Spanish Train Crash

[davieddy]

Assuming the video shown on BBC was real time, the question that needs to be answered is not "Was it was going too fast?" but "Why?".

Deepest sympathy for anyone affected by this tragedy,

David

[tha]

Since humans cannot be expected to be failsafe the technical systems should be. The disaster happened where the new system (ERTMS) ends and the old local system takes over. That raises the suspicion that the implementation does not hand over control in a failsafe way.

[chalsall]

Originally Posted by tha

Since humans cannot be expected to be failsafe the technical systems should be. The disaster happened where the new system (ERTMS) ends and the old local system takes over. That raises the suspicion that the implementation does not hand over control in a failsafe way.

A reasonable presupposition.

Unfortunately, the "technical systems" still tend to be implemented by humans. With little or no real testing in "real-world" and "unexpected" simulated situations.

Tangentially, while such systems can always be monitored by AI systems (and/or other humans) to try to qualify and quantify their sanity, this is usually not done (because humans are "too expensive").

Please remember that the initial mirror on the Hubble Space Telescope was launched ground incorrectly, because it was considered too expensive (read: a few million bucks) to test the mirror on the ground....

[ewmayer]

Turns out the train driver in question was a speed freak - which does not imply that there should have been a fail-safe; rather it reinforces the need for such safeguards against human stupidity and recklessness.

Originally Posted by chalsall

Please remember that the initial mirror on the Hubble Space Telescope was launched ground incorrectly, because it was considered too expensive (read: a few million bucks) to test the mirror on the ground....

Not to divert the thread, but the Hubble story (starting at post #16) is especially rich in incompetence, payola and irony - short story, NASA had 2 completely figured primary mirrors - one (bad) made by P/E, one (good) by Eastman/Kodak in storage for nearly a decade before the mirror-installation, no one thought it might be prudent to compare the 2 and pick the better one. (Which would have quickly revealed the major issue with the P/E mirror).

[NBtarheel_33]

Originally Posted by ewmayer

Turns out the train driver in question was a speed freak - which does not imply that there should have been a fail-safe; rather it reinforces the need for such safeguards against human stupidity and recklessness.

Indeed. The driver *blithely* admitted to the emergency operator that he had been traveling 190 km/h in an 80 km/h zone.

Even under the comparatively lenient (to the USA, anyway) sentencing laws of Europe, 80+ counts of reckless negligence/manslaughter ought to net Mr. Speedy an honest-to-goodness lifetime behind bars, along with a thorough cleanout of any and all assets he might possess.

Seems to me that it would be simple enough to link some kind of GPS location awareness to a speed governor that restricts the train from traveling any faster than a preset limit for a given location as reported by the GPS.

I honestly thought that many of these trains were actually on "autopilot" for most, if not all, of their journeys. It seems to work for airplanes, after all...as long as the pilots do not become confused about what the plane should be doing - "mode confusion", as has been suspected in the recent Asiana Airlines incident.

[Brian-E]

What the driver may have said immediately after the disaster is obviously something to be considered in the enquiry but should not, I suggest, be used to jump to conclusions now. Note that he would have been in shock at that time and not in any state to think about the possible mitigating circumstances and reasons why the train was travelling at that speed.

[chalsall]

Originally Posted by NBtarheel_33

Seems to me that it would be simple enough to link some kind of GPS location awareness to a speed governor that restricts the train from traveling any faster than a preset limit for a given location as reported by the GPS.

I would like to suggest that until we're very comfortable with our "digital overlords", that we use as much technology as possible to collect and report relevant data to the humans responsible as is possible, but that we keep the humans in the loop (and fully responsible) for as long as is demonstrably reasonable.

This position might change in ten years or so...

Just my opinion....

[xilman]

Originally Posted by chalsall

I would like to suggest that until we're very comfortable with our "digital overlords", that we use as much technology as possible to collect and report relevant data to the humans responsible as is possible, but that we keep the humans in the loop (and fully responsible) for as long as is demonstrably reasonable.

This position might change in ten years or so...

Just my opinion....

Too late, far too late, by several decades.

Since the 1980s at least, and arguably earlier, we've become so dependent on computers that if they magically disappeared the result would be a massive die-back running to the billions.

The modern economy, transportation systems and many other essential mechanisms could not function on 1950's technology.

[kladner]

Originally Posted by xilman

The modern economy, transportation systems and many other essential mechanisms could not function on 1950's technology.

The US Air Traffic Control System is only recently crawling out of 50's tech at its core. But your point is well taken overall.

[ewmayer]

A max-speed-limiting system need not even be terribly hi-tech: For speed limited by local track conditions (curvature, closness to 'real world' external hazards, etc.) one could implant local-speed cues at intervals in or around the track itself, perhaps via magnetic strip or by some other means. That would serve as a robust "last resort" in case of computer outage or sabotage. The fancier GPS or remote-control system could still be used for normal conditions.

I wonder if Spanish rail central even has a way of *monitoring* train speed remotely - it would be grossly incompetent to not have that capability. And assuming such capability, an excessive-speed-alert functionality in the monitoring system also seems a blatantly obvious thing to do. Even if you don't have the ability to control train speed remotely - and there are good reasons to not have that - you don't have the ability to call the train driver and say "hey, we just got a speed alert for you - what's going on?" Seriously? This ain't bloody rocket science here, these are basic modern mass-transportation kinds of safeguards, and unlike passenger aircraft, rail makes such things easier to do in so many ways. You simply don't even *design* a high-speed rail system without such safety features. Operator go to jail, fine - but I want to hear from the planners of this system, too.

[kladner]

If the Chicago Transit Authority has had that ability for a decade or two on its El trains, it is hard to imagine that a High Speed Rail system would not. Besides trains, the buses now have GPS so that would-be riders can look on a web page and see where the buses are.

The "cab signal" has been around for a long time. I'm not sure about the current setup, but it used to be a thing which would beep when an operator was supposed to slow down. If the operator did not do so in a brief time period, the system would apply significant braking automatically.

There is an even older pneumatic system which is tied into the track signals. When a train passes some kind of sensor at a signal, it changes the signal light to red. This happens about the time the second car is passing the signal. At the same time a small metal arm pops up beside one of the rails. This arm is placed to engage and trip a lever on the front of the lead car. If a following train goes past the signal and gets tripped, it puts on full emergency braking. The driver also gets in trouble if this happens.