5 ms·
"Simulated engine failures were performed in just about every conceivable situation, including take-off roll, before nose rotation, during rotation, after rotat
by barrkel 14y ago
"Simulated engine failures were performed in just about every conceivable situation, including take-off roll, before nose rotation, during rotation, after rotation, during the climb, during the landing and during both compensating and aggravating crosswinds. The result is an airplane that is as boring as possible to fly, even in emergencies."
I wonder how the system reacts when it gets false information about an engine failure. I mean, it's great when the computer is able to react to actual failures; but I'm also worried about how it reacts to bad sensor data. Sure, it's good engineering practice to defend in depth and eliminate single points of failure, but reading about the Air France crash, cascading errors don't seem to be properly dealt with.
- 51Cards 14y agoI agree with the concern for cascading failures but I'm not sure I agree with the comment re: Air France. There was no error cascade there as far as I can see. The pitot tubes froze and the when the aircraft lost airspeed data it did what it was designed to do, turned off auto-pilot and gave control back to the pilot. The pilots were the ones who failed to understand and manage the situation from there on, relying on a computer that was unable to provide data. Perhaps (said with a little tongue in cheek) it wouldn't be a bad thing if the computers flashed up a big "I don't know!" on the screens so the humans would know to look elsewhere for their information. Just a thought. Then again the humans need to know how to get the data in other ways too. As my auto racing instructor used to say, "It was a failure of the squishy bit between the steering wheel and the seat".
- epochwolf 14y agoThere is a big "I don't know" warning. It's called alternate law. And there is another mode called direct law for when the computer really doesn't know.
- barrkel 14y agoThere were also inconsistent stall warnings. If the pitot tubes freeze over, I expect windspeed will be underreported, which will make it look like a stall. IIRC, the pitot tubes were only out of commission for a minute or two, but the co-pilots ignored the (now correct) stall warning all the way into the ocean (until the pilot woke up and came back, but then it was too late). I guess the real problem here is having only a single modality for detecting windspeed. It doesn't matter how many pitot tubes you have if they all have the same failure mode; quantitative redundancy isn't as good as qualitative.
- genwin 14y agoDo you know why the plane doesn't use GPS as a backup method for airspeed measuring?