDo on Mon Nov 28th, 2005 at 04:12:58 PM EST
I remember the Russian disaster. I helped organize relief efforts from Tokyo then.

Last April, we had another disaster of a commuter train. The report says the train conductor exceeded the prescribed speed of 95km (!), and it caused the derailing and wreck, as the train was slammed to the nearby apartment building. So, we are not that high-speed, except for Shinkansen, but are rather similar to Brit Rail.

An excellent analysis, DoDo.

I will become a patissier, God willing.

by tuasfait on Mon Nov 28th, 2005 at 08:03:19 PM EST
[ Parent ]
This blog series has been most interesting.  

I liked your analysis today.  It is typical, almost an inevitability, that "high tech" disasters are made up of a convergence of multiple low tech mistakes--many of them noticed and even documented in advance.  Usually the folk most involved with the technical issues aren't even surprised.  


The Fates are kind.

by Gaianne on Wed Nov 30th, 2005 at 03:21:48 AM EST


Display:
Go to: [ European Tribune Homepage : Top of page : Top of comments ]